Advances in Insect Physiology, Volume 7

Advances in

INSECT PHYSIOLOGY

VOLUME 7

This Page Intentionally Left Blank

Advances in

Insect Physiology Edited by

J . W. L. BEAMENT, J. E. TREHERNE and V. B. WIGGLESWORTH Department of Zoology, The University, Cambridge, England

VOLUME 7

1970

ACADEMIC PRESS London and New York

ACADEMIC PRESS INC. (LONDON) LTD Berkeley Square House Berkeley Square London, W1X 6BA

US.Edition published by ACADEMIC PRESS INC. 111 Fifth Avenue, New York, New York 10003

Copyright @ 1970 by Academic Press Inc. (London) Ltd

All Rights Resewed NO PART OF THIS BOOK MAY BE REPRODUCED IN ANY FORM BY PHOTOSTAT, MICROFILM, OR ANY OTHER MEANS, WITHOUT WRITTEN PERMISSION FROM THE PUBLISHERS

L i b r a v of Congress Catalog Card Number: 63-14039 SBN: 12-024207-9

Printed in Great Britain by The Whitefriars Press Ltd London and Tonbridge

List of Contributors toVolume 7 M. ASHBURNER, Department of Genetics, University of Cambridge, England (P. 1) L. J . GOODMAN, Department of Zoology, Queen Mary College, University of London, England (p. 97) G. HOYLE, Department of Biology, University of Oregon, Eugene, Oregon, U.S.A. (p. 35 1) P. A. LAWRENCE, Department of Genetics, University of Cambridge, England (P. 197) B. SACKTOR, Gerontology Research Center, National Institutes of Child Health and Human Development, National Institutes of Health, Baltimore, Maryland, U.S.A. (p. 267)

V

This Page Intentionally Left Blank

Contents LIST OF CONTRIBUTORS TO VOLUME 7

. . . . . . . . . .

V

FUNCTION AND STRUCTURE OF POLYTENE CHROMOSOMES DURING INSECT DEVELOPMENT MICHAEL ASHBURNER I. I1. 111. IV .

Introduction . . . . . . . . . . . . . . . . . Polytene Chromosome Structure . . . . . . . . . . . Polyteny and Endopolyploidy during Insect Development . . The Occurrence of Polytene Chromosomes . . . . . . . A . Larval Tissues of Diptera . . . . . . . . . . . B. Pupal and Adult Tissues of Diptera . . . . . . . . C . Polytene Chromosomes in other Organisms . . . . . V . Puffs: Introduction and Biochemistry . . . . . . . . . A . The Unit Hypothesis of the Chromomere . . . . . . B . Puffs . . . . . . . . . . . . . . . . . . C. RNA Synthesis and Puffs . . . . . . . . . . . D . RNA Transport from the Nucleus and Nucleocytoplasmic Interactions . . . . . . . . . . . . . . . . E . The “Puff Proteins” . . . . . . . . . . . . . . F . Histones and Puffs . . . . . . . . . . . . . G . DNA Synthesis and Puffing . . . . . . . . . . . H . “DNA Puffs” . . . . . . . . . . . . . . . . VI . Puffs: Developmental Physiology . . . . . . . . . . A . Developmental Specificity . . . . . . . . . . . B. Intragland Variation . . . . . . . . . . . . . C . Tissue Specificity of Puffs . . . . . . . . . . . . VII . The Endocrine Control of Puffing Activity . . . . . . . . A . Ecdysone . . . . . . . . . . . . . . . . B . Mechanism of Ecdysone Action . . . . . . . . . C. JuvenileHormone . . . . . . . . . . . VIII . The Experimental Modification of Puffing Activity . . . . . IX . Modification of Polytene Chromosome Structure and Function as the Result of Infection . . . . . . . . . . . . . . X . The Physiology of Nurse Cell Polytene Chromosomes . . . . XI . The Physiology of Epidermal Cell Polytene Chromosomes . . XI1. The Physiological and Functional Significance of Puffing . . . A . Salivary Gland Function . . . . . . . . . . . B . Specific Correlations of Puffs and Salivary Gland Function C . General Considerations on the Physiological Function of Puffs . . . . . . . . . . . . . . . . . . XI11. Conclusions and Outlook . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . Notes Added in Proof . . . . . . . . . . . . . .‘ . . . vii

2 2 4 7 7 8 9 10 10 11 12 17 18 19 21 21 24 24 28 31 32 32 40 46 47 51 54 56 59 59 64 68 69 71 93

...

Vlll

CONTENTS

THE STRUCTURE AND FUNCTION OF THE INSECT DORSAL OCELLUS LESLEY J . GOODMAN

I . Introduction . . . . I1. Distribution and Structure A.

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

Distribution and Development of Dorsal Ocelli within the Class Insecta . . . . . . . . . . . . . . . . B. Structure . . . . . . . . . . . . . . . . C. The Visual Field of the Ocelli . . . . . . . . . . . . . . . . . . . . . . . . D. Form Perception 111. Behavioural Studies on the Role of the Ocelli . . . . . . . A . Early Work . . . . . . . . . . . . . . . . B. The Ocelli as Stimulatory Organs . . . . . . . . . C . The Ocellar Contribution to Phototactic Orientation . . D . Detection of thePlaneofPolarized Light . . . . . . E . Registration of Intensity Level and of Changes of Intensity IV . Electrical Activity in the Ocellus . . . . . . . . . . . . . . . . A . The Electrical Response of the Visualcells B. Electrical Activity in the Second Order Neurons . . . . C. Sensitivity, Light and Dark Adaptation and Flicker Fusion Frequency . . . . . . . . . . . . . . . . D . The Spectral Sensitivity of Ocelli . . . . . . . . V . Ocellar Units in the Brain and Ventral Nerve Cord . . . . . . . . . . . . . . . . A . Ocellar Units in the Brain . . . . . . B. Ocellar Unitsin the Ventral Nerve Cord C. The Effect of Ocellar Input on Compound Eye Units . . D. The Influence of the Ocelli on Motor Activity in the Thoracic Ganglia . . . . . . . . . . . . . . . . . . VI . Conclusion . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .

97 99 99 101 131 131 132 132 133 141 147 148 152 152 161 164 170 171 171 173 182 184 188 190

POLARITY AND PATTERNS IN THE POSTEMBRYONIC DEVELOPMENT OF INSECTS PETER A . LAWRENCE I. I1 .

Introduction . . . . . . . . . . . . . . . . . Cell Polarity . . . . . . . . . . . . . . . . . A . Insect Segmental Gradients . . . . . . . . . . B. Origin of the Segmental Gradient . . . . . . . . C . Polarized Transmission of Information During Growth and Regeneration . . . . . . . . . . . . . . . D. Growth and Gradients . . . . . . . . . . . . E . Gradients and Patterns . . . . . . . . . . . . F. Insect Segmental Gradient: a Summary of the Working Hypothesis . . . . . . . . . . . . . . . . G . Gradient Phenomena in Other Organisms: a Comparison .

197 198 199 206 207 212 216 220 22 1

CONTENTS

111.

Pattern Formation . . . . . . . . . . . . . . . A . The Development of Spaced Bristles and Hairs in Rhodnius and Oncopeltus . . . . . . . . . . . . . . . B. Genetic Mosaics . . . . . . . . . . . . . . IV . Determination and Regulation . . . . . . . . . . . A . Introduction . . . . . . . . . . . . . . . B. Experiments w i t h c u t Discs-Regulation . . . . . . . C. Dissociation Experiments . . . . . . . . . . . D . Changes in the Determined State . . . . . . . . . V . Cellular Differentiation . . . . . . . . . . . . . . VI . Outlook . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References

ix 224 224 231 235 235 237 247 254 257 259 260 260

REGULATION OF INTERMEDIARY METABOLISM. WITH SPECIAL REFERENCE TO THE CONTROL MECHANISMS IN INSECT FLIGHT MUSCLE BERTRAM SACKTOR I. I1.

111.

IV .

V.

Introduction . . . . . . . . . . . . . . . . . 268 Physiological Properties and Structural Organization of Insect Flight Muscle . . . . . . . . . . . . . . 269 A . Utilization of Oxygen during Flight . . . . . . . . 269 . . . 269 B. Supply of Oxygen and Fuel t o the Flight Muscle C . Nature of the Substrate Consumed during Flight . . . 271 D . Properties of the Contractile Proteins . . . . . . . 271 E . Morphological Organization of Flight Muscle . . . . . 275 F. Structural-Functional Correlates . . . . . . . . . 281 Regulation of Carbohydrate Metabolism . . . . . . . . 281 A . Glycogenolysis . . . . . . . . . . . . . . . 283 B. Phosphorylase b Kinase . . . . . . . . . . . . 295 C. Glycogen Synthetase . . . . . . . . . . . . . 295 D . Trehalase . . . . . . . . . . . . . . . . 296 E . Biosynthesis of Trehalose . . . . . . . . . . . 300 F . Glycolysis . . . . . . . . . . . . . . . . 303 . 310 G . Identification of Other Loci of Control of Metabolism Regulation of Fat Metabolism . . . . . . . . . . . 312 A . Fatty Acid Catabolism . . . . . . . . . . . . 313 B . The Role of Carnitine . . . . . . . . . . . . 314 C. Biosynthesis of Fat . . . . . . . . . . . . . 316 D . Mobilization and Transport of Fat . . . . . . . . 319 Regulation of Mitochondria1 Metabolism . . . . . . . . 322 . 323 A . The Respiratory Chain and Oxidative Phosphorylation . . . . . . . . . 325 B. Control of Pyruvate Oxidation C. Control of Proline Oxidation . . . . . . . . . . 330 D . Control of cwGlycero-P Oxidation . . . . . . . . 332 . . 333 E . The Energy-Dependent Accumulation of Ca2+and Pi

A.1.P.-I*

X

CONTENTS

F.

Interactions of Metabolic Effectors with the Respiratory Chain . . . . . . . . . . . . . . . . . . 334 VI . Conclusions . . . . . . . . . . . . . . . . . . 336 References . . . . . . . . . . . . . . . . . . . . . 338 CELLULAR MECHANISMS UNDERLYING BEHAVIOR -NEUROETHOLOGY GRAHAM HOYLE Introduction . . . . . . . . . . . . . . . . A . Insects and Ethology . . . . . . . . . . . . B. Behavior Amenable to Analysis: Defining Neuroethology I1 . Neural Architecture and Physiology . . . . . . . . . A . Anatomy . . . . . . . . . . . . . . . B. Cellular Physiology of Motorneurons . . . . . . C. The Electrical Activity of Neuropil . . . . . . . D . Mechanism of Habituation . . . . . . . . . . E. Memory and “Learning” Machinery . . . . . . . 111. Motor Mechanisms . . . . . . . . . . . . . . A . General Aspects . . . . . . . . . . . . . B. Respiration . . . . . . . . . . . . . . . C. Perambulatory Locomotion . . . . . . . . . D. Flight . . . . . . . . . . . . . . . . . E . Insect Song: Crickets . . . . . . . . . . . F . Courtship Behavior . . . . . . . . . . . . IV . Modelsof Neural Activity and Terminology . . . . . . V . Discussion . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . Note Added in Proof . . . . . . . . . . . . . . . . . I.

AUTHOR INDEX . . . . . . . . . . SUBJECT INDEX . . . . . . . . . . . CUMULATIVE LIST OF AUTHORS . . . . CUMULATIVE LIST OF CHAPTER TITLES . .

.

. . .

.

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

349 349 353 356 356 361 375 387 392 398 398 401 403 408 412 417 420 425 430 431 439

. . . . . . . . . . 445 . . . . . . . . . 457

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

471 473

ERRATUM TO VOLUME 6

Page 241, lines 32 and 33 should read:

This Page Intentionally Left Blank

Function and Structure of Polytene Chromosomes During Insect Development MICHAEL ASHBURNER Department of Genetics, University of Cambridge, England Introduction . . . . . . . . . . . . . . . . . Polytene Chromosome Structure . . . . . . . . . . . Polyteny and Endopolyploidy during Insect Development . . The Occurrence of Polytene Chromosomes . . . . . . . A. Larval Tissues of Diptera . . . . . . . . . . . B. Pupal and Adult Tissues of Diptera . . . . . . . . C. Polytene Chromosomes in other Organisms . . . . . V. Puffs: Introduction and Biochemistry . . . . . . . . . A. The Unit Hypothesis of t h e Chromomere . . . . . . B. Puffs . . . . . . . . . . . . . . . . . . C. RNA Synthesis and Puffs . . . . . . . . . . . D. RNA Transport from the Nucleus and Nucleocytoplasmic Interactions . . . . . . . . . . . . . . . E. The “Puff Proteins”. . . . . . . . . . . . . . F. Histones and Puffs . . . . . . . . . . . . . G. DNA Synthesis and Puffing . . . . . . . . . . . H. “DNA Puffs . . . . . . . . . . . . . . . . VI. Puffs: Developmental Physiology . . . . . . . . . . A. Developmental Specificity . . . . . . . . . . . B. Intragland Variation . . . . . . . . . . . . . C. Tissue Specificity of Puffs. . . . . . . . . . . . VII. The Endocrine Control of Puffing Activity . . . . . . . . A. Ecdysone . . . . . . . . . . . . . . . . B. Mechanism of Ecdysone Action . . . . . . . . . . C. Juvenile Hormone . . . . . . . . . . . . . VIII., The Experimental Modification of Puffing Activity . . . . . IX. Modification of Polytene Chromosome Structure and Function as the Result of Infection . . . . . . . . . . . . . X. The Physiology of Nurse Cell Polytene Chromosomes . . . . XI. The Physiology of Epidermal Cell Polytene Chromosomes . . . XII. The Physiological and Functional Significance of Puffing . . . A. Salivary Gland Function . . . . . . . . . . . B. Specific Correlations of Puffs and Salivary Gland Function c. General Considerations on the Physiological Function of Puffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIII. Conclusions and Outlook References . . . . . . . . . . . . . . . . . . . . Notes Added in Proof . . . . . . . . . . . . . . . . . . I. 11. 111. IV.

1

2 2 4 7

7 8 9 10 10 11 12 17 18 19 21 21 24 24 28 31 32 32 40 46 47 51 54 56

59 59 64 68 69 71 93

2

M. ASHBURNER

I. INTRODUCTION “How do the adult characteristics lie latent in the germ cell; and how d o they become patent as development proceeds? This is the final question that looms in the background of every investigation of the cell.” E. B. Wilson: “The Cell in Development and Inheritance”, 1896.

With characteristic insight Edmund Wilson, over half a century ago, posed a question that is still unanswered and still remains central to many investigations into the biology of cells and organisms. It is now widely recognized that an answer to this question must be sought through an understanding of the mechanisms of control of gene activity. A wide variety of “model systems” are currently under close scrutiny with an equally wide variety of experimental techniques. Only in exceptional circumstance has nature provided us with a “system” in which differential gene activity, and its control, can be analysed directly at the level of the genes themselves. The giant chromosomes of Diptera afford one such example and their contribution to the problems of gene control and of the control of development is the subject of the present review. For earlier reviews the reader is referred to: Beermann (1956, 1959, 1962, 1965b); Berendes and Beermann (1969); Clever (1966a, 1968); Kroeger and Lezzi (1966); Laufer (1968) and Swift (1 962).

11. POLYTENE CHROMOSOME STRUCTURE

Balbiani (1881), during a histological study of the nuclei of the salivary glands of larval chironomids, was impressed by the presence, in these nuclei, of “un corps pile, qui a la forme d’un cordon cylindrique diversement enroulC sur lui-mCme A la manicre d’un intestin”. The significance of the structures observed by Balbiani was to remain enigmatic for over 50 years, despite occasional researches (see Tanzer, 1922; Makino, 1938). In 1933 Heitz and Bauer, and Painter independently rediscovered Balbiani’s “cordon nuclCaire” in the Malpighian tubules of Bibio hortulanus” and in the larval salivary glands of Drosophila melanogaster respectively. The intranuclear structures observed by Balbiani and subsequent cytologists were shown, by both investigations, to be homologous with mitotic chromosomes. Considerable controversy developed as to the exact *In all cases the taxonomic nomenclature of the original publications has been retained.

PUFFS

3

structure of these “giant chromosomes” and their relationship to regular mitotic chromosomes but the “polytene hypothesis”, originally proposed by Koltzoff ( 1 934) and modified by Bauer (1935), is now fully accepted (Bauer and Beermann, 1952; Beermann and Pelling, 1965). The reviews of Alfert (1954) and Beermann (1962) may be consulted for a complete consideration of the evidence as the following is only a brief account. Polytene chromosomes develop from the chromosomes of diploid nuclei by successive duplication of each chromosome element. Usually the homologous chromosomes pair early during this process and remain paired throughout polytenization. The chromosome duplication involved differs from that typical of a normal mitotic cycle in many respects: (i) the homologues remain paired, (ii) the chromosomes fail to participate in the normal mitotic cycle of coiling and uncoiling, (iii) daughter chromosomes do not segregate but remain intimately paired with each other at the end of each replication cycle, and (iv) the nuclear membrane and nucleoli remain intact throughout the replication cycles. The final product of the 9 or 10 cycles of replication characteristic of the salivary gland nuclei of Drosophila, for example, is a nucleus containing the haploid number of giant chromosome elements each composed of from 1000-2000 individual “chromosomes” in intimate apposition-as if a multistranded cable. Along the linear axis of each constituent chromosome (often known, in this context as “chromonemata”) variation in the extent of coiling (tertiary structure) of the DNA and its associated histones leads to variation in the concentration of these substances. Regions of high concentrations are known as chromomeres. For each chromosome the pattern of chromomeres is highly specific so that in the polytene chromosome homologous chromomeres align alongside each other exactly and usually appear to fuse as a band across the polytene element. The banding pattern of polytene chromosomes is such a stable and specific feature of their organization that the individual bands can be recognized, mapped and assigned reference numbers. In the interchromomeric, or interband, regions the DNA and histone concentration is lower than in the bands; the DNA content of bands and interbands being in the ratio of 20 : 1 in Drosophila rnelanogaster (Paul and Mateyko, 1967). In contrast to the highly-coiled state of the DNA-protein fibres in the bands they are aligned more or less parallel to the long axis of the chromosomes in the interband regions (MacInnes and Uretz, 1966; but see Wetzel

4

M. ASHBURNER

et al., 1969), although small local variations in coiling do occur here also (Sorsa and Sorsa, 1967). Claims (e.g. Steffenson, 1963) that DNA is absent from the interbands are now known to be incorrect (Swift, 1962; Lezzi, 1965; Wolstenholme, 1965, 1966). The pattern of bands and interbands of each polytene chromosome is not only specific for the species but is characteristic of that particular chromosome no matter in which tissue of the animal it is found. The “Istanbul hypothesis” of tissue specific variability in banding patterns (e.g. Kosswig and Sengun, 1947; Sengun, 1954) is incorrect (Berger, 1940; Beermann, 1950; Dreyfus et al., 195 1; Slizinski, 1950) at least insofar as the larval tissues of Diptera are concerned. The level of polyteny attained in any particular tissue is usually characteristic-for example by the end of the third larval instar the fat body of D. melanogaster possesses polytene chromosomes with a 64C DNA value whilst the salivary gland chromosomes reach the 1024C level (Alfert, 1954; Rodman, 1967b). As a general rule the number of duplications is identical for each chromosome of a particular nucleus; exceptions have been described by Melland (1942) for the chironomid Anatopynia varius and by White (1948) for the cecidomyiid Camptomyia sp. In both of these species one chromosome (the smallest of the complement in each case) appears to be of a higher level of polyteny than the other chromosomes.

111. POLYTENY AND ENDOPOLYF’LOIDY DURING

INSECT DEVELOPMENT

Growth resulting from an increase in the size of relatively few cells, rather than an increase in cell number through cell division, is a phenomenon well known in the Insecta (Trager, 1937). Such growth is, of course, accompanied by a parallel increase in nuclear size and DNA content. In most groups of the Insecta, other than the Diptera, the mechanism of nuclear. growth involves the reduplication of each chromosome within an intact nuclear membrane, the individuality of each chromosome being retained. Such a process is known as endomitosis and has been reviewed by Geitler (1953) and more recently studied by Nur ( 1968). Endopolyploid nuclei attain DNA contents often of the same order of magnitude as polytene nuclei. Endopolyploidy and polyteny are not mutually exclusive

PUFFS

5

mechanisms of nuclear growth as the following discussion will illustrate. Furthermore under certain circumstances transitions from endopolyploidy to polyteny, and vice versa, occur during development. The functional significance of polyteny vs. endopolyploidy is obscure; presumably the available surface area of chromatin in an endopolyploid nucleus of a given ploidy is greater than that in a polytene nucleus of similar DNA content. A further factor may be that extensive invagination and lobulation of the nuclear membrane of a polytene nucleus may lead to such compartmentalization of the nucleus that the transfer of substances between different chromosomes and different chromosome regions would be impeded. In an endopolyploid nucleus with each chromosome represented individually many times no such barrier would exist. While polytene nuclei are usually spherical or approximately so in form, endopolyploid nuclei often assume bizarre shapes (for example the spinning gland nuclei of Lepidoptera and Trichoptera), a consequence of which is an extensive increase in their surface area. The interrelationship between endopolyploidy and polyteny is perhaps best studied in the salivary gland nuclei of larval Cecidomyiidae. In the predacious genus Lestodiplosis the larval salivary gland possesses, in its anterior region, one or two “super giant” cells containing from 16 to 32 representatives of each of the four chromosomes-each of these elements themselves being polytene, i.e. these nuclei are endopolyploid polytene nuclei. In Lestodiplosis a single super giant cell is present containing numerous chromosomes, while the nuclei of the main region of the gland have four normal polytene elements (White, 1946). Henderson (1967a) has described another example of endopolyploid polytene nuclei in the super giant cells of the salivary gland of Lestodiplosis pisi, whilst Ashburner and Henderson ( 1969) have studied further examples. From a study of the development of these nuclei Henderson ( 1967a) concludes that the endopolyploidy results from the progressive splitting of the four chromosome elements of a standard polytene nucleus with the haploid number of chromosomes. In addition to this he suggested that the fusion of two cells into a single one may have contributed to the growth and development of the single super giant cell found by White. Coincident endopolyploidy and polyteny in larvae of Dmyneuru urticue (Cecidomyiidae) parasitized by platygasterids has been studied by Matuszewski (1964). The nuclei of the cells which surround the developing hymenopteran embryos possess up to 240

6

M. ASHBURNER

polytene chromosome elements, suggesting that they are either 60n or 120n nuclei ( n = 4). In this particular case it is likely that polytenization has followed endopolyploidy by the polytenic replication of a large number of individual chromosome elements. The disintegration, by fibrillation, of polytene chromosomes resulting in typical reticular endopolyploid nuclei appears to be a sequence normal to salivary-gland nuclear development in many Cecidomyiid larvae. This occurrence was first suggested by White (1948) on the basis of the nuclear morphology of the “transitional zone” in salivary glands of Dasyneura affinis. It has been subsequently confirmed and studied by Matuszewski (1 965) in D. urticae, by Henderson (1967b) in D. crataegi and by Ashburner and Henderson (1969) in a variety of Cecidomyiinae. While the range of chromosome behaviour is extraordinary, the developmental sequence is basically as follows. In first instar larvae the linear posterior region of the gland (approximately 6 4 cells) is very small in relation to the wide anterior region of 8-32 cells. All gland nuclei possess polytene chromosomes, usually four elements. As growth proceeds the posterior region becomes much larger than the anterior region. Indeed in many species the anterior region collapses, and its nuclei degenerate, near pupation. During this growth phase the two smallest polytene chromosomes of the posterior region may fibrillate into bundles of smaller and smaller multiplicity and finally into their constituent haploid elements. Whether or not this process is accompanied by chromosome replication is unknown. When the fibrillation of these small elements is well advanced the larger chromosomes (which usually bear nucleoli) may also. go through a similar process and the nuclei become completely endopolyploid. In some species (e.g. Putoniella marsupialis) a similar sequence of events may occur, but later during development, in the anterior gland nuclei. The speed or completeness of these cytological changes, relative to gross morphological development, varies greatly between different species. They may also be modified by, for example, secondary polytenization of endopolyploid posterior gland nuclei (in Perrisia ulmariae, Ashburner and Henderson, 1969) or by transverse chromosome fragmentation, described in Mikiola fagi and Rhabdophaga saliciperda by Kraczkiewicz and Matuszewski ( 1958). The organization of chromosomes in polyploid nuclei can also be influenced by environmental circumstances. The salivary-gland nuclei of larval Aphiochaeta xanthina (Phoridae) are typically polytene only when the larvae are grown under optimal nutritional conditions.

PUFFS

7

When either protein or fat are in short supply these nuclei are endopolyploid (Barigozzi and Semenza, 1952). Transitions between endopolyploidy and poly teny in the nurse cell nuclei of Diptera were first described by Bauer (1938) and have been studied recently by Bier. This work will be discussed below. IV. THE OCCURRENCE OF POLYTENE CHROMOSOMES A. LARVAL TISSUES OF DIPTERA

Polytene chromosomes have been reported in many families of Diptera in both larval tissues destined for histolysis at metamorphosis (e.g. salivary gland, midgut, rectum, fat body and muscle cells) and in larval tissues that persist intact into the imago (e.g. Malpighian tubules, brain) (Makino, 1938). These polytene chromosomes have been studied in the following families: NEMATOCERA; Psychodidae (Pavan et ul., 1968), Mycetophilidae (including Sciarinae) (e.g. Metz, 1 9 3 9 , Bibionidae (Heitz and Bauer, 1933), Cecidomiidae (e.g. White, 1948), Chironomidae, Scatopsidae (Mainx, 1949), Ptychopteridae (Mainx, 1949), Limnobiidae (Mainx, 1949), Culicidae (e.g. Sutton, 1942; Kitzmiller, 1968; Kiknadze, 1967), and Simuliidae (e.g. Painter and Griffen, 1937; Rothfels and Dunbar, 1953); BRACHYCERA; Leptidae (Mainx, 1949), Scenopinidae (Mainx, 1949); ASCHIZA; Phoridae (Mainx, 1949; Barigozzi and Semenza, 1953); ACALYPTRATAE; Coelopidae (Philip, 1966), Trypetidae (Frizzi and Springhetti, 1953; Krimbas, 1963), Drosophilidae, and Agromyzidae (Mainx, 195 1 ; Block, 1969). The various larval tissues of Drosophilu virilis were surveyed by Makino ( 1938) who found polytene chromosomes of varying sizes in: salivary gland, oesophagus, proventriculus, gastric caecae, midgut, Malpighian tubules, fat body, trachael wall cells, muscle cells and posterior ganglion. More recent investigations allow the addition to this list of: nerve cells and anal gill nuclei in Chironomids (Melland, 1942), rectal gland nuclei and the nuclei of the cell complex associated with the humeral imaginal disc in D. rnelunogaster (Lamprecht and Remensberger, 1966). Not all larval tissues of Diptera have polytene chromosomes; for example the binucleate “Guirlandzellen” of Weismann have endopolyploid nuclei in Drosophilu (Ashburner, unpublished), while the larval salivary gland nuclei of some families (e.g. Tipulidae) are endopolyploid.

8

M. ASHBURNER

B. PUPAL AND ADULT TISSUES OF DIPTERA

In their study of bristle differentiation of Drosophila melanogaster, Lees and Waddington (1942) noted that the nuclei of both trichogen and tormogen cells “become very large, and contain deeply staining threads which show much the same structure as the salivary gland nuclei”. Although these authors appear to be in two minds as to the exact nature of these nuclei-whether polytene or endopolyploid-there is no doubt from the recent work of Ribbert and of Whitten that, in the Calliphoridae, at least, large polytene chromosomes develop in the nuclei of many epidermal cells during the period of secretion of the adult cuticle and its appendages. These studies, with Calliphora erythrocephala (Ribbert, 1967), C. stygia (Thomson, 1969),Surcophaga bullata (Whitten, l963,1964a, 1969d), and Lucilia cuprina (Childress, 1969) will be fully considered below (Section XI). The occurrence of polytene chromosomes in the ovarian nurse cells of Diptera, and their discovery by Bauer (1938) have already been mentioned. Bauer (1938) and later Painter and Reindorp ( 1939) have described polytene chromosomes of low multiplicity during the early growth period of the nurse cell nuclei of Lucilia and Drosophila melanogaster respectively. Later during development these poly tene elements fall apart into their constituent chromosomes, a process recalling that described for the salivary-gland nuclei of Cecidomyiidae, and the nuclei subsequently grow by endomitotic duplication. This process has been studied in detail by Bier (1957, 1958, 1959, 1960) in the nurse cell nuclei of Calliphora erythrocephala. Up to the 16C-32C stage these nuclei possess well-developed polytene chromosomes. Normally these chromosomes fibrillate and subsequent growth is, as described by Bauer and by Painter and Reindorp, by endoduplication. Under certain conditions however nuclei of 512C to 2048C develop typical polytene chromosomes from endopolyploid nuclei. These polytene chromosomes are known as secondary polytene chromosomes. Functional aspects of nurse cell chromosome behaviour will be discussed in Section X. In Drosophila melanogaster the nurse cell nuclei of certain female sterile mutants develop polytene chromosomes during advanced growth stages. They have been described by King et al. (1961), and Klug et al. (1 968), and further studied by Schultz (1965). Various experimental treatments also result in the formation of polytene nurse cell chromosomes in D. melanogaster, e.g. X-rays (King, 1957

PUFFS

9

quoted in Jacob and Sirlin, 1959) or the feeding of 5-aminouracil. Nurse cell polytene chromosomes have been described from a number of families of Diptera by Stalker (1954), e.g.: Cecidomyiidae, Empididae, Dolichopodae, Agromyzidae, and Tachinidae. In Rhynchosciara angelae (Sciarinae) Basile (quoted by Pavan, 1965b) claims that the nurse cell nuclei are of the endopolyploid polytene type. Other tissues of Diptera with polytene nuclei include the adult Malpighian tubules, the vesiculae seminales (in Rhynchosciara angelae, Pavan and Breuer, 1955), the pupal heart (in Sarcophaga bullata, Whitten, 1963), and adult salivary glands in the Simuliid Boophthora erythrocephala (Welsch e t al., 1968). C. POLYTENE CHROMOSOMES IN OTHER ORGANISMS

With the exception of the well-known instances in higher plants (e.g. Nagl, 1969a, b), it was, until recently, considered that polytene chromosomes were not to be found outside the Diptera. We now know that they occur in the macronuclei of some Ciliates (Golikova, 1964; Alanso and P6rez-Silva, 1965; Ammermann, 1965; Radzikowski, 1967) and also in another insect group, the Collembola. The polytene chromosomes of the salivary gland cells of adult Collembolla were first described by Prabhoo ( 1961) in Womersleya sp., but have been more closely studied by Cassagnau in various species of Neonura and in Bilobella massoudi (Cassagnau, 1966, 1968). In the last-named species large polytene chromosomes, with characteristic banding patterns, occur in the distal region of the posterior salivary gland. These polytene chromosomes possess some unique morphological features, most strikingly the “echarpe hCtCrochromatique” which are regions of the chromosomes swollen to 3-5 times the normal chromosome diameter whilst retaining a typical banding pattern. Cassagnau ( 1968) is of the opinion that these specializations of the chromosomes represent the centromeric regions. In contrast to the typical Dipteran condition of complete pairing between homologues (although this is not always so in the Simuliids) the polytene chromosomes of the Collembolla are unpaired and the nuclei contain the diploid number (2n = 14 in Bilobella massoudi) of polytene elements. In other families of Collembolla (e.g. Pseudachorutini) the salivary gland nuclei are endopolyploid while an intermediate condition in the evolution of Collembolla polytene chromosomes is found in the Protanurini (Cassagnau, 1969).

10

M. ASHBURNER

V. PUFFS: INTRODUCTION AND BIOCHEMISTRY A. THE UNIT HYPOTHESIS OF THE CHROMOMERE

The pattern of bands (chromomeres) and interbands (interchromomeres) along polytene chromosomes is a constant enough character to allow these patterns to be mapped (Painter, 1933) and for each band to be assigned a reference number (Bridges, 1935). A very major advance in the study of these chromosomes was the demonstration of the colinearity of these cytogenetic maps and of the standard genetic map of mutant loci in D. melanogaster (Bridges, 1937). By a variety of genetic techniques it has proven possible to map mutant loci to small regions or even to single polytene chromosome bands (summarized in Lindsley and Grell, 1968). For technical reasons this analysis has been most extensive for the X chromosome of D. melanogaster (Sutton, 1943), and in the region of the white and Notch loci the phenotypic effects of the deletion or duplication of any one of a number of bands has been studied (Beermann, 1967a, b reviews the literature; see also Rayle and Green, 1969). The “one-to-one correspondence” (Beermann, 1967b) of bands and classical genes is a popular hypothesis that has considerable heuristic value. That it can only be an approximation to the true state of affairs is illustrated by the following considerations. Estimates of the DNA content of individual bands are technically extremely tedious and involve a number of assumptions. Despite this the results of two independent investigations have been published and agree remarkably well. Rudkin ( 1965) used microspectrophotometry to estimate that an average band of the D. melanogaster polytene chromosome has a DNA content of 0.3 x g/haploid unit. The smallest measurable bands contained 0.05 x g DNA/haploid unit. By chemical means Daneholt and Edstrom (1 967) estimated that the average Chironomus tentuns salivary gland chromosome band had a DNA content of 1 x g/haploid unit. Both groups of data are approximations in part due to the difficulties of experimental technique and in part to assumptions about the actual number of bands and about the DNA content of the interband and heterochromatic regions. However they are unlikely to be in error by more than one order of magnitude. g as the average DNA content per haploid unit of a Taking band this is equivalent to a linear length of DNA of 40 1-1 corresponding to lo5 base pairs (Daneholt and Edstrom, 1967; Panitz, 1968), sufficient to code for 30,000 amino acids.

PUFFS

11

Wolstenholme e t al. ( 1 968) and Berendes (cited in Berendes and Beermann, 1969) have measured the lengths of DNA strands prepared from salivary gland nuclei by the Kleinschmid t technique. In Chironomus tentans the mean strand length was 46.7 p (Wolstenholme e t al., 1968) and in Drosophila melanogaster the mean length was 37 p. In both studies the range of sizes of DNA molecules varied over an order of magnitude or more. These data are in apparent conflict with the cytogenetic evidence which leads to the conclusion “one band; one gene”. The paradox has been succinctly stated by Daneholt and Edstrom (1967): “this g) is an unlikely high number for a cistron and suggests that bands either represent operational units of great complexity or that a large part of the DNA is non-functional in the informational sense”. Evidence for the unitary function of bands extends beyond the cytogenetic data. Bands also behave as units with respect to both RNA and DNA synthesis (Pelling, 1965). It is unlikely that bands are units of “great functional complexity” yet despite considerable speculation (Beermann, 1965a; Callan, 1967), and some elegant observations (Keyl, 1965) the organization of the DNA within the bands, and the relation between its organization and its function, remains obscure (see discussion in Berendes, 1969). This fact should be borne in mind during the following discussion. B. PUFFS

Polytene chromosome puffs may be defined as structural modifications of the DNA and its associated substances of single bands which involve an uncoiling of the chromosome fibres allowing the synthesis of specific RNA species and the accumulation of RNA and other substances at that site. This accumulation often results in the enlargement of the specific chromosome region. This definition excludes the DNA puffs of the Sciarinae. The DNA puffs will be discussed more completely below but due t o their historical importance the following facts are of particular significance: DNA puffs are found only in this one family of Diptera and, like normal puffs, arise from individual polytene chromosome bands. The puffed chromosome region enlarges largely as the result of accumulation of DNA at the puff site not, as in the case of normal puffs, as the result of the accumulation of RNA and protein at the puff site. It should be pointed out that in addition t o DNA puffs the polytene chromosomes of the Sciarinae also form RNA puffs (often known as “bulbs” in this group).

