Advances in Food Research, Volume 6

ADVANCES IN FOOD RESEARCH VOLU,ME V I

This Page Intentionally Left Blank

ADVANCES IN FOOD RESEARCH VOLUME VI Edited by

E. M. MRAK

G. F. STEWART

University of California Davis, California

University of California Davis, California

Editorial Board E. C. BATE-SMITH

S. LEPKOVSHY

W. H. COOK

B. E. PROCTOR

W. F. GEDDES

EDWARD SELTZER

M. A. JOSLYK

P. F. SHARP

A. J. KLUYVER

W. M. URBAIN 0. B. WILLIAMS

1955

ACADEMIC PRESS INC., PUBLISHERS NEW YORK, N. Y.

Copyright 1955, by ACADEMIC PRESS INC. 125

EAST

2 3 ~ 0STREET 10, N. Y.

S E W YORK

ALL RIGHTS RESERVED NO PART OF THIS BOOK MAY B E REPRODUCED I N A N Y FORM, BY

PHOTOSTAT,

MICROFILM,

OR

ANY

OTHER

MEANS,

WITHOUT WRITTEX PERXIISSION FROM T H E PUBLISHERS.

Library of Congress Catalog Card Number: 48-7808

PRINTED I N THE UNI TED STATES OF AMERICA

CONTRIBUTORS TO VOLUMEVI JOHNC. AYRES,Food Processing Laboratory, Iowa Agricultural Experimental Staiion, Iowa State College, Ames, Iowa GEORGBORGSTROM, Swedish Institute for Food Preservation Research ( S I K ) , Goteborg, Sweden

H.T. H. FARRER, Research Laboratories, Kraft Foods Ltd., Melbourne, Australia

W. 0. HARRINGTON, Western Utilization Research Branch, Agrzcultural Research Service, U . S. Department of AgrLculture, Albany, California P. W. KILPATRICK, Western Utilization Research Branch, Agricul..ural Research Service, U . S . Department of Agriculture, Albany, California E. LOWE, Western Utilibation Research Branch, Agricultural Research Service, U . S. Department of Agriculture, Albany, California L. F. MARTIN, Sugarcane Products Division, Southern Regional Research Laboratory, U . S . Department of Agriculture, New Orleans, Louisiana

R. L. OLSON,Western Utilization Research Branch, Agricultural Research Serrice, U . S. Department of Agriculture, Albany, California W. B. VAN ARSDEL, Western Utilization Research Branch, Agricultural Research Seruice, U. S. Department of Agriculture, Albany, California REESEH. VAUGHX,Department of Food Technology, University of California, Davis, California

This Page Intentionally Left Blank

FOREWORD I n keeping with the editorial policy of Advances in Food Research, seven articles covering five commodity and three functional areas are included in this volume. The contributions are exhaustive, critical, integrating, and of fundamental importance t o the development of food research in the wide areas of the food industries. I n general, they are concerned with microbiology, processing, retention of nutrients, and engineering. Several of the reviews cover rather neglected areas and one presents a new and interesting analysis of the much discussed subject of vitamin destruction. A classical example of a neglected area is that concerned with applications of research t o problems of candy manufacture. Although the candy industry in the United States involves the use of approximately 1% million tons of materials annually, candy manufacture has developed as an art and continues primarily as an art, employing “rule of thumb’’ methods. Research in this field is in its infancy, although it is being conducted t o a n increasing extent in industry and research institutions. Nevertheless, most of the information applicable t o the field has been developed as a result of research in related fields, hence pertinent literature is widely scattered in a great variety of journals and public,ations. I n his review, Dr. Martin has brought together this widely distributed information in a n excellent manner, has outlined clearly the wide diversity and complexity of the fundamental problems of candy manufacture, and shows the great opportunities for fundamental research in this field. The review by Vaughn on the “Bacterial Spoilage of Wines” is timely, since his interpretations clarify many points of confusion. Although the wine industry is a very old one, of importance in many countries throughout the world, confusion exists concerning nature, causes, and nomenclature of microbial spoilage of wines. Taxonomic interpretations frequently have been in error, and in many cases have persisted over a period of many years. Furthermore, common terminology for the microbiological spoilage of wines is often inappropriate although it has persisted t o a large extent because of the historical background of this field. It is, therefore, most appropriate t h a t Dr. Vaughn should review the literature of the last 100 years in a critical manner and interpret existing concepts in light of recent knowledge. I n previous volumes of Advances in Food Research, the physioIogy vii

...

Vlll

FOREWORD

and chemistry of rigor mortis with special reference t o the aging of beef and the oxidative rancidity and discoloration in meat were considered. The review by Ayres on the microbiology of meat animals covers a third area of scientific problems of primary interest t o this large industry. Meat provides many, if not all, nutrients desired by most microorganisms. Preventing their growth on meat, therefore, is a serious and important problem t o the industry. The load of organisms on the living animal contributes its share t o the total count on the carcass. This count, however, also depends on factors involving the method of handling, the interrelationship between defensive mechanisms of the animal, and the enormous microbial populations which gain access t o it. Dr. Ayres has discussed the significance and sources of microbial contamination with respect t o the animal defenses against these invasions, ante-mortem changes in microbial populations, death agonies, and after death. Dr. Ayres appropriately concludes his review by pointing out the needs for further research and suggesting certain improvements in processing. Although certain frozen foods have been on the market for a number of years, the industry at present is involved in the production of greater quantities and more diversified products than ever before. This is particularly true of frozen precooked foods, the production of which is increasing at a spectacular rate. Microbial problems are always present, insofar as frozen foods are concerned, but are particularly important now because of the great variety of frozen cooked foods. Hence, Borgstrom’s inclusive review on the microbiology of frozen foods (except ice cream) is timely and in keeping with processing advances. I n Volume I of Advances i n Food Reseurch, a chapter was devoted t o the deterioration of processed potatoes. At that time, specific reference t o the processing and stability of any particular potato product was avoided, because of the need for a general discussion of the stability of all potato products. Since then, however, a large amount of work has been done on the development of potato granules. The production of potato granules coincides with the present trend toward producing foods which are easy to use, the so-called “instant ” foods. The review by Olson and Harrington is inclusive and brings together all the literature on this subject in a critical and useful manner. A very extensive literature exists relating to the destruction or stability of Vitamin B1 during cooking, processing, and handling of various food commodities. The destruction (or retention) of this vitamin is of great importance because of nutritional implications. In spite of the vast literature however, Farrer appears t o be the first t o attempt t o relate results of vitamin B1 destruction t o other food stuffs or t o any particular set of conditions. I n his review, Farrer has considered available

FOREWORD

ix

data on factors influencing thermal destruction of Vitamin B1,has shown a reliable and satisfactory approach by the use of kinetics, and has critically analyzed published data in the field-a most valuable contribution a t this time. Volume I1 of Advances in Food Research contained an article on the spray drying of foods. I n this volume, Kilpatrick, Lowe and Van Arsdale have reviewed another aspect of dehydration pertaining t o tunnel dehydrators for fruits and vegetables. While the “germ” of the idea of the tunnel dehydrator is at least a century old, and while considerable literature and information is available on these types of driers, this information has not been heretofore brought together in a critical and satisfactory manner. The material included considers type of tunnel, mechanical elements of tunnel construction, theory of tunnel dehydrators, operating procedures, and recent trends in tunnel dehydration of fruits and vegetables. This review will serve as an excellent companion article for that which appeared on spray drying in Volume I1 and as a contribution in the field of engineering and production.

October, 1955

E. M. MRAK G. F. STEWART

This Page Intentionally Left Blank

CONTEXTS Contributors t o Volume VI . . . . . . . . . . Foreword . . . . . . . . . . . . . . . . . . . .

. .

.

.

.

v

.

vii

Applications of Research to Problems of Candy Manufacture B YL. F MARTIN,Sugarcane Products Division, Southern Regional Research Laboratory. U S Department of Agriculture. New Orleans. Louisiana I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Desirable Properties of Candy . . . . . . . . . . . . . . . . . . . I11. Properties and Reactions of Sugars in Candymaking . . . . . . . . . . IV . Modification of Sugar Properties by Minor Ingredients . . . . . . . . . (V . Major Ingredients Other than Sugars . . . . . . . . . . . . . . . . VI . Production Methods . . . . . . . . . . . . . . . . . . . . . . . . VII Summary . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20 29 48 50 55

Bacterial Spoilage of Wines with Special Reference to California Conditions BY REESEH VAUGHN, Department of Food Technology, University of California, Davis, California I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Historical Consequences . . . . . . . . . . . . . . . . . . . . . . 111. Types of Wine Spoilage Caused by Bacteria . . . . . . . . . . . . . . IV . Factors Affecting the Growth of Bacteria in Wines . . . . . . . . . . V . Characteristics of the Bacteria Found in California Wines . . . . . . . VI Additional Research Needs . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

67 68 75 82 90 100 101 101

.

. .

.

1 6 15

.

.

Microbiological Implications in the Handling, Slaughtering, and Dressing of Meat Animals BY JOHN C . AYRES,Food Processing Laboratory, Iowa Agricultural Experimental Station, Iowa State College, Ames, Iowa I Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 I1 Defensive Mechanisms . . . . . . . . . . . . . . . . . . . . . . . 111 111 Ante-Mortem . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 IV Slaughter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 V. Post-Mortem . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 VI Improvements in Processing Practices . . . . . . . . . . . . . . . . 149 154 References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

Microbiological Problems of Frozen Food Products BY GEORGBORGSTROM, Swedish Institute for Food Preservation ( S I K ) , Goteborg, Sweden I. Introduction . . . . . . . . . . . . . . . . . . . . . . . I1. The Influence of Freezing Temperatures on Microorganisms . I11. The Influence of the Freezing Rate . . . . . . . . . . . . IV. The Freezing Death of Bacteria . . . . . . . . . . . . V . Occurrence of Bacteria in Frozen Foods . . . . . . . . . xi

Research

. . . .

163 . . . . . 164 . . . . . 168 . . . . . 170 . . . . . 172

xii

CONTENTS

VI . Pathogenic Bacteria in Frozen VII . Defrosting Problems . . . . . VIIJ . Packaging Problems. . . . . I X . Cooking . . . . . . . . . . X . Hygienic Aspects . . . . . . X I . Practical Aspects . . . . . . References . . . . . . . . .

Foods . . . . . . . . . . . . . . . . 201 . . . . . . . . . . . . . . . . . . . 204 . . . . . . . . . . . . . . . . . . . 209 . . . . . . . . . . . . . . . . . . . 210 . . . . . . . . . . . . . . . . . . . 210 . . . . . . . . . . . . . . . . . . . 211 . . . . . . . . . . . . . . . . . . . 213

Potato Granules, Development and Technology of Manufacture BY R . L . OLSONA N D W. 0. HARRINCTON, Western Utilization Research Branch, Agricultural Research Service, U . S . Department of Agriculture, Albany, California I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 I1. Related Nongranular Products . . . . . . . . . . . . . . . . . . . 233 I11. Spray Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 . . . . . . . . . . . . . . . . . . 235 IV . Direct Dehydration-Two-Stage V . “Freeze and Squeeze” Method . . . . . . . . . . . . . . . . . . . 237 VI . Cold Tempering . . . . . . . . . . . . . . . . . . . . . . . . . . 237 VII . Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . 238 VIII . Add-Back Method . . . . . . . . . . . . . . . . . . . . . . . . . 238 I X . General Considerations . . . . . . . . . . . . . . . . . . . . . . . 243 X . Qiality Evaluation of Potato Granules . . . . . . . . . . . . . . . . 249 References . . . . . . . . . . . . . . . . . . . . . . . . . 253 The Thermal Destruction of Vitamin B, in Foods BY X . T . H . FARRER,Research Laboratories, Kraft Foods Ltd., Melbourne, Australia I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 I1. Factors Influencing the Thermal Destruction of Vitamin B1 . . . . . . 258 I11. Thermal Losses in Cereals . . . . . . . . . . . . . . . . . . . . . 264 IV. Thermal Losses in Meats . . . . . . . . . . . . . . . . . . . . . . 275 V. Losses in Processing Vegetables . . . . . . . . . . . . . . . . . . . 285 VI . Thermal Losses in Other Foodstuffs . . . . . . . . . . . . . . . . . 293 VII . Thermal Losses on Storage . . . . . . . . . . . . . . . . . . . . . 294 VIII . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 Tunnel Dehydrators for Fruits and Vegetables

BY P. W . KILPATRICK, E . LOWE,A N D W . B. VAN ARSDEL,Western Utilizztion Research Branch, Agricultural Research Service, U . S . Department of Agriculture, Albany, California I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Classification of Tunnel Dehydrators. . . . . . . . . . . . . . . . 111. Mechanical Elements of Tunnel Construction . . . . . . . . . . . . IV . Typical Commercial Tunnel Dehydrators . . . . . . . . . . . . . . V. Criteria for Selection of Tunnel Dehydrators . . . . . . . . . . . . VI Basic Theory of Tunnel Dehydrators . . . . . . . . . . . . . . . VII . Operating Procedures for Tunnel Dehydrators . . . . . . . . . . . VIII . Recent Trends in Tunnel Dehydration of Fruits and Vegetables . . . IX . List of Symbols Used . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . .

314 . 316 . 326 . 339 . 345 . 347 . 363 . 367 369 369 373 390

Applications of Research to Problems of Candy Manufacture

BY L. F. MARTIN Sugarcane Products Divisivn, Southern Regional Research Laboratory, U.S. Department of Agriculture, New Orleans, Louisiana Page I. Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 11. Desirable Properties of Candy. . . . . . . .... . . . . . . . . . . . . . 6

111. IV.

V.

VI. VII.

1. Physical Properties.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Chemical Composition and Properties.. . . . . . . . . . . . . . . . . . 3. Preservation of Desirable Properties. . . . . . . . . . . . . . . . . . . . . . . . . Properties and Reactions of Sugars in Candymaking.. . . . . . . . . 1. Effect of Heat upon Sugars.. . . . . . . . . . . . . . . . . . . . . . . . . . Modification of Sugar Properties by Minor Ingredients. . . . . . . . 1. Protein Whipping Agents.. . . . . . . . . . . . . . . . . . . . . 2. Gelatin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Pectin . . . . . . . . . . . . . . . . . . . . . . . . .. 4. Natural G u m s . . . . . . . . . . . . . . . . . . . . . . . . . Major Ingredients Other than Sugars. . . . . . . . . . . 1. Milk Products.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Fats Other than Dairy Butter and Cocoa Butter. . . . . . . . . 3. Starch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Cocoa and Chocolate.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Nutmeats and Fruits.. . . . . . . . . . . . . . . . . . . . . . . . . Production Methods. . . . . . . . . . . . . . . . . Summary.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 9 10 15 16 20 20 24 26 29 29 30 32 33 37 45

48 50 55

I. INTRODUCTION Candymaking continues t o be primarily an art t o a greater extent than most other modern food-processing operations. The popularity of candies is attributable largely t o their great variety and the scope afforded for originality in developing a wide diversity of qualities that do not lend themselves readily to standardization by a simple set of scientific principles. Despite increasing use of precise methods for control of raw material and product quality and extensive adoption of improved machinery and techniques of mass production, the candymaker’s skill in applying “rule of thumb” procedures remains the principal guide in formulation and cooking. Research is being conducted to an increasing extent within the industry, but far more information applicable t o its 1

2

L. F. MARTIN

operations has been developed by research in related fields. Major improvements are the result of advances in the science and technology of the food materials used as ingredients of candy. A review of recent investigations bearing upon the solution of candymaking problems must include such related research as well as studies devoted specifically t o various aspects of candy manufacture. The confectionery industry, including all types of candy and chocolate goods, produces about three billion pounds of these products annually t o rank as one of the largest food industries in the United States. It is one of the largest industrial consumers of sugar, t o which it adds some eighty or ninety other ingredients in making more than a hundred distinct items. Table I presents a broad classification of major confectionery raw materials together with statistics relative to the quantities used and their value. The numerous individual ingredients may be grouped conveniently into six major classes of food materials in addition t o the miscellaneous pectic and protein products used as gel-forming or whipping agents. Ingredients in each group possess common properties and present similar problems when used in making different types of candy. One property common t o all candies is their high sugar content, attained by cooking t o relatively high temperatures. The effective use of any particular ingredient depends more upon its properties and behavior under these conditions of heating and concentration than upon the type of candy in which it happens t o be incorporated. Chemical reactions of sugars are fundamentally the same, but differ in the extent t o which they proceed in the preparation of marshmallows a t relatively low temperatures, or in cooking to high temperatures for the production of hard candy. Interactions of ingredients of different classes are primarily important in many cases such as, for example, the combination of sugars and milk solids t o produce caramel. The recently acquired knowledge of the chemistry of each class of ingredients will be considered in the following sections, with emphasis upon its practical application in improving the methods of production and the quality of candies of all types. It should not be inferred th at chemical aspects of candy manufacture have been neglected entirely until recent years in favor of more immediately profitable developments in engineering of mass production methods, mechanizing the handling of materials, automatic process control, and machine packaging. The excellent reference book of Jordan (1930) summarizes the practical knowledge of candy chemistry acquired prior t o its publication. I n earlier studies Paine (1924, 1928) sought t o explain some of the processes involved in terms of the properties of sugars as they were known at t ha t time. Subsequent discoveries have made it evident th a t reactions which may occur in the production of even the simplest, types

3

CANDY MANUFACTURE

TABLE I Estimated Amounts and Average Cost of Ingredients Used by the Candy Industrye Quantity ( X 1000 pounds)

___ Ingredients

1947

1950

Sugar and Sweetenrrs Cane and bept sugar 1,020.055 1,218,032 14,517 Corn sugar 30,061 Corn sirtip 676,864 678,683 Other (honey, maple, etr.) 3,368 20,837 30,936 Corn starch 20,074 Cocoa products 198,666 178,75:3 Cocoa beans Cocoa powder 5878 9776 Cocoa butter 20.989 23,098 8993 Chocolate liquor 27,430 Coatings 2!i6. 882 249,718 \[ilk products 258,241 Fluid milk 194,177 Condenscd and evaporated 96.343 72,268 4141 Butter 3164 Other (dried, cream, 33,562 46,949 etc.) Eggs and egg products 6735 6368 Fats and oils 39,062 31.921 Essential oils and 3495 flavors 3965 I’eanrits 199.975 136.337 Almonds and other 39 ,:3xi 14,660 nuts Coconut 18,570 64,201 Fruit products 12,543 15,193 Miscellaneous other 34.521 56,807 ingredients Total ingredicnts

2 990 548

3,175,617

Cost

(x1000 dollars)

Average cost per pound

__.

1947

85,138 2207 35,096

1950

1947

1950

99,272 SO. 084 073 1030 052 32,642

$0.082 071 053

2357 1307

1439 1980

.069 .065

.069 .064

49,377 1082 20,099 11,507 90,220

53,130 2274 13,548 3685 80,770

.276 184 670 420 338

.267

7200

9741

.037

.038

12.018 2332

763 1 2691

125 737

106 650

11,124 2893 11,261

8307 1607 7141

237 430 288

247 252 221

5779 33.660

45G8 25,111

1 653 168

1 152

28,514 5316 3933

20,345 12,467 4924

638 286 314

714

233 587 410 323

184

194 32-1

5290

7062

153

124

427,710

402,365

0 143

0 127

of candies are extremely complex, giving rise t o problems that challenge the utmost ingenuity and skill of the chemist. Earlier knowledge of sugars, fats, proteins, starches, and other complex ingredients was far from adequate for the Rolutioii of these problems. Some essential details of the structures and properties of these substances have heen determined only

4

L. F. MARTIN

within the past two decades and much remains t o be learned. The progress that has been made possible by modern research methods in explaining the changes that occur in the drying of milk (Coulter et al., 1951) provides but one example. Structures and simpler reactions of the common sugars, sucrose, glucose, and fructose, have been fairly wellestablished for some time, but changes produced by heating solutions of the pure sugars are still very imperfectly understood. Obviously, the complete and accurate description of all of the reactions that occur when sugars, milk, and fats are cooked together at high temperatures t o produce caramels must await the results of future and more difficult research. Candy appeals t o the universal liking for sweetness, modified or varied by less accurately definable yet tangible qualities of flavor and texture. Unlike many other food products th a t are preserved or processed under mild conditions t o minimize alteration of their natural composition, texture, and flavor, the qualities of most candies are created by more or less drastic modification of the properties of their ingredients. These qualities must be maintained for considerable periods of time, not only for distribution but also after the products reach the consumer. Organic chemistry is the principal field in which the scientific guides to better production methods and improved quality must be sought. Improvement or control of texture calls for the application of the techniques of physical chemistry, including particularly colloid chemistry. Biochemical factors affect the composition, stability, and quality of the important natural products used as ingredients. Many of the latest methods of analytical chemistry can be applied advantageously in process control as well as in establishing and maintaining standards for the raw materials and finished products. Chemical engineering must devise the practical methods and equipment required for commercial application of improvements developed by research in the other fields of chemistry. Recent contributions of each of these fields of investigation t o progress in the candy industry will be reviewed and, of possibly greater importance, applicability of some of the more recent discoveries and techniques t o research on unsolved problems of the industry will be indicated. Many of the ingredients listed in Table I are complex systems of unstable compounds that may undergo involved reactions even under mild processing conditions. Only the sugars are comparatively simple, well-defined, organic compounds and it will be seen that they also give rise t o complex reaction systems under the conditions of concentration and temperature reached in cooking candies. When several such ingredients of different classes are combined and cooked t o final temperatures of 130" C. (266" F.) or higher, the resulting system defies complete

CANDY MANUFACTURE

5

chemical analysis by any methods yet devised. Consideration of the present knowledge of the “browning” products formed by reaction of proteins with sugars, despite the very extensive research devoted t o the study of this reaction in recent years, exemplifies the complexity of the composition of such reaction mixtures. Nevertheless, a chemical approach t o the problem of candy composition is far from hopeless. Practical results of the int.ricate processes involved in candy production can be explained only in terms of the fundamental reactions of carbohydrates, proteins, and fats revealed by the latest investigations of their chemistry. The dependence of final quality upon properties of the ingredients and processing conditions can thus be understood better, and formulation will be placed upon a more logical basis. Individual types of candy are too numerous and varied t o serve as a practical classification for systematic review of their chemistry. For this purpose they fall conveniently into three large categories, each of which possesses certain distinctive chemical characteristics. First are candies made entirely of sugars, with or without flavors or colors, such as hard candies of both the plain and pulled types, stick candies, and crystallized creams. The second category includes those made largely of sugars, the properties of which are modified by the incorporation of small proportions, not exceeding 575 and usually much less, of nonsugar ingredients. Typical examples of this category are pectin jellies, marshmallows, and nougats. The third group comprises candies such as fudges, caramels, starch jellies, chocolates, and others that contain large proportions of ingredients other than the sugars. The logic of this classification is evident from the obvious fact that the all-sugar candies are relatively the simplest, a t least in formulation if not in chemical constitution, and the introduction of increasing amounts of nonsugar ingredients modifies the behavior of the sugars, in many cases superimposing additional reactions upon those which the sugars undergo. The most fruitful contribution of research in this field will continue t o be, as in the past, the gradual introduction of more exact and scientific guides for formulation and processing. Its objective should be t o provide explanations and better understanding of the art of candymaking, rather than an attempt t o supplant that art entirely by exact science. It should not be aimed a t such rigid standardization that the play of originality in devising novel products would be restricted; on the contrary, increased knowledge of the principles underlying the chemical processes involved should lead t o development of a n even greater variety of candies made from a wider range of raw materials. Reliable methods of measuring many properties associated with candy quality are still urgently needed for use in systematic research. The effects of variations of composition

G

L. F. MARTIN

and cooking procedures on texture, consistency or body, color, and flavor must be determined quantitatively for rapid, sure progress to be made in experimental work. Such methods are also invaluable for standardization and control of commercial production once optimum conditions have been established experimentally. Specifications for the quality of raw materials required to produce the best results can be improved as our knowledge increases regarding the importance in candymaking of specific properties and reactions of these materials. Storage stability, or “shelflife,” is of such great economic importance in the marketing of candy that it is imperative to consider the gradual changes in composition and properties that occur after it has been packaged. There is much yet to be learned of the nature and causes of changes in stored candies so that effective means of retarding them may be devised.

11. DESIRABLE PROPERTIES OF CANDY The variety of confections generally accepted by custom as candy is so great that a precise definition of the word is both impossible and unnecessary. Individual types may be described broadly, but cannot be defined strictly by narrow limits of chemical composition or physical properties. The only truly characteristic quality common to all types is that of sweetness. The British term sweets, applied to all except chocolate candies, is more accurately descriptive than our word derived from the Arabic Khandi. It is desirable, nevertheless, to consider briefly the range of variations in composition, physical properties, and organoleptic qualities that good candies are expected to possess. The ultimate goal of the application of research reviewed here is to improve and maintain these properties. 1. Physical Properties The important quality of many candies designated as texture is the sum or resultant of several physical properties including density, hardness, plasticity or elasticity, and consistency. It varies in different types from the soft, “tender” texture of marshmallows or chocolate cream centers to the glass-like hardness of the clear varieties of hard candy. The particular property or properties of primary importance vary according to the kind of texture desired in different candies. Quantitative measurements of physical properties have been employed to a very limited extent and texture is still evaluated qualitatively or described by the candymaker’s terms, “short,” “tender,” “firm,” “chewy,” etc. Whatever the texture, uniformity and smoothness are desirable almost invariably, as grainy or gritty candies are generally unpopular. The problem of translating these qualities into precisely defined, measurable

CANDY MANUF-4CTURE

7

properties has been solved in very few cases. Adaptation of methods th a t have been applied t o other products should prove effective for determination of the intrinsic properties upon which the texture of candies depends. This is necessary particularly in experimental work on improvement of texture. a. Density. The true density or specific gravity of all sugar hard candies does not vary significantly. Apparent density can be determined readily and is of more importance in relation t o the textures of many types of candy, but n o data have been published on this property. Variations are greatest in aerated candies made with frappes such as nougats and certain grades of fudge. Nougat texture in particular ranges from light, “short” types almost like fudge t o dense, “chewy” sorts approaching the density and qualities of caramels. Gelatin marshmallows vary in apparent density with differences in gel structure and moisture content. b. Hardness. This property, associated with elasticity and brittleness, is obviously of primary importance in relation t o the texture of those all-sugar candies that have minimum moisture contents. No suitable methods have been devised or adapted for measuring the hardness or brittleness of such candies. A major difficulty in marketing them is their tendency t o become sticky because of the hygroscopicity of the sugar reaction products formed in cooking a t high temperatures. This hygroscopicity is probably related t o hardness or other physical properties, as it is not simply dependent upon a low initial moisture content which is readily attained in modern vacuum cooking equipment. c. Plasticity. The texture of a large variety of candies is governed by this property. The quality described by candymakers as ‘(tenderness” is essentially dependent upon plasticity. The maximum permissible degree of tenderness is a prime attribute of the best grades of creams, caramels, nougats, fudges, and marshmallows. Both pectin and starch jellies are rated higher in proportion t o the extent t o which they possess and retain this quality. It can be developed t o the highest degree in any of these candies prepared for chocolate coating, but some tenderness must be sacrificed in less expensive grades th at must be durable enough to be shipped and sold in bulk. Loss of moisture decreases plasticity with resultant toughening of nougats, jellies, and marshmallows which is the principal cause of their deterioration in storage. Fudges, creams, and caramels are more prone t o grain and harden on drying. Quantitative measurements of plasticity are used for control of quality, particularly in large scale manufacture of jellies and gum candies, such as gum drops. The Humboldt penetrometer is modified for this purpose by adjusting the weight of the plunger to give a penetration of

8

L. F. MARTIN

1 cm., or 100 units, with samples of predetermined desirable plasticity (Alikonis, 1952a). This makes it possible t o detect deviations with greater speed and certainty. The standard Humboldt instrument has been used t o obtain data on the effect of emulsifiers upon the texture of starch jellies (Martin et al., 1952a). A test of the effectiveness of tempering in solid chocolate based upon measurements with a sensitive penetrometer has been suggested by Neville et al. (1950), whereas a less sensitive modification is proposed by Mickevicz (1950a) for determination of its surface hardness. d. Viscosity. Efficient tempering and application of liquid chocolate coatings are critically dependent upon viscosity, particularly in modern, continuous, enrobing machinery. The textures of finished coatings and candies are governed by plasticity of the solidified chocolate, but the two properties are related. Specifications for different grades of chocolate include the viscosity of the liquid material determined at temperatures somewhat above its melting point. The MacMichael (1915) viscosimeter has been employed generally for this purpose for many years (Stanley, 1939), although the need for more precise standardization of the procedure for making the determinations has been emphasized recently by Kempf (1949a). e. Consistency. The smoothness of texture essential for highest quality in most candies is approximated by the physical property of consistency. The exact definition of this property by Bingham (1930) includes plastic as well as fluid materials, and the properties of some candies are intermediate between the plastic and fluid states. Examples are the creamy fudges in which the formation of very small, uniform sugar crystals is induced by the use of separately creamed fondant, and the soft cream centers produced by invertase action th a t increases the ratio of the sirup to the crystal phases after they have been coated. No attempts t o measure the consistency of such candies have been reported. Neither methods used for fluids nor those applicable t o strictly plastic materials are readily adaptable for quantitative determinations in this intermediate range. Campbell (1940) has made an interesting study of the consistency of chocolate and its relation t o viscosity, employing methods and theoretical principles developed earlier by Williamson (1929). f. Color. Attractive colors are essential because color is known t o affect sales appeal directly, and t o influence the organoleptic response t o flavors, indirectly affecting their acceptability. This property of candies is assessed invariably by personal judgment. It can be standardized for many candy items with the available range of certified food colors. White candies, and those which depend upon reactions of the ingredients t o develop desirable colors, are not so readily standardized. Accurate,

CANDY MANUFACTURE

9

objective methods of measuring color have been applied t o sugar and other products, but there are no published reports of their use in candy manufacturing or research. g. Flavor. Candy flavors, like their colors, are judged subjectively. This property cannot be measured in any other way, yet it is undoubtedly the most important single physical property of the vast majority of candies. Fortunately, it is possible t o introduce almost any desired flavor by the use of either natural or approved synthetic flavoring materials. Standardization is more difficult for individual products such as caramels, chocolates, or fudges, the flavor of which depends upon variable natural ingredients or their reactions with sugars. 2 . Chemical Composition and Properties

A sufficiently high concentration of sugar t o make the finished products self-preserving is the only basic requirement of the chemical composition of all candies. This is consistent with providing the desired sweetness. Excessive sweetness is undesirable in many types of candy, requiring increased proportions of corn sirup or other less sweet ingredients t o modify the sugar present. It is also advantageous t o incorporate fats, proteins, minerals, and other natural nutritional values, although high-sugar contents make candies primarily an energy food. The range of proportions of the principal ingredients entering into the formulation of important types of candy is presented in Table 11. The composition of each type of candy varies within rather wide limits so that its optimum chemical requirements can be defined only in a general manner. Ingredients other than sugar are used for their various effects in modifying the properties of sucrose or reacting with it t o produce the desirable physical properties described in section 11, 1. All of these nonsucrose substances are effective in either preventing crystallization, or limiting it t o the formation of extremely small, uniform crystals under suitable processing conditions. Clear, white, or bright,ly colored candies must be produced under conditions that minimize possible reactions of the ingredients, whereas the flavors of caramels, taffies, brittles, and butterscotches are developed by controlled reactions leading t o the formation of complex end products. The final chemical composition of hard candies, crystallized creams, and jellies must be as nonhygroscopic as possible so that they will not become sticky. On the other hand, candies such as fudges and nougats develop undesirable textures rapidly unless they are sufficiently hygroscopic t o prevent loss of moisture. As the specific effects of individual ingredients will be considered in subsequent sections, details are superfluous a t this point.

10

L. F. MARTIN

TABLEI1 Range of Cooking Temperatures, Moisture Contents, and Proportions of Sugars of Principal Types of Candy

Type of candy

Final Final cooking moisture temperature content range range (OF.) (%)

Sugar ingredients" range (%I

Other principal Ingredientsa

Corn ________ sirup Range Sucrose Invert solids Ingredient (%)

Hard Plain Butterscotch Brittle Creams Fondant Cast Butter

275-338 240-265 290-295

1.0-1.5 1.5-2.0 1.0-1.5

40-100 40-65 25-55

0-10

235-244 235-245 235-247

10.0-11.5 9.5-10.5 9.5-11.0

85-100 65-75 50-65

Fudge

24s250

8.0-10.5

30-70

Caramel

240-265

8.0-11.5

0-50

Nougat Marshmallow Grained Soft Jellies Starch Pectin

255-270

8.0-8.5

20-50

0-10 Starch 0-1 25-40 Egg albumin 0-0.05 - 25-40 Butter 1-5 10-20 0-17 12-40 Milk solids 5-15 Fat 1-5 0-15 0-50 Milk solids 15-25 Fat 0-10 0-15b 30-60 F a t 0-5

-

-

0-60 35-60 Butter 20-50 -

1-7 -

5-10

-

240-245 12.0-14.0 50-78 225-230 15.0-18.0 25-54

0-5 0-10

15-40 Gelatin 40-60 Gelatin

1.5-3 2-5

230-235 220-230

0-10 -

28-65 Starch 30-48 Pectin

7-12 1.5-4

14.5-18.0 18.0-22.0

25-60 40-65

.0 Adapted from "Confectionery Analysis and Composition," Jordan, S., and Langwill, K. E., &fan,,facturing Confectioner Publishing, Chicago (1946). b Honey is often used.

3. Preservation of DesiraEle Properties

The marketing of candy involves storage for considerable periods because of the manner in which it is consumed. Demand fluctuates both seasonally and on occasions such as Christmas and Easter. Even after i t reaches the consumer it may be expected t o last longer than other foods without any special storage precautions such as those normally taken with perishable items. Candies have generally been stored and distributed under the most adverse conditions. They are practically immune t o microbiological spoilage and have been assumed t o be entirely stable in other respects. Efforts are being made t o devise better storage practices that will ensure the marketing of products that retain all of their desirable properties. Research has been in progress t o demonstrate

CANDY MANUFACTURE

11

the advantages of controlling both temperatures and humidities in storing candy for extended periods of time (Heaton and Woodroof, 1952). Experience with candies required for military rations and the necessity of stockpiling such items in times of emergency has given impetus t o these investigations (Cosler, 1951). As the widespread adoption of adequately conditioned storage and marketing facilities will require considerable time and investment, other research has sought means of extending the shelf-life of candies by suitable modifications of formulation and processing conditions (Martin et al., 1952b). a. Stabilization of Texture. Alterations of texture result in the most obvious deterioration of many types of candy. Marshmallows may become tough and rubbery; fudges, creams, and mints become hard and gritty because of excessive graining; pectin jellies mag either lose moisture and become cloudy and tough, or undergo syneresis and become sticky. Loss of moisture can be prevented by judiciously increasing the extent of inversion, or by using larger proportions of corn sirup for its hygroscopicity, up t o the maximum proportions shown in Table 11. Addition of smaller percentages of more hygroscopic ingredients has been resorted t o for the purpose of still further increasing moisture retention. The use of glycerin has been noted (Anon., 1937) 10% being incorporated in fudge and from 5 t o 15% in other candies. Alikonis (1952b) found that fudge and nougat are improved by sorbitol added in amounts of about 10% of the batch weight. It appeared t o be beneficial in gum drops only after storage for 9 months t o a year, but 5% and 10% additions of sorbitol had an adverse effect on the quality ratings of both pectin and starch jellies stored for 6 months according t o Heaton and Woodroof (1952). Detrimental effects of the use of 15% concentrations of sorbitol on the strength, texture, and sweating tendency of pectin jellies during 40 days storage under various conditions have been described by Poulsen (1953). The most important contribution on the effect of moisture changes upon texture and keeping quality was that of Grover (1947) who showed that the equilibrium vapor pressures of candies may be calculated approximately from their compositions. He established conversion factors for expressing the concentrations of other ingredients in the sirup phase in terms of equivalent sucrose, and tabulated data on the relative vapor pressures for total equivalent sucrose concentrations. This makes it possible to calculate, from a given formula, whether the candy will be in equilibrium with surrounding atmosphere at a particular relative humidity, or will gain or lose moisture by exposure. Measurements on actual candies were found t o agree well with calculated values. Equations for calculating the equilibrium humidities for various candies have been published more recently by Money and Born (1951).

12

L. F. MARTIN

The tenderness of starch jellies can be maintained for longer storage times by addition of small amounts of surface active compounds t h a t delay aging of the gel which increases its firmness. Naturally occurring traces of stearic acid have this effect as shown by Hamer (1947) who found that complete defatting increased the rate of aging of gels of corn and wheat starches. The effect of polyoxyethylene stearates in retarding this change in wheat starch gels was studied quantitatively by Lord (1950).

110

90 % u)

3

E

E

+ 0

.2

70

0

w

2 50

30

0

1

2

3

4

5

6

Storage time, months

FIG.1. Effect of polyoxyethylene stearate, in 0.5% concentration in delaying the aging (toughening) of starch jellies. (Martin et al., 1953a.) Curves 1 and 2 for jellies of 10% starch, 3 and 4 for jellies of 12% starch; 1and 3 show penetrometer readings with addedpolyoxyethylene stearate, 2 and 4 readings for controls without added emulsifier.

Results obtained by Martin et al. (1953a) for starch gum candies containing 0.5 % ’ additions of polyoxyethylene stearate based on the weight of starch used are shown in Fig. 1. After 6 months storage the jellies made with the emulsifier were as tender, by penetrometer measurement, as the regular controls that had been stored for about 2 months. Monoglycerides produce similar results, but both types of emulsifiers introduce the practical disadvantage of making the jellies too tender initially t o be removed from the starch molds within 48 hr., which is essential in commercial production a t modern equipment capacities. b . Rancidity. Vegetable and animal fats are important ingredients of caramels, fudges, some types of nougat, butterscotch, chocolate, and

CANDY MANUFACTURE

13

butter creams. Martin et al. (1951a, 1952c) have investigated the stability of fats incorporated in various typical candies during storage at 30" C. (86" F.) with respect t o oxidation and hydrolysis. It was established t h a t fat could be extracted from the candies without significant alteration of peroxide or free fatty acid contents t h a t had been determined prior t o its incorporation into standard formulas. This made i t possible t o subject butter and other fats t o storage tests, with and without added stabilizers, under the exact conditions of exposure t o air and moisture prevailing in normal candies. Rates of deterioration and the effectiveness of antioxidants or other stabilizers may differ under these conditions from those that would be observed in experiments carried out with isolated fats at higher temperatures (Mayberry, 1949). Extensive studies of the stabilization of the fats themselves are reported in the literature. All of the fats tested in candies hydrolyzed slowly, yielding measurable amounts of free fatty acids in a few weeks. Hydrolysis continued thereafter at increasing rates during the remainder of the storage periods of 4 t o 6 months. The number of experiments conducted was insufficient t o indicate possible relationship of the rate of hydrolysis t o the nature of the fat and moisture content of the candies. A few preliminary tests of candies containing butter or coconut fat gave results showing that the reaction may be retarded somewhat by addition of glycerol in amounts about equal t o t h a t of the f a t present, but no effective means of delaying hydrolysis sufficiently t o be of practical value has been found (Martin et al., 195313). Oxidative rancidity, as measured by the amounts of peroxide formed, was not detected in processed vegetable fats extracted from these candies until they had been stored for 6 t o 12 weeks at 30" C. (86" F.). Oxidation of this type of fat proceeded so slowly that i t was of no practical significance at the end of the 24 weeks during which the tests were continued. Animal fats proved t o be much less resistant t o oxidation. Martin et al. (1951b) found t h a t butter incorporated in cream fondant was stable during a short induction period of only 2 t o 4 weeks, after which peroxides were formed rapidly. The onset of oxidation could be delayed for a t least 5 months by adding the prescribed amounts of any of the widely used antioxidants such as NDGA (nordihydroguaiaretic acid), BHA (butylated hydroxyanisole), or propyl gallate t o the butter prior t o mixing it in the candy batch. These synthetic or chemically purified stabilizers may not be employed under existing Pure Food and Drug Act regulations prohibiting the use of non-nutritive ingredients in confectionery; however, inactive, dried yeast or concentrates prepared from oat flour were found t o be equally effective. The antioxidant properties of oat flour had been reported by both Peters and Musher (1937) and Conn and Asnis (1937),

14

L. F. MARTIN

and similar uses of yeast had been described by Musher (1942). Concentrates prepared from extracts of oat flour were used t o stabilize fats by Musher (1944). The vitamins of the B complex contained in both oats and yeast do not possess antioxidant properties according t o Gyorgy and Tomarelli (1943), and the nature of the specific compounds responsible for stability imparted t o fats by these products is unknown. Being natural, nutritive food products, dried brewer’s yeast or oat concentrates may be added t o candy in amounts equal t o about 3% of the total batch weight, which provides sufficient antioxidants t o protect the butter or other animal fat present. c . Chocolate Bloom. Chocolate does not become rancid very readily because of the inherent stability of cocoa butter and the low moisture content of coatings or solid chocolate. If it is used t o flavor fudges or similar candies containing moisture, Sjostedt and Schetty (1946) found the liberation of free fatty acid9 in storage t o be proportional t o the percentage of water present. A form of deterioration peculiar t o chocolate is the graying or discoloration of its surface accompanied by loss of gloss, described as “bloom.” Flavor may not be affected perceptibly, but the appearance of bloomed chocolates is so unattractive as t o render them unsaleable. Sugar bloom is a white or light gray coating of sucrose crystals, readily distinguishable by their birefringence in microscopic examination. As it usually results from condensation of moisture on the candy, i t can be prevented b y proper packaging and storage t o avoid exposure t o high humidity or sudden changes of temperature. Fat bloom is more unsightly and of far greater economic importance than sugar bloom. The discoloration in this case is known t o be formed of surface crystals of fat that has exuded from the chocolate, and is observed in different forms under widely varying conditions. Improper tempering or storage a t high temperatures increase the incidence of bloom, but the fundamental causes of the phenomenon are still obscure. Present knowledge of this problem remains essentially that summarized two decades ago by Whymper (1933a). Recent information on the nature of fat bloom has been disclosed largely in patents covering proposed methods for its prevention. Most of these methods involve modification of the fat intended t o convert it t o more stable forms. Cook and Light (1940) proposed elaidizing the fat by treatment of chocolate or the separated cocoa butter with sulfur dioxide or oxides of nitrogen in order t o isomerize the oleic acid present t o the higher-melting, elaidic trans acid. Extracted cocoa fat can be transformed similarly by the action of nitrous and nitric oxides in a mixture of acetone, amyl acetate, and a small amount of water according t o Eipper (1948). Attempts t o prevent fat bloom by adding surface-active

CANDY MANUFACTURE

13

materials to improve and stabilize the degree of dispersion of fat throughout the chocolate have met with but limited success. An early suggestion of this approach appeared in a patent on the use of oleodistearin or, preferably, an emulsifier produced by oxidation and polymerization of cocoa butter (Clayton et al., 1937). It was soon followed by a patent describing nonblooming chocolate produced by incorporating esters of sorbitol (Eipper, 1938). Stabilization against bloom by small additions of the aminoethyl esters, or carbobenzoxyaminoethyl esters of diacylglycerophosphoric acid has been claimed by Rose (1948a,b). None of these methods is known t o be in practical, commercial use. Various other additives or modifications of the fat have been tried, but all have the disadvantage of affecting flavor or texture or both adversely (Clay, 1953). Recent work on this problem has been summarized by Neville et al. (1950) in reporting the degree of success attained by their experiments with polyoxyethylene sorbitan stearate and sorbitan stearate in concentrations of about 0.5%. These authors point out the need for a more reliable and rapid test for bloom as a measure of the degree of stability imparted by these surface-active agents. They also believe that the complete solution of the problem may require modification of the fat itself by selective hydrogenation, isomerization, or fractionation. Several patents have been issued on this application of various modifications or combinations of fatty acid esters of sorbitan and polyoxyethylene as bloom inhibitors (Mayberry, 1951; Cross, 1952, 1953). An exception to the general conviction that bloom can be prevented by altering the condition of the fat is a process proposed by Rubens (1944) requiring the separate grinding of dry, fat-free cocoa solids t o reduce their particle size, the finest dust being recombined with fat. 111.

PROPERTIES AND

REACTIONS OF

S U G A R S I N CANDYMAKING

Candies made solely of sugars with small additions of flavoring arid coloring substances may be supposed to be the simplest from a superficial consideration of their formulas. This simplicity is only apparent, as finished products of this class, such as hard candy, may be quite complex chemically. High temperatures are necessary to produce them with sufficiently low final moisture contents. These conditions are favorable for dehydration and reversion of all of the sugars present in addition t o inversion of sucrose, and at least a portion of the simple sugars undergo profound changes in structure. The occurrence of such reactions accounts for the fact that the candy can be produced as a supercooled glass, as the presence of glucose or invert sugar alone would not prevent crystallization indefinitely. Controlled crystallization is of more importance than

16

L. F. MARTIN

chemical reactions of the sugars in creams and other all-sugar types made by cooking a t lower temperatures. T o understand the problems of hard candy manufacture it is necessary t o consider in some detail the present knowledge of the chemistry of individual sugars used and their reactions in concentrated solutions a t elevated temperatures.

1. Efect of Heat upon Sugars

a. Dry Sugars. Extensive investigations of sugar decomposition reactions have succeeded in establishing only the initial steps and intermediates with certainty. The products obtained by heating “dry ” sugars are not entirely distinct from those formed in extremely concentrated solutions a t comparable temperatures a s water is liberated in early stages of the process. Numerous studies of caramelisation and other colorforming reactions of sugars have been summarized by Zerban (1947). The simple, stepwise dehydration postulated by Gelis (1857, 1859) from results of his early experiments on sucrose has been shown t o be a n inadequate explanation of the behavior of this sugar. His “ caramelan ” and ‘‘ caramelin” were not definite compounds, and the structure of the first product formed by loss of one molecule of water, which he termed 1 1 caramelen,” is more complex than G&s supposed. Pictet and his coworkers (Pictet and Adrianoff 1924; Pictet and Stricker, 1924) employed vacuum t o remove the water formed with better control of the rate of heating and succeeded in isolating the initial products of sucrose dehydration. They determined the structures of these compounds as completely as was possible a t the time of their work. Glucose and levulosan, the anhydride of fructose, are formed first without loss of water in the following reaction : CIZHZZOII 4 C6H1206 Sucrose

+C~HIOO~

Glucose

Levulosan

The levulosan is identical with the anhydride previously obtained by Pictet and Reilly (1921) on heating dry levulose. I n the next step of sucrose decomposition, water is eliminated from the glucose which Pictet and Castan (1920) had shown t o form an anhydride in the reaction:

+

C6H1206 4 C6Hl0O6 HzO Glucose Glucosan

Pictet and Adrianoff (1924) found that the glucosan and levulosan next react t o form isosaccharosan, the anhydride of sucrose. C6HloOs Giucosan

+ C6H1006

+

Levulosan

C~2HzoOlo Isosaccharosan

CANDY MANUFACTURE

17

This is the product of the first step of dehydration which Gelis termed (‘caramelen,” but which he was unable t o purify sufficiently for determination of its exact composition and structure. Further changes are complicated by the tendency of glucosan t o polymerize t o dilevoglucosan and higher polymers (Pictet and Ross, 1922), whereas levulosan is even more readily converted t o diheterolevulosans by dimerization (Pictet and Chavan, 1926). The importance of these sugar anhydrides, or reversion products, in candy cooking will be appreciated from the fact that high yields of diheterolevulosans are obtained by simply refluxing an 80% solution of fructose, as demonstrated later by Wolfrom and Blair (1948). b. Concentrated Sugar Solutions. It is not possible t o limit chemical changes on heating either dry sugars or their solutions t o the relatively simple, initial stages of reaction. Subsequent reactions proceed simultaneously with the consecutive steps of dehydration. Dark-colored polymerization products are formed by further loss of water from the sugar anhydrides (Wolfrom et al., 1951) together with simple end products of decomposition such as formaldehyde and hydroxymethylfurfural (Joszt and Molinski, 1935; Wolfrom and Blair, 1948). A key t o the transformations of sugars in solution under a wide range of conditions is provided by the rearrangement reactions of glucose and fructose discovered and first studied in detail by Lobry deBruyn and van Ekenstein (1895, 1896). These transformations are involved also when sucrose alone is used in candymaking, as tartaric acid or similar “doctors” aregenerally added t o provide sufficient acidity for its inversion t o the simpler hexoses. From any of these 3 sugars, or combinations of them, a system is established in which some glucose, fructose, and mannose are present and tend t o establish equilibrium with one another. Alkalis are particularly effective catalysts for the transformation of glucose into fructose and mannose, but the Lobry deBruyn rearrangement takes place throughout a wide range of pH. It has been found t o occur a t p H 6.4-6.6 in the presence of phosphates (Englis and Hanahan, 1945) as well as in solutions of either glucose or fructose buffered a t p H 7.5 and heated t o 90” C. (194” F.) (Schneider and Erlemann, 1951). Mathews and Jackson (1933) determined that fructose is most stable at p H 3.3, and Singh et al. (1948) found glucose t o have maximum stability in the same range at about p H 3.0. The rearrangement and interconversion of the simple sugars are invariably accompanied by side reactions leading t o the formation of sugar anhydrides and sugar acids, together with products of further degradation very similar t o those obtained by heating dry sugars. I n strongly acid solutions these side reactions predominate, but they occur to some extent under any conditions. Pictet and Chavan (1926) prepared fructose anhydrides by the action of cold, concentrated hydrochloric acid

18

L. F. MARTIN

on the sugar, and Wolfrom and Blair (1948) proved by chromatographic methods of purification that the products formed are diheterolevulosans. The rearrangement of sugars and the formation of by-products proceed t o various extents depending upon conditions of concentration, temperature, pH, and the catalysts present such as various bases used in the alkaline range. The effects of variations in all of these conditions upon the transformation of glucose in the Lobry deBruyn reaction have been investigated most thoroughly by Gottfried and Benjamin (1952). Their results show that the maximum yield of fructose attainable is 2175, accompanied by the formation of 8.5 % of unfermentable sugar anhydrides or reversion products, and 3 % of sugar or simpler organic acids, estimated as saccharic acid. At high temperatures and concentrations more glucose was transformed, largely into the side reaction products. It is interesting that the same reactions proceed slowly a t low temperatures under very mild conditions. Earlier results reported by Wolfrom and Lewis (1928) for the very slow conversion of glucose in 1 M solution by 0.035 N calcium hydroxide a t 35" C. (94.5' F.) are predictable from the data and constants of Gottfried and Benjamin, obtained from concentrated solutions at their boiling points. c. Importance I n Candy Manufacture. Consideration of the reactions described in the preceding sections makes it apparent t h a t even simple hard candies are much more complex than mere supercooled solutions of partially inverted sugar. The physical-chemical principles governing inversion as applied in candymaking have been treated comprehensively and precisely by Heiss et al. (1953), who point out the difficulty of exact control of this comparatively simple reaction. Inversion is but one of several possible initial steps in the sequence of reactions leading t o the formation of anhydrides, sugar acids, reversion, and caramelizatiori products, all of which may be derived from a single starting material, sucrose. Moreover, a minor proportion of all-sugar candies are made of granulated sugar alone, substantial proportions of invert or corn sirups being used in the bulk of the production. Corn sirup itself is a complex mixture of glucose with maltose, dextrins, sugar anhydrides, and polymers which have not been resolved completely. Less than one-fourth of the solids in the most widely used grade of corn sirup, with a dextrose eyuivalent of 42%, are actually glucose. This will be seen in Table 111, which gives approximate compositions of four different grades as compiled by Meeker (1950). Chromatographic methods are required for the separation of the noriglucose fraction into its constituents. This was accomplished in the case of a specially prepared hydrolyzate of the amylose fraction of starch by Dimler et al. (1952). Their results illustrate the complexity of starch hydrolyzates in general. The greater part of this fraction con-

19

CANDY MANUFACTURE

TABLEI11 Approximate Compositions and Properties of Corn Sirupsa ~

Types of Sirups

Composition Moisture Total solids Dextrose Maltose Higher sugars Dextrins Ash Properties Density, “Be. Dextrose equivalent Approx. relative viscosity at 100” F., poises PI3

Acid-enzyme High-acid Medium conversion conversion Regular conversion (%) (%I (%) (%) Total Solids Total Solids Total Solids Total Solids

19.7 80.3 18.0 17.0 16.0 30.0 0.3

19.0 81.0 22.0 26.0 21.0 21.0 20.0 11.0 37.0 23.0 0.37 0.3 43a 42

160 4.7-5.0

-

32.0 26.0 13.5 28.0 0.37 43” 52

104 4.7-5 .O

18.5 81.5 30.5 28.0 13.0 10.0 0.3

-

37.5 34.0 16.0 12.0 0.37

18.5 81.5 33.0 4 0 . 5 23.0 28.0 6.5 8.0 19.0 23.0 0.3 0.37

43” 63

43” 60

60 4.7-5.0

80 4.7-5.0

” From Meeker (1950).

sists of trioses, tetroses, and higher polymers of glucose, somewhat less than half of it being accounted for as maltose hydrate. Maltose, like sucrose, was shown by Pictet and Marfort (1923) t o yield an anhydride, maltosan, that may be formed when it is heated t o the temperatures required for hard candy production. Variations in the quality of granulated sugar for candymaking purposes can be explained by the effect of traces of catalytic substances on the reactions of sucrose a t elevated temperatures. The empirical “candy test” developed for use in the sugar industry has been widely used by candymakers t o determine the “strength” of sugars. This property has been defined as the resistance of the sugar t o inversion under standardized conditions of cooking t o finished hard candy (Ambler, 1927). The shorter the time required for the test candy t o crystallize, the stronger the sugar. As granulated sugar is a highly refined and exceptionally pure product, variations in quality must be caused by extremely small traces of impurities. It has been noted that a “weak” sugar may be strengthened by addition of as little as 0.001 % of soda (Ambler, 1927). Dehydration and caramelization of sugar are more sensitive t o changes in p H than inversion in the range of this test. Pucherna (1950) studied the effects of various catalysts upon the formation of color in an empirical caramelization test.

20

L. F. MARTIS

His procedure was improved by Kalyanasundaram and Rao (1951) who carried out the test by heating sugar in anhydrous glycerine. They investigated the effects of addition of 0.1 % of various substances on both color formation and sugar decomposition. Sodium chloride had least effect in destroying sugar and developing color, although it produced a high percentage of inversion. I n contrast, potassium chloride caused very little inversion but Catalyzed the destruction of 4.2% of the total sugar. Water added at a concentration of 0.1% caused both inversion and destruction of sugar. As would be expected, ammonium sulfate and chloride not only inverted the sucrose almost completely, but caused destruction of substantial percentages of the reducing sugars formed with the maximum development of color. IV. MODIFICATION OF SUGARPROPERTIES BY MINORINGREDIENTS Nearest in simplicity of formulation t o hard candies and other allsugar types are those in which the required properties are developed by incorporating small percentages of ingredients other than sugars. Reference t o Table I1 shows t h a t nougats, marshmallows, and pectin jellies are the principal examples of this class. Taffies, kisses, and molasses types in which flavor and color are the result of reactions of the sugars catalyzed by traces of nonsugar ingredients should be included in this category. Creams are not included because the optional small additions of starch or albumin occasionally used are not essential for the control of crystallization upon which their properties depend. Butter is used in butter creams primarily for flavor, rather than for modification of texture, and is usually added under conditions that minimize reaction with the sugar. The principal minor ingredients employed t o modify the properties of sugar are egg albumin, gelatin, and pectin, with less general current use of agar, tragacanth, and similar natural gums. Nougats are cooked a t high temperatures so that chemical reactions of the sugars are as important in their preparation as is the modification of texture by aeration with whipped egg albumin or other protein whipping agents. Physical properties are more important than chemical changes in the production of marshmallows and pectin jellies that are processed a t much lower temperatures. Their textures depend upon the formation of gels of sugars with the added gelatin or pectin under suitable conditions t o provide sufficient body with maximum tenderness. These candies must be produced with high moisture contents that must be retained in storage t o prevent deterioration of texture. 1. Protein Whipping Agents

a. Egg Albumin. The function of protein whipping agents in candy production is t o provide lightness of texture and low density in all-sugar

CANDY MANUFACTURE

21

formulations that, without such modification, would be essentially hard candies. The procedure generally followed is to prepare a portion of the batch of sugar sirup separately by whipping with albumin to produce frapp6 or nougat cream. From 10% t o 15% of this preparation is added with whipping to the balance of the cooked sugar sirup after it has cooled to a suitable temperature. Maximum foam volume and stability are the important requirements of proteins employed for this purpose. Some studies have been made of the whipping capacity of egg albumin prepared under different conditions, but there are no published results of research

PH

FIG.2. Dependence upon p H of the foaming of solutions of pure (crystallized) egg albumin, (1) dissolved in pure water, and (2) dissolved in 0.001 N KOH initially. (Thuman et al., 1949.)

on the optimum properties required for candy manufacture. The dependence of whipping quality upon p H has been determined for both fresh and frozen egg whites by Bailey (1935). Her results show th a t freezing and storage in the frozen state have a negligible effect upon the volume or stability of the whipped material. Egg albumin purified by crystallization was used by Thuman et al. (1949) in their precise study of foam volumes produced by bubbling under various conditions. The results which they obtained a t different values of p H are shown in Fig. 2 . Pure albumin dissolved in water yields maximum foam at p H 4.0,

22

L. F. MARTIN

slightly on the acid side of its isoelectric point, but no foam in the range from p H 7-11 adjusted by adding potassium hydroxide, as shown by curve 1 (solid). Curve 2 (dashed) is a plot of the foam volumes determined when the albumin was first dissolved in alkali, 0.001 N KOH, in which case foams are produced in the alkaline range as well as a t about p H 4 (the maximum production). Alkaline earth salts were found t o increase foaming in the alkaline range but caused a decrease in acid solutions. Dried egg albumin is used in practically all candymaking applications of the protein, as it can be stored and weighed conveniently. Drying has little effect on the foam-forming characteristics, freshly dried albumin being almost equivalent t o whole egg white in this respect; however, it tends t o develop color and off-flavor with loss of whipping power during storage unless it is properly prepared. Stewart and Kline (1941) discovered that small amounts of glucose normally present in egg white are the principal cause of this deterioration in storage. As little as 0.02% of glucose, or of other reducing sugars substituted for it, were shown t o produce rapid coloration, a decrease in solubility, and loss of whipping capacity. Practical methods were developed for eliminating the glucose prior t o drying. This is accomplished by natural or spontaneous fermentation, by inoculation with specific fermenting organisms, or more effectively by fermentation with yeast (Ayres and Stewart, 1947; Stewart and Kline, 1948). Research on the development of this process also has added t o knowledge of the control of temperature and p H in the drying operation necessary t o produce dried albumin of the best whipping quality (Kline and Stewart, 1948). The practical superiority of fermenfed and dried egg white has been established by Carlin and Ayres (1951) by comparison with unfermented material in angel cake baking. The fermented product retained 9501, of its effectiveness for this purpose for 2 weeks, and 90% for 12 weeks in storage a t 40" C. (104" F.), whereas the untreated egg white lost from 50% t o 66 % of its cake-volume producing quality under these conditions. Similar practical storage tests of the improved whipping agent for candy frappe production have not been carried out, but obviously would be of value in this application of dried albumin. b . Soy Protein. Effective whipping agents have been developed from soy flour following the determination of the essential conditions for their preparation by Monaghan-Watts (1937). F a t must be extracted completely by petroleum ether or other hydrocarbon solvents, as alcohol extraction has an adverse effect upon whipping quality. The product can be deflavored by heating for twenty minutes a t 130" C. (266' F.) under vacuum. It produced foam volumes equal t o those obtained with egg white whipped under the same standardized conditions. Maximum

CANDY MANUFACTURE

23

volumes were observed just above and below the isoelectric point of the protein a t p H 4.1, which is approximately the same as that of egg albumin. Addition of salt in concentrations up t o 2% increased the whipping capacity of this protein by as much as 20%. Perri and Hazel (1947) determined the stabilized foam volumes produced by bubbling nitrogen through solutions of soy protein at different p H values in the range from 2.0 t o 12.0. These measurements were made by a n adaptation (Perri and Hazel, 1946) of the method and apparatus developed by Clark and Ross

PH

FIG.3. Dependence upon pH of the stabilized foam volumes produced by bubbling nitrogen through solutions of soy protein. (Perri and Hazel, 1947.)

(1940) with the results shown in Fig. 3. The similarity of the behavior of soy protein t o that of egg albumin with respect t o p H will be evident from comparison of these results with curve 1 of Fig. 2. Maximum foamforming capacity in the practical p H range is observed a t values near the isoelectric points of both proteins. The application of commercial soy protein whipping agents t o production of frapp6 for use in nougats and other candies has been described by Butler (1942). Turner (1946) states that this protein does not coagulate on heating in the same manner as egg albumin in order t o maintain the body of nougat candies, but that it dissolves more readily and produces maximum volumes of frappe with the limited amount of water in concentrated sugar sirups. Soy protein may be used in frappe in proportions

24

L. F. MARTIN

of 2 % t o 3 % of the weight of the sugars, as with egg albumin, but modified methods of preparation are recommended t o keep the moisture content below 18%. Unless used promptly, such frapp6s shrink in volume during storage t o a much greater extent than those produced with egg albumin (Clay, 1953). c. Milk Protein. Preparation of a whipping agent from the protein of skim milk or whey has been described by Leviton (1938) who employed extraction with aqueous alcohol t o obtain a soluble, undenatured product. Oberg et al. (1951) treated skim milk with acid, followed b y digestion with trypsin a t p H 8.0 and 50" C. (122" F.) t o obtain material that produced stable whips. Their properties have not been studied as extensively as those of egg albumin or soy protein, and they have been recommended generally only as a partial replacement for egg albumin in frapp6s. 9. Gelatin Gelatin is used principally for the production of marshmallows, ranging from the soft, standard type t o harder-grained candies molded in various novelty shapes. Both kinds are usually molded in starch that must be dried sufficiently t o form a durable outer skin while absorbing moisture from the candies t o produce proper gel formation. Soft regular marshmallows must retain the relatively high moisture content shown in Table I1 t o have good storage quality. Measurements of gel strength with the instrument devised by Bloom (1925) are the generally accepted basis for grading and standardizing the quality of gelatin. The procedure for making the determinations has been described by Richardson (1923). Bronson (1951) considers gelatin of 225 Bloom most suitable for making regular marshmallow, and lower strengths of 75 t o 125 Bloom best for producing grained novelties, but i t has been demonstrated by Gorfinkle (1953) and others that the final texture of the candy depends on other properties as well as Bloom. Whipping power, which is not related t o gel strength, is deemed t o be equally important by Clay (1953). Steigmann (1944) investigated gel setting times and the viscosity of gelatin solutions as possible measures of quality. The utility of these measurements, in which observed differences were magnified by protein dispersing agents such as urea and Teepol, remains t o be established by their correlation with the gel strength and whipping properties that are essential in candy production. The strength and stability of gels formed by gelatin depend upon its molecular size and the distribution of acid and basic groups in its structure. Friedman et al. (1939) attempted t o establish the molecular weight o€ gelatin by determinations of the acid and base combining capacities. More accurate estimates were obtained by Mosiman and Signer (1944)

CANDY MANUFACTURE

25

from data on the sedimentation equilibrium in the ultracentrifuge, which show that there are two distinct protein fractions with molecular weights of approximately 16,000 and 89,000. The effect of p H upon the disaggregation, or average molecular size of the larger protein molecules, was studied by Friedman and Klemm (1939) by measurement of their diffusion rates in solution. If ion effects were suppressed by potassium chloride in concentrations of 0.1 N , there was little change in molecular

Concentrattor molar

FIG.4. Variation of the setting time of gelatin gels with increasing concentrations of sucrose and leviilose. (Curves plotted from data of Friedman and Shearer, 1939.)

size between ~ € 1 2 . 0and 6.4. The variation of setting time of a gelatin gel caused by low concentrations of sugars was investigated by Friedman and Shearer (1939) and is of particular interest with respect t o use of the protein in candies. Their results, shorvn in Fig. 4, extend only to 0.2 JL solutions containing approximately 6.75 % sucrose and 3.6 % levulose, but it would be desirable t o determine whether the trend of decreasing setting times holds for concentrated solutions of these and other sugars. Invert sugar and glucose might be expected t o act like levulose in bringing about much more rapid setting than granulated sugar a t equal concentrations, at least within the range covered by these experiments. The addition of sodium hexametaphosphate or other complex phosphates in concentrations of 0.5% t o 5.0% is claimed by Grettie (1940) t o improve the whipping properties of gelatin in marshmallow compositions without decreasing their viscosity or gel strength. Collins (1940) patented the use

26

L. F. MARTIN

of sodium bitartrate, tartrate, lactate, acetate, or citrate in concentrations sufficient t o develop a final p H of 3.0 t o 4.7 t o produce smoother, fastersetting gels. 3. Pectin

Pectin is a valuable by-product of fruit processing as well as an essential ingredient of a variety of food products, including candies. Although it is found almost universally in plants of all species, suitable grades for use in the production of sugar jellies are obtained commercially from only two sources, apple pomace from cider presses and citrus peel remaining after juice extraction for canning. Innumerable investigations have been carried out on the theoretical as well as practical aspects of the varied uses of pectin. An excellent comprehensive treatise on the subject is now available (Kertesz, 1951a), and Baker (1948) has summarized some of the major recent developments in pectin chemistry more concisely. Most of this research has dealt with problems in the production of fruit jellies having lower solids contents and weaker gel structures than are required for pectin jelly candies. Some fundamental studies have included the range of high-solids concentrations. Recent research has established the details of the structure of pectic substances completely enough t o provide explanations of their behavior in terms of chemical constitution. Although the data obtained in many practical investigations of the low-solids gels may be extrapolated as a guide for the use of pectin in candy, there is urgent need of similarly thorough research on gels of high-sugar content. The principal value of this brief review of the subject will be t o indicate some of the lines that such research might. follow. a. Chemical Nature of Pectin. The source of pectin is an insoluble plant constituent, protopectin, which can be solubilized and extracted only by treatment with acids, alkalis, or enzymes. Some hydrolysis is necessary to convert it t o soluble forms and this inevitably alters the product in two respects: the size and homogeneity of the polymer are reduced, and some of the methoxyl groups are removed. Evidence summarized by Percival (1950a) has established the structure of unaltered, ideal pectin t o be completely methylated, branched chain polymers of galacturonic acid as first suggested by Morel1 et al. (1934). Such an ideal pectin should contain 16.3% methoxyl, but it has not been isolated as yet and all of the substances actually studied and used are pectinic acids. The general term pectin (or pectins) designates those water-soluble pectinic acids of varying methyl ester content and degree of neutralization which are capable of forming gels with sugar and acid under suitable conditions (Kertesz, 1951b). Complete demethylation produces pectic acid. The unit linear chains have been shown by Jansen et al. (1949) t o be composed of a

CANDY MANUFACTURE

27

minimum of about 32 galacturonic acid groups, and these units are joined by branching t o form polymers of very high molecular weights. Average molecular sizes estimated by various methods range from 16,000 t o 50,000 according t o Svedberg and Gralhn (1938), and were found t o be above 100,000for some pectin preparations studied by Owens et al. (1946). The foregoing very brief description of the structure of pectinic acids provides some explanation of the variability of commercial pectins. The polymer sizes or molecular weights are averages determined for particular pectin preparations which may be mixtures of polygalacturonide esters of widely varying sizes. This heterogeneity and the fact t h a t different fractions of the material may differ also in degree of demethylation, account for variations in gel-forming properties (Speiser and Eddy, 1946). This has been explained in detail by Kertesz (1951~).I n addition, significant amounts of araban and galactan are so closely associated with the pectin that they were believed t o be essential units of its structure until evidence t o the contrary was obtained by Hirst and Jones (1939). They may be present in amounts ranging from 5% t o as much as 70% of the weight of total pectic substance (Kertesz, 1951d), and are termed “ballast.” The practical use of pectin has been established and simplified only by long and continuing efforts t o perfect methods of standardization and grading (Wilson, 1926; Baker and Woodmansee, 1941), without which it would be impossible t o depend upon the performance of commercial preparations. h. Applzcations i n Candy Production. As pectin jelly candies necessarily have a high-sugar content, they are produced with regular, highmethoxyl pectin, generally of 150 grade. One pound of such pectin will suffice t o produce jelly of prescribed strength from 150 lb. of sugar under standardized conditions. Details of the various methods for determining jelly grade are described by Kertesz (1951e). Such high-methoxyl products are all demethylated t o a considerable extent during extraction, ranging from a minimum of about 7 % t o slightly over 11% methoxyl content. Variation of the degree of esterification within this range and differences in the size and homogeneity of the polygalacturonide molecules affect the gel-forming properties. Production of slow-setting pectins by patented acid treatment procedures is believed by Baker and Goodwin (1944) t o depend primarily upon reduction of methoxyl content, although all high-methoxyl pectins are not necessarily quick-setting and those of high jelly grade tend t o set slowly. Slow-set types that start t o gel a t about 54” C. (130” F.) in 65% sugar solution are recommended for ease of manipulation and development of better body in cast jellies. If the pectin contains more than about 7 % methoxyl, it will not produce a jelly unless sugar is present in concentrations of at least 40%,

28

L. F. MARTIN

or preferably higher. This distinguishes the low-methoxyl pectins from ,, the regular, less demethylated products that have been standardized for sugar jelly production. Myers and Baker (1934) first showed that pectins containing as little as 3% methoxyl can produce useful gels if they have been demethylated under conditions that do not cause extensive depolymerization. Speiser et al. (1947) have presented evidence t h a t the gel in this case is of a different type, being formed by ionic bonding in the presence of calcium ions through the preponderance of free carboxyl groups. With ester groups predominating in the high-methoxyl pectins, gel formation depends upon hydrogen bonding for which the hydroxyl groups of sugar are essential. Although the high-sugar concentrations of candy are more than ample t o form regular pectin jellies by hydrogen bonding, a n interesting range of properties can be developed by various modifications of the low-methoxyl products with appropriate adjustment of concentrations, pH, and calcium ions used. Demethylation has been effected by acid (Baker and Goodwin, 1941), by alkali (McDowell, 1951), or by the enzyme, pectinesterase (Willaman et al., 1944), either during extraction or by treatment of isolated high-methoxyl pectin. The different methods yield products having dissimilar gel-forming characteristics, thus affording a wide range of possible variations of jellies. Owens and Maclay (1946) have investigated the properties of 65 % sugar gels formed by pectins ranging from 10.8% methoxyl t o as low as 4.5% methoxyl. The low-methoxyl products used were variously demethylated by acid, alkali, or the enzyme. Gel strengths were measured by a rigidometer based upon the principles of the B.A.R. jelly tester (Lampitt and Money, 1936, 1939; Campbell, 1938) and, although they were less than the strengths desired for candies, the results are of interest in indicating the possibilities of using the modified products in high-sugar content jellies. The.limiting p H for gel formation decreased from p H 4.5 for the undemethylated pectins t o p H 2.8 for those of lowest methoxyl content. Below the limiting values, no maxima were observed and gel strength for either type of pectin remained constant down t o p H 1.4. Increased sensitivity t o the presence of calcium was observed as the methoxyl content decreased, as reported by others (Hills et al., 1942; Baker and Goodwin, 1944). An especially interesting observation in this study was the dependence of gel strength upon the molecular weight of the high-methoxyl pectins of approximately the same methoxyl contents. The enzyme-demethylated product produced weaker gels than those demethylated by acid or alkali under comparable conditions. The application of regular grades of pectin in candymaking has been described by Cruess (1946), and Cruess et al. (1949) have published a number of formulas for candies made from fruit juices, employing 150

29

CANDY MANUFACTURE

grade pectin. If alkali-demethylated, low-methoxyl pectin is used, satisfactory jellies can be made with larger proportions of corn sirup substituted for sugar according to Hall and Fahs (1946); and Martin et al. (1952b) described the preparation of candies containing large percentages of honey with a similar type of pectin. Demethylated pectins form gels of high as well as low-sugar content according t o Angermeier (1953), but the former have not been studied as extensively as the low-solid gels for which these pectins are uniquely suited. Kertesz (1951f) states that, “It is also clear that the commercial value of low-ester pectins must be defined in a different manner than the old-fashioned high-ester types.” The gel-forming power of low-ester products has not been standardized in terms of grade, and acid, alkali, or enzyme de-esterified pectins cannot be used interchangeably in jelly formulas. Poulsen (1953) has presented evidence that standard, candy-type jellies made with regular pectin are superior to similarly formulated low-methoxyl pectin jellies in texture and resistance to sweating.

4. Natural Gums The gel-forming substances classified as natural gums are now used t o a negligible extent in candy production. Only scattered references t o their use in any types of candy appear in a recent reference book on these gums by Mantell (1945), and no research on candy applications of gum arabic or tragacanth has been published. Gum drops, originally produced with gum arabic, are currently made with either starch or pectin. The use of gum tragacanth as a binding agent in lozenge paste has been superseded by gelatin or dextrin (Clay, 1953). Combination of gum tragacanth with lactates improves its dispersion in a product described by Buchanan (1945) for use in confectionery.

v. MAJORINGREDIENTS

OTHER

THAN

SUGARS

Many important types of candies derive their characteristic properties from nonsugar ingredients that are used in such substantial proportions that they not only act as modifiers of the sugars, but impart their own properties to the confections. The principal examples are fudges, caramels, and starch jellies. As shown in Table 11, the milk solids, fats, and starch constitute from at least 5% to a maximum of 35% of the finished batch weights of these candies. Cocoa solids and cocoa butter are major ingredients of chocolate goods, imparting the flavor and desirable texture. Cooking procedures and temperatures may be regulated primarily t o achieve optimum physical blending or, in some cases, they may be such as t o cause extensive chemical reaction and modification of the ingredients, Nuts, peanuts, and fruits that are incorporated physically in candies

30

L. F. MARTIN

are in this category. Pertinent results of research on each of these diverse food materials are summarized in the following sections. 1. Milk Products

Milk introduces proteins, fat, mineral salts, and a n additional sugar, lactose, into the composition of candies in which it is used. Concentrates such a s sweetened condensed or evaporated milk are generally employed rather than whole milk. Skim milk and nonfat milk concentrates or solids find wide application. The dry milk powders are used extensively in fudge and chocolate, and suitable grades, properly reconstituted, can be used satisfactorily in making caramels. Typical caramel flavor is developed by the proper reaction of milk and sugars. Additional butter is incorporated in limited quantities of these candies in lieu of the vegetable fats used for their large-scale production. The use of butter t o develop the characteristic flavor of butter creams has already been noted, and its reactions with sugars in cooking a t higher temperatures t o a low final moisture content produces butterscotch. Milk solids products of various kinds are constituents of milk chocolate candies and coatings. Numerous reactions in addition t o those of the sugars previously described in section I11 become possible in complex formulations involving constituents of milk. The predominant chemical changes are those resulting from ‘ I browning” reactions of the milk proteins with sugars. Attention was first directed t o such phenomena by the work of Maillard (1912, 1913), and they have been the subject of increasingly extensive investigation since that time. The mechanism of the interaction of natural products containing aldehyde and amino groups is not definitely established, although it has been studied in experiments far too numerous t o be described here. The most significant results of this research and their interpretation will be found in comprehensive recent reviews such as that by Danehy and Pigman (1951). The “browning” reactions in milk itself are particularly important in the production and storage of dry milk products, and they have been investigated very thoroughly in recent years. A pertinent study of the development of color in lactose solutions was carried out b y Webb (1935) with the conclusion that lactose caramelization alone does not account for the formation of color in evaporated milk. Phosphate in the buffer used in some of these experiments had a specific color-producing effect. Addition of amino compounds led t o formation of more color than that attributable t o caramelization of the sugar. From a later study of the effect of lactose concentration, pH, and other variables upon the formation of color in casein-lactose solutions, Kass and Palmer (1940) held t o the conclusion that lactose caramelization is responsible for changes observed on heating milk. More

CANDY MANUFACTURE

31

recent evidence from determinations of the decrease in the amounts of certain amino acids (Patton et al., 1948) and of amino groups (Mohammed et al., 1949) of proteins upon heating with glucose supports the view that browning is a protein-sugar reaction. Lea (1948) showed that the amino groups of milk proteins are lost in a 1: 1 ratio t o the disappearance of reducing sugar in the course of the reaction. Tarassuk (1947) has summarized the results of studies of the proteinsugar reaction bearing upon the practical problem of alteration of color and flavor in production of evaporated milk. He added evidence that the presence of oxygen at sterilizing temperatures is a n important factor in accelerating the adverse effects of further decomposition of the initial condensation products. Oxygen is consumed and carbon dioxide is formed in these later stages of the Maillard reaction. Similar changes occur, although more slowly, in the “ d r y ” state t o which Lea (1948) and his co-workers have directed their attention (Henry et al., 1948; Lea and Hannan, 1949, 1950). Knowledge of the chemistry of “browning” is of fundamental importance in several respects for the production of candies cmtaining any form of milk solids. It affects the quality of condensed and dried milk products employed as ingredients. Conditions must be regulated t o limit the extent t o which this reaction takes place in the production of fudge or milk chocolate, in which i t is desirable t o retain unaltered milk flavor. Improved caramel flavors can be developed by taking advantage of conditions t h a t have been found t o accelerate reaction of the milk proteins with sugars, or t o alter the course of the browning reaction. Storage life of the candies is affected by changes resulting from the slow interaction of the ingredients when at a lowmoisture content, which has been the subject of investigation by the Cambridge (England) group under Lea. Pertinent results have been reported by Lewis and Lea (1950) in their study of the loss of amino nitrogen of casein in its reactions with glucose, fructose, maltose, and lactose a t low temperatures and a low-moisture content. This change proceeded most rapidly with glucose, a t lower and approximately equal rates with either maltose or lactose, and most slowly with fructose. Candymaking would profit by the extension of such research t o the effects of excess proportions of the sugars used upon the course of the browning reaction. Butter is used for its flavor in some fudges and caramels, and in butter creams. It is incorporated in such candies under conditions t h a t minimize reaction or alteration. Means of retarding the development of oxidative rancidity have been described in section 11, 3. Richmond (1953) reports t h a t the keeping quality is generally known t o be improved markedly by adding the butter t o the batch at the highest possible temperature that

32

L. F. MARTIN

will not impair its flavor, probably by inactivation of the enzymes. Martin et al. (1951b) have reported an experiment in which freshly churned butter was found t o produce inherently longer lasting candies than butter of apparently equal quality that had been stored under refrigeration for a month. Similar observations had been made by Holm and Greenbank (1923) on the keeping quality of butterfat in milk powders. The fat was more stable in powders prepared after holding the milk for only 12 hr. than in those produced from milk held for 1 day. These authors consider the deterioration of butterfat to be primarily the result of enzyme action. Butter is stabilized by cooking t o the high temperatures and final concentrations required t o develop the flavor of butterscotch candy, but no research has been reported on the chemical changes occurring under these conditions. 2. Fats Other than

Dairy Butter and Cocoa Butter

Large quantities of processed vegetable fats are consumed in candy manufacture, principally in fudges, caramels, and nougats, with smaller quantities being required for a variety of other candies, such as creams and jellies. Formulas for fudge and nougat call for amounts u p t o 5% of the batch weight and as much as 15 % is used in some caramels ; however, nougat and caramel may be made without vegetable fat. The larger proportions of fat effectively control or prevent the crystallization of sugar and impart plasticity and smoothness of texture required in the highest quality products. An extremely wide range of grades with different melting points, plasticities, and other properties suitable for various candyproduction requirements are available, and modifications meeting virtually any specifications are possible. Even a cursory review of research so extensive as that which has been applied t o the technology of fats is beyond the scope of this article, and recent comprehensive texts such as that by Bailey (1951) should be consulted for important details. Four of the principal classes of fats listed by th a t author are important in candy production. They are: (1) vegetable butters, typified by cocoa butter, which will be considered in relation t o chocolate; (2) lauric fats, such as coconut, palm, and babassu that contain 40-50% lauric acid; (3) oleic-linolenic fats represented by peanut and cottonseed, with less than 20% saturated acids; and (4) linolenic fats, of which soy is the principal example that finds use in candy. Cocoa butter has been established as the ideal candymaking fa t because of its availability from cocoa and chocolate production, and its unique plastic properties with a narrow melting range below body temperature. Such research as has been done on other fats with candy applications as its objective has dealt with their modification t o simulate

CANDY MANUFACTURE

33

the properties of cocoa butter. Hydrogenation of low-melting, lauric-type coconut f a t provides products melting in the same range as cocoa fat. Removal of the low-melting glycerides by fractionation of coconut fat a t 22’ C. (72’ F.) produces vegetable stearines suitable for candy production. Selective hydrogenation has been employed by Ziels and Schmidt (1949) to convert peanut fat into high iso-oleic forms having properties similar t o those of cocoa butter. Various types of fat may be used in caramel production, but Clay (1953) reports that the refractive index is important in affecting the translucency of the finished candy, high refractive indices giving a more desirable creamy appearance t o caramels or toffees. 3. Starch

There are two distinct ways in which starch is used in the production of candies-as a major ingredient of starch gums or jellies, and as the molding medium for the large variety of candies that are shaped by casting in starch. Different properties are required for these two applications, and manufacturers furnish special grades for each purpose. Corn starch of suitable regular or modified grades is used almost exclusively because of its availability and low cost. Results of research applicable t o jelly production and t o molding will be considered separately, and it will be evident that the unique properties of corn starch are better adapted t o candymaking than the corresponding qualities of other starches such as wheat, rice, tapioca, or cassava. a. Starch Jellies. These candies are made by forming rigid starch gels in which high concentrations of sugars happen t o be present. The sugars alter the gel-forming properties of the starch which is a major ingredient that must be present in amounts of a t least 7 t o 12% or more. This distinguishes the starch jellies from those made with pectin which, in relatively minor proportions, serves t o modify the properties of sugar and t o form gels in which the sugar is the principal structural material. The qualitative requirements of starches and processing techniques for gum candy production have been described by Kooreman (1952) who states that the ideal starch for this purpose would be one that “boils as thin as water” at the concentration necessary t o produce proper gel strength of the finished product. The nearest practical approach t o this unattainable ideal is t o use modified, thin boiling starches in the range of 40 t o 70 nominal fluidity. The viscosity of the gelatinized starch dispersion and the strength of the gel formed on cooling after concentration t o between 76 and 80 % solids are the most important properties of starches governing the quality of the candies produced in this way. Kooreman points out that it is essential to gelatinize the starch as completely as possible, and

34

L. F. MARTIN

that p H 5-6 produces maximum viscosities in cooking that will result in stronger gels and better “body” or firmness of the finished candy. Starch gums are generally molded by casting in dry starch which assists the setting of the gel by reducing t h e final moisture content t o between 15% and about 18% as a maximum. The importance of rapid setting t o a sufficiently firm body for handling in continuous production, coupled with adequate moisture retention t o maintain the desired tenderness in storage, has already been explained in section 11, 3. The fundamental chemistry of the structure of the amylose and amylopectin constituents of starch as they are now known has been summarized most concisely and clearly by Percival (1950b). Amylose is the long-chain polymer of glucose that constitutes approximately 20 % of corn starch, whereas the balance consists of the branched-chain, amylopectin polymer. Some understanding of the nature of these two fractions is essential for interpretation of the results of practical experiments on the formation of gels under conditions employed in producing starch gum candies. Interesting and useful data on starch gel formation were reported by Seck and Fisher (1940a,b), although some of their theoretical interpretations have not been substantiated. The ability t o form useful gels, defined as both “form-elastic ” and “volume-elastic,” was proportional t o the degree of hydration of the starch. The degree of swelling was proportional to the viscosity of starch dispersions, but these properties were not related t o gel formation. Their conclusion that seed starches, such as corn or wheat, are capable of forming gels whereas root starches are not is not valid, at least for modified starches. Clay (1953) reports that high-quality gum drops are produced advantageously with tapioca and other root starches modified by the chlorination process described by Fuller (1943). Seck and Fisher did find that starches modified by oxidation formed stronger gels. The deduction from this observation that the amylopectin fraction was principally responsible for the formation of rigid gels is a t variance with measurements by Hamer (1947), which show that gel strengths of various starches are roughly proportional t o their amylose contents as reported by Bates et al. (1943). Hamer’s (1947) results are particularly applicable t o the production of starch jelly candies. He developed and standardized a method for determining the breaking strength of jellies prepared under reproducible conditions from a variety of starches. Corn and wheat starches formed the strongest gels in concentrations of 10 t o 15%, the practical range for candy production. Arrowroot starch also formed strong gels a t these concentrations. Other starches formed gels of comparable strength only when used in much larger amounts. At 12% concentrations, completely defatted corn and wheat starches formed gels with two t o three times the

CANDY MANUFACTURE

35

breaking strengths of those prepared from the same starches before extraction t o remove traces of fat or adsorbed fatty acids. Hamer also determined the effect of increased cooking times u p t o a maximum of 30 min. in producing stronger gels. Similar techniques were used by Bechtel (1950) t o measure both the rigidity and breaking strength in a detailed study of modified as well a s unaltered corn starch gels. Some of his data were obtained on starches in the range of fluidities used in candy production. Increasing degrees of acid modification reduced the consistency of the cooked paste proportionately more than the strength of the finished gel. At a fluidity of 40 the consistency was only about 10% of that of pastes of unmodified starch, whereas the corresponding gel was 30% as rigid and retained more than 2001, of the breaking strength. Extensively altered starch of 60 fluidity formed increasingly firm gels as the cooking temperatures were increased from 88" C. (190.4" F.) t o the maximum of 96" C. (205" F.). Gels of unmodified starch exhibited maximum rigidity and breaking strength when formed by cooking a t 94" C. (201" F.). The behavior of an intermediate grade of 30 fluidity was less consistent, showing a slight maximum in breaking strength a t 94" C. (201" F.), but minimum rigidity at this temperature. Changes in the properties of the gels on aging were included in this investigation by Bechtel. He found that those prepared from unmodified starch increased only about 20% in both breaking strength and rigidity in the period between 4 and 24 hr. after they were formed. Slight modification by acid altered this behavior markedly. Starches of 10 t o 20 fluidity formed gels whose breaking strength and rigidity increased as much as 90% in the same period of aging. I n view of the increased gel strengths reported by Hamer (1947) for defatted corn and wheat starches, this difference may be partly the result of removal of some of the adsorbed fatty acid during the treatment of the starch with mineral acid for modification. Starch normally contains about 0.8% free fatty acid which Schoch (1942) found t o be removed only by repeated extraction with lyophilic solvents such as dioxane or methanol. Favor and Johnston (1947) reported that this impurity prevented the increase in firmness of starch gels on aging. Although Lord (1950) questioned the soundness of this conclusion, he did find that addition of 1% stearic acid t o defatted wheat starch raised the temperature required for complete gelatinization from 76" C. (169" F.) t o 80" C. (176" F.). The lowering of gelatinizing temperatures and formation of clearer and firmer gels has been reported also by Caesar (1944). Maximum gel strength and firmness are not necessarily the optimum properties required for the manufacture of starch gum candies. Rapid initial setting t o jellies that, are sufficiently firm t o be handled mechani-

36

L. F. MARTIN

cally is essential, but subsequent increase of rigidity and strength causes toughening and loss of desirable quality. Proper balance of the grade or quality of starch, of time and temperature of cooking, and of the concentration of starch in relation to the sugars used is essential for the production of tender, high-quality jellies th at will have adequate storage life. Further research is needed on the quantitative effects of high concentrations of sugars upon the measurable properties of the gels. Certain inferences may be drawn from the results of investigations of pure starch gels, but this work is of value principally as an indication of lines of experimentation to be followed in determining the properties of starch over the range of conditions and effective concentrations used in candymaking. Scientifically sound and accurate methods will have t o be developed for measuring the strengths and rigidities of the high-sugar content jellies, as only roughly reproducible, empirical data are provided by penetrometer determinations of tenderness now used for routine control of quality (Martin et al., 1952a). b. Starch Molding. Unmodified corn starch possesses the required physical properties for retaining mold impressions and, if suitably dried, of absorbing moisture from the candies cast in it. It has the additional advantages of being edible, and low in cost as it may be sifted and reused repeatedly. Many different types of candy are shaped by depositing in starch and much ingenuity has been applied to the mechanization of this operation in ((mogul,” continuous production equipment. Hard or crystallized creams as well as those in which invertase is used t o develop fluidity do not require drying of the starch as they become sufficiently solid by cooling and crystallization. The starch is dried somewhat below its normal moisture content of l O - l Z % for starch jellies and t o a greater extent for pectin jelly production. The extreme case is that of marshmallow, for which starch must be dried t o about 6 % moisture content to absorb sufficient moisture t o effect proper setting of the gelatin. Addition of very small amounts of mineral oil is widely used to improve the definition and stability of mold impressions. Material that has been reused a number of times acquires superior molding properties t o those of fresh starch, which is usually added in small amounts t o replace losses. This conditioning is attributed by Liebig (1953) t o a polishing action that makes the granules more spherical. Starch has the disadvantage of being susceptible t o contamination and it also presents an explosion hazard when dried, although only one serious starch dust explosion in a candy plant has been reported. Martin et al. (1949) experimented with various U.S.P. food grades of calcium carbonate in attempts t o overcome these disadvantages. This inert material was suitable for molding hard creams on a laboratory scale, but could

CANDY MANUFACTURE

37

not be used for jellies or marshmallow as it has no moisture absorbing properties. Steel (1949) proposed the use of mixtures of carbonate and starch that proved satisfactory for all types of cast candies, except that in proportions of 50% or more of the carbonate the mold impressions were unstable. As Hartmann et al. (1950) found that a t least 60% calcium carbonate is required in mixtures with starch t o prevent spark ignition, and as much as 90% t o prevent ignition at heated surfaces, their use does not provide a practical solution of the problem. Hartmann and Nagy (1949) listed 38 major industrial starch dust explosions during the past 50 years as evidence of the urgent need for practical means of eliminating this hazard in other uses of starch as well as in candy production.

4. Cocoa and Chocolate Candies consisting entirely of chocolate in its different flavor modifications and a wide variety of forms, including those in which nutmeats are incorporated, constitute one of the major types of confectionery. Large amounts of chocolate are used in the coatings applied t o candies of almost every other type, and substantial quantities of cocoa powder and liquor chocolate are used t o flavor fudges, creams, hard candies, and many other items. The total consumption of products of the cocoa bean can be estimated from the amounts of these materials given in Table I t o have been approximately 350,000,000 lb. in 1950. This is substantially greater than the usage of any other candy ingredients except sugar and corn sirup. The unique properties and importance of chocolate have made it the subject of specialized study, and comprehensive reference books such as those of Whymper (1921) and of Jensen (1931) should be consulted for details of its production and uses. Whole nibs of cocoa beans, after fermentation, decortication, and roasting, are ground t o produce a basic material that is neither cocoa nor chocolate. Cocoa is a product of lower fat content obtained by pressing the ground beans t o remove a large proportion of the cocoa butter. Ample quantities of cocoa butter are usually available from this operation t o increase the fat content of ground nibs that are combined with sugar t o produce chocolate. Milk solids as well as sugar are added t o milk chocolate types. Refining and adjustment of the fat content converts the ground nib material into the bitter liquor chocolate employed as flavoring and as the basis of various chocolate compositions. a. Composition of Cocoa Products. Cocoa beans vary in composition, as would be expected, with differences in varieties and the localities in which they are grown. The manner in which they are fermented and the extent of decortication and separation of shell material introduce further variations. Pressing may be carried out t o remove more or less of the fat,

38

L. F. MARTIN

and roasting conditions are adjusted to develop different flavor qualities. It is not surprising that the few fairly complete analyses of cocoa beans or products processed from them that have been reported give somewhat widely divergent results. Examples of such analyses are assembled in Table IV. The composition of beans when freshly harvested and at TABLE IV Partial Composition of Cocoa Beans and Nibs Whole Beansa Cotyledonsa Cotyledonsa Commercialb Fresh Unfermented Fermented Beans (%) (%) (dry) (dry) Moisture Fat Total nitrogen Protein nitrogen Protein Total carbohydrates (sugar, starch, cellulose, pectins, gums, etc.) Sugar (glucose) Pectins Cellulose Pentosans Gums Fiber Tannins Pigments Theobrom in e Caffeine Ash 0

b

33.0 37.0 -

-

3.65 53.05 2.28 1.50

2.13 54.68 2.16 1.34

6.0

12.4

-

1.8 2.1

-

1.2 0.3 2.2

6.23 26.69 -

-

18.34

-

-

0.30 2.25 1.92 1.27 0.38 2.09 2.24 5.30 1.71 0.08 2.63

0.10 4.11 1.90 1.21 1.84 2.13 1.99 4.16 1.42 0.07 2.74

26.32

-

4.48 -

1.15 0.16 5.49

Condensed from Knapp (1937). Anonymous (1938).

different stages of preparation (as determined by Knapp, 1937) is compared with analytical data on average commercial beans and processed chocolate (Anonymous, 1938). As the fat content is of primary importance next to flavor, the fact that the West African beans analyzed by Knapp had about double the fat content of the commercial product of unstated origin and method of preparation is particularly noteworthy. A larger percentage of shell included in the commercial sample analyzed may account for part of the difference and is suggested by the higher values for ash, non-nitrogenous residue, and for cellulose which was separately determined in this case. Adequate data are not available to establish the typical composition or

39

CANDY MANUFACTURE

the range of variations to be expected for cocoa products. With the exception of a few major constituents such as the cocoa butter, starch, theobromine, and caffeine, individual chemical compounds present have not been identified and determined separately, and no complete investigations of composition have been reported. The characteristic pigments, cocoa red or cocoa purple as they are designated by Knapp (1937), have been classified as anthocyanins by Forsyth and Rombouts (1952). Fincke (1932) extracted fat-soluble fractions which he considered more important than the tannins or volatile constituents such as diacetyl in giving chocolate its desirable flavor and aroma. Three tannin fractions were separated by Aasted (1941), who speculated as t o the role of their oxidation in flavor development. Cocoa butter is an important industrial fat in applications other than candymaking, and its composition has been determined by Lea (1929), by Hilditch and Stainsby (1936), and most recently by Meara (1949). Results of the two latest investigations are in good agreement, as shown in the following tabulation: Composition of Cocoa Fato

Molecular percentages

Fatty acids

Weight per cent (Hilditch and Stainsby)

Palmitic Stearic Oleic Linoleic

24.4 35.4 38.1 2.1

Glycerides Palmitostearin Oleodipalmitin Oleopalmitostearin Oleodistearin Palmitodiolein Stearodiolein Triolein

(Hilditch and Stainsby) (Meara) 2.5 6.5 51.9 18.4 8.4 12.0 -

2.6 3.7 57.0 22.2 7.4 5.8 1.1

" Iodine number: 36.7.

Distribution of the acids is less random than in the glycerides of most other fats. The high molecular percentages of the individual glycerides provides the chemical homogeneity that accounts for the narrow melting range and other important properties of the fat. Its behavior on crystallization governs the changes produced in physical properties of chocolate by tempering. Vaeck (1950) distinguished three fractions by thermal analysis of cocoa butter. Further investigation by this author (Vaeck, 1951a,b) produced evidence for an unstable polymorphic form obtainable by rapid cooling which was transformed into a modification melting a t 20" C. (68" F.). Storage a t this temperature for a month resulted in complete transformation t o a stable form melting at 33.7-35.7" C. (92.7-96.3" F.). There is evidence for the presence of a larger number of

40

L. F. MARTIN

individually crystallizable, component glycerides, most of which exhibit polymorphism. I n his study of the composition of cocoa fat, Meara (1949) separated a t least eleven fractions by crystallization and determined-the melting ranges of stable and unstable modifications. b. Chocolate Viscosity. The importance of viscosity in molding, and particularly in the application of coatings, has been noted in the general discussion of physical properties. Viscosity of melted chocolate depends largely upon the properties of cocoa butter, but Stanley (1941) notes that it is affected by the previous processing history of the material, by the nature and particle size distribution of other cocoa constituents or added solids, and by the presence of moisture, air, and lecithin. For effective molding or coating work, chocolate must be used a t temperatures sufficiently near t o that of solidification, a t which it is not fluid enough t o be handled efficiently unless it is modified. Lecithin is the most widely used additive for reducing and stabilizing the viscosity of chocolate. Earlier research on the problem of altering or controlling this physical property has been summarized by Whymper (1933a). Results of more recent investigations appear in patents covering proposed treatments or additives devised t o reduce the viscosity of molten chocolate. Thurman (1937) described the addition of phosphatides from soap stock a t about the same time that foreign patents (Hansa-Muhle, 1937) claimed the effectiveness of similar phosphatides as viscosity reducers. Lanolin, or similar unsaturated fats were phosphorylated and polymerized by treatment with phosphorus pentoxide t o produce materials claimed t o be effective for this purpose by both Ross and Rowe Inc. (1939), and by Jordan (1941). Cocoa butter itself, or the fatty acids derived from it, were used by Jordan (1940) in preparing synthetically phosphorylated materials for addition t o chocolate. Sulfonated as well as phosphorylated diglycerides are claimed by Harris (1939) t o produce the required reduction in viscosity. These are typical examples of the numerous patent disclosures of empirical solutions of the problem, which is evidently complex and not yet understood on a scientific basis. Both the apparatus and procedure for determining the viscosity of chocolate have been studied with the object of standardization or improvement. Modifications of the MacMichael (1915) viscosimeter were developed by Kampf and Schrenk (1929) and by Eckstein (1937) specifically for use with chocolate. These are based upon the principle of measuring the torsion produced in a standardized wire by rotation of a cylinder immersed in the chocolate. A thorough study of the conformity of observations on chocolate samples t o the theoretical equation expressing the torsion as a function of dimensions of the wire and bob, and of the size and speed of rotation of the cup containing the sample, was reported

CANDY MANUFACTURE

41

by Stanley (1941). He determined factors for converting measurements made with various wire and bob sizes t o MacMichael degrees based upon the recommended standard use of a No. 27 wire with a bob 2 cm. in diameter, immersed t o a depth of 4 cm. in the melted chocolate a t 38' C. (looo F.) in a 7-cm. cup rotating at 20 r.p.m. It is thus possible, within the limits of accuracy of the conversion factors, t o cover the extreme range from less than 50' MacMichael for ice cream coatings t o the viscosities of heavy-bodied chocolate approaching 2000' MacMichael by employing bobs of either 1-cm. or 2-cm. diameter with 3 wire sizes, Nos. 26, 27, and 30. Zenlea (1938) also recommended standardization using the No. 27 wire with conditions the same as above, but a No. 26 wire with a 3-cm. immersion of the 2-cm. diameter bob and a speed of rotation of 15 r.p.m. was preferred by a committee of the American Association of Candy Technologists (Kempf, 1949a) as representative of practice in the industry. These conditions were used by Mitchell (1951) in a statistical study of the reproducibility of the method applied by two different operators t o numerous samples of a uniform lot of chocolate. The range of determinations was k l l ' from the value of 178.5" MacMichael which was found to be the viscosity of the samples; the standard deviation was 3.61'. A method capable of giving more accurate and reproducible results is much to be desired. The validity of converting measurements with other wire and bob sizes t o MacMichael viscosities using prescribed standard conditions has been questioned by Freundlich (1937, 1939). He advocates determination of chocolate viscosity a t 35" C. (95' F.) by using heavier wires a t a speed of rotation of 35 r.p.m. Variations observed in his measurements may arise partly from the rapid heating t o 40.5-46.1' C. (105-115" F.) employed in melting the samples, as Stanley (1941) has shown the effect of previous heat treatment upon the results of viscosity determinations and stresses the importance of duplicating conditions exactly, preferably by slow heating t o the same final temperature t o melt the samples. Viscosities determined a t a lower temperature, just above the melting point of the chocolate, have the merit of being more representative of its behavior in practical use, and Freundlich (1937) applied his procedure t o obtain a measure of the body, or "covering power" of coatings. Similar information was obtained by Stanley (1941) from curves of viscosities at various rotational speeds, measured at the standard temperature of 37.8" C. (100' F.). Extrapolation of such curves t o zero speed of rotation gave values for the yield stress which, expressed a s a percentage of viscosity, provides a measure of the body of the material. This work included experiments on the effect of lecithin upon both the viscosity and body of both cocoa fat and chocolate.

42

L. F. MARTIN

The pure fat behaves like an ideal liquid and has zero body when measured in this manner. Chernenko et al. (1951) used a cone penetrometer t o measure changes in plasticity of solid chocolate produced during treatment or on storage. There was an increase with time of the plastic modulus measured in this way. Their results indicate that small amounts of lecithin reduce the plastic modulus only temporarily. Clay (1953) reports that the most promising innovation, based upon a new principle of measuring viscosity, is the application t o chocolate of the Ultraviscoson (Roth and Rich, 1953), although Stanley (1941) believes t h a t it requires further development for practical use and general adoption. c. Conching and Tempering. Development of the velvety texture and final modification of the flavor of chocolate are accomplished by heating a t carefully regulated temperatures for periods of 36 t o 72 hr. or longer in machines called "conches." The name refers t o the shell-like shape of the container in which the chocolate is kept in motion during the process. Conching temperatures may be anywhere between 110 and 210" F. (43.3 and 98.8' C.) depending upon the initial quality of the chocolate and the intensity of flavor modification desired. No studies have been reported of the chemical changes that occur during this operation. When it is completed t o the satisfaction of the operator, the untempered product may be molded for storage or run directly t o tempering kettles or machines for the final step of processing prior t o its use in molding or coating finished candies. The purpose of tempering is t o develop desirable physical properties in the chocolate by promoting crystallization of the fat in very small crystals of the most stable modifications. Proper tempering prevents the development of internal stresses t h a t would result from unnecessary crystallization after cooling and thus definitely reduces the tendency t o bloom. The crystalline forms of cocoa fat obtained by effective processing in this step also possess the important characteristic of providing for sufficient contraction on setting t o release the candies from molds. Whether carried out by hand in ordinary kettles or in automatic tempering machines, the steps of the process are the same. The chocolate is heated under constant agitation t o 46-49' C. (115-120' F.) t o melt all crystals of fat. The temperature is then reduced gradually t o 29-30' C. (84-86" F.) t o recrystallize the component having the highest melting point together with some of the next highest melting glyceride fraction. As crystallization proceeds the material becomes a thick paste of extremely small fat crystals nearing solidification. When seeded in this manner with stable crystalline forms of the fat, chocolate retains its temper while reheated for use a t 31.7-32.2" C. (89-90" F.). Alternatively, the melted chocolate may be cooled only t o this temperature range at

CANDY MANUFACTURE

43

which i t is used, seeded with 10% of well-tempered chocolate, and stirred until the entire batch is tempered.l The tempering process has been described in detail by Kempf (1949a,b), and discussed most recently by Koch (1952). T h a t the postulated explanations of the results obtained, based upon various interpretations of the underlying phenomena, are controversial is evident in the criticism of the latter author’s deductions by Whymper (1952a). There is general agreement upon the primary importance of proper crystallization of the cocoa butter, which has long been recognized as the major purpose of tempering. No such agreement has been reached upon the utility of methods proposed for determination of the effectiveness of the process, or the degree of temper obtained under different conditions. Cooling curves determined b y plotting temperatures at various times of cooling under standard conditions were used b y Pichard (1923) as a means of detecting adulteration of cocoa butter. This method was applied t o chocolate by Jensen (1931) t o determine its temper. Pichard (1932) also extended his method t o chocolate products as well as cocoa butter. Its application t o control of the tempering process was later developed more completely by Pichard (1937). All of the earlier attempts t o adapt such cooling curves t o the practical control of the process have been summarized by Whymper (195213). Additional modifications have been proposed recently by Easton et al. (1951), and by Meyers and Graham (1952). Criticism of the latter workers’ procedure b y Whymper (195%) emphasizes the need for a new approach t o the development of a practical method for measuring and controlling temper. Utility of the test procedure advocated by Easton and his co-workers has been questioned by Clay (1953). The possibility of using ultrasonic vibrations t o determine the temper of chocolate has been suggested by Mickevicz (1951). I n its present state of development, the method based upon this principle may be suitable for measurernent of viscosity (Roth and Rich, 1953), but this property has no simple or direct relation t o the state of crystallization and temper of chocolate. Better tempering was claimed t o result as an incidental benefit of the method of conching in stages patented by Aasted (1939). The primary purpose of this conching procedure is t o enhance the development of flavor, and its effect upon subsequent crystallizat,ion of fat during tempering must be secondary. d . High Melting Modifications of Chocolate. The natural melting range of cocoa butter below body temperature provides one of the most desirable qualities of chocolate, but it makes storage and distribution difficult during summer months. High-melting-chocolate is essential for the proIndebtedness to Mr. Clifford Clay for these details of conching and tempering is gratefully acknowledged.

44

L. F. MARTIN

duction of candy items for use in military rations. Numerous methods have been proposed to overcome the impairment of quality that results from substitution of the cocoa butter by hardened fats or t o raise the melting point of cocoa fat in conformity with identity standards. Penn (1941) hydrogenated chocolate liquor as well as added fat with Raney nickel, increasing the melting point of the fat component naturally present. A process similar t o the winterizing of vegetable oils was applied by Carver (1943) to cocoa butter to obtain the higher melting glyceride fractions for use in making heat-stable chocolate. The same end was sought, in addition t o stability against bloom formation, in the patents of Cook and Light (1940) and of Eipper (1948) on elaidizing the oleic acid present in cocoa butter. Increasing or modifying the sugar content has been described as another method of formulating high-melting chocolate candies. Bars made with dextrose, lactose, or other sugars with similar properties are claimed by Sarotti (1943) t o be more resistant t o melting t.han those of chocolate made ent,irely with sucrose. A higher sucrose content can be employed if the sugar is dispersed in the chocolate in the form of extremely fine crystals like those in fondant, according t o McGee (1948). As extra-fine grades of sugar are generally employed in chocolate manufacture and are very thoroughly dispersed during refining and tempering, the result of this proposed modification is primarily a reduction in the proportion of lom-melting cocoa butter in the product. Kempf and Hoben (1949) give formulas for milk chocolate compositions th a t are sufficiently stable t o heat for summer distribution. These contain, in approximately 100 lb.: chocolate liquor of 52% f a t content, 14.75 lb.; sugar, 49.75 lb.; dry skim milk, 12.25 lb.; cocoa butter, 23.25 lb.; water, 2.5 lb., and vanilla flavoring. Emulsifiers of the type th at have been patented as preventives of f a t bloom (Eipper, 1938; Mayberry, 1951; Cross, 1953) have been used also for improvement of the palatability of chocolate made with high melting, noncocoa fats to withstand the extreme temperatures t o which ration candies must be subjected. Although the fats used melt above body temperature, the emulsifiers are reported t o minimize the unpleasant, waxy texture sensation which makes such fats objectionable. Alikonis and Farrell (1951) recommend the use of 1%of a mixture of equal parts of sorbitan monostearate and polyoxyethylene sorbitan monostearate for this purpose in chocolate bars and coatings to meet ration specifications. High lauric acid fats are used, and coconut oil hydrogenated to a melting point of 43.3-45.6' C. (110-114° F.) is preferred as the replacement for cocoa butter. The coatings are of the cocoa type, containing only 7.5% to a maximum of about 18% of cocoa powder, with coconut fat replacing

CANDY MANUFACTURE

45

92-94 % of the cocoa butter content of regular chocolate coating. Approximately S-9% of cocoa butter is retained in the chocolate bar for which these authors give the following composition: sugar, 50%; chocolate liquor, 17%; whole milk solids, 16%; hydrogenated coconut fat (110114" F.), 16%; and 1% of the mixture of emulsifiers. These modified and stabilized formulations are permissible for chocolate required for manufacturing military ration candies, but they do not conform t o the standards of identity for chocolate or regulations prohibiting the use of non-nutritive substances in confectionery for the civilian market. 5. Nutmeats and Fruits

The principal method of using nutmeats and fruits in candies is by incorporating forms in the finished batches just prior t o molding or forming the desired pieces. Diced or whole fruits may be added as a step in molding items such as cordial fruit candies or cordial maraschino cherries. The cream centers of this type are initially firm enough t o be chocolate coated mechanically or by hand, but contain sufficient invertase t o produce the required amount of sugar sirup by inversion after they have been coated. Peanut butter, almond paste or marzipan, jams, and similar materials that may be cooked into the candy or dispersed throughout it are used in smaller quantities. Reference t o Table I shows that peanuts are most important, comprising more than half the volume and about 40% of the total value of all ingredients in this group utilized in 1950. Coconut and almonds are the other major items, accounting for most of the remainder of the total consumption of fruits and nutmeats combined. Comparison of the figures for 1950 with those for 1947 indicates the extent t o which utilization of individual products may be altered by fluctuating prices or availability. Very little research has been conducted specifically upon candymaking uses of these products, and details of their production and processing for general food use would be superfluous t o this review. Useful summaries of existing information on the composition and chemistry of each of the fruits and nuts used in candy production will be found in the comprehensive treatise by Jacobs (1951). Reference will be made only t o particularly suggestive or pertinent investigations of problems of processing and storage of the individual ingredients of this class. a. Peanuts. Different types of peanuts have been found t o possess significantly different storage qualities. Development of rancidity was more rapid in Spanish peanuts than in Virginia or Runner types stored under the same conditions in the extensive study carried out by Pickett and Holley (1951). Organoleptic methods were used by these workers t o determine the rates of deterioration. The same relative stabilities are

46

L. F. MARTIN

exhibited by the oils extracted from the peanuts. In a study of oils extracted under comparable conditions from 16 different varieties, Fore et al. (1953) found that those from the Spanish types were less stable to oxidation determined by chemical tests. Flavor can be preserved for much longer storage times by refrigeration according to Woodroof et al. (1949), who recommend the use of activated carbon t o minimize absorption of odors. Cecil and Woodroof (1951a) tested a variety of antioxidants that were effective in protecting flavors of peanuts and other nutmeats in storage. Owen (1950) claimed that 1-2% of salt, containing 5 g. per lb. of ascorbic acid, was effective in stabilizing the flavor of peanuts to which it was applied. An antioxidant salt preparation containing propyl gallate, citric or other acids, and polyhydric alcohols, was patented by Hall and Sair (1950) for this and similar food applications. Reznikova (1941) states that antioxidants, not specified in the available report of her work, were ineffective, but that peanut flavor could be stabilized by alcohol, sugar sirup, or oatmeal. Higher alkyl or alkylene esters of polyhydric alcohols were applied to various nutmeats including peanuts t o prevent separation of oil by Neal et al. (1949). Sugar-amine browning reactions were the most important changes observed by Pickett and Holley (1951) in their extensive study of roasting. The recorded information on the composition of peanuts has been reviewed critically and summarized in a comprehensive bibliography by Guthrie et al. (1949). The earliest study of their composition by Payen and Henry (1825) noted similarities to that of almonds. Total unsaturated acids, oleic and linoleic, were found in significantly lower percentages in Spanish than in Virginia-type peanut oils by Jamieson et al. (1921), an observation at variance with the later determined greater susceptibility of the Spanish peanuts t o development of rancidity. Mean results of more recent analyses of a large number of samples by Stansbury et al. (1944) showed the oil of Spanish type peanuts to have a somewhat higher iodine number than the Runner or Virginia. Effectiveness of low-temperature storage was demonstrated by the preservation of samples for analysis for 2 years in sealed cans a t 1' C. (33.8' F.) without appreciable changes in total nitrogen, oil content, free fatty acids, or iodine number of the oils, as reported by Stansbury and Guthrie (1947). Changes in the extracted oil detectable by spectrophotometric methods were found by Pons et al. (1948) t o be much less at 1' C. (33.8' F.) than at 27" C. (80.6" F.) after 4 years storage. b. Almonds. The composition of almonds has received relatively little attention and no studies of changes in storage have been reported. The very early analyses by Payen and Henry (1825) have been supplemented by Pavlenko's (1940) determinations of the 15 to 18-fold increase in

CANDY MANUFACTURE

4i

benzaldehyde content and other changes on ripening, and by comparison of the compositions of domestic and imported almonds initiated by Hart (1930) and completed by Pitman (1930). These investigators developed a shearing test for the objective measurement of texture. They also obtained some data on the composition of different varieties in relation t o seasonal changes and locality of production. The average oil content of whole kernels was 51-54%, a maximum of 59% being found in almonds of the Drake variety grown in the Sacramento Valley. Drake almonds gave lower shear test values than varieties of lower oil content. The oil is highly unsaturated according t o Heiduschka and Weisemann (1930) who reported the fatty acid composition t o be oleic, 77.0%, linoleic, 19.9%, and palmitic acid only 3.1%. Almonds are known t o be rich in a variety of enzymes and arc a source of emulsin which hydrolyzes glycosides. c. Pecans. Changes affecting the quality of pecans after they have been incorporated in candies have not been investigated, but their handling and storage has been studied extensively in order t o provide nutmeats of the highest quality for confectionery and other uses. The work of Brison (1937, 1945), extending over a 10-year period, established the importance of prompt refrigeration and storage a t low temperatures. Pecans stored a t - 15" C. (5" F.) did not become rancid in 2 years, whereas those held just below 0" C. (32" F.) developed rancidity in 11 months. Brison also determined certain chemical changes that occur during storage, but found that neither the phloroglucinol test nor the free fatty acid content parallels the development of rancidity determined by organoleptic methods. His results show that storage of pecans a t 1.7" C. (35" F.) and 90-92% relative humidity as recommended by Baker (1938) is not adequate for long holding times. Coating with a 40% sugar sirup was found by Godkin et al. (1951) t o improve the keeping quality of pecans. Cecil and Woodroof (1951b) showed that pecans are stabilized by antioxidants applied by addition t o the roasting oil or, in the case of salted nuts, by admixture with the salt. A combination of butylated hydroxyanisole, propyl gallate, and citric acid in total concentrations of 0.02% of the oil or 0.2% of the salt used, increased stability in the Shaal oven test 133 %. Interesting experiments were conducted by McGlamery and Hood (1951) on heat treatments intended t o inactivate the enzymes responsible for changes t h a t produce rancid flavors. Either hot air circulation, or immersion in oil after rapid air heating, was used t o bring the internal temperature of the nutmeats t o 80" C. (176" F.) for periods of 1 t o 12 min., followed by rapid cooling. Subsequent storage and organoleptic tests of treated and untreated pecans indicated beneficial results of this treatment applied a t any time u p t o 4 months after harvest. d . Coconut. Despite its importance as a flavor-imparting ingredient

48

1,.

F. MARTIN

and the considerable quantities of coconut used in cream or in shredded forms in candymaking, no research on this application of the material has been reported in the scientific literature. The oil, which has been studied thoroughly, has applications in candy previously discussed, but only the carbohydrates of dried copra have been investigated by Caray (1924). Glycerol is used t o improve moisture retention of much of the dried coconut used in candymaking. Recent patents describe the use of purified, carbonyl-free propylene glycol and butylene glycol as humectants (Kaufmann and deMaya, 1952), and the addition of sorbitol for the same purpose (Welker et al., 1953). A rapid method of analyzing copra, applicable t o determining the content and quality of oil in coconut, has been developed by Pinto and Enas (1949). e. Other Nutmeats. Relatively small quantities of brazil nuts, cashews, filberts, walnuts, and hazelnuts are utilized in candy manufacture. If these present any special problems in storage or preparation for use as candy ingredients, or in candy formulation, no published research has dealt with the subject. Available data on their composition, relating principally t o the oils and proteins, have been assembled by Jacobs (1951). f. Fruits. The not inconsiderable volume of fruits, jams, and fruit products used by the confectionery industry is made u p of a large variety of ingredients too numerous to be considered individually. A summary of published information on the composition, storage, processing, and preservation of quality of these products will be found in the treatise by Jacobs (1951). Fundamental research on enzymatic browning, one of the most important factors in deterioration of stored or processed fruit quality, has been reviewed b y Joslyn (1951). Quantities of dried fruits are used in candy production, and whole fruits are often subjected t o heat and desiccation by incorporation in candies. For this reason, the changes produced by chemical, or nonenzymatic browning are important ;progress in research on this problem has been summarized by Stadtman (1948). Kirchner (1949) assembled and reviewed available information on the composition and chemical properties of fruit flavors. Concentrated natural fruit essences recovered by a process developed by Milleville (1948, 1950) and described by Milleville and Eskew (1944) have been used in pectin jellies by Martin et al. (1946). Conditions for making high quality glace fruits were studied by Tressler (1942), and Pentzer et al. (1942) patented an apparatus and method for their production.

VI. PRODUCTION METHODS More attention has been given t o mechanization than t o the fundamental processes of candy production. The voluminous patent literature on the subject is beyond the scope of this review, and deals almost exclu-

CANDY MANUFACTURE

49

sively with mechanical improvements in carrying out the same basic operations that have been employed traditionally. Even the so-called "continuous" vacuum cookers widely used for mass production of hard candy are actually high speed, intermittent batch cookers. This is evident in the description of one of the most modern installations for hard candy production by Ziemba (1950). Very recent innovations have been described in the development of truly continuous processes and equipment for cooking various types of candy. These provide the advantage of rapid, more uniform, and precisely controlled heating, in addition to the saving of time and labor. Most of these continuous flow methods, employing machines such as the "Votator " and "Turba-film" evaporators, are in the experimental stage, although commercial installations have been reported in a few cases. The shorter heating times made possible in such equipment alter the extent of reactions that are important in producing desired qualities. Limited reaction of the ingredients is advantageous in producing hard candy or marshmallow, but is ineffective in developing flavor of caramel which requires longer cooking times. A better understanding of the chemical changes occurring in any particular cooking process, and their dependence upon temperature and time relationships, will be essential in working out the most effective applications of these novel methods of candy manufacture. I n truly continuous cooking only a small volume of the material is heated for a short time t o the maximum temperature required t o convert it t o the final candy composition. Dissolving of ingredients and precooking in batch kettles are usually necessary, but the most critical stages of concentration and reaction are effected within seconds in a continuous stream of material flowing t o the forming or molding operations. An example is the process and apparatus recently patented by Leach (1953) to produce hard candy. A reservoir of sirup, precooked t o approximately 132" C. (270" F.), feeds the charge through a heating tube of small volume in which the temperature of the sirup is raised rapidly t o 149" C. (300" F.) or higher. A specially designed valve ejects quantities of the cooked material into die-molds on a continuously moving chain. I n the starch jelly process described by Bolanowski et al. (1952), the prepared batch is adjusted t o the moisture content desired in the finished candy t o be molded in starch, and preheated t o 76.7-82.2' C. (170-180" F.). The starch gel is formed by rapid heating t o 140" C. (285" F.) in a Votator unit. A second unit cools the gel promptly t o the temperature suitable for continuous depositing in starch molds. Among advantages claimed for this process are uniform cooking and elimination of the long holding period in hot rooms t o dry the candies t o the proper moisture content. Premixed ingredients for caramel are rapidly cooked and concentrated

50

L. F. MARTIN

by evaporation of water in a process employing the Turba-film evaporator of Swiss design (Anonymous, 1952). Mechanical rotary blades produce the turbulence and maintain the film conditions for rapid heating of the viscous material. I n this case, additional cooking is required to develop satisfactory caramel flavor. Alikonis (1953) has described an ingenious method of cooking and aerating marshmallow continuously by a n air jet th a t directs the cooked fluid against a target, after which it passes through a porous alumina dispersing sleeve. The same author (Alikonis, 1949a,b) has patented a one-step process with continuous features for manufacturing chocolate liquor, as described in more detail by Slater (1952). I n a different line, Visenjou (1952) has developed a n inclined tube apparatus t o roast nuts in oil continuously and has applied it t o the processing of cashew nuts. The radically different conditions of processing necessitated or made possible by continuous operation of the cooking steps in candy manufacture, the possibilities of which are just beginning t o be explored, may be the next major development in the evolution of the industry. Wide adoption of such methods for making all types of candies would follow the trend in modern food processing toward more uniform conditions and results, automatically and precisely controlled without dependence upon human judgment and with further reduction of labor costs. Next t o the investigation and understanding of the chemical or physical changes in combining various ingredients t o produce candies, means of accurately controlling the conditions upon which such changes depend are of greatest importance in manufacturing candies of high and uniform quality. These two lines of development should go hand in hand as the development and perfect:on of new processing methods will depend upon the progress of research in discovering the fundamental nature of the processes involved and establishing the conditions for obtaining optimum results.