12

M. ASHBURNER

The first study of puffs was that of Poulson and Metz ( 1 938) with Sciara ocellaris. They observed that the DNA puffs were modifications of particular bands or groups of bands and that they apparently underwent cyclical changes in size so that the appearance of the post-puff chromosome region differed only slightly from the pre-puff chromosome region. Poulson and Metz made the observation that within any single salivary gland all nuclei possessed the same spectrum of puffs yet there was considerable variability in the sites puffed between different individuals. They concluded that this variability was probably due to environmental differences, e.g. “extreme conditions, such as food, humidity or temperature”. Largely through the work of Beermann (1952a, b) Mechelke (1952) and of Pavan and Breuer (1952) it is now known that the reason for the variation in puffing observed by Poulson and Metz is the fact that there is very marked developmental specificity in the time of appearance and regression of puffs. Indeed, anticipating a later discussion, the DNA puffs of Sciara d o not form until very late in the last (fourth) larval instar (Breuer and Pavan, 1955; Gabrusewycz-Garcia, 1964). As the result of his study of structural modifications of the polytene salivary gland chromosomes of Chironomus tentans Beermann ( 1952b) proposed the hypothesis that “puffing indicates changes, most probably increases, in the activity of gene loci” (Beermann, 1956). We must now consider this hypothesis in the light of the results of recent investigations into the biochemistry and developmental physiology of puffs.

C. RNA SYNTHESIS AND PUFFS

Our current understanding of gene action demands, as a consequence of Beermann’s hypothesis, the synthesis of RNA at puff sites. Beermann himself showed that puffs, unlike unpuffed chromosome regions, stained metachromatically with toluidine blue (Beermann, 1952b) and Breuer and Pavan (1955) found that the DNA puffs of Rhynchosciara angelae exhibited an RNA reaction with methyl-green pyronin. Following the demonstration of a high rate of nuclear RNA synthesis in Drosophila salivary gland nuclei by Taylor (1953), Gross (1957) showed in the chironomid Metriocnemus hygropetricus that puffs selectively incorporated RNAse sensitive P3’. With the advent of more specific tritium labelled precursors Pelling (1959, 1964) analysed in great detail the

PUFFS

13

chromosomal incorporation of H3-uridine by salivary gland chromosomes of Chironomus tentans. Pelling confirmed the earlier finding of Beermann that the puffs and Balbiani Rings (a name given to the few extraordinarily large puffs of Chironomids) of C. tentans stained red with toluidine blue while unpuffed regions stained blue. After treatment of the chromosomes with RNAse only the blue staining, due to DNA, remained. Like puffs, the RNA containing sites, as determined by toluidine blue metachromasia, were identical in each nucleus of an individual salivary gland, yet differed between the glands of different animals. In vivo incorporation of tritiated uridine into chromosomes was completely restricted to those sites that showed the RNA reaction with toluidine blue. Only a proportion of these sites were visibly puffed. This fact is important since it emphasizes that there is a continual gradation between condensed, inactive, bands and fully enlarged puffs. Cytological studies of puffing very often ignore many small puffs, or unpuffed active sites, due to the difficulty of unambiguous classification. It is usually assumed that the sample of sites observed cytologically as puffs is a random selection, at least with respect to physiological function, of all the active sites. The total number of active sites can only be determined by very detailed cytochemical and autoradiographic procedures such as those used by Pelling (1964). The use of these techniques allowed Pelling to estimate that, of the 1900 or so bands of the C. tentans salivary gland chromosomes, only 272 (1 5%) were ever active in the synthesis or accumulation of RNA during the last (fourth) larval instar. In general Pelling observed a good correlation between puff size and rate of incorporation of RNAse sensitive label after exposure to H3-uridine. In the salivary glands of C. tentans the smallest chromosome, chromosome IV, is notable in possessing three Balbiani rings (BR); incorporation of label into these three puffs accounted for approximately 25% of the total nuclear incorporation (which included label into two nucleoli). The largest of the BR was the most heavily labelled. In contrast the number of grains due to all of the puffs on the long chromosome I only accounted for some 20% of the total nuclear grain count. Puff formation was accompanied by an increase in synthesis, and puff regression by a decrease and eventual cessation of locus specific incorporation. Although the general relationship between puff size and rate of RNA synthesis is readily interpretable, Pelling did observe that between different individuals there were sometimes marked differences in the number

14

M. ASHBURNER

of grains over given BR despite an equivalence in the size of these puffs in the different animals. Similar observations were made by Gaudecker ( 1967). Such differences may be due to variations in, for example, pool size or in the accessibility of label, as suggested by Pelling (1964), but it is apparent that factors other than the rate of locus specific RNA synthesis may well be involved in the control of puff size (see Section D). The observation that puffs are sites of rapid and specific incorporation of RNA precursors has been amply confirmed, although not as thoroughly studied, in other Dipteran polytene chromosomes (Rudkin and Woods, 1959; Sirlin, 1960; Ritossa, 1962a; Fujita and Takemoto, 1963; Clever, 1964a; Berendes, 1967a, 1968a). Rudkin ( 1962) showed, in Drosophila melanogaster, a parallel increase in puff size and RNA content as determined by cy tophotometric techniques. He also showed the locus specific incorporation of H3-cytidine into puff RNA. The antibiotic actinomycin D is commonly used to suppress DNA dependent RNA synthesis. At concentrations between 0.1 and 1 pg/ml the incorporation of H3-uridine into puffs and the formation of new puffs is inhibited by actinomycin D (Ritossa and Pulitzer, 1963; Clever, 1967). However, not all puffs are affected equally by the drug (Beermann, 1965b; Clever, 1967). In Chironomus concentrations of actinomycin D of the order mentioned cause the actual regression of the BR, in addition to stopping RNA synthesis (Beermann, 1965b; Lauferetal., 1964; Kiknadze, 1965; Clever, 1967) yet the smaller puffs remain expanded (Clever, 1964a; but see Clever, 1967). After removal of actinomycin RNA synthesis slowly recovers in parallel with the reformation of the BR (Clever, 1964b; Clever and Romball, 1966). The process of normal puff regression is also sensitive to actinomycin D (Clever, 1967). High (e.g. 20 pglml) concentrations of actinomycin D do not result in BR regression in Chironomus despite the inhibition of RNA synthesis (Beermann, 1965b; Kiknadze, 1965). This is probably due to the binding of the actinomycin to the guanine residues of the DNA to such an extent that recoiling of the DNA is no longer possible; the DNA is thus “frozen” in its extended position. Actinomycin D has one further effect on polytene chromosomes. Both Kiknadze (1965) and Serfling (personal communication) have observed “puff” induction in Chironomus salivary gland and Malpighian tubule chromosomes after the treatment of larvae with actinomycin D. This occurs especially at the heterochromatic

PUFFS

15

centromere regions of the chromosomes and involves the accumulation, but apparently not the synthesis, of RNA and protein. It is presumably due in some way to disturbances in the normal mechanisms of transport of RNA and protein from the chromosomes (see p. 18). I t cannot necessarily be assumed that the RNA accumulating in these “centromeric puffs” reflects the activity of these chromosome regions themselves. In C. thurnmi identical puffs are induced by several acridines (Serfling, personal communication). The rate of RNA synthesis a t puff sites has not been accurately determined. Clearly it is rapid in view of the results of, for example, Ritossa’s experiments showing the appearance of label at puffs after only 30 s of in vitro availability (Ritossa, 1962a; also Gay, 1964; Arnold, 1965). Recently two groups of workers have published information pertinent to the problem of RNA stability in salivary glands of Chironornus. It has been known for some time that intermoult Chironornus larvae are relatively insensitive t o actinomycin D, metamorphic stages having a much greater sensitivity (Laufer et al., 1964). Clever et al. ( 1 969b) and Doyle and Laufer (1969b) have both now shown that protein synthesis continues in salivary glands of intermoult C. tentans larvae after considerable periods (up to 48 hr in the experiments of Clever et al.) of actinomycin D treatment. Furthermore Clever ( 1969) demonstrated that this protein synthesis, which continued in the absence of detectable RNA synthesis, included the synthesis of a specific secretory product-the “SZ” granules of C. palliduvitatus (see p. 64). Young fourth instar larvae and prepupae are more sensitive t o actinomycin D and drug treatment results in a reduction of protein synthesis and the inhibition of developmentally specific increases in the rate of protein synthesis (Clever et a l , 1969b). It can be concluded from these experiments that in intermoult larvae salivary gland secretion synthesis proceeds on relatively stable RNA templates yet during periods of rapid change in gene (puffing) activity there is some indication for the existence of more unstable species of template. Technical difficulties have so far prevented detailed analysis of the species and properties of RNA synthesized at individual puff loci. Such results that are available are from analyses of C. tentans RNA, and especially of the RNA of isolated fourth chromosomes and its Balbiani rings. The standard method of chromosome and BR isolation is by microdissection. Edstrom and Beermann (1962) subjected the RNA’s of various nuclear and cytoplasmic fractions t o

16

M. ASHBURNER

base frequency analysis by the microelectrophoretic techniques developed by Edstrom (1964b). They found that the base composition of the RNA’s from different chromosomes differed significantly from each other and from nucleolar and cytoplasmic RNA’s. Moreover three regions of the fourth chromosome, each carrying a single BR, had RNA’s of characteristic base composition, which in some instances differed from each other. The G C ratios of chromosomal RNA’s were unlike that of chromosomal DNA, and this fact, coupled with the lack of base symmetry of the RNA’s, suggested that they might be transcribed from one or other DNA chains, but not from both, along any particular nucleotide sequence (Edstrom, 1964a). To conclude that the RNA species studied by Edstrom and Beermann are messenger RNA’s would be premature on the basis of available evidence. Messenger RNA’s can only be identified operationally if they can be shown to be translatable into specific sequences of amino acids. Edstrom has recently extended his analysis of RNA’s from Chironomus salivary glands by a study of its physical characteristics through sedimentation behaviour in sucrose gradients (Edstrom and Daneholt, 1967). The RNA of the largest BR of chromosome IV of C. tentuns (BR2) is polydisperse in a sucrose gradient with sedimentation constants from 10s t o 70s. These results d o not warrant the conclusion that the RNA synthesized at the site of BR2 is similarly heterogeneous. Even assuming that the RNA profiles observed on the gradients represent species of RNA made at this site during the in vitro incubation period it is unlikely that a single BR can be isolated by microsurgery without the inclusion of adjacent active sites. Although Edstrom and Daneholt do not record the exact chromosome regions included in their “BR2” fraction this can be estimated from the data of Edstrom and Beermann ( 1 962). Figure 2(b) of this paper would indicate that at least six of Pelling’s (1964) active sites, in addition t o BR2, were included and may, therefore, have contributed RNA species to the material analysed. Analysis of the experiment of Edstrom and Daneholt is complicated further by our ignorance of (a) the effects of in vitro incubation on the types of RNA synthesized by the glands (see Greenberg, 1967) and (b) of the effects of fixatives (e.g. formalin used during the procedure of Edstrom and Daneholt) on the macromolecular structure of RNA. Although the identity of the RNA made at puff sites is far from being rigorously established, as is clear from the above discussion, we

PUFFS

17

can at least exclude the possibility that it is ribosomal RNA. In Drosophila melanogaster experiments of Ritossa and Spiegelman (1965) clearly indicate that the ribosomal RNA is synthesized under the direction of the DNA of the nucleolar organizer region, located at the base of the X and Y chromosomes. It seems rather unlikely (at least to this reviewer), that the remaining characterized class of major RNA’s, the transfer RNA’s, are the exclusive product of puffed sites. Several ancillary questions concerning the nature and function of puff RNA remain unanswered, for example: does the DNA of one puff code for one or more than one species of RNA? does all of the RNA made at puff sites reach the cytoplasm and function as template for protein synthesis? (see Harris, 1968); does some puff RNA have an intranuclear regulatory function? and so on. D. RNA TRANSPORT FROM THE NUCLEUS AND NUCLEOCYTOPLASMIC INTERACTIONS

Assuming that at least a fraction of puff RNA serves as messenger RNA it is necessary to postulate some mechanism for the transfer of this RNA from its site of synthesis (the chromosomes) to its site of function (the endoplasmic reticulum of the cytoplasm). Several studies of the ultrastructure of puffs have noted the presence, in these regions, of electron dense granules which have often been implicated in this transport process (Beermann and Bahr, 1954; Schurin, 1959; Swift, 1959, 1962, 1963; Berendes and de Bruyn, 1963; Jacob and Sirlin, 1963; Stevens, 1964; Kalnins et al., 1964a, b; Abramyan and Reingol’d, 1965; Stevens and Swift, 1966; Berry and Dietz, 1968; Smith, 1969; Sorsa, 1969). In Chironomus the diameter of the granules associated with the BR is of the order of 300-500 A. In Drosophila virilis Swift ( 1962, 1963) claims to have evidence that the size of the puff granules is a characteristic of particular puff sites. The attachment of the granules to thin ( 100 A ) chromatin fibres has been reported by a number of workers (Beermann and Bahr, 1954; Jacob and Sirlin, 1963; Kalnins et al., 1964a; Stevens and Swift, 1966). Enzyme digestion studies have led to the conclusion that chemically the granules are ribonucleoproteins (Swift, 1963). After actinomycin D treatment of salivary glands (1 pg/ml) alterations in granule morphology were found by Stevens ( 1964). Although predominant in puffed regions granules of very similar morphology are also present in the nuclear sap, in particular in

18

M . ASHBURNER

association with the nuclear membrane (Beermann and Bahr, 1954; Jacob and Sirlin, 1963; Abramyan and Reingol’d, 1965; Stevens and Swift, 1966; Beermann, 1966; Berry and Dietz, 1968). Stevens and Swift ( 1966), studying the granules of Chironomus thummi salivary gland nuclei describe a phenomenon interpreted as granule transfer from the nucleus through the membrane into the cytoplasm. After actinomycin D treatment the association of the BR granules with the nuclear membrane could no longer be detected in Chironomus by Stevens et al. (1965). Since naked informational RNA would presumably be vulnerable to hydrolytic breakdown within the cell it has been suggested that a complex of this RNA with the granules may be important for its protection and transport (see also the discussion in Spirin, 1966). Neither the site of manufacture of puff granules, nor their relation to the ribosomes (if any) is known. Lezzi (1967a) has discussed the possible role of ribosomes in RNA transport from puff sites. 0ther ultrastructural features of the Dipteran salivary gland nucleus have been described and considered to play a role in nucleocytoplasmic transport. Gay ( 1956, 1959) describes outpocketings of the salivary gland nuclear membrane of D. melanogaster and other species and suggests that the frequency of these “nuclear blebs” is correlated with the secretory activity of the gland, although no quantitative data have been published. In support of the idea that these blebs are concerned with nucleocytoplasmic transport Gay finds that they are often associated with chromosomal material. Similar blebs have been described by Berendes and de Bruyn (1963) in D. hydei, by Swift (1959) in D. virilis and by Kimoto (1958) in Chironomus dorsalis. Both Swift and Kimoto report an association of chromosomal material with the blebs (see also Section XI, below). E. THE “PUFF PROTEINS”

In addition to the accumulation of RNA-largely the result of “on site” synthesis-puffs are also characterized by the accumulation of a family of proteins which may be stained with fast green or light green at acid pH (Rudkin, 1955; Swift, 1959, 1962; Schurin, 1959; Clever, 1961, 1962a, 1964a). Newly activated puffs appear to contain especially large quantities of these “puff proteins” (Berendes, 1968a; Clever, 1965b) but their nature is far from being fully understood. The “puff proteins” differ from the chromosomal histones not only in the pH optimum for their stainability with fast

PUFFS

19

green, but also by the fact that, unlike histones, they are insoluble in 0.2 M HCI (Swift, 1964, see Black and Ansley, 1964). “Puff proteins” are apparently low in the proportion of cyclic amino acids (Rudkin, 1964). During puff induction the accumulation of puff proteins precedes any visible change in chromosome structure and detectable RNA synthesis (Berendes, 1968c, 1969; Pettit and Rasch, 1966). Lezzi (1967b) studied the effects of various enzymes on the puff proteins and as the result of these, and other cytochemical experiments, concluded that, as a class, the puff proteins closely resembled the proteins of ribosomes. It is probable that much of the acid protein detected in puffs is, in fact, associated with the ribonucleoprotein granules described above (Swift, 1964). Some part of the puff protein fraction, however, presumably plays other roles, e.g. RNA polymerase and regulatory functions. Ribonucleoprotein granules are not detectable until 20 min after puff induction although “puff proteins” are found within 3 min (Berendes, 1969). Unlike puff RNA the puff proteins are not synthesized at the puff site (Pelling, 1959; Sirlin, 1960; Ritossa and Pulitzer, 1963; Ritossa et al., 1965; Clever, 1964a; Berendes, 1967b; Cave, 1968). Puff induction does not involve protein synthesis (Clever and Beermann, 1963; Clever and Romball, 1964; Ritossa and Pulitzer, 1963; but see Pettit and Rasch, 1966), and will occur during the experimental inhibition of protein synthesis by antibiotics (loc. cit. and Ashburner, unpublished). Localized protein synthesis at puff sites has been described in isolated salivary gland chromosomes of Chironomus tentans by Lezzi but is clearly a result of very abnormal physiological conditions perhaps leading to messenger translation at the puff site (Lezzi, 1967a). Recently Berendes (1968a) has published the results of an interesting experiment describing apparent puff induction whilst RNA synthesis was inhibited by actinomycin D. In fact it appears that most of the induction observed by Berendes in these experiments involved locus specific accumulation of puff proteins. F. HISTONES AND PUFFS

The histones of polytene chromosomes have been reviewed by Swift (1964) and in other studies from the same laboratory. The distribution of histones along polytene chromosomes parallels the distribution of DNA (Horn and Ward, 1957; Black and Ansley, 1964; Swift, 1964); that is the histone concentration is high in the bands

20

M. ASHBURNER

and low in the interbands. During the transformation of a band into a puff the histone concentration, like the DNA concentration, decreases as the chromatin involved expands to cover a larger area. The actual amount of histones remains unchanged during this process however. Gorovsky and Woodward (1966, 1967) measured the histone : DNA ratios of various bands and puffs in Drosophila virilis. This ratio was almost identical (0.36) in both condensed bands and expanded puffs. The theory that histones are important factors in the control of gene activity has been widely considered since its proposal by Stedman and Stedman (195 1) (e.g. Bonner, 1965 and de Reuck and Knight, 1966). The failure to find any substantial qualitative or quantitative variation in the histones of differentiated nuclei [for example the electrophoretic pattern of histones from Drosophila salivary gland nuclei and imaginal disc nuclei cannot be distinguished (Cohen and Gotchel, 1969)] has led recently to the proposal that perhaps more subtle changes of the histones are important for their supposed regulatory role. In particular Allfrey and his colleagues have suggested that the acetylation of histones is of great functional significance and claim a positive correlation between the extent of histone acetylation and the capacity for RNA synthesis of the DNA : histone complex (Allfrey et al., 1966). Of direct relevance to this interesting proposal are the experiments of Clever (1967) and Ellgaard (1967) who have both failed to find any selective chromosomal or puff incorporation of H3 -acetate during puff formation. Allfrey et al. (1 968) contested the significance of these results on the grounds that the fixation methods used (acetic acid/alcohol) would have dissolved the histones away. Despite this objection the cytochemical detection of histones is very satisfactory after treatment of the chromosomes with 45% acetic acid (Swift, 1964; Black and Ansley, 1964). Using acetic-alcohol as a fixative Allfrey et al. (1968) claim to have demonstrated the chromosomal incorporation of H3 -acetate into Chironomus thummi salivary gland chromosomes, but their data are by no means unequivocal. Benjamin et al. (1968) have recently published a preliminary report which suggests that the phosphorylation of polytene chromosome histones (in Sciara) may be important in the supposed regulatory role of histones. Robert and Kroeger (1965) have been successful in injecting trypsin into individual living salivary gland nuclei of Chironomus thummi. The effect of the enzyme is to increase the size, and rate of

PUFFS

21

H3-uridine incorporation into RNA, of preexisting puffs by a factor of 3. These authors interpret these results to indicate that in the puffs the tertiary structure of the histones is altered in some way so that it is more susceptible to trypsin attack than it is in unpuffed regions. The experimental results have been confirmed by Desai and Tencer (1968), after injection of trypsin into larvae. They also claim that injection of histones into the haemolymph of C. thummi larvae inhibits RNA synthesis in the salivary gland chromosomes. The rate of RNA synthesis in these experiments was followed autoradiographically with H3-uridine but in the absence of some measure of the ribonuclease activity of the histone preparations used the results are not readily interpretable. G. DNA SYNTHESIS AND PUFFING

It is not the aim of this review t o consider the large body of data concerning DNA synthesis and chromosome replication in polytene chromosomes. However a few remarks about the relationship between DNA synthesis and puffing are relevant. Excluding from consideration the DNA puffs of the Sciarinae (see below) no correlation between the time of DNA synthesis and puffing has ever been observed (Pelc and Howard, 1956; Schurin, 1957, quoted in Pavan, 1958; Rudkin and Woods, 1959; Ritossa, 1962a, 1964b; Schultz, in Pavan, 1965b; Sebeleva et al., 1965; Sebeleva, 1968) in either Drosophila or Chironomus. In a brief abstract Schultz and Rudkin (1960) published preliminary indications that there was disproportionately high early DNA synthesis at sites which form puffs in D. melunoguster but this observation was never confirmed. The report of Stich and Naylor (1958) of DNA accumulation at puff sites in the chironomid Glypotendipes was not sustained by Key1 ( 1963). H. “DNA PUFFS”

The sub-family Sciarinae appears to be unique amongst the Diptera in possessing “DNA puffs”. Studied by Poulson and Metz (1938) in Sciaru ocellaris and later by Pavan’s group in Rhynchosciura angelae and R. millen’ these puffs are characterized by the accumulation, at the sites of certain bands of polytene chromosomes, of large quantities of DNA. This accumulation results in the expansion of these chromosome regions. Typically DNA puffs exhibit a high degree of developmental specificity-only appearing at the end of the final (fourth) larval instar (often called the “prepupal A.I.P.-2

22

M. ASHBURNER

period” but not homologous with the prepupal stage of the cyclorrhaphan Diptera) (Dreyfus e t ul., 195 1 ; Pavan and Breuer, 1952; Breuer and Pavan, 1955; Pavan and Breuer, 1955; Mattingly and Parker, 1968a; Pavan and da Cunha, 1969b). The accumulation of DNA at puff sites in Rhynchosciuru was deduced on purely histochemical grounds by Breuer and Pavan (1955). This proposal met, originally, with some scepticism as it contravened the so-called rule of DNA constancy. Two types of experiment have, however, confirmed that DNA puffing does indeed result from the localized disproportionate synthesis of DNA. Cytophotometric measurements indicated “at least a doubling” of the DNA of puffed regions in R. ungelue during puff formation (Rudkin and Corlette, 1957). Later Rasch (quoted by Swift, 1962 and Rasch, 1966) measured the DNA content of individual Sciuru coprophilu salivary gland nuclei by similar methods. Before DNA puff formation the DNA measurements for the nuclei fell into a geometric series of classes with 256C : 512C; 1024C and 2048C DNA values. After the formation of the DNA puffs, late in the fourth larval instar, these classes were still discrete yet the distribution had become considerably skewed due t o the extra DNA synthesis at the puff loci. By measuring the relative DNA content of individual loci before, and during, puff formation Rasch confirmed the observations of Rudkin and Corlette. More recently Crouse and Key1 (1968) found, in Sciuru coprophilu, that disproportionate DNA synthesis at DNA puff sites involves complete rounds of replication-that is the DNA content of these puffs increases geometrically during puff formation. The second approach to this problem was that of Ficq and Pavan (1957, 1961) and Ficq e t ul. (1958), who demonstrated the local incorporation of radioactive thymidine into the DNA puff loci during puff formation. In addition to being sites of disproportionate DNA synthesis the DNA puffs are also sites of RNA synthesis (Ficq e t u l , 1958; Pavan, 1965a, b) and of histone (Swift, 1962) and non-histone protein synthesis (Ficq and Pavan, 1961; Ficq et ul., 1958; Mattingly, 1966). In the latter respects the DNA puffs differ from the normal, RNA, puffs. The RNA synthesized at the DNA puff sites may be eliminated from the chromosomes as small globules known as “micronucleoli” (Swift, 1962; Pavan, 1965a, b; Gabrusewycz-Garcia and Kleinfeld, 1966). The latter authors made a careful study of micronucleolar formation in Sciara coprophila. They found that

PUFFS

23

approximately 18% of the polytene chromosome bands were potential sites of micronucleolar formation, but that most of the sites were those involved in the formation of DNA puffs or were “heterochromatic” chromosome regions. With respect to RNA synthesis, greatest activity was found just prior to, and during, the early stages of DNA puff formation. At their maximum expansion the DNA puffs were not very active in RNA synthesis (but see Gabrusewycz-Garcia, 1964). Granules, similar to those seen in Balbiani rings, are present at DNA puff sites (Gabrusewycz-Garcia, 1968). Prior to pupation the DNA puffs regress. No transport of DNA from the puff sites has been observed in Rhynchosciara or Sciara sp. (Pavan, 1965b; Rasch and Pettit, 1967; Crouse and Keyl. 1968; Mattingly and Parker, 1968b), yet the majority of the DNA puffs do return to the normal banded chromosome condition. In the case of the largest puffs the post-puff band has a higher DNA content than before puff formation. In Hybosciara fragilis extrusion of DNA and associated ribonucleoprotein from DNA puff sites has been reported (da Cunha et al., 1969). This results in the formation of “micronucleoli” in the nuclear sap. These observations raise the question as to the exact nature of the “micronucleoli” of other species-perhaps these too contain DNA (see Ribbert and Bier, 1969). The exact relationship of the various “micronucleoli” of Sciarids to discrete chromosomal loci and to the nucleolus proper is an open question which should be resolved. It is clear that DNA puffing is a mechanism involving the amplification of localized regions of the genome. It is a phenomenon, therefore, which may well be functionally similar to nucleolar DNA amplification in the oocytes of amphibia (e.g. Gall, 1968) and insects (Bier et al., 1967; Gall et al., 1969). Perhaps related to DNA puff formation in the Sciarinae is the so-called “heterochromatin proliferation” described by Keyl and Hagele (1966). In the salivary gland chromosomes of some individuals of Chironomus melanotus large masses of DNA (and presumably of associated protein) were found t o be “budding off” from the heterochromatic kinetochore regions. These DNA bodies often severed their connections with the chromosomes to lie free in the nucleoplasm. Their subsequent fate was not clear. A similar process has been described in the cecidomyiid Dasyneura crataegi by Henderson ( 1967b), and in the epidermal cell polytene chromosomes

24

M. ASHBURNER

of Sarcophaga by Whitten ( 1 965) and Roberts (1 968) (see Section

XI). VI. PUFFS: DEVELOPMENTAL PHYSIOLOGY A. DEVELOPMENTAL SPECIFICITY

Puffs are transient modifications of polytene chromosome regions. Many, if not all, of them appear at specific stages during development, are active for varying periods of time, and then regress. These characteristics of puffs may be well illustrated by a general description of the changes in puffing that occur during larval development of Drosophila melanogaster (Becker, 1959; Ashburner, 1967a, 1969a). T o set this description in context it will first be necessary to give an abbreviated account of the development of Drosophila. After eclosion from the egg the larvae proceed through three larval stages of which the third is the longest and the period of most rapid growth. Near the end of the third larval instar the larvae migrate from the food and begin a period of wandering. Eventually they settle down (usually when 120 hr old), become immobile and form a puparium by the contraction and subsequent tanning of their cuticle. Within this puparium the epidermis soon separates from the larval cuticle to secrete a new, extremely delicate, cuticle. This event occurs 4 hr after puparium formation and the animals are now pharate fourth instar larvae (Snodgrass, 1924). Eight hours later the pharate larva moults again into the true pupa. The period between puparium formation and pupation is commonly, although loosely, called the “prepupal stage”. The salivary gland chromosomes of young third instar larvae possess few-perhaps five-obvious puffs. Not until some 10 hr before puparium formation d o any clear changes in the puffing pattern occur. Then, quite suddenly, puffs form, in a characteristic and regular sequence at a number of loci whilst some of the puffs active in younger stages regress. These changes continue, with new puffs appearing and some puffs regressing until, at puparium formation, some 70 sites are puffed. After puparium formation the actual number of sites puffed decreases although new puffs continue to appear. The period between 3 and 5 hr after puparium formation is one of generally low puffing activity. From 6 hr after puparium formation the number of puffs increases again reaching a peak in 8 hr prepupae. This is followed by a slow decline in puffing as pupation is approached. In D . melanogaster peak periods of puffing activity precede the

PUFFS

25

subsequent moults by 4 hr. With respect to these moults the puffs may be grouped into three classes: (i) those active at both periods, yet inactive in the intervening interval (44% of puffs), (ii) those active exclusively during one or other period (42%), and (iii) those puffs whose activity is not clearly related to the moults ( 14%). Similar changes in puffing activity with development have been described in other species of Drosophila, e.g. D. simulans (Ashburner, 1969a, b), D. hydei (Berendes, 1965a), D. busckii (Kroeger, 1960; Ashburner, unpublished), D. anannassae and D. takahashii (Ashburner, unpublished). The occurrence of very active periods of puffing activity during moulting has been studied in detail by Clever in Chironomus tentans. In contrast to the Acalyptrate Diptera (e.g. Drosophila) some Nematocera have a mobile pupal instar. In Chironomus there are four larval instars; the morphological changes, especially in the thoracic region, of old fourth instar larvae have often resulted in this stage being called a “prepupa”. This stage is not, however, completely homologous with the prepupal stage of Drosophila. Clever analysed changes in puffing activity of salivary gland chromosomes occurring during the moult from the third to four larval instars (Clever, 1963a) and during the larval/pupal moult (Clever, 1962a). He has also (Clever, 1962b) studied the puffing patterns of diapausing fourth instar larvae. From an analysis of the data presented by Clever for Chironomus tentans and of the Drosophila data, one major difference in the developmental behaviour of puffs in these two groups is apparent. The puffing patterns of three species of Drosophila have been more or less completely analysed ( 0 . melanogaster, D. simulans and D. hydei) and the results clearly show that the great majority (over 80%) of the puffs are active only at specific stages of development. In Chironomus, on the other hand, Clever stresses, and his data certainly confirm, that only a minority (perhaps 1Wo) of the puffs show such developmental specificity. In Chironomus the remaining puff sites are active more or less continuously and fail to show any variation in their activity that can be correlated with developmental events. Clever ( 1962a) thus distinguishes two classes of puffs: “entwicklungsspezifische Puffs” and “entwicklungsunspezifische Puffs”. It should be pointed out that this classification is by no means absolute: some of the “unspecific” puffs (e.g. 1-17B, I-19A) do show marked changes in size at certain developmental stages. This difference in puff behaviour between Drosophila and Chironomus is

26

M. ASHBURNER

probably of little physiological significance, and indeed may be more apparent than real. Clever routinely uses a sensitive cytochemical technique (orcein-light green) for the detection and scoring of puffs while workers with Drosophila rely, for technical reasons, largely on morphological criteria. Clever’s method probably detects a much higher proportion of the total number of active chromosome sites than would be seen on morphological criteria alone. In Drosophila there may well be a large number of puff sites, which are very small and not obvious by their effect on chromosome diameter, and whose activity is, in Clever’s sense developmentally unspecific. This point can only be settled by a very detailed analysis of puff sites in Drosophila, preferably by high resolution autoradiographic studies after uridine incorporation. Cytologically the first changes detectable in the puffing patterns of the salivary gland chromosomes of Chironomus tentans before either the larval/larval or larval/pupal moults is the appearance of a puff at 18C on chromosome I, followed by puffing at 2B on chromosome IV and at 14A on chromosome I. During the moult from third to fourth instar larvae these are the only major puffing changes, the activity of most of the other puffs remaining unchanged throughout the moult. With the completion of ecdysis the puffs at I-l8C, IV-2B and I-14A regress. Only one puff, II14A2 was found by Clever to be specific for the larval/larval moult, although even this puff does show some activity at the subsequent larval/pupal moult. Whilst initially the changes in puffing that precede pupation are similar to those that preceded the larval/larval moult (i.e. activation of I-18C, IV-2B and I-14A) puffs appear later at at least four sites (1-1 1B2, I-A1, 11-14A and 111-9B) that have not been active during previous development. In addition Clever describes two groups of puffs active in late fourth instar larvae. The first group (I-17B and 1-19A) are puffs that were active since the third larval instar but did not increase in size during the larval/larval moult. They enlarge in late fourth instar larvae yet regress before pupation itself. The second group (e.g. I-13B and 11-6A) were not active previously, enlarge in the prepupal stage and continue active throughout the moult. After pupation the salivary glands of both Drosophila and Chironomus histolyse. In Chironomus, but not in Drosophila (Ashburner, 1967a), there is a very great reduction in puffing activity (especially at the “developmentally unspecific” sites) just prior to pupation. In diapausing fourth instar larvae the spectrum of potentially

PUFFS

27

active puff sites is identical t o that of non-diapause animals. There is, however, a marked reduction in the frequency of sites actually puffed. On initiation of metamorphosis in the diapause larvae the pattern of moult specific puff changes does not differ from that found in larvae which develop without a diapause (Clever, 1962b). Data comparable t o that of Clever are not available for any other species of Chironomid although some information has been published concerning the puffing patterns of two species, C. thummi and C. dorsulis. For C. thummi Kroeger (1964) described the activities of seven puffs during the pupal moult. Four of these puffs regress before the moult while three become active. One of these (puff IIIdl) woiild appear to be homologous with the puff I - l K of C. tentuns. According t o information published in Kroeger and Lezzi (1966) of 27 puffs investigated in C. thummi the activity of seven was developmen tally specific. In C. dorsulis Kiknadze and Filatova (1963) studied a total of 28 puffs on chromosomes I and 111. Ten of these were active in prepupae, but not earlier in the fourth instar, 14 larval puffs regressed before pupation, while the activities of the remaining three puffs remained unchanged during these periods. All puffs except one regressed after pupation (see Clever, 1962a). The exceptional puff ( 12A on chromosome I) was active only after the pupal ecdysis. The puffing patterns of another chironomid, Acricotopus lucidus, have been analysed by Mechelke (1952, 1958, 1962) and Panitz (1964). The salivary gland of this species is of great interest for, as will be discussed in detail below, it consists of three lobes, each lobe possessing a unique pattern of Balbiani rings. The BR3 and BR4 are present only in the anterior lobe of the gland and are active throughout the fourth larval instar. Just prior to pupation they regress (Mechelke, 1952, 1958). These BR are also active in third @star larvae and show some slight reduction in their activity during the moult to the fourth instar (Panitz, 1964). The BR of the other lobes of the gland (the side and main lobes) (i.e. B R l , BR2, and BR6) d o not show changes in activity correlated with gross developmental events. The DNA puffs of Sciarids show a developmental specificity similar to that described above. In Rhynchosciuru ungelue three major DNA puffs (on chromosomes B and C) form about midway through the prepupal stage (i.e. late fourth instar larvae). One puff on chromosome B reaches its maximum size first and is followed by the puff on chromosome C. By pupation both DNA puffs have

28

M. ASHBURNER

regressed (Breuer and Pavan, 1952). Similar events were found in R . milleri (Pavan and Breuer, 1955). More recently Mattingly and Parker (1968a) have studied a third species in this genus and analysed the behaviour of 13 DNA puffs through the fourth larval instar and pupation. These puffs appear and regress in an ordered fashion in a manner familiar from studies on Drosophila and Chironomus. There is a marked peak in puffing activity in 54-day-old larvae-that is the day prior to pupation. In passing it may be noted that Rhynchosciara larvae develop gregariously and that the larvae of any one group (the progeny of a single female) exhibit an almost incredible synchrony in their development throughout the long (50-60 day) larval period. Extensive DNA puff formation prior to pupation is typical also of Sciara ocellaris and S. coprophila (Gabrusewycz-Garcia, 1964; Crouse, 1968). Gabrusewycz-Garcia ( 1964) also noted the developmental specificity of normal RNA puffs (“bulbs”) in S. coprophila. DNA puff formation at larval moults in Sciarids remains to be investigated . B. INTRAGLAND VARIATION

One approach in the study of puffs and their functional significance has been the comparison of puffing activities in the various regions of the salivary glands of those species whose glands show some morphological complexity. Dipteran larval salivary glands are paired ectodermal structures usually with a relatively low cell number that remains constant throughout the functional life of the gland. Frequently (e.g. in the drosophilids) the gland is asymmetric, one lobe being larger than the other. Even in those species where this asymmetry is particularly evident [ e.g. D. victoria (Wharton, 1943), D. lebanonensis (Ward, 1949)] no difference in puffing activity has been found between the two lobes (Berendes, personal communication). In the chironomid Prodiamesia olivacea the left and right lobes of the gland differ greatly in their shape (Melland, 1941) although no puff studies have been undertaken with this species. In many Chironomids, however, the individual lobes show some morphological specialization and the different regions of the gland differ in their pattern of puffing activities. The external form of the salivary glands of the closely related species Chironomus tentans and C. palliduvitatus is very similar. Yet in C. palliduvitatus a group of four cells (Sonderzellen) close to the duct differ from the remainder of the cells in possessing a particular type of secretion granule. In C. tentans the homologous group of cells do not differ in their

PUFFS

29

histological structure from the major part of the gland in this way. Beermann (1956, 1959) surveyed various species of Chironomus and discovered glands with Sonderzellen in representatives of the following genera: Cryptochironomus, Trichocludius (see Beermann, 1952a). Acricotopus and Zavrelia. In the subgenus Chironomus sensu strictu Beermann (1961) notes that all species studied, except C. tentans, possess from 4 to 6 Sonderzellen in the anterior region of the gland. Salivary gland organization in another chironomid (Orthocladid), Acricotopus lucidus is more complex: the gland has three distinct lobes. The anterior lobe accumulates beta-carotinoids during the prepupal stage (Mechelke, 1952; Baudisch, 1960, 1963a, 1967). The main and side lobes both differ from the anterior lobe in their high content of hydroxyprohne (Baudisch, 1960, 1963c, 1967). These differences in cell morphology and function within the salivary glands of Chironomids are reflected in differences in the pattern of activities of the Balbiani rings. In the nuclei of the Sonderzellen of C. palliduvitatus a small BR on chromosome IV is puffed; this locus is inactive in the other nuclei of the gland and is not present in any nucleus of the gland of C. tentans. Hybridization of the two species is possible in the laboratory and by virtue of species specific inversions their IVth chromosomes are distinguishable. Four Sonderzellen are seen in salivary glands of hybrid larvae but the number of cytoplasmic granules is lower than in C. palliduvitatus. Of the two fourth chromosomes in the hybrid the special BR SZ is active only on the C. palliduvitatus homologue. In segregating generations the ability to make the Sonderzellen granules and the inheritance of the SZ Balbiani ring parallel each other exactly. The Mendelian factor responsible for granule production in the Sonderzellen maps genetically at a site coincident with the location of the BR SZ (Beermann, 1961). There are, therefore, good reasons for proposing a causal relationship between the activity of the special BR and the differentiation of the Sonderzellen as witnessed by granule production (Beermann, 1961, see also Grossbach, 1968 and Section XII). Trichocladius vitripennis (Chironomid : Orthocladid) is another species with a group of special salivary gland cells. In this species the six Sonderzellen (which possess a granular cytoplasmic secretion) differ from the main gland in the activities of two BR. Furthermore a third BR, active in the “normal” cells is inactive in the special cells (Beermann, 1952a). A.I.P.-2*

30

M. ASHBURNER

Changes in the activities of the BR with the development of fourth instar larvae and prepupae complicates the pattern of intragland variation of BR in Acricotopus lucidus. In the carotin accumulating anterior lobe two BR (BR3 and BR4) are active in fourth instar larvae, but regress before pupation. Three other Balbiani rings show no obvious change in their activities with development. Two (BRl and BR2) are common to both the main and side lobes, whilst the third (BR6) is active only in the side lobe (Mechelke, 1952, 1958, 1962). Intragland variation in puffing activity has also been described in species which lack any clear morphological differentiation of their salivary glands. The sac-like gland of the Drosophilids is a case in point. In the repletu species group the distal salivary gland cells of mid third-instar larvae often possess cytoplasmic granules clearly visible in living preparations under phase contrast. With increasing age the boundary between these granule containing cells and the cells lacking the granules moves anteriorly and eventually nearly all of the cells contain granules (McMaster-Kaye, 1962; Berendes, 1965a). In D. hydei there is a correlation between the activity of a particular puff (47B) and the presence, in the cytoplasm, of the granules (Berendes, 1965a). Although salivary gland differentiation in D. melunoguster is not as clear as in D. hydei Becker (1 959) recorded that the X chromosome puff 15BC is larger in the proximal cells of the gland. The cytoplasm of these proximal cells is considerably more basophilic than that of the more distal, larger, cells. The salivary gland of Sciuru coprophilu is linear in form consisting of two rows of cells surrounding the lumen. The cells of the anterior region of the gland are larger than those of the posterior region, from which they are separated by a narrow neck. The DNA puffs of prepupae are active in both anterior and posterior regions (Gabrusewycz-Garcia, 1964) but not in the neck cells (Crouse, 1968). Differences in the spectrum of DNA puffs active in the two main regions were noted by GabrusewyczGarcia ( 1964). In addition one RNA puff (bulb 111-15B) is only active in the anterior cells. Despite its enormous length (circa 2 cm) the salivary gland of Rhynchosciuru angelue is not differentiated and the DNA puff pattern of all nuclei is identical. The formation of the DNA puffs does, however, start slightly earlier during development in the anterior cells (Breuer and Pavan, 1955; Pavan, 1965a). Considerable intragland variation in DNA puff activity is found in certain abnormal individuals of this species (Pavan, 1965a).