VII. SUMMARY If this review has outlined clearly the wide diversity and complexity of the fundamental problems of candy manufacture and the almost limitless opportunities for further research in this field, it will have served its primary purpose. Scientific principles have been applied too seldom t o investigation of particular candymaking processes. With few exceptions, past attempts t o apply such principles have been based upon oversimplifications of the problems involved. The real complexity of some of these problems has become apparent only recently as a result of progress in research on sugaxs, starches, proteins, fats, and other ingredients. Theoretical as well as experimental methods had to be developed before many of the difficulties inherent in studying the chemical and

CANDY MANUFACTURE

51

physical processes of candy production could be overcome. Efforts of earlier workers were undoubtedly limited by the inadequacy of available methods. There was reason t o believe t h a t the only sure guide t o progress was the accumulated experience and “know-how ” of the candymaker. Fortunately, these restrictions no longer apply, as the detailed knowledge of chemical structures, reactions, and properties of the natural products used t o make candies has expanded rapidly in recent years. Precise, effective methods are being devised or improved continually and applied t o increase this store of knowledge. The urgent need is to extend fundamental investigations with the tools of modern science into the specific areas of candymaking conditions and requirements. We have noted the extent t o which this has been done, but have cited many more examples of‘ obvious applications of newer research methods or recently acquired knowledge that remain t o be explored. It is fair t o state that the candy industry has made continual, significant technological progress, but it is also accurate t o observe that this progress has been one-sided. Knowledge of the products themselves and of the basic processes of manufacture has not kept pace with developments in mechanization, production engineering, and merchandising. The undesirable consequences of such unbalanced progress are clear t o leaders of the industry, and Adelson (1953) has stated that: (‘In the final analysis, we have got t o have good candy in order t o sell it. . . . I n the past we may have given too little attention t o the candymaking end of our industry.” Attention t o the candymaking end calls for a greatly expanded, well-coordinated research program t o be carried on within and by the industry to bring it fully abreast of other food-processing enterprises. Such research can make rapid strides with the scientific knowledge and means now available. Any programs of research, whether carried on by individual firms or supported cooperatively by the entire industry, must produce tangible improvements t h a t can he translated into increased profits. This requires concentration on major problems with emphasis on practical solutions, but i t also calls for sound, fundamental investigation of the basic principles of candy production. Future research should include proportionately more inquiries of a fundamental nature than has been the case in past work summarized in this review. Approximately, but simply defined, practical or applied research deals with “how” processes operate, whereas fundamental research seeks t o discover ( ( w h y ” they operate as they do: the results of applied research usually are limited t o the solution of one particular problem, whereas those of fundamental research invariably are applicable t o the solution of many problems. T o illustrate this with reference t o candy, merely determining how variations of the proportions

52

L. F. MARTIN

of sugars and of times and temperatures of cooking affect the tendency of hard candy t o become sticky would be of less practical value than discovering why such candies are hygroscopic a t all. Intelligent research in this field, including an effective balance of fundamental and applied studies, will not attempt t o supplant the a r t or experience of the candymaker that has brought many types of candy t o a high state of perfection. Nevertheless, the possibility of further improvement can be ascertained only by acquiring a better understanding of the chemical and physical changes involved in production methods now in use. Opportunities for immediate and almost certainly profitable investigation are afforded by a variety of problems related t o storage quality in order t o lengthen the shelf-life of candies of every type. Section 11, 3 was devoted t o a summary of efforts t o correct deficiencies in storage quality that have been the cause of substantial losses throughout the industry. Winger (1952) has estimated allowances for returned goods ranging from 0.45% t o as much as 0.81% of total sales during recent years. These figures were based upon reports of some 170 firms representative of every line of candy manufacture, and represent total losses b y all manufacturers amounting t o between $4,500,000 and $8,000,000 annually when applied t o the total wholesale value of approximately $1,000,000,000 of candy production. This is only the direct, tangible cost of goods t h a t become unsaleable. Less complete deterioration impairs the quality of large amounts of candy t o a n extent sufficient t o cause even greater loss from reduced sales volume, although this indirect loss cannot be estimated accurately. Improvement in texture stability has been most significant, and some progress has been made in retarding the oxidative deterioration of fat in butter creams and similar candies. Further improvement of the keeping qualities of marshmallows, jellies, and gums can be brought about by attention t o the fundamental properties of the gelatin, pectin, and starch used in their production. Each of these modifiers of the sugars behaves differently in forming gels that will retain desirable properties. Extensive practical experimentation has made negligible progress in solving the problems of chocolate bloom, or the hygroscopicity of hard candy. Modernization of candy production has been accomplished largely by mechanizing materials handling, and operations such as molding, forming, depositing, coating, and packaging. Revolutionary advances in production techniques are promised by the high-speed, strictly continuous cooking methods that are being introduced on a pilot plant or limited commercial scale. Such methods require far more precise control than the slower cooking in batch kettles under the constant surveillance of experienced candymakers. The conditions that have been found empirically

CANDY MANUFACTURE

53

t o give the best results in batch cooking provide a very imperfect guide for developing optimum conditiocs for the new procedure. It is essential t o obtain more precise, fundamental data on the chemical and physical chemical changes involved in order t o take full advantage of the efficiency and uniformity of operation of the new processes. The chemistry of the sugars, in the range of conditions used in candymaking, is of foremost importance. It is evident from the results of research described in section I11 t h a t candies may no longer be considered simply as solid solutions or dispersions of sugars that have undergone no changes more profound than inversion of sucrose. There are no recent reports of investigations of the behavior of sugars in candy production, and even the obvious applications of the latest discoveries in this field t o candy problems have been neglected. Major lines of future investigation should deal with the role of sugar anhydrides or reversion products and caramelization reactions in governing the properties of many import a n t types of candy. The effect of heat upon sugars in the highly concentrated solutions formed in candy cooking has not been studied in a manner comparable t o the investigations of chemical changes produced by heating the dry sugars or their dilute solutions. Results of the extensive work of van Hook and Bruno (1949) on the kinetics of sucrose crystallization in sugar manufacturing may be applied advantageously in developing more effective control of its crystallization in candies. The different modifications of sugar properties produced by albumin or other protein whipping agents, by gelatin, and by pectin will be understood better in the light of fundamental principles developed by the research described in section IV. It has been possible here t o include brief descriptions of only the most significant or pertinent work on each of these ingredients. Careful study of the background and details of the investigations cited will suggest many more extensions of this experimental work to candy applications than i t has been possible t o note in review. The estimates in Table I show that the quantity of milk products used in confectionery is exceeded only by those of cocoa and chocolate products and of the sugars. The reactions which the constituents of milk undergo in various candymaking processes are more complex than those of other ingredients, and no results of experimental work on this subject have been published. Changes produced by heating, drying, or storing the milk products themselves are beginning t o be understood only recently through the latest and most fundamental research on this important food. The ultimate objective of this work is t o devise processing methods or conditions that will minimize or prevent alteration of natural qualities. A uniquely interesting subject for study by the same fundamental methods would be the production of caramels in which reactions of the

54

L. F. MARTIN

constituents under controlled conditions are necessary t o develop the desired flavors. Results of such a study would not only provide a basis for improving caramel-cooking procedures, but would serve t o complement knowledge gained from research on stabilizing unaltered milk flavor for application in making other types of milk candies. Important properties for candymaking applications have been determined in considerable detail for fats and starches. Significant recent progress has been made in stabilizing animal fats, particularly butter, against oxidative rancidity, but the problem of hydrolysis of both vegetable and animal fats in candies having a high-moisture content remains for future investigation. Recent experimental work on the nature of starch gels and factors affecting their properties is especially useful in providing practical guides for starch gum candy production. Quantitative methods of measurement make it possible t o determine how gel strength and rigidity are affected by variations of the fluidity of the starch and the conditions under which it is cooked in making jellies. These methods should be applied t o more thorough investigation of starch gels formed in the presence of high concentrations of sugars. Changes that occur on aging are particularly important for maintaining quality of the candies during storage. Cocoa and chocolate have been studied more extensively for their specific applications in candymaking than have any of the other ingredients of candies. Knowledge of the constitution and properties of cocoa butter is more complete than that of the nonfat fractions of these materials, but their compositions have been determined as accurately as the older methods of analysis permitted. Major problems are still unsolved, notably the causes and prevention of chocolate bloom. Little is known of the chemical changes responsible for flavor development in conching. Useful contributions may be expected from research employing the latest precise, effective methods t o investigate the composition of chocolate and the phenomena of both conching and tempering. Cocoa products have not been analyzed by chromatographic or similar modern techniques for separation and identification of the individual chemical constituents of complex natural substances. A systematic study of the composition of cocoa nibs and of the materials obtained in successive stages of processing into various types or grades of chocolate mill be essential t o provide data for a rational, scientific approach t o the development of better processing methods and improved products. The background and methods required for such a study are now available. Measured in terms of expenditures, the confectionery industry has applied a smaller percentage of its gross income of almost $1,000,000,000 per year t o research than any industry of comparable size. Its progress

CANDY MANUFACTURE

55

has been achieved t o a great extent by the application of research performed by the industries supplying its principal raw materials, including chocolate, t o the improvement (for particular uses) in candy production. There is evidence that this complacency is yielding t o an awareness of the possibilities of research within the industry on specific candy production problems. The variety of these problems, and the complexity of some of those suggested by this review, afford manifold opportunities for fundanierital investigation by the most modern scientific techniques. TO the extent that the industry avails itself of these opportunities, it will continue t o progress and keep pace with the development of scientific food technology. ACKNOWLEDGMENTS This outline of problems in candy manufacture and the present status of research applicable to their solution would not have been possible without the continual advice and encouragement of many friends and associates in the candy industry. It is a sincere pleasure to acknowledge indebtedness to Mr. Philip P. Gott, President, and members of the Research Committee, progressive leaders of the confectionery industry, for an understanding of these problems acquired during a decade of cooperation in research on a modest scale between the National Confectioners’ Association and the Department of Agriculture. I am particularly grateful to Mr. Clifford Clay, of Stephen F. Whitman & Son, for his invaluable review of most of the manuscript and generous information on many topics besides his authoritative treatment of chocolate tempering and problems in chocolate candy production.

REFERENCES Aasted, I<. C. S. 1939. Conching of chocolate and similar operations. U. S. Patent 2,147,184. (Chem. Abstr. 33, 4340) Aasted, K. C. S. 1941. Investigation of the working of chocolate in the conche. Kemzsl, 22, 173. (Chem. Abstr. 37, 4816) Adelson, C. 1953. Natl. Confectioners’ Assoc. Bull. 38(2), 2. Alikonis, J. J. 1947. Apparatus for making confectionery. U. S. Patent 2,424,950. Alikonis, J. J. 1949a. Method for making confectioneries. U. S. Patent 2,459,908. Alikonis, J. J. 1949b. Method of making chocolate confectioneries. U. S. Patent 2,465,828. Alikonis, J. J. 1951. Method and apparatus for making confectionery. U. S. Patent 2,536,310. Alikonis, J. J. 1952a. Private communication. Alikonis, J. J. 1952b. Sorbitol in confections. Mfg. Confectioner 32(12), 27. Alikonis, J. J. 1053. Diffusion problem whipped. Food Eng. 26, 56. Alikonis, J. J., and Farrell, K. T. 1951. Improvement of cocoa-type coating for use in army rations. Food Technol. 5, 288. Ambler, J. A. 1927. The candy test for sugars. SIfg. Confectioner 7(1), 17. Angermeier, H. F. 1953. Demethoxylated pectins. Proc. 8th Prod. Conf., Penn. Mfg. Confectioners Assoc. M f g . Confectioner 33(6), 25. Anonymous. 1937. Uses of glycerin in confectionery. Chemist and Druggist 127, 601. Anonymous. 1938. Cacao as a winter food. Znd. dolcaana 3, 182. (Chert. Abstr. 33, 3477)

56

t. F. MARTIN

Anonymous. 1952. A continuous caramel production line. Mfg. Confectioner 3 2 ( l l ) , 20. Ayres, J. C., and Stewart, G. F. 1947.Removal of sugar from egg white by yeast before drying. Food Technol. 1, 519. Bailey, A. E 1951. “Industrial Oil and F a t Products” (See particularly Chapters V and VI, pp. 87-IN), Interscience, New York. Bailey, M. I. 1935. Foaming of egg white. Znd. Eng. Chem. 27, 973. Baker, C. T. 1938. Cold storage of pecans. Refrig. Eng. 36, 28. Baker, G. L. 1948. High polymer pectins and their deesterification. Advances i n Food Research 1, 395. Baker, G. L.,and Goodwin, M. W. 1941. Pectin jellying composition. U. S. Patent 2,233,574. Baker, G. L., and Goodwin, M. W. 1944.Effect of methyl ester contents of pectinates upon gel characteristics at different concentrations of sugar. Delaware State Agr. Expt. Sta. Bull. KO.246. Baker, G. L., and Woodmansee, C. W. 1941. Grading pectins. Food Technol. 3, 23. Bates, F. L., French, D., and Rundle, R. E. 1943. Amylose and amylopectin content of starches determined by their iodine complex formation. J . Am. Chem. SOC.66, 142. Bechtel, W. G. 1950. Measurements of properties of corn starch gels. J . Colloid Sci. 6, 260. Bingham, E.C. 1930. Fundamental definitions of rheology. J . Rheol. 1, 509. Bloom, 0. T. 1925.Penetrometer for testing jelly strength of glues, gelatins, and the like. U. S. Patent 1,540,979.(Chem. Abstr. 19, 2280) Bolanowski, J., White, T. A., and Ciccone, V. R. 1952. Pilot plant experiments on starch jellies. Candy I n d . 17(7), 15. Brison, F. R. 1937. The storage of pecans. Proc. Texas Pecan Growers’ Assoc. Ann. Meeting 17, 21. Brison, F. R. 1945. The storage of shelled pecans. Texas State Agr. Expt. Sta. Bull. No. 667. Bronson, W. F. 1951. Technology and utilization of gelatin. Food Technol. 6, 55. Buchanan, B. F. 1945. Dispersion of gums. U. S. Patent 2,376,656. (Chem. Abstr. 39, 3375) Butler, H. G. 1942. Soybean produces whipping agent. M f g . Confectioner 22(4), 12. Caesar, G. V. 1944.The hydrogen bond in starch as a basis for interpreting its behavior and reactivity. I n “Chemistry and Industry of Starch” (R. W. Kerr, ed.), pp. 178-182. Academic Press, New York. Campbell, L. E. 1938. The calibration of jelly testers. J . SOC.C h e w Ind. 67,413. Campbell, L. E. 1940. Pseudo-plastic properties of molten chocolate. J . SOC. Chem. Znd. 69, 71. Caray, E. M. 1924. Isolation and identification of some of the sugars in copra meal and coconut water. Philippine J . Agr. 13, 229. (Chem. Abstr. 19, 547) Carlin, A. F., and Ayres, J. C. 1951. Storage studies on yeast-fermented dried egg white. Food Technol. 6, 172. Carver, F. S. 1943. Chocolate. U. S. Patent 2,336,346.(Chem. Abstr. 38, 3037) Cecil, S. R., and Woodroof, J. G. 1951a. Butylated hydroxyanisole as an antioxidant for salted peanuts, salted pecans, and peanut butter. Ga. State Expt. Sta. Bull. No. 265. Cecil, S.R., and Woodroof, J. G. 1951b. BHA ups shelf-life of salted nuts. Food I n d . 23(2), 81. Cherenko, L. E.,Markovich, V. E., and Votkina, L. S. 1951. Structure formation

CANDY MANUFACTURE

57

during treatment of chocolate mass and the effect of liquefiers. Kolloid Zhur. 13, 379. (Chein. Abstr. 46, 2202) Clark, G. L., and Ross, S. 1940. Measurement of static and dynamic foams in characteristic units. Znd. Eng. Chem. 32, 1594. Clay, C. 1953. Stephen F. Whitman’s Sons Co., Philadelphia, Pa. Private communications. Clayton, W., Back, S., Morse, J. F., and Johnson, R. I. 1953. Chocolate. Brit. Patent 469,112. (Chem. Abstr. 32, 674). Collins, W. R. 1940. Gelatin desserts. U. S. Patent 2,196,146. Conn, R. C., and Asnis, R. E. 1937. Oat flour as an antioxidant. Ind. Eng. Chem. 29,951. Cook, L. R., and Light, J. H. 1940. Retarding fat bloom a t the surface of chocolate and chocolate-coated products. U. S. Patent 2,216,660. (Chem. Abstr. 36, 820) Coder, H. B. 1951. Better quartermaster rations. Candy Znd. 16(9), 188. Coulter, S. T., Jenness, R., and Geddes, W. F. 1951. Physical and chemical aspects of the production, storage, and utility of dry milk products. Sect. IV. Chemical changes in dry milk products during storage. Advances in Food Research 3, 93. Cross, S. T. 1952. Bloom inhibited chocolate. U. S. Patent 2,586,615. (Chem. Abstr. 47, 2905) Cross, S. T. 1953. Bloom inhibited chocolate. U. S. Patent 2,626,216. (Chem. Abstr. 47, 3492) Cruess, W. V. 1946. Jellied fruit candies. Fruit Products J . 26, 166. Cruess, W. V., Frian, C., Jang, R., Lawrence, D., Miller, G., and Cytron, B. 1949. Fruit Products J . 29, 15. Danehy, J. P., and Pigman, W. W. 1951. Reactions between sugars and nitrogenous compounds and their relationship to certain food problems. Advances in Food Research 3, 241. Dimler, R. J., Schaefer, W.C., R7ise, C. S., and Rist, C. E. 1952. Quantitativepaper chromatography of d-glucose and its oligosaccharides. Anal. Chem. 24, 1411. Easton, h-.R., Kelly, D. J., and Bartron, L. R. 1951. The use of cooling curves as a method of determining the temper of molten chocolate. Food Technol. 6, 521. Eckstein, G. R. 1937. Viscometer suitable for testing materials such as melted chocolate. U. S. Patent 2,079,247. (Chem. Abstr. 31, 4170) Eipper, IT. R. 1938. Stabilizing chocolate, butter, and other products. U. S. Patent 2,137,667. (Chem. Abstr. 33, 1834) Eipper, W.R. 1948. Isomerization of fats and oils. U. S. Patent 2,456,691. (Chenz. Abstr. 43, 2341) Englis, D. T., and Hanahan, D. J. 1945. Changes in autoclaved glucose. J. Am. Chem. SOC.67, 51. Favor, H. H., and Johnston, N. F. 1947. Effect of polyoxyethylene stearate on the crumb softness of bread. Cereal Chem. 24, 346. Fincke, H. 1932. Aroma-producing substances of cacao. Kazert 21, 381. Fore, S. P., Morris, N. J., Mack, C. H., Freeman, A. F., and Bickford, W. G. 1953. Factors affecting the stability of crude oils of 16 varieties of peanuts. J . Am. 0 2 1 Chemists’ SOC.30, 298. Forsyth, W. G. C., and Rombouts, E. 1952. Extraction of cacao pigments. J . Sci. Food Agr. 3, 161. Freundlich, L. 1937. Determining coverage value of chocolate coatings. Food Inds. 9, 630. Freundlich, L. 1939. Clearing u p misconceptions about chocolate viscosities. Food Znds. 11, 612.

58

L. F. MARTIN

Friedman, L., and Klemm, K . 1939. Diffusion velocity and molecular weight. 11. The effect of p H upon particle size in gelatin solutions. J . Am. Chem. SOC.61, 1747. Friedman, L., and Shearer, W. N. 1939. The effect of nonelectrolytes upon the time of setting of gelatin gels. J . Am. Chem. Soc. 61, 1749. Friedman, L., Klemm, K., and Thompson, F. K . 1939. Combination of gelatin with acids and bases. J . Phys. Chem. 43, 1183. Fuller, A. D. 1943. Modified starch suitable for use in foods or other uses. U. S. Patent 2,317,752. (Chem. Abstr. 37, 6152) GBlis, A. 1857. De l’action de la chaleur sur les matieres organiques neutres. Compt. rend. 46, 590. Gelis, A. 1859. Recherches sur le sucre fondu et sur un principe nouveau, la saccharide. Ann. chim. anal. et Chim. appl. (3) 67, 455. Godkin, W. J., Beattie, H. G., and Cathcart, W. H. 1951. Retardation of rancidity in pecans. Food Technol. 6, 442. Gorfinkle, W. I. 1953. Gelatin-the body control agent in marshmallow. Mfg. Confectioner 33(6), 74. Gottfried, J. B., and Benjamin, D. G. 1952. Ketose formation in alkaline dextrose systems. Znd. Eng. Chem. 44, 141. Grettie, D. P. 1940. Gelatin. U. S. Patent 2,196,300. (Chem. Abstr. 34, 5196) Grover, D. W. 1947. Keeping qualities of confectionery as influenced by its water vapor pressure. J . SOC.Chern. Ind. 66, 201. Guthrie, J. D., Hoffpauir, C. L., Stansbury, M. F., and Reeves, W. A. 1949. Survey of the chemical composition of cotton fibers, cottonseed, peanuts, and sweetpotatoes. A literature review. U. S. Dept. Agr. Bureau Agr. and Ind. Chem., Mimeographed Circ. AIC-61, p. 50. Gyorgy, P., and Tomarelli, R . 1943. Antioxidant activity in sources of the B vitamins. J . Biol. Chem. 147, 515. Hall, H. H., and Fahs, F. J. 1946. Modified pectins make possible new type candies. Confectioner 31(6), 10. Hall, L. A., and Sair, L. 1950. Antioxidant salt. U. S. Patent 2,511,804. (Chem. Abstr. 44, 8562) Hamer, W. J. 1947. Improved method for measurement of gel strength and data on starch gels. J . Research Natl. Bur. Standards 39(1), 29. Hansa-Muhle, A . G. 1937. Chocolate. German Patent 645,286. (Chem. Abstr. 31,5895). Harris, B. R. 1939. Phosphate esters for various uses. U. S. Patents 2,177,983-4. (Chem. Abstr. 34, 1416) Hart, C. V. 1930. Composition of European and California almonds. Znd. Eng. Chem. 22, 1128. Hartmann, I., and Nagy, J. 1949. The explosibility of starch dust. Chem. Eng. News 27, 2071. Hartmann, I., Cooper, A. R., and Jacobson, M. 1950. Recent studies on the explosibility of corn starch. U . S. Bur. Mines Rept. Invest. No. 4725. Heaton, E. K., and Woodroof, J. G. 1952. Extending candy shelf life. Food Eng. 24(10), 95. Heiduschka, A., and Weisemann, C. 1930. Composition of almond oil; comparison of almond oil with oil of apricot kernels. J . prakt. Chem. 124, 240. Heiss, R., Schachinger, L., and Bartusch, W. 1953. Physical and chemical principles of sugar inversion. Mfg. Confectioner 33(8), 19. Henry, K. M., Kon, S. K., Lea, C. H., and White, J. C. D. 1948. Deterioration on storage of dried skim milk. J. Dairy Research 16, 292.

CANDY MANUFACTURE

59

Hilditch, T. P., and Stainsby, W. J. 1936. Component glycerides of cacao butter. J . SOC.Chem. I n d . 66, 95T. Hills, C. H., White, J. W., and Baker, G. L. 1942. Low-sugar jellying pectinates. Proc. Znst. Food Technol., 3rd Conf. Minneapolis, Conf. 1942, p. 47. Hirst, E. L., and Jones, J. K. N. 1939. Pectic substances. 111. Composition of apple pectin and the molecular structure of the araban component of apple pectin. J . Chem. SOC.,p. 454. Holm, G. E., and Greenbank, G. R. 1923. The keeping quality of butterfat with special reference to milk powder. Proc. World’s Dairy Congr. 2, 1253. Jacobs, M. B., Ed. 1951. “The Chemistry and Technology of Food and Food Products,” 3 vols. Interscience, New York. Jamieson, G. S., Baughman, W. F., and Brauns, D. H. 1921. Chemical composition of peanut oil. J. Am. Chem. Soc. 43, 1372. Jansen, E. F., MacDonnell, L. R., and Ward, W. H. 1949. Minimum size for the structural unit of pectin. Arch. Biochem. 21, 149. Jensen, H. R. 1931. “The Chemistry, Flavoring, and Manufacture of Chocolate Confectionery and Cocoa.” J . & A. Churchill, Ltd., London. Jordan, S. 1930. “Confectionery Problems.” Mfg. Confectioner Publishing Co., Oak Park, Ill. Jordan, S. 1940. Viscosity-lowering condensation products from higher fatty acids and their esters. U. S. Patent 2,185,592. (Chem. Abstr. 34, 2966) Jordan, S. 1941. Product for lowering the viscosity of chocolate coatings. U. S. Patent 2,237,441. (Chem. Abstr. 36, 4514) Joslyn, M. A . 1951. Enzyme-catalyzed oxidative browning of fruit products. Advances in Food Research 3, 1. Joszt, A., and Molinski, S. 1935. The caramelization of sucrose. Biochem. 2. 282, 269. Kalyanasundaram, A., and Rao, D. L. N. 1951. Studies on the caramelization test. Sugar 46(3), 40. Kampf, A., and Schrenk, 0. 1929. A new viscometer. Kunstseide 11, 376. Kass, J. P., and Palmer, L. S. 1940. Browning of autoclaved milk. Znd. Eng. Chem. 32, 1360. Kaufmann, C. W., and de Maya, C. B. 1052. Coconut. U. S. Patent 2,615,812. (Chetri. Abstr. 47, 1308) Kempf, N. W. 1949a. Chocolate viscosity standardization. Mfg. Confectioner 29(8), 24. Kempf, N. W. 1949b. Crystallization of cocoa butter and its effect on the properties of chocolate. Proc. 3rd Ann. Production Conf., Penn Mfg. Confectioners Assoc. Candy Ind. 11(132), 23; (133), 10. Kempf, N. W., and Hoben, H. H. 1949. Chocolate products. U. S. Patent 2,480,935. (Chem. Abstr. 44, 246) Kertesz, Z. I. 1951a. “The Pectic Substances.’’ Interscience, New York. Kertesz, Z. I. 1951b. “The Pectic Substances,” Chapter 1, p. 6. Interscience, New York. Kertesz, Z. I. 1951c. “The Pectic Substances,” Chapter 5, p. 89. Interscience, New York. Kertesz, Z. I . 1951d. “The Pectic Substances,’’ Chapter 4, p. 82. Interscience, New Tork. Kertesz, Z. I. 1951e. “The Pectic Substances,” Chapter 23, p. 480. Interscience, Kew York. Kertesz, Z. I. 1951f. “The Pectic Substances,” Chapter 24, p. 512. Interscience, New York

60

L. F. MARTIN

Kirchner, J. G. 1949. The chemistry of fruit and vegetable flavors. Advances in Food Research 2, 259. Kline, R. W., and Stewart, G. F. 1948. Factors influencing rate of deterioration in dried egg albumin. f n d . Eng. Chem. 40, 916. Knapp, A. W. 1937. “Cacao fermentation.” Bale, London. Koch, J. 1952. Modern chocolate tempering. Mfg. Confectioner 32, No. 1, 16. Kooreman, J. A. 1952. Raw materials and processing techniques in gum candies. Confectioners’ J. 78(296), 44. Lampitt, L. H., and Money, R. W. 1936. Measurement of the strength of dilute gelatin gels. J . SOC.Chem. f n d . 66, 88T. Lampitt, L. H., and Money, R. W. 1939. Pectin from various sources. Determination of the strength of gels. J. SOC.Chem. I n d . 68, 29T. Lea, C. H. 1929. Component glycerides of cacao butter. J. SOC.Chem. f n d . 48, 41T. Lea, C. H. 1948. The reaction between milk protein and reducing sugars in the dry state. J. Dairy Resezrch 16, 369. Lea, C. H., and Hannan, R. S. 1949. Studies of the reaction between proteins and reducing sugars in the dry state. I. The effect of activity of water, of pH, and of temperature on the primary reaction between casein and glucose. Biochim. et Biophys. Acta 3, 313. Lea, C. H., and Hannan, R. S. 1950. Studies of the reaction between proteins and glucose in the dry state. 11. Further observations on the formation of the caseinglucose complex. Biochim. et Biophys. Acta 4, 518. Leach, J. M. 1953. Process for manufacturing confections. U. S. Patent 2,651,573. Leviton, A. 1938. Soluble protein powder from whey. U. S. Patent 2,129,222. (Chem. Abstr. 32, 8618) Lewis, V. M., and Lea, C. H. 1950. Note on the relative rates of reaction of several reducing sugars and sugar derivatives with casein. Biochinz. et Biophys. Acta 4,532. Liebig, A. W. 1953. Physical changes in molding starch used for fondant cream casting. M f g . Confectioner 33(1), 21. Lobry de Bruyn, C. A., and van Ekenstein, W. A. 1895. Action des alcalis sur le sucre. 11. Rec. trav. chim. 14, 203. Lobry de Bruyn, C. A., and van Ekenstein, W. A. 1896. Action des alcalis sur le sucre. 111. Rec. trav. chim. 16, 92. Lord, D. D. 1950. The action of polyoxyethylene monostearate upon starch with reference to its softening action in bread. J . Colloid Sci. 6, 360. McDowell, R. H. 1951. Improvement in or relating t o the hydrolysis of pectin. British Patent 541,528. McGee, E. F. 1948. Sweet chocolate. U. S. Patent 2,451,630. (Chem. Abstr. 43, 787) McGlamery, J. B., and Hood, M. P. 1951. Effect of two heat treatments on rancidity development in unshelled pecans. Food Research 16, 80. MacMichael, R. F. 1915. A new direct-reading viscosimeter. I n d . Eng. Chem. 7, 961. Maillard, L. C. 1912. Action of amino acids on sugars. Formation of melanoidins in a methodical way. Compt. rend. 164, 66. Maillard, L. C. 1913. General reaction of amino acids on sugars and its biological consequences. Compt. rend. SOC. biol. 72, 599. Mantell, C. L. 1945. “The Water-Soluble Gums.” Reinhold, New York. Martin, L. F., Hall, H. H., and Fahs, F. J. 1946. Progress in candy research. U. S. Dept. Agr. Rept. No. 8, p. 8. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1949. Progress in candy research. U. S. Dept. Agr. Rept. No. 18, p. 3.

CANDY MANUFACTURE

61

Martin, L. F., Robinson, H. M., and Fahs, F. J. 1951a. Progress in candy research. U. S. Dept. Agr. Rept. No. 24, p. 14. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1951b. Progress in candy research. U. S. Dept. Agr. Rept. No. 23, p. 12. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1952a. Progress in candy research. U. S. Dept. Agr. Rept. No. 25, p. 8. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1952b. Progress in candy research. U. S. Dept. Agr. Rept. No. 26, p. 2. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1952c. Progress in candy research. U. S. Dept. Agr. Rept. No. 25, p. 6. Martin, L. F., Robinson, H. M., and Fahs, F. J. 1953a. Progress in candy research. U. S. Dept. Agr. Rept. No. 27, p. 3. Martin, L. F., Robinson, H. M., and Fahs, F. J. 195313. Progress in candy research. U. S. Dept. Agr. Rept. No. 27, p. 12. Mathews, J. A., and Jackson, R. F. 1933. Stability of levulose in aqueous solutions of varying p H values. J . Research Natl. B u r . Standards 11, 619. Mayberry, M. G. 1949. Chemicals combat rancidity. Candy I n d . 11, 5. Mayberry, M. G. 1951. Stabilized chocolate material. U. S. Patent 2,539,518. (Chem. Abstr. 46, 4376) Meara, M. L. 1949. The configuration of naturally occurring mixed glycerides. V. The configuration of the major component glycerides of cacao butter. J. Chem. Soc., p. 2154. Meeker, E. W. 1950. Confectionery sweeteners. Food Technol. 4, 361. Meyers, E. W., and Graham, A. S. 1952. Temper and its evaluation. M f g . Confectioner 32, No. 6, 37. Mickevicz, M. J. 1950a. The measurement of chocolate gloss. Confectioners’ J. 76(905), 58. Mickevicz, M. J. 1951. Investigating chocolate with ultrasound. Confectioners’ J . 77, No. 921, 36. Milleville, H. P., and Eskew, R. K. 1944. Recovery and utilization of natural apple flavors. Fruit Products J . 24, 48. Milleville, H. P. 1948. Volatile-flavor recovery process. U. S. Patent 2,457,315. (Chem. Abstr. 43, 3541) Milleville, H. P. 1950. Recovering volatile flavors. U. S. Patent 2,513,813. (Chem. Abstr. 44, 8561) Mitchell, D. G. 1951. Measurement of chocolate viscosity. M f g . Confectioner 31, No. 12, 15. Mohammed, A., Fraenkel-Conrat, H., and Olcott, H. S. 1949. The browning reaction of proteins with glucose. Arch. Biochem. 24, 157. Monaghan-Watts, B. 1937. Whipping ability of soybean proteins. I n d . Eng. Chein. 29, 1009. Money, R. W., and Born, R. 1951. Equilibrium humidity of sugar solutions. J . Sci. Food Agr. 2, 180. illorell, S., Baur, L., and Link, K. P. 1934. The methylglycosides of the naturally occurring hexuronic acids. 111. Polygalacturonic acid-methyl glycosides derived from pectin. J . Biol. Chem. 106, 1. Mosiman, H., and Signer, R. 1944. Sedimentation equilibrium and molecular constitution of gelatin fractions. The Svedberg (Mem. Vol.) 464. (Chem. Abstr. 39, 1096). Musher, S. 1942. Stabilization of glyceride oils, U. S. Patent 2,282,792. (Chem. Absfr. 36, 6033)

62

L. F. MARTIN

Musher, S. 1944. Oat extract useful as a n antioxidant with foods and beverages of various kinds. U. S. Patent 2,355,097. (Chem. Abstr. 38, 6420) Myers, P. B., and Baker, G. L. 1934. Delaware State Agr. Expt. Sta. Bull. No. 187. Neal, R. H., Valteich, H. W., and Gooding, C. M. 1949. Stabilization of nutmeats. U. S. Patent 2,485,636. (Chem. Absfr. 44, 1729) Neville, H. A., Easton, N. R., and Bartron, L. R. 1950. Chocolate bloom. Food Technol. 4, 439. Oberg, E. B., Nelson, C. E., and Hankinson, C. L. 1951. Whipping agent from casein. U. S. Patent 2,547,136. (Chem. Abstr. 46, 5841). Owen, J. T. 1950. Peanut products. U. S. Patent 2,494,717. (Chem. Abstr. 44, 4164) Owens, H. S., and Maclay, W. D. 1946. Effect of methoxyl content of pectin on the properties of high solids gels. J . Colloid Sci. 1, 313. Owens, H. S., Lotzkar, H., Schultz, T. H., and Maclay, W. D. 1946. Shape and size of pectinic acid molecules deduced from viscometric measurements. J . Am. Chem. SOC.68, 1628. Paine, H. S. 1924. Constructive chemistry in relation to confectionery manufacture. Znd. Eng. Chem. 16, 513. Paine, H. S. 1928. Research in the confectionery industry. Znd. Eng. Chem. 20, 1325. Patton, A. R., Hill, E. G., and Foreman, E. hl. 1948. Amino acid impairment in casein heated with glucose. Science 107, 623. Pavlenko, 0. N. 1940. The chemical composition of the almond kernel during the ripening process. Biokhim. Kul’tur Rastenii 7, 461. (Chem. Abstr. 36, 3287) Payen, A., and Henry, 0. 1825. Sur la composition des semences de l’arachis hypogea e t son analogie avec celle des amandes douces. J . Chim. M i d . 1, 431. Penn, F. H. 1941. Chocolate of high melting point. U. S. Patent 2,244, 569. (Chem. Abstr. 36, 5590) Pentzer, D. J., Miller, H. G., and Moon, L. F. 1942. Evaporating syrups for concentration, as in treating fruits and vegetables. U. S. Patent 2,270,138. (Chem. Abstr. 36, 3284) Percival, E. G. B. 1950a. “Structural Carbohydrate Chemistry,” Chapter 8, p. 151. Prentice-Hall, New York. Percival, E. G. B. 1950b. “Structural Carbohydrate Chemistry,” Chapter 8, p. 133. Prentice-Hall, New York. Perri, J. M., Jr., and Hazel, F. 1946. Lime hydrolysis of soybean protein-foaming properties of the hydrolyzate. Znd. Eng. Chem. 38, 549. Perri, J. M., Jr., and Hazel, F. 1947. Effect of electrolytes on the foaming capacity of alpha-soybean protein dispersions. J . Phys. & Colloid Chem. 61, 661. Peters, F. N., and Musher, S. 1937. Oat flour as an antioxidant. Znd. Eng. Chem. 29, 146. Pichard, M. 1923. A method of analysis of cacao butter, and of its mixtures with vegetable butters. Compt. rend. 176, 1224. Pichard, M. 1932. Analysis of milk chocolate. Bull. oficl. ofice intern. cacao et chocolat 2, 65. Pichard, M. 1937. Cocoa butter. Bull. oficl. ofice intern. cacao et chocolat 7, 333. Pickett, T. A., and Holley, K. T. 1951. Susceptibility of types of peanuts to rancidity development. J . Am. Oil Chemists SOC.28, 478. Pictet, A., and Adrianoff, N. 1924. De l’action de la chaleur sur le saccharose. Helv. Chim. Acta 7, 703. Pictet, A., and Castan, P. 1920. Sur la glucosane. Helv. Chim. Acta 3, 645. Pictet, A., and Chavan, J. 1926. Sur une h6t6ro-16vulosane. Helv. Chim. A c f a 9 809.

CANDY MANUFACTURE

63

Pictet, A., and Marfort, A. 1923. Sur la maltosane. Helv. Chim. Acta 6, 129. Pictet, A., and Reilly, J. 1921. Sur la lkvulosane. Helv. Chim. Acta 4, 613. Pictet, A., and Ross, J. H. 1922. D e I’influence de la pression sur la polym6risation de la 1Cvoglucosane. Helv. Chim. Acta 6, 876. Pictet, A., and Stricker, P. 1924. Constitution et synthirse de l’isosaccharosane. Helv. Chim. Acta 7, 708. Pinto, A. F., and Enas, J. D. 1949. Rapid method of copra analysis and its application to the various oil seeds. J. Am. Oil Chemists’ Sac. 26, 723. Pitman, G. 1930. Comparison of California and imported almonds. Znd. Eng. Cheni. 22, 1129. Pons, W. S., Jr., Murray, M. D., O’Connor, R. T., and Guthrie, J. D. 1948. Storage of cottonseed and peanuts under conditions which minimize spectrophotometrir changes in the extracted oil. J . Am. Oil Chemists’ Sac. 26, 308. Poulsen, S. D. 1953. Is sorbitol needed in pectin confectionery jellies? Xfg. Confectioner 33(5), 47. Pucherna, J. 1950. The action of nonsugars in sugars upon the caramelization test. Listy Cukrovar. 49, 13-20. Reenikova, S. B. 1941. Rancidity of fats in stored walnuts and peanuts. T r u d y Vsesoyuz. Nauch. Issledovatel. Inst. Konditers Roi Prom. 4, 1 15-54. Richardson, W. D. 1923. A new instrument for testing glue and gelatin jellies. Chem. & Met. Eng. 28, 551. Richmond, W. L. 1953. Norris Candy Co., Atlanta, Ga. Private communication. Rose, W. G. 1948a. Nitrogenous diacylglycerophosphates. U. S. Patent 2,436,699. (Chem. Abstr. 42, 6843) Rose, W. G. 1948b. Aminoethyl esters of diacylglycerophosphoric acid. U. S. Patent 2,447,715. (Chem. Abstr. 43, 1052) Ross and Rowe, Inc. 1939. Fat-soluble condensation products. British Patent 514,848. (Chem. Abstr. 36, 4879) Itoth, W., and Rirh, S. R. 1953. A new method for continuous viscosity measurement. General theory of the ultra-viscoson. J . A p p l . Phys. 24, 940. Rubens, J. G. 1944. Chocolate. U. S. Patent 2,356,181 (Chem. Abstr. 39, 135) Sarotti, A.-G. 1943. Chocolate resistant to elevated temperatures. German Patent 744,862. (Chem. Abstr. 39, 2828) Schneider, F., and Erlemann, G. A. 1951. Developing reagents for sugars in paper chromatography. Zucker-Beih. 3, 40. Schoch, T. J. 1942. Non-carbohydrate substances in the cereal starches. J . A m . Chern. Sac. 64, 2954. Seck, and Fisher, G. 1940a. The condition of starch solutions a t extremely high shear velocities. Kolloid-2. 90, 51. Seck, W., and Fisher, G. 1940b. Starch gels. Kolloid-2. 93, 207. Singh, B., Dean, G. R., and Cantor, S. 1948. Role of 5-(hydroxymethyl) furfural in the discoloration of sugar solutions. J . Am. Chem. SOC.70, 517. Sjostedt, Ph., and Schetty, 0. 1946. Changes in the fatty material used in the chocolate industry. Mitt.Lehensm. Hyg. 37, 142. Slater, L. E. 1952. Radical pulverizing technique short-cuts chocolate refining. Food Eng. 24(7), 62. Speiser, R., Copley, M. J., and Nutting, G. C. 1947. Effect of molecular association and charge distribution on the gelation of pectin. J . Phys. & Colloid Chem. 61, 117. Speiser, R., and Eddy, C. R. 1946. Effect of molecular weight and method of deesterification on the gelling behavior of pectin. J. Am. Chem. SOC.68, 287.

mi.,

64

L. F. MARTIN

Stadtman, E. R. 1948, Nonenzymatic browning in fruit products. Advances in Food Research 1, 325. Stanley, J. 1939. Viscosity of chocolate. Confectionery Production 1939,249. Stanley, J. 1941. Viscosity of chocolate. Development of standard method. Znd. Eng. Chem. Anal. Ed. 13, 398. Stansbury, M. F., and Guthrie, J. D. 1947. Storage of samples of cottonseed and peanuts under conditions which minimize changes. J . Agr. Research 76, 49. Stansbury, M. F., Guthrie, J. D., and Hopper, T. H. 1944. Analysis of peanut kernels with relation to U. S. standards for farmer’s stock peanuts. Oil & Soap 21, 239. Steel, K.A. 1949. West Virginia Pulp & Paper Co., Industrial Chemical Sales Div., Chicago, Ill. Private communication. Steigmann, A. 1944. New methods for differentiating gelatins and glues. 6.Sac. Chem. Znd. 63, 96. Stewart, G. I?., and Kline, R. W. 1941. Dried egg albumin. I. Solubility and color denaturation. Proe. Inst. Food Technol. 2nd Conf. 1941, p. 48. Stewart, G. F., and Kline, R. W. 1948. Factors affecting rate of deterioration in dried egg albumin. Znd. Eng. Chem. 40, 916. Svedberg, T.,and Gralen, N. 1938. Carbohydrates of well-defined molecular weight in plant juices. Nature 142, 261. Tarassuk, N. P. 1947. Effect of oxygen on color and flavor of heated milk. Food Inds. 19, 781. Thuman, W. C., Brown, A. G., and McBain, J. W. 1949. Studies of protein foams obtained by bubbling. J. Am. Chem. Sac. 71,3129. Thurman, B. H. 1937. Phosphatides from soapstock. U. S. Patent 2,078,428. (Chern. Abstr. 31,4410) Tressler, D. K. 1942. Enzyme-converted corn sirup produces better glace fruit. Food Inds. 14(6), 60. Turner, J. R. 1946. Advancements in development and use of soy albumen. Mfg. Confectioner 26(6),36. Vaeck, S. V. 1950. Quelques considerations sur l’analyse thkrmique du beurre de cacao. C & N Congr. intern. indust. agr., Brussels 4, 104. (Summarized in Congr. Intern. Indust. Agr. Rap. 8(1), 551). Vaeck, S. V. 1951a. The polymorphism of some natural fats. I. Thermal and microscopic study of cacao butter. Rev. intern. chocolat. 6, 12. Vaeck, S. V. 1951b. The polymorphism of some natural fats. 11. Calorimetric studies on cacao butter. Rev. intern. chocolat. 6, 350. van Hook, A., and Bruno, A. J. 1949. Nucleation and growth in sucrose solutions. Discussions Faraday Sac. 6, 112. Visenjou, L. B. 1952. Apparatus for processing cashew nuts. Indian Patent 44,562. (Chem. Abslr. 47, 3492) Webb, B. H. 1935. Color development in lactose solutions during heating with special reference to the color of evaporated milk. J. Dairy Sci. 18, 81. Welker, P. L., Boddington, R. J., and Woodruff, S. J. 1953. Sorbitol as a humectant for coconut. U. S. Patent 2,631,104. (Chem. Absfr. 47, 5580) Whymper, R. 1921. “Cocoa and Chocolate.” J. & A. Churchill, London. Whymper, R. 1933a. “The Problem of Chocolate F a t Bloom,” p. 86. Manufacturing Confectioner Publishing, Chicago. Rhymper, R. 193313. “The Problem of Chocolate Fat Bloom,” pp. 93-104. Manufacturing Confectioner Publishing, Chicago. Rhymper, R. 1952a. A criticism of some writers in the fields of tempering and chocolate fat-bloom. Mfg. Confectioner 32(7), 17.

CANDY MANUFACTURE

65

Whymper, R. 1952b. A further note on temper and its evaluation. Mfg. Confectioner 32(12) 21. Willaman, J. J., Mottern, H. H., Hills, C. H., and Baker, G. L. 1944. Pectinate. U. S. Patent 2,358,430. (Chem. Abstr. 39, 1480) Williamson, R. V. 1929. The flow of pseudoplastic materials. Ind. Eng. Chem. 21, 1108. Wilson, C. P. 1926. Citrus Pectin. Am. Food J . 21, 279. Winger, E. L. 1952. Confectionery sales and distribution (1950). Business Information Service, U. S. Dept. of Commerce. Feb. 1952. Wolfrom, M. L., and Lewis, W. L. 1928. The reactivity of methylated sugars. 11. The action of dilute alkali on tetramethyl glucose. J . Am. Chem. SOC.60, 837. Wolfrom, M. L., and Blair, M. G. 1948. Action of heat on D-fructose. Isolation of diheterolevulosan and a new di-o-fructose dianhydride. J. Am. Chem. SOC.70, 2406. Wolfrom, M. L., Binkley, W. W., Shilling, W. L., and Hilton, H. W. 1951. Action of heat on u-fructose. 11. Structure of diheterolevulosan. J. Am. Chem. SOC.73,3553. Woodroof, J. G., Thompson, H. H., and Cecil, S. R. 1949. How refrigeration protects the quality of peanuts. Food Inds. 21, 16. Zenlea, B. J. 1938. Determining the viscosity of chocolate coatings. Food Inds. 10, 37. Zerban, F. W. 1947. The color problem in sucrose manufacture. Technol. Rept. No. 2. Sugar Research Foundation, N. Y. Ziels, N. W., and Schmidt, W. H. 1949. Hydrogenation of oils. U. S. Patent 2,468,799. (Chem. Abstr. 43, 5613) Ziemba, J. V. 1950. Charms hard candy production. Bottlenecks licked by mechanizing. Food Inds. 22, 2047.

This Page Intentionally Left Blank

Bacterial Spoilage of Wines with Special Reference to California Conditions'

BY REESE H. VAUGHN Department of Food Technology, University of California, Davis, California Page 67 68 75 78 80 82

I. Introducti ......... ................................ 11. Historical .................. ................ 111. Types of Wine Spoilage Caused by Bacteria. ................ 1. Aerobic (Oxidative) Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Anaerobic (Fermentative) Re ................. IV. Factors Affecting the Growth of B es . . . . . . . . . . . . . . . . . . . . . 1. Acidity and p H . . . . . . . . . . . . .................... 2. Sugar Content.. . . . . . . . . . . . ............................. 3. Concentration of Alcohol.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............................... ................................. ...............................

.........................................

84 84 85 87 89 89 90 90 92 92 97 99 100 101

8. Effect of Air ........................ .............. V. Characteristics of the Bacteria Found in Cal ............... 1. Differentiation of the Genera.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Acetobacter Species. . . . . . . . . . . . . . . . . ................ 3. Lactobacillus Species.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Taxonomic Status of the Cocci of Wines.. . . . . . . . . . . . . . . . . . . . . . . . . . 5. Taxonomy of Tartrate-Fermenting Lact .............. VI. Additional Research Needs. . . . . . . . . . . . . . . . .............. Acknowledgments.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

I. INTRODUCTION Pasteur and his students, in their efforts t o firmly substantiate the microbial theory of putrefaction, fermentation, and decay, made frequent and substantial contributions t o the microbiology of wines. Many of these utilitarian studies had a marked effect on establishment of the foundations for the present science of microbiology. The studies on wine culmiAlthough the author is thoroughly familiar with California wine production and has observed the other United States production areas, his knowledge of other world areas is limited t o his reading of the literature in Dutch, French, German, Italian or Spanish and to correspondence with individual foreign workers. 67

68

REESE H. VAUGHN

nated by publication of Pasteur’s famous treatise “Rtude sur le vin” in 1866 mark the real beginning of our knowledge of the bacteriology of wines. I n spite of these early and auspicious beginnings, which stimulated the activity of many workers, interest in the microbiology of wines has waned in more recent times. One natural result of this lack of interest has been that confusion has persisted and, in some instances, has been magnified either through lack of understanding of recent knowledge or refusal or inability t o give adequate attention t o taxonomy and nomenclature of the microorganisms responsible for the decomposition of wines. It is the author’s purpose t o present a critical review of the literature pertaining t o the bacteriology of mines and t o attempt to interpret existing concepts in the light of recent knowledge. It is hoped that such treatment will aid in the reduction of some of the confusion that exists both in the literature and in the industry concerning the bacterial spoilage of wines. 11. HISTORICAL CONSEQUENCES Because Pasteur contributed so much t o the beginnings of the experimental science of microbiology he exerted a very profound influence on others. Some of the results of this extraordinary influence have persisted. Even today, terminology in common usage in some of the enological literature dealing with the bacteriology of wines originated with or was popularized by Pasteur, his students, and other contemporaries. This perpetuation has been common practice despite the fact that the knowledge of bacteriology has increased to the point where much of the older terminology is without meaning. Such terms perhaps were of some descriptive value during the formative years. Yet, with the development of techniques that permitted pure culture studies, it soon became evident that different bacteria grew in wines and caused different chemical changes therein. Then, instead of discarding the old terminology for more appropriate nomenclature, different authors used the same outmoded names for describing different chemical types of wine spoilage caused by quite similar types of bacteria. For example, i t still is customary t o speak or write of the ‘(diseases” of wines and t o differentiate these as ‘(tourne,” ‘ I pousse,” “amertume,” etc. Persistent use of “diseases” of wines is ill-advised. All of the bacteria known t o spoil wines are saprophytes. Furthermore, whimsical thinking notwithstanding, wines are inanimate. As such, they are spoiled or decomposed by microorganisms that can attack the fermentable or oxidizable constituents under the conditions that exist in the various kinds of wines. Therefore correct terminology obviously should include reference t o the alteration, decomposition, defects or spoilage of wines

BACTERIAL SPOILAGE O F WINES

69

caused by bacteria or, more appropriately and specifically, fermentation or oxidation of wines caused by bacteria rather than “diseases” of wines. Perhaps it is somewhat more rational t o continue t o use ‘‘tourne,” 16 pousse,” “arnerturne,” etc., or translated equivalents, as descriptive nomenclature for types of wine spoilage. These terms originally were used in France t o denote simple changes in wine and meant the equivalent of sourness (turning), gassiness, and bitterness of wines, respectively. Yet carelessness has resulted in much unnecessary confusion concerning their respective meanings in the more recent enological literature. “Tourne” probably has been used the most indiscriminately of all. As already mentioned, “tourne” was first used by the French simply to denote mine spoilage characterized by obvious sensory changes including development of turbidity, sourness, insipid or bitter after-taste, and, in the case of red wines, loss of pigment. Evolution of other meanings was gradual and came as the result of natural divergence of opinions accompanying chemical and bacteriological progress. Probably the first attempt t o connect “tourne” with a definite chemical decomposition of wines was made by Sick1i.s. Nickks in 1862 claimed that tartrate disappeared in wines so affected. BBchamp (1862) and G l h a r d (1862) working independently also came t o the conclusion that tartrate was fermented in cases where ‘‘ tourne” spoilage was involved. Pasteur (1866), the first t o point out the relationship of microorganisms t o spoilage of wines, recognized several specific bacterial defects of wines which he called (‘maladies des vins”: 1. Maladie de l’acescence d u vin.-Vins piquks, aigres, etc. 2. Maladie des vins tournks, montes que ont la pousse, etc. 3. Maladie de la graisse.-Vins j l a n t s , V i n s huileux. 4. Maladie de l’amertume, De l’arner. D u goiit de vieux, etc. Here, however, Pasteur paid no specific attention t o the fermentation of tartrate when he referred to ‘ L f o u r n e l and f “pousse.” Later (Pasteur, 1873) he mentioned that tartrate decomposition occurred in ‘‘uins tournes ” but did not attach any special significance t o that fermentation. Ile did, however, stress the similarity between the bacteria seen in “vins tournes” and the “ferment lactique” that he had studied previously. It is also clear that he considered (‘pousse” only as a gassy stage of “tourne.” Since many of the spoiled samples which Pasteur observed and described were charged with carbon dioxide as well as having other characteristics of ‘(vinstournes,” on the basis of present knowledge it may be assumed by inference a t least that he may have seen gas-forming (heterofermentative) lactobacilli in his microscopic preparations of the wines. Thus, a t the very beginning, the meaning of “tourne” was confused by three concepts:

70

REESE H. VAUGHN

(1) Concept of enologists. A spoilage of wine characterized by turbidity, sourness, insipid or bitter after-taste, and, in the case of red wines, loss of color. (2) Concept of Nicklds and contemporary chemists. A spoilage of wine characterized chiefly by the fermentation of tartrate. (3) Concept of Pasteur. A spoilage of wine characterized by turbidity and sourness caused by rod-shaped bacteria which in some cases produced gas from fermentable substances of the wine. The study of the end-products of the tartrate fermentation per se as well as in spoiled wines proceeded. The investigations, dating from the work of Nollner in 1841, were culminated by the experiments of Fitz in 1879 which established that propionic acid might originate from the fermentation of tartrate.2 Detection of propionic acid in wines that had undergone tartrate fermentation was claimed by Bkchamp (1862)) Duclaux (1874), Nickl&s (1862), Schultz (1877) , Semichon (1905)) and others. Duclaux and Semichon, who were especially emphatic in their claim that propionic acid resulted from the tartrate fermentation, materially aided the confusion because they insisted th a t tartrate fermentation with production of propionic acid was characteristic of “ v i n pousse ’’ rather than “vin tourne.” Thus, the chemical concept of “tourne” was modified to include production of propionic acid from tartrate and, a t the same time, became relegated to synonomy with (‘pousse.” I n the meantime Gautier (1878) claimed to have recovered tartronic acid as well as acetic, propionic, and lactic acids from ( ( v i n s tournes.” Others likewise claimed to have recovered tartronic acid from wines that had undergone tartrate fermentation. Although to the present writer’s knowledge the work of these investigators has never been verified in detail, this concept of “tourne” has been carefully preserved in some of the enological literature. Those interested in the bacteriological aspects of wine spoilage had been handicapped immeasurably by lack of an adequate method for pure culture studies. Once, however, the solid media developed by Koch (1881) became known generally, the bacteria of spoiled wines were subjected to intensive study. It soon was possible to unequivocably associate definite groups of bacteria with certain types of wine spoilage having definite chemical characteristics. One of the first instances of proof of this nature resulted from the study of mannitol formation in wines. It was evident from the reports of Carles (1891), Portes (1892), Roos (1892), and others that mannitol 2 Van Niel (1928) has presented a n excellently detailed discussion of some of the general historical aspects of this problem.

71

BACTERIAL SPOILAGE O F WINES

occurred as an end-product of a specific wine spoilage apparently peculiar t o the warm climates of southern France, northern Africa, Italy, and Spain. Gayon and Dubourg (1894, 1901) isolated, described, and demonstrated the causal organism t o be a short, rod-shaped bacterium which they called the (‘ferment mannitique.” Gayon and Dubourg maintained that the bacteria that caused mannitol production in wines were quite different from those that caused “vin tourne.” Despite the contentions of some of their contemporaries, the “maladie de la mannite” soon was accepted as a specific spoilage of wine, thus adding a new “disease” t o the group of defects already recognized. The malic acid decomposition in certain very acid wines of Switzerland and Germany was also investigated during this period. Both spherical and rod-shaped lactic acid bacteria had been found t o cause a desirabIe reduction in the total acidity of these tart mines. The bacteria decompose malic acid with the formation of lactic acid and carbon dioxide as major end-products, thus effecting a beneficial reduction in the acidity (Koch, 1900; Kunz, 1901 ; Moslinger, 1901 ; Muller-Thurgau, 1891, 1908; Seifert, 1901). However, in wines normal or deficient in acidity, the activity of lactic acid bacteria can hardly be considered beneficial. It seems logical therefore that the decomposition of malic acid may also be considered a defect of wines. At the same time the causes of mannitol production and malic acid fermentation were being investigated, other workers were very busy applying the newer bacteriological techniques t o other types of wine spoilage. L-nfortunately, many of the workers did not describe enough of the essential characteristics of the bacteria with which they worked. Others were more thorough in this respect but failed t o test the ability of their isolates t o cause the same defect when inoculated into sound, sterile wine.3 Still others described bacteria in detail which did not cause any change when inoculated into sound, sterile wine. It is also probable that many of the wines examined during this period were deficient in alcohol and acid content; otherwise many of the bacteria (members of the genus Bacillus and the coliform bacteria in particuIar) could not have been recovered from the wines. It is certain that many of these bacteria would not survive the concentrations of alcohol and acidity found in bona fide table wines, much less spoil them. Nonetheless, between about 1890 and 1910, the main groups of bacteria responsible for the spoiling of wines had been isolated and described in enough detail t o make them recognizable even today. The bacteria responsible for acetification (‘( maladie de l’acesence of wine, the Mycoderma aceti of Pasteur, had been studied in pure culture by

”>

8

These criticisms are especially valid in the light of present knowledge.

72

REESE H. VAUGHN

Henneberg (1898), Kramer (1892), and others. The rod-shaped lactic acid bacteria also had been associated with (‘maladie de l’amertume.” These accomplishments, however, did not nullify the confusion that had existed previously; if anything, they intensified it, particularly with respect t o the role of the rod-shaped lactic acid bacteria in the spoilage of wines. As already stated, these rod-shaped bacteria had been definitely associated with four specific wine defects. Careful consideration of the original descriptions of these bacteria in the light of present knowledge makes i t clear that the heterofermentative lactobacilli could cause changes characteristic of any one or all four defects (“tourne,’’ ‘Lpousse,ll mannitol formation, and malic acid fermentation) depending upon conditions that might prevail initially in wines they infected. I n some manner, however, the concept developed that specific saprophytic bacteria had definite habitats and caused specific alterations just as different parasitic bacteria caused specific diseases of animals and humans. It soon was common practice among enologists t o write of “ tourne l 1 bacteria, “ pousse” bacteria, “mannite” bacteria, (‘malolactic l 1 bacteria, ‘(ropy” bacteria, etc. This concept has persisted t o the present, despite the fact it never was tenable.4 An understanding of the ability of lactic acid bacteria t o grow and produce chemical changes in wines best illustrates the indefensibility of such a concept. All bona fide table wines naturally contain some glycerol, malic, and tartaric acids. (Citric acid may be added t o wines to correct natural deficiencies in acidity but never has entered seriously into any discussions of spoilage although Charpenti6 et al. (1951) have investigated its decomposition in wine.) Sweet table wines also contain glucose and fructose, glycerol, and malic or tartaric acid5 which may be attacked, depending upon the type of wine and the individual characteristics of the bacteria. All lactobacilli and cocci known t o spoil wines can decompose glucose and fructose if these sugars are present. If the invading lactic acid bacteria are gas-forming types (heterofermentative) , they will ferment these sugars with production of significant quantities of carbon dioxide and acetic acid in addition t o lactic acid and also cause reduction of fructose t o mannitol. If they are nongas-forming species (homofermentative), lactic acid will be the major product formed from the sugars. The 4 A t the time this concept was being established (about 1890 until 1910) the bacterial decomposition of appetizer and dessert wines had not been investigated, or at least, had not been recorded in the literature. 6 Until now the ability of lactic acid bacteria to decompose tartrate is somewhat controversial despite the claims of Arena (1936), Mvller-Thurgau and Osterwalder (1919), Osterwalder (1952), and RibCreau-Gayon (1947). The discrepancies will be discussed in the following pages.

BACTERIAL SPOILAGE O F WINES

73

same bacteria may or may not be able t o decompose glycerol, and malic and tartaric acids. Both types produce a “mousey” taste under proper conditions. Thus, Lactobacillus brevis, a common heterofermentative species capable of spoiling table wines, might cause chemical changes characteristic of the different defects (propionic acid production excepted), all in the same wine. Obviously, also, this species might cause different chemical changes when growing in different wines. When one considers that a number of different species of Lactobacillus can do likewise, the concept obviously is without much meaning. The specific habitat concept also becomes indefensible when it is realized that lactobacilli isolated from fermenting sauerkraut, cucumbers, or olives, as well as from saliva, feces, milk, and many other sources can be trained to spoil musts and wines. The first serious attempts t o clarify and correct the anomalous situation with respect t o defects of mines caused by lactic acid bacteria were made by Muller-Thurgau and Osterwalder. I n a series of papers (MullerThurgau ttnd Osterwalder, 1912, 1918, 1919), these authors presented results of their exhaustive studies of the bacteria that they found responsible for defects in Swiss grape and fruit wines. Their basic contributions to the knowledge of the bacteriology of wines were substantial. Every effort was made t o determine the chemical changes caused by the different cultures in wines as well as in media containing various fermentable substances. The fermentations of glucose, fructose, glycerol, and malic, citric, and tartaric acids were investigated in detail. Furthermore, and rightly so, meticulous attention was given t o the taxonomy and nomenclature of each culture. Muller-Thurgau and Osterwalder (1912, 1918) recognized four species of “niannite bacteria.” Thus, for the first time it was clear that the ferment mannitique” of Gayon and Dubourg (Zoc. cit.) was only one of a number of closely related types of heterofermentative lactobacilli capable of forming mannitol from fructose. Furthermore, the importance of the rod-shaped lactic acid bacteria in the fermentation of malic acid was clarified. It also became certain as a result of their work that many lactobacilli as well as cocci could cause the decomposition of malic acid in Swiss wines. The investigations concerning the fermentation of sugars and malic acid in wines were followed by studies of the bacteria purported t o cause decomposition of glycerol and tartaric acid in wines. Miiller-Thurgau and Osterwalder (1919) isolated a new bacterium they claimed was responsible for the fermentation of glycerol and tartaric acid in wines. The organism was named Bacterium tartarophthorum. Muller-Thurgau and Osterwalder allocated the species t o the mannite bacteria and compared it

74

REESE H. VAUGHN

morphologically t o other bacteria now known to be heterofermentative lactobacilli. Bacterium tartarophthorum, grown in wine or laboratory media, caused the decomposition of tartrate with the production of acetic acid and carbon dioxide and fermented glycerol with the formation of acetic, propionic, and lactic acids. Thus, for the first time, the origin of propionic acid in spoiled wines was linked experimentally with the substrate, glycerol, rather than tartrate as had been claimed previously by Duclaux (1874) and Semichon (1905) in particular. Miiller-Thurgau and Osterwalder (loc. cit.) concluded that since Bacterium tartarophthorum formed propionic acid from glycerol rather than tartaric acid, as so commonly maintained in the enological literature, continued use of “tourne” or “pousse” as designations for types of wine spoilage was not justified. They suggested that a more exact nomenclature be used t o indicate different types of wine spoilage; for example, catabolism of tartaric acid” or “decomposition of glycerol” etc. The desirability of adherence t o such nomenclature t o avoid unnecessary confusion is obvious. Arena (1936) was one of the first to verify the work of Miiller-Thurgau and Osterwalder (Zoc. cit.) in some detail and, in addition, present new knowledge concerning the bacteriology of wines. One of Arena’s principal contributions dealt with a study of tartrate decomposition by a bacterium he named Bacterium acidovorax. Arena considered Bacterium acidovorax together with Bacterium tartarophthorum t o be the causes of ((tourne” (“torcido” in Spanish) of Argentinian wines. Arena’s new species perhaps is a homofermentative type of Lactobacillus, whereas Bacterium tartarophthorum may be a representative of the heterofermentative lactobacilli. Ribereau-Gayon also apparently was able t o confirm the observations of Miiller-Thurgau and Osterwalder in some detail. He found tartratefermenting as well as malate-decomposing bacteria among those he isolated. Ribkreau-Gayon was particularly interested in the malic acid fermentation of the red wines of Bordeaux (Ribbeau-Gayon, 1936, 1946) and was assisted by Peynaud in some of these studies (Ribkreau-Gayon and Peynaud, 1938a,b). Emphasis was placed on the chemical aspects of the problem of malic, citric, and tartaric acid fermentations, and balances were published. Unfortunately, however, in all of this work the taxonomy and nomenclature of tartrate-fermenting bacteria isolated were not treated extensively or in the light of present knowledge. It is significant, however, that Ribbreau-Gayon (1947) considered the controversial “tourne” t o be a wine spoilage characterized by the fermentation of tartrate and glycerol although he did not emphasize propionic acid as a breakdown product of either substrate.

BACTERIAL S P O I L A G E O F WINES

- F

la

Perhaps the most striking reason for the seeming equivocation on the part of the present author results from an attempt t o explain the role of propionic acid in the spoilage. As already stressed, the majority of writers have included propionic acid formation together with tartrate fermentation as distinguishing features of “maladie d u vin tourne” and have incriminated lactobacilli as causes of the spoilage. I n view of the present state of our knowledge such a view is untenable for the lactobacilli do not produce propionic acid from the major fermentable constituents of wines, although they probably are responsible for most cases of wine spoilage. Other investigators, equally interested in the bacteriology of spoiled wines, did not encounter bacteria isolated from spoiled wines that were able t o decompose tartrates either in mines or in culture media. Kor did they report propionic acid as a spoilage product. These investigators include Fornachon (1943), Fornachon et al. (1940, 1949), Olsen (1948), and Vaughn et al. (1949). I n phases other than tartrate decomposition, and propionic acid formation the results reported by these latter investigators, for the most part, confirm the studies previously reported by Arena, Muller-Thurgau and Osterwalder, and Ribereau-gay on. The need for additional detailed and critical study was stressed by Muller-Thurgau and Osterwalder but, in the intervening years, too little attention has been paid t o the problem. The majority of the “modern” writers have continued t o use the old nomenclature and probably, in many cases, were not even aware of the important work of the Swiss investigators until comparatively recently. Unfortunately, some of those familiar with the research of Muller-Thurgau and Osterwalder have accepted their conclusions in toto without any apparent effort t o obtain any additional confirmatory data. As a result, the recent enological literature pertaining t o the “maladie de la tourne” may contain only the outmoded knowledge of the past century, only the somewhat newer knowledge based on the critical work of Muller-Thurgau and Osterwalder, or a very confusing admixture of both the older and newer concepts. Therefore, both the meaning and cause of “tourne” present a perplexing riddle to beginning students and, in particular, develop unnecessary and unjustified prejudices among trained scientists. 111. TYPESOF WINE SPOILAGE CAUSEDBY BACTERIA That which is considered t o be a reasonable approach t o the nomenclature of bacterial wine spoilage is presented in Table I. The original suggestion for this type of nomenclature derived from the work of Muller-Thurgau and Ostermalder (1919). The abnormalities are dividcd into two groups, depending upon whether the chemical reactions induced

76

REESE H. VAUGHN

TABLEI Chemical Abnormalitiw of Table Wines Induced b y Bacteria Bacteria causing Substrate decomposition attacked

Main products of decomposition

Common manifestations of abnormality

Enological nomenclature

AEROBIC(OXIDATIVE) REACTIOXS Acetobacter

Gluconic and ketogluconic acids; on complete oxidation, COZ and

Glucose

H2

Increase in total acidity; “mousiness”

Not specific but obviously might be confused with “pousse” or “tourne” because of increase in total acidity

Maladie de I’acescence” ; vinegar fermentation; acetification No specific name

0

Fructose

Not commonly attacked by vinegar bacteria found in wines

Ethyl alcohol

Acetic acid; on complete oxidation, C o n and HZO

Increase in volatile acidity

Glycerols

Dihydroxyacetone and COz

Increase in reducing substances

(vinegar bacteria)

Not known to be attacked in wines Malic acid Only COPhas been identified

‘I

Lactic acid6

Citric acid Only CO, has been ident.ified

Decrease in total acid content Decrease in total acid content

Only COz has been identified

Decrease in total acid content

Tartaric acid

Might be confused with malo-lactic fermentation Might be confused with malo-lactic fermentation Might be confused with malo-lactic fermentation

ANAEROBIC (FERMENTATIVE) REACTIONS Glucose

Homofermentative lactic acid bacteria Lactic acid Increase in total acidity; “mousiness”

Commonly but perhaps incorrectly called “tourne“ when caused by rodshaped forms

77

BACTERIAL SPOILAGE O F WINES

TABLE I (Continued) Bacteria causing Substrate decomposition attacked

dficrococcusb Fructose (spherical forms)

Ethyl alcohol

Main products of decomposition

Lactic acid

Common manifestations of abnormality

Increase in total acidity; “mousiness”

Enological nomenclature

Commonly but perhaps incorrectly called “tourne” when caused by rodshaped forms

Not attacked

or Lactobacillus Glycerol” Lactic acid Increase in total (rod forms) acidity Lactic acid- Not attacked Pvlalic acid Lactic acid and Decrease in carbon dioxide total acidity Citric acid Lactic and acetic Decrease in acids, carbon total acidity dioxide, diacetyl, acetoin, 2,3butanediol Tartaric Lactic acid and Decrease in total acid carbon dioxide acidity Heterofermentative Lactic Acid Bacteria Lactic acid, ethyl Glucose Increase in fixed and volatile alcohol, glycerol, acidity; carbon dioxide effervescence ; ‘ I mousiness” Presence of Fructose Mannitol in addimannitol in tion to products addition to mentioned for those abnorglucose malities mentioned for glucose Leuconostoc Ethyl Not attacked (spherical alcohol

Malo-lactic fcrmentation Not gcnerally considered in the literature on wine spoilage

“Tourne”(?)

Commonly called “tourne” or “pousse”

“

Tourne,” “pousse” or mannite fermentation

forms)

Glycerolo

Not commonly attacked but propionic acid and CO, claimed products

Questionable increase in volatile acidity

Tourne”(?)

78

REESE H. VAUGHN

TABLE I (Continued) Bactcria causing Substrate decomposition attacked Lactic acida

Main products of decomposition

Common manifestations of abnormality

Enological nomenclature

Not commonly attacked

or Lactic acid and Lactobacillus hfalic carbon dioxide (rod acid forms) Citric acid Lactic and acetic acids, carbon dioxide, diacetyl, acetoin, 2,3bu tanediol Tartaric Not commonly acid attacked. Acetic acid and carbon dioxide claimed products Sucrosec Poly saccharides, (dextrans)

Decrease in fixed acidity

Malo-lactic fermentation

Decrease in fixed acidity

Not generally considered in the literature on wine spoilage

Questionable increase in volatile acidity

“Tourne”(?)

Ropy, viscous fruit wines

“Maladie de la graisse ”

Both glycerol and lactic acid are formed by wine yeasts (Saccharomyces cereuisiae var. eZZipsoideus) in varying quantities during the fermentation of grape sugars. As will be noted both compounds also may be produced by the heterofermentative lactic acid bacteria from several normal substrates. Few glycerol-fermenting lactic acid bacteria have been isolated from spoiled n-ines. As will be shown these bacteria are not members of the genus Micrococcus b u t probably belong with the genus Streptococcus. Sucrose is not added t o California grape wines but is used in fruit wines. Ropy. viscous fruit wines result when infected u i t h Leucoiiostoc species which form polysaccharides from sucrose.

by the bacteria are aerobic (oxidati-ve) or anaerobic (fermentative) in nature. I . Aerobic (Oxidative) Reactions The compounds found in wines that are oxidized by the acetic acid bacteria (genus Acetobacter) are shown in the table. The vinegar bacteria supposedly utilize ethyl alcohol or the intermediate acetaldehyde in preference t o the other oxidizable substrates. Consequently, the conversion of alcohol t o acetic acid is the most common oxidative abnormality of bacterial origin observed in wines. This conversion of alcohol (acetification) i s a strictly aerobic process which m a y occur in fermenting musts or finished wines, as are the other oxidations discussed in this section. The nomenclature used for this particular oxidation has not been greatly confused in the literature. It is known t o the French as “maladie de l’ascescence ” and in the United States is variously called acetic souring

BACTERIAL SPOILAGE O F WINES

79

or spoilage, vinegar souring or spoilage, vinegar fermentation, and acetification. However, the alcohol oxidation is not a fermentation in the true sense. Therefore, it is not correct t o refer t o the abnormality as a fermentation, common usage notwithstanding. Acetification is a much more appropriate term and has the distinct advantage of ease of translation. Glucose oxidation is one of the conversions caused by the acetic acid bacteria which may, t o a degree, be confused with fermentative reactions caused by the lactic acid bacteria. Glucose, if present in a wine, may not be oxidized because, as already stated, the intermediates acetaldehyde or alcohol are claimed t o be preferred as substrates by the vinegar bacteria. However, such is not the case with fermenting musts. Here, if for any reason aldehyde or alcohol cease t o become available, the bacteria concentrated in the surface area of the must begin t o oxidize the glucose t o gluconic acid. If the total fixed acidity of the must increases only moderately as the result of gluconic acid formation it may or may not be accompanied by the development of a (‘mousey” taste. T h i s change has been confused with simple souring caused by species of the genus Lactobacillus. The confusion has resulted because the organoleptic changes indicate only souring and perhaps ‘(mousiness.” Furthermore, the usual, conventional, microscopic analyses cannot be relied upon t o detect genera and species of bacteria which may grow in grape musts. If, however, the oxidation of gIucose proceeds far enough, there is a very abnormal increase in fixed acidity, the must is certain t o taste ‘(mousey” and, in addition, develops a very characteristic “ sweet-sour ” taste because the acetic acid bacteria have not oxidized the fructose a t all. I n most instances, however, the acetic acid bacteria finally oxidize both alcohol and glucose in musts and cause increases in volatile (acetic) as well as in fixed (gluconic) acids. Formerly, this spoilage, known as rapid acetification, was quite common in the interior valleys of California. Vaughn (1938) showed that the rapid acetification mas caused by rather unusual strains of Acetobacter aceti which could grow in association with the wine yeasts and oxidize the alcohol as it was formed by the yeasts. When enough acetic acid accumulated, or the temperature interfered, the yeasts ceased to function. Then the bacteria began t o oxidize the glucose or the acetic acid which had already accumulated. Musts so affected contained excessive amounts of both volatile and fixed acidity, generally were “mousey” and always had a “sweet-sour ” taste. Although the acetic acid bacteria also oxidize glycerol and citric, malic, and tartaric acids in vitro, there is no evidence t o indicate they normally attack these compounds in wine per se because of their claimed preference for alcohol. It should be stressed, however, that the strains of Acetobacter aceti described by Vaughn (1938) have the power t o almost

80

REESE H. VAUGHN

completely deacidify acetous table wines on prolonged incubation under aerobic or partial aerobic conditions. Furthermore, A . xylinum, although apparently unable t o grow in table wines of normal alcohol content, can cause deacidification of wine vinegar if the acidity is of the concentration used for retail channels (4.0 t o 5.0%) and aerobic or partial aerobic conditions prevail. 2. Anaerobic (Fermentative) Reactions

With the exception of alcohol, all of the simple carbohydrate constituents found in table wines are fermented by the lactic acid bacteria. As shown in Table I, these fermentations fall into two groups: homofermentative or heterofermentative, depending upon the manner in which the bacteria decompose the hexoses. The bacteria causing homofermentative breakdown of the hexoses produce a preponderance of lactic acid, whereas those causing heterofermentative decomposition of glucose or fructose produce appreciable quantities of acetic acid, alcohol, carbon dioxide, and glycerol as well as lactic acid. I n addition, some fructose is reduced t o mannitol by the heterofermentative species. It is believed that failure t o acknowledge these fundamental differences in fermentation of the hexoses by the lactic acid bacteria is responsible for much of the serious confusion in the enological literature pertaining t o the bacteriology of wines. Additional confusion undoubtedly results from the failure t o realize that both homofermentative and heterofermentative species of lactic acid bacteria can ferment the organic acids in wines, particularly citric and malic acids, although some species do ferment tartaric acid. It is true that some of the lactic acid bacteria prefer malic acid but others prefer citric acid and apparently will not attack malic acid at all. It should be clear, therefore, that a decrease in titratable acidity (the conventional method of detecting deacidification) is not necessarily evidence for the presence of “malo-lactic ” bacteria or the (‘malo-lactic ” fermentation in wines but may indicate fermentation of any one or all of the acids present. It should be clear, too that there can be no real difference between the rod-shaped “mannitic ” bacteria and the rod-shaped “malo-lactic ” bacteria of wines. The heterofermentative lactobacilli must, by definition, produce mannitol from fructose and also may attack the organic acids as well. The same is true for the heterofermentative cocci of the genus Leuconostoc (consult Breed, Murray, and Hitchens, 1948 for more detail). The fermentation of glycerol also is claimed t o occur in wines (MullerThurgau and Osterwalder, 1919; Osterwalder, 1952). However, there is some confusion concerning the end-products of glycerol fermentation in wines by the heterofermentative lactic acid bacteria. Miiller-Thurgau and

BACTERIAL SPOILAGE O F WINES

.81

Osterwalder (1919) and Osterwalder (1952) reported these products t o be propionic acid and carbon dioxide. Nevertheless, i t is certain that, if propionic acid actually was identified by these investigators as a n endproduct of the glycerol fermentation, the lactic acid bacteria were not responsible for its formation. Neither the heterofermentative cocci (Leuconostoc) nor rods (Lactobacillus) commonly ferment glycerol (Breed, Murray, and Hitchens, 1948). Furthermore, glycerol-fermenting strains of either group are extremely rare in wines (Fornachon, 1943; Fornachon et al., 1949; Muller-Thurgau and Osterwalder, 1918; Olsen, 1948; Vaughn et al., 1949) and, t o the present author’s knowledge, species of Propionibacterium or other bacteria capable of forming propionic acid from glycerol, viz. Clostridium propionicum (Cardon and Barker, 1946) have never been isolated from wines. However, from the work of Gunsalus (1947), Gunsalus and Sherman (1943), and Gunsalus and L-mbreit (I 945) with streptococci, one may speculate that the lactobacilli also may decompose glycerol t o lactic acid or t o lactic and acetic acids, depending upon the nature of the individual cultures, the kind and concentration of hydrogen acceptors, and accessory growth factors. Glycerol decomposition in wines also has been claimed t o cause bitterness, “ maladie de l’amertume,” by Pasteur (1866)) Duclaux (1901)) and Voisenet (1910a)) among others. As the result of a detailed study Voisenet (1910a,b, 1911a,b; 1913, 1918) claimed the causal organism to be a new facultative aerobic, sporeforming type which he named Bacillus amaracrylus. This bacterium which causes the formation of acrolein from glycerol is a probable synonym of Bacillus polymyxa (Breed, Murray, and Hitchens, 1948). The exact cause of bitterness is by no means clear. As stressed, Voisenet and others have claimed i t resulted from the decomposition of glycerol. Voisenet also presented some evidence t o show that it was caused by a specific bacterium. However, others have claimed bitterness t o originate by other means, as pointed out by Muller-Thurgau and Osterwalder (1912). Bitterness has not been encountered in California wines, according t o Cruess (1943). The spoilage probably is of very rare occurrence or it would have received more attention in California and elsewhere (Cruess, 1947; Fornachon, 1943; Ribkreau-Gayon, 1947). There is some question, too, whether Bacillus polymyxa can grow in table wines which meet the legal requirements of the United States for acidity and alcohol content. The author and his associate, Professor G. L. Marsh (unpublished data), have encountered this species in distilling material (alcohol content 3 t o 6 % by volume; p H 4.0 t o 4.8) for brandy manufacture. Although the cultures isolated would grow in distilling materials, stillage, and tartrate infusion liquors, attempts t o grow them in table

82

REESE H. VAUGHN

wines were unsuccessful. However, many of the wines of Europe are low in acidity as well as in alcohol (vin ordinaire and fruit wines in particular). Therefore, it is plausible to expect that bacteria of the Bacillus polymyxa group might grow and produce bitterness in such wines. Vaughn and Stadtman (1946) showed that the related species Bacillus macerans could grow in acid surroundings having initial p H values of 3.8. Slimy or ropy wines (“maladie de la graisse”) also are of very infrequent occurrence in California. The legal requirement which prevents the use of supplementary sugar (sucrose) probably explains the absence of descriptions of such spoilage in the California literature. On the other hand, ropy wines apparently occur rather frequently in Europe, particularly in fruit wines (Kayser and Manceau, 1909; Luthi, 1949, 1950, 1953; Olsen, 1948; Rentschler, 1948). As indicated in Table I, the ropy, viscous wines result from the production of polysaccharides, probably dextrans (Hassid and Barker, 1940) or levulans (Niven et al., 1941) from sucrose or possibly dextrin (see review by Hehre, 1951). Olsen (1948) who investigated ropy Danish fruit wines considered the organism he isolated (Cetacoccus arabinosaceus, a synonym for Leuconostoc mesenteroides) to be identical with the “ferment de la graisse” described b y Kayser and Nanceau (1909). Ropy wines of Switzerland have been studied in some detail by Hochstrasser (1955), Liithi (1949, 1950, 1953), Martin (1948) and Rentschler (1948). Luthi (1953) has described the species Streptococcus nzucilaginosus var. vini which he believes is responsible for production of ropiness in Swiss wines. Buchi and Deuel (1954) have reported some of the characteristics of the slime produced by this species. The possibility also exists that the heterofermentative species of Lactobacillus described by Mayer (1939) and Perquin (1939-1940) may be found to cause ropiness in certain wines. Recent investigations by Hehre and Hamilton (1949) indicate that species of Acetobacter might, under the right conditions, produce ropiness in wines. Shimwell (1947, 1948a,b) already has associated this genus with ropiness in beers. The alcohol contained in California table wines is not attacked by the lactic acid bacteria. However, as will be shown, its presence and, particularly its concentration, have a profound effect on the growth of bacteria in wines. IV. FACTORS AFFECTINGTHE GROWTHOF BACTERIA IN WINES

It is common knowledge among enologists that wines vary widely in their susceptibility to attack by spoilage bacteria. It is not so well-known nor widely appreciated how differences in the chemical composition of wines are related to differences in susceptibility and how such differences

BACTERIAL SPOILAGE O F WINES

83

may, t o a degree, be controlled t o prevent spoilage. Factors which are known t o influence the susceptibility of wines include the following: (1) Acidity and pH. (2) Sugar content. (3) Concentration of alcohol. (4) Presence and concentration of accessory growth substances (vitamins and amino acids). (5) Concentration of tannins. (6) Presence of sulfur dioxide. (7) Temperature of storage. (8) Relation of air (aerobic vs. anaerobic storage).

It is obvious that these various factors may be manipulated t o a certain extent to increase or decrease the resistance of a \vine t o spoilage. The synergistic effects of acidity, alcohol, and sulfur dioxide are well known. However, the degree to which these effects can be manipulated usefully will depend upon the kind of wine and the particular type of spoilage organism as well as the other factors involved in the over-all resistance or susceptibility of the wine. 1. Acidity and p H Wines with a high total acidity are less liable t o undergo spoilage than those with a low total acidity. However, the p H value rather than the total acidity of the wine determines for the most part the effect of acidity 011 bacterial growth. Therefore, in general, the lower the p H value (the higher the total acidity) the less susceptible the wine will be to spoilage by bacteria. However, because of the limitations of the natural acidities of wines, and the legal requirements which determine the amounts of extraneous acids (generally citric) which may be added, n-ines never have pH values low enough to prevent the growth of acetic acid bacteria. Therefore, these bacteria must be limited by other means (sulfur dioxide and anaerobic conditions). On the other hand, the growth of the lactic acid bacteria is influenced by the pH values of different wines. The limiting p H for the growth of lactic acid bacteria in table wines of California is in the range 3.3 t o 3.5 depending upon the type of wine and the concentration of alcohol (unpublished data of the present author). The pH value a t which Lactobacillus trichodes fails t o grow in appetizer and dessert wines is about 3.5 (Fornachon, 1943; Fornachon et al., 1949). Although the p H values of California wines vary from as lorn as 3.2 to as high as 4.2, the majority range between 3.5 and 4.0, even after corrections have been made t o offset naturally occurring deficiencies in acidity (Amerine, 1954; Amerine and Joslyn 1940, 1951 ; Joslyn and Amerine,

84

REESE

H.

VAUGHN

1941). It is obvious, therefore, that many California wines do not have p H values which necessarily will limit the growth of lactic acid bacteria if fermentable and accessory growth substances are available in the wines.

2. Sugar Content Lack of sugar, rather than variations within the normal limits, is probably the factor which retards the growth of bacteria in wines the most. I n the dry table wines which contain a minimum of sugar (about 0.1 %) bacteria rarely cause spoilage unless they can initiate deacidifaction. However, increasing amounts of residual sugar support more vigorous growth. Unmistakable, serious evidence of spoilage in flavor and odor changes usually occur when 0.5 t o 1.07, or more residual sugar is available. That natural sugar content alone does not limit growth of wine spoilage bacteria is shown by the fact that they grow well in grape musts (Vaughn, 1938; Vaughn and Douglas, 1938). 3 . Concentration of Alcohol

The inhibition of bacteria by ethanol is well-known. Legal requirements control the alcohol content of California wines (State of California, Department of Public Health, 1946). California table wines may contain from a minimum of 10% t o a maximum of 147G of alcohol by volume. California appetizer and dessert mines may contain from a minimum of 14% t o not more than 21 % of alcohol by volume. The minimum alcohol requirements for various wine types also are specified. The bacteria capable of spoiling California wines vary in their tolerance for alcohol. The limiting concentration of alcohol even varies from species to species and from strain to strain. The maximum concentration of alcohol in wines tolerated by the acetic acid bacteria is between 14 and 15% by volume (Vaughn, 1942). However, there are species and strains of Acetobacter which are unable to tolerate more than 8 t o 10% of alcohol by volume. These include strains of Acetobacter x y l i n u m and Acetobacter melanogenum as well as some strains of Acetobacter aceti and Acetobacter oxydans, even though these latter species apparently are among the most alcohol-tolerant types of the acetic bacteria (unpublished data of the present author). T o prevent acetification of Spanish sherry, Bobadilla (1943) recommended a minimum of 14.5% alcohol in the wine, whereas Cruess (1948) recommended fortification of the sherry stock to 15.5 % alcohol by volume t o prevent activity of the vinegar bacteria. There also is a great variation in alcohol tolerance among the lactic acid bacteria. The deacidifying cocci including Micrococcus variococcus, Microccoccus acidovorax, and Microccoccus multivorax, according to Arena

BACTERIAL SPOILAGE O F W I N E S

85

(1936) and Muller-Thurgau and Osterwalder (1912), are among the least resistant t o alcohol. These species are inhibited by 8.5% alcohol and will not grow when the alcohol concentration is increased t o 11.85% by volume (Arena, 1936). The species of cocci which cause ropiness and deacidification, Leuconostoc mesenteroides and Leuconostoc dextranicus, are strongly inhibited by 10% alcohol by volume, according t o Olsen (1948). However, the publications of Luthi (1949, 1950, 1953) and Rentschler (1948), as well as unpublished data of the present author, indicate that the slime-forming cocci can tolerate the usual concentrations of alcohol found in California table wines. T o the present author’s knowledge, none of the cocci cause spoilage in appetizer or dessert wines. The rod-shaped forms, particularly the heterofermentative species of Lactobacillus, exhibit the greatest resistance t o alcohol of any of the wine spoilage bacteria. All of the species that can spoil table wines will grow in wines containing at least 15% alcohol. Some can tolerate 18% alcohol in wines. One species, Lactobacillus trichodes, can grow in wines containing more than 20% alcohol by volume (Fornachon, 1943; Fornachon et al., 1949; Olsen, 1948; Vaughn et al., 1949). The homofermentative species are not so alcohol tolerant. Although all of the homofermentative species of Lactobacillus commonly recognized will grow in grape musts and wines of 5 t o 8 % alcohol (unpublished data of the present author), their growth is markedly limited when the alcohol content is increased t o 10% by volume (Arena, 1936; Olsen, 1948). Lactobacillus plantarum is the only homofermentative species the present author and his students have encountered in California table wines. None of the homofermentative lactobacilli have been reported t o spoil fortified wines (Fornachon, 1943; Fornachon et al., 1949; Olsen, 1948; and Vaughn et al., 1949).

4. Accessory Growth Factors As already stressed, wines vary widely in their susceptibility t o attack by spoilage bacteria. I n all probability, the principal reason for the observed differences in susceptibility of similar wine types results from variations in the presence and concentration of vitamins, amino acids, and other organic compounds required for growth of the bacteria. The nutritional requirements for the spoilage bacteria vary from the very simple t o the most complex. The vinegar bacterium (Acetobacter aceti) requires only a n inorganic source of nitrogen plus the conventional inorganic salts and alcohol or acetic acid for its growth in laboratory media (Hoyer, 1989; Visser’t Hooft, 1925) and will grow in a11 CaIifornia table wines under suitable conditions of aerobiosis and lack of sulfur dioxide (Vaughn, 1942). On the other hand Lactobacillus trichodes, one of the most fastidious of all lactobacilli, will not grow in laboratory media

86

REESE H. VAUGHN

or wines unless these are supplemented with yeast autolysate (Fornachon, 1943; Fornachon et al., 1949). Thus, unless protected, wines containing less than about 14y0alcohol by volume are suscept,ible t o acetification if air is present although the same wines may present varying degrees of resistance t o attack by the lactic acid bacteria. The lactic acid bacteria all require a utilizable source of energy and certain inorganic salts for growth. I n addition, because they belong t o that group of microorganisms (heterotrophs) which have lost the ability t o synthesize many specific compounds required for formation of their protoplasm, the lactic acid bacteria also require a few or many preformed compounds including vitamins, amino acids, purine and pyrimidine bases, and other organic compounds (Cheldelin and King, 1953; Dann and Satterfield, 1947; Dunn, 1947; Hendlin, 1954; Hutchings and Peterson, 1943; Knight, 1945; Moller, 1939; Orla-Jensen et al., 1936; Peterson and Peterson, 1945; Schweigert and Snell, 1947; Snell, 1945a,b, 1946, 1948, 1950; Snell et aZ., 1937,1938; Stokes, 1952; Wood el al., 1940). It is known that the amino acid and vitamin content of wines vary widely (Amerine, 1954; Amerine and Joslyn, 1951; Cailleau and Chevillard, 1949; Castor, l950,1953a,b,c; Castor and Guymon, 1952; Genevois and Flavier, 1938,1939; Luthi and Vetsch, 1952; Perlman and Morgan, 1945; Peynaud, 1951 ; Ribkreau-Gayon and Peynaud, 1952a; Watt and Merrill, 1950). Thus, it is obvious that variations in susceptibility of wines to spoilage must result in part a t least from differences in availability of nutrients required for growth of the lactic acid bacteria. (Contrariwise, there is no good evidence t o indicate a lack of accessory nutrients (vitamins, etc.) for yeasts in California musts as might be inferred from some of the recent publications. I n reality, yeast fermentations in California musts generally proceed too rapidly for the best quality wine production unless controlled by cooling.) It is easy to demonstrate that the wine yeasts supply most of the required preformed nutrients. VCTines, unless protected by chemical preservatives, or where the acidity (pH) is a limiting factor, can be rendered susceptible t o attack by bacteria simply by the addition of a small amount of yeast autolysate. The yeast autolysate contains most, if not all, of the nutrient accessory materials (vitamins, amino acids, etc.) required for growth of the lactic acid bacteria (Carter, 1950; Peterson, 1950; Vosti and Joslyn, 1954a,b). Wines always contain carbohydrates which can serve as a source of energy for the lactic acid bacteria. It is believed that wines contain enough of the required inorganic constituents for growth of the spoilage bacteria (Amerine, 1955). The susceptibility or resistance of wines t o spoilage then must depend in large measure on whether the yeast is allowed to remain in contact with the fermented wine

BACTERIAL SPOILAGE OF WINES

87

long enough t o undergo autolysis (Fornachon, 1943; Niehaus, 1932; Olsen, 1948). (It is assumed here that p H values, alcohol contents, total acidities, etc., are not limiting factors.) 5. Effect of Tannin

Tannin is commonly added t o musts and wines t o facilitate the subsequent fining (clarification) with gelatin. Tannin also may be added to improve the flavor and protect the color of wines which have a natural deficiency in tannin. It is not t o be denied that tannins retard the growth of bacteria in wines (Cruess, 1935; Fornachon, 1943). It should be understood, however, according t o the work of Fornachon (1943), that the amount of tannin which can be added without adversely affecting the flavor of wine is below the amount required t o prevent the growth of all of the spoilage bacteria. Fornachon concluded that it was unlikely that many wines could be protected from spoilage by addition of tannin alone. 6. Effect of Sulfur Dioxide

Sulfur dioxide6 (SOs) is used in the wine industry t o prevent the growth of bacteria already present on the grapes or in the musts and wines. It is also used t o protect wines from subsequent attack by bacteria. Its general utility is based upon the fact that i t selectively inhibits most of the undesirable microorganisms whereas the wine yeasts inherently are more resistant t o its effects. Sulfur dioxide sometimes is used also because of its reducing or antioxidative properties or because it exerts an acidifying, clarifying, and dissolving influence (Amerine and Josyln, 1940, 1951; Joslyn and Braverman, 1954). The antiseptic power of sulfur dioxide is influenced by the composition of the must or wine, the species, numbers, and activity of the microorganisms, and the temperature. Sulfur dioxide reacts readily with sugars and aldehydes, and in a short time a considerable portion of that added t o a must or wine will become “bound.” In this combined form sulfur dioxide is considerably less active as a germicide. According t o Bioletti and Cruess (1912) “bound” SO2was only %.50 t o 360 as antiseptic as the “free” (uncombined) sulfur dioxide. Wine yeasts are considerably less sensitive t o sulfur dioxide than are most of the undesirable bacteria, molds, and some yeasts which are found on grapes or in wines. Furthermore, the inherent resistance of the wine yeasts t o sulfur dioxide can be markedly increased through intentional training of pure strains used for starters (Porchet, 1931), or unintentional 6In the wine industry the terms “sulfur dioxide” and “SO,” are used interchangeably to include pure sulfur dioxide as well as aqueous solutions of sulfites which occur in musts and wines treated with sulfur dioxide, sulfurous acid, or metabisulfites.

88

REESE

H.

VAUGHN

acclimatization through faulty sanitation practices (Scheffer and Mrak, 1951). As pointed out by the latter the yeasts most resistant to sulfur dioxide in wines are not true wine yeasts (Saccharomyces cerevisiae var. ellipsoideus) in the classical enological sense. These investigators found the species S. chevalieri, S. carlbergensis var. monacensis, S. oviformis, S. cerevisiae, Pichia alcoholophila, and Candida rugosa to cause either clouds or sediments or both in California white table wines. The isolates of S. chevalieri possessed characteristics very similar to those of the Spanish sherry yeasts described by Hohl and Cruess (1939), Marcilla e2 al. (1936), and Schanderl (1926). Since the California wine industry has been producing Spanish-type sherry on a n experimental to full commercial scale for some time it may be inferred that these sherry yeasts have become widely distributed. It appears that the use of 350 p.p.m. total SOZ, the maximum concentration allowed by regulation (U.S. Treasury Department, Bureau of Internal Revenue, 1948), is necessary to prevent the growth of these yeasts in wine. There also is some evidence to show that well acclimatized yeasts will tolerate even more than the maximum allowable concentration of SO2 in certain commercial wines. On the other hand, the spoilage bacteria are sensitive to much less sulfur dioxide. Various authorities (Amerine and Joslyn, 1940, 1951; Bioletti and Cruess, 1912; Cruess, 1943, 1947, 1948; Fornachon, 1943; Joslyn and Amerine, 1941; Ribereau-Gayon, 1947) agree, in the main, that from about 50 t o as much as 150 p.p.m. of SO2 are required to control the spoilage bacteria in wines. As already stressed, the amount of sulfur dioxide required to control bacteria is determined in part by the chemical composition of the wine. More SOz is required t o protect a wine which has a high p H value and contains much sugar and other reactive substances than a wine which has a low p H value and little reactive material to dissipate the SOe. The amount to be added to the wine, and the concentration a t which it is t o be maintained, must be as small as possible to protect it adequately and yet not render the wine unpalatable. It is generally agreed that the sulfiting of musts results in sounder, cleaner fermentations. The wines have a lower volatile acidity, the total acidity is more uniform, and the yield of alcohol is increased when the musts are sulfited. Furthermore, such wines keep much better than those obtained by natural fermentations (Amerine and Joslyn, 1940, 1951; Cruess, 1943, 1947; Fornachon, 1943; Joslyn and Amerine, 1941; Rib6reau-Gayon, 1947). The amount of sulfur dioxide recommended by Amerine and Joslyn (1940) for use in musts under various conditions of commercial operation in California is shown in Table 11. It is mandatory that the SOz be

89

BACTERIAL SPOILAGE OF WINES

TABLE I1 Recommended Amount of Sulfur Dioxide t o Be Added to Grapes under Various Conditions of Qualityo

Condition and temperature of grapes

Amount of SO2 recommended for use (approximate p.p.m.)

Clean, sound and underripe; cool Sound, mature, cool Moldy, bruised, overripe, low in acid; hot

75 115 275

“ Adapted

from Amerine and Joslyn (1940)

thoroughly mixed into the must. Unless this mixing is done, portions of unsulfited must may become infected and reduce the quality of the resultant wine. It is also necessary t o allow for some time t o elapse between the sulfiting operation and the inoculation of the must with the yeast starter. Freshly added sulfur dioxide is quite toxic to the yeast because most of it is “free.” After 2 t o 4 hr. have elapsed enough of the added SO2 will have combined with reactive compounds of the must t o permit inoculation without undue retardation of the yeast. In any event, routine sulfiting should be practiced with judgment and care for, if proper sanitary precautions are used, sulfur dioxide is not required in objectionable amounts. (For a more detailed discussion of the role of sulfur dioxide in winery practices consult Amerine and Joslyn, 1951; and Joslyn and Braverman, 1954.) 7 . Temperature of Storage

There is no evidence to indicate that normal storage temperatures for wine have any great effect on the ultimate growth of spoilage bacteria (Breed et al., 1948; Fornachon, 1943; Fornachon ef al., 1949; RibBreauGayon, 1947; Schanderl, 1950; Vaughn, 1942; Vaughn et al., 1949). It is probable that most of the spoilage bacteria grow most rapidly in musts and wines when the temperature is between 20 and 35°C. (68 t o 95°F.). However, mines stored a t 10 to 15°C. (50 to 59°F.) undergo significant spoilage over a long period of time. Spoilage of musts is accelerated a t temperatures of 35 t o 40°C. (95 t o 104°F.) because of the retardation of the wine yeast. High temperatures during fermentation and storage of new wines are detrimental because of the resultant acceleration of autolysis of the yeast left in the wine until the first racking (Fornachon, 1943). 8. Effect of Air

It is a universal practice t o store wines under conditions as nearly airfree as possible, t o protect them from undesirable oxidation-not for any

90

R E E S E H. VAUGHN

effect anaerobiosis might have on specific spoilage organisms. However, since the acetic acid bacteria are aerobic forms i t is obvious that storage of wines under strictly anaerobic conditions will prevent their growth. Some enologists have attempted t o classify the lactic acid bacteria found in wines as obligate anaerobes, but such is not the case. They will grow in wines either in the presence or absence of air although some species apparently have a preference for a reduced oxygen supply even when grown in pure culture. Because of this microaerophilic tendency of some of the types, storage of wines under anaerobic conditions will tend t o favor the development of these bacteria unless the wine is protected.

v. CHARACTERISTICS

BACTERIA FOUND I N CALIFORNIA WINES As already indicated, both the acetic acid and lactic acid bacteria cause spoilage of California wines. However, the species of either group able to grow are restricted t o those types which can tolerate the low p H values, withstand the effects of alcohol, and utilize the comparatively poor source of nutrients found in California wines. The following discussion of the genera Acetobacter, Lactobacillus, and Leuconostoc is confined t o a description of those species which have been isolated and studied in detail by the author and his associates. OF THE

I . Differentiation of the Genera Generic differentiation is the first fundamental step in the accurate classification of the bacteria which cause spoilage of wines. Pure cultures of bacteria of the genera Acetobacter, Lactobacillus, and Leuconostoc have TABLEI11 Principal Generic Characteristics of Bacteria Known to Cause Spoilage of California Wines Genus Differential character Cell morphology

A cetobacter

Lactobacillus

Leuconostoc

Rods, ellipsoidal to long -

Rods, ellipsoidal to long

Cocci, spherical to ellipsoidal

Facultative aerobic

Facultative aerobic

Fermentative 25 to 35°C. (77 to 95°F.)

Fermentative 20 to 35°C. (68 to 95°F.)

Gram stain Catalase production Relation to oxygen Obligate aerobic Oxidation of alcohol Oxidative Glucose utilization Optimum tempera- 20 to 35°C. ture for glucose (68 to 95°F.) utilization

+ +

+-

+

91

BACTERIAL SPOILAGE O F WINES

characteristics which permit their separation with comparative ease. These characters are shown in Table 111. Under normal conditions determination of the cell morphology and gram stain, detection of the presence or absence of catalase, determination of the oxidative or fermentative utilization of glucose including the end-products of its decomposition, and determination of the ability t o oxidize ethanol t o vinegar are sufficient t o separate these three genera. There is such a marked difference between the genus Acetobacter and the other two genera that members of this genus should never be confused with Lactobacillus or Leuconostoc types. However, the chance for confusion between Lactobacillus and Leuconostoc types is greater, particularly as regards the cell morphology of the heterofermentative species of Lactobacillus as compared with Leuconostoc types from wine. Although the majority of the lactobacilli are definite rods, sometimes strains are observed that have very short, ellipsoidal cells that occur singly or in pairs. The cells of Leuconostoc types, although supposed t o be spherical, very frequently are ellipsoidal in shape and also occur singly or in pairs. Since the cells of both genera are gram-positive and cataiase-negative, the chance for confusion is obvious. Therefore, it is advisable t o use the additional physiological characters shown in Table I V TABLEII’ Additional Characteristics Useful for Separation of Lactobacillus and Leuconostoc Types from Wines Genus

Lactobacillus Differential character

.-

homofrrmmtativr heterofermentative

Leuconostoc

-~

Products of glucose utiliz:~ tion

Lactic acid

Products of fructose utilization

Lactic acid

l’olysaccharide production from sucrose Alcohol tolerance, per cent by volume, maximum Salt tolerance, grams per 100 ml., maximum

Lactic acid, acetic acid, carbon tlioxide, ethanol Mannitol in addition to products from glucose

Lactic acid, acetic acid, carbon dioxide, ethanol Mannitol in addition to products from glucose

+ 10 to 14

15 to 21

10 to 14

> 10

5 to 6

> 10

92

REESE H. VAUGHN

to complete the differentiation of the genera Lactobacillus and Leuconostoc. If this additional care is practiced, the three genera can be differentiated without any of the confusion already mentioned. 2. Acetobacter Species

Only two species of vinegar bacteria capable of causing acetification of California wines have been encountered by the author. These are Acetobacter aceti and Acetobacter oxydans. Their characteristics are quite different as is shown in Table V. The two species may be differentiated with a minimum of effort by three simple tests. Acetobacter aceti has a n optimum temperature of 30 t o 35°C. (86 t o 95°F.) for growth and oxidation of ethanol or glucose, causes complete oxidation of either compound, and grows in Hoyer's medium (see Hoyer, 1898; Visser't Hooft, 1925; Vaughn, 1942). I n contrast, Acetobacter oxydans has a low optimum temperature [20"C. (68"F.)] for growth and oxidation of ethanol or glucose, cannot cause complete oxidation of either compound, and does not grow in Hoyer's medium. It is to be emphasized that the possibility exists that other species of vinegar bacteria may be encountered in the spoilage of California wines. The present author and his students have repeatedly recovered other species from fresh grapes and musts. The species Acetobacter x y l i n u m has been isolated regularly from these sources, whereas Acetobacter melanogenum, Acetobacter suboxydans, and Acetobacter roseurn have been encountered very infrequently. It is likely that differences in alcohol tolerance and nutritional requirements are responsible for their apparent inability t o grow in California table wines. 3. Lactobacillus Species

Lactobacillus plantarum is the only homofermentative species the present author and his associates have recovered from California wines. I n contrast, 5 different species of heterofermentative lactobacilli have been isolated. These are Lactobacillus brevis, Lactobacillus fermenti, Lactobacillus buchneri, Lactobacillus hilgardii, and Lactobacillus trichodes. The characteristics which differentiate these species are shown in Table VI. The separation of the heterofermentative lactobacilli from wines into groups which may be justified as species is a very difficult task. The fermentation of pentoses and optimum temperature relationships have been used most frequently for separation of the heterofermentative species of Lactobacillus (Orla-Jensen, 1919, 1943; Pederson, 1938, 1939, 1948). If these are accepted as valid characters t o use for primary differentiation of the heterofermentative species, then it is obvious that the cultures of

TABLEV Characteristics of Acetobacter Species Known to Cause Acetification of California Table Winesa ~~

~

Species

A . aceti

Film formation

Temperature for growth and oxidation of ethanol

Oxidation of ethanol

Dry, thin, 30 to 35°C. (86 to 95°F.) Complete ascending (CO, HZO)

+

A . o q d a n s Dry, thin, asrrnding

20°C;. (68°F.)

Incomplete (Only as far as acrtic acid)

Compounds commonly oxidized Ethanol, propanol, glycol, glucose

Ethanol, propanol, glycol, arabinose, glucose, fructose, galactose, sucrose, maltose, raffinose, dextrin, erythritol, mannitol, glycerol

Differential characters* Growth in Hoyer’s medium; complete oxidation of ethanol; high optimum temperature

No growth in Hoyer’s medium; incomplete oxidation of et,hanol; low optimum temperature

m

P

0

H

r

E*

0

M

o 4

A

5

s

Adapted from Vaughn (1942). This medium is used to detect the ability of Acetobmter cultures to utilize ammonium salts as sole sourccs of N. See Hoyer (1898); Visser’t Hooft (1925); Vaughn (1942). a

b

W

W

94

REESE H. VAUGHN

TABLE Over-all Characteristics of Various Cultures of

Species Lactobacillus plantarum

Lactobacillus brevis

Lactobacillus buchneri

Lactobacillus fermenti

Lactobacillus hilgardii

-

Morphology in wines

Optimum conditions for glucose fermentation

Cell arrangement

Temperature range Initial pH ("C.) range

Single, pairs some chains Single, pairs some chains Single, pairs some chains Single, pairs some chains Single and short chains Long chains

Filamentation

5

b c

d

++ ++

Frequent

30 to 35 4.5 to 6.5 (86 to 95°F.)

Occasional

30 to 35 4.5 to 6.5 +-t (86 to 95°F.)

Frequent

30 to 35 4.5 to 6.0 (86 to 95°F.)

Frequent

30 to 35 4.5 to 5.5 (86 to 95°F.)

Frequent

30 to 35 4.5 t o 5.5 (86 to 95°F.)

++

Marked 25 t o 30 4.5 to 5.5 inter(77 to 86°F.) twined hairlike clumps Leuconostoc mesenleroides Single, pairs Marked 20 to 30 4.5 to 6.5 elonga- (68 to 86°F.) tion of rrlls in pairs Leucmostoc deztranicunc Single, pairs 3farkc.d 20 to 25 4.5 to 6.5 elonga- (68 to 77°F.) tion

Lactobacillus trichodes

..

m

++ ++

+ ++

-

-

-

+ ++ -

-

It is to be stressed that individual cultures vary in their ability to ferment the different carbo-, N o activity; f.0.5 to 1.5 ml. 0.1 N NaOH to neutralize 10 ml. of culture; 1.6 to 4.0ml. Same relative activity for all cnltures. based on decrease in titratable acidity. Alcohol tolerance based on growth of "trained" cultures in wine with 1 ml. yeast autolyaate per (-), No activity when initial pH of medium was 6.5 to 7.0.

+.

95

BACTERIAL SPOILAGE O F W I N E S

VI Lactic Acid Bacteria Found in California Wines" Fermentationh (Initial pH of basal medium 5.0 to 5.5) Alcohol

++ ++

F

+ ++ ++ ++ ++ - - ++

+ + + + + + + + ++ F

++- + + +

(-)e

I

(-1" k

+

-

-

++

F

-

-

++

(-P (-P

+++++ +++++ + + - -

-

15to18

f (-18

-

15to18

f

-

1 5 t o 18

-

10 to 14

-

10 to 14

?c

(-I*

(-18

++I (-1'

+ + + + + + + + + + + + - - - ++

f (-10

hydrates. However. all cultures thus f a r encountered ferment glucose and fructose. 1 N NaOH t o neutralize 10 ml. of culture; 5 or more mlO.1 N NaOH t o neutralize 10 ml. of culture

+ +,

100 nil. of nine.

96

REESE H. VAUGHN

of Lactobacillus fermenti described in Table VI may not be properly allocated. According t o the generally accepted definition, based on the work of Pederson (1938), representatives of the species Lactobacillus fermenti should not ferment pentoses or, a t most, xylose and should have a n optimum temperature ranging from 35 t o 42°C. (95 t o 107.6”F.) (Pederson, 1939, 1948). The cultures called Lactobacillus fermenti (Table VI) fermented xylose (an allowable variation for the species) but had a n optimum temperature for glucose fermentation ranging from 30 to 35°C. (86 t o 95°F.). Furthermore, these cultures had a n optimum temperature range of 20 t o 25°C. (68 t o 77°F.) for growth in table wines. Fornachon (1943) encountered similar cultures among those he isolated from Australian dessert wines. The close similarity between the cultures called Lactobacillus fermenti and those of the “inactive ” species Lactobacillus hilgardii is obvious. It is believed that the former cultures are but more active strains of the “inactive ” species Lactobacillus hilgardii (Vaughn et al., 1949). The species Lactobacillus buchneri is not too different from the more common Lactobacillus brevis as shown in Table VI. Some might consider these two species t o be similar enough t o identify both as Lactobacillus brevis. It may be that some of the lactic acid bacteria have become acclimatized t o the lower temperatures associated with storage of wines and thus have altered their inherent preference for a specific temperature range. One fact disputes this assumption. Fornachon et al. (1940) have already demonstrated that the initial p H of the medium has a marked effect on the ability of heterofermentative lactobacilli t o utilize glucose and other energy sources. The characteristics of the cultures differentiated in Table VI were determined in such a manner. Those fermentations designated as (-) were negative when the initial p H of the test medium was between 6.5 and 7.0. Consequently, the present author is of the opinion that the fermentation characteristics of all of the heterofermentative species of lactobacilli, regardless of source (milk, olives, pickles, wine, etc.) should be re-evaluated. Care should be taken t o establish the optimum initial pH for fermentation of each energy source, and its relation t o the optimum temperature range for growth should be determined. The investigations of Fornachon (1943) and Fornachon et al. (1949) appear t o have established the authenticity of the species Lactobacillus trichodes, the only type known t o cause spoilage of California appetizer and dessert wines. This species, commonly known as the “hair bacillus” in the wine industry of California, is the most alcohol-tolerant but is the most inactive in other respects. It ferments only glucose and fructose well; i t will not grow in wines unless yeast autolysate is present. It must be trained t o grow in alcohol-free media and, once acclimatized t o an

BACTERIAL SPOILAGE OF WINES

97

alcohol-free environment, must be trained again t o grow in fortified wines. It always grows very slowly and has a low optimum temperature for growth, even in media within the optimum initial pH range of 4.5 to 5.5. The present author also is of the opinion that Lactobacillus hilgardii is a valid species. However, until the optimum initial p H relationships of a large group of related species have been determined, there may be some obvious doubt as t o its authenticity. I n the experience of the present author, Lactobacillus hilgardii is next in importance t o Lactobacillus plantarum and Lactobacillus brevis in the spoilage of California table wines. As seen from Table VI, the few cultures of Leuconostoc which were studied presented no problems in differentiation. However, as will be shown, the presence of these bacteria in California wines prompted speculations concerning the taxonomy of the cocci isolated from wines by other investigators.

4, Taxonomic Status of the Cocci of W i n e s As already emphasized, several investigators have described and named species of cocci which, for many years, have been associated with malic acid decomposition in wines. Invariably these species have been placed in the genus Micrococcus. The four species commonly recognized in the enological literature include Micrococcus malolacticus Seifert (1903) ; Micrococcus acidovorax and Micrococcus variococcus Miiller-Thurgau and Osterwalder (1912) ; and Micrococcus multivorax, Arena (1936). The taxonomic status of these species is not clear. On the basis of published information concerning their morphology and physiology it is questionable that the four species should be placed in the genus Xicrococcus a t all. On morphological grounds the four species are very similar. The cells orcur singly, in pairs, in triads, tetrads, and sometimes in irregular masses, but division is never in three planes (Arena, 1936; Miiller-Thurgau and Ostcrivalder, 1912; Seifert, 1903). The cells are not motile (Arena, 1936; hIuller-Thurgau and Osterwalder, 1912; Seifert, 1903). The cells are gram-positive (Arena, 1936; Muller-Thurgau and Osterwalder, 1912). The physiological similarities also are striking. All species are assumed t o be homofermentative, lactic acid-producing cocci even though the characteristics of the fixed acidity produced from the hexoses has not been determined. This assumption is based on several facts. The species do not produce significant quantities of carbon dioxide or volatile acid (acetic) from glucose and do not produce mannitol from fructose (Arena, 1936; Muller-Thurgau and Osterwalder, 1912). The species ferment malic acid in wines and produce carbon dioxide and lactic acid (Arena, 1936;

98

REESE H. VAUGHN

Muller-Thurgau and Osterwalder, 1912). The species tolerate 10 to 14% alcohol by volume (Arena, 1936; Muller-Thurgau and Osterwalder, 1912). Furthermore, it is known that two of the species, Micrococcus varicoccus and Micrococcus multivorax, do not form catalase (Arena, 1936). The failure of these bacteria to produce catalase may be significant. It indicates the fundamental physiological relationship of these wine cocci t o other catalase-negative species of the genus Streptococcus. The consistent occurrence of some, but not all of the cells of the individual species in tetrads, is also considered significant because it indicates a close morphological relationship between these wine cocci and other similar cocci which cause “beer sickness” and fermentation of vegetables (Henneberg, 1926; Lindner, 1927, 1928; Mees, 1934; Pederson, 1949; Shimwell, 1940, 1948a; Shimwell and Kirkpatrick, 1939). It might be concluded that these catalase-negative, tetrad-f orming cocci, regardless of origin, do not belong in the catalase-positive genus Micrococcus. However, Felton, et al. (1953) have reported that catalase is produced by pedicocci if the medium contains little sugar. On this basis it would appear that the tetrad-forming cocci of wines are more closely related to members of the genus Micrococcus than to Streptococcus. It is a matter for additional study before deciding whether the tetradforming cocci from the wine should be placed in the genus Pediococcus or in the genus Streptococcus. Pederson (1949) concluded from a study of similar bacteria that the “genus should be included in the tribe Streptococceae of the family Lactobacteriaceae with the genera Diplococcus, Streptococcus, and Leuconostoc, rather than in the family Micrococcaceae, as in Bergey’s Manual (1949).” Pederson would place these bacteria in the genus Pediococcus which he defined as follows: “the genus should include those gram-positive, nonmotile, nonspore-forming cocci that occur in tetrads and sometimes singly or in pairs, show poor surface growth because they are microaerophilic, are high acid homofermentative lactic acid producers and do not reduce nitrates, liquefy gelatin or produce catalase.” Five cultures isolated by the author in 1949 did have characteristics which would permit them t o be placed in the genus Pediococcus. They were catalase-negative, gram-positive cocci which occurred singly, in pairs, triads, or tetrads; grew poorly on the surface of agar media; produced a homofermentative lactic acid fermentation of glucose; and always decomposed malic acid with the formation of lactic acid and carbon dioxide. However, the tendency for tetrad formation was not marked except when the cultures were grown in wine. I n media containing glucose and tryptone (Difco) or freshly prepared yeast autolysate the cells most commonly occurred in pairs, unless the initial p H of the medium was low. For this reason, it might be that the wine cocci described above should be placed

BACTERIAL SPOILAGE O F WINES

99

in the genus Streptococcus. Shimwell and Kirkpatrick (1939) and Walters (1940) already have placed simiIar tetrad-forming beer cocci in the genus Streptococcus because they felt t h a t tetrad formation was a character that marranted no more than specific consideration. Shimmell (1940, 1941, 1948b, 1949) has persisted in this view, whereas hilees (1934) and Pederson (1949) have preferred t o use the genus Pediococcus. It is also questionable whether the type of lactic acid formed from sugars (ix. dextrorotatory, levorotatory, or inactive) should merit more than specific importance, although both hlees and Pederson have stressed its importance in the description of Pediococcus. As described in Bergy’s Manual of Determinative Bacteriology, the genus Leuconostoc (Hucker and Pederson, 1948) provides for either D- or L-acid formation, and in the genus Lactobacillus (Pederson, 1948) the type of lactic acid formed is used for species differentiation. Confusion also may arise concerning classification of some of the other cocci that cause wine spoilage. One may, with reason, question the correctness of the generic allocation of Streptococcus mucilaginosus var. vini Luthi (1953). The same questions arise with respect t o the species Streptococcus ma!olacticus and Xireptococcus malolacticus var. mticilaginosus Hochstrasser (1955). It is not clear from the description whether Streptococcus muci:aginosiis var. iini produced a significant quantity of carbon dioxide from glucose. Carbon dioxide production from glucose, not malic acid, is the criterion for heterofermeiitative activity according t o definition. ill1 true lactic acid bacteria, cocci and rods, probably produce some carbon dioxide in addition to lactic acid if they decompose malic acid or malates. One thing is certain, until the generic position of these different cocci is firmly established, species differentiation of the three ecological groups (beer, fermenting vegetables, and wine) will remain confused. 5 . T a x o n o m y of Tartrate-Fermenting Lactobacilli

As indicated earlier in this review, some confusion still exists with respect t o the taxonomy of the tartrate-fermenting lactobacilli of wines. Recent work by the present author and his students has dispelled doubts previously expressed (Vaughn, 1949) concerning the existence of tartratefermenting lactic acid bacteria. Tartrate-fermenting cultures of Lactobacillus plantarum were isolated from spoiled red wines arid lees (Berry and Vaughn, 1952). These cultures, carefully trained for alcohol tolerance, could be grown in table wines enriched with yeast autolysate if the pH values or the sulfur dioxide contained in the wines were not limiting factors. Lactic acid and carbon dioxide were the major end-products of tar-

100

REESE H. VAUGHN

trate decomposition by these cultures of L. planiarum (Krumperman, Berry, and Vaughn, 1953). These observations are considered t o be ample verification of Arena’s (1936) work with the homofermentative species, Bacterium acidovorax. Unfortunately, the taxonomic position of Bacterium tartaropthorum’ Muller-Thurgau and Osterwalder (1919) is not clear. The catalase reaction was not given in the original description. No mention was made of this importa.nt reaction in the recent description of this species by Ostermalder (1952). I n view of the fact t h a t both descriptions stress the ability of this species t o decompose glycerol with formation of propionic acid, i t is obvious B. tartarophthorum may not belong with the lactic acid bacteria a t all. Additional studies are indicated.

VI. ADDITIONAL RESEARCH NEEDS The research needs already mentioned, directed as they are toward the accumulation of additional scientific knowledge, are largely academic in scope. Of a more practical nature is the search for a preservative t o replace sulfur dioxide in finished, bottled wines in order t o prevent undesirable yeast growth or bacterial spoilage. The “sterilization filtration ” technique has been used in an attempt to overcome the objectionable flavor and odor of sulfur dioxide in bottled table wines of high quality. The technique and its limitations are discussed in some detail by Amerine and Joslyn (1951). Studies also have been made of the use of antibiotics and other chemicals in the preservation of wines. For a summary of information pertaining t o the possible use of antibiotics and other chemical preservatives t o protect bottled wines against microbial activity the reader should consult Dal Cin (1948, 1950), Kielhofer (1953), Peynaud (1952), Peynaud and Lafourcade (1953a,b, 1954a,b), P r a t t et al. (1950), Ribbreau-Gayon and Peynaud (1952b,c), Ribbreau-Gayon et al. (1952a,b), Verona (1952, 1953), and Verona and Picci (1953) among others. No attempt has been made t o review the subject in detail. As yet antibiotics have not been approved for use in wines in the United States and chemical preservatives such as the halogenated fatty acids (monochloracetic acid, etc.) are not permitted (Federal Food and Drug Administration, 1941). Furthermore, sulfur dioxide, because of its time honored use in the wine industry, is exempt from label declaration. Therefore, it is without point t o discuss this subject in detail in these pages. 7 Conformance with accepted nomenclature would require t h a t the genus name Bacterium be replaced with Lactobacillus, at least in the United States.