PUFFS

31

C. TISSUE SPECIFICITY O F PUFFS

Differences in the activity of Balbiani rings distinguishing different larval tissues of Chironomus tentans was first studied by Beermann (1 952b). Although similar tissue specificity has been noticed by other workers (e.g. for the DNA puffs of Rhyncosciara by Breuer and Pavan, 1955) only Berendes (1965b, 1966) has studied this phenomenon in detail. Berendes compared the puffing patterns of the salivary gland, Malpighian tubule and midgut chromosomes of Drosophila hydei. In the first of these papers (Berendes, 1965b) he compared the puffing activities in salivary gland and stomach cells during the period 10-6 hr before puparium formation. A total of 116 different puffs were scored and of these 14 were active in the salivary gland, but not the stomach and eight were active in the stomach cells but not the salivary gland. Analysis of the puffing patterns of the chromosomes of these tissues at different ages, however, demonstrated that several of these 22 puffs were active in both tissues but that they differed in the time of their activity relative to the time of puparium formation. In fact only five puffs (four in the salivary gland and one in the stomach cells) were truly tissue specific when such timing differences were taken into account. In D. hydei the activity of those puffs common to both salivary gland and stomach cells commences earlier during development in the latter tissue. This is especially true of those puffs which, in the salivary gland, are most active at puparium formation (Berendes, 1965b). This emphasizes the difficulty in rigorously determining whether or not a particular puff is tissue specific. Indeed many of the puffs originally considered by Beermann (1952b) to be tissue specific are now known to be not so (Clever, 1962a). Berendes later (1966) extended his analysis to include the Malpighian tubules of both larval and adult D. hydei. His data indicate that the great majority (7%) of puffs are common to the three tissues studied. Only 14/116 puffs are specific to a single tissue and a further 19 puffs are specific to two of the three tissues. Of 22 salivary gland puffs considered by Berendes to be specific for the period of puparium formation only six showed similar changes in their activity at puparium formation in stomach and Malpighian tubule cells. There were no puffs specific for the adult Malpighian tubules although many puffs (22) active in this tissue in the larvae were not present in adult chromosomes. In Chironomus tentans Beermann (1952b) noted differences in the

32

M. ASHBURNER

pattern of puffing activities between the distal and proximal regions of the larval Malpighian tubules. Recently Buetti (1968) has followed the Malpighian tubule chromosomes of C. thummi throughout pupal development and described changes in the puffs correlated with certain morphological stages of pupal differentiation. It appears that the midpupal period is one of low puffing activity and that a dramatic increase in the size of certain puffs occurs just prior to the adult ecdysis. The general conclusions of this study are in accord with those drawn from the analysis of puffing patterns during earlier developmental stages (see Section VI(A)). In Rhynchosciuru ungelue the DNA puffs are highly specific-no single puff having been identified in more than one tissue type (Breuer and Pavan, 1955). VII. THE ENDOCRINE CONTROL OF PUFFING ACTIVITY A. ECDYSONE

The correlation between periods of intense puffing activity and moults led Clever and Karlson (1 960) and Becker (1959,1962a) to implicate the moulting hormone ecdysone in the control of this puffing activity. Becker (1962a) ligatured third instar D. mehogaster larvae behind the ring gland in such a way that ligature divided the salivary gland into two regions. When the ligature was applied before the time of moulting hormone release by the ring gland (see Fraenkel, 1935) the posterior part of the salivary gland was separated from the hormonal source. Analysis of the puffing patterns from such larvae showed that the nuclei of the anterior part of the gland possessed puffing patterns characteristic of the time of puparium formation yet the puffing pattern of the posterior nuclei had not advanced beyond the typical intermoult stage. I have confirmed these results (Ashburner, unpublished). In a second series of experiments salivary glands were transplanted between animals of differing age (Becker, 1962a, b), and puffs characteristic of the host animals induced. The role of ecdysone in the control of puffing was more directly studied by Clever and Karlson ( 1960) and later Clever published the results of his extensive experiments in a series of papers (Clever, 1961, 1962b, 1963b, c). In the preliminary communication Clever and Karlson ( 1960) injected 10-20 Calliphora Units (C.U.) (see

PUFFS

33

Karlson, 1966); 1 C.U. = 0.01 pg) of an ecdysone preparation into intermoult fourth instar Chironomus tentans larvae. As a result of this treatment there was a marked induction of puffing at one site-I-18C while the puff I-19A present at the time of injection regressed. In a further series of experiments Clever ( 196 1 ) found that the reaction of puff I-18C to ecdysone was rapid-the puff was visible within 10-15 min after injection. Closely following the appearance of I-18C a second puff, IV-2B, was induced by ecdysone. It will be recalled that the two puffs I-18C and IV-2B of C. tentans are the first to appear, during normal development, prior to the moults (see page 26). The sequence of their induction by ecdysone in intermoult larvae is identical to the sequence seen during normal development. At moderate doses of ecdysone (e.g. l0C.U.) both I-1% and IV-2B reach their maximum activity within a few hours of the injection and then regress. At higher doses (e.g. 45 C.U.) they remain active for considerably longer periods (48 hr in the case of IV-2B and over 72 hr in the case of I-18C). If, after the regression of the two puffs following a 10 C.U. injection the animals are reinjected with ecdysone both puffs are induced a second time. The dose response relationships for the ecdysone induction of I-18C and IV-2B were investigated by Clever in a later paper (Clever, 1963b). Intermoult fourth instar C. tentans larvae were used as the test animals and were injected with ecdysone with the doses ranging from 2.10-7 pg to 2.10-' pg/animal. The mean wet weight of the larvae being in the order of 20 mg these doses correspond to and pg/mg respectively. The reaction thresholds of the two puffs differed; that for I-18C was pg/mg and that for IV-2B pg/mg. Higher doses were necessary if the puffs were to attain their maximum size: lo-' to pg/mg for I-18C and pg/mg for IV-2B. The threshold doses were calculated to correspond to some 10-100 molecules of ecdysone per haploid genome. These results were interpreted by Clever to indicate that during normal development the two puff loci respond to a gradual increase in ecdysone titre which precedes the moult, the puffs reacting as their respective threshold concentrations are reached. These conclusions were tested in a further series of experiments (Clever, 1964a; Clever and Beermann, 1963). Haemolymph was taken from animals at four ages, (i) fourth instar larvae immediately after the moult from the third instar, (ii) intermoult fourth instar larvae, (iii) young prepupae just before the activation of 1-18C, and (iv) old prepupae. These haemolymph samples were injected into intermoult

34

M. ASHBURNER

fourth instar larvae and their effect on puffing at I-18C and IV-2B assessed. The results of these experiments were that late prepupal haemolymph always induced both puffs whilst young prepupal haemolymph and haemolymph from immediately post-moult fourth instar larvae sometimes induced 1-18C but never IV-2B. Haemolymph from intermoult larvae was without effect on puffing at these loci. The injection of ecdysone into either post-moult fourth instar larvae or into young prepupae enlarged (if possible) the size of the puff at 1-18C and, in animals of an appropriate age, induced puffing at IV-2B which, during normal development regresses earlier than the 1-18C puff. The situation found after the injection of ecdysone into older prepupae is more complicated. In contrast to the pattern found after the third larval moult the puff I-18C does not regress in old prepupae. This observation is consistent with the finding that the ecdysone titre in these animals is high (the haemolymph experiments). Despite this IV-2B does regress in old prepupae and this fact seemingly contradicts the hypothesis that the activity of these two puffs is dependent upon the haemolymph ecdysone titre. Injection of ecdysone into old prepupae does not induce IV-2B (Clever, 1961, 1964a). These results suggest that some other factor controls the regression of IV-2B prior to pupation. Since treatment of old prepupae with cycloheximide, an inhibitor of protein synthesis, can result in IV-2B induction in these animals, Clever ( 1966b) has suggested that the regression of IV-2B at this stage is due to the action of some unspecified substance, whose metabolic survival is dependent upon protein synthesis. Ecdysone will also induce puffing at I-18C and IV-2B in vitro but only when the glands are incubated in relatively complex media, and even then the results are erratic (Clever, 1965b). So far only the initial chromosomal response to ecdysone has been considered. However apart from I-18C and IV-2B a large number of puff loci react and d o so in a sequence similar to that found during normal prepupal development (Clever, 196 1). The reaction of these loci is, however, considerably delayed; the first (I-8A) being activated more than 5 hr after the injection of 45 C.U. of ecdysone and the last group (e.g. III4A) some 7 2 hr after injection. Some of the puffs that react are not classified by Clever as moult specific puffs (e.g. I-17B, 11-16A). The puff I-19A is of interest as within 2-8 hr after ecdysone injection it regresses (see Clever and Karlson, 1960) yet later, at 15 hr, its activity increases. The effects of ecdysone on puffing have been examined in two

PUFFS

35

other species of Chironomus, C. palliduvitatus and C. thummi, although less comprehensively than in the C. tentans. In C. palliduvitatus, a sibling species of C. tentans, the puff homologous to I-18C of C. tentuns shows a similar ecdysone response (Clever, 1961). The transplantation of intermoult fourth instar larval salivary glands of C. thummi into prepupae results in the induction of the puff IIId 1. During normal development IIId 1 displays a similar pattern of activity to I-18C of C. tentans and the two puffs are presumably homologous (Kroeger, 1964; Pelling, 1965). Induction of IIId 1 can also be achieved in vitro by adding ecdysone to cultured salivary glands of C. thummi (Kroeger, 1966). This response, however, is very sensitive to such factors as oxygen availability in the medium-an observation that may account for Clever’s (1965b) difficulty in achieving a reliable effect of ecdysone in vitro. C. Davis (personal communication) has also found difficulty in inducing puffing with ecdysone in vitro, perhaps suggesting the necessity of other substances for the ecdysone response. No effect of ecdysone on puffing in isolated salivary gland nuclei of D.hydei could be found by Berendes and Boyd (1969). The regression, in prepupae, of two Balbiani rings (BR3 and BR4) of the anterior salivary gland lobe of Acricotopus lucidus has already been mentioned (page 30). Panitz (1964) used a variety of experimental techniques to demonstrate that this inactivation is dependent upon a high ecdysone titre. Interestingly he demonstrated that an in terspecific transplant of larval salivary glands in to prepupae of Calliphora would result in BR regression. Similarly the incubation of the glands in the presence of Calliphoru prothoracic glands (i.e. the side limbs of the ring gland) would produce the same result. Recently Panitz (personal communication) has shown the inactivation of BR3 and BR4 following the injection of ecdysone into larvae. In addition to causing regression of BR3 and BR4 the rising ecdysone titre in Acricotopus prepupae also leads to the regression of puff I/H-l8 and the induction of puff II/Q-32. Induction of specific puffing patterns by ecdysone in Drosophilids has been studied by Berendes ( 1967a) in D. hydei and by Ashburner (1969f) in D. melanogaster. In D.hydei the consequence of injecting 33 C.U. of ecdysone into 136-140 hr third instar larvae (i.e. 20 hr before puparium formation) is to produce changes in puffing activity which “show only minor differences” from those changes that normally occur before puparium formation (Berendes, 1967a). The changes are first evident within 15-30min of the injection and

36

M. ASHBURNER

include the de novo induction of puffs (3 puffs), and both the enlargement ( 18 puffs) and regression ( 12 puffs) of preexisting puffs. A group of five puffs do not react within the first 60 min after injection of ecdysone but do react several hours later. The results of the short term ecdysone induction experiments of Berendes (1967a, Table 1) should be compared closely with the normal behaviour of the puffs prior to puparium formation (Berendes, 1965a, Table 5). Fourteen puffs are listed by Berendes as increasing in activity before or during puparium formation; only five of these are ecdysone inducible. Of 15 puffs that decrease in activity before or during puparium formation four are ecdysone inducible, and four regress after ecdysone injection. It appears that in D. hydei there are considerable differences between the normal puffing patterns and those induced by ecdysone. One further paradox is apparent from the data of Berendes (1967a). A total of 2 1 puffs either appear de novo or enlarge shortly after the injection of ecdysone into 136-140 hr larvae; 14 of these show considerable activity in the polytene chromosomes of the Malpighian tubule nuclei of adult flies, a stage expected to have a very low ecdysone titre (see e.g. Shaaya and Karlson, 1965). Because of the ambiguities of the D. hydei data it is important t o know the ecdysone response of puffs in other, related, species. So far only the incomplete data from a study of D. melanogaster are available (Ashburner, 1969f). In this study 90 hr old third instar larvae were injected with the 2 x pg of ecdysone and the puffing response followed for several hours subsequent t o the time of injection. The results clearly indicate that at this dose ecdysone initiates the complete sequence of changes in puffing activity that would normally occur before puparium formation some 20-30 hr later. The first visible effects are evident within 15 min with the induction of puffing at four loci: 2B5-6, 23E, 74EF and 75B. These puffs are the first to appear during normal development (Ashburner, 1967a, 1969a). Although the changes in puffing activity induced by ecdysone injection in 90 hr larvae are identical to the normal developmental sequence they become out of step with the gross developmental physiology of the animals. The amounts of ecdysone injected do not accelerate puparium formation and 10-14 hr after ecdysone injection third instar larvae, of normal external appearance, possess puffing patterns typical of prepupal animals. Many of the puffs inducible by ecdysone in both D. hydei and D. melanogaster are, during normal development, active at both

PUFFS

37

puparium formation and, later, prior to pupation. Yet attempts to induce them in prepupae by ecdysone injection have not been completely successful in either species (Berendes, 1967a; Ashburner, 19690. Furthermore in D. melanogaster I have found that very young third instar larvae are refractory to ecdysone (contra Berendes, 1967a). The failure of prepupae and young third instar larvae to react to ecdysone may be suggested to be due to the presence in these animals of enzymes destroying ecdysone (see Ohtaki et al., 1968; Karlson and Bode, 1969). It is not unreasonable to assume that such enzymes would be active immediately following ecdysis. The central role of ecdysone in the control of puffing activity is further exemplified by the mutant lethal-giant-larvae (Zgl) of D. melanogaster. This mutant is a larval lethal, the animals failing to form normal puparia or to pass through the prepupal moult. The animals live as larvae for several days and either die as larvae or as very abnormal animals known as “pseudopupae”. The results of ring gland transplant experiments (Scharrer and Hadorn, 1938) and of the injection of ecdysone preparations (Karlson and Hanser, 1952), together with the phenotypic characteristics of the mutant suggest that lgl is defective in some part of its ecdysone endocrine system. Mutant larvae almost completely lack puffing activity even during pseudopuparium formation (Becker, 1959; Ashburner, unpublished). They can, however, be induced to form salivary gland chromosome puffs after various experimental treatments, such as temperature shock. After such treatments the reaction of lgl larvae is qualitatively similar to that of normal larvae (see page 50 and Ashburner, 1969d). Attempts at inducing puffs in lgZ third instar larvae with ecdysone have so far been unsuccessful (Ashburner, unpublished). The kinetics of puff induction and the dose response relationships for several puffs were analysed by Berendes (1967a) in D. hydei. Loci differed in their reaction thresholds but these were generally in C.U. per larva (i.e. 5 x the order of pglmg). The dose response curves for the several puffs tested were not linear-in particular high concentrations of ecdysone (e.g. 33 C.U. per larva) had less effect than concentrations lower by several orders of magnitude [contrast with the linear response of I-18C and IV-2B in C. tentans (Clever, 1963b)l. The effect of ecdysone on puffing appears to be similar in all larval tissues of an organism so far studied, i.e. salivary gland, midgut, Malpighian tubules, of D. hydei (Berendes, 1967a), and salivary

38

M. ASHBURNER

gland, rectum and Malpighian tubules of C. tentans (Clever, 1961). The DNA puffs of the Sciarinae are also ecdysone inducible; as might perhaps have been expected in view of their appearance before pupation. Cannon (1 965) cultured the salivary gland of S. coprophila in vitro and provided circumstantial evidence of the inducibility of the DNA puffs. Recently Crouse ( 1968) injected S. coprophila larvae with 3.3C.U. of ecdysone and within 24 hr observed DNA puff formation, and marked DNA synthesis at these sites. After ecdysone injection the puffing patterns became uncoupled with gross developmental events (see above). The mode of action of ecdysone in inducing specific puffing activity is still unclear and will be discussed below. But one approach to this problem could possibly be by a comparison of the effects of a range of related compounds on puffing. Since the elucidation of the structural formula of ecdysone by Karlson’s group (see Karlson, 1966) many different “ecdysones” have been isolated, especially from plants. While few studies have as yet been completed available information indicates that there are no qualitative differences in the effects on puffing of the compounds tested. For example Berendes (1 968b) tested ecdysterone (20-OH ecdysone) and 2p 30 dihydroxySa-choleston-( 6) with D. hydei, Ashburner (unpublished) 20-OH ecdysone with D. melanogaster, and Panitz (personal communication) 20-OH ecdysone and inokosterone (20,26-dihydroxy25-deoxy-ecdysone) with Acricotopus lucidus; in all instances the effects of these steroids was the same as found with ecdysone itself. The compound 20 30 14a trihydroxy-5P-pregn-7-ene,6,20dione, a possible candidate in the metabolism of ecdysone (Horn et al., 1966), does not induce puffing in D. melanogaster (Ashburner, unpublished). There is clearly one great difficulty in the interpretation of the results of in vivo experiments such as those described. The conversion of the test compounds into a common substance which exerts the physiological effect may confound the results. Indeed there is little evidence to suggest that ecdysone itself does not become converted after injection. Recently King and Siddall ( 1969) demonstrated the conversion of ecdysone to 20-OH ecdysone in Calliphora larvae, clearly indicating that such events are possible. In Drosophila, at least, it is clear from the results of the experiments described above that ecdysone does not induce puffing activity at particular individual loci; rather it induces a whole sequence of changes in puffing pattern and this sequence at least approximates to that observed prior to puparium formation. The

PUFFS

39

question clearly arises as t o the causal relationship, if any, between the activities of early reacting loci and the activities of late reacting loci. Clever ( 1964b) and Clever and Romball ( 1966) have approached this problem experimentally. The induction, in C. tentuns, of puffs I-18C and IV-2B by ecdysone is sensitive to actinomycin D. Pretreatment of larvae with actinomycin D for 6 hr (at 0.2 pg/ml in the culture water) results in the inhibition of salivary gland RNA synthesis for about 24 hr. Injection of ecdysone immediately following actinomycin D treatment is without effect on puffing as long as the inhibition of RNA synthesis persists. With the gradual recommencement of RNA synthesis the puffs I-18C and IV-2B appear in their normal sequence to be followed by puffing at the later reacting sites. In some ways this is a surprising result as it implies either a long survival of the injected ecdysone in the haemolymph (in contrast to the results of Ohtaki et ul., 1968) or perhaps the stimulation by the injected ecdysone of the host prothoracic glands t o release further quantities of the hormone. Compare the results of actinomycin D inhibition experiments with those using inhibitors of protein synthesis such as puromycin (Clever, 1964b) or cycloheximide (Clever and Romball, 1966). In the first series of experiments larvae were treated with puromycin for 1 hr and then injected with ecdysone. This was followed by an apparently normal induction of 1-18C and IV-2B but the later reacting puffs failed to appear, at least within 24 hr (Clever, 1964b). Essentially the same results were obtained using cycloheximide (Clever and Romball, 1966) although in these experiments a third puff (I-8A) was induced following I-18C and IV-2B. Although the interpretation favoured by Clever is “that as one of the early cellular reactions to ecdysone, specific genes are induced, and that it is by the reaction of these genes that the subsequent changes of cell metabolism (i.e. the induction of the later puffs) are controlled” (Clever and Romball, 1966) there are several factors which complicate the interpretation of the results of these and similar experiments. One such complication is the observation that treatment of fourth instar larvae with cycloheximide (and other inhibitors of protein synthesis) alone results in considerable changes in puffing activity including the induction of puff 1-18C (Clever, 1967). Similarly in Drosophilu melunoguster the injection of 90 hr third instar larvae with either cycloheximide or puromycin alone results in alteration of the puffing pattern including the induction of

40

M. ASHBURNER

puffs at novel sites and the induction of puffs at 74EF and 75B-two of the earliest puffs to react to ecdysone (Ashburner and Mitchell, unpublished). The recent results of Berendes (1968a) have also confused the interpretation of actinomycin D experiments. He finds that it is possible to induce puffs, or at least acid protein accumulation at puff sites, by ecdysone and other agents even under conditions where the incorporation of uridine into RNA (as detectable by autoradiography) is inhibited by actinomycin D (see Clever, 1968; Clever et al., 1969a). A great handicap to our interpretation of experiments using metabolic inhibitors is our ignorance of the precise metabolic consequences of their presence within the cells. For example it is difficult to discount the possibility that treatment of Chironomus larvae with cycloheximide results in the blocking of particular metabolic processes, unrelated to the action of puffs I - l 8 d and IV-2B, and that the failure of subsequent puff induction by ecdysone is a consequence of this effect. As an illustration we can imagine that the cycloheximide stops the synthesis of RNA polymerase and of the ribonucleoprotein puff granules-both presumably necessary components for puff induction. Subsequent to the induction of I-18C and IV-2B the cells may run short of these components and further puff induction would not take place. The causal relationship between the induction of I-18C and IV-2B and the induction of the later puffs, or the causal relationship between the activities of any groups of puffs cannot therefore be assumed. As an alternative hypothesis the complete sequence of gene activations may be pre-programmed within the chromosomes themselves, the activities of individual loci being essentially independent of each other. If such were the case then ecdysone or any other external inducer may only be acting as the trigger for the initiation of this programme (see Kroeger, 1963a, 1964). B. MECHANISM OF ECDYSONE ACTION

Considerable controversy has arisen concerning the molecular basis of ecdysone action, particularly in respect of the induction of specific puffing activity. Two main views have been proposed and discussed by Karlson (e.g. 1965, 1966) and Clever ( 1 9 6 4 ~ )on the one hand and by Kroeger (e.g. 1968) on the other. The publication of the initial finding of ecdysone induction of puffing in Chironomus tentans coincided with the announcement of

PUFFS

41

a general theory of gene control by Jacob and Monod (1961a, b). The mechanism of ecdysone action was interpreted within the framework of this model by Karlson and Clever. In the terms of Jacob and Monod’s hypothesis ecdysone was thought to act as an effector molecule which bound to repressor molecules resulting in their detachment from specific DNA regions allowing subsequent DNA transcription (see Beermann, 1963). Rigorous proof that ecdysone acts in this way would require extensive genetic and biochemical analysis; analysis which is so far not forthcoming. However, the model makes one prediction which should be easily testable with the availability of labelled ecdysone. That is the site of action of ecdysone will be intranuclear. The model does not predict, as is often assumed (e.g. Clever, 1964a; Kroeger, 1964) the chromosomal binding of ecdysone. Karlson et al. (1964) followed the fate of unspecifically tritium labelled ecdysone in Calliphora larvae with particular attention to the epidermal cells. Although label was recovered from the nuclear fraction a very considerable (60%) proportion remained in other cell components. More recently Milkman ( 1 968, quoted by Williams and Robbins, 1968) and Emmerich ( 1969) have made use of the recently available specifically labelled ecdysones t o follow the cellular distribution of the hormone. Neither group have been able to report any marked preferential incorporation of ecdysone into either nuclei or chromosomes of Drosophila salivary glands. In the absence of experimental data it is difficult to evaluate the hypothesis of Karlson further. The interaction of physiologically active steroids with cellular membrane systems and their effects on the permeability of these membranes to small ions is well known (see Willmer, 1961). The hypothesis that ecdysone is acting at the level of the cellular membranes and changing the selective permeabilities of these membranes to small ions (Kroeger, 1963b, 1966, 1967, 1968) has, therefore, a general basis although it has met with considerable scepticism and controversy. Specifically Kroeger suggests that ecdysone acts on the plasma/nuclear membranes and results in a rise in the intranuclear concentration of potassium ions. As a direct consequence of this effect induction of puffing activity at specific chromosome sites results. Explantation of the salivary glands from late fourth instar Chironomus thummi larvae into haemolymph or saline led to specific changes in the puffing patterns. In particular Kroeger ( 1963b)

42

M. ASHBURNER

observed that the puffing pattern underwent an apparent reversionexisting puffs regressing and puffs characteristic of younger stages in development appearing in a regular sequence. This reversion of puffing patterns on explantation was called rejuvenation by Kroeger (1963b) and is probably similar to the phenomenon of “rollback” (Ruckwarstslaufen) noted some years earlier in D. melanogaster by Becker ( I 959). [Miss C. Davis ( I 967, unpublished observations) has been unable to confirm “rollback” of puffing in D. melanoguster.] The important fact concerning Kroeger’s rejuvenation phenomenon is that, in C. thummi at least, it involves a more or less precise reversal of the ecdysone initiated puffing patterns and that during reversal, as during normal development, the puffs appear in progression. Subsequently Kroeger ( I 963a, 1964) analysed the cellular components whose presence was necessary for rejuvenation to occur. In fact it occurred with equal facility in isolated nuclei and in whole glands (Lezzi, 196 1 ). Kroeger ( 1964) then demonstrated that the activation or regression of any particular puff locus during rejuvenation was independent of the presence of any specific chromosome region other than, naturally, the puff locus itself. He concluded, therefore, that the “control system’’ responsible for gene activation and repression during rejuvenation lies in the nuclear sap (Kroeger, 1963a, 1964) and assumed that this control system would be identical to that operating during normal development in the intact animal. What is this control system? Although demonstrating that a number of compounds (e.g. ZnCI, , urethane) could, in explanted glands, counteract rejuvenation and even advance the puffing pattern to that characteristic of older animals (Kroeger, 1964) these were clearly a-physiological agents. [The effect of ZnCl, on puffing could not be confirmed by Clever (1965b) using C. tentuns but has been confirmed by Berendes (personal communication) with C. thummi. ] According to results first published in 1963 (Kroeger, 1963b) the control system is the intranuclear concentration of particular species of ions-especially potassium and sodium ions. The replacement of sodium by potassium in the explant medium led to a progression of the puffing pattern closely resembling that thought in normal animals to be due to the rising titre of ecdysone. One puff, IIId1, which can be induced rapidly with ecdysone treatment (see page 35) was also inducible, in intermoult fourth instar larvae, by

PUFFS

43

explantation of salivary glands into a medium containing a relatively high (e.g. 120 mM) potassium concentration. The induction of puff IIId1 b y potassium ions has been confirmed b y Berendes (personal communication). Similarly puffs characteristic of younger stages were inducible by particular concentrations (e.g. 100 mM) of sodium ions. On this basis the rejuvenation phenomenon observed on explantation into regular (i.e. high sodium) media o r on wounding of the glands was viewed as the result of an influx of sodium ions into the gland nuclei (Lezzi and Kroeger, 1966). During normal development the sodium concentration within the nuclei would be high during the intermoult stages, b u t with the approach of a moult the rising ecdysone titre would change the membrane permeabilities in favour of a high intranuclear potassium ion concentration. Specific activation of moult specific puffing patterns would result. Clever ( 1965b) attempted to repeat Kroeger’s observations and explanted the salivary glands of C. tentans into a variety of media. Although he observed many changes in puffing activity as a consequence of such experiments he was unable to find any specific puff induction by particular ions. Clever attributes the effects observed by Kroeger to unspecific changes in osmotic pressure since he observed identical effects of isotonic NaCl and KCI solutions. Alteration of the osmotic pressure of the medium by ions or by larger molecules certainly does produce drastic changes in puffing activity in explanted salivary glands of C. thummi (Kroeger, 1967) yet there remain a small number of loci which only respond t o changes in the concentration of particular ions. Four puffs react to a high sodium concentration, one t o a high potassium concentration and one to magnesium ions (see also Lezzi, 1967c for an account of BR induction in C. thummi and C. tentans by magnesium ions). These particular ions will also produce identical effects with isolated salivary gland nuclei (Lezzi, 1966, 1967c) of C. thummi. Furthermore, Lezzi has been able to induce puffing at I-18C and IV-2B of C. tentans by incubation of isolated nuclei in a potassium-rich medium (Lezzi, 1966, 1967), and a larval puff, 1-19A, with a sodium-rich medium. One puff (IV-5C) of C. tentans also reacts to magnesium ions. There appear to be real differences in the membrane permeabilities of the two species which account for the differences of the experimental results obtained using whole salivary glands. Such differences have also been found between C. thummi

44

M. ASHBURNER

and C. tentans with respect to uridine permeability (Berendes, personal communication). Berendes et al. ( 1965) also tested the effects of K'ions on puffing in D. hydrei. Their experimental results were ambiguous and have been interpreted both in favour (e.g. Berendes et al., loc. cit.) and against (e.g. Pelling, 1965; Clever, 1965b; Berendes, 1968a) Kroeger's ideas. If ecdysone is indeed causing a redistribution of ion species across the cellular membranes then this should be reflected in the electrophysiological parameters of these membranes. Kroeger ( 1966) measured the potential difference across the plasma membrane in explanted salivary glands of C. thurnrni. After explantation he found a progressive depolarization of the membrane; the extent of this process apparently correlated with the degree of puff rejuvenation. More importantly on addition of ecdysone to the explant medium the plasma membrane potential difference increased and only after this reached -35 mV was the puff IIId1 induced. Kroeger argues that the potential difference across the plasma membrane is a result of the asymmetrical distribution of potassium ions be tween the cells and their surrounding medium. Ito and Loewenstein ( 1 9 6 3 , on the other hand, although finding a change, attributable to ecdysone, in the resistance of the nuclear membrane of C. thurnmi salivary gland cells, failed to find evidence for related changes in the plasma membrane resting potential. It is important to recall in this discussion that studies by Loewenstein and his co-workers have shown that the membrane resistance between adjacent cells of the Drosophila salivary gland is so low that virtually no barrier exists to the free diffusion of ions and larger molecules (e.g. flourescein labelled serum albumin, mol. wt 69,000) between cells (Loewenstein and Kanno, 1964; Kanno and Loewenstein, 1966). This permeability depends upon free calcium ions and is sensitive to agents that uncouple oxidative phosphorylation (Politoff et al., 1969). There is, however, a considerable barrier to the diffusion of ions between the outside surface of the cells and the external environment of the gland. In view of this it is unlikely that differences in ion concentration can be maintained between different cells of the same gland, and unless special permeability barriers within the gland exist, it is hard to see just how differences in ion concentration could account for intragland variation in puffing activities (see page 28). It would be

PUFFS

45

interesting t o study the nature of the cellular junctions between, for example, the Sonderzellen and the main gland cells of C. palliduvitatus (see Wiener et al., 1964). One of Kroeger’s objections to Karlson’s theory was that “it is difficult to see how such a small molecule (i.e. ecdysone) could harbour the specificity for singling out a specific base sequence in order t o activate it” (Kroeger, 1963a). Despite the fact that such an event is not necessitated by the theory the objection applies with equal validity t o the proposal that particular ion concentrations control specific gene activities. Kroeger ( 1 967) argues that they do so by displacing histones from particular chromosome regions. It is well known that high salt concentrations will disassociate the DNAhistone complex. However, even if the role of histones in the regulation of gene action were fully acceptable the recognition mechanism is neither defined nor obvious. At present we may summarize this rather unsatisfactory state of affairs by saying that the mode of ecdysone action at the puff level is unknown. Neither of the predominant ideas have a firm enough theoretical or experimental basis for acceptance. Before leaving the subject of the mode of action of ecdysone mention should be made of some recent reports of the effects of vertebrate corticosteroids on puffing activity. Gilbert and Pistey (1966) claim to have induced specific puffs in isolated D. melanogaster salivary glands after incubation in the presence of hydrocortisone (250 pg/ml). However, there is no evidence from the data published to substantiate their claim since untimed larvae were used and since an essential control (incubation of sister salivary glands) was omitted. The effects of a related steroid, cortisone, on DNA puffing have also been studied recently. Addition of cortisone acetate (70 mg/g) or hydrocortisone to the food medium of Sciara coprophila for the complete period of larval development did not, apparently, influence either developmental rate or the viability and subsequent fertility of the flies (Goodman et al., 1967). Yet, it was claimed, this treatment resulted in the complete absence of the normal DNA puffing cycle characteristic of prepupal development. Prior t o and during the period of DNA puffing there normally occurs an increase in the nuclear DNA content. This increase could not be detected in those larvae fed cortisone acetate. It must be stressed, however, that there is no evidence that RNA synthesis was affected by the treatment. In

46

M. ASHBURNER

fact the published autoradiographs give little indication of any difference in uridine incorporation over the DNA puff loci between the control and cortisone-fed animals. The conclusion of Goodman e t al. that “the significance of large chromosome puffs for normal development is questionable” cannot be justified by the data published. Crouse ( 1 968) has been unable to repeat the essential observations. She both injected and fed Sciara larvae with cortisone acetate and examined its effect on DNA puffing. Her results indicate that (a) the injection of cortisone had no effect on DNA puff induction, (b) that cortisone injected simultaneously with ecdysone did not modify the chromosomal response to ecdysone (see page 38), and (c) that DNA puff formation was normal in larvae fed cortisone acetate under either of two different conditions. In contrast to the experience of Goodman e t al., Crouse finds that at the concentrations used by the first authors cortisone does delay development by several days. Crouse concludes that the failure of Goodman e t al. to detect DNA puff formation after cortisone feeding is the result of this developmental delay so that the animals scored by them were too young for puff formation to have occurred. Similarly Rasch and Lewis (1968) fed cortisone to S. coprophila larvae and failed to observe any effect on DNA puff formation. Agreeing with the results of Crouse these authors did find that cortisone treatment resulted in delayed larval development. In D. melanogaster, Smith e t al. ( 1 968) and Sang ( 1 968) have failed to produce any modification of larval/prepupal puffing patterns following cortisone feeding. Both studies note that cortisone feeding results in delayed development. Similar findings were briefly reported by Midelfort and Rasch ( 1 968) working with D. virilis. C. JUVENILE HORMONE

Comparison of moult specific puffing patterns at larval/larval and larval/pupal moults has lead to the implication of juvenile hormone in the control of puffing activity (Becker, 1962a: Clever, 1964a; Ashburner, 1967a; Kroeger, 1968). Although many puffs are common to both moults there remain several whose activity is restricted t o one or other stages. For example in D. melanogaster Ashburner ( 1967a) found 30/87 puffs active at puparium formation (i.e. before the moult of the third instar larvae to the pharate fourth instar), and again before the pupal moult, 12/87 puffs active only at puparium formation and 13/87 only before the pupal moult.

PUFFS

47

Interestingly the relative sequence of activation of many of the puffs common to both processes differed before the pupal moult from that found a t puparium formation. Despite the frequent suggestion that these differences in puffing activity are d u e t o differences in juvenile hormone titre the results of experimental investigation of this point have been disappointing. Attempts t o modify puffing patterns by injection of juvenile hormone preparations (Clever, I963a), by the injection of juvenile hormone analogues (Ashburner, unpublished) o r by the implantation of corpora allata (e.g. Kroeger and Lezzi, 1966) were for a long time without success. Recent experiments briefly reported by Laufer and Holt ( 1 969) are likely t o lead t o a reappraisal of this situation. These authors fed prepupae of C. thummi with chlorinated farnesoic acid derivatives and, after prolonged treatment, observed both abnormalities of metamorphosis and effects on puffing. The effects on puffing were of two types: (a) an inhibition of the development of some late prepupal puffs and (b) the enlargement of a specific BR (BR IVb) which, during normal development, regresses in the late prepupal stage. No general return of the puffing pattern t o that characteristic of younger animals was observed (contra Kroeger, 1963a). The many negative results found previous to the report of Laufer and Holt were probably due t o inadequacies of experimental design. Many of the experiments used intermoult larvae, a period of low ecdysone activity. Since juvenile hormone is without effect, in the immature insect, in the absence of ecdysone it is of little surprise that n o effects of juvenile hormone could be detected. Significantly Laufer and Holt (1969) have found their effects in prepupaeanimals with a high ecdysone titre. It will be of great interest now t o know how juvenile hormone modifies the ecdysone response of the chromosomes when both compounds are injected into intermoult larvae. VIII. THE EXPERIMENTAL MODIFICATION OF PUFFING ACTIVITY

A large variety of experimental agents, other than those discussed previously, have been found t o influence puffing activity. Beermann (1952b) subjected larvae of C. tentans to cold treatment (5°C)for several hours and observed, after the return of the animals t o room temperature for 1 hr or more, the accumulation of droplets in the nucleoplasm adjacent to the puff sites. These droplets contain both RNA and protein (Beermann, 1959). Similar droplets have been found

Table I Puff induction by miscellaneous agents Agent

Species

Effect

Reference

Remarks

1. Ribonuclease

D. busckii

Specific puff induction after in vitro treatment of glands.

Ritossa and von Borstel, 1964; Ritossa e t al., 1964; Ritossa e t al., 1965.

2 . Thioacetamide

C. dorsalis

Specific puff induction after feeding.

Kiknadze and Filatova, 1963.

-

3. Tryptophan

D. melanogaster

Specific puff induction. I n vitro treatment of glands.

Federoff and Milkman, 1964. 1965.

Milkman (personal communication) and Ashburner (unpublished) unable t o confirm.

4. Dicyandiamide

D. melanogaster

General increase in puffing after in vitro incubation of glands.

Mukherjee, 1968.

5. Gibberellin A

Acricotopus lucidus

Specific puff regression after feeding.

Panitz, 1967.

Not confirmed by Prostakova (1968) with C. plumosus, by Paul (1965) with D. melanogaster or by Smith (1969) with D. virilis.

See page 65. Rasch and Lewis, 1968 did not observe any effect after feeding to Sciara coprophila.

6. Actinomycin D

Chironomus sp.

Specific induction of puffs.

Kiknadze, 1965; Serfling (personal communication).

See page 15.

7. Acridines (e.g. proflavin)

Chironomus thummi

Specific puff induction.

Serfling (personal communication).

See page 15.

8. Oxytetracycline

C. thummi

Specific puff induction and general puff enlargement after feeding.

Serfling et al., 1969.

About 30 puffs induced many heterozygous in different wild strains. C. strenzkei does not react.

9. I n vivo culture in adult abdomens.

D. melanogaster

Specific induction of puffs.

Staub. 1969.

Actual puffs induced dependent upon age of glands at transplantation.

10. I n vivo culture D. busckii in D. melanogaster egg cytoplasm.

Specific puff induction Kroeger, 1960. and regression in isolated salivary gland nuclei.

1 1. Various Ringertype solutions.

D. melanogaster

Specific puff induction after in vitro treatment of glands or after injection of larvae.

Becker, 1959; Ashburner (unpublished observations).

12. Various media

C. tentans

Puff induction.

Clever, 1965b.

13. Puromycin and cycloheximide

D. melanogaster

Specific puff induction after injection.

Ashburner and Mitchell (unpublished).

Exact pattern depends on age of donor glands and developmental stage of eggs used.

See also Clever, 1967, and page 39.

50

M. ASHBURNER

by Paul (1965) after cold treatment of D. melanogaster larvae and this author also concluded that the droplets have a high RNA content. From the published descriptions these droplets appear to resemble the micronucleoli of Sciarids (see p. 23) and are probably the result of a disturbance of the normal process of transport of RNA and protein from puff sites. Kiknadze (1965) reportedly induced a new puff in C. dorsalis as the result of low temperature treatment. The effect of heat shock on puffing is very much more dramatic than that of cold shock. The specific induction, and regression, of puffs due to a 12°C heat shock (25-37°C) was observed initially by Ritossa ( 1962b) in Drosophila busckii and has been studied subsequently by Berendes and Holt (1964), Berendes et al. (1965) and van Breugel (1966) in D. hydei, and by Ritossa (1963a), Ashburner (1969d) and Ellgaard (1969) in D. melanoguster. In D. melanoguster a continuous heat shock (2537°C) results in the induction of puffs at nine loci within 5-10 min. These puffs reach their maximum sizes rapidly and are active for 30-60 min and then regress. The shock also causes the regression of all puffs, apart from these nine, present at the time of transfer to the high temperature. Of the nine puffs induced four are at sites not seen to be active during normal development (Ashburner, 1969d). Similar changes in puffing activity also occur during recovery from anoxia (Ritossa, 1964a; Burdette and Anderson, 1965 ; van Breugel, 1966; Ashburner, 1969d). During the period of anoxia itself no changes in puffing take place, but after the return of the animals to air or oxygen specific induction at a limited number of chromosome sites begins within minutes. The puffs remain active for several hours after a 3 hr period of anoxia. Initially the induced puffs are at an identical spectrum of sites to those active during heat treatment (Ritossa, 1964a; van Breugel, 1966) but later additional puffs appear (Ashburner, 1969d), some of which are at loci not active during normal development. The pattern of puff induction by heat treatment is identical in viuo, in uitro (Ritossa, 1964a; Berendes and Holt, 1964; Ashburner, 1969d) and in larval salivary glands transplanted to adult abdomens (Berendes and Holt, 1964). The induction is not dependent on protein synthesis (Ashburner, 1969d; Ellgaard, 1969) but is inhibited by actinomycin D (Ritossa and Pulitzer, 1963, but see Berendes, 1968a and p. 19). The mechanism of induction of puffing by temperature treatment

PUFFS

51

or during recovery from anoxia is not understood. Experiments of Ritossa ( 1964a) are, however, suggestive. He induced the same puffs with heat shock and with a variety of chemicals whose common characteristic was their ability to uncouple oxidative phosphorylation (e.g. sodium azide, 24,dinitrophenol, dicoumarol). Since both heat shock and anoxia are expected to result in some disturbance of the respiratory metabolism the changes in gene activity may be a consequence of this effect. However, since such a disturbance would be expected to have manifold effects on general cellular metabolism the causal connection between it and the puff induction may be tenuous. In addition to heat shocks and anoxia a miscellaneous collection of agents have been found to produce qualitative changes in puffing activity but there is little indication of either the mechanism of action of any of these agents or any clue as to the functional significance of the observed changes. Some of these agents and their effects are listed in the accompanying table (see also Table IV of Kroeger and Lezzi, 1966). IX. MODIFICATION OF POLYTENE CHROMOSOME STRUCTURE AND FUNCTION AS THE RESULT OF INFECTION

The infection of Dipteran larvae by various organisms has frequently been found to have an effect on polytene chromosome morphology and activity. Most of our knowledge of the results of infection on giant chromosome structure and function is due to the researches of Pavan and his colleagues. Two types of infection of last instar larvae of Rhynchosciara angelae have been studied. Infection of larval tissues (e.g. mid-intestine) by a polyhedral virus results in considerable hypertrophy of both cells and polytene chromosomes. Although measurements of the DNA content of the nuclei of these cells is complicated by the presence of the viral DNA it is clear that the enlargement of the chromosomes has involved extensive DNA synthesis (Diaz and Pavan, 1965). Considerable degenerative changes occur in infected cells including chromosome fragmentation at specific loci and eventual pycnosis. The chromosomes of viral infected cells have few puffs (Pavan and Basile, 1966a, b; Pavan and da Cunha, 1969a). Microsporidia are responsible for a second type of infection which also results in cellular and chromosome hypertrophy (Diaz and

52

M. ASHBURNER

Pavan, 1965; Pavan and Basile, 1966a). The great increase in size of the polytene chromosomes occurring in infected cells is due t o extra rounds of DNA synthesis and these chromosomes have DNA contents that are 2” times the DNA content of normal chromosomes, the value of n varying from two to five (Roberts et al., 1967). Some control is clearly still exercised over these extra rounds of DNA replication since the proportion of the total nuclear DNA content accounted for by each individual chromosome remains constant. In striking contrast to the hypertrophied chromosomes in viral infected cells those in microsporidian infected cells are highly active with respect to RNA synthesis and have been said t o show “generalized puffing” (Roberts et al., 1967). This is especially striking on the X-chromosome of male larvae. Although a detailed analysis of the puffing activities of these chromosomes has not been published there is considerable specificity as to the site of the puffs and the chromosomes from different infected tissues (e.g. salivary gland, gut) show some differences in their puffing patterns (Pavan and Basile, 1966a). The DNA puffs normally active in Rhynchosciara prepupae are not seen in infected cells (Pavan and Perondini, quoted by Jurand et al., 1967; Pavan and da Cunha, 1969a, b). The effects of protozoan infection have been studied in other Sciarinae by Pavan and Perondini (1966) and by da Cunha e t al. ( 1968). In addition to polytene chromosome hypertrophy the infection of a sciarid Trichosia sp. by a gregarine eventually results in the complete fibrillation of the polytene chromosomes, a process recalling the chromosome behaviour of cecidomyiids (see page 6 ) . Jurand et al. (1967) examined the changes in ultrastructure in the salivary glands of Sciara ocellaris infected by Thelophania sp. (microsporidian). Widespread changes were found especially in the endoplasmic reticulum, its associated ribosomes and in the mitochondria. In addition to the examples studied by Pavan et al., chromosome hypertrophy resulting from infection has been described in Simuliids (Debaisieux and Gastaldi, 19 19), Chironomus melanotus (Keyl, 1960), and in Dasyneura urticae by Matuszewski (1964, see page 5). Severe abnormalities of polytene chromosome organization were found in individuals of Drosophila willistoni suffering from unidentified infections (da Cunha e t al., 1967). It has generally been accepted that the cellular and chromosomal hypertrophy in infected tissues results from an increase in gene activity which supplies the needs of the multiplying parasites

PUFFS

53

(Roberts et al., 1967). The molecular basis of these changes is unknown. However, it is probably of significance that at least one microsporidian, Nosema sp., which infects insects, is known to produce juvenile hormone or a substance with juvenile hormone activity. Infected animals fail to metamorphose or show derangements of metamorphosis such as the production of larval/pupal intermediates (Finlayson and Walters, 1957; Fisher and Sanborn, 1962). The possibility that the microsporidia infecting Pavan’s cultures also produce juvenile hormone should be investigated. It is interesting to note, however, that Pavan and Basile (1966a) have found that infected Rhynchosciara often “give rise to phenocopies, many of which d o not pass the beginning of metamorphosis”. The suggestion might be made that the observed changes in chromosome structure and function result from severe hormonal imbalance in the infected individuals, specifically from a high juvenile hormone titre. In this context it is relevant t o consider other examples of polytene chromosome hypertrophy which are not the result of infection. In Drosophila two different situations have been found to result in polytene chromosome hypertrophy. Firstly, several mutant stocks in which metamorphosis is blocked and the larval period greatly extended have chromosomes that are much shorter, but also much wider, than normal. Examples are the tu-h stock studied by Rodman and Kopac ( 1964) and Rodman (1964, 1967a) and the related lethal-tumerous (ZtZ) stock studied by Kobel and van Breugel (1 968). Rodman and Kopac (1 964) were inclined to suggest that the increase in chromosome diameter in tu-h larvae was due to “swelling” of the chromosomes but Rodman ( 1967a) later demonstrated that at least one extra round of DNA replication takes place in these larvae and that an increase in the degree of polyteny occurs. Rodman (1 964) claims that the puffing patterns of the tu-h larvae are abnormal but her data are not unequivocal. The polytene chromosomes of the It1 mutant of D. melanogaster have a similar morphology to those of the tu-h strain (Kobel and van Breugel, 1968). These authors note that shortened polytene chromosomes occur in other tissues of the It1 stock, e.g. Malpighian tubules and mid-gut. Larvae of D. hydei injected with 5-fluorouracil or larvae of the pci mutant stock of this species also have hypertrophied salivary gland chromosomes-both the analogue-treated larvae and the mutant stock have an extended larval development. Not all conditions which result in extended larval development or failure of metamorphosis result in chromosome hypertrophy. For example the A.1.P.-3

54

M. ASHBURNER

lgl mutant of D. melunoguster fails to pupate due to a deficiency in ecdysone (see page 37) and the larvae remain alive for up to 15 days. The polytene chromosomes of lgl do not show abnormalities of structure similar to those described (Ashburner, unpublished). The mutant giant of D. melunoguster also has an extended larval development yet normal salivary gland chromosomes and puffing patterns (Becker, 1959). The second situation that leads to polytene chromosome hypertrophy in Drosophilu is the culture of larval salivary glands in the abdomen of adult flies (Hadorn et ul., 1963). After prolonged culture (e.g. 20 to 50 days) many nuclei of the glands possess “supergiant” chromosomes and Hadorn e t ul. (1 964) demonstrated the occurrence of extra DNA replication under these conditions. These results were confirmed by Berendes and Holt (1965) who transplanted the larval salivary glands of D. hydei into adult hosts and studied patterns of DNA synthesis. Garcia-Bellido (1965) discovered that if the head region of a newly enclosed first instar larva of Drosophilu is implanted into an adult abdomen considerable growth and differentiation of larval tissues occurs including the development of polytene salivary gland chromosomes. Staub ( 1969) made a further study of this phenomenon and obtained successful polytene chromosome differentiation from the implantation of half embryos as early as 4 hr old. Many of the salivary gland nuclei developing under such conditions had supergiant chromosomes resembling those previously described. One common feature to all of these experiments, both those of nature and those performed in the laboratory, appears to be that the nuclei are developing in an abnormal hormonal environment. Circumstantial evidence indicates that, in particular, levels of juvenile hormone are high. Just how this hormonal imbalance produces its effect at the level of the chromosomes is a question that would repay study. X. THE PHYSIOLOGY OF NURSE CELL POLYTENE CHROMOSOMES

The ovaries of the higher Diptera are meroistic-that is each oocyte is provided with a number of nurse cells, usually IS. A single oocyte and its 15 nurse cells are the mitotic products of a particular oogonial cell. Due to incompleteness of cell divisions canals connect all of the nurse cells with the oocyte thus enabling the transfer of substances from the nurse cells to their oocyte (see King, 1964; Bier,

PUFFS

55

1967). In Drosophila the growth of the nurse cells during oogenesis is initially accompanied by the development of small polytene chromosomes. Later these chromosomes fall apart and the individual chromosomes replicate, presumably by endomitosis. In some mutants of D. melanogaster this process is disturbed and aberrant nurse cells with small (as in the mutant su2-Hw)or large polytene chromosomes (as in the mutant fes) occur (King et al., 1961; Klug et al., 1968). The polytene chromosomes of the fes mutant were studied by Schultz (1965) who noted that although their banding patterns differed in many details from those of the salivary gland chromosomes the individual chromosome arms were identifiable. On cytological grounds the X chromosome of fes nurse cell nuclei appears to be very active since, in contrast with the autosomes, it is very diffuse and is the major site of H3 -uridine incorporation. In Calliphora erythrocephala secondary polytenization of the endopolyploid nurse cell nuclei occurs spontaneously at low frequency during late growth stages (Bier, 1957). The frequency can be increased by low culture temperature, by inbreeding of the cultures and by the correct nutritional conditions (Bier, 1959). Secondary poly tene chromosomes have a well-defined banded structure, characteristic puffed regions and Balbiani rings (Bier, 1960). The banding pattern of the polytene chromosomes of nurse cells cannot, however, be homologized with that of the epidermal cell polytene chromosomes of the same species. The secondary nurse cell polytene chromosomes of Calliphora erythrocephala differ in one fundamental respect from the polytene chromosomes of larval tissues. Even after short periods of incorporation of H3-uridine into RNA the nurse cell chromosomes show a continuous pattern of labelling-the total length of the chromosome is covered with silver grains and puffed chromosome regions are not necessarily more heavily labelled than structurally unmodified regions (Bier et al., 1967; Ribbert and Bier, 1969). Furthermore, numerous irregular ribonucleoprotein aggregations adhere to several chromosome loci. Functionally the nurse cell polytene chromosomes bear a greater similarity to the lampbmsh chromosomes of oocytes of vertebrates (Callan, 1963) and of the oocytes of insects with panoistic ovaries (Kunz, 1967a, b) than to the polytene chromosomes of the larval and pupal tissues of Diptera. The difference in behaviour of the nucleoli of the polytene nuclei of these two cell types in Calliphora emphasizes this point. In

56

M. ASHBURNER

polytene nurse cell nuclei the nucleolus organizer fragments in some way to produce multiple nucleoli each containing DNA and synthesizing RNA. In the trichogen cells of pupae only a single nucleolus is visible. The proportion of the total nuclear RNA synthesized in the nucleoli of nurse cell nuclei is 72% and in trichogen nuclei 13%(Ribbert and Bier, 1969). In the polytene nurse cell nuclei, as in the oocytes of insects with panoistic ovaries, ribosomal RNA synthesis is a, if not the, major cellular function. XI. THE PHYSIOLOGY OF EPIDERMAL CELL POLYTENE CHROMOSOMES

The development of polytene chromosomes in epidermal cell nuclei during pupal differentiation was first described by Whitten (1 963) from Sarcophagn bullata. As Whitten has emphasized, the behaviour of these chromosomes is of particular importance since they occur in tissues whose function-that is the secretion of the adult cuticle and its appendages-is well known and to a certain extent understood. Furthermore, the epidermis itself is a “target organ” of ecdysone. The analysis of puffing patterns in the nuclei of the giant cells which secrete the dorsal cuticle of the pulvilli may prove to be very important for our understanding of the regulation and functional significance of puffing. The development of the foot pads (the pulvilli of the pretarsus) of Sarcophaga bullata has been described in detail by Whitten in a series of papers (Whitten, 1968, 1969a, b, c). A resume of the more important events is, however, essential for an understanding of the behaviour of the giant chromosomes. After eversion of the leg imaginal disc (on day 1 of pupal development) the foot is a sac of uniform epidermal cells filled with haemolymph and blood cells. These epidermal cells secrete a pupal cuticle from which they later retract on day 2. Cell divisions cease on the dorsal surface of the foot on day 2, yet continue on the ventral surface for a further day. Dorsally, on day 3, two groups of cells on each half of the foot can be distinguished by their size and by their migratory movements. The first group, of five cells, will form the pretarsal claws while the second group, of two cells, grow enormously through days 3 and 4 to cover the entire dorsal surface of the foot. The nuclei of the giant cells start to become polytene on day 2, and by day 4, when growth has ceased, they contain well-developed polytene chromosomes, with a 1024C DNA value (Roberts, 1968). Starting on day 5 secretion of

PUFFS

57

cuticle begins; on the dorsal surface of the foot pad by the two giant cells, and on the ventral surface by the many thousand small “tenant” cells. At least five layers of cuticle are synthesized by day 7. They are, in order of laying down (1) ecdysial membrane (day 5), (2) cuticulin (day 6), (3) dense exocuticle (days 6-7), (4) homogeneous exocuticle (day 7), and (5) “mesocuticle” (day 7). On day 9 histolysis of the tenant cells begins, aided by the phagocytic activity of the haemocytes. The tenant cells have disappeared by days 1O/1 1 and many thousands of hairs are all that remain witness to their activity. On day 11 endocuticle deposition and sclerotization commences and the giant cells begin to be histolysed. By emergence on day 11-12 the giant cells have gone and the adult pulvillus with its ridged dorsal cuticle and the many thousand tenant hairs on its ventral and lateral surfaces is completed (Whitten, 1968, 1969a, c). Growth of the flat dorsal giant cells and polytenization of their chromosomes is essentially complete by day 4 although further cycles of replication d o occur late in development (Roberts, 1968). From the time that cytological analysis of the chromosomes is possible many puffs are visible in these chromosomes and throughout days 4 t o 11 a large number of puffs appear, regress and in some cases reappear in a sequence that is highly co-ordinated with such events as the phases of cuticle deposition and cell histolysis. The majority of the 80 or so puffs studied by Whitten (1969d) would appear to be typical RNA puffs. The number of active sites increases to a maximum at day 5-that is just prior to the maximum phase of cuticle synthesis-and then decreases. The nucleolus is also largest on day 5 (Clever et ul., 1969a). From the diagrams published by Whitten (1969d, Figs 15-19) it is clear that the number of puffs active on day 5 is about three times the number active from days 9 to 11. However, as Whitten points out even during histolysis of the giant cells many puffs continue to be active, and new puffs appear. This situation resembles that described by Ashburner ( 1967a) during the breakdown of the salivary gland of Drosophilu melunoguster. Similar observations on puffing patterns in foot pad nuclei of S. bullutu have been made by Bultmann (Clever et ul., 1969a). Over 600 puffs have been recorded by Bultmann, the activity of each changing in a regular manner during the various stages of foot pad development. It is of great interest that although the puffing pattern of each of the 24 foot pad nuclei of an individual are identical they are not exactly synchronous-the cells of the hind legs lagging behind the cells of the front legs and the outer cells of each foot pad lagging

58

M. ASHBURNER

behind the inner ones. This observation may be of significance to speculation concerning the endocrine control of puffing (see p. 40). In addition t o the RNA puffs typical of polytene chromosomes the foot pad chromosomes also show a phenomenon related to DNA puffing (see Section V(H), p. 23) (Whitten, 1965). Although obvious DNA puffs are relatively few, disproportionate DNA synthesis, at certain loci, is common and is characterized by the budding off, from these chromosome regions, of DNA granules (Whitten, 1965). A close association of these granules and the nuclear membrane has been noted (Whitten, 1965; Roberts, 1968). DNA granule production occurs at a limited number of chromosome loci and the frequency of granules at each site is subject to variation correlated with development. The granules themselves appear to have a DNA core surrounded by a matrix of ribonucleoprotein. They incorporate both H3 -thymidine and H3 -uridine, although it is not clear whether either DNA or RNA can be synthesized by the granules after their detachment from the chromosomes. The extra DNA synthesis involved in the production of the granules results in a skewing of the DNA values measured from foot pad giant cell nuclei (Roberts, 1968). Puffing patterns in the foot pad giant cell nuclei of S. bullutu can be compared with puffing patterns of the polytene chromosomes of the bristle initial cells. The trichogen and tormogen cells are also modified epidermal cells concerned with the synthesis and secretion of cuticle. In S. bullutu preliminary observations (Whitten, 1969d) indicate that the banding patterns of the chromosomes of foot pad and trichogen cells are identical and that many puffs are active in both tissues. The similarity of puffing patterns of foot pad and trichogen chromosomes is especially evident during the early stages of differentiation. Later, however, many puffs are active in the foot pad that are absent from the trichogen nuclei. In particular many puffs active in the foot pad cells during their histolysis are not active in the trichogen nuclei. Significantly the trichogen cells d o not histolyse on emergence of the adult (Whitten, 1969d) and the differences in puffing activity may well reflect the differences in developmental fate of the two cell types. The polytene chromosomes of the thoracic trichogen cells of Culliphoru erythrocephula have been extensively studied by Ribbert ( 1967); Ribbert and Bier ( 1969). The polytene chromosomes of the tormogen (bristle socket) cells are, in this species, not very suitable for detailed analysis. In C. erythrocephala polytenization of

PUFFS

59

trichogen chromosomes accompanies growth of the cell and its nucleus; final DNA contents are in the 2048C range. Maximum nuclear size is accompanied by maximum puffing activity of these chromosomes. A number of puffs have been followed and they exhibit considerable developmental specificity. All of these puffs appear to be RNA puffs. Neither DNA puffs nor DNA granule formation were described by Ribbert ( 1 967). In general the frequency of active chromosome sites correlated with the changes in ecdysone titre as measured by Shaaya and Karlson (1965). Towards the end of imaginal differentiation the puffing frequency, and the ecdysone titre, decrease. Several puffs are, however, characteristic of these late stages in development. The results of experimental induction of puffing in these chromosomes by hormones have not yet been published. XII. THE PHYSIOLOGICAL AND FUNCTIONAL SIGNIFICANCE OF PUFFING A. SALIVARY GLAND FUNCTION

Despite its name the salivary gland of larval Diptera is not usually regarded as playing a major role in digestion. At least in Drosophila recent studies have shown that digestion occurs in the mid-intestine and that many enzymes characteristic of digestion are virtually absent from the larval salivary gland [e.g. amylases (Doane, 1969), alkaline DNAse (Boyd and Boyd, 1969), and trypsin-like proteases (Waldner-Steifelmeier, 1967)], but are very active in the midintestine. Similar findings were reported for the salivary gland of Acricotopus lucidus by Baudisch ( 1 963a). Many histological studies of the salivary gland of Drosophila larvae have noted the accumulation in the cytoplasm of large granules or vacuoles during the period of the late third instar. These have been given various names by, for example Painter (1946, secretory globules), Hsu (1 948, food vacuoles), Lesher (195 1, 1952, deuteroplasmic substances) and assigned an equally wide variety of functions, e.g. digestive, food storage, as a precursor of chitin or as evidence of histolytic changes. The true nature and significance of the secretion of late larval salivary glands of Drosophila was only realized by Fraenkel and Brookes ( 1 953). They demonstrated that the secretion is stored in the lumen of the salivary gland until puparium formation when it is expectorated and serves as a glue to affix the puparium to its substrate. These authors also noted that the salivary

60

M. ASHBURNER

glands of cyclorraphous Diptera which d o not fix their puparium in this way d o not show the massive accumulation of secretion prior to puparium formation characteristic of, for example, Drosophila. Although suitable ultrastructural studies have not been carried out it is clear that during the last half of the third larval instar a secretion accumulates in large membrane bound cytoplasmic vesicles which migrate to the luminal cell border and discharge their contents into the gland lumen (Gay, 1956; Swift, 1962; Berendes, 1965a; Rizki, 1967). The light microscope study of Hsu (1948) implicated the Golgi apparatus in the formation of these vesicles in common with the secretory process of many other cell types (see below). Histochemical studies of the cytoplasmic granules demonstrated that, like the glue, they were PAS positive (Lesher, 1952; Swift, 1962; Berendes, 1965a) and it was suggested that the material of the granules was neutral mucopolysaccharide (Lesher, 1952) or mucoprotein (Berendes, 1965a) in nature. A biochemical study of the secretion by Kodani (1948) characterized the glue as a mucoprotein; Kodani detected several amino acids in hydrolysates of the secretion and in addition the amino sugar glucosamine. Recent studies (Ashburner and Blumenthal, unpublished) have confirmed much of Kodani’s analysis. The glue of D. rnelunogaster is a mucoprotein or complex of mucoproteins with a composition 70% amino acids and 30% sugars. The amino acid analysis is notable in the very high proportion of serine, threonine, glutamate, aspartate and proline and the low amounts of cyclic and sulphur amino acids. It is interesting that serine, threonine and aspartate are the three major amino acids involved in the linkage of the carbohydrate moeties to the protein backbone of mucoproteins (Gottschalk, 1966). A complex mixture of sugars was detected including several hexoses, aminohexoses, N-acetyl aminohexoses and hexuronic acids. A partial biochemical and immunological analysis of the glue of D. virilis was published by Perkowska (1 963a). The salivary glands of many other families of Diptera also appear to be largely involved in the production of one major secretion. The function of this secretion is varied: tube building and feeding (e.g. Chironomidae), slime and coccoon formation (e.g. Mycetophilidae, Cecidomyiidae) or as a “life line” (e.g. Simuliidae). Very little information is available concerning the compositions of these secretions, with the exception of the work of Laufer and his group which will be discussed below (page 66) (see Kato and Sirlin, 1963; Kato e t al., 1963; Yoshimatsu and Uehawa, 1968).