BACTERIAL S P O I L A G E O F W I N E S

101

ACKNOWLEDGMENTS Special thanks are due to Professor W. V. Cruess who was the first to stimulate the author’s interest in the bacteriology of wines. Thanks also are due to many of the author’s graduate students as well as his associates; of the latter, especially Professors M.A. Amerine, M. A. Joslyn, and G. L. Marsh, who have continually stimulated his interest in the subject. The author also is very grateful for the critical reading of the manuscript by Mrs. Marjorie Barton Vaughn and Mr. H. W. Berg and is indebted to the many members of the library staff who so carefully traced obscure references. Errors or omissions that remain are the author’s responsibility.

REFERENCES 11. -4.1954. Composition of wines. I. Organic constituents. Advances in Research 6, 353-510. 11. A. 1955. Personal communication. 11. A., and Joslyn, M. A. 1940. Commercial production of table wines. Calif. Agr. E z p t . Sta. Bull. No. 639. Amerine, 11. A,, and Joslyn, I f . A. 1951. “Table Wines: The Technology of Their Production in California.” Univ. of California Press, Berkeley. Arena, A. 1936. Alteraciones bacterianas de vinos argentinos. Rev. agr. vet. (Buenos A i r e s ) 8, 155-320. BBchamp, 9. 1862. Sur les variations dans la quantiti. de certains principes imm6diats du vin, et sur les transformations que ces principes subissent par suite de certaines altirations spontan6es. Compt. rend. 64, 1148-1 152. Berry, J. M., and Vaughn, R. H. 1952. Decomposition of tartrates by lactobacilli (an abstract). Proc. Am. Soc. Enologists, pp. 135-138. Bioletti, F. T., and Cruess, W. V. 1912. Enological investigations. Cahj‘. A y r . E z p t . Sta. Bull. No. 230. de Bobadilla, G. F. 1943. Aplicaciones industriales de las levaduros de flor. Agricultura (Alladrid) 12, (133), 203-207. Brred, R. S., Murray, E. G. D., and Hitchens, A. P. 1948. “Bcrgey’s Manual of Drterniinative Bacteriology,” 6th ed. Williams and Wilkins, Baltimore. Uiiehi, IT.,and Deuel, H. 1954. n b e r die Lindstoffe fadenziehender Weine. Mitt. Gebiete Lebensm. u. Hug. 46, 222-229. Caillcau, It., and Chevillard, L. 1949. Teneur de quelqucs vins franCais en aneurine, riboflavine, acide nicotinique e t acide pantothhique. Ann. agron. 19, 277-281. Cardon, B. P., and Barker, 13. A. 1946. Two new amino-acid fermenting bacteria, Clostridium propionicunt and Diplococcirs glycinophilus. J. Bacteriol. 62, 629-634. Carles, P. 1891. Sur la caractbristique des vins de fique. Compt. rend. 112, 811-812. Carter, H.E. 1950. Kitrogen compounds of yeast. “Yeasts in Feeding, Proceedings of the Symposium” (S. Brenner, ed.), pp. 5-8. Garrard Press, Champaign, Ill. Castor, J . G. B. 1950. Biochemical events during vinous fermentation. Proc. Am. Soe. Enologists, pp. 21-37, 104. Castor, J . G. B. 1953a. B-complex vitamins of musts and mines as microbial growth factors. A p p l . Microbiol. 1, 97-102. Castor, J. G. B. 1953b. The free amino acids of musts and wines. I. Microbiological rstimation of fourteen amino acids in California grape musts. Food Eesearch 18, 139-145. Castor, J. G. B. 1953c. The free amino acids of musts and wines. 11. The fate of amino acids of musts during alcoholic fermentation. Food Research 18, 146-151. Amerine, Food Amerine, Amerine,

102

REESE H. VAUGHN

Castor, J. G. B., and Guymon, G. F. 1952. On the mechanism of formation of higher alcohols during alcoholic fermentation. Science 116, 147-149. Charpenti6, Y., Ribbreau-Gayon, J., and Peynaud, E. 1951. Sur la fermentation de l’acidc citrique par lrs bactbries malolactiques. Bull. SOC. chim. biol. 33, 13691378. Cheldelin, V. H., and King, T. E. 1953. Nutrition of microorganisms. Ann. Rev. Microbiol. 7, 113-142. Cruess, W. V. 1935. Tannin in wines. W i n e s and V i n e s 16, 5-7. Cruess, W. V. 1943. The role of microorganisms and enzymes in wine making. A d vances in Enzymol. 3, 349-386. Cruess, W. V. 1947. “The Principles and Practice of Wine Making,” 2nd ed. Avi Publishing, New York. Cruess, W. V. 1948. Investigations of the flor sherry process. Calif. Agr. E x p t . Sta. Bull. No. 710. Dal Cin, G. 1948. Gli antibiotici in enologia c loro particolari applicazioni. La Bioamicina come regohtore della fermcntazione. Eiv. viticolt. e en01 (Conegliano) 1, 335-340. Dal Cin, G. 1950. I derivati alogenati dell’acido acetic0 in enotecnia. Riu. viticolt. e enol (Conegliano) 3, 357-361, 387-393, 419-428. Dann, W. J., and Satterfield, G. H. 1947. “Estimation of the Vitamins.” Cattell Press, Lancastcr, Pa. Duclaux, E. 1874. Recherches sur les vins. Arm. chim. et phys. 2[51,289-324. Duclaux, E. 1901. “Traite de microbiologie,” Val. 4. Masson, Paris. Dunn, M. S. 1947. Amino acids in food and analytical methods for their determination. Food Technol. 1, 269-286. Federal Food and Drug Administ,ration. 1941. Monochloracetic acid an adulterant in foods. Trade Correspondence Letter No. 377, Washington, D. C., December 29, 1941. Felton, E. A,, Evans, J. B., and Niven, C. F., Jr. 1953. Production of catalase by the pediococci. J . Bacteriol. 66, 481-482. Fitz, A. 1879. Ueber Spaltpilzgahrungen. V. Mittheilung. Ber. 12, 474-481. Fornachon, J. C. M. 1943. Bacterial spoilage of fortified wines. Australian Wine Board, Adelaide. Fornachon, J. C. M., Douglas, H. C., and Vaughn, R. H. 1940. The pH requirements of some heterofermentative species of Lactobacillus. J. Bacterial. 40, 649-655. Fornachon, J. C. M., Douglas, H. C., and Vaughn, R. H. 1949. Lactobacillus trichodes nov. spec. A bacterium causing spoilage in appetizer and dessert wines. Hilgardia 19, 129-132. Gautier, A. 1878. Sur une maladie non encore diicrite des vins du midi de la France dits vins tourniis. Compt. rend. 86, 1338-1341. Gayon, U., and Dubourg, E. 1894. Sur les vins mannitks. Ann. inst. Pasteur 8, 108-16. Gayon, U., and Dubourg, E. 1901. Nouvellcs recherches sur le ferment mannitique. Ann. inst. Pastezcr 16, 526-569. Genevois, L., and Flavicr, H. 1938-1939. Dosage des vitamines B, et Bz dans quelques vins de la Gironcle. Proc. verb. sdan. SOC. sci. phv. nut. Bordeaux, pp. 72-73. Glenard, A. 1862. Note sur la fermentation tartrique du vin. L y o n SOC.Agr. Ann. 6, 141-160. Gunsalus, I. C. 1947. Products of anaerobic glycerol fermentation by Streptococcus faeccclis. J . Bactetiol. 64, 239-244. Gunsalus, I. C., and Sherman, J. M. 1943. The fermentation of glycerol by streptococci. J. Bacteriol. 46, 155-162.

BACTERIAL SPOILAGE O F WINES

103

Gunsalus, I. C., and Umbreit, W. W. 1945. The oxidation of glycerol by Streptococcus faecalis. J . Bacteriol. 49, 347-357 Hassid, W. Z., and Barker, H. A. 1940. The structure of dextran synthesized from sucrose by Betacoccus arabinosaceus, Orla-Jensen. J . Biol. Chern. 134, 163-1 70. Hehre, E. J. 1951. Enzymic synthesis of polysaccharides: a biological type of polymerization. Advances in Enzymol. 11, 297-337. Hehre, E. J., and Hamilton, D. M. 1949. Bacterial conversion of dextrin into a polysaccharide with the seriological properties of dextran. Proc. Soc. Exptl. Biol. M e d . 71, 336-339. Hendlin, D. 1954. The nutrition of microorganisms. Ann. Rev. Microbiol. 8, 47-70. Henneberg, W. 1898. Weitere Untersuchungen iiber Essigbakterien. Zentr. Bakteriol. Parasitenk. Abt. I I 4, 11-20, 67-73, 138-147. Henneberg, W. 1926. “Handbuch dcr Giiirungs-Bakteriologie,” T’ols. 1 and 2. Parry, Berlin. Hochstrasser, R. 1955. Cber einige Bedingungrn beim Lindrerden der Weine. Dissertation, Zurich, Ebner, Zurich. Hohl, L. A., and Cruess, W.V. 1939. Observations on certain film forming yeasts. Zentr. Bakteriol. Parasitenk. Aht. ZI 101, 65-78. Hoyer, D. P. 1898. Bijdrage tot de Kennis van de Azijnbactcrien. Dissertation, Tkden, Waltman, Delft. Hucker, G. J., and Pederson, C. S. 1948. “Bergey’s Manual of Determinative Bacteriology,” 6th ed., pp. 346-348. Williams and Wilkins, Baltimore. Hutchings, B. L., and Peterson, W. H. 1943. Amino acid requirement of Lactobacillus casei. Proc. SOC.Exptl. B i d . M e d . 62, 36-38. Joslyn, X. A., and Amerine, M. A. 1941. Commercial production of dessert wines. Calif. Agr. Ezpt. StJ. Bull. KO.661. Joslyn, hl. A., and Braverman, J. B. S. 1954. The chemistry and technology of the pretreatment and preservation of fruit and vegetable products with sulfur dioxide and sulfites. Advances in Food Research 6 , 97-100. Kayser, E., and Manceau, E. 1939. “Les ferments de la graisse des vins.” Henri Villers, Epernay. Kirlhofrr, E. 1953. Die Wirkung antibiotischrr StoEe auf die Weingiirung. Deutsche Weinzeitung. W‘ein 11. Rebe 89 (35), 638, 640, 642. Knight, B. C. J. G. 1945, Growth factors in microbiology. Some wider aspects of nutritional studies with microorganisms. T’itclmins and Hormones 3, 105-228; 228 a,b.

Koch, A4.1900. Uebar die Ursachen dcs Vcrschwinclcns der SBnrc bai Gahrung und Lag-erung des Weines. Tl’ein,btru 71. TVcinhandeZ. 18, 40. (Original not seen : cited by illullar-Thurgau and Ostrrwalder, 1919.) Koch, R. 1881. Zur Zuchtung von pathogenen Mikro-organismen. Mitt. Kaiserl. Gesundheitsamte 1, 1-48. Kramer, E. 1892. “Die Bakteriologie in ihren Bezicchungen zur Landmirtschaft und den landwirtschafts technischen Gewerben,” Vol. 2, C. C-erold’s Sohn, Vienna. Krumperman, P. H., Berry, J. M., and Vaughn, R. H. 1953. Utilization of tartrate by Lactobacillzis plantarum. Bacteriol. Proc., 1963, 24. (An abstract.) Kunz, R. 1901. Uber Vorkomnien und Bestimmung der Milchsaure im Weine. 2. Untersuch. Nahr. Genwsm. 4, 673-683. Lindner, P. 1927, 1928. “Atlas der Mikroskopischen Grundlagen der Garungskunde, Vols. 1 and 2. Parey, Berlin. T,iithi, H. 1949. ’iiber das Lindwerden der Weine und Obstweine. Schweiz. 2. 0bst.-u. Weinbau 68, 266-272.

104

REESE H. VAUGHN

Luthi, H. 1950. uber das Lindwerden der Weine. Schweiz. Z . 0bst.-u. Weinbau 69, 149-153, 165-169. Liithi, H. 1953, Beitrag zur Kenntnis fadenziehender (linder) Weine und Obstweine. Mitt. Gebiete Lebensm. Hyg. 44, 30-72. Luthi, H., and Vetsch, U. 1952. Papierchromatische Bestimmung von Aminosauren in Weinen. Schweiz. 2. Obst.-u. Weinbau 61,39-394, 405-408. Marcilla Arrazola, J., Alas, G., and Feduchy Mariiio, E. 1936. Contribucion a1 estudio de la levaduras que forman velo sobre ciertos vinos de elevado grado alcoholico. Anales centro invest. vinicol. 1, 1-230. Martin, L. 1948. Beitrag Zur Kenntnis der linden Weine. Dissertation, Ziirich, Neographik, Zurich. Mayer, H. D. 1939. Das “Tibi” Konsortium nebst einem Beitrag zur Kenntnis der Bakterien-Dissoaiation. Thesis, Delft. Mees, R. H. 1934. Onderzoskingen over de Biersarcina. Thesis, Delft. Moller, E. F. 1939. Das Wuchstoffe-System der Milchsaure-Bakterien. 2. physiol. Chem. 260, 246-256. Moslinger, R.1901. n e r die Sauren dcs Weines und den Sauerungsgang. Z. 7yntersuch. Nahr. Genussm. 4, 1120-1130. Miiller-Thurgau, H. 1891. u b e r die Ergebnisse neuer Untersuchungen auf dem Gebiete der Weinbereitung. Ber. XII. Deut. Weinbaukong (Worms), p. 128. (Original not seen; cited by Muller-Thurgau and Osterwalder, 1919.) Muller-Thurgau, H. 1908. Bakterienblasen (Bakteriocysten). Zentr. Bakteriol. Parasitenk Abt. ZI 20, 353-400, 449-468. Miiller-Thurgau, H., and Osterwalder, A. 1912. Die Bakterien im Wein and Obstwein und die dadurch verursachten Veranderungen. Zentr. Bakteriol. Parasitenk. Abt. ZZ 36, 129-338. Miiller-Thurgau, H., and Osterwalder, A. 1918. Weitere Beitrage zur Kenntnis der Mannitbakterien im Wein. Zentr. Bakteriol. Parasitenk. Abt. I I 48, 1-35. hluller-Thurgau, H., and Osterwalder, A. 1919. Ueber die durch Bakterien verursacht e Zersetzung von Weinsaure und Glyzerin im Wein. Landwirtsch. Jahrb. Schweiz 33, 313-361. Kickl&s,J. 1862. Sur le vin tourn6. Compt. rend. 64, 1219-1220. Niehaus, C. J. G. 1932. Mannitic bacteria in South African sweet wines. Farming in S. Africa 4, 443-444. Niven, C. F., Jr., Srniley, K. L., and Sherman, J. M. 1941. The polysaccharides synthesized b y Streptococcus salivarus and Streptococcus bovis. J . Biol. Chem. 140, 105-109. Nollner, C. 1841. Die Pseudo-Essigsaure. Ann. Chem. Pharm. 38, 299-307. Olsen, E. 1948. Studies of bacteria in Danish fruit-wines. Antonie uun Leeuwenhoek J . d4icrobiol. Serol. 14, 1-28. Orla-Jensen, S. 1919. The lactic acid bacteria. Kql. Danske Videnskab. Selskabs. Skrifter Naturvidenskab. math. Afdel 6[8], 81-196. Orla-Jensen, S. 1943. The lactic acid bacteria. Die echten Milchsaurebakterien. Erqanqungsband D. Kgl. Danske Videnskab. Selskab. Biol. Skrifter 2(3), 1-145. Orla-Jensen, S., and Faulenborg, G. 1940. Die fur das Wachstum der Milchsaurebakterien optimale Wasserstoffionenkonzentration. Zentr. Bakteriol. Parasitenk. Abt. II 102, 289-295. Orla-Jensen, S., Orla-Jensen, A. D., and Kjaer, A. 1947. On the ensiling of lucerne by means of lactic acid fermentation. Antonie van Leeuwenhoek J . Microbiol. Serol. 12, 97-114.

BACTERIAL SPOILAGE O F WINES

105

Orla-Jenscn, S., Otte, N. C., and Snog-Kjarc, A. 1936. Vitamin and nitrogen requircmcnts of the lactic acid bacteria. Kgl. Danske Videnskab. Selskabs. Skrijter Naturvidenskab. math. Afdel[9] Rackke 6(5), 1-52. Ostcrwaldcr, A. 1952. ijbcr die durch Bakterien verursachtc Zcrsctzung von Weinsaurc und Glyzcrin im Wein. Die Baktericn dcr Wcinstein-und Glyzerinzcrsctzung. Landwirtsch. Jahrb. Schweiz. 66, 181-197. Pastcur, L. 1866. “ fitude sur le vin.” L’imprimeric imperialc, Paris. Pastcur, L. 1873. “ e t u d e sur le vin.” 2nd cd. Savy, Paris. Pedcrson, C. S. 1938. The gas-producing species of the genus Lactobacillus. J . Bacteriol. 36, 95-108. Pcderson, C. S. 1939. Genus IV. Lactobacillus Beijcrinck. “Bcrgey’s Manual of Determinative Bacteriology,” 5th ed., pp. 362-378. Williams and W-ilkins, Baltimore. Pcdcrson, C. S. 1948. Genus I. Lactobacillus Bcijerinck. “Bergcy’s Xanual of Detcrminative Bacteriology,” 6th ed., pp. 349-364, Williams and Wilkins, Baltimore. Pcdcrson, C. S. 1949. The genus Pediococcus. Bacteriol. Revs. 13, 225-232. Pcrlman, L., and Morgan, A. F. 1945. Stability of B vitamins in grape juices and wines. Food Research 10, 334-341. Pcrquin, L. H. C. 1939-1940. On the incidental occurrence of rod-shaped, dcxtran producing bacteria in a beet-sugar factory. Antonie van Leeuwenhoek J . Microbiol. Serol. 6, 227-249. Peterson, W. H. 1950. Vitamins and minerals of yeasts. In “Ycasts in Feeding, Proceedings of the Symposium” (S. Brenner, ed.), pp. 26-33. Garrard Press, Champaign, Ill. Peterson, W. H., and Peterson, M. S. 1945. Relation of bacteria to vitamins and other groiyth factors. Bacteriol. Revs. 9, 49-109. Pcynaud, E. 1951. Experiences avec un nouvel activeur de la fermentation. Vignes et V i n s (16), 9-12. Pcynaud, E. 1952. Etude de l’inhibition de Saccharomyccs ccrevisiae par l’actidione. Cornpt. rend. 236, 1163. Pcynaud, E., and Lafourcade, S. 1953a. Etude d’antibiotiques et d’antiseptiques nouveaux actifs sur gcnrc Saccharomyces. Compt. rend. 236, 1924-1 925. Pcynaud, E., and Lafourcade, S. 195313. Action dc quelques antibiotiques sur les lcvures alcooliques. Rev. ferment. inds.-aliment. 8, 228-241. I’eynaud, E., and Lafourcadc, S. 1954a. Etudes rbccntes de la station oenologique dc Bordeaux sur l’application des antibiotiqucs en oenologic. 10th Congr. intern. inds. agr. aliment., M a d r i d , J u n e , 1954,pp. 1-10. (A preprint.) I’eynaud, E., and Lafourcadc, S. 1954b. Etudes r6ccntes de la station oenologique dc Bordeaux sur les facteurs dc croissance. 10th Congr. intern. inds. agr. aliment., Madrid, J u n e , 2954, pp. 1-10. (A preprint.) Porchct, V. 1931. Contribution A l’etude de l’adaptation des levures B l’acidc sulfurcaux. Ann. agr. Suisse 32(2), 135-154. Portcs, M. 1892. (No title given) J. Pharm. Chem. 26, 382-384. Pratt, R., Dufrcnoy, J., Sah, P. P. T., Oneto, J., Brodic, D. C., Riegleman, S., and Pickcring, V. L. 1950. Vitamin K as a n antimicrobial medicament and preservative. J. Am. P h a r m . Assoc. 39, 127-134. Rentschler, H. 1948. h e r das Lindwerdcn der Wcine und Obstwcinc. Schweiz. 2. Obst-u. W e i n b a u 67, 7-11. RibCreau-Gayon, J. 1936. Sur la dksacidification biologiquc dcs vins. Proc. verb. skan. SOC. sci. phy. nut. Bordeaux, pp. 23-25. Ribkrcau-Gayon, J. 1946. Sur la fermentation dc l’acidc maliquc dans lcs grands vins rouges. Bull. oflc. intern. vin 1946(182), 26-29.

106

REESE H. VAUGHN

RibEreau-Gayon, J. 1947. “Trait6 d’oenologie.” Beranger, Paris. Rib6reau-Gayon, J., and Peynaud, E. 1938a. La dCsacidification des vins par les bactCries. Compt. rend. acad. agr. France 24, 600-605. RibCreau-Gayon, J . , and Peynaud, E. 193813. Bilan de la fermentation malo-lactique. Ann. ferment 4, 559-569. RibEreau-Gayon, J., and Peynaud, E. 1952a. Sur l’emploi en vinification de quelques activeurs vitarniniques de la fermentation. Compt. rend. acad. agr. France 38, 444-448. Rib6reau-Gayon, J., and Peynaud, E. 195213. Action inhibitrice sur les levures de la vitamine K E et de quelques antibiotiques. Compt. rend. acad. agr. France 38, 479-481. Ribbreau-Gayon, J., and Peynaud, E. 1952c. Antiseptiques nouveaux et antibiotiques des levures alcooliques. Vignes et V i n s (22), 25-27; (23), 13. Rilibreau-Gayon, J., Peynaud, E., and Lafourcade, S. 1952a. Sur la formation de substances inhibitriccs de la fermentation par Botrytis cinerea. Compt. rend. 234, 478-483. RihCreau-Gayon, J., Peynaud, E., and Lafourcade, S. 1952b. Formation d’inhibiteurs et d’activeurs de la Fermentation alcoolique par diverses moisissures. Compt. rend. 234, 251-253. Roos: L., 1892. (KO title given) Ilfbms. Extraits Proc.-verb. sban. soc. sci. phy. nct. Bordeaux. S6ance de 28 July, 1892, pp. 59-63. Schanderl, H. 1926. Gntersuchungen iiber sogenannte Jerez-Hefen. W e i n u. Rebe 18, 16-25. Schanderl, H. 1950. “Handbuch der J(ellerwirtschaft,” Vol. 11. Ulmer, Stuttgart. Scheffer, W. R., and hlrak, E. K.1951. Characteristics of yeast causing clouding of dry white mines. Mycopalhol. et Mycol. A p p l . 6 , 236-249. Schultz, A. 1877. Cber das Umschlagen der Rotweine. Weinlaube 1877,KO.11. (Original not seen; cited by Nuller-Thurgau and Osterwalder, 1919.) Schweigert, B. S., and Snell, E. E. 1947. hliciobio og‘cal methods for the estimation of amino acids. Kutrition Abstr. & Revs. 16,497-510. Seifert, W. 1901. Cber die Saureabnahme im Wein und dcm dabci sich vollziehenden Garungsprozess. M i t t . gurungsphysiol. Lab. Klosterneuburg, 1901. (Original not seen; cited by Seifert, 1903.) Seifert, W. 1903. Cber die Siiureabnahme im Wein und dcm dabei stattfindenden Giirungsprozess 11. 2. landwirtsch. Versuchs. Deut. Oesterr. 6, 567-585. Semichon, L. 1905. “Maladies des vins.” Coulet, hlontpellier. Shimwell, J. 1,. 1940. Bacteria. in “Brewing Science and Practice” (H. C. Hind, cd.), Vol. 2, pp. 910-931. Wiley, New York. Shimwell, J . L. 1941. The lactic acid bacteria of beer. Wallerstein Labs. Communs. 4(11), 41-48. Shimwell, J. 1,. 1947. A study of ropiness in beer. J . Znst. Brewing 63, 280-294. Shimwell, J. L. 1948a. Brewing bacteriology IV. The Acetic acid bacteria (Family Acetobacterisceac; Genus Acetobacter). Wallerstein Labs. Communs. 11 (32), 27-39. Shimurell, J. L. 1948b. A rational nomenclature for the brewery lactic acid bacteria. J. Znst. Brewing 64 [n.s. 451, 100-104. Shimwell, J. I,. 1949. Brewing bacteriology VI. The lactic acid bacteria (family Lactobaeteriaceae). W a l l c r s t ~ i nLabs. Comrnuns. 12(36), 71-88. Shimwell, J. I,., and Kirkpatrick, W. F. 1939. New light on the “sarcina” question. J . Znst. Brewing 46, 137-1-25. Snell, E. E. 1945a. The microbiological assay of amino acids. Advances in Protein Chem. 2,85-118.

BACTERIAL SPOILAGE O F WINES

107

Snell, E. E. 194Sb. The nutritional requirements of the lactic acid bacteria and their application to biochemical research. J . Racteriol. 60, 373-382. Snell, E. E. 1916. Microbiological methods in amino acid analysis. Ann. A:. P. Acad. Sci. 47(2), 161-179. Snell, E. E. 1948, Xutritional requirements of the lactic acid bacteria. Wallerstein Labs. Co/n,nILns. 11(33), 81-104. Snell, E. E. 1950. Microbiological methods in vitamin research. in “\-itamin Methods” (1’. Gyiirgy, ed.), Vol. 1, pp, 327-421. Academic Press, Kew York. Snell, I:. I<., Strong, F. M., and Peterson, W. H. 1937. Growth factors for bacteria. VI. Fractionation and properties of a n accessory factor for lactic acid bacteria. Bl‘ochc~n.J . 31, 1789-1’799. i c and nicotinic Snell, F:. E., Strong, F. M., and Peterson, W.H. 1938. P a n t o t h e ~ ~acid acids as growth factors for lactic acid bacteria. J . Am. Chem. SOC.60, 2825. State of California, Department of Public Health. 1946. Regulations establishing standards of identity, quality, purity, sanitation, labeling and advertising of wine. Adopted May 23, 1942, and amended as of May 8, 1946. An excerpt from the California Administrative Code, Title 17, Public Health, 385. San Francisco. Stokes, J. L. 1952. Nutrition of microorganisms. Ann. Rev. Microbiol. 6 , 29-48. U. S. Treasury Dept., Bureau of Internal Revenue. 1948. Regulations KO.4 relating to laheling and advertising of wine. U. S. Govt. Printing Office, Washington, 1).C . Van Niel, C. B. 1928. The propionic acid bacteria. Dissertation, Technische Hoogeschool, Boissevain and Co., Haarlem. Vaughn, It. H. 1938. Sonic effects of association and competition on Acetobacter. J . Rncforiol. 36, 357-36’7. Vaughn, 11. 11. 1942. The acetic acid bacteria. T4-allerstein Labs. Commiins. 6(14),5-26. Vaughn, R . H. 1949. The Bacteriology of Wines. Review of the Meaning and Possiblc Causes of “Maladie de la Tourne.” C’nio. Cal
108

REESE H. VAUGHN

Voisenet, E. 1911b. Consideration nouvelles sur la maladie de l’amertume des vins dans scs rapports avec la fermentation acrylique de la glyckine. Campt. rend. 163, 898-900. Voisenet, E. 1913. Nouvelles recherches sur un ferment des vins amers. Compt. rend. 166, 1181-1182. Voisenet, E. 1918. Sur une bacterie de l’eau vegetant dans les vins amers capable de deshydrater la glyckrine. Glycero-reaction. Ann. inst. PasteuT 32, 476-510. Vosti, D. C., and Joslyn, M. A. 1954a. Autolysis of baker’s yeast. A p p l . Micrabiol. 2, 70-78. Vosti, D. C., and Joslyn, M. A. 195413. Autolysis of several pure culture yeasts. A p p l . Microbial. 2, 79-84. Walters, L. S. 1940. A note on the characters of a coccus isolated from South Australian stout. J . Inst. Brewing 46, 11-14. Watt, B. K., and Rlerrill, A. L. 1950. Composition of foods-raw, processed, prepared. U . S. Dept. Agr. Handbook 8, 1-147. Wood, H. G., Geiger, C., and Werkman, C. H. 1940. Kutritive requirements of the heterofermentative lactic acid bacteria. Iowa State Call. J . Sci. 14, 367-378.

Microbiological Implications in the Handling. Slaughtering. and Dressing of Meat Animals

BY JOHN C . AYRES Food Processing Laboratory. Iowa Agricultural Experimental Station. Iowa State College. Ames. Iowa

Page 110 111 111 1 . Primary Lines of Defense . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . a . Skin and Mucous Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 b . Hair and Cilia . . . .............................. 111 c. Gastric Juices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 112 d . Digestion in the Intestines . . . . . . . . . . . . . . . . . . . . . . . . . . e . Urine . . . 112 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Inflammatory ....................................... 112 3 . Humoral Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 111. Ante-Mortem ........................................ 114 s, Hoofs, and Hair . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 a . Microbial Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 b . Water Spraying of the Live Animal . . . . . . . . . . . . . . . . . . . . . . . 114 115 c . Effect of Climate on Microorganisms . . . . . . . . . . . . . . . . . . . . . . . ............................... 115 2 . Microorganisms in the Gut . . . . a. Rumen Bacteriology . . . . . ............................. 116 b . hlicroorganisms in the Intestines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 3 . Microorganisms in the Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 4 . Effects of Fatigue on Spoilage . . . . . . . . . . . . . . . . . . . . . . . 120 IV. Slaughter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 1 . Killing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 2. Stick-Knife . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 3 . Scald Tank . . . . . . .................................. 127 V . Post-Xortem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 1 . Skinning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 2 . Dehairing, Shaving, arid \Vax Dipping . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 3 . Evisceration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 4 . Chill Room . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 140 a . Off-odor and Slime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . b . Ham Souring, Bone Stink, or Bone Taint . . . . . . . . . . . . . . . . . . 141 c . Black Spot, Whiskers, and Other Mold Discolorations . . . . . . . . . 142 d . F a t Rancidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 143 5 Cutting and Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Defensive Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

110

J O H N C. AYRES

VI. Improvements in Processing Practices. . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Desired Additional Studies.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Suggested Improvements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page 149 149

151 154

I. IKTRODUCTION Meat provides many, if not all, of the nutrients desired by most microorganisms. Consequently, the development of satisfactory methods t o prevent or retard spoilage of meats is a very serious problem. Measures t o minimize bacterial invasion and growth on and in animal tissue must begin prior t o slaughtering the animal and care must be exercised in handling the carcass throughout dressing operations. Whether or not the load of microorganisms associated with the living animal contributes a significant share of the contamination occurring in and on the carcass depends not only upon the methods of handling that the meat receives but is determined also by the interrelation between the defensive mechanisms of the animal and the enormous microbial populations which gain access t o the animal. With these considerations in mind i t seems appropriate, then, t o discuss the bacteriological implications in four separate categories. The first of these deals with the animal’s defensive mechanisms ; the second considers the antemortem microbiological problems t h a t are involved in the live animal. The third describes the contamination introduced during the death agonies; while the last relates the contamination of the animal after death. Enormous numbers of organisms are associated with the hide, hoofs, and hair, with the gut, and with excretions voided by the animal. I n the healthy animal only occasional bacteria penetrate the gut and enter the blood stream and tissues; the fate of these organisms that survive within the flesh is of no small consequence. There are several important sources of contamination during slaughter which might not necessarily be considered as post-mortem phenomena even though the opportunities for invasion are never encountered under normal conditions by the healthy animal. One of these is the contamination which may be ascribed t o the stick-knife; the other is t h a t which may be encountered through use of a scald tank. After the animal dies and the shackled carcass is hoisted for cleaning and dressing, there no longer is a n equilibrium between the defense mechanisms of the animal and the organisms gaining access t o the tissues. Much of the flora that comes in contact with the carcass during evisceration is derived from workmen, from cutting tools, and from water and air in the dressing, cooling, and cutting rooms. After the completion of the dressing operation, the storage life depends on whether or not these

MICROBIOLOGY O F MEAT ANIMALS

111

organisms find the nutrients, humidity, and temperature of their new found residence satisfactory. This review will attempt t o discuss the significance of the several sources of microbiological contamination outlined above. The discussiori mill be directed largely t o the microbiology of the intact carcass and of fresh wholesale cuts. Owing t o limitations of space, studies which relate t o salting, curing, and smoking and t o the bacteriological quality of sausage products or fats will not be considered. Also, no attempt will be made t o include the bacteriological implications of the infected animal or of imperfections in sanitation generally recognized as such by the industry as a whole. Only those avenues of infection and contamination which treat with normal animals and carcasses and cuts receiving handling procedures in accordance with commonly accepted sanitary practices will be considered in this report.

11. DEFENSIVEMECHANISMS

I. Primary Lines of Defense a. Skin and Mucous Membranes. I n the living animal the skin acts largely in a mechanical way t o prevent the introduction of foreign agents such as bacteria and other antigens. It is aided in its action by secretions of oil and sweat. The oral and intestinal mucosa, as well as other mucous membranes lining the alimentary, respiratory, and genitourinary tracts also serve as barriers t o prevent the entrance of microorganisms into the deep tissues. However, these obstacles are not impenetrable t o all microbes ; hair follicles, oil glands, and sweat glands are particularly vulnerable. Also, most exposed areas of the normal skin and mucous membranes support a microbiological flora adapted t o the environmental conditions indigenous t o the area. From time t o time, pathogenic members of this resident population may initiate infection by passing into or beyond these lines of defense through a n injury or defect in the skin or membrane. Secretions, such as saliva, nasal secretions, sweat, tears, oil, and gastric secretions, serve t o reduce internal contamination. I n ruminants there is a copious flow of saliva during the time of food ingestion. In the saliva of some animals as in certain other secretions such as nasal mucus and tears, a n enzyme called lysozyme (Fleming, 1922) is present (Goldsworthy and Florey, 1930) which hydrolyzes a mucopolysaccharide (Hartscll, 1948) in the cell walls of many gram-positive bacteria (Dickman and Proctor, 1952). This fluid (Fraenkel-Conrat, 1950; Kern et al., 1951) not only lubricates the food and aids digestion but has protective action as n-ell. b. Hair and Cilia. Hair, particularly that lining the intricate nasal

112

JOHN C . AYRES

passages, entraps many of the microorganisms before they have an opportunity to invade the respiratory tract. The waving action of cilia moving in unison sweeps bacteria lodged on dust particles from the finer branches in the trachea and bronchi into the back of the mouth or pharynx, and the foreign agents are removed from the body during slavering, coughing, snorting, and sneezing. c. Gastric Juices. When food and water are ingested, microorganisms associated with these substances pass into the stomach where they are subjected to the action of acidic gastric secretions. Their numbers diminish rapidly in the stomach and very few bacteria are found in the materials emptied from the stomach into the small intestine. The stomach enzymes, rennin and pepsin, which aid in the digestion of protein foods, may have a slightly inhibitory function on microorganisms. d . Digestion in the Intestines. The powerful enzymes secreted by the glands lining the walls of the small intestine, together with those introduced from the pancreatic and bile ducts, not only are able to complete carbohydrate digestion but attack lipids and protein as well. Microorganisms which have succeeded in “running the gauntlet ” of enzymes in the saliva and the acidity of the stomach must, in order to survive, be refractive to the bile salts and to the ropy, viscid, alkaline bile which enter the small intestine from the liver. According to Sarles e2 al. (1951) the efficacy of these barriers may be demonstrated by the fact that 1000 to 100,000 times more cells of certain Salmonella are needed to cause infection when introduced by way of mouth than when injected into the peritoneal cavity. The lower small intestine and the colon contain enormous numbers of bacteria which ordinarily are held in check by the thick mucous membranes lining the intestines and by phagocytic mechanisms. Materials introduced into the gastrointestinal tract still remain outside of the physiological interior of the body proper. Consequently there is little concern for these agents unless they are able to gain entrance parenterally. e . Urine. The genitourinary tract is protected against most microorganisms by thick mucous membranes, by leucocytes, by lysozyme, and by the flushing action of urine. Urine normally contains few, if any, bacteria. 2. Inflammatory Processes If and when organisms penetrate the deep tissues, a number of devices operate to localize the invasion. Inflammation is a response to a local injury such as is commonly caused by bacteria or other antigens. The blood supply to the infected area is augmented by a dilatation of the arterioles and capillaries so that fluids diffuse into the tissue space. Cells

MICROBIOLOGY O F MEAT ANIMALS

113

primarily associated with defense against invading microorganisms are widely distributed through the animal in the blood and lymph, cartilage, bone reticular (blood-forming) tissue of the myeloid and lymphatic organs, and in the loose connective tissue associated with the skin, omentum, liver, lungs, etc. Cellular response t o a n inflammatory stimulus is characterized by a n initial migration of short-lived polymorphonuclear leucocytes from the blood vessels t o the point of injury where they inactivate the invading antigens (Menkin, 1940). More important t o local defense are the lymphocytes and monocytes which also migrate from the blood and lymph but are long-lived and may be transformed t o macrophages. These, together with “fixed macrophages” (already present in the area) digest the bacteria. When large amounts of antigenic agents are present, the macrophages may fuse t o form foreign body giant cells which may effectively wall off the invasive agent. If the local inflammatory reaction is not sufficiently prompt to eliminate the microorganisms before they have had an opportunity t o multiply, they will spread through the lymphatic capillaries and may reach the lymph nodes which drain the region. These sites of contamination are considered by Lepovetsky et al. (1953) t o be the locale of most of the bacteria in the deep tissues. The lymph nodes, comprised essentially of a network of connective tissue, act as filters and collect many bacteria which are then engulfed and digested by the “fixed” macrophages. If the microorganisms entering the lymphatic vessels are too numerous or invasive, some may escape destruction and continue t o multiply, causing a swelling and soreness of the node. Bacteria also are effectively screened out by several organs, particularly the spleen, lungs, and liver. For example, when Petersen et al. (1927) injected 5-10 million cells per min. of B . coli ( E . coli) intravenously into the dog, the majority of these were removed from circulation right up t o the time of maximal injury (sometimes for as long as 12 hr.). 3 . Humoral Antibodies

The process of inflammation, the collecting and screening out of bacteria by the lymphatics and in the spleen and liver, and the digestion of bacteria by phagocytic cells are normal defensive mechanisms. I n addition t o these reactions there are others which depend upon previous contact with the infectious agents; these latter, called humoral antibodies, provide mechanisms of resistance t o infection. The presence of an infection or prior contact with the infectious agent tends t o accentuate the effectiveness of these defensive agents. According t o Burrows (1949) the cooperative role of the humoral antibodies with the phagocytic cells is of considerable significance ; a n antigen is localized by agglutination, if cellular,

114

JOHN C. AYRES

or by precipitation, if in solution, and in either case the material is made more readily phagocytable by opsonization. 111. ANTE-MORTEM 1. Microbiology of Hides, Hoofs, and Hair a. Microbial Loads. Although the intact skin and mucous membranes present a barrier which most microorganisms are unable t o penetrate, bacteria remaining on the hides, hoofs, and hair at the time of slaughter constitute an important source of contamination of the meat surface and deep tissues. For this reason, the resident and adventitious flora of the animal are of great concern. Empey et al. (1934) and Mallman et al. (1940) indicated that flora developing on cold-storage beef were largely soil organisms deriving from the hoofs and hides of the animals. Empey and Scott (1939a) and Jensen and Hess (1941) made extensive studies of the numbers and types of microorganisms encountered on various areas of unwashed hides and skins of cattle and hogs, respectively. Empey and Scott (1939a) found from 100,000 t o 31 million bacteria growing aerobically per sq. cm. of surface on the unwashed hides of cattle whereas Jensen and Hess recovered from 100,000 to 1.5 trillion aerobes and from 10,000 to 2 trillion anaerobes on 2 sq. in. of neck skin of hogs in the area where the jugular vein is cut. Mean numbers per sq. em. for the two studies were 3.3 and 91 million aerobes, respectively. Mean numbers for the anaerobes were 110 million. Also, the mean counts of yeasts and molds were 580 and 850, respectively. b. Water Spraying of the Live Animal. Empey and Scott (1939a) considered t h a t a reduction in the initial contamination was brought about by subjecting the animals t o a forced spray of cold water over the whole area of the hide. However, they did riot think that the total populations present on the hides of cattle were altered appreciably as the result of exposure t o the light showering with cold water used prior t o the stunning of the animals. Also, they stated that the efAuent water supplied a foot bath about 12 in. in depth through which the animals were driven before being sprayed and this bath facilitated the removal of much of the soil carried on the hoofs. Mean microbial populations of the water used t o wash the hides were found t o be 330,000 per ml.; between 20 t o 25 gallons of water were used for spraying each animal and the net effect was considered equivalent t o the reduction of about one-half of the total initial populations of the hides. Since the adding of moisture was found t o favor the proliferation of the microflora of the hair and soil, spraying was done immediately prior t o the slaughter of the animal. According t o Empey and Scott (1939a) considerable variation occurred in the extent of contamination of hides from different animals and also in that of various areas of

MICROBIOLOGY OF MEAT ANIMALS

115

hide from the same animal. The explanation for these differences is not precisely known, but they are probably influenced by such factors as moisture, amount of available food material, and the extent of contamination acquired by the hides through contact with soil. It has been demonstrated by means of small-scale experiments that the microbial populations of relatively dry areas of hide may be increased from five t o ten-fold as the result of the addition of sufficient moisture, and the subsequent maintenance of these samples at temperatures ranging between 20" C. (68" F.) and 27" C. (80.6" F.). Empey and Scott (1939a) suggested the use of chlorine solutions as germicidal agents approximately 15 min. before the knocking of the animals. They had not performed any experiments other than the immersion of bits of hide attached t o feet and subsamples of hair for short times. Under these conditions mean per cent kills for .025, 0.05, 0.10, and 0.15% chlorine were 82, 90, 94.3, and 95.3, respectively. c. Effect of Clamate on Microorganisms. I n the subtropical regions, beef prepared in the summer period will have a potential storage life a t least several days in excess of beef prepared in the winter period. Similarly, beef prepared in tropical climates may be expected t o have a potential life some two or three weeks in excess of beef prepared in temperate zones. Workers in Australia (Empey and Scott, 1939a,b) found not only that the numbers of "soil-type bacteria diminished in tropical soils when compared with those growing in the temperate zones but that the numbers found on the skins of cattle and on meat followed similar trends. In essence, they considered meat from animals living in hot climates t o be better suited for cold storage than that from beasts living in regions having more moderate temperatures. They reported that beef prepared in tropical zones may be expected t o have a potential storage life some two or three weeks in excess of beef prepared in temperate zones. Empey and Scott (1939a) were of the opinion that dirt is a n important source of the contaminants t h a t bring about spoilage of meat during storage. )'

2. Microorganisms in the Gut

Prior t o birth, animals are considered t o be maintained under germfree conditions (Reyniers, 1943, 1946; Reyniers, et al., 1946). As early as 1895, h'uttall and Thierfelder demonstrated that i t is possible t o rear animals free from bacteria by adopting adequate precautions. Much later, Reyniers and his associates (1946, 1949) succeeded in raising numerous germ-free animals including fish, chickens, and pigs. Ordinarily, a bacterial flora becomes established in the digestive tract

116

JOHN C. AYRES

a few hours after the animal is born and, from t h a t time throughout life, the animal continues t o be a n involuntary host and possibly a symbiont with countless numbers of microorganisms. Xutritionists nom tend t o regard intestinal microorganisms as sources of nutrilites indispensable t o the animal host. Great numbers of many different varieties of bacteria are swallowed with food and water. It is t o be expected that, since the dietary of the herbivores differs from t h a t of the omnivores and carnivores, the numbers and kinds of microorganisms that are ingested among the several domestic animals will vary. a. R u m e n Bacteriology. Herbivorous animals are dependent t o a considerable extent upon the action of the rumen bacteria in the digestive process. The animal host must rely on its symbiotic microorganisms t o carry out the essential predigestion. The cellulose of plant cells effectively imprisons the stored food of the cells until i t is degraded by bacterial action. Insofar as is presently known the orginary enzymes of the digestive tract of ruminants and, for that matter of most mammals, do not include the enzyme cellulase. Lohnis and Kuntze (1908) estimated 0.074-11.6 million bacteria per gram of straw, and these authors noted a considerable decrease in bacterial population when the straw was dried and preserved in clean surroundings. I n hay, one group of workers (Duggeli, 1906) reported 10-400 million bacteria per gram whereas, in another study, Gutierrez (1953) obtained counts of 300-600 million organisms having colony appearance similar t o those of identified Propionibacterium. The types of organisms carried on hay and straw include thermophilic bacteria, sporeforming aerobes, butyric-acid bacteria, cocci, members of the genera Pseudomonas and Achromobacterium, Actinomycetes, and various higher fungi. The digestive habits of the animal also have marked influence on the numbers and types of microbial flora which persist in the gastrointestinal tract. Food of the ruminant animals (cattle, sheep, goat) is swallowed without sufficient mastication and later is regurgitated in small quantities and thoroughly chewed. According t o Carroll and Hungate (1954), who quote and extrapolate values for a number of workers (Sisson, 1914; Winogradowa-Fedorawa and Winogradow, 1929; Hoflund, 1940; Elsden rt al., 1946; and Blamire, 1952), 70 kg. would represent the average weight of rumen contents in a 500 kg. cow. This food is soaked by saliva and then taken into the rumen where i t is milled by repeated contractions of the walls of the rumen, and fermented by protozoa and bacteria. Following the ingestions of feed, the number of microorganisms in the rumen of healthy adult animals increases rapidly within 2 hr. after eating (Bortree et al., 1946). There high counts are maintained or increased for

MICKOBIOLOGY O F M E A T A N I M A L S

117

several hours and then gradually return t o the range observed prior t o feeding. As much as 12 t o 24 hr. may elapse between the time that the food is first taken into the rumen and the time it passes back into the mouth as cuds t o be thoroughly ground by the molar teeth before passing into the reticulum, psalterium, and abomasum (the true stomach). According t o Grant (1953) it has been estimated that, a t the peak of the digestion cycle, the bacteria and protozoa constitute approximately 10% of the volume of rumen material. Grant states that it seems likely that the nitrogen transfer from all sources within the rumen t o the protoplasm of microbic organisms is sufficient t o meet the protein requirements of the ruminant. Some investigators suggest that the cow is not only nourished directly by the grasses and grains that i t ingests, but in part a t least, by the microorganisms and the products that they produce from the feeds eaten by the cow. Empey and Scott (1939a) made observations of the microbial contents in the bovine rumen and found mean populations on a dry weight basis of 53 million bacteria, 180,000 yeasts, and 1,600 molds per gram. Gutierrez (1953),found numbers of organisms varying from250,OOOt o 7 billion per ml. of rumen fluid, with a n average of 1.3 billion. Although there is not complete agreement, the nature of the food is considered by some workers t o be an important factor in establishing the flora of the digestive tract. The type of microorganisms present in the rumen of animals fed a ration high in roughage feeds differ noticeably from that found in the rumen if the ration is rich in grains (Pounden and Hibbs, 1948; Masson, 1950; Gall et al., 1953). Further confirmation of these relationships has been demonstrated with rumen microorganisms in vitro (Pearson and Smith, 1943; Arias et al., 1951). Bortree et al. (1946) reported that changes in counts were not great when animals were changed from hay t o pasture, but when 3 lb. of glucose was given in addition t o the usual feeding of hay, counts were 100% greater than those observed for a ration of hay alone. Pounden and Hibbs (1948) observed a drastic change in the rumen microbiology of dairy calves when grain instead of hay was used in the ration. Moderate numbers of protozoa and flora of the varieties associated with hay accompanied the ingestion of hay without grain. On the other hand, limited numbers of protozoa were accompanied by great numbers of bacteria of the grain groups, but no organisms of the hay groups were present when the ration consisted of almost all grain. With sheep, Masson (1950) found that the bacterial types mere few and their numbers and size were small. In sheep that were fed hay and moderate amounts of concentrates, the populations were rich in starch-digesting sarcinae, diplococci, and streptococci; the protozoa were large in size. Microorganisms of the rumen of

118

JOHN C. AYRES

these animals were largely disintegrated in the abomasum. I n sheep fed a large proportion of flaked maize, protozoa and yeasts were absent; grampositive spore forming rods were predominant (Clostridium butyricum) . Gall et al. (1953) found that rumen, flora, and environment are closely related. The two basically different types of ruminant feeds-high grain rations, and high roughage rations-produce characteristic rumen flora patterns. High roughage rations were found t o encourage obligate anaerobic bacterial action on fiber and t o produce propionic, butyric, and acetic acids and some gasses. On the other hand the rumen bacteria of animals fed a high grain ration were of less strictly anaerobic type, most of which acted on readily available carbohydrates t o produce lactic acid. Whereas the p H of the rumen of animals fed on roughage was near neutrality, that of high-grain-ration animals was distinctly acid. 0. &ficroorganisms in the Intestanes. I n general, the majority of the bacteria are destroyed in the stomach. In some cases, however, when the material does not remain long in the stomach, the bacteria may escape destruction and pass further clown the canal. I n the lower small intestine, bacteria grow abundantly so that the microbial population increases progressively until, in the colon and rectum, enormous numbers are present. From five observations made of freshly voided bovine feces, Empey and Scott (1939a) reported 90 million bacteria, 200,000 yeasts, and 60,000 molds. Lissauer (1906) estimated that the feces of cattle contained 15-19% by weight of bacteria, corresponding in the case of the cow t o about 60-80 million bacterial cells per milligram of dried feces. Not all of these bacteria are viable. Hutterman (1905) found about one billion bacteria in one ml. of the intestinal contents of the ox. Gruber (1909) estimated that manure from stall-fed cattle contained 1-120 million bacteria per gram whereas that from cattle on pasture was said t o contain 1-4 million bacteria per gram. 3. Microorganisms in the Tissues

Early workers were not in agreement regarding the source of tissue contaminants. One school of thought, largely the German (von Fodor, 1886; Hauser, 1886; Wyssokowitsch, 1886; Neisser, 1896; Opitz, 1898; Messner, 1910; Amako, 1910; Thole, 1912; Haas, 1922), contended that the organs and musculature of healthy animals were free from microorganisms whereas other workers (Adami et al., 1899; Carriere and Vanverts, 1899; Ford, 1901; Boni, 1901; Nicholls, 1904; Norris and Pappenheimer, 1905; Selter, 1906; Conradi, 1909; Wolbach and Saiki, 1909; Bierotti and Machilda, 1910; Horn, 1910; Bugge and Kiessig, 1911; Zwick and Wiechel, 1911; Grunt, 1912; Savage, 1918, 1920; Spray, 1922; Boyer, 1926; Reith, 1926) indicated that bacteria commonly were con-

MICROBIOLOGY OF MEAT ANIMALS

119

taminants of animals that were considered normal. According t o Bulloch (1930) many of the early studies led t o faulty conclusions owing to the use of defective methods. More recently, the data in the literature support the concept that a few bacteria may be present in certain of the living tissues (Boyer, 1926; Reith, 1926; Burn, 1934a,b). However, there is controversy concerning the portal of entry for these contaminants. Some workers claim that the microorganisms which gain entrance t o tissues penetrate the intestinal wall and are carried t o various parts of the body by the blood. Desoubry and Porcher (1895) were of the opinion that feeding aided the passage of bacteria into the blood stream (of dogs), particularly if the meat cont>ainedmuch fatty substances. According t o Haines (1937) Cobbett and Graham Smith (personal communication) did not approve of Desoubry and Porcher’s techniques, but did find that when guinea pigs were fed wet cabbage, bacteria entered into the organs and muscle and the animals had diarrhea. I n another early study Nocard (1895) concluded that the serum of fasting horses was almost always sterile whereas that collected from horses shortly after they had been fed often contained bacteria. Ficker (1905) starved rabbits, dogs, cats, and mice from 3 t o 17 days and then fed them with infected food which contained a n easily identified organism. Four hours after the animals had ingested the contaminated food, they were sacrificed. The bacteria were found in the blood and organs. Ficker concluded that prolonged starvation decreased the resistance of the gut t o invasion. Griffith (1911), using monkeys, pigs, goats, dogs, and cats, showed that tubercle bacilli were able t o pass through the mucous membrane of the alimentary tract and t o reach the adjacent lymphatic glands within a few hours of their ingestion. Only a small proportion of the bacilli ingested were able t o penetrate the mucosa, the rest passing out with the feces. Of those t h a t did penetrate, the majority mere arrested in the glands but some, after a period the duration of which appears t o vary with the animal used, were carried into the lungs or even beyond; the shortest period of time between their ingestion and their demonstration in the lungs was 4 days. Hulphers (1934) described several organisms found in the lungs of slaughtered hogs. He indicated that most of the microorganisms appear t o be soil flora. Tarozzi (1906a,b) and Canfora (1908) injected spores of Clostridizim tetani into animals. After a considerable length of time (up t o 55 days) a leg or bone of each animal was broken or injured and the animals were observed for tetanus which usually occurred following this treatment, showing the longevity of spores in the normal body. Jensen and Hess (1941) conducted extensive studies t o support their contention that the

120

JOHN C. AYRES

muscle tissue and bone marrow of living hogs are generally sterile but that many kinds of bacteria can be isolated from these sites after the animals have been slaughtered. A study conducted by Adamson (1949), although made with human cadavers, has direct bearing on the subject of tissue invasion. His study revealed that a large proportion of 804 lymph nodes removed from human corpses contained a variety of bacteria, with coliforms, micrococci, and streptococci predominating. He suggested that the bacteria found in the nodes might have gained access to the body through skin abrasions, became localized in the nodes and survived there for some time. Lepevetsky et al. (1953) obtained samples from 11 chucks and 12 rounds of beef from 23 cattle. The prescapular lymph node, the humerus, and muscle tissue bordering that bone were removed from each check while the popliteal lymph node, the femur, and neighboring muscle tissue were taken from each round. Bacteria were isolated from 15 of 23 lymph nodes, from 3 of 23 marrow samples, and from 2 of 23 muscle samples. The numbers recovered from infected lymph nodes ranged from 80-764,000 per gram whereas only 2 samples each for bone marrow and muscle tissue showed a sufficient number of organisms to be counted. Bacteria isolated from the various samples represented the following 12 genera : Aerobacter, Alcaligenes, Bacteroides, Clostridium, Corynebacterium, Escherichia, Flavobacterium, Micrococcus, Proteus, Pseudomonas, Serratia, and Streptococcus. (Of the 93 organisms isolated, 31 were classified in the genus Streptococcus, 18 in Escherichia, 10 in Aerobacter, and 8 in Pseudomonas.) I n comparing their results with those of Jensen and Hess (1941) and of Adamson (1949), the Ohio workers considered that the morphology of the flora studied in their investigation had similarities t o those from human lymph nodes and that the muscle tissue and bone marrow had more freedom from bacteria than t h a t which was removed from the slaughtered hogs as reported by Jensen and Hess. Further, VC'eiser et al. (1954) considered that the results confirmed observations on sour rounds that the spoilage obtained appears t o propagate from lymph nodes.

4. Effects of Fatigue o n Spoilage It has been long recognized that fatigue exerts a pronounced effect upon the defensive mechanisms of animals. Charrin and Roger (1890) found that exercise in a revolving drum increased the susceptibility of rats to anthrax and quarter evil (blackleg). Ficker (1905) did not believe that fatigue alone facilitated the passage of bacteria into the tissues from the gut of animals but that fatigue in conjunction with starvation promoted their entrance from the intestines. He postulated that decreased

MICROBIOLOGY O F MEAT ANIMALS

121

gastric secretion, diminished peristalsis, and acclerated flow of blood and lymph during muscular exercise, aided the penetration of bacteria, and Ficker suggested that this would explain the observation that the flesh of animals slaughtered after having been driven long distances decomposes very quickly. I n his opinion, fatigue helped t o facilitate the passage of organisms through the walls of the gastrointestinal tract more rapidly than would have been the case from hunger alone. Boycott and Price-Jones (1926), working with considerable numbers of rats, found t h a t fatigue usually led t o a lowering of the rectal temperature by 2.8 t o 8.2" C. (5 t o 15" F.) and, occasionally, by more than 11.1" C. (20" FA).They used as a n infecting organism, Gaertner's bacillus (S'almonella enteritidis), a natural parasite of the rat. These workers concluded that, after feeding, S. enteritidis found its way into the spleen of both normal and fatigued rats; however, fatigue increased illness and mortality. Boycott and Price-Jones did not believe that the permeability of the intestine t o ingress of microorganisms was dependent upon fatigue, a s indicated by Charrin and Roger and by Ficker, since almost all of the controls (only 1 of 28 of which was ill) as well as the survivors (14 of 27) of the fatigued animals had S.enteritidis in their spleens when they were killed at the end of the experiment. However, they felt that their obserrations clearly demonstrated that, under certain circumstances, infection was promoted by fatigue. Meat packers have learned that, in order t o maintain high quality in meat, it is necessary t o delay the slaughter of animals when they are fatigued, hungry, or thirsty; also, it has been long known that animals killed after they had become exhausted during a hunt, in fighting, or when struggling violently during slaughter undergo rigor mortis early and putrefy rapidly (Jensen, 1915). Various explanations have been offered to account for this phenomenon. One has been previously stated in this paper (i.e., that fatigue decreased the resistance of the gut t o invasion with subsequent entry of the organisms in the deep tissues); a second is that blood is retained in the vessels when the animal goes into shock. For example, Morrison and Hooker (1915) found that the outflow of blood from the perfused organs of a shocked animal was less than that from similar organs under normal conditions. Further discussion relating t o the efficiency of bleeding in the shocked or dying animal is given in a later section of this article. A third factor to be considered is the modification of globulin. Antibodies are thought t o be modified serum globulins. It may be of interest t o note that Burrows (1949; pp. 307-8) states that "there is a feeling on the part of a number of competent investigators that the serum proteins including globulin do not exist as such in the body

122

JOHN C. AYRES

but are artifacts arising as a consequence of laboratory manipulation.” Still another factor involving fatigue as a causal agent for putrefaction refers t o the ultimate’ p H and keeping quality of the carcass. The relation of ante-mortem condition to post-mortem changes in glycogen and lactic acid content of muscles was reviewed by Bate-Smith (1948). According to Bate-Smith (1938b, 1939) the pH of muscle steadily falls from the moment that circulation stops; the precise pH of the muscle immediately prior to that time depends on the recent history of muscular activity. If the muscles have been completely inert and well supplied with oxygen before death, the lactic acid content of the tissue is very low. If the animal is given time to rest, the lost muscle glycogen is restored according t o Callow (1936) and Callow and Boa2 (1937). The p H of muscle with minimal lactic acid is about pH 7.4 (Bate-Smith, 1 9 3 8 ~ ) . Since the glycogen content is approximately 10/0,the lactic acid formed is about 1.1% and the pH reached in full rigor about p H 5.6 (Bate-Smith, 1948). However, if the muscle is exercised (fatigued) shortly before the animal is slaughtered, much of its glycogen reserve has been converted to intermediates of the glycolytic cycle and may be lost during slaughter (Bate-Smith, 1936, 1938a). I n fact, Best et al. (1926) have shown that it is possible t o eliminate glycogen from the muscles by inducing death by convulsion through the injection of insulin. Callow (1939) found that the glycogen reserves of the hog were easily depleted and believed th at feeding was necessary to restore muscle glycogen after exercise. In other experiments reported by Callow (1939), hogs were shipped one mile by truck and then walked a quarter of a mile before slaughter. The ultimate pH in the psoas muscle of a group of animals which was rested overnight after this treatment, but not fed, averaged pH 5.80 as against pH 5.79 for the unrested group; whereas a group that was both rested and fed had an average pH of pH 5.58 as against pH 5.87 for the unrested control group. Earlier, Callow (1936) pointed out t ha t “resting” must, in fact, be resting in order to be effective. He said t ha t pigs tend t o fight when strange groups are mixed together in resting pens and, under such circumstances, recovery of glycogen does not take place. Since the quantity of acid present in the dead muscular tissue depends principally upon the quantity of glycogen a t the moment of death, the final acidity of the meat derived from animals depleted of glycogen is less 1 The p H after all residual glycogen has been converted. Although i t is generally assumed that the ultimate p H of muscle is the result of the post-mortem accumulation of lactic acid arising from glycogen, there are many other normal muscular constituents t h a t are acidic in nature.

MICROBIOLOGY O F MEAT ANIMALS

123

than t h a t for animals having adequate reserve glycogen for normal activity. Callow (1949) indicated that the following ranges in ultimate p H were obtained for beef, lamb, and pork during the course of his investigations: beef p H 5.1-6.2; lamb p H 5.4-6.7; pork p H 5.3-6.9. The importance of this relationship was illustrated by Ingram (1948, 1949) who presented data showing that a slight acidification of the medium markedly reduced the rate of growth of a wide variety of bacterja. Also, Bate-Smith (1948) points out t h a t “if meat has reached its ultimate pH with a n excess of glycogen still remaining, the growth of microorganisms may not cause an alkaline shift, because production of base will only result in further breakdown of glycogen to lactic acid and i t will not be until the glycogen is completely exhausted that the p H can begin t o rise. I n other words, from the point of view of resistance t o bacterial growth there cannot be too much glycogen in the muscles.” Callow (1935) also attributed t o fatigue the difference in electrical resistance between the muscles of hogs butchered on the farm and those slaughtered in packing plants. Subsequently, he (1936, 1938) demonstrated a correlation between a high resistance of flesh and its degree of acidity. Banfield (1935) and Callow (1935, 1936, 1938) regard this relationship as due t o swelling of the fibers, with a subsequent narrowing of the channels through which ions can move freely. Callow (1937) used the terms “open” and “closed” t o differentiate structure of meats; open structure is considered t o have low p H values, moist feel, firm texture, and pale color; closed structure is associated with a sticky dry feel, flabby texture, and darker color. Bate-Smith (1933) considers the stickiness with the high pH due not only t o the swelling of the fibers but also due t o the dissolution of myosin. Ingram (1948, 1949) demonstrated that the rate of growth of bacteria can be greatly diminished by a fall of 0.1 pH unit and, therefore, intimated that growth becomes progressively less probable as the p H falls below pH 6.0. Dry, salt-cured hams made from fatigued pigs are thus likely t o be tainted, and the incidence of taint will diminish as the pig is rested and fed (Callow, 1937). Moreover, it is not only the growth of the anaerobic bacteria which cause tainting in hams that is affected by the state of the live pig, but also the growth of contaminating bacteria on the surface of the meat. Madsen (1943) found that the formation of slime due t o the growth of aerobic bacteria on the surface of sides of bacon is retarded and the storage life prolonged by feeding and resting pigs before slaughter. Thus the growth of aerobic as well as of anaerobic bacteria on and in the flesh, post-mortem, are affected by the state of the animal a t the time of slaughter.

124

JOHN C . AYR E S

IV. SLAUGHTER 1 . Killing

With the exception of the animals slaughtered for the Kosher trade, the method of killing cattle and calves differs from t h a t used for hogs and sheep. As a consequence of these differences in operation, the microbiological changes taking place in the tissues during the dying period may vary with the type of animal being considered. Haines (1937) lists several methods of slaughter:

(1) Simple bleeding after sticking in the thorax or cutting the throat. (2) Bleeding after previous mutilation of the medulla, e.g. pithing. (3) Bleeding after previous stunning with a blow from a hammer, killing mask, or captive-bolt. (4) Bleeding after stunning electrically. S o t in Haines’s list, but a commonly used method on the farm is killing the animal by shooting. Also, an immobilizing procedure has been introduced recently (Murphy, 1950) which may revolutionize the method of handling animals preparatory t o bleeding them. Ordinarily, cattle and calves are stunned by a blow from a hammer t o the animal’s head. With cattle, several animals are crowded into knocking pens (the number varies among plants and also with the size of the animals) t o prevent excessive movement. After all of the animals have been stunned, a shackling chain is put around each animal’s two hind legs and it is hoisted t o a n overhead rail for bleeding. Depending upon the number of animals and the proficiency of the knocker, several minutes may elapse from the time that the first animal is knocked down and the last animal is in position t o be bled. There is opportunity during this period for the hide t o become impregnated with filth from the walls and floor of the killing pen. A slit is cut through the hide down the middle line of the throat and into this a knife is inserted. The arteries and veins of the throat are then severed by a slight sidewise cut. (In most cases stunned cattle cease breathing quickly.) Although stunning suspends the heart action so that bleeding is largely mechanical under ideal conditions, some of the animals recover partial consciousness during these operations. Under such circumstances, the microbiological implications discussed later for the hog also apply. I n England cattle are often killed by the use of a captive bolt pistol. Although this method may have some humanitarian values, hair and hide, blood and bone, and the bolt as well are embedded in the brain. The con-

MICROBIOLOGY O F MEAT ANIMALS

125

taminative load of bacteria introduced with these agents quickly attack the damaged tissues rendering them unfit for food usage in a short time. Hogs and sheep are driven into shackling pens where they are shackled by one of the hind legs and hoisted on t o a n inclined rail. Immediately after the animal is suspended, i t is “stuck” by a quick thrust of a sharppointed, two-edged knife into the throat severing the large blood vessels leading to the head. Often the animal struggles violently on the shackle rail and great care must be exercised by the sticker t o prevent injury t o himself or t o the flesh of the animal. After sticking, the blood quickly drains from the carcass. In England pigs are stunned electrically before shackling, for humane reasons, but this method has not been received very favorably on this continent because changes induced in the lungs and pleura make veterinary inspection difficult (Gibbons, 1953). According t o Haines (1937) there are numerous reports in the literature attempting t o show that one method gives more complete bleeding than another; he criticizes most of these as involving inadequate numbers of animals. Jensen and Hess (1941) criticize the concept t h a t faulty exsanguination is responsible for increased bacterial spoilage, and cite, as evidence, many tests made over a period of 4 years wherein hogs were bled from large and from small incisions and dropped into scalding vats immediately or 15 min. after sticking. They stated that no conclusive evidence was presented t o show the superiority of one method of bleeding over another. Recently a process whereby carbon dioxide is administered in controlled quantities in tunnels through which the animals are moved, has been developed for immobilizing hogs (Slater, 1952). The method was also suggested for cattle, sheep, and poultry (Anonymous, 1953). Sufficient gas concentration is inhaled by the animal t o asphyxiate it during most of the shackling and sticking period. Although the technique was developed primarily t o reduce injury t o the workmen and t o the flesh of the animal, it has several values of microbiological importance as well. For example, immobilizing (1) diminishes animal fatigue; ( 2 ) reduces the amount of dirt on the animals and workers; (3) assures more accurate and sanitary sticking; and (4) allows animal respiration and thereby improves bleeding. 6. The Stick-Knife

Sticking of life animals in this country is restricted t o hogs and sheep; cattle and calves are stunned by a blow to the head before bleeding. However, with any animal in which the heart is still beating, changes cannot be considered post-mortem. For this reason, it is best t o consider

126

JOHN C. AYRES

the physiological changes and the immediate microbiological aspects which attend them as being agonal. As previously mentioned, Jensen and Hess (1941) found the skin of a hog t o be heavily contaminated. They considered the use of the stickknife a very important avenue for the introduction of these microorganisms into the tissue. Jensen and Hess did not feel th a t the tissue of the live animals were contaminated but that the bacteria are introduced during slaughter and while the animal was dying, they are circulated by way of the blood stream. They suggest the following sequence of events after a n animal is stuck: The stick-knife passes through the skin severing the jugular vein and sometimes the carotid artery; then arterial and venous blood sweeps over the contaminating blade, returns to the heart and is pumped throughout the arterial system. According t o Jensen and Hess (1941) the heart of a hog may beat from 2 t o 9 min. after the stick wound is made. Visible movement is detectable for at least 40 sec. after the animal has been stuck. I n addition, some of the hogs contract their heads in the direction of their forelegs and in this manner withhold a good deal of the blood flow for 5 to 15 sec. by constriction and hematoma, allowing more arterial and venous blood from this area t o reach the heart and eventually t o be circulated in the arteries. Also the severed or pierced vessels may be under reduced pressure owing to the labored breathing resulting from oxygen starvation during exsanguination. The flow of the pooled blood and blood within the vein passes toward the heart. These workers also point out th at the stick-knife ordinarily is wet and heavily contaminated with spores. They theorized that, on the basis of the amount of serum or liquid blood in a 250 Ib. hog, a knife blade introducing only 50,000 bacteria could infect the circulating blood with over a dozen bacteria per milliliter since phagocytosis would not necessarily eliminate many bacteria in the short span of life after sticking. I n a series of biopsy studies which Jensen and Hess (1941) made of blood, bone, bone marrow, and muscle tissue, with one exception (an animal which harbored Hemophilus sp. in all tissues while alive and also after dressing), none of the hogs showed the presence of microorganisms in the tissue or blood. After the surgical fields were closed, sutured, and heavily covered with celloidin, the animals were immediately taken to the killing floor, hoisted, stuck and bled, washed, dehaired, butchered, and dressed. Post-mortem findings indicated that bacteria entered the animal during these manipulations. Jensen and Hess considered th a t fewer bacteria were t o be found in the blood of the heart of sterilely stuck hogs than in those t hat had been septically stuck. I n both cases, they found considerable numbers of aerobes and anaerobes in the incoming blood up to the time that the heart stopped beating. They considered that the

MICROBIOLOGY O F MEAT ANIMALS

127

sterile sticking method and long bleeding mere not adequate t o prevent contamination of blood. They concluded that the presence of bacteria in the blood, muscle, and bone marrow of these aseptically bled hogs indicates that the organisms enter during the dying period. Jensen and Hess (1911) also placed pure cultures of bacteria on the blade of the knife just prior t o sticking and then used this instrument t o sever the vessels in the hog’s neck; they found that the specific bacteria introduced could be isolated from the marrow of the tibia and other long bones. It should be mentioned that the largest percentage of blood remaining in a slaughtered animal is found in the muscle. The stick-knife contamination and the sequence of events following the severing of the neck vessels should favor the entry of coliform and other bacteria into the blood and marrow. However, some factor prevents their reaching the marrow or surviving therein if they do reach it. It has been long known (Adami, 1899) that undiluted, fresh mammalian blood is bactericidal. Boyer (1926) called attention t o “the absence of the Bacillus coli (Escherichia coli) group of organisms from . . . hams (although) members of the group . . . are almost invariably found on the surfaces of the carcasses which are exposed during killing floor operations. Their absence is of special significance in that it goes far t o eliminate the possibility that the organisms present in the hams gain access during killing floor operations.” Jensen and Hess (1941) thought that bactericidal action of the blood was responsible for the absence of coliforms from heart blood or from ham. I n order t o test this theory arid t o find a clue to the presence of only certain bacterial species in the bone marrows, they conducted the following tests. Blood was withdrawn aseptically from the tail of a live hog and was allowed t o drip directly into sterile flasks containing glass beads t o defibrinate the blood. Various freshly isolated strains of bacteria were added t o this blood t o give a level of about 50,000 viable cells per ml. and subcultures were made a t short intervals up to 24 hr. It was found t h a t the suspensions of Staphylococcus aureus (Micrococcus pyogenes var. aureus), and E. coli often were sterile after 2 t o 5 hr. whereas the bacteria which Jensen and Hess termed the ham-souring types, such as certain strains of Serratia, Achromobacter, Clostridium putrefaciens of McBryde (1911) (C. lentoputrescens), and Pseudomonas, were more or less resistant t o the bactericidal effects of hog blood. 3. Scald Tank

Use of a scald tank is restricted t o hogs since the hides of the other meat animals except for those of calves and “hothouse” lambs are removed by skinning. These latter are dressed with their skins or pelts remaining on the carcass t o prevent excessive dehydration of the meat.

128

JOHN C . AYRES

Ordinarily the water in the scald tank is maintained a t a temperature of 62.8 t o 65.6" C. (145 t o 150" F.) and contains soap and sal soda t o aid in the loosening of hair and scurf from the body of the hog. The numbers of organisms that are contributed t o this water during the process of a day's operation are almost astronomic. If one considers the bits of soil, fecal material, blood, and other foreign material that may be on the feet, hair, and hides of these animals, it is easy t o realize that millions and even billions of bacteria may be present in every milliliter of the water after the first few carcasses have traversed the scald tank. It has been pointed out by a number of workers (Empey and Scott, 1939a,b; Empey and Vickery, 1933; Mallman et aE., 1940; Haines, 1937), that the flora which ultimately develops on the meat come principally from soil organisms associated with the hoofs and hides. Unfortunately, no studies are available t o indicate the role of the scald tank as a cross-contaminating medium. Depending somewhat on the speed of movement of the process line, hogs move through the tank in approximately 3 t o 5 min. Most of the carcasses float; occasionally, however, one will sink t o the bottom and remain there for some time before i t is rescued. The Meat Inspection Division of the Bureau of Animal Industry specifies that a hog shall hang on the bleeding rail for a t least 6 min. This interval has been required in order that the animal be lifeless before being dipped into the scald tank. However, Jensen and Hess (19$1) observed that the hearts of hogs stuck by using the large incision method (5-in. slits) continued to beat for 6 t o 9 min. Under such conditions, and with the sticking and scalding limitations listed, it is easily possible for the circulatory system of a n occasional animal t o be functioning while i t is in the scald tank. Also, it is doubtful t h a t the temperature of the scald water is sufficiently high t o destroy spores or even all of the vegetative cells. However, the heat is sufficient t o destroy many of the psychrophilic bacteria. Few studies are available which relate t o the presence of bacteria in the circulatory or respiratory tissues of animals immediately after slaughter. However, government regulations forbid the use of hog lungs for edible purposes owing t o the amount of tank water which enters during scalding (White and Patrick, 1916). Spray (1922) reported that, in several instances, microorganisms were recovered from apparently normal lungs of pigs slaughtered in a commercial packing plant. Jensen and Hess (1941) found that the lungs of hogs, stuck in the ordinary manner and sent through the scald tank and dehairing machine with the wound open, were highly contaminated. I n addition, they found t h a t when a sterile technique was used in sticking, followed by scalding in a hot spray cabinet, samplings of the washings and swabs taken from the lungs showed no

MICROBIOLOGY OF MEAT ANIMALS

129

more than a few colonies. The bacteria which were identified were representatives of the genera Bacillus, Achromobacter, Micrococcus, Pseudomonas, Clostridum, Pasteurella. Alcaligenes, and a miscellany. Although Jensen and Hess reported that substitution of a continuous spray in steaming cabinets instead of the scalding bath does not aid materially in reducing the bacteria in the carcass, they do not present data substantiating this statement. Recently , a t least in one commercial installation, use of the scalding tank has been replaced by a steaming cabinet. Ti. POST-MORTEM 1. Skinning

Ordinarily cattle and sheep are skinned before they are eviscerated. With cattle, the usual procedure is t o perform part of the operation while the carcass is resting on the floor of the eviscerating room or ‘(bed” on its back or side. There are some recent innovations in this dressing procedure wherein the animals are skinned in the same manner as that used for hogs, i.e., suspended from a rail. Recently, a device has been placed in operation which functions mechanically t o partially separate the hide from carcass of cattle. Also, in some installations the hide is removed by applying pressure from air hammers instead of by the use of knives. With sheep, the skin is removed while the animal is suspended from the rail. The pelt is partly removed by cutting and by pressing the fist between the pelt and the thin white membrane covering the flesh. Empey and Scott (1939a) found that the transfer of microorganisms from the hide t o the underlying tissues begins with the first stage of skinning. They found populations ranging between 10,000 and 100,000 per sq. em. of the superficial tissues of the carcass. Numbers of organisms in the tissues were highest in the region below the initial incision through the hide and lowest in the areas that were furthest removed from this region. The population of bacteria on knives used for incising pieces of hide of various lengths were found t o range between 80,000 and 40 million organisms per blade. Empey and Scott state that, in addition t o the numbers that may eventually be transferred during separation of hide from carcass, there are some that are planted directly by the blade when it touches the exposed tissues during the initial incision through the hides. Further transfer was found t o occur when parts of the hide that had been separated again came in contact with the carcass while the skinner was changing position. Experiments showed that when such areas of direct contact were made, contamination equal t o one-third that of a n equal area of hide was transferred. Additional organisms were transplanted to the carcass from the hands, arms, legs, and clothing of workmen. How-

130

JOHN C . AYHES

ever, Empey and Scott were unable t o estimate the number of organisms tha t were planted in this manner. The clothing of the workmen who were engaged in removing the hide and which came in contact with the hair were found t o acquire a high level of contamination. I n the presence of moisture, blood, and tissues, these organisms found the substrate suitable for multiplication or survival and often reach levels as high as 3 billion per gram or 30 million per sq. cm. in the material that could be scraped from the clothing of workmen who had participated in the skinning of approximately 100 carcasses during a 6-hr. period. The population acquired by the hand of a workman during the handling of the hair on about 100 sq. cm. of hide reached 2 million. I n addition to what may be regarded as a permanent microflora of the skin and hair, the hide and hoofs of the animal carry varying amounts of soil. Soil was found to contribute t o the contamination of the carcass in about the same manner as did the hide, but Empey and Scott (1939a) state t ha t it is not possible to accurately determine the number of organisms originating from this source. Populations of microorganisms vary considerably in different soils. According t o Sarles et al. (1951) as few as 1000 or as many as 10 billion bacteria may be found in a gram of soil; the usual range was stated to be from 1-10 million. Then too, the hoofs and lower parts of the legs usually carry a considerable amount of soil and, if the workman engaged in handling the feet subsequently touches other areas of the carcass, soil provides a high proportion of the contamination. 2. Dehairing, Xhaving, and W a x Dipping

As with scalding, the dehairing, shaving, and wax dipping operations are restricted t o hogs. After the animal has been properly scalded to loosen the hair and scurf, it is lifted from the scald tank b y a hook placed under the tendons of one of the hind legs. Then the hog passes through a dehairing machine where the hair is beaten or scraped from the carcass. Although this device “polishes” the hog by removing most of the hair and scurf, from a sanitary viewpoint, bacteria or spores may be pounded or scratched into the skin of the animal. Following dehairing, the carcass passes through a hot water shower and then parts not properly dehaired are shaved with sharp knives. The hog may or may not be singed t o remove the remaining hair and to change the coloration of the skin. If this processing step is used, the head and two front feet generally receive most of the heat. The final processing steps before the animal is sent t o the chill room

MICROBIOLOGY O F MEAT ANIMALS

131

are those involving shaving and wax dipping. Shaving is done with sharp knives which remove the remaining hairs on the surface of the carcass. I n many plants in the United States the animal is then immersed in a vat of moulten wax which adheres t o the carcass. Then the hog is withdrawn and the wax quickly congeals, enmeshing the remaining hair and practically all of the hair roots. After the wax is congealed, i t is pulled from the carcass. The adhering hair and scurf are separated by screening and the wax remelted and reused. There is no information available concerning the introduction of bacteria during the dehairing operation or the removal of these organisms later in the shaving and the wax-dipping processes. Also, no data could be found which related to the destruction of bacteria or spores in the molten wax. 3. Evisceration

I n general, all meat animals are eviscerated in approximately the same manner. However, there are some rather important differences in the techniques that are involved in removing the viscera from these several animals. The animals arrive on the evisceration floor, shackled by their hind feet and hanging head down. The hog is opened from tail t o throat by the use of a knife and a cleaver. An electric saw sometimes is used to help divide the belly side along the median line. The bung is loosened and the entire viscera is removed in one continuous operation. The fat is pulled and then the hog is split through the center of the spinal column. The hams are faced by removing fat and skin from the flank and cushion sides of the ham. The carcass is again sprayed and sent t o the cooler. In cattle, the breast bone and aitchbone are sawed through exposing thoracic, abdominal, and pelvic cavities. The esophagus, bung, and bladder of beefs are tied to prevent regurgitation or contamination of the carcass with fecal matter or urine. The bung is loosened and dropped within the pelvic cavity permitting the viscera t o fall free. After evisceration the carcass is split in half usually with the aid of an electric saw, and the sides washed and scrubbed with hot water and covered with wet muslin sheets a t the time the carcass is sent to the chilling room. These cloths usually are moistened in hot brine or hot water before they are draped closely over the surface of the carcass. Pressure is applied in the draping of the shroud in order to give the carcass a smoother appearance and more desirable conformation. Shrouding also is credited with preventing some loss of moisture and with bleaching of the fat. However, Jensen (19$5) states that microbial growth is induced a t the interface of textile and water film and, to a smaller extent, on the beef. He also mentions that pins used t o fasten the shroud t o the carcass may a t times cause

132

JOHN C. AYRES

the formation of dark, discolored spots or even areas several inches in width and depth. I n sheep, the body is opened by splitting the breast bone and, after loosening the bung, the viscera is removed without dividing the aitchbone. Following this operation a spreadstick is placed inside the thoracic cavity t o permit air circulation when the carcass is sent t o the chill room. The feet and lower leg of cattle and sheep are not saved for edible purposes; owing to their unclean condition, i t is almost impossible to remove them so as to prevent contamination. An extensive study of the microbial contamination encountered in a beef packing plant was made by Empey and Scott (1939a,b). Although they accredited a considerable inoculum of organisms on the carcass as having come from the soil and dirt on the hide and hair and from the skinner, they reported th at freshly voided feces or intestinal content which accrued before the completion of skinning and the accidental puncture of various sections of the gastrointestinal tract by knives by the workmen were responsible for some of the contamination. Principal sources of microorganisms on the eviscerating floor include contamination that is airborne, from the water used in washing and rinsing the carcass, from cloths and brushes used for wiping tools or carcasses, from bandsaw races and other tools such as knives, saws, cleavers, and hooks, and from the hands of the workmen. I n many abattoirs and packing plants, the sticking and scalding area for hogs, and the knocking and bleeding enclosure for cattle and sheep, are separated from the eviscerating and cutting floors. Although these operational steps have been separated principally for aesthetic reasons, they also serve to reduce airborne contamination on the eviscerating floor which otherwise would accumulate from dust and dirt derived from the milling and thrashing of the animals while they are on the killing rail or in the knocking pens. Empey and Scott (1939a,b) presented data t o show the microbial deposits from air of slaughter floors; they found a n immediate plate count of about 30 bacteria and about 2 molds per sq. cm. per hr. Haines (1933a) surveyed the bacterial flora in two types of slaughter houses : the small privately-owned killing shed and the large modern processing plant. He found the numbers of organisms in the former were proportional t o the number of animals being slaughtered and the climatic conditions. This did not hold when the small killing plant was compared with the larger, more modern, abattoir. Haines stated th at although each animal is to be regarded as a potential source of a given load of bacteria which will be scattered during handling, the degree of infection was less owing to the better system of ventilation in the large abattoir.

MICROBIOLOGY O F MEAT ANIMALS

133

Also, in the two types of processing plants studied by Haines (1933a)) analysis of the types of flora present in air samplings showed striking differences which he attributed t o the diverse methods of handling the blood, inedible viscera, and hides. I n the more modern abattoirs in England where chutes and drains were provided t o carry away offal, there were only 9% of the flora which were “intestinal’) types compared with 19% for the other type of slaughterhouse. Haines reported the following types in air; Staphylococci, Micrococci, Bacillus, Pseudomonas, Achromobacter, Flavobacteria, Azotobacter, Proteus, Aerobacter, Coliforms, and molds. I n modern American packing houses elaborate conveying and processing devices are employed for handling viscera after they have been removed from the animal. Generally, stainless steel trays are used t o catch the internal organs as they are removed or drop from the carcass. Pans and carcasses move together in sets in the processing line until both the animal and its viscera have passed post-mortem inspection. Trays are emptied into appropriate chutes, rinsed, and, after cleaning, ret]urned to service. During eviscerating operations, workmen often use their hands t o make incisions, t o wash or brush inner surfaces, and t o remove blood and other contamination. Also, the veterinary inspector may touch edible parts of the animal during the course of his inspection of the head, heart, liver, lungs, kidneys, stomach, intestines, membranes of the thorax and abdomen, various groups of lymph glands, internal and external surfaces of the body, and exposed bones. From time t o time, bacterial transfer from one part of the animal t o another is a n expected consequence of these manipulations. Also, during such contact, the workman or inspector may transplant part of the flora that is on his hands t o the meat. Horwood and Minch (1951) examined 34 hand-washing samples which they had obtained from 22 food handling establishments. They isolated large numbers of bacteria from the hands of food handlers; also, Escherichia coli, hemolytic staphylococci and streptococci, and aerobic spore forming bacteria were frequently recovered. Horwood and Minch concluded that (1) food handlers frequently bring the hands in contact with food when the us? of a n implement is indicated and ( 2 ) hands are frequently soiled with the discharges from the nose and mouth and in other ways. After the animal has passed inspection, the carcass is sent t o the chilling room but the edible viscera and offal receive additional handling. Ordinarily, the brain, heart, liver, and kidneys (not removed from sheep but sold with the carcass) are saved. I n addition, the stomach of hogs and the rumen and reticulum of sheep and cattle are often processed and used

134

JOHN C. AYRES

in the manufacture of tripe. The intestines of all three of the meat animals commonly are processed and used for sausage casings. Before animal st,omachs or intestines can be utilized for food purposes i t is necessary that they be turned or cut open t o permit the ejection of their contents and a thorough washing of their interiors. Also, before intestines are satisfactory for sausage casings, fat must be removed from the outer surface and the intestines turned with the inside out and run through the sliming machine. Inedible components of the viscera, together with material flushed from the gut of the animal, are conveyed into appropriate chutes for disposal as tankage or fertilizer. Water, when applied with force directly from the t a p or hose, does not of itself contribute many microorganisms. The wash water dislodges much of the blood, tissue debris, hair, and scurf t h a t accumulates on the carcass during the skinning and dehairing operations. I n all probability i t also reduces the total amount of contamination even though it more uniformly redistributes through other tissues bacterial populations from local areas where their numbers may be large. Similarly, microorganisms may be transferred t o carcasses from the hands and clothing of the operator or splashed on the animal from the walls and floor of the washing stall. Haines (1933a) reported t h a t when walls were covered with a rough plaster, material which splashed on them was readily retained and absorbed. He also pointed out that walls of this type were often covered with mold growth and that carcasses brushing against such walls became contaminated. Empey and Scott (1939a) cautioned against the use of plaster or wood. Rooms in most modern meat packing plants have smooth concrete floors, tile walls and steel girders, which are easily washed down. Empey and Scott (1939a) exposed sterile cloths t o the fine mist produced from the contact of water with floor. They estimated the extent of contamination dislodged from the slaughter floor t o range between 1000 and 4000 organisms per sq. cm. for a 2-sec. exposure. Numbers of microorganisms found in water collected in utensils or for the washing of tools depend upon the previous history of the water and of that for the utensil or tool. If water is recycled or reused, its microbial content increases rapidly. Also the blood, grease, and bits of tissue introduced into the water when the tool is washed or immersed provide nutrients t o support the growth of microorganisms. Haines (1933a) reported some extremely high counts (2-25 million per ml.) for water that was used in swabbing carcasses under rather poor conditions and concluded that such practice served merely t o inoculate the flesh instead of t o clean it. The chief types of bacteria associated with this water were : Aerobacter cloacae, Staphylococcus epidermis (Micro-

I\IICROBIOLOGY OF MEAT ANIMALS

135

coccus tardior), Micrococcus candicans, Escherichia communior (coli var.), Proteus vulgaris, Micrococcus luteus, Pseudomonas, and Micrococcus perjavus. On the other hand, under cleaner working conditions and with water changed at frequent intervals, the water used for swabbing carcasses contained 1500 organisms per ml. principally of the following types: Xarcina lutea, Escherichia communior (coli var.) , Flavobacterium, and Achromobacter. Water used for the washing of cloths and brushes employed in wiping or cleaning parts of the carcass was found by Empey and Scott (1939a) to become heavily populated with organisms. The loads reached an equilibrium soon after the beginning of the day’s operation. Numbers per milliliter of fluid ranged between 20,000 and 350,000 with a median of 140,000, These workers found, also, that the populations on the cloths ranged bet\%-een-20,000 and 100,000 per sq. cm. They considered that the question regarding whether or not organisms were transferred from cloth to carcass or were removed by wiping depended on the relative populations of both cloth and tissues prior to the operation. Material adhering t o the blades of saws, knives, cleavers, and to the platform of conveyor tables often is contaminated and mechanically inoculates the newly exposed cut flesh of the animal with organisms derived from air, water, handlers, or miscellaneous sources. Portable and stationary electric saws are receiving increased usage in modern packing plants. The blades move at high speed and the blade-guards quickly accumulate tissue materials which are removed only periodically. Large loads of microorganisms inhabit this material and, from time to time, are caught on the revolving blade to inoculate new tissue.

4. Chill Room Present practice in the handling of meat is to send the carcass to the chill room as quickly after evisceration as possible. Several reasons have been advanced for the adoption of this procedure in order t o prolong the storage life of meat. Among these are: (1) to free it of body heat, (2) to firm the flesh, (3) to delay undesirable bacterial and chemical changes, and (4) to prevent shrinkage. This procedure has not always been in vogue as may be witnessed by the comments of Moran and Smith (1929) decrying the general practice of the trade at that time of allowing the animal to remain at room temperature after slaughter until the animal heat had dissipated. As they pointed out, and as others have indicated since (Jensen, 1945),it is imperative to chill the carcass quickly in order t o avoid early spoilage of the meat.

136

JOHN C. AYRES

Haines (1937) stated that, even under the best conditions of chilling, the temperature in the region of the 0s pubis of beef will not drop below 5" C. (41" F.) in less than 48 hr. I n most packing plants in the United States the old method of allowing beef t o hang for several days or weeks in the packing plant is no longer used. Although just 10 years ago beef commonly was held about 5 days at 2 t o 3" C. (36 t o 38" F.) after slaughter before it was removed for final disposition (Jensen, 1944), now the ordinary holding period is reduced to 2 t o 3 days. For pork and mutton the usual chilling time is overnight. I n America, no specific schedule for aging is set other than the time given in the chill room, in transit, and in the distributing house. The use of ultraviolet light in the chill room also serves to shorten the aging time since it permits the maintenance of a higher room temperature in order t o obtain more rapid meat tenderization while inhibiting microbial growth (Christensen, 1940; McIntosh et al., 1942; Ewell, 1943; Sotola et al., 1943; Volz et al., 1949). The air and walls of the chilling room are subjected t o contamination from workmen and materials with which they might come in contact. Further, organisms can be transferred t o the carcass by the circulating air and by the feet of the workers; areas of the animal th a t are near the floor are especially receptive. Richardson et al. (1954) state th at cooling rooms in meat packing plants undergo a rigid sanitary clean-up a t the end of the day. During this clean-up, where sawdust is used, it is stirred up so th a t the air becomes permeated with mold spores which remain suspended for many hours. They considered th at this practice provided a source of contamination of the carcasses for the following day. Fresh sawdust is used in chill rooms and elsewhere in packing plants t o absorb blood and fluid from beef, pork and sheep carcasses as it drips t o the floor. Sawdust is cheap, absorbs moisture readily, is a good insulator, and is readily available in most localities. According t o Empey and Scott (1939a), the number of organisms deposited from the air was invariably higher in those rooms in which sawdust was used. They reported that fresh sawdust used in chill rooms contained, on the average, 3 million organisms per gram; of these about 1 % were considered to be psychrophilic bacteria. Sawdust, after exposure on floors of chill rooms, showed evidence of increased moisture content and of blood stains. Microbial populations of chill-room sawdust were predominately bacterial, with counts ranging between 60 and 450 million low-temperature types per gram of moist material. Also, the extent of microbial contamination deposited in the air of the chill room indicated counts of about 8-1100 bacteria, 10-100 yeasts, and 2-250 molds per

MICROBIOLOGY O F MEAT ANIMALS

137

sq. em. of agar surface per day. Percentages of psychrophilic bacteria were 29 % a t 20 days and 43 % a t 40 days of incubation. Air samplings made in a beef cooler by Richardson et al. (1954), using a n electrostatic sampling device, indicated mold spore contamination ranging from 882 t o 4500 spores per plate exposed for 2 min. in afternoon tests. Three t o ten men were working in the area (8 f t . X 40 ft. X 150 ft.). Richardson et al. considered that the activities of these individuals kept the mold spores suspended and in motion. Empey and Scott (1939a) found that plaster walls in processing rooms usually were cleaner than were wooden walls. Modern chilling rooms in this country utilize compressed corkboard as a n insulation material, these establishments report that this type of insulation remains in excellent condition for many years (Anonymous, 1952). Several investigators (Talayrach, 1901; Klein, 1908; hIcBryde, 1911; Massee, 1912; Monvoisin, 1918; Bidault, 1921; Brooks and Kidd, 1921; Brooks and Hansford, 1923; Haines, 1931, 1933a,b; Moran et al., 1932; Empey and Vickery, 1933; Gorovitz-Wlassova and Grinberg, 1933; Empey and Scott, 1939a; Mallmann et al., 1940; Kirsch et al., 1952; Ayres, 1954) have identified organisms isolated from fresh meat held a t chill room temperatures (Tables I and 11). The taxonomic distribution of the various bacteria, molds, or yeasts isolated indicates that the following genera are represented : (bacteria) Pseudomonas, Azotobacter type, Chromobacterium, Micrococcus, Gaffkya, Sarcina, Diplococcus, Xtreptococcus, Lactobacillus, Achromobacter, Flauobacterium, Xerratia, Proteus, Bacterium, Bacillus, Clostridium, Diphtherozds, Streptomyces, Actinomyces; (molds) Rhixopus, Phycomyces, Mucor, Thamnidium, Geotrichum, Monilia, Aspergillus, Penicillium, Xporotrichum, Botrytis, Verticillium Torula, Cladosporium, Alternaria, Stysanus; (yeasts) Rhodotorula, Wardomyccs, Saccharomyces. Although a rather heterogeneous flora has been found on chilled meats by various workers, many of the organisms recovered may be transient or adventitious. The majority of these were not identified with specific defects in the stored meat; only a few of them were associated with spoilage. Various explanations have been offered for this restricted residential bacterial load on the surface of the intact carcass. One opinion held among workers in the meat industry seems t o be that some reduction in humidity facilitates preservation of all types of meat. For example, Jensen (1945) states that certain mold growths are troublesome on cold-storage meats if humidities of coolers are not controlled. However, there is evidence in the literature (Scott, 1936; Ogilvy and Ayres, 1951b) which indicates that the influence of relative humidity has little effect in delaying microbial

138

JOHN C. AYRES

TABLEI Genera of Bacteria Viable on Refrigerated Beef

Genera represented Pseudomonas Azotobacter type Chromobacterium Micrococcus Gaffkya Sarcina Diplococcus Streptococcus Lactobacillus Achromobacter Flavobacterium Serratia Proteus Bacterium Bacillus Clostridium Diphtheroids Streptomyces Actinomyces

Empey Gorovitzand Wlassova Empey Mall- Jensen Haines, Vickand and mann and Kirsch 1931, ery, Grinberg, Scott, et al., Hess, et al., Ayres, 1933a,b 1933 1933 1939a,b 1940 1941 1952 1954

+ + + + +

+

+

+

+

+

+ +

+

+

+

+

+ +

+

+

+ + +

+

+ + +

+

+

+

+ +

+

+

+ + + + +

+

+

+ +

+ + + +

+ +

+

+ +

+ + + +

growth on meat surfaces. Certainly in a dry atmosphere, the diffusion of water from the interior of meat helps t o maintain a higher moisture content near the surface than is indicated by the relative humidity of the surrounding air. Haines (193313) offered a possible explanation for the growth on uncut surfaces being limited: namely, t h a t the carcass surface of the whole or quartered animal is covered by a layer of fat and connective tissue and has poor nutrient qualities for most organisms. Also, many of the organisms coming in contact with the meat are mesophiles which grow poorly a t low temperatures (Haines, 1934) and, when the animal is chilled, die before conditions in and on the meat are again favorable for their growth (Ayres, 1951). If the carcass is properly and quickly chilled, putrefaction is minimized inasmuch as many of the organisms responsible for this type of decomposition are inhibited. The principal types of changes of refrigerated fresh red meats are four in number. These are: (1) off-odor and slime, (2) ham souring or bone stink or taint, (3) black spot, whiskers, and other mold discoloration, and (4) fat rancidity.

T m i , L I1 Genera of Molds and Yt,nsts ViaLlc. on Refrigcrated Meats

Genera represented

Talayrach, 1901

Massee, 1912

Monvoisin, 1918

Bidault, 1921

Brooks and

Kidd, 1921

Brooks and Hansford, 1923

Moran et al., 1932

Empey and Scott, Ayres, 1939a,b 1954

~

Molds Rhizopus Phycomyces Mucor Thamnidium Geotrichum Monitia Aspergillus Penicillium Sporotrichum Botr ytis Verticil tium Torula. Cladosporium Alternaria Stysanus Yeasts Rhodotorula Wardomyces Saccharomyces

+ + + +b

(+P

+ + + + + +

+ ++ ++

++ + + + + +

++ +

+

++

++

++

+ +

+

5 g 0

$c

z

F

0 0

+ +

+ +

4

+ +

+ +

s 5

%

+ +

+ + +

+

+ +

+ +

$

z F

E:

+

a Most mycologists use the name Torula t o refer to certain of the Dematiaciae (Skinner et al., 1947) and, in particular, the genus Cladosporium. Brooks and Ransford (1923) did not consider the organism that they had isolated to be a species of Cladosporium. b AIso inclndes Dernatiaeium and Hormodendrum. Described a pink yeast.

F w cc

140

JOHN C. AYRES

a. Off-odor and Slime. Haines (193313) considered that the cut flesh is subjected t o an increase in numbers of microorganisms even while i t is stored a t refrigeration temperatures. Ingram (1949) pointed out that the low-temperature group of bacteria is very greatly affected by humidity, tending t o form a slime more readily on cut surfaces than on skin, connective tissue, or fat. Microorganisms appear first in damp pockets, such as folds between the foreleg and breast of a carcass, and their spread is greatly promoted by the condensation which occurs when a cold carcass is exposed t o warm, damp air. I n 1901, Glage isolated slime-forming bacteria, which he called "Aromobakterien," from the surfaces of meat stored a t low temperature and high humidity. He considered t h a t there were 7 species, one of which predominated. The bacteria were oval t o rod shaped with rounded ends and occurred occasionally in chains. They were motile, aerobic, liquefied gelatin slowly, and turned litmus milk alkaline; they grew well a t 2" C. (35.6" F.) but poorly at 37" C. (98.8" F.) ; the optimum temperature was thought t o be a t 10 t o 12" C. (50 t o 53.8" F.). On fresh meat t,hese organisms produced a gray coating which later became yellow. Glage noted that a characteristic aromatic odor, which he considered rather pleasant in the early stages, accompanied the growth of these organisms. As they grew, the surface of the meat became covered with tiny drop-like colonies which increased in size and finally coalesced t o form a slimy coating. Haines ( I 933b) thought that Glage's " Aromobakterien" were identical with, or closely related to, the organisms that he considered responsible for the characteristic "slime" which appears on meat surfaces when stored a t low temperatures. Haines considered that, ('with the exception of a certain number of organisms of the Pseudomonas group and a few Proteus, the bacteria growing on lean meat stored in the range 4-0" C. (40-32" F.) almost all belong t o the Achromobacter group." At the same time, but independently, Empey and Vickery (1933) observed that 95 % of the initial flora of beef, capable of growth a t - 1" C. (30.2" F.), consisted of members of the genus Achromobacter, the remainder were species of Pseudomonas and Micrococcus. During storage the relative numbers of Achromobacter and Pseudomonas increased while those of Micrococcus decreased. Later Empey and Scott (1939a) found that less than 1%of the microbial populations growing on the surfaces of beef at 20" C. (68" F.) were viable a t - 1" C. (30.2" F.) and, although bacteria represented 97 % of the contamination acquired by beef surfaces a t the higher temperature, yeasts and molds made up a greater share of the population a t -1" C. (30.2" F.). These workers considered the four principal genera of lowtemperature bacteria comprising the initial flora of the carcass t o be: Achromobacter 90 % ; Micrococcus, 7 %; Flavobacterium 3 % ; and Pseudomonas, less than 1%.

MICROBIOLOGY O F MEAT ANIMALS

141

More recent reports (Jensen, 1944; Ayres, 1951; Kirsch et al., 1952) have indicated that species of Pseudomonas have a relatively greater importance in causing off-odor and slime than that assigned them by the earlier workers. However, Ayres et al. (1950) pointed out that, owing t o changes in classification made in the most recent edition of Bergey’s Manual (Breed et al., 1918), a number of types of organisms which were previously reported as being Achromobacter, no doubt would be classified as members of the genus Pseudomonas in the present schema since the organisms in question were reported t o have polar flagellation. Later. Kirsch et al. (1952) came t o this same conclusion. b. H a m Souring, Bone S t i n k , or Bone Taint. Ham souring, bone stink, and bone taint are terms used in the industry t o indicate a putrefactive spoilage of particular importance in the deep tissue of large thick pieces such as the hind quarters of pork and beef. The joint fluid and bone marrow, as well as the flesh of hams from dressed hog carcasses, seldom were sterile as early as 45 min. after slaughter (Boyer, 1926). I n 1908, Klein reported a n anaerobic nonspore-forming bacillus from ‘(miscured hams ” which he called Bacillus foedans (Eubacterium foedans). Three years later, McBryde (1911) reported a n entirely different type of organism, Bacillus putrefaciens (C. putrefaciens), t o be the etiologic agent involved. Boyer (1926) isolated both aerobic and anaerobic bacteria; among the spore-forming anaerobes which he identified were Bacillus putrefaciens (Clostridium putrefaciens)) B. histolyticus (C. histolyticum) , B. sporogenes (C. sporogenes), B. tertius (C. tertium), and a n unidentified organism resembling B. oedematicus (C. novyi) . Tucker (1929) implicated C. putriJcum (C. lentoputrescens), whereas Moran and Smith (1929) showed that C. sporogenes could cause ham souring. The unidentified bacterium isolated by Boyer (1926) resembling R. oedematicus (C. novyi), is probably the same as that described by Haines and Scott (1940) and associated with bone taint of beef. Haines (1937) indicated that there were at least two types of bone taint: (1) true “souring,” a n anaerobic production of a volatile, evilsmelling fatty acid or related compound, and ( 2 ) true putrefaction or “green bone.” A (‘sewage-like” odor was considered t o emanate from the synovial fluid from an example of the first type. A large variety of proteolytic anaerobes were implicated as being responsible for the second (Howarth, 1917; Savage, 1918). Jensen and Hess (1941) catalogued various types of ham souring and asserted t h a t salt-tolerant bacteria which grow a t 0 t o 3.3” C. (32 t o 38’ F.) in bone marrow can cause any kind of souring. They named these bacterial genera t o be : Achromobacter, Bacillus, Pseudomonas, Proteus group, Serratia, Clostridium, Micrococcus, streptobacilli, and a miscellaneous group.

142

JOHN C. AYRES

Moran and Smith (1929) pointed out that the experience of the trade has shown that bone taint can be reduced if the animals are well rested and watered and if, after slaughter, the beef sides are cooled as rapidly as possible. c. Black Spot, Whiskers, and Other Mold Discolorations. Talayrach (1901) was one of the first workers to identify molds responsible for the contamination of chilled meats. A few years later Massee (1912) found that discolored black spots on chilled beef and mutton shipped to England from Argentina were due to the fungus threads of Cladosporium herbarum which penetrated the superficial layers of the meat. Brooks and Kidd (1921) made a detailed study of the “black spot” fungus and concluded that, except for its disagreeable appearance, the meat was firm and safe for human consumption. The infection was considered to be entirely superficial. They associated the organism Cladosporium herbarum with the defect and reported that, in addition to beef and mutton, lamb, veal, and rabbit meats were also affected. The mold was shown to be able to grow and produce black spots on meat kept at a temperature several degrees below the freezing point. In addition t o Cladosporium, Brooks and Kidd (1921) mentioned several other molds which could develop on chilled meats. Later, Brooks and Hansford (1923) described the morphology of several strains of molds able to grow on meats and, in certain instances, determined their minimum and maximum temperature requirements for germination and sporulation. In chill rooms having high humidities, the mycelia of molds of the family Mucoraceae growing on the surfaces of meats produce “whiskers.” Rhizopus, Mucor, and Thamnidium commonly are the genera incriminated (Brooks and Kidd, 1921; Brooks and Hansford, 1923; Bidault, 1921; Jensen, 1945). Tomkins and Smith (1932) refer to a sticky condition on the surface of meat caused by extensive growth of molds prior t o their development of aerial hyphae. This condition is not t o be confused with bacterial sliming. I n this country, discolorations caused by the growth of molds on chilled meats are less common than formerly owing to the fact that animal carcasses are seldom aged now for long periods of time in the packing plant. d . Fat Rancidity. Certain of the unpleasant tastes and odors in fat of stored beef were considered by Lea (1931) and by Haines (1933b) to be caused by microorganisms growing either in fatty tissue or in the adjacent muscle. Lea found that the fat of beef carcasses stored in still air a t 0’ C. (32” F.) was good after 25 days but somewhat tainted at 42 days; a tainted odor was present a t 15 days. Later (Lea, 1938) stated that tainted

MICROBIOLOGY O F MEAT ANIMALS

143

fat may contain several million bacteria per gram. Haines (193310) and Vickery (1936a,b) showed t h a t some of their strains of Achromobacter and Pseudomonas and all of the yeasts isolated were lipolytic. Coccoid forms (similar t o Alcaligenes viscosus except that they were almost spherical, nonmotile, and did not produce ropiness) commonly were encountered (Ayres et al., 1950; Ogilvy and Ayres, 1951a) on off-odor or slimy chicken. Although these organisms were comparatively inactive on artificial biochemical media, they showed pronounced ability t o hydrolyze fat and so may play some role in spoilage of meat. Of the Pseudomonas cultures isolated by Sulzbacher and McLean (1951) from fresh pork sausage, 70 % were lipase-forming organisms. 5. Cutting and Storage

After carcasses leave the chill room they are quartered, cut, and trimmed. The sources of contamination-knives, saws, conveyers, tables, air, water, workmen-are similar t o those which the meat came in contact with during evisceration. However, there are a t least three important differences which accentuate the bacteriological loads in the cutting room: (1) cut surfaces and juices support the growth of large bacterial populations, ( 2 ) the microorganisms which have become entrenched and have multiplied during evisceration and in the chill room are redistributed by cutting, and (3) much larger amounts of surface are exposed t o potential contaminants. T h a t some organisms do find cut flesh a satisfactory environment for abundant multiplication even though i t is kept a t refrigeration temperatures was forcefully illustrated by Weinzirl and Newton (1914, 1916). These workers proposed a bacteriological standard of 10 million organisms per gram for ground meats after observing that the earlier standard of not over 1 million suggested by Marxer in 1903 (Weinzirl and Newton, 1914) would result in the condemnation of most ground meats. This situation, according t o Tanner (1944) results from the fact that, since ground meats “may be made from meat scraps and handled carelessly, they are subject t o marked bacterial development because grinding thoroughly distributes bacteria, releases juices, and provides a much larger surface for the bacteria.” According t o Moran (1935) and Mallmann et al. (1940) most of the problems associated with the spoilage of meat are surface problems. Moran and Smith (1929) reported a negligible increase in numbers of organisms in the deep flesh of beef stored for 2 weeks at 5” C. (41” F.), which led Moran t o conclude t h a t spoilage by bacteria in the deeper parts of the flesh is unimportant compared with that a t the surface. This opinion was confirmed by Ayres and Adams (1953) who recovered

144

JOHN C . AYRES

about as many organisms from fresh beef by swabbing a square centimeter of surface with sterile moistened cotton as by mincing a gram of meat having a n equivalent amount of surface. They considered the swab method t o give more duplicable results and t o be easier t o use; difficulty was experienced in obtaining cuts of uniform depth and surface area. On the other hand, these workers obtained larger counts after storage when cut sections of tissue were analyzed. Earlier, Ayres et al. (1950) had pointed out t h a t bacteria may perfuse chicken flesh during storage. Various workers refer t o certain minimal concentrations of bacteria a t the time that incipient spoilage becomes apparent. Schmid (1931) associated the limit of saleability of beef with a bacterial count of 50 t o 100 million per sq. cm. of surface; when bacterial growth was in the advanced stage, meat was slimy t o the touch. Haines (1933b) and Empey and Vickery (1933) reported that the slime point was attained when the surface loads were 32 million and 50 million, respectively. The minimum number of organisms required for sliminess of beef was reported by Moran (1935) t o be 3 million. Although Lea (1931) reported that chilled beef in carcass form could be stored for 60 days a t 0" C. (32" F.), Haines demonstrated that small pieces of lean meat consistently developed "slimes" in from 8 t o 18 days at ' 0 C. (32" F.) even when the surrounding air had a comparatively low humidity. When Haines (1933a) analyzed the '(slime" which developed on meat from animals killed in a poor type slaughter house after storage for about a week in the range 5-0" C. (41-32" F.), he found loads of almost 28 billion organisms per sq. cm. The predominating organisms were Achromobacter (71 %), Pseudomonas (1 1 %), Proteus (1 1 %), and Actinomyces (7 %) . Haines and Smith (1933) presented data in graphical form showing the effect of the initial contamination on the time required for the development of slime on beef stored at 0" C. (32" F.). Ayres (1954) showed the effect of resident populations on the ultimate storage life of meats stored a t 0, 4.4 and 10" C. (32, 40, and 50" F.) (see Fig. 1). It can be noted t h a t the time intervals which elapse before slime develops a t 0" C. (32' F.) in the later study almost coincide with those reported by Haines and Smith. Also, it can be seen that the level of initial contamination has a marked influence on the ultimate storage life of the meat. Initial counts on meats cut or prepared in accordance with sound sanitary practice had relatively small numbers of bacteria when compared with similar items handled less carefully. Ayres (1951) reported that during the first day or two, bacterial counts for meats stored a t 4.4" C. (40" F.) show a n initial decline before microorganisms begin t o proliferate (see Fig. 2). Presumably, this temperature

145

MICROBIOLOGY OF MEAT ANIMALS

may be unsuitable for survival or growth of organisms other than the spoilage types and, in these early hours, insufficient time has elapsed for the psychrophilic organisms t o replace losses of the adventitious flora. Figure 3 illustrates the relationship which exists between temperature of storage and the length of time elapsing before the appearance of slime

,

4

1

1

8

12

1 16

I

20

2d

KEEPING TIME ( D A Y S )

FIG.1. Relation of initial bacterial counts to time required for appearance of slime on fresh beef a t three temperatures (from Ayres, 1954).

on pieces of lean meat (Ayres, 1954). Note that the slope of the curve representing results of recent tests indicates that temperature differentials of only 1-2" C. (1.8-3.6" F.) below 7-8" C. (44.6-46.4' F.) result in significantly longer keeping times for the meat. Also, the improvement in the storage life of beef processed under present sanitary conditions over that of pieces of meat studied by Haines and Smith (1933) is readily seen if the curves for slime points in the two investigations are compared. The longer keeping times of meat sampled in recent trials over those

146

JOHN C . AYRES

analyzed 20 to 25 years ago may be attributed to refinements now in use during the slaughtering and dressing procedure and t o the speed in handling the carcass and of getting cuts into the chill room and cooler under adequate refrigeration. I og--

I00

80

60 w

u

a 40

,"

20

0

TIME

IN D A Y S

FIG.2. Effect of type of initial flora of frozen ground beef trimmings (I) and experimental fresh ground beef (11) on their storage lives. Percentage distribution of organisms b y types at different storage times (from Ayres, 1951).

Samplings made from the native surfacw of 37 matched sets of knuckle, inside round, and outside round of cutter and canner grade beef (Ayres and Adams, 1953) shipped to the Iowa State College laboratory from a Chicago meat packing plant indicated that aerobic loads usually ranged from 10,000 t o 1 million bacteria per sq. cm. or from 100,000 to 10 million per gram. For areas that had been sliced at the packing plant just before the meat was shipped, microbial populations ordinarily varied from 10,000 to 10 million per sq. cm. or per gram. It should be pointed out

147

MICROBIOLOGY OF MEAT ANIMALS

that the surface flora on the native areas was permitted t o develop from the time the animal was skinned whereas organisms on sliced portions may not have been introduced until after the quarter was dissected. The number of aerobes differed considerably among samplings. For example, two surface tests indicated the presence of less than 100 bacteria per sq. cm. whereas two others had 10 million. That i t is not only the size of the initial contamination which determines the time interval before incipient off-odor or slime points become

I

2ok Y

I5

, \\

l

oc

-5 0

I

8 12 I6 K E E P I N G TIME ( D A Y S )

o;

-G

FIG.3. Relation between temperature of storage and the time which elapses before slime is observed on beef (from Ayres, 1954).

apparent is shown in Fig. 2 . Two samples of ground beef, one prepared from trimmings and stored in the frozen state and the other made t o order from chunks of chuck and plate, had almost identical starting loads. Examination of 50 representative colonies isolated from subculture of the frozen ground trimmings (sample I) indicated 78% t o be of the typical spoilage (Achromobacter-Pseudomonas) type while only 18% of the isolates from the freshly ground meat (sample 11) was comprised of typical spoilage forms. Many gram-positive micrococci, sarcinae, and bacilli and short, gram-negative rods as well as other miscellaneous forms made up over 80% of the flora. The two types of comminuted meats were stored in cellophane bags a t 4.4" C. (40" F.) until, by organoleptic

148

JOHN C. AYRES

evaluation, each was considered spoiled, The trimmings (sample I) spoiled in 3 days whereas the fresh ground beef (sample 11)kept an additional 3 days. At the time of spoilage, more than 98% of the Aora from both meats was comprised of typical spoilage forms. The load of cells growing anaerobically was consistently less than that of organisms growing in an atmospheric environment; the proportion of cells which would reproduce without free air represented 15-200/, of the total flora. From 81 samplings, 70% of the determinations showed from 1 to 10 aerobes to each anaerobe; there was an over-all ratio of aerobes to anaerobes of 6 :1. The number of cells or spores that survived a 20-min. heating period at 80" C. (176" F.) and which were still capable of reproducing when grown under atmospheric conditions, differed widely among samples. This group, tentatively classified as aerobic (or facultative) spores, may play a significant role in the spoilage of canned meats and merit considerable study in the future. It would not be advisable to consider that all of the surviving organisms were spores. Various workers (Sulzbacher and McLean, 1951; Ayres, 1951) have demonstrated the presence of species of Microbacterium in meats. Should certain heat-resistant species or strains of this genus be represented, the organisms probably could have survived the heat treatment given. Also, it is quite possible that heat tolerant micrococci, actinomycetes, and other vegetative cells can withstand 80" C. (176" F.) for 20 min. in a meat substrate. Since the technique used for detecting putrefactive anaerobic spores diff ered from the other methods for determining viable organisms, results are not directly comparable. Burke et al. (1950) made several determinations of anaerobic spore loads in pork trimmings and found numbers of spores to average less than 1 spore per gram of meat. Later, Steinkraus (1951) extended this work to include samplings from four commercial packing plants in Iowa and reported that 90% of the samples of fresh pork trimmings which he obtained contained less than 3 spores per gram. The maximum spore count found in any sample tested was 51 spores per gram. About the same number of spores were recovered from meat obtained from various plants. Ninety-nine separate samplings were made of packaged raw beef. I n no case did the putrefactive anaerobic spore count exceed 1.4 spores per gram; in only 11 samples were there more than 0.06 spores per gram. Screening tests of seventy-five cultures of putrefactive sporeforming organisms isolated from fresh pork trimmings and pork luncheon meat indicated 21 of the isolates to be obligate anaerobes (Steinkraus and Ayres, 1951). Species were tentatively identified t o be similar t o the following organisms in Bergey's Manual (6th edition) : Clostridium tetano-

MICROBIOLOGY OF MEAT ANIMALS

149

morphum, C . novyi, C . carnis, C. paraputrijcum, C . tetani, C . histolyticum, and C. sporogenes. Also, an organism having biochemical characteristics similar t o Cameron’s Putrefactive Anaerobe No. 3679 was isolated.

VI. IMPROVEMENTS IN PROCESSING PRACTICES I . Desired Additional Studies

It seems paradoxical, but none the less true, that one of the values which derives from a review article is that, in evaluating the work that has been accomplished, attention is called t o a much larger task which still remains to be done. This is particularly true as it refers t o bacteriological studies which concern the slaughtering and dressing of meat animals. As a general example, it should be mentioned that very little information is available regarding the incidence and types of microorganisms found on pork, mutton, lamb, or veal. Inasmuch as the several animals are exposed to somewhat different environmental conditions, feeding habits, and slaughtering procedures, certain differences in amount and t>ypeof flora remaining on the carcass are to be expected. Many additional studies are required t o determine the role of the animal’s defensive mechanisms in preventing gross contamination of its musculature. Also, there is need for fundamental studies which concern the activity of hyaluronidase, collagenase, and capsular substance in aiding or limiting microbial invasion. Information is desired regarding numbers and types of organisms (1) entering the tissue a t the wound site, (2) coursing through the circulatory system through heart action, and ( 3 ) inhaled during final respiratory activity. There is considerable information concerning the effects of feeding but some of it is controversial. For example, Madsen (1943) mentioned that the intestines of sugar-fed pigs were more easily washed than were those of unfed animals but Gibbons (1953) reported that, in his studies, packing plant personnel found no difference. It would be of interest to know how the types of flora of the fed and unfed animals compared. What influence does acidity produced by metabolic activities of organisms during the degradation of the carbohydrate have on the deposition of the intestinal wall’s slime-layer ? Although Callow (1949) and Ingram (1949) have indicated that fatigue in hogs may render carcasses of these animals more susceptible to bacterial decomposition than they would be otherwise, these conclusions have been deduced or extrapolated from a knowledge that the glycogen reserves in fatigued animals are depleted and that slight acidification of media greatly reduce bacterial growth rates. Although there is good

150

JOHN C. AYRES

probability that the train of events which Ingram outlines does take place, more direct evidence is needed. Also, information concerning the difference in the microbiology of starved ruminant and nonruminant meat animals might be of value in determining the reason that cattle and sheep can be starved longer than can the hog without ill effects. British workers (Bate-Smith, 1937, 193%; Callow, 1949) have explained this phenomenon as being ‘‘expected because the pig is badly out of training, whereas both sheep and cattle usually get plenty of exercise.” No consideration has been given to the enormous reservoir of food in the rumen, t o the digestive lag for these animals, or t o the possibility that microorganisms in the rumen may serve as sources of nutrilites. The use of detergents and sanitizers for cleaning and aiding in mechanically removing of filth and microorganisms from the live animal is deserving of study. Although Empey and Scott (1939a)b) made a n exploratory study of the use of chlorine solutions under laboratory conditions, no further report could be found of its value when tried under field conditions. Since that time, “in-plant chlorination” has been used for many of the other food processing fields and its value either on the hide of the live animal or on the carcass after dressing should be determined. The use of antibiotics in meat processing has been suggested by Weiser et al. (1953, 1954) as a means of preventing deep spoilage, especially prior t o dressing out. I n several exploratory studies that they made regarding the effect of aureomycin, they found th a t the antibotic delayed spoilage of infused meats when they were compared with controls. However, they posed several questions which, among others, will need t o be answered before the use of antibiotics can be considered. For example, the consumer must be assured that the material added is not toxic even though it may be used for longer periods of time. Also, it will be necessary to know whether or not some organisms develop resistance t o the antibiotic upon its continued usage in a foodstuff. Then, too, questions arise as t o the manner and nature of modification of the flora in meats. There is urgent need for published information regarding the incidence of the several types of aerobic and anaerobic microorganisms found in and on meats. I n the past, some question has been raised regarding the value of such observation. Certainly mere counts and taxonomic cataloguing of species encountered on meat does not present sufficient evidence to incriminate or exonerate the product. However, studies which determine not only the numbers of microorganisms present but their contribution to its ultimate spoilage should prove of considerable worth. The construction of some of the pieces of equipment used for eviscerating or cutting should be made more satisfactory from a sanitary

MICROBIOLOGY O F MEAT ANIMALS

151

st,andpoint. For example, band saws now in use harbor much tissue debris. Reciprocating saws (e.g. jig saws) might be devised in such manner t h a t tissue debris does not accumulate and serve t o provide a n inoculum for successive pieces of meat. The design of all equipment should be such that easy and rapid cleaning is possible. More study is needed t o ascertain the significance of pathogens in meats. McCullough et al. (1951a) present evidence t o show that the exposure of packing plant personnel t o Brucella through the handling of hog carcasses presents a n important problem. Further, they (McCullough et al., 1951b) point out that, since Bang’s reactor cattle probably are sent for slaughter because they were no longer profitable t o the owner, such infections may have been of long-term durations. As a mat,ter of record, programs have been initiated which encourage the sale and slaughter of reacting animals. Since Brucella were recovered from numerous sites in a significant number of the animals examined (42 of 100 reactor cattle), these workers considered processing of such carcasses hazardous. Probably of even more grave consequence, is the possibility that meat of reactor animals may pass a clinical inspection and be considered safe for human consumption whereas in fact i t is not. 2 . Suggested Improvements

It is universally recognized that animals should not be fatigued, frightened, chilled, or overheated just prior t o slaughter. Immobilization by asphyxiating with carbon dioxide is a t present in use for hogs; its use for sheep and cattle is suggested as well. Before the animals regain consciousness and prior t o bleeding, the site where the stick-knife is t o enter should be sterilized by flooding or swabbing with a n approved germicidal solution (e.g. a quaternary ammonium compound). Also, the knife blade should be cleansed and sanitized by rinsing in water t o free it of blood, hair, and other debris and immersed in a sanitizing agent before sticking the next animal. After sticking, the practice of allowing the carcass to remain hanging on the rail, free of contact with other carcasses until bleeding is completed and all heart and respiratory action ceases, has much t o recommend it. (Six minutes generally is considered adequate.) It would seem that these operations could be expedited and simplified in such manner t h a t they would not interfere with the routine now in practice. Removal of hair or hide depends upon the type of animal considered. The microbial loads associated with dirty hides, hoofs, and hair provide ample witness t o the need for more adequate cleansing before such animals are killed. Force spraying over the whole area of the hide with a sanitizing agent would serve t o facilitate the removal of loosely adhering

152

JOHN C. AYRES

soil and filth. However, on some animals fecal material gets “caked on” the flanks and feet so tenaciously that it is almost impossible t o wash free unless a laborious and expensive soaking operation is undertaken. Often the animals are in this condition before they reach the packing plant; in many instances livestock accumulate filth owing t o poor bedding on the farm or in the feed lots. It would seem that one way to discourage such insanitary practices would be to downgrade animals if they have noticeable deposits of scurf. An alternative suggestion when there are only a few animals, would be to segregate these animals before they reach the killing floor and t o make further attempts t o clean them or t o handle them separately a t the end of the day’s run immediately before time for the “rlean up.” Cattle and sheep skinning should be done in such a manner that the workmen minimize touching of the exposed carcass with the hair or wool. The hide, after it is cut free, should be removed immediately from the skinning floor. Running water should be available and sanitizing of hands and tools should be enforced when accidental contact is made with dirty hair or wool. Since the hands of the workman are a very important sourcc of contamination, the skinner should be trained not to touch the carcass whenever such contact can be avoided. Prevention of contact with the musculature after handling the lower legs and feet is particularly important. Also, insofar as possible, evisceration of cattle and sheep should be handled separately from skinning or fisting. Wherever possible, mechanical devices such as hooks should be employed and sanitized between animals. Use of the scald tank t o loosen hair and scurf on the hog has little t o recommend i t from a sanitary point of view. Hot-water spraying of individual carcasses in steam cabinets is advocated a t as high a temperature as possible without scalding or reddening the skin. The bung should be plugged and the plug should remain with the viscera or be sterilized before reuse. If the hog is t o be wax dipped after scalding, dehairing, and shaving, the wax should not only be settled free of hair but should be strained or filtered and possibly heated thoroughly before reuse. I n many plants, scalding, dehairing, and waxing are completed before the hogs are permitted on the eviscerating floor. This practice should be required universally. The eviscerator should avoid alternate touching of the meat of the carcass and intestinal or fecal matter. If the esophagus, bladder, bung, or rumen are t o be tied, this should be done by machine or by gloved hand with sanitization between animals. It should be ascertained by examination that all personnel coming into direct contact with any portion of the

MICROBIOLOGY OF MEAT ANIMALS

153

animal which is t o be used for edible purposes are not carriers of any of the enteric pathogens; preferably, the workers should be licensed food handlers. Federal or plant inspectors should not be permitted t o handle or probe viscera and carcass interchangeably. During protracted examination of the organs of the intestinal tract, a ruptured or infected tissue is a n almost certain eventuality. When and if, such occurrences take place, superficial rinsing of the hands is but small safeguard against the contaminative load th a t is transferred to the inspector's hands. If no other method of examination is feasible, the inspectors should work in pairs-the one examining and handling the carcass, the other probing the viscera. Walls of eviscerating rooms should be constructed of glazed tile while floors should be made of smooth concrete or cemented brick or tile. Supporting structures should be of concrete or painted steel. All conduits, water, and heating pipes should be separated from the interior of these floors in order t o permit and expedite force spray washing. Adequate drainage facilities are requisite. Any points wherein off a1 might collect should be kept separate from the lines where meat is moving. Implements used in the eviscerating and cutting rooms should be sanitized whenever practical and possible during operation. Except for chopping or cutting blocks, tables, benches and conveying devices should be made of metal. Wood, when required, should be edge-grained maple or wood of similar porosity. I n any case, surfaces over which the meat must move should be kept as free as possible from bacterial infusion. Tissue debris should not accumulate on saw races. Where practical, ultraviolet irradiation should be employed t o reduce surface contamination and t o permit more rapid tenderization. Circulation of air in the chill room must be assured; there should be no dead pocket areas. Sawdust, when used, should have a n approved fungicide incorporated t o reduce contamination deriving from drippings and from feet of workmen. Use of antibiotics should be evaluated by the industry as a means for reducing contamination on meat surfaces. Finally, extrapolation of the information presented in this review indicates that the nature and number of bacteria found on meat cuts can be used, not only t o determine if the meat is processed under good sanitary conditions, but also to predict its expected storage life. Examination of Fig. 1 reveals that meat cuts with initial loads of 100 bacteria will keep only 5 t o 6 days a t 10" C. (50" F.) but a t 0" C. (32" F.) will not become slimy until stored for more than 14 days. On the other hand, if meats have high bacterial loads, their storage life is only slightly prolonged by the use of lower temperatures. The chill room cannot be regarded as a

154

JOHN C. AYRES

means of controlling the flora on the carcass of animals if prior handling of the carcass has been carelessly done. The significance of the kind and amount of Aora which resides on meat when it leaves the packing plant prompts the author of this article to insist that standards for bacterial numbers are needed both for wholesale and for retail meats. This recommendation is made with full realization that similar suggestions advocated for ground beef by Marxer (1903) (see Weinzirl and Newton, 1914) and by Weinzirl and Newton (1914,1915) failed to gain many proponents. As a matter of record, Hoffstadt (1924) and even Weinzirl (1924) concluded that there was no correlation between freshness of meat and its bacterial content. However, Haines (1937) criticized the views taken by Hoff stadt and Weinzirl since the organisms they isolated and studied (staphylococci and proteolytic anaerobes) do not grow a t temperatures below about 10" C. (50" F.). Workers (Schmid, 1931; Haines, 1933a; Empey and Vickery, 1933; Ayres et al., 1950), who have independently estimated populations on flesh of different animals at the time of slime formation, have obtained counts showing remarkably good agreement with regard to numbers, viz. 32-100 million bacteria per sq. cm., and two types of organisms predominating; Pseudomonas and Achromobacter. Also, since it has been demonstrated that most of the flora is found a t or near the surface, samplings can be performed quite simply through the use of (1) direct smears on glass slides, (2) spot plates, or (3) cotton swabs. Tolerance limits for loads of organisms on and in meats have not been established. However, little imagination is required to estimate the storage life of meats having initial loads of from 1-10 million bacteria per gram. Until such time as the industry has made a concerted study of the customary loads on and in meats, the following counts for wholesale cuts are considered by the author to be reasonable and are proposed for consideration: (1) aerobic population: 10,000 to 100,000 bacteria per sq. cm.; (2) anaerobic population: 5,000 to 50,000 bacteria per gram; (3) most probable number of cells (M.P.N.) surviving 20 min. of heating at 80" C. (176" F.) and growing aerobically: 10-100 per gram; (4) M.P.N. of cells surviving 20 min. of heating at 80" C. (176" F.) and growing anerobically: less than 1 spore per gram. REFERENCES Adami, J. G. 1899. On latent infection and subinfection, and on etiology of hemochromatosis and pernicious anemia. J . Am. Med. Assoc. 33, 1509. Adami, J. G., Abbott, M. E., and Nicholson, F. J. 1899. On the diploccoid form of the colon bacillus. J. Exptl. Med. 4, 349. Adamson, C. A. 1949.A bacteriological study of lymph nodes.Actu Med. Scand. 337,l.

MICROBIOLOGY OF MEAT ANIMALS

155

Amako, T. 1910. Untersuchungen uber das Conradi’sche Oelbad und den Bakteriengelialt der Organe gesunder Tiere. 2. Hyg. Infectionskrankh. 66, 166. Anonymous. 1952. What changed meat packing from a trade to an industry. Natl. Provisioner 126 (4), 197. Anonymous, 19L3. Progress cast in bronze. Food E n g . 26, 43. Arias, C., Burroughs, IV., Gerlaugh, P., and Bethke, R. 31. 1951. The influence of different amounts and sources of energy upon i n oitro urca utilization by rumcn microorganisms. J . A n i m a l Sci. 10, G83. Ayres, J. C. 1951. Some bacteriological aspects of spoilage of self-service meats. Iorcn State College J. Sci. 26, 31. Ayrcs, J. C. 1954. Manuscript in preparation. Ayres, J. C., and Adams, A. T. 1953. Occurrence and nature of bacteria in canned berf. Food Technol. 7, 318. Ayres, J. C., Ogilvy, W.S., aud Stewart, G. F. 1950. Post-mortem changes in storetl meats. I. Microorganisms associated with dewlopment of slinic on eviscrratcti cut-up poultry. Food Technol. 4, 199. Hanfield, F. H. 1935. The electrical resistance of pork and bacon. J . Soc. C h e m I n d . 64, 411T. Fate-Smith, E. C. 1933. Physiology of muscle protein. A n n . Kept. Food Invest. Board ( B r i t . ) , p. 19. Bate-Smith, E. C. 1936. The effect of fatigue on post-mortem changes i n muscle. Dept. Sci. I n d . Research Ann. Rept. Food Invest. Board (Brit.), p. 21. 13:rte-Smith, 13. C . 1937. The special metabolism of the pig. D e p t . Sci. 1nd. Kcwtrrrh ii7an. Rept. Food Invest. Board ( B r i t . ) , p. 22. Ihtc43rnith, 12. C. l938a. The buffering of muscle in rigor: ProttLin, phosphate, : i r i ( l .I. I’hysiol. (London) 92, 336. . C. 1938b. Physiology of rigor mortis. Dept. Sci. I n d . Research .1iin. Rept. Food Invest. Board ( B r i t . ) , p. 15. Bate-Smith, E. C. 1938~.The carbohydrate metabolism of slaughterhouse animals. Dcpt. Sci. I n d . Research dnn. Rept. Food Invest. Board (Brit.),p. 22. Bate-Smith, E. C. 1939. Changes in elasticity of mammalian muscle undergoing rigor mortis. J . Ph?ysiol. (London) 96, 176. Hate-Smith, E. C. 1948. The physiology and chemistry of rigor mortis, with special reference t o thr aging of beef. L4dvances in Food Research 1, 1. I
Bierotti, E., and Machilda, S. 1910. Untersuchungen uber dm Keimgehalt normaler Organe. Miinch. med. Wochschr. 67, 636. Blnmire, R. V. 1962. The capacity of the bovine “stomachs.” V e t . Record 64, 493. I
156

JOXN C. AYRES

Brooks, F. T., and Hansford, B. A. 1923. Mould growths upon cold-store meat. Dept. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No. 17. Brooks, F. T., and Kidd, M. N. 1921. The “black spot” of chilled and frozen meat. Dept. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No. 6. Bugge, A., and Kiessig, A. 1911. Ueber den Keimgehalt der Muskulatur gewerbmiissig geschlachteter, normaler Rinder. 2. Fleisch-u. Milchhyg. 22, 69. Bulloch, W. 1930. History of Bacteriology. A system of bacteriology. Med. Research Council (Brit.) No. 1, 15. Burke, M. V., Steinkraus, K. H., and Ayres, J. C. 1950. Methods for determining the incidence of putrefactive anaerobic spores in meat products. Food Technol. 4, 21. Burn, C. G. 1934a. Post-mortem bacteriology. J . Infectious Diseases 64, 388. Burn, C. G. 193413. Post-mortem bacterial invasion. J . Infectious Diseases 64, 395. Burrows, W. 1949. “Jordan-Burrows Textbook of Bacteriology,” 15th ed., p. 981. Saunders, Philadelphia. Callow, E. H. 1935. The electrical resistance of muscular tissue and its relation to curing. Dept. Sci. Ind. Research Ann. Rept. Food Invest. Board (Brit.), p. 57. Callow, E. H. 1936. Transport by rail and its after-effects on pigs. Dept. Sci. I n d . Research Ann. Repl. Food Invest. Board (Brit.), p. 81. Callow, E. H. 1937. The “ultimate p H ” of muscular tissue. Dept. Sci. I n d . Research Ann. Rept. Food Invest. Board (Brit.), p. 48. Callow, E. H., 1938. The after-effects of fasting. Dept. Sci. I n d . Research Ann. Rept. Food Invest. Board (Brit.), p. 54. Callow, E. H. 1939. The p H of muscular tissue. Dept. Sci. I n d . Research Ann. Rept. Food Invest. Board (Brit.), p. 27. Callow, E. H. 1919. Benjamin Ward Richardson Lecture. Part 1. Hygiene and storage. J . Roy. Sanit. l n s t . 69,35. Callow, E.H., and Boaz, T. G. 1937. Transport by rail and its after-effects on pigs. Dept. Sci. I n d . Research Ann. Rept. Food Invest. Board (Brit.), p. 53. Canfora, M. 1908. Ueber die Latenz der Tetanussporen im tiersichen Organismus. Zentr. BakterioE. orig. 46, 495. Carriere, G., and Vanverts, J. 1899. fitudes sur les 16sions produites par la ligature experimentale ties vaiseaux de la rate. Arch. med. erptl. anat. Pathol. 11, 498. Carroll, E. J., and Hungate, R. E. 1954. The magnitude of the microbial fermentation in the bovine rumen. A p p l . Microbial. 2, 205. Charrin and Roger. 1890. Contribution a 1’ etude experimentale du surmenage; son influence sur l’infection. Arch. Physiol. Norm. et Pathol. 2, 273. Christensen, P. B. 1940. Modern beef chilling. Refrig. Eng. 39, 296. Cobbett, L., and Graham-Smith, G. S. Unpublished observations. Depl. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No. 46, 5. Conradi, H. 1909. Ueber den Keimgehalt normaler Organe. Miinch. med. Wochschr. 66, 1318. Desoubry, M. G., and Porcher, M. C. 1895. D e la presence de microbes d a m le chyle normal chez le chien. Compt. rend. SOC. biol. 47, 101. Dickman, S. R., and Proctor, C. M. 1952. Factors affecting the activity of egg white lysozyme. Arch Biochem. and Biophys. 40, 364. Duggeli, M. 1906. Beitrag zur Kenntnis der Selbsterhitzung des Heues. Naturw. 2. Forst-u. Landwirtsch. 4, 473. Elsden, S. R., Hitchcock, M. W. S.,Marshall, R. A., and Phillipson, A. T. 1946. Volatile acid in the digesta of ruminants and other animals. J . Exptl. Biol. 22, 191. Empey, W. A,, and Scott, W. J. 1939a. Investigation on chilled beef. Part I. Microbial

MICROBIOLOGY OF MEAT ANIMALS

157

contamination acquired in the meatworks. Australia Council Sci. and I n d . Research Bull. No. 126. Empey, W. A., and Scott, W. J. 1939b. Investigation on chilled beef. Part 11. Cooling and storage in the meatworks. Australia Council Sci. and Ind. Research Bull.

No. 129. Empey, W.A., and Vickery, J. R. 1933. The use of carbon dioxide in the storage of chilled beef. Australia J . Council Sci. and I n d . Research 6,p. 233. Empey, W. A., Scott, W. J., and Vickery, J. R. 1934. The export of chilled beef. The preparation of the “idomenous ” shipment at the Brisbane abattoir. Australian J. Council Sci. I n d . Research 7, 73. Ewell, A. W. 1943. Allies of refrigeration in food preservation. Refrig. Eng. 43, 159. Ficker, M. 1905. Ueber den Einfluss des Hungers auf die Bakteriendurchlassigkeit des Intestinaltraktus. Arch. Hyg. 64, 354. Fleming, A. 1922. Remarkable bacteriolytic element found in tissues and secretions. Proc. Roy. SOC.93B,306. von Fodor, J. 1886. Bakterien im Blute lebender Thiere. Arch. Hyg. 4, 130. Ford, W. W. 1901. On the bacteriology of normal organs. J . Hyg. 1, 277. Fraenkel-Conrat, H. 1950. The essential groups of lysozyme. Arch. Biochem. 27, 109. Gall, L. S., Huhtanen, L. C., Saunders, R., and Schmidt, W. 1953. Comparison of rumen flora and environment in roughage us. grain-fed animals. J. Dairy Sci. 36, 587. Gibbons, N. E . 1953. Wiltshire bacon. Advances in Food Research 4, 1. Glage, F. 1901. Ueber die Bedeutung der Aromabakterien fur die Fleischhygiene. 2. Fleisch-u. Milchhyg. 11, 131. Goldsworthy, N. E., and Florey, H. 1930. Some properties of mucus with special reference to its antibacterial function. J . Exptl. Pathol. 11, 192. Gorovitz-Wlassova, L. M., and Grinberg, L. D. 1933. On cryophilic microbes. Experiments. (transl. title Advances R e f r i g . Techno1 (Russian) 1, p. 19; table 5, diagram 2). Abstr. Bull. Intern. Inst. Refrig. 16, p. 2.25, 1934. Grant, M. P. 1953. “Microbiology and Human Progress,” p. 295. Rinehart, N e v York. Griffith, A. S. 1911. The cultural characters and pathogenicity of the cultures isolated. Final Report of the Royal Committee on Human and Animal Tuberculosis. Parts 1 and 2, 1-2, p. 20. Gruber, T. 1909. Die Bakterienflora von Runkelruben, Steckruben, Karrotten von milch wahrend der Stallfutterung und des Weideganges einschliesslich der in Streu, Gras und Kot vorkommenden und deren Mengenverhaltnisse in den 4 Medien. Zentr. Bakteriol. Parasitenk. Abt. 11 22, 401. Grunt, 0. 1912. Beitrag zur Frage des physiologischen Vorkommens von Bakterien im Fleische gesunder Schlachtrinder. 2. Fleisch-u. Milchhyg. 23, 193. Gutierrez, J. 1953. Numbers and characteristics of lactate utilizing organisms in the rumen of cattle. J . Bacterial. 66, 123. Haas, W. 1922. Ueber den Bakteriengehalt des Pfortaderblutes und die Entstehung von Leberabszessen. Langenbecks Arch. klin. Chir. 173, 239. Haines, R. B. 1931. The growth of microorganisms on chilled and frozen meat. J . Sac. Chem. I n d . 60, 2231’. Haines, R. B. 1933a. Observations on the bacterial flora of some slaughterhouses. J . Hyg. 33, 165. Haines, R. B. 1933b. The bacterial flora developing on stored lean meat, especially with regard t o “slimy” meat. J. H?jg. 33, 175.

158

JOHN C. AYRES

Haines, R. B. 1934. The minimum temperatures of growth of some bacteria. J. Hyg. 34, 277. Haines, R.B. 1937. Microbiology in the preservation of animal tissues. Dept. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No.46. Haines, R. B.,and Scott, W. J . 1940. Anaerobic organism associated with “bone taint” in beef. J . Hyg. 40, 154. Haines, R. B., and Smith, E. C. 1933. The storage of meat in small refrigerators. Dept. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No.43. Hartsell, S. E. 1948. The newer knowledge of lysozyme and bacteria. Proc. Indiana Acad. Sci. 67, 44. Hauser, G. 1886. Ueber das Vorkommen von Micro-Organismen im lebendem Gewebe gesunder Thiere. Naunyn-Schmiederberg’s Arch. exptl. pathol. Parmakol. 20, 162. HofTstadt, R. E. 1924. The bacteriological examination of ground beef. Am. J. Hyg. 4, 33. Hoflund, S. 1940. Untersuchungen iiber Storungen in den Funktionen des Wiederkiiuermagens durch Schiidigungen des n. vagus verursacht. Svensk. Veterinartidskr. Suppl. 45. Horn, A. 1910. Ein Beitrag zur Frage des Bakteriengehaltes des Muskelfleisches gesunder und kranker Schlachttiere. 2. Hyg. Infectionskrankh. 8, 424. Horwood, M. P., and Minch, V. A. 1961. The number and types of bacteria found on the hands of food handlers. Food Research 16, 133. Howarth, W. J . 1917. Some matters of interest in the inspection of imported meat. Proc. Cold Storage Ice Assoc. 14, 46. Ilulphers, G. 1934. Skallvatten floran i svinlungan. Nordiska veterinarmotet (Helsingf o r s ) p. 800. Hutterman, W. 1905. Beitrage zur Kenntnis der Bakterienflora im normalen Darmtraktus des Rindes. Kochs Jber. Garungs. 16, 402. Ingram, M. 1928. Fatigue musculaire, p H et proliferation bacterienne dans la viande. Ann. inst. Pastezir. 76, 139. Ingram, M. 1949. Benjamin Ward Richardson Lecture. Part 11. Hygiene and Storage. J. Roy. Sanit. Znst. 69, 39. Jensen, L. B. 1944. Microbiological problems in the preservation of meat. Bacteriol. Revs. 8, 161. Jensen, 1,. B. 1945. “Microbiology of Meats,” 2nd ed., p. 389. Garrard Press, Champaign, Ill. Jensen, L. B., and Hess, W. R. 1941. Ai study of ham souring. Food Research 6, 273. Kern, R. A., Kingkade, M. J., Kern, S. I?., and Bchrens, 0 . K. 1951. Characterization of the action of lysozyme on Staphylococc?ts aureus and on Micrococcus lysodeikticus. J. Bacteriol. 61, 171. JGwch, R. H., Berry, F. E., Baldwin, G . I,., and Foster, E. M. 1952. The bacteriology of refrigerated ground beef. Food Research 17,495. Klein, E. 1908. On the nature and causes of taint in miscured hams. Lancet 174,1832. Lea, C. H: 1931. Chemical changes in the f a t of frozen and chilled beef. 11. Chilled beef. J . SOC.Chem. Znd. 60, 215T. I,ea, C. H. 1938. Rancidity in edible fats. Dept. Sci. I n d . Research Food Invest. Board (Brit.) Special Rept. No. 46. I,rl:evctsky, B. C., Weiser, H. H., and Deatherage, F. E. 1953. A microbiological study of lymph nodes, bone marrow, and muscle tissue obtained from slaughtered cattle. A p p l . Microbiol. 1, 57. Lissauer, M. 1906. Ucber den Rakteriengehalt menschlicher und tierischer Faeces. Arch. Hyg. 68, 136.

MICROBIOLOGY OF MEAT ANIMALS

159

Lohnis, F., and Kuntze, W. 1908. Beitriige zur Kenntnis der Mikroflora des Stalldiingers. Zentr. Bakteriol. Parasitmk., Abt. I I 20, 676. RlcBryde, C.S . 1911. A bacteriological study of ham souring. U . S. Dept. A p . Bureau of Aniiizal I d . Bull. Yo. 132. McCullough, N. B., Eisele, C. W., and Pavelchek, E. 1951a. Survey of brucellosis in slaughtered hogs. Public Health Repts. (U.S.) 66, 205. McCullough, N. B., Eisele, C. W., and Byrne, A. F. 1951b. Incidence and distribution of Brucella abortus in slaughtered Bang’s reactor cattle. Public Health R e p i s . (U.S.) 66, 341. McIntosh, J. A., Sotola, J., Boggs, M. M., Roberts, J., Prouty,‘C. C., and Ensmingrr, M. E. 1942. The relation of ultra-violet light and temperature during aging t o quality of beef. Washington Agr. E x p t . Sta. Bull. 422, 26. Madsen, J. 1943. Investigation on the keeping qualities of meat from sugar-fed pig.. Nord. Jordbrugsforskn. (5-6), 340. Mallmann, W.L., Zaikowski, L., and Ruster, M., 1910. The effect of carbon dioxide on bacteria with particular reference t o food poisoning organisms. M i c h . Agr. E’xpl. Sta. Bull. No. 489,25. Massec, G . 1012. On the discoloured spots sometimes present on chilled beef, with special reference to “black spot.” J . Hug. 12, 489. Jlasson, M. 1953. Microscopic studies of the alimentary microorganisms of the sheep. Brit. J . A-utrilion. 4, 2-3, viii-ix. Mcnkin, V. 1910. “The Dynamics of Inflammation.” Xacmillan, S e w York. Mcssnrr, A,, 1910. Ein Beitrag zur bakteriologischen Fleischuntersuchung mit Berucksichtigung der Verhaltnisse in der praktischen Fleischbeschau. Tiertirztl. Zentr. 28, p. 31. Monvoisin, M. 1918. Les moisissures des viandes congelces. IZec. m6d. vBt. 94, 306. hloran, T. 1935. Post-mortem and refrigeration changes in meat. J . SOC.Chenz. I n d . 64, 149T. Xloran, T., and Smith, E. C. 1929. Post-mortem changes in animal tissues. The conditioning or ripening of beef. Dept. Sci. Ind. Research Food Invest. Board ( B r i t . ) Special Rept. No. 36. Moran, T., Smith, E. C., and Tomkins, R. G. 1932. The inhibition of mould growth on meat by carbon dioxide. J . SOC.Chem. I n d . 61, 114T. Morrison, R. h.,and Hooker, D. R. 1915. The vascular tone and the distribution of blood in surgical shock. Am. J . Physiol. 37, 86. Murphy, H. W. 1950. Process for immobilizing livestock prior t o slaughtering. U. S. Patent 2,526,037. Seisser, 11. 1896. Ueber die DurchgCngigkeit der Darmwand fur Bakterien. Z . H y g . Iufeclionskrankh. 22, 12. Sicholls, -4.G. 1904. A simple method of demonstrating the presence of bacteria in the mesentery of normal animals. J . E x p t l . M e d . 11, 455. Socard, 31. 1895. Influence des repas sur la penetration des microbes dans le sang. Seniaine m8d. 16, 63. Norris, C., and Pappenheimer, M. 1905. A study of pneumococci and allied organisnis . 7, 450. in human mouths and lungs after death. J . E x ~ L ZMed. Suttall, G. H. F., and Thierfelder, H. 1895. Thierisches Leben ohne Bakterien im T’erdauungskanal. Z . Physiol. Chem. 21, 109. Ogilvy, li-.S., and Ayres, J. C. I95la. Post-mortem changes in stored meats. 11. T h e effect of atmospheres containing carbon dioxide in prolonging the storage life of cut-up chicken. Food Technol. 6, 97.

160

J O H N C. AYRES

Ogilvy, W. S., and Ayres, J. C. 1951b. Post-mortem changes in stored meats. 111.The effect of atmospheres containing carbon dioxide in prolonging the storage life of frankfurters. Food Technol. 6, 300. Opitz, E. 1898. Beitrage zur Frage der Durchgangigkeit von Darm und Nieren fur Bakterien. 2. Hyg. Infectionskrankh. 29, 505. Pearson, R. M., and Smith, J. A. B. 1943. The utilization of urea in the bovine rumen. 3. The synthesis and breakdown of protein in rumen ingesta. Biochem. J. 37, 153. Petersen, W. F., Miiller, E. F., and Boikan, W. 1927. The reaction of the dog to continuous intravenous injection of B. coli. J. Infectious Diseases 41, 405. Pounden, W. D., and Hibbs, J. W. 1948. The influence of the ratio of grain to hay in the ration of dairy calves on certain rumen microorganisms. J . D u ~ Sci. T ~ 31,1051. Reith, A. F. 1926. Bacteria in the muscular tissues and blood of apparently normal animals. J. Bacteriol. 12, 367. Reyniers, J. A. 1943. “ Micrurgical and Germ-Free Methods.” C. C Thomas, Springfield, Ill. Reyniers, J. A. 1946. Germ-free life applied to nutrition studies. Lobund Repts. No. 1, 87. Reyniers, J. A. 1949. Some observations on rearing laboratory vertebrates germ-free. PTOC. N . Y . State Assoc. of Public Health Labs. 28, 60. Reyniers, J. A., Trexler, P. C., Erwin, R. F., Wagner, M., Luckey, T. D., and Gordon, G. A. 1946. Germ-free life studies. Lobund Repts. KO.1, 120 Reyniers, J. A., Trexler, P. C., Erwin, R. F., Wagner, M., Luckey, T. D., and Gordon, G. A. 1949. Germ-free life studies. Lobund Repts. KO.2, 162. Richardson, J. H., Del Guidice, V. J., and Wiesman, C. K. 1954. Hazards of spores in floor coverings-sawdust. A p p l . Microbiol. 2, 177. Sarles, W. B., Frazier, W. C., Wilson, J. B., and Knight, S. G. 1951. “Microbiology: General and Applied.” Harper, New York. Savage, W. G. 1918. Further investigations on the distribution of Gaertner group bacilli in domestic and other animals. J . Hyg. 17, 34. Savage, W. G. 1920. “Food Poisoning and Food Infections.” Cambridge Univ. Press, London. Schmid, W. 1931. Einfluss von Temperatur und Feuchtigkeit auf das Bakterienwachstun auf gekiihltem Fleisch. 2. ges Kalte I n d . 38, 1. Scott, W. J. 1936. The growth of microorganisms on ox muscle. I. The influence of watcr content of substrate on rate of growth at -1“ C. J. Council Sci. Ind. Research 9, 177. Selter, H. 1906. Bakterien im gesunden Korpergewebe und deren Eintrittspforten. 2. Hyg. Infectionskrankh. 64, 363. Sisson, S. 1914. “The Anatomy of the Domestic Animal,” 2nd ed. Saunders, Philadelphia.

Skinner, C. E., Emmons, C . W., and Tsuchiya, H. XI. 1947. “Henrici’s Molds, Yeasts and Actinomycetes,” 2nd ed. Wiley, New York. Slater, L. E . 1952. Hormel perfects painless kill. Food Eng. 24 (lo), 90. Sotola, J., McIntosh, J. A., Prouty, C. C., Dobie, J. B., Ensminger, M. E., and McGregor, M. A. 1943. The relation of ultra-violet light and temperature during aging to quality of beef. 11. Utility grade short loins. Wash. Agr. E z p t . Sta. Bull. No. 431, 16. Spray, R. S. 1922. The bacteria in normal and diseased lungs of swine. J. Infectious Diseases 31, 10. Steinkraus, K.H. 1951. Putrefactive anaerobic spores in meats. Unpublished Ph. D. Thesis. Iowa State College Library, Ames, Iowa.

MICROBIOLOGY O F MEAT ANIMALS

161

Steinkraus, K. H., and Ayres, J. C. 1951. Biochemical and serological relationships of putrefactive anaerobes isolated from meat. Bacterial Proc., p. 19. Sulzbacher, W. L., and McLean, R. A. 1951. The bacterial flora of fresh pork sausage. Food Technol. 6, 7. Talayrach, XI. J. 1901. Conserves de viandes par les procedes frigorifiques. A n n . hyg. publ. el mbd. lbgale 46, 166. Tanner, F. W. 1944. “The Microbiology of Foods,” 2nd ed. Garrard Press, Champaign, Ill. Tarozzi, G. 1906a. Ueber das Latentleben der Tetanussporen im tierischen Organismus. Zentr. Bakteriol. Parasitenk. Abt. I . 40, 305. Tarozzi, G. 1906b. Ueber das Latentleben der Tetanussporen im tierischen Organismus. Zentr. Bakteriol. Parasitenk. Abt. I . 40, 451. Thole, F. 1912. Die verletzungen der Leber und der Gallenwege. Neue Deut. Chir. 4, 1. Tomkins, R. G., and Smith, E. C. 1932. “Stickiness” and growth of moulds. Cold Storage 36, 56. Tucker, W. H. 1929. Studies on C2. putrijicum and Cl. putrefaciens. Institute of American Meat Packers, Chicago, Ill. Vickery, J. R. 1936a. The action of microorganisms on fat I. The hydrolysis of beef fat by some bacteria and yeasts tolerating low temperatures. J . Council Sci. Znd. Research 9, 107. Vickery, J. R. 1936b. The action of microorganisms on fat 11. A note on the lipolytic activities of further strains of microorganisms tolerating low temperatures. 1. Council Sci. I n d . Research 9, 196. Volz, F. E., Gortner, W. A., Pits, C. W., and Miller, J. I . 1949. The effect of ultraviolet light in the meat cooler on the keeping quality of frozen pork. Food Technol. 3, 4. IVeinzirl, J. 1924. Concerning the relation of the bacterial count to the putrefaction of meat. Am. J . Public Health 14, 946. Weinzirl, J., and Newton, E. B. 1914. Bacteriological analyses of hamburger steak with reference t o sanitary standards. Am. J . Public Health 4, 413. Weinzirl, J., and Newton, E. B. 1915. The fate of bacteria in frozen meat held in cold storage and its bearing on a bacteriological standard for condemnation. Am. J . Public Health 6, 833. Weiser, H. H., Goldberg, H. S., Cahill, V. R., Kunkle, L. E., and Deatherage, F . E. 1953. Observations on fresh meat processed by the infusion of antibiotics. Food Technol. 7, 495. Weiser, H. H., Kunkle, L. E., and Deatherage, F. E. 1954. The use of antibiotics in meat processing. A p p l . Microbiol. 2, 88. TVhite, J. H., and Patrick, R. 1946. Quartermaster Food and Container Institute for the Armed Forces. 6 Part 11. Fresh Meats, p. 28. Winogradowa-Fedorawa, T., and Winogradow, M. 1920. Zalungsmethode der Gesamtzahl der im Wiederkauermagen lebenden Infusorien. Zentr. Bakteriol. Parasitenk. Abt. I I . 78, 246. Wolbach, S. B., and Saiki, T. 1909. A new anaerobic sporebearing bacterium commonly present in the livers of healthy dogs, and believed t o be responsible for many changes attributed to aseptic autolysis of liver tissues. J . Med. Research 21, 267. Wyssokowitsch, W. 1886. Ueber die Schicksale der ine Blut injizierten Xfikroorganismen im Korper der Warmbltitler. 2. Hyg. 1, 1. Zwick, A.,and Wiechel, A. 1911. Zur Frage des Vorkommensvon Bakterien im Fleische normaler Schlachttiere und zur technik der bakteriologischen Fleischbeschau bei Notschlachtungen. Arb. kaiserl. Gesundh. 38, 327.

This Page Intentionally Left Blank

Microbiological Problems of Frozen Food Products

BY GEORG BORGSTROM Sicedish Institute f o r Food Preservation Research ( S I K ) ,Goteborg, Sweden Page I. Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 . . 164 11. The Influence of Freezing Temperatures on Microorganisms. 111. The Influence of the Freezing R a t e . . . . . . . . . . . . . . . . . . . . . . . . 168 170 IV. The Freezing Death of Bacteria.. . . . . . . . . . . . . . . . . . . 172 V. Occurrence of Bacteria in Frozen Foods.. . . . . . . . . . . . 1. Survey of Problems.. . . . . . . . . . . . . . . . . . . . . . . . . 172 2. Microbiological Methods of Analysis. . . . . . . . . . . . . . . 173 a. Methods for Frozen Vegetables.. . . . . . . . . . . 174 b. Methods for Frozen Fruits.. . . . . . . . . . . . . . . 1 ii c . Methods for Eggs.. . . . . . . . . . . . . . . . . . . . . . . 157 179 3. Results of Examinations oE Frozen F o o d s . . . . . . . . a. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 b. Fruits and Fruit Juices.. . . . . . . . . . . . . . . . 179 c. Vegetables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 d. Summary of Discussion on Fruits and Vegetables . . . . . . . . . . 187 e. Meat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 f. Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 g. Whalemeat.. . . . . . . . . . . . . . . . . . . . . . . . 191 h. Poultry.. . . . . . . . . . . . . . . . . . . . . . . . . . . 192 i. Eggs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 197 j . Dairy Products and Miscellaneous. . . . . . k. Bread Dough.. . . . . . . . . . . . 1 !I7 1. Precooked Frozen Foods. . . . . . . . . . . . 198 VI. Pathogenic Bacteria in Frozen Foods.. . . . . . . . . 201 VII. Defrosting Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . 204 V I I I . Packaging Problems. . . . . . . . . . . . . . . 209 I X . Cooking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 X . Hygienic Aspects.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 XI. Practical Aspects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 1 R.eferences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 ,

I. INTRODUCTIOS Frozen foods are constantly gaining new markets. An account of the microbiological problems raised hy this kind of food is, therefore, appropriate a t this time. Frozen foods are not sterile in the same way as are canned products, for they are not subjected t o the same lethal tempera163

164

GEORG BORGSTROM

ture treatment as are foods sterilized by heat, though a n absolute sterility, as a rule, is not even reached through this treatment. The large number of bacteria in a frozen product is of great importance, especially while thawing. This report will be concerned mostly with frozen products with one important exception: ice cream, which has been excluded from this survey, for it requires special consideration. 11. THE INFLUENCE OF FREEZING TEMPERATURES O N MICROORGANISMS

It is a common belief that microorganisms are effectively killed by freezing. They are destroyed t o a certain extent but far from completely. Bacteria and fungi in the vegetative stage are sentitive t o low temperatures and generally succumb. This effect however is not always immediate, and may require long periods of freezing storage. On the other hand, resistance is greatest in the spore stage. One decisive factor is the freezing temperature. Especially sensitive are yeasts and molds, which occur commonly on berries and vegetables. Soil bacteria, which as a rule survive the freezing temperatures occurring in nature, are, however, not very susceptible t o cold. Several species are extremely psychrophilic and even manage to grow at -7" C. (20" F.) (Prescott et al., 1932), which otherwise exerts a definite lethal effect on most bacteria. A unique observation, t o which there is as yet no explanation, is the ability of intestinal bacteria t o endure the acid medium of orange juice better a t -4" C. (25" F.) than a t room temperature (Beard and Cleary, 1932). It is a n often overlooked fact that temperature alone is not the decisive factor with respect t o the degree of destruction. It is the formation of ice crystals (actual freezing of the product) which has the more profound and disturbing influence on bacterial growth. Supercooling may take pIace without any substantial change in the colloidal condition or tissue structure. Studies on supercooling both in fish and in meat have shown very slight effects, irrespective of the retardation of growth. It may be the desiccation of the substrate and not temperature as such, which sets the lower limit for the growth of microorganisms in tissue. This is indicated in studies on meat and fish, and from many observations in which growth has been observed in media supercooled t o as low as -20" C. (-4" F.). When meat is frozen and is allowed t o reach equilibrium at a certain temperature, only a certain proportion of the water separates as ice. The amount of bound water and the degree of protein denaturation determines the extent to which ice is formed. Moran (1931) reported the following:

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

Temperature -3" C. (27" F.) -5" C. (23" F.) -10' C. (14" F.)

165

Percentage of total water present as ice 70 82 94

Most papers in this field make no clear distinction between freezing rate and freezing temperature. The most efficient way of accelerating freezing rate-the advance of the freezing front in the product-is t o use low temperature. This implies the need of distinguishing between the freezing temperature as i t is measured outside the product-the external temperature a t which freezing takes place on one hand, and on the other the internal temperature a t which actual freezing takes place in the product. A similar distinction should be made regarding thawing temperature, which in publications has a two-fold meaning: either the internal temperature a t which defrosting of the product actually takes place, or the external temperature condition under which thawing is carried through. A critical scrutiny of available publications in the field showed that very few studies could be accepted as providing information on the effert of freezing (thawing) temperature and freezing (thawing) rates. The corresponding increase in the concentration of the solutes may on the other hand obviously exercise some kind of a protective influence on surviving bacteria enabling them to avoid actual freezing. More studies are needed t o clarify this fundamental point. Several studies have also indicated that far more microorganisms are destroyed a t -4" C. (25°F.) than a t - 15" C. (5" F.) or - 24" C. ( - 11"F.). In most cases for temperatures below - 10" C. (14" F.) the lower the temperature, the less effectively are the bacteria killed. It can also be said that temperatures below -24" C. (-11' F.) and even as low as -193" C. (-315" F.) have no additional effect (Luyet and Gehenio, 1940; Swift, 1937). The finding that low freezing temperatures -20" C. (-4" F.) are less harmful t o microorganisms than the medium range of temperatures such as -10" C. (14" F.) (Campbell, 1932; Berry, 1934; Haines, 1934-38; McFarlane, 1940a,b; Weiser, 1951; Gotlib, 1951; Hucker et al., 1952) is highly important t o an understanding of the microbiology of frozen food (see Tables I and 11). Thus for example, less than 1% of microorganisms in beans survived a storage temperature of -10" C. (14" F.) while 6% survived -21" C. (-6" F.). This also implies the strange consequence t h a t freezing is bactericidal t o the greatest extent if it takes place slowly and the products are afterwards stored a t a comparatively high temperature (Haines, 1938). These findings are quite contrary t o the demand of

166

GEORG BORGSTROM TA4BLE 1

Per Cent Survival of Bacteria a t Different Subzero Temperatures" Length of freezing storage (days) Storage temperature

0

115

-10" C. (14" F.) -15' C. (7" F.) -20" C. (-4' F.)

100 100 100

6.1 16.8 50.7

(1

178

192

206

220

per cent survival 2.1 3.6 3.9 10.4 57.4 61.0

2.1 10.0 55.0

2.5 8.2 53.2

From Gotlib (1951).

TABLE I1 Survival Time of Certain Pathogenic Bacteria in Sliced, Sweetened Strawberries Stored a t -18" C. (-4" F.)a Bacteria per gram before freezingb

Bacteria per gram after freezing

Culture hr' culture Eberthella typhosa Salmonella paratyphi Salmonella sehottmulleri Salmonella aert rycke Staphylococcus aureus a

b c

I 1 I week

8 m A t h m 2 t h e months

---

1040"

1420

3650

5100

3700

5150

980

2600

310OC

2100

112,000

194,000

33, OOOc

248,000

214,000

152,000

8900

89,000

232,000

20O,00Oc

900W

151,000

335,000

200,000~ 41,000" 47,00OC

13,50OC 625W

135,0003,810,0001,140,000 485,000 86,000 42,000 24,000

From McCIeskey and Christopher, 1941. Plate counts u-ere inade after the berries had hecn kept in crushed ice for 5 hrs. Showed the existewe of pathogenic bacteria in the berries.

modern freezing technique, which dictates a rapid freezing and subsequent storage a t low temperature. I n spite of the pronounced lethal effect of temperatures in the range of 0' C. (32" F.) down t o almost -10" C. (14" F.), there are substantial data t o testify t o the capacity of several microorganisms t o grow even under these conditions. The cause for this dualism in the reaction of the living protoplasm has so far not been unveiled. A great number of barteria, generally termed psychrophilic, can grow a t subzero temperatures even down t o -8" C. (ISo F.) (Haines, 1934; Tschistjakow and Botscharowa, 1938). I n most cases, growth is checked by temperatures

MICROBIOLOGICAL

167

PROBLEMS OF FROZEN FOOD PRODUCTS

TABLEI11 The Effect of Increased Sucrose Concentration on Survival of Bacteriaa Temperatures Duration of experi- -18" C. (0" F.) ment in days 2% 20%

10 20 30 40 50 '6

10,390 15,000 3,560 1,030 2,780 1,580 4,495 836 1,900

-

-12" c. -8" c. (10" F.) (10" F.) Percentage of sugar in bouillon 2% 20% 2% 20% Xumber of bacteria in 100 ml. 2,485 1,130 203 76 13

6,245 5,975 2,192 1,940 2,457

3,240 11,640 622 5,150 240 11,700 285 5,510 890 240

-5" C. (23" F.)

2%

20 %

211,000 3,128,000 62,000 3,620,000 24,500 2,202,000 3,500 28,500 850 54,900

Prom Gotlib (1951).

below - 10" C. (14" F.). Some molds even manage t o grow a t this temperature (Bidault, 1922). In 1941, McCleskey and Christopher proved that Staphylococcus nureus as well as some Salmonella species still survived in unsliced strawberries after being stored for fourteen months a t - 18" C. (0" F.). Freezing storage consequently does not iipasteurize" a product, if it contains bacteria. A vast number of papers prior t o this discovery, and subsequently, have given overwhelming proof of the wide range of organisms which manage t o survive in frozen storage, even for extremely long periods. The reason why temperatures higher than - 10" C. (14" F.) are more lethal t o microorganisms is most likely connected with the processes of protein deiiaturation being far less disastrous a t lower temperatures. I n frrrzing mammalian and fish muscles, Moran (1935) and Reay (1933) drew attention t o the rapid denaturation of the protein which occurs a t temperatures just below the freezing point of the muscle and in the region of - 2 t o 4 O C. (28-39" F.). They attributed the rapid death of microorganisms in this temperature range t o the fact that, within 8 days a t - 2 " C. (28O F.), half of the coagulable protein of the organisms was precipitated. At - 20" C . ( - 4" F.) no such change took place. Death of frozen microorganisms consequently may be ascribed t o the denaturation of the protein and subsequent flocculation of the cellular proteins. This concept entirely refutes the idea t h a t death is due t o the mechanical action of ice crystals. Nor is it likely t h a t intracellular ice crystals cause destruction as death is most effective a t -2" C. (28" F.) where such

168

GEORG BORGSTROM

intracellular crystals should not form because of the salts in the cell contents (Weiser and Osterud, 1945). 111.

INFLUENCE OF

THE

FREEZING RATE

When freezing extremely rapidly (e.g., by immersion in liquid air), the effect on bacteria is so small that their count is at the most only slightly diminished (van Eseltine et al., 1948). Some sensitive species of certain Pseudomonas may in this respect represent important exceptions (Ingram, 1951; Borgstrom, unpublished data). This is explained by the discovery that even a living tissue can be kept in a frozen state because of the fact that no ice crystals are formed when freezing is sufficiently rapid (Luyet and Gehenio, 1940). The protoplasm is transformed into a vitreous state. I n water solutions a slow freezing may sometimes cause an unusual result. During this process practically pure water in the form of ice crystals is extracted from the cell solution, and the bacteria become concentrated in the remaining solution. I n this case the microorganisms survive and even continue t o grow in the remaining unfrozen solution. Slow freezing consequently might allow far more bacteria t o survive than rapid freezing, in clear contradiction t o what happens in most cases (Hess, 1934a,b). Along such lines, a n explanation may be given t o why Haines (1938) has been able in his studies t o establish t h a t the freezing rate is of no importance. The rate of freezing a t temperatures as low as -20' C. (-4' F.) affect the mortality rates of bacteria and yeasts, but there are certain differences with respect t o the effect. Escherichia coli and Lactobacillus casei were found t o be more readily destroyed by slow than quick freezing, whereas the opposite effect was observed with Micrococcus pyogenes v. aureus and some yeasts (Devik and Ulrich, 1949). The freezing rate has a great influence on the number of bacteria developing before the temperature inhibits further growth. Packaging in highly insulated material or freezing in still air, as often happens in lockers or home freezers, might give products with a higher bacterial count (van Eseltine et al., 1948). Generally, however, this is no reason for alarm. The bacterial count is seldom sufficiently high t o seriously impair the quality. Mortality curves for frozen foods, stored at a constant, defined temperature, indicate that during the initial stage, when the bacterial count is decreasing comparatively quickly, there is a definite relationship between death rate and viable bacteria count. During later stages of frozen storage the bacterial count decreases less rapidly. Gunderson and Rose (1948b) found t h a t the number of bacterial colonies (bacterial populations) remained more or less constant when stored at -255' C. ( - 13' F.). It is of special interest that pathogenic bacteria seem t o have

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

169

less resistance t o freezing storage than saprophytic types (Anonymous, 1946a). Greater numbers of viable cells are destroyed in the first 24 hr. of storage after freezing than in any other similar period, except for Staphylococcus aureus which may remain unchanged for at least 55 days (Ulrich and Halvorson, 1947). Of greatest importance in connection with precooked frozen foods is the circumstance t h a t substrates containing fats and sugar seem t o protect bacteria (James, 1933a; McFarlane, 1942; Gotlib, 1951). Table 111. shows the protective effect of 20% sucrose as compared with 2%. I n an unsweetened frozen orange juice 97% of the organisms were destroyed within 48 hr., whereas large numbers of cells were still viable after 26 weeks of storage in juice containing 4Ck50% sucrose. I n frozen eggs bacteria are able t o survive for a very long time (McFarlane and Goresline, 1943; Wallace and Baumgartner, 1936). Freezing temperatures, although not necessarily lethal t o microorganisms, may markedly influence the action of deleterious agents present. There is a n appreciable enhancement of the bactericidal effect of the hydrogen ion by low temperatures (Beard and Cleary, 1932). Thus in an acid reaction the mortality rate is higher than in a n alkaline one (Stille, 1950). This is favorable in so far as frozen fruits, berries and their juices are concerned (Berry, 1932a,b and c ; Hahn and Appleman, 195213). Of course the more concentrated the syrup or the juice, the lower is the freezing point, but more bacteria survive the lower freezing temperatures (Woodroof, 1931; Lutz et al., 1932; McFarlane, 1940a,b). Keith (1913) showed that if certain substances such as sugar, milk, and glycerol were added t o the suspending medium, bacteria were partially protected from the damage of freezing at -20' C. Recent observations on a similar effect of glycerol on spermatozoa (Polge et al., 1949; Smith and Polge, 1950) is the basis of the present new developmental trend in artificial insemination. Glycerol is also able t o prevent the damage t o bacteria (4 species studied) in freezing and thawing, but does not seem t o exert any influence during storage (Hollander et al., 1954). This protective effect may be due t o its interference with the expansion of water, when freezing t o ice (Hollander e2 al., 1954). The recent discovery that bacteria, particularly E. coli, inactivated by ultraviolet radiation, may be reactivated later seems t o be of little importance t o the frozen food industry even though it has been proved that such an inactivation may take place a t subfreezing temperatures. I n this Case reactivation occurs only in the liquid phase (Heinmets and Taylor, 1951). These peculiar reactions may in special cases affect small samples

170

GEORG BORGSTROM

treated in the daylight. I n most cases, however, all light influences can be ruled out with respect t o frozen food packages. I n summary, it may be stated that the bacteria and spore killing effect of low temperatures implies that frozen foods generally contain fewer bacteria than the corresponding fresh products. This, however, is valid only if the frozen products are actually stored a t a temperature guaranteeing an unthamed condition.

IT.FREEZING DEATH

OF

BACTERIA

Critical reviews on the death of bacteria a t low temperatures are few, and much data remain widely scattered in the literature among other subjects. Only a brief summary of some of the most important publications will be given here, for an extensive review of the subject is beyond the scope of this paper. Some of the authors who have presented the most extensive bibliographies on the subject are Prudden, 1887; Sedgwick and Winslow, 1902; Smith and Swingle, 1905; Keith, 1913; Hilliard and Davis, 1918; Vass, 1919; Hampil, 1932; Wallace and Tanner, 1933; Turner and Brayton, 1939; Kyes and Potter, 1939; Luyet and Gehenio, 1940; and Stille, 1942. Different phases of frozen food niicrobiology have been reviewed by Mattick (1952), Ingram (1951), and Davis (1951). As the result of investigations involving experiments on supercooling versus freezing, repeated freezing, and storage of frozen water suspensions of various bacteria, Prudden (1887) concluded that the marked initial killing following freezing is due to the immediate killing of the “more feeble” bacteria. He reported that supercooling was even more destructive than freezing and that very low temperatures were more destructive than the higher freezing temperatures. He also noted the gradual death of organisms during storage, a higher mortality with repeated freezing, and also a low resistance of old cultures of Staphylococcus aureus t o freezing. Most of these observations are contrary to present conceptions and may be due t o deficient experimental conditions. The most common belief for a long period of time was that bacteria died when crushed by the formation of extracellular ice crystals, that is t o say, mechanical death (Keith, 1913). This was disproved chiefly by the studies of Haines (1938) showing that a slow freezing rate, giving larger ice crystals, was not more destructive than rapid freezing. Only the temperature of subsequent storage (in the frozen state) had a decisive influence on survival. This mechanical theory received a late supporter in Gaebelein (1940) claiming that ice crystals a t -4” C. (25’ F.) caused a maximum disruption of the cells. Recently evidence was presented t o the effect that the death of bac-

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

171

teria by freezing involves a rapid action or "immediate" death, caused by freezing and thawing per se, and a "storage" death, which is a direct function of time and temperature. The immediate death seems t o result principally from the mechanical action of extracellular ice (Weiser and Osterud, 1945). No ice is formed within the bacterial cells, even at very low temperatures. I n such cases water solidifies t o a vitreous state. Hollander et al. (1954) suggest that death which occurs during the freezing and thawing of bacteria is due t o mechanical compression, a consequence of the fact that water expands 9 % when it changes into ice. The killing effect of freezing temperature is complicated by the diverse reactions of water t o subzero temperatures. Under ordinary circumstances it crystallizes t o form ice. However, if cooled t o very low temperatures under special conditions, it may solidify without crystallizing, a state analogous t o glass and termed vitreous. The vitrification of pure water is very difficult t o accomplish, but has been reported by Hawkes (1929) and by Burton and Oliver (1935). Any extensive discussion of the factors influencing the crystallization, vitrification, and devitrification of water is not within the scope of this review. These subjects are well treated in such works as those by Luyet and Gehcnio (1940) and Dorsey (1940). Briefly, vitrification of any aqueous solution can only be accomplished by reducing the temperature through the zone a t which freezing occurs so rapidly that there is insufficient time for crystals t o form. With ordinary water, the velocity of the formation of ice crystal nuclei, and of crystal growth, is so great that vitrification is seldom accomplished. The vitrification point is a t - 110" C. ( - 166" F.) and consequently has no relevance t o ordinary frozen foods, lout most likely explains the survival even a t temperatures quite close t o the absolute limit (120" K.) as was proved by Becquerel (1950). As already stated denaturation of the protein must be taken into serious consideration as a causative factor of the death of the microorganisms (van den Broek, 1949) whatever the mechanism of this denaturation is. If the destruction of bacteria proceeds in the same way, in subsequent long-range storage, it is so far not established. An indication that this might be so is the fact that most frozen products, when stored, show a greater reduction in viable counts when held a t temperatures between -2" C. (28" F.) and -10" C. (14" F.) than between -15" C. (5" F.) and -20" C. (-4" F.) (Hartsell, 1949; Hucker et al., 1952). This gradual destruction may have importance t o sanitary control and t o the hygienic evaluation of a product. I n the following, many examples will be given of survival of bacteria during freezing storage. There are several indications of a differing resistance t o such conditions. At any rate freezing storage

172

GEORG BORGSTROM

is now used as a method of maintaining stable infectious bacterial collections, which are still fully viable after two years of such storage (Yurchenco et al., 1954).

V. OCCURRENCE OF BACTERIA IN FROZEN FOODS 1. Survey of Problems

The factors influencing the bacterial content of prepackaged frozen foods may be summarized in the following manner: (1) The treatment of the raw product including the number of bacteria originally present in the raw product, the manner and rapidity of handling between harvesting (or slaughter) and processing, the processing methods, and the hygienic conditions in factories with machines and equipment. ( 2 ) The freezing rate. (3) The amount of oxygen present in the package. (4)The microbiological conditions in packages for frozen foods. (5) The storage temperature. (6) The p H of the product. (7) The presence of osmotic substances in frozen foods. (8) Defrosting. An unexpected high bacterial content in frozen foods may under certain conditions be apprehended. On the whole this may be due t o the following four causes: (1) High original content of bacteria, possibly a n initial decomposition in the raw products. ( 2 ) Delay in the freezing process. (3) Slow or imperfect freezing. (4) Thawing or partial heating above 0" C. (32" F.) induced by fluctuating temperatures reaching dangerous levels [heating t o temperatures in the vicinity of -2 t o -6" C. (28 t o 21" F.); should nevertheless diminish the bacterial count (p. 165). In principle, we may distinguish between two categories of investigations on these problems: studies on the microbiological flora of frozen foods and its character under different conditions, and direct infection studies concerned with bacterial inoculation, particularly pathogenic forms. When discussing the microbiology of frozen foods it may be appropriate t o point out that there are essentially three groups of microbes which are of significance, namely: pathogenic, toxigenic, and saprophytic types. This classification serves all practical purposes although certain exceptions may be easy t o state, as Cl. botulinum being a pathogenic saprophyte (Frobisher, 1953). It is particularly important that the methods for the examination of frozen foods primarily aim at detecting the organisms t h a t are dangerous t o the gastro-intestinal system.

MICROBIOLOGICAL

PROBLEMS

OF FROZEN FOOD PRODUCTS

173

2. Microbiological Methods of Analysis

It is likely that too much attention has been given t o the quantitative point of view in the bacterial control of frozen foods. The number of colonies on a plate, as a matter of fact, does not always give an adequate picture of the quality or of the risks. This does not only refer t o frozen foods but may be valid for foods in general (Wilson, 1935). I n some cases a wrong conception of the bacterial condition is obtained by relying on per cent decrease or increase figures. As pointed out above large changes may mean little from the bacteriological point of view. For example Nickerson (1943) found that the bacterial count in different packages of frozen broccoli might vary within a range of 200% and still not be very different in quality from the bacteriological point of view. Consequently no fixed figures corresponding t o quality grades can be established or should be promulgated. The standard methods valid for bacterial control of other preserved foods have been elaborated on the basis of long experience, principally from the standpoint of the errors that can occur. These accepted methods, however, can be used without modifications for frozen foods only in exceptional cases. Frozen foods vary t o such an extent from other types of preserved foods that different sets of standards must be established (Humphrey, 1950). The sampling technique may influence considerably the result of a determination of the microbial content, particularly if the sample is not thawed under defined conditions. The substrate always influences the result obtained. The methods in this field have been inadequately elaborated (Humphrey, 1950). I n particular, observations made on frozen meat, using the plate count method, have shown the necessity for emphasizing great care in arriving a t conclusions with respect to the lethal effect of freezing. The occurrence of conglomerates which, according t o temperature conditions, become separated into smaller units might thus produce an increase in the number of colonies without any real growth taking place. Contrary t o what might be supposed, taking samples for bacteriological investigation from the unthawed product is in most cases t o be preferred. If the product has thawed, the water frozen in the product melts, as well as the water which has been retained as a coating during the preparation. This water generally contains more bacteria per unit weight than the product itself. Therefore, a sample which includes too much water will result in erroneous counts. It is essential, therefore, t o get proportionate amounts of water; this however is very difficult. Hence, sampling of unthawed products is preferable. Special sample-taking instru-

174

GEORG BORGSTROM

ments have been invented which cut out representative samples from the frozen product (Anonymous, 194613). I n the United States a committee (Committee on Microbiological Examination of Foods) has been developing methods intended t o be used for the preparation of samples of frozen foods. The following is a summary of the recommendations of this committee (Anonymous, 194613): A mechanical blender (Waring Blender, Oster Blender, Turmix, etc.) should be used in the preparation of samples of frozen fruits and vegetables. This conclusion has been arrived a t after trying many different methods of preparation with various frozen products. I n comparison with hand grinding, mechanical disintegration as a rule yields higher bacterial counts, since conglomerates are more efficiently separated. At any rate more uniform results are obtained and the method is less laborious. Similarly, the choice of tryptone-glucose agar (with meat extract) rests on the demonstrated superiority of this medium over nutrient agar or glucose agar as a medium for frozen peas. The influence of the plating media is profound. I n many cases negative results do not indicate the absence of microorganisms. Hartsell (1951) warns against this risk, particularly due t o his experience with pathogenic Salmonella. The coliform test appears t o be more efficient for detecting contamination in foods prior t o freezing and storage, while fecal Streptococci are superior indicators in frozen foods, since coliform organisms seem less able t o survive the low storage temperatures. The presumptive enterococcus test, using the "SF" medium of Hajna and Perry (1943), seems t o be reliable and practical for the routine examination of frozen foods (Burton, 1949a and b). Larkin et al. (1955) do, however, recommend other media as more reliable indicators of the number of enterococci. Nickerson (1943) modified the Frost little-plate technique and on comparison with the Petri-plate method found counts t o be of the same general magnitude. Wolford (1943) described a modification of the directmicroscopic method stained by the Gram method, thereby achieving a differentiation between live and dead organisms. This opens the possibility of presenting a more complete sanitary history of the sample being examined. Berry (194613) is of the opinion t h a t a direct microscopic count is a better gauge of the sanitary history than is the culture method.

a. Methods for Frozen Vegetables: These methods --ere taken from Anonymous, 1946b and Goresline, 1948. Select from the lot to be examined a suitable number of packages (3 or 4 from each pack). Transport in dry ice to the laboratory for analyses and place the samples in a refrigerated [--18 to -20" C. (0 t o -4' F.)] storage chest until they are to be analyzed. The temperature preferably should not rise above 18" C. (0" F.) in the handling of the material.

-

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

175

Samples of peas, green beans, etc. are prepared in the following way for microscopic examination. Open the package and note the color, structure and consistency. Especially state if ice crystals are present on the inner wall of the package and if the vegetables show a tendency toward shriveling. Such a condition is indicative of thawing and subsequent refreezing. Record observations and the occurrence of abnormalities such as unnatural color or odor or the occurrence of pink colonies of Torulae, which are indicative of improper handling. Samples are then taken for microbiological analyses. The sample, if not loosefrozen, should be broken up into small units. This can be done by tapping the unopened package sharply against the table edge or by striking the package with a dull instrument, being careful not to break it open. After opening remove portions with a sterile spoon from various parts of the package ( k , center and corners) in order to obtain a composite sample. Al 50 g. sample is aseptically weighed into a sterile, glass, mechanical blender jar: 150 rnl. of sterile water are added and the contents blended for 2 min. If the blender is equipped with a variable transformer it is advisable to increase the speed of the motor gradually. Allow the samples to stand for 2 to 3 min. to permit the foam to subside. Pipette 1 ml. of the mixture into a 99 ml. sterile water blank. Replace the cap on the dilution bottle and shake the bottle briskly. Pipette 1 ml. aliquots of this mixture into each of 2 Petri dishes (1: 1000 dilution) and also 0.1 ml. aliquots into each of 2 more Petri dishcs (1: 10,000 dilution). A 1: 100 dilution may be obtained by pipetting 0.1 ml. aliquots of the original mixture into each of two Petri dishes. Pour melted tryptone-glucose extract agar (pH 7.0) cooled to 45" C. (113" F.) into the Petri dishes immediately, and thoroughly mix the dilution water with the agar by gently rotating the plates in a figure 8 motion with slight tilting of the Petri dish. Cool to harden, and incubate a t 32" C. (90" F.) for 4 days. Dilutions of 1: 100, 1: 1000, and 1 : 10,000 will usually suffice for commercially packed frozen vegetables, although further dilutions should be made if the history or the appearance of the samples warrant it. It is of primary importance that the agar be poured immediately after the inoculuni is introduced; otherwise many bacteria will adhere to the glass and an inaccurate caount will result.

The direct microscopic method has certain advantages over the plate count method; it is quicker and requires less equipment and glass. It also detects dead microorganisms, and indicates sanitary history, irrespective of t,he viable count. Wcigh 50 g. of the vegetable into a 250 ml. flask. Add 100 ml. water, stopper the flask and shake it briskly 50 times through a wide arc. Using a Breed pipette, transfer 0.01 ml. of the washings to a microscope slide, and with a needle, spread the drop rverily over a 1 sq. cm. area of the slide. Dry and fix with heat or methyl alcohol. Stain with Gray's double dye stain or with North's aniline oil-methylene blue stain, rinse, dry, and examine under the microscope, using oil immersion. Use a n ocular micrometer (such as a Whipple or Howard disk) with the microscope tube so adjusted t h a t the side of the graduations is equal to 0.1 mm. (area of field 0.01 sq. mm.). Count the cells in 100 fields and multiply the number by 20,000 to bring to a gram sample basis. Express results as "direct microscopic estimate in microorganisms per gram." For refwence to staining methods see Gray (1943) and North (1945).

In the direct method the following assumptions are made: all microorganisms are removed from the vegetable surface by washing; the sus-

176

GEORG BORGSTROM

pension of bacterial cells is uniform; and the drop of liquid is evenly spread over 1 sq. cm. A convenient and frequently used method is Nickerson's (1943) modified small-plate count: Ten grams of sample are placed in 90 ml. of sterile water and thoroughly shaken. A sterile slide is placed on a warming plate [metal plate regulated to 45" C. (113" F.) by clamping on a ring stand above a hot plate] and 1 ml. of the dilution water is delivered to the raised portion. The measuring pipette used (capacity 0.1 milliliter) is cleaned with water several times by filling and emptying the pipette, and the lower portion is wiped with sterile Kleenex * before the culture sample is delivered. Four drops of molten nutrient agar (33.5 grams Bacto-nutrient agar in 1000 ml. of distilled water) is then placed on the raised portion of the slide and the material mixed thereon. Mixing is accomplished by running a sterile wire through the culture 15 times, first backwards and forwards, then from right to left, and left to right. The culture is spread t o the edges of the raised portion by slanting the needle and following the edge of this area. These slides are marked for identification and placed in a moist chamber for incubation [I6 hr. at 25" C. (77" F.)1. After this time the cultures are removed, heated on a hot plate at about 80" C. (176" F.) until dried, treated with 1% aqueous ferric sulphate solution for 20 sec., washed and then treated with a 5 % aqueous solution of hematoxylin for 15 to 30 sec. The agar film is then washed and dried. The prepared slides can then be examined microscopically to detect and count bacterial colonies.

Plate counts of more than 4,000,000 per gram of peas or direct microscopic counts of over 1,000,000 per gram may be considered indicative of poor sanitary conditions in the plant or poor handling in transit or in warehousing. I n vegetable freezing plants, plate counts of over 500,000 per gram are generally not encountered in newly frozen products unless there is some degree of carelessness in the factory, such as faulty clean-up practice or prolonged holding of the material after blanching. In the preparation of samples of frozen spinach it is important to obtain a certain degree of thawing t o enable proper comminution: Allow the package to thaw a t room temperature for 135 or 2 hr. Then the package is opened and 50 g. of the contents are weighed into a sterile, borosilicate glass, mechanical blender jar. The samples are then assembled from various portions of the package, taking care to select petiole and blade portions in about the same ratio as occurs in the whole sample. Sterile water (450 ml.) is added and the mixture is blended for 2 min., then analyzed in the same way as samples of peas and beans, etc.

The procedure for broccoli and cauliJEower is somewhat different: Using a sterile scalpel, cut portions from the curd and stem of several representative pieces of the vegetable. This can be done after the broccoli has been allowed to thaw partially at room temperature. Aseptically transfer 50 g. of these portions into the sterile, borosilicate glass, mechanical blender jars, add 450 ml. sterile water and proceed as directed for frozen peas.

* (Kleenex is a trade-marked soft cleaning paper.)

blICROBIOLOGICAL PROBLEMS O F FROZEN FOOD PRODUCTS

177

Asparagus also is partly thawed a t room temperature before sampling. The spears are cut into short lengths with a sterilized scalpel. Then 50 g. of the segments are transferred aseptically into a sterile, borosilicate glass, mechanical blender jar. Sterile water (450 ml.) are added after which the procedure used for the analyses of peas is followed. When weighing out the sample portions, a number of spears should be selected and the proper proportion of butts and tips maintained.

The presence of psychrophilic organisms in frozen vegetables can best be determined by holding the sample a t 0" C. (32" F.) for 30 days prior t o plating with plates incubated subsequent t o this holding, a t 20" C. (68" F.) for 48 hours and 0' C. (32" F.) for 30 days (Hucker, 1954).

6. Methods for Frozen Fruits: In the preparation of samples of frozen fruits the package should always be held at room temperature for 1 t o 2 hr. before opening, in order to partially defrost the contents (Goresline, 1948). While the fruit is still somewhat frozen, portions are cut from various parts of the contents of the package with a sterilized scalpel. The proportion of fruit t o syrup should approximate that of the whole package. Weigh 50 g. of fruit and syrup into a sterile, borosilicate glass blender jar. Add 450 ml. sterile water and blend for two min. Make a 1: 1000 dilution by adding 1 ml. of the blender mixture to 99 ml. sterile water, and make further dilutions in the usual manner. Plate 1 ml. portions from the various dilutions on tryptone-glucose extract agar. Incubate for 3 days at 32" C. (90" F.). Colonies are counted under a suitable colony counter and the results recorded as "plate count of microorganisms per gram." Direct microscopic counts for molds and yeasts are made according to the methods given in Official Methods (A.O.A.C.) for microscopic analyses of tomato products. Care must be taken to distinguish between mold hyphae and fruit setae.

c. Methods for Eggs: In the bacterial analyses of frozen eggs the following procedure is recommended : A well-mixed sample of 10 g. is weighed on a torsion scale into a flask containing 90 ml. of sterile water. The flask is stoppered and well shaken. Sterile glass beads are placed therein t o comminute frozen yolk to a homogenous mixture. The initial dilution, 1: 10, is used for making other dilutions and direct microscopic counts. Plates are prepared in 1: 1000 and 1: 10,000 and other concentrations which may be required.

For lower qualities of fresh frozen eggs it may be necessary to use higher dilutions of 1 : 100,000 or 1 : 1,000,000. As a substrate, Bacto-tryptone-glucose agar is used and is prepared according to standard methods. The plates are incubated a t approximately 32" C. (90" F.) for 72 hr.

As a rapid method, the direct microscopic count can be used: In this case 0,0001 ml. of the 1: 10 egg-mixture mentioned above is spread by means of a pipette on a clean microscopic slide over a 1 sq. cm. area. The smear is airdried, defatted in xylol for 2-5 min. and fixed in 95% alcohol for 2-5 min. The smear is then stained for 2 min. in North's aniline oil-methylene blue stain, washed in a beaker with distilled water, drained and then dried. The result of the direct microscopic count is recorded as count of bacteria per gram. The microscope is adjusted according to standardized methods (A.O.A.C.) and the estimates of bacterial popula-

178

GEORG BORGSTROM

tions are made according to approved methods. Generally 60 microscope fields of vision are counted. Escherichia coli tests are made in the following manner: portions of the 1 : 100 and 1 : 10,000mixtures are pipetted into a fermentation tube containing lactose broth. The inoculated lactose-broth tube is held in an incubator a t 37" C. (99" F.) for 48 hr. All tubes showing gas in 48 hr. are streaked on eosin-methylene blue (EMB) agar or Endoagar plates. The presence of Escherichia coli is determined according to growth characteristics (dark colonies with a greenish metallic sheen) on eosin-methylene blue agar after 18-24 hr. of incubation a t 37" C. (99" F.). Count of coli bacteria (presumptive test) is made as follows: 1 ml. portions of suitable dilutions of the egg-water mixture are placed in sterile Petri dishes. Ten to 15 ml. of desoxycholate agar, melted and cooled to a temperature of 40 to 44" C . (104 to 111" F.) are then added. When the agar and the test mixture have congealed a second thin layer of the desoxycholate agar is poured over the surface. The purpose of this is to prohibit the occurrence of surface colonies of coli organisms as the appearance of such colonies is often so atypical that they are not easily recognizable. When the agar coating has set, the plates are invertcd and incubated at 37" C. (99" F.) for 20-24 hr. I n counting the plates after the incubation, record as coli only typical dark red colonies of a t least 0.5 mm. diameter.

As a rapid method, the applicability of the resazurin decoloration is not t o be overlooked. This method has been tried in a variety of foods by Proctor and Greenlie (1939) and has given remarkably good indications of the bacterial content. The tests for coliforms are preferable t o the E. coli presumptive test a t the 0.001 g. level for assessing the quality of frozen egg products (Winter et al., 1951). These workers analyzed 112 samples of commercial frozen egg products by the standard plate bacterial count. Ninety-two were also analyzed for coliform and 45 for direct bacterial count. The range in standard plate and coliform bacterial count was wide and far more variable than the direct count. Since enterococci usually occur in small numbers and are lost in the usual plating procedure, selective media are essential for their detection. Failure t o isolate enterococci can probably be attributed t o this fact. For routine examination a modified Winter's medium has been suggested (Brown and Gibbons, 1950). Yeast and mold counts are made in the following way: One ml. portions of the 1: 100 solution of the above-mentioned egg liquid and any further dilution required are pipetted into sterile Petri dishes. Fifteen ml. melted and cooled (45" C. (113" F.)) potato dextrose agar (pH 3.5 k 0.1) are poured into each of the Petri dishes. The plates are incubated for 5 days a t 25" C . (77" F.). I t is advisable to examine the plates for molds after 3 days and once again after 5 days. The count obtained on the fifth day constitutes a measure of the extent of yeast and mold contamination. The potato dextrose agar is adjusted with 10% sterile tartaric acid to pH 3.5 k 0.1. The amount of 107, tartaric acid, of course, depends on the buffering degree of the substrate. In case of doubt as to whether the colonies are yeasts, molds, or bacteria, a microscopic examination of the colony is necessary.

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

179

The sampling method profoundly influences not only the amount of solids but also the bacterial count. This has been comprehensively studied by Kahlenberg et al. (1951). It was concluded that the common procedure for drilling frozen eggs t o obtain shavings does not give representative samples. If satisfactory devices can be developed, drilling of plugs would be preferable, or an even better procedure is t o allow a complete defrosting of the entire can before sampling. Freezing of small samples simultaneously with the large size can6 for later use as samples also has been suggested. 3. Results of Examination of Frozen Foods

a. General. An extensive investigation on the occurrence of microorganisms in frozen berries and vegetables was made by Wallace and Tanner (1934, 1935). More than 2000 packages of cherries, strawberries, peaches, beans, peas, etc., were studied. The number of microorganisms continuously increased during pretreatment and packaging, but diminished substantially during freezing and subsequent storage. E’etists decreased more rapidly than molds, and bacteria most slowly. After 1 year of storage, however, the microbial population did not diminish any further. Even after 3 years of storage the proportion of microorganisms remained fairly unchanged. The longer the frozen foods are stored, the more rapidly, however, microorganisms develop during and after thawing. I n other words, the longer the frozen products have been kept, the more important i t is t o consume them immediately after thawing, and not t o keep them in a n unfrozen state. As a rule, the microorganisms which develop are not dangerous t o the health, but the frozen products spoil and become unpalatable. An equally extensive study was made by Smart (1934, 1935). He observed in strawberries a dying-off of 99.3% of the microorganisms after 1 year of storage; however, 7 varieties of fungi, 1 yeast, and about 30 species of bacteria were viable even after 3 years. I n spite of the high mortality rate, no less than 1,000,000 bacteria per gram of frozen berries remained. There is a preferential killing of more freeze-sensitive varieties and a survival of more resistant ones. Micrococci and Flauobacterium belong t o this last group (Lochhead and Jones, 1936) whilst Achromobacter steadily declines in the frozen pack. Coliforms also die off more rapidly (Larkin et a[., 1955) whereas fecal streptococci remain constant. 6. Fruits and Fruit Juices. On arrival a t the factory, fruit carries 011 its surface large numbers of microorganisms. Molds of the genera A spergillus, Penicillium, Mucor, Rhizopus, and Sterigmatocystis, and yeasts such as Saccharomyces and Torula are the most common. Certain bacteria, such as Staphylococcus aureus, Bacillus termo, and B. subtilis, also occur in abundance.

180

GEORG BORGSTROM

I n order t o avoid unnecessary microbial development on berries, all present day textbooks recommend the greatest speed possible in preparation and quick chilling (preferably ice water) immediately after picking. Furthermore, as Magoon (1931) has already proved, half of the microorganisms on the surface of the berries can be removed by thorough washing. Bactericidal and fungicidal substances may be added in washing. Direct investigations have shown the favorable influence of propionates (sodium and calcium salts) on berries and vegetables (Wolford and Andersen, 1945). These chemicals, however, exert their influence only for a short period of time. I n freezing by immersion such protection is particularly important, since the freezing liquid is being continuously contaminated with spores and bacteria. The addition of benzoic acid (0.04%), acetic acid (0.07%), or propionic acid (0.06%) will also hold the bacteria in check. According t o Lenhart and Cosens (1949), the use of ultraviolet radiation also has yielded good results. Special importance is attached t o the influence of sugar which, because of its enzyme-inhibiting effect, results in a better quality, but at the same time protects microorganisms against the killing effect of the freezing temperature (McFarlane, 1942). The acidity in fruits and berries also has some influence in maintaining good quality, due t o its inhibitory effect on microorganisms. Obold and Ilutchings (1947) indicated that fruits and vegetables, after packing and before freezing, should not be stored for longer than 24 hr. a t 4.5" C. (40" F.), 5 hr. a t 10" C. (50"F.) or 2 hr. a t 27" C. (80" F.). I n the case of berries, which are often frozen in large containers, there is danger of fermentation occurring in the innermost parts of the mass before the entire contents freeze. Ireland (1941) recommended on account of this t h a t the fruit be precooled before barreling for freezing and that the barrels be rolled frequently when in the freezer. During slow freezing, the growth of certain fungi has been noted even in frozen tissues (Young, 1947). However, frozen fruits and berries, as well as frozen concentrates, all have such low p H values t h a t the substrate is unfavorable for most organisms that might be factors in spoilage during the freezing. Several scientists have studied the behavior of microorganisms in frozen juices and have stated t h a t such factors as acidity, temperature, and content of dry matter influence the number surviving (Irish and Joslyn, 1929; Tanner and Wallace, 1931; Beard and Cleary, 1932; Wallace and Park, 1933; Berry and Diehl, 1934; Shrader and Johnson, 1934 ; McFarlane, 1940b). During the first few weeks of storage after freezing a rapid decrease in the number of bacteria has been noted by Lochhead and Jones (1936).

MICROBIOLOGICAL PROBLEMS O F FROZEN FOOD PRODUCTS

181

After the eighth month, however, the rate of killing was remarkably slow. There is a notable reduction in the per cent of living microorganisms in the concentrate at all temperatures below -4" C. (25" F.) (Du Bois and Kew, 1951). Depending on methods of handling and conditions of culture, strawberries in a fresh state carry comparatively large numbers of surface microorganisms, 7500 t o 100,000 per gram of fruit, according t o Berry (1934) and Gilbert and Wiegand (1950) and 19,000 t o 800,000 per gram, according t o Magoon (1931). The microorganisms were mainly molds (Penicillium, Rhizopus, Mucor, Botrytis, Stemphylicum, and Fusarium), yeasts, and sporeforming bacteria. Recently it was reported an effective reduction of mold content was achieved with a synthetic detergent in the washing water (Haynes et al., 1953). The number of microorganisms varies as well as the type of flora. Magoon (1931, 1932), observed on an average the following ratio of flora types; 65% bacteria, 23% molds, and 12% yeasts. The bacteria generally originate from the soil and are transferred t o the product by the hands and containers. The yeasts are chiefly air-borne ; mold spores are ubiquitous. Mundt (1950) studied the microbiology of strawberries during harvest and handling and indicates that the following factors affect tlhe number of microorganisms : climatic conditions, ripeness, temperature, and delay in preparation. By use of the Howard mold-count procedure and plating techniques it was proved that the coating on containers used for conveying fruit from the field t o the plant favors the growth of yeasts, possibly by preventing the escape of juice. It has recently been observed that the bacterial count in fruit, juice is somehow causally related t o the occurrence of fungi. Individual fruits contaminated by certain molds give an abnormally high bacterial count (Proctor and Nickerson, 1948a). This is probably due t o the fact that a fruit damaged by fungi easily succumbs t o bacterial infection. Remarkably high numbers of bacteria developed in these particular fruits. The so-called soft rot of oranges, caused by bacteria, results in an abnormal increase in bacterial content (Wolford and Berry, 1948a,b ;Beisel, 1951). The total count is often many thousand times higher than in sound fresh fruit. Even the count of coliform bacteria rises under these circumstances. Ordinarily coliforms are very rare in fresh, uncontaminated juice, and the death-rate is comparatively high. I n citrus juice as in strawberries there is every evidence t o indicate the presence of specific bactericidal substances (Jakovliv, 1948). The decrease in the bacterial count which occurs in the juice of oranges contaminated b y soft rot even during several months freezing storage is less important because the predominating kind of bacteria belongs t o the

182

GEORG BORGSTROM

genus Aerobacter, rather than Escherichia, whereas fresh juice only under exceptional conditions contains Aerobacter. Because of these circumstances careful grading of the fruit is important. However, this cannot be done easily owing t o the fact that the initial attacks of soft rot are very hard t o detect. Only random bacteriological tests can be used for guidance in this connection (Wolford and Berry, 1948a). Beisel (1951) also recommends a thorough washing of the fruit surface with the use of detergents. A few years ago a troublesome off-odor spoilage was encountered in several plants used for frozen concentrates. This was traced t o 1act)icacid organisms in the genera Leuconostoc and Lactobacillus (Hays, 1951 ; Murdock et al., 1952; Hays and Riester, 1952). Rapid heating [I sec. a t 71" C. (160" F.)] prior to freezing is gradually becoming a general procedure in order to eliminate these risks. This commercial pasteurization is more effectively applied t o the single-strength juice than t o the concentrate. The resistance of these microorganisms was greater in concentrates than in single-strength juice, possibly due t o the increased sugar content (Murdock et al., 1953). Lactose-fermenting -yeasts occurring in orange juice cause false positive results with standard (American) lactose broth and brilliant green bile and lauryl sulphate-tryptose broths when large innocula are used. Colonies of some of these yeasts cultured on eosin-methylene blue (EMB) agar closely resemble colonies of E. coli. I n 62 samples of commercially packed orange juice showing presumptive positive tests in one of the above media no true coliforms were found by Martinez and Appleman (1949). It has been shown in the case of frozen juices or berries and other fruits that some of the most important sources of infection occur along the production line. Bacteria carried on the raw material are, as a rule, of less importance than coliforms which are common and occur abundantly in slime, on conveyer belts, elevators, washing tubs, waste, etc. More than 3,000,000 bacteria per gram have been recorded frequently (Teunisson and Hall, 1947; Wolford and Berry, 1948b). Coliforms and above all Aerobacter, easily survive in orange juice according t o Wolford (1950). Shrader and Johnson (1934) obtained results a t variance with those reported above, claiming that three organisms ( E . coli, Lactobacillus acidophilus, and B. subtilis) failed t o multiply in orange juice a t temperatures ranging from 37 t o 10" C. (99 t o 50" F.). It was even concluded that organisms of the coli type fail t o survive longer than 2 weeks in frozen orange juice but that spore-forming bacteria would probably survive for a considerable period of time. There are also strains of yeasts th a t grow slowly a t -18" C. (0" F.) and rapidly a t 0" C. (32" F.). Recent reports from Florida reveal th a t extensive sanitation pro-

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

183

cedures, have been introduced in citrus juice factories. Fruit surfaces are washed t o reduce the bacterial load (80 t o 95%) and all preparation equipment is regularly cleaned during operation (Brokaw, 1952). Continuous plate counts and microscopic examinations of the juice are used t o check the bacterial levels and determine the time schedule for regular clean ups. The number of bacteria originally present in fruit juices vary in extremely wide ranges from 0 t o more than 1000 colonies per milliliter (Nolte and Loesecke, 1940). I n concentrated frozen juices bacteria are killed considerably more efficiently (Faville and Hill, 1932; Faville et al., 1951; Hahn and Appleman, 1952a,b). This is remarkable, since most of this juice is not pasteurized. With longer storage periods there is a notable reduction in the percentage of living microorganisms in the concentrate at all temperatures below -5" C. (23" F.) (Miller and Marsteller, 1952). I n concentrates, however, yeasts seem t o have greater possibilities of survival (Patrick, 1949). This is of great importance from the standpoint of the storage of thawed concentrates. Coliforms and other nonsporeforming bacteria generally do not survive long range storage in frozen juice or concentrates (Patrick, 1953). The recent observation t h a t enterococci frequently appear in citrus juice is somewhat of a riddle. The organisms have been identified as Streptococcus fecalis and S. liquefacians. Whether they originate from the soil, are conveyed by water or birds, or are carried inside the fruit tissue during growth and development is yet t o be determined (Kaplan and Appleman, 1952). These research workers even suggest that enterococci showing greater survival capacity might serve better than E. coli, which is rapidly eliminated, as a n indicator of pollution in frozen food. This has been confirmed by Larkin et al. (1955). c. Vegetables. The bacterial flora in frozen vegetables has proved mainly t o include the following genera: Achromcbacter, Aerobacler, Alcaligenes, Bacillus, Cellulomonas, Chromobacterium, Erwirbia, Flavobacterium, Lactobacillus, Leuconostoc, Micrococcus, Mycobacterium, Neisseria, Phytomonas, Pseudomonas, Sarcina, Serratia, Staphylococcus, Streptococcus (fecalis), and Vibrio (Sanderson and Fitzgerald, 1940; Nickerson, 1943; Hucker et al., 1952; Hucker, 1954). Many of these bacteria are typical soil organisms (Smart, 1937b). They have been isolated from vegetables in the frozen state or immediately after thawing and consequently they are not present as a result of infection. I n peas, species of Lactobacillus dominate (Berry, 193313) but spores of yeasts and molds are also common (Nickerson, 1943). Hucker et al. (1952) state that there are indications of a basic cold-resistant flora of bacteria in frozen peas, beans, and corn. Coliform bacteria in vegetables have been studied by Burton (1949a)

184

GEORG BORGSTROM

who found by use of the IMVIC tests t h a t 53 % were Aerobacter aerogenes and 25% intermediate types. The ratio of the different bacterial species did not change after a year of storage at -20" C. (-4" F.) and the enterococci were still dominant. Elrod (1942) showed that Erwinia-bacteria may give reactions similar t o coliform bacteria, which means that a fecal contamination may be erroneously indicated. The numbers and types of organisms found in frozen vegetables stored over long periods and the presence of types which grow a t 0" C. (32" F.) mere recently studied by Hucker (1954). Snap beans and Lima beans were found t o contain a significant freeze-resistant flora which remained viable after storage a t -17.5" C. (0" F.) for 2-10 years. Facultative psychrophiles mere found in asparagus, brussels sprouts, broccoli, and cauliflower but were not generally present in peas, squash, corn, and snap beans. Obligate psychrophilic types were not encountered in any of the samples. I n frozen vegetables held a t - 17.5" C. (0" F.) for 2 years or longer, the predominating organism was a large, paired biscuit-shaped coccus when the isolations were made from plates incubated a t 32" C. (90" F.). When the isolations were made from plates incubated a t 0" C. (32" F.). for 30 days, the predominating type was a gram-negative rod, resembling the Flaz~obacteriumestereoromaticum types. I n blanching vegetables the bacterial content may decrease as much as 90% and in some cases as much as 99% (see Table IV) (Smart and TABLE IV Reduction in Microorganisms Effected by Commercial Blanchinga Total microorganisms per gram Vegetables Potatoes Carrots Cabbage Beets Cauliflower Broccoli Peas Beans (Lima) Corn

before blanching

after blanching

7,050,000 435,000 50,000 3,150,000 150 000 450,000 1,250,000 2,500,000 495,000

170 60 25

100 80 150 175 50 75

From Koelensinid ( l ! E O )

Brunstetter, 1936, 1937; Vaughn and Stadtman, 1946; Jones and Pierce, 1946; Koelensmid, 1950; Pederson, 1947; Hucker and Robinson, 1950). This however does not mean sterilization since the sanitarily important

M I C R O B I O L O G I C S L P R O B L E M S O F FROZEN FOOD P R O D U C T S

185

decrease occurs only within the last 1%. Properly speaking, the killing of the bacteria should always be recorded exponentially. Rather few such determinations have been made, however. Besides other important technical effects, blanching nevertheless has a hygienic implication. Accordingly it seems, as a rule, justified t o conclude that a high bacterial count in frozen vegetables is due t o contamination during the time between blanching and freezing (Vaughn et al., 1946; Hucker et al., 1952). Proper sanitary procedure subsequent t o blanching is an important factor in determining the number of organisms in the final package. This was recently confirmed by Larkin et al. (1954) tracing the occurrence of fecal bacteria. The flora which is built u p prior t o blanching is predominantly mesophilic. Small pieces of vegetable material which become lodged in the conveyer belt, and cell sap which exudes from the cut tissue provide a better medium for the growth of microorganisms than the raw product itself (Pederson, 1947). For this reason the regular cleaning of conveyer belts and other equipment is so very important (Hucker and Robinson, 1950; Hucker et al., 1952). Pederson (1947) made a study of the extent of contamination encountered in frozen vegetables and on the effect of various methods of handling on the bacterial count. The greatest danger is when vegetables get contaminated by organisms in the active growth phase. Counts of 10,000 t o 100,000 per gram may be expected in frozen vegetables, but when they exceed 1,000,000 this is a clear indication of careless handling. During the freezing storage [-20" C. (-4" F.)] of vegetables the microbial flora gradually diminishes as various organisms are destroyed. This process slows down with time, but even after 8 months sufficient bacteria may remain so the vegetables may spoil after thawing (Lochhead and Jones, 1936). Greater reduction in counts were recorded a t - 12" C. (10" F.) than a t - 18" C. (0" I?.) and -23" C. (-9" F.) and was most pronounced in samples with higher initial counts (Hucker et al., 1952). Above all, coliforms and Lactobacillus are moPt apt t o survive (Weiser, 1951). Actually, with respect t o frozen vegetables, attention has been primarily directed t o the occurrence and danger of botulism. The possibility that there might be sufficient time for the botulinus toxin t o develop during preparation before freezing and thus be left in the product even after long-term freezing storage has been considered. There is of course also the possibility that this organism might develop during transport and distribution of frozen vegetables, particularly if some thawing has been permitted t o occur. Since vegetables usually have p H values which are quite favorable t o bacterial development, it is fortunate t h a t most frozen

186

GEORG BORGSTROM

vegetables are cooked before serving. If they are heated sufficiently, the destruction of viable organisms and their toxins is insured, thereby rendering the cooked product safe for consumption. According t o experience in the United States, it is usually possible t o produce frozen peas with bacterial counts less than 100,000 per gram. On the other hand, neither frozen corn nor spinach can be prepared with such a low count. Humphrey (1950) and Hucker (1950) arrived at the conclusion that there is no direct relation between quality and bacterial counts. He is, however, of the opinion th at bacterial counts do reflect the sanitary conditions under which the products have been prepared. I n other words, the hygienic conditions of the factories are important with respect t o quality and certain maximum values for bacterial counts should be stipulated. $Peas should contain less than 50,000 per gram, cut green beans less than 100,000, and corn less than 60,000 organisms per gram on entering the final freezer. Samplings made in a number of factories showed the very wide range of counts from 1300 to 870,000 organisms per gram. I n the United States most vegetables have low bacterial counts upon arrival a t the freezing plant because of the methods of handling and transportation used. The methods of handling various products during the different steps in preparation in the plant is the factor determining the final load of bacteria. On conveyer belts for beans and for cauliflower respectively, as many as 10,000,000 and 15,000,000 organisms per sq. cm. have been found after a few hours (Koelensmid, 1950). I n this connection it is of interest t o note some figures collected by workers of the Western Regional Laboratory in California (1944) a t 13 pea freezing plants. The counts, given below, represent averages of results obtained in all plants: 0

TIMEOF SAMPLING

BACTERIA PER GRAM OF PEAS( XlOOO)

On arriving a t the factory After washing After blanching After rinsing and transport From inspection belt before packaging When entering the freezer After freezing

11,346 1,090 10 239 410 736 560

Further information in this area may also be found in another paper (Anonymous, 1945b). Hucker et al. (1952) were unable to observe any correlation between bacterial content and the nutritive value of peas or beans. Spinach must be chilled immediately after harvest; otherwise the bacterial count will increase rapidly and washing becomes relatively

MICROBIOLOGICAL PROBLEMS OF FROZEN

FOOD PRODUCTS

187

ineffective. If large quantities are blanched at one time the hot water fails t o penetrate properly, and conglomerates are not removed. These experiences have been reported from California by Weiser (1951). I n spite of what has been said above on the possibility of drawing conclusions with respect t o the quality of frozen foods, it is in order t o point out that, according t o Tressler and Evers (1947), low values are indicative of good handling practices. The latter are of the opinion t h a t if counts as low as 10,000 per gram are found in a frozen vegetable the following conclusions are possible: (1) The raw product has been adequately blanched. (2) The product has been cooled rapidly and effectively after blanching. (3) The factory and particularly the production line are in a good sanitary condition. (4) The packaging material was in excellent condition from the standpoint of bacterial contamination. (5) Freezing took place rapidly. (6) The product has not had an opportunity t o thaw. A study of the bacterial content of frozen foods has been carried out in Canada (Anonymous, 1950a). A large number of commercial samples of vegetables (376) were investigated for fecal bacteria, primarily from the standpoint of coli-types and enterococci. Coli analyses proved t o be more effective for the detection of infection in ram material, whereas the determination of fecal Streptococci was better for frozen products since coli-bacteria are less liable t o survive freezing. Hajna-Perry’s SF-substrate for the diagnosis of enterococci was used (Burton, 1949b). According t o Diehl (1945) practical experience in the United States has shown that peas and spinach in particular offer difficulties since damage usually appears in the stored product only after a considerable period of storage. Recently Larkin et al. (1955) found both coliform bacteria and fecal Streptococci in many commercial samples of frozen vegetables. If not rightly handled, they constitute a possible health menace. d. Summary of Discussion o n Fruit and Vegetables. The number of microorganisms in frozen fruits and vegetables may reflect the quality of the raw material, act as a measure of the extent of bacterial contamination of the raw product, and reflect the sanitary condition in the processing plant and the speed with which the product was processed. Bacterial infection takes place principally during transportation and storage of the raw material under unfavorable conditions and along thc production line during preparation. All raw products including fruits, berries, and vegetables should be properly washed in potable water in order t o remove soil bacteria as thoroughly as possible.

188

GEORG BORGSTROM

e. Meat. The presence of microorganisms in frozen meat may not be important from a pathogenic viewpoint, but they are of primary influence t o the storage life of these products. The occurrence of large numbers of bacteria in ,meat may be related t o the method of handling prior t o freezing, such as careless management or improper ageing. Blood is particularly sensitive and easily shows high bacteria counts if strict hygienic controls are not maintained in the slaughter houses. The chilled storage of meat for ripening or other purposes is definitely detrimental t o the frozen product in so far as bacterial growth is concerned. Psychrophilic bacteria propagate well on the surface during storage so it is most advisable t o freeze rapidly immediately after slaughtering. At temperatures just below 0" C. (32" F.), molds may develop but their growth rapidly ceases at lower ranges of temperature (Brooks, 1924; Wright, 1923). Geer et al. (1933) have observed a n 80% decrease in the bacterial count of meat during freezing and a slight further decrease (84%) after 1 month of freezing storage. I n a number of cases an increase immediately after freezing has been observed (Anonymous, 1946a). It is not yet possible t o give a satisfactory explanation t o this phenomenon. There have been suggestions t o the effect that i t might be due t o bacterial populations (conglomerates) breaking u p during the freezing process. I n studies on frozen pork, Sulzbacher (1950) observed a decrease in the microbial count in all samples regardless of the storage temperature or the protective procedures used. One requirement is said to be that the storage temperature be kept below -20" C. (-4" F.). The freezing of pork is naturally of special importance in destroying Trichinella spiralis, but freezing has t o take place a t extremely low temperatures (Anonymous, 1953). Observations by Hendrickson and Miller (1950) in Kansas indicated that nearly 75% of the organisms naturally present died during the first two months of storage of pork sausage but t h a t the death-rate for the remaining 25 % was much slower. The vegetative and nonsporeforming bacteria were most susceptible t o low temperatures and therefore died early in the storage period. The spore formers, being more resistant t o freezing, were still present in large numbers after 310 days in storage. The results of Sulzbacher (1950), on the other hand, show t h a t several bacteria manage t o multiply even a t freezing temperatures. Several low temperature strains of Pseudomonas and Achromobacter were isolated from both frozen lamb and pork. All of them developed well a t temperatures as low as -4 t o -6" C. (25 t o 21" F.). Perhaps the protective effect of fat may account for the survival of so many coliforms for long periods of storage a t - 18" C. (0"F.). Even a few molds succeed in growing on meat at - 5 t o -6" C. (23 t o 21" F.) according t o Brooks (1924) and

MICROBIOLOGICAL

PROBLEMS

OF FROZEN FOOD PRODUCTS

189

Griffiths et al. (1932). When temperatures decrease t o about -10" c . (14' F.), microbial development is inhibited provided freezing has taken place. Some bacteria may grow under supercooled conditions a t temperatures as low as -20' C. (-4' F.). Bacterial studies on a variety of frozen foods containing chopped meat have been made in military laboratories in the United States. The average colony counts were 16,000 per gram on pork, 263,000 per gram on beef and 63,000 per gram on lamb. After cooking, these counts decreased considerably, t o between 30 and 720 per gram. Various thawing and cooking procedures had no effect on the counts (Anonymous, 1950b). Bacteria in hamburgers showed a remarkably low rate of mortality, even when stored a t -10' C. (14' F.). On an average the bacterial count decreased from 200,000,000 t o 10,000,000 per gram in one year (Weinzirl and Newton, 1915). On pork sausagc: large numbers of bacteria are conducive t o the development of off-flavors (particularly rancidity) according t o Hendrickson (1949). Microbiological problems have become increasingly relevant as a result of the development of self-service of meat in retail stores. Contamination during packing, microbial flora of the packing material, and quality of the primary products are factors of great importance. Unfortunately, no investigations of significance have as yet been reported in this field. Various prepared meat products such as sausages should, in fact, not be stored in frozen condition since deterioration in quality proceeds too rapidly when compared t o the fresh material. I n the United States substantial amounts of sausage are held in locker plants, but few bacterial studies were made. Special attention has been given t o the contaminating influence of seasoning ingredients. The work of Yesair and Williams (1942, 1945) showed that nearly all the common spices contain large numbers of bacteria. When these ingredients were used in the ratio of 1 g. of seasoning t o 1 lb. in pork sausage, pepper was found t o contribute 30,000 bacteria per gram of meat, sage 60 per gram, and salt fewer than 1 per gram (Hendrickson and Miller, 1950). f . Fish. Fish, unlike most other foods, carries from the start a markedly psychrophilic flora. As t o its composition reference is made t o Griffiths (1937), Zobell (1946), Aschehoug and Vesterhus (1946) and Castell and Anderson (1948). Anaerobic sporeformers are encountered less frequently (Castell, 1947; Castell and Anderson, 1947). I n order t o effectively check the psychrophilic flora, freezing and freezer storage must take place a t particularly low temperatures. These temperatures must be - 12' C. (10' F.) according t o Stewart (1934). The familiar fish contaminating organism, Pseudomonas jluorescens, retains its motility and even grows a t -7" C. (19" F.) (Bedford, 1931, 1933) and

190

GEORG BORGSTROM

also retains its chemical activity at still lower temperatures (Hess, 1934a,b). Slow freezing, which is quite often used for fish, is therefore an evident inconvenience, if only from a qualitative point of view. Several bacteria may then develop in freezing, and t o a much greater extent than in other frozen products. According t o Birdseye (1929) the death of bacteria on fish during freezing is as a rule unimportant. During freezing, Fellers (1932) has indicated that growth ceases completely. The bacterial flora of mackerel is greatly reduced by rapid freezing (Kiser and Beckwith, 1942). I n spite of the considerable movement of frozen fish in commerce, only one case of poisoning from such products has been reported (Fitzgerald, 1947a). The death rate of Pseudomonasfluorescens held a t low temperature has been studied by Hess (1934a,b). When subjected t o various temperatures for 30 min., the count of viable bacteria decreased as follows: 0" C. (32" F.), 26%; -3" C. (27" F.), 27%; -6.5" C. (20" F.), 35%; -10" C. (14" F.), 95%; and - 16" C. (3" F.), 90%. Refreezing at - 16" C. (3" F.) reduced the bacterial content still further. The experimental time range was presumably too short to give a clear picture of the killing effect of various temperature levels. The high figure a t - 10" C. (14" F.) may indicate the more disastrous influence of the temperature range between zero and minus ten below. Due t o growth concentrating psychrophilic bacteria slow freezing counterbalanced destruction and resulted in a lower rate of mortality than did rapid freezing. The p H has a marked effect on the survival of halophilic bacteria. Hess (1934a,b) observed a minimum effect in the p H range 6.0-6.3. More bacteria survive on halibut fillets in freezing storage a t -20" C. (-4" F.) than a t -15" C. (5" F.) or -10" C. (14" F.). This corresponds t o the observations made by Haines and is related to the fact that the effect of low temperature on proteins is far less pronounced a t -30" C. (-22" F.) than at -10 or -15" C. (14 or 5" F.), Handling of the raw product decides the bacterial load according to Proctor and Nickerson (194813). Studies underway in various countries, therefore, are aimed at reducing the risk of bacterial contamination. With this object in view, the merits of thorough disinfection on board ship, by use of new storage procedures and the addition of bactericidal substances t o ice are being considered. The improvement of methods for storage of the raw material in harbor is also under consideration. Investigations on the effect of a n initial and immediate freezing at sea should be regarded as part of this work. This procedure definitely gives a superior product (Fieger, 1950). Fish should be thoroughly washed before filleting in order to avoid a n infection by bacteria in slime-coatings. Good results are achieved by this procedure with cod (Dyer and Dyer, 1949).

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

191

Escherichia coli is the most common bacterial species in Caribbean shrimps according t o Holmes and McCleskey (1949). The number of viable bacteria decreased more rapidly in peeled shrimps than in unpeeled ones. Generally, freezing storage a t - 12" C. (lo" F.) effected more killing than a t -40" C. (-40" F.). Shrimp frozen immediately after catching even after one year of freezing storage retained considerably fewer bacteria (22,000 per gram) than those stored in ice for several days before freezing ( > 1,000,000 per gram) (Green, 1949). Bedford (1934) studied organisms on frozen fish and found a t the higher temperatures of storage the decrease was more rapid than a t the lower temperatures. I n fish a preferential killing effect of thawing was observed on certain bacteria e.g. F1a:obacterium in comparison t o other varieties more resistant on Achromobacter (Castell and Mapplebeck, 1952). According to these authors Pivnick studied the composition of the flora of cod when freezing and storing in frozen condition and found the percentage of Flavobacterium remarkably constant. g. Whalemeat. Whalemeat has t o an increasing degree been processed as food, particularly as frozen products. Chiefly British studies have been published as t o the bacteriological and hygienic problems related t o such products. The bacterial flora found in whalemeat is of either an intrinsic or extrinsic origin. The Streptococci and Clostridia probably originate in the intestine of the whale, and they are distributed throughout the tissues a t the time of death, constituting most of the intrinsic flora of whalemeat. These bacteria originating from the body of the whale are largely mesophilic. The source of the aerobic sporing bacilli is unknown. The evidence available a t present indicates that the remaining aerobic species, Micrococci and gram-negative bacilli, are probably mostly contaminants from the extrinsic flora of which they form the predominant part (Sharp and Marsh, 1953). These are mostly psychrophilic and are picked up in the freezing plant. The number of bacteria added here depends almost entirely on the conditions of hygiene observed in working the plant but, even with regular and efficient cleaning, about 100 times more bacteria are added, in practice, a t this stage than were present in the meat beforehand. As a result, the proportion of psychrophilic bacteria is a t the same time increased. The over-all result of processing whalemeat, from the carcass t o the frozen block, is t o increase its bacterial load 10 t o 100 times. I n effect, this is comparable with normal slaughterhouse practice, where even with the best technique a 100-fold increase in the numbers of bacteria on exposed surfaces may be expected. Special measures can greatly diminish the

192

GEORG BORGSTROM

amount of bacterial contamination, but they do not eliminate i t entirely. These extrinsic contributions of bacteria thus increase the proportion of bacteria in the meat able t o grow a t temperatures below 15" C. (59" F.), a n important factor in relation t o the effects of freezing. Freezing kills many bacteria, and after frozen storage the number of bacteria in whalemeat was found t o be relatively small again, being of the order of lo2 to lo3per gram. Freezing made no great difference, however, t o the proportions of the various types of bacteria; hence the variety of psychrophilic types was greater than that found among comparable bacterial counbs in fresh meat. The bacterial count is in every instance lower in the post-rigor than the pre-rigor meat, although there is no appreciable difference between them a t the time of removal from the carcass. The meat becomes contaminated, presumably, during preparation for freezing (Robinson et al., 1953). This contamination is in the end (i.e. after freezing) less in postrigor meat, possibly either because pre-rigor meat is more sticky or because some of the contaminating bacteria drain away in the "weep" from post-rigor meat while it is being cut up and frozen (Robinson et al., 1953). Taken over-all, however, there is little bacteriological difference in the frozen and stored products prepared from pre- and post-rigor whalemeat (Robinson et al., 1953). When storing frozen whalemeat for a period of 2 t o 3 months a t -10" C. (14" F.), there was a decrease in the number of aerobic bacteria and Clostridia in every case. I n the pre-rigor meat, the number of aerobes with a 37" C. (99" F.) optimum decreased 12-fold, while those with a 20" C. (68" F.) optimum decreased 63-fold; whereas in the post-rigor meat, the numbers decreased 1%- and %fold, respectively. The proportion of Clostridia killed on storage was not influenced by whether the meat was pre- or post-rigor, the numbers decreasing 6-fold in both cases. The rate of increase of aerobic bacteria in defrosted whalemeat at room temperatures of 15" C. (59" F.) t o 20" C. (68" F.) is rapid, largely because of the amount of free fluid available in the form of drip. Spoilage shows up within 24 t o 48 hr. afterthawing (Robinson et al., 1953). For this reason whalemeat is not suitable for the ordinary process of retail meat distribution but it can be kept in the frozen state, or for a week or two a t temperatures not exceeding 0" C. (32" F.) t o -3" C. (27" F.) until a few hours before use; however, i t is no more perishable than some other foodstuffs, such as milk or egg pulp. h. Poultry. There have been several investigations relating t o the bacterial flora on poultry and its behavior in the production of frozen foods (Sair and Cook, 1938; Gunderson et al., 1947). Bacterial growth

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

193

totally ceases a t - 10" C. (14" F.) (Pennington, 1945). It has been impossible t o enlist here the many species of bacteria found in poultry. Even in the bone marrow several have been observed, esp. Micrococcus. It has not yet been completely clarified whether the discoloration of bones cannot, also have a bacteriological cause. Coliforms are extremely common in freshly drawn poultry. For the rest, the flora is dominated by Achromobacter, Alcaligenes, Flavobacterium, and Micrococcus, which are all fairly common in processing plants. Their number decreaees in freezing and subsequent storage, but all types of bacteria survive and are thus represented in the frozen product after thawing. Sair and Cook (1938) studied the effect of freezing rate using temperatures between - 5 and -70" C. (23 and -94" F.) in freezing poultry, and found t h a t the number of surviving bacteria does not depend on the freezing rate. Nevertheless, Heitz and Swenson (1933) have been able t o show t h a t the number of bacteria on ducks which are frozen slowly turns out t o be 1000 times greater than in a rapidly frozen product. According t o the results obtained by Sair and Cook, it seems justified t o presume that this latter result is t o be attributed t o the fact that the products have been frozen so slowly that they have been allowed t o keep above the limit of bacterial growth temperature during a long period of time. I n the tests performed by these two investigators, it was not possible t o state any effect of the freezing rate, or a t the most a very slight one. In practice, the commodities are frozen as soon as possible after slaughtering or the product is stored in a cold-storage warehouse or refrigerator a t about 0" C. (32" F.). As bacteria grow very slowly a t this temperature, there is no reason for supposing that this might have essentially changed the results. If handling practices are satisfactory, the bacterial content can be kept a t a reasonably low levcl. Above all, removing the meat from the carcasses is a procedure which may involve heavy bacterial contamination. By boiling the eviscerated carcasses before the bones are removed (the most common practice) the bacterial flora is, of course, essentially reduced, although it always increases in subsequent handling (Gunderson et al., 1947). The original bacterial load of the poultry carcasses, when being frozen, depends on the cooling procedure, the use of a sterilizing wash prior to packing, and the time lag from slaughter t o freezing. Of great interest are the distinct differences in composition of the flora and the number of bacteria connected with the diet and the microbiology of the intestines (Sisler et al., 1940). Schneider and Gunderson (1949) concluded that freezing only partly eliminated the surface bacteria, chiefly Salmonella, on eviscerated chicken even after long storage. Various Salmonellae have been recovered from

194

GEORG BORGSTROM

the skin of frozen turkey even after long time storage (Cherry et al., 1946; Browne, 1949). Considerable numbers of bacteria remain in poultry meat even after 3 years of freezing storage (Harshaw et al., 1941). The various bacteria in the dressing of frozen stuffed chicken showed no significant changes during storage for one year at -23.3' C. (- 10' F.) (Esselen and Levine, 1954). By these investigations it seems t o have been proved that even for poultry meat the number of bacteria constitutes a good indicator of the hygienic conditions in the processing plants. For judging the risk of bacterial contamination in poultry meat it may be pointed out that research workers have traced organisms of the paratyphoid type in drawn poultry (Gunderson and Schneider, unpublished data). Typhus bacilli can survive for a considerable period of time in frozen products (Gunderson et al., 1947; Gunderson and Rose, 1948a,b ; McCleskey and Christopher, 1941). Owing t o these circumstances it is necessary t o maintain strict bacteriological control both in eviscerating and freezing plants. i. Eggs. The bacterial flora of frozen egg products is important because it influences the wholesomeness of end products, their keeping quality, and functional properties in the preparation of food. Haines (1939), Gibbons and Moore (1944), and Winter et al. (1946) have presented data which indicate that more than 80% of freshly laid eggs are bacteriologically sterile. However, nearly all commercial liquid, frozen, and dried egg products contain several hundred to several million bacteria per gram, according to reports by Redfield, 1920; Johns, 1948; McFarlane et al., 1945; and Winter and Wrinkle, 1949a. This stresses the importance of a regular sanitary control of frozen egg products. Special methods for bacterial counts in frozen eggs and egg products have been developed (Schneiter, 1940; Schneiter et al., 1943). Hartsell (1949) is of the opinion th a t better plating media need t o be developed for testing frozen eggs for pathogenic bacteria. T ha t Escherichia coli and the coliform group can survive in frozen eggs has been shown by many investigators (Brownless and James, 1939; Colien, 1942; Holtman, 1943; Johns and Berard, 1946; Nielsen and Garnatz, 1940; Pennington, 1948; Quinn and Garnatz, 1943; Schneiter et al., 1943; Sherman and Naylor, 1942; and Wallace and Park, 1933). I n previous studies it has been observed that a rapid reduction in viable cells occurs in the early periods of exposure and storage a t subfreezing temperatures. However, viable bacteria were present in some samples even after 18 months. Enterococci, invariably present in chicken feces, also survive and are considered a more reliable index of fecal contamination (Brown and Gibbons, 1950).