PUFFS

61

Electron microscopic studies of salivary gland structure have been published for several species: Simulium niditifrons (Macgregor and Mackie, 1967), Brudysia mycorum (Sciarinae) (Jacob and Jurand, 1963), Sciuru coprophilu (Phillips and Swift, 1965), S. ocelluris (Jurand e t ul., 1967), Chironomus tentuns (Schin and Clever, 1968a), C. thummi (Kloetzel and Laufer, 1968), C. dorsulis (Yoshimatsu, 1962) and Smittiu purthenogenetica (Chironomidae) (Jacob and Jurand, 1965). A full account of these studies will not be given in this review since there has been little correlation of the ultrastructural observations with the physiology of the gland itself. However in all species examined a variety of “secretory granules” have been figured and the Golgi apparatus implicated in their genesis. Typically these granules release their contents into the gland lumen, presumably through reverse pinocytosis. The lumenal border of the gland cells is usually modified as a brush border with numerous microvilli. The salivary gland of Sciuru coprophilu, studied by Phillips and Swift (1965) is of interest in possessing a wide variety of different secretory granules each with a corresponding type of Golgi apparatus. Marked changes in the frequency of two of the three main types of granule were found to occur before pupation. The electron-lucid granules, common in early fourth instar larvae, disappear during the prepupal period, while the electron dense granules are only found in the late fourth instar larvae and prepupae. Phillips and Swift suggest that the electron dense granules are the precursors of the acid mucoprotein used for coccoon formation in this species. In contrast to the situation described for Sciuru only one type of secretory granule was seen in the salivary glands of the larvae of Simulium niditifrons by Macgregor and Mackie ( 1967). In addition to the major secretory products of the larval salivary gland a wide variety of enzymes of the salivary glands have been studied and changes in enzyme activity have often been found to be correlated with development. Patterson e t ul. ( 1949a, b) were the first to study salivary gland enzymes in Drosophilu melanoguster. In particular they followed the changes in activity of various peptidases. A peptidase which hydrolysed alanyl-glycine with a pH optimum of pH 7.65 was found to increase in activity throughout the first 6 hr of the prepupal period and then to decrease. The function of this enzyme is obscure since its optimal pH is not that expected for an enzyme concerned with salivary gland histolysis (see below). Laufer and his colleagues have followed the activities of eight enzymes in the salivary gland of Chironomus thummi. Three A.1.P.-3

*

62

M. ASHBURNER

enzymes (RNAse, “esterase” and “phosphatase”) showed n o change in specific activity during the development of fourth instar larvae and prepupae. The activities of malate dehydrogenase, trehalase, hyaluronidase, and a protease (see Rodems e t al., 1969) all doubled from mid-fourth instar larvae to early prepupae and then decreased in the late prepupae (Laufer, 1963, 1965). These increases in activity were sensitive to both inhibitors of protein synthesis (puromycin, Laufer, 1965) and inhibitors of RNA synthesis (actinomycin D, Laufer et al., 1964). In addition to being present in the salivary gland the trehalase, protease and hyaluronidase activities have also been detected by Laufer in the salivary gland secretion itself (Laufer and Nakase, 1965a). In the salivary gland of Sciara coprophila the activity of glucose-6-phosphatase shows complex changes with the development of the larvae (Terner e t al., 1965, 1967; Goodman e t al.. 1968). Three peaks of activity, in 12-, 16- and 19day-old animals, have been described and these authors have studied the effect of cortisone feeding on this activity. Apparently the different peaks are affected in dissimilar ways; the 13day peak is unaffected, the 16day peak increased and the 19day peak reduced (Goodman et al., 1968). Parallel changes in the activity of RNA polymerase were also reported (Terner et al., 1967). In the salivary glands of Drosophila changes in the distribution of beta-galactosidase were followed histochemically by Ritossa ( 1963b). Present in all cells of young third instar larvae, activity disappeared from the distal cells in older larvae only to reappear in these cells just before pupation (sic). The study of DNA% activity in salivary glands of C. thummi by Laufer is of great interest as it highlights a function of the salivary gland which has yet to be discussed; i.e. its histolysis. In common with most larval tissues the dipteran salivary gland is completely histolysed shortly after pupation although both biochemical and cytological changes preparative for this event can be detected many hours earlier. In C. thummi DNAse (pH optimum 7.0 with native DNA as substrate) activity increases seven-fold in late prepupae (Laufer and Nakase, 1965b). The activity of all other enzymes studied decreases at this time. Following the increase in DNAse activity the DNA content of the individual salivary gland nuclei decreases (Laufer et al., 1968). Laufer draws the reasonable conclusion from these data that the increase in DNAse activity at this time is intimately concerned with the process of salivary gland

PUFFS

63

histolysis. However the changes in DNAse activity are probably very much more complex than described. This is apparent from a detailed analysis of the DNAse of Drosophila by Boyd ( 1969) and Boyd and Boyd (1969). Larval salivary glands of D. hydei have one major DNAse activity; the enzyme has a low pH optimum ( p H 4 ) and migrates on electrophoresis with a Rm of 0.64. The prepupal salivary glands, on the other hand, have five bands of DNAse activity with acid pH optima but differing in their electrophoretic mobilities and other characteristics. The prepupal salivary gland also has a sixth DNAse with an alkaline pH optimum. It is of great interest, and this point will be discussed below, that very similar changes in DNAse activity, after puparium formation, are found even in tissues that do not undergo histolysis, e.g. Malpighian tubules (Boyd and Boyd, 1969). The role of the lysosomes in tissue destruction has been reviewed by de Duve (1963) and their importance in tissue histolysis in insects emphasized by the studies of Lockschin and Williams ( 1965a, b) and of Osinchack (1966). Misch (1962) was the first author to note an increase in the activity of lysosomes, as visualized by the standard Gomori reaction for acid phosphatase, prior t o metamorphosis in the salivary glands of Sarcophaga. Later Rasch and Gawlik ( 1964) noted a similar phenomenon in the salivary glands of Sciara coprophila prepupae and further found two types of “acid phosphatase” granule in an electron microscope study. Jacob and Jurand (1965) described lysosome-like bodies in the gland of Smittia parthenogenetica. The only comprehensive account of salivary gland histolysis is, however, that of Schin and Clever (1965, 1968a) for Chironomus tentans. A variety of cytoplasmic bodies showed a positive acid phosphatase reaction, but two main classes were distinguished (see Rasch and Gawlik, 1964). The larger bodies (interpreted as lysosomes) were 0.5 p to 1 p in diameter and were electron lucid. They were bounded by a single membrane and were not sensitive to cycloheximide. The lysosomes were found mainly in the basal region of the cells (an observation confirmed by Rasch and Gawlik, 1964). The small bodies, on the other hand, were sensitive to cycloheximide, were electron dense and were frequently associated with the Golgi complex. They were considered, by Schin and Clever, to be a precursor of the secretory material. The frequency of lysosomes, per unit area of cytoplasm, was found to change dramatically with development. Rare in third instar or young fourth instar larvae they became more frequent as the time of pupation approached and were

64

M. ASHBURNER

most frequent just prior to the start of cytolysis. The actual breakdown of the salivary gland cells in young pupae involves not only the lysosomes but- also other cell organelles. In histolysing glands considerable free acid phosphatase activity was found. Schin and Clever suggest that the lysosomes may be functional in the salivary glands of younger animals also, for example they may be involved in such activities as the uptake of foreign proteins (see Schin and Clever, 1968b and below). A third enzyme, a protease with an optimal pH of 3.5, probably involved in tissue destruction has been studied in the salivary glands of C. tentans by Rodems et al. ( 1969). During prepupal development the specific activity of this enzyme increases three-fold. B. SPECIFIC CORRELATIONS OF PUFFS AND SALIVARY GLAND FUNCTION

The correlation of the activities of specific puffs with particular salivary gland functions has only been thoroughly analysed in three instances. The first example is in the sibling species Chironomus palliduvitatusltentans where the production of the special secretory granules of the “Sonderzellen” of C. palliduvitatus is dependent upon the activity, in these cells, of a fourth chromosome Balbiani ring “BR sz” (Beermann, 196 1). Extensive cytogenetic investigation of the BR sz and special secretion production in the Sonderzellen by Beermann established that the genetic difference responsible for variation in the ability to make the secretion and the locus of the BR sz mapped coincidentally near the end of chromosome IV (see page 29). The BR sz is not active in the normal cells of the salivary gland of C. palliduvitatus nor in any of the cells of the gland of C. tentans. These cell types lack the special secretion. Comparison of the proteins of the normal cells and the Sonderzellen of the palliduvitatus salivary gland has shown a unique protein fraction in the Sonderzellen (Grossbach quoted by Beermann, 1967b). Biochemical analysis of the secretion of the salivary glands of C. palliduvitatus and C. tentans has further implicated the BR in controlling the nature of the main gland secretion (Grossbach, 1968). Electrophoresis of the secretion of C. tentans salivary glands separated six main protein components each of high molecular weight (above 5 x l o 5 daltons). These components can be dissociated, by the reduction of S-S bonds and by alkylation, into a greater number of subunits each of which is probably equivalent to a single polypeptide.

PUFFS

65

The electrophoretic pattern of reduced proteins from the main gland secretion of C. tentans and C. palliduvitatus is similar except that an extra protein band is found in the secretion of the latter species. The presence of this protein is dependent upon the presence in the salivary gland nuclei of at least one palliduvitatus fourth chromosome and preliminary evidence suggests strongly that only the distal end of this chromosome-which bears a Balbiani ring (BR3)-is necessary for the production of this particular component of the secretion. The structure of the salivary gland of Acricotopus lucidus and the intragland variation in Balbiani ring formation in this species has been discussed in a previous section of this review (page 29). Of the three lobes of the salivary gland of this species two, the main lobe and the side lobe, are characterized by the presence of hydroxyproline detected after hydrolysis and chromatography. This amino acid is not found in hydrolysates of the anterior lobe nor is it present in other organs of the larvae (Baudisch, 1960, 1963b) nor in the food (Baudisch, 1964). Salivary glands take up C14 labelled proline from the food (Baudisch, 1964) and will convert labelled proline into hydroxyproline during in vitro incubation (Baudisch, 1967). The two lobes also to convert proline to hydroxyproline differ from the anterior lobe in the possession of the two Balbiani rings BRl and BR2. Panitz (1967) discovered that the plant hormones gibberellin A3 and gibberellin A4, on addition to the culture medium of fourth instar larvae, caused regression of both BR1 and BR2 but not of BR3 and BR4 of the anterior lobe. At the concentrations of gibberellin used ( 1500 pg/ml) the effect took many hours to become apparent. The regression of these BR due to gibberellin treatment was paralleled by a reduction of H3 -uridine incorporation into RNA at these loci. The two Balbiani rings differed in their response to the gibberellins, BR2 being considerably more sensitive. Baudisch and Panitz ( 1968) studied in vitro proline hydroxylation in salivary glands from animals pretreated for 48 hr with gibberellin A3. The BR2 was inhibited in the main and side lobes of these glands. In contrast to salivary glands from untreated larvae those from gibberellin treated larvae were unable to convert proline to hydroxyproline. Since proline hydroxylation occurs after the incorporation of proline into polypeptides these authors conclude that BR2 is involved in the control of production of a secretory protein containing hydroxyproline residues (see Panitz, 1968). These two examples are the only well studied correlations of

66

M. ASHBURNER

particular salivary gland function and the activity at particular puff loci. In neither case has the causal relationship between the activity of the puff and the presence in the secretion of a particular protein been rigorously proved, although the published data are very suggestive that such causal relationships do exist. In addition t o these examples Berendes ( 1965a) has demonstrated a correlation between the activity of the puff 47B in Drosophila hydei and the presence in the cells of PAS positive secretion granules. The discovery that the salivary gland secretion of Chironomus thummi possessed several enzyme activities and the failure to find any antigens specific to either salivary glands or their secretion led Laufer (1965) and Laufer and Nakase (1964, 1965a) t o suggest that the salivary gland does not synthesize its secretion itself but acts as a transport organ sequestering the components of the secretion from the haemolymph and elaborating them into the final secretory product. The Balbiani rings and other puffs were, it was proposed, not coding for the synthesis of the secretion but for the “permeases” necessary for the uptake by the gland of macromolecules. Six enzymes (hyaluronidase, DNAse, RNAse, trehalase, esterase, and protease) were detected in the salivary gland secretion, in the salivary gland itself, and in the haemolymph (Laufer and Nakase, 1965a) (recently Rodems et al., 1969 have been unable to detect the presence of any significant protease activity in C. tentuns haemolymph). The six antigens detected in the secretion were also common to the haemolymph although later a secretion specific antigen was discovered (Laufer, 1968). Electrophoresis of salivary gland proteins led to the separation of 11 major bands of which 5 appear to be haemoglobin proteins. All of these are also found in the haemolymph although their relative proportions differ (Doyle and Laufer, 1968). Furthermore C14 labelled haemolymph proteins injected into larvae were secreted by the salivary glands. Human albumin protein injected into larvae was also detectable in the secretion with its antigenic properties conserved. Uptake of macromolecules by the salivary gland does not depend on either RNA or protein synthesis (Doyle and Laufer, 1969b). In later papers Doyle and Laufer (1968, 1969a) modify their position and demonstrate that some proportion of the salivary gland secretion is synthesized in the gland itself; the major portion however appears to be synthesized in other tissues. Ultrastructural studies of the salivary gland by Laufer and Goldsmith (1965) and by Kloetzel and Laufer (1968) have been used

PUFFS

67

to support the hypothesis that the gland is a transport organ. The mechanism of transport of macromolecules into the gland is claimed to be a pynocytosis-like process occurring along the haemal cell borders, the substances being sequestered in the salivary gland cytoplasm in small membrane bound vesicles. Macgregor and Mackie (1967) have challenged this interpretation of the structure of the gland claiming that similar specializations of the plasmalemma are to be found in tissues that are manifestly synthesizing proteins for export. Protein tracer molecules (e.g. ferritin, peroxidase), but not non-proteinaceous substances such as colloidal gold, are also taken up by a similar process and are to be found, within the cell, in a variety of types of vesicles including lysosomes and the Golgi apparatus (Kloetzel and Laufer, 1968) and eventually the secretion itself (Laufer, 1965). Similar results were reported by Schin and Clever (1 96813) although these authors failed to find any evidence that ferritin was ever associated with the Golgi or secreted into the gland lumen. Indeed the amount of ferritin found to be taken up by the salivary gland was small compared to the amount taken up by the larval fat body. Pertinent to Laufer’s hypothesis are the recent studies of Wobus (1969). Wobus has followed the changing patterns of proteins with development in the haemolymph and salivary gland of various chironomids by the technique of disc electrophoresis. While the proteins of the haemolymph showed considerable, and characteristic, changes with development of the fourth instar larvae and prepupae no parallel changes could be observed in the proteins of the salivary glands. There is no indication that Laufer’s hypothesis can be generalized to other groups of Diptera. Indeed the significance of the ability of the Chironornus salivary gland to take up macromolecules from the haemolymph may perhaps be seen from a consideration of the feeding biology of Chironomus larvae. In species such as C. tentans and C. thumrni the larvae inhabit tubes in mud etc.; these tubes are coated with salivary gland secretion. To feed the larvae spin a net of secretion across the mouth of their tube and use this to trap their food. Water is drawn across the net by movements of the larvae. The larvae then proceed to eat the net and the entrapped food (Walshe, 1947). At least under laboratory conditions this process is repeated with extreme rapidity and the possibility arises that some of the secretion can be recycled without degradation through the gut and back to the salivary gland for re-use. The gland may lack the suitable

68

M. ASHBURNER

specificity to avoid taking up other haemolymph proteins as well. Otherwise it is not clear why minor haemoglobins should accumulate in the gland, or why the gland should take up completely foreign proteins such as human albumin. There is no evidence at all that the Balbiani rings and the puffs are directly involved in this process of macromolecule uptake by the salivary glands. C. GENERAL CONSIDERATIONS ON THE PHYSIOLOGICAL FUNCTION O F PUFFS

There is no need to attempt a general hypothesis concerning the function of puffs. Doubtless puffs are concerned with all of the manifold aspects of cellular and tissue function. There are, however, some aspects of the developmental physiology of puffs which require closer examination. In particular the very striking increases in puffing activity that occur during the moults and prior to metamorphosis. In many species a major secretory product is synthesized by the salivary gland at the end of larval development and functions, for example as the puparial glue in Drosophila or as coccoon material in many Nematocera. But in Drosophila considerable puffing activity continues during the prepupal period, after release from the gland of the puparial glue, and many puffs active at this time were also active, earlier in development, in third instar larvae. In D. rnelanogaster, and probably other species, the situation is even more striking since regular changes in puffing pattern occur after the initiation of salivary gland histolysis (Bodenstein, 1943). Indeed the salivary glands of animals just prior to pupation are in an advanced stage of histolysis with many pycnotic nuclei. Yet the functional nuclei of these glands have a characteristic puffing pattern (see discussion in Laufer, 1968). What is the function of this gene activity? I t was suggested (Ashburner, 1967a) that during the prepupal period at least the salivary gland, and other larval tissues, synthesize proteins and other materials of n o particular functional significance for the tissues themselves. These materials were considered t o be exported from the tissues and to be utilized by the imaginal tissues during their histogenesis. At that time this idea was put forward largely as the result of the studies of Mitchell (1966) and Geiger and Mitchell (1966) on the S protein component of the Drosophila phenol oxidase enzyme. This protein, essential for phenol oxidase activity, could only be detected in the salivary glands of young and old prepupae; in the salivary glands from midprepupal animals no activity was found. The changes in S protein activity with time

PUFFS

69

closely resembled the changes in activity of a large group of moult specific puffs. No other tissue was found to possess S protein activity and it is difficult t o avoid the conclusion that this protein must be exported from the gland since by far the most important site of action of phenol oxidase is the cuticle. Support for this idea is gained from the studies of Whitten (1 964b, 1969a, c) who has emphasized the role of the haemocytes in the transfer of materials from histolysing larval tissues to the developing imaginal tissues. During histolysis the granular haemocytes send out cytoplasmic processes which attach to and ingest fragments of histolysing larval tissues. These granular haemocytes subsequently change their form to the spherules which are often closely associated with differentiating imaginal tissues (Whitten, 1964b). The very close association of the spherules with the development of the leg, and evidence for the transfer of materials from the blood cells, has been described by Whitten (1969a). The fact that the puffing patterns of the Malpighian tubules of Drosophilu were very similar to those of the salivary glands provided an objection to these general ideas since the Malpighian tubules do not histolyse at metamorphosis. However considerable reorganization of these structures does take place after puparium formation. Boyd and Boyd (1969) have shown that in one respect at least, that is changes in the spectrum of DNAses, the behaviour of the Malpighian tubules after puparium formation closely resembles that of the salivary glands. For example larval salivary glands and larval Malpighian tubules have identical DNAse patterns-that is a single enzyme with an acid pH optimum. After puparium formation both tissues possess at least four enzymes with acid pH optima. These enzymes differ in their electrophoretic mobility and in other ways. Rasch and Gawlik (1964) also find an increase in acid phosphatase granules in the Malpighian tubules of Sciuru coprophilu prepupae.

XIII. CONCLUSIONS AND OUTLOOK

In addition to summarizing the extant literature I have attempted, in this review, t o highlight some of the more important lacunae in our knowledge of polytene chromosome puffs, their physiology and biochemistry. While there can be little doubt that Beermann’s hypothesis “that puffs represent active gene loci” is essentially correct and that puffs are regions of RNA synthesis, the molecular

70

M. ASHBURNER

basis of puff information is still obscure. Several processes are presumably involved, (i) the recognition of specific chromosomal sites by controlling elements resulting in (ii) a change in the tertiary structure of the DNA-protein complex of the chromosome such that (iii) specific RNA synthesis can now take place, (iv) the control of the transport of the RNA from the puff site perhaps involving some change in the molecular structure of the RNA itself, (v) the cessation of RNA synthesis as the result of specific signals and (vi) the return of the DNA-protein complex to the pre-active state. The precise molecular events involved in these processes, and even, in some cases, the relative order of the events, is unknown. New techniques are continually being involved for an attack on these problems; two may, perhaps, be mentioned. First methods for the mass isolation of salivary glands, their nuclei and their polytene chromosomes from Dipteran larvae (Sirlin and Schor, 1962; Karlson and Loffler, 1962; Ristow and Arends, 1968; Boyd e t ul., 1968; Boyd and Berendes, 1968; Berendes and Boyd, 1969; Cohen and Gotchel, 1969). The availability of nuclei and chromosomes in quantities sufficient for biochemistry can only lead to a furthering of our knowledge of the puffs. The second technique, the in vitro hybridization of RNA to polytene chromosome DNA, now under active investigation in several laboratories, may well answer many of our queries concerning the nature and behaviour of the RNA synthesized at the puff loci. However neither of these approaches is likely fully to satisfy our curiosity as t o the nature of the control systems, and their integration, which result in the very precise co-ordination of activities of many individual puff loci during development. Another approach, used with success in other “systems”, involves the genetic analysis of the control of puffing activity. This has already met with some success in Chironomus (Beermann, 196 1 ; Grossbach, 1968), Drosophilu (Ashburner, 1967b, 1969b, c, e; Rayle, 1967) and Sciuru (Perondini and Dessen, 1969), with the identificatibn of various mutants lacking individual puffs and an analysis of their genetic behaviour. The importance of understanding the phenomenon of puffing in all its aspects lies in the general applicability of the conclusions. There are no reasons to believe that puffs are in any way exceptional phenomena except for the fact that they are studied in chromosomes of a high degree of polyteny. Polyteny itself is probably more related to the functional problems in producing a large amount of secretion

PUFFS

71

for export by an organ with a relatively small number of cells than anything else. Were they visible by standard microscopical techniques doubtless all active gene loci of interphase chromosomes would show structural modifications directly comparable with the puffs of polytene chromosomes.

REFERENCES Abramyan, K. S . and Reingol’d, W.N. (1965). Electron microscopic study of the interrelations of the nucleolus and cytoplasm. Dokl. Akad. Nauk. S.S.S.R. 161, 160-163 (English Translation Edition (1966)). Alanso, P. and PBrez-Silva, J. (1965). Giant chromosomes in protozoa. Nature, Lond. 205, 313-314. Alfert, M. (1954). Composition and structure of giant chromosomes. Int. Rev. Cytol. 3, 131-175. Allfrey, V. G., Pogo, B. G. T., Littau, V. C., Gershey, E. C. and Mirsky, A. E. (1968). Histone acetylation in insect chromosomes. Science, N . Y. 159, 3 14-3 16. Allfrey, V. G., Pogo, B. G. T., Pogo, A. O., Kleinsmith, L. J. and Mirsky, A. E. ( 1966). The metabolic behaviour of chromatin. In “Histones: Their Role in the Transfer of Genetic Information” (A. V. S . de Reuck and J. Knight, eds). J. & A. Churchill, London. Ammermann, D. (1965). Cytologische und genetische Untersuchungen an dem Ciliaten Stylonychia mytilus Ehrenberg. Arch. Protistenk. 108, 109-152. Arnold, G. (1965). An autoradiographic study of RNA synthesis in isolated salivary glands of Drosophila hydei. I. Autoradiographic studies. J. Morph. 116, 65-88. Ashburner, M. (1967a). Patterns of puffing activity in the salivary gland chromosomes of Drosophila. I. Autosomal puffing patterns in a laboratory stock of D. melanogaster. Chromosoma 21,398428. Ashburner, M. (1967b). Gene activity dependent on chromosome synapsis in polytene chromosomes of Drosophila melanogaster. Nature, Lond. 214, 1159-1 160. Ashburner M. (1969a). Patterns of puffing activity in the salivary gland chromosomes of Drosophila. 11. X chromosome puffing patterns in D. melanogaster and D. simulans. Chromosoma 2 7 , 4 7 4 3 . Ashburner, M. (1969b). Patterns of puffing activity in the salivary gland chromosomes of Drosophila. 111. A comparison of the puffing patterns of the sibling species D . melanogaster and D. simulans. Chromosoma 27, 64-8 5 . Ashburner, M. ( 1 9 6 9 ~ ) . Patterns of puffing activity in the salivary gland chromosomes of Drosophila. IV. Variability of puffing patterns. Chromosoma 27, 156-177. Ashburner, M. (1969d). Patterns of puffing activity in the salivary gland chromosomes of Drosophila. V. The responses to environmental shocks. (In preparation).

72

M. ASHBURNER

Ashburner, M. (1969e). The genetic control of puffing in polytene chromosomes. In “Chromosomes Today” (C. D. Darlington and K. R. Lewis, eds). Vol. 2, pp. 99-106. Oliver and Boyd, Edinburgh. Ashburner, M. (1969f). Ecdysone induction of puffing in the salivary gland chromosomes of Drosophila melanogaster. (In preparation.) Ashburner, M. and Henderson, S. A. (1969). Chromosome organisation and behaviour in the salivary glands of larval Cecidomyiidae (Diptera). (In preparation.) Bahr, G. F. and Beermann, W. (1954). The fine structure of the nuclear membrane in the larval salivary gland and midgut of Chironomus. Expl. Cell. Res. 6, 5 19-522. Balbiani E. G. (1881). Sur la structure du noyau des cellules salivaires chez les larves de Chironomus. 2001.Anz. IV, 637-641 and 662-666. Barigozzi, C. and Semenza, L. (1952). A preliminary note on the biology and chromosome cycle of Aphiochaeta xanthina Sp. Am. Nut. 86, 123-124. Baudisch, W. ( 1960). Spezifisches Vorkommen von Carotinoiden und Oxyprolin in den Speicheldriisen von Acricotopus lucidus. Nahtnvissenschaften 47, 498-499. Baudisch, W. ( 1963a). Chemisch-physiologische Untersuchungen an den Speicheldrusen von A cricotopus lucidus. 100 Jahre Landwirtschaftl. Institut der Universitat Halle, pp. 152-159. Baudisch, W. ( 1963b). Untersuchungen zur physiologischen Charakterisierung der einzelnen Speicheldriisenlappen von Acricotopus lucidus. In “Struktur und Funktion des genetischen Materials”. Bd 111, pp. 231-233. ErwinBaurGedachtnisvorlesungen.Akademie-Verlag, Berlin. Baudisch, W. ( 1963c). Aminosaurezusammensetzung der Speicheldriisen von Acricotopus lucidus. Biol. Zbl. 82, 35 1-36 1. Baudisch, W. (1964). Einbau von W-Prulhi in die Speicheldriisen von Acricotopus lucidus (Chkonomidae). In “Second International Congress of Histo- and Cytochemistry”, p. 236. Springer-Verlag, Berlin. Baudisch, W. ( 1967). Spezifische Hydroxyprolinsynthese in den Speicheldriisen von Acricotopus lucidus. Biol. Zbl. 86 (Suppl), 157-162. Baudisch, W. and Panitz, R. (1968). Kontrolle eines biochemischen Merkmals in den Speicheldriisen von Acricotopus lucidus durch einen Balbiani-Ring. Expl. Cell Res. 49,470-476. Bauer, H. (1935). Der Aufbau der Chromosomen aus den Speicheldrusen von Chironomus thummi Kiefer (Untersuchungen an den Riesenchromosomen der Diptera I). Z. Zellforsch. Mikrosk. Anat. 23, 280-313. Bauer, H. ( 1938). Die polyploide Natur der Riesenchromosomen. Naturwissenschaften 26,77-78. Bauer, H. and Beermann, W. (1952). Die Polytanie der Riesenchromosomen. Chromosoma 4,6304548. Becker, H. J. ( 1959). Die Puffs der Speicheldrusenchromosomen von Drosophila melanogaster. I. Beobachtungen zum Verhalten des Puffmusters im Normalstamm und bei zwei Mutanten, giant und lethal-giant-larvae. Chromosoma 10,6544578. Becker, H. J. ( 1962a). Die Puffs der Speicheldriisenchromosomen von Drosophila melanogaster. 11. Die Auslosung der Puffbildung, ihre Spezifitat und ihre Beziehung zur Funktion der Ringdriise. Chromosoma 13, 341-384.

PUFFS

73

Becker, H. J. ( 1962b). Stadienspezifische Genaktivierung in Speicheldrusen nach Transplantation bei Drosophila melanogaster. Zool. Anz., Suppl. 25, 92-101. Beermann, W. ( 1950). Chromomerenkonstanz bei Chironomus. Naturwissenschaften 37,543-544. Beermann, W. (1 952a). Chromosomenstruktur und Zelldifferenzierung in der Speicheldriise von Trichocladius vitripennis. Z. Naturf. 7b, 237-242. Beermann, W. ( 1952b). Chromomerenkonstanz und spezifische Modifikation der Chromosomenstruktur in der Entwicklung und Organdifferenzierung von Chironomus tentans. Chromosoma 5, 139-198. Beermann, W. ( 1956). Nuclear differentiation and functional morphology of chromosomes. Cold Spring Harb. Symp. Quant. Biol. 2 1 , 2 17-232. Beermann, W. ( 1959). Chromosomal differentiation in insects. In “Developmental Cytology” (D. Rudnick, ed.), pp. 83-103. Ronald, New York. Beermann, W. (1961). Ein Balbiani-ring als Locus einer Speicheldriisen-Mutation. Chromosoma 12, 1-25. Beermann, W. ( 1962). Riesenchromosomen. Protoplasmatologia VI/D. SpringerVerlag, Vienna. Beermann, W. (1963). Cytological aspects of information transfer in cellular differentiation. A m . Zool. 3, 23-32. Beermann, W. ( 1965a). Operative Gliederung der Chromosomen. Nafurwissenschaften 52, 365-375. Beermann, W. (1965b). Structure and function of interphase chromosomes. In “Genetics Today”. Proc. XI Internat. Congress o f Genetics, Vol. 2 (S. J. Geerts, ed.), pp. 375-384. Beermann, W. (1 966). Differentiation at the level of the chromosomes. In “Cell Differentiation and Morphogenesis”. North Holland, Amsterdam. Beermann, W. (1967a). Gene action at the level of the chromosome. In “Heritage from Mendel” (R. A. Brink, ed.), pp. 179-210. University of Wisconsin, Madison. Beermann, W. ( 1967b). Gene-Regulation in Chromosomen hohere Organismen. Jahrbuch 1966 der Max-PlanckGesellschaft, pp. 69-87. Beermann, W. and Bahr G. F. (1954). Submicroscopic structure of the Balbiani-ring. Expl. Cell R es. 6, 195-20 1. Beermann, W. and Pelling, C. (1965). H3-Thymidin-markierung einzelner Chromatiden in Riesenchromosomen. Chromosoma 16,l-2 1. Benjamin, W., Goodman, R. M. and Gellhorn, A. (1968). Studies on the phosphorylation of Dipteran chromosomes and rat liver chromatin. J. Cell Biol. 39, 12A. Berendes, H. D. (1965a). Salivary gland function and chromosomal puffing patterns in Drosophila hydei. Chromosoma 17,3577. Berendes, H. D. (1965b). The induction of changes in chromosomal activity in different polytene types of cell in Drosophila hydei. Devl. Biol. 11, 37 1-384. Berendes, H. D. (1966). Gene activities in the Malpighian tubules of Drosophila hydei at different stages. J. exp. Zool. 162,209-218. Berendes, H. D. (1967a). The hormone ecdysone as effector of specific changes in the pattern of gene activities of Drosophila hydei. Chromosoma 22, 274-293.

74

M. ASHBURNER

Berendes, H. D. (1967b). Aminoacid incorporation into giant chromosomes of D. hydei. Drosoph. Znf. Serv. 42, 102-103. Berendes, H. D. (1968a). Factors involved in the expression of gene activity in polytene chromosomes. Chromosoma 24,418-437. Berendes, H. D. (1968b). The effect of ecdysone analogues on the puffing pattern of D . hydei. Drosoph. Znf. S e n . 43, 145. Berendes, H. D. (1968~ ) .Controlled induction of gene activity in polytene chromosomes. Atti del “Le basi molecolari del differenziamento”. Accademia Nazionale dei Lincei. Berendes, H. D. (1969). Induction and control of puffing. Ann. Embryol. Morph. (In press.) Berendes, H. D. and Beermann, W. (1969). Biochemical activity of interphase chromosomes. Zn “Handbook of Molecular Cytology” (A. Limade-Faria, ed.). North Holland, Amsterdam. Berendes, H. D. and Boyd, J. B. (1969). Structural and functional properties of polytene nuclei isolated from salivary glands of Drosophila hydei. J. Cell Biol. 41,591-599. Berendes, H. D. and de Bruyn, W. C. (1963). Submicroscopic structure of Drosophila hydei salivary gland cells. Z. Zellforsch. Mikrosk. Anat. 59, 142-152. Berendes, H. D. and Holt, Th. K. H. (1964). The induction of chromosomal activities by temperature shocks. Genen. Phaenen 9 , 1-7. Berendes, H. D. and Holt, Th. K. H. (1965). Differentiation of transplanted larval salivary glands of Drosophila hydei in adults of the same species. J. exp. Zool. 160,299-318. Berendes, H. D., Breugel, F. M. A. van and Holt, Th. K. H. (1965). Experimental puffs in salivary gland chromosomes of Drosophila hydei. Chromosoma 16,3546. Berger, C. A. (1940). The uniformity of the gene complex in nuclei of different tissues. J. Hered. 31, 3-4. Berry, S. J. and Dietz, W. (1 968). The action of dimethyl sdphoxide and DNAse on the fine structure of Chironomus salivary gland cells. J. Insect Physiol. 14,847-854. Bier, K. (1957). Endomitose und Polyttinie in den Ngihrzellkernen von Calliphora erythrocephala Miegen. Chromosoma 8,493-522. Bier, K. ( 1958). Beziehungen zwischen Wachstumsgeschwindigkeit, endometaphasischer Kontraktion und der Bildung von Riesenchromosomen in den Nihrzellen von Calliphora. Z. Naturf. 13b, 85-93. Bier, K. (1959). Quantitative Untersuchungen uber die Variabilitat der Niihrzellkernstruktur und ihre Beeinflussung durch die Temperatur. Chromosoma 10,619453. Bier, K. (1960). Der Karyotyp von Calliphora erythrocephala Meigen unter besonderer Beriicksichtigung der Nihrzellkernchromosomen im gebundelten und gepaarten Zustand. Chromosoma 11,335-364. Bier, K. ( 1967). Oogenese, das Wachstum von Riesenzellen. Naturwissenschaften 54, 189-195. Bier, K., Kunz, W. and Ribbert, D. (1967). Struktur und Funktion der Oocytenchromosomen und Nukleolen sowie der Extra-DNS wihrend der Oogenese panoistischer und meroistischer Insekten. Chromosoma 23, 2 14-254.

PUFFS

75

Bier, K., Kunz, W. and Ribbed, D. (1969). Insect oogenesis with and without lampbrush chromosomes. In “Chromosomes Today” (C. D. Darlington and K. R. Lewis, eds), Vol. 2, pp. 107-1 15. Oliver and Boyd, Edinburgh. Black, M. M. and Ansley, H. R. (1964). Histone staining with ammoniacal silver. Science, N.Y. 143, 693-695. Block, K. (1969). Chromosomal variation in Agromyzidae. I. Phytomyza abdominalis Zeitt. Two incipient species and their natural hybrid. Hereditas 62, 131-152. Bodenstein, D. (1943). Factors influencing growth and metamorphosis of the salivary gland in Drosophila. Biol. Bull. Mar. Biol. Lab., Woods Hole 84, 13-33. Bonner, J. ( 1965). “The Molecular Biology of Development”. Oxford University Press, Oxford. Boyd, J. B. ( 1969). Drosophila deoxyribonucleases. I. Variation of deoxyribonucleases in Drosophila melanogaster. Biochim. biophys. Acta 17 1, 103-112. Boyd, J. B. and Berendes, H. D. (1968). Nuclei isolated from salivary glands of Drosophila hydei-structural and functional aspects. Genetics 60, 164. Boyd, J. B. and Boyd, H. (1969). Drosophila deoxyribonucleases. 11. Deoxyribonucleases of the organs of Drosophila hydei at the onset of metamorphosis. (In press.) Boyd, J. B., Berendes, H. D. and Boyd, H. (1968). Mass preparation of nuclei from the larval salivary glands of Drosophila hydei. J. Cell, Biol. 38, 369-376. Breuer, M. E. and Pavan, C. (1954). Salivary chromosomes and differentiation. Proc. IX Internat. Congress Genetics, Caryologia (Suppl.) 2, 778. Breuer, M. E. and Pavan, C. (1955). Behaviour of polytene chromosomes of Rhynchosciara angelae at different stages of development. Chromosoma 7, 37 1-386. Breugel, F. M. A. van (1966). Puff induction in larval salivary gland chromosomes of Drosophila hydei Sturtevant. Genetica 37, 17-28. Bridges, C. B. (1935). Salivary chromosome maps. J. Hered. 26,6064. Bridges, C. B. (1937). Correspondences between linkage maps and salivary chromosome structure, as illustrated in the tip of 2R of Drosophila melanogaster. Cytologia, Fujii Jubilee Vol. pp. 745-755. Buetti, E. (1968). Einfluss von Elektrolyten auf das Puff-Muster der Riesenchromosomen in den Malpighigefassen von Chironomus thummi. Diploma-Thesis Eidgenossischen Technischen Hochschule, Zurich. Burdette, W. J. and Anderson, R.(1965). Puffing of salivary gland chromosomes after treatment with carbon dioxide. Nature, Lond. 208.409410. Callan, H. G. (1963). The nature of lampbrush chromosomes. Int. Rev. Cytol. 15, 1-34. Callan, H. G. (1967). The organisation of genetic units in chromosomes. J. Cell Sci. 2 , 1-7. Cannon, G. B. (1964). Culture of insect salivary glands in a chemically defined medium. Science, N . Y. 146, 1063. Cannon, G. B. (1965). Puff development and DNA synthesis in Sciara salivary gland chromosomes in tissue culture. J . cell. Comp. Physiol. 65, 163-182. Cassagnau, P. (1966). Presence de chromosomes geants dans les glandes salivaires des Neanura (Collembole suceurs). C.r. hebd. Seanc. Acad. Sci., Paris 262, 168-170.

76

M. ASHBURNER

Cassagnau, P. (1968). Sur la structure des chromosomes salivaires de Bilobella massoudi Cassagnau (Collembola: Neanuridae). Chromosoma 24,42-58. Cassagnau, P. (1969). Phyloghie et chromosomes polytknes chez les Collemboles Neanuridae. (In press.) Cave, M. D. (1968). Chromosomal replication and synthesis of non-histone proteins in giant polytene chromosomes. Chromosoma 25,392-40 1. Childress, D. ( 1969). Polytene chromosomes and linkage group-chromosome correlations in the Australian Sheep Blowfly Lucilia cuprina (Diptera: Calliphoridae). Chromosoma 26,208-214. Clever, U. ( 1961). Genaktivitaten in den Riesenchromosomen von Chironomus tentans und ihre Beziehungen zur Entwicklung. I. Genaktivierungen durch Ecdyson. Chromosoma 12,607-675. Clever, U. ( 1962a). Genaktivitaten in den Riesenchromosomen von Chironomus tentans und ihre Beziehungen zur Entwicklung. 11. Das Verhalten der Puffs wiihrend des letzen Larvenstadiums und der Puppenhautung. Chromosoma 13,385436. Clever, U. ( 1962b). Genaktivitaten in den Riesenchromosomen von Chironomus ten tans und ihre Beziehungen zur Entwicklung. 111. Das Aktivitatsmuster in Phasen der Entwicklungsruhe. J. Insect Physiol. 8,357-376. Clever, U. (1963a). Genaktivitaten in den Riesenchromosomen von Chironomus tentans und ihre Beziehungen zur Entwicklung. IV. Das Verhalten der Puffs in der Larvenhautung. Chromosoma 14,65 1-675. Clever, U. ( 1963b). Von der Ecdysonkonzentration abhangige Genaktivitatsmuster in den Speicheldriisenchromosomen von Chironomus tentans. Devl. Biol. 6,73-98. Clever, U. (1963~).Einige Bemerkungen uber die Regulation von Genaktivaten in Riesenchromosomen. In “Funktionelle und Morphologische Organisation der Zelle”. Wjssenschaftliche Konferenz der Gesellschaft Deutscher Naturforscher und Arzte 1962. Springer-Verlag, Heidelberg. Clever, U. (1964a). Puffing in giant chromosomes of Diptera and the mechanisms of its control. In “Nucleohistones” (J. Bonner and P. Ts’o, eds) Holden-Day, San Francisco. Clever, U. (1964b). Actinomycin and puromycin: Effects on sequential gene activation by ecdysone. Science, N.Y. 146,794-795. Clever, U. (1964~). Genaktivitaten und ihre Kontrolle in der tierischen Entwicklung. Nahtrwissenschaften 5 1,449-459. Clever, U. (1965a). The control of gene activity as a factor in cell differentiation in insect development. In “Developmental and Metabolic Control Mechanisms and Neoplasia”, Nineteenth Annual Symposium on Fundamental Cancer Research 1965. Williams & Wilkins, Baltimore. Clever, U. (1965b). Puffing changes in incubated and in ecdysone treated Chironomus tentans salivary glands. Chromosoma 17,309-322. Clever, U. (1966a). Gene activity patterns and cellular differentiation. Am. 2001. 6,3341. Clever, U. (1966b). Induction and repression of a puff in Chironomus. Devl. Biol. 14,42 1-438. Clever, U. (1967). Control of chromosome puffing. In “The Control of Nuclear Activity” (L. Goldstein, ed.). Prentice Hall,Englewood Cliffs, New Jersey. Clever, U. (1968). Regulation of chromosome function. A. Rev. Genetics 2, 11-30.

PUFFS

77

Clever, U. (1969). Chromosome activity and cell function in polytenic cells. 11. The formation of secretion in the salivary glands of Chironomus. Expl. Cell Res. 55, 3 17-322. Clever, U. and Beermann, W. (1963). Studies of nucleo-cytoplasmic interrelations in giant chromosomes of Diptera. Proc. XVZ Znternat. Congress Zoology, Washington, 3, 21 0-21 5. Clever, U. and Karlson, P. (1960). Induktion von Puff-veranderungen in den Speicheldriisenchromosomen von Chironomus ten tans durch Ecdyson. Expl. Cell Res. 20,623-626. Clever, U. and Romball, C. G. (1966). RNA and protein synthesis in the cellular response to a hormone, ecdysone. Proc. natn. Acad. Sci. U S A . 56, 1470-1476. Clever, U., Bultmann, H. and Darrow, J. M. (1969a). The immediacy of genomic control in polytenic cells. In “RNA in Development”. First International Park City Symposium on Problems in Biology (E. W. Hanly, ed.). (In press.) Clever, U., Storbeck, I. and Romball, C. G. (1969b). Chromosome activity and cell function in polytenic cells. I. Protein synthesis at various stages of larval development. Expl. Cell Res. 55,306-316. Cohen, L. H. and Gotchel, B. V. (1969). Histones and other proteins in Drosophila polytene nuclei. Fedn. Proc. Fedn. A m . SOCS.exp. Biol. 28, 600. Crouse, H. V. (1968). The role of ecdysone in DNA-puff formation and DNA synthesis in the polytene chromosomes of Sciara coprophila. Proc. natn. Acad. Sci. U.S.A. 61, 971-978. Crouse, H. V. and Keyl, H. G. ( 1968). Extra replications in the “DNA-puffs” of Sciara coprophila. Chromosoma 2 5 , 357-364. da Cunha, A. B., Franca, Z. M., GonGalves, A. M. A., Hitelman, A. and Garrido, M. C. (1967). Chromosomal diseases in Drosophila willistoni Sturtevant. Revta. bras. Biol. 2 7 , 1 13-124. da Cunha, A. B., Morgante, J. S., Pavan, C. and Garrido, M. C. (1968). Studies on the cytology and differentiation in Sciaridae. I. Chromosome changes induced by gregarine in Trichosia sp. (Diptera; Sciaridae). Caryologia 21, 27 1-282. da Cunha, A. B., Pavan, C., Morgante, J. S. and Garrido, M. C. (1969). Studies on cytology and differentiation in Sciaridae. 11. DNA redundancy in salivary gland cells of Hybosciara fragilis (Diptera; Sciaridae). Proc. Symp. Nuclear Physiology and Differentiation. Belo Horizonte, Genetics, 6 1, Suppl. 1, 335-349. Daneholt, B. and Edstrom, J.-E. (1967). The content of deoxyribonucleic acid in individual polytene chromosomes of Chironomus tentans. Cytogenetics 6, 35 0-356. Debaisieux, P. and Gastaldi, L. (1919). Les microsporidies parasites des larves de Simulium. La Cellule 30, 187-2 13. de Duve, C. (1963). General properties of lysosomes. Zn “Lysosomes”. CIBA Foundation Symposium (A. V.S. de Reuck and M. Cameron, eds). J. & A. Churchill, London. de Reuck, A. V. S. and Knight, J. (eds) (1966). “Histones: Their Role in the Transfer of Genetic Information”. CIBA Foundation Study Group. J. & A. Churchill, London.

78

M. ASHBURNER

Desai, L. and Tencer, R. (1968). Effects of histones and polylysine on the synthetic activity of the giant chromosomes of salivary glands of Dipteran larvae. Expl. Cell Res. 52, 185-197. Diaz, M. and Pavan, C. (1965). Changes in chromosomes induced by micro-organism infection. Proc. natn. Acad. Sci. U.S.A. 54, 1321-1327. Doane, W . (1969). Drosophila amylase and problems in cellular differentiation. In “RNA in Development”. First International Park City Symposium on Problems in Biology (E. W.Hanly, ed.). (In press.) Doyle, D. and Laufer, H. (1968). Analysis of secretory processes in Dipteran salivary glands. In “In Vitro”, Vol. 3; Differentiation and Defense Mechanisms in Lower Organisms. Symposium of the Tissue Culture Association (M. M. Siegel, ed.). Doyle, D. and Laufer, H.(1969a). Sources of larval salivary gland secretion in the Dipteran Chironomus tentans. J. Cell Biol. 40,61-78. Doyle, D. and Laufer, H. (1969b). Requirements of ribonucleic acid synthesis for the formation of salivary gland specific proteins in larval Chironomus tentans. (In press.) Dreyfus, A., Nonato, E., Breuer, M. E. and Pavan, C. (1951). Cromosomas politenicos em virios 6rgHos de Rhynchosciara angelae Nonato and Pavan (Diptera). Revta bras. Biol. 11,439-450. Edstrom, J.-E. (1964a). Chromosomal RNA and other nuclear RNA fractions. In “Role of Chromosomes in Development” (M. Locke, ed.), pp. 137-152. Academic Press, New York. Edstrom, J.-E. ( 1964b). Microextraction and microelectrophoresis for determination and analysis of nucleic acids in isolated cellular units. In “Methods in Cell Physiology”, Vol. I (D. M. Prescott, ed.). Academic Press, New York. Edstrom, J.-E. and Beermann, W. (1962). The base composition of nucleic acids in chromosomes, puffs, nucleoli and cytoplasm of Chironomus salivary gland cells. J. Cell Biol. 14, 37 1-379. Edstrom. J.-E. and Daneholt, B. (1967). Sedimentation properties of the newly synthesised RNA from isolated nuclear components of Chironomus tentans salivary gland cells. J. molec. Biol. 28, 331-343. Ellgaard, E. G. (1967). Gene activation without histone acetylation in Drosophila melanogaster. Science, N . Y . 157, 1070-1072. Ellgaard, E. G. (1969). In preparation. Emmerich, H. (1969). Distribution of tritiated ecdysone in salivary gland cells of Drosophila. Nature, Lond. 22 1, 954. Federoff, N. and Milkman, R. D. (1 964). Specific puff induction by tryptophan in Drosophila salivary chromosomes. Biol. Bull. Mar. biol. Lab., Woods Hole 127, 369. Federoff, N. and Milkman, R. D. (1965). Induction of puffs in Drosophila salivary chromosomes by amino acids. Drosoph. In$ Serv. 40,48. Ficq, A. and Pavan, C. (1957). Autoradiography of polytene chromosomes of Rhynchosciara angelae at different stages of development. Nafure, Lond. 180, 983-984. Ficq, A. and Pavan, C. (1961). Mktabolisme des acides nuclkiques et des protkines dans les chromosomes gkants. Path. Biol., Paris 9,756-757. Ficq, A., Pavan, C. and Brachet, J . (1958). Metabolic processes in chromosomes. Expl. Cell Res. Suppl. 6 , 105-1 14.

PUFFS

79

Finlayson, L. H. and Walters, V. A. (1957). Abnormal metamorphosis in Saturniid moths infected by a microsporidian. Nature, Lond. 180, 7 13-7 14. Fischer, F. M. and Sanborn, R. C. (1962). Production of insect juvenile hormone by the microsporidian parasite Nosema. Nature, Lond. 194, 1193. Fraenkel, G. (1935). A hormone causing pupation in the blowfly Calliphora erythrocephala. Proc. R . SOC. B118, 1-12. Fraenkel. G. and Brookes, V. J. (1953). The process by which the puparia of many species of flies become fixed t o a substrate. Biol. Bull. Mar. biol. Lab., Woods Hole 105,442-449. Frizzi, G. and Springhetti, A. (1953). Prime richerche citogenetiche sul Dacus oleae Gmel. La Ricerca Scientifica 23, 1613-1620. Fujita, S. and Takemoto, K. (1963). Synthesis of messenger RNA on the polytene chromosomes of Dipteran salivary gland. Nature, Lond. 200, 494-495. GabrusewyczGarcia, N. ( 1964). Cytological and autoradiographic studies in Sciara coprophila salivary chromosomes. Chromosoma 15, 3 12-344. GabrusewyczGarcia, N. (1968). RNA metabolism of polytene chromosomes of Sciara coprophila. J. Cell Biol. 39,49A. Gabrusewycz-Garcia, N. and Kleinfeld, R. G. (1966). A study of the nucleolar material in Sciara coprophila. J. Cell Biol. 29, 347-359. Gall, J. G. (1968). Differential synthesis of the genes for ribosomal RNA during amphibian oogenesis. Proc. natn. Acad. Sci. U.S.A. 60, 553-560. Gall, J. G., MacGregor, H. C. and Kidston, M. E. (1969). Gene amplification in the oocytes of Dytiscid water beetles. Chromosoma 26, 169-187. Garcia-Bellido, A. (1965). Larvalentwicklung transplantierter organe von Drosophila melanogaster im Adultmileu. J. Insect Physiol. 11, 1071-1078. Gaudecker, B. von (1967). RNA synthesis in the nucleolus of Chironomus thummi as studied by high resolution autoradiography. Z . Zellforsch. Mikrosk. Anat. 82, 536-557. Gay, H. ( 1956). Nucleocytoplasmic relations in Drosophila. Cold Spring Harb. Symp. quant. Biol. 21, 257-269. Gay, H. (1959). In “The Nature of the Materials of Heredity” (B. P. Kaufman, H. Gay, M. K. Dutt, A. K. Bal and J. Buchanan, eds). Yb. Carnegie Instn. Wash. 58,440-449. Gay, H. (1964). Chromosome structure and function. Y b . Carnegie Instn. Wash. 63,605-614. Geiger, H. R. and Mitchell, H. K. (1966). Salivary gland function in phenoloxidase production in Drosophila melanogaster. J. Insect Physiol. 12,747-754. Geitler, L. ( 1953). Endomitose und endomitotische Polyploidisierung. Protoplasmatologia VI/C. Springer-Verlag, Vienna. Gilbert, E. F. and Pistey, W. R. (1966). Chromosomal puffs in Drosophila induced by hydrocortisone phosphate. Proc. SOC. exp. Biol. Med. 121, 831-832. Golikowa, M. N. (1964). Polytene chromosomes in the developing macronucleus of a ciliate. Tsitologia 6, 250-253. Goodman, R. M., Goidl, J. and Richart, R. M. (1967). Larval development in Sciara coprophila without the formation of chromosomal puffs. Proc. natn. Acad. Sci. Wash. 58, 553-559.

80

M. ASHBU RN ER

Goodman, R. M., Terner, J. Y. and Spiro, D. (1968). The effect of cortisone on glucose-6-phosphatase during development. Expl. Cell Res. 49, 504-5 12. Gorovsky, M. A. and Woodward, J. (1966). Hivtone : DNA ratios in five regions of a polytene chromosome of Drosophila virilis. J. Cell Biol. 31,41A42A. Gorovsky, M. A. and Woodward, J. (1967). Histone content of chromosomal loci active and inactive in RNA synthesis. J. Cell Biol. 33,723-728. Gottschalk, A. ( 1966). “Glycoproteins”. Elsevier, Amsterdam. Greenberg, J. R. (1967). Sedimentation studies on Drosophila virilis salivary gland RNA. J. Cell Biol. 35,49A-5OA. Gross, J. D. ( 1957). Incorporation of phosphorus-32 into salivary-type chromosomes which exhibit puffs. Nature, Lond. 180,440. Grossbach, U. (1968). Cell differentiation in the salivary glands of Camptochironomus tentans and C. palliduvitatus. Ann. Zool. Fennici 5 , 37-40. Hadorn, E., Gehring, W. and Staub, M. (1963). Extensives Grossenwachstum larvaler Speicheldrusenchromosomen von Drosophila melanogaster im Adultmilieu. Experientia 19,530-531. Hadorn, E., Ruch, F. and Staub, M. (1964). Zum DNS Gehalt in Speicheldriisenkernen mit “iibergrossen Riesenchromosomen” von Drosophila melanogaster. Experientia 20, 566. Hams, H. (1968). “Nucleus and Cytoplasm”. Clarendon Press, Oxford. Heitz, E. and Bauer, H. (1933). Beweise fur die Chromosomennatur der Kernschleifen in den Knauelkernen von Bibio hortulanus L. Z. Zellforsch. Mikrosk. Anat. 17,67-82. Henderson, S. A. (1967a). A second example of normal coincident endopolyploidy and polyteny in salivary gland nuclei. Caryologia 2 0 , 181-186. Horn, D. H. S., Middleton, E. J. and Wunderlich, J. A. (1966). Identity of the moulting hormone of insects and crustaceans. Chem. Comm. 1966, 3 39-34 1. Henderson, S . A. (1967b). The salivary gland chromosomes of Dasyneura crataegi (Diptera: Cecidomyiidae). Chromosoma 23, 38-58. Horn, E. C. and Ward, C. L. (1957). The localization of basic proteins in the nuclei of larval Drosophila salivary glands. Proc. natn. Acad. Sci. U.S.A. 43,776-779. Hsu, W. S. (1948). The Golgi material and mitochondria in the salivary glands of the larva of Drosophila melanogaster. Q. Jl. microsc. Sci. 89,401414. Ito, S . and Loewenstein, W. R. (1965). Permeability of a nuclear membrane: Changes during normal development and changes induced by growth hormone. Science, N. Y. 150, 909-910. Jacob, F. and Monod, J. (1961a). Genetic regulatory mechanisms in the synthesis of proteins. J. molec. Biol. 3 , 3 18-356. Jacob, F. and Monod, J. (1961b). On the regulation of gene activity. Cold Spring Harb. Symp. quant. Biol. 26, 193-2 11. Jacob, J. and Jurand, A. (1963). Electron microscope studies on salivary gland cells of Bradysia mycorum Frey (Sciaridae). 111. The structure of the cytoplasm. J. Insect Physiol. 9,849-857. Jacob, J. and Jurand, A. (1965). Electron microscope studies on salivary gland cells. V. The cytoplasm of Smittia parthenogenetica (Chironomidae). J. Insect Physiol. 11, 1337-1343.

PUFFS

81

Jacob, J. and Sirlin, J. L. (1959). Cell function in the ovary of Drosophilu. I. DNA classes in nurse cell nuclei as determined by autoradiography. Chromosoma 1 0 , 2 10-228. Jacob, J. and Sirlin, J. L. (1963). Electron microscope studies on salivary gland cells. I. The nucleus of Brudysia mycorum Frey (Sciaridae) with special reference to the nucleolus. J. Cell Biol. 17, 153-165. Jacob, J . and Sirlin, J. L. (1964). Electron microscope studies on salivary gland cells. IV. The nucleus of Smittiu parthenogeneticu (Chironomidae) with special reference to the nucleolus and the effects of actinomycin thereon. J. Ultrastruct. Res. 11, 315-328. Jurand, A., Simoes, L. C. and Pavan, C. (1967). Changes in the ultrastructure of salivary gland cytoplasm in Sciara ocellaris (Comstock, 1882) due to microsporidian infection. J. Insect Physiol. 13,795-803. Kalnins, V. I., Stich, H. F. and Bencosme, S. A. (1964a). Fine structure of nucleoli and RNA-containing chromosome regions of salivary gland chromosomes of Chironomus and their interrelationships. Can. J. Zool. 42, 1147-1155. Kalnins, V. I., Stich, H. F. and Bencosme, S. A. (1964b). Fine structure of the nucleolar organiser of salivary gland chromosome of Chironomids. J. Ultrustruct. Res. 11, 282-29 1. Kanno, Y. and Loewenstein, W. R. (1966). Cell-to-cell passage of large molecules. Nature, Lond. 212, 629-630. Karlson, P. (1965). Biochemical studies of ecdysone control of chromosomal activity. J. cell. comp. Physiol. 66, 69-76. Karlson, P. ( 1966). Ecdyson, das Hautungshormon der Insekten. Nuturwissenschuften 53,445453. Karlson, P. and Bode, C. (1969). Die Inaktivierung des Ecdysons bei der Schmeissfliege CUlliphOrU erythrocephdu Meigen. J. Insect Physiol. 15, 11 1-118. Karlson, P. and Hanser, G. (1952). Uber die Wirkung des Puparisierungshormons bei der Wildforms und der Mutante lgl von Drosophilu melanogaster. Z. Nuturforsch. 7b, 80-83. Karlson, P. and Loffler, U. (1962). Isolierung von Speicheldrusenchromosomen durch differentiefie Zentrifugation. Hoppe-Seyler 's 2.physiol. Chem. 321, 28 6-288. Karlson, P., Sekeris, C. E. and Maurer, R. (1964). Zum Wirkungsmechanismus der Hormone. I. Verteilung von Tritium-markiertem Ecdyson in Larven von Calliphoru erythrocephdu. Hoppe-Seyler 's 2. physiol. Chem. 336, 100-106. Kato, K.4. and Sirlin, J. L. (1963). Aspects of mucopolysaccharide production in larval insect salivary cells. J. Histochem. Cytochem. 11, 163-168. Kato, K.-I., Perkowska, E. and Sirlin, J. L. (1963). Electro- and immunoelectrophoretic patterns in the larval salivary secretion of Chironomus thummi. J. Histochem. Cytochem. 11,485488. Keyl, H. G. (1960). Ehrohung der chromosomalen replikationsrate durch Mikrosporidieninfektion in Speicheldrusenzellen von Chironomus. Naturwissenschaften 4 1 , 212-213. Keyl, H.G. (1963). DNS-Konstanz in Heterochromatin von Glyptotendipes. Expl. Cell Res. 30, 245-247. Keyl, H.G. (1965). Duplikationen von Untereinheiten der Chromosomalen DNS

82

M. ASHBURNER

w a r e n d der Evolution von Chironomus thummi. Chromosoma 17, 139- 180. Keyl, H.G. and Hagele, K. ( 1966). Heterochromatin-proliferation an den Speicheldriisen-chromosomen von Chironomus melanotus. Chromosoma 19,223-230. Kiknadze, I. I. (1965). Functional changes of giant chromosome under conditions of inhibited RNA synthesis. Tsitologia 7 , 3 1 1-3 18 (Russian with English summary). Kiknadze, I. I. (1967). Chromosomes of Diptera. I. Culicidae (evolutionary and practical importance of studying the karyotypes). Genetika (U.S.S.R.) 1967( 1 l) , 145-165 (Russian with English summary). Kiknadze, I. I. and Filatova, I. T. (1963). Changes in the RNA of giant chromosomes of Chironomus dorsalis during metamorphosis and under experimental conditions. Dokl. Acad. Nauk. S.S.S.R. 152, 1276-1279 (English Translation Edition). Kimoto, Y. (1958). An attachment of the salivary gland chromosomes to the nuclear membrane found in Chironomus dorsalis. Cytologia 23, 478-484. King, D. S. and Siddall, J. B. (1969). Conversion of @cdysone to pecdysone by crustaceans and insects. Nature, Lond. 221,955-956. King, R. C. (1964). Studies on early stages of insect oogenesis. In “Insect Reproduction”: Symposia of the Royal Entomological Society, 2 (K. C. Highnam, ed.). London. King, R. C., Sang, J. H. and Leth, C. B. (1961). The hereditary ovarian tumours of the fes mutant of Drosophila rnelanogaster. Expl. Cell Res. 23, 108-1 17. Kitzmiller J. B. ( 1967). Mosquito cytogenetics. In “Genetics of Insect Vectors of Disease” (J. W. Wright and R. Pal, eds). Elsevier, Amsterdam. Kloetzel, J. A. and Laufer, H. (1968). Fine structural analysis of larval salivary gland function in Chironomus thummi. J. Cell Biol. 39,74A. Klug, W. S., Bodenstein, D. and King, R. C. (1968). Oogenesis in the su2-Hw mutant of Drosophila melanogaster. I. Phenotypic characterization and transplantation experiments. J. exp. 2001.167, 15 1-156. Kobel, H. R. and Breugel, F. M. A. van (1968). Observations on It1 (lethal tumorous larvae) of Drosophila melanogaster. Genetica 38,305-327. Kodani, M. (1948). The protein of the salivary gland secretion in Drosophila. Proc. natn. Acad. Sci. U.S.A. 34, 131-135. Koltzoff, N. (1934). The structure of the chromosomes in the salivary glands of Drosophila. Science, N . Y . 8 0 , 312-313. Kosswig, C. and Sengiin, A. (1947). Neuere Untersuchungen uber den Bau der Riesenchromosomen der Dipteren. Rev. Fac. Sci. Univ. Istanbul B12, 107- 12 1. Kraczkiewicz, Z. and Matuszewski, B. (1958). Observations sur une structure atypique des chromosomes salivaires de certains cbcidomyides. Chromosoma 9,484-492. Krimbas, C. B. (1963). A contribution to the cytogenetics of Dacus oleae (Gmel) (Diptera, Trypetidae): The salivary gland and mitotic chromosomes. Caryologia 16,37 1-375. Kroeger, H. (1960). The induction of new puffing patterns by transplantation of salivary gland nuclei into egg cytoplasm of Drosophila. Chromosoma 11, 129-145. Kroeger, H. (1963a). Experiments on the extranuclear control of gene activity in

PUFFS

83

Dipteran polytene chromosomes. J. cell. comp. Physiol. 62 (Suppl. l), 45-59. Kroeger, H. (1963b). Chemical nature of the system controlling gene activities in insect cells. Nature, Lond. 200, 1234-1235. Kroeger, H. ( 1964). Zellphysiologische mechanismen bei der Regulation von Genaktivitaten in den Riesenchromosomen von Chironomus thummi. Chromosoma 15, 36-70. Kroeger, H. ( 1966). Potentialdifferenz und Puffmuster. Electrophysiologische und Cytologische Untersuchungen an der Speicheldriisen von Chironomus thummi. Expl. CellRes. 41, 64-80. Kroeger, H. (1967). Hormones, ion balances and gene activity in Dipteran chromosomes. Mem. SOC.Endocr. 15, 5566. Kroeger, H. (1968). Gene activities during insect metamorphosis and their control by hormones. Zn “Metamorphosis: A Problem in Developmental Biology” (W. Etkin and L. I. Gilbert, eds). North Holland, Amsterdam. Kroeger, H. and Lezzi, M. (1966). Regulation of gene action in insect development. A . Rev. Ent. 11, 1-22. Kunz, W. (1967a). Funktionsstrukturen im Oocytenkern von Locusta migratoria. Chromosoma 20, 332-370. Kunz, W. ( 1967b). Lampenburstenchromosomen und multiple Nukleolen bei Orthopteren. Chromosoma 2 1,446462. Lamprecht, J. and Remensberger, P. (1966). Polytane Chromosomen im Bereiche der Prothoracal-dorsal-Imaginalscheibe von Drosophila melanogaster. Experientia 22, 293. Laufer, H. (1963). Hormones and the development of insects. Proc. XVZ Znt. Congr. Zool. 4 , 2 15-220. Laufer, H. (1965). Developmental studies of the Dipteran salivary gland. 111. Relationships between chromosomal puffing and cellular function during development. Zn “Developmental and Metabolic Control Mechanisms and Neoplasia”. Nineteenth Annual Symposium of Fundamental Cancer Research. Williams and Wilkins, Baltimore. Laufer, H. (1968). Developmental interactions in the Dipteran salivary gland. A m . Zool. 8,257-271. Laufer, H. and Goldsmith, M. (1965). Ultrastructural evidence for a protein transport system in Chironomus salivary glands and its implication for chromosomal puffing. J. Cell Biol. 27,57A. Laufer, H. and Holt, Th. K. H. (1969). Arch. d’anat. Microsc. (In press.) Laufer, H. and Nakase, Y. (1964). Chromosomal puffing as an expression of protein transport by the Dipteran salivary gland. J. Cell Biol. 23,52A. Laufer, H. and Nakase, Y.(1965a). Salivary gland secretion and its relation to chromosomal puffing in the Dipteran, Chironomus thummi. Proc. natn. Acad. Sci. U.S.A. 53,511-516. Laufer, H. and Nakase, Y. (1965b). Developmental studies of the Dipteran salivary gland. 11. DNAse activity in Chironomus thummi. J. Cell Biol. 25, 97-102. Laufer,H., Nakase, Y. and Vanderberg, J. (1964). Developmental studies of the Dipteran salivary gland. I. The effects of Actinomycin D on larval development, enzyme activity and chromosomal differentiation in Chironomus thummi. Devl. Biol. 9 , 367-384. Laufer, H., Rao, B. and Nakase, Y. (1965). Dipteran salivary gland metamorphosis, a self-regulating system. A m . Zool. 5 , 642.

84

M. ASHBURNER

Laufer, H., Rao, B. and Nakase, Y. (1968). Developmental studies of the Dipteran salivary gland. IV. Changes in DNA content. J. exp. Zool. 166, 7 1-76. Lees, A. D. and Waddington, C. H. (1942). The development of bristles in normal and some mutant types of Drosophila melanogaster. Proc. R . SOC. B131,87-110. Lesher, S . (1951). Studies on the larval salivary gland of Drosophila. I. The nucleic acids. Expl. Cell Res. 2, 577-585. Lesher, S. (1962). Studies on the larval salivary gland of Drosophila. 111. The histochemical localization and possible significance of ribonucleic acid, alkaline phosphatase and polysaccharide. Anat. Rec. 114,63345 1. Lezzi, M. (1961). Das Verhalten des Puffmusters der Speicheldriisenchromosomen von Chironomus unter dem Einfluss von DrosophilaEinhalt. Diploma-Thesis, Eidgenossischen Technischen Hochschule, Zurich. Lezzi, M. (1965). Die Wirkung von DNAse auf isolierte Polytlinchromosomen. Expl. Cell Res. 39, 289-292. Lezzi, M. (1966). Induktion eines Ecdyson-aktivierbaren Puff in isolierten Zellkernen von Chironomus durch KC1. Expl. Cell Res. 43, 57 1-577. Lezzi, M. ( 1967a). RNS-und Protein-synthese in Puffs isolierter Speicheldriisenchromosomen von Chironomus. Chromosoma 21,72-88. Lezzi, M. (1967b). Cytochemische Untersuchungen an Puffs isolierter Speicheldriisen-chromosomen von Chironomus. Chromosoma 2 1, 89-108. Le'zzi, M. ( 1 9 6 7 ~ ) .Spezifische Aktivitatssteigerung eines Balbianiringes durch Mg2+ in isolierten Zellkernen von Chironomus. Chromosoma 2 1, 109-122. Lezzi, M. and Kroeger, H. (1966). Aufnahme von 22Na in die Zellkerne der Speicheldriisen von Chironomus thummi. Z. Naturf. 21 b, 274-277. Lindsley, D. L. and Grell, E. H. (1968). Genetic variations of Drosophila melanogaster. Publs. Carnegie Instn 627. Lockschin, R. A. and Williams, C. M. (1965a). Programmed cell death. I. Cytology of degeneration in the intersegmental muscles of the pernyi silkmoth. J. Insect Physiol. 11, 123-133. Lockschin, R. A. and Williams, C. M. (1965b). Programmed cell death. V. Cytolytic enzymes in relation to the breakdown of the intersegmental muscles of silkmoths. J. Insect Physiol. 11,831-844. Loewenstein, W. R. and Kanno, Y.(1964). Studies on an epithelial (gland) cell junction. I. Modifications of surface membrane permeability. J. Cell Biol. 22, 565-586. MacGregor, H. C . and Mackie, J. B. (1967). Fine structure of the cytoplasm in salivary glands of Simulium. J. Cell Sci. 2 , 137-144. MacInnes, J. W. and Uretz, R. B. (1966). Organization of DNA in Dipteran polytene chromosomes as indicated by polarized fluorescence microscopy. Science, N.Y. 151,689-691. Mainx, F. (1949). The structure of the giant chromosomes in some Diptera. Proc. VZI Internat. Congress Genetics, Hereditas (Suppl.), 622-623. Mainx, F. (195 1). Die Verbeitung von Chromosomendislokationen in Natiirlichen Population von Liriomyza urophonna MIK. Chromosoma 4, 5 2 1-534. Makino, S . (1938). A morphological study of the nucleus in various kinds of somatic cells of Drosophila virilis. Cytologia 9,272-282.

PUFFS

85

Mattingly, E. (1966). Variations in metabolic activity during the larval development of Rhynchosciara. J. Cell Biol. 31,74A. Mattingly, E. and Parker, C. (1968a). Sequence of puff formation in Rhynchosciara polytene chromosomes. Chromosoma 23,255-270. Mattingly, E. and Parker, C. (196813). Nucleic acid synthesis during larval development of Rhynchosciara. J. Insect Physiol. 14, 1077-1083. Matuszewski, B. (1964). Polyploidy and polyteny induced by a hymenopteran parasite in Dasyneura urticae (Diptera; Cecidomyiidae). Chromosoma 15, 31-35. Matuszewski, B. (1965). Transition from polyteny to polyploidy in salivary glands of Cecidomyiidae. Chromosoma 16,22-34. McMaster-Kaye, R. (1962). The metabolism of nuclear ribonucleic acid in salivary glands of Drosophila repleta. J. Histochem. Cytochem. 10, 154- 16 1. Mechelke, F. ( 1952). Reversible Strukturmodifikationen der Speicheldriisenchromosomen von Acricotopus lucidus. Chromosoma 5, 5 11-543. Mechelke, F. (1958). The timetable of physiological activity of several loci in the salivary gland of Acricotopus lucidus. Proc. X Znt. Congress Genet. 2, 185. Mechelke, F. (1962). Spezielle Funktionszustande des genetischen Materials. In “Funktionelle und Morphologische Organisation der Zelle”. Wissenschaftliche Konferenz der Gesellschaft Deutscher Naturforscher und Arzte. Springer-Verlag, Heidelbefg. Melland, A. M. (1941). The salivary glands of Prodiamesia olivacea. J. Hered. 32, 165-167. Melland, A. M. (1942). Types of development of polytene chromosomes. Proc. R . SOC.Edinb. B61, 316-327. Metz, C. W. (1935). Structure of the salivary gland chromosomes in Sciara. J. Hered. 26, 176-188. Midelfort, M. E. and Rasch, R. W. (1968). The effect of exogenous cortisone acetate on development and “chromosome puffing” in Drosophila virilis. J. Cell Biol. 39, 91A. Misch, D. W. (1962). Localisation of acid phosphatase in tissues of metamorphosing flesh fly larvae. J. Histochem. Cytochem. 10,666. Mitchell, H . K. (1966). Phenol oxidases and Drosophila development. J. Insect Physiol. 12,755-765. Mukhejee, A. S. (1968). Effect of Dicyandiamide on puffing activity and morphology of salivary gland chromosomes of Drosophila melanogaster. Znd. J . Exp. Biol. 6,49-5 1. Nagl, W. (1969a). Banded polytene chromosomes in the legume Phaseolus vulgaris. Nature, Lond. 22 1,70-7 1. Nagl, W. (1969b). Puffing of polytene chromosomes in a plant (Phaseolus vulgaris). Naturwissenschaften 56, 22 1-222. Nur, U. (1968). Endomitosis in the Mealy Bug, Planococcus citri (Hornoptera: Coccoidea). Chromosoma 24,202-209. Ohtaki, T., Milkman, R. D. and Williams, C. M. (1968). Dynamics of ecdysone secretion and action in the flesh fly Sarcophaga peregrina. Biol. Bull. Mar. biol. Lab., Woods Hole 135,322-334. Osinchack, J. (1966). Ultrastructural localisation of some phosphatases in the prothoracic gland of the insect Leucophaea maderea. 2. Zellforsch. Mikrosk. Anat. 72,236-248. A.I.P.4

86

M. ASHBURNER

Painter, T. S . (1933). A new method for the study of chromosome rearrangements and the plotting of chromosome maps. Science, N. Y. 78, 585-586. Painter, T. S . (1946). Nuclear phenomena associated with secretion in certain gland cells with especial reference to the origin of cytoplasmic ribonucleic acid. J. exp. 2001.100, 523-547. Painter, T. S . and Griffen, A. B. (1937). The structure and the development of the salivary gland chromosomes of Simulium. Genetics 2 2 , 6 12-633. Painter, T. S . and Reindorp, E. C. (1939). Endomitosis in the nurse cells of the ovary of Drosophila melanogaster. Chromosoma 1, 276-283. Panitz, R. ( 1964). Hormonkontrollierte Genaktivitaten in den Riesenchromosomen von Acricotopus lucidus. Biol. Zbl. 83, 197-230. Panitz, R. ( 1967). Funktionelle Verlinderungen an Riesenchromosomen nach Behandlung mit Gibberellinen. Biol. Zbl. 86 (Suppl.), 147-156. Panitz, R. (1968). Uber die Rolle der Chromosomen bei der Informationsubertragung in hoheren Organismen. Biol. Zbl. 87,545-565. Patterson, E. K., Dackerman, M. E. and Schultz, J. (1949a). Peptidase activities of extracts of salivary glands of Drosophila melanogaster. J. gen. Physiol. 32,607-622. Patterson, E. K. Dackerman, M. E. and Schultz, J. (1949b). Peptidase increase accompanying growth of the larval salivary gland of Drosophila melanogaster. J. gen. Physiol. 32, 623-645. Paul, J. S . (1965). Quantitative biophysical and cytochemical studies of polytene chromosomes. Ph.D. Thesis, New York University. University Microfilms, Ann Arbor. Paul, J. S. and Mateyko, G. M. (1967). Quantitative cytochemical studies on polytene chromosomes. J. Cell Biol. 35, 183A. Pavan, C. (1958). Morphological and physiological aspects of chromosomal activities. Proc. X Znt. Congress Genet. 1, 321-336. Pavan, C. (1965a). Nucleic acid metabolism in polytene chromosomes and the problem of differentiation. Brookhaven Symp. Biol. 18, 222-24 1. Pavan, C. (1965b). Chromosomal differentiation. Natn. Cancer Znst. Monogr. 18, 309-323. Pavan, C. and Bade, R. (1966a). Chromosome changes induced by infections in tissues of Rhynchosciara angelae. Science, N.Y. 151, 1556-1558. Pavan, C. and Bade, R. (1966b). Invertebrate pathology, cytology, and development. J. Invert. Path. 8 , 131-132. Pavan, C. and Breuer, M. E. (1952). Polytene chromosomes in different tissues of Rhynchosciara angelae. J. Hered. 43, 150-157. Pavan, C. and Breuer, M. E. (1955). Polytene chromosomes of Rhynchosciara milleri sp. n. (Diptera; Mycetophilidae). Revta. bras. Biol. 15,329-339. Pavan, C. and da Cunha, A. B. (1969a). Chromosome activities in normal and infected cells of Sciaridae. Nucleus (In press). Pavan, C. and da Cunha, A. B. (1969b). Chromosomal activities in Rhynchosciara and other Sciaridae. A . Rev. Genetics 3, 425450. Pavan, C. and da Cunha, A. B. ( 1969~ ) .Gene amplification in ontogeny and phylogeny of animals. Proc. Symp. Nuclear Physiology and Differentiation-Belo Horizonte. Genetics 6 l , Suppl. l , 289-304.