MICROBIOLOGICAL PROBLKMS OF FROZEN FOOD PRODUCTS

195

Of special significance in freezing plants are the sanit,ary conditions prevailing in areas intended for breaking the eggs. The bacterial flora of the egg shell influences the count of the frozen product (Winter and Wrinkle, 1949a; Winter, 1952). This is why a washing with germicides generally is applied before breaking, a t least for soiled eggs (Penniston and Hedrick, 1944, 1947). Close control is necessary. A single bad egg may contain billions of bacteria and spoil a whole batch of liquid egg. Frozen yolks as a rule contain more bacteria than frozen egg whites (Winter and Wrinkle, 1949a). I n freezing, the bacterial content in liquid egg immediately decreases. Coliforms generally die within 3 months in subsequent freezing-storage, but a further destruction of bacteria generally does not take place, even during a fairly long time. The egg evidently possesses a protective substance of some kind (Johns and Berard, 1946). Many of the species of bacteria found in frozen eggs multiply even a t the low temperature found in freezing and defrosting processes. Even rapid freezing gives a certain concentration of bacteria t o the central part of 30 lb. packages (Anonymous, 1953). By pasteurizing liquid egg a t 61 t o 62" C. (143 t o 144" F.) for 3 t o 4 min., more than 99% of the bacteria (appearing on standard plates) are killed, including nearly all coliforms and gram-positive cocci as well as the pathogenic bacteria (Winter et al., 1948; Winter and Wrinkle, 1949a; Wrinkle et al., 1950; Winter, 1952). Particularly valuable is the complete elimination of Salmonella through this procedure. It has also been suggested (van Oijen, 1940) that liquid egg can be pasteurized by heating a t 65" C. (149" F.) for 20 min. if 1 t o 2 % trisodium citrate were added as anticoagulant. This method is practiced commercially (Ingram and Hrooks, 1952). The egg white cannot be pasteurized a t a higher temperature than 56 t o 57" C. (133 t o 135" F.) without damage taking place. Pasteurizing at this temperature for 4 min. results in destroying 92% of the bacteria. As a whole, bacteria are killed more readily in egg white than in yolk or whole egg. Freezing and storing both nonpasteurized and pasteurized samples of egg gave rise t o some destruction of bacteria. The number of bacteria killed in freezing and storing a t -22 t o -23" C. (-8 t o -9" F.) for 2 t o 3 weeks, however, was by far not as great as pasteurizing for 3 t o 7 min. a t 61 t o 62" C. (142 t o 144" F.). Freezing and storage of liquid egg products at -18 t o -28" C. (0 t o -18" F.) resulted in a n average destruction of 55% in 12 days, 62% in 30 days, 87% in 60 days, and 90% in 100 days (based on plate count). Therefore, the destruction is most rapid in the first four weeks (Holtman, 1943; Lepper et al., 1944; Winter and Wrinkle, 1949b; Winter et al., 1951; Wrinkle et al., 1950). The ratio of direct plate count t o standard plate count averaged 2 : 1

196

GEORG BORGSTROM

but was much higher in frozen samples than in unfrozen (Winter et al., 1951). Possibly a methodological difficulty might have given rise t o an erroneous picture, as the frozen product could not be dissolved in water as easily when sampling for bacteriological tests (formation of clots was noticed). Alcaligenes, Flavobacterium, Proteus, and Pseudomonas predominated among the 13 species found in nonpasteurized eggs. Freezing reduced the number of bacteria, but not the number of species, which afterwards remained unchanged (Wrinkle et al., 1950). On thawing there was always a considerable increase in the number of coliform bacteria or grampositive cocci, but seldom a great increase of both types in the same sample. (A determination on agar plates from pasteurized samples showed that Alcaligenes, Escherichia, Flavobacterium, and Proteus mere predominant among the 13 strains from nonpasteurized eggs.) Pasteurization on the other hand always reduced the number of genera present, in this case from 13 t o 6 (Wrinkle et al., 1950). The viscosity is, however, reduced (about 17%) through pasteurization in the subsequently defrosted whole egg (Miller and Winter, 1951). Alcaligenes and Escherichia, the two most frequent species in liquid egg, gave a sour smell within 60 hr. after incubation on a bacteriological medium. Aerobacter also gave a taint odor and often showed coagulation; Proteus gave souring and occasionally coagulation, and Pseudomonas only souring (Wrinkle et al., 1950). I n earlier investigations a correlation had been stated between the number of bacteria and the amount of reducing sugar in liquid egg (Pearce and Reid, 1946). The higher the bacterial count, the lower the residual reducing-sugar content. This was, however, not confirmed in later studies (Johns, 1948) and may in fact not be expected, as a high bacterial count may be attributed t o two different Causes: either it, is due t o the bacteria having developed in the liquid egg mass and consequently contributed t o a high degree of breakdown and a corresponding consumption of sugar, or the product may have been exposed t o a heavy bacterial infection which need not necessarily have developed through growth prior t o freezing in the egg mass. I n this latter case there is no direct relation t o the amount of sugar but rather t o the degree of sanitation in the handling of the eggs and their content. The bacteriological count of commercial frozen whole egg varies widely. A standard plate count of less than 50,000 per gram is unusual, a figure below 500,000 per gram is probably better than the average and counts of 1,000,000 to 5,000,000 per gram are often encountered (Ingram and Brooks, 1952). Recently analyses have been made of commercial frozen whole egg in

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

197

Canadian laboratories. More than 10% had a bacterial count of over 10,000,000. During 3 years (1944-47) bacterial count,s were reduced by using precooled eggs, rapid freezing of the melange, breaking shell egg stock immediately, and improvement in equipment and plant sanitation (Fletcher and Johns, 1951). A new technique similar t o t h a t employed by the British Ministry of Food was tested. A plug was drawn and used for measuring texture emulsification and color. It was found that the bacterial counts from the plug were approximately twice as high as the sample taken by the electric drill method. The original Canadian frozen egg standards were rather lenient from the standpoint of bacterial counts. These have been reduced t o 2,500,000 for Grade A; a count of 500,000 is considered t o be practical. During 1949 the Burri slant technique mas introduced t o check the plate counts on all samples analyzed and any sample which exceeded 500,000 was then directed to a n official laboratory, where the count was determined by the official plate method. j . Dairy Products and Miscellaneous. I n butter, the bacterial count decreases by ten or twenty times in freezing storage in spite of the protection given by fats. On the whole the number of bacteria shows a slow decrease in dairy products (Wilster, 1946). Ice cream. as pointed out in the introduction, has not been included in this survey. A few words should, however, be said about cream (Fabian and Trout, 1943). Pasteurization a t 85" C. (185' F.) for 5 min. most effectively brought down the bacterial content of the original product and subsequent freezing storage for 1 year resulted in a continuous decrease in the number of viable organisms. A more effective homogenization multiplied the bacterial counts obtained. Frozen milk also shows a decrease in bacterial content during storage (Babcock et al., 1947). Both Lactobacillus casei and Escherichia ccli die off more rapidly in milk at p H 7.0, when the storage temperature is -2" C. (28" F.) than when it is -21" C. (-6" F.) (Ulrich and Halvorson, 1947). Hiscox (private communication) found viable cells of Streptococcus fecalis and of aerobic sporeformers in Cheddar cheese after 7 years of storage a t -255" C. ( - 13" F.). No other organisms survived. Butter appears t o protect effectively microorganisms. Even after 9 t o 12 months a t -26" C. ( - 15" F.) it is not uncommon t o find appreciable numbers of certain cocci, Pseudomonas sp., and aerogenes bacteria. Salt does not always destroy these bacteria, chiefly because of its uneven distribution. The concentration of salt in the water phase may reach 15% but there may still be droplets almost free of salt (Mattick, 1952). k. Bread Dough. Freezing storage of dough raises a great many special problems owing t o the particular nature of this product. A com-

198

GEORG BORGSTROM

plete account of the special reactions in the dough-mass a t low temperatures will not be given here. Briefly, it may be stated that investigations have proved that freezing storage of the unfermented dough is more favorable. The fermenting ability of the yeast will be better retained thereby, compared t o storing semifermented products. On the other hand freezing rate is also of importance. When a dough containing yeast is frozen slowly, the dying-off will be considerably higher than when it is frozen rapidly. The biological activity of yeast is retained during 1 t o 3 weeks of storage a t -18 t o -23" C. (0 t o -9" F.). Even if yeasts lose their ability t o grow, their enzyme systems may still be active. This is why a count of the number of surviving cells does not const,itute a reliable measure of the fermenting power of yeast (Godkin and Cathcart, 1949). 1. Precooked Frozen Foods. A special problem is offered by these products. They are very easily contaminated and invaded by microorganisms because the structure of the tissues is soft as a result of cooking when compared with those of frozen raw products. Cooking may destroy some of the bacteria but there are numerous chances for contamination during subsequent handling. Precooked frozen foods in many cases also offer a better medium for the development of microorganisms. Cleanliness and sanitation are imperative, since freezing does not sterilize the food. Rapid cooling is less likely t o result in spoilage, since the foods remain for a shorter time within the optimal temperature range for growth of bacteria. When leftovers are frozen it is particularly important t o be careful since they may have become contaminated during holding. Changes in bacterial count should be observed continuously during manufacture in order t o locate the most important sources of infection. This is said t o have been done in the United States after i t was observed t h a t precooked frozen foods contained more bacteria than similar preparations made in the home. If there exists a risk of infection with pathogenic organisms during preparation i t may lead t o disastrous consequences. I n direct incubation tests with foods consisting of fish or shellfish i t was possible t o establish t h a t several such bacteria were not killed by freezing or during subsequent storage (Hutchings and Evers, 1946). Furthermore, precooked frozen foods are cooked and prepared for eating in a much . shorter time than is customary for similar foods prepared entirely in the home so an effective sterilization is much more difficult t o achieve. Most of the reports in the field of frozen cooked foods have been directed toward problems concerned with home or commercial production. There is, however, little if any information on the freezing of cooked foods for institutional use, where freezer space may be available for storing cooked foods prepared in advance for peak load periods, or as a method of storing leftovers until they can be served. Information is needed about the

MlCROBlOLOGICAL PROBLEMS O F FROZEN FOOD PRODUCTS

199

wholesomeness of these foods after freezing, storage, thawing, and reheating. A recent study of meat loaves stresses the importance of avoiding long keeping periods after preparing the food (Ice et al., 1952). How long the precooked frozen foods will keep after thawing depends on the ingredients contained in the products. Control of raw products is fundamental in preparing frozen foods. In order t o insure a low bacterial count in desserts, special standard methods have been developed for different ingredients (such as evaporated milk, dry milk, sugar, eggs, and egg products) which may be used for this purpose (Anonymous, 1948). The microbiological standard of meat loaves also may be determined by the bacteriological count of the constituents (Ice et al., 1952). Poultry meat intended for the manufacture of chicken chow-mein, according t o investigations by Gunderson and Rose (1948a) contained 9,000,000 bacteria per gram, which was reflected in the bacterial count of the frozen product. The effect of prefreezing delay a t 25" C. (77" F.) on survival and multiplication of a Micrococcus sp. in creamed chicken was detrimental when exceeding 2 hr. (Straka and Combes, 1952). Tests a t the Western Regional Research Laboratory in the United States on samples of pork and beans failed t o show bacterial growth in 18 days a t 4.5' C. (40" F.) but after 21 days, growth could be noted. On the other hand, a t room temperature the bacterial count rose considerably on the second day (Hutchings and Evers, 1946). Apparently this product contained no psychrophilic types of bacteria. I n samples of the same food from commercial packages, more than 500,000 bacteria per gram were found. Whether this implies risks is, of course, totally dependent on the kinds of bacteria developing. It is also evident that similar phenomena may be observed in all precooked foods. I n spite of the fact that precooked frozen foods have been on the market in the United States for more than 10 years no case of food poisoning has been attributed t o this category of foods. Investigations on precooked frozen foods of this kind have been performed by Proctor and Philips (1947) and Hussemann (1951). They determined both the total bacteria count, using the plate count method, and the frequency of coliforms. More than 100 such frozen products were examined. The flora of surviving bacteria varied considerably. Certain products contained more than 1,000,000 colonies per gram. Among products examined were fish stews, meat, poultry, and soups. The average number of colonies in the different products of each category as a rule constituted less than 50,000 per gram. I n only 2 products was the average number of colonies above 100,000. As for the separate samples, it was found that 21 samples of fish stews (9.5% of the total number) contained more than 100,000 colonies. The highest figure obtained was 154,000. I n 23 samples of frozen

200

GEORG BORGSTROM

meat products (17.4% of the total) this count was surpassed, but the highest figure was only 250,000. Five hundred samples were examined for coliforms; and in most samples the number of colonies was below 50,000. Fitzgerald (1947a,b) arrived a t the conclusion that for most precooked frozen foods there should be fixed an upper limit of about 100,000 colonies per gram. The total permissible number of colonies, including coliforms, should be a t most 1,000,000 per gram. Furthermore, i t should be established that no lots be rejected, unless the number of colonies exceeds 500,000 per gram. This means that a distinction is made between a qualitatively desirable figure, or in other words, from a sanitary point of view, a preferable figure, and a dangerous limit. I n extensive investigations (Proctor and Phillips, 1947; Proctor and Nickerson, 1948a,b; Buchbinder et al., 1949; Hussemann, 1951) the total bacterial count in precooked frozen products was taken into consideration as was the taxonomic composition of the bacterial flora. Great variations in the number of colonies on plates from different samples of the same food were noted. In some cases more than 1,000,000 bacteria per gram were found, although the average count was as low as 50,000. Proctor and co-workers examined 64 different foods; fish foods showed a tendency for higher counts, but these frozen products were in most instances good from a bacteriological point of view. Marked exceptions were chicken stews, such as “chicken a la king,” the number of bacteria here was essentially higher than in other dishes. The same results were obtained a t the University of Wisconsin by Hussemann (1951) and Logan et al. (1951). Of 39 samples examined more than half had a bacterial count exceeding 1,000,000 per gram, 7.7% had more than 100,000, but in all cases less than 500,000 coliforms. This stands in marked contrast t o the results of Proctor and Phillips (1947) who showed that, in 68 samples of creamed meat and poultry dishes, only 14.7% contained more than 100,000 organisms per gram. Buchbinder et al. (1949) stated that toxic Staphylococci were present in 12 out of 39 samples studied. Eight of these gave Staphylococcus counts between 1,000 and 100,000 per gram, 2 gave 200,000 t o 400,000 per gram, and in 1 case, more than 2,000,000. I n 9 of the samples, Xtreptococcus was absent; in 3, the number of colonies was below 20,000; in 1, i t was 90,000 and in another 300,000. Four samples showed a count of more than 1,000,000. The dressing of frozen stuffed poultry is an important source of contamination (Esselen and Levine, 1954). I n order t o obtain a comparison for the evaluation of these counts Buchbinder and co-workers undertook an investigation in which samples were taken from restaurants in New York. I n no case were more than 100,000 bacteria obtained. Neither coliforms nor enterococci were found

MICROBIOLOGICAL PROBLEMS O F FROZEN FOOD PRODUCTS

201

in any of the twenty samples checked. I n this connection, the effect of heating frozen chicken stew (“chicken a la king”) was also studied. It was found that heating for 8 min. t o about 66” C. (151” F.) was sufficient t o kill most bacteria. This temperature is evidently sufficient t o destroy enteric pathogenic bacteria, but would of course only exert but slight effect on the Staphylococcus toxin already formed. The great difference in the bacterial counts for frozen “chicken a la king” and a recently prepared dish of the same kind as served in restaurants is surprising. The infection probably takes place when the poultry meat is removed by hand from the boiled chicken for preparation of the stew. Such meat is not sterilized again so enterococci may be transmitted easily. When the meat is later combined with a special sauce consisting of starch, f a t , milk, etc. possibilities for the development of bacteria are increased. It is therefore likely t h a t the infection occurs before freezing. The heating of the frozen product before it is consumed will, of course, destroy the majority of the microorganisms, but offers little protection against the heat-stable toxins formed by Staphylococci. Due t o this risk, very strict requirements with respect t o the bacterial content of frozen precooked foods should be established. The very high figures for counted colonies are due primarily t o the high degree of infection, chiefly during the preparation. This clearly proves the importance of good sanitation in the factories and during processing and handling of the products. These conditions are decisive with respect t o the bacteriological condition of the product. The number of bacteria, and consequently the number of colonies obtained in a plate count, decreases steadily during freezing. I n spite of this there are high counts. This makes the indicated causal relationship with hygienic conditions very obvious. Fitzgerald (1947b) stated t h a t frozen cooked foods should be regarded as a potential health hazard. Sanitation controls should be aimed a t two objectives: first, t o keep contamination t o a minimum and t o avoid the possibility of the development of off quality during processing, storage, distribution, and consumption; second, t o keep the “most probable number” (M.P.N.) value for coliform organisms low t o diminish the possibility of including pathogenic organisms and hence prevent the possibility of illness in cases where subsequent reheating might not kill all such organisms. VI. PATHOGENIC BACTERIAIN FROZEN FOODS Hartsell (1951) prepared packs of beef and peas inoculated wit.h various strains of Salmonella. He showed that these pathogens would survive for many months a t -9” C. and -17” C. ( 1 6 O F. and 1” F.). The cells

202

GEORG BORGSTROM

seem t o have their metabolic requirements markedly altered, since with a highly nutritive medium much larger numbers were recovered than on selective media. A greater destruction of cells on frozen beef and on peas was observed at -9" C. (l6O F.) than at -18" C. (0" F.). The order of resistance, from greatest t o least, of cultures on stored frozen beef or peas, was Staphylococcus aureus, Salmonella oranienburg, 8.typhosa, and S. dysenteriae. Wallace and Park (1933) determined the survival rate for pathogenic bacteria inoculated into frozen cherries and cherry juices. Similar studies were performed later for strawberries by McCleskey and Christopher (1941) concerning various intestinally active species of Staphylococcus, Eberthella, and Salmonella. A certain per cent of the inoculated bacteria survived even after 8 months of storage (see Table 11). See also p. 167. In this connection the remarkable observation was made th a t the destruction of these bacteria was far more effective at room temperature and even at 0" C. (32" F.) than in the frozen product. It was not indicated whether this was dependent on cell structure or on definite bactericidal effects of the strawberry juice. Similar destruction of bacteria has been observed in apple juice. A strong germicidal effect has also been demonstrated in orange juice concentrates. Even stock cultures of Escherichia coli, Salmonella typhcsa, and Shigella paradysenteriac inoculated into the concentrate, which was subsequently frozen, could not be recovered after 24 hr. of storage a t -17" C. (1" F.). This effect has been attributed t o p H and the organic acid molecule. T o a lesser degree, small amounts of the orange peel oil exert a toxic effect on these pathogenic bacteria (Hahn and Appleman, 1952a,b). Fecal coliforms are particularly sensitive and do not survive. Of special interest is the botulinum problem. Studies on the behavior of the Clostridium botulinum organism, when inoculated into frozen vegetables, yielded extremely contradictory results. The fears uttered by experts concerning the risk of the formation of dangerous toxins b y this particular bacteria was finally dispelled by an investigation published in the late thirties (Prescott and Geer, 1936). It was shown that Clostridium botulinum endured freezing well, but did not grow. I n a few specific cases the formation of toxins was noticed when the thawed product was stored a t room temperature (Tanner and Wallace, 1931 ;Strakaand James, 1932, 1933; Wallace and Park, 1933). Other investigators could not find toxin (Berry, 1938; Edmondson et al., 1922; Prescott and Geer, 1936). Defrosted frozen foods, inoculated with Clostridium botulinum proved t o contain toxins, often after 4 days at 20" C. (68" F.)(Tanner and Oglesby, 1936; Tanner et al., 1940). At 15" C. (59" F.) the formation of toxins took place a t an appreciably slower rate, e.g., spinach and lean beef developed

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

203

toxins in certain cases in 4 days a t temperatures as low as 10" C. (50"F.). The competition with other bacteria, however, generally is responsible for the fact that C1. botulinum does not have sufficiently favorable conditions for development. An extensive investigation has recently been reported in this field in which several important vegetables were examined (Perry et al., 1948). I n no cases were toxins found, not even after storing the thawed product for 10 days a t 21" C. (70" F.). Although the commercial freezing of vegetables in the United States has increased from 4994 tons a year in 1930 t o 454,000 tons in 1952, and although considerable quantities of vegetables have been frozen in home freezers and locker plants, no case of botulism caused by frozen foods has as yet been recorded. This is even more remarkable when one considers that industrial, as well as home freezing and storing was, in many cases, done in an unsatisfactory way. One case of botulism is reported from locker frozen fish kept a t home 24 hr. prior to freezing (Dolman et al., 1950). Many investigators have stated that other bacteria inhibit the development of Cl. botulinum or possibly destroy any toxin that may form. Among such antagonistic forms are: Aerobacter aerogenes, Bacillus subtilis, Clostridium sporogenes, Escherichia coli, Lactobacillus casei, Proteus vulgaris, Streptococcus lactis, and Streptococcus thermophilus (Hall and Peterson, 1923; Jordan and Dack, 1924; Dack, 1926; Sherman et al., 1927; Stark el al., 1928; Kayukova and Kremer, 1940; Ramon et at., 1944, 1945). When thawed vegetables were stored a t room temperature by Berry (1933b), acid-forming bacteria developed most rapidly. Liquid eggs were experimentally inoculated with pathogenic bacteria, then frozen a t -25" C. (-13" F.), and stored a t -lo C. (30" F.) and a t - 18" C. (0" F.). The following organisms survived storage for periods up t o 10 months a t - 18" C. (0" F.) : Salmonella typhosa, Salmonella oranieaburg, Escherichia coli, Salmonella aertrycke, and Staphylococcus aurerrs. Sensitivity of the organisms increased in the order given above. Destruction was greater at, -1" C. (30" F.) than a t -18" C. (0" F.) (Hartsell, 1949). Schneider and Gunderson (1949) concluded that freezing only partly eliminated the surface bacteria, chiefly Salmonella, on eviscerated chicken even after long storage. Schoening et al. (1949) showed that lyophilizing S. cholera-suis had no effect on virulence. Cherry et al. (1946) report the recovery of a nonmobile Salmonella from the skin of frozen turkeys and Browne (1949) 8. fyphi-murium after 13 months of frozen storage. Six strains of Salmonella were studied and inoculated into frozen chicken stew by Gunderson and Rose (1948a). The death rate was directly proportional t o the concentration of organisms, which emphasizes the great importance of plant sanitation. The mortality of Escherichia coli

204

GEORG BORGSTROM

and Aerobacter aerogenes also has been studied under similar conditions. In neither case mas a full pasteurizing effect reached. I n 2 or 3 cases the bacterial content remained high, even after 5% months of storage a t -25 to -30" C. (-13 to -22" F.). Similar observations were made by Proctor and Phillips (1947) in studies on cooked foods, experimentally inoculated with pathogenic strains of Staphylococcus, Streptococcus, and Salmonella. These foods (which were chicken a la king, creamed salmon, beef stew, and cooked shrimp) were frozen and stored at - 18" C. (0" F.). I n a few cases 10% or more of the organisms survived storage for 6 months.

PROBLEMS VII. DEFROSTING When food is defrosted, the bacteria are liberated and immediately begin multiplication and hence decomposition i.e. chemical breakdown of the food. Under industrial and institutional conditions, spoilage may be considerable. It is thus essential t o maintain control of the microorganisms after thawing. The problem of how bacteria grow and multiply after defrosting a frozen food apparently has not been investigated extensively. A voluminous literature is available with reference to the total number of viable cells in frozen foods immediately after defrosting, but little attention has been given to later intervals and t o the growth requirements of the surviving cells. Even during the very thawing of the frozen foods, bacteria start growing within the temperature limits characteristic for the development of respective strains. The higher the external temperature is kept, the more favorable are the growth conditions for most bacteria. Thus defrosting a t a low temperature is the best method for keeping the bacterial count a t a low level (Winter and Wrinkle, 1 9 4 9 ~ ). Rapid thawing can be readily accomplished by the use of highfrequency heating (Gilb, 1945; Sherman, 1946; Cathcart and Parker, 1946; Cathcart, 1946; Bartholomew, 1948). It must, however, be borne in mind that there is no evidence that dielectric defrosting or heating has a specific "bactericidal effect even though this has been often claimed. Of this type of radiation, only the ultraviolet has any specific bactericidal effect. The earlier assumption of the existence of a point-heating effect sufficient t o kill microorganisms has been disproved by Higasi according t o Asami (1950). High-frequency energy has no effect on a material other than that of inducing heat in it. It is, therefore, necessary to use another medium than high-frequency energy for bactericidal effects. But the short-time defrosting obtained with this method does indirectly imply that few microorganisms manage to develop. Actually, frozen foods deteriorate principally in the same manner as l1

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

205

fresh products, however, they decompose more rapidly. Consequently thawed frozen products may be regarded as slightly more perishable than fresh ones. Particularly this refers t o fruits and vegetables, where the tissue disintegrates and the open and shattered cells permit a rapid growth of microorganisms. The bacteria invade the partly softened tissue easier than fresh intact tissue. Whether or not freezing storage influences the subsequent growth rate of the bacteria physiologically is still an open question. Hartsell (1951) noted that E . coli was stimulated by being held in freezer storage and showed a greatly shortened lag phase after freezing and thawing. Sulzbacher (1952), however, found no evidence for the belief that frozen meats are more perishable after thawing than fresh meat. Studies on viability and the effect of longer freezer storage remain t o be made. A group of bacteria of great hygienic importance are the species of Staphylococcus. Canadian research workers have shown t h a t if frozen foods are stored after thawing at temperatures below 10" C. (50" F.) bacteria belonging t o this group do not develop. Frozen vegetables sometimes get sour. This is particularly the case with asparagus, peas, and yellow and green beans. The souring agent is a special bacteria (Streptococcus fecalis) which develops luxuriantly immediately when defrosted or prior t o freezing (Sanderson and Fitzgerald, 1940). This species has been found on peas in the pod. It is a strongly acid-forming species, rendering the product inedible long before other microorganisms have had time t o form toxins. It has been proposed t h a t frozen foods be inoculated with this or some other acid-forming bacteria as a safety factor, by means of which the frozen foods would be given a sour taste before toxicogenic forms could produce any injurious enterotoxins (Prescott and Geer, 1936 ; Berry, 1933c; Prescott et al. 1932; Wallace and Park, 1933; Smart, 193.2, 1939a,b; Wallace and Tanner, 1935). Since poisoning by frozen foods occurs so rarely, such an inoculation with spoilage organisms may, however, be considered a very doubtful procedure. The rich bacterial flora which is generally found in frozen products offers the best protection against infections. If stored for a long time a t a temperature favorable for bacterial development, these bacteria are most likely t o develop taint or odor-producing substances long before toxins may possibly appear. In other words, there is a warning long before there is danger. The risk of a dangerous bacterial development in fruits and vegetables in connection with thawing may as a rule have been overestimated. I n samples containing one of the acid-forming bacteria from peas, the number of bacteria grew in 24 hr. from 1600 t o 5000 per g. a t 4.5" C. (40' F.), whereas a t 21" C. (70" F.) the increase was t o 2,000,000 (Berry, 1946c). After 6 days the number of bacteria a t 4.5" C. (40" F.) was still less than

206

GEORG BORGSTROM

100,000. If the thawed foods are stored in a refrigerator the risks of both bacterial development and impaired quality are slight, if the storage period is not extended for more than a week. Refreezing is, however, as a rule not advisable as the number of bacteria has increased during thawing, and the structure and nutritive value are also adversely affected b y enzymes contained within the plant cells themselves. There is also the risk th a t the thawing or storage may be carried out in a careless manner and a t a higher temperature than in a refrigerator [about 2 t o 4" C. (36 t o 39" F.)]. Often thawed packages are stored in the kitchen or in a shop. According t o the investigations made by Sanderson (1941) storing for 7 days a t zero centigrade implies no risk, not even for vegetables, but at 30" C. (86" F.) the product spoiled in 12 hr. Fruit thawed a t room temperature keeps for-24 hr.; however, vegetables deteriorate under the same conditions (Fellers, 1932). It is well known that bacteria do not grow nearly as rapidly in the acid media of fruits as under the more neutral conditions prevailing in vegetables. Recently figures were published as to the keeping quality of frozen orange concentrate after defrosting. It was preferable to hold the concentrate below 5' C. (41" F.) (Miller and Marsteller, 1952). Yeasts, which constitute the principal surviving organisms, will grow in concentrated orange juice stored at 6" C. (43" F.) and will cause spoilage (Patrick, 1949). With regard t o frozen eggs, an important practical point is that they cannot be thawed rapidly by methods normally available in bakeries. At the same time the length of the defrosting period is a most decisive factor in determining the number of bacteria developing. Plate counts often increase 100% during the thawing of large size cans (30 lbs.) of frozen whole eggs. Egg yolk thaws more rapidly than both egg white and mixtures of the two. The white thaws most slowly because heat conduction is lowest for the whites, A repeated shaking of the containers accelerates thawing and hence the bacterial content may be kept low during defrosting (Winter and Wrinkle, 1 9 1 9 ~ ) . I n thawing egg a t 15" C. (59" F.) bacteria grow more slowly than a t 20" C. (68" F.). Remarkably enough, i t has been stated that, if liquid egg is kept a t temperatures lower than 13" C. (55" F.), so few bacteria on the whole develop that it is of no practical importance (Winter and Wrinkle, 1949~).Hence is derived the assumption that psychrophilic types are almost absent. This implies the practical advantage th a t liquid egg can without any inconvenience be kept in ordinary cold storage for a couple of days. Nor does the transport of frozen egg require a strict maintenance of fixed limits of temperatures, as the product can without risk be allowed

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

207

t o thaw and the temperature raised t o about 12" C. (54" F.) for as long as 3 days. On an average, yolk contains more bacteria than whites whereas a mixture of both yields medium values. This is in accordance with the fact that initially they show similar differences in bacterial contents. As thawing proceeds from the surface inward, a higher bacterial count is found in the exterior areas than in the interior areas (Winter and Wrinkle, 1949c). Liquid egg can be refrozen after thawing without risk if the initial count of bacteria is sufficiently low (Stiles and Bates, 1912). Based on the results of investigations, United States authorities have issued special but fairly mild regulations for the handling of frozen eggs t o be used as raw material for subsequent manufacture into egg powder (Pennington et al., 1914; IicFarlane et al., 1945). They imply t h a t thawing in transit is permissible under the following conditions: The maximum is 4.5" C. (40" F.) for 3 days or 10 t o 25" C. (50 t o 77" F.) for a t most 24 hr. If thawed, they must be held below a maximum of +8" C. (46" F.) for no longer than 16 hr. These regulations may eventually be accepted for the handling of frozen eggs in maritime and railway transport. Esselen and Levine (1954) established a marked increase in aerobic and anaerobic bacteria counts in stuffed chicken after 20 hr. during thawing and holding a t room temperature. This was accompanied by an increase in acidity and development of off-odor. Causey and Fenton (1951) studied the bacterial flora of home-frozen creamed chicken and rice, chicken paprika with gravy, spaghetti and meat balls, and ham patties, before and after 5 methods of reheating. The initial counts were low, less than 2000 per gram in all products. Thawing a t room temperature in most cases resulted in an increase in counts. Organisms occurring were found t o be strains of Bacillus and Staphylococcus. Bacterial pathogens such as Staphylococcus, Streptococcus, and Salmonella in frozen cooked foods increased in number after storage a t defrosting temperatures of 30 t o 37" C. (86 t o 99" F.) for 6 t o 8 hr. At lower thawing temperatures the number of these bacteria increased less rapidly (Proctor and Phillips, 1947). Some increase did occur a t 10" C. (50" F.) storage. All these observations lend emphasis t o the need for proper storage and thawing of all such foods. Straka and Combes (1952) claim t o have demonstrated t h a t creamed chicken containing a pathogen (Micrococcus pyogenes v. u w e u s ) could be defrosted a t 25" C. (77" F.) over a 10 t o 11-hr. period without encountering excessive counts. It has often been reiterated t h a t the number of microorganisms surviving is primarily influenced by the freezing rate. But it is equally true that the length of time for defrosting has a profound influence. Slow

208

GEORG BORGSTROM

freezing and thawing give the lowest counts for surviving organisms, although as a general rule, rapid freezing and defrosting result in less destruction. There are, however, certain aspects of the defrosting which require further studies. Thus far we have too little information t o indicate what thawing might do toward diminishing the number of organisms (Stille, 1942). As temperatures between -4 and -10' C. (25 and 14" F.) destroy bacteria and spores rather quickly (as has been well established when studying counts during the freezing process) it would be worth while t o elucidate what actually happens when thawing products pass this temperature zone considered so dangerous t o microorganisms in the freezing stage (p. 165). It is possible, although not likely, t h a t in passing through this zone from a lower t o a higher temperature there is a corresponding effect on the protein structure t o what occurs during freezing. It is most likely that the process of denaturation is continued as thawing generally takes place less rapidly than the freezing. Perhaps prolonged freezing storage also may affect the physiological reactivity in some way or other, similar t o what often takes place in higher plants when cold treatment stimulates subsequent growth. Such destruction during defrosting along with a subsequent growth stimulation has now (1954) been experimentally established a t SIK (in press). I n order t o be certain t h a t frozen foods do not thaw during distribution a number of indicators have been invented (mostly in the United States) which show by coloring of strips, or other reactions, whether or not the product in question has a t some time undergone thawing (Andersen, 1949; Ramstad and Volz, 1950). The earlier indicators were based mainly on the principle that a printed statement of some kind was eradicated by coming into contact with a thawing liquid. These could not give sufficient indication of for how long a period the product had been in a thawed state; this is the reason why new t8ypeshave been developed. The most reliable one consists of a gelatin gel, enclosed in a n oxygen-permeable transparent envelope. A gel containing a n enzyme system (an oxidase), a colorless phenolic compound, and a suitable amount of ascorbic acid is packed in a plastic envelope. Intensity of the color which develops upon thawing by darkening resulting from oxidation of the phenolic compound measures the time during which liquid from the thawed product has been in contact with the indicator (Ramstad and Volz, 1950). The ascorbic acid regulates the rate of the color development. The color developed is formed by reactions similar t o those taking place in enzymatic browning of fruit tissue. By means of these indicators we are able t o tell whether or not a package has been kept frozen during the entire period of storage and distribution. Furthermore, we may also estimate the length of time a package may have been thawed.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

209

Frozen foods should always be stored on a dunnage and not adjacent t o exterior walls, due t o the risk of heat penetration causing local defrosting. According t o experiences in the United States, the danger of packages thawing is greatest during transit and in retail shops. VIII. PACKAGING PROBLEMS On purely theoretical grounds it might be inferred that, from a hygienic point of view, paper is safer than a tin can for frozen foods. Such a package cannot be sealed hermetically. It is generally not possible for anaerobic bacteria t o develop even in the interior of food not exposed t o air in paper packages. More important, however, is the fact t h a t the consumer need not risk mistaking frozen foods for canned ones. This eliminates the danger of storing unsterile frozen food for a long period of time a t temperatures which permit bacterial development. The use of tin cans, however, offers great advantages; among others, more rapid freezing, more efficient filling, and easier handling. I n addition, the effective method of immersion freezing may be practiced. This is certainly the reason why the United States frozen food market now offers several products packed in cans, particularly juice concentrates and berries. But in order t o avoid any error some frozen fish products in cans are overwrapped with a n extra paper package. As mentioned above, tests with several vegetables have shown that there is very little danger of the development of Clostridium botulinum in sealed cans, even after several days of storage in a thawed state a t room temperature (Perry et al., 1948). It is well known that frozen foods are packed in a great many different packaging materials. From a microbiological point of view, i t is of course desirable that any material used should be free from microorganisms and their spores. I n other sectors of the food industry the microbiology of the package has received much attention. This has been the case only t o a limitcd scale in the frozen foods industry. I n the United States packaging materials specially treated against bacteria have been designed for frozen foods (White, 1951). Such packaging materials are used primarily for frozen fish fillets in order t o destroy surface bacteria. It is highly probable t h a t in the future greater attention will be given t o the microbiological problems concerned with the packaging of frozen foods. Of great interest, in point of principle, are the observations reported by Weiser (1951) t o the effect that microorganisms in frozen berries, which are not hermetically packed, rapidly increase a t -2 t o -4' C. (28 t o 25" F.) whereas those properly sealed show a heavy decrease. This attributes greater significance t o the use of air-tight packages. This observation becomes particularly important when one considers that, during the period when freezing heat is released and the temperature is

210

GEORG BORGSTROM

stabilized a t a level immediately below 0" C. (32" F.), aerobic bacteria may grow freely if air is admitted. No significant effect due to packaging material was revealed through bacterial studies on dressings of frozen stuffed poultry (Esselen and Levine, 1954).

IX. COOKING As already pointed out, the preparation of several frozen foods before they are eaten also has a prophylactic significance. All preserved vegetables are thoroughly heated in boiling water (spinach, cauliflower, peas, etc.) . Bakery products, stews, and ready-to-serve dishes of fish and chicken are not always effectively heated through, yet the temperature is sufficient to call this procedure a pasteurization (even if it does not always have a completely sterilizing effect) greatly reducing the number of bacteria. Observations have been made on the bacterial flora of commercially frozen precooked chicken a la king, beef stew, and creamed fish as they reached the kitchen and at different stages during subsequent kitchen procedures, such as cookery and refrigeration. Although cooking reduced the number of all kinds of microorganisms, it did not eradicate any type. Continued multiplication of bacterial cells was observed in chicken a la king and beef stew under conditions of household refrigeration (Hussemann, 1951). X. HYGIENICASPECTS The reason for this survey is, of course, to collect as much material as possible for estimating the hygienic dangers connected with frozen foods, and, if possible, to specify under what conditions safety may be considered to the same degree as with other foods in the market. This problem received early attention in the United States (Fellers, 1932; James, 1932), and the progress of frozen foods in the thirties and forties was watched with some uneasiness, particularly since adequate quality control did not exist. Especially disturbing were the obvious risks of toxin formation by Clostridium botulinum and species of Staphylococcus. But nothing happened despite the fact that the raw products originated from all parts of the continent and preparation methods varied considerably (Tressler, 1946; Fitzgerald, 1947a). During the 6 to 7 years that ready-to-serve foods have been in the United States market, not a single case of food poisoning caused by frozen products has been recorded or established with certainty. As for other frozen foods, only one case of poisoning (from fish) has been reported (White, 1951). The bacterial count, therefore, is more of an indication of the standard of factory sanitation rather than of the unsanitary condition of the

MICHOBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

21 1

product. As in other foods, the important aspect is concerned with the kinds of bacteria which develop and t o what extent these cause intestinal infections or affect consumers in other ways. Water bacteriology is of prime importance according t o experiences in the United States. Frozen food factories, therefore, have been installing chlorination systems for water used in plants. This has resulted in a considerable reduction of bacterial counts in the frozen foods (Scarlett and Martin, 1948). I n the United States, 90% of the water supplied by the cities is chlorinated. As is the case in certain sectors of the canned food industry, there has been a tendency within the frozen food industry t o underestimate the importance of strict sanitary handling of the products during preparation for preservation. The subsequent heating or freezing was supposed t o kill the bacteria rather effectively and the risks were regarded a s very slight. This is not recommendable for either hermetically sealed products or frozen foods. The statements made above show clearly how important it is t o keep the bacterial content a t as low a level as is possible. This is particularly true for frozen foods which will never become sterile, not even after prolonged freezing storage. From the point of view of quality, i t is consequently desirable t o make rapid estimations of bacterial counts regularly during manufacture, using the methods described earlier in the review. The bacteria occurring in frozen foods are, as a rule, harmless t o man. From a public health point of view, however, i t is important t o test the products from time t o time for the presence of fecal bacteria (above all the Escherichia group) SO that the possibility of such contamination ma\be eliminated. The simplest preventive measure in factories is, however, t o undertake regularly a thorough cleaning of all equipment with bactericidal substances. The sanitary problems in the processing and distribution of frozen foods have been well reviewed by Evers (1950). X definite health hazard is the one of frozen fruits and vegetables in ram salads which could easily, through carelessness, be a source of infection.

XI. PI~ACTICAL ASPECTS The practical consequences of investigations on the bacteriology of frozen foods may be summarized briefly in the following way: Most frozen products are not sterile, i.e. they contain natural forms of bacteria (above all spores) which may develop as soon as the external conditions are favorable. The same is also true of spores of certain molds or yeasts. Cleanliness and good hygiene, consequently, must be maintained in all aspects of the handling of foods intended for freezing. The handling of r u w products in factories must, therefore, be done carefully, and, in reality, the requirements in this respect are greater than for other methods

212

GEORG BORGSTROM

of preservation. A regular cleaning of plant and equipment should be considered as routine in a frozen food plant. Furthermore, preparation of products should be done as rapidly as possible. The longer it requires t o get the products ready for freezing, the more bacteria will develop to impair quality. The demand for cleanliness, hence, is ultimately a prerequisite for good quality. The time between harvest of fruits and vegetables, catching of fish, slaughtering of meat and poultry, and freezing must be as short as possible, and the product should be kept as near 0" C. (32" F.) as possible in order t o assure a minimum rate of bacterial growth, If a reassuring sanitary control of plants is provided for, the risk of infection with pathogenic bacteria may be of minor importance. Sanitary measures, in this case, must be extended even to the personnel. Strict demands concerning personal hygiene will have t o be maintained. All water used for cleaning containers, preparation, and rinsing must be clean and, as far as possible, free from bacteria. All equipment which is used regularly should be washed. A strict schedule should hereby be observed. I n this connection it must be remembered that there is a n essential difference between freezing and such preservation where most of the bacteria are killed by heating. Only a fraction is rendered harmless in freezing as is t o be seen from this review. I n accordance with the discussion given in the section on sanitation, every freezing plant should have regular bacteriological controls run on the finished products. Raw products also should be subjected t o microbiological examination. The most crucial risk occurs during thawing, when conditions are such that the chances for the development of surviving bacteria and their spores are greatest. In the preparation of vegetables for freezing, the tissues are so affected th at bacteria, yeasts, and molds can develop considerably faster than is possible in canned products. As has already been pointed out, vegetables are exposed t o a far-reaching sterilization in blanching. Consequently, the bacterial contamination takes place between blanching and freezing. It is, therefore, particularly important to avoid any delay and to freeze the products as rapidly as possible after blanching. Precooked frozen foods are subject t o similar risks. I n connection with the cooking and preparation of the raw products large numbers of bacteria may be destroyed, but subsequent contamination may take place. Even though dangerous bacteria might not be involved in this case, generally there are bacteria th at may cause a deterioration in quality. It is entirely justified to conclude th at most spoilage encountered in the frozen food field is due to improper handling of the product either prior t o freezing or after it is thawed. Freezing and storage conditions play a secondary role.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

213

Due t o a break in the electric current or t o pure neglect, i t may happen that frozen foods are unintentionally allowed t o thaw. This risk is also present during transit and in retail shops. If the thawing has not proceeded sufficiently t o be complete, the package may in exceptional cases be refrozen, but such packages should be used as soon as possible and previous t o other unthawed ones. The keeping quality is no longer the same. If the thawing has been complete and the temperature of the products has risen above 0” C. (32” F.), the measures t o be taken depend on the type of food in question. Meat, poultry, and fish which still have a n unchanged fresh odor (determined by opening some packages) may be refrozen if the product appears fresh upon examination and if a rapid estimation of bacterial load gives a reassuring result. But even in this case i t is essential t h a t such a product be consumed within a short time. It must, however, be taken into consideration t h a t enzymic activity is stimulated by such a procedure and this inevitably leads t o a n inferior product. Fruit can also be refrozen, but as a rule i t is better t o make jam, compote, jelly, etc. of such lots. Shellfish and vegetables should neuer be refrozen. They should be boiled without delay or used in cooking. If they have been standing a t temperatures higher than 10 t o 15”C. (50 t o 59” F.), they should not be used a t all for dangerous bacteria may have had a chance t o develop, and no such risks should be taken. Finally, consumer information on the correct procedure for handling frozen foods is of great importance for eliminating risks. The experiences gained up till now in all countries of the world show that this seems t o have met with considerable success.

REFEREXCES Andersen, A. A. 1949. A defrosting indicator for frozen foods. Food Technol. 3, 357-58. Anonymous. 1933. Microbiology of frozen foods. Ice and Refrig. 86, 75-76. Anonymous. 1940. Public health aspects of frozen foods with particular reference to t h e products frozen in cold storage lockers and farm frcwwrs. A m . Public Health Assoc. Y e a r Book 30(2), 77-83, 10th Year (1940-41). Suppl. to Ana. J . Public Health. Anonymous. 1941. The microbiological examination of eggs and egg products used in frozen desserts. in “Standard Methods for the Examination of Dairy Product,” 8th ed., p. 191-205. American Public Health Association, New York. Anonymous. 1944. Egg products (frozen). Chicago Quartermaster Depot, Subsistence Research and Development Laboratory. CQD No. 169 -4, Dec. 12. Anonymous. 1945a. Microbiological methods: Examination of eggs and egg products. in “Official and Tentative Methods of Analysis,” 6th ed., p. 750-52. Association of Official Agricultural Chemists, Washington, TI. C. Anonymous. 1945b. Controlling bacteria in freezing peas. Food Packer 26(5), 68-70. Anonymous. 1945c. Operating standards for egg storage, breaking, freezing and drying plants. Chicago Quartermaster Depot, Subsistence Research and Development Laboratory. Tech. Bull. No. 2A, amended July 13, 1945.

214

GEORG RORGSTROM

Anonymous. 1946a. Bacteriology of frozen foods. in “Refrigerating Applications Data Book,” 2nd ed., p. 77-85. American Society of Refrigerating Engineers, New York. Anonymous. 1946b. Microbiological examination of foods. Am. J . Public Health 36, 332-35. Anonymous. 1946c. Storage of canned foods. in “Refrigerating Applications Data Book,” 2nd ed., p. 283. American Society of Refrigerating Engineers, New York. Anonymous. 1948. Microbiological methods for frozen dessert ingredients. i n “Standard Methods for the Examination of Dairy Products,” 9th ed., p. 171-98. American Public Health Association, Kew York. Anonymous. 1949a. Destruction de la trichine dans la viande de porc. Rejrigeration, Cold Storage and A i r Conditioning Nov., p. 22. Anonymous. 1949b. Microbiology of low temperature food preservation. Australian Food Mfr. 18(9), 18. Anonymous. 1949c. Recommended specifications for standard packages and packs for frozen eggs. U.S. Dept. Agr., 24 pp. Anonymous. 1949d. Canada Live Stock and Live Stock Products Act, 1949. The frozen egg regulations. p. 590. Anonymous 1950a. Frozen foods. Public Health Repts. (77.8.)66, 1645-46. Anonymous. 1950b. Defense Department tests frozen meat palatability. Quick Frozen Foods 12(12), 69-74. hnonymous. 1953. Long-freezing of pork stops painful trichinosis. Science ,liews Letter 63, 59. Appleman, M.D., and Hahn, S. S. 1951. Factors limiting survival of enteric organisms in frozen orange concentrate. Bacteriol. Proc. p. 35. Asami, Y. (ed.) 1950. “Theory and Application of High-Frequency Phenomena.” Hokkaido University, Sapporo, Japan. Aschehoug, V., and Vesterhus, R. 1946. Bacteriological investigation on spoilage of winter herring during storage. Food Research 12(1), 57-76. Babcock, C. J., Roerig, R. N., Stabile, J. N., Dunlap, W. A,, and Randall, R. 1947. Frozen homogenized milk. 11. Effect of freezing and storage temperatures on the chemical and bacteriological properties of homogenized milk. J . Dairy Sci. 30, 49-54. Barnes, H. T. 1925. Collodial water and ice. Colloid Symposium ikronograph 3, 103-11 I. Bartholomew, J. W. 1948. Utility of high frequency heating in the frozen food industry. Quick Frozen Foods 11(4), 59-61. Beard, P. J., and Cleary, J. P. 1932. The importance of temperature for the survival time of bacteria in acid foods. J . Preventive Med. 6, 141-44. Beckwith, T. D. 1936. hlolds in cold storage. Ice and Refrig. 90, 159-60. Becquerel, P. 1950. La suspension de la vie au-dessous de 450 “K absolu par dCmagnetisation adiabatique de l’alun de fer dans le vide le plus ClevC. Compt. rend. 231, 261-63. Becquerel, P.1952. La suspension de la vie aux confins du zero absolu. Scientia ( M i l a n ) Ser. VI, 46, 243-247. Bedford, R. H. 1931. The growth of some marine and other bacteria at low temperatures. Fisheries Research Board Can., Progr. Repts. Pacific Coast Sta. No. 11, pp. 14-17. Bedford, R. H. 1933. Marine bacteria of the Northern Pacific Ocean. The temperature range of growth. Contribs. Can. Biol. and Fisheries 7(34), 433-38. Bedford, R. H. 1934. A note on bacteria and low temperature storage. Fisheries Research Board Can., Progr. Repts. Pacific Coast Sta. 22, 14.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

215

Beisel, C. G. 1951. Controlling contamination in a citrus plant. Canner 113(23), 16, 18;113(24), 19-20. Beisel, C. G., and Troy, V. S. 1949. The Vaughn-Levine boric acid medium as a screening presumptive test in the examination of frozen concentrated orange juice. Fruit Products J . 28, 356-57. BelehradBk, J. 1935. “Temperature and Living Matter,” 277 pp. Borntraeger, Berlin. Berry, J. A. 1932% Bacteria question in cold packing. Western Canner and Packer 23(10), 17-18. Berry, J. A. 1932b. How freezing affects microbial growth. Food I n d s . 4, 205. . of the frozen pack. Cawning Age 13, 251-54. Berry, J. A. 1 9 3 2 ~Microbiology Berry, J. A. 1932d. Strax-berriesfrozen 15 months maintain quality. F r u i t Products J . 11, 183. I3erry, J. A. 1933a. Destruction and survival of microorganisms in frozen pack foods. J . Bacteriol. 26, 459-70. Berry, J. 8.1933b. Lactobacilli in frozen pack peas. Science 77, 350-51. Berry, J. A. 1933c. Microbiology of frozen pack berries and vegetables. Ice and 12efrzy. 84, 204-05. Berry, J. A. 1933d. Some findings on microbiology of the cold pack. Canning Age 14, 445-46, 463. Berry, J. A. 1934. Cold-tolerant micro-organisms and frozen pack. Canner 78(11) 13-14. Berry, J. A. 1936. Microbiological studies of frozen pack berries, with special reference to effects of carbonation. Proc. Am. SOC.Hort. Sci. 33, 224-26. Berry, J. A. 1037. Microbiology of frozen pack vegetables. Western Canner and Packer 29(5), 14-16. Berry, J. A. 1938. No danger from botulinus in frozen fruits or vegetables. Western Canner and Packer 30(4), 32. Berry, J. A. 1941. The fewer the bacteria, the better the frozen pack. Canner 94(4), 13-14. Berry, J. A. 1946,. Sanitation of frozen foods. U.S. Dept. Agr., Bureau Agr. and Ind. Chem. AIC-120, 4 pp. Berry, J. A. 194613. Bacteriology of frozen foods. Food Packer 27(10), 51-52. Berry, J. A. 1946c. Bacteriology of frozen foods. J . Bacteriol. 61, 639. Berry, J. A, and Diehl, H. C. 1934. Freezing storage in relation to microbial destruction and retention of quality in sweet cider. Proc. Am. Soc. Hoit. Sci.31, 157-59. Berry, J. A,, and Magoon, C. A. 1934. Growth of microorganisms a t and below 0” C. Phytopathology 24, 780-96. Bidault, C. 1922. The molds of frozen meat. Rev. ge‘n. froid 3, 246-58. Birdseye, C. 1929. Some scientific aspects of packaging and quick-freezing perishable flesh products. 111. Sanitary measures in a fish dressing plant. I n d . Eng. Chem. 21, 824-57. Bland, F. 0. S. 1047. Frost and food. Modern Refrig. 60, 108. Brokaw, C. H. 1952. The role of sanitation in quality control of frozen citrus concentrates. Food Technol. 6, 344-49. Brooks, F. T. 1924. Molds on frozen meats. J . SOC.Chem. I n d . (London) 43, 306T. Brooks, F. T., and Kidd, M. N. 1921. The “black spot” of chilled and frozen meat. Dept. Sci. I n d . Research (Brit.) Food Invest. Board Special Repts. No. 6 , 6 pp. Brown, E. B. 1933. Bacterial studies of defrosted peas, spinach, and Lima beans. J . Home Econ. 26, 887-92. Brown, H. J., and Gibbons, N. E. 1950. Enterococci as an index of fecal contamination in egg products. Can. J . Research F28, 107-17. ~

216

GEORG BORGSTROM

Browne, A. S. 1949. The public health significance of Salmonella on poultry and poultry products. Ph.D. Thesis. University of California, Berkeley. Brownless, D. S., and James, L. H. 1939. Bacterial contamination of frozen whole eggs and a n improved method of defrosting. Proc. 7th World's Poultry Congr. pp. 488-92. Buchbhder, L., Osler, A. G., and Steffen, G. I. 1948. Studies in enterococcal food poisoning; isolation of enterococci from foods implicated in several outbreaks of food poisoning. Public Health Repts. (U.S.) 63, 109-118. Buchbinder, L.,Loughlin, V., Walter, M., and Dangler, G. 1949. A survey of frozen precooked foods with special reference to Chicken & la King. J . Mzlk and Food Technol. 12, 209-13, 231. Burton, hl. 0. 1949a. The types of coli-aerogenes organisms occurring in frozen vegetables. Can. J . Public Health 40, 361-63. Burton, M. 0. 1949b. Comparison of coliform and enterococcus organisms as indices of pollution in frozen foods. Food Research 14, 434-38. Burton, E. F., and Oliver, W. F. 1935. The structure of ice a t low temperatures. Proc. Roy. Soc. A163, 166-72. Campbell, H. 1932. The viability and growth of certain microorganisms common t o frozen pack fruits and vegetables a t temperatures below zero. Masters Thesis. University of Washington, Seattle. Castell, C. H. 1947. Growth of Clostridiurn in seaweeds and marine fish. J . Fisheries Research Board Can. 7(2), 62-69. Castell, C. H., and Anderson, G. W. 1947. Effect of salting and smoking on survival and growth of Clostridiurn in fish. J . Fisheries Research Board Can. 7(2), 70-73. Castell, C. H., and Anderson, G. W. 1948. Bacteria associated with the spoilage of cod fillets. J . Fisherzes Research Board Can. 7(6), 370-77. Castell, C. H., and McDermott, L. A. 1942. Multiplication of bacteria in water and its significance in food spoilage. Food Research 7, 244-53. Castell, C. H., and hlapplebcck, E. G. 1952. The importance of flavobacterium in fish spoilage. J . Fisheries Research Board Can. 9(3), 148-56. Cathcart, W. H. 1946. Frozen foods defrosted by electronic heat. Food Inds. 18, 1524-25. Cathcart, W. H., and Parker, J. J. 1946. Defrosting frozen foods by high-frequency heat. Food Research 11, 341-44. Causey, K., and Fenton, F. 1951. Effect of reheating on palatability, nutritive value and bacterial count of frozen cooked foods. I. Vegetables. J . Am. Dieiet. Assoc. 27, 390-95. 11. Meat dishes. Ibid. 27, 491-95. Cherry, W. B., Barnes, L. A., and Edwards, P. R. 1946. Observations on a monophasic Salmonella variant. J. Bacteriol. 61, 235-243. Colien, F. E. 1942. The microbiology of frozen eggs. I. Presence and significance of coliform organisms in frozen eggs. J . Bacteriol. 42, 47. Dack, G. M. 1926. Influence of some anaerobic species on toxin of Cl. BotuZinunz with special reference to Cl. Sporogenes. J. Infectious Diseases 38, 165-73. Dangler, G., and Steffen, G. I . 1948. Studies on enterococcal food poisoning. A study of the incidence of enterococci and staphylococci in suspected food in outbreaks of food poisoning. J. M i l k and Food Technol. 11, 242-43. Davis, J. G. 1951. The effect of cold on micro-organisms in relation to dairying. J . A p p l . Bacteriol. Proc. 14(2), 216-42. Dervichian, D.G. 1949. Lipoproteins. Discussions Faraday SOC.6, 7-15. Devik, O.,and Ulrich, J. A. 1949. Investigations of changes in foods during storage in the frozen condition. M i n n . Univ. Hormel T m t . Ann. Rept. 1948-49, pp. 40-45.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

217

Diehl, H. C. 1945. Refrigeration research foundation reports on microbial spoilage of frozen foods. Ice and ReJrig. 108(5), 49-50. Dolman, C. E., Chang, H., Kerr, D. E., and Shearer, A. R. 1950. Fish-borne and type E Botulism: Two cases due to home-pickled herring. Can. J . Public Health 41(6), 215-29. Dorsey, N. E. 1940. Properties of ordinary water-substance in all its phascs: Types of ice. Am, Chem. Soc. Monograph No. 81, pp. 395-97. D u Bois, C. W., and Kew, T. J. 1951. Storage temperature effects on frozen citrus concentrates. Refrig. Eng. 69, 772-75, 812. Dyer, F. E. 1947. Microorganisms from Atlantic cod. J.Fisheries Research Board Can. 7(3), 128-36. Dyer, W. J., and Dyer, F. E. 1949. Amines in fish muscle. IV. Spoilage in freshly cut cod fillets. J . Fisheries Research Board Can. ?(lo), 580-84. Dykstra, K. G., and Smith, S. E. 1950. Sanitation practices in thc frozen food industry. Food Technol. 4, 419-22. Edmondson, R. B., Giltner, L. T., and Thom, C. 1922. Bacillus botulinus: relation of toxin production to temperature. Abstr. Bacteriol. 6, 23. Elrod, It. P. 1942. The Erwinia-coliform relationship. J . Bacteriol. 44, 433-440. Esselen, W. B., and Levine, A. S. 1954. Bacteriological investigations on frozen stuffed poultry. J . Milk and Food Technol. 17(8), 245-50, 255. Evers, C. F. 1950. Sanitary problems in the processing and dist.ribution of frozen foods. J . Milk and Food Technol. 13, 35-39. and Trout, G. M. 1943. Influence of various treatments on the bacteria Fabian, F. W., content of frozen cream. J. Dairy Sci. 26, 959-65. Fabian, J. R. 1946. Microbial growth and quality control. Frozen Food Ind. 2(7), 8-9, 30-41. Faville, L. W., and Hill, E. G. 1951. Incidence and significance of microorganisms in citrus juices. Food Technol. 6, 423-25. Faville, L. W., and Hill, E. C. 1952. Acid-tolerant bacteria in citrus juices. Food Research 17, 281-87. Faville, L. W., Hill, E. C., and Parish, E. C. 1951. Survival of microorganisms in concentrated orange juice. Food Technol. 6, 33--36. Fellers, C. R. 1932. Public health aspects of frozen foods. Am. J. Public Health 22, 601-11. Fenton, F., and Darfler, J. 1946. Foods from the freezer, precooked or prepared. Cornell Ext. Bull. 692, 100 pp. Fieger, E. A. 1950. Problems in handling fresh and frozen shrimp. Food Technol. 4, 409-11. Finn, D. B. 1932. Denaturation of proteins in muscle juice by freezing. Proc. Roy. So?. B111, 396-411. Fitzgerald, G. A. 1946. Quality grades are essential for frozen food packs. Food I d s . 18, 874-76. Fitzgerald, G. A. 1947a. Are frozen foods a public health problem? Am. J . Public Health 37, 695-701. Fitzgerald, G. A. 1947b. How to control the quality of frozen cooked foods. Food Inds. 19, 623-25, 730. Fitzgerald, G. A. 1947c. Quality control of frozen foods. Food Technol. 1, 575-79. Fletcher, D. A., and Johns, C. K. 1951. Three years experience in grading frozen egg in Canada. Poultry Sci. 30, 445-48. Frobisher, hi. 1953. “Fundamentals of Bacteriology,” 5th ed., 824 pp. Saiinders, Philadelphia.

218

GEORG BORGSTROM

Funk, E. M. 1938. Improving the keeping quality of eggs by cleaning with sodium hydroxide. Missouri Agr. Expt. Sta. Research Bull. 277, 40 pp. Funk, E. M. 1943. Pasteurization of shell eggs. Missouri Agr. Expt. Sta. Research Bull. No. 364, 28 pp. Gaebelein, U. 1940. Zur Wirkung niederer Temperaturen auf Bakterien und Beobachtungen bei der Lebensmittelkonservierung mittels Kiilte. Vorratsp$ege u. Lebensmittelforsch. 3, 176-91. Geer, L. P., Murray, W.T., and Smith, E. 1933. Bacterial content of frosted hamburg steak. Am. J . Public Health 23, 673-76. Gibbons, N. E. 1934. A bacteriological study of “ice fillets.” Contribs. Can. Bid. and Fisheries 8(24), 303-10. Gibbons, N. E., and Moore, R. L. 1944. Dried whole egg powder XII. The effect of drying, storage and cooking on the salmonella content. Can. J . Research F22, 58. Gibbons, N. E., Fulton, C. O., and Reid, M. 1946. Dried whole egg powder. XXI. Pasteurization of liquid egg and its effect on quality of the powder. Can. J . Research F24, 327-37. Gilb, H. W. 1945. Electronic defrosting of frozen foods. Food in Canada 6 ( 8 ) , 9-10, Gilbert, P. E., and Wiegand, E. H. 1950. Quick method for the determination of strawberry mold contamination. Quick Frozen Foods 12(11), 52-53. Godkin, W. J., and Cathcart, W. H. 1949. Fermentation activity and survival of yeast in frozen fermented and unfermented doughs. Food Techno2. 3, 139-46. Goresline, H. E. 1939. Micro-organisms in foods and food preservation. Yearbook 4 g r . U.S. Dept. Agr., pp. 341-49. Goresline, H. E. 1912. Report on frozen fruits and vegetables. J . Assoc. O$c. Agr. Chemists 26, 736. Goresline, H. E, (chairman). 1946. Microbiological examination of foods. Tentative methods for the microbiological examination of frozen foods. Am. J . Public Health 36, 332-35. Goresline, H. E. 1948. Report on microbiological methods for frozen fruits and vegetables. J . Assoc. Ofic. Agr. Chemists 31, 519-21. Goresline, H. E. 1951. Chlorine washes control bacteria on poultry items. Food Field Reporter 19, NO. 5, 49. Gotlib, M. A. 1951. Influence du froid sur la vie microbienne des aliments. Rev. Cons. (France) 3(6), 45-46. Gray, P. H. H. 1943. Two stain method for direct bacteria count. J . M i l k Techno/. 6(2), 76. Green, M. 1949. Bacteriology of shrimp. 111. Quantitative studies on frozen shrimp. Food Research 14, 384-94. Griffin, A. M., and Stuart, C. A. 1940. An ecological study of the coliform bacteria. J . Bacteriol. 40, 83-100. Griffiths, E., Vickery, J. R., and Holmes, N. E. 1932. The freezing, storage and transport of New Zealand lamb. Dept. Sci. Ind. Research (Brit.) Food Invesf. Board Special Repts. No. 41, 178 pp. Griffiths, F. P. 1937. A review of the bacteriology of fresh marine-fishery products. Food Research 2, 121-34. Gunderson, M.F., and Gunderson, S. D. 1945. Eggs can be washed clean. U.S. Egg & Poultry Mag. 61, 533, 560-62. Gunderson, M. F., and Rose, K. D. 1948a. Survival of bacteria in a precooked freshfrozen food. Food Research 13, 254-63. Gunderson, M. F., and Rose, K. D. 1948b. Cold storage won’t “pasteurize” frozen food. U.S. Egg & Poultry Mag. 64(9), 9-11, 18.

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

219

Gundersen, M. F., Rose, K. D., and Henn, M. J. 1947. The bacteriology of eviscerated poultry. Boned chicken and chicken products. Food Znds. 19, 1515. Gunderson, M. F., Rose, K. D., and Henn, M. J. 1947. Poultry boning plantsneed bacteriological control. Food Znd. 19(11), 105-06. Hahn, S. S., and Appleman, hl. D. 1952a. Microbiology of frozen orange concentrate. I. Survival of enteric organisms in frozen orange concentrate. Food Technol. 6, 156-55. Hahn, S. S., and Appleman, M. D. 1952b. Microbiology of frozen orange concentrate. 11. Factors influencing the survival of microorganisms in frozen orange concentrate. Food Technol. 6, 165-67. IIaines, R. B. 1931. The growth of microorganisms on chilled and frozen meat. J . Soc. Chem. Znd. 60, 223-271'. Haines, R. B. 1934. The minimum temperatures of growth of some bacteria. J . Hyg. 34, 277-82. Haines, R. U. 1935. The freezing and death of bacteria. Dept. Sci. Znd. Research (Brit.) Food Invest. Board Special Rept. 1934, 47-48. Haines, R. €3. 1937a. Microbiology in the preservation of animal tissues. Dept. Sci. I n d . Research (Brit.) Food Invest. Board Special Rept. No. 45, 85 pp. Ilaines, R. B. 193713. Report on freezing studies with bacteria. Proc. 2nd Intern. Congr. illicrobiol., London, 1936, pp. 45-46. Haines, R. B. 1938. The effect of freezing on bacteria. Proc. Roy. SOC.B124, 451-63. Haines, R. B. 1939. Microbiology in the preservation of the hen's egg. Dept. Sci. I n d . Research (Brit.) Food Invest. Board Special Rept. No. 47; Low Temp. Research Sta. Cambridge 34, 277. Hajna, A. A., and Perry, C. A. 1943. Comparative study of presumptive and confirmative media for bacteria of the coliform group and for fecal streptococci. Am. J . Public Health 33, 550-56. Hall, I. C., and Peterson, E. 1923. The effect of certain bacteria upon the toxin production of Bacillus botulinus in vitro. J . Bacteriol. 8, 319-41. Hall, J. E., and Blundell, C. C. 1946. The use of break point chlorination and sterilized water in canning and freezing plants. Natl. Canners Assoc. Inform. Letter No. 1073. Halvorson, H. 0. 1949. Food microbiology. Ann. Rev. Microbiol. 3, 357-70. Hampil, U. 1932. The influence of temperature on the life processes and death of bacteria. Quart. Rev. Biol. 7, 172-96. Hankins, 0. G., and Hiner, R. L. 1941. Quality of meat as affected by freezing temperatures. Refrig. Eng. 41, 185-89. Harshaw, H. hf., Hale, W. S., Swenson, T. L., Alexander, L. M., and Slocum, R. R. 1941. Quality of frozen poultry as affected by storage and other conditions. U.S.Dept. Agr. Tech. Bull. 768, 20 pp. Hartsell, S. E. 1949. The testing of frozen eggs for pathogens. J. M i l k and Food Technol. 12, 107-08. Hartsell, S. E. 1951. The growth initiation of bacteria in defrosted eggs. Food Research 16, 96-106. Hartsell, S . E. 1952. The longevity and behavior of pathogenic bacteria in frozen food: the influence of plating media. Am. J . Public Health 41, 1072-77. Havighorst, C. R. 1949. Cooked-frozen soup by aseptic process. Food Znds. 21, 1234-35. Havighorst, C. R. 1951. Aseptic canning in action. Food Znds. 23, 72-74. Hawkes, L. 1929. Super-cold water. Nature 123, 244. Haynes, R. D., Harlin, H., Mundt, J. O., and Stokes, R. 1953. Xew washing method could reduce mold. Science News Letter 63, 137.

220

GEORG BORGSTROM

Hays, G. L. 1951. The isolation, cultivation and identification of organisms which have caused spoilage in frozen concentrated orange juice. Proc. Florida State Hort. SOC., 64th Ann. Meeting, pp. 135-37. Hays, G. L., and Riester, D. W. 1952. The control of “off-odor’’ spoilage in frozen concentrated orange juice. Food Technol. 6, 386-89. Hegarty, C. P., and Weeks, 0. B. 1940. Sensitivity of Escherichia coli to cold-shock during the logarithmic growth phase. J . Bacteriol. 39, 475-84. Heinmets, F., and Taylor, W. W. 1951. Photobiological studies on Escherichia coli a t low temperatures. J . Bacteriol. 62, 477-85. Heitz, T. W., and Swenson, T. L. 1933. The quick freezing of dressed poultry. Ice and Refrig. 86, 163-65. Hendrickson, R. L. 1949. Bacterial studies related to the keeping quality of frozen pork sausage. J. Animal Sci.8, 613. Hendrickson, R. L.? and Miller, W. A. 1950. Some observations on the bacteria in frozen pork sausage. Trans. Kansas Bead. Sci. 63, 70-74. Hess, E. 1934a. Cultural characteristics of marine bacteria in relation to low temperature and freezing. Contribs. Can. Biol. and Fisheries 8(32), 461-74. Hess, E. 1934b. The effects of freezing on marine bacteria. J . Biot. Board Can. 1, 95-108. Higley, H. A. 1907. Cold storage poultry fallacies exploded. Food Law Bull. 2(1), 6-8; (2), 21-23. Hill, E. C., and Faville, L. W. 1950. Comparison of plating media used for the estimation of microorganisms in citrus juices. Proc. Florida State Hort. SOC.6Srd Ann. Meeting, 146-49. Hill, E. C., and Faville, L. W. 1951. Studies on the artificial infection of oranges with acid tolerant bacteria. Proc. Florida State Hort. Soc. 64th Ann. Meeting, 174-77. Hill, E. C., Wenzel, F. W., and Barreto, A. 1951. Colorimetric method for detection of microbiological spoilage in citrus juices. Food l‘echnol. 8 ( 3 ) , 168-71. Hilliard, C. M., and Davis, M. A. 1918. The germicidal action of freezing temperatures upon bacteria. J . Bacteriol. 3, 423-31. Hilliard, C. M., Torossian, C., and Stone, R. P. 1915. Notes on the factors involved in the germicidal effect of freezing and low temperatures. Science 42, 770-71. Hollander, D. N., and Nell, E. E. 1954. Improved preservation of Treponema pallidum and other bacteria by freezing with glycerol. A p p l . Microbiol. 2, KO. 3, 164-169. Holmes, D., and McCleskey, C. S. 1949. Bacteriology of shrimp. V. Effect of peeling, glazing and storage temperature on bacteria in frozen shrimp. Food Research 14, 401-04. Holtman, D. F. 1943. The bacteria of fresh and frozen eggs in relation to public health. J . Bacteriol. 46, 50. Hucker, G. J. 1950. Relation of quality of frozen vegetables t o number of bacteria present. Canner 111(22), 12. Hucker, G. J. 1954. Low temperature organisms in frozen vegetables. Food Technol. 8, 79-82. Hucker, G. J., and Brooks, R. F. 1948. Relation of quality of frozen foods to microbiological flora. N . Y . State Agr. Expt. Sta. (Geneua) 67th Ann. Rept., 17 pp. Hucker, G. J., and Robinson, W. B. 1950. Bacteria in frozen vegetables. Farm Research 16(4), 5; Canner 111(20), 16-17; Food in Canada 11, 14-15. Hucker, G. J., Brooks, R. F., and Emery, A. J. 1952. The source of bacteria in processing and their significance in frozen vegetables. Food Technot. 6, 147-55.

MICROBIOLOGICAL PROBLEMS O F FROZEN FOOD PRODUCTS

221

Hucker, G. J,, Pederson, C. S., and Brooks, R. F. 1944. The sanitation of food plants. N . Y . State Agr. Expt. Sta. 63rd Ann. Rept., pp. 12-13. Humphrey, H. J. 1936. Freezing makes rapid progress. Food I n d s . 8, 612-13. Humphrey, H, J. 1950. The bacteriology of frozen foods. Quzck Frozen Foods 13(5), 50-51, 104; Canner 111(25), 10-11; Western Canner and Packer 43(3), 45-46. Hunter, A. C. 1939. Uses and limitations of the coliform group in sanitary control of food production. Food Research 4, 531-38. Hussemann, D. L. 1951. Effect of cooking on the bacteriologic flora of selected frozen precooked foods. J . Am. Dietet. Assoc. 27, 855-58. Hutchings, B. L., and Evers, C. F. 1946. Research and quality control of precooked frozen foods. Refrig. Eng. 61, 26-29, 61, 78, 82; Quick Frozen Foods 8(6), 71. Hutchings, B. L., and Evers, C. F. 1947. Problems in the production of precooked frozen foods. Food Technol. 1, 421-26. Ice, R. PII., LongrBe, K., Fenton, F., and Harris, K. W. 1952. Effect of holding on bacterial count and palatability of meat loaves befor? and after freezing and after reheating. J . Am. Dietet. Assoc. 28, 325-28. Iljin, W.S. 1933. u b e r den Kaltetod der Pflanzen und seine Ursachen. Protoplas?na 20, 105-24. Ingram, RII. 1951. The effect of cold on micro-organisms in relation to food. J . A p p l . Bacteriol. Proc. 14(2), 243-60. Ingram, M., and Brooks, J. 1952. Bacteriological standards for perishable foods. (d) Eggs and egg products. J . Roy. Sanit. Znst. 72, 411-19. Ireland, R.1941. Freezing and shipping of cold pack strawberries sugared in barrels. Ice and Refrig. 100, No. 3, 239-40. Irish, J. H., and Joslyn, hl. A. 1929. The viability of yeast. The effects of freezing storage in grape juice. Fruit Products J . 8(12), 11. Jakovliv, G. 1948. De la microbiologie des conserves de fruits e t des confitures. A n n . Gembloux 66, 9-13; Rev. Cons. (France) 1 (Oct.) 2 pp. James, L. H. 1932. The microbiology of frozen foods. Fruit Products J . 12, 110-14. James, L. H. 1933a. Frozen foods from the health standpoint. J . Bacteriol. 26, 74. James, L. H. 1933b. Effects of freezing on the spores and toxin of Clostridium botulznum. J . Infectious Diseases 62, 236-41. Jensen, L. B. 1943. Bacteriology of ice. Food Research 8, 265-72. !repsen, A. 1945. Biological problems in connection with chilling and freezing of meat. I, 11. Tek. Tidskr. 69(2), 31-34. Johns, C. K. 1948. The reducing sugar content of frozen egg as an index of the bacterial content. Can. J . Research F26, 18-23. Johns, C. K., and Berard, H. L. 1946. The effect of freezing and cold storage upon the bacterial content of egg melange. Sci. Agr. 26, 34-42. Johnson, V. T. 1942. Frozen and dried eggs. J . Am. Vet. M e d . Assoc. 100, 479-83. Jones, A. H., and Lochhead, A. G. 1939. A study of micrococci surviving in frozenpack vegetables and their enterotoxic properties. Food Research 4, 203-16. Jones, A. H., and Pierce, M. E. 1946. The significance of microorganisms in frozen fruits and vegetables. Paper presented a t Canadian Public Health Association, Dec. 17. Jordan, E. O., and Dack, G. 11. 1924. The effect of Clostridium sporogenes on Clostrzdium botulinum. J . Infectious Diseases 36, 576-80. Kahlenberg, 0. J. 1949. Recommended procedure for defrosting frozen eggs. Baker’s Dig. 23, 63-64. Kahlenberg, 0. J., Lally, hl. M., and Gorman, J. M. 1950. What is the right way to drill frozen eggs. 7i.S. Egq & Poultry l f a g . 66, 14-15.

222

GEORG BORGSTROM

Kahlenberg, 0. J., Lally, M. M., and Gorman, J. M. 1951. What’s the best way to sample frozen eggs. U.S. Egg & Poultry Mag. 67, 14-15, 25. Kaplan, M. T., and Appleman, M. D. 1952. Microbiology of frozen orange concentrate. 111. Studies of Enterococci in frozen concentrated orange juice. Food Technol. 6, 167-70. Kayukova, PIT. I., and Kremer, T. A. 1940. Development and toxin formation of Bacillus botulanum in mixed culture. Mikrobiologiya 9, 585-93. Keith, S. C., Jr. 1913. Factors influencing the survival of bacteria at temperatures in the vicinity of the freezing point of water. Science 37, 877-79. Kiser, J. S. 1943. A quantitative study of rate of destruction of a n Achromobacter sp. by freezing. Food Research 8, 323-26. Kiser, J. S., and Beckwith, T. D. 1942. Effect of fast-freezing upon bacterial flora of mackerel. Food Research 7, 255-59. Koelensmid, W. A. A. B. 1950. Bestrijding von het bederf van levensmiddelen. Voeding 11, 465-86. Kyes, P., and Potter, T. S. 1939. The resistance of avian tubercle bacilli to low temperatures with especial reference to multiple changes in temperature. J . Infectious Diseases 64, 123-134. Larkin, E. P., Litsky, W., and Fuller, J. E. 1955. Fecal Streptococci in frozen foods. I. A bacteriological survey of some commercially frozen foods. Appl. Microbiol. 3(2), 98-101. 11. Effect of freezing storage on Escherichia coli and some fecal streptococci inoculated into green beans. 102-04. 111. Effect of freezing storage on Escherichia coli, Streptococcus fuecalis, and Streptococcus liquefaciens inoculated into orange concentrate. 104-06. IV. Effect of sanitizing agents and blanching temperatures on Sireptococcus .fuecalis. 107-10. Lea, C. H., and Hawke, J. C. 1951. Relative rates of oxidation of lipovitellin and of its lipide constituent. Biochem. J . 60, 67-73. Lea, C. H., and Hawke, J. C. 1952. Lipovitellin. 2. The influence of water on the stability of lipovitellin and the effects of freezing and of drying. Biochem. J. 62, 105-14. Lenhart, J., and Cosens, K. W. 1949. Molds and bacteria curbed in immersion-freezing. Food Inds. 21, 442-44. Lepper, H. A., Bartram, M. T., and Hillig, F. 1944. Detection of decomposition in liquid, frozen and dried eggs. J . Assoc. Ofic. Agr. Chemists 27, 204-23. Levitt, J. 1941. “Frost Killing and Hardiness of Plants.” Burgess, Minneapolis. Lochhead, A. G., and Jones, A. H. 1936. Studies of numbers and types of microorganisms in frozen vegetables and fruits. Food Research 1, 29-39. Lochhead, A. G., and Jones, A. H. 1938. Types of bacteria surviving in frozen-pack vegetables. Food Research 3, 299-306. Lochhead, A. G., and Landerkin, G. B. 1935. Bacteriological studies of dressed poultry. I. Preliminary investigation of bacterial action a t chill temperatures. Sci. Agr. 16, 765-70. Logan, P. P., Harp, C. H., and Dove, W. F. 1951. Keeping quality of precooked frozen chicken A la King, a bacteriological evaluation of hot and cold packs. Food Technol. 6, 193-98. Lutz, J. M., Caldwell, J. S., and Moon, H. H. 1932. Frozen pack: studies on fruits frozen in small containers. Ice and Refrig. 83, 111-13. Luyet, B. J., and Gehenio, P. M. 1940. Life and Death a t Low Temperatures. Biodynamica, Normandy, Mo. McCleskey, C. S., and Christopher, W. N. 1941. Some factors influencing the survival of bacteria in cold-pack strawberries. Food Research 6, 327-33.

MICROBIOLOGICAL

PROBLEMS OF FROZEN FOOD PRODUCTS

223

McFarlane, V. H. 1938. Behavior of microorganisms a t sub-freezing temperatures. Ph.D. Thesis. University of Washington, Seattle. McFarlane, V. H. 1940a. Behavior of microorganisms a t subfreezing temperatures. I. Freezing redistribution studies. Food Research 6, 43-57. McFarlane, V. H. 1940b. 11. Distribution and survival of microorganisms in frozen cider, frozen syrup-packed raspberries and frozen brine-packed peas. Food Research 6, 59-68. McFarlane, V. H. 1941. 111. Influence of sucrose and hydrogen ion concentrations. Food Research 6, 481-92. McFarlane, V. H. 1942. Behavior of microorganisms in fruit juices, and in fruit juicesucrose solutions stored a t -17.8' C. (0" F.). Food Research ?, 509-18. McFarlane, V. H., and Goresline, H. E. 1943. Microbial destruction in buffered water and in buffered sugar syrups a t -17.8" C. (0" F.). Food Research 8, 67-77. McFarlane, V. H., Watson, A. J., and Goresline, H. E. 1945. Microbiological control in the production of spray dried whole egg powder. U.S. Egg & Poultry M a g . 61, 250-57, 270-86. Magoon, C. A. 1931. Spoilage in frozen pack fruits. Glass Container 9(5), 48-49. Magoon, C. A. 1932. Microorganisms as affecting frozen foods. I n d . Eng. Chem. 24, 669-7 1. Mallmann, W. L. 1945. Sanitation in frozen food locker plants. Quick Frozen Foods ? ( 6 ) , 80-81, 88. Marshall, R. E. 1945. Plant sanitation conference. Fruit Products J . 26, 83-84, 9192. Martin, R. B., and Dobson, J. G. 1945. Identification and characteristics of typical slime-forming organisms. Paper Trade J. 121, 143-50. Martinez, J. 1950. The significance of the coliform group in frozen orange juice. M.A. Thesis, University of Southern California. Martinez, J., and Appleman, M. D. 1949. Certain inaccuracies in the determination of coliforms in frozen orange juice. Food Technol. 3, 392-94. Mattick, A. T . R. 1952. The effect of cold on microorganisms: general introduction. Proc. Soc. A p p l . Bacteriol. 14, 211-15. Miller, C., and Winter, A. R. 1950. The functional properties and bacterial content of pasteurized and frozen whole eggs. Poultry Sci. 29, 88-97. Miller, C., and Winter, A. R. 1951. Pasteurized frozen whole egg and yolk for mayonnaise production. Food Research 16, 43-49. and Marsteller, R. L. 1952. The storage life of frozen orange concentrate Miller, E. IT., from the standpoint of the consumer. Food Technol. 6, 119-22. Moran, T. 1931. Rapid freezing of meat. Dept. Sci. 2nd. Research (Brit.) Food Invest. Board Kepts. pp. 14-21. Moran, T. 1935. Post-mortem and refrigeration changes in meat. J . Soc. Chem. 2nd. (London) 64, 149-51. Morris, T. N., and Barker, J. 1935. The freezing of vegetables. Dept. Sci. I n d . Research (Brit.) Food Invest. Board Repts. pp. 149-51. Mundt, J. 0. 1950. Microbiology of strawberries during harvest. Quick Frozen Foods 12(9), 53-56. Murdock, D. I., Folinazzo, J. F., and Troy, V. S. 1952. Evaluation of plating media for citrus concentrates. Food Technol. 6, 181-85. Murdock, D. I., Troy, V. S., and Folinazzo, J. F. 1952. Development of off-flavor in 20" Brix orange concentrate inoculated with certain strains of Lactobacilli and Leuconostoc. Food Technol. 6, 127-29.

224

GEORG BORGSTROM

Murdock, D. I., Troy, V. S., and Folinazzo, J. F. 1953. Thermal resistance of lactic acid bacteria and yeast in orange juice and concentrate. Food Research 18, 85-89. Nickerson, J. T. R. 1943. A modified little plate method for bacterial counts in vegetable freezing plants. Food Research 8, 163-68. Nielsen, F. A,, and Garnatz, G. F. 1940. Obscrvations on death rates for bacterial flora of frozen egg products. Proc. 1st. Food Con$ Inst. Food Technol., July 1940, pp. 289-94. Nolte, A. J., and von Loesecke, H. W. 1940. Types of organisms surviving in commercially pasteurized juices in Florida. Food Research 6, 73-81. North, W. R. 1945. Aniline oil methylene blue stain for the direct microscopic count of bacteria in dried milk and dried eggs. J . Assoc. Ofic. Agr. Chemists, 28(2), 425-27. Novick, J. 1946. Fruits and vegetables. Quick-frozen fruits and vegetables, Vol. 1, Part V. U.S. Quartermaster. Food and Container Institute for the Armed Forces (PB 40728). Obold, W. L., and Hutchings, B. L. 1947. Sanitation technic in the frozen food industry. Food Technol. 1, 561-64. Osler, A. G., Buchbinder, L., and Steffen, G. I. 1949. Experimcntal enterococcal food poisoning in man. Proc. SOC.Exptl. B i d . Med. 67, 456-59. Owen, R. F., and Van Duyne, F. 0. 1950. Comparison of the quality of freshly baked cakes, thawed frozen baked cakes, and cakes prepared from batters which had been frozen. Food Research 16, 169-78. Panassenko, W. T., and Tatarenko, J. S. 1940. Psychrotolerant fungi on foodstuffs (in Russian). Mzkrobiologiya 9, 579-84. Patrick, R. 1949. The role of microorganisms and storage temperatures on the quality of orange concentrates. Proc. Florida Slate Hort. SOC.6dnd Ann. Meeting, pp. 174-77; 1950 Citrus Ind. 31, 8-9, 13. Patrick, R. 1953. Coliform bacteria from frozen orange concentrate and damaged oranges. Food Technol. 7, 157-59. Pearce, J. A., and Reid, M. 1946. Liquid and frozen egg. 111. Some factors affecting the quality of stored frozen egg. Can. J . Research F24, 437-44. Pederson, C. S. 1947. Significance of bacteria in frozen vegetables. Food Research 12, 429-38. Pennington, 111. E. 1945. Refrigeration, the first essential in quality preservation. U.S. Egg & Poultry Mag. 61, 498-500. Pennington, M. E. 1948. The freezing of eggs. Rejrig. Eng. 66(5) Refrig. Eng. Application Data 20R. Pennington, M. E., Jenkins, M. K., St. John, E. Q., and Hicks, W. 13. 1914. A bacteriological and chemical study of commercial eggs in the producing districts of the central west. G.S. Dept. Agr. Bull. No. 51. Penniston, V. A., and Hedrick, L. R. 1944. Use of germicide in mash water to reduce bacterial contamination on shell. U.S. Egg & Poultry Mag. 60, 26-27, 47-48. Penniston, V. A., and Hedrick, L. R. 1947. The reduction of bacterial count in egg pulp by use of germicides in washing dirty eggs. Food Technol. 1, 240-44. Perry, H., Townsend, C. T., Andersen, A. A., and Berry, J. A. 1948. Studies on Clostridzum botulinum in frozen pack vegetables. Food Technol. 2, 180-90. Phillips, A. W., Jr., and Proctor, B. E. 1947. Growth and survival of a n experimentally inoculated Staphylococcus aureus in frozen precooked foods. J. Bacteriol. 64, 49. Plagge, H. H. 1938. Fruits and vegetables-studies relating to their freezing and storage. Ice and Refrig. 94, 220-23.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

225

Polge, C., Smith, A. U., and Parkes, A. S. 1949. Revival of spermatozoa after vitrification and dehydration a t low temperatures. Nature 164, 666. Prescott, S. C. 1932a. Bacteria as affected by temperature. Refrig. Eng. 23, 91-96. Prescott, S. C. 193213. Numbers of bacteria in frozen foods. Ice and Refrig. 82, 311-13. Prescott, S. C., and Geer, L. P. 1936. Observations on food poisoning organisms under refrigeration conditions. Refrig. Eng. 32, 21 1-12, 282-83. Prescott, S. C., and Tanner, F. W. 1938. Microbiology in relation to food preservation. Food Research 3, 189-97. l’rescott, S. C,, Bates, P. K., and Highlands, M. E. 1932. Yumbers of bacteria in frozen food stored a t several temperatures. Am. J . Public Health 22, 257-62. Proctor, B. E. 1941. Advances in refrigeration biology. Refrig. Eng. 41, 244-46. Proctor, B. E. 1948a. Frozen precooked foods. Am. J. Public Health 38, 44-49; Refrzg. J . (Australia) 1, 560-63. Proctor, B. E., 1948b. Research will improve the edible quality of frozen foods. Canner 106(13), 15-32. Proctor, B. E., and Greenlie, D. G. 1939. Redox potential indicators in quality control of foods. I. Correlation of resazurin reduction rates and bacteria plate counts as indices of the bacterial condition of fresh and frozen foods. Food Research 4, 441-46. Proctor, B. E., and Nickerson, J. T. R. 1948a. Improving frozen foods through microbiological control. Food Inds. 20, 682-83, 820-21. Proctor, B. E., and Nickerson, J. T. R. 1948b. Sanitation-key to frozen food quality. Refrag. Eng. 66, 456-59. Proctor, B. E., and Phillips, A. W. 1947. Microbiological aspects of frozen precooked foods. Refrig. Eng. 63, 30-33. Prudden, T. E. 1887. On bacteria in ice and their relation t o disease. Med. Record 31, 341-354, 369-378. Quinn, H., and Garnatz, C. G. 1943. A study of the increase of bacterial counts of frozen whole eggs, using three methods of defrosting. J . Bacteriol. 46, 49-50. Radabaugh, J. H. 1939. Economic aspects of the frozen egg industry in the United States. Proc. 7th World’s Poultry Congr., pp. 364-67. Rahn, 0. 1945. Physical methods of sterilization of microorganisms. Bacteriol. Revs. 9, 1-47. Rahn, O., and Bigwood, F. M. 1938. The effect of subminimal tcmperatures upon Streptococcus lactis. J. Bacteriol. 36, 64. Rahn, O.,and Bigwood, F. M. 1939. The effect of low temperature on Streptococcus lactis. Arch. Mikrobiol. 10, 1-5. Ramon, G., Richou, R., and Ramon, P. 1944-45. De l’anta gonisme microbien en g6nBral et en particulier des propri6ti.s antagonistes des filtrats de culture du B. subtilis. Rev. immunol. 9, 161-217. Ramstad, P. E., and Volz, F. E. 1950. Did it thaw? Gel will tell. Food Inds. 22, 2091, 2189-91. Reay, G. A. 1933. Influence of freezing temprratures on haddock muscle. J . SOC. Chem. I n d . (London) 62, 265-70. Redfield, H. W. 1920. Examination of frozen egg products and interpretation of results. U.S. Dept. Aqr. Bull. No. 846. Robinson, R. H. M., Ingram, M., Case, R. A. M., and Benstead, J. G. 1953. Whalemeat: Bacteriology and hygiene. Dept. Sci. I n d . Research (Brit.)Special Repts. No. 59, 55 pp. Rosser, F. T. 1942. Preservation of eggs. 11. Surface contamination on egg-shell in relation to spodage. Can. J . Research D20, 291-96.

226

GEORG BORGSTROM

Sair, L., and Cook, W. H. 1938. Effect of precooling and rate of freezing on the quality of dressed poultry. Can. J . Research D16, 139-52. Sanborn, R. E. 1946. What sanitation means t o the processor. Food Inds. 18, 26-28. Sanderson, N. H., Jr. 1941. The bacteriology and sanitation of quick frozen foods. Refrig. Eng. 42, 228-32; Canner 92(18), 11-12, 24, 28. Sanderson, K.H., Jr., and Fitzgerald, G. A. 1940. Eacteriological flora during spoilage of frozen vegetahles. Proc. Srd Inlern. Congr. Microbiol., New York, 1939, p. 710. Scarlett, W. J., and Martin, R. D. 1948. Value of chlorination in frozen food processing. Canner 106(13), 12. Schneider, M. D., and Gunderson, M. F. 1949. Investigators shed more light on Salmonella problems. U.S. Egg & Poultry Mag. 66, 10-11, 22. Schneiter, R. 1940. Microbiological examination of frozen egg products. J . Assoc. Ofic. Agr. Chemists 23, 613-17. Schneiter, R,, Bartram, M. T., and Lepper, H. A. 1943. Bacteriological and physical changes occurring in frozen eggs. J . Assoc. Offic. Agr. Chemists 26, 172-82. Schoening, H. W., Dale, C. N., Mott, L. O., and Habermann, R. T. 1949. Pathogenicity of Salmonella cholera-suis; maintaining virulence by lyophilization. Amer. J . Vet. Research 10, 101-110. Sealey, J. Q. 1939. Eacteria and other biological factors in relation t o frozen foods. Refrig. Eng. 37, 310-12. Sedgwick, W. T., and Winslow, C. E. A. 1902. Experiments on the effect of freezing and other low temperatures upon the viability of the bacillus of typhoid fever, with considerations regarding ice as a velucle of infectional disease. M e m . Am. Acad. Arts. Scz. 12, 471-577. Sharp, J. G. and Marsh, B. B. 1953. Whalemeat: Production and preservation. Dept. Sci. Ind. Research Repts. No. 58, 47 pp. Sherman, J. M., and Naylor, H. B. 1942. Aging without reproduction and the viability of young bacterial cells at low temperatures. J . Bacteriol. 43, 749. Sherman, J. M., Stark, C. N., and Stark, P. 1927. The destruction of botulinum toxin by intestinal bacteria. Proc. SOC.Exptl. Bid. Med. 24, 546-47. Sherman, J. M., Stark, C. N., and Stark, P. 1928. Destruction of botulinum toxin b y milk bacteria. J . Bacteriol. 16, 35-36. Sherman, V. W. 1946. Electronic heat in the food industries. Food Znds. 18, 506-09, 628-30. Shillinglaw, C. A., and Levine, M. 1943. Effect of acids and sugar on viability of Escherichia coli and Eberthella typhosa. Food Research 8, 464-76. Shrader, J. H., and Johnson, A. H. 1934. Freezing orange juice. Ind. Eng. Chem. 26, 869-77. Sisler, F. D., Ravenburg, R. F., and James, L. H. 1940. Microbiology of the intestines of chickens as related to frozen poultry. Proc. 3rd. Intern. Congr. Alicrobiol., N e w York, 1939, pp. 710-11. Smart, H. F. 1934. hiicroorganisms surviving the storage period of frozen-pack fruits and vegetables. Phytopathology 24, 1319-31. Smart, H. F. 1935. Growth and survival of microorganisms a t subfreezing temperatures. Science 82, 525. Smart, H. F. 1937a. Microbiological studies on cultivated blueberries in frozen pack. Food Research 2, 429-34. Smart, H. F. 1937b. Types and survival of some microorganisms in frozen packed peas, beans and sweet corn grown in the east. Food Research 2, 515-28.

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

227

Smart, H. F. 1939a. Further studies on the behavior of microorganisms in frozen cultivated blueberries. Food Research 4, 287-92. Smart, H. F. 193913. Microbiological studies on commercial packs of frozen fruits and vegetables. Food Research 4, 293-98. Smart, H. F., and Brunstetter, B. C. 1936. Lima beans in frozen pack. I. Planting tests. 11. Microbiological studies. Canner 83, (10) 14-16. Smart, H. F., and Brunstetter, B. C. 1937. Spinach and kale in frozen pack. I. Scalding tests. 11. Microbiological studies. Food Research 2, 151-63. Smith, A. U., and Polge, C. 1950. Survival of spermatozoa a t low temperatures. Nature 166, 668-69. Smith, E. F., and Swingle, D. B. 1905. The effect of freezing on bacteria. Science 21, 481-83. Stark, C. N., Sherman, J. M., and Stark, P. 1928. The influence of the filtrates of Clostridium botzclinum and Clostridium sporogenes upon the growth of each of these organisms. J . Bacferiol. 16, 18. Stark, C. N., Sherman, J. M., and Stark, P. 1929. Destruction of botulinum toxin by Bacillus subtilis.Proc. SOC.Exptl. Biol. Med. 26, 343-44. Stewart, J. J., and Edwards, A. L. 1948. “ Foods-Production, Marketing and Consumption,” 2nd ed. Prentice Hall, New York. Stewart, M. M. 1934. Effect of exposure to low temperatures on the numbers of bacteria in fish muscle. J . SOC.Chem. Ind.(London) 63, 273-78. Stiles, G. W., and Bates, C. A. 1912. A bacteriological study of shell, frozen, and desiccated eggs. U.S. Dept. Agr. B u r . Chem. Bull. 168. Stille, B. 1942. Uber die Haltbarkeit von Gefrierfleisch nach dem Auftauen. 2. ges. Kalteindustr. 49, 36. Stille, B. 1950. Untersuchungen uber den Kaltetod von Mikroorganismen. Arch. Microbiol. 14, 554-87. Straka, R. P., and Combes, F. M. 1951. The predominance of micrococci in the flora of experimental frozen turkey meat steaks. Food Resezrch 16, 492-93. Straka, R. P., and Combes, F. M. 1952. Survival and multiplication of Micrococcus pyogenes v. aureus in creamed chicken under various holding, storage and defrosting conditions. Food Research 17, 448-55. Straka, R. P., and James, L. H. 1932. A health aspect of frozen vegetables. Am. J . Public Health 22, 473-92. Straka, R. P., and James, L. H. 1933. Frozen vegetables. Am. J. Public Healfh 23, 700-03. Straka, R. P., and James, L. H. 1935. Further studies on frozen vegetables. J. Bacteriol. 29, 313-22. Stuart, L. S., and McNally, E. H. 1943. Bacteriologiral studies on the egg shell. U.S. Egg & Poultry M a g . 49, 28-31, 45-47. Sulzbacher, W. L. 1950. Survival of microorganisms in frozen meat. Food Technol. 4, 386-90. Sulzbacher, W. L. 1952. Effect of freezing and thawing on the growth rate of bacteria in ground meat. Food Technol. 6, 341-43. Swenson, T. L., and James, L. H. 1935. A comparison between eggs frozen a t 0” F. and eggs frozen at -109” F. U.S. E g g & Poultry Mag. 41(3), 16-19. Swift, H. F. 1937. The preservation of bacterial cultures by freezing and drying. Proc. 2nd Intern. Congr. Microbiol. London, 1936, pp. 40-41. Tanner, F. W. 1932. “The Microbiology of Foods.” 1st ed., Twin City Printing Co., Champaign, Ill. ; 1944. 2nd ed., Garrard Press, Champaign, Ill.

228

GEORG BORGSTROM

Tanner, F. W. 1934. Microbiological examination of fresh and frozen fruits and vegetables. Am. J . Public Health 24, 485-92. Tanner, F. W., and Oglesby, E. W. 1936. Influence of temperature on growth and toxin production by Clostridium botulinum. Food Research 1, 481-94. Tanner, F. W., Beamer, P. R., and Rickher, C. J. 1940. Further studies on development of Clostridium botulinum in refrigerated foods. Food Research 6, 323-33. Tanner, F. W., and Wallace, G. I. 1931. Effect of freezing on microorganisms in various menstra. Proc. SOC.Exptl. Biol. Med. 29, 32-34. Tanner, F. W., and Williamson, B. W. 1928. The effect of freezing on yeasts. Proc. Soe. Exptl. Biol. Med. 26, 377-81. Teunisson, D. J., and Hall, H. H. 1947. Study of bacteria from citrus processing operations in relation to orange juice quality. Fruit Products J . 26, 199-203. Tressler, D. K. 1938. Bacteria, enzymes and vitamins-indices of quality in frozen vegetables. Refrig. Eng. 36, 319-21. Tressler, D. K. 1945. Quality control in the frozen foods industry. Proc. Inst. Food Technol., pp. 146-53. Tressler, D. K. 1946. “Recent Advances in Frozen Food Technology,” pp. 3-5. Frozen Food Foundation, Syracuse, New York. Tressler, D. K., and Evers, C. F. 1947. “The Freezing Preservation of Foods,” 2nd ed., 932 pp. Avi Publishing, New York. Trout, G. M., and Scheid, M. V. 1943. The influence of several factors upon the flavor of frozen sweet cream. J . Dairy Sci. 26, 609-18. Tschistjakow, F. M., and Botscharowa, Z. Z. 1938. The effects of low temperatures on the development of moulds (in Russian). Mikrobiologiya 7, 498-524. Tschistjakow, F. M., and Moskowa, G. L. 1938. Die Wirkung niedriger Temperatur auf die Mikroorganismenentwicklung. 111. Die Wirkung niedriger Temperatur auf die Entwicklung von Bakterien und Hefen (in Russian). Mikrobiologiya 7, 565-78; Zentr. Bakteriol. Parasitenk. Abt 11, 101, 471. Turner, T. B., and Brayton, N. L. 1939. Factors influencing the survival of spirochetes in the frozen state. J . Exptl. Med. 70, 639-50. Ulrich, J. A., and Halvorson, H. 0. 1947. Changes in foods during storage in frozen conditions. Hormel Inst. Univ. M i n n . Ann. Rept. 48, 48-52. van den Broek, C. J. H. 1949. Het bederf van dierlijke voedingsmiddelenlage temperaturen. I. Chem. Weekblad. 46, 777-84. Van Eseltine, W. P., Nellis, L. F., Lee, F. A,, and Hucker, G. J. 1948. Effect of rate of freezing on bacterial content of frozen vegetables. Food Research 13, 271-80. Van Oijen, C. F. 1940. “Gepasteurigeerde” bevroren Eieren. Een helanggrijke hygienische Verbetering. Tijdschr. Diergeneesk. 67, 686. Vass, A. F. 1919. The influence of low temperatures on soil bacteria. Cornell Agr. Expt. Sta. Mem. 27, 1043-74. Vaughn, R. H., and Stadtman, T. C. 1945. A method for control of sanitation in food processing plants. Am. J . Public Health 36, 1292-96. Vaughn, R. H., and Stadtman, T. C. 1946. Sanitation in the processing plant and its relation to the microbial quality of the finished products. Food Freezing 1, 334-36, 364. Vaughn, R. H., Stadtman, T. C., and Keunemann, R. W. 1946. Control of microorganisms in frozen food plants. Quick Frozen Foods 9, 76-78, 114. von Schelhorn, M. 1951. Control of microorganisms causing spoilage in fruit and vegetable products. Advances Food Research 3, 429-32. Wallace, G. I., and Park, S. E. 1933. Microbiology of frozen foods. IV. Longevity of

MICROBIOLOGICAL PROBLEMS OF FROZEN FOOD PRODUCTS

229

certain pathogenic bacteria in frozen cherries and frozen cherry juice. V. The behavior of Clostridium botulinzina in frozen fruits and vegetables. J . Znfectious Diseases 62, 146-49, 150-56. Wallace, G. I., and Tanner, F. W. 1933. Microbiology of frozen foods. I. Historical review and summary of results. Fruit Products J . 13(2), 52-54; (a), 109-113. Wallace, G. I., and Tanner, F. W. 1934. 11. Studies on frozen fruits and vegetables. Fruit Products J . 13(9), 274-77; (12), 366-69, 377. Wallace, G. I., and Tanner, F. W. 1935. 111. Longevity of pure cultures of microorganisms frozen in various menstra. Fruzt Products J . 14(8),235-37. Wallace, M. D., and Baumgartner, J. G. 1936. The destruction of microorganisms in the presence of sugars. 11. The influence of sugars in chemical disinfection. J . Soc. Chern. Znd. (London) 66, 37-40. Weinzirl, J., and Newton, E. B. 1915. The fate of bacteria in frozen meat, held in cold storage. Am. J . Public Health 6, 833-35. Weiser, H. H. 1946. Some fundamental observations on the activities of microorganisms in frozen foods. Frozen Food Znd. 2(8), 8-9, 36. Weiser, H. H. 1951. Survival of certain microorganisms in selected frozen foods. Quick Frozen Foods 13(7), 50-52. Weiser, R. S., and Hargiss, C. 0. 1946. Studies on the death of bacteria a t low temperatures. 11. The comparative effects of crystallization, vitromelting, and devitrification on the mortality of Escherichia coli. J . Bacteriol. 62, 71-79. Weiser, R. S., and Osterud, C. M. 1945. Studies on the death of bacteria at low temperatures. I. The influence of intensity of the freezing temperature, repeated fluctuations of temperature, and the period of exposure to freezing temperatures on the mortality of Escherichia coli. J . Bacteriol. 60, 413-39. Western Regional Laboratory, Albany, Calif. 1944. Information sheet on factors that affect quality in freezing preservation of peas. U.S. Depl. Agr. Bur. Agr. Znd. Chem. White, L. S. 1951. Microorganisms and frosted foods; applications of bacterial findings controlling food quality. Food in Canada 11(2), 22-26. Wilkin, RI., and Winter, A. R. 1947. Pasteurization of egg yolk and white. Poultry S C ~26, . 136-42. Wilster, G. H. 1946. Bacteria. Creamery J . 67(10), 5. Wilson, G. S. 1935. The Bacteriological Grading of Milk. Great Britain. H. M. Stationery Office, London. Winter, A. R. 1952. Production of pasteurized frozen egg products. Food Technol. 6, 4 14- 15. Winter, A. R., and Wilkin, M. 1947. Holding, freezing, and storage of liquid egg products to control bacteria. Food Freezing 2, 338-41. Winter, A. R., and Wrinkle, C. 1949a. Frozen egg quality depends on clean eggs, clean equipment and quick cooling. U.S. Egg & Poultry Mag. 66(1), 7-9, 18-19. Winter, A. R., and Wrinkle, C. 194913. Fast freezing a t low temperatures protects frozen egg quality. U.S. Egg & Poultry Mag. 66(2), 20-23, 30-32. Winter, A. R., and Wrinkle, C. 1949c. Proper defrosting methods keep bacterial counts low in frozen egg products. U.S. Egg & Poultry Mag. 66(3), 28-31. Winter, A. R., Burkhart, B., and Wrinkle, C. 1951. Analysis of frozen egg products. Poultry Sci. 30, 372-80. Winter, A. R., Greco, P. A., and Stewart, G. F. 1946. Pasteurization of liquid-egg products. I. Bacteria reduction in liquid whole egg and improvement in keeping quality. Food Research 11, 229-45.

230

GEORG BORGSTROM

Winter, A. R., Stewart, G. F., McFarlane, V. H., and Solowey, M. 1946. Pasteurization of liquid-egg products. 111. Destruction of Salmonella in liquid whole egg. Am. J. Public Health 36, 451-60. Winter, A. R., Stewart, G. F., and Wilkin, M. 1948. Pasteurization of liquid-egg products. IV. Destruction of coliforms. Food Research 13, 11-18. Wolford, E.R. 1943. A direct microscopic method to estimate the sanitary history of frozen pack peas. Wesiern Canner and Packer 36(13), 58; Food Research 8, 163. Wolford, E. R. 1950. Bacteriological studies on frozen orange juice, stored at -10” F. Food Technol. 4, 241-45. Wolford, E. R., and Andersen, A. A. 1945. Propionates control microbial growth in fruits, vegetables. Food Inds. 17, 622-24, 726-32. Wolford, E. R., and Berry, J. A. 1948a. Condition of oranges as affecting bacterial content of frozen juice with emphasis on coliform organisms. Food Research 13, 172-78. Wolford, E.R.,and Berry, J. A. 1948b. Bacteriology of slime in a citrus processing plant with special reference to coliforms. Food Research 13, 340-46. Woodroof, J. G. 1931. Preservation freezing, some effects on quality of fruits and vegetables. Ga. Agr. Expt. Sta. Bull. 168. Wright, A. M. 1923. Molds on frozen meats. J . SOC.Chem. Znd. (London) 42, 488-90. Wrinkle, C., Weiser, H. H., and Winter, A. R. 1950. Bacterial flora of frozen egg products. Food Research 16, 91-98. Yesair, J., and Williams, 0. B. 1942. Spice contamination and its control. Food Research 7, 118-26. Yesair, J., and Williams, 0. B. 1945. Food M a n u f . 20(1), 13-16. Young, S. 1947. Much to venture, much to win in food technology. Food in Canada. 7(10), 23-26. Yurchenko, J. A., Piepoli, C. R., and Yurchenco, M. C. 1954. Low temperature storage for maintaining stable infectious bacterial pools. A p p l . Microbiol. 2, 53-5. Zagaesky, J. S., and Lutikova, P. 0. 1944. Sanitary measures in the egg breaking plant. U.S.Egg & Poultry Mag. 60, 17-20, 43-46, 75-77, 88-89, 121-23. Zobell, C. E. 1946. (‘Marine Microbiology, a Monograph on Hydrobacteriology.” Chronica Botanica, Waltham.

Potato Granules, Development and Technology of Manufacture

BY R. L. OLSON

AND

W. 0. HARRINGTON

Western Utilization Research Branch, Agricultural Research Service, U . S . Department of Agriculture, Albany, California

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Related Nongranular Products. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Precooked Potato Shreds (Riced Potatoes). . . . . . . . . . . . . . . . . . . . . . . 2. Quick-Cooking Potato Shreds. . . . . . . ................... 111. Spray Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Direct Dehydration-Two-Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. “Freeze and Squeeze” Method ................................ VI. Cold Tempering.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V I I . Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V I I I . Add-Back Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX. General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Raw Material., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Predrying Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Drying Methods and Equipment.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Product Quality. . . . ....................................... X. Quality Evaluation of Potato Granules.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Subjective Appraisal.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Objective Measurements.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page 231 233 233 233 234 235 237 237 238 238 213 2-13 213 245 247 2-19 249 251 2.53

I . INTRODUCTION With regard t o nut,ritive value and extent of cultivation, the white potato is the most important single vegetable. As an energy source i t ranks second only t o cereals among the products of the vegetable kingdom used for human food. I t s relatively easy cultivation, high yield, pleasant flavor, and excellent nutritive qualities have made it popular almost wherever it has been introduced (Boswell and Bostelman, 1949). Although potatoes store reasonably well a t cool temperatures and can he held for several months before consumption, decay and other storage losses, including losses of nutritional values, are significant even under the best conditions. Therefore, preservation processes have been developed t o conserve the substance and value of potatoes over long periods of time. 23 1

232

R. L. OLSON AND W. 0. HARRINGTON

Preservation of potatoes by drying had its origiii in South American prehistory, antedating by centuries the modern techniques of canning, freezing, and starch manufacture. Archeological investigations date potato drying back 2000 years or more (Salaman, 1949, 1953). The earliest chroniclers of the Spanish Conquest of South America record the existence of a well-established commerce in dried potato products. The preserved potatoes could be stored for use in times of famine and provided a stable, concentrated food for military supply under the Inca regime. The value of dried potatoes was illustrated in the looting of communal and imperial stores for sale a t the Potosi silver mines by the Conquistadores, a n action that established fortunes comparable t o some of those provided directly by the mined metal (Boswell and Bostelman, 1949; Salaman, 19-19, 1953; Tjomsland, 1950). Dehydrated potatoes have had a place in the military stores of more recent times and have also filled the needs for a food staple in times and places where fresh tubers could not be conveniently obtained because of high shipping costs and storage losses. The development of new and improved processes for dehydrated potato products has been, t o a large extent, stimulated by militaxy requirements in recent wars. However, t o a n increasing degree, dehydrated potatoes, and dehydrated and concentrated foods in general, have been entering domestic markets because of convenience of preparation, reduced costs, and the consistent high quality of the product. I n this communication the authors will trace the historical development of potato granules-a dehydrated product that appears to be headed for an important role in potato utilization. Potato granules are precooked, dehydrated potatoes produced in particulate form consirting substantially of whole tissue cells or small aggregates of cells. The product is characterized by high bulk density (about 0.9 that of water) and great convenience of preparation (requiring only a simple mixing with hot liquid’,. The general characteristics of potato granules have been described by Barker and Burton (1944), Rendle (19454, Morris (1947), King (1948), Olson and Harrington (1951), Cooley et al. (1954), and others, as will be brought out later in this review. On reviewing the feeding problems of the British Army during World War 11, King (1948) described potato granules (also called mashed potato powder) as the most generally acceptable of all dehydrated vegetables, having application both t o military and civil needs. Although commercial production was not started until 1942, the British Ministry of Food procured over 6000 tons of potato granules during t h a t emergency (Rendle, 194513).

233

POTATO GRANULE MANUFACTURE

Table I contains data from Rendle (1945b) and Morris (1947) t o illustrate characteristic differences among various methods of preserving potatoes. It will be seen that, in space and weight conservation, dehydrated potatoes have a distinct advantage and that granules rank higher in these respects than strips. A balancing factor for canned potatoes is the convenience of preparation which, as Rendle (1945b) pointed out, is the easiest of the 3 preserved products t o prepare for serving, with granules a very close second. TABLEI Comparative Weight, Space Requirement, and Food Value of Potatoes Preserved by Different Methods

Type of potatoes

Type of pack

Volume (including packing case) per ton edible potato

Gross weight

~

Calories per pound of edible food plus package and waste ~~~~

Fresh potatoes Sack No. 235 can Canned potatoes Potato granules 14 lb. can Dehydrated potato 9 lb. can strips

48 cu. f t . 94 cu. f t . 1 2 . 5 cu. f t .

1 ton 1 . 8 6 tons 400 lbs.

280 190 1320

40.3 cu. f t .

400 Ibs.

1060

In the course of reviewing the literature and describing experimental arid commercial developments in the preparation of potato granules some unpublished findings of the authors and their associates will be included. These investigations have been carried out in the Western Utilization Research Branch of the Agricultural Research Service of U. S. Department of Agriculture and include participation of G. H. Neel, M. W. Cole, W. R. Mullins, A. L. Potter, Jr., G. S. Smith, J. Miers, E. Wood, and M . D. Nutting, as well as the authors. 11. RELATEDNONGRANULAR PRODUCTS 1. Precooked Potato Shreds (Riced Potatoes)

Before the development of potato granules, other processes were developed t o improve product quality and convenience of preparation of dehydrated potatoes and t o reduce bulk density of the product. Over a century ago, Edwards (1845) was granted a patent on a process for preserving precooked potatoes as dried particles. Through the years minor changes were proposed for the process (Cooke, 1912; Remmers, 1918; Allen, 1921; Potato Corporation of Idaho, 1939; Gano, 1940; Stamberg

234

R. L. OLSON AND W. 0. HARRINGTON

and Beresford, 1943; and Nixon, 1944), but the essentials remained and have been used for commercial production. Potato shreds are produced by drying riced, or extruded, cooked potatoes on trays in tunnel dehydrators (Anonymous, 1943a,b, 1948; Morris, 1947). By this method, large quantities of potato shreds were produced for military supply during World War 11. According t o United States Military Specification (MIIJ-P-1073A, 12 December 19.50) preparation of mashed potatoes from potato shreds requires addition t o boiling water, a 10-min. cook over low heat, and whipping. Cooke (1912), Gano (1940), and Nixon (1944) understood the importance of maintaining the tissue cell structure intact t o prevent the development of “free” soluble starch in the product. Starch that is retained by unbroken cell walls will rehydrate satisfactorily with hot water. An unpalatable stickiness will result if undue amounts of starch are released from ruptured cells. Engineering studies of the drying characteristics of potato shreds were made by Brown and Kilpatrick (1943) and Van Arsdel et al. (1947) as an aid in the design and operation of dehydrators in the production of potato shreds. 2. Quick-Cooking Potato Shreds

A further development in dehydrated precooked potatoes was the process for making a dehydrated potato shred that would not require the 10 min. of cooking that ordinary shreds require in order t o reconstitute. Stoddard (1922) described a method t h a t followed usual operations but included high-temperature drying, which formed a porous product of filaments with a tubular bore. The dried product could be prepared for serving with little more effort than mixing with hot (boiling) liquid. Kaufman et al. (1949) used a similar method, modified by the addition of liquid (such as reconstituted dry milk solids) t o cooked potatoes, followed by thorough whipping before extrusion and drying t o aid in the production of a very porous product. Market testing of the product has been conducted extensively in metropolitan areas. The investigators have recognized the relationship of cell damage t o the texture or” reconstituted product. Their raw material has been selected and operations performed in a manner intended t o minimize such damage. 111. SPRAYDRYING The earliest attempts t o produce dehydrated precooked potatoes in the form of intact but separate tissue cells involved spray drying. Heimerdinger (1926) outlined such a method, which was similar t o those described later by Bowen (1931), Barker cf al. (1943), Burton (1944a,b), and Morris (1947).

POTATO GRANULE MANUFACTURE

235

A commercial venture was described by Bowen (1931) in which potatoes mere washed, cooked in steam, extruded, and made into a slurry with water. The slurry was sprayed through 5 centrifugal nozzles into a spray chamber 40 ft. in diameter and 60 ft. high. Swept from the chamber floor by an air stream, the product was collected in cyclone-type separators and screw-conveyed t o sizing screens and then t o packaging equipment. The venture did not continue, although sales of product for restaurant and institutional use were reported as late as 1937 (Anonymous, 1937). Barker et al. (1943) and Burton (1944a,h) discussed the importance of designing and operating equipment for spray drying potatoes so as t o avoid undue cell rupture. They reported production in trial runs with commercial equipment of potato granules of good textural quality after reconstitution. Gravity feed t o a centrifugal atomizer was used and damage t o the product was reported when helical rotor or centrifugal pumps were used. Rivoche (1951b) suggested that freezing and thawing cooked potatoes before making the slurry would improve quality of spray-dried potato granules by rendering the tissue cells less susceptible t o rupture. He also described a method in which frozen cooked potatoes would be broken by a hammer mill into a snow mist that could be spray dried (Rivocht.. 1951a). Limited research at Massachusetts Institute of Technology (Proctor and Sluder, 1944; Campbell et al., 1945) on certain types of spray-drying equipment did not lead t o the production of a satisfactory product. The limitations of standard-make spray driers, particularly with respect t o the available spray nozzles, were recognized as the principal source of cell damage and consequent lorn product, quality.

IV. DIRECTDEHYDRITIOZ,-TWO-ST.~GE An important advance in the development of a potato granule process was the observation that, if the moisture content of cooked potatoes i5 reduced t o about 40%) a moist friable powder can he produced by the action of crushing or grating (Bunimovitch and Faitelowitz, 1936). The patent of Bunimovitch and Faitelowitz does not clearly outline a method t o be used, but it does reveal the partial drying of cooked potatoes. followed by crushing or grating t o a moist powder which can be dried with continual agitation t o produce granules without substantial ccll rupture. Difficulties are encountered in trying t o follow such general directions. Pilot-plant studies on direct drying in two stages i n thib laboratory did not result in high-quality potato granules. Sonuniformity of drying on trays and in a rotary drier, even a t lorn drying potenlialq,

236

R. L. OLSON AND W. 0. HARRINGTON

led t o a high percentage of coarse material t h a t subsequently would not break down into a granular condition. T o overcome the problem of inadequate granulation Jones and Greer (1940, 1943) used a hammer mill t o reduce the partiallv dried material t o a granular condition. A further modification was initial dehydration on a roller drier. It was expected that resiliency of the potato cells would be adequate t o avoid undue rupture of cell walls throughout this process. Final drying was accomplished in a pneumatic system in which hot dry air passed through the hammer mill and carried the product through a duct as it was being dried. The product was separated from the air stream by a cyclone-type separator. In pilot-plant studies, in the laboratory represented by the present writers, partial drying on roller driers a t temperature of 212" F. (100" C.) or above was found t o be harmful t o product quality (texture of reconstituted product was pasty and a relatively high percentage of course material was produced) even when only about 10% of the moisture was removed from cooked potatoes. It appeared t h a t the action of the roller, per se, did not degrade quality of product. It was possible t o mash cooked potatoes through the roller a t about 140" F. (60' C.) or lower temperature without substantial effect on product quality. Barker (1941) and Barker et al. (1943) describe conditions of partial drying on a roller drier whereby moisture content of cooked potatoes was reduced t o 50% without excessive product damage. They found that underdrying or overdrying would result in damage. I n trials on commercial scale equipment they found that damage occurred due t o the action of the small diameter feed roll in the type of roller drier normally used in the manufacture of potato flour. With a double-drum drier, however, they reported successful operations. I n these studies Barker et al. used a low-temperature conditioning period (see section VI) t o equilibrate moisture content and reduce the tendency toward stickiness in the partially dried product. This step was found essential in the production of a satisfactory product. Other processes for manufacture of potato granules, utilizing direct dehydration of cooked potatoes as a first step, were described by Volpertas (1937, 1944) and Rivoche (1948, 1950). The former indicated in one case a crushing, stirring, and scraping action t o accompany the drying t o a fine, damp powder of homogeneous consistency, and in the other case prescribed a partial drying of 1-inch cubes of cooked potatoes t o below 50 % weight reduction without agitation, optionally under reduced pressure in a retort, followed by cooling and very gentle stirring. The cool cubes mere subjected t o mechanical pressure t o produce a light, moist powder for final drying. Rivoche's methods involved evaporative cooling

POTATO GRANULE MANUFACTURE

237

or freezing of cooked potatoes, partial drying with mild heat (with vacuum or by other means), and thorough mixing of the cooled product t o produce a fine, damp powder that could then be dried by application of heat. V. “FREEZE AND SQUEEZE” METHOD Another development in potato-granule manufacture involves a freezing operation. This has been described in several patent specifications (Willets and Rendle, 1946, 1948; Poulton and Noel, Ltd., and Bostock, 1946; Rivoche, 1948, 1950, 1951b; and Greene et al., 1949) and publications (Greene et al., 1947, 1948; Werts, 1947; and Hall, 1933). Further details are t o be found in unpublished research manuscripts (Olson, 1947; Erskine, 1950; Hall, 1951). Essentially, this method consists of removal of liquid by application of pressure following cooking, freezing, and thawing of cooked potatoes. The press-cake then consists of cooked potato tissue with a moisture content a t about 40%. This can be granulated and dried like the moist friable powder in the methods involving direct dehydration. A portion of the soluble solids is removed from the potatoes in the process. Although such loss detracts somewhat from the riatural flavor of the product and reduces yield, there is a substantial improvement in the storage stability of the product. By removal of soluble browning reactants, potato granules made by the “freeze and squeeze” method have been stored for several years a t room temperature without color degradation. Much of the early developmental work on this method for potato granule manufacture and related engineering problems was supported by contract research of the U. S. Army Quartermaster Corps a t Kansas State College. A limited commercial operation existed in Pennsylvania before 1950 but was discontinued. Commercial application of the process has also been reported in Europe (Erskine, 1950). 1’1. COLDTEMPERING Other proposed methods of producing potato granules do not necessitate the two-stage drying. I n one process potatoes are cooked, mashed, and held a t a low temperature until they lose their tendency t o be sticky. The mash is then extruded and dried under conditions whereby agglomeration of cells is prevented (Barker, 1941; Proctor and Sluder, 1943, 194-1: Campbell et al., 1945). In his patent on this method, Barker (1941) points out that up t o 20 hr. of low-temperature conditioning are required. Extrusion is achieved by rubbing the tempered material through a screen. The low-temperature holding reduces stickiness, and it was postulated that retrogradation of starch is responsible for or connected with the physical changes in the

238

R.

L.

OLSON .4ND W. 0. HARRINGTON

cookcd and mashed potatoes. The independently developed method reported by Proctor and Sluder (1943, 1944) and Campbell et al. (1945) is similar, although differences in equipment design are noted. In these investigations the importance of raw-material quality and of temperature to the minimum tempering time required was demonstrated. A minimum tempering time of 15 hr. was reported but with poorer-quality raw material, tempering times u p t o 96 hr. mere necessary in some instances t o produce an acceptable product. The principal application of the tempering procedure in potato granule production is its use in connection with other processes in which the moisture content is reduced in freshly cooked potatoes and the tempering operation is introduced t o aid in the granulation of the product before the final drying operations.

VII. SOLVENT EXTRilCTION A novel method for the production of potato granules has been suggested by Heisler et al. (1953) in which the water is extracted from potatoes with an organic liquid. Freshly cooked mashed potatoes are suspended in a water-miscible solvent, The slurry is filtered and the filter cake dried (removing residual moisture and solvent) by application of mild heat. Most efficient use of solvent is obtained by multiple extractions in which the first suspension is made with filtrate from a second and third extraction. Fresh solvent is then used for the second and third extractions. Soluble solids are lost in the filtrate creating a situation similar t o such loss as mentioned in the section on the “freeze and squeeze” process. Although yield and flavor losses occur in the process, the storage stability of the product is enhanced by removal of browning reactants. The selectioii of solvent t o be used in this process involves a number of factors. Of principal importance are (a) the efficiency with which water is extracted relative t o the amount of solvent required and soluble solids removed, (b) the distillation characteristics of the solvent for recovery from filtrate and filter cake, (c) the ease of removal of residual solvent from product, and (d) the effect of the solvent on the flavor of product. Ethanol is considered a relatively good solvent for this process. Continuous liquid extraction and vapor-phase extraction procedures have been explored also in connection with solvent extraction investigations (Heisler et al., 1953). VIII. ADD-BACKMETHOD At present the only significantly successful commercial process for potato-granule manufacture in the United St,atesis the add-back method.

POTATO GRANULE M A S U F A C T U R K

239

In this method the desired partial dehydration (as explained in previous sections) is achieved by recycling a sufficient quantity of dried product with freshly cooked mashed potatoes t o prepare a friable mixture that can be dried t o a powder. I n an early development of this process Volpertas (1939, 1944) specified a recycling of dry product in connection with his process involving direct dehydration in two stages. T o improve yield and uniformity of product he suggested t h a t the coarser product could be separated by screening and softened by mixing with freshly mashed potatoes in the continued operation of the process. Rivoche (1948, 1950) described an add-back process, also as a n alternative operation, in the specification of a direct dehydration process for manufacture of potato granules. I n this he described the addition of a sufficient amount of previously dried material t o reduce the moisturc content of mashed potatoes t o below 50% as a suitable method in the preparation of potato granules. I n patent specifications, Rendle (1943, 1915a) gave an adequate description of this add-back method of potato granule manufacture. The process can be described briefly as one in which an approximately equal weight of previously dried product is added t o freshly mashed potatoci t o produce a friable moist powder of about 40% moisture content. Thc> mixture is sieved through a 12-to-18-mesh screen and dried with hot air t o 6-8 % moisture content. Bar (1943) specified certain modificatioiii in the process, notably the introduction of a cooling and holding step t o increase the friability of the moist powder. Morris (1947) indicated that both the Rendle and the Bar process have been used in commercial production of potato granules in Great Britain. The former process na. brought t o the United States following World War I1 and potato granule< appeared in domestic markets in America about 1947. Stimulated by military requirements following the outbreak of h o t tilities in Korea in 1950, three additional processors began potato granult. operations. All used the add-back method but with considerable variation in details of operations. Because the add-back method has been dernoirstrated t o have comniercial feasibility, a more detailed description of thli process will be given, outlining several factors of importance. Inherent in a process that involves recycling of a finished product 1. the difficulty of identifying and measuring the influence of processing variables in the process. I n the add-back method about 8553, (on a solids basis) of the friable moist mixture has already passed through the system. The appraisal of a sample of product collected in an investigation of a variable in the process may not demonstrate an important effect becaiw, as the process approaches a steady state, only about 15% of the material will bear the initial impact of the causative action.

240

~

R. L. OLSON AND W. 0. HARRINGTON

The effect on quality of a particular element in the process can sometimes be determined by comparable runs, except for the elimination in one case of the subject element. T o be meaningful the product should be appraised as the processes approach a steady state. As an example of this approach an investigation conducted a t a commercial plant will be described. For a period of operation, potato granules were produced in parallel processing lines except t h a t in one case a mechanical conveying system was by-passed. Ten cycles of the two operations were completed. The product a t the end of each cycle was appraised for texture after reconstitution as mashed potatoes. Blue Values (see section X, 3 below on objective evaluation of consistency by chemical determination of easily extracted starch) were also determined. A damaging action of the conveying equipment was demonstrated (Table 11). TABLEI1 Evaluation of Effect of Conveying System on Quality of Potato Granulesa Normal run Conveyor usedScries 1

Cycle 1 2 3 4 5 6 7 8 9 10 11 0

b

Texturec 4 5 3 5 5 6

5 5

Blue value Index

104 102 109

120 112 130 202 232

Series 3-granules passed through conveyorb

Conveyor-by-passedSeries 2

Texturec 3 5 3 5 6 7 6 7 6 7

7

Blue value Index 109 108

'

94

86 82 67 78 71 67 76 70

Texturec

5 3 6 4 6

4 4 1 2 2 2

Blue value Index 88 108 140 132 143 162 144 178 192 206 218

From Mullins et al. (1955). Feed is cycle 11, series 2. Subjective texture score. 7 = good, not pasty or rubbery, 1 = slightly pasty and rubber>

By means of another device, a further verification of this conclusion was obtained. The product from the run in which the conveyor was bypassed was, after 11 cycles, passed through the conveyer system 11 times. At the end of each passage, a sample was taken for evaluation. Results indicated damaging action (Table 11). The authors are indebted to R. W. Kueneman and J. Conrad of the J. R. Simplot Co., Caldwell, Idaho, for the results of this investigation.

POTATO GRANULE MANUFACTURE

24 1

Often misleading results are obtained in such investigations, because the elimination or modification of a processing step may in itself impose a qualitative change in the product. Thus, in the determination of the effectof a mixer on the quality of product, comparable runs may be made with and without the mixer in the processing line. I n this case, the elimination of the mixing action would adversely affect the degree of granulation (relative coarseness of the friable mixture) which in turn affects the drying rate, the particle-size classification of the dried product, the package density, and the rehydration rate of the product. These and probably other actions and reactions are related. Thus, no comparable material is produced as a standard for judging the mixer in question. Uncontrolled factors may be introduced by any change in the system, and such factors may complicate studies of the process or adjustment of processes toward improvement of product quality. Experimental progress must depend in large part on relative values obtained in experimental runs in which the introduction of uncontrolled elements is recognized and conclusions are modified accordingly. I n general, trends in quality attributes that persist over wide ranges of operating conditions are better criteria for establishment of process improvements than absolute appraisal values of the experimental product. I n the wide raw-material variability encountered iii potatoes and in the difficulties inherent in a process with a large percentage of finished product recycling through the system may be found the explanation of the problems expressed by Rendle (1945b). (‘ . . . from start t o finish, the manufacture calls for such delicate adjustment that while the full process may be described in a few words, the know-how has proved even more difficult in the factory production of this product than is normally thc case in the translation of laboratory into works practice.” Despite the conditions imposed by the nature of the add-back process, forn a d steps have been taken. In the short history of commercial potato granule production, a trend in product quality improvement has been evidcnt. The difficulties inherent in the add-back method are being overcome. Of great advantage is the fact that the method lends itself readily t o simple mechanical processes involving mixing, conveying, and screening operations, all of which can be carried out in continuous mechanical equipment. Some data relating processing variables t o product quality in the addback method are available. Olson et al. (1953) reported that over a wide range of drying temperatures and tempering times, the moisture content of the friable mixture was inversely related t o the degree of granulation. When no tempering time was allowed, an increase in free starch (related t o the pastiness of the mashed potato on reconstitution) was observed

242

R. L. OLSOX AND W. 0. HARRINGTON

with increasing moisture content of the mix. Further, a n inverse relationship was found between moisture content and bulk density if a 3-hr. tempering time was allowed in the process. I n these studies the cooked potatoes were mixed directly with the seed material and mashed and mixed in a vigorous, planetary-type batch mixer. Cooling the cooked potatoes and mashing them before mixing with the seed material has been found successful but much depends upon the mashing and mixing equipment t h a t is used. When mashed hot, potatoes can withstand vigorous mashing without undue damage and consequent liberation of soluble starch by rupture of tissue cells. With reduced prodiict temperature during mixing, increased mechanical damage was observed (Table 111).However, such a decrease in temperature during tempering mill result in a desirable increase in degree of granulation. TABLE I11 Effect of Temperature of Mixing and Tempering on Potato Granule Qualitye Temperature of mixing (OF.)