PUFFS

87

Pavan, C. and Perondini, A. L. P. (1966). Changes in chromosomes induced by infection in cells of Sciara. Genetics 54, 353. Pavan, C., Perondini, A. L. P. and Amabis, J. M. (1968). Heterozygous puffs, bands, heterochromatin and nucleolus in natural population of Diptera. Proc. XII Int. Congress Genet. 1, 14. Pelc, S. and Howard, A. (1956). Metabolic activity of salivary gland chromosomes in Diptera. Expl. Cell Rex 10, 549-552. Pelling, C. (1959). Chromosomal synthesis of ribonucleic acid as shown by incorporation of uridine labelled with tritium. Nature, Lond. 184, 655-656. Pelling, C. ( 1964). Ribonucleinsaure-synthese der Riesenchromosomen. Autoradiographische Untersuchungen an Chironomus ten tans. Chromosoma 15, 71-122. Pelling, C . (1965). A replicative and synthetic chromosomal unit-the modern concept of the chromomere. Proc. R. Soc. B164, 279-289. Perkowska, E. (1963a). Some characteristics of the salivary gland secretion of Drosophila virilis. Expl. Cell Res. 32, 259-27 1. Perkowska, E. (1963b). The salivary gland secretion in Bradysia. Expl. Cell Res. 30,432-436. Perondini, A. L. P. and Dessen, E. M. (1969). Heterozygous puffs in Sciara ocellaris. Proc. Symp. Nuclear Physiology and Differentiation-Belo Horizonte. Genetics 6 1, Suppl. 1, 25 1-260. Pettit, B. J. and Rasch, R. W. (1966). Tritiated-histidine incorporation into Drosphila salivary chromosomes. J. Cell. Physiol. 68, 325-334. Philip, U. ( 1966). Chromosomal polymorphism in Coelopa frigida (Fabricus) (Diptera). Heredity, 21, 167-168. Phillips, D. M. and Swift, H. (1965). Cytoplasmic fine structure of Sciara salivary glands. I. Secretion. J. Cell Biol. 27, 395-409. Politoff, A., Socolar, S. J. and Loewenstein, W.R. (1969). Permeability of a cell membrane junction: Dependence on energy metabolism. J. gen. Physiol. 53,498-5 15. Poulson, D. and Metz, C. W. (1938). Studies on the structure of nucleolus forming regions and related structures in the giant salivary gland chromosomes of Diptera. J. Morph. 63,363-395. Prabhoo, N. R. (1961). A note on the giant chromosomes in the salivary glands of Womersleya sp. (Collembola: Insecta). Bull. Ent. Madras No. 2, 21-22. Prostakova, T. M. (1968). A study of the action of ribonuclease on RNA synthesis in giant chromosomes of Tendipes plumosus. Tsitologia 10, 157-161. Radzikowski, S. (1967). Nuclear behaviour during conjugation and polytene chromosomes in the exconjugants of Chilodonelia cucullalius (0. F. Muler). Bull. Acad. pol. Sci. C1. II S b . Sci. biol. 15, 749. Rasch, E. M. (1966). Developmental changes in patterns of DNA synthesis by polytene chromosomes in Sciara coprophila. J. Cell Biol. 31,91A. Rasch, E. M. and Gawlik, S. (1964). Cytolysosomes in tissues of metamorphosing Sciarid larvae. J. Cell Biol. 23, 123A. Rasch, E. M. and Lewis, A. L. (1968). Effects of cortisone, gibberellic acid and starvation on puffing patterns in polytene chromosomes of Sciara. J. Histochem. Cytochem. 16,508.

88

M. ASHBURNER

Rasch, E. M. and Pettit, B. J. (1967). Nucleoprotein metabolism in salivary gland chromosomes of Sciara during pupation. J. Cell Biol. 35, 110A. Rayle, R. E. (1967). A new mutant in Drosophila melanogaster causing an ecdysonecurable interruption of the prepupal molt. Genetics, 56, 583. Rayle, R. E. and Green, M. M. (1969). A contribution to the genetic fine structure of the region adjacent to white in Drosophila melanogaster. Genetica 39,497-507. Ribbert, D. ( 1967). Polytihchromosomen der Borstenbildungszellen von Calliphora erythrocephala. Unter besonder Beriicksichtigung der geschlechtsgebundenen Strukturheterozygotie und des Puffmusters wahrend der Metamorphose. Chromosoma 21,296-344. Ribbert, D. and Bier, K . (1969). Multiple nucleoli and enhanced nucleolar activity in the nurse cells of the insect ovary. Chromosoma 27, 178-197. Ristow, H. and Arends, S. (1968). A system in vitro for the synthesis of RNA and protein by isolated salivary glands and by nuclei from Chironomus larvae. Biochim. biophys. Acta 157, 178-186. Ritossa, F. (1962a). Attivita’ sintetiche al livello dei puffs in Drosophila busckii. A t t i Ass. genet. ital. 7, 147-156. Ritossa, F. (1962b). A new puffing pattern induced by temperature shock and DNP in Drosophila. Experientia 18, 571-573. Ritossa, F. (1963a). New puffs induced by temperature shock, DNP and salicilate in salivary chromosomes of D. melanogaster. Drosoph. Inf Serv. 37, 122-123. Ritossa, F. (1963b). Beta-galactosidase distribution in various organs of D . busckii and melanogaster. Drosoph. Inf. Serv. 37, 122. Ritossa, F. (1964a). Experimental activation of specific loci in polytene chromosomes of Drosophila. Expl. Cell Res. 35,601-607. Ritossa, F. (1964b). Behaviour of RNA and DNA synthesis at the puff level in salivary gland chromosomes of Drosophila. Expl. Cell Res. 36, 5 15-523. Ritossa, F. and Pulitzer, 5 . F. (1963). Aspects of structure of polytene chromosome puffs of Drosophila busckii derived from experiments with antibiotics. J. Cell Biol. 19, 6QA. Ritossa, F. and Spiegelman, S. (1965). Localisation of DNA complementary to ribosomal RNA in the nucleolus organiser region of Drosophila melanogaster. Proc. natn. Acad. Sci. U.S.A. 53,737-745. Ritossa, F. and von Borstel, R. C. (1964). Chromosome puffs in Drosophila induced by ribonuclease. Science, N. Y. 145, 5 13-5 14. Ritossa, F., von Borstel, R. C. and Swift, H. (1964). The action of ribonuclease on salivary gland cells of Drosophila. Genetics 50,279. Ritossa, F., Pulitzer, J. F., Swift, H. and von Borstel, R. C. (1965). On the action of ribonuclease in salivary gland cells of Drosophila. Chromosoma 16, 144-15 1. Rizki, T . M. (1967). Ultrastructure of the secretory inclusions of the salivary gland cell in Drosophila. J. Cell Biol. 32,531-534. Robert, M. (1964). Activation de genes chez Chironomus thummi par injection de Trypsine dans le noyau cellulaire. Diploma-Thesis Eidgenossischen Technischen Hochschule. Zurich. 27 pp. Robert, M. and Kroeger, H. (1965). Lokalisation zusatzlicher RNS-Synthese in

PUFFS

89

Trypsin-behandelten Riesenchromosomen von Chironomus thummi. Experien tia 2 1, 3 26. Roberts, B. (1968). DNA granule synthesis in the giant foot pad nuclei of Sarcophaga bullata. J. Cell Biol. 39, 112A. Roberts, P. A., Kimball, R. F. and Pavan, C. (1967). Response of Rhynchosciara chromosomes to microsporidian infection. Increased polyteny and generalized puffing. Expl Cell Res. 47,408-422. Rodems, A. E., Henderson, P. A. and Clever, U. (1969). Proteolytic enzymes in the salivary gland of Chironomus tentans. Experientia 25,686-687. Rodman, T. C. (1964). The larval characteristics and salivary gland chromosomes of a tumorigenic strain of Drosophila melanogaster. J. Morph. 115, 4 19-446. Rodman, T. C. (1967a). Control of polytenic replication in Dipteran larvae. I. Increased number of cycles in a mutant strain of Drosophila melanogaster. J. Cell. Physiol. 70, 179-186. Rodman, T. C. (1967b). DNA replication in salivary gland nuclei of Drosophila melanogaster at successive larval and prepupal stages. Genetics 55, 375-386. Rodman, T. C. and Kopac, M. J. (1964). Alterations in morphology of polytene chromosomes. Nature, Lond. 202,876-877. Rothfels, K. M. and Dunbar, R. W. (1953). The salivary gland chromosomes of the blackfly Simulium vittatum Zett. Can. J. 2001.31,226-241. Rudkin, G. T. (1955). The ultraviolet absorption of puffed and unpuffed homologous regions in the salivary gland chromosomes of Drosphila melanogaster. Genetics 40,593. Rudkin, G. T. (1962). Nucleic acid metabolism in giant chromosomes of Drosophila melanogaster. Annls Histochim. Suppl. 2 (1962), 77-84. Rudkin, G. T. (1964). The proteins of polytene chromosomes. In “Nucleohistones” (J. Bonner and P. Ts’o, eds). Holden-Day, San Francisco. Rudkin, G. T. (1965). The relative mutabilities of DNA in regions of the X chromosome of Drosophila melanogaster. Genetics 52,665-68 1. Rudkin, G. T. and Corlette, S. L. (1957). Disproportionate synthesis of DNA in a polytene chromosome region. Proc. natn. Acad. Sci. U.S.A. 43,964-968. Rudkin, G . T. and Woods, P. S. (1959). Incorporation of H3cytidine and H3-thymidine into giant chromosomes of Drosophila melanogaster during puff formation. Proc. natn. Acad. Sci. U.S.A. 45, 997-1003. Sang, J. H. (1968). Lack of cortisone inhibition of chromosomal puffing in Drosophila melanogaster. Experientia 24, 1064. Scharrer, B. and Hadorn, E. (1938). The structure of the ring gland (corpus allatum) in normal and lethal larvae of Drosphila melunogaster. Proc. natn. Acad. Sci. U.S.A. 24,236-242. Schin, K. S. and Clever, U. (1965). Lysosomal and free acid phosphatase in salivary glands of Chironomus tentans. Science, N. Y. 150, 1053-1055. Schin, K. S. and Clever, U. (1968a). Ultrastructural and cytochemical studies of salivary gland regression in Chironomus tentans. Z . Zellforsch. Mikrosk. Anat. 86,262-279. Schin, K. S. and Clever, U. (1968b). Ferritin-uptake by salivary glands of Chironomus tentans and its intracellular localisation. Expl. Cell Res. 49, 208-2 1 1.

90

M. ASHBURNER

Schultz, J. ( 1965). Genes, differentiation, and animal development. Brookhaven Symp. Biol. No. 18, 116-147. Schultz, J. and Rudkin, G. T. (1960). Direct measurement of deoxyribonucleic acid content of genetic loci in Drosophila. Science, N . Y . 132, 1499-1500. Schurin, M. F. ( 1959). Localised cytochemical and submicroscopic differentiation in Drosophila virilis salivary gland chromosomes. Genetics 44, 534. Sebeleva, T. E. (1968). Quantitative determination of DNA content during formation of one of the puffs and one of the Balbiani rings in Chironomus dorsalis. Tsitologia 10, 765-769 (Russian with English summary). Sebeleva, T. E., Sherudilo, A. I. and Kiknadze, I. I. (1965). Quantitative estimation of DNA in puff formation in Chironomus dorsalis. Genetika (U.S.S.R.)(1965) 2, 101-105. Sengiin, A. (1954). Variability of the banding patterns of giant chromosomes. J. Hered. 45, 119-122. Serfling, E., Panitz, R. and Wobus, U. (1969). Die experimentelle Beinflussung des Puffmusters von Riesenchromosomen. I. Puffinduktion durch Oxytetracyclin bei Chironomus thummi. Chromosoma 28, 107-1 19. Shaaya, E. and Karlson, P. (1965). Die Ecdysontiter wihrend InsektenEntwicklung. 11. Die Postembryonale Entwicklung der Schmeissfliege Calliphora erythrocephala Meig. J. Insect Physiol. 11, 6569. Sirlin, J. L. (1960). Cell sites of RNA and protein syntheses in the salivary gland of Smittia (Chironomidae). Expl. Cell Res. 19, 177-180. Sirlin, J. L. and Schor, N. A. (1962). Macromolecular syntheses in isolated polytene nuclei. Expl. Cell Res. 27, 165-167. Slizynski, B. M. (1950). Chironomus versus Drosophila. J. Genet. 50,77-78. Smith, P. D., Koenig, P. B. and Lucchesi, J. C. (1968). Inhibition of development in Drosophila by cortisone. Nature, Lond. 217, 1286. Smith, W. J. (1969). Effects of ribonuclease on nuclear ultrastructure in larval Drosophila salivary glands. J. Ultrastruct. Res. 27, 168-18 1. Snodgrass, R. E. (1924). Anatomy and metamorphosis of the apple maggot, Rhagoletis pomonella Walsh. J. agric. Res. 29, 1-36. Sorsa, M. (1969). Ultrastructure of puffs in the proximal part of chromosome 3R in Drosophila melanogaster. Ann. Acad. Sci. Fennicae, A IV Biol. 150, 1-21. Sorsa, M. and Sorsa, V. (1967). Electron microscopic observations on interband fibrils in Drosophila salivary chromosomes. Chromosoma 22, 32-4 1. Spirin, A. S. (1966). On “masked” forms of messenger RNA in early embryogenesis and in other differentiating systems. In “Current Problems in Developmental Biology’’, Vol. 1, 2-38. Stalker, H. D. (1954). Banded polytene chromosomes in the ovarian nurse cells of adult Diptera. J. Hered. 45,259-264. Staub, M. (1969). Veranderungen im Puffmuster und das Wachstum der Riesenchromosomen in Speicheldriisen von Drosophila melanogaster aus spatlarvalen und embryonalen Spendern nach Kultur in vivo. Chromosoma 26,76-104. Stedman, E. and Stedman, E. (1951). The basic proteins of cell nuclei. Phil. Trans. R. SOC.Lond. B235,565-595. Steffenson, D. M. (1963). Evidence for the apparent absence of DNA in the interbands of Drosophila salivary chromosomes. Genetics 48, 1289-1301.

PUFFS

91

Stevens, B. J. (1964). The effect of actinomycin D on nucleolar and nuclear fine structure in the salivary gland cell of Chironomus thummi. J. Ultrastruct. Res. 11, 329-353. Stevens, B. J. and Swift, H. (1966). RNA transport from the nucleus to cytoplasm in Chironomus salivary glands. J. Cell Biol. 31,55-77. Stevens, B. J.. Swift, H. and Adams, B. A. (1965). Fine structure and cytochemistry of Balbiani ring granules in Chironomus J. Cell Biol. 27, 1OOA . Stich, H. F. and Naylor, J. M. (1958). Variation of deoxyribonucleic acid content of specific chromosome regions. Expl. Cell Res. 14,442-445. Sutton, E. (1942). Salivary gland type chromosomes in mosquitoes. Proc. natn. Acad. Sci. U.S.A. 28, 268-272. Sutton, E. (1943). A cytogenetic study of the yellow-scute region of the X chromosome in Drosophila melanogaster. Genetics 2 8 , 2 10-2 17. Swift, H. (1959). Studies on nuclear fine structure. Brookhaven Symp. Biol. 12, 134-152. Swift, H. (1962). Nucleic acids and cell morphology in Dipteran salivary glands. In “Molecular Control of Gene Activity” (J. Allen, ed.). McGraw Hill, New York. Swift, H. (1963). Cytochemical studies on nuclear fine structure. Expl. Cell Res. Suppl. 9 , 5 4 6 7 . Swift, H. (1964). The histones of polytene chromosomes. In “Nucleohistones” (J. Bonner and P. Ts’o, eds). Holden Day, San Francisco. Swift, H. (1965). Molecular morphology of the chromosomes. “In Vitro”, Vol. 1. “The Chromosomes: Structural and Functional Aspects”. Tissue Culture Association. Tanzer, E. (1922). Die Zellkeme einiger Dipterenlarven und ihre Entwicklung. Z. wiss. Zool. 119, 114-153. Taylor, J. H. ( 1953). Intracellular localisation of labeled nucleic acid determined by autoradiographs. Science, N . Y . 118, 555-557. Terner, J. Y., Goodman, R. M. and Spiro, D. (1965). Glucose-6-phosphatase in the salivary gland of Sciara coprophila: A histochemical and biochemical study. J. Histochem. Cytochem. 13, 168-181. Terner, J. Y., Goodman, R. M. and Spiro, D. (1967). The effect of cortisone on ribonucleic acid polymerase and ribonuclease during development : Coincidental evidence for the identity of ribonucleic acid polymerase with the “operator” gene. Expl. Cell Res. 45, 550-558. Thomson, J. A. (1969). The interpretation of puff patterns in polytene chromosomes. Currents in Modern Biology 3 (In press). Trager, W. (1937). Cell size in relation to the growth and metamorphosis of the mosquito, Aedes aegypti. J. exp. Zool. 76,467489. Waldner-Stiefelmeier, R. D. (1967). Untersuchungen uber die Proteasen im Wildtyp und in den Letalmutanten (lme und Itr) von Drosophila melanogaster. Z. vergleich. Physiol. 56, 268-289. Walsche, B. M. (1947). Feeding mechanisms of Chironomus larvae. Nature, Lond. 160,474. Ward, C . L. (1949). Karyotype variation in Drosophila. Univ. Tex. Publs. 4920, 70-79. Welsch, U., Wachtler, K. and Riihm, W. (1968). Die Feinstruktur der

92

M. ASHBURNER

Speicheldriise von Boophthora erythrocephala (Simuliidae, Diptera) vor und nach der Blutaufnahme. 2.Zellforsch. Mikrosk. Anat. 88,340-352. Wetzel, R., Buder, E., Schiilike, W. and Zirwer, D. (1969). Linearer Dichroismus bei Riesenchromosomen von Chironomus. Chromosoma 26,201-207. Wharton, L. T. (1943). Analysis of the metaphase and salivary gland chromosome morphology with the genus Drosophila. Univ. Tex. Publs. 4313, 282-319. White, M. J. D. (1946). The cytology of the Cecidomyidae (Diptera). I. Polyploidy and polyteny in salivary gland cells of Lestodiplosis spp. J. Morph. 78, 201-220. White, M. J. D. (1948). The cytology of the Cecidomyidae (Diptera). IV. The salivary gland chromosomes of several species. J. Morph. 87, 53-80. Whitten, J. M. (1963). Giant polytene chromosome development in the cells of the hypodermis and heart of fly pupae. Proc. XVZ Znt. Congr. Zool. 2, 276. Whitten, J. M. (1964a). Giant polytene chromosomes in hypodermal cells of developing foot pads of Dipteran pupae. Science, N. Y. 143, 1437-1438. Whitten, J. M. (1964b). Haemocytes and the metamorphosing tissues in Sarcophaga bullata, Drosophila melanogaster, and other cyclorrhaphous Diptera. J. Insect Physiol. 10,447-469. Whitten, J. M. (1965). Differential deoxyribonucleic acid replication in the giant foot-pad cells of Sarcophaga bullata. Nature, Lond. 208, 1019-102 1. Whitten, J. M. (1968). Metamorphic changes in insects. Zn “Metamorphosis. A Problem in Developmental Biology” (W. Etkin and L. I. Gilbert, eds). North Holland, Amsterdam. Whitten, J. M. (1969a). Haemocyte activity in relation to epidermal cell growth, cuticle secretion, and cell death in a metamorphosing cyclorrhaphan pupa. J. Insect Physiol. 15, 763-778. Whitten, J. M. (1969b). Cell death during early morphogenesis: Parallels between insect limb and vertebrate limb development. Science, N.Y. 163, 1456-1457. Whitten, J. M. (1969~).Coordinated development in the fly foot: Sequential cuticle secretion. J. Morph. 127,73-104. Whitten, J. M. (1969d). Coordinated development in the foot pad of the fly Sarcophaga bullata during metamorphosis: Changing puffing patterns of the giant cell chromosomes. Chromosoma 26,215-244. Wiener, J., Spiro, D. and Loewenstein, W. R. (1964). Studies on an epithelial (gland) cell junction. 11. Surface structure. J. Cell Biol. 22,587-598. Williams, C. M. and Robbins, W. E. (1968). Conference on insect-plant interactions. BioScience 18,791-792,797-799. Willmer, E. N. (1961). Steroids and cell surfaces. Biol. Rev. 36,368-398. Wobus, U. ( 1969). Chromosomale Differenzierung und Proteinmuster bei Chironomiden. Verh. dt. Ges. explt. Med. (In press). Wolstenholme, D. R. (1965). The distribution of DNA and RNA in salivary gland chromosomes of Chironomus ten tans as revealed by fluorescence microscopy. Chromosoma 1 7 , 2 19-229. Wolstenholme, D. R. (1966). Direct evidence for the presence of DNA in interbands of Drosophila salivary gland chromosomes. Genetics 53, 357-360. Wolstenholme, D. R., Dawid, I. B. and Ristow, H. J. (1968). An electron

PUFFS

93

microscope study of DNA molecules from Chironomus tentans and C. thummi. Genetics 60,759-770. Yoshimatsu, H. (1962). Electron microscopy of salivary gland cells of Chironomus larvae. Ann. Zool. Japan 35,89-94. Yoshimatsu, H. and Uehawa, M. (1968). Some fibrin-like properties of the secretion of the salivary gland of Chironomus larvae. Ann. Zool. Japan 4 1, 107-112. NOTES ADDED IN PROOF The following notes detail some of the papers received since the completion of the review. P. 9 Basile (1969) describes the development of the polytene chromosomes of the endopolyploid nurse cells of Rhynchosciara angelae. P. 16 Four recent publications from Edstrom’s group extend our knowledge of the RNA metabolism of Chironomus tentans salivary glands. Egyhlzi et al. (1969) report further on the 4s RNA of chromosomes, nuclear sap and cytoplasm, and find it to be methylated, probably derived from a larger precursor and to be distinct from the 4s RNA of the nucleolus. Daneholt et al. (1969a) provide the controls to show that the heterodisperse (H) chromosomal RNA( 10s-90s) is composed of unbroken molecules and that neither aggregation nor complex formation contribute significantly to its heterogeneity. The kinetics of chromosomal RNA synthesis and metabolism are considered in Daneholt et al. (1969b). The H RNA of chromosomes I, I1 and I11 has a rapid turnover (less than 45 m) and leaves the chromosomes for the nuclear sap intact. Some of this H RNA reaches the cytoplasm but the majority is degraded within the nucleus. The RNA of the BR2 of chromosome IV is heterogeneous with respect to size but the amount of very heavy material is larger than that found from chromosomes I, I1 and 111. Within salivary glands the rate of H RNA synthesis on the chromosome IV BR correlates with the actual size of the BR (Daneholt et al., 1 9 6 9 ~ ) . P . 3 8 The induction of DNA puffs in Sciara by 20GH ecdysone has been reported by GabrusewyczGarcia and Margles (1 969). P. 38 Further metabolic conversions of ecdysones are reported by Thomson et al. ( 1969) and by Kaplanis et al. ( 1969). P . 4 1 Weirich and Karlson (1969) have also failed to find any marked heterogeneity in the distribution of H3-ecdysone in salivary glands of Chironomus or Rhynchosciara. However from Emmerich’s recent paper (1969) there is evidence that in larval salivary glands of D. hydei more ecdysone is bound by the nucleus than by the cytoplasm. No marked chromosomal binding was detected in either study. P. 47 Laufer and Greenwood (1969) confirm that Cecropia juvenile hormone, like the farnesoic acid derivative, alters puffing when fed to late prepupae of Chironomus thummi at a concentration of 3000 T.U./ml. After 6 hr treatment the activity of at least 7 puffs is altered and BR IVB not only fails to regress, as would normally occur at this stage of development, but actually increases in size. P. 52 The salivary gland chromosomes of Sciara ocellaris respond to microA.1.P.-4*

94

M. ASHBURNER

sporidian infection in a manner similar to that described above. In the fat body infection may result in coincident polyploidy and polyteny and the development of highly condensed “brachy polytene” chromosomes. Infected salivary gland cells survive metamorphosis whilst uninfected cells of the same gland proceed to their usual fate and histolyse (Pavan et al., 1969). P. 57 In the giant food pad nuclei of Sarcophaga bullata studied by Bultmann and Clever (1969) the peak puffing activity occurs on days 6-7 of pupal development. The autonomy of the timing of puffing patterns of the foot pad nuclei and the implication of this finding for theories of the hormonal control of puffing is further discussed in this paper. p. 61 The salivary gland of Bradysia sp. is differentiated into four regions each unique with respect to chromosome morphology, patterns of nucleic acid synthesis, and cytoplasmic secretory inclusions (da Cunha et al., 1969). P. 64 The role of the BR in the control of secretory protein synthesis in the salivary gland of Chironomus tentans and C. palliduvitatus is considered fully by Grossbach (1969). P. 67 Grossbach (1969) claims that the conditions used by Laufer will not dissolve the major secretory proteins of C. tentans. P. 67 Martin e t al. (1969) studying patterns of protein synthesis in larval salivary glands and other tissues of Calliphora stygia were unable t o find any evidence for the large scale transport of proteins into the salivary gland. P. 68 Bultmann [Ph.D. Thesis quoted by Bultmann and Clever (1969)l finds that degenerating foot pad nuclei have a specific puffing pattern and continue to incorporate H3uridine. P. 70 Successful RNA-DNA hydridization in cytological preparations is reported by Gall (19691, Gall and Pardue (1969) and by John et al. (1969). The latter paper reports preliminary work by Jones and Robertson on the hybridization of RNA to Drosophila salivary gland chromosomes. Basile, R. (1969). Nucleic acid synthesis in nurse cells of Rhynchosciara angelae Nonato and Pavan, 195 1. Proc. Symp. Nuclear Physiology and Differentiation. Belo Horizonte. Genetics 61, Suppl. l , 261-273. Bultmann, H. and Clever, U. (1969). Chromosomal control of foot pad development in Sarcophaga bullata. I. The puffing patterns. Chromosoma 28, 120-135. da Cunha, A. B., Morgante, J. S., Pavan, C. and Garrido, M. C. (1969). Studies on cytology and differentiation in Sciaridae. 111. Nuclear and cytoplasmic differentiation in the salivary glands of Bradysia sp. Univ. Tex. Publs. 6918, 1-1 1. Daneholt, B., Edstrom, J-E., Egyhizi, E., Lambert, B. and Ringborg, U.(1969a). Physico-chemical properties of chromosomal RNA in Chironomus tentuns polytene chromosomes. Chromosoma 28, 379-398. Daneholt, B., Edstrom, J-E., Egyhlzi, E., Lambert, B. and Ringborg, U. (1969b). Chromosomal RNA synthesis in polytene chromosomes of Chironomus tentans. Chromosoma 28,399417. Daneholt, B., Edstrom, J-E., Egyhizi, E., Lambert, B. and Ringborg U. ( 1 9 6 9 ~ ) . RNA synthesis in a Balbiani Ring in Chironomus tentans salivary gland cells. Chromosoma 28,418429.

PUFFS

95

Egyhazi, E., Daneholt, B., Edstrom, J-E., Lambert, B. and Ringborg, U. (1969). Low molecular weight RNA in cell components of Chironornus tentans salivary glands. J. Molec. Biol. 44, 517-532. Emmerich, H. (1 969). Anreicherung von tritiummarkiertem Ecdyson in den Zellkernen der Speicheldriisen von Drosophila hydei. Expl. Cell Res. 58, 261-270. GabrusewyczGarcia, N. and Margles, S. (1969). Induction of DNA puffs by ecdysterone. J. Cell Biol. 43,41A. Gall, J. G. ( 1969). The genes for ribosomal RNA during oogenesis. Proc. Symp. Nuclear Physiology and Differentiation. Belo Horizonte. Genetics 6 1, Suppl. 1, 121-132. Gall, J. G. and Pardue, M. L. (1969). Formation and detection of RNA-DNA hybrid molecules in cytological preparations. Proc. natn. Acad. Sci. U.S.A. 63, 378-383. Grossbach, U. ( 1969). Chromosome-Aktivitat und biochemische Zelldifferentzierung in der Speicheldriisen von Carnptochironornus. Chrornosorna 28, 136-187. John, H. A., Birnsteil, M. L. and Jones, K. W. (1969). RNA-DNA hybrids at the cytological level. Nature, Lond. 223, 582-587. Kaplanis, J. N., Robbins, W. E., Thompson, M. J. and Baumhover, A. H. (1969). Ecdysone analog: conversion to alpha ecdysone and 20-hydroxyecdysone by an insect. Science, N.Y. 166, 1540-1541. Laufer, H. and Greenwood, H. (1969). The effects of juvenile hormone on larvae of the Dipteran Chironomus thurnrni. A m . Zool. 9, 603. Martin, M. D., Kinnear, J. F. and Thomson, J. A. (1969). Developmental changes in the late larvae of Calliphora stygia. 11. Protein synthesis. Aust. J. biol. Sci. 22, 935-945. Pavan, C., Perondini, A. L. P. and Picard, T. (1969). Changes in chromosomes and in the development of cells of Sciara ocellaris induced by microsporidian infection. Chrornosoma 28,328-345. Thomson, J. A., Siddall, J. B., Galbraith, M. N., Horn, D. H. S. and Middleton, E. J. (1969). The biosynthesis of ecdysones in the blowfly Calliphora stygia. Chem. Cornrn. 1969, 669-670. Weirich, G. and Karlson, P. (1969). Distribution of tritiated ecdysone in salivary glands and other tissues of Rhvnchosciara and Chironornus larvae. An autoradiographic study. Wilhelm Roux’ Arch. EntwMech. Org. 164, 170-18 1.

This Page Intentionally Left Blank

The Structure and Function of the Insect Dorsal Ocellus LESLEY J. GOODMAN Department of Zoology, Queen Mary College University of London, England Introduction . . . . . . . . . . . . . . . . . . 97 Distribution and Structure . . . . . . . . . . . . . 99 A. Distribution and Development of Dorsal Ocelli within the . . . . . . . . . . . . . . . . 99 Class Insecta B. Structure . . . . . . . . . . . . . . . . . 101 C. The Visual Field of the Ocelli . . . . . . . . . . . 131 D. Form Perception . . . . . . . . . . . . . . . 131 111. Behavioural Studies on the Role of the Ocelli . . . . . . . . 132 A. Early Work . . . . . . . . . . . . . . . . . 132 B. The Ocelli as Stimulatory Organs . . . . . . . . . 133 C. The Ocellar Contribution to Phototactic Orientation . . . 141 . . . . . . . 147 D. Detection of the Plane of Polarized Light E. Registration of Intensity Level and of Changes of Intensity . 148 IV. Electrical Activity in the Ocellus . . . . . . . . . . . 152 A. The Electrical Response of the Visual Cells . . . . . . 152 B. Electrical Activity in the Second Order Neurons . . . . 161 C. Sensitivity, Light and Dark Adaptation and Flicker Fusion Frequency . . . . . . . . . . . . . . . . . 164 D. The Spectral Sensitivity of Ocelli . . . . . . . . . . 170 V. Ocellar Units in the Brain and Ventral Nerve Cord . . . . . . 171 A. Ocellar Units in the Brain . . . . . . . . . . . . 171 B. Ocellar Units in the Ventral Nerve Cord . . . . . . . 173 C. The Effect of Ocellar Input on Compound Eye Units . . . 182 D. The Influence of the Ocelli on Motor Activity in the Thoracic Ganglia . . . . . . . . . . . . . . . . . . . 184 VI. Conclusion . . . . . . . . . . . . . . . . . . . 188 References . . . . . . . . . . . . . . . . . . . . . 190

I. 11.

I. INTRODUCTION

In addition to their large compound eyes adult insects typically possess simple cup-shaped eyes on the dorsal region of the head composed of a few hundred visual cells beneath a common lens. Known as the dorsal ocelli, they are of distinct ontogenetical origin 97

98

LESLEY J . GOODMAN

from the compound eyes and are innervated directly from the ocellar lobes located in the protocerebrum between the mushroom bodies. Generally three are present, a median receptor on the frons and two lateral receptors on the suture between the frons and the vertex, although one, two, four and more are characteristic of some species. In the Plecoptera all the ocelli are borne on the frons, in the Blattoidea they have become somewhat degenerate and are known as fenestrae, whilst in the higher insects they frequently form a compact triangle or row on top of the vertex. This article is concerned solely with the functions of the dorsal ocelli and excludes any consideration of other simple eyes such as the lateral ocelli or stemmata which may sometimes be retained in adult insects or certain atypical eyes, such as those of Pediculus, which have an ocelliform structure but whose relationship to the compound eyes or to the dorsal ocelli is in doubt. The role of the dorsal ocelli in insect behaviour has never been satisfactorily established. N o clue is afforded by examining their distribution within the Class and no very obvious differences of behaviour be tween ocellate and anocellate species have been reported. Once the optical system of a number of dorsal ocelli had been examined it became apparent that they were not important in form vision since the image formed by the lens system did not fall within the retinal layer. Behavioural studies confirmed this and showed their inability to mediate photic orientation reactions although it was increasingly noted that the accuracy of such reactions mediated by the compound eyes was reduced if the ocelli were occluded. Many authors found that the ocelli had a positive photokinetic effect on locomotory activity and they were thus classed as “stimulatory organs” under which title they are still briefly dismissed in many texts. Several electrophysiological studies have yielded interesting information on the way in which certain ocelli respond to light and on the type of information which they signal. Recently it has become apparent that the ocelli probably have distinct functions in different groups of insects. Behavioural and electrophysiological studies have given some insight into the way in which compound eyes and ocelli can interact to influence motor activity. However we are still far from understanding the full role of the dorsal ocelli as photoreceptors. It is hoped that by focusing attention on the many gaps in our knowledge of the physiological properties of rhe ocellus and by indicating current areas of interest the deficiencibs may be made good and the possibility of an ocellar

99

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

contribution may automatically be considered when any visual dependent behaviour in insects is being examined. 11. DISTRIBUTION AND STRUCTURE A. DISTRIBUTION AND DEVELOPMENT O F DORSAL OCELLI WITHIN THE CLASS INSECTA

Distribution of the dorsal ocelli within the Class is erratic. Some insect orders lack ocelli completely, others, such as beetles, have few ocellate species and within an order members of the same species, genus or family may be ocellate or anocellate. Ocelli are common amongst good fliers and, indeed, a connection between wings and ocelli has been made in insects (Priesner, 1928; Kalmus, 1948). If the orders of insects are grouped according to the presence or absence of ocelli and wings and the resulting table (Table I) treated as a Table I The distribution of wings and ocelli in the insect orders (from Kalmus, 1945) Wings Present Ocelli present

Ephemerop tera 0don at a 2

Ocelli present or absent

Neuroptera Trichop tera

Present or absent

Orthoptera Dermaptera Plecop tera Isoptera Psocoptera Thysanoptera

2 Ocelli absent

-

Hemiptera Mecoptera Lepidoptera Coleoptera Hymenoptera Diptera

Absent

Thysanura Collembola

12

2

2

Protura Anoplura Aphanip tera 3

Em bioptera Strepsip tera

contingency table, a strong association between wings and ocelli is indicated. N o insect order is completely ocellate and wingless or anocellate and winged, neither is there an order in which all species are ocellate but some wingless. The correlation can be seen very

100

LESLEY J GOODMAN

clearly in families with sexually dimorphic forms (Table 11). However, not only the ocelli but the complete visual system can be linked with wings since size and facet number in the compound eye is frequently associated with wing development. Kalmus has pointed out that n o single gene difference can be responsible for the presence or absence of wings and ocelli or for the good or poor differentiation Table I1 The distribution of wings and ocelli among 135 species of the Thysanoptera (from Kalmus, 1945). The terms macropterous, brachypterous and micropterous refer to the relative development of the wings '

Monomorphic forms

Dimorphic forms

Ocelli

Wings Monomorphic: Macropterous Micropterous Apterous Sexually dimorphic: Macropt. ? apt. d Macropt. ? micropt. d Macropt. and brachypt. ? apt. d Brachypt. ? micropt. d Dimorphic : Macropt. and brachypt. 0 Macropt. and brachypt. 0 and d Macropt. and micropt. ? Macropt. and apt. 0 Trimorphic: Macropt., brachypt. and micropt. 9

83 1

1 1

1 7

4 1 4

1 1

1

1

1 1 2

1

1

1

1

of wings and compound eyes; indeed he finds no chromosome mechanism which could affect the many co-variant intraspecific differences in wingedness and degree of eye development which d o occur. Since the full development of the ocelli and the wings and the

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

101

complexity of the lateral eye are all imaginal characters, Kalmus has suggested that neotony might provide an explanation of the wing eye relationship; an insect which showed some degree of neotony having less well developed compound eyes and wings and lacking, or having poorly developed, ocelli. Development of the ocelli does vary between groups which are good fliers, suggesting that they may have different roles in different insects. The Hymenoptera are the only order which have been examined in any detail from this point of view. Gotze (1927-28) found that the ocelli were normally larger and the lenses more convex in males. In night flying Hymenoptera where the compound eyes are rudimentary, the ocelli have also become much reduced and sunk into the cuticle. However, reduction of the compound eyes is not always accompanied by reduction of the ocelli, in Dorylus, for example, where the compound eyes are completely rudimentary, the ocelli are fairly well developed. In some Hymenoptera groups, Ichneumonids and Braconids for example, the ocelli appear particularly well developed in forms which fly at twilight having large, highly convex lenses (Fig. 1). This is not true of the Lepidoptera where ocelli are not often present in twilight and night flying moths and are notably absent in the swiftly flying Sphingids. Ocelli appear to be equally well developed in day flying and in dusk flying dragonflies, but no detailed studies have been made. In the Diptera they are commonly well developed in good daylight fliers but are oddly absent from the strongly flying Tabanids. An investigation of the relationship between the development of the ocelli and the flight habits of the insects might indicate a starting point in the search for the role of the ocelli within a particular group. B. STRUCTURE

1. Dioptric Apparatus The dioptric apparatus is commonly a transparent layer of cuticle above the layer of retinal cells which becomes thickened to varying degrees in different ocelli by the underlying layer of corneageneal cells. In some Collembola the cuticle remains thin and differs from the surrounding cuticle only in that it lacks pigment and so provides a transparent window for the entry of light to the few retinal cells lying underneath. In all Pterygote ocelli the cuticular window is thickened to some extent by additional chitin laid down by the corneageneal cells (Fig. 3). In the Blattoidea, whose ocelli are regarded as degenerate, the cuticular window is simply thickened

102

LESLEY J . GOODMAN

K

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

103

(Fig. 1 ) but more commonly it forms a domed shape or a planoconvex or biconvex lens as in Schistocerca gregaria, Lucilia sericata, Zygaena sp. and the majority of Lepidoptera with ocelli (Figs 1 and 2). After secreting the lens the corneagenous cells frequently become reduced to a thin transparent layer which alone separates the cuticular lens from the retinal cells beneath (Fig. 1). In other ocelli the corneagenous cells provide an auxiliary refractory organ which can take a variety of forms. In Vespa crabo there is a crystalline body beneath the lens. In Schistocerca and Lucilia the cup-shaped region between lens and retinal cells is filled with large, elongated vitreous cells (Figs 2 and 4) or more rarely with a vitreous fluid. Occasionally a cuticular lens is absent and a crystalline body, formed by the corneagenous cells, lies beneath a thin, transparent cuticle as in Cloeon (Fig. 1).