136 133 33 (L

tempering

Yield of minus-70-mesh particle size

(OF.)

136 39 33

(%I

Package density (g./ml.)

20 62 87

0.86 0.88 0.91

Blue Value 123 94

249

From Olson and Harrington, unpublished.

The cooked potatoes must be adequately mashed and mixed so t h a t they are reduced essentially t o individual tissue cells. However it seems evident that any mixer will in time cause cells t o break and total mixing time should be reduced t o a minimum. After the cooked potatoes are mashed and mixed, they should be held for a period of time before drying, before agglomerates are loosened by a secondary mix or by screening. During the tempering time the physical characteristics of the potato starch change (Potter, 1954). The solubility of the starch and its swelling capacity decrease. The rate of change in these properties is affected inversely by a reduction in temperature and by a reduction in moisture content t o about 29%. The rate of change decreases with decrease in moisture content below t h a t point and changes are not found below about 15% moisture. The changes in these physical characteristics of the potato starch during tempering are paralleled by changes in degree of friability of the moist mixture. With increased holding time, a n increase in degree of

243

POTATO GRANULE MANUFACTURE

granulation was observed with mixtures at 35, 40, and 45% moisture content (Olson et al., 1953). With reduced temperature during tempering an increase in the degree of granulation was observed for a given holding time (Table HI). These facts lead one t o the conclusion th a t some attributes of potato granule quality may well be related t o changes that take place in the starch gel during the process of manufacture, and, further, t ha t the tempering period serves a function beyond the simple equilibration of moisture between the freshly cooked and previously dried material. IX. GENERALCONSIDERATIONS 1 . Raw Material

In many investigations it has been found th a t potatoes of the nonwaxy type are more suitable than the waxy type for the production of high-quality potato granules (Barker et al., 1943: Barker and Burton, 1944; Burton, 1944a, Proctor and Sluder, 1944; Campbell et al., 1945; Rendle, 194513). I n studies a t this labora.tory, the specific gravity of the raw material (an indicator of degree of waxy character of cooked potato tissue) was shown t o influence quality of product. I n one experiment a single lot of Russet Burbank potatoes was separated by brine flotation into 4 specific-gravity groupings. By comparable processes, potato granules were produced from the 4 lots. Results, summarized in Table IV, indicate t hat with an increase in specific gravity an increased yield of TABLEIV Effect of Specific Gravity of Raw Material on Potato Granule Quality" Yield, minus-70

a

Specific gravity

(%I

Package density (g./ml.)

< I ,074 1.074 t o 1.084 1.084 to 1.090 >1.090

86 87 90 90

.90 .91 .92 .93

Blue Value 248 257 153 117

From Olson and Harrington, unpublished.

fine particle-sized material, an increased package density, and a decreased Blue Value (reflecting a decrease in the degree of pastiness of the reconstituted product) were obtained. 2. Predrying Operaticns

Of economic importance in the manufacture of potato granules is the fact t ha t peeling, in the usual sense, is not an essential operation. All but

244

R. L. OLSON AND W. 0. HARRINGTON

a negligible portion of the peel fragments can be removed from the product by screening after the product is dried. Of advantage, in addition t o the elimination of peeling, is the potential increase in yield because screening removes a smaller portion of potato flesh than do the conventional peeling methods. On the other hand the peel may impart a flavor t o the product, resembling t h a t of baked potatoes, which may or may not, be desired. Further, the presence of skin on the potatoes may make more difficult the detection of bruised and spoiled tissue that should be removed by trimming before the potatoes are cooked. I n the absence of peeling, vigorous mechanical washing must be used t o insure removal of dirt. Peel may contribute adversely t o product color. I n the production of good mashed potatoes it is essential that the cooking time be appropriate t o the tubers used. Different raw materials require different cooking times. I n general, the higher the total solids content of potatoes, the shorter the time required for cooking. A properly cooked potato will break u p into a mealy mash when mixed without undue rupture of tissue cells. If potatoes are undercooked, the tissue cannot be satisfactorily mashed and one of two conditions will generally prevail. If the mashing action is mild, lumps of undercooked tissue remain in the mash; if vigorous, the lumps may be subdivided but only with disintegration of much of the cellular structure. This will liberate soluble starch and contribute a pasty consistency t o the mash. On the other hand, if potatoes are overcooked, there appears t o be a weakening of the cell-wall structure, releasing solubilieed starch or, a t least, rendering the individual cells less resistant t o mechanical damage and subsequent release of starch during the process or rehydration operations. A manifestation of undercooking is the presence, in the coarser screen fractions of the product, of particles of dried, vitreous potato tissue appearing much like rice grains. The effects of overcooking are not obvious. Probably the principal effect is one of lower textural quality in the reconstituted product. I n the presence of complicating raw material and processing variables, i t is not simple t o isolate the effect of overcooking of potatoes in potato granule manufacture. However, i t is desirable t o maintain minimum cooking times. Operations are generally controlled by adjusting the cooking time so as t o just avoid the obvious manifestations of undercooking. Because of the time required t o transfer heat through potato tissue, potatoes should be sliced before cooking t o insure uniformity in cooking. Cutting the potatoes t o a maximum heat-transfer depth of about 3.8 in. (potatoes cut t o a 3i-in. slab thickness) appears t o provide the necessary uniformity. As tissue cells are disrupted along the cut potato surfaces, starch is

POTATO G R A N U L E MANUFACTURE

245

released. To avoid undue material loss, slicing and peeling should be held t o a minimum. T o avoid a contribution t o undesirable pastiness in the reconstituted product, the released starch should be thoroughly washed from cut surfaces. Mashing of the cooked potatoes is a critical operation in the production of high-quality granules. The individual tissue cells should be separated by mechanical action but remain whole with cell walls enveloping the swollen, heat-solubilized starch. The type of mixer used, the quality of raw material, and the temperature during mashing are very important. A vigorous mashing action, as with a high-speed, planetary-type mixer, must be of short duration and with a relatively high product temperature t o prevent mechanical damage t o product. I n the add-back method, the immediate addition of previously dried material, as the mashing commences, appears t o reduce product damage. It is possible t o mash potatoes after the cooked product has cooled [to around 140" F. (60" C.)] b y the gentle action of pressing between revolving drums. A 0.050-in. separation between drums appears t o be appropriate. With greater spacing, the mashing is inadequate, with smaller, the flow rate of the product is reduced. The drums should revolve a t the same angular velocity t o prevent shearing action t h a t might damage cell structure. T o self-feed properly in mashing potatoes, drums should be a t least 18 in. in diameter. The tempering operation in the manufacture of potato granules is used following the mashing operakion and converts the mash from a gummy or sticky t o a friable condition. As outlined above, the temperature and product moisture content affect the rate of change in friability during the tempering operation and have a n important effect on the degree of granulation of the product. It appears advisable t o do preliminary mashing of potatoes while they still retain cooking heat and a relatively high temperature [above 140" F. (60" C.)] in order t o achieve adequate cell separation without undue cell damage. On the other hand, i t is advantageous t o provide a mixing operation following product cooling [to below 90" F. (32" C.)] and tempering in order t o adequately granulate or subdivide the product into individual cells. 3. Drying Methods and Equipment

In commercial operations potato granules have been dried in tunnel, rotary-turbo, kiln, and various types of pneumatic dehydrators. The advantages of pneumatic drying systems include efficient utilization of heat and the prevention of agglomeration of separated granules. A dis-

246

R. L. OLSON AND W. 0. HARRINGTON

advantage is the susceptibility of the product t o abrasion and impact damage as i t is conveyed in the duct system and separated from the air stream in the various standard-type collectors. Descriptions of several early developments in pneumatic dehydrators designed specifically for potato granule operations have been published (Fison, 1943; Bar, 1941; Anonymous, 1946). Proctor and Sluder (1943) described a batch-type pneumatic drier for laboratory studies. Studies on certain engineering aspects of duct drying were described by Olson (1947). These studies were not exhaustive but probably provide a useful introduction t o such investigations. A pneumatic drier designed to reduce abrasion and impact damage consisted of a vertical riser and a novel collector that provided for a single change in direction of product flow under decelerating conditions of the air stream (Olson et al., 1953; Neel et al., 1954). Constructed for use in laboratory investigations, the design principles of this drier have had limited commercial application for finish drying, cooling, and conveying of product. One advantage of a vertical duct drier is that it makes possible the use of relatively low air velocities (1500 t o 2000 ft. per min.). Inhorizontal ducts it. is necessary t o maintain higher velocities to prevent the settling of the product on the bottom of the duct. Cooley et al. (1954) indicated that air velocities of 70 ft. per sec. (4200 ft. per min.) were necessary t o keep moist potato powder in suspension in a horizontal pipe. I n this investigation it was found th at product damage was greater (measured by chemical analysis of the free starch) with higher air velocity in the duct. Olson et al. (1953) presented data relating bulk or package density t o the inlet air temperature in a pneumatic drier. In the range of 212 t o 390" F. (100 t o 199" C.) a more dense product was obtained with lowertemperature drying air. Cooley et al. (1954) confirmed this finding and presented data showing that the differences in air temperature (or perhaps drying rate) did not materially affect the particle size distribution found by screen analysis of the product. Although particle size was found t o be correlated with bulk density, the magnitude of this factor was slight compared t o the correlation of drying-air temperature with density. Neel et al. (1954) described a finish drier as a n improved means for reducing moisture content of potato granules from about 12% t o about 4%, a level a t which quality is reasonably stable against browning, even under adverse temperature conditions. This drier utilizes the principle of a bed of granular solid material fluidized b y a uniform, low-velocity flow of air, directed vertically upward from the bottom of a drying chamber. When the drying chamber is filled, the introduction of material

POTATO G R A N U L E MAXUFACTURE

247

at one end causes an overflow of product a t the exit end and a continuous drying operation is achieved. Other dehydrators used for finish drying have been the tray, pneumatic, rotary-kiln, and rotary-turbo dehydrators. Questions of economic feasibility and maintenance of acceptable product quality have been raised concerning finish drying to below 6 % moisture content. It remains t o be seen whether or not fluidized-bed drying or some other method will make i t commercially practicable for the interested trade t o reduce moisture-content specification to below 6 %.

4. Product Quality I n the appraisal of potato granules and in the development of a process for their manufacture, a number of factors concerning product quality are considered. Certainly of prime importance is the particle size distribution of the finished product. If in the process tissue cells are completely separated from each other, after drying they will pass without difficulty through a screen of about 60 mesh. By screening t o such a size classification, a uniform appearing material is obtained. Within this material considerable differences exist because of the natural variability of cell size. If the add-back method is used, a major portion of the material that does not pass a 60-mesh screen can be recycled in the process. However, as the particle size gets larger, the effectiveness of the material in mixing and tempering operations is lessened, and a higher percentage of dried product must be added back t o allow successful continuance of the process. Further, the conveying and drying of large particles in pneumatic driers imposes problems connected t o product flow and drying rates. Therefore, all material that will not pass through a 16 t o 20-mesh screen is removed from the system (this includes most of the peel fragments if such are present). This coarse material has salvage value as stock feed but should not exist in excessive amounts because of economic considerations. A specification that all material packed must pass a 70-mesh screen is quite feasible for the add-back process, because larger material can be recycled. It is desirable that about 50% of the material to be recycled should also be of such a fine particle size (minus-70-mesh) as t o insure high product quality in a continuing operation. With methods not involving add-back, the product usually contains appreciable quantities of larger material. In such cases modification in mixing operations to achieve a greater degree of granulation, or use of a partial add-back may be desirable. I n laboratory operations, good yields of fine particlesized granules have been reported for the cold-tempering method (Campbell et al., 1945) and for the solvent-extraction method (Treadway, 1954).

248

R. L. OLSON A N D W. 0. HARRINGTON

Considerations of bulk density, stratification of product in packages due t o variation in particle density, and rehydration rate would be determining factors in particle-size specifications. Related t o particle-size distribution and affected by several factors in the process is the bulk density of potato granules. The bulk density of potato granules has been specified as 6 lb., 2 oz. per No. 10 can for military procurement. This requirement establishes potato granules at about 0.9 the density of water. Although this specification can be met without undue difficulty if the add-back process is used, other methods of manufacture, such as the "freeze and squeeze" and solvent-extraction processes, produce a product with a density that is somewhat lower. The quality attributes of potato granules of greatest importance t o the consumer are flavor, color, and texture of the product when prepared for serving. Factors in the process can materially affect all three. Flavor is delicate and subject t o loss or change. Highly volatile flavor components are lost in dehydration. Heat damage may impart a scorched flavor. The raw material itself may be responsible, as in the case of severe sunburning, which makes potatoes bitter and turns them green. The color of potato granules is the result of natural pigment (yellow t o yellow-green) and processing effects. Most prominent is the browning that may result from heat damage in the process. A graying is sometimes observed that is probably related t o holding at high temperature levels while the product is fairly moist. Perhaps consistency is the quality attribute least satisfactorily controlled by present-day practices, although great improvement has been made by all of the commercial dehydrators in recent years. Raw material certainly has significance in this regard. However, nearly every unit operation in any potato granule process has also a n effect on the freeing of solubilized starch from the cooked potato cells and this affects the texture of the reconstituted product. The temperature of the liquid used t o reconstitute potato granules has a marked effect on the texture of product. The temperature must be a t least 160" F. (71" C.) t o satisfactorily reconstitute potato granules.2 An upper temperature limit cannot be defined although it may be generally stated that the higher the temperature of the water of rehydration, the greater the tendency toward an undesirable pastiness. This factor is relative and it is found that an exceptionally high-quality product is better when rehydrated a t nearly boiling temperature than a poor sample rehydrated a t 160" F. (71" C.). An improvement in texture of reconstituted potato granules t o more

* An exception is the selvent-extracted potato granules, which can be rehydrated at lower temperatures.

249

POTATO GRANULE MANUFACTURE

closely resemble freshly mashed potatoes is probably a prerequisite t o the establishment of a domestic market of considerable size. I n an earlier volume of this series the storage characteristics of potato products have been reviewed (Ross, 1948). As demonstrated b y Burton (1945, 1949) and confined by Hendel et al. (1951), the two major deteriorative processes that occur in storage are browning, accompanied by a scorched flavor development and an oxidative staling that causes an ‘(off” flavor attributed to deteriorative changes in the fatty constituents of the potato. The browning can be checked by reducing moisture content, adding sulfite, and reducing storage temperature. The reduction of product moisture content, however, enhances the development of oxidative staling. Even a t near freezing temperature the stale, off-flavor may develop. Control of this factor is best achieved by packaging in the absence of oxygen as in an inert atmosphere of nitrogen. The nutritive value of potato granules as served is probably equivalent t o that of dehydrated potato dice. Although processing losses of vitamin C are greater for the granules, the instantaneous reconstitution contributes no further substantial loss whereas considerable loss occurs in the reconstitution of dice (Barker et al., 1943; Green et aZ., 1947, 1948, 1949). Data in Table V suggest that the principal vitamin C loss in granule production occurs in the cooking operation. TABLEV Vitamin C Retention in Potato Granules during Process and Reconstitution“

Type of potatoes

Average of two separate experiments with same raw material (mg./100 g. at moisture content as served)

Vitamin C in raw potatoes, Russet Burbank variety about 3 months after harvest Potatoes, cooked Potato granules (add-back process) Reconstituted potato granules Q

10.6

7.0 8.4 8.5

From Olson and Harrington, unpublished.

X. QUALITYEVALUATION OF POTATO GRANULES 1 . Subjective Appraisal

Quality evaluation of potato granules, not unlike that of other food products, is involved in the development of new and better processing methods and in the control of manufacturing operations. I n development

250

R. L. OLSON AND W. 0. HARRINGTON

work it is important t o determine the effect of new or modified processes on quality attributes of the product. Results are obtained demonstrating significant differences in certain quality attributes and the relationship of such differences t o specific variations in the process. On the other hand, for control of plant operations the objective of appraisal is t o establish and maintain a suitable range of product uniformity and minimum level of acceptability. Techniques for determining differences in foods have been established, using trained test panels and following accepted statistical procedures for determining significance of results. Boggs and Hanson (1949) have provided a general review of the subject. For potato granule appraisal, methods of sample preparation, presentation, and scoring must be developed with consideration not only of product but of the particular quality attribute being judged. For example, flavor judgments should not be influenced by differences in sample color, which can be neutralized by use of colored (amber or red) light in the test area. Color should be judged under controlled conditions of uniform lighting. Off-color and off-flavor can probably be best scored against a standard control sample. Consistency of reconstituted potato granules can also be scored relative t o standard controls. Wood et al. (1955) described a method of ranking mashed potatoes for rubberiness with two coded control samples. The relative ranking placed unknown samples into categories which enabled indirect comparison of samples tested at different times. This has been a valuable tool for research and has eliminated much tedious and costly cross comparison of experimental samples. The evaluation of samples and establishment of levels of acceptability for production control is a problem of greater magnitude than the determination of differences between samples. The final answer t o consumer acceptability is found in the sales data of a commercial concern. Continuing success often will depend upon approximating a level of acceptability, training analysts t o recognize the level, and appraising routine production samples t o provide a control over processing operations. More frequently than not the standard of acceptability is established as a mental concept rather than as a selected material that can be used for comparisons. The reliance on memory and the momentary diligence of the judge during the test may well lead t o an unnecessary lack of precision. Although the replications necessary for statistical analysis in the accepted difference measurements are too cumbersome for routine production control, it is probably advantageous t o establish physical standards for comparison. These standard samples, coded and presented with production samples for judgment of difference, would not give the precision of

POTATO GRANULE MANUFACTURE

251

test panel procedures with 30 t o 50 replications and statistical treatment of data t o establish levels of probability of correctness. However, with trained analysts, the method should achieve a higher degree of product uniformity than would be possible with reliance entirely on the analyst’s memory. 2. Microscopy I n recognition of the relationship of broken cell structure, released soluble starch, and pasty or rubbery texture of reconstituted potato granules, microscopic observations have been used as an aid in the appraisal of potato granules. Bunimovitch and Faitelowitz (1936) reported such observations. Greene et al. (1947, 1948, 1949) counted ruptured potato cells in a microscopic field, and established a correlation between subjective evaluation and percentage of cells ruptured. If less than 6% were ruptured, the product was superior; lO-l2%, average; and 20%, a very pasty product. Proctor and Sluder (1944) and Campbell et al. (1945)found counts of broken cells in a microscopic field t o be a suitable method for evaluating texture of reconstituted potato granules. They reported use of this method as a control in experimental investigations. It was possible t o appraise samples taken from the process line and, within limits, predict the general area of acceptability of the product when finished. Hall (1953) has discussed various appraisal methods and postulates t h a t a better value can be obtained by count of broken cells than by measurement of released starch (see below, Objective Mcasurements) . It is noted that, in these references, methods other than the add-back process were involved. [Bunimovitch and Faitelowitz (1936) used the direct two-stage dehydration; Greene et al. (1947,1948, 1949) and Hall (1953)used the “freeze and squeeze” process; and Proctor and Sluder (1943, 1944) and Campbell et al. (1945) used a cold-tempering method.] T o date, no demonstration of successful application of microscopic examination in quality evaluation has been found where the add-back method is used, probably because soluble starch released by cell-wall rupture and then recycled adheres t o intact cells. Microscopic identification of the degree of starch liberation and total cell breakage could not be very precise under such conditions. 3. Objective Measurements

The ultimate appraisal of food products is the response of individuals t o quality attributes. However, cost and lack of precision of subjective appraisals make it desirable t o develop objective measurements that can be correlated with the subjective evaluations. A number of attempts have

252

R. L. OLSON AND W. 0. HARRINGTON

been made t o do so with reconstituted potato granules. I n addition there are objective physical measurements of granules that have descriptive value in themselves. The measurement of viscosity of supernatant liquid obtained by centrifuging a mixture of water and reconstituted potato granules was found a suitable method for control of a potato granule process (Barker et al., 1943; Barker and Burton, 1944). Free solubilized starch extracted from the mashed potatoes was the principal factor related t o the viscosity. Campbell et al. (1945) also used a viscosity measurement but found the method less valuable in their work than microscopic examination. I n preliminary studies a t the Western Utilization Research Branch it was found that viscosity measurement of a starch extract of potato granules is a more difficult analytical procedure than a chemical test for the starch. Viscosity measurements were thus abandoned. Direct measurement of viscosity of potato mash with a Brookfield viscometer3 was not found t o be satisfactorily reproducible. Cooley et al. (1954) reported successful use of viscosity measurement of reconstituted potato granules with a Brookfield viscometer with product supported on a helipath stand. However, i t was found necessary t o rehydrate the granules with appreciably more water than would be used for food preparation. I n no case have viscosity measurements been found t o provide an absolute judgment of potato granule quality. Differences in processing method and probably in raw material influence the relationship of the viscosity of supernatant liquid from potato granule slurry t o the consistency of granules as prepared for serving. However, when raw material is uniform and processing operations do not vary widely, viscosity measurements demonstrate a correlation with subjective appraisal values and appear t o provide a useful purpose in investigations of potato granule processes. Chemical determination of starch extracts from potato granules, as an objective measure of product consistency, is also limited t o similar raw materials and processing methods. A convenient method is extraction of potato granules in an excess of water a t about 150" F. (66" C.). Iodine is added t o a portion of a filtrate from the slurry. The intensity of the starch-iodine color is determined photometrically and designated the Blue Value. A method for determination of the Blue Value has been developed (Olson et al., 1953; Mullins et al., 1955) which has been useful in research and development investigations of potato granule technology. 3 Mention of product name does not imply endorsement by the U. S. Department of Agriculture.

POTATO G R A N U L E MANUFACTURE

253

Important quality attributes of potato granules that can be measured objectively include bulk density (weight of sample per unit volume), color, and particle-size classification. Adequately reproducible determinations of bulk density can be obtained by dividing the weight of a sample by its volume, as measured by tamping or jarring t o minimum volume in a graduated cylinder (Olson et al., 1953). Color can he readily measured photometrically by difference comparison with standard color plates in reflectance-type photometers. Such objective data are particularly useful in following changes due t o storage degradation of uniform material stored under different conditions. The establishment of color standards for production cont(ro1 also appears possible, but, of course, depends upon subjective evaluation and judgment t o establish a base point and an allowance for deviations. Particle-size classification is accomplished by screen analysis. I n the add-back process it is perhaps more important t o follow particle size in connection with seed used and total dried product rather than the material that is packaged for sale. For processes where recycling is used, the particle-size classification and the effects of oversized material on rehydration character and bulk density may be a very important indication of operational success.

REFERENCES Allen, A. E. 1921. Process of dehydrating potatoes. U. S. Patent 1,377,172. Anonymous. 1937. Packaged baked potato flour. Food I n d s . 9, 324. Anonymous. 1943a. Potato shreds, Rogers Bros. Pioneered development of dehydrated potato shreds. Western Canner and Packer 36(2), 37, 39, 41. Anonymous. 194313. How potato shreds are made. Food I n d s . 16(3), 47-49. Anonymous. 1946. Simplot shifts production set-up. Food Packer 27(7), 39-40. Anonymous. 1948. Pre-cooked potatoes. Food M a n u f . 23, 295. Bar, P. J. 1941. Process and apparatus for treating solids in gases. Brit. Patent KO. 546,088.

Bar, P. J. 1943. Dehydrated potato product and method of making same. Brit. Patent No. 566,498. Barker, J. 1941. Improved process for producing starchy vegetable products in dry powdered form. Brit. Patent No. 542,125. Barker, J., and Burton, W. G. 1944. Mashed potato powder. I. General characteristics and the “brush-sieve” method of production. J. SOC.Chem. I n d . 63, 169-172. Barker, J., Burton, W. G., and Cane, R. 1943. Mashed potato powder. I. Properties and methods of production, Great Britain Dept. of Scientific and Industrial Research and Ministry of Food. Dehydration; United Kingdom Prog. Rept., Sec. VI, part 6, 10 pp. Boggs, bl. M., and Hanson, H. L. 1949. Analysis of foods by sensory difference tests. Advances in Food Research 2, 219-258. Bosncll, V. R., and Bostelman, E, 1949. Our vegetable travelers. Natl. Geographic Mag. 46(2), 145-153.

254

R. L. OLSON AND W. 0. HARRINGTON

Bowen, W. S. 1931. Spray-drying Idaho’s surplus potatoes. Food Znd. 3, 380-383. Brown, A. H., and Kilpatrick, P. W. 1943. Drying characteristics of vegetables-riced potatoes. Trans. Am. Soc. Mech. Engrs. 66, 837-842. Bunimovitch, M., and Faitelowitz, A. 1936. An improved method of reducing potatoes and other starch containing vegetables to the form of a dry powder. Brit. Pat. No. 457,088. Burton, W. G. 1944a. Mashed potato powder. 11. Spray-drying method. J . Soc. Chem. Znd. 63, 213-215. Burton, W. G. 1944b. Improvements in or relating to the production of dried potatoes. Brit. Patent No. 566,828. Burton, W. G . 1945. Mashed potato powder. 111. The high temperature browning of mashed potato powder. J . SOC.Chern. Ind. 64, 215-218. Burton, W. G. 1949. Mashed potato powder. IV. Deterioration due t o oxidative changes. J. Soc. Chem. Ind. 68, 149-151. Campbell, W. L., Proctor, B. E., and Sluder, J. C. 1945. Dehydration of precooked white potatoes in granular form. Research Reports on Quartermaster Contract Projects, Massachusetts Institute of Technology, July 1, 1944 to June 30, 1945, pp. 248-273. Cooke, E. W. 1912. Dehydrated potatoes and process of preparing same. U. S. Patent No. 1,025,373. Cooley, A. M., Severson, D. E., Peightal, D. E., and Wagner, J. R. 1954. Studies on dehydrated potato granules. Food Technol. 8(5), 263-269. Edwards, C. S. 1845. Improvement in preserving potatoes. U. S. Patent No. 4,337. Erskine, H. L., Jr. 1950. Development of a slurry process for the production of dehydrated mashed potato granules. Master’s Thesis, Kansas State College of Agriculture and Applied Science. Fison, F. J. 1943. Improvements in or relating t o a’method of and a means for drying powdered or other more or less finely divided material. Brit. Patent No. 566,170. Gano, 0. 1940. Treatment of potatoes to produce dried mashed potatoes. U. S. Patent No. 2,190,063. Greene, J. W., Marburger, G. C., and Rohrman, F. A: 1947. Instant mashed potatoes from dehydrated granules. Food Znds. 19, 1622-1625, 1749-1751. Greene, J. W., Rohrman, F. A., Marburger, G. C., Honstead, W. H., Messenheimer, A. E., and Olson, B. E. 1948. Development of a potato granule process. Chem. Eng. Progr. 44(7), 547-552. Greene, J. W., Conrad, R. M., and Rohrman, F. A. 1949. Dehydration process for starchy vegetables, fruits and the like. U. S. Patent No. 2,490,431. Hall, R. C. 1951. Mechanisms of potato cell, rupture resulting from dehydration process, Master’s Thesis. Kansas State College of Agriculture and Applied Science. Hall, R. C. 1952. Cell-rupture closeups aid development of better dehydrated potatoes. Food Eng. 24(8), 91, 221. Hall, R. C. 1953. Better potato dehydrating by slow freezing. Food Eng. 26(3), 90-91, 150, 152. Hall, R. C., and Fryer, H. C. 1953. Consistency evaluation of dehydrated potato granules and directions for microscopic rupture count procedure. Food Technol. 7(9), 373-377. Heimerdinger, H. M. 1926. Food product. U. S. Patent No. 1,571,945. Heisler, E. G., Hunter, A. S., Woodward, C. F., Siciliano, J., and Treadway, R. H. 1953. Laboratory preparation of potato granules by solvent extraction. Food Technol. 7(8), 299-302.

POTATO G R A N U L E MANUFACTURE

255

EIendel, C. E., Burr, H. K., and Boggs, Mildred h l . 1951. Factors affecting storage stability of potato granuIes. U. S. Bureau of Agr. and Ind. Chem., Agr. Research Admin., U. S. Dept. Agr. AIC 303, 8 pp. Jones, C. R., and Greer, E. N. 1940. Improvements in or relating to the reduction of potatoes and other starch-containing vegetables t o dry powder. Brit. Patent No. 537,669. Jones, C. R., and Greer, E. X. 1943. Mashed potato powder. 11. Hammer mill process. Great Britain Dept. of Scientific and Industrial Research and Ministry of Food Dehydration. United Kingdom Progr. Rept., Sec. VI, part 6 (11). Kaufman, C. W., Burgess, N. M., and Hollis, F., Jr. 1949. Process of preparing dehydrated mashed potato. U. S. Patent No. 2,481,122. King, J. 1948. Scientific problems in feeding a modern army in the field. Chemislry R: Industry (47), 739-743. Morris, T. N. 1947. “The Dehydration of Food,” Chapman and Hall, London. Mullins, W. R., Harrington, W. O., Olson, R. L., Wood, E. R., Nutting, M.-D. 1955. Estimation of free starch in potato granules and its relation to consistency of reconstituted product. Food Technol. I n press. Neel, G. H., Smith, G. S., Cole, M. W., Olson, R. L., Harrington, W. O., and Mullins, W. R. 1954. Drying problems in the add-back process for production of potato granules. Presented at 1953 Ann. Inst. Food Technol. Convention. Food Technol. in press. Nixon, E. L. 1944. Preservation of potatoes, fruit and vegetables. U. S. Patcnt No. 2,339,028. Olson, B. E. 1947. A study of the fundamentals of parallel flow drying in ducts. Master’s Thesis, Kansas State College of Agriculture and Applied Science. Olson, R. L., and Harrington, W. 0. 1951. Dehydrated mashed potatoes-A review. Bureau of Agr. and Ind. Chem., Agr. Research Admin. U. S. Dept. Agr. AIC 297, 23 pp. Olson, R. L., Harrington, W. O., Neel, G. H., Cole, M. mi.,and Mullins, W. R. 1953. Recent advances in potato granule technology. Presented a t 1952 Ann. Inst. Food Technol. Convention. Food Technol. 7(4), 177-181. Potato Corporation of Idaho. 1939. Treatment of potatoes to produce dried mashed potatoes. Brit. Patent 526,474. Potter, A. L., Jr. 1954. Changes in physical properties of starch in potatogranules during processing. Agr. and Food Chem. 2(10), 516-519. Poulton and Noel, Ltd., and Bostock, B. R. 1946. Improvements in or relating t o the dehydration of potatoes. Brit. Patent No. 589,830. Proctor, B. E., and Sluder, J. C. 1943. Dehydration of pre-cooked white potatoes in granular form. Research Reports on Quartermaster Contract Projects, Massachusetts Institute of Technology, July 1, 1942 t o June 30, 1943. pp. 140-151. Proctor, B. E., and Sluder, J. C. 1944. Dehydration of pre-cooked white potatoes in granular form. Research Reports on Quartermaster Contract Projects, Massachusetts Institute of Technology, July 1, 1943 to June 30, 1944. pp. 227-243. Remmers, B. 1918. Method of preparing pre-cooked food products. U. S. Patent hTo. 1,258,047. Rendle, T. 1943. Improvements in and relating to the preparation of cooked starchy vegetables in powder form. Brit. Patent No. 566,167. Rendle, T. 1945a. Preparation of cooked starchy vegetables in powder form. U. S. Patent No. 2,381,838. Rendle, T. 194513. The preservation of potatoes for human consumption. Chemistry & Industry (45), 354-359.

256

R. L. OLSON AND W. 0. HARRINGTON

Rivoche, E. J. 1948. Improvements in and relating to the drying of vegetables. Brit. Patent No. 601,151. Rivoche, E. J. 1950. Drying of starchy foodstuffs. U. S. Patent No. 2,520,891. Rivoche, E. J. 1951a. Method and technique of food drying. U. S. Patent No. 2,572,761. Rivoche, E. J. 1951b. Process of preserving moisture-containing cellular foodstuffs. U. S. Patent No. 2,572,762. Ross, A. F. 1948. Deterioration of processed potatoes. Advances in Food Research 1, 257-290. Salaman, R. N. 1949. “The History and Social Influence of the Potato,” Cambridge Univ. Press, England. Salaman, R. N. 1953. Potatoes as a crop and a n industrial raw material. The potato’s influence in shaping society. Chemistry & Industry (35), 907-912. Stamberg, 0. E., and Beresford, H. 1943. Potatoes baked, then dehydrated to avoid loss of product. Food. Inds. 16(9), 78-79. Stoddard, E. S. 1922. Precooked food. U. S. Patent No. 1,402,108. Tjomsland, Anne. 1950. The white potato. Ciba Symposia 2(6), 1254-1284. Treadway, R. H. 1954. Personal communication. Van Arsdel, W. B., Brown, A. H., and Lazar, M. E. 1947. Drying-rate nomographs. I. Riced white potatoes. Bureau of Agr. and Ind. Chem., Agr. Research Admin., U. S. Dept. of Agr., AIC 31-1 (revised). Volpertas, Z. 1937. Improvement in process and device for reducing vegetables containing starch t o dry powder. Brit. Patent No. 496,423. Volpertas, Z. 1939. Process for reducing vegetables containing starch to a dry powder. Brit. Patent No. 525,043. Volpertas, Z. 1944. Art of dried starch bearing food. U. S. Patent No. 2,352,670. Werts, M. 1947. They’ve lost their eyes. Kansas Agr. Student 23(3), 11. Willets, A. K., and Rendle, T. 1946. Improvements in and relating to the production of mashed potato powder. Brit. Patent No. 589,876. Willets, A. K., and Rendle, T. 1948. Production of mashed potato powder. U. S. Patent No. 2,439,119. Wood, Elizabeth R., Olson, R. L., and Nutting, M.-D. 1955. A method for the comparison of consistency in potato granule samples appraised at different times. Food Technol. 9(4), 164-168.

The Thermal Destruction of Vitamin B1 in Foods BY K . T. H . FARRER Research Laboratories. Kraft Foods Ltd., Melbourne. Australia Page I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 I1. Factors Influencing the Thermal Dcstruction of Vitamin B1. . . . . . . . . . . . . 253 1 . Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 3. 4 5. 6. 7.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrolyte System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heavy Metals . . . .................................... Concentration of Electrolytes-the Salt Effect . . . . . . ...... Nonelectrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Form of the Vitamin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 . Concentration of Thiamine and Cocarboxylase . . . . . . . . . 10. Oxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Moisture Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Losses in Cereals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Loss during Baking (or Toasting) of Bread . . . . . . . . . . . . . . . . . . . 2 . Stability of Vitamin BI from Different Sources during Baking . . . . . . 3 . Loss during the Cooking of Breakfast Cereals . . . . . . . . . . . . . . . . . . . . 4 . Effect of Various Baking Powders on Stability in Baked Products . . . 5 . Losses in Other Cereals.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Losses in Meats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Losses in Processing Vegetables . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Losses in Other Foodstuffs . . . . . . . . . . . . . . . . . . . . . 1. Dairy Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Peanuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Losses on Storage . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Unprocessed Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Canned Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Dehydrated Products . . . . . 4. Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

111.

I V. V. VI .

VII .

VIII .

259 259 259 260 260 261 261 262 262 263 261 264 268 269 271 274 275 285 293 293 293 294 294 294 296 299 301 303 306

I . INTRODUCTION I n the course of the last ten to fifteen years. a considerable amount of information concerning the thermal destruction of vitamin B1 during the 257

258

K. T. H. FARRER

cooking, processing, and storage of foodstuffs has been published. However, by far the greater part of these data has been concerned with the retention of the vitamin during the treatment of a particular foodstuff under certain specified conditions and, generally speaking, no attempt has been made to relate the results obtained t o any other foodstuff or any other set of conditions. During the same period, some work has been done to determine the fundamental principles underlying the thermal destruction of vitamin Bl* and several factors are now known to be involved. This paper aims: (1) t o survey the data available on factors influencing the thermal destruction of vitamin B1, and the thermal losses of vitamin B1 from foodstuffs; (2) t o show that the most reliable and satisfactory method of approaching vitamin B1 destruction is through simple reaction kinetics; (3) by means of the kinetic approach, to show: (a) that of the data available, much is worthless and the rest can be correlated and simplified, (b) that information available can be used through the first order reaction and the Arrhenius equation t o predict the behavior of vitamin Bl under specified conditions. The term "thermal destruction" is used t o differentiate this hydrolysis from the results of enzymic or microbiological activity, and covers losses of vitamin B1on storage as well as a t the higher temperatures associated with cooking or processing.

11. FACTORS INFLUENCING THE THERMAL DESTRUCTION OF VITAMINB1

I. Temperature Most workers have understood the importance of varying temperatures in considering the retention of vitamin B,, but very little has been done t o study quantitatively the effect of temperature from the point of view of reaction kinetics. Rice and Beuk (1945) systematically studied thiamine decomposition in pork a t different temperatures. Farrer and Morrison (1949) have shown that there is no deviation from the Arrhenius equation for thiamine destruction in buffer solutions a t temperatures between 50 and110" C. (122 and 230" F.). Farrer (1950, 1953a) has further used the Arrhenius equation for studies on processed cheese and yeast extract at storage temperatures, while Bendix et al. (1951) have also used kinetic methods for a study of the mechanism of thiamine destruction in peas, corn, Lima beans, and tomato juice from 104.5 to 132" C. (220 t o 270" F.). *For the purposes of this discussion "vitamin B," is considered to consist of thiamine, cocarboxylase, and "bound " thiamine.

THE THERMAL DESTRUCTION O F VITAMIN B1 I N FOODS

25'3

2. Time Early in the history of this work, Sherman and Burton (1926) showed t h a t destruction of vitamin B1 was greatly increased if the time of heating was increased from 1 t o 4 hr. The effect of time was clearly demonstrated by Farrer (1941), and much work since then has underlined the importance of this factor. Feaster et al. (1948) have shown clearly the interrelationship of time and temperature in thiamine retention in canned foods. 3. p H Here again early work pointed t o the importance of the p H (Sherman and Burton, 1926; Guha and Drummond, 1929), but i t was not until the thiochrome method (Jansen, 1936) for the determination of vitamin B1 became available that the full importance of p H was demonstrated (Farrer, 1941; Booth, 1943; Farrer, 1945,). Roy (1953) has emphasized the destructive effect of high p H in a study of losses from various samples of water used for cooking in Western Bengal.

4. Electrolyte System Beadle et al. (1943) first made the suggestion that the buffer system of the solutions under examination is important. This idea was greatly extended by Farrer (1945a) t o reconcile a variance between the results obtained earlier b y him (Farrer, 1941) and by Booth (Zoc. cit.). As a result of the study of thiamine destruction a t 100' C. (212" F.) in 4 buffer systems (citric acid-phosphate; phosphate; succinic acid-borate ; phosphate-borate) Farrer (1945a) believed that the rate of destruction of thiamine changes as the ionic constitution of the buffer changes with rising pH, and t h a t where the latter change is accompanied by a marked change in pH, there is a marked change in rate of destruction of thiamine. Similarly, where the pH is only slightly affected by the changing ionic constitution of the solution, there is a correspondingly slight change in the rate of thiamine decomposition. This close association of anionic constitution with p H in catalyzing the pseudo-unimolecular decomposition of thiamine has been underlined by subsequent work (quod vide) on various forms of thiamine under various conditions, and it is the author's belief that this same effect operates in the more complicated systems met with in footstuffs. I n the meantime, McIntire and Frost (1944) had claimed that a- and @-aminoacids decrease the rate of thiamine destruction. This fact, too, would be in accord with the recognition of the special part played by the electrolyte system as amino acids are well-known as buffer agents.

260

K. T. H. FARRER

I n 1944 some Japanese workers (Watanabe et al., 1944) published a study on the effect of acids on the stability of vitamin B1 solutions. This was not brought t o notice until 1951 (Chem. Abstr. 46,3566).' They used 0.001 M solutions of thiamine and 0.001 M-O.1 M concentrations of acids (hydrochloric, sulphuric, oxalic, maleic, acetic, and boric), and determined reaction velocities for thiamine destruction at 100" C. (212" F.). They found that all acids increased thiamine stability, although acetic and boric acids were less effective than the others. They appear t o have noted the general pH effect while overlooking the significance of the effect of the individual buffer anions. The same group (Watanabe et al., 1949) reported t ha t lithium, sodium, and potassium chlorides and potassium nitrate had no effect, and th at sodium sulphate had a slight accelerating action on thiamine destruction. Watanabe and Marui (1949) stated th a t although the decomposition was strongly accelerated b y sodium sulphite, borate, thiosulphate, acetate, carbonate, and monohydrogen phosphate as well as potassium dihydrogen phosphate, the effect was the same as that of alkali. Again, it would appear from the abstract th a t the significance of the individual buffer anions was overlooked. Ache and Ribeiro (1945) have claimed that sodium sulphite and chloride accelerate decomposition and Murphy and Goodyear (1949) report th a t potassium iodide is without effect. Much earlier, Escudero and de Alvarez Herrero (1942) had shown that potassium bromate (a bread improver) rapidly destroys thiamine in solution and in doughs. This is almost certainly a n oxidation effect. Similarly, Tavares and Rodrigues (1947), while showing that the stability of thiamine in pharmaceutical solutions is unaffected by sulphate, chloride, iodide, sodium, or magnesium ions, were emphatic that thiosulphate ions caused considerable destruction. This, too, is not a true ionic effect but, in this case, a reduction. 5. Heavy Metals

Farrer (1947a) has shown that heavy metals which can form complex anions with constituents of the medium can influence thiamine destruction very significantly. This is probably a special case of section 11, 4. 6. Concentration of Electrolytes-the

Salt Eflect I n certain circumstances, a t least, the concentration of the buffer salts can influence the rate of destruction of thiamine (Farrer, 1947b; 1949b) and cocarboxylase (Farrer, 1949b). Whereas almost all the papers cited in this review have been read in their original

form, most Japanese and some obscure foreign journals are unobtainable in Australia.

THE THERMAL DESTRUCTION

OF VITAMIN

~1

IN FOODS

261

7. Nonelectrolytes Melnick et al. (1941) suggested t h a t a t low temperatures thiamine is more stable in digestive juices than in synthetic solutions. Greenwood et al. (1943) have reported that both gelatin and egg albumin considerably retard the rate of aneurin destruction at 98" C. (208" F.). Atkin et al. (1943) also refer t o the protective effect of gelatin and albumin (at atmospheric temperature) and also of gums, dextrin, and soluble starch. Rice et al. (1943, 1944), too, have described the stabilizing effect of certain cereals on thiamine during the storage of dehydrated pork and, in addition, McIntire and Frost (1944) have described the effects of a number of nonelectrolytes (e.g. certain amines) on thiamine destruction. Watanabe and Sakaki (1944) studied the effect of sucrose, lactose, and glucose on thiamine destruction a t 110" C. (230" F.) and claimed t o show a slight accelerating effect. Watanabe (1951) used 0.1% of thiourea t o stabilize a 1% solution of thiamine, which is then claimed t o be stable for 1 hr. a t 100" C. (212" F.) or 30 min. a t 115" C. (221" F.). Terao and Kawaziri (1951) use methionine for a similar purpose and Watanabe et al. ( I 952) additionally reported that thioglycollic and thiolactic acids also stabilized thiamine solutions. Ache and Ribeiro (1945) have stated that glycine, xanthine, chloretone, and riboflavin are without effect on thiamine destruction, but that fructose, invertase, and inositol in 5 % solutions retard decomposition. Murphy and Goodyear (1949) claim that glucose and ethyl alcohol are without effect. An interesting report is that of Inagaki and Takeda (1950, 1951) whose figures suggest t h a t thiourea has a considerable protective effect on thiamine over a wide range of pH. No clear picture emerges from this brief survey, but it is apparent that nonelectrolytes cannot be ignored. 8. F o r m of the V i t a m i n

It is well-recognized that vitamin B1 can occur in three forms. Free thiamine and the pyrophosphate ester (cocarbouylase) are well-known ; the occurrence of protein-bound thiamine in milk (Houston et al., 1940) and cheese (Dearden et al., 1945; Evans et al., 1946 has also been described. Greenwood et al. (1943) stated that thiamine is less stable than cocarboxylase. However, Booth (1943), Lincoln et al. (1944), Farrer (1945b), and others later, agree t h a t cocarboxylase is a good deal less stable than thiamine. Kandutsch and Baumann (1953) have referred t o the apparently greater thermal stability of protein-bound thiamine in brewers' yeast and pig kidney a s added t o laboratory diets. Thiamine mononitrate is being used in the reinforcement of certain

262

I<. T. €I. FARRER

foods, especially flour, and its thermal stability has been studied in this connection by Hollenbeck and Obermeyer (1952) who reported it t o be more stable than the hydrochloride; already studies carried out on pharmaceuticals (Taub et al., 1949; Macek et al., 1950; Bird and Shelton, 1950; Dutta et al., 1952) had been generally in favor of the mononitrate. Farrer (1953a) has compared the thermal behavior of the mononitrate and hydrochloride in buffer solutions and yeast extract solutions, and of the mononitrate and natural thiamine in processed cheese and yeast extract. Whereas the mononitrate is sensitive to pH, buffer anions, and concentration of electrolyte in the same way as the hydrochloride in buffer solutions, it is less stable under these conditions. I n yeast extract solutions, the 2 salts are destroyed at the same rate, and in cheese and yeast extract at storage temperatures differences between moninitrate and natural thiamine are scarcely discernible. Waibel, et al. (1954) found the mononitrate t o be rather more stable than the hydrochloride in certain purified chick diets. 9. Concentration of Thiamine and Cocarboxylase McIntire and Frost (1944) suggested that the concentration of thiamine is important. They found th at in unbuffered solutions (which changed in pH during heating from 6 t o about 5.2) the recovery of thiamine after 4 hr. a t 100" C. (212" F.) was 8,35, and 73 %, respectively, for solutions containing 1, 10, and 100 pg./ml. These results show a virtually linear relationship between recovery and the log of the concentration. However, a study of work published, in which rates of destruction have been determined with solutions of different concentrations u p to 5 pg./ml., and work of the present author (Farrer, 1948) show th a t the concentration of thiamine and cocarboxylase below 10 pg./ml. is virtually without effect on the rate of destruction of these compounds a t 100" C. (212" F.). At higher concentrations there is a tendency for the reaction velocity t o decrease, the exact effect of the increased concentration varying with the buffer solution. As few foodstuffs contain more than 10 pg./ml. (or per gram) one may, for all practical purposes, neglect the effect of concentration of vitamin B1 present. 10. Oxygen

Farrer and Morrison (1949) have found th a t oxygen can accelerate the thermal destruction of thiamine a t temperatures greater than 70" C. (158' F.). The effect was more marked in phosphate buffer solutions than in citric acid-phosphate or phosphate-phthalate buffers. It was absent in solutions of yeast extract. The susceptibility of thiamine to certain oxidizing agents is well known, e.g. excess potassium ferricyanide and potassium bromate, referred to above.

THE THERMAL DESTRUCTION

OF VITAMIN

nl

I N FOODS

263

11. Moisture Content

It seems clear th at the moisture content of dried products can affect vitamin B1destruction. Rice et al. (1944) and Nymon and Gortner (1948) refer t o its importance in dehydrated pork. Hollenbeck and Obermeyer (1952) recognized it in flour, and Kandutsch and Baumann (1953) have shown t ha t complete desiccation of laboratory rations can overcome losses found when they are merely “dry.” Of these factors, time, temperature, and p H are the most important and have t o be considered in every case of vitamin B1 destruction. The electrolyte system, heavy metals, nonelectrolytes, and the state of the vitamin could probably be grouped together as being not quite so important as the first group but still capable of playing a part. The predominating buffer anion in foodstuffs is probably HPO,-- although organic acids (in vegetables and fruits) and amino acids (in meats, etc.) are likely t o have a modifying effect on the rate of destruction of the vitamin. Under certain circumstances, as the author has found, heavy metals can have a profound influence on the rate constant for thiamine breakdown, viz., in yeast extract solutions (Farrer, unpublished data), but normally it is very doubtful whether they would play a significant part. The total concentration of heavy metals is very small in foodstuffs, and it is most unlikely that all of any one metal present is in the aqueous phase as free ions capable of forming complex anions, which seems t o be the necessary condition for these elements t o have an influence on thiamine destruction. It seems that some nonelectrolytes, e.g. proteins and starch, do have the power of decelerating the destruction of vitamin B,. This will be referred t o in more detail later, but it may be pointed out here that, a s with the other three factors grouped with this one, the actual conditions prevailing are likely t o be relatively constant in any one group of products (e.g. in meat, cereals, or vegetables). Cocarboxylase is less stable to heat than thiamine, protein-bound thiamine possibly more so, but the actual proportion of each form present varies little within the various food groups; e.g. Melnick and Field (1939) found 89% of vitamin B1 in wheat germ t o be in the free form, Slater (1941) found 91 %, whereas Myrback and Vallin (1944) found 75% of the thiamine in pork to be free and 25% phosphorylated, a proportion generally confirmed b y the present author. Under the conditions found in foodstuffs, variations in the concentration of electrolytes or of thiamine are not likely t o be very important except in the case of the actual concentration of a solution to yield a paste or a solution of greater total solids content. Under such circumstances, as the author has found with yeast extracts, the rate of destruction of the vitamin is likely t o increase as the product becomes more concentrated

264

K. T. H . FARRER

[cf. the influence of higher roncentrations of buffer salts (Farrer, 1947b)l. On the other hand, there will be a tendency for the increasing concentration of the thiamine itself to retard its own rate of destruction. One may say that, under some conditions, the presence of oxidizing agents may accelerate thiamine decomposition but th a t the effect of atmospheric oxygen will be very small, if it is there a t all, except possibly in dehydrated products. It is the author’s belief th at when BI is destroyed thermally in food processing, this destruction may be ascribed t o these factors, and that the results of any study can be and indeed are best expressed by the application of simple reaction kinetics. Although most of the work done on this subject is useless so far as adding t o our knowledge of fundamentals is concerned, a few papers have used kinetic methods, and from others it is possible to calculate rate constants2 and t o draw certain conclusions therefrom.

111. THERMAL LOSSESIN CEREALS I n attempting to correlate information on the loss of thiamine in processing foodstuffs with fundamental studies on thiamine destruction, theeleast complicated picture would appear t o be th a t given by studies on cereals. Work published may be divided into five main sections as follows: (1) Loss during baking (or toasting) of bread. (2) Stability of vitamin B1 from different sources during baking. (3) Loss during the cooking of breakfast cereals. (4) Effect of various baking powders on stability in baked products. ( 5 ) Losses in other cereals. i. Loss during Baking (or Toasting) of Bread

Relevant data are set out in Table I and do not require detailed treatment. Generally speaking, the figures clearly show the influence of the two well-recognized factors, time and temperature, on the loss of vitamin B1 [as do the figures of Down and Meckel (19-23) which show increasing losses with increasing toasting time]. However, it would be misleading t o attempt t o calculate reaction rates for the thermal destruction of vitamin B1under the conditions of dough making because of the high temperature gradients existing. So far as this is concerned, several authors refer to the high losses in the crust and the corresponding low losses in the crumb; Heupke and Kittelman (1943) report up t o 90% loss from the crust, but Vogel (1944) strongly criticized their results. 1 Wherever

k, the rate constant, is referred to in this paper, the unit used is rnia-1.

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

265

TABLE I Loss of Vitamin B, in Baking Bread

Product Regular white bread High vitamin B1 bread \Vhole wheat bread Whole wheat bread White bread White bread White bread enriched with thiamine White bread enriched with high B1 yeast 85 % wheatmeal bread 98% wholemeal bread 100 % wholemeal bread Hovis (wheat germ) bread JVhite bread enriched with high B1 yeast White bread enriched with thiamine JVhole wheat bread LVhite bread, l>ilb. loaf LVhite bread, 10 lb. loaf 1% hite bread, 10 lb. loaf JVhite bread, 4 lb. loaf JVhite bread, 4 Ib. loaf British 85% extraction bread British 80% extraction bread South African 95% extraction bread Canned Bread White bread

Baking time

Baking temperature

Per cent loss

Reference

8 Hoffmann et al., 1940

-

-

45 min. 20 min. 30 min.

190-218" C. 425" F. 475" F.

5 9 15 6-8 29

45-48 min.

510-520" F.

22

Dawson and Martin, 1942

45-48 min.

510-520" F.

3-14

Dawson and Martin, 1941

45-48 min.

510-520" F.

27

Dawson and Martin, 1942

45-48 min.

510-520" F.

33

Dawson and Martin, 1942

45-48 min.

510-520" F.

35 Dawson and Martin, 1942

45-48 min.

510-520" F.

19 Dawson and Martin, 1942

commercial conditions" "commercial conditions " "commercial conditions" 35 min. 440" F. 44 min.

440" F.

80 min.

440-360" F.

75 min.

440-360" F.

80 min.

475-360" F.

22

IIoffmann et al., 1940 Hoffmann et al., 1940 Aughey and Daniell, 1940 Harrel et al., 1941 Harrel et al., 1941

Schultz et al., 1942

21 Schultz et al., 1942 26 Schultz et al., 1942 17 Meckel and Anderson, 1945 14 Meckel and Anderson, 1945 21 Meckel and Anderson, 1945 19 Meckel and Anderson, 1945 24 Meckel and Anderson, 1945 16 Goldberg and Thorp, 1946

15 Goldberg and Thorp, 1046

-

-

30 min.

400" F.

29 Goldberg and Thorp, 1946 15 Brenner et al., 194% 17 Zaehringer and Personius, 1949

266

K. T.

n.

FARRER

TABLEI (Continued)

Product

'

Baking time

Baking temperature

White bread

40 min.

400" F.

White bread

50 min.

400" F.

Bread rolls (various)

15 min.

400" F.

Bread rolls (various)

20 min.

400" F.

Bread rolls (various)

25 min.

100" F.

Wheat bread Rye bread Hard Tack

30 min. 45 min. 20 min.

225-40" F. 240-60" F. 230-80" F.

Per cent loss

Reference

Zaehringer and Personius, 1949 26 Zaehringer and Personius, 1949 7-1 2 Zaehringer and Personius, 1949 8-16 Zaehringer and Personius, 1949 12-22 Zaehringer and Personius, 1949 11-20 Pulkki et al., 1950 15-25 Pulkki et al., 1950 11-15 Pulkki et al., 1950 20

Pelshenke and Schulz (1951) found only small losses on heat sterilization of bread, but Bukin et al. (1953) report 30% loss on baking rye bread, and 12-200/, loss for wheat bread. The average over-all loss to be expected in baking a loaf would appear to be of the order of rather less than 20% whatever the actual conditions used. However, Dawson and Martin (1941, 1942) have shown that, in baking, high-extraction flours show greater loss of vitamin Bl than normal white flours. Goldberg and Thorp (1946) also found greater destruction in baking doughs from high-extraction flours but attributed it to the longer baking time required. This explanation does not hold, however, a s Dawson and Martin's experiments were carried out under constant conditions. These authors suggest two possible explanations for these different rates of destruction of which the first was pH. However, they could find little difference in the p H of the different breads (the doughs were not tested), and the second suggestion was that, in the fermentation, the yeast absorbed free thiamine from the dough, and protected it in the subsequent baking (quod vide). A likely explanation for this greater destruction with higher-extraction flours is t o be found in the mineral contents of the flours used, for, if one plots percentage ash in the flours used by Dawson and Martin (Zoc. cit.) against the percentage destruction of vitamin Bl, one obtains a linear relationship. The present author has repeated this with Australian flours (Farrer, 1949a) and obtained the same relationship, but with this difference, that, when the p H was measured (with the glass electrode), it was

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

267

found t o increase with the percentage extraction from p H 5.68 with 75% extraction t o p H 5.98 with 100% extraction. However, in view of the slight increase in k between p H 5 and 6 found in phosphate buffer solutions (Farrer, 1945a) (phosphate seems t o be the predominating anion in cereals, as, indeed, in most foods), it seems likely t h a t this increase in destruction is not solely a p H effect, but may also be related t o the increase in inorganic constituents. This may be hard t o conceive, for even in white flours the minerals present are very greatly in excess of the vitamin BI, but, although the latter increases a t a greater rate than the mineral content as the percentage extraction of the flour rises, the increase in the mineral concentration in the aqueous phase is likely t o play a much more important part in influencing vitamin B1 destruction than the increase in concentration of vitamin B1 in the aqueous phase. The only other explanation for the phenomenon discussed would appear t o require a n increase in the amount of the more vulnerable cocarboxylase in the higher extraction flours. This is discounted by Booth (1940), who stated t h a t (‘wheat contains little or no protein-bound vitamin B1 or cocarboxylase,” and by other workers since then who agree on the virtual absence of cocarboxylase (Shetlar and Lyman, 1941; Obermeyer et al., 1944; and Teysseire, 1945). The percentages of thiamine as cocarboxylase in the 4 samples of flour examined by the present author (Farrer, 1949a) were 8.2, 12.9, 11.3, and 10% in the 74.9,85,98, and 100% extraction flours, respectively. More recently, van der Mijll Dekker (1952) has taken a different view. He carried out a series of baking experiments with flours of 100, 85, and 75% extractions, and, contrary t o the findings of most other workers, found smallest losses with highest extraction. He claims that the loss of thiamine during bread manufacture occurs during the fermentation of the dough and scarcely a t all during baking. Further, he claims t h a t the losses, which may be from 0 t o 60% of the thiamine present, depend on the brand of yeast used, being linked with the intensity of the yeast metabolism. It should be noted that the assays were carried out with the diazotization procedure. If the thiochrome method were used, scarcely any loss of thiamine was noted a t the fermentation stage. van der Mijll Dekker himself raises the question of the different assay procedures and postulates the conversion of thiamine during fermentation t o a form which cannot be determined by the diazotization method, but which may be oxidized to thiochrome. Baking, he says, then converts this compound t o a form which cannot yield thiochrome. His argument is based on the premise that “it is improbable that vitamin B1 is destroyed during the

268

K. T. H. FARRER

baking process, in which the temperature inside the bread does not exceed 98” C.” (208’ F.). On the basis, not merely of observed losses during baking (as determined by thiochrome, yeast fermentation, and rat growth methods), but also of what is now known of the conditions governing the thermal destruction of thiamine, it is more likely t h a t the reverse is the case. Furthermore, it appears from the literature, that the diazotization method for thiamine is subject t o greater interference (Gyorgy and Rubin, 1950) and is less sensitive (Association of Vitamin Chemists, 1951) than the thiochrome procedure. It is therefore the present author’s view th a t thiamine losses of the order indicated above (i.e. about 20%) may be expected when bread is baked. 2. Stability of V i t a m i n B1 j r a m Diflerent Sources during Baking

Some attention has been given to the relative stability in cereals of vitamin B, derived from various sources. One would, of course, expect thiamine itself t o behave in the same way whether it is derived “naturally,” or is added as a synthetic compound. Williams and Fieger (1944) found baking losses in scones of 18.3 and 18.1% when enrichment was by synthetic thiamine and rice polishings (“natural” vitamin Bl) , respectively. The only way in which the source of the vitamin could affect the stability would be to vary the form of the thiamine. It has already been mentioned that there is no protein-bound thiamine (Booth, 1940) and little cocarboxylase (Obermeyer et al., 1944; Teysseire, 1945). Dawson and Martin (1941) claim that vitamin B, added as high-vitamin yeast is much more stable than that added as synthetic thiamine. Petit et al. (1945) on a priori grounds, suggest the reverse. It is well known th at vitamin B1 occurs in yeast in the phosphorylated form and would therefore be expected to be less stable. However, how much of it is in the protein-bound form is not clear. Kandutsch and Baumann (1953) seem t o think th at the vitamin B1of yeast is less vulnerable in laboratory rations than the pure compound. There is one point which, although probably of little significance, has been neglected in considering this matter, viz. that the p H inside the yeast cell is almost certainly lower than that of the surrounding dough. The p H of bakers’ yeast cell contents is 4.6 t o 5.0. Values of this order have been obtained repeatedly in the present author’s laboratory where it has, moreover, been demonstrated th at i t is very difficult indeed to shift the p H of the cell contents by immersing the cells in a medium of different pH (Farrer, 1 9 5 3 ~ ) .This condition would tend t o retard the destruction of the more vulnerable cocarboxylase, but even so it is almost certain that this form of the vitamin in the yeast would be destroyed

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

269

more rapidly than the free thiamine in the aqueous phase of the dough. Support is lent t o this belief by the statement of Petit (1950) that, in the course of breadmaking, part of the thiamine is converted into cocarboxylase (in accordance with the known tendency of yeast t o collect thiamine), and that during baking, this cocarboxylase is destroyed more readily than the thiamine in the aqueous phase of the dough. T o balance this effect, there is some evidence for the existence in yeast cells of proteinbound vitamin B1 which is probably less sensitive t o heat than thiamine (cf. Kandutsch and Baumann, Zoc. cit.). T h a t the various factors operating may balance out is apparent from the report of Schulta, Atkin and Frey (1942) who found virtually the same destruction of vitamin B1 in breads baked under idential conditions from whole wheat flour (26%), white flour plus high-vitamin yeast (22%), and white flour plus synthetic thiamine (21%). [The higher destruction in the whole wheat flour confirms the observations of Dawson and Martin (1942), Goldberg and Thorp (1946) and Farrer (1949a) already referred to.] Similarly, Brenner et al. (1948a) have reported that thiamine added either as the synthetic or in the form of dried yeast is decreased by 15% in baking canned bread. Mention has already been made of the claim that thiamine mononitrate is more stable in flour and bread than the hydrochloride (see Merck Pamphlet L-1739 and Hollenbeck and Obermeyer, 1952). 3. Loss during the Cooking of Breakfast Cereals

Some attention has been paid t o losses on cooking cereals but without providing any specific information which could assist in any understanding of the fundamentals involved. Thus Hanning (1941) discussed wheat and commercial cereals, Slater and Rial (1942) mentioned rolled oats, and Eklund and Goddard (1945) were concerned with breakfast cereals of various pH values. All reports record excellent retention of thiamine. Fortunately, more complete data are available. Lincoln et al. (1944) and Cooperman and Elvehjem (1945) have studied the losses sustained in cracked wheat, farina, and oatmeal on boiling for various lengths of time, and it is possible t o calculate from their results approximate values of Ic (the rate constant) for the thermal destruction of vitamin B1. The p H of the farina, which was enriched with thiamine, was altered by adding varying amounts of sodium monohydrogen phosphate. I n Fig. 1 the values of log k (calculated from the data of the above authors) are plotted against pH, and compared with those for thiamine destruction in phosphate buffer (Farrer, 1945a). It is apparent t h a t the course of the destruction of thiamine in these cereals is essentially that found in phosphate buffer solutions, but the slower rate

270

K. T. H. FARRER

0

Farina

0 Wheat A Oatmeal

x Phosphate buffer

9:

g

-2.0

J

-3.01

5

7

6

PH FIG. 1. pH/log k curves for vitamin BI destruction in phosphate solutions and cereal preparations (from Lincoln et al., 1944; Cooperman and Elvehjem, 1945).

in the cereals tends to confirm the statement of Atkin et aZ. (1943) that starch decelerates the destruction of vitamin B,. The high value for wheat is very likely due to experimental error. Lincoln et al. (Zoc. n't.) also investigated the destruction of cocarboxylase in enriched farina a t various p H values, and again it was possible to calculate approximate values of k and thence of A log k (Farrer, 194513) which is the difference between log k for thiamine and log k for cocarboxylase under the same conditions (see Table 11). TABLE I1 Log k of Thiamine and Cocarboxylase a t Various pH Values pH 5.8 6.6

6.9 7.2 7.5

Log k thiamine -

3.17 3.42 3.75 2.07 2.47 -

Log k cocarboxylase -

3.72 2.02 2 -. 2 3 2.38 2.58 -

A Log k

0.55 0.60 0.48 0.31 0.11

If these values for A log k are compared with the curve pH/A log lc for buffer solutions (Farrer, 1945b), it is seen that the values for p H 5.8 and 6.9 are rather higher and lower, respectively, than would be expected if A log k depends only on p H and is not influenced by the medium. However, the value for p H 6.6 is quite close t o the expected value and

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

271

those of p H 7.2 and 7 . 5 show the sharp fall which seems t o be usual above p H 7 . It is, therefore, quite likely that the “abnormal” values for A log k at pH 5.8 and 6.9 are due t o cumulative errors. Subsequently, White et al. (1946) published data on losses of thiamine in cooking breakfast cereals. They did not add anything t o the over-all picture, but did underline the effect of time, and especially pH. Thus whole wheat cooked for 5 min. a t p H 6.2 lost 8.5% of its thiamine, and a sample of enriched farina a t p H 6.9 lost 33.997,. This large increase in destruction is t o be expected from the pH/log k curve for phosphate buffer (Fig. 1) which shows that a substantial increase in the velocity coefficient for thiamine destruction occurs between p H 6.2 and 6.9. On the other hand, the percentages destroyed a t p H values of 5.94 and 6.20 (where k is approximately constant, see Fig. 1) show anomalies. However, the two greatest discrepancies, viz. 14.3% loss in 2% min. and 2.6% loss in 10 min. were obtained with the samples lowest in thiamine, in which cases experimental errors are likely t o be greatest. It seems very reasonable, therefore, t o ascribe all these anomalies t o experimental errors in the thiamine assays, as the present author has encountered just such cumulative experimental errors when examining cereals.

4. E$ect

of Various Baking Powders on Stability in Baked Products

Studies on losses of vitamin B1 in the baking of scones have been concerned with the influence of the p H and the inorganic ingredients of the dough (i.e. the electrolytes). Here again, as with bread, it is not possible t o compare values of k because of differing baking conditions. However, it is possible t o consider the effects of various factors individually. Barackman (1942) has studied the recovery of thiamine added t o selfrising flour scones, the p H values of which were fixed by the addition of various amounts of calcium acid phosphate (monohydrate or anhydrous) and sodium bicarbonate. His results may all be interpreted in terms of pH, lower recoveries with the monohydrate being attributable t o a higher maximum p H during baking than that obtained with the anhydrous phosphate. Wilson (1942) has referred t o the greater loss of vitamin B1 from bread raised with tartrate baking powder than from bread leavened with yeast. Williams and Fieger (1944) investigated, inter alia, the effects of different chemical leavening agents on the thiamine content of scones baked from the same flours. Although the range of p H was very small (6.3 t o 7.1) they found t h a t the higher the pH, the lower the thiamine content of the scones, but there are not enough data t o calculate percentage recovery. Two leavening agents were phosphate powders (calcium acid phosphate and a sodium aluminum sulphate-calcium acid phosphate

272

K. T. H. FARRER

mixture), and the other two were a sodium bicarbonate-lactic acid mixture and cream of tartar. It is unfortunate that there are insufficient data t o enable comparisons of the relative effects of the different anions present t o be compared. Briant and Hutchins (1946) also studied losses in baking scones. I n their experiments with sodium aluminum sulphate-calcium acid phosphate they got much the same recovery of thiamine (80-90%) in each experiment, variations being attributable t o variations in the p H of the dough and/or finished scone, or the use of water instead of milk (resulting in a longer baking time). The lower recoveries with tartrate (72%) and phosphate (75%) are not in line with the recoveries obtained by Barackman, the phosphate recovery being particularly hard t o understand, as one would expect results similar t o those obtained with the sulphatephosphate mixture which, in general, confirm the figures obtained by Barackman. The retentions obtained with mixtures containing lactic acid, sodium bicarbonate, and baking powder, if plotted against pH, give a much steeper curve than that obtained b y plotting the same curve for phosphate powders. This would seem t o indicate a n ionic variation. Briant and Hutchins also record a significant fall in thiamine retention when the amount of baking powder is increased from 12 g. (the normal amount) t o 18 g. This is a substantial increase, especially if viewed from the point of view of the aqueous phase, and i t is not difficult t o link this with variations in the rate of destruction of thiamine brought about by variations in the concentration of buffer salts (Farrer, 1947b). Similar studies have been carried out on muffins. Arny and Hanning (1917), using a sodium aluminum sulphate-phosphate baking powder, obtained 85 % retention of thiamine but also obtained similar retentions with cream of tartar, tartrate baking powder, and calcium monophosphate. Substitution of cultured buttermilk, with its lactic acid unneutralized, gave a poorer quality product, but increased the retention of thiamine t o 93%. Briant and Klosterman (1950) have gone into the matter rather more thoroughly. With various milk mixtures, retention increases with falling pH between 7.1 and 6.8. With a constant amount of lactic acid equivalent t o that of sour milk, and varying amounts of soda and sodium aluminum sulphate-phosphate baking powder, the results shown in Table I11 were obtained. These results fit in well with the cereal curve in Fig. I ; pH 6.0 and 6.4 are on the relatively flat part of the curve, 6.9 is around the corner onto the steep part, 7.5 is associated with a much accelerated rate of destruction, and 9.1, although the phosphate is absent, is too alkaline for much thiamine t o remain. The same authors record results for different baking powders. They obtained lowest retentions with tartrate, 64-68%;

T H E THERMAL DESTRUCTION OF VITAMIX B 1 I N FOODS

273

slightly better with phosphate, 70% (even though the p H was slightly higher, 6.9 a s against 6.6); and 78% with sodium aluminum sulphatephosphate (16 g., p H 7.1). When the latter was reduced (11 g., p H 6.9), the retention rose sharply t o 88%. These results are substantially the same as those obtained by Briant and Hutchins (1946) for scones, and the same comments apply, as the effects of p H and electrolyte are clearly evident. TABLE I11 Retention of Thiamine in Muffinsa

Leavening agent Soda 0.0 g., baking Soda 1.0 g., baking Soda 2.2 g., baking Soda 3.6 g., baking Soda 7.1 g., baking a

powder powder powder powder powder

16.0 g. 14.1 g. 10.7 g. 8.0 g. 0.0 g.

PH

Thiamine retention (%)

6.0 6.4 6.9 7.5 9.1

83 84 73 46 6

From Briant and Klosterman (1950).

Kawasaki and Kani (1949) have considered thiamine decomposition by baking powder in bread, prepared by steaming small balls of dough. They examined 14 commercially available powders of unstated composition, and 16 prepared powders of known composition. The latter consisted of sodium bicarbonate with varying proportions and combinations of tartrate, tartaric acid, phosphates, and sodium, potassium, or ammonium aluminum sulphates. They showed a very fair linear relation between the percentage retention of thiamine and the p H of the bread. Comparison with the p H values of the powders themselves was less satisfactory. There were several obvious deviations from linearity, but there is not sufficient information in the paper t o permit any general conclusion t o be drawn about the influence of specific anions. The large losses of thiamine a t p H values greater than 7 are clearly evident, even in samples a t these p H values, allowed t o stand for some hours after steaming. Pace and Whitacre (1953a, 1953c) in studies of corn bread and corn muffins have linked the p H of the batter, the higher p H of the bread or muffins, and the thiamine retention. Their results confirm the well recognized increased destruction of thiamine with rising pH, and are very similar t o those of Kawasaki and Kani (1949). They show significant differences in thiamine retention with different types of baking powder which yield almost exactly the same p H in the final bread. This fits in with the idea of ionic variations in the baking powder affecting thiamine retention.

274

K. T. H. FARRER

5. Losses in Other Cereals

Only general statements appear t o be available about losses in other cereals. Fritsch (1944) has written of losses in rye and wheat foods, and some attention has been paid to rice. Swaminathan (1942) was mainly concerned with leaching, but his figures did show a larger thermal loss of thiamine from raw milled rice than from parboiled milled rice. This is not altogether unexpected, as parboiling ensures greater protection for the vitamin, both mechanically in insulating it and b y virtue of the starch effect reported by various authors. Iiik (1945) in a comprehensive study of three B vitamins in rice, considered cooking losses as well. Rice cooked in a double boiler in such a way that all the water added was absorbed, and hence no thiamine was lost by leaching, lost 4.3% of its thiamine. Leong and Strahan (1952) have reported the results of large-scale steam cooking of enriched rice. When cooked on a perforated tray, the thermal destruction was 12% of the total thiamine present. When a nonperforated tray was used, 5% mas destroyed by heating. I n this experiment, the cooking time was only about half th a t used when the perforated plate was employed, and the results obtained approximate to those of Kik (Zoc. cit.) Ahmad et al. (1948) reported th at 21-69% of the vitamin B1 was lost during the cooking of rice, but this included leaching losses. Roy (1953) showed losses of 60-70% when it was found that the water used gave p H values of 9-10 on boiling. If the water was acidified to p H 4-5, the losses fell t o 558%. Kennedy and Tsuji (1952) found considerable losses in fried and toasted rice. Pace and Whitacre (195313) confirm the effect of temperature in a study of corn loaves, muffins, and sticks baked under identical conditions. Crust formation increased with the surface area of the batch and losses were 15%, 23% and 34% respectively. It is apparent that such information is largely qualitative and points only t o the obvious causes of thiamine loss without relating them precisely t o specific conditions. To sum up, there is very little in the data available on losses of vitamin B1in processing cereals which cannot be interpreted by reference t o the work on the destruction of vitamin B, in buffer solutions. Losses sustained in baking bread are related t o the time and temperature of baking, and also possibly t o the mineral content of the flour used, increasing as the percentage increases. The form of the vitamin can be important. The differing stabilities of free thiamine, cocarboxylase, and thiamine mononitrate are apparent, but there is evidence for believing th a t in bread

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

275

baking the rate of destruction of thiamine is very similar in its various forms, the mononitrate probably being the most stable. The varying catalytic effects of different anions on vitamin B, destruction are evident, but more systematic information is required on the anions of leavening agents. Evidence available for a study of the destruction of thiamine in baking scones may also be interpreted in terms of time, temperature, pH, and the electrolyte system. Data on losses from boiling breakfast cereals fit in well with the picture of thiamine and cocarboxylase destruction obtained from studies in buffer solutions and show the influence of phosphate and of starch t o be important. Information on other cereals, e.g. rice, is scanty. IV. THERMAL LOSSESIN MEATS As early as 1936, Christensen et al. using the rat-growth assay, found losses of 12 and 20% of thiamine from pork and beef, respectively, during cooking at about 90" C . (194" F.). Autoclaving for 70 min. a t 10 lbs. pressure or 116" C . (241" F.) caused a loss of 21 % from pork but led t o an almost complete loss of thiamine from beef. Mickelson et al. (1939) also used the rat-growth assay for studies on the cooking of meat. Their figures suggest the following losses: roasting, 40-6097,; broiling and baking, 50%; stewing, about 50%; and frying, very variable, 0 4 5 %. These figures provide a very interesting comparison with the general conclusions set out below and based on the literature reviewed in Table IV. A great deal of work on the thermal losses of thiamine from meats (and vegetables, too, p o d vide) has been done in recent years. Unfortunately, most of it, although possibly of some value as a guide t o the home economist, is worthless t o anyone seeking t o arrive a t fundamentals. This objection t o the mass of practically unrelated data available is clearly brought out by a consideration of Table IV (and also Table Y which relates t o vegetables). I n the great majority of cases, information on the time and temperature of processing is either wholly or partly lacking. 0nly.about 35% of the entries in Table IV record both time and temperature conditions, and of these, several give only the final internal temperature of the meat, and the total cooking time. As in the baking of bread, there are large temperature gradients, and quantitative considerations of the data are out of the question. There is a general agreement on the loss of thiamine from similar cuts of meat although some anomalies occur. Thus, for example, different authors record losses of 23, 25, and 29% on roasting pork loins, whereas two others give the loss as 41 and 43 %. Fried ham is said by two authors t o lose 37 and SO%, respectively, by a third, 8%; and loss in roasting mutton is variously given as 40%, 42%) 39%, 32%) 22% and 12%.

276

I(.

T. H. FARRER

TABLEIV Loss of Vitamin B1 on Processing Meats

Meat and process

Time of processing (minutes)

Temperature of Per processing cent loss ("C.1

Reference

Pork

Baked chop Baked ham

-

-

-

-

Until 70" internal temp. Boiled ham (cured) Until 70" internal temp. Boiled shoulder Until 70" internal temp. Boiled shoulder Until 70" internal (cured) temp. Baked spareribs Until 70" internal temp. Braised chop 13 Boiled ham

Braised loin Braised heart (whole)

45 150

Braised heart (sliced)

120 11 45 30

90 90 90

90 90 -

96-98 internal temp. 96-99 internal temp. -

70 Lane el a / . , 1942 50 Sarett and Cheldelin, 1945 66 Eriksen and Boyden, 1947a 53 Eriksen and Boyden, 1947a 60 Eriksen and Boyden, 1947a 57 Eriksen and Boyden, 1947a 58 Eriksen and Boyden, 1947a 15 Aughey and Daniell, 1940 30 McIntire et al., 1943b 23 Grismold et al.. 1949 26 2

Broiled bacon Broiled ham Broiled loin Pressure-cooked heart Canned pork

110

Canned (6 lb.)

200

110

51

Canned (6 lb.)

200

113

57

Canned (12 0 2 . )

60

111

33

Canned ham

60

111

32

Canned loin

85

121

73

Canned ham

85

121

82

Canned shoulder

85

121

77

45

21 15-43 95-96 internal 42 temp. 121 80

Griswold et al., 1949 McIntire McIntire McIntire Griswold

et et et et

al., al., al., al.,

1944 1944 1944 1949

Arnold and Elvchjem, 1939 Reedman and Buckby. 1943 Rice and Robinson, 1944 Rice and Robinson, 1944 Rice and Robinson, 1944 Eriksen and Boyden, 1947a Eriksen and Boyden, 1947a Eriksen and Boyden, 1947a

THE THERMAL DESTRUCTION OF VITAMIN

~1

IN FOODS

277

TABLEIV (Continued)

Meat and process Canned liver

Time of processing (minutes) 85

Temperature of Per processing cent ("C.) loss 121

Reference

75 Eriksen and Boyden, 1947a 80 Sarett and Cheldelin, Fried bacon 1945 Fried chop 7, 9, 10 14, 23, 58 Campbell et al., 1946 Fried chop 28 "LON heat" 41 Hartzler et al., 1949 50 Sarett and Cheldelin, Fried ham 1945 47 Lane et al., 1942 Fried ham 8 Schweigert et al., 1943 Fried ham Fried liver 23 "LON heat" 11-24 Hartzler et al., 1949 Fried pork 10, 15, 20 163 22, 42, 51 Jackson et al., 1945 Fried ham Until "done" 53 Eriksen and Boyden, 1947a Fried ham (cured) Until "done" 55 Eriksen and Boyden, 1947a Fried shoulder Until "done" 43 Eriksen and Boyden, 1947a Until "done" ..Fried shoulder 49 Eriksen and Boyden, 1947a (cured) 30 Schweigert et al., 1943 Roast ham 180-210 32 McIntire et al., 1943b Roast ham To 84 internal 43 Aughey and Daniell, 43 Roast loin 1940 temp. Oven 177" until 85" internal 41 Westerman et al., Roast loin 1952b temp. 29 Mclntire et al., 19431, 120-150 Roast loin To 85 internal 25 Brady et al., 1944 Roast loin temp. 85, 108 To 83 internal 18,23 Campbell et al., 1946 Roast loin temp. At 149" or 177" until internal 53 Eriksen and Boyden, Roast loin 1947a temp. is 71" At 149" or 177" until internal 58 Eriksen and Boyden, Roast ham 1947a temp. is 71" 64 Eriksen and Boyden, Roast ham (cured) At 149" or 177" until internal 1947a temp. is 71" At 149" or 177" until internal 48 Eriksen and Boyden, Roast shoulder 1947a temp. is 71" At 149" or 177" until internal 62 Eriksen and Boyden, Roast shoulder 1947a temp. is 71" (cured) A t 149" or 177' until internal 47 Eriksen and Boyden, Roast spareribs 1947a temp. is 71" To 85 internal 60 Hartzler et al., 1949 Roast shoulder temn. ~

278

K. T. H. FARRER

TABLE IV (Continued)

Meat and process

Time of processing (minutes)

Temperature of Per processing cent (“C.) loss

Reference

9 45 70

55, 42, 54 68, 83, 88 63 121 51 (‘Boiling ” 50 90-5 58 121 72 “Boiling” 73 90-5 76

Lane et al., 1942 Jackson et al., 1945 Cover and Smith, 1952 Cover and Smith, 1952 Cover and Smith, 1952 Cover and Smith, 1952 Cover and Smith, 1952 Cover and Smith, 1952

Roast shoulder Roast pork Stew, browned Stew, browned Stew, browned Stew, unbrowned Stew, unbrowned Stew, unbrowned Reef Boiled plate Roast blade Roast chuck Roast rib Roast rib Roast rib

102 Oven 150”

Roast rib

Oven 260”

Roast Roast Roast Roast

11 60 110 -

102 .-

-

114 192

rib rib round round

-

Roast shoulder

94

Stew, neck Pressure-cooked round Pressure-cooked stew Broiled Patties

__

17 2 35

Broiled beef Canncd No. 2 cans Canned No. 2 cans Canned pint jar Canned pint jar No. 2 Canned quart jar No. 3 Braised beef Braised kidneys Braised heart

85 65 85 60 80-90

75 “Till tender”

-

15 1.3 37 To 71 50 To 74 10.2 To 80 internal 36-46 temp. To 80 internal 46-58 temp. Oven 150 25 Oven 150 31 55 To 82 internal 66 temp. To 74 internal 5.5 temp. 74.8 To 82 internal 57 temp. 120 46 8-14

To 74

To 58-70 internal temp. 120 116 120 121 121 To 93 internal temp.

-

177

Lane et al., 1942 Campbell et al., 1946 Lane et al., 1942 Noble et al., 1946 Campbell et al., 1946 Mayfield and Hedrick, 1949 Mayfield and Hedrick, 1949 Cover et al., 1944 Cover et al., 1944 Lane et al., 1942 Clark and van Duyne, 1949 Campbell et al., 1946

Campbell et al., 1946 Clark and van Duyne, 1949 Cover and Smith, 1948 Clark and van Duyne, 1949 20 Tucker et al., 1946

76 Cover et al., 63 Cover et al., 86 Cover et al., 65 Cover et al., 78 39

1944 1944 1944 1949

Cover et al., 1949 Tucker et al., 1946

36 Griswold et al., 1949 67 Noble and Gomez, 1951

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

279

TABLEIV (Continued)

Meat and process Fried slices Fried liver Stewed kidneys Simmered heart

Time of processing (minutes) 2 1?4 120 180

Mutton and Lamb Broiled loin

30

Broiled chops

-

Roast breast Roast leg Roast leg Roast leg

180-270 113, 130 -

Itoast neck

127

Fried lamb chops Stewed lamb Stewed lamb

120 15

Stewed lamb chops Veal Braised rib Braised loin Braised heart Roast leg Roast shoulder Stewed chuck Stewed veal Simmered heart Cutlets” Braised sweetbreads Parboiled sweetbreads Parboiled fried sweetbreads Parboiled creamed sweetbreads

b’ -

10, 10, 12 60 60 l 1 Till tender”

200-300 140-250 -

120 120 14 -

~

-

Temperature of Per processing cent (“CJ loss

120-150 “Simmer” “Simmer”

10 11 47 58

Reference Tucker et al., 1946 Meyer et al., 1947 Griswold et al., 1949 Noble and Gomea, 1951

To 77 internal 20 McIntire et al., l943a temp. 110-130 32-37 Wilcox and Galloway, 1952 40 Lane et al., 1942 42-39 BlcIntire el al., 1943a 22,12 Campbell et al., 1946 To 82 internal 37-51 Wilcox and Galloway, temp. 1952 To 80 internal 32 Campbell et al., 1946 temp. 50 Lehrer et al., 1952 85 49 McIntire et al., 1943a 121 50 Wilcox and Galloway, 1952 16, 10, 23 Campbell et al., 1946 ~

~~

177 85

-

To 90 internal temp. To 90 internal temp. To 90 internal temp. To 90 internal temp.

45 McIntire et al., l943a 46 McIntire et al., l943a 56 Noble and Gomea, 1951 40 McIntire et al., 1943a 32 McIntire et al., 1943a 65 Lane et al., 1942 50 McIntire et al., 1943a 53 Noble and Gomez, 1951 67 Campbell et al., 1946 52 Griswold et al., 1947 58 Griswold et al., 1947 72

Griswold et al., 1947

59

Griswold et al., 1947

280

K. T. H. FARRER

TABLEIV (Continued)

Meat and process

Poultry Braised chicken Roast chicken Roast turkey Roast turkey (breast) Roast turkey (skin, leg) Roast turkey (white meat) Roast turkey (dark meat) Stewed fowl Broiled chicken Canned chicken (No. 2 cans) Canned chicken (after roasting) Canned chicken (after curing) Fish Baked oysters Baked halibut Fripd flounder

Time of processing (minutes)

Temperature of Per processing cent ("(2.) loss

22 80 105

-

180-240

168

180-240 180 180 155 70 85 75

-

Reference

78 Campbell et al., 1946 45 Campbell et al., 1946 30 Campbell et al., 1946 38-43

Cook et al., 1949

168 67-82 Cook et al., 1949 135, to 88 55 Eriksen and Boyden, internal temp. 1947b 48 Eriksen and Boyden, 135, to 88 internal temp. 1947b 76 Campbell et al., 1946 Up to 85 20-40 Morgan et al., 1949 internal temp. 121 67 Millares and Fellers, 1949 121 55 Millares and Fellers, 1949 121 70 Millares and Fellers, 1949

-

-

-

-

-

-

40 Marks and Nilson, 1946 45 Lane et al., 1942 40 Lane el al., 1942

* Method followed was t h a t set out by The Committee on Preparation Factors, National Co-operative Meal Investigation (Meat Cookery Book), National Live Stock and Meat Board, Chicago, Ill, 1942.

T o draw any firm conclusions is impossible. Trends only can be noted. Canning offers relatively constant conditions, and of those that are comparable [85 min., 121' C. (250' F.)] the mean of the losses is 78% (S.D., 1k4.4). Roasting will usually lead t o a thermal loss of 40-60%, and stewing t o a t least 50%, with losses as high as 75% recorded. Braising will on the average give a loss of 40-45%, going as high as 78% (recorded for turkey), but pork, as recorded by three different authors, seems t o be able t o resist this process better than other meat, the loss being only 20-30%. One cannot even make a general statement on frying, losses ranging from 8 t o 80%. Procedures are subject t o such wide variation as are other factors, such as thickness of section of the meat. A rough sort of guide for broiling is the expectation of a 20-40% loss. The figures I

THE THERMAL

DESTRUCTION

OF VITAMIN

B~ IN FOODS

281

quoted can only be the broadest possible indication of what has been observed. However, the well-known general effects of time and temperature are confirmed, viz. shorter cooking time and lower cooking temperatures favor vitamin B1 retention, although a combination of high temperature, short time may lead t o smaller losses than lower temperatures for longer times. This aspect has been discussed by Feaster et al. (1948) with particular reference t o vegetables. I n 1943 Elvehjem and Pavcek published a table summarizing the then scanty information on the cooking losses of thiamine from foods. They rightly referred t o the danger of attaching too much significance t o average figures arrived a t without due regard t o cooking conditions, but it is interesting t o compare the general conclusions drawn above from Table IV with the values quoted by Elvehjem and Pavcek for the loss of thiamine from meats, via. boiling, 15%; broiling, 25%; braising, 45%; roasting, 35 % ; and stewing 43-65 %. For a critical study of the reaction rates of thiamine destruction in meat, there are only five papers available and four deal with pork. This is not surprising as pork is by far the richest meat source of the vitamin. It would appear from the work of Myrback and Vallin (1944) that the vitamin B1 in pork is made up of something of the order of 75% free thiamine and 25% cocarboxylase; there is no bound thiamine. Although the proportion of cocarboxylase is variable, this distribution has been confirmed in the present author's laboratory, as has the absence of bound thiamine as shown by trichloroacetic acid extraction of the meat. Greenwood et al. (1913) studied the loss of vitamin B1 on heating lean pork muscle at pH 6 in the presence of substances used in meat curing, and recorded the percentage losses after 1hr. a t 98" C. (208" F.). Approximate values of lc (the rate constant) have been calculated from these single readings and are found t o vary between 0.0025 and 0.0033, although the average is 0.0030. The deviations from the mean are readily attributable t o experimental error, and i t seems evident that the three salts added, sodium chloride, sodium nitrite, and sodium nitrate, are without effect in the presence of protein. Greenwood et al. (1944) have made a n attempt t o study systematically the rate of destruction of vitamin B1 in meat under controlled conditions but without calculating the rate constant. They showed conclusively that, as one would expect, there exists a vitamin gradient from the outside of a processed can where the heat treatment was least severe and the vitamin concentration is highest. This underlines the difficulty of obtaining precise data on the thermal destruction of vitamin B1 in normally processed meats. However, these authors, by using very small cans (2>$ in. x 3/4 in.) were able t o obtain very rapid heating and cooling of

282

K . T. H. FARRER

the meat, and so studied the rate of destruction of vitamin B1 in pork a t temperatures of 99, 110, 118.5, and 126.5" C. (i.e. 210, 230, 245, and 260" F.). The rate constants, calculated from the data presented by the methods previously used by the present author (Farrer, 1945a), are tabulated (Table V). TABLEV Rate of Destruction of Vitamin BI ( k ) in Pork"

,, From

Temperature, ("C.)

k

99.0 (210" F.) 110.0 (230" F.) 118.5 (245' F.) 126.5 (260" F.)

,0025 ,0056 ,011 .022

Greenwood et al. (1944).

The value of k a t 99" C. (210" F.) compares very favorably with that calculated from Greenwood et al. (1943). The lower value found in this case is more likely t o be correct as it is based on observations at several different time intervals and, in addition, the earlier result is likely to be higher because of abnormally rapid destruction frequently met with in early stages of heating in thiamine solutions and of which there is evidence in the data of Greenwood et al. (1944). No account could be taken of this short-lived abnormality in calculating k from only one time interval, but in a series of readings, the correct destruction rate is readily assessed. Rice and Beuk (1945) determined the rate constant for thiamine destruction in pork a t 9 different temperatures from 49 t o 121" C. (120 to 250" F.). At 99" C. (210" F.), k , calculated from their data, is 0.0048. (The actual value tabulated by these authors is 80 X 10W from which it seems that they have taken t in seconds instead of minutes, as is usually the case in the application of the first order reaction equation.) This value is roughly double that of Greenwood et al. (1943). Rice and Beuk (loc. cit.) applied the Arrhenius equation t o their data which may be expressed as log, k = I - E / R T (where k = rate constant, T = temperature in degrees Absolute, and I , E and R are constants) and found that the curve log k 1/T was a straight line down t o 77" C . (171" F.). There are sufficient data in the paper of Greenwood et al. (1944) t o permit the application of the Arrhenius equation to this work also, and in Fig. 2, log k is plotted against l / T for the work of Rice and Beuk (the values of k are calculated from time in minutes, not seconds) and Greenwood et al. Two curves result,

T H E THERMAL DESTRUCTION O F VITAMIN B1 IN FOODS

283

from which it is apparent that vitamin B1 is destroyed more rapidly at each temperature in the work of Rice and Beuk than in that of Greenwood et al. The reason for this is not apparent. It is most unlikely that pH differences were such as t o cause the different values of k , but it could be t h a t the pork used by Rice and Beuk contained a higher proportion of cocarboxylase than that used by Greenwood et al. However, the two curves are approximately parallel, i.e. the temperature coefficient is approximately the same in each case, as would be expected.

Greenwood el 01. (1944) hlyrback & Vallin ( 1 9 4 4 ) x Rice & Beuk ( 1 9 4 5 )

0

-3.OL 0.00250

, 0.00260

I

0,00270

%\

0.00280

l/T' abs.

FIG.2. Effect of temperature on log k for thiamine in pork.

The only direct information permitting the application of this reasoning t o other meats is that supplied by Myrback and Vallin (1944) and Cover and Dilsaver (1917). Myrback and Vallin examined beef before and after processing for 2 hr. a t 115" C. (239" F.) and found a loss of vitamin B1 similar t o that found in pork. Cover and Dilsaver compared average values obtained for losses in stewing beef and lamb (12 samples of each) with losses obtained by Rice and Beuk (1945) in their examination of pork. Their table, however, records varying cooking times for cach meat, and the present author has plotted the logarithm of the percentage retention [in fact, log (a - .)] against time (Fig. 3) for the 3 temperatures mentioned, indicating which points are for pork, beef, and Iamb. As the curves must be linear, i t is very easy t o see that the values for pork and beef are in line a t each temperature (thus confirming Myrback and Vallin's finding), but that those for lamb are somewhat lower than would be anticipated. Nor is the explanation for this t o be found in the cooking procedures which are described by Cover et al. (1947) and

284

K. T. H. FARRER

which would, if anything, tend to favor a more rapid destruction in the lamb. This discrepancy also can most logically be attributed to the state of combination of the vitamin, as the variations in chemical composition, etc., from one type of meat to another are such as to make it very likely that observations on one type of meat are applicable t o another. If this is accepted, it simplifies very greatly the fundamental considerations underlying the thermal destruction of vitamin B1 in meats reducing them to t'ime, temperature, and state of the vitamin as, indeed, the other factors remain constant.

121°C. 1.6

I

0

60

I

I

120 180 Time, min.

240

FIG.3. log ( a - z ) / ' t curve for beef, pork, and lamb (from results of Rice and Beuk (1945);Cover and Dilsaver, 1947).

At least one attempt has been made to make use of the reported protective effect of starch on thiamine. Cover and Smith (1948) found a loss of 46% of the thiamine in a beef stew but only 8% from potatoes. When, however, they added potatoes to a beef stew and boiled for 15 min., they could find only the thiamine which could be calculated from the known behavior of the separate ingredients. This is scarcely surprising, as the meat had been pressure cooked for 17 min. a t 15 lbs. pressure before the addition of the potatoes. One may say, then, that most of the data available on losses on cooking meats are indefinite, lacking precise information regarding time and temperature, and taking no account of temperature gradients. The data which do permit a systematic study of vitamin B1 losses in meats point to a similar set of conditions in each case, with the main variable likely to be the proportion of cocarboxylase present. This generalization applies particularly to pork and beef, but less definitely to lamb.

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

285

V. LOSSESIN PROCESSING VEGETABLES

A large amount of information on losses of vitamin B1 in vegetables during processing has been published. It relates both t o thermal destruction and t o leaching of the vitamin into the cooking water. Only the former is considered in this report, and it must be recorded t h a t not only is by far the greater part of the information useless so far as study of the fundamentals of thiamine destruction are concerned, but that many of the data are mutually contradictory. Available figures are summarized in Table VI. Only the most general conclusions can be shown regarding the effect of the different forms of processing. Boiling, steaming, simmering, i.e. any treatment a t or near 100" C. (212" F.) rarely causes a greater loss of thiamine than 20%. Much smaller losses are frequently found. Losses on canning are of the order of 40% where the conventional procedures are used. High-temperature, short-time methods, as are described by Feaster et al. (1948), will reduce such losses t o less than 20% as will agitation, leading t o shorter processing times a t normal temperatures. Pressure cooking gives results similar t o those obtained in canning, although McIntosh and Jones (1947) found no destruction (within experimental error) in 6 vegetables pressure cooked a t pressures of from 0 t o 20 p.s.i. The losses on baking are analogous t o those found on boiling, probably because the comparatively low internal temperatures rarely exceed 95" C. (203" F.). Stevens and Fenton (1951) examined the comparative effects of dielectric and stew-pan cooking of peas on thiamine retention and found no differences. The numerous exceptions t o these general conclusions may be att,ributahle to the following factors which do not seem t o have been taken into account: (1) Variations (of pH etc.) between varieties, and within the same variety. This is not a likely source of error because of the relative constancy of composition of vegetables, etc. ( 2 ) Variations in the size of the pieces cooked, leading to variations in (a) rate of heat penetration, (b) penetration of cooking water, (c) leaching away of vitamin from "native tissues" thus possibly making it more vulnerable, (d) leaching away of electrolytes, etc. (3) Variations in strength and permeability of cellulose envelope of peas, beans, etc., which must cause the same sort of variations as are listed above. Some support for the belief in the importance of these factors is found in the work of Ashikaga (1916) on soybeans and green peas, the report of which has only just become available.

286

K. T. H. FARRER

TABLEVI Loss of Vitamin BI on Processing Vegetables

Vegetable

Method of processing

TemTime peraof ture Loss proco€ of essing proc- thi(min- essing amine utes) ("C.) (%)

Reference

45 70 0.63 90 30 55 0.63 30 20

18 Fellers et al., 1940 36 Clifcorn and Heberlein, 194413 Nil Gleim et al., 1944 100 100 13, nil Hinman et al., 1944 100 2 1 , s Connolly et al., 1947 18 Aughey and Daniell, 100 1940 100 Nil Aughey and Daniell, 1940 100 59 Aughey and Daniell, 1940 Nil Aughey and Daniell, 100 1940 115.5 40 Fellers et al., 1940 115.5 38 Feaster et al., 1948 140 14 Feaster et al., 1948 . 204 9, 12 Hinman et al., 1944 100 8 Hinman et al., 1944 115.5 41 Feaster et al., 1948 140 22 Feaster et al., 1948 100 Nil, 12 Hinman et al., 1944 115.5 46 Guerrant et al., 1946

35

115.5

33 Guerrant et al., 1946

20

115.5

24

30

115.5 2 9 4 2

Beans, Lima

Canned, (No. 2) Boiled

25

100

22

Beans, Lima

Canned

45

116

38

Beans, Lima

Bottled

45

116

44

Beans, Lima Beans, Green

Canned Canned

-

-

-

-

78 38

Asparagus Asparagus

Canned Canned

23 14

Asparagus Asparagus Asparagus Beans, snap

Boiled Boiled Boiled Boiled, (PH 5.8) Boiled, (PH 5.7) Boiled, (PH 6.6) Boiled, (PH 6.0) Canned Canned Canned Baked Boiled Canned Canned Boiled Canned, (No. 2) Canned, (No. 10) Canned

8 30 15 40

Beans, snap Beans, snap Beans, snap Beans, Lima Beans, Lima puree Beans, Lima puree Beans, Lima puree Beans, green cut Beans, green cut Beans, green cut Beans, Lima Beans, Green, Wax, Lima Beans, Green, Wax, Lima Beans, Green Beans, Lima

85 40 53

115.5 120

Clifcorn and Heberlein, 1944b Clifcorn and Heberlein, 1944b Guerrant and O'Hara, 1953 Guerrant and O'Hara, 1953 Guerrant and O'Hara, 1953 Ingalls et al., 1950 Ingalls et al. 1950.

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

287

TABLEVI ( C o n t i n u e d )

Method

of Vegetable Beans* Beans, snap Beans, green Beet (dehydrated) Beet (dehydrated) Beet Broccoli Broccoli Brussel Sprouts Cabbage Cabbage, dehydrated Cabbage, dehydrated Cabbage, fresh Cabbage, stored Carrots Carrots Carrots Carrots Carrots Carrots, dehydrated Carrots Carrots Carrots Carrots Carrots, puree Carrots, puree Cauliflower cauliflower Cauliflower Cauliflower Corn Corn Corn

processing Boiled Boiled Canned Boiled Steamed Boiled Boiled Pressure cooked Pressure cooked Pressure cooked Pressure cooked Steamed Boiled Boiled Boiled Pressure cooked Boiled Boiled Steamed Simmered Simmered Simmered Canned Boiled Canned Canned Boiled Boiled Steamed Canned Boiled Boiled Canned

TemTime peraof ture Loss of of processing proc- thi(min- essing amine Utes) ("C.) (%)

Reference

100 2 Connolly et al., 1947 100 41, 71 Lane et al., 1942 38 Ingalls et al., 1950 100 Nil Fenton et al., 1943 100 Nil Fenton et al., 1943 80 Lane et al., 1942 100 5 Oser et al., 1943 15 100 5-13 100-121 Nil Trefethen et al., 1951

20 25 25

20

100-121

-

-

10

-

25

Connolly et al., 1947

17 Lane et al., 1942 Nil

Fenton et al., 1943

Nil Fenton et at., 1943 27 Connolly el al., 1947 5 Connolly et al., 1947 Nil Aughey and Daniell, 1940 Nil Aughey and Daniell, 121 13 1940 33 Lane ec al., 1942 100 30 5 Hinman et al., 1944 100 18 Gleim et al., 1944 15 100 10 Gleim et al., 1946 30 25 37 Gleim et al., 1946 -~ 20 Nil Gleim et al., 1946 115.5 Nil Guerrant et al., 1946 25 15-20 100 Nil Connolly et al., 1947 23 Feaster et al., 1948 115.5 60 0.63 140 13 Feaster et al., 1948 100 15 48 Connolly et al., 1947 12>< 100 13 Lunde el al., 1940 1235 100 17 Lunde et al., 1940 10.5 Lunde et at., 1940 110 10 Nil Barnes et al., 1943 100 10 100 22 Hinman et al., 1944 30 40-50 115-120 75-66 Guerrant et al., 1946

10 20 25 23

-

~~

288

K. T. H. FARRER

TABLEVI (Continued)

Vegetable Corn Corn Corn, cream style Corn, cream style Corn, cream style Kale Onion Onion Parsnip Parsnip Peas Peas Peas, sugar Peas, sugar Peas, sugar Peas, split Peas, split Peas, split Peas, green Peas, green Peas, green Peas, green Peas, green Peas, dried Peas, fresh Peas. frozen Peas. frozen Peas, frozen Peas, frozen Peas, frozen

Method of processing

TemTime peraof ture Loss of of processing proc- thi(min- essing amine utes) ("(3.1 (%)

10 Boiled Canned Canned 70 Canned 85 Canned 0.63 Boiled 30 Boiled 15 Boiled Boiled 10 Boiled 15 Roiled 20 Canned 35 Boiled 1235 Steamed 1234 Canned 10 Boiled 30 Boiled 15 120 min. then baked 1 Pressure cooked Canned 15 Boiled, 12 (PH 6.4) Boiled, 12 (PH 7.4) Boiled 20 Boiled 5 Boiled Boiled 8 Simmered, 6 (PH 7.7) Simmered, 4 (PH 8.8) Simmered, 6 (PH 7.7) Simmered, 4 (PH 8.7) Pressure 9 cooked

100

-

115.5 115. 5 140 100 100 100 100 100 100 115.5 100 100 110 100 375

Nil 62 68 64 2-3 Nil 11 27 3 3-11 30 40 4 16 Nil 50 27

121

32

119 100

9 9

100

22

100 100 100 100 100

13.2 2.2 48 Nil Nil

Reference Connolly et al., 1947 Ingalls et al., 1950 Feaster el al., 1948 Feaster el al., 1948 Feaster e l al., 1948 Connolly et al., 1947 Connolly et al., 1947 Lane et al., 1942 Connolly et al., 1947 Connolly et al., 1947 Connolly et aZ., 1947 Fellers et al., 1940 Lunde et al., 1940 Lunde et nl., 1940 Lunde et al., 1940 Murray, 1948 Murray, 1948 Murray, 1948 Lunde et al., 1940 Aughey and Daniell, 1940 Aughey and Daniell, 1940 Ashikaga, 1946 Ashikaga, 1946 Lane et al., 1942 Barnes et al., 1943 Johnston et aZ., 1943

100

8 . 5 Johnston et al., 1943

100

3 Johnston et al., 1943

100

38 Johnston et al., 1943

100-121

2-7

Trefcthen and Fenton, 1951

T H E THERMAL DESTRUCTION OF VITAMIN B1 I N FOODS

289

TABLEVI (Continued) Temperaof ture Loss of of processing proc- thi(min- essing amine Utes) ("C.) (%) Time

Vegetable l'cas, fresh

hlethod of processing

Peas " Peas, Alaska

Simmered, (PH 7.3) Simmered, (PH 8.8) Boiled, (pH 7.5) Boiled, (PH 9.4) Boiled Canned

Peas, sweet Peas, sweet Peas, sweet

Peas, fresh Peas, fresh

Peas, fresh

17

100

Nil

Johnston et al.. 1943

8

100

Nil

Johnston et al., 1943

-

45 Johnston et al., 1943 -

60 Johnston et al., 1943

15 35

100 115.5

6 37

Canned Canned

35-40 35

115.5 115.5

39 33

Canned

45

120.5

40

35 35 70 0.63 25

120 42 120 35 115.5 21-37 13-14 140 46-50 7 100

Peas, sweet Canned Peas, sweet (immature) Canned Peas, puree Canned Peas, puree Canned Peas, puree Canned Peas, green Boiled

Reference

Peas, green

Canned

35

116

45

Pens, grern

Bottled

35

116

34

Potato Potato Potato, whole

Boiled Boiled Boiled

15-20 15 36

100

21 8 20

Potato, whole

Baked

63

190

16

Potato" Potato, dehydrated Potato, dehydrated Potato * Potato, Green Mountain Potato, Green Mountain

Boiled Boiled Steamed Steamed Steamed Steamed

15 15 20 60 90

-

-

107 70

19 Nil Nil 4 15.5 7.2

Oser et al., 1943 Clifcorn and Heberlein, 1914b Guerrant et ul., 1946 Clifcorn and Heberlein, 194413 Clifcorn and Heberlein, 1944b Heberlein et al., 1950 Heberlein et al., 1950 Feaster et al., 1948 Feaster et al., 1948 Ingalls et al., 1950 Guerrant and O'Hara, 1953 Gurrrarit and O'Hara, 1953 Gucrrarit and O'Hara, 19.53 Connolly et al., 1947 Cover and Smith, 1948 Aughey and Daniell, 1910 Aughey and Daniell, 1940 Lane et al., 1942 Fenton ez al., 1943 Fenton et al., 1913 Oser et al., 1943 Wertz and Weir, 1914 Wertz and Weir, 1944

290

K. T. H. FARRER

TABLE VI (Continued)

Vegetable

Method of processing

TemTime peraof ture Loss of of processing proc- thi(min- essing amine utes) ("C.) (%)

Pumpkin, fresh Pumpkin, stored Rutabagas, dehydrated Rutabagas, dehydrated Spinach

Boiled Boiled Boiled Steamed Boiled

15 20 25 45 7

100 100 100 100

Nil 49 Nil Nil 22

Spinach Spinach Spinach Spinach Spinach Spinach Soybeans

5 5 8 30 45-60 60

100 100 100 100 100 122 121

40 Nil Nil Nil Nil 75 74

Squash Sweet potatoes Sweet potatoes

Boiled Boiled Steamed Boiled Boiled Canned Pressure cooked Boiled Baked Baked

10

100

Sweet potatoes

Boiled

Taro

Commercial > 4 hr. steaming 121 45 Pressure cooked 35 Boiled 60 Steamed 30 Boiled 30 Boiled 40-60 Canned 110-130 Canned Boiled 15-20 Boiled

Taro Taro Taro, leaves Tomatoes Tomatoes Tomatoes Tomatoes Turnips Turnips

* Variety not given.

-

-

Reference Connolly et al., 1947 Connolly et aZ., 1947 Fenton et al., 1943 Fenton et al., 1943 Aughey and Daniell, 1940 Lane et al., 1942 Cutlar et al., 1944 Cutlar et al., 1944 Gleim et aZ., 1944 Hinman et al., 1944 Guerrant et al., 1946 Miller, 1945

11 Connolly et al., 1947 75 Lane et al., 1942 25 Pearson and Luecke, 1945 8 Pearson and Luecke, 1945 40-60 Miller et al., 1952

20-40 11 20 29 6 Nil Nil 5.5 40

Miller ef al., 1952 Miller et al., 1952 Miller et aZ., 1952 Hinman et aZ., 1944 Guerrant et al., 1946 Guerrant et al., 1946 Lane et aZ., 1942 Connolly et al., 1947

THE T H E R M A L DE S T RUCT ION O F VITA MI N B1 I N FOODS

291

Only a few of the many papers appearing have made any attempt t o consider fundamental conditions. Jackson et al. (1945), in a discussion of the effect of canning procedures on the retention of vitamins in canned products, included in their survey a consideration of the effect of the thermal processing on vitamin B1. They were able t o shorn that, in commercial practice where the first essential is t o obtain the highest degree of bacterial mortality, retention of vitamin B1,other things being equal, is favored by higher temperatures for shorter times rather than by lower temperatures for longer times. They point out that this does not hold for slow-heating products such as minced meats, but applies under conditions which permit more rapid heating of the mass of the product, viz. fruit and vegetable packs. Feaster et al. (1948) have added results obtained by the conventional canning [55-70 min. a t 115.5" C. (240" F.)] and flash sterilization [0.63 min. a t 139" C. (282" F.)] of pea, carrot, green Lima bean, and green bean purees. The average loss of thiamine in the conventional procedure was 32%. I n the high-temperature, short-time process i t was 15%. Similarly, in comparing the conventional method for corn [60 min. a t 121" C. (250" F.)] with the agitated-can technique [17 min. a t 121" C. (250" F.)] losses were much lower, as would be expected over the shorter time. Bendix et al. (1951) have discussed the factors influencing the stability of thiamine during heat sterilization. They used kinetic procedures in a study of thiamine destruction in peas, corn, tomato juice, and Lima bean puree over the temperature range 104.5 t o 132" C. (220 t o 270" F.). I n peas, the first order reaction curves are linear and the rate constant k , was calculated for 104.5, 118, 126.5, and 132" C. (220, 245, 260, and 270" F.). A straight line for the graph log k 1

ToAbs. showed that the Arrhenius equation was valid for thiamine destruction in peas. With the other vegetables these authors found a sharp initial drop in the log (a - z ) / t curve over the first 10-20 min., followed b y a much more gradual linear relation. Booth (1943) and the present author have sometimes encountered this in thiamine solutions buffered with inorganic buffers, and there appears t o be no obvious explanation for it. Experience has shown that the second part of the curve yields values of k which are comparable with those obtained when no such inflected curves are met with. However, Bendix et a1 (1951), even using log k calculated in this way, could not satisfy the Arrhenius equation for thiamine in the other

292

K . T. H . FARRER

three products, nor could they obtain a satisfactory lead from the study of mixtures buffered with either phosphate or citrate. They rightly conclude "that the rate and nature of thiamine destruction in foodstuffs is governed by the combined influence of numerous factors," and go on to say t ha t high-temperature, short-time processing will favor thiamine retention. This, of course, follows from the results recorded by Feaster et al. (1948). Farrer (1953b) has, in some preliminary experiments, obtained the following rate constants for thiamine destruction in purees a t 100" C. (212" F.): peas, 0.0021; carrots, 0.0022; cabbage, 0.0027; potatoes, 0.0020 and 0.0032. No difficulties such as Bendix et al. (1951) report were encountered, but the gel formed by the potato starch gave rise t o temperature gradients and, therefore, doubtful results. Basu and Malakar (1946) attributed variations in the destruction of thiamine in different Indian vegetables t o variations in pH, electrolyte systems, and possibly protein systems in the tissues. At the same time, they presented figures t o show that there was concurrently virtually no loss of cocarboxylase in the same vegetables, whereas thiamine losses ranged from 50 to 100%. This is completely a t variance with studies in buffer solutions (Farrer, 1915b), cereals (Lincoln et al., 1914), and meat (Myrback and Vallin, 1944) which show cocarboxylase t o be more vulnerable than thiamine. Kohman and Rugala (1949) have considered the enzymic production of thiamine in sweet potatoes during cooking, and conclude th a t the temperature optimum for this enzymic process is probably as low as 40" C. (101" F.). They refer t o a loss of thiamine with continued heating a t relatively low temperatures, but their use of the fermentometric assay limits the value of the figures presented. Ashikaga (1951) has examined the behavior of thiamine during the cooking of 10 vegetables. He noted leaching of the vitamins into the cooking water and the relative losses during steaming (smallest loss), boiling, parching, frying, and baking (largest loss, 25 %). Loss was much influenced by time and temperature of cooking. These observations are in accord with those made by a number of authors. The present author believes that the conditions which govern thiamine destruction will be found t o be very similar in all types of vegetables. If this is so, i t will greatly facilitate the prediction of thiamine losses during processing and cooking. As has already been pointed o u t by Elvehjem and Pavcek (1943), it cannot be too greatly emphasized th a t full details of conditions used should always be included with the results of thiamine retention studies. Even so, more specific information can only be expected when "homogeneous" purees are used for proper kinetic studies, as the

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

293

misleading effects of temperature gradients, size of vegetable portions, etc., are only too evident in the literature.

VI. THERMAL LOSSESIN OTHER FOODSTUFFS I . Dairy Products The loss of thiamine on the heat treatment of milk has been studied to some extent. Weckel (1938) reported loss of one-third of the thiamine in milk during commercial sterilization. Houston et al. (1940) state that 10% is lost during pasteurization, and up t o 50% on sterilization. Kon (1941) gives the following percentage losses during the processing of milk : Pasteurizing 10% Sterilizing 30 % Spray drying 10% Roller drying 15 % Condensing 40 % The figure quoted by Elvehjem (1941) for loss on pasteurization is 20%, but Holmes et al. (1943) agree with the earlier workers in quoting a loss of 9 % on pasteurization. Hitherto, attention was directed towards holding pasteurization, but Holmes et al. (1945) found a loss of not more than 3 yo during H.T.S.T. (high-temperature, short-time) pasteurization. Randoin and Perroteau (1950) have reported a loss of 23 % on pasteurization of human milk for 20 min. a t 65" C. (149' F.) and van der Mijll Dekker and Engel (1952) record a 14-30% loss of vitamin B1 on sterilizing milk, which is in general agreement with Kon (1941). Dearden and co-workers (1945) found t h a t 6-20% of the thiamine in the milk could not be accounted for in the curd and whey during the manufacture of Cheddar, Cheshire, and Stilton cheese in the winter months. I n the summer this loss was 6-7% for Stilton and vanishingly small for Cheddar and Cheshire. No further losses occurred during ripening up t o 42 weeks. The present author (Farrer, unpublished data) has shown that there is no loss of thiamine during the processing of cheese. 2. Peanuts

Pickett (1944) examined 60 samples of peanut butter and related lower thiamine contents t o the dark color which indicates more severe roasting conditions. When the temperature of the nuts reaches 147" C. (300" F.), large percentages of thiamine are lost. Fournier et al. (1949) showed that losses on roasting for 20 min. a t

294

K. T. H. FAHRER

155-160' C. (311-320" F.) were over 90%. A log (a - z)/time curve from tabulated results shows clearly the lag in thiamine destruction as the peanuts heated up over the first 10 min., but thereafter the destruction was equivalent to a rate of k = 0.152, i.e. a loss of well over 50% in 70 min. A cooling gradient is also apparent from the loss of more thiamine (about one-third of th at remaining) after roasting has ceased and while the nuts cool. Storage losses were also shown. French et al. (1951) and Willich et al. (1952) have also reported large losses of thiamine during the roasting of peanuts, the latter confirming Pickett's correlation of darkest color with greatest loss. 3. Miscellaneous

Stamberg and Peterson (1946) found an average thiamine loss of 15% on cooking eggs by any one of a number of ways. Farrer (1946) has shown th at no thiamine is lost during the autolysis of yeast for up to 24 hr. a t 50" C. (122" F.). Further work by the same author (Farrer, unpublished data) has shown that it is possible t o make yeast extract commercially without thermal loss. Farrer (19534 has further recorded the effect of pH on the rate of destruction of thiamine in yeast extract solutions a t 100" C. (212" F.). VII. THERMAL LOSSESO N STORAGE The storage of food for varying periods is commonplace. Fresh, unprocessed foodstuffs, e.g. meat, fruit, vegetables, and cheese, are sometimes held for months a t a time a t freezing temperatures, and canned products can be stored for years. I n the former, enzyme and microbiological activity still proceeds and can cause changes in vitamin potencies. Furthermore, vitamin losses can occur by release of fluids during the thawing process prior to cooking or processing. It is not with these losses th a t this discussion is primarily concerned, but, because some of them are undoubtedly thermal, some account must be given of them. I . Unprocessed Foods

Those who have studied the behavior of thiamine during storage a t low temperatures generally agree th at losses either do not occur, or are very small. Cook et al. (1949) found no significant loss from turkey tissues in 3 t o 9 months a t -23" C. (-9" F.). Morgan et al. (1949) reported similarly for chicken held frozen for 8 months. Hartzler et al. (1919) stored 16 samples of various pork cuts for 5-9 months a t -21 t o -23" C. (- 6 t o - 9" F.) and found a very slight loss of thiamine from shoulder and loin, but no loss from liver. Westerman et al. (1952a) studied the effect of storage time u p t o 72 weeks and temperatures from -29 t o

295 ' - 12" C. (-20 t o 10" F.) on thiamine in pork and concluded from their statistical analysis of the results that neither factor had any significant effect on the thiamine content. The same authors (Westerman et al., 195213) showed that there was no correlation between storage time and temperature, and the subsequent loss of thiamine on roasting. Results rather less satisfactory were obtained by Lehrer et al. (1951) who claim losses of 21% and 40% after the frozen storage of pork chops for 3 and 4 months, respectively. Chops quick frozen at -26" C. ( - 15" F.) or held a t - 18" C. (0" F.) gave substantially the same results. A similar study on lamb chops (Lehrer et al., 1952) showed larger amounts of thiamine in those held a t -26' C. (-15" F.) for 48 hr. than in those held a t -28" C. (-18" F.). There was no difference in 3 t o 6 months between those held a t the 2 different temperatures, although it was claimed that there was a loss of 1.2 pg per gram of tissue u p t o 3 months. Rice et al. (1946) showed very good retention of thiamine in pork a t 4" C. (39" F.) but much higher losses a t room temperature, accompanied by putrefaction. It is apparent t h a t in frozen storage there is a balance between three processes : (1) Slow "thermal" destruction over a long period of time. This is scarcely perceptible. (2) Thiamine destruction by enzyme and microbiological activity. (3) Thiamine synthesis by enzyme and microbiological activity. These same factors are possibly important during the storage of unprocessed foods a t atmospheric temperatures, although, under these conditions, the thermal destruction will assume greater importance. Connolly et al. (1947) studied the cool storage of certain raw vegetables for various periods from 100 t o 300 days, and obtained confusing results which nevertheless led them t o conclude that storage did not decrease the thiamine contents appreciably. Even cucurbits stored a t 14.5-19" C. (58-66" F.) for 21 weeks gave contradictory results. Ingalls et al. (1950) were concerned with the short storage of vegetables (corn, spinach, and legumes) in the shade (up t o 48 hr.), cold (up t o 120 hr.), and iced (up t o 120 hr.) prior t o canning, and their results showed that, under the conditions studied, either the precanning loss was very small, or a n apparent gain occurred. Carrots were subjected t o longer storage periods of up t o 90 days and losses of 63% (held out of doors) and 44% (cold storage) were recorded. It is obvious that under these conditions enzyme activity must also play a part. Ashikaga and Chachin (1951) report a fairly rapid loss of thiamine from orange and radish juices (17 and 22%, respectively) after 24 hr. a t 9" C . (48" F.). At 10" C. (50" F.) grated radish lost 27%, and a t 14" C. (5'7" F.) grated sweet potato lost 14.5% in the same time. These losses are T H E T H E R M S L DESTRUCTION O F VITAMIN B1 I N FOODS

296

K. T. H. FARRER 8

greater than would be normally expected, and the influence of oxygen or enzymes either separately or together seems t o be the likely explanation for the high figures, especially as the authors report that the losses from radish and sweet potato were greater when the samples were chopped finely than when chopped coarsely. According t o Caileau et al. (1945) brown rice, bran, and polishings held for 6 months a t a n average room temperature of 20" C. (68" F.), lost from 0 t o 30% of thiamine. After 24 months, the loss was 50-67%. Canned and parboiled brown and parboiled undermilled rice showed little or no loss over 6 and 3 months, respectively. Kik (1945) after storing rice a t high room temperatures for u p t o 245 years, reported a n average loss in 3 months of 5-7oJo; after 9 months, the loss was 10-15%. The loss after 235 years storage in the dark was about 30%. I n the same period, rice stored a t -10" C. (14" F.) lost only about 1% of its thiamine. Pelshenke and Schulz (1951) found only small losses from bread after storage for six months. Pearce (1943) found no loss from wheat germ after 6 months storage a t 15.5" C. (60" F.). On the other hand, Bayfield and O'Donnell (1945) found losses of 32% in 5 months and 40% in 7 months from wheat stored a t atmospheric temperatures. 2. Canned Products

Guerrant et al. (1945, 1948) have studied losses of thiamine from fruit and vegetable products a t temperatures of from -1" t o 43.5" C . (30 t o 110" F.) for periods of up t o 2 years. They found very limited losses a t the two lowest temperatures, - 1 and 5.5" C. (30 and 42" F.), but very definite losses a t the higher temperatures. Clifcorn and Heberlein (1944b) obtained results a t room temperature on asparagus, Lima beans, corn, and peas which, in some cases, were contradictory. Moschette et al. (1947) found loss of thiamine from canned fruits and fruit juices a t 26.5" C. (SO" F.), but not a t 18.5" C. (65" F.) or less. They were concerned with warehouse storage and considered temperatures likely t o be encountered in practice. They found losses, after 1 year a t 26.5" C. (80" F.), of from 9 % with citrus fruits and fruit juices, t o 19% for peaches. Pineapples and tomatoes were intermediate. Sheft and her co-workers (1949) carried out a similar study over 18 t o 24 months with a number of canned fruits and fruit juices (citrus, pineapple, tomatoes, peaches). Losses in 2 years a t 10 t o 18.5" C. (50 t o 65" F.) did not exceed 13%; a t 26.5" C. (80" F.) they were higher. The authors state that losses of thiamine seemed to depend more on temperature than time of storage. This is hardly in accord with the Arrhenius equation which relates temperature with rate of loss. Feaster et al. (1949) have recorded results for

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

297

tomato juice in warehouse stacks. Losses were very small a t 4 and 21" C. (39 and 70" F.)) and it was concluded that thiamine is not seriously affected during stack cooling. Cover et al. (1949) showed high losses from canned beef and veal in 3 months a t summer temperatures [35 t o 36.5" C. (95 t o 98' F.)]. None of these investigations were kinetic studies, nor do the results permit any such calculations t o be made. They are therefore of limited value. Although they cannot be used for purposes of strict comparison or prediction of thiamine behavior, they do have a cumulative effect in showing the trend of thiamine stability in stored canned products, especially fruit and vegetables. Freed et al. (1949) have presented nomographs for thiamine retention in canned apricots, green beans, Lima beans, tomato juice, peas, and orange juice. They are based on data obtained on these products over 18 months storage a t 21, 32, and 38" C. (70, 90, and 100" F.) and were used t o compare the results obtained by Guerrant et al. (1945) and Moschette et al. (1947) for the same products, with predictions obtained from the nomographs. These comparisons were very promising, and there is no doubt that these nomographs are valuable in predicting the storage life of thiamine in the products studied and in other similar ones. Storage temperatures of 29.5" C. (85" F.) for tomato juice and peas, and 35' C. (95" F.) for orange juice led t o losses not greater than 20%, whereas the temperatures required for the same retention in Lima beans, apricots, and green beans were 18, -9, and -2" C. (64, 16, and 28" F.), respectively. The following table is taken from Freed et al. (1949) : Retention for 12 months storage at

38" c. (1000 F.) 1.5" C. (35' F.) %7

Apricots Grcen beans Lima beans Tomato juice Peas Orange juice

35

8 48 60

68 78

70 72 76 92 100 100 100

Although the paper of Freed et al. is very valuable and is soundly based on kinetic considerations, i t still leaves unexplored the factors affecting the thermal stability of thiamine. It does, however, permit a comparison with the results obtained by Guerrant et al. (1945) and Moschette et al. (1947). Farrer (1950) approached the problem in a different way by suggesting that the results of storage studies should be used t o calculate the rate

298

K. T. H. FARRER

constant, k,, where time is measured in weeks, not minutes. ( k , very nearly equals k X lo4.) The advantages ascribed t o this are: (1) the ability t o compare quickly and conveniently results obtained in different laboratories, and (2) the means of predicting, through the Arrhenius equation, the life of the thiamine under given conditions of temperature and time. I n assessing the latter conditions, the paper of Monroe et al. (1949) is of considerable assistance. Brenner et al. (1948b) have published thiamine retention curves for the storage of certain canned fruit and vegetable products. They make the general claim that 60% of the thiamine, on the average, is retained in these products after storage for 6 months at 38" C. (100" F.) or for 12 months a t 32" C. (90" F.). Farrer (Zoc. cit.) showed th a t their results could be treated kinetically, and presented rate constants for thiamine destruction in the various products and Arrhenius curves for apricots and orange juice. I n the same way, Farrer demonstrated the identity of the results of Feaster et al. (1946) and Rice and Robinson (1944) for thiamine losses on storage of canned pork. Farrer's own results give rate constants and Arrhenius equations for thiamine destruction in processed cheese (Farrer, 1950) and, more recently (Farrer, 1953a), in yeast extract as well. I n the latter paper the behavior of the mononitrate is compared with th a t of the naturally occurring thiamine. A similar approach to storage losses from dehydrated products was also found t o be possible. A further example of the usefulness of the kinetic approach may be found in the results of Guerrant and O'Hara (1953). They summarized many data obtained on the storage a t 20-29.5" C. (68-85" F.) of canned and bottled peas and Lima beans. If the results are examined by kinetic methods, i t is seen, firstly, that some sets show the initial unexplained rapid rate of destruction which has already been referred to, secondly, that the rate of destruction was the same as far as can be determined, in all samples of peas and in all samples of beans, and thirdly, that destruction was faster in beans than in peas. Approximate values of k, are 0.0049 for peas and 0.015 for Lima beans. It is possible now t o compare these results with those of Brenner et al. (1948b), from which Farrer (1950) calculated k , values of 0.013 [at 38" C. (100" F.)] and 0.0097 [at 32" C. (90" F.)] for beans. It is apparent that Guerrant and O'Hara's results for peas are of approximately the same order, having regard t o the lower temperatures, but are quite different for Lima beans. There is no doubt that the suggestions of both Freed et al. and Farrer are far better than the unrelated results which have appeared in the literature hitherto. The nomographs of Freed et al. are possibly simpler, whereas Farrer's approach is possibly more precise. Both, however, are based on fundamental kinetic considerations.

THE THERMAL DESTRUCTION

B~ IN FOODS

OF VITAMIN

299

Fournier et aE. (1949) have given results for losses of thiamine on storage of peanut butter a t temperatures ranging from -29 to 54" C. (-20 to 129" F.). At -29 and ' 4 C. (-20 and 39" F.). There was no loss in 5 months, but losses at 18, 27, 37, and 54" C. (64, 81, 99, and 129" F.) were significant and can be treated kinetically. The present author has calculated Ic, for thiamine destruction a t each temperature, and from the values obtained the Arrhenius curve (Fig. 4) has been drawn.

-2.5

I

I

0.00300

I

I

I

0.00320 1IT" abs.

I

0.00340

FIG.4. Arrhenius equation curve for loss of thiamine on storage of peanut butter at different temperature (from data of Fournier et al., 1949).

It is clear that the pattern of thiamine destruction is the same in different samples of the same product, and there is good reason for believing that it is very similar within groups of products. 3 . Dehydrated Products

Schultz and Knott (1939) and Knott (1942) have studied the storage loss of thiamine from evaporated milk. They show considerable losses over a period of some months. Knott (1942) claims that destruction on storage eventually reaches equilibrium, and that the rate of destruction varies with each lot of milk. By applying simple kinetics to the values given, it can be shown that neither statement is true. Photography and projection onto graph paper were used by the present author to measure the position of the points in Knott's curves for thiamine destruction. The rate of destruction k,, (Farrer, 1950) (where time is measured in weeks) was determined, and for 2 curves of 4 points each the values obtained were 0.0172 and 0.0205. Applying the same method t o Schultz and

300

K. T. H. FARRER

Knott's (1939) values gave 0.0182 and 0.0209 for 2 other milks, whereas another sample recorded by Knott (1942) gave the value, at one time interval only, of 0.0204. One other sample gave anomalous results, the same amount of thiamine remaining after two different time intervals and corresponding with k , values of either 0.044 or 0.097. However, it is clear that equilibrium is never reached as a simple time retention curve must be logarithmic, and that the rate of destruction of thiamine is, generally speaking, the same in each sample, as would of course be expected in products of similar composition. I n spray-dried whole egg, Klose ef al. (1913) found virtually no loss of thiamine a t -9.5" C. (15" F.) in 9 months. At 21" C. (70" F.) only 14% was lost in 3 months, but 48% in 9 months. At 37" C. (99" F.) the losses were 25, 32, and SO%, respectively, a t 3, 6, and 9 months. Olsen et al. (1948) showed that in spray-dried whole egg the thiamine retained after 20 and 57 weeks' storage a t 37" C. (98.5" F.), varied inversely with the moisture content in the range 1% t o 6%. A kinetic interpretation of their results suggests that at all moisture levels studied the rate of destruction is the same from 20 to 57 weeks and that the rates vary in the initial stages. This change of rate constant has been observed in buffer solutions and other food products. Tressler et al. (1943) report no loss of thiamine from dehydrated vegetables (rutabagas, beets, cabbage, and potatoes) on storage under air or carbon dioxide, or in cellophane or pliofilm bags for 3 to 4 months a t -40, 0.5, 14.5, and 24" C. (-40, 33, 58 and 75" F.) The Research Staff of the Continental Can Co. (1944, 1945) have published figures for similar samples incubated for longer times a t higher temperatures [up to 54.5" C. (130" F.)]. Losses up to 76% in 3 months a t 54.5" C. (130" F.) were found, and the effects of time and temperature were clearly shown. Rice et al. (1944) made a thorough study of the destruction of thiamine in dehydrated pork a t -29,3,27,37,49, and 63" C. (-20,37,81,98,120, and 145" F.). There was no detectable loss in 3 weeks a t the 2 lowest temperatures but Farrer (1950) showed that the results for the 4 highest temperatures permitted a kinetic study which, conforming with the Arrhenius equation, gave a clear-cut relationship between rate of destruction and temperature. Nymon and Gortner (1948) in a similar study, investigated the effect of temperature, incorporation of soya flakes, container, and moisture content on the retention of thiamine in dehydrated pork loaves. Farrer has also treated their results kinetically giving quantitative expression t o Nymon and Gortner's conclusions; via. that soya flakes retard thiamine destruction, and th at the loss of thiamine from dehydrated pork, with and without soya flakes, is faster in moistureproof packages than in cellophane. Nymon and Gortner attribute this t o

THE THERMAL DESTRUCTION OF VITAMIN B I IN FOODS

301

further loss of moisture through the cellophane, and this agrees with the emphasis laid by Rice et al. on the importance of the moisture content of dehydrated pork on the stability of thiamine. The importance of moisture content has also been emphasized by Hollenbeck and Obermeyer (1952) who worked with enriched flour. Storage was carried out for 4 months a t 38" C. (100" F.) and a t room temperature. Losses of thiamine hydrochloride were 40 and 27 %, respectively, but not more than 5% of thiamine mononitrate in flour of 14.5% moisture. Increasing moisture content favored destruction of thiamine hydrochloride which the authors attributed t o its greater hygroscopicity. Kandutsch and Baumann (1953) have studied the storage loss of thiamine from laboratory rations and confirmed that it follows a first order reaction. Losses in basal diets amounted t o about 70% in 1 week a t room temperature and attempts were made to modify the rations t o avoid these losses. They were reduced by removal of fat or salts, by the addition of antioxidants, such as ascorbic acid, or by storage under nitrogen. However, desiccation in air gave greater protection than storage under nitrogen, and desiccation in vacuo gave complete stability for 2 weeks a t room temperature. Thiamine was completely stable when mixed with sucrose, or with sucrose and other vitamins. Kandutsch and Baumann argued that destruction occurred in the aqueous film on the ration particles and that it was primarily a n oxidation. The formation of a film of glycerol in the particles successfully prevented losses. Waibel et al. (1954) in a similar study found dipotassium hydrogen phosphate t o be the ingredient most responsible for destruction. Calcium carbonate and manganese sulphate also contributed. Alodification of t h e formula and the addition of 1% glycerol greatly reduced losses. The importance of humidity was emphasized. That oxygen can accelerate thiamine destruction under certain conditions was shown by Farrer and Morrison (1919). Although these authors worked with buffer solutions from 50 t o 110" C. (122 t o 230" F.) there is, nevertheless, support for Kandutsch and Baumann's contention. The catalytic effect of salts is only t o be expected from the work already outlined earlier in this paper, and the protective effect of drying follows from what is known of the destruction of thiamine in aqueous solutions.

4. Pharmaceuticals It is generally recognized that thiamine is very stable in acidified solution, especially if stored in the cold. It is not possible, however, t o acidify pharmaceuticals t o the same extent as, for example, standard solutions, and losses of thiamine potency on storage of these materials is also recognized. Nevertheless, Watanabe et al. (1942) reported

302

K. T. H. FARRER

the high stability on storage of sterilized, aqueous solutions of thiamine hydrochloride, and Myrback et al. (1942) found no loss of thiamine hydrochloride in 9 months from solutions a t pH 2-6.5 and containing sodium pyrophosphate. Taub et al. (1949) found maximum stability of thiamine in parenteral solutions with riboflavin and niacinamide under nitrogen a t pH 4. They carried out storage tests at 22 and 45" C. (72 and 113" I?.) and found the mononitrate to be rather more stable than the hydrochloride. In iron compounds, ferrous gluconate favored thiamine stability more than iron peptonate or ferric ammonium citrate. The effects of time, temperature, pH, anions, thiamine salt, and, t o a lesser extent, air are evident in this work without there being sufficient data for the calculation of reaction rates. Macek et aZ. (1950) have compared the relative stability of the mononitrate and hydrochloride in a number of pharmaceutical products a t room temperature and 40" C. (104" F.). Quite clearly, their results are capable of kinetic treatment. They show the hydrochloride to be less stable than the mononitrate, losses of up to SO% of the former in 6 months at room temperature being recorded. Use of the mononitrate, and attention to methods of preparation and changes of formula, reduced storage losses to 10% or less. Bird and Shelton (1950) report the results of collaborative studies on the stability of thiamine mononitrate in tablets, capsules, and solutions also at room temperature and a t 40" C. (104" F.). Losses in dry preparations were very small (6-80/, after 1 year) except for those from capsules containing tartaric acid and held at 40" C. (104" F.). I n this case the loss was about 20%. Losses in solution were 7 and 12%, respectively, at room temperature and 40" C. (104" F.) in a solution adjusted to pH 3.8 with nitric acid, but in an elixir at pH 4.0, 15% was lost over a year a t room temperature and 42% a t 40" C. (104" F.). These figures illustrate wellknown variations in thiamine stability, but no comparisons with the hydrochloride were recorded. Partington and Waterhouse (1953) also studied losses on storage of dry pharmaceutical preparations at 15.5-21' C. (60-70" F.). I n formulas which give pH values of 4-5 in solution they showed losses of about 10% after 2-3 years. This confirms Bird and Shelton's work. I n addition, Partington and Waterhouse showed that formulae which gave solutions of higher pH suffered much greater loss on long storage, practically the whole of the vitamin disappearing. Copper and cobalt had little effect on thiamine loss under the conditions studied. Dutta et al. (1952) have studied the stability of thiamine in oral preparations at 25-32" C. (77-90' F.) and concluded that thiamine mono-

THE THERMAL DESTRUCTION OF VITAMIN B I IN FOODS

303

nitrate is more stable than the hydrochloride, but that in water or aqueous alcohol both forms are subject t o considerable losses over 40 weeks in ferrous sulphate solution (93 %). Their results allow for kinetic treatment. For the mononitrate in water and alcohol, k, has been calculated t o be 0.00345; for the hydrochloride, k , is approximately 0.0057. I n copper sulphate solutions, k , for both forms was approximately 0.01. I n all solut,ions the log (a - $)/time curves showed a very fast initial rate of destruction which quickly gave place t o a slower rate over almost all the time period. As already stated, this was noted by Booth (1943) and Rendix et al. (1951) and has frequently, but certainly not always, been encountered by the present author. The cause is still obscure. I n 2 cases in the paper of Dutta et al., the log ( a - $)/time curves were clearly 2 intersecting straight lines representing 2 distinct rates of reaction. Both applied t o the hydrochloride, and in each case (ferric sulphate and ferric ammonium citrate) the more rapid rate of reaction corresponds with th a t for the mononitrate, whereas the slower rate, attained after 12 weeks, was very much slower than th at for the mononitrate. These points are clearly evident from the kinetic treatment of the results, but are not apparent from the more usual time/retention curves or tables. VIII. CONCLUSION It is clear from the evidence available that the thermal destruction of vitamin B1 in foods can be followed by simple kinetic methods. First order reactions are relatively easy to follow, and the result may be expressed as a rate constant, k , in set.-', min.-' or weeks-', or as a halflife time, 2 t0.5 = log, k Whichever method is preferred, the result may then be compared with other results obtained under the same conditions. I n addition, it is a neat and quantitative expression of the percentage retention (or loss) after a given time, but, most of all, it is the first step in correlating losses with temperature as well as with time. The temperature coefficient of first order reactions is of the order of 2 or 3, i.e. the rate constant, k , approximately doubles or trebles for each 10" C. rise in temperature. The Arrhenius equation, log, k

=

I -

E RT

-

which is so useful in the study of first order reactions, has been shown, theoretically and experimentally, t o be of great value in the prediction of

304

K. T. H. FARRER

thiamine retention in foods, both in the canning process and in the storage of canned or dried foods. I n the graphical representation of the Arrhenius equation, where log k is plotted against 1 / T , the slope of the line is proportional to E , the energy of activation ( R = gas constant), and E can thus be calculated easily. Values of E from all the data which can be treated kinetically are given in Tables VII and VIII, and wide variations are evident. There is a TABLEV I I The Energy of Activation for Thiamine Destruction At Storage Temperatures Product

E in calories

Reference

Yeast extract (natural B,) Yeast extract (mononitrate) Cheese (natural BI) Cheese (mononitrate) Pork, canned Pork, canned Pork, dehydrated Peanut butter Apricots Orange juice Yeast extract, I % , solution a t 100-110" C.

10,200 9900 8450 8040 7440 7440 11,180 6520 6920 6210 6760

Farrer, 1953a Farrer, 1953a Farrer, 1950 Farrer, 1953a Rice and Robinson, 1944 Feaster et al., 1946 Rice et al., 1944 Fournier et al., 1949 Brenner et al., 194810 Brenner et al., 194813 Farrer, unpublished data

tendency for E to be similar in the high-protein products to the values obtained in buffer solutions, but whether or not this is a real trend cannot be determined from the limited information available. The big difference between the value obtained for canned pork and that for dehydrated pork is noteworthy. This may be related t o the claim of several authors that moisture content in dehydrated products is important in this connection and t o the belief of Kandutsch and Baumann (1953) that oxidation plays a part in such products. Some support for this is found in Table VIII which shows that E for thiamine destruction in buffer solution with oxygen is close t o the value obtained for dehydrated pork. I n addition, the great increase in E found in phosphate buffer solutions sealed in a n atmosphere of oxygen and heated in the range 90-110" C. (194230' F.) suggests that here there is some other reaction besides hydrolysis. The very low values for E obtained with several foods calls for comment, since i t is evident th at the reaction involved cannot be the same asmgivenin buffer solutions, yeast extract, etc., with values of the order of 10,000 cal. There appear t o be two possible reactions. One is the breaking

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

305

TABLEVIII E, the Energy of Activation for Thiamine Destruction in Buffer Solutionse Temperature range Buffer

(“C.1

Other conditions

E in calories

Citric acid-phosphate Citric acid-phosphate Citric acid-phosphate Citric acid-phosphate Phosphate Phosphate Phosphate Phosphate Phosphate Phosphate Phosphate Phosphate-phthalate Phosphate-phthalate Phosphate-phthalate Phosphate-phthalate

70-110 70-110 90-110 80-100 50-100 50-70 90-110 70-110 80-90 90-110 80-100 70-100 70-100 70-100 80-100

Sealed with air Sealed with O2 Sealed with n - 2 Reflux Reflux Sealed with air Sealed with Nz Sealed with air Sealed with 0 2 Sealed with 0 2 Reflux, with 0 2 stream Reflux, with 02 stream Reflux Sealed with air or Nz Sealed with O2

11,460 11,460 11,460 11,460 8630 8630 8630 11,380 11,380 20,060 11,380 10,100 8190 10,750 11,800

From Farrer and hloriison (1‘349)

of the CH, “bridge” leaving the pyrimidine and thiazole moieties, and the other involves the breakdown of the thiazole ring with production of HZS. Attempts are now being made t o obtain information on the thermal destruction products of thiamine, and from these and from further studies on foods, specifically designed t o yield kinetic data, the explanation will no doubt be obtained. Another point of interest is the discrepancy between the values obtained for yeast extract and yeast extract solutions in Table V I I . _4t storage temperatures, a commercial intermediate (about 40 % moisture) was used, but a t the higher temperatures a 1% solution mas studied. I n addition, the pH was lower in the latter case. The big difference in E; confirms the observation made by the present author (unpublished datda) that thiamine becomes more vulnerable as the concentration of yeast extract solutions increases. The effect of buffer salts on thiamine destruction is now well recognized, and i t is t o be expected that the observed rate constant is, as a result of the primary salt effect, related t o the ionic strength. Glasstone (1943, for example, gives the following relationships : (1) in noncatalytic reactions log k

=

log ko

+ 1.02ZAZ8-&

306

K. T. H. FARRER

where Z A and Z B are the charges carried by the reactants and p is the ionic strength. (2) in catalytic reactions, as between a neutral molecule and an ion.

IC =

kO(1

+

PP*)

where P8 is a constant. Since thiamine forms salts (hydrochloride and mononitrate), one would expect the first relationship to hold, but there is clear evidence for the catalytic effect of various ions on thiamine destruction so that the second relationship, Ic being a linear function of ionic strength, may apply. So far as the author is aware, no account has been taken of ionic strengths in studies on thiamine. However, all basic work on the pure substance has been done in buffer solutions and calculations will show that the changes in ionic strength from solution to solution are not great enough t o alter the general conclusions. This follows, too, from the work of the present author with buffer solutions of different concentrations (Farrer, 1947b; 194913). I n foods, the actual concentrations of ions in the aqueous phase are likely to be high, and ionic strengths difficult to calculate and of little practical importance. From studies on cereals, particularly, the catalytic effect of various ions appears to be more important, and, in any case, conditions within a food group will tend to be similar from sample to sample. From the many data available, a fairly clear picture is emerging, and the value of the kinetic approach is evident. If further work is done along these lines, one may expect the rapid solution of problems still outstanding. REFERENCES Ache, L., and Ribeiro, 0. F. 1945. Rev. fac. med. vet., univ. Sao Paulo 3, 27 (Chem. Abstr. 40, 7525). Ahmad, B., Mehra, S. L., and Bharihoke, G. 1948. Ann. Biochem. Exptl. Med. (India) 8, 89. Arnold, A., and Elvehjem, C. A. 1939. Food Research 4, 547. Amy, E., and Hanning, F. 1947. J . Am. Dietet. Assoc. 23, 690. Ashikaga, C. 1946. J . Fermentation Technol. (Japan) 24, 85 (Chem. Abstr. 47, 5039, 1953). Ashikaga, C. 1951. Vitamins (Japan) 4, 23. (Chem. Abstr. 46, 10428). Ashikaga, C., and Chachin, T. 1951. Vitamins (Japan)3,285 (Chem. Abstr. 46,10428). Ashikaga, C., and Koshimizu, M. 1951. Vitamins (Japan) 3, 110 (Chem. Abstr. 46, 10428). Association of Vitamin Chemists. 1951. “Methods of Vitamin Assay,” p. 110. Interscience, New York. Atkin, L., Schultz, A. S., and Frey, C. N. 1943. U.S. Patent 2,322,270. Aughey, E., and Daniell, E. P. 1940. J . Nutrition 19,285.

THE THERMAL DESTRUCTION

OF VITAMIN

B~ IN FOODS

307

Barackman, R. A. 1942. Cereal Chem. 19, 121. Barnes, B., Tressler, D. K., and Fenton, F. 1943. Food Research 8, 420. Basu, K . P., and Malakar, M. C. 1946. I n d i a n J . M e d . Research 34, 39. Bayfield, E.G., and O’Donnell, FV. W. 1945. Food Research 10, 485. Beadle, 13. W., Greenwood, D. A, and Kraybill, H. R. 1943. J . Biol. Chem. 149, 339. Bendix, G.H., Heberlein, D. G., Ptak, L. R., and Clifcorn, L. E. 1951. Food Research 16, 494. Bird, J. C., and Shelton, R. S. 1950. J . Am. Pharin. Assoc. Sci. E d . 39, 500. Booth, R. G. 1940. J . SOC.Chem. I n d . 69, 181. Booth, R. G. 1943. Biochem. J. 37, 518. Brady, 0 . E., Peterson, W.J., and Shaw, A. 0 . 1944. Food Mesearch 9, 400. Brenncr, S., Dunlop, S. G., and Wodicka, V. 0. 1948a. Cereal Chens. 26, 367-76. Brenner, S.,IVodicka, V. O., and Dunlop, S. G. 1948b. Food Technol. 2, 207. Briant, A. hl., and Hutchins, M. R. 1946. Cereal Chem. 23, 512. Uriant, A. >I., and Klosterman, A. M. 1950. Trans. Am. Assoc. Cereal Chem. 8, 69. Hiikin, V. >I., Auerman, I,. Ta., Zaitseva, Z.I., Kutseva, L. S., Pashovkin, V. F., and Sherbatenko, V. V. 1953. Voprosy Pitaniya 12, 29; Chem. Abstr. 48, 10293b. Chileail, It., Ridder, I,. E., and Morgan, A . F. 1945. Cereal Chem. 22, 50. Campbell, R., Ililtz, 31.C., and Robinson, A. D. 1946. Can. J . Research, F24, 140. Christensen, F. FV., Latzke, E., and Hopper, T. H. 1936. J. Agr. Research 63, 415. Clark, R. K., and van Duyne, F. 0. 1949. Food Research 14, 221. Clifcorn, I,. E., and Heberlein, D. G. 1943a. Canner 98(17) 12. Clifcorn, L. E., and Heberlein, D. G. 1944b. I n d . Eng. Chem. 36, 168. Connolly, F.,Hiltz, M. C., and Robinson, A . D. 1947. Can. J . Research C26, 4 3 . Cook, €3. B., Morgan, A. F., and Smith, M. B. 1949. Food Research 14, 449. Cooperninn, J. M., and Elvehjem, C. A. 1945. J . Am. Dietet. Assor. 21, 155. Cover, S.,and Dilsaver, E. M.1947. J . Am. Dietet. Assoc. 23, 613. and Hays, R. M. 1947. J . Am. Dietet. Assoc. 23, 501. Cover, S., Dilsavcr, E. IN., Cover, S., Dilsaver, E . M., and Hays, R. bf. 1949. Food Research 14, 104. Cover, S.,.\IcLaren, R. A., and Pearson, P. B. 1944. J . N u f r i t i o n 27, 363. Cover, S., and Smith, \V. H. 1948. Food Research 13, 475. Cover, S.,and Smith, W . H. 1952. Food Research 17, 148. Ciitlar, K. L., Joncs, J . R., Harris, K. W., and Fenton, F. 1944. J . A m . Dietet. ilssor. 20, 757. Dawson, E. It., and Martin, G. LV. 1921. J. Sot. C h e w I d . 60, 211. Dawson, E. R., and Martin, G. TV, 1942. J . Sor. (Ihem. Z , n d 61, 13. Dearden, D. V., Henry, K. M., Houston, J., Kon, S. K., and Thompson, S. Y. 1915. J . Dairy Research 14, 100. Down, D. E., and Meckel, R. B. 1943. Cereal Chem. 20, 352. Dutta, N. K., Mehta, C. J., and Narayanan, K. G. A. 1952. I n d i a n J . Pharm. 14, 53. Eklund, A. B., and Goddard, V. R. 1945. Food Research 10, 365. Elvehjem, C.A. 1941. Milk Plant Monthly, p. 26. Elvehjem, C. A., and Pavcek, P. L. 1943. Modern Hosp. 61, 110. Eriksen, S. E., and Boyden, R. E. 1947a. Kentucky Agr. E x p t . Sta. Bull. 603. Eriksen, S . E., and Boyden, R. E. 1947b. Kentucky Agr. E x p t . Sta. Bull. 604. Escudero, :L, and de Alvarez Herrero, H. G. 1942. Publs. inst. nacl. nutricion (Buenos Aires) Puhls. cient. C N P N o . 6, 7. Evans, E. V.,Irvine, 0. It., and Bryant, 1,. It. 1946. J . Nutrition 32, 227. Farrer, K.T . H. 1941. Australian Chem. Inst. J . & Proc. 8, 113. Farrer, K. T. H. 1945a. Biochem. J . 39, 128.

308

K. T. H. FARRER

Farrer, K. T. H. 1945b. Biochem. J. 39, 261. Farrer, K. T. H. 1946. Australian J . Exptl. Biol. Med. Sci. 24, 213. Farrer, K. T. H. 1947a. Biochem. J . 41, 162. Farrer, K. T. H. 1947b. Biochem. J . 41, 167. Farrer, K. T. H. 1948. Brit. J . Nutrition 2, 242. Farrer, K. T. H. 1949a. Australian J . Ezptl. Biol. Med. Sci. 27, 157. Farrer, K. T. H. 1949b. Australian J . Exptl. Biol. Med. Sci. 27, 591. Farrer, K. T. H. 1950. Australian J . Exptl. Biol. Med. Sci. 28, 245. Farrer, K. T. H. 1953a. Australian J . Expal. Biol. Med. Sci. 31, 247. Farrer, K. T. H. 1953b. Australian J . Sci. 16, 62. Farrer, K. T. H. 1953c. Australian J . Ezptl. Biol. Med. Sci. 31, 577. Farrer, K. T. H., and Morrison, P. G. 1949. Australian J . Exptl. Biol. Med. Sci. 27, 517. Feaster, J. F., Jackson, J. M., Greenwood, D. A., and Kraybill, H. R. 1946. Ind. Eng. Chem. 38, 87. Feaster, J. F., Tompkins, M. D., and Ives, M. 1948. Food Inds. 20, 14. Feaster, J. F., Tompkins, M. D., and Pearce, W. E. 1949. Food Research 14, 25. Fellers, C.R., Esselen, W. B., and Fitzgerald, G. A. 1940. Food Research 6, 495. Fenton, F., Barnes, B., Moyer, J. C., Weeler, K. A., and Tressler, D. K. 1943. Am. J . Publ. Health 33, 799. Fournier, S. A., Beuk, J. F., Chornock, F. W., Brown, L. C., and Rice, E. E. 1949. Food Research 14, 413. Freed, M., Brenner, S., and Wodicka, V. 0. 1949. Food Technol. 3, 148. French, R.B., Abbott, 0. D., and Townsend, R. 0. 1951. Florida Agr. Expt. Sta. Bull. 482, 5. Fritsch, E. 1944. Deut. Mullerei 33. Glasstone, S. 1948. “Textbook of Physical Chemistry,” 2nd. ed., p. 1116, 1138. Macmillan, London. Gleim, E. G., Tressler, D. K., and Fenton, F. 1944. Food Research 9, 471. Gleim, E.G., Albury, M., Visnyei, K., McCartney, J . R., and Fenton, F. 1946. Food Research 11, 61. Goldberg, L., and Thorp, J. M. 1946. Nature 168, 22. Greenwood, D.A., Beadle, B. W., and Kraybill, H. R. 1943. J . Biol. Chem. 149,349. Greenwood, D.A., Kraybill, H. R , Feaster, J. F., and Jackson, J. M. 1944. Ind. Eng. Chem. 36, 922. Griswold, R. M., Jans, L. M., and Halliday, E. G. 1947. J. Am. Dietet. Assoc. 23, 600. Griswold, R. M., Jans, L. M., and Halliday, E. G. 1949. J . Am. Dietet. Assoc. 26, 866. Guerrant, N.B., Vavich, M. G., and Dutcher, R. A. 1945. Ind. Eng. Chem. 37, 1240. Guerrant, N. B., Vavich, M. G., Fardig, 0. B., Dutcher, R. A., and Stern, R. M. 1946. J . Nutrition 32, 435. Guerrant, N. B., Fardig, 0. B., Vavich, M. G., and Ellenberger, H. E. 1948. Ind. Eng. Chem. 40, 2258. Guerrant, N. B., and O’Hara, M. B. 1953. Food Technol. 7, 473. Guha, B. C., and Drummond, J. C. 1929. Biochem. J . 23, 880. Gyorgy, P., and Rubin, S. H. 1950. in “Vitamin Methods” (Gyorgy, ed.), p. 223. Academic Press, New York. Hanning, F. 1941. J . Am. Dietet. Assoc. 17, 527. Harrel, C. G., Sherwood, R. C., Sullivan, B., and Rainey, W. L. 1941. Bakers’ Digest 16, 42. Hartzler, E., Ross, W., and Willett, E. L. 1949. Food Research 14, 15.

THE THERMAL DESTRUCTION OF VITAMIN B~ IN FOODS

309

Heberlein, D. G., Ptak, L. R., Medoff, S., and Clifcorn, L. E. 1950. Food Technol. 4, 109. Heupke, W., and Kittelman, J. 1943. Mtinch. med. Wochschr. 90, 423 (Chem. Abstr. 38, 5603). Hinman, W. F., Brush, M. K., and Halliday, E. G. (1944). J . Am. Dietet. Assoc. 20, 752. Hofmann, C., Schweiteer, T. R., and Dalby, G. 1940. Cereal Chem. 17, 737. Hollenbeck, C. M., and Obermeyer, H. E. 1952. Cereal Chem. 29, 82. Holmes, A . D., Jones, C. P., Wertz, A. W., and Kuzmeski, A. W. 1943. J . Nutrition 26, 337. Holmes, A. D., Lindquist, H . G., Jones, C. P., and Wertz, A. W. 1945. J . Dairy Sci. 28, 29. Houston, J., Kon, S. K., and Thompson, S. Y. 1940. J . Dairg Research 11, 151. Houston, J., Thompson, S. Y., and Kon, S. K. 1940. J . Dairy Research 11, 87. Inagaki, C., and Takeda, K. (1950-51). J . Agr. Chem. SOC.J a p a n 24, 486. (Chem Abstr. 47, 2293.) Ingalls, R. L., Brewer, W. D., Tobey, H. L., Plummer, J., Bennett, B. B., and Paul, P. 1950. Food Technol. 4, 264. Jackson, S. H., Crook, A., Malone, V., and Drake, T. G. H. 1945. J . Nutrition 29, 391. Jackson, J. M., Feaster, J. F., and Pilcher, R. W. 1945. Proc. Inst. Food. Technol. 6 t h Conf., Chicago, p. 81. Jansen, B. C. P. 1936. Rec. trav. chim. 66, 1046. Johnston, C. H., Schauer, L., Rapaport, S., and Deuel, H. J., Jr. 1943. J . Nutrition 26, 227; Science 97, 50. Kandutsch, A. A., and Baumann, C. A. 1953. J . Nutrition 49, 209. Kawasaki, C., and Kani, T. 1949. J . Pharm. SOC.J a p a n 69, 216. Kennedy, B. M., and Tsuji, F. 1952. J. Am. Dietet. Assoc. 28, 1145. Kik, M. C. 1945. Arkansas Agr. Expt. Sta. Bull. 468. Klose, A. A., Jones, G. I., and Fevold, H. L. 1943. I n d . Eng. Chem. 36, 1203. Knott, E. RI. 1942. Am. J . Public Health 32, 1014. Kohman, E. F., and Rugala, A. A. 1949. Food Research 14, 72. Kon, S. K. 1941. Nature 148, 607. Lane, R. L. Johnson, E., and Williams, R. R. 1942. J . Nutrition 23, 613. Lehrer, W. P., Wiese, A. C., Harvey, W. R., and Moore, P. R. 1951. Food Research 16, 485. Lehrer, IT. P., Wiese, A. C., Harvey, W. R., and Moore, P. R. 1952. Food Research 17, 24. Leong, P. C., and Strahan, J. K. 1952. Med. J . Malaya 7 , 3 9 . Lincoln, H., Hove, E. L., and Harrel, C. G. 1944. Cereal Chem. 21, 274. Lunde, G., Kringstad, H., and Olsen, A. 1940. Angew. Chem. 63, 123. Macek, T. J., Feller, B. A., and Hanus, E. J. 1950. J . Am. Pharm. Assoc. 39, 365. McIntire, F. C., and Frost, D. V. 1944. J . Am. Chem. SOC.66, 1317. McIntire, J. M., Schweigert, B. S., and Elvehjem, C. A. 1943a. J . Nutrition 26, 621. McIntire, J. M., Schweigert, B. S., Henderson, L. M., and Elvehjem, C. A. 194313. J . Nutrition 26, 143. McIntire, J . M., Schweigert, B. S., Herbst, E. J., and Elvehjem, C. A. 1944. J . Nutrition 28, 35. McIntosh, J. A., and Jones, E . 1947. Food Technol. 1, 258. Marks, A. L., and Nilsen, H. W. 1946. Com. Fisheries Rev. 8 (12), 12. Mayfield, H. L., and Hedrick, M. T. 1949. J . Am. Dietet. Assoc. 26, 1024.

310

K. T. H. FARRER

Meckel, R. B., and Anderson, G. 1945. Cereal Chem. 22, 429. Melnick, D.,and Field, H., Jr. 1939. J . Biol. Chem. 127, 531. Melnick, D.,Robinson, W. D., and Field, H., Jr. 1941. J . Bid. Chem. 138, 49. Meyer, B. H., Hinman, W. F., and Halliday, E. G. 1947. Food Research 12, 203. Mickelson, O.,Waisman, H. A., and Elvehjem, C. A. 1939. J . Nutrition 17, 269. van der Mijll Dekker, L. P. 1952. Voeding 13, 89. van der Mijll Dekker, L. P., and Engel, C. 1952. Voeding 13, 152. Millares, R., and Fellers, C. R. 1949. Food Research 14, 131. Miller, C.D. 1945. J . Am. Dietet. Assoc. 21, 430. Miller, C. D., Bauer, A., and Denning, H. 1952. J . Am. Dietet. Assoc. 28, 435. Monroe, K. H., Brighton, K. W., and Bendix, G. H. 1949. Food Technol. 3, 292. Morgan, A. F., Kidder, L. E., Hunner, M., Sharokh, B. H., and Chesbro, R. M. 1949. Food Research 14, 439. Moschette, D. S.,Hinman, W. F., and Halliday, E. 1947. I n d . Eng. Chem. 39, 994. Murphy, H.W., and Goodyear, J. M. 1949. J . Am. Pharm. Assoc. 38, 174. Murray, H.C. 1948. Food Research 13, 397. Myrback, K., Vallin, I., and Kihlberg, B. 1942. Suensk Kem. Tidskr. 64,97. Myrback, K., and Vallin, I. 1944. Suensk K e m . Tidskr. 66, 175. Noble, I., and Gomez, L. 1951. J . Am. Dietet. Assoc. 27, 752. Noble, I., Turnbull, F., and Spear, J. 1946. M i n n . Agr. Expt. Sta. Misc. Ser. 664. Nymon, M. C.,and Gortner, W. A. 1948. Food Research 12, 83. Obermeyer, H.G., Fulmer, W. C., and Young, J. M. 1944. J . Bid. Chem. 164,557. Olsen, A. L., Weybrew, J. A., and Conrad, R. M. 1948. Food Research 13, 184. Oser, B.L., Melnick, D., and Oser, M. 1943. Food Research 8, 115. Pace, J. K., and Whitacre, J. 1953a. Food Research 18, 231 Pace, J. K., and Whitacre, J. 1953b. Food Research 18, 239. Pace, J. K., and Whitacre, J. 1953c. Food Research 18, 245. Partington, H.,and Waterhouse, C. E. 1953. J . Pharm. and Yharmacol. 6, 715. Pearce, J. A. 1943. Can. J. Research C21, 57. Pearson, P. B., and Luecke, R. W. 1945. Food Research 10, 325. Pelshenke, P., and Schulz, A. 1951. Brot u. Gebuck 6, 183. Petit, L. 1950. Ann. inst. natl. recherche agron. Ser. A Ann. agron. 1, 41. Petit, L., Guillot, A,, and Guillemet, R. 1945. Bull. SOC. chim. biol. 27, 529. Pickett, T. A. 1944. Georgia Agr. Expt. Sta. Circ. 146,p. 8. Pulkki, L. H., Puutula, K., and Laurila, U. R. 1950. Maataloustieteellinen Aikakauskirja 22, 164. Randoin, L., and Perroteau, A. 1950. Lait, 30, 622. Reedman, E. J. and Buckby, L. 1943. Can. J. Research C21, 260. Research Staff, Continental Can Co. 1944, 1945. Food In&. 16,171,366, 458, 542, 635, 702, 903, 991; 17, 147. Rice, E. E., Beuk, J. F., Kaufman, F. L., Schulte, H. W., and Robinson, H. E. 1944. Food Research 9, 491. Rice, E. E., Beuk, J. F., and Robinson, H. E. 1943. Science 98, 449. Rice, E.E., and Beuk, J. F. 1945. Food Research 10, 99. Rice, E. E., Fried, J, F., and Hess, W. R. 1946. Food Research 11, 305. Rice, E.E., and Robinson, H. E. 1944. Am. J . Public Health 34, 587. Roy, J. K., 1953. J . I n d i a n Chem. SOC.,I n d . N e w s E d . 16, 50. Sarett, H. P., and Cheldelin, V. H. 1945. J . Nutrition 30, 25. Schultz, A. S., Atkin, L., and Frey, C. N. 1942. Cereal Chem. 19, 532. Schultz, H. W., and Knott, E. M. 1939. Proc. SOC.Exptl. Biol. Med. 40, 532.

THE THERMAL DESTRUCTION OF VITAMIN ~1

IN FOODS

311

Schweigert, B. S., McIntire, J. M., and Elvehjem, C. A. 1943. J . Nutrition 26, 73. Sheft, B. B., Griswold, R. M., Tarlowsky, E., and Halliday, E. G. 1949. Ind. Eng. Chem. 41, 144. Sherman, H. C., and Burton, G. W. 1926. J . Biol. Chem. 70, 639. Shetlar, M. R., and Lyman, J. F. 1941. Cereal Chem. 18,666. Slater, E. C. 1941. Australian J . Exptl. B i d . Med. Sci. 19, 29. Slater, E.C., and Rial, J. 1942. Med. J . Australia 29, 231. Stamberg, 0.E., and Peterson, C. F. 1946. J . Am. Dietet. Assoc. 22, 315. Stevens, H. B., and Fenton, F. 1951. J . A m . Dietet. Assoc. 27, 32. Swaminathan, M. 1942. Indian J . Med. Research 30, 409. Taub, A.,Katz, I., and Katz, M. 1949. J . A m . Pharm. Assoc. 38, 119. Tauber, H.1937. Proc. Soc. Exptl. Biol. Med. 37, 541. Tavares, A., and Rodrigues, L. D. 1947. J . farm (Lisbon) 6, 153. Terao, K., and Kawaziri, S. 1951. Japanese Patent 7696 (Chem. Abstr. 47, 5641a). Teysseire, Y. 1945. Compt. rend. soc. biol. 139, 822. Trefethen, I., Causey, K., and Fenton, F. 1951. Food Research 16, 409. Trefethen, I., and Fenton, F. 1951. Food Research 16, 342. Tressler, D.K., Moyer, I. C., and Wheeler, K. A. 1943. Am. J . Public Health 33,975. Tucker, R. E., Hinman, W. F., and Halliday, E. G. 1946. J . Am. Dietet. Assoc. 22,877. Vogel, M. 1944. Mehl u. Brot 44, 180. (Chem. Abstr. 41, 7547.) Waibel, P. E., Bird, H. R., and Baumann, C. A. 1954. J . Nutrition, 62, 273. Watanabe, A. 1951. Japanese Patent 7695. (Chem. Abstr. 47,5640i.) Watanabe, A.,Kamio, H., and Tada., M. 1942. J . Pharm. SOC.Japan 62, 1-4. Watanabe, A., Marui, T., and Tada, M. 1944. J . Pharm. Soc. Japan 64, 23 (Chem. Abstr. 46, 3566). Watanabe, A., and Marui, T. 1949. Ann. R e p t . Takeda Research Lab. 8, 11 (Chem. Abstr. 46, 11587a). Watanabe, A., and Sakaki, M. 1944. J . Pharrn. Soc. Japan 64,322. (Chem. Abstr. 46, 824). Watanabe, A,, Tada, M., and Marui, T. 1949. Ann. Rept. Takeda Research Lab. 8, 5 (Chem. Abstr. 46, 11587~ ). Watanabe, A., Terao, M., and Sakashita, T. 1952. Ann Repts. Takeda liesearch Lab. 11, 124 (Chem. Abstr. 47, 6603g). Weckel, K. G. 1938. Can. Dairy Ice Cream J . 17, 49-50. Wertz, A. W., and Weir, C. E. 1944. J . Nutrition, 28, 255. Westerman, D. B., Vail, G. E., Kalen, J., Stone, M., and Mackintosh, D. L. 1952a. J . Am. Dietet. Assoc. 28, 49. Westerman, D. B., Vail, G. E., Kalen, J., and Stone, M. 1952b. J . Am. Dietet. Assoc. 28, 33 1. White, E. G., Murray, M., and Maveety, D. J. 1946. 1. Am. Dietel. Assoc. 22, 770 Wilcox, E. B., and Galloway, L. S. 1952. Food Research 17, 67. Williams, V. R., and Fieger, E. A. 1944. Food Research 9,328. Willich, R. K., Murray, M. D., O’Connor, R. T., and Freeman, A. F. 1952. Food Technol. 6, 199. Wilson, E. C. G. 1942. New Zealand J . Sci. Technol. B24, 35. Zaehringer, M.V., and Personius, C. J. 1949. Cereal Chem. 26, 384.

This Page Intentionally Left Blank

Tunnel Dehydrators for Fruits and Vegetables BY P . W . KILPATRICK. E . LOWE.

AND

W . B . VAN ARSDEL

.

Western Utilization Research Branch. Agricultural Research Service. U S. Department of Agriculture. Albany. California

Page I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314 1. The Development of Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . . . . . . . 314 2. Production of Dehydrated Fruits and Vegetables . . . . . . . . . . . . . . . . . . . 315 I1. Classification of Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . 316 d Arrangements . . . . . . . . . . . 316 1. General Discussion, Characteris 318 2. Longitudinal Air Circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . a . Counterflow Circulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 b . Parallel-Flow Circulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 c. Two-Stage Tunnels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 3 Transverse Air Circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324 a Combination Compartment and Tunnel . . . . . . . . . . . . . . . . . . . . 324 ........................ 325 4 Other Tunnel Arrangements . . . . . . . . I11. Mechanical Elements of Tunnel Construction . . . . . . . . . . . . . . . . . . . . . . . . . 326 326 1. Fans and Blowers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Heating Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332 3 . Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 4 . Materials of Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 337 5. T r a y s a n d T r u c k s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Typical Commercial Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 1 . Twin-Tunnel Counterflo-w Dehydrator . . . . . . . . . . . . . . . . . . . . . . . . . . 339 342 2. The Miller Tunnel Dehydrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . The Carrier Compartment Drier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 V. Criteria for Selection of Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . . . . 345 VI . Basic Theory of Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 350 1. Theoretical Tunnel Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Optimum Tray-Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 ............................. 357 3 Optimum Recirculation of Air . 4 . Product Temperature in the D rator . . . . . . . . . . . . . . . . . . . . . . . . . 359 362 5. Departures from Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Operating Procedures for Tunnel Dehydrators . . . . . . . . . . . . . . . . . . . . . . . . . 363 VIII . Recent Trends in Tunnel Dehydration of Fruits and Vegetables . . . . . . . . . 367 ........................................ 369 IX List of Symbols Used . References . . . . . . . . . . . .................................. 369 313

. .

.

.

.

314

P. W. KILPATRICK, E. LOWE, A N D W. B. VAN ARSDEL

I. INTRODUCTION This discussion of tunnel dehydrators, as used to dehydrate certain fruits and vegetables, is intended to provide a general introduction t o the subject, primarily for the use of students, food technologists, and engineers. Although it goes into a number of matters concerned with the design and operation of this kind of equipment, it is in no sense a manual either of design or of operation. Emphasis is laid on discussion of underlying principles and the more recent advances in application of these principles. An effort has been made to bring together published information from many different sources, some of which are not widely available. Unfortunately, it has not been possible for the authors t o survey publications in other than English-language journals and books. 1. The Development of Tunnel Dehydrators

The germ of the idea of the tunnel dehydrator is a t least a century old. Various features of present-day, typical tunnels were undoubtedly added one a t a time t o as simple a basic idea as th at described b y Yule (1845) in an English patent for “improvements in preserving animal and vegetable matters.” Yule, a “preserved provision manufacturer,” placed the cooked or uncooked animal or vegetable product on shelves in “ a chamber of oblong form,” and passed through the chamber a current of air which had been dried by passage through a receptacle of lump calcium chloride or other chemical absorbent of moisture. Yule says nothing in his patent about heating the air stream, but Prescott and Proctor (1937) say that Eisen, in 1795, dried vegetables on racks arranged around a stove in a dry-room, so Yule undoubtedly was acquainted with warm-air drying. No records have come t o light about the kind of equipment used to dry the dehydrated vegetables used by the Union Army in the Civil War. Something strongly resembling a counterflow tunnel dehydrator was being used later in the 19th century t o dry glue, according to Thorp (1905). Cruess (1938) says the Oregon tunnel drier was invented by Allen about 1890; this drier, which consists of a long sloping box with spaced ledges on the interior vertical sides to support trays of fruit, mas originally ventilated only by convection of warm air from a furnace room a t the lower end, but this was later supplanted by fan ventilation. The trays of fruit were pushed downhill through this tunnel, counterflow t o the air movement. Still later, as described b y Wiegand (1923)) the design was modified t o permit recirculation of part of the air, and finally the tunnels were constructed level and wheeled trucks replaced the sliding trays. During the First World War a considerable flurry of interest in vegetable dehydration had led t o the construction and operation of a number

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

315

of dehydration plants. Some of the driers were extremely elaborate and complex, and were abandoned at the end of the war. Soon thereafter, however, untimely rains in the prune-drying area of California led t o a great demand for practical farm fruit dehydrators. Within the next few years the work of investigators like Cruess (1919) and Cruess and Christie (1921a,b), engineers like Ridley (1921), and designers and builders like Chapman (1922a,b), Rees (1922), Puccinelli (1923), and Pearson (1923) resulted in the rapid development of simple tunnel dehydrators basically similar t o those used today. The design of the Oregon tunnel had converged toward a similar pattern. Many hundreds of these dehydrators went into regular use in central California, Oregon, and Washington for drying fruits. A few plants gradually built up a steady business in dehydrated onions, garlic, peppers, and several other vegetables. A number of years later Eidt (1938) described two-stage tunnels which had been designed and built in the Canadian Maritime Provinces for use in dehydrating apples. With the outbreak of the European war in the following year the British Ministry of Foods (1946), after intensive investigation, decided upon a two-stage tunnel dehydrator for its emergency vegetable dehydration plants. The same pattern was followed in most of the British Commonwealth countries. When the United States entered upon its own wartime dehydration program, individual operators were left free t o select the dehydration system they thought most suitable. Some of them adapted existing fruit dehydrating tunnels t o the faster evaporation rates obtainable from cut vegetables. Many new tunnels of the same basically simple design were built. Several of the larger plants installed two-stage tunnels, others purchased the more elaborate multistage transverse-flow “compartment tunnels.” At the end of the war, all three of these types were operating successfully in the United States. Today simple counterflow tunnels handle most of the prune and raisin dehydration and a considerable proportion of the apple dehydration; counterflow or two-stage tunnels are used for most of the current vegetable dehydration. The authors do not know of any count of the dehydration tunnels now in use in the United States, but the number must be several thousand. The vast majority of these are used for drying fruit. 2. Production of Dehydrated Fruits and Vegetables

Raisins, prunes, and apples make up by far the bulk of the dried fruits. Although most raisins are still sun-dried in California, golden-bleached raisins are dehydrated, and during the war there was extensive building of dehydrators for the Thompson Seedless grapes which are the main source of ordinary raisins. The annual production of raisins varies from about 150,000 t o about 400,000 tons.

316

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

Substantially all of the American prunes are dehydrated. The production has declined somewhat since the early 1930’s, but is fairly steady around 200,000 tons. The production of dehydrated apples, peaches, and pears has gradually declined from peaks reached in the 1920’s and ’30’s. About 15,00020,000 tons of dried apples, something less than 10,000 tons of dried peaches, and 1000-2000 tons of dried pears have been produced during recent years. All of the dried apricots, and most of the dried peaches, figs, and pears are processed by sun-drying, rather than dehydrating; on the other hand, all of the dried apples are dehydrated. The production of dehydrated vegetables, in contrast to the production of dried fruits, has fluctuated widely in response to demands brought about by wartime emergencies. Total United States production in 1941 is estimated by Rasmussen and Shaw (1953) to have been 13,000,000 lb.; only 3 years later it had increased 15-fold1to 209,000,000 lb. In another 2 years it dropped back to 55,000,000 lb. The European crisis of 1948 boosted it steeply, and 3 years later the Korean war produced another upsurge. The production of 60,000,000 lb. in 1950 was composed 36 of potatoes, 15% of onions and garlic, 13 % of peppers, and the remainder distributed between many vegetables. Except for the fairly large proportion of mashed potato powder (“potato granules”) in this 1950 production, nearly all of the product was made in tunnel dehydrators. A growing demand for high-quality dehydrated vegetables in a variety of processed foods (canned hashes and stew, catsup, cottage cheese, “ a la king” products, meat pies, etc.) has resulted in a steady and diversified post-war growth of the civilian market. 11. CLASSIFICATION OF TUNNEL DEHYDRATORS

I. General Discussion, Characteristics, and Arrangements The fruit and vegetable dehydration industries, both in the United States and the British Commonwealth countries, have used the tunnel drier far more extensively than any other type of dehydrator. Tunnel dehydrators, as a class, are frequently called tunnel-and-truck, truck-andtray, or simply tunnel driers. Principles pertaining to their use have been discussed by Van Arsdel (195la,b) and Perry and associates (1946). Abstracts from these sources have been freely incorporated into the material contained in this part of the chapter. Likewise, Figures 1through 7 have been reproduced from these same sources. All tunnel-and-truck dehydrators for fruits and vegetables have a common feature which distinguishes them from other kinds of drier. This characteristic is the method of handling the commodity. Normally, the

T U N N E L DEHYDRATORS FOR FR UIT S AND VEGETABLES

317

prepared commodity in piece form is spread thinly on special trays fabricated of wood or metal. Depending on the commodity, tray loading for vegetables may range from 1 t o 3 lbs. per sq. ft.; for fruit, theloading may be in the range of 1 t o 5 lbs. per sq. ft. Loaded trays are then stacked, one above another, on a portable low-bed truck or dolly. Height of the stacked trays may range from about 5 t o 7 ft., depending on operating conditions.

I WET TRUCKS INSERTED

DIRECTION OF AIR FLOW

,FOR

SIDE EXIT DRY TRUCKS

TRUCKS PROGRESS I N THIS DIRECTION

BLOWER’

FIG. 1. Simple counterflow tunnel (elevation) (from Iran Arsdel, 1951b, Fig. I).

The trays are so designed that, when loaded and stacked, there is a clear air passage left between successive trays. The loaded trucks are pushed, one a t a time, into one end (usually called the “wet end” or loading end) of the dehydrator’s drying section. The drying section or “tunnel” is a straight passageway with a cross-section just large enough t o accommodate the loaded trucks. Tunnel lengths vary; some may hold only 4 or 5 SIDE ENTRANCE FOR WET TRU

EXHAUST

TRUCKS PROGRESS IN THIS DIRECTION

FIG.2. Simple parallel-flow tunnel (elevation) (from Van Arsdel, 1951b, Fig. 2).

trucks, whereas others may contain as many as 15 t o 20. During operation, a truck of dried material is removed from the “dry end” of the tunnel, the remaining trucks are pushed forward one truck length, and a truck of wet material is rolled into the vacant space a t the “wet end” of the tunnel. It is obvious that operation is only quasi-continuous (not truly continuous as it would be in the case of most conveyer driers), and this is known as “progressive” operation. Primarily, the flow of hot air used for drying is across the horizontal surface of the layer of wet material. Very little air circulates through the layer of wet material while it is in a tunnel drier. This is‘distinctly different from the air flow in most conveyer driers in which the air flows up (or down) through the product layer. I n commercial use, there are 3 basic arrangements of tunnel driers,

318

P. W. KILPATRICK, E. LOWE, AND W. B . VAN ARSDEL

plus several different combinations of these basic arrangements. The essential difference between the various types is mainly the direction of air flow relative t o truck movement through the tunnel. The three basic types are illustrated by the simplified sketches shown in Figs. 1 to 3, and will be referred t o later. Actually, all three types are much more complicated than indicated, since working units have provisions for recirculating H,E ATE TS

0

WET TRUCK ___.

0-

-HEATER

EXHAUST AIR

TRUCKS PROGRESS I N T H I S DIRECTION

- F R E S H AIR INLET

‘BLOWER

FIG. 3. Simple combination compartment and tunnel (plan view) (from Van Arsdel, 1951b, Fig. 3).

part of the drying air. The paragraphs which follow describe the counterflow, parallel-flow, and compartment tunnels, as well as combinations of the parallel and counterflow units (two-stage tunnels) and other tunnel arrangements. 2. Longitudinal A i r Circulation a. CounterJlow Circulation. I n the counterflow tunnel (Fig. l ) ,the hot drying air is blown into the dry end of the tunnel and moves straight through it, in a direction opposite t o the movement of material being dried. The (‘wet” air is discharged a t the wet end of the tunnel where the prepared fruit or vegetable enters. I n actual operation, in order t o increase fuel economy, or to raise the air humidity in the tunnel, provisions are made for recirculating a part of the air discharged from the wet end. As the hot air passes through the line of loaded trucks, it picks u p moisture from the fruit or vegetables on the trays, and in so doing the air becomes cooler. I n the counterflow tunnel, the warmest, driest air comes in contact with the nearly dry product while the cooler, more humid air is in contact with the wet material entering the tunnel. The maximum air temperature which can be used is determined by the commodity being dried, and is that temperature which the nearly dried product will tolerate for several hours without perceptible damage. I n the counterflow tunnel, the best conditions for drying are a t the end of the tunnel where the product is nearly dry. Reasonably good drying conditions can be secured at the wet

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

319

end of the tunnel if “light tunnel loading” is used. (Light tunnel loading refers t o a suitable balance between mass air flow rate, air temperature, and total water evaporated per unit of time, so that air a t the wet end of the tunnel has a reasonably high evaporative capacity, i.e., a wet-bulb depression of a t least 15 t o 25” F. for most commodities.) Industry has used several different arrangements of the counterflow tunnel. If production capacity requires more than one such tunnel, the TRUCKS

J

RECIRCULATION DAMPERS

I

TRUCKS PROGRESS IN THIS DIRECTION

FIG. 4. Direct-fired twin counterflow tunnels (plan view) (from Van Arsdel, 1951b, Fig. 4). FRESH AIR

\SIDE ENTRANCE FOR WET TRUCKS

\SIDE E X I T FOR DRY TRUCKS

FIG. 5. Side-entrance counterflow tunnel (elevation) (from Van Arsdel, 1951b, Fig. 5 ) .

initial investment in tunnel cost can be kept a t a minimum by using a common blower, heater, and recirculation return for two tunnels. Such an arrangement (see Fig. 4) is known as the “twin tunnel.” Arrangements such as the one illustrated in Fig. 4 sometimes prove unsatisfactory because of uneven air distribution. To correct this difficulty, a further modification of the basic counterflow design (using doors in the side walls of the dehydrator near the tunnel ends through which the trucks are pushed) is shown in Fig. 5. This side entry principle was

320

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

used for many two-stage dehydrators operating in Great Britain during World War 11. However, the side entrance and exit method adds t o the complexity of truck movement, requires additional tunnel length, and increased floor space. b. Parallel-Flow Circulation.. I n the parallel-flow tunnel (Fig. 2), the air moves straight through the tunnel in the same direction as that of truck movement or progression. The name “concurrent,” instead of parallel-flow, is usually used in the British Commonwealth countries for this type of tunnel arrangement. The parallel-flow tunnel is very similar to the counterflow unit in the general arrangement and layout. Basically, the only difference is that the product loading and unloading ends are interchanged, resulting in a reversal of the direction of truck travel with respect to air movement. With these changes in mind, Figs. 4 and 5 are applicable t o either parallelflow or counterflow tunnels. There are marked differences in the behavior of parallel-flow and counterflow tunnels. For example, if prepared vegetables are dehydrated in a parallel-flow tunnel, difficulty may be encountered in drying the product sufficiently t o assure satisfactory stability in subsequent storage. On the other hand, in dehydrating a whole fruit, for example prunes, a parallel-flow tunnel may cause cracking of the skins and excessive loss of juice. These problems do not arise in proper counterflow tunnel operation. Nevertheless, the parallel-flow tunnel can be and is used very satisfactorily if operated in conjunction with an auxiliary or finishing drier. The following facts explain why it is impractical t o use the parallel-flow tunnel by itself for the dehydration of fruits and vegetables. The hot drying air entering the parallel-flow tunnel comes in contact with the very wet product at the loading end. As drying progresses, the wet product is warmed up by contact with the hot air. At the discharge end of the tunnel, the relatively dry product is in contact with moisture laden air which has been greatly cooled and has a very low evaporative capacity. Thus it is difficult t o dry a commodity to reasonably low moisture levels in a parallel-flow tunnel. Let us now re-examine the loading end of the parallel-flow tunnel. The hot drying air which enters the tunnel will always have a wet-bulb temperature very much lower than its dry-bulb temperature. If the commodity being dehydrated is a prepared vegetable, the temperature of the wet material (near the loading end) will not exceed the wet-bulb temperature of the hot drying air. This condition will prevail for an appreciable length of time. During this initial drying period, the evaporation process is quite similar to that which would occur if a wick were moistened with water and placed in the air stream. As the prepared vegetable dries down

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

321

t o a moisture content somewhere in the region of 50 to 65%, the water has more and more difficulty in traversing the internal structure t o the outer surface of the pieces where i t vaporizes. The surface of the product then becomes relatively dry. At this stage, the temperature of the vegetable pieces slowly rises above the wet-bulb temperature of the air. Therefore, the hot air supplied to a parallel-flow tunnel can be raised t o a higher temperature than would be safe for counterflow tunnel operation. A different situation arises if the commodity being dehydrated is uncut fruit. I n this case, since the moisture diffuses slowly t o the surface, the temperature of the wet material will rapidly rise above the wet-bulb temperature of the air. Drying under excessively high temperature conditions will tend t o make the fruit crack and bleed, and there may be a n appreciable loss of juice, When in proper use, evaporation is very rapid in the wet-end zone of the parallel-flow tunnel drier. Compared t o the counterflow tunnel, evaporation in this zone is at least 3 times as fast. Therefore, while the material is still very wet, excellent drying conditions prevail in the parallel-flow tunnel. Due t o the evaporative cooling effect, it is possible t o use a relatively high hot-end temperature without scorching the product. Consequently, parallel-flow tunnels have a high potential evaporative capacity. On the other hand, if a very dry product must be produced without aid of an auxiliary drier or finisher, the parallel-flow tunnel must be operated in such manner that heat economy and capacity are very low. Nonuniformity of drying may also be a serious difficulty in the use of a parallel-flow t*unneldrier. The tray edge closest t o the hot-air end of any tunnel is always exposed to more severe drying conditions than the down-stream edge. I n the parallel-flow tunnel this condition is most pronounced. c. Two-Stage Tunnels. The fruit dehydration industry has sometimes modified the design of the standard counterflow tunnel to take advantage of the good characteristics of the parallel-flow unit. One of these modifications, called the "hot-center " arrangement, has been used successfully in the drying of prunes. I n a typical unit, the hot drying air is blown into the center of a long tunnel. The air stream divides and moves toward both ends. This modification provides counterflow operation in the wet half and parallel-flow in the dry half of the tunnel. However, it does not take full advantage of the high wet-end evaporative capacity of the parallelflow tunnel. Another tunnel arrangement, with a parallel-flow wet end followed by a counterflow dry end, capitalizes on both the high wet-end evaporative capacity of the parallel-flow tunnel and the good final drying characteristics of the counterflow unit. This arrangement is best suited for the

322

P.

w.

KILPATRICK,

E. LOWE, AND

w.

B. VAN ARSDEL

dehydration of commodities which are able t o withstand the high wet-end temperature characteristics of a parallel-flow first stage. There is ample evidence t o indicate that most vegetables are in this category; also some cut fruits, for example, apples. I n general, the determining criterion is that the prepared commodity shall be a relatively fast-drying material. The dehydration industry has also used a n arrangement called the “ center-exhaust ” dehydrator, illustrated by the simplified diagrammatic sketch of Fig. 6. This type unit operates b y drawing heated air into both ends of the tunnel. The hot air (under a slight negative pressure) travels through the loaded trucks and is sucked out of the tunnel chamber (near FRESH AIR TO FIRST STAGE HEATER, I

I

EXHAUST AIR

EXHAUSTEI~

RECIRCULATION

I

DAMPERS

h l R EXIT BELOW TRUCK L E V E L

FRESH AIR TO STAGE

SEC\oND

,HEATER

TRUCK) DOOR c

TRUCKS PROGRESS IN THIS DIRECTION

FIG.6. Center-exhaust tunnel dehydrator (plan view) (from Van Arsdel, l95lb, Fig. 11).

its center) by a blower acting as an exhauster. Although called a centerexhaust arrangement, the exhaust port is usually located about one-third of the tunnel length from the loading end. An advantage of the “centerexhaust” system lies in the fact th at trucks can be pushed straight through the tunnel, and therefore do not require rehandling during the transition from first to second stage drying. However, there are some serious design and operating problems. It is hard t o balance the air flow through the two ends, particularly if the two sections contain a n unequal number of trucks. It is also difficult t o secure good air-flow distribution through the trucks as they approach and leave the vicinity of the air exhaust section of the tunnel. Another arrangement of the two-stage type of dehydrator is shown diagrammatically in Fig. 7. Essentially this consists of separate blowers and heaters a t each end of a single tunnel. A sliding partition near the tunnel center divides the parallel-flow section from the counter-flow section, but permits truck movement from one t o the other. I n this type of divided single-tunnel, the trucks are pushed straight through from the parallel-flow section to the counterflow section. Sometimes, to economize

323

TUNNEL DEHYDRATOHS FOR F R U I T S AND V E G E T A B L E S

on labor, provisions are made for automatic synchronization of the truckadvancing and partition-moving mechanisms. I n the United States, a widely used two-stage arrangement has the parallel-flow and counterflow stages physically separated. This requires handling of the trucks during their transfer from the parallel-flow tunnel to the counterflow unit. The usual practice is t o have a bank of parallelflow tunnels arranged side by side. Special transfer tracks with turntables permit manual transfer of the trucks from the “dry end” of the parallelflow units t o the “wet end” of the counterflow tunnels. The latter are also arranged in banks side by side, and usually located behind the bank of BLOWER,

\ HEATER,

FRESH AIR INLETS

BLOWER ,HEATER /’

I

I

TRUCK) DOOR TRUCKS PROGRESS IN THIS DIRECTION

FIG.7. Two-stage, single-tunnel dehydrator (plan view) (from Van Arsdel, 1951b, Fig. 12).

first-stage driers. This arrangement permits considerable flexibility in the dehydration plant’s operation and drier load capacity. The number of first- and second-stage driers in operation ran be varied, so as to gear the plant’s drying capacity to the commodity being processed in the preparation line ahead of the driers. Trucks from any of the operating first stage tunnels can be routed to any of the counterflow second stage driers. Since most of the evaporative load takes place in the parallel-flow or first-stage tunnels, there is only a relatively light evaporative load requirement in the counterflow sections. The heater and blower capacities in the two sections will usually differ accordingly. The air exhausted from the counterflow stage is relatively warm and dry. For reasons of heat economy, i t is desirable t o use this air as a part of the supply t o the parallel-flow section. This requires duct-work of fairly large dimensions when the two drying stages are physically separated. The two-stage arrangement, standardized by the British Ministry of Foods (1946)) uses primary and secondary stage tunnels of equal length placed side by side. Trucks progress in one direction through the parallelflow tunnel. As they leave that section, they are turned around and pro-

324

P.

w.

KILPATRICK, E. LOWE, AND

w.

B. VAN ARSDEL

gress in the opposite direction through the counterflow tunnel. I n this arrangement, the loading and unloading of a group of dehydrators are done at the same end. Air recirculation is controllable in each stage. Normally, the exhaust air from the secondary tunnel furnishes the entire "fresh-air '' supply for the parallel-flow first-stage unit. Extensive performance data and plant experience are given in the bulletin issued by the Great Britain, Ministry of Foods (1946). The two-stage method of dehydration offers some distinct advantages over a single-stage drier. The reversal of air-flow direction, with respect to movement of material, tends to give a more uniformly dried product. Drying times are shorter, and good drying conditions prevail a t both ends of a two-stage unit. These conditions tend to favor product quality. The shorter drying time also allows an increased output from a dehydrator of a given size. Three-stage tunnels have been used successfully in a t least one plant, but the dehydration industry generally favors one- or two-stage driers. Perhaps the chief advantage of the three-stage unit is its flexibility which permits the drying (under nearly optimum conditions) of a large variety of different commodities. As the number of stages increases, control becomes rather complex, and more labor is needed for operation (unless truck handling is completely automatic). 3. Transverse Air Circulation

a. Combination Compartment and Tunnel. I n direct contrast to the dehydrators previously mentioned, the combination compartment and tunnel drier operates with the drying air moving back-and-forth through the trucks transversely to the axis of the tunnel. The principle is illustrated in simplified form by Fig. 3. As is evident, the material advancing through the tunnel is subjected to reversals of air-flow direction, an advantage which tends to equalize drying. There are many possible variations of the basic arrangement. The combination compartment and tunnel drier can be equipped with controlled air reheaters and provisions for air recirculation for each cornpartment (each truck position). The large number of independent controls makes the combination compartment tunnel very flexible in operation. As the material is passed through the tunnel, the commodity can be subjected to almost any desired time-temperature-humidity drying condition. Such units are also well suited as general-purpose, experimental, continuous driers, and can produce the relatively large quantities of dehydrated material necessary for storage studies. For example, at the Western Regional Research Laboratory such a unit has been used for this purpose,

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

325

and counterflow, parallel-flow, and various combinations of two-stage drying conditions simulated. The design and construction of a compartment tunnel requires the greatest of care to avoid operating difficulties. Elaborate provisions must be made for straightening out and equalizing the air flow across the trayed material and through the trucks. If air is forced to make sharp turns, it tends t o hug the outside of the curve. The ordinary arrangement of guide vanes or “splitters” will not control this tendency sufficiently when the high velocity air stream is turned through 180”. The use of a system of perforated plates, the small holes acting as orifices or nozzles, has met with some success when properly designed and installed. However, this requires a substantial increase in the power necessary for air movement. Short-circuiting of air from one compartment t o another, without going through a truck, is another difficulty, unless provisions are made for a permanent and reasonably tight seal between the trucks and tunnel walls. The general drying characteristics of the compartment type tunnel equipped with auxiliary heaters may be briefly characterized as follows: Each time that the air passes through a heater, both the wet-bulb and drybulb temperatures increase. If general movement of the air is toward the loading end of the tunnel, the wet-bulb temperature will progressively rise, unless additional fresh air is introduced a t each of the reheating stages. I n this regard, the compartment unit differs from the simple counterflow tunnel, for in the latter, the wet-bulb air temperature remains substantially constant throughout the tunnel. The commercial fruit and vegetable dehydration industry has used comparatively few compartment type tunnel driers. Preference has been shown to the counterflow and two-stage tunnels because of their relative simplicity and comparative freedom from the difficulties inherent in the compartment type.

4. Other Tunnel Arrangements Guillou and Moses (1943) developed a modified form of cross-flow fruit dehydrator for farm use, and presented plans, construction details, and operating instructions. This is a modified, simple form of the compartment type drier and has been used in a number of California orchards for dehydration of prunes. There are other possibilities of tunnel arrangements. Only two will be mentioned briefly, and neither of these apparently has progressed beyond the pilot-plant study stage. The closed-cycle system dehydrator does not exhaust to the atmosphere. It operates by partially dehumidifying the exhaust air from the

326

P.

w.

KILPATRICK,

E. LOWE, AND

w.

B. VAN ARSDEL

drier and this air is returned for reuse. Proposals have been made to use the system for the dehydration of onions and garlic in order to overcome the normally obnoxious exhaust from such dehydrators. The plan has also been used as a part of a pilot plant system for dehydrating food commodities in an atmosphere of oxygen-free gas. A combination blancher-dehydrator has received some study. Essentially the arrangement consists of an isolated compartment a t the wetend of a tunnel drier, with air temperature and humidity in the compartment under independent control. Humidity and temperature are controlled, so that the raw, wet commodity can be rapidly elevated t o temperatures in the range of 180 to 210" F. with little drying and no condensation taking place. The wet product is held at the high temperature for a short period, as determined by the time-temperature requirement to inactivate its enzyme system completely, and thereby securing a full blanch. The hot product then immediately enters the tunnel drier section. Initial drying of the hot product is extremely fast, and its temperature rapidly drops to the vicinity of the ambient wet-bulb air temperature. Further dehydration proceeds in the usual manner. One of the system's advantages is the minimized loss of nutrient material from the product. During the conventional blanching procedure, there is a loss of such material due to leaching. This is obvious if water blanching is used, but it is also true to somewhat a lesser extent, if steam is used as the heat transfer medium. In the latter case, steam condenses on the product as it is heated, and the hot condensate has a tendency to leach out soluble material rapidly from the product. Theoretically, it is possible to adjust the wet-bulb and dry-bulb temperature of the circulating air in the blanching compartment, so that the product temperature can be rapidly elevated without either condensation or evaporation of moisture taking place. However, a t the high temperatures involved, there is little margin between the conditions required for condensation and those for extremely rapid drying and scorching. Informal reports and observations made by the authors indicate that control is difficult. The method offers potential advantages, and perhaps some future investigator will develop a practical procedure. 111. MECHANICAL ELEMENTS O F TUNNEL CONSTRUCTIOX 1. Fans and Blowers

The introduction of forced air movement is probably the most important single contribution in the development of the modern dehydrator. Prior to the use of fans, air movement in dehydrators was entirely dependent upon natural circulation of a rising current of warm drying

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

327

air. Drying was slow and drier performance was poor, with the net result that a large number of driers was required for a given plant capacity. The modern dehydrator is a high-performance unit primarily because powerdriven fans make possible t,he movement of very large quantities of drying air. The performance of a drier is affected in two different ways by the velocity of the air moving through it. First, because an increase in air velocity past a moist body increases the rates of heat transfer and mass transfer, the rate of drying of the moist material increases. Second, and of greater practical importance, the mass of air moving through the drier is proportional to the velocity of the air, and the evaporative capacity of the drier is proportional t o this mass velocity. The effect of air velocity on drying rate is complex. The rate of evaporation from a free water surface is known to be proportional to the 0.8 power of the air velocity across the surface, but Guillou (1942) found that the drying rate of prunes increases only as the 0.2 power of the air velocity. Brown and Kilpatrick (1943) showed that the effect of air velocity on the drying rate of vegetables gradually decreases as the moisture content falls; below about 15 to 20% moisture content the drying rate is substantially independent of air velocity. High air velocity is effective in accelerating evaporation near the wet end of a vegetable dehydrator, but not near the dry end. Increasing the mass flow-rate of air through the tunnel increases the evaporative capacity of the tunnel, essentially by supplying additional heat which is available t o produce more evaporation. As is shown in a later section of this article (equation 2, p. 349) the temperature of the air falls and its humidity rises as it passes over the moist material in the drier, but the extent of these changes is inversely proportional to the massvelocity of the air. At a very high rate of air flow a good drying potential can be maintained even near the cool end of a very long tunnel. Air velocities in commercial fruit or vegetable tunnel dehydrators range from about 300 to over 1000 ft. per min., based on the entire crosssection of the tunnel, empty of trucks and trays; actual lineal velocity across the material in the trays will be from 50% to 100% greater. This is a range established by practical experience. To the writers’ knowledge, the only effort that has been made t o arrive a t an economic optimum air movement through application of information about the drying characteristics of a specific product was the design of Guillou and Moses (1943) of a farm fruit dehydrator. I n this case the drying characteristics of prunes and the importance of keeping capital costs and operating costs low led to choice of a relatively low air velocity. In vegetable dehydrators, on the other hand, the air velocity in the empty cross-

328

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

section usually is as much as 600 t o 800 ft. per min. For tunnels of dimensions which are widely used (about 6.5 ft. wide by 7 f t . high) the air-handling requirement of the fan or blower ranges from about 10,000 cu. ft. per min. to about 40,000 cu. ft. per min.; if a single fan supplies air to two tunnels, these figures are, of course, doubled. I n considering fan performance it is customary t o standardize the conditions to air weighing 0.075 lb. per cu. ft., corresponding to dry air at a temperature of 70“ F. and a barometric pressure of very nearly 29.90 inches. The performance of a fan whose rating is known in terms of this “standard air” can be readily computed for other temperatures and pressures by means of the well-known fan laws given in engineering handbooks. Increasing the air flow in a dehydrator by increasing the speed of rotation of the fan is subject to a very drastic law of diminishing returns, because the power absorbed by a fan of given size varies as the cube of its rotational speed. For example, under otherwise identical conditions the time required for drying potato half-dice to 16.7% moisture content in a parallel-flow tunnel can be reduced from 3% hr. to 3 hr. by increasing the air velocity from 400 ft. per min. to 600 f t . per min. If the same size fan is used to obtain the higher air velocity in the same size dehydrator, the power consumed by the fan will increase 3.38 times. Whether the increase in power cost will offset the decreases in other costs occasioned by the 17% increase in output can be determined only by an analysis of all the other cost items. The resistance to air flow, or static pressure drop in tunnel driers ranges from a minimum of about 56 inch water gauge to a maximum of about 145 inches water gauge (standard air conditions), the magnitude of the resistance depending upon the length of the drier, the number of trucks in the drier, the air velocity, and the air-flow path. I n some dehydrators, especially those of earlier vintage, the air-flow resistance is lower than might be expected because much of the air flows around instead of through the space between the trays. This happens when there is excessive clearance between the trucks in the drier and the walls, floor, and ceiling of the dehydrator. The air, seeking the path of least resistance, tends to bypass the drying trays by traveling down the clearance paths. I n spite of a high rate of air circulation drying is slow because the air bypassing the trays has little influence on the drying. The size and type of fan used in a tunnel dehydrator depend upon a number of factors, the air-handling requirement, the air-flow resistance, the permissible noise level in the plant, the space available for mounting the fan, the need for a fan with nonoverloading characteristics, and last but not least, the relative importance of minimum equipment cost as compared with minimum operating cost.

T U N N E L DEHYDRATORS FOR F R U I T S AND VEGETABLES

329

Space and cost play more important roles in the selection of a fan for use in a tunnel dehydrator than are generally recognized. For a given type of fan, there is usually more than one fan size t h a t can be used for a given air volume and air-flow resistance. Of the several sizes involved, one size will be more efficient than the others, and therefore would be the size selected unless space requirements or the importance of keeping equipment cost a t a minimum dictate the use of a smaller, less efficient fan. Selecting a fan on a compromise basis is not at all uncommon, especially when large size fans are involved. The need for a fan with nonoverloading characteristics depends a great deal upon the manner in which the tunnel is operated. If the drier is

k FIG.8. Siniple propeller-type fan.

operated so that the fan is at times discharging against considerably less than normal resistance pressure, then a fan with nonoverloadiiig characteristics is needed t o prevent the fan motor from being temporarily overloaded. On the other hand, if the fan is always discharging against a fixed resistance, then the nonoverloading characteristic is not essential. Nonoverloading type fans are used in most tunnel dehydrators because the flow resistance is considerably less than normal when the drier is only partly loaded, or when the end doors are open. Three different types of fans or blowers are commonly used in food dehydrators, namely propeller, axial flow, and centrifugal fans. I n general, simple propeller fans of the type shown in Fig. 8 are seldom found in tunnel dehydrators because they are used only in applications involving very low air-flow resistance, usually under $4 inch water gauge static pressure. T o improve their ability t o discharge against pressure, propeller-type fans are equipped with a special ring housing (see Figure 9*). When so equipped, they are used in tunnel dehydrators with moderate

* Mention of a specific manufacturer in the caption in Fig. 9, and at other places in the text, does not imply that the equipment shown or mentioned is recommended by the U. S. Department of Agriculture over similar equipment of other manufacture not mentioned or shown.

330

P.

w.

KILPATRICK,

E. LOWE,

AND

w.

B. VAN ARSDEL

air-handling requirement and nominal flow resistance (e.g. 20,000 cu. ft. per min. against a resistance pressure of 1% inches water gauge). Operating efficiency is improved if the fan discharge is equipped with an expanding conical duct connection to convert the velocity energy of the

FIG.9. Propeller-type fan with special ring housing (courtesy of Hartzell Propeller Fan C o . ) .

fan to pressure energy. Propeller fans equipped with the special ring housing are usually nonoverloading. The chief advantages of using propeller-type fans are simplicity of installation due to the compactness inherent in a piece of equipment i n which the air enters and leaves in the same direction, and comparatively

UPEL LER WHEEL OR

,/

AIR FLOW

VANES FIG.10. Cut-away view of vaneaxial fan (courtesy of Hartzell Propeller Fan

CO.).

low equipment or first cost. The principal disadvantage is the high operating noise level, an important factor where driers are located in populated areas. Axial flow fans can be divided into two general classifications, tubeaxial and vaneaxial (see Fig. 10). Both types resemble a propeller fan in that a rotating impeller moves air through the fan, with the air entering and leaving the fan in the same direction.

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

331

A tubeaxial fan consists essentially of an impeller or wheel with airfoil blades, rotating within a cylinder. Tubeaxial fans are used to move air over a wide range of volumes a t medium pressures. In tunnel driers, for example, tubeaxial fans are used to move from 30,000 to 75,000 cu. ft. of air per min. against a resistance pressure of from 1 t o 1% inches water gauge. Both tubeaxial and propeller-type fans discharge air traveling with a rotating or screw motion. When the rotating air stream enters the stack of drying trays in the drier, some of the tray surfaces will be exposed t o an air stream approaching from the top while other tray surfaces will be exposed to an air stream approaching from the bottom, below the wood slats. The result is that drying will not be uniform because of differences in the air velocity over the trays. To correct this difficulty, air straighteners of the egg-crate type are often used in tunnel driers equipped with tubeaxial or propeller-type fans. A vaneaxial fan is essentially a tubeaxial fan with air-guide vanes located either before or after the impeller. The guide vanes improve the performance of the axial flow fan, especially when discharging against pressure. When used in tunnel dehydrators, for example, vaneaxial fans are generally somewhat more efficient than tubeaxial fans of equivalent size and rating. The vanes also straighten the air leaving the fan, eliminating the rotating or screw motion characteristic of the air stream leaving a tubeaxial or propeller-type fan. Vaneaxial fans are capable of delivering against higher pressures than tubeaxial fans, a factor important in drier applications only if the air-flow resistance in the drier is abnormally high. Both tubeaxial and vaneaxial fans are available with nonoverloading characteristics. Like propeller-type fans, tubeaxial fans and vaneaxial fans are more efficient when the fan discharge is equipped with an expanding conical duct connection to convert the velocity energy of the fan t o pressure energy. A centrifugal fan consists essentially of a fan rotor or wheel rotating within a scroll shaped housing. Centrifugal fans are capable of moving air over a wide range of volumes and pressures, and are commonly used in tunnel dehydrators of all types and sizes. When equipped with backwardly inclined wheel blades, they are nonoverloading. Because the air enters from the side, centrifugal fans must be installed with ample air-flow clearance on one or both sides of the fan, depending upon whether the fan is single or double entry. This requirement, combined with the fact that centrifugal fans are, in general, relatively large in size, results in very large space requirements within the drier to accommodate the fan. This is especially true in high-performance driers using centrifugal fans

332

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

of the most efficientsize. To conserve space in such cases, fan size is often compromised by using a smaller fan of lower efficiency. Axial flow fans are competitive with centrifugal fans in most tunnel dehydrator applications. When compared with centrifugal fans, axial flow fans are more compact, easier, and therefore less expensive to install and, in general, lower in first cost. On the other hand, they are much noisier in operation and are somewhat less efficient if the fan is selected for maximum efficiency, independent of space requirements. 2. Heating Systems

Heating systems used in tunnel dehydrators are of two basic types, direct combustion heating and indirect heating. I n a direct combustion heating system the gaseous products of combustion are mixed and circulated with the drying air and hence come in direct contact with the product in the drier. An open flame in the main air stream of the dehydrator is an example of this type of heating system. I n an indirect heating system, the products of combustion are not circulated with the drying air. Heating surfaces are used to transfer the heat from the primary source to the drying air. A dehydrator using steam-air heating coils is an example of this type of heating system. Direct combustion heaters are widely used in tunnel dehydrators. Because there are no transmission losses, heat efficiency is at a maximum. The fuel used is usually either natural or manufactured gas, fuel oil, or bottled gas such as butane. A gaseous fuel is usually preferred to fuel oil because of the simplicity of the control equipment, the ease of handling, and the fact that the products of combustion are unlikely t o affect the quality of the dried fruit or vegetable. Gas burners are almost always of the “premixed” type, installed directly in the drier air stream with the flame shielded from the cooling effect of the surrounding air currents by a simple unlined sheet metal combustion chamber. Although not essential when using gaseous fuels, refractory-lined combustion chambers are sometimes used with gas burners to insure complete and therefore more efficient burning of the fuel. Oil burners are of many types-rotary, atomizing, centrifugal, pressure, etc. A basic rule in connection with the use of oil is that combustion must occur in a relatively high temperature zone. If the flame is chilled so that some of the oil particles are cooled below their ignition point, smoke and soot will be formed which will contaminate the product in the dehydrator. A common way of burning fuel oil in food dehydrators is in a refractory-lined steel sheet combustion chamber such as that shown in Fig. 11. The combustion chamber is divided into two zones by a refractory

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

333

checker brick partition, the partition serving t o confine the radiant heat t o the primary zone. I n operation, this primary zone becomes incandescent so t ha t combustion of the fuel occurs a t incandescent temperatures. The checker screen also serves as a baffle t o prevent the escape of unburned oil droplets since impingement of the droplets on the incandescent screen results in surface combustion of the fuel. The secondary zone is a n added precaution against smoking due to incomplete combustion. Large unburned particles of oil that escape from the primary zone will burn a t an accelerated rate when they come into contact with the high-velocity, high-temperature gases flowing through the checker wall restrictions. The particles are then given additional time in the secondary zone t o burn completely before coming in contact with /CYLINDRICAL

STEEL SHELL\

BURNER PRIMARY ZONE

\FIREBRICK

LINING)

[FIREBRICK CHECKER WALL’

FIG.11. Refractory-lined combustion chamber.

the drying air. Without the secondary zone, combustion may not be complete enough t o eliminate smoking. Considerable care must be exercised in the selection of a fuel oil for use in a direct combustion food drier. I n most cases, oils with high sulfur content cannot be used satisfactorily because the product absorbs a n excessive amount of the sulfur dioxide liberated during combustion. Indirect heating systems for dehydrators usually involve steam-to-air heaters although combustion gas-to-air heaters are used, particularly in apple dehydrators. The principal advantage of using an indirect heating system is t ha t there is no possibility of contaminating the material being dried with the products of combustion. The principal disadvantages are the additional equipment required and the lower heat economy. 3 . Instrumentation

The dry-bulb temperature of the air entering the drying tunnel is automatically controlled in virtually all tunnel dehydrators. I n a very few plants the wet-bulb temperature of the entering air is also automatically controlled. Dry-bulb temperature is controlled by regulating the flow of heat into the drier, ordinarily by means of a valve in the fuel or steam supply

334

P. W. KILPATRICK, E . LOWE, AND W. B. VAN ARSDEL

line. Wet-bulb temperature is controlled by regulating the position of the recirculating air damper. The instruments used to control dry-bulb temperature in tunnel dehydrators are of two basic types, on-off and modulating or proportional control. The latter is by far the most common. Wet-bulb temperature can be controlled satisfactorily only with modulating or proportional type instruments. On-off instruments range in complexity from simple thermostatic switches to industrial type controllers that indicate and, if so desired, also record the temperature. On-off control of dry-bulb temperature is practical only if the air-heating system in the drier involves a large amount of thermal capacitance or heat inertia. I n a tunnel dehydrator this usually means the use of a combustion chamber large enough to serve as a heat reservoir, storing heat while the burners are on and releasing the stored heat to the air when the burners are off. With an on-off control, air temperature will fluctuate to some extent, the amplitude and frequency of fluctuation depending upon the instrument, its adjustment, the size of the thermal capacitance, and the size of the heating load. On-off control without thermal capacitance is unsatisfactory because air temperature will fluctuate excessively. Fluctuations can be minimized by by-passing the control valve with a manually operated valve adjusted to maintain, without help from the controller, an air temperature slightly lower than the correct temperature, and depending upon the controller to supply only the additional heat required to bring the air to the proper temperature. The practice is not a good one, however, because the air temperature can rise to damaging if not dangerous levels if the heating load is reduced much below normal, for example, by a slackening in the rate of supply of wet material to the dehydrator. The simplest on-off temperature control system would consist of a thermostatic switch opening and closing an electric solenoid valve in the fuel supply line. A more elaborate system would consist of a pneumatic or air-operated controller opening and closing an air-operated control valve. Modulating or proportional control is best exemplified by an ordinary float valve, wherein the valve opening is a function of the liquid level, the lower the level the greater the valve opening. In a modulating or proportional temperature control system, valve opening or damper position is a function of temperature, either dry- or wet-bulb. The control valve or damper is normally neither fully open nor fully closed but is modulating a t some intermediate position. As a result, the controlled temperature does not cycle between limits as it does with the on-off control, but, remains steady once the system is stabilized.

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

335

Modulating control is applicable t o all types of heating systems commonly used in tunnel dehydrators. Proportional controllers are either electrically or air operated or self-acting. Fig. 12 shows a simple potentiometer type used in conjunction with a n electrically energized control valve. A self-acting controller is shown in Fig. 13. I n most plants the dry-bulb temperature is recorded continuously, either by use of a recording type controller or a n independent recording thermometer. Ordinary mercury thermometers are frequently used t o indicate both wet- and dry-bulb temperatures.

adjustment spring

y)Be'lows assembly

FIG. 12. Simple electrically-operated (potentiometer type) proportional temperature controller (courtesy of Minneapolis-Honeywell Regulator Co.).

4. Materials

of Construction

For obvious reasons, modern tunnel dehydrators heated by direct combustion are almost invariably built of fire-proof material such as hollow concrete block, hollow tile, sheet metal, or asbestos-cement sheeting. Most of the tunnels built on the Wrest Coast in recent years have been of hollow concrete or pumice block construction. Some of the early direct-fired driers still in use are built of wood, but they are fast disappearing. As might be expected, the use of wood and other flammable materials of construction increases the cost of insuring the structure against loss due t o fire.

336

P.

w.

KILPATRICK,

E. LOWE, AND

w.

B. VAN ARSDEL

F I G . 13. Self-acting temperature controller (courtesy of Taylor Instr'ument Companies).

Indirect-fired dehydrators are less vulnerable to damage by fire! and, therefore, can be built from a wider variety of materials. Usually, however, the materials of construction used are the same as those used in their direct-fired counterpart. Compartment type tunnel dehydrators are frequently of panel. construction, with wood or metal structure frame, and wood, asbestos-ce ,merit

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

337

board, or sheet-metal panels. The panels are usually insulated, with the thermal insulation applied between the two faces of the panels. 5. Trays and Trucks

Despite numerous attempts by designers and operators of tunnel dehydrators to find a better material of construction, drying trays used in most fruit and vegetable dehydration plants in this country are still made from wooden slats. The reasons are rather simple. A good drying tray

FIG.14. Wooden drying trays (court,esy of Gentry Division, Consolidated Fo as

Corp.).

must be easy t o fabricate, inexpensive (in terms of its probable USE ul life), easily scraped clear of adhering dried material, and light but strong and rigid. Furthermore, the tray material must not contaminate the product. Few materials besides wood can be used t o build trays which will possess a majority of the desired characteristics. All-metal trays, for example, are expensive, not easily fabricated, heavy in order t o be strong and rigid, and if fabricated with wire-mesh drying surfaces, tend t o develop a permanent sag with use. In some plants wooden trays enjoy an added advantage in that they can be readily fabricated by plant personnel during the off season and a t other times when the permanent members of the operating staff would otherwise be idle. A typical dehydrator tray is shown in Fig. 14. Most trays are 3 ft. by

338

P. W. KILPATRICK, E . LOWE, AND W. B . VAN ARSDEL

6 ft., made from Ponderosa Pine, Douglas Fir, or a combination of both woods. Wooden trays suffer from one serious drawback. Material drying on the trays tends to stick t o the wood surface. During the de-traying operation some wooden splinters may pull loose and stick to the product as it is scraped off of the trays. Most of the splinters are removed during final inspection of the finished product, but unless this inspection is painstaking, enough may remain in the dried material to pose a serious contamination problem. Elimination is especially difficult in the case of leafy

F

;ed

vegetables such as cabbage. Tooden slivers and pieces of dry produce with splinters adhering to them are usually removed from the dried product by hand, an expensive operation. To minimize product sticking and consequent pulling off of splinters, some operators oil or wax their wooden trays. Leafy vegetables such as cabbage are blanched on the drying trays. Moisture absorbed by the trays during the blanching operation increases the drying load in the dehydrator. To minimize the amount of water absorbed, some plants use wooden frame trays with wire-mesh drying surfaces for the blanching-drying operation. Drying trays are conveyed through the tunnels and to the tray loading and unloading stations on either of two types of vehicles-a flanged-wheel type that runs on steel rails (see Fig. 15), or a caster-wheel type that runs either on flat surfaces or in channel irons through the tunnels.

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

339

The rails for the flange-wheel trucks are laid flush with the concrete floor. To move trucks at right angles to their line of movement in the tunnels, turntables (see Fig. 16), or transfer cars and rails are used (see Fig. 17). Transfer rails are recessed so that the rails on the top of the transfer cars are flush with the tunnel rails. To change direction of movement, tray trucks are pushdd onto the transfer cars and moved at right angles to their previous direction of travel.

FIG. 16. Turntable for tray trucks (courtesy of Gentry Division, Consolidated Foods Corp.).

IV. TYPICAL COMMERCIAL TUNNELDEHYDRATORS 1. Twin-Tunnel CounterJlow Dehydrator

Popularly known as a Puccinelli dehydrator (R. L. Puccinelli, prominent in the development of prune dehydration in California during the 1920's and still actively engaged in the business), the simple twin-tunnel counterflow drier is widely used for drying both fruits and vegetables, par-

340

P. W. KILPATRICK,

E.

LOWE, AND W. B . VAN ARSDEL

ticularly on the West Coast. The basic elements of the drier are shown diagrammatically in Fig. 18, consisting essentially of a direct-fired combustion chamber and a blower, both located in an air passage between the two drying tunnels. Air enters the drier through openings surrounding the front of the combustion chamber, is heated to the proper drying temperature by the direct combustion heater, and is discharged by the blower into the two drying tunnels. Upon leaving the tunnels, part of the air is recirculated via the central air passage while the balance is exhausted to the atmosphere via overhead discharge ducts.

FIG.17. Transfer car and tracks (courtesy of Gentry Division, Consolidated Foolds Gorp.).

A variation of this twin-tunnel arrangement places the direct-fired heater and the blower in an air passage located above the two drying tunnels, which are arranged side by side. Being custom-built in most cases, Puccinelli-type dehydrators vary somewhat in size. A typical unit would have drying tunnels about 6 ft. 4 in. wide by 7 f t . high (inside dimensions), with a central air passage of equivalent height, and a width of 9 f t . A drier accommodating in each drying tunnel 12 truckloads of trays measuring 3 ft. by 6 ft., would have an over-all length of approximately 50 ft. This allows about 7 f t . for the air stream t o straighten out before entering the trays, after making the turn from the central air passage into the drying tunnels. About 5 f t . of space is left a t the opposite end of each tunnel for the air to enter the

P

3 /-FIG.18. Twin-tunnel counterflow dehydrator.

342

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

exhaust duct or t o be recirculated back into the central air passage. A shutter or sliding door is usually installed a t each dry-end air opening t o shut off the flow of hot drying air into the tunnels during loading and unloading operations. Total air movement and heating capacity vary from about 30,000 cu. ft. per min. and 3,000,000 B.t.u. per hr. for a fruit drier, t o 50,000 cu. f t . per min. and 5,000,000 B.t.u. per hr. for a vegetable drier. A typical up-to-date Puccinelli-type dehydrator would have hollow concrete or pumice-block walls, prestressed hollow concrete block roof slabs, and wooden-frame, metal-clad, center opening end doors. The drier would be equipped with a natural gas burner complete with a simple unlined sheet-metal combustion chamber, a direct drive tubeaxial blower with the motor cooled by a suction duct, wooden trays measuring 3 ft,. by 6 ft., flanged-wheel tray trucks running on rails set flush with the concrete floor, and a modulating dry-bulb temperature recording-controller. 2. The Miller Tunnel Dehydrator

The Miller dehydrator (L. N. Miller Dehydrator Company, Eugene, Oregon) is widely used in the Pacific Northwest for drying fruits such as

FIG.19. Miller tunnel dehydrator (elevation).

apples and prunes. The drier (see Fig. 19) is basically an indirect-fired, counterflow tunnel dehydrator, but of a special type. To create conditions believed t o be desirable for fruit drying, the Miller dehydrator is equipped with shutters or louvers located above the tray trucks in the middle third of the drier. The shutters are adjustable, t o vary the amount of air bypassing successive truckloads of product. The adjustable shutters make possible some measure of control of the humidity of the drying air in the various parts of the drying tunnel. By opening the shutters, for example, the relative humidity of the air a t the wet end of the drier can be increased. The resulting decrease in the rate of drying prevents the prune skins from

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

343

cracking due t o “case-hardening,” with consequent loss of juice from the fruit. Drying trays are usually of wood, measuring 3 ft. by 3 ft., and stacked 24 t o a truck. Truck arrangement within the drier varies. Depending upon drier capacity, trucks are either 2 abreast or 3 abreast, with 10 t o 12 trucks per row. A typical tunnel would, for example, hold 36 trucks, 3 wide by 12 long. The cross-sectional dimensions of such a tunnel would be 7 ft. 7 in. high by 9 ft. 8 in. wide. Gross air velocity in the empty tunnel, with the shutters closed, is usually not more than 500 ft. per min. Trucks are of the caster-wheel type, running on steel tracks while inside the drier and directly on the concrete floor when outside. The drier is of panel construction, with the insulated, galvanized, sheet-metal-clad panels bolted t o an angle iron frame. The indirect air-heating system is commonly of the combustion gasto-drying air type, consisting of a combustion chamber and flue pipes t o transfer the heat from the combustion gases t o the drying air. Steamto-air heaters are used, but are less common. 3. The Carrier Compartment Drier

During World War 11, the Carrier Corporation of Syracuse, New York manufactured a vegetable drier which in many ways is typical of compartment-type tunnel dehydrators used commercially. The drier consists essentially of a steam-to-air preheater section, 6 drying compartments or sections each equipped with its own blower and steam-to-air heater, and a n exhaust air fan section, arranged so that truckloads of trayed material are progressively pushed through the tunnel formed by the 6 compartments in series (see Fig. 20). Although the air flow across the trays is in a direction transverse t o the direction of truck movement, the drier is essentially a counterflow unit, with the wet material entering the drier a t the end where the exhaust air is discharged from the tunnel. Fresh air enters a t the opposite end, through the steam-to-air preheater. From the preheater, the partially heated air enters the steam-to-air heater and blower units of the individual drying sections. Perforated baffles progressively decrease the amount of fresh air taken in a t each compartment (thereby progressively increasing the amount of air recirculated a t each drying section), starting a t the dry end of the dehydrator and ending a t the wet end. The mixture of fresh and recirculated air is reheated a t each stage of drying, the amount of reheating automatically controlled by a separate dry-bulb temperature controller a t each compartment. The amount of air discharged from the drier by the exhaust fan, and consequently the amount of fresh make-up air entering the drier, is controlled by the wet-bulb temperature of the exhaust air.

344

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

I n operation, the product moving through the drier is subjected t o a different temperature condition at each compartment, and the direction of air flow over the material is reversed at each succeeding drying section. Although basically a counterflow unit, the Carrier dehydrator has some of the characteristics of a two-stage drier. Both air flow and drying temperature are higher at the 2 wet-end drying sections than they are a t the other 4 sections. This is made possible by using larger heater and blower units at the 2 wet-end compartments. EXHAUST

EXHAVST FAN UNIT

H

PLAN VIEW

STEAM-AIR P ~ E H E ~ T E R S

PERFORATED PLATE BAFFLE

TYPICAL CROSS SECTION FIG.20. Diagrammatic sketch of Carrier Compartment Drier.

Air velocities in the four dry-end and two wet-end sections are 700 and 1200 ft. per min., respectively, through the free area of the loaded trucks. When operated at maximum capacity under high ambient moisture conditions, fresh make-up air enters the drier a t a rate of approximately 20,000 cu. ft. per min. Steam consumption under these conditions is about 5500 Ib. per hr. Trays are either of wood or metal, measuring 36 in. by 36 in., and are stacked forty to a truck. The Carrier dehydrator has a nominal rating of 30 tons of wet vegetables per 24 hr., based on a tray loading of 1.2 lb. per sq. ft. and 2 hr. drying time to reduce the moisture content of the product t o a level suitable for bin finishing.

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

345

V. CRITERIAFOR SELECTIONOF TUNNEL DEHYDRATORS Numerous factors govern the selection of the drying system, or type of dehydrator, chosen for a particular plant. Economic factors, such as initial and installation costs and operating labor costs, although important, are not the only criteria for proper selection. Some of the other important considerations are listed below : (1) Flexibility with respect t o integrated plant operation. (2) Ability t o handle a wide range of commodities, if required or if the current economic demand changes. (3) Ability t o dry the product to meet current specifications for moisture content, product damage tolerances, etc. (4) Adaptability t o meet future contract specification changes for the product. (5) Floor-space requirements. (6) Capacity requirements, current and future; and the possible use of other final driers, such as finishing bins. (7) Availability of critical materials of construction and precision machine parts in case of emergencies. (8) Mechanical reliability and foolproofness t o guard against complete plant shut-down. Truck-and-tray type tunnel dehydrators are generally satisfactory for drying most of the various fruit and vegetable commodities which are processed in piece form. This type of drier can be used t o dehydrate a wide range of products and can be operated continuously or intermittently as desired. These factors add t o the flexibility of plant operation. The type of drier used influences, to some extent, the characteristics of the finished product. I n general, multistage driers permit the use of higher temperatures during the initial part of the drying cycle when the product’s surface is still moist, and consequently the product dries faster. This combination of higher temperatures and shorter drying time often produces a more porous and bulky material. The greater porosity of the finished product makes reconstitution faster and easier. However, the greater bulk may cause difficulty in meeting some contract specifications, i.e., getting the required weight in the containers. The counterflow tunnel is, perhaps, the most versatile of the truckand-tunnel driers for the dehydration of fruits and vegetables. These units are relatively easy t o operate and are of comparatively simple design. I n the fruit dehydration industry, the counterflow truck-andtunnel drier is the type most widely used, and the design has been more or less standardized (Perry, 1947, and Perry and associates, 1946). Tunnel-and-truck dehydrators with one, two and three stages may

346

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

be found operating satisfactorily in the vegetable dehydration industry, and they have been built by many different people. As previously mentioned in section I1 (Classification of Tunnel Dehydrators), the design has not been universally standardized. Broadly speaking, two-stage tunnel driers are preferable t o singlestage units for vegetable dehydration ij" the commodity has a very high percentage of water that can be removed rapidly (such as cabbage), or, if the commodity cannot tolerate a high final drying temperature (onions, 135 to 140" F.). The choice between one-stage and two-stage tunnel driers becomes more or less an arbitrary decision for operators drying commodities which have a relatively low initial moisture content and which can tolerate a reasonably high final drying temperature, for example, carrots. For a given capacity, the single-stage drier would occupy substantially the same floor space as the two-stage unit, provided the stages in the latter were not physically separated. Multiple-stage construction permits the use of drying conditions which change in a predetermined manner as the material progresses through the tunnel. This flexibility offers distinct advantages in operation. Moreover, the multistage unit, when compared t o a single-sbage dehydrator, provides more rapid drying of the product and somewhat better heat economy. Against the advantages for the multistage unit, there should be weighed the higher capital cost, the increased labor cost (unless an additional investment is made for an automatic mechanism to handle the trucks between stages), and the increased complexity of operation. Some of the justification for two- or three-stage operation is. also being weakened by the increased reliance upon finishing bins to accomplish the late stages of drying. There are also certain general basic considerations which should govern the selection of the drying system. Driers installed in multiples are less likely to cause a complete plant shutdown due to mechanical failure. Drying systems of proved or unquestioned performance have advantages since there is a calculated risk for each installation and the individual must decide how much of a pioneer he can afford to be. Initial cost of the equipment is often not as important in determining production costs, as is the effectiveness of the equipment chosen. For example, assume that two drying systems are available, one of which involves a sizeable capital investment, but has been proved t o be efficient and foolproof, and the other involves a modest investment, but is of doubtful efficiency and surety of operation. There is little question but that, generally, the more expensive unit should be given preference. Comparison of the truck-and-tray tunnel dehydrator with the continuous belt-conveyer dehydrator for a specific drying operation should take into account a t least the following general considerations.

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

347

Assuming both units have the same drying capacity under the specified conditions, the truck-and-tunnel arrangement has the following advantages : (1) The installed cost of the system will probably be less. (During the war, the cost of a simple tunnel was considerably less.) (2) Smaller quantities of critical materials are required for construction, and fewer precision-fabricated parts are needed (an important consideration in wartime). (3) Operation can be either intermittent or continuous. (4) A wider range of commodities can be dried satisfactorily. (For example, whole or halved fruit, such as prunes or peaches, are unsuited t o conveyer-belt drying because of the long drying time involved. Shredded cabbage is likewise unsuited because of the tendency for the blanched material t o mat and not dry uniformly.) On the other hand, the conveyor drier has certain advantages over the truck-and-tunnel dehydrator of' a similar drying capacity, for example: (a) Less floor space is needed. (b) The dryiiig time can be made shorter, and product quality may thereby be improved. (c) Less operating labor is required since the conveyer drier is fundamentally automatic in operation. OF TUNNEL DEHYDRATORS VI. BASIC THEORY

The quantitative theory of drying, applicable t o the design and control of dehydrators for fruits and vegetables, is the work of many investigators, nearly all within the past 50 years. Advances in development of the theory, especially pertinent t o the subject of this chapter, were made by Grosvenor, 1908; Carrier, 1911, 1921 ; Hausbrand, 1912; Tiemann, 1917; Lewis, 1921; Cruess and Christie, 1921b; Sherwood, 1929-1932, 1936; Newman, 1931a,b; McCready and McCabe, 1933; Bateman et al., 1939; Hougen et al., 1940; Van Arsdel, 1942, 1947, 1951a; Marshall, 1942, 1923; Brown and Kilpatrick, 1923; Cruess and hlackinney, 1943: Perry, 1944; Perry et al., 1916; Ede and Hales, 1948; Marshall and Friedman, 1950; Broughton and Mickley, 1953; and Hendel et al., 1954. Several of the earlier discussions of tunnel design, in the absence of quantitative information about the effects of temperature, humidity, air velocity, and other factors on the drying rates of specific commodities, simply assumed that the time required to dry a commodity was wellestablished by practice and on that basis computed the requisite air flow and heat input by methods established in the field of heating and ventilating engineering. This procedure of course forswore any adventure off well-trodden paths. Lewis (1921), followed by Sherwood and several of his other colleagues at the Massachusetts Institute of Technology and by a number of other chemical engineering investigators, noted that many wet materials exhibit two sharply distinct phases of drying be-

348

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

havior under constant drying conditions: an initial phase of constant rate of loss of water per unit surface exposed to the air, and a final phase during which the drying rate falls steadily toward zero. Generalized correlations of drying rate with vapor pressure relations and air velocity were derived, and these in turn were applied to the heat-balance and mass-balance equations characteristic of practical dehydrators. Good correspondence of prediction with experience was obtained in the drying of several industrial materials, and Perry (1944) and his co-workers used the same principles successfully in designing prune dehydrators. Van Arsdel (1942) and Brown and Kilpatrick (1943), concerned a t that time with the dehydration of vegetables to quite low levels of moisture content, found that drying rates determined experimentally could not be represented satisfactorily by any simple mathematical formula. They and their colleagues at the Western Regional Research Laboratory published a series of bulletins (AIC-3 1, I-VIII, 1943-1947) which summarized the drying behavior of a number of common vegetable materials in the form of nomographs readily applicable to dehydrator calculations. Broughton and Mickley (1953) have made the final step in the retreat from Lewis’ highly idealized picture of drying behavior by basing dehydrator design upon an actual analog” drying experiment in which the temperature and moisture-content history of the experimental material serves directly as the basis for the design. Their procedure obviates the difficulty, recognized but not fully overcome by earlier investigators, that the drying rate of a hydrophilic material at any instant depends to some extent upon the previous drying history of the sample (the internal distribution of moisture within the pieces is determined by that history). Some of the later AIC-31 nomographs contain a correction factor intended to deal approximately with this situation. Van Arsdel (1942), noting that fruit and vegetable dehydration presented a special case in which the heat absorbed in evaporation of water far outweighed all other causes of heat usage, proposed the following theorem: I n any section of a tunnel dehydrator where no reheating of the air takes place, the change in air temperature i s proportional to the change in moisture content of the material, if moisture content is expressed on the ,‘dry basis.” This is, expressed in differential form, dt dx

-

=

b -dw” dz

Under these special conditions the tunnel acts substantially like an adiabatic humidifier. It was already well-known that in such an adiabatic system the wet-bulb temperature of the air remains constant, and that, at

* Refer to list of symbols on p. 369.

TUNNEL DEHYDRATORS FOR FRUITS -4ND VEGETABLES

349

constant wet-bulb temperature, the fall in temperature of the air is very nearly a linear function of its rise in absolute humidity. A mass balance relating the loss of moisture by the material t o the uptake of water vapor by the air then led directly t o equation 1. Wet-bulb temperature lines on a humidity chart using temperatures as abscissas and absolute humidities as ordinates are slightly curved and differ slightly in slope. The actual variations within the general range of interest in fruit and vegetable dehydration are as follows: Wet-bulb temperature 90" F. : Fall in air temperature per 0.001 rise in absolute humidity: Air temperature 120' F., 4.28" Air temperature 180" F., 4.35" Wet-bulb temperature 120" F. : Fall in air temperature per 0.001 rise in absolute humidity: Air temperature 140" F., 3.81" Air temperature 200" F., 3.96" Van Arsdel (1942) suggested that for general exploratory computations a temperature change of 5" F. per 0.001 change in humidity be used; this would correspond t o a lumped total of about 20% for all heat losses; a t the same time, however, he proposed that wet-bulb temperature be taken as constant in spite of heat losses. If the 5" F. figure is accepted, the coefficient b in equation 1 can be evaluated readily, and the equation becomes,

The plus sign will be used for a parallel-flow arrangement, the minus sign for counterflow. Brown (1943) and Lazar (1944) examined this approximation critically, and concluded that for the range of conditions encountered in parallel-flow and counterflow tunnel dehydrators for vegetables the errors t o be expected from i t are smaller than the other inherent uncertainties of such systems. The wet-bulb temperature in practical tunnels will usually fall less than 1" F. between the hot end and the cool end of the tunnel. Perry (1947) computed the heat balance for a typical counterflow prune dehydrator, with a hot-end air temperature of 165" F . and wet-bulb temperature of 110" F . ; the wet-bulb temperature of the exhaust air was calculated t o be 109.6" F. The British Ministry of Foods, in its bulletin ((VegetableDehydration" (1946) described the performance of two-stage tunnels in potato dehydration; in the first stage the fall in air temperature * Refer to list of symbols on p. 369.

350

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

averaged 4.6" F. for each increase of 0.001 in absolute humidity, and the fall in wet-bulb temperature through the tunnel was somewhat less than 0.5" F. The following discussion of theoretical tunnel behavior is based on computations made by use of the approximate relationship of equation (2). I . Theoretical Tunnel Behavior Figure 21 shows the computed course of moisture content of prunes being dehydrated in a counterflow tunnel and the change in air temperature as i t passes through the tunnel. The conditions assumed were those 70 I-

z 0

60

(r:

W

a v, cn

c

3

I-

w

- I70

50

40

3

u

W

LT 3

0

30

20

2

130

10

0

6

12

18

TI ME, HOURS FIG.21. Moisture content of fruit and temperature of air in counterflow tunnel drying prunes, computed from drying-rate expression derived by Guillou (1942). given by Perry et al. (1946), in Figs. 14 and 15, and the drying rate expression used in the computation was t h a t published by Guillou (1942), which correlates the drying rate with air temperature, humidity, and velocity and with the size of the fruit. The computed curves agree well with the curves of Fig. 14 in the cited publication by Perry and his associates, which represent data secured from commercial counterflow tunnels. Figure 22 is a similar diagram showing computed conditions in a counterflow tunnel dehydrating white potato half-dice in. x in. x in. in the wet state). The conditions assumed for the example, which is taken from Van Arsdel's (1951b) Fig. 7, are substantially those followed in commercial dehydration. The drying rate expression used in the com-

x~

(x

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

351

putation is given by the group of nomographs in AIC-31-VII, Western Regional Research Laboratory (1945). The pictures presented by Figs. 21 and 22 exhibit the characteristic behavior of counterflow tunnels: maintenance of a relatively high air temperature through much of the tunnel, somewhat slow initial drying of the wet material, and good drying conditions a t the dry end of the tunnel. If the drying rate characteristics of a product have been determined, a single composite diagram can be used t o summarize the behavior of that product in counterflow tunnels. Figure 23 is such a diagram for the drying of prunes, published by Perry et al. (1916), Fig. 15. Figure 24 is a similar L

AIR TEMPERATURE

150 140

130 h IT

c

120

W

a E

110 W

2 0

E

0' 0

I 2

4

6

8

TIME, HOURS FIG.22. Moisture content of material and temperature of air in counterflow tunnel drying potato half-dice, computed from drying-rate nomographs of AIC-31-VII.

diagram for the counterflow drying of potato half-dice, published by Van Arsdel (1951b), Fig. 8. The general similarity of the two diagrams is striking, only the time scales being substantially different. The characteristics of a parallel-flow tunnel are, of course, the reverse of those of a counterflow tunnel; a relatively low air temperature prevails throughout much of the tunnel, very rapid evaporation occurs a t the wet end, and drying conditions are poorest a t the dry end of the tunnel. Van Arsdel (1951b), Fig. 10, illustrates the contrast in a n example reproduced in Fig. 25, showing the behavior of counterflow and parallelAow tunnels of the same length and same air flow, operated in such a way as t o produce the same hourly output of dehydrated potato half-dice. Van Arsdel points out t h a t whereas the most striking contrast is in the heat consumption of the two arrangements (almost 50% greater for the parallel-flow tunnel) there will also certainly be differences in the quality characteristics of the products turned out. Material dried in the parallel-

352

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

Net tunnel length, ft.

FIG.23. Relation between air velocity and drying time, with final exhaust temperature also given, for counterflow prune dehydrators of various lengths. Initial air temperature 165" F., wet-bulb temperature not over 105" F. Initial prune moisture content 70%, final prune moisture content 16.7%. Prune size, dry count of 50 per pound (from Perry et al., 1946, Fig. 15).

4

5

6

7 8

9

10 I I

12 13 14 15

16

Number of active trucks in tunnel

FIG.24. Relation between air velocity and drying time, with final exhaust temperature also given, for counterflow potato half-dice dehydrators of various lengths. Initial air temperature 150" F., wet-bulb temperature 85" F. Initial moisture content of material 76% final moisture content 6%. Tray-loading, 1.50 lb. per sq. ft., trucks contain 540 sq. ft. active tray surface (adapted from Van Arsdel, 1951b, Fig. 8).

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

353

flow tunnel will be the more bulky because shrinkage stresses will open up internal voids in the pieces. Some difference in the extent of "heat damage " suffered during dehydration may also be apparent, damage probably being slightly greater in the counterff ow-dried material. If only the first stage in a multiple-stage dehydration system is being considered, there can be no doubt that the parallel-flow arrangement offers substantial advantages over counterflow. I n such a system product I90

180 W'

5

5

5 5 IQ

170 160 150 140

130 120 110

80

cn v,

70

a

m

60

I-+

50

w z

40

30

I-

cn0

I

20 10 n

" 0

I

2

3

4

5

6

7

T IME, HOURS FIG.25. Comparison of counterflow and parallel-flow drying; moisture content of material and temperature of air in tunnels drying potato half-dice, under conditions chosen to make outputs equal. Twelve trucks in each tunnel, 540 sq. ft. per truck, loading 1.50 lb. per sq. ft. Air velocity 1000 ft. per min. between trays. In counterflow drying, initial air temperature 150" F., wet-bulb temperature 85" F. In parallel-flow drying, initial air temperature 185" F., wet-bulb temperature 90" F. (from Van Arsdel, 1951b, Fig. 10).

is discharged from the first-stage tunnel still somewhat moist; the relatively poor drying conditions a t the dry end do not matter so much. The evaporative capacity of the tunnel can be increased substantially by raising the temperature of the inlet air t o a point far above that which would be safe for a counterflow tunnel. Van Arsdel (1951b) presents an example of first-stage driers for potato half-dice in which a parallel-flow tunnel, operated a t an inlet temperature of 200' F., turns out almost 50%

354

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

more material dried to 16.7% moisture than a counterflow tunnel of the same length, operated a t an inlet temperature of 145" F. Two-stage tunnels now generally consist of a parallel-flow first stage and counterflow second stage, the arrangement first suggested by Eidt (1938) for the dehydration of apples. The course of moisture content and air temperature in such a two-stage tunnel, as used for the dehydration 220 0 -

2

200

180

.=I

160

5 n 2

120

I-

100

W

m

Lg

-0 3w W E

5: Is

0 I

140

80 80

70 60 50 40

30 20

10

0

0

1

2

3

4

5

6

7

TIME, HOURS FIG.26. Moisture content of material and temperature of air in a, two-stage tunnel drying potato half-dice. Eight trucks (540 sq. f t . of tray surface in each) in primary stage, 16 trucks in secondary. Tray loading, 1.50 Ib. per sq. ft. Air velocity between trays, 1000 ft. per min. (from Van Arsdel, 1951b, Fig. 15).

of potato half-dice, are shown in Fig. 26, taken from a n example computed by Van Arsdel (195lb), Fig. 15. I n this case the secondary tunnel is twice as long as the primary. Van Arsdel showed that the combination, carrying altogether 24 truckloads of material, should produce about 7 % more product dried t o 6 % moisture than two 12-truck counterflow tunnels and use about 5 % less fuel. T o offset these advantages, the capital cost of the two-stage dehydrator would be somewhat greater, especially if truck transfer between stages were handled automatically, and the control of the unit would be more complex and critical with respect to maintenance of product quality.

T U N N E L DEHYDRATORS FOR F R U I T S AND VEGETABLES

355

The ratio of lengths of the two stages of this type of dehydrator apparently can vary within rather wide limits without much effect on performance. Commercially they have ranged between equality in length and 2 t o 1. Theoretically, for the example of Fig. 26, maximum output would have been realized if the total of 24 trucks were divided 7 t o 1, that is, 3 trucks in the parallel-flow primary, 21 trucks in the counterflow secondary; however, the gain in output would have been negligible, and operation of the 3-truck primary would make extreme demands on control and scheduling of the system. None of the theories referred t o above is directly applicable t o study of the operation of transverse-flow compartment tunnels. I n commercial forms of these dehydrators there may be as many as 6 or 8 compartments in which the air temperature is individually controlled a t desired levels. A truckload of material standing a t any one of these positions is exposed t o unvarying drying conditions until i t is shifted t o the next position, where the direction of air flow is reversed and a new set of drying conditions is maintained. No satisfactory mathematical formulation has been proposed for this kind of system. Prediction of the performance of a transverse-flow dehydrator, with given values for the air flow and air temperature a t each compartment, could presumably be accomplished by means of a n analog type of drying experiment somewhat similar t o t h a t proposed by Broughton and Mickley (1953). 2. Optimum Tray Loading

Drying-rate experiments have invariably shown that the rate of drying of wet materials spread on trays decreases as the load of material on the trays increases, once the load exceeds appreciably that of a single layer of pieces on the tray. The drying time is therefore shortest for light tray loadings. But for a fixed area of tray surface in a dehydrator, the output of produce is proportional t o the tray loading. Net dehydrator capacity is the resultant of these two effects. Quantitative estimations have been made of the effect of tray loading on dehydrator output, for 2 quite different vegetable products, potato half-dice and cabbage shreds. The drying-rate data are those of the ATC-31-VII and AIC-31-IV nomographs. Trucks in the tunnel postulated for these examples each contain 400 sq. ft. of useful tray surface. The mass air flow through the tunnel is 2000 lb. of dry air per min. and the air velocity between trays a t the wet end of the tunnel will be 900-1000 ft. per minute. For cabbage, a hot-end temperature of 140" F. and a wet-bulb temperature of 90" F. are assumed in counterflow drying, 180" F. and 95" F. in parallel-flow drying; for potatoes, a hot-end temperature of 150' F. and a wet-bulb temperature of 90" F . in counterflow drying,

356

P.

w.

KILPATRICK,

20 -

E. LOWE, AND

w.

B. VAN ARSDEL

PARALLEL-

CABBAGE SHREDS

0.5 I.o 1.5 2.0 TRAY LOADING, LBS./S 0. FT. FIG. 27. Effect of tray-loading upon tunnel capacity in dehydrating cabbage shreds; counterflow tunnel, 9 trucks, drying to 4.75 % moisture; parallel-flow tunnel, 12 trucks, drying only to 9.1 % moisture. 0

35

t

I

PARALLEL-

-

POTATO HAL F-DICE

0" 0

I

2

3

T RAY L 0AD 1NG, L BS./ S 0.F T FIG.28. Effect of tray-loading upon tunnel capacity in dehydrating potato halfdice; counterflow tunnel, 10 trucks, drying to 6.55 % moisture; parallel-flow tunnel, 12 trucks, drying only to 16.7% moisture.

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

357

200' F. and 100" F. in parallel-flow drying. Final moisture content, for counterflow operation, is 4.75% (wet basis) for cabbage, 6.55% for potatoes; for parallel-flow operation (used as the first stage in multiplestage dehydration), 9.1 % for cabbage, 16.7% for potatoes. Figures 27 and 28 show estimated tunnel capacity for dehydration of cabbage shreds and potato half-dice (tons of prepared material per 24-hr. day) as a function of tray loading, pounds of prepared material per square foot. Tunnel capacity reaches substantially a maximum at a loading of about 1.5 lb. per sq. ft. for the cabbage shreds, something over 3 lb. per sq. ft. for the potato half-dice. The figures for drying time marked on the curves indicate that, if the tunnel is being operated near its maximum loading, a material shortening of drying time may be accomplished with only minor sacrifice of capacity by lightening the tray loading. If heat damage is being experienced, this may be a n important remedial measure. 3 . Optimum Recirculation of

Air

The reasons for, and advantages and disadvantages of recirculation of air in a tunnel dehydrator have been discussed by Van ArsdeI (195la, pp. 80-84). Recirculation raises the wet-bulb temperature of the air, returns some heat t o the system and thereby saves fuel, but a t the same time reduces the drying rate of most materials. Purely from a drying cost standpoint, the saving in fuel cost must be balanced against the decrease in production and the attendant increases in other unit costs. Ramage and Rasmussen (1943) noted that there should be some proportion of recirculation t h a t would give the minimum drying cost per pound of product, and they computed this optimum for one simple set of conditions. So many combinations of the numerous variables are possible that no single general principle has been established. The following example illustrates the procedure and typifies the kind of results that map be computed A counterflow tunnel long enough t o hold'a maximum of eleven %foot trucks (400 sq. ft. of useful tray surface on each truck) is t o be used t o dehydrate potato half-dice t o 6.55% moisture (wet basis) or cabbage shreds t o 4.75% moisture. Trays will be loaded with 1.40 lb. of prepared cabbage or 2.50 lb. of prepared potato per sq. f t . The fan supplying the air flow through a pair of these tunnels is of the "limit-load" type, double width, with a standard air rating of 54,000 cu. ft. per min. against 1.5 in. static pressure, when operated at 423 r.p.m. and absorbing approximately 17 h.p. The hot-end temperature for the cabbage dehydration will be 140" F., for the potato dehydration 150" F. The power absorbed by the fan a t a temperature of 140-150" F. is approximately 15 h.p. The outside fresh air temperature is 60" F., absolute hu-midity is 0.0100 lb. of water

358

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

vapor per lb. dry air. The tunnel will be operated with a minimum of a 15"-wet-bulb depression a t the cool end, and, if a n increase in recirculation would result in a lower wet-bulb depression than that, the load on the tunnel will be reduced by decreasing the number of trucks in it. The cost of heat is taken as 35# per million B.t.u. transferred t o the air stream, the cost of power 26 per kw.-hr., operating labor cost for each side of the double tunnel $1.45 per hr. regardless of output within the range considered in the example, and all other costs (plant overhead and fixed charges) $1.35 per hr.

c

t

t

7 OTHER

LABOR

POWER

n "

00

90

100

W E T-BULB

0

Q20

0.40

0.60

0.00

1.00

PROPORTlON CF AIR REClRCULATED

FIG.29. Effect of air recirculation upon drying costs: shredded cabbage, counterflow tunnel.

The computations involve estimation of drying time for selected values of the wet-bulb temperature, using the AIC-31 nomographs, and then deriving the corresponding values of tunnel output, proportion of air recirculated, and necessary heat input. The approximations described by Van Arsdel (1951a) were used. The results of the computations are shown graphically in Figs. 29 and 30. It is evident from these curves that, for the combination of conditions chosen, the minimum total cost for drying both vegetables would be realized by employing little or no recirculation of the air; the saving in cost of heat through recirculation would be more than offset by the increases in other costs. Quite a different result would have appeared if a material like prunes were being dehydrated. According to Guillou (1942) and Perry (1944), the drying rate of prunes is substantially independent of the relative humidity of the drying air unless the relative humidity exceeds about 35 %. T ha t being the case, raising the wet-bulb temperature by increasing

359

TUNNEL DEHYDRATORS F O R F R U I T S AND VEGETABLES

the proportion of recirculation within reasonable limits will decrease the cost of heat materially without a t the same time increasing other costs.

/

8 TRUCKS

TOTAL

8 TOTAL

z

w o 0

I

POWER

80

100

120

WET- BULB TE MP ERAT URE

0

0.20

0.40

0.60

a00

1.00

PROPORTION OF AIR REClRCULATED

FIG.30. Effect of air recirculation upon drying costs: potato half-dice, counterflow tunnel.

No data are available t o indicate whether other fruits, such as grapes and sliced apples, will behave more like the cut vegetables than like prunes.

4. Product Temperature in the Dehydrator Perry (1944) and Perry et al. (1946) have published measurements of the internal temperature of prunes during dehydration. Figure 31, taken

I-

100 } ~

0

6

I2

18

24

TIME, HOURS FIG.31. Air temperature and fruit temperature in a typical counterflow prune tunnel (from Perry et al., 1946, Fig. 14).

from the second of these references, illustrates the typical course of fruit temperature during counterflow dehydration, with the air temperature

360

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

also shown for comparison. There appears to be no tendency for the fruit temperature t o hold for a period of time a t or near 105" F., the wet-bulb temperature of the air. No similar measurements of material temperature of vegetable pieces undergoing dehydration have been published, in spite of the obvious importance of the time-temperature relationship t o the quality of the dry product. The following procedure leads t o a rough approximation t o the material temperature under typical tunnel conditions. There is some evidence that during early stages of drying the shrinkage in volume of vegetable pieces very nearly equals the volume of water lost by evaporation, but in later stages the volume shrinkage is less, and no substantial further decrease in volume occurs as the pieces dry below about 15 t o 20% moisture. If the density of the dry substance is 1.25 g. per ml., the piece area during early stages should change as follows:

(3) Study of drying rates for potatoes shows th at this relationship holds down t o about w = 1.50 lb. of moisture per lb. dry solids (60% moisture) for potato pieces. A graph of area versus moisture content can be extrapolated with only moderate curvature t o an expected final dry area of 43 % of the original area. If it is now assumed that the drying rate in very early stages maintains the material a t the wet-bulb temperature of the air by convective heat transfer (radiative and conductive transfer being neglected), a convective heat transfer coefficient can be computed from the following relationship:

-H,

~.Lo

w,,+ 1

dw -

(4)

d6

The transfer coefficient k is expressed here in terms of unit area of tray surface. Now if it be further assumed th at this transfer coefficient will remain unchanged throughout the entire tunnel, allowance being made for the area shrinkage of the material, the piece temperature will be given approximately by the following equation : T = t f

HL,

dw*

-+ 1) '2Z

i%a(w,

(5)

Material temperatures computed in this way, and the corresponding * Refer to list of symbols on p. 369.

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

160

g

L

,

4 120

E

220

I \

2oo-l \

PRIMARY

7

180 -

w

[L

3

160-

\ \ \

SECONDARY

STAGE

','

STAGE

AIR TEMPERATURE

WET-BULB I

TEMPERATURE

361

362

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

can easily lead to the serious quality defect variously known as “heat damage,” ‘‘ caramelization,” L‘scorching,’l or simply “browning.” A product which has been damaged in this way by the drying step may be unacceptable as a food product because of off-color and off-flavor. The phenomenon, by whatever name called, is the result of a highly complex system of reactions in and among the natural constituents of the fruit or vegetable. Its chemistry is beyond the scope of this article. T h e two points of importance here are these: (a) the reactions leading t o (‘browning” or ‘(scorching” have a very large positive temperature coefficient; (b) the reaction rate increases as the concentration of the components increases (i.e., as the moisture content of the product is reduced), but water itself must be involved in the reactions in some way, because when the moisture content is reduced far enough the rate of browning becomes very low. The result is that, in some intermediate range of moisture content, the rate of browning a t any given temperature goes through a maximum. Studies of the kinetics of this system have been reported by Hendel et al. (1954). For white potato the maximum browning rate occurs in the range of 15 t o 20% moisture content. When sufficient quantitative information becomes available, it may for the first time be possible t o devise dehydrator-operating procedures for an optimum combination of high output and minimum heat damage. 5. Departures from Theory The foregoing discussion of tunnel theory ignores several complications of actual dehydrator operation. One of these is inherent in the nature of a tunnel-and-truck drier, whereas the others may be classed as inevibable imperfections of design and operation. The simple theory th at has been presented assumes that the tunnel is loaded continuously and operates in the steady state so that the material moisture content and the air temperature and humidity a t any given point along the length of the tunnel will remain unvarying. The fact that truckloads of material are introduced a t finite intervals of time and are advanced through the tunnel in discrete steps, introduces a complication for which no mathematical description has been proposed. The temperature t o which a product is exposed in its progress through the tunnel changes discontinuously, not along a smooth curve. This so-called “sawtooth effect” has been discussed by Van Arsdel (1951a). Practical experience suggests t hat its effect cannot be very great, except possibly in short parallel-flow tunnels. The imperfections of construction and operation, on the other hand, sometimes may have very serious effects. The commonest causes of trouble are poor air distribution, temperature stratification in the air,

T U N X E L DEHYDRATORS FOR FRUITS AND YEGETM3LES

363

uneven loading of trays, and poorly designed or sagging trays. Even with good design and careful maintenance and operation, these troubles will always be present t o some extent. Since they obviously cannot be embodied in the theory in any quantitative way, the dehydrator operator must learn by experience how much of a safety factor he must apply t o the predictions of dehydrator theory.

VII. OPERATINGPROCEDURES FOR TUNNEL DEHYDRATORS Methods for dehydrating fruits and vegetables have been and are constantly being improved, for competition is keen and i t is necessary t o increase product quality and general efficiency. Although the dehydration a r t is well established, the drying characteristics of most vegetables and fruits are known only in rather general terms. Conditions and operation procedures differ from plant t o plant, and general recommendations may not apply t o a particular dehydrator. Therefore, only certain items of general interest will be brought t o the reader's attention, and details of operating a tunnel drier will be described only briefly. From the discussion which follows, it will become apparent that, t o achieve best results, each operator must take into consideration the physical setup and conditions present in his plant and use his initiative to secure the best combinations for high product quality, capacity, and over-all efficiency. As illustrated by the series of curves (Figs. 23 and 24), and as previously mentioned, if the tunnel length, air velocity, and initial temperature are fixed, then the characteristics of the material being dried will determine the drying time and final air temperature. The original authors have pointed out that even though the curves apply quantitatively only t o a particular set of conditions, the character of the curves is such that broad generalities can be deduced. Some of these generalized facts are pertinent to the operation and proper use of tunnel driers and merit discussion. If the commodity being dehydrated is a slow-drying material, the temperature drop per foot of tunnel length will be small and long tunnels can be used. If the tunnel is relatively short, the temperature drop through i t will be small, and unless air recirculation is used, the heat efficiency will be low. On the other hand, if the commodity is a fastdrying material, the temperature drop per foot of tunnel length will be large. I n this case, if a long tunnel is used, it should not be completely filled with trucks, otherwise drying conditions will be unsatisfactory a t the cool humid end. However, if the number of trucks in the tunnel is increased, the tunnel output capacity will be increased (assuming there is no change in the air velocity), but a t the possible expense of injury t o the product. The latter is particularly true, since both the drying time and

364

P. W. KILPATRICK, E. LOWE, AND W. B . VAN ARSDEL

possibility of other unfavorable drying conditions will be increased. Therefore, it is apparent there is an optimum balance for each different operating condition. The series of curves referred t o previously, also illustrate another fact. If the number of trucks is increased, and a t the same time the air velocity is increased just enough to keep the wet-bulb depression unchanged a t the wet end of the tunnel, two desirable things occur. The product drying time becomes shorter (a tendency that favors product quality), and the output of the tunnel increases. Thus the operator must use good judgment and balance the gains in increased capacity and better product against the substantial increase in the power required t o run the blower. It should be remembered that power consumption increases approximately eight-fold if the air velocity is doubled, even though the number of trucks remains the same (see pp. 326-332, Fans and Blowers). Power consumption will be even greater if the number of trucks is increased simultaneously. Although several tunnels in a dehydration plant may be of identical design and be drying the same commodity t o a given moisture content, the drying time will inevitably differ somewhat from tunnel to tunnel. Also there may be even a day t o day difference in the drying time for a given tunnel. Small variations in drying time can be caused by slight differences in the amount of recirculated air, effectiveness of air distribution in the tunnel, hot-air temperature, uniformity of tray loading, atmospheric conditions at the fresh air intake, and last, but not the least important, are the variations in the nature of the prepared commodity. The effect of these variables has been discussed in TJnited States Department of Agriculture Miscellaneous Publication No. 540 (1944), and by Van Arsdel (1951a). Although the drying times will differ as indicated, there will be a mean time interval for each truck handled. The preparation line must be geared t o that rate, and even then there will occasionally be either a shortage or an excess of loaded trucks at the dehydrators. Several tunnels may be needing trucks at the same time, and a little later other loaded trucks may begin t o accumulate because no tunnel is ready to accommodate them. Regardless of this situation, a crew a t the dry end of the tunnel must be ready t o remove the trucks of dried product whenever the cars are scheduled t o be pulled. It is necessary t o keep the dry-end foreman informed as t o the exact time each wet truck load is placed in a tunnel, so that he can add the probable cycle times and schedule the pull times. This calls for some kind of message or signal system plus a running log sheet a t the dry end of the tunnel on which the probable pulling times can be entered. If the preparation line must be slowed down for a number of hours, the drier foreman needs advance notice, for he may have t o change

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

365

drier conditions to compensate for partly empty tunnels. A more complex scheduling system is needed in a plant which operates two-stage tunnel dehydrators, than in one which uses only counterflow units. Operation of a tunnel dehydrator under normal conditions presents few difficulties. At the beginning and end of a processing run, certain modifications of tunnel drying conditions are usually necessary t o prevent product injury during the starting-up and shutting-down periods. There are various procedures for starting and shutting down tunnel dehydrators, as discussed by Perry and associates (1946), and as explained in the United States Department of Agriculture Miscellaneous Publication No. 540 (1944). Exact procedures will, of course, vary from plant t o plant. Essentially, the methods consist of readjusting either the amount of recirculation air, the air temperature, or rate of tunnel loading, or adjusting and balancing combinations of these factors which affect the drying conditions within the tunnel. The procedures are a necessary operating requirement, not only t o minimize excessive heat damage t o the product, but also t o avoid other undesirable drying conditions, which may in turn, cause other types of quality deterioration before the product is dried. As a n example of the latter condition, let us assume th a t several trucks or cars of wet material are placed in a counterflow tunnel on start-up. The evaporation rate from the first truck would be great, and the drop in temperature of the air passing through th a t car would therefore be high. The air moving through the last truck would be nearly saturated. The product in th at car would be heated u p rapidly t o the wet-bulb temperature of the air, and little evaporation would take place until the first cars of material had lost a considerable part of their moisture content. Such drying conditions would be very undesirable, and probably adversely effect the product quality. For example, under some conditions, there might be a n excessive loss of ascorbic acid, or rapid growth of microorganisms and attendant spoilage. On the other hand, as the leading truckload of product progressed through the tunnel, it would be subjected t o a n abnormally hot, untempered blast of air during its entire time in the tunnel. If the product happened t o be heat sensitive, then heat damage might result. Furthermore, each succeeding truck would be subjected to nonuniform drying conditions as the tunnel was progressively filled. In order t o avoid the difficulties mentioned, one of the start-up procedures for counterflow tunnels uses a modified hot-end air temperature schedule, and loaded trucks are pushed into the tunnel a t the normal or regular scheduled rate. Another method fixes the hot-end air a t the normal operating temperature, and a t first the loaded trucks are rolled into the tunnel a t a faster rate than the normal schedule. The time interval between introduction of cars is progressively increased until normal

366

P. W. KILPATRICK, E . LOWE, *4ND W. B . VAN A R S D E L

operating conditions are reached. If the latter method is used, scheduling can become rather complex if there are numerous tunnels. Regardless of the starting method used, as the pull-time approaches for the first car, the dried material is usually examined and the drying schedule modified if necessary. Experience shows t h a t a trained operator can estimate the moisture content of the warm, nearly dry material rather closely b y “feel.” One of the authors, during the course of research work on vegetable dehydration, developed this knack and could readily estimate moisture contents of dehydrated products, such as cabbage, carrots, white potatoes, sweet potatoes, onions, spinach, corn, peas, and green beans. Normally, the uncertainty of the estimates, as shown by subsequent assays, was about 1% moisture content. Terminating a run in a counterflow dehydrator is relatively simple. As the evaporative load in the tunnel decreases, there will be a decline in the wet-bulb temperature of the air if recirculation is being used. Ordinarily this condition will not result in product injury, but may carry the material t o a lower moisture content than desired. To prevent this, some operators readjust the recirculation damper, and by doing so also save some fuel. Sometimes the tunnel temperature is lowered slightly, or the last trucks are removed ahead of schedule. I n the first stage tunnel (or parallel-flow section) of a two-stage dehydrator, drying conditions are rather critical during the start-up and shutdown periods. During normal operation, the controlled air temperature a t the hot end is usually above 180” F., and, in some cases, may be set in the region of 250” F. Most of the original moisture is removed from the product while it is in the first-stage drier. The material is usually dried t o a moisture content of about 50% in that section of the dehydrator. This corresponds t o a moisture removal of roughly 75 t o 90% for most commodities. However, the evaporative load in the tunnel is obviously light during the start-up period. To retard excessive drying of the product’s outer surface and prevent the consequent possibility of case hardening, heat damage, or even scorching, it is necessary t o keep the wet-bulb temperature of the air a t a reasonably high level. Therefore, during the starting-up period, it is necessary t o readjust the recirculation damper carefully, so as to maintain normal operating wet-bulb conditions in the parallel-flow tunnel section. Readjustment of the damper may be required after each truck enters that tunnel section, and until the section is full and normal operating conditions are reached. Likewise, when operations are being terminated, good judgment must be used in shutting down the first stage. Near the end of a run, as filled trucks are replaced by empties, the evaporative load decreases, and consequently, since recirculation is normally used, the wet-bulb temperature of the air will

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

367

begin t o drop. Instrumentation usually controls the dry-bulb temperature of the air. Unless the instrument is manually reset, the air will remain at its constant high temperature level. If changes are not made, the last filled truck passing through the first stage will be subjected t o a constant high dry-bulb temperature and a continually decreasing wet-bulb temperature. Unless controlled, there will be a cumulative effect t h a t will produce higher and higher drying rates within each car of the last tunnel load. This may result in case hardening, discoloration, scorching, or other product injury. Corrective measures are relatively simple ; for example, the hot-end, dry-bulb temperature can be progressively decreased stepwise, as empty trucks replace full ones. Theoretically, each temperature drop should be equal t o the average temperature drop across a truck. The latter can be estimated from the known operating data for that particular first-stage tunnel (normal hot-end temperature minus normal cold-end temperature divided by the number of trucks in t h a t tunnel section). I n a similar manner, the wet-bulb can be kept a t a relatively constant level by adjusting the recirculation damper. VIII. RECENTTRENDS IN TUNNEL DEHYDRATION OF FRUITS AND VEGETABLES The vegetable dehydration industry has gradually come t o rely more and more on the use of bin-type finishing driers. Their operation, general use, and design are discussed in United States Department of Agriculture Miscellaneous Publication No. 540 (1944), and Bureau Publication AIC-15 (1943, Revised 1944). The advantages obtained through the use of finishing bin driers are numerous. The bins provide a low-cost method for removing the moisture from the product during the slow-drying stage near the end of the dehydrating process, and they permit close control of the product’s moisture content during the final stage of drying. The bins also provide a n economical means for holding the product and equalizing its moisture content before the material is packaged. Dried apples and prunes are commonly stored for some time in bins after drying for this same purpose. I n addition, a properly designed bin finishing system can materially increase tunnel capacity and improve operation flexibility. I n some plants the bins are made portable for convenience. additional flexibility of operation, and also, they reduce conveyer equipment requirements. Sometimes dehumidified air is supplied t o the bins to expedite the drying process. The theory of through-flow drying in the low-moisture region, as applied t o bin finishing driers, has been discussed by Van Arsdel (1953). I n order t o make the dehydration operation less critical with respect t o possible heat damage or other injury t o the product, morKattention is

368

P.

w.

KILPATRICK, E. LOWE, AND

w.

B . VAN ARSDEL

being directed to the quality and preparation of the raw material. For example, consideration is now being given to the control of the sugar content in potatoes. Campbell and Kilpatrick (1945) demonstrated that sensitivity to heat damage during dehydration is a function of the reducing sugar content in the potatoes. It has been shown that blanching is necessary during the preparation of most vegetables that are to be dehydrated. The two primary reasons for this requirement are: (I) to prevent or check the development of undesirable colors, flavors, and loss of vitamins which occur during dehydration and subsequent storage, and (2) to obtain a finished product which will rehydrate readily, cook rapidly, and yield a cooked product of desirable texture and flavor qualities. There are no data available which directly correlate the degree of enzyme inactivation and quality retention of dehydrated products during storage. However, most investigators agree that blanching to a degree sufficient to inactivate the peroxidase enzyme systems present in the various commodities is sufficient for quality retention. Blanching much beyond this point is undesirable due to the loss of soluble nutrients as a result of leaching caused by extended exposure of the product to the blanching medium. I n order to prevent undesirable changes during dehydration and storage, there is a recent trend to treat the raw prepared material with additives. Sulfur dioxide has, of course, been used for this purpose for many years. Some of the new additives may have certain advantages over sulfur dioxide. As examples of the new additives, starch has been used to coat prepared carrots. Compared with sulfur dioxide, the starch gives carrots equal or better protection from loss of carotene and color during subsequent storage, according to Masure and associates (1950). Treatment of potatoes with calcium chloride to control sloughing and reduce heat damage during dehydration was suggested by Campbell and Kilpatrick (1945). Experimental confirmation of the favorable results of this treatment has recently been presented by Simon et al. (1954). The same authors (1953) proposed the use of thinner pieces, treated with calcium chloride to control sloughing, in order to speed up the drying and thus reduce heat damage. In effect, most of these procedures increase the drying capacity of the dehydrators by permitting the use of higher drying temperatures. The mounting cost of labor has caused a trend toward the use of labor-saving devices in the operation of dehydrators. Semi-automatic and automatic tray loaders, stackers, and unloaders have been recently developed for use in conjunction with tunnel drier operations in the dehydration industry, and they have been used with apparent success. Automatic truck transfer equipment is being used to handle cars between staged tunnels. Reducing the complexity of operation tends to alleviate

T U N N E L DEHYDRATORS FOR FRUITS AND VEGETABLES

369

the labor problem. Therefore, there is a tendency t o favor the simple single-stage tunnels, rather than the more complex multistage units. Some vegetable dehydration plants even prefer t o operate their tunnels without recirculation of air, sacrificing fuel economy in order t o simplify the job for the tunnel crew. As pointed out in another section of this chapter (pp. 357-359), the drying characteristics of cut vegetables are such that this practice probably entails little or no increase in total drying cost. The vast majority of the fruit and vegetable dehydration plants in the United States use truck-and-tunnel driers in preference t o other type dehydrators. A small number of conveyer driers is in use, and the popularity of this type of dehydrator appears t o be increasing. However, for many years t o come, the tunnel drier will continue t o serve as the dependable workhorse of the fruit and vegetable dehydration industry. The authors believe it will continue t o occupy a foremost place in the further development of the industry.

IX. LIST OF SYMBOLS USED A-Evaporating

area of a piece, square feet.

A A 83XL, -Coefficient in tunnel equation, = G(w, + I ) . G--Mass air flow, pounds dry air per minute. 11-Latent heat of evaporation, B.t.u. per pound. /<-Heat transfer coefficient,,B.t.u. per hour per degree F. per square foot of tray surface. L-Tray loading, pounds per square foot. ,\'--Total tray area in the tunnel, square feet. ?'-Material temperature, degrees F. t --Air temperature, degrees F. w-Moisture content of the material, pounds water per pound dry solids. %-Distance along tunnel, feet. 0-Time, hours. Subscripts : o-Initial conditions. -At wet-bulb temperature. a-Area

as a proportion of the original area,

=

0

REFERENCES Bateman, E., Hohf, J. P., and Stamm, A. J. 1939. Unidirectional drying of wood. Ind. Eng. Chem. 31, 1150-1154. Broughton, D. C., and Mickley, H. S. 1953. Design of full-scale continuous-tunnel drier. Chem. Eng. Progr. 49, 319-324.

370

P. W . KILPATRICK, E. LOIVE, A N D W . B . V A N ARSDEL

Brown, A. H. 1943. Heat balance in tunnels. Unpublished report. U. S. Dept. Agr., Bureau Agr. Ind. Chem., Western Regional Research Lab., Albany, Calif. Brown, A. H., and Kilpatrick, P. W. 1943. Drying characteristics of vegetables-riced potatoes. Trans. Am. SOC.Mech. Engrs. 66, 837-842. Campbell, H., and Kilpatrick, P. W. 1945. Effect of storage temperatures on sensitivity of White Rose potatoes to processing heat. Fruit Products J . 26(4), 106-108, 120-12 1. Carrier, W. H. 1911. Rational psychrometric formulae. Trans. Am. SOC.Mech. Engrs. 33, 1005-1053. Carrier, W. H. 1921. The theory of atmospheric evaporation-with special reference to compartment dryers. Ind. Eng. Chem. 13, 432-438. Chapman, F. C. 1922a. U. S. Patent No. 1,404,369. Chapman, F. C. 192213. U. S. Patent No. 1,422,416. Christie, A. W. 1926. The dehydration of prunes. Univ. Calif. Agr. Expt. Sta. Bull. No. 404, 10-19, 19-21, 21-23, 23-25. 25-26, 26-32. Revised by P. F. Xichols, Dec. 1929. Cruess, W. V. 1919. Evaporators for prune drying. Univ. Calif. Agr. Expt. Sta. Circ. No. 213. Cruess, W. V. 1935. “Commercial Fruit and Vegetable Products,” 2nd ed., pp. 462501, 502-513, 463-479, 479-490. McGraw-Hill, New York. Cruess, W. V., and Christie, A. W. 1921a. Dehydration of fruits (A Progress report). Univ. Calij. Agr. Expt. Sta. Bull. No. 330. Cruess, W. V., and Christie, A. W. 1921b. Some factors of dehydrator efficiency. Univ. Calif. Agr. Expt. Sta. Bull. No. 337. Cruess, W. V., and Mackinney, G. 1943. The dehydration of vegetables. Cniv. Calif. Agr. Expt. Sta. Bull. No. 680, 9-19, 19-22, 41-48. Ede, A. J., and Hales, H.C. 1948. The physics of drying in heated air, with particular reference to fruits and vegetables. Dept. Sci. I n d . Research Food Incest. Board (Brit.) Special Repts. No. 63. Eidt, C. C. 1935. Principles and methods involved in dehydration of apples. Can. Dept. Agr. Publ. No. 626, Tech. Bull. No. 18. Great Britain, Ministry of Foods. 1946. Vegetable dehydration. (Scientific and Technical Series.) Grosvenor, W. M. 1908. Calculations for dryer design. Trans. Am. Inst. Chem. Engrs. 1, 154-202. Guillou, R. 1942. Developments in fruit dehydrator design. Agr. Eng. 23, 313-316. Guillou, R., and Moses, B. D. 1943. Farm fruit dehydrator. University of California Agricultural Engineering and Agricultural Extension, Farm Building Plan C-214: 1-22. (Mimeo.) Hausbrand, E. 1912. “Drying by Means of Air and Steam.” Scott, Greenwood & Sons, London, Hendel, C. E., Silveira, V. G., and Harrington, W. 0. 1954. Non-enzymatic browning of white potato. Presented June 1954, Institute of Food Technologists, Los Angeles, Calif. Hougen, 0. A., McCauley, H. J., and Marshall, W. R., Jr. 1940. Limitations of diffusion equations in drying. Trans. Am. Inst. Chem. Engrs. 36, 183-209. Lazar, M. E. 1944. Deviations from adiabatieity in tunnel dehydrators. Unpublished report. U. S. Dept. Agr., Bureau Agr. and Ind. Chem., Western Regional Research Lab., Albany, Calif. Lewis, W. K. 1921. The rate of drying of solid materials. Ind. Eng. Chem. 13,427-432.

TUNNEL DEHYDRATORS FOR FRUITS AND VEGETABLES

371

McCready, D. W., and McCabe, W. L. 1933. The adiabatic air drying of hygroscopic solids. Trans. Am. Inst. Chem. Engrs. 29, 131-160. Marshall, W. R., Jr. 1942. The drying of food. Heating, Piping Air Conditioning 14, 527-531, 588-591, 671, 673, 724-728. Marshall, W. R., Jr. 1943. The drying of food. Heating, Piping Air Conditioning 16, 10-12, 567-572. Marshall, W. R., Jr., and Friedman, S. J. 1950. I n “Chemical Engineers Handbook” (Perry, ed.), 3rd ed., pp. 800-884. McGraw-Hill, New York. Masure, M. P., Bohart, G. S., Eastmond, E. J., and Boggs, M. M. 1950. Value of starch coating in the preservation of quality of dehydrated carrots. Food Technol. 14, 9G97. Newman, -4.B. 1931a. The drying of porous solids: Diffusion and surface emission equations. Trans. Am. Inst. Chem. Engrs. 27, 203-220. Newman, A. B. 1931b. The drying of porous solids: Diffusion calculations. Trans. Am. Inst. Chem. Engrs. 27, 310-333. New Zealand Department of Science and Industrial Research 1944. Vegetable dehydration. (Mimeo.) Pearson, J. W. 1923. U. S. Patent No. 1,461,224. Perry, R. L. 1944. Heat and vapor transfer in the dehydration of prunes. Trans. Am. SOC.Mech. Engrs. 66, 447-456. Perry, R. L. 1947. Designing a counterflow tray drier from laboratory data. Unpublished report, Apr. 15, 1947. Univ. Calif. Agr. Expt. Sta., Davis, Calif. Perry, It. L., Mrak, E. M., Phaff, H. J., Marsh, G. L., and Fisher, C. D. 1946. Fruit dehydration. I. Principles and equipment. Dec. 1946. Univ. Calif. Agr. Ezpt. Sta. Bull. No. 698. Prescott, S. C., and Proctor, B. E. 1937. “Food Technology,” 1st ed., pp. 488-508. McGraw-Hill, New York. Puccinelli, R. L. 1923. U. S. Patent No. 1,464,338. Raniagr, IT-. D., and Rasmussen, C. L. 1943. This is what i t costs t o dehydrate vegetables. Food Inds. 16(7), 64-71, 137-138; (8), 66-67, 118-119; (9), 75-77, 126. Rasmussen, C. L., and Shaw, W. L. 1953. Preliminary planning for vegetable dehydration. Bureau Publication AIC 356, U. S. Dept. Agr., Bureau Agr. and Ind. Chem., Western Reg. Research Lab., Albany, Calif., June 1953. Itees, C. 1922. U. S. Patent No. 1,413,135. Itidley, G. B. 1921. Tunnel dryers. Ind. Eng. Chem. 13, 453-460. Sherwood, T. K. 1929. The drying of solids, I, 11. Ind. Eng. Chem. 21,12-16,97&-980. Sherwood, T. K. 1930. The drying of solids, 111. Mechanism of the drying of pulp and paper. Ind. Eng. Chem. 22, 132-136. Sherwood, T. K. 1931. Application of theoretical diffusion equations to the drying of solids. Trans. Am. Inst. Chem. Engrs. 27, 190-202. Sherwood, T. K. 1932. The drying of solids: Application of the diffusion equations. Ind. Eng. Chem. 24, 307-310. Sherwood, T. K. 1936. The air drying of solids. Trans. Am. Inst. Chem. Engrs. 32, 15&168. Simon, M., Wagner, J. R., Silveira, V. G., and Hendel, C. E. 1953. Influence of piece size on production and quality of dehydrated Irish potatoes. Food Technol. 7, 423-428. Simon, M., Wagner, J. R., Silveira, V. G., and Hendel, C. E. 1954. Calcium chloride as a non-enzymatic browning retardent for dehydrated white potatoes. Presented June 1954, Institute of Food Technologists, Los Angeles, Calif.

372

P. W. KILPATRICK, E. LOWE, AND W. B. VAN ARSDEL

Tiemann, H. D. 1917. The theory of drying and its application t o the new humidityregulated and recirculating dry kiln. U. S. Dept. Agr. Bull. No. 609. Thorp, F. H. 1905. “Outlines of Industrial Chemistry,” 2nd ed., pp. 545-546. Macmillan, New York. United States Western Regional Research Laboratory. 1943 (Revised 1944). Information sheet on bin-type finishing driers in vegetable dehydration. U. S. Bureau of Agricultural and Industrial Chemistry, Albany, Calif. AIC-15. United States Bureau of Agricultural and Industrial Chemistry. 1944. Vegetable and Fruit Dehydration-A Manual for Plant Operators. U. S. Dept. Agr. M i x . Publ. NO. 640, 46-74, 77-88, 96-104, 114-115. United States Western Regional Research Laboratory. 1943-1947. The application of drying-rate nomographs to the estimation of tunnel dehydrator drying capacity. U. S, Bureau Agr. Ind. Chem., Albany, Calif., AIC-31. I. Riced white potatoes (Rev. June 1947). 11. Blanched sweet corn (Nov. 1943). 111. White potato strips, vertical airflow (Jan. 1944). IV. Shredded cabbage (Feb. 1944). V. Onion slices (Apr. 1944). VI. Sweetpotato strips (Sept. 1944). VII. White potato half-cubes (Mar, 1945). VIII. Carrot pieces (May 1947). Van Arsdel, W. B. 1942. Tunnel dehydrators and their use in vegetable dehydration. Food Inds. 14(10), 43-46, 106; (ll), 47-50, 103; (12), 47-50, 108-109. Van Arsdel, W. B. 1947. Approximate diffusion calculations for the falling-rate phase of drying. C h e m Eng. Progr. 43, 13-24: (also U. S.Bureau Agr. and Ind. Chem., AIC-152, issued a t Albany, Calif.). Van Arsdel, W. B. 1951a. Principles of the drying process, with special reference to vegetable dehydration. U. S. Bureau Agr. and Ind. Chem., Albany, Calif., AIC-300. Van Arsdel, W. B. 1951b. Tunnel-and-truck dehydrators, as used for dehydrating vegetables. U. S. Bureau Agr. and Ind. Chem., Albany, Calif., AIC-308. V-an Arsdel, W. B. 1953. Simultaneous heat and mass transfer in nonisothermal systems: Through-flow drying in the low-moisture range. Presented December 1953, American Institute of Chemical Engineers, St. Louis, Mo. Wiegand, E. H. 1923. Recirculation driers. Oregon Agr. Expt. Sta. Circ. No. 40. Yule, W. T. 1845. English patent, Jan. 28, 1845. Abstract in J . Franklin Znst. Ser. 111, 11, 179-180 (1846).

Author Index Numbers in italics indicate the page on which the reference is listed a t the end of the article.

A Aasted, K. C. S., 39, 43, 55 Abbott, M. E., 118, 154 Ahhott, 0. D., 294, 308 Ache, L., 260, 261, 306 Adami, J. G., 118, 127, 154 Adams, A. T., 143, 146, 155 -4damson, C. A., 120, 154 Adelson, C., 51, 55 Adrianoff, S., 16, 62 Ahmad, B., 274, 306 Alas, G., 88, lo4 Albury, M.,287, 308 Alexander, L. M., 194, 219 Alikonis, J. J., 8, 11, 44, 50, 55 Allen, A. E., 233, 253 Amako, T., 118, 155 Ambler, J. 8., 19, 55 Amerine, M. &4., 83, 84, 86, 87, 88, 89, 100,101, 103

Andersen, A. A., 180, 203, 208, 209, 213, 224, 230 Anderson, G., 265, S10 Anderson, G. W.,189, 216 Angermeier, H. F., 29, 55 Appleman, W1. D., 169, 182, 183, 202, 214, 219, 222, 223 Arena, A., 72, 74, 84, 85, 97, 98, 100, 101 Arias, C., 117, 155 Arnold, A., 276, 306 Amy, E., 272, 306 Asami, Y., 204, 214 Aschehoug, V., 189, 214 Ashikaga, C., 285, 288, 292, 296, 306 Asnis, R. E., 13, 57 i2tkin, L., 261, 265, 269, 270, 306, 310 Auerman, L. Ya., 266, 307 Aughey, E., 265, 276, 277, 286, 287, 288, 289, 290, SO6

hyres, J. C., 22, 56, 137, 138, 139, 141, 143, 144, 145, 146, 147, 148, 150, 154, 155, 156, 159, 160, 161

B Bahcock, C. J., 197, 214 Back, S., 15, 57 Bailey, A. E., 21, 32, 56 Baker, C. T., 47, 56 Baker, G. L., 26, 27, 28, 56, 59, 62, 65 Baldmin, G. L., 137, 138, 141, 158 Banfield, F. H., 123, 155 Bar, P. J., 239, 246, 253 Barackman, R. A., 271, SO7 Barker, H. A., 81, 82, 101, 103 Barker, J., 223, 232, 234, 235, 236, 23i, 243, 249, 252, 253

Barnes, B., 287, 288, 289, 290, 307, 308 Barnes, H. T., 214 Barnes, L. A., 194, 203, 216 Barreto, A., 220 Barthoiomew, J. W., 204, 214 Bartram, M. T., 194, 195, 222, 226 Bartron, L. R., 8, 15, 43, 57, 61 Bartusch, W.,18, 68 Basu, K. P., 292, 307 Bate-Smith, E. C., 122, 123, 150, 155 Bateman, E., 347, 369 Bates, C. A., 207, 227 Bates, F. L., 34, 56 Bates, P. K., 164, 205, 225 Bauer, A., 290, 310 Baughman, W. F., 46, 59 Baumann, C. A., 261, 262, 263, 268, 269, 301, 304, 309, 311 Baumgartner, J. G., 169, 229 Baur, L., 26, 61 Bayfield, E. G., 296, 307 Beadle, B. W., 259, 261, 281, 282, 307, 308

373

374

AUTHOR INDEX

Beamer, P. R., 202, 228 Beard, P. J., 164, 169, 180, 214 Beattie, H. G., 47, 58 BBchamp, A., 69, 70, 101 Bechtel, W. G., 35, 56 Beckwith, T . D., 190, 214, 226 Becquerel, P., 171, 214 Bedford, R. H., 189, 191, 214 Behrens, 0. K., 111, 158 Beisel, C. G., 181, 182, 215 Belehradgk, J., 215 Bendix, G. H., 258, 291, 292, 298, 303, 307, 310 Benjamin, D. G., 18, 58 Bennett, B. B., 286, 287, 288, 289, 295, 309 Benstead, J. G., 192, 265 Berard, H. L., 194, 195, 621 Beresford, H., 233, 234, 256 Berry, F. E., 137, 138, 141, 158 Berry, J. A., 165, 169, 174, 180, 181, 182, 183, 202, 203, 205, 209, 215,224,230 Berry, J. M., 99, 100, 101, 105 Best, C. H., 122, 155 Bethke, R. M., 117, 155 Beuk, J. F., 258, 261, 263, 282, 283, 284, 293, 299, 300, 304, 508,310 Bharihoke, G., 274, 306 Bickford, W. G., 46, 57 Bidault, C., 137, 130, 142, 155, 167 Bierotti, E., 118, 155 Bigwood, F. M., 225 Bingham, E. C., 8, 56 Binkley, W. W., 17, 65 Bioletti, F. T., 87, 88, 101 Bird, H. R., 262, 301, 311 Bird, J. C., 262, 302, 307 Birdseye, C., 190, 215 Blair, M. G., 17, 18, 65 Blamire, R. V., 116, 155 Bland, F. 0. S., 215 Bloom, 0. T., 24, 56 Blundell, C . C., 219 Boaz, T. G., 122, 156 Boddington, R. J., 48, 64 Boggs, M. M., 136, 169, 249, 250, 253, 255, 368, S71 Bohart, G. S., 368, 371 Boikan, W., 113, 160 Bolanwoski, J., 49, 56

Boni, A., 118, 156 Booth, R. G., 259,261,267,268,291,303,

sor

Born, R., 11, 61 Bortree, A. L., 116, 117, 165 Bostelman, E., 231, 232, 255 Boswell, V. R., 231, 232, 253 Botscharowa, Z. Z., 166, 228 Bowen, W. S., 234, 235, 254 Boycott, A. E., 121, 155 Boyden, E. R., 276, 277, 280, 307 Boyer, E. A., 118, 119, 127, 141, 155 Brady, 0. E., 277, 307 Brauns, D. H., 46, 59 Braverman, J. B. S., 87, 89, 103 Brayton, K. L., 170, 228 Breed, R. S., 80, 81, 89, 101, 141, 155 Brmner, S., 265, 269, 297, 298, 304, 307, SO8

Brewer, W. D., 286, 287, 288, 289, 295, 309

Briant, A. &I.,272, 273, SOT Brighton, K. W., 298, S10 Brison, F. R., 47, 56 Brodie, D. C., 100, 105 Brokaw, C. H., 183, 215 Bronson, W. F., 24, 56 Brooks, F. T., 137, 139, 142,166, 188,216 Brooks, J., 195, 196, 221 Brooks, R. F., 165, 171, 185,186,220,221 Broughton, D. C., 347, 348, 355, 369 Brown, A. G., 21, 64 Brown, A. H., 234, 254, 256, 327, 347, 348, 349, 570 Brown, E. B., 215 Brown, H. J., 178, 194, 215 Brown, L. C., 293, 299, 304, 308 Browne, A. S., 194, 203, 216 Brownless, D. S., 194, 216 Bruno, A. J., 53, 64 Brunstetter, B. C., 184, 227 Brush, M. K., 286, 287, 290, 309 Bryant, L. R., 261, 307 Buchanan, B. F., 29, 56 Buchbinder, L., 200, 216, 224 Buckby, L., 276, S10 Biichi, W., 82, 101 Bugge, A., 118, 156 Bukin, V. M., 266, 307 Bulloch, W., 119, 166

AUTHOR INDEX

Bunimovitch, M., 235, 251, 254 Burgess, N. M., 234, 255 Burke, M. V., 148, 156 Burkhart, B., 178, 195, 196, 229 Burn, C. G., 119, 156 Burr, H. K., 249, 255 Burroughs, W., 117, 155 Burrows, W., 113, 121, 156 Burton, E. F., 171, 216 Burton, G. W., 259, 511 Burton, M. O., 174, 183, 187, 216 Burton, W. G., 232, 234, 235, 236, 243, 249, 252, 25S, 254 Butler, H. G., 23, 56 Byrne, A. F., 151, 159 C

Caesar, G. V., 35, 56 Cahill, V. R., 150, 161 Caillcau, R., 86, 101, 296, 307 Caldwell, J. S., 169, 222 Callow, E. H., 122, 123, 149, 150, 156 Campbell, H., 165, 216, 368, 370 Campbell, L. E., 8, 28, 56 Campbell, R., 277, 278, 2T9, 280, 507 Campbell, W. L., 235, 237, 238, 243, 247, 251, 252, 264 Canfora, M., 119, 156 Cantor, S., 17, 6s Caray, E. M., 48, 56 Cardon, B. P., 81, 101 Carles, P., 70, 101 Carlin, A. F., 22, 56 Carrier, W.H., 347, 370 Carriere, G., 118, 156 .q Carroll, E. J., 116, 166 Carter, H. E., 86, 101 Carver, F. S., 44, 56 Case, R. A. M., 192, 225 Castan, P., 16, 62 Castell, C. H., 189, 191, 216 Castor, J. G. B., 86, 101, 102 Cathcart, W. H., 47, 68, 198, 204, 216, 218

Causey, K., 207, 216, 287, 311 Cecil, S. R., 46, 47, 56, 65 Chachin, T., 295, 506 Chang, H., 203, 217 Chapman, F. C., 315, 570

375

CharpentiB, Y., 72, 102 Charrin, 120, 156 Chavan, J., 17, 62 Cheldelin, V. H., 86, 102, 276, 277, 310 Cherenko, L. E., 42, 56 Cherry, W. B., 194, 203, 216 Chesbro, R. M., 280, 294, 310 Chevillard, L., 86, 101 Chornock, F. W., 293, 299, 304, SO8 Christensen, F. W., 276, 307 Christensen, P. B., 136, 156 Christie, A. W.,315, 317, 370 Christopher, W. X., 166, 167, 194, 202, 222 Ciccone, V. R., 49, 66 Clark, G. L., 23, 57 Clark, R. K., 278, 307 Clay, C., 15, 24, 29, 33, 34, 42, 43, 57 Clayton, W., 15, 57 Clrary, J. P., 164, 169, 180, 214 Clifcorn, L. E., 258, 286, 280, 291, 292, 296, S07, 309 Cole, M. W., 241, 242, 246, 252, 253, 2.55 Colien, F. E., 194, 216 Collins, W. R., 25, 57 Combes, F. M., 199, 207, 227 Conn, R. C., 13, 57 Connolly, F., 286, 287, 288, 289, 290, 295, 507 Conrad, R. M., 237, 249, 251, 254, 300, 510 Conradi, H., 118, 156 Cook, B. B., 280, 294, 307 Cook, L. R., 14, 44, 57 Cook, W. H., 192, 193, 226 Cooke, E. W., 233, 234, 254 Cooley, A. M., 232, 246, 252, 254 Cooper, A. R., 37, 58 Cooperman, J. M., 269, 270, 307 Copley, M. J., 28, 63 Cosens, K. W., 180, 222 Coder, H. B., 11, 57 Coulter, S . T., 4 , 57 Cover, S., 278, 283, 284, 289, 297, 507 Crook, A,, 291, 509 Cross, S. T., 15, 44, 57 Cruess, W. V., 28, 57, 81, 84, 87, 88, 101, 102, 103, 314, 315, 347, 370 Cutlar, K. L., 290, 307 Cytron, B., 28, 67

376

AUTHOR INDEX

D Dack, G. M., 203, 216, 221 Dalby, G., 265, 309 Dal Cin, G., 100, 102 Dale, C. N., 203, 226 Danehy, J. P., 30, 67 Dangler, G., 200, 216 Daniell, E. P., 265, 276, 277, 286, 287, 288, 289, 290, 306 Dann, W. J., 86, 102 Darfler, J., 217 Davis, J. G., 170, 216 Davis, M. A., 170, 620 Dawson, E. R., 265, 266, 268, 269, 307 de Alvarez Herrero, H. G., 260, 307 Dean, G. R., 17, 63 Dearden, D. V., 261, 293, 307 Deatherage, F. E., 113, 120, 168, 161 de Bobadilla, G . F., 84, 101 Del Guidice, V. J., 136, 137, 160 de Maya, C. B., 48, 59 Denning, H., 290, 310 Dervichian, D. G., 216 Desoubry, M. G., 119, 156 Deuel, H., 82, 101 Deuel, H. J., Jr., 289, 30.9 Devik, O., 168, 216 Dickman, S.R., 111, 156 Diehl, H. C., 180, 187, 215, 217 Dilsaver, E. M., 278, 283, 284, 297, SO7 Dimler, R. J., 18, 57 Dobie, J. B., 136, 160 Dobson, J. G., 223 Dolman, C. E., 203, 217 Dorsey, N. E., 171, 217 Douglas, H. C., 75, 81, 83, 84, 85, 86, 89, 96,109, 10Y Dove, W. F., 200, 222 Down, D. E., 264, 307 Drake, T. G . H., 291, 309 Drummond, J. C., 259, 308 Du Bois, C. W., 181, 217 Dubourg, E., 71, 102 Duclaux, E., 70, 74, 81, 102 Dufrenoy, J., 100, 105 Duggeli, M., 116, 156 Dunlap, W. A., 197, 214 Dunlop, S. G., 265, 269, 298, 304, SO7 Dunn, K. M., 116, 117, 165

Dunn, M. S., 86, 102 Dutcher, R. A., 286, 287, 289, 290, 296, 297, 308 Dutta, N. K., 262, 302, 307 Dyer, F. E., 190, 217 Dyer, W. J., 190, 217 Dykstra, K. G., 21 7 I

E Eastmond, E. J., 368, 371 Isaston, N. R., 8, 15, 43, 67, 62 Eckstein, G. R., 40, 67 Eddy, C. R., 27, 63 Ede, A. J., 347, 370 Edmondson, R. B., 202, 217 Edwards, A. L., 227 Edwards, C . S., 254, ,954 Edwards, P. R., 194, 203, ,916 Eidt, C. C., 315, 354, 370 Eippcr, W. R., 14, 15, 44, 57 Eisele, C. W., 151, 159 Eklund, A. B., 269, 307 Ellenberger, H. E., 296, 308 Elrod, R. P., 184, 217 Elsden, S. R., 116, 156 Elvehjem, C. A., 269, 270, 275, 276, 277, 279, 281, 292, 293, 306, 307, 309, 310,312

Ely, R. E., 116, 117, 155 Emery, A. J., 165, 171, 183, 185, 186, 220 Emmons, C. W., 139, 160 Empey, W. A., 114, 115, 117, 118, 128, 129, 130, 132, 134, 135, 136, 137, 138, 139, 140, 144, 150, 154, 156, 257 Enas, J. D., 48, 63 Engel, C., 293, 310 Englis, D. T., 17, 67 Ensminger, M. E., 136, 159, 160 Eriksen, S.E., 276, 277, 280, 307 Erlemann, G. A., 17, 63 Erskine, H. L., Jr., 237, $54 Erwin, R. F., 115, 160 Escudero, A., 260, SO7 Eskew, R. K., 48, 61 Esselen, W. B., 194, 200, 207, 210, 217, 286, 288, 308 Evans, E. V., 261, SO7 Evans, J. B., 98, 102

37’7

AUTHOR I N D E X

Evers, C. F., 187, 198, 199, 211, 217, 221, 228

Ewell, A. W., 136, 167

F Fabian, F. W., 197, 217 Fabian, J. R., 217 Fahs, F. J., 8, 11, 12, 13, 29, 32, 36, 48, 58, 60, 61 Faitelowitz, A., 235, 251, 254 Fardig, 0. B., 286, 287, 289, 290, 296, 308

Farrell, K. T., 44, 56 Farrer, K. T. H., 258, 259, 260, 261, 262, 264, 266, 267, 268, 269, 270, 272, 282, 292, 294, 297, 298, 299, 300, 301, 304, 305, 306, 307, 308 Faulenborg, G., 104 Faville, L. W., 183, 217, 220 Favor, H. H., 35,57 Feaster, J. F., 259, 277, 278, 281, 282, 283, 285, 286, 287, 288, 289, 291, 292, 296, 298, 304, 307, 308, SO9 Feduchy Mariiio, E., 88, lo4 Feller, B. A., 262, 302, 309 Fellers, C. R., 190, 206, 210, 217, 280, 286, 288, 308, 310 Felton, E. A., 98, 102 Fenton, F., 199, 207, 216, 217, 221, 285, 286, 287, 288, 289, 290, SOT, 308,311 Fevold, H. L., 300, 309 Ficker, &I.,119, 120, 157 Fieger, E. A., 190, 217, 268, 271, 311 Field, H., Jr., 261, 263, 310 Fineke, H., 39, 57 Finn, D. B., 217 Fisher, C. D., 316, 345, 347, 350, 351, 352, 359, 365, 371 Fisher, G., 34, 6s Fison, F. J., 246, 254 Fitz, A., 70, 102 Fitxgerald, G. A., 183, 190, 200, 201, 205, 210, 217, 226, 286, 288, 308 Flavier, H., 86, 102 Fleming, A., 111, 157 Fletcher, D. A., 197, 217 Florey, H., 111, 157 Folinazzo, J. F., 182, 223, 224 Ford, W. W., 118, 167

Fore, S. P., 46, 67 Foreman, E. M., 31, 62 Fornachon, J. C. M., 75, 81, 83, 85, 86, 87, 88, 89, 96, 102, 107 Forsyth, W. G . C., 39, 57 Foster, E. M., 137, 138, 141, 158 Fournier, S. A., 293, 299, 304, 308 Fraenkel-Conrat, H., 31, 61, 111,157 Frazier, W. C., 112, 130, 160 Freed, M., 297, 308 Freeman, A. F., 46, 67, 294, 311 French, D., 34, 66 French, R. B., 294, 308 Freundlich, L., 41, 57 Frey, C. N., 261, 265, 269, 270, 306, 310 Frian, C., 28, 67 Fried, J. F., 295, 310 Friedman, L., 24, 25, 58 Friedman, S. J., 347, 371 Fritsch, E., 274, 308 Frobisher, M., 172, 217 Frost, D. V., 259, 261, 262, SO9 Fryer, H. C., 254 Fuller, A. D., 34, 58 Fuller, J. E., 174, 179, 183, 185, 187, 226 Fulmer, W. C., 267, 268, 310 Fulton, C. O., 218 Funk, E. M., 218

G Gaebelein, U., 170, 218 Gall, L. S., 117, 118, 157 Galloway, L. S., 279, 311 Gane, R., 234,235,236,243,249,252,253 Gano, O., 233, 234, 254 Garnatz, C. G., 194, 225 Garnatz, G. F., 194, 224 Gautier, A., 70, 102 Gayon, U., 71, 102 Geddes, W. F., 4, 57 Geer, L. P., 188, 202, 205, 218, 225 Gehenio, P. M., 165, 168, 170, 171, 222 Geiger, C., 86, 108 Gklis, A., 16, 58 Genevois, L., 86, 102 Gerlaugh, P., 117, 155 Gibbons, N. E., 125, 149, 157, 178, 194, 215, 218 Gilb, H. W., 204, 218

378

AUTHOR INDEX

Gilbert, P. E., 181, 218 Giltner, L. T., 202, 217 Glage, F., 140, 157 Glasstone, S., 305, SO8 Gleim, E. G., 286, 287, 290, 308 GlQnard, A., 69, 102 Goddard, V. R., 269, 307 Godkin, W. J., 47, 68, 198, 218 Goldberg, H. S., 150, 161 Goldberg, L., 265, 266, 269, 308 Goldsworthy, N. E., 111, 157 Gomez, L., 278, 279, 310 Gooding, C. M., 46, 6d Goodwin, M. W., 27, 28, 56 Goodyear, J. M., 260, 261, 310 Gordon, G. A., 160 Goresline, H. E., 169, 174, 177, 194, 207, 218, 22s Gorfinkle, W. I., 24, 58 Gorman, J. M., 179, 221, 222 Gorovitz-Wlassova, I,. RI., 137, 138, 157 Gortner, W. A., 136, 161, 263, 300, 310 Gotlib, M. A., 165, 166, 167, 169, 218 Gottfried, J. B., 18, 68 Graham, A. S., 43, 61 Gralen, N., 27, 64 Grant, M. P., 117, 157 Gray, P. H. H., 175, 218 Greco, P. A., 194, 229 Green, M., 191, 218 Greenbank, G. R., 32, 59 Greene, J. W., 237, 249, 251, 254 Greenlie, D. G., 178, 225 Greenwood, D. A., 259, 261, 281, 282. 283, 298, 304, 307, 308

Guhtrie, J. D., 46, 58, 63, 64 Guillemet, R., 268, SIO Guillot, A., 268, 310 Guillou, R., 325, 327, 350, 338, 370 Gunderson, M. F., 168, 192, 193, 194, 199, 203, 218, 219, 226

Gunderson, S . D., 218 Gunsalus, I. C . , 81, 102, 103 Gutierrez, J., 116, 117, 157 Guymon, G. F., 86, 102 Gyorgy, P., 11,58, 268, 308 H

Haas, W., 118, 157 Habcrmann, R. T., 203, 226 Hahn, S. S., 169, 183, 202, 214, 219 Haines, R. B., 119, 124, 125, 128, 132, 133, 134, 136, 137, 138, 140, 141, 142, 143, 144, 145, 154, 157, 158, 165, 166, 168, 170, 194, 219 Hajna, A. A., 174, 219 Hale, W. S., 194, 219 Hales, K. C., 347, 370 Hall, H. H., 29, 48, 58, 60, 182, 228 Hall, I. C., 203, 219 Hall, J. E., 219 Hall, L. A., 46, 58 Hall, R. C., 237, 251, 254 Halliday, E., 296, 297, 310 Halliday, E. G., 276, 278, 2i0, 286, 287, 290, 308, 509, 310, 311 Halvorson, H. O., 169, 197, 219, 228 Hamer, W. J., 12, 34, 35, 58 Hamilton, D. M., 82, 103 Hampil, B., 170, 219 Hanahan, D. J., 17, 57

Greer, E. N., 236, 255 Grettie, D. P., 25, 58 Griffin, A. M., 218 Griffith, A. S., 119, 157 Griffiths, E., 189, 218 Griffiths, F. P., 189, 218 Grinberg, L. D., 137, 138, 157 Griswold, R. M., 276, 278, 279, 308, 311 Grosvenor, W. M., 347, 370 Grover, D. W., 11, 58 Gruber, T., 118, 157 Grunt, O., 118, 157 Guerrant, N. B., 286, 287, 289, 290, 2%.

Hankins, 0. G., 219 Hankinson, C. L., 24, 62 Hannan, R. S., 31, 60 Hanning, F., 269, 272, 306, SO8 Hansford, B. A., 137, 139, 142, 156 Hnnson, H. L., 250, 253 Hanus, E. J., 262, 302, SO9 Hargiss, C. O., 229 Harlin, H., 181, 219 Harp, C. H., 200, 222 Harrel, C. G., 261, 265, 269, 270, 292,

297, 298, 308 Guha, B. C., 259, 308

Harrington, W. O., 232, 240, 241, 242, 246, 252, 253, 255, 347, 362, S70

308, 309

AUTHOR INDEX

Harris, B. R., 40, 58 Harris, K. W., 199, 221, 290, SO7 Harshaw, H. M., 194, 219 Hart, C. V., 47, 58 Hartmann, I., 37, 58 Hartsell, S. E., 111, 158, 171, 174, 194, 201, 203, 205, 219 Hartzler, E., 277, 294, SO8 Harvey, W. R., 279, 295, SO9 Hassid, W.Z., 82, 103 Hausbrand, E., 347, S70 Hauser, G., 118, 158 Havighorst, C. R., 219 Hawke, J. C.,222 Hawkes, L., 171, 219 Haynes, R. D., 181, 219 Hays, G. L., 182, 220 Hays, R. M., 278, 283, 297, SO7 Hazel, F., 23, 62 Heaton, E. K., 11, 58 Heberlein, D. G., 258, 286, 289, 291, 292, 296, 303, 507, SO9 Hedrick, L. R., 195, 224 Hedrick, RI. T.,278, SO9 Hegarty, C . P., 220 Hehrc, I?. J., 82, 103 Heiduschka, A., 47, 58 Heimerdinger, H. M., 234, 254 Heinmets, F., 169, 220 Heisler, E. G., 238, 254 Hciss, R., 18, 58 Heitz, T. W., 193, 220 IImdel, C . E., 249, 255, 347, 362, 368, 370, 371 Henderson, L. M., 276, 277, SO9 Hendlin, D., 86, 103 Hendrickson, R. L., 188, 189, 220 Henn, hf. J., 192, 193, 194, 219 Henneberg, W., 72, 98, 103 Henry, K. M.,31, 58, 261, 293, SO7 Henry, O., 46, 62 Herhst, E. J., 276, 309 IIess, E., 168, 190, 220 Hess, W '. R., 114, 119, 120, 125, 126, 127, 128, 138, 141, 158, 295, 310 Heupke, W.,264, SO9 Hibbs, J. W.,117, 160 Hicks, W. B.,207, $24 Highlands, M. E., 164, 205, 225 Higley, H. A., 220

3'79

Hilditch, T.P., 39, 59 Kill, E. G.,31, 62, 183, 217, 220 Killiard, C. M., 170, 620 Hillig, F., 195, 626 Hills, C . H., 28, 59,65 Hilton, H. W., 17, 65 Hiltz, M. C.,277, 278, 280, 286, 287, 288, 289, 290, SO7 Hiner, R. L., 219 Hinman, W.F., 278, 279, 286, 287, 290, 296, 297, 509, 310,311 Hirst, E. L., 27, 59 Hitchcock, M. IT. S., 116, 156 Hitchens, A. I?., 80, 81, 89, 101, 141, 155 Hoben, H. H., 44, 59 Hochstrasser, It., 82,99, 103 Hoet, J. P., 122, 155 Hoffpauir, C. L., 46, 58 Hoffstadt, R. E., 154, 158 Hoflund, S., 116, 158 Hofmann, C., 265, 309 Hohf, J. P., 347, 369 Hohl, L. A., 88, 103 Hollander, D. N., 169, 171, 220 Hollenbeck, C . M., 262, 263, 269, 301, SO9 Holley, K. T., 45, 46, 62 Hollis, F., Jr., 234, 255 Holm, G. E., 32, 59 Holmes, A. D., 293, SO9 Holmes, D., 191, 220 Holmes, N. E., 189, 218 Holtman, D. F., 194, 195, 220 Honstead, W. H., 237, 249, 251, 254 Hood, M. P., 47, 60 Hooker, D. R., 121, 159 Hopper, T.H., 46, 64,275, SO7 Horn, A., 118, 158 Horwarth, W. J., 141, 158 Horwood, M. P., 133, 158 Hougen, 0. A,, 347, 370 Houston, J., 261, 293, 307,SO9 Hove, E. L., 261, 269, 270, 292, SO9 Hoyer, D. P., 85, 92, 93, 103 Hucker, G. J., 99, 103,165, 168, 171, lii, 183, 184, 185, 186, 220, 221, 228 Hiilphers, G.,119, 158 Huffman, C. F., 116, 117, 155 Huhtanen, L. C., 117, 118, 157 Humphrey, H. J., 173, 186, 221 Hungate, R. E., 116, 156

380

SUTHOR INDEX

Hunner, M., 280, 294, 310 Hunter, A. C., 221 Hunter, A. S.,238, 254 Hussemann, D. L., 199, 200, 210, 221 Hutchings, B. L., 86, 103, 180, 198, 199, s21, ss4 Hutchins, M. R., 272, 273, SO7 Hutterman, W., 118, 158 1

Ice, R. M., 199, 221 Iljin, W. S., 221 Inagaki, C., 261, SO9 Ingalls, R. L., 286, 287, 288, 289, 295, 309 Ingram, M., 123, 140, 149, 168, 168, 170, 192, 195, 196, 221, 225 Ireland, R., 180, 221 Irish, J. H., 180, 221 Irvine, 0. R., 261, SO7 Ives, M., 259, 281, 285, 286, 287, 288, 289, 291, 292, SO8

J Jackson, J. M., 277, 278, 281, 282, 283, 298, 304, 308, SO9 Jackson, R. F., 17, 61 Jackson, S. H., 291, SO8 Jacobs, M. B., 45, 48, 59 Jacobson, M., 37, 58 Jakovliv, G., 181, 621 James, L. IT., 169, 193, 194, 202, 210, 216, 221, 226, 227 Jamieson, G. S., 46, 59 Jang, R., 28, 57 Jans, L. M., 276, 278, 279, SO8 Jansen, B. C. P., 259, 309 Jansen, E. F., 26, 59 Jenkins, M. K., 207, 224 Jenness, R., 4, 67 Jensen, H. R., 37, 43, 59 Jensen, L. B., 114, 119, 120, 121, 125, 126, 127, 128, 131, 135, 137, 138, 141, 142, 168, 221 Jepsen, A., 221 Johns, C. K., 194, 195, 196, 197, 217, 221 Johnson, A. H., 180, 182, 226 Johnson, E., 276, 277, 278, 279, 280, 287, 288, 289, 290, 509

Johnson, R. I., 15, 57 Johnson, V. T., 221 Johnston, C. H., 289, SO9 Johnston, N. F., 35, 57 Jones, A. H., 179, 180, 184, 185, 221, 222 Jones, C. P., 293, SO9 Jones, C. R., 236, 265 Jones, E., 285, 309 Jones, G. I., 300, 309 Jones, J. B., 290, SO7 Jones, J. K. N., 27,69 Jordan, E. O., 203, 221 Jordan, S., 2, 10, 40, 69 Joslyn, M. A., 48, 59, 83, 84, 86, 87, 88: 89, 100, 101, 103, 108, 180, 221 Joszt, A., 17, 59

K Kahlenberg, 0. J., 179, 221, 222 Kalen, J., 277, 294, 295, 311 Kalyanasundaram, A., 20,59 Kamio, H., 301, 311 Kampf, A., 40, 59 Kandutsch, A. A., 261, 263, 268, 269, 301, 304, SO9 Kani, T., 273, 309 Kaplan, M. T., 183, 222 Kass, J. P., 30, 59 Katz, I., 262, 302, 311 Katz, M., 262, 302, 311 Kaufman, C. W., 48, 59, 234, 255 Kaufman, F. L., 261, 263, 300, 304, 310 Kawasaki, C., 273, SO9 Kawaziri, S., 261, 311 Kayser, E., 82, 103 Kayukova, N. I., 203, 222 Keith, S. C., Jr., 169, 170, 222 Kelly, D. J., 43, 67 Kempf, N. W., 8, 41, 43, 44, 59 Kcnnedy, B. M., 274, 309 Kern, R. A., 111, 158 Kern, S. F., 111, 158 Kerr, D. E., 203, 217 Kertesz, Z., 26, 27, 29, 59 Keunemann, R. W., 185, 228 Kcw, T. J., 181, 217 Kidd, M. N., 137, 139, 142, 156, 215 Kidder, L. E., 280, 294, 310 Kielhofer, E., 100, 10s Kiessig, A., 118, 156

381

AUTHOK. INDEX

Kihlberg, B., 302, S10 Kik, M. C., 274, 296, SO9 Kilpatrick, P. W., 234, 254, 327, 347, 348, 368, S70 King, J., 232, 255 King, T. E., 86, 102 Kingkade, R l . J., 111, 158 Kirchner, J. G., 48, 60 Kirkpatrick, W. F., 98, 99, 106 Kirsch, R. H., 137, 138, 141, 158 Kisrr, J. S., 190, 222 Kittelman, J., 261, 30.9 Kjaer, -k., lo4 Klein, E., 137, 158 Klemm, K., 24, 25, 58 Kline, R. W., 22, 60, 64 Klose, A. A., 300, SOB Klosterman, A . M., 272, 2T3, SO7 Knapp, A. W., 38, 39, 60 Knight, B. C. J. G., 86, 103 Knight, S. G., 112, 160 Knott, E. lI.,299, 300, SO.Y, 310 Koch, A , , 71, 103 Koch, J., 43, 60 Kocli, It., 70, 103 Koelensniid. W.A. A. B., 184, 186, 622 Kohman, E. F., 292, 309 Kon, S.I<., 31, 58, 261, 293, 307, 30.9 Koorenian, J. A., 33, 60 Koshimizu, M., SO6 Kramer, J.; , i 2 , 103 Kraybill, H. 12., 259, 261, 281, 282, 283, 298,304, 307, 308 Kremer, T. A , 203, 222 Kringstad, H., 287, 288, SOY Krumperman, P. H., 100, 103 Kunkle, L. E., 120, 150, 161 Kuntzr, \I7.. 116, 16.9 Kunz, It., 71, 103 Kutseva, L. S., 266, 307 Kuzmeski, A. W., 293, SOY Kyes, P., 170, 222 Lafoureade, S.,100, 105, 106 Lally, M. M., 179, ,921, 222 Lampitt, L. H., 28, 48 Landerkin, G. B., 222 Lane, R. L., 276, 277, 278, 279, 280, 287, 288, 289, 290, 309

I,angwill, K. E., 10 Larkin, E. P., 174, 179, 183, 185, 187,222 Latzke, E., 275, 307 Laurila, U. R., 266, 510 Lawrence, D., 28, 57 Lazar, M. E., 234, 256, 349, S70 Lea, C. H., 31, 39, 58, 60, 142, 144, 158, 222 Leach, J. M., 49, 60 Lee, F. A., 168, 228 Lehrer, W. P., 279, 295, 309 Lenhart, J., 180, 222 Leong, P. C., 274, SO9 Lepevetsky, B. C., 113, 120, 158 Lepper, H. A,, 194, 195, 222, 226 Levine, A. S., 194, 200, 207, 210, 217 Levine, M., 226 Leviton, A., 24, 60 Lcvitt, J., 222 Lewis, V. M., 31, 60 Leivis, W. K., 337, 370 Lewis, W. L., 18, 65 Liebig, A. W., 36, 60 Light, J. H., 14, 44, 51 Lincoln, H., 261, 269, 270, 292, 30.9 Lindner, P., 98, 103 Lindquist, H. G., 293, 309 Link, K. P., 26, 61 Lissauer, M., 118, 158 Litsky, W., 174, 179, 183, 185, 187, 266 Lobry de Bruyn, C. 8., 17, 60 Lochhead, A. G., 179, 180, 185, 221, 262 Logan, P. P., 200, 222 Lohnis, F., 116, 159 Longrke, K., 199, 221 Lord, D. D., 12, 35, 80 Lotzkar, H., 27, 62 Loughlin, V., 200, 216 Luckey, T. D., 115, 160 Luccke, R. W., 290, 310 Luthi, H., 82, 85, 86, 99, 10S, 104 Lunde, G., 287, 288, 309 Lutikova, P. O., 230 Lutz, J. M., 169, 222 Luyet, B. J., 165, 168, 170, 171, 222 Lyman, J. F., 267, S l l

M XcBain, J. W., 21, 64 McBryde, C. N., 127, 137, 141, 159

382

AUTHOR INDEX

McCabe, W. L., 347, 371 McCartney, J. R., 287, 308 McCauley, H. J., 347, 370 McCleskey, C. S., 166, 167, 191, 194, 202, 220, 222 McCready, D. W., 317, 371 RlcCullough, N. B., 151, 15.9 McDermott, L. A., 216 MacDonnell, L. R., 26, 59 RIcDowell, R. H., 28, 60 Alacek, T. J., 262, 302, 309 McFarlane, V. H., 165, 169, 180, 101, 207, 223, 230 &Gee, E. F., 44, 60 McGlamery, J. B., 47, 60 McGregor, M. A., 136, 160 Machilda, S., 118, 155 hlchtire, F. C., 259, 261, 262, SOY McIntire, J. M., 276, 277, 279, 309, 311 Mack, C. H., 46, 57 Mackinney, G., 347, 370 Mackintosh, D. L., 294, 3lf hlcIntosh, J. A., 136, 159, 160, 285, 309 McLaren, B. A., 278, SO7 Maclay, W. D., 27, 28, 62 McLean, R. A,, 143, 148, 161 MacMichael, R. F., 8, 40, 60 McNally, E. H., 227 Madsen, J., 123, 149, 159 Magoon, C. A , , 180, 181, 615, 623 Maillard, L. C., 30, 60 Malakar, M. C., 292, 307 hkllmann, W. L., 114, 128, 137, 138, 113, 159, 223

Malone, V., 291, SO9 Manceau, E., 82, 103 Mantell, C. L., 29, 60 Pvlapplebeck, E. G., 191, 216 hlarburger, G. C., 237, 219, 251, 254 hlarcilla Arrazola, J., 88, 104 Marfort, A., 19, 63 Markovich, V. E., 42, 56 Marks, A. L., 280, 309 Marks, H. P., 122, 155 Marsh, B. B., 191, 226 Marsh, G. L., 316, 345, 347, 350, 351, 352, 359, 365, 371 Marshall, R. A., 116, 156 Marshall, R. E., $63 Marshall, W. R., Jr., 347, 370, S71

Marsteller, R. L., 183, 206, 223 Martin, G. W., 265, 266, 268, 269, 307 Martin, L., 82, 104 hlartin, L. F., 8, 11, 12, 13, 29, 32, 36, 48, 60, 61 Martin, R. B., 223 Martin, R. D., 211, 626 Martinez, J., 182, 223 Marui, T., 260, 311 Marxer, 143, 154, 169 Massee, G., 137, 139, 142, 159 Masson, M., 117, 159 hhsure, M. P., 368, 371 Mathews, J. A., 17, 61 Mattick, A. T. R., 170, 197, 223 Maveety, D. J., 271, 311 Mayberry, 31.G., 13, 15, 44, 61 Mayer, H. D., 82, 104 Mayfield, H. L., 278, SO9 Meara, M. L., 39, 40, 61 Meckel, R. B., 264, 265, 307, 310 Medoff, S., 289, 309 Meeker, E. W.,18, 19, 61 Mees, R. H., 98, 99, 104 Mehra, S. L., 274, 306 Mehta, C. J., 262, 302, 307 Melnick, D., 261, 263, 287, 289, 310 Menkin, V., 113, 159 Merrill, A. L., 86, 108 Messenheimer, A. E., 237, 249, 251, 254

Messner, A., 118, 159 Meyer, B. H., 279, 310 Meyers, E. W,, 43, 61 Mickelson, O., 275, 310 Mickevicz, M. J., 8, 43, 61 Mickley, H. S., 347, 348, 355, 369 Milares, R., 280, 310 Miller, C., 196, 223 Miller, C. D., 290, 310 Miller, E. V., 183, 206, 225 Miller, G., 28, 57 Miller, H. G., 48, 62 Miller, J. I., 136, 161 Miller, W. A., 188, 189, 220 Milleville, H. P., 48, 61 Minch, V. A., 133, 168 Mitchell, D. G., 41, 61 Moller, E. F., 86, lo4 Moslinger, R., 71, lo4

383

AUTHOR INDEX

Mohammed, A., 31, 61 Molinski, S., 17, 59 Monaghan-Watts, B., 22, 61 Money, R. W., 11, 28, 60, 61 Monroe, K. H., 298, 310 Monvoisin, M., 137, 139, 159 Moon, H. H., 169, 222 Moon, L. I?., 48, 62 Moore, P. R., 279, 295, SO9 Moore, R. L., 194, 218 Moran, T., 135, 137, 139, 141, 142, 143, 144, 159, 164, 167, 223 Morell, S., 26, 61 Morgan, A. F., 86, 105, 280, 294, 296, 307, 310 Morris, N. J., 46, 57 Morris, T. N., 223, 232, 233, 234, 239,255 Morrison, P. G., 258, 262, 301, 305, 308 Morrison, R. A., 121, 159 Morse, J. F., 15, 67 Moses, B. D., 325, 327, 370 Moschette, D. S., 296, 297, 310 Mosiman, H., 24, 61 Illoskowa, G. L., 228 hfott, L. O., 203, 226 Mottern, H. H., 28, 65 Moyer, J. C., 287, 289, 290, 300, 308, 311 Mrak, E. M., 88, 106, 316, 345, 347, 350, 351, 352, 359, 365, 371 hliiller, E. F., 113, 160 Miiller-Thurgau, H., 71, 72, 73, 75, 80, 81, 85, 97, 98, 100, 104 Mullins, W. R., 240, 241, 242, 246, 255 Mundt, J. O., 181, 219, 223 Murdock, D. I., 182, 263, 224 Murphy, H. W., 260, 261, 310 Murphy, L. W., 124, 159 Murray, E. G. D., 80, 81, 89, 101, 141, 165

Murray, H. C., 288, 310 Murray, M., 271, 311 Murray, M. D., 46, 63, 294, 311 Murray, W. T., 188, 218 Musher, S., 13, 14, 61, 62 Myers, P. B., 28, 62 M y r b k k , K., 263,281,283,292, 302,310

N Nagy, J., 37, 58 Narayanan, K. G. A,, 262, 302, SO?’

Xaylor, H. B., 194, 296 Neal, R. H., 46, 62 Xeel, G. H., 241, 242, 246, 252, 253, 255 Neisser, M., 118, 159 Nell, E. E., 169, 171, 220 Nellis, L. F., 168, 228 Xelson, C. E., 24, 62 R’eville, H. A., 8, 15, 62 Newman, A. B., 347, 371 Newton, E. B., 143, 154, 159, 161, 189, 229

Xcholls, A. G., 118, 169 Xicholson, F. J., 118, 154 Nickerson, J. T. R., 173, 174, 176, 181, 183, 200, 224, 225 Xicklhs, J., 69, 70, 104 Niehaus, C. J. G., 87, lo4 Nielsen, F. A., 194, 224 Nilsen, H. W., 280, SO9 Niven, C. F., Jr., 82, 98, 102, 104 Nixon, E. L., 234, 255 Soble, I., 278, 279, 310 Nocard, M., 119, 159 Nollner, C., 70, 104 Nolte, A. J., 183, 224 Norris, C., 118, 169 S o r t h , W.R., 175, 224 Novick, J., 224 Xuttall, G. H. F., 115, 159 Sutting, G. C., 28, 63 Sutting, M. D., 240, 250, 252, 265, 2S6 Nymon, M. C., 263, 300, 310

0 Oberg, E. B., 24, 62 Obermeyer, €I. G., 262, 263, 267, 268, 269, 301, 309, 310 Obold, W. L., 180, 224 O’Connor, R. T., 46, 63, 294, 311 O’Donnell, W. W.,296, SO7 Ogilvy, W. S., 137, 141, 143, 144, 154, 155, 159, 160

Oglesby, E. W., 202, 228 O’Hara, M. B., 286, 289, 298, 308 Olcott, H. S., 31, 61 Oliver, W. F., 171, 216 Olsen, A., 287, 288, SO9 Olsen, A. L., 300. 310 Olsen; E., 75, 81; 82, 85, 87, 104

384

AUTHOR INDEX

Olson, B. E., 237, 246, 254, 256 Olson, R. L., 232, 240, 241, 242, 246, 250, 252, 253, 255, 256 Oneto, J., 100, 105 Opite, E., 118, 160 Orla-Jensen, A. D., 86, 92, 104, 105 Orla-Jensen, S., 104 Oser, B. I,., 287, 289, S10 Oser, M., 287, 289, 310 Osler, A. G., 216, 224 Osterud, C. M.,168, 1i1, 269 Osterwalder, A, 72, 7 3 , 75, 80, 81, 85, 95, 98, 100, 104, 105 Otte, N. C., 86, 105 Owen, J. T., 46, 62 Owen, R. F., 224 Omens, H. S., 27, 28, 62

P Pace, J. K., 273, 274, 310 Painc, H. S., 2, 62 Palmer, L. S., 30, 59 Panassenko, W.T., 224 Pappenheimer, bl.,118, 159 Parish, E. C., 183, 217 Park, S. E., 180, 194, 202, 205, 268 Parker, J. J., 204, 216 Parkes, A. S., 169, 226 Partington, IT., 302, 310 Pashovkin, V. F., 266, 307 Pasteur, L., 68, 69, 81, 105 Patrick, B., 128, 161 Patrick, R., 183, 206, 224 Patton, A. R., 31, 62 Paul, P., 286, 287, 288, 289, 295, SO9 Pavcek, P. L., 281, 292, 307 Pavelchek, E., 151, 159 Pavlenko, 0. N., 46, 62 Payen, A,, 46, 62 Pearce, J. A., 196, 224, 296, 310 Pearce, W.E., 296, 308 Pearson, J. W., 315, 371 Pearson, P. B., 278, 290, SOT, 310 Pearson, R. M., 117, 160 Pederson, C. S., 92, 96, 98, 99, 103, 105, 184, 185, 190, 221, 224 Peightal, D. E., 232, 246, 252, 254 Pelshenke, P., 266, 296, 310 Penn, F. H., 44, 62

Pennington, M. E., 193, 194, 207, 224 Penniston, V. A,, 195, 224 Pentzer, D. J., 48, 62 Percival, E. G. B., 26, 34, 62 Perlman, L., 86, 1.96 Perquin, L. H. C., 82, 105 Perri, J. M., Jr., 23, 62 Perroteau, A., 293, 310 Perry, C. A., 174, 219 Perry, H., 203, 209, 224 Perry, R.L., 316, 345, 347, 348, 349, 350, 351, 352, 358, 359, 365, 371 Personius, C. J., 265, 266, 311 Peters, F. N., 13, 62 Petersen, W.F., 113, 160 Peterson, C. F., 294, 311 Peterson, E., 203, 219 Peterson, M. S., 86, 195 Peterson, W.H., 86, 103, 105, 107 Peterson, W. J., 277, 507 Petit, L., 268, 269, 310 Peynaud, E., 72,74,86, 100,102,106,106 Phaff, H. J., 316, 345, 347, 350, 351, 352, 359, 365, 371 Phillips, A. W.,199, 200, 204, 207, 224, 665

Phillips, A. W.,Jr., 694 Phillipson, A. T., 116, 156 Pioci, G., 100, 107 Piohard, M., 43, 62 Pickering, 1 ' . L., 100, 106 Pickett, T. A., 45, 46, 62, 293, 310 Piotet, A., 16, 17, 19, 62, 6.5' Fiepoli, C. R., 172, 230 Picrce, M. E., 184, 221 Pigman, W. W.,30, 57 Pilcher, R. W.,257, 278, 309 Pinto, A. F., 48, 65 Pitman, G., 47, 63' Pitz, C. W.,136, 1/31 Plaggc, H. H., 224 Plummer, J., 286, 287, 288, 289, 295, 309 Polgc, C., 169, 225, 227 Pons, W. S., Jr., 46, 63 Porcher, M. C., 119, 166 Porchet, V.,85, 195 Portes, M., 70, 105 Potter, A. L., Jr., 242, 255 Potter, T. S., 170, 222 Poiilsen, S. D., 11, 29, 65

AUTHOR INDEX

Pounden, W.D., 117, 160 Pratt, R., 100, 105 Prescott, S. C., 164, 202, 205, 225, 314, 571

Price-Jones, C., 121, 165 Proctor, B. E., 178, 181, 190, 199, 200, 204, 207, 224, 225, 235, 237, 238, 243, 246, 247, 251, 252, 254, 255, 314, 371 Proctor, C. M., 111, 166 Proiity, C. C., 136, 159, 160 Prudden, T. E., 170, 225 Ptak, L. R., 258, 289, 291, 292, 303, 307, 309 Puccinclli, R. L., 315, 371 Pucherna, J., 19, 63 Pulkki, L. H., 266, 310 Puutula, I<., 266, 510

Q Quinn, H., 194, 22.6

R Radabaugh, J. II., ,325 Rahn, O., 225 Rainey, IT-.L., 265, 308 Ramage, IT-. D., 357, 371 Ramon, G., 203, 225 Ramon, P., 203, 225 Ramstad, P. E., 208, 226 Randall, R., 197, $14 Randoin, L., 293, 310 Rao, D L. N., 20, 69 Itapaport, S., 289, 309 Rasmussen, C . L., 316, 357, S l l Itaveiiburg, R. F., 193, bd6 Itcay, G. h.,167, 265 Rcdfield, H. W., 194, 225 Reedman, I!:. J., 276, 510 Rees, C., 315, 371 Reeves, W. A,, 46, 58 Reid, &I., 196, 218, 224 Reilly, J., 16, 63 Rcith, h. F., 118, 119, 160 Remmers, B., 233, 255 Rcndle, T., 232, 233, 237, 239, 241, 243, 255, 256

Rentschler, H., 82, 85, 105 Reyniers, J. A,, 115, 160 Reznikova, S. B., 46, 63

385

Rial, J., 269, 311 Riheiro, 0. F., 260, 261, 506 RibBreau-Gayon, J., 72, 74, 81, 86, 88, 89, 100, 102, 105, 106 Rice, E. E., 258, 261, 263, 276, 282, 283, 284, 293, 295, 298, 299, 300, 304, 508,310 Rich, S. R., 42, 43, 63 Richardson, J. H., 136, 137, 160 Richardson, W. D., 24, 63 Richmond, W. L., 31, 63 Richou, R., 203, 225 Rickhcr, C. J., 202, 228 ltidder, L. E., 296, 307 Ridlry, G. B., 315, 571 RirgIemaIi, S., 100, 105 Riester, D. W., 182, 280 Rist, C. E., 18, 57 Rivoche, 14:. J., 235, 236, 237, 239, ,356 Roberts, J., 136, 159 Robinson, A. D., 277, 278, 279, 280, 286, 287, 288, 289, 290, 295, 307 Robinson, H. E., 261, 263, 276, 298, 300, 304, 310 Robinson, H. hl., 8, 11, 12, 13, 29, 32, 36, 60, 61 Eobinson, It. H. A f . , 192, 225 Kobinson, W. B., 184, 185, 260 Robinson, W. D., 261, 310 Rodrigues, L. D., 260, 311 Roerig, R . K., 197, 214 Roger, 120, 166 Rohrman, F. A., 237, 249, 251, 254 Rombouts, E., 39, 57 Roos, L., 70, 106 Rose, I<. D., 168, 192, 193, 194, 199, 203, 218, 219 Rose, W. G., 15, 63 Ross, A. F., 249, 256 Ross, J. H., 17, 63 Ross, S., 23, 57 Ross, W.,277, 294, 508 Rosser, F. T., 225 Roth, W.,42, 43, 63 Roy, J. K., 259, 274, 310 Rubens, J. G., 15, 63 Rubin, S. H., 268, 308 Rugala, A. A,, 292, 309 Rundle, R. E., 34, 56 Ruster, M., 114, 128, 137, 138, 143, f69

386

AUTHOR INDEX

S Sah, P. P. T., 100, 105 Saiki, T., 118, 161 St. John, E. Q., 207, 224 Sair, L., 46, 58, 192, 193, 226 Sakaki, M., 261, 311 Sakashita, T., 261, 311 Salaman, R. N., 232, 256 Sanborn, R. E., 226 Sanderson, N. H., Jr., 183, 205, 206, 226 Sarett, H. P., 276, 277, 310 Sarles, W. B., 112, 130, 160 Satterfield, G. H., 86, 102 Saunders, R., 117, 118, 157 Savage, W. G., 118, 141, 160 Scarlett, W. J., 211, 226 Schachinger, L., 18, 58 Schaefer, W. C., 18, 57 Schanderl, H., 88, 89, 106 Schauer, L., 289, 309 Scheffer, W. R., 88, 106 Scheid, M. V., 228 Schetty, O., 14, 63 Schmidt, W., 117, 118, 144, 154, 157, 160 Schmidt, W. H., 33, 65 Schneider, F., 17, 63 Schneider, M. D., 193, 203, 226 Schneiter, R., 194, 286 Schoch, T. J., 35, 63 Schoening, H. W., 203, 226 Schrenk, O., 40, 59 Schultz, A., 70, 106 Schulte, A. S., 261, 265, 269, 270, 306, s10

Schulta, H. W., 261, 263, 299, 300, 304, 310 Schulte, T. H., 27, 62 Schule, A., 266, 296, 310 Schweigert, B. S., 86, 106, 276, 277, 279, 309, 311 Schweitzer, T. R., 265, SO9 Scott, W. J., 114, 115, 117, 118, 128, 129, 130, 132, 134, 135, 136, 137, 138, 139, 140, 141, 150, 156, 157, 158,160 Sealey, J. Q., 226 Seck, W., 34, 63 Sedgwick, W. T., 170,126 Seifert, W., 71, 97, 106 Selter, H., 118, 160

Semichon, L., 70, 74, 106 Severson, D. E., 232, 246, 252, 254 Sharokh, B. H., 280, 294, 310 Sharp, J. G., 191, 226 Shaw, A. O., 277, 307 Shaw, W. L., 316, 371 Shherbatenko, V. V., 266, SO1 Shearer, A. R., 203, 217 Shearer, W. N., 25, 58 Sheft, B. B., 31 1 Shelton, R. S., 262, 302, $07 Sherman, H. C., 259, 31 1 Sherman, J. M., 81, 82, 102, 104, 194, 203, 266, 227 Sherman, V. W., 204, 226 Sherwood, R. C., 265, 308 Sherwood, T. K., 347, 371 Shetlar, M. R., 267, Sf1 Shilling, W. L., 17, 65 Shillinglaw, C. A., 226 Shimwell, J. L., 82, 98, 99, 106 Shrader, J. H., 180, 182, 226 Siciliano, J., 238, 254 Signer, R., 24, 61 Silveira, V. G., 347, 362, 368, 3;U, 37f Simon, M., 368, 371 Singh, B., 17, 63 Sisler, F. D., 193, 226 Sisson, S., 116, 160 Sjostcdt, Ph., 14, 63 Skinner, C. E., 139, 160 Slater, E. C., 263, 269, 3f 1 Slater, L. E., 50, 63, 125, f 6 0 Slocum, R. R., 194, 2f9 Sluder, J. C., 235, 237, 238, 243, 246, 217, 251, 252, 254, 256 Smart, H. F., 179, 183, 184, 205, 866, 267 Smiley, K. L., 82, 104 Smith, A. U., 169, 225, 221 Smith, E., 188, 218 Smith, E. C., 135, 137, 139, 141, 142, 143, 144, 145, 158, 169, 161 Smith, E. F., 170, 227 Smith, G. S., 246, 255 Smith, J. A. B., 117, 160 Smith, M. B., 280, 294, 307 Smith, S. E., 217 Smith, W. H., 278, 284, 289, 30; Snell, E. I?., 86, 106, 107 Snog-Kjare, A., 86, 105

AUTHOR INDEX

Solowey, M., 194, 250 Sotola, J., 136, 169, 160 Spear, J., 278, 310 Speiser, R., 27, 28, 63 Spray, R. S., 118, 128, 160 Stabile, 6. N., 197, 214 Stadtman, E. R., 48, 64 Stadtman, T. C., 82, 107, 184, 185, 228 Stainsby, W. J., 39, 59 Stamberg, 0. E., 233, 234, 294, 256, 311 Stammn A. J., 347, 369 Stanley, J., 8, 40, 41, 42, 64 Stansbury, M. F., 46, 58, 64 Stark, C. S . , 203, 226, 227 Stark, P., 203, 226,227 Steel, K. A., 37, 64 Steffen, G. I., 216, 224 Steigmann, A,, 24, 64 Steinkraus, K. H., 148, 156, 160, 161 Stern, It. M.,286, 287, 289, 290, 308 SteL ens, H. B., 285, 3 1 1 Stewart, G. F., 22, 56, 60, 64, 141, 143, 141, 154, 1&5, 194, 195, 229, 230 Gteaart, ,J J., 227 Stewart, 31. M.,189, 887 Stilrs, G. IT.,207, 2% Stillr, B., 169, 170, 208, 227 Stoddard, E. S., 234, 256 Gtohes, J. L., 86, 107 Stokcs, R., 181, 219 Stone, M., 277, 294, 295, 311 Stone, 11. P., 220 Gtrahan, J. IT, 274, 36.9 Gtraka, K. P., 199, 202, 207, 227 Stricher, P., 16, 63 Strong, F. 31., 86, 10; Stuart, C. A., 218 Stuart, L. S., 227 Sullivan, B., 265, 308 Sulzbacher, \V. L., 143, 148, 161, 188, 205, 227 Svedberg, T , 27, 64 Snaniinathan, M., 274, 311 Smenson, T. L., 193, 194, 21.9, 220, 297 Swift, H. F., 165, 227 Sainglc, D. B., 170, 227

T Tada, M., 260, 301, 311 Takedn, K., 261, 309

387

Talayrach, M. J., 137, 139, 142, 161 Tanner, F. W., 143, 161, 170, 179, 180, 202, 205, 225, 227, 228, 229 Tarassuk, N. P., 31, 64 Tarlowsky, E., 31 1 Tarozzi, G., 119, 161 Tatarenko, J. S.,224 Taub, A., 262, 302, 311 Tauber, H., 31 1 Tavnres, A,, 260, 311 Taylor, W. W., 169, 220 Terao, K., 261, 311 Terao, M., 261, 311 Teunisson, D. J., 182, 228 Teysseire, Y., 267, 268 Thierfelder, H., 115, 159 Thole, F., 118, 161 Thorn, C., 202, 217 Thompson, F. K., 24, 58 Thompson, H. H., 46, 65 Thompson, S. Y., 261, 293, 307, 309 Thorp, F. H., 314, 312 Thorp, J. M., 265, 266, 269, 308 Thuman, W. C., 21, 64 Thurman, B. H., 30, 64 Tiemann, H. D., 347, 372 Tjomsland, A., 232, 266 Tobey, R.L., 286, 287, 288, 289, 295, 30.9 Tomarelli, R., 14, 58 Tombins, R. G., 139, 142, 159, 161 Tompkins, M. D., 259,281,285,286,287, 288, 289, 291, 292, 296, 308 Torossian, C., 220 Townsend, C. T., 203, 209, 824 Townsend, R. O., 294, 308 Treadway, R. H., 238, 247, 254,35cia Trefethen, I., 287, 288, 311 Tressler, D. K., 48, 64, 187, 210, 828, 286, 287, 288, 289, 290, 300, 307, 308, 311 Trexler, 1’. C., 115, 160 Trout, G. M., 197, 217, 628 Troy, V. S., 182, 215, 223, 224 Tschistjakow, F. M., 166, 298 Tsuchiya, H. M., 139, I60 Tsuli, F., 274, 309 Turker, R. E., 278, 279, 311 Tucker, W. H., 141, 161 Turnbull, F., 278, 310

388

AUTHOR INDEX

W

Turner, J. R., 23, 64 Turner, T. B., 170, 228

U Ulrich, J. A., 168, 169, 197, 216, 225 Umbreit, W. W., 81, 103

V Vaeck, S. V., 39, 64 Vail, G. E., 277, 294, 295, 311 Vallin, I., 263, 281, 283, 292, 302, 310 Valteich, H. W., 46, GR Van Arsdel, W. B., 234, 266, 316, 317, 318, 319, 322, 323, 347, 348, 349, 350: 351, 352, 353, 354, 357, 358, 362, 364, 367, 872 van den Broek, C. J. H., 171, 228 van dcr Mijll Dekker, L. P., 267, 293,310 Van Duyne, F. O., 224, 278, 507 van Ekenstein, W. A., 17, 60 Van Ekeltine, W. P., 168, 298 van Hook, A., 53, 64 Van Wiel, C . B., 70, I07 Van Oijen, C. F., 195, 228 Vanvertx, J., 118, 156 Vass, A. F., 170, 228 Vaughn, R. IT., 75, 79, 81, 82, 83, 84, 85, 86, 89, 92, 93, 96, 99, 100, 101, 102, 103, 107, 184, 228 Vavich, M. G., 286, 287, 289, 290, 206, 297, 508 Verona, O., 100, 107 Vesterhus, R., 189, 214 Vetsch, U., 86, lo4 Vickery, J. It., 114, 128, 137, 138, 140, 143, 154, 157, 161, 189, 218 Visenjou, L. B., 50, 64 Visnyei, K., 287, 308 Visser't Hooft, F., 85, 92, 93, 107 Vogel, M., 264, $11 Voisenet, E., 81, 107, 108 Volpertas, Z., 236, 239, 256 Volz, F. E., 136, 161, 208, 226 von Fodor, J., 118, 157 von Loesecke, H. W., 183, 224 von Schelhorn, M., 228 Vosti, D. C., 86, 108 Votkina, L. S., 42, 56

Wagner, J. R., 232, 246, 252, 254, 368, 371 Wagner, M., 115, 160 Waibel, P. E., 262, 301, 311 Waisman, H. A,, 275, 310 Wallace, G. I., 170, 179, 180, 194, 202, 205, 228, 229 Wallace, M. D., 169, 229 Walter, M., 200, 216 Walters, L. S., 99, 108 Ward, W. H., 26, 59 Watanabe, A., 260, 261, 301, 311 Watcrhouse, C. E., 302, 310 Watson, A. J., 194, 207, 223 Watt, B. K., 86, 108 Webb, B. H., 30, 64 Weckel, K. G., 293, 311 Weeks, 0. B., 220 Weinzirl, J., 143, 154, 159, 161, 183, 229 Weir, C . E., 289, 311 Weisemann, C., 47, 58 Weiser, H. H., 113, 120, 150, 158, 161, 165, 185, 187, 195, 196, 209, 229, 230 Weiser, R. S., 168, 171, 229 Welker, P. L., 48, 64 Wenzel, F. W., 220 Werkman, C. H., 86, 108 Werts, M., 237, 256 Wertz, A. W., 289, 293, 309, 511 Westerman, D. B., 277, 294, 295, 311 Weybrew, J. A., 300, 310 Wheeler, K. A., 287, 289, 290, 300, 508, 31 1

"hitacre, J., 273, 274, 310 White, E. G., 271, 511 White, J. C. D., 31, 58 White, J. H., 128, 161 White, J. W., 28, 56, 59 White, L. S., 209, 210, 229 White, T. A., 49, 56 Whymper, R., 14, 37, 40, 43, 64, 66 Wiechel, A., 118, 161 Wiegand, E. H., 181, 218, 314, 375' Wiese, A. C., 279, 295, 309 Wiesman, C. K., 136, 137, 160 Wilcox, E. B., 279, 311 Wilkin, M., 195, 629, 250 Willaman, J. J., 28, 66

389

AUTHOR INDEX

Willets, A. K., 237, 266 Willett, E. L., 277, 294, 308 Williams, 0. B., 189, 230 Williams, R. R., 276, 277, 278, 279, 280, 287, 288, 289, 290, 309 Williams, V. R., 268, 271, 311 Williamson, B. W., 228 Williamson, R. V., 8, 65 Willich, R. K., 294, 311 Wilson, C. P., 27, 65 Wilson, E. C. G., 271, 311 Wilson, J. B., 112, 130, 160 Wilson, G. S., 173, 229 Wiister, G. H., 197, 229 Wingrr, E. L., 3, 52, 65 Winogradom, RI., 116, 161 Winogradowa-Fedorawa, T., 116, 161 Winslom, C. E. A., 170, 226 Winter, A. R., 178, 194, 195, 196, 204, 206, 207, 293, 229, 230 Wise, C. S., 18, 57 Wodicka, V. O., 265, 269, 297, 298, 304, 307, 308

Wolbach, S. B., 118, 161 Wolford, E. R., 174, 180, 181, 182, 230 Wolfrom, M. L., 17, 18, 65 Wood, E. R., 240, 250, 252, 255, 256 Wood, H. G., 86, 108 Woodmansee, C. W., 27, 56

Woodroof, J. G., 11, 46, 47, 56, 58, 65, 188, 230 Woodruff, S. J., 48, 84 Woodward, C. F., 238, 254 Wright, A. M., 188, 230 Wrinkle, C., 178, 194, 195, 196, 204, 206, 207, 229, 630 Wyssokowitsch, W., 118, 161

Y Yesair, J., 189, 230 Young, J. M., 267, 268, 310 Young, S., 180, 230 Yule, W. T., 314, 312 Yurchenco, M. C., 172, 25'0 Yurchenko, J. A., 172, 230 Z

Zaehringer, M. V., 265, 266, 311 Zagaesky, J. S., 230 Zalkowski, L., 114, 128, 137, 138, 143, 159 Zaitseva, Z. I., 266, 307 Zenlea, B. J., 41, 65 Zerban, F. W., 16, 65 Ziels, N. W., 33, 65 Ziemba, J. V., 49, 65 Zobell, C. E., 189, 230 Zwick, A., 118, 161

Subject Index A

Antioxidant, butylated hydroxyanisole, 13 propyl gallate, 12 Apples, dehydrated, 316 Apricots, thiamine retention or storage, 297, 304 Asparagus, thiamine loss, 286, 296

Acetobacter, species in wine, 92, 93 Acid acetic, in wines, 74, 76 amino, decrease thiamine destruction, 259 ascorbic, loss, 301 citric, in spoiled table wines, 76 B galacturonic, 27 gluconic, in spoiled table wines, 76 Babassu, lauric fats in, 32 lactic, formation of, 7 1 Bacon, loss of thiamine on processing, from glycogen, 123 276 in spoiled table wines, 91 Bacteria in wines, 74 freezing death of, 170 lauric, in candy production, 32 in frozen eggs, 194 replacement for cocoa butter, 44 on frozen fish, 190 linoleic, in cocoa fat, 39 frozen meat, 188 malic, bacteria decomposition of, 7 1 frozen poultry, 192 in wines, 73 growth factor for, in wine, 85 rnonochloracetic, in wines, 100 lactic acid, in California wines, 95 nordihydroguaiaretic, antioxidant, 13 in meat animals, 112 oleic, elaidizing, 44 pathogenic, survival time in frozen in cocoa fat, 39 strawberries, 166 propionic, bacteria capable of forming, on refrigerated beef, 138 81 slime-forming on meat, 140 in spoiled wines, 74 survival of, in frozen sucrose aolristearic, in cocoa fat, 39 tions, 167 tartaric, in candy making, 17 vinegar, in spoiled table wines, 76, 92 in table wines, 72 wine spoilage caused by, 75 Albumin, foaming of solutions, 21 in wines, 70, 90, 92, 97, 99 in candy production, 20 Beans modifications of sugar properties profrozen green, 175, 179 duced by, 53 green cut, loss of thiamine on processAlcohol, concentration in wines, 84 ing, 286 in California table wines, 82 lima, decrease in bacterial content by tolerance by wine spoilage bacteria, 91 blanching, 184 Almond paste, in candies, 45 thiamine destruction in, 258, 286, Almonds, composition of, 46 296 used by the candy industry, 3 snap, loss of thiamine on processing, 286 Amylopectin, structure of, 34 thiamine retention on storage, 2 9 i Amylose, structure of, 34 Beef Antibiotics, used in meat, 150 bacteria in refrigerated, 138 390

391

SUBJECT INDEX

fat rancidity in, 142 ground trimmings, 146 prepared in temperate zones, 115 slaughter of, 124 slime on fresh, 145, 147 thiamine loss in, 275, 276, 283 Beer, ropiness of, 82 Beets decrease in bacterial content during blanching, 184 dehydrated, loss of thiamine on processing, 287, 300 Berries frozen, 16'3 microorganisms in, 209 sources of infection of, 182 microbial development on, 180 Blood, bacteria count of, 188 Brain, of meat animals, 133 Bread baking canned, 269 dough, freezing storage of, 197 thiamine in, 266 loss of thiamine in, 262, 265, 271, 296 Broccoli decrease in bacterial content by blanching, 184 frozen, 176 Brussel sprouts, loss of thiamine on processing, 287 Butter bacterial count in freezing storage, 193 flavor in candies, 31 Butter creams butter in, 20 stability of fats in, 13 Buttermilk, in muffins, 272 Butterscotch, stability of fats in, 12 C Cabbage decrease in bacterial content by blanching, 181 dehydrated, thiamine loss in, 300 loss of thiamine on processing, 287 puree, thiamine destruction, 292 Caffeine, in cocoa beans and nibs, 38 Calcium carbonate, for molding hard creams, 36

Candy chemical composition of, 10 color, 8 made from fruit juices, 28 manufacture, 1 physical properties, 6 production methods, 48 Caramelan, product of sugar decomposition, 16 Caramels chemical composition and properties of, 10 drying, 7 ingredients in, 29 stability of fats in, 12 Carbohydrates in cocoa beans, 3 8 in wine, 86 Carbon dioxide in spoiled table wines, 76 in wines, 71 Carrots decrease in bacterial content b y blanching, 184 loss of thiamine, 287, 295 puree, thiamine destruction, 292 Cattle digestive process of, 116 skinning, 152 slaughter of, 125 Cauliflower decrease in bacterial content by blanching, 184 frozen, 176 loss of thiamine, 287 Cellulose, in cocoa beans and nibs, 38 Cereal breakfast, thiamine loss, 269 human food, 231 protective effect on thiamine, 261 thiamine loss in, 26-2, 274 Cheese thiamine in, 261, 262, 293, 304 in processed, 298 Cherries maraschino, in candies, 45 microorganisms in frozen, 179 Chicken B la kine. bacterial count of frozen, 200 chow-mein, frozen, 199 Yl

392

SUBJECT INDEX

home frozen creamed, 207 loss of thiamine in cooking, 280 salmonella in eviscerated, 193 thiamine retention in frozen, 294 Chocolate bloom, 14 candies, 30, 37 creams, stability of fats in, 12 milk, 44 physical properties of, 39 types of candy, 2 Chow-mein, bacterial count of frozen, 199 Citrus, loss of thiamine in, 296 Clostridium botulinum in frozen foods, 202 toxin formation by, 210 in vegetables, 185 Cocoa beans, ground, 37 composition of, 37 butter, in candies, 29, 37 properties of, 40 stability of, 14 fat, composition of, 39 effect of lecithin upon, 41 nibs, composition of, 38 products, composition of, 37 used by the candy industry, 3 solids, in candies, 29 Coconut in candies, 47 lauric fats in, 32 used by the candy industry, 3 Coconut oil, replacement for cocoa but,ter, 44 Cod, bacterial flora of, 191 Corn loaves, thiamine in, 274 loss of thiamine on processing, 287 Corn sirup composition and properties of, 19 proportions of, 11 Corn starch aging of gels of, 12 used by the candy industry, 3 Cottonseed, oleic and linolenic fats in, 32 Cream, bacterial content during freezing storage, 197 Creams, crystallized, 36

D Dairy products, bacterial count, 197 Dehydraters, tunnel basic theory of, 347 classification of, 316 compartment, 321 criteria for selection, 345 development of, 311 fruit, 315 heating systems, 332 instrumentation, 333 materials of construction, 335 mechanical elements of, 326 operating procedures, 363 optimum tray loading, 355 Oregon tunnel, 314 potato granule, 234 static pressure drop, 328 typical commercial, 339 Dextrin, binding agent in lozenge paste, 29 Dextrose, in chocolate candies, 44 Dihydroxyacetone, in spoiled table wines, 76

E Egg albumin, protective action on thiamine, 261 spray, dried whole, 300 yolk, thawing of frozen, 206 Eggs and egg products used by the candy industry, 3 Emulsifiers, in chocolate candies, 44 Esters, aminoethyl, 15 csarbobenzoxyaminoethyl, 15 Ethyl alcohol in table wines, 17

F Fans, axial flow in dehydraters, 330 Farina, enriched with thiamine, 269 Fats in candies, 29, 32 in cocoa beans and nibs, 38 cooked a t high temperatures, 4 ingredients of candy, 12

SUBJECT INDEX

3 93

and oils, used by the candy industry, 3 thiamine loss in, 296 oleic-linolenic in candy production, 32 Fruit juices, for candies, 28 protection of bacteria during freezing Fruit products, used by the candy industry, 3 by, 167 rancidity of meat, 138, 142 Fudge stability of, 12 chemical composition and properties Fatty acids of, 10 in cocoa fat, 39 drying, 7 esters of sorbitan and polyoxyethylene, improved by sorbitol, 11 15 ingredients in, 29 iiberation of free, 14 stability of fats in, 13 Fiber, in cocoa beans and nibs, 38 Fish G bacteria in frozen, 190 psychrophilic flora of, 18!1 Garlic, dehydrated, 315 stews, bacterial count of frozen, 199 Gelatin supercooling of, 164 binding agent in lozenge paste, 29 Flavor modification of sugar properties proin candies, 9, 31 duced by, 53 caramel, 31 in production of marshmallows, 24 Flounder, loss of thiamine in, 280 protective action on thiamine, 261 Flour, thiamine in, 262, 263, 267 Gels Fondant, chemical composition and propfinal pH, 26 erties of, 10 properties of, 35 Foods, losses of thiamine in, 282 Glucosan, from glucose, 16 Frozen foods Glucose cooking of, 210 effect on thiamine destruction, 261 defrosting of, 201 formed from sucrose, 16 hygienic aspects of, 210 in spoiled table wines, 76 microbiological methods of analysis, structure and simple reactions of, 4 173 in table wines, 72 occurrence of bacteria in, 172 utilization by spoilage bacteria in packaging problems, 209 winc, 91 pathogenic bacteria in, 201 Glycerides, in cocoa fat, 39 precooked, bacteria in, 169, 198 Glycerol Fructose anhydrous, 20 anhydrides, 17 in candies, 13 in spoiled table wines, i(i i n laboratory rations, 301 structure and simpler reactions of, 4 in table wines, 72 in table wines, 72 use of, 11 utilization by spoilage bacteria in wine, in frozen food, 169 91 Glycogen, in muscular tissue, 122 Fruit Grapes in candies, 29, 45, 48 recommended amount of sulfur diordehydration of, 367 ide added to musts, 89 dried, used by confectionery industry, Gum 4s arabic, candy applications of, 29 frozen, 160, 174, 177 drops, 29 microorganisms on, 179 produced with tapioca, 34 thawing of frozen, 206 tragacanth, candy applications of, 29

394

SUBJECT I N D E X

Gums in cocoa beans and nibs, 38 natural, 29 protective action on thiamine, 261

H Halibut, Ioss of thiamine in, 280 Ham patties, home frozen, 207 salt-cured, 123 souring, 138, 141 thiamine loss in, 275, 276 Heart, of meat animals, 133 Hogs butchered on the farm, 123 carcasses of, 149 slaughter of, 125 Hydroxymethylfurfural, end products of, 17

I Invertase, used to develop fluidity in candies, 36 Isosaccharosan, anhydride of sugar, 16

J Jams, in candies, 45 Jellies chemical composition and properties of, 10 fruit, 26 toughening of, 7 Juices frozen fruit, 169 sources of infection of, 182 fruit, microorganisms on, 179

K Kale, loss of thiamine on processing, 288 Keto-gluconic acids, in spoiled table wines, 76 Khandi, Arabic candies, 6 Kidneys loss of thiamine on processing, 278 of meat animals, 133

L Lactose in candies, 30, 45 effect on thiamine destruction, 261 Lamb organisms in frozen, 188 retention of thiamine in, 283 Lanolin, viscosity reducer in chocolate, 40 Lecithin, added to chocolate, 40 Levulosan, formed from sucrose, 16 Linolenic fats, use in candy, 32 Liver loss of thiamine on processing, 277 in meat animals, 113, 133 Lozenge paste, binding agent in, 29

M Maltose, hydrate, 19 bfannitol, in spoiled table wines, 77, 91 Mannose, presence in candies, 17 Marshmallows chemical composition and properties of, 10 gelatin in production of, 25 setting of, 36 toughening of, 7 Marzipan, in candies, 45 Meat animals, microbiology of, 109, 111 bacterial count during freezing, 188 chilling of, 135 contamination of in self-service stores, 189 frozen, 164, 188 molds and yeasts on, 139 quality in, 121 spoilage of, 110, 115 thermal losses of thiamine in, 275 Microorganisms in frozen meat, 188 influence of freezing rate, 168 of freezing temperatures on, 164 in meat animals, 115 Milk bacterial content during freezing storage, 197 evaporated color of, 30 in candies, 30

395

SUBJECT INDEX

in precooked frozen foods, 199 in muffins, 272 in precooked frozen foods, 199 products, in candies, 30 used by the candy industry, 3 protection of bacteria during freezing by, 169 protein-bound thiamine in, 261 solids, addition t o potatoes, 234 combination of sugars and, 2 in milk chocolate type candies, 37 percent in candies, 29 thiamine loss in, 293 whipping agent from protein of, 2-2 Mineral salts, in candies, 30 Molds discolorations on meat, 142 on refrigerated meats, 139 Monoglycerides, type of emulsifier, 12 Muffins, thiamine retention in, 272, 273 Mutton, thiamine loss in, 275, 276

N Niacinamide, 302 Nitrogen, in cocoa beans and nibs, 38 Nougats chemical composition and properties of, 10 improved by sorbitol, 1 1 stability of fats in, 12 toughening of, 7 Nuts in candies, 29, 45, 47, 48 used by the candy industry, 3

0 Oat flour, extracts of, 14 Oils in coconut, 48 in nutmeats, 48 used by the candy industry, 3 vegetable, winterizing of, 44 Oleodipalmitin, in cocoa fats, 39 Oleodistearin in cocoa fats, 39 use of, 15 Oleopalmitostearin, in cocoa fat, 39

Onions dehydrated, 315 loss of thiamine, 288 Orange, soft rot of, 181 Orange juice frozen, 169 bacterial count of, 181 germicidal effect, 202 lactose-fermenting yeast occurring in, 182 loss of thiamine, 295, 304 survival of bacteria in, 164 Oysters, loss of thiamine in, 280

P Palm oil, lauric fats in, 32 Palmitic acid, in cocoa fat, 39 Palmitostearin, in cocoa fats, 39 Parsnips, loss of thiamine on processing, 288 Peaches dehydrated, 316 loss of thiamine in canned, 296 microorganisms in frozen, 179 Peanut butter, in candy, 45 Peanuts in candies, 29, 45 oleic-linolenic fats in, 32 thiamine content of, 293, 299, 304 used by the candy industry, 3 Pears, dehydrated, 316 Peas bacterial counts during preparation for freezing, 186 decrease in bacterial content by blanching, 184 frozen, 175 microorganisms in, 179, 186 thiamine destruction in, 258, 285, 288, 296 Pectin chemical nature of, 26 in cocoa beans and nibs, 38 in gum drops, 29 high and low methoxyl, 28 Pentosans, in cocoa beans and nibs, 38 Pepper, bacterial contamination of pork sausage by, 189 Peppers, dehydrated, 315

396

SUBJECT INDEX

PH effect on thiamine destruction, 259, 272 of muscle of meat animals, 122 of wines, 83 Pigments, in cocoa beans and nibs, 38 Pineapple, loss of thiamine in, 296 Polyoxyethylene sorbitan monostearate, in chocolate bars, 44 Polysaccharides in spoiled table wines, 78 utilization by spoilage bacteria in wine, 91 Pork dehydrated, thiamine destruction, 261, 263 thiamine in, 300 loss of thiamine in, 275, 281, 282, 304 microbial count of frozen, 188 thiamine decomposition in, 258 retention in frozen, 294 Pork and beans, frozen, bacterial count of, 199 Poultry, bacterial flora on, 192 Potato granules add-back method of drying, 238 calories per pound, 233 characteristics of, 232 color of, 218 drying methods and equipment, 245 effect of conveying system on quality of, 210 of specific gravity of, 243 of temperature of mixing and tcnipering on, 242 nutritive value of, 249 processes for manufacture, 236 product quality of, 217 quality evaluation of, 249 weight, space requirement, 233 Potato shreds, 234 Potatoes in beef stew, thiamine retention, 281 calories per pound, 233 canned, calories per pound, 233 weight, space requirement, 233 cooked, freezing of, 237 decrease in bacterial content b y blanching, 184 dehydrated, as food staple, 232 calories -per -pound, 233

thiamine in, 300 weight, space requirement, 233 loss of thiamine on processing, 289 mashed, thiamine in, 292 non-waxy type, 243 preservation of, 231 riced, development of, 233 spray drying, 234 sweet, loss of thiamine, 290 processing of, 290, 292 vitamin C content of, 249 weight, space requirement, 233 Protein in candies, 30 in cocoa beans and nibs, 38 denaturation by freezing, 161, 167 milk, loss in browning, 31 modifications of sugar properties produced by, 53 in nutmeats, 48 soy, whipping agent, 22 whipping agents, 20, 53 Pumpkin, loss of thiamine on processing, 290

R Radish, loss of thiamine in, 296 Radish juice, loss of thiamine, 295 Rancidity in candies, 12 oxidative, 13 Riboflavin, 261, 302 Rice home frozen, 207 thiamine loss in, 274 Rutabagas, dehydrated, thiamine in, 300

S Salmonella in frozen foods, 201 in frozen poultry, 193 Salt bacterial contamination of pork sausage by, 189 tolerance by wine spoilage bacteria, 91 Sausage casings, 134 pork, organisms in, 188 Scones, losses of thiamine in baking, 271

397

SUBJECT INDEX

Sheep digestive process of, 116 skinning, 152 slaughter of, 125 Shellfish, frozen, 213 Sherry wine, Spanish-type, 88 Shrimps, bacteria species in, 191 Sorbitan monostearate, in chocolate bars, 12 Sorbitol, fudge and nougat improved by, 11

Soups, bacterial content of frozen, 199 Soy, use in candy, 32 Soya flakes, in dehydrated pork loaves, 300 Soybeans loss of thiamine on processing, 299 thiamine retention in, 285 Spaghetti and meat balls, home frozen, 207 Spinach bacterial count of, 186 frozen, bacterial count of, 186 loss of thiamine on processing, 290, 295 Squash, loss of thiamine on proressing, 290 Staphylococcus in frozen food, 200 toxin formation by, 210 Starch in candies, 29, 33 in cocoa beans and nibs, 3 8 corn and wheat, 12 dust explosions, 37 jellies, ingredients in, 29 made by, 33 molding of, 36 tenderness of, 12 molds in, 12 potato, physical characteristics of, 242 rehydration of, 234 soluble, protective action on thiamine, 261, 284 in reconstituted potato granules, 251 trioses in hydrolysates, 19 Strawberries microbiology of, 181 microorganisms in frozen, 179 survival time of pathogenic bacteria in frozen, 166

Stearates, polyoxyethylene, 12 Sucrose in chocolate candies, 44 dehydration products of, 16 effect on survival of bacteria, 167 on thiamine destruction, 261 and stability of thiamine, 30 structure and simpler reactions of, 4 Sugar in candy making, 15 chemical reactions of, 2 in chocolate, 37 combination of milk solids and, 2 effect of heat upon, 16 modification of properties, 20 in precooked frozen foods, 199 protection of bacteria during freezing by, 169 structures and simpler reactions of, 4 in wines, 73, 84 Sulfur dioxide effect of, in wines, 8 i recommended amount in wine making, 89 Sweets, British term, 6

T Tannins in cocoa beans and nibs, 38 effect on bacteria in wine, 87 Tapioca, gum drops produced with, 34 Taro, loss of thiamine on processing, 290 Theobromine, in cocoa beans and nibs, 38 Thiamine, see Vitamin B, Tomato juice, thiamine destruction in, 258, 297 Tomatoes, loss of thiamine on processing, 290, 296 Triolein, in rocoa fat, 39 Turkey salmonellae on skin of, 194 thiamine loss in, 280 thiamine retention in frozen, 294 Turnips, loss of thiamine in on processing, 290

U Urea, as dispersing agent, 24

398

SUBJECT INDEX

V Vegetables dehydrated, 300, 315, 367 frozen, 174 bacterial flora in, 183, 185, 187 danger of botulism in, 185 thawing of, 206 Indian, thiamine destruction in, 292 losses of thiamine during processing, 285 thiamine loss in, 296 Vitamin B, in baked products, 271 in breads, 265 in dehydrated products, 299 form of the vitamin, 261 loss in canned products, 296 in dairy products, 293 in meats, 275 in processing vegetables, 285 pharmaceuticals containing, 301 thermal destruction in foods, 257, 291 Vitamin C, retention in potatoes, 2-19

W Whalemeat, bacterial flora in, 191 Wheat germ, Vitamin B, in, 263, 296 Wheat starch, aging gels of, 12 Wine bacteria in California, 90 chemical abnormalities induced by bacteria, 76 Danish fruit, 82 effect of air on, 89 growth of bacteria in, 82 microbiology of, 67 slimy or ropy, 82 spoilage caused by bacteria, 75, 76 storage temperature, 89

Y Yeast dried stabilizer in candy, 13 extract, thiamine destruction in, 294, 304 thiamine in, 262 on refrigerated meats, 130 wine, 87

This Page Intentionally Left Blank