2. The Screening Pigment Dark screening pigment is entirely absent in some ocelli as in Periplaneta americana (Ruck, 1958a) and variously distributed in others. In S. gregaria the pigment granules are confined to the peripheral cells of the corneagenous layer just above the distal ends of the retinal cells. This thin ring of pigment cells, some six to seven cells thick and ten to twelve cells deep, is called the “iris” since it is presumed to limit the angle of light acceptance of the ocellus. Within the pigment cells of S. gregaria formation of vesicles from the endoplasmic reticulum can be observed (Fig. 5). The vesicles increase in size as material is deposited in them until they reach the order of 0.6 p, when they appear densely packed with what is believed to be Fig. 1. A-E. A, a-f shows differences in the development of the ocelli in certain Hymenoptera. (From Gotze, 1927.) (a) Astutu sp. the lateral ocelli, though well developed, are triangular in shape and have no frontal view. (b) In Apis sp. all three ocelli are particularly well developed in drones. (c) In Tuchysphex pectinepes the median ocellus is reduced and the lateral ocelli are spindle shaped and flattened. This has been carried even further in (d) Tuchytes europaeu where the lateral ocelli form a flat crescent shaped slit. (e) Moneduh sp. have large, circular ocelli with a covering of fine hairs. In Steniolu sp. (f) lateral ocelli are much reduced and the elongated median ocellus forms a slit parallel to the long axis of the head. B, C and D show differences in the dioptric apparatus of ocelli. In B there is simply a flattened transparent cornea above the layer of retinal cells in Periphnetu umericunu. (After Ruck, 1957.) C. Aphrophoru spumariu possesses a well developed corneal lens (1) with a thin layer of corneageal cells beneath (c.n.) seperating it from the retinal layer (rh). D. CZoeon lacks a cuticular lens having a thin transparent layer of cuticle (c) overlying a crystalline body (1) secreted by the underlying corneagel cells (v). E illustrates the partial decussation which takes place as the ocellar nerves enter the brain, in this case in the Hymenoptera. (After Miiller, 1931) L 0, Lateral ocelli. M 0, Median ocellus. TI, Trachea. Oe. Oesophagus. P, Corpora pedunculata.

Fig. 2. L.S. through the median ocellus of a newly fledged locust, Schisrocercagregaria. x 240. ch. transparent chitin being laid down by corneageal cells (see Fig. 3) will gradually thicken to form the corneageal lens. cor, corneageal cells v. elongated vitreous cells with prominent nuclei. p.c, pigment cells. ret, visual layer showing retinular cells at different depths. tap, tapetal layer. ret. ax, axons of retinula cells descending through the tapetal layer. memb. comp, membrane complex. fib. cap, fibrous capsule. oc, ocellar nerve. as, air sacs. (From Goodman and Kirkham, 1970).

Fig. 3. Formation of the cuticular lens in the median ocellus of Schistocercugregaritz. c, cuticle. cor, corneageal cells. mv, microvilli of corneageal cells lying beneath newly secreted granular endocuticle. g, golgi apparatus. gr., granular material. m. tub., microtubules. x 30,000. (From Goodman and Kirkham, 1970.)

Fig. 4. Cells from the vitreous body lying between the lens and the retinal layer in the median ocellus of Schistocercu greguria. v, very large fluid filled vitreous cells. cell memb., convoluted cell membrane. nuc, large nucleus. Nuclei of these cells are concentrated in one zone (see Fig. 2). x 12,000 (From Goodman and Kirkham, 1970.)

Fig. 5. Ocellar pigment cells in S. gregariu. P.c., pigment cell. v., neighbouring vitreous cells. p.p.g, primary pigment granule. d.p.g, developing pigment granule. m.p.g, mature pigment granule. er., endoplasmic reticulum. nuc., nucleus. x 9000. (From Goodman and Kirkham, 1970.)

I08

LESLEY J. GOODMAN

melanin (Goodman and Kirkham, 1970). In some ocelli the retinal cells themselves contain dark pigment granules usually proximal to the rhabdomere region, as in Helophilus and Syconastes marginatus (Hesse, 1901) and Zygaena (Link, 1909). Others have such granules present in the sheath cells through which the retinal axons run as in Sympetrum (Ruck and Edwards, 1964). In Cloeon, Ceratopsyllus canis and Agrion a densely pigmented epithelium invests the whole eye and a rapid movement of the pigment in this sheath has been reported in response to changing light intensity (Lammert, 1925). Pigment migration has not been reported in other ocelli, certainly no migration of granules or movement of cells has been observed in the pigment ring of Schistocerca in response to changes of light intensity. The principal function of the pigment granules of the ocelli thus seems to be to prevent light from entering the photoreceptor except through the frontal lens aperture.

3. The Retinular Cells There are commonly between 500 and 1000 retinular cells lying in a shallow layer or arranged around a cup-shaped depression beneath the dioptric apparatus. The cells may occur at different depths within the retinal cell layer (Blaberus craniifer, Fig. 1, Schistocerca, Fig. 2 ) but in no case do they appear to be organized into separate distal and proximal layers as was once suggested for the dragonfly, Agrion, by Hesse (1901). The unipolar sense cells are grouped together in twos, threes, fours and fives, the number can vary within a given ocellus. The proximal ends of these primary sense cells narrow into short axons which synapse with the dendrites of second order neurons whose cell bodies are situated in the pars intercerebralis of the brain. Some portion of the inner limiting membrane of each cell is modified to form a rhabdomere. In insect eyes there is a certain amount of variety in the relationship of the rhabdomere to the photoreceptor of which it is a part. In ocelli the rhabdomere is commonly a narrow V-shaped strip on the inner surface of the cell embedded within the cytoplasm (Schistocerca, Fig. 6). Usually this strip extends to the distal limit of the cell and in Fig. 6. T.S.through the visual cell layer of the light adapted median ocellus of S.gregurk showing retinular cells. rh., rhabdomere. mit., mitochondria. ax., group of four retinular cell axons descending from a higher level surrounded by a glial sheath, gls. ur. sp., urate spheres within the glial membranes. I . er., rough endoplasmic reticulum, probably site of retinular cell axon formation. g, golgi apparatus, generally found between the nucleus and the cell membrane in a position removed from the rhabdom. n., nucleus. x 8000. (From Goodman and Kirkham, 1970.)

110

LESLEY J GOODMAN

a number of cases the rhabdomere completely surrounds the distal end of the cell, Eristalis, Syrphus (Redikorzew, 1900) and Drosophila (Hertweck, 193 1 ). In cross-sections through the distal region of these ocelli the rhabdomeres have the appearance of a continuous hexagonal mesh. Similar effects are seen in some insect compound eyes and in spider eyes (Ruck, 1964; Curtis, 1969). The syrphid Helophilus, has retinal cells with the rhabdomeres confined solely to the distal margin of the cell in one region of the ocellus whilst in other regions the rhabdomeres form “sleeves” around the distal end of the cell open at both ends. The significance of these differences in rhabdomeric disposition is not known, possibly where the rhabdomere surrounds the distal end of the cell this may serve to increase the sensitivity of the photoreceptor. In those ocelli so far examined the retinular cells form compact groups with the rhabdomeres fused and forming a central rayed rhabdom (Fig. 6). The rhabdomeres resemble those of other arthropod eyes being composed of closely packed parallel microvilli, finger-like evaginations of the cell membrane. In Schistocerca the microvilli are between 0.4 and 0.5 p long with an overall diameter of 900 A” and an internal diameter of 300 A ” , (Fig. 8). No basal retinular cells lacking or having a very small rhabdomere have been found in E.M. studies. A central process is seen occasionally in the centre of the retinulae of Schistocerca which is believed to be a downward extension of the vitreous cells between the distal ends of the retinular cells (Fig. 7). The cell nuclei are large ovoid structures situated distally in Schistocerca and proximally in Libellula. The cells are comparatively rich in cytoplasmic organelles including numerous mitochondria, endoplasmic reticulum, multi-vesicular bodies, Golgi complexes and “onion bodies”.

4. Membrane Systems Within the Cell At present a threedimensional picture of the cytoplasmic contents of the retinular cells is lacking and there are conflicting reports on the disposition of membrane systems within the cell, probably due in part to real differences between the eyes studied. Endoplasmic reticulum is conspicuous in the distal rhabdomere bearing ends of the Fig.7 . T.S.through a group of four light adapted retinular cells from the median ocellus of S. greguria showing the rhabdomere regions. rh, rhabdom. int. sp, intercellular space, possibly a downward extension of the vitreous cells. er, endoplasmic reticulum occurring near the microvilli of the rhabdomeres and extending through the cell towards the site of the developing axons. ax. st, site of developing axon. mit mitochondria x 15,000. (From Goodman and Kirkham, 1970.)

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

111

Fig. 8. Rhabdomere region of dark adapted retinular cells from the median ocellus of S. gregaria. rh, rhabdom cut longitudinally. mit, mitochondria. cist, cisternae, fluid filled intracellular cavities. x 20,000. (From Goodman and Kirkham, 1970.)

Fig. 9. T.S. through the retinular cell region of a dark adapted ocellus showing the palisade layer of cisternae adjacent to the rhabdomeres (cf. Fig. 7). rh, rhabdom. pl, palisade layer. mit, mitochondria. ax, descending retinular cell axon. glm, glial membranes. x 12,000. (From Goodman and Kirkham, 1970.)

114

LESLEY J . GOODMAN

cells in both ocelli and compound eyes. Ruck and Edwards ( 1 964) have suggested that there is a connection between the endoplasmic reticular system and the extracellular space surrounding the cell in the ocellar retinular cells of the dragonfly, Libellulu. They were able to trace pairs of membranes in the endoplasmic reticulum some few microns away from the rhabdomere delimiting channels which extended uninterruptedly for some distance in longitudinal sections. Below the level of the rhabdomere the channels, whose widths are reported as varying from between 100 A"-1 500 A" and more, wrap around the interior of the cell like concentric cylinders. This system extends as far as the region of the origin of the retinular cell axon where it becomes restricted to the area just inside the outer limiting membrane. Here Ruck and Edwards find difficulty in making a distinction between the channels of the endoplasmic reticulum and extracellular space and they believe that the channels are in fact confluent with extracellular space. Ruck ( 1 964) cites as a further example in which internal cytoplasmic membranes appear to be confluent with conventional extracellular spaces the membrane bound tubules of Dissosteiru, Anux and Apis, tentatively described as ultra-tracheoles by Fernandez-Morin ( 1958). In the skipper butterfly these tubules, approximately 250-300 A" in diameter, extend from the rhabdomeres through the cytoplasm to the outer membrane of the retinular cells through rows of mitochondria. In Musca similar structures are reported to connect the rhabdomeres with spaces between contiguous retinular cells. Since the microvilli of the rhabdomeres are open to the cellular cytoplasm at 'their basal ends Ruck claims that the membrane system derived from the endoplasmic reticulum could put the rhabdomeres in close proximity to a significant volume of extracellular space within the cell. In many cases extracellular space is very limited around insect photoreceptor cells. In the ocellus of Blaberus the outer limiting membranes of neighbouring cells are only about 100 A" apart and Ruck has questioned the sufficiency of this space as an ion store for maintaining a prolonged positive ion entry phase and suggests that the intracellular membrane system could greatly supplement the volume of extracellular space and might prove the most likely site of the generator potential. A number of objections can be made to Fig. 10. Site of retinular cell axon formation. er, endoplasmic reticulum forming the site of axon development. Note proximity to rhabdomere and to extracellular space. f.rib, free ribosomes. at.rib, attached ribosomes. des, desmosomes. x 30,000 (see also Fig. 11, A, B and C). (From Goodman and Kirkham, 1970.)

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

117

Ruck’s proposals. In the ocellar retinular cells of Schistocercu the site of axonal development is clearly marked by concentric rings of endoplasmic reticulum (Figs 10, 1 1 A, B and C) but although the channels appear to be confluent in some cases with larger intracellular vacuoles or lacunae no continuity with extracellular space can be found (Fig. 1 1 C) and no such continuity has been reported from other studies of retinular cells. Possibly treatment with ferratin might establish whether such continuity does in fact exist. Even if continuity existed it is difficult to see that it could be an effective way of increasing the useful volume of extracellular space since it is the volume per unit area of membrane which is of significance in the ionic exchange and this is not increased by proliferating relatively narrow tubules within the cell. In any case in view of a comparable lack of extracellular space in other parts of the central nervous system it is probable that Ruck’s doubts about the sufficiency of available space between the retinal cells are unfounded. Horridge and Barnard ( 1965) have provided an alternative suggestion as to the function of some of the larger vacuoles or lacunae present in the retinular cell cytoplasm including those of Dissosteiru and Muscu. They have reported a zone of vacuoles, 2-4 p across, in the cytoplasm adjacent to the rhabdomeres in the dark adapted compound eye of Locustu forming a clear zone or palisade around the rhabdom and they have found instances in which the central spaces of these vacuoles are continuous with the space between the paired membranes of the endoplasmic reticulum. In the light adapted eye the lacunae are distributed in the cytoplasm and the mitochondria are crowded close to the rhabdomeres. The same effect has been noted in the median ocellus of Schistocercu (Figs 6 , 8 and 9). These endoplasmic lacunae or cisternae appear to be filled with fluid and because comparisons of cross-sections show that the volume of lacunae plus palisade remains constant Horridge and Barnard believe that the lacunae move centrally towards the rhabdomeres in bulk as the eye becomes dark adapted, displacing the ~~~

~

Fig. 1 1 . A, L.S. through the base of a retinular cell. ax, retinular cell axon descending through tapetal layer. tap, tapetal layer. rh, rhabdom. nt., neurotubules. x 8000. B, Oblique section through lower portion of a retinular cell. er, concentric rings of endoplasmic reticulum surrounding early stage axon material. x 8000. C. Oblique section through a retinular cell at a lower level showing concentric rings of endoplasmic reticulum, er, surrounding axon material which includes mitochondria, mit. Space bounded by endoplasmic reticulum membranes appears to be in continuity with intracellular cistemae. x 1800. (From Goodman and Kirkham, 1970.) A.1.P.-5

118

LESLEY J . GOODMAN

mitochondria towards the periphery of the cell. A loose structure of lacunae is thus formed around the rhabdom transversed by a thin cytoplasmic bridge connecting the rhabdomeres with the rest of the cell. Horridge and Barnard suggest that the clear zone around the rhabdomeres of Dissosteira and Musca in fact represent palisade layers. No mechanism for palisade forma tion and mitochondria1 movements has yet been suggested. The effect of palisade formation in the dark adapted eye is to bring about an increase in sensitivity. Measurements of the refractive index in the region immediately surrounding the rhabdom in light and dark adapted eyes of Locusta indicated that the palisade layer must increase the difference in refractive index between the rhabdom and the surrounding cytoplasm, thus tending to keep light within the rhabdom by internal reflection. The angle of acceptance of light of individual cells is increased so that in the dark adapted eye more light enters and is kept within the rhabdom. In the light adapted eye the palisade is replaced by a layer of mitochondria having a high refractive index similar to that of the rhabdom, thus tending to prevent internal reflection and to decrease the angle of light acceptance of the cell. Horridge and Barnard point out that these changes would scarcely be sufficient to account for the changes in sensitivity of three to four orders of magnitude that accompany light and dark adaptation in the compound eye of Locusta. The limited amount of pigment movement that occurs in response to illumination would probably also be insufficient. The change in sensitivity accompanying adaptation of the ocelli of Schistocercu is at least as great as that in the compound eye of Locusta but no changes in cellular organization which might promote sensitivity changes have been noted other than palisade formation. 5. Multivesicular Systems Spherical multivesicular bodies containing many small vesicles of the order of 500-600A" have been found in various types of arthropod retinular cell and the ocelli are no exception (Fig. 6). Fig. 12. A, B and C show the multivesicular system found within the glial membranes. A. ret, retinular cell. er, endoplasmic reticulum. glm, glial membranes. ax, retinular cell axons. nt. neurotubules. mit mitochondria. x 11,000. B. shows the glial membranes associated with the ends of the rhabdomeres. glm, glial membranes. des, desmosomes. er, endoplasmic reticulum within the glial membranes adjacent to the rhabdomeres. rh, rhabdom. x 3400. C. rh, rhabdom. des, desmosome. glm, glial membranes. ax, axon. ves, vesicles within the glial membranes (see also Fig. 13). x 48,000. (From Goodman and Kirkham, 1970.)

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

121

Though commonly bounded by a membrane, groups of vesicles may sometimes be seen lying free in the cytoplasm and individual vesicles may be dispersed throughout it (Fig. 8). The multivesicular bodies are usually closely associated with the rhabdomeres, indeed Waddington and Perry (1 96 l), who first reported them in Drosphila, suggested that they participate in the formation of the membrane of the rhabdom tubules; later workers however have not supported this view (Baccetti and Bedini, 1964; Rutherford and Horridge, 1965; Trujillo€en6z, 1965). Curtis (1 969) has recently found both vesicular and tubular elements forming a peri-rhabdomeric reticulum in the retinular cells of the phalangid, Mitopus morio, with vesicles occurring within the microvilli. He has demonstrated the presence of a presumed acetylcholinesterase in these vesicles and suggests that acetylcholinerase activity in vesicles so closely associated with the rhabdomere indicates that the ace tylcholine/acetylcholinesterase system is involved at some stage in the transduction of photic stimulation into a receptor potential. In Schistocerca an interesting association between the membraneous sheaths surrounding the visual cells and the rhabdomeres has been observed. The lower half of the retinulae and their axons are supported by a meshwork of cells having the extensive and intricate shape characteristic of glial cells (Fig. 14). The glial cells continue upwards to form membraneous sheaths around the retinular cells. They extend between adjacent cells to the ends of the rhabdomeres, becoming intimately associated with the retinular cells at this point (Figs 6, 12 A, B and C, and 13). Frequently a desmosome region is seen between the retinular cell membrane and the bounding membrane in the region of the rhabdomere. The bounding membranes adjacent to the rhabdomeres can be traced as far as neighbouring bundles of axons where they form part of the glial sheath around the bundle. The bounding membranes can be seen to enclose endoplasmic reticulum at the base of the rhabdomere in some regions (Fig. 12 B) and in others small, membrane bound vesicles. Vesicles of the same order of size may be seen within the adjacent microvilli (Fig. 12 C), and in large numbers enclosed within the membranes of the supporting cells (Figs 12 A, B and C, and 13) extending down through the tapetal layer, synaptic layer and within Fig. 13. H.P. section through retinular cell axons descending between retinular cells. ret, retinular cell. ax, axon. n.tub, neurotubules. ob, onion body. glm, gllal membranes. yes, unit membrane vesicles within the bounding glial membranes of the same order of magnitude as those seen near the ends of the rhabdoms. x 60,000. (From Goodman and Kirkham, 1970.)

STRUCTURE A N D FlJNCTION OF INSECT DORSAL OCELLUS

123

the supporting cells of ocellar nerve (Fig. 18). The close association of the bounding membranes with the rhabdomere regions at the junction of retinular cells is a constant feature between each pair of cells and it may be that it is purely of a supporting nature since arrays of vesicles occur in other supporting cells in insects. However the presence of such a multivesicular system, extending from regions adjacent to the rhabdomere down t o and along the ocellar nerve, is possibly of significance in view of the belief of some authors that the ocellus has a neurosecretory function (Bullock and Horridge, 1965). 6. ‘‘011 ion Bodies ” Spherical organelles composed of concentric membranes occur in the peripheral cytoplasm of ocellar retinular cells as in other arthropod eyes. In the ocellus they are also found in the retinal cell axon (Fig. 13). In Locustu compound eyes they are of the order of 4-8 p but in Schistocercu ocelli they are much smaller being approximately 0.4 1.1 in diameter. Their function is entirely unknown. I t has been suggested (Bernstein, 1961) that they might be receptor organelles for pigment migration or other possible cytoplasmic changes occurring during adaptation, but Horridge and Barnard were unable to detect differences in “onion bodies” of light and dark adapted locust compound eyes.

7. The Retiniilur Cell Axon The retinular cell axon arises towards the base of the rhabdomere bearing region of the cell. It is first noted in E.M. sections as a region of the cell in which the endoplasmic reticulum becomes increasingly prominent and finally appears to be arranged in concentric circles of double membranes (Figs 10 and 1 1 A , B and C). Deeper in the cell both neurotubules and neurofibrils become visible within the cylinder of double membranes. Ruck and Edwards state that the substance of the neurotubules appears to be in continuity with membranes of vesicular portions of the endoplasmic reticulum in Libellula. As the differentiated axon cylinder descends through the retinular cell it leaves the proximity of the rhabdomere and comes to lie adjacent to the opposite wall of the cell. In Schistocerca the axon runs centrally through the supporting layer of glial cells that invests the lower margins of the retinular cells and between retinular cells of a lower level (Figs 6 and 1 1). The axons from each retinula come to Fig. 14. Section through the tapetal layer of the median ocellus of S. gregariu. ax. gr, group of retinular cell axons surrounded by @idcells, glm, containing urate spheres, ur. sp. x 30,000. (From Goodman and Kirkham, 1970.)

124

LESLEY J . GOODMAN

lie side by side and are invested with a common glial sheath. In some cases individual axons appear to have their own sheaths, in others a single glial cell appears to invest the separate axons of a bundle, probably both conditions occur. In Libellula the sheath cells surrounding the retinular axons include pigment cells (Ruck and Edwards, 1964). Although the axons of a retinula descend together as a group through the tapetal layer they d o not necessarily remain together until the synaptic region is reached. In many cases one or two of the fibres will split off from the bundle and join with another descending bundle of axons. Some of the descending bundles themselves join to form larger bundles. Groups of two or more axons occasionally split off from these large bundles and join others. No regularity in the arrangement and synaptic connections of the descending fibres such as is seen in the descending fibres of some compound eyes (Calliphora, Trujillo€en6z, 1965) can be found in ocelli (Goodman and Kirkham, 1970).

8. The Supporting Cells and Tapetal Layer The proximal region of the retinular cells, the axons and the dendrites of the synaptic region are supported and sheathed by glial cells. In many ocelli there is a tapetal layer at this depth forming a white reflecting backing to the sensory cells. In Schistocerca this takes the form of the inclusion of spherical granules within the glial cells packed tightly within their convolutions and extensions (Figs 11 and 14). It appears that the tapetal material is composed of granules of urate since the tapetal layer gives a positive reaction when stained for uric acid and urates. This reflecting layer is 40-5Op deep around the base of the ocellar cup in Schistocerca and is extended upwards at the sides of the cup by the presence of granules and crystals, resembling those of uric acid in structure, in the peripheral corneageal cells. This region, approximately 10 cells thick, lies outside the supporting layer and extends from its limits upwards as far as the peripheral ring of dark pigment. The large membrane complexes situated just within the lateral bounding membrane of the ocellus, adjacent to the tapetal layer, are the probable sources of the granules within the glial cells (Goodman and Kirkham, 1970). They consist of concentric layers of double membranes having the Fig. 15. Cells, containing what appear to be urate crystals, u.r.c, apparently extending the tapetal layer from the region of the supporting cells to the pigment ring. x 12,000. (From Goodman and Kirkham, 1970.)

A.1.P.-5*

126

LESLEY J. GOODMAN

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

127

appearance of endoplasmic reticulum. Unit membrane vesicles appear to be formed within the double membranes of the complex and to pass to the outside where they are enclosed within the glial membranes (Fig. 16). N o connection has been seen between the membrane complex and the peripheral corneageneal cells containing urate material. Since urate and uric acid deposition in epidermal cells is known in locusts (Bursell, 1967) and these corneageal cells are adjacent to haemolymph channels in the tissue supporting the ocellus it seems likely that deposition occurs directly here. The membrane complex may in fact be a mechanism for introducing urate material into the relatively inaccessible glial tissue. In some ocelli, notably those of the Diptera, the tapetal layer is formed by the presence of numerous tracheoles present in the supporting cells which form tracheoblasts, in others there is a combination of tracheoles and granules. In Libellula and Sympetrum the retinular cells themselves are invested with tapetal sheath cells (Ruck and Edwards, 1964). The distal ends of the tapetal cells bearing the nuclei invest the distal ends of the retinular cell and extend over the major part of the cell. The tapetal reflecting material, again spherical granules probably of urate deposits, is concentrated between the distal end of the cell and the level of the retinular cell nuclei.

9. Connections with Higher Order Neurons The retinular cell axons are short, of the order of 100-200 p in Schistocerca, and synapse with second order neurons whose cell bodies lie in the pars intercerebralis of the protocerebrum in an anterior medial position. The ocellar nerve commonly contains a few large fibres and a number of very small ones. In the ocellar nerve of P. americana Ruck (1957) finds about 25 fibres, four large ones of diameter 6-10 p , 20 small ones, 0.5-2 p , and two intermediate in diameter. In the lateral ocelli of Sympetrum there are three “giant” fibres, the largest being 30 p in diameter and probably some smaller ones (Ruck and Edwards, 1964). Hoyle (1955) has reported six large axons in L. migratoria, two of 8 p diameter and four of 3-4 p. Although these six axons occupy only a small part of the total volume of the nerve cylinder no smaller axons could be found and Fig. 16. Section through part of the membrane complex lying at the base of the tapetal layer (see Fig. 2), showing vesicles within the glial membrane system, glm, developing into urate spheres. d.ur.sp. x 7500. (From Goodman and Kirkham, 1970.)

Fig. 17. A and B. The synaptic region. A. g.ax, one of the “giant” axons of the ocellar nerve. ret.ax, dendrite of retinal cell axon. syn.reg, synaptic region. ves, vesicles within the giant axon. n.tub, neurotubules. x 40,000. B. g.ax, “giant” axon, ret.ax, retinular cell axon. syn.reg, synaptic region. x 43,000. (From Goodman and Kirkham, 1970.)

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

129

the rest of the space appeared to be occupied by fluid and cellular material resembling the sheath cells. In the desert locust, S. greguriu, there are five large axons of the order of 10 I.( in diameter, and a number of very much smaller fibres, about 0.5 I.( (Goodman and Kirkham, 1970, Fig. 18). These axons are invested by sheath cells, and the whole bundle is surrounded by a neural lamella. The large ocellar axons begin to branch before the nerve enters the ocellar capsule. Repeated branching into finer and finer “twigs” continues, the branches spreading out and coming into contact with the descending axons of many retinal cells from all regions of the ocellus. The retinular cell axons appear to have short lateral branches in the synaptic region. The connections of any one retinular cell have not been established but it does appear from silver stained sections of Schistocerca that twigs from more than one “giant axon” synapse with an individual retinular cell. Obviously with such a relatively small number of second order neurons and some 500-1000 retinular cells there must be a high degree of convergence. The connections of the smaller second order neurons have not yet been established. Since the ocellar first order synapse is believed to be inhibitory the nature of the synaptic vesicles is of some interest. At present all that can be said is that some synapses containing spheroid vesicles (S type), which are presumably excitatory (Uchizono, 1 9 6 9 , are certainly present. Work in progress suggests that synapses containing flattened vesicles (F type), presumably inhibitory, may also be present (see Fig. 17 A and B) but the detailed morphology of the synaptic region is still obscure. One interesting feature in Schistocerca is that occasionally junctions are found in which the second order neuron appears to be the presynaptic unit. The ocellar nerves enter the brain in the median antero dorsal region of the protocerebrum. A partial decussation takes place just above and in front of the protocerebral bridge and tracts from each lateral ocellus and the median ocellus pass back on either side to two neuropile regions in the dorsal protocerebrum, possibly the calyces (Muller, 193 1, Fig. 1 E; Jawlowski, 1958). Ocellar association areas and pathways of higher order neurons have not been described in detail but a branch from the ocellar tract on each side of the brain is reported to run in the posterior optic tract of the compound eye (Power, 1943). Ocellar fibres are also reported to descend in the ventral connectives of Drosophilu and Schistocerca as far as the thoracic ganglia although it is not clear whether these are third or higher order neurons (Power, 1948; Goodman, 1968). This somewhat sketchy histological picture is confirmed by behavioural

130

LESLEY J. G O O D M A N

Fig. 18. Part of a T.S.of the ocellar nerve of the median ocellus of S. greguriu. Showing four of the five “giant” axons. g.ax, “giant” axons. s.ax, small axon. glm, ghal membranes. tr, trachea. d.mit, degraded mitochondria. n.tub, neurotubules. x 9000. (From Goodman and Kirkham, 1970.)

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

131

and electrophysiological studies but a detailed study of the central connections of the ocellar nerves is needed. C. THE VISUAL FIELD OF THE OCELLI

It has been suggested that the ocelli might serve to increase the visual field of the insect by either having a wider field of view than the compound eyes or by viewing a part of the visual field unseen by the compound eyes. However measurement of the field of view of representative ocelli by reflection from the tapetal layer shows that although their field of view is wide it is generally not as wide as that of the neighbouring compound eyes. In L. migratoria, in the horizontal plane, for example, the field of view of the median ocellus is 104" ; that of the lateral ocelli is 137" and of the compound eye is 180" (Cornwell, 1955; Jander and Barry, 1968). In female Calliphora erytkrocepkala the field of view is 330" in a vertical plane for the two compound eyes and 206" for the three ocelli. Moreover in these insects and in Hymenoptera (Gotze, 1927-28), the principal axis and field of vision of the lateral ocelli is approximately the same as that of the compound eyes. The insect's frontal visual field is composed of the overlapping visual fields of the two compound eyes, the two lateral ocelli and the median ocellus. Behavioural studies have shown that this approximation of the visual fields of the compound eyes and the ocelli has a functional significance in the phototactic orientation of the insect. D. FORM PERCEPTION

Form perception depends upon the ability of the lens system to form an image, the location of the image relative to the retinal layer, little or no convergence on to the higher order neurons and the ability of the central nervous system to make use of the information received. Several workers (Hesse, 190 1 ; Caesar, 1913; Link, 1908, 1909; Homann, 1924; Wolsky, 1930; Parry, 1947; Cornwell, 1955) have determined the focal length of the lens system of the ocelli of insects from a number of different orders, including Panorpa communis, Agrion sp., Cicada concinna, Locusta migratoria, Zygaena sp., Pempkigus fraxini, Perla abdominalis, Helopkilus sp., Calliphora ery throcepkala, and Formica pratensis. In every case the principal focal plane was found to lie deeper in the eye than in the retinal layer. Despite some criticisms of the methods of measuring refractive indices of the lenses and depths of the retinal layers employed by earlier workers (Wolsky, 1930; Cornwell, 1955) corrected measurements still show the principal focal plane to fall just beyond the retinal layer. In Calliphora erytkrocephala for example, the focal length of the lens is 0.12 mm whilst the retinal

132

LESLEY J . GOODMAN

space lies between 0.04 mm and 0.10 mm below the outer lens surface. In addition to this there is a high degree of convergence on to the second order neurons, commonly 500-1000 cells converging on to a maximum of 25-30 fibres in the ocellar nerve. The ocellus appears not to be capable of form perception and this conclusion is borne out by behavioural tests. 111. BEHAVIOURAL STUDIES ON THE ROLE OF THE OCELLI A. EARLY WORK

Early workers suggested a variety of roles for the ocelli but they generally assumed the capacity for form vision and their speculations were not supported by experimental evidence. Kolbe (1 893), Hesse (1 908) and Link (1 909a b) concluded that they were used for the perception of distant objects. Muller (1 826), Lowne ( 1 870) and Lubbock (1889), on the other hand, thought them only to be capable of perceiving nearer objects and Lowne supposed them to be capable of stereoscopic vision as well. Demo11 and Scheming ( 1912) thought that the ocelli together could function to assess distance. After Homann (1 924) had demonstrated that the image formed by the cuticular lens and the corneal cells never falls on the retinal cells and had suggested that the main function of the ocelli was to register brightness changes, several authors began to compare the normal behaviour of insects in simple responses t o light stimuli with that of ocellar blinded insects. Bozler (1926) came to the conclusion that the ocelli influenced the speed of movement in Drosophilu and that they were functioning as stimulatory organs. Gotze ( 1927-28) noted that ocellar blinded bees showed a loss of tonus and a decrease in the accuracy of positive phototactic orientation. Muller (1 93 1 ) claimed that bees showed no light reactions at all when the ocelli were the sole visual organs. He too noted a decrease in phototactic orientating ability at all light intensities once the ocelli were covered and thought they chiefly functioned as auxiliary organs for phototactic behaviour. Wolsky (1930, 1931) also noted loss of tonus and of speed of walking in several insects when the ocelli were occluded and claimed their sole function was to act as stirnulatory organs. Subsequent authors have repeated this type of examination in more detail over a wider range of insects and a number of common observations have emerged from their work, namely: occluding the ocelli causes loss of tonus and an increased latency period before the insect will respond t o light stimuli; the ocelli have a positive

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

133

photokinetic effect at high intensities in the majority of cases; the ocelli are rarely responsible for initiating a motor reaction on their own but appear to participate in phototactic reactions mediated by the compound eyes. Only recently has the manner of interaction of the ocelli and compound eyes begun to be understood and the need to analyse all visual responses carefully to determine the exact contribution made by the ocelli been appreciated. B . THE OCELLI AS STIMULATORY ORGANS

Some authors hold the view that, in addition t o their specific sensory function, certain sense organs have an especial function to maintain the central excitatory state or tonus by sending continuous stimuli to the central nervous system and so raising the level of non-specific reactivity of the organism. Wolsky ( 1933) thought that such impulses from the “Stimulationsorgane” were “necessary for the production and maintenance of muscular tone and for the ability of muscles to contract normally. The former effect is known as tonus producing and the latter as the kinetic effect . . . they (the ocelli) are considered to be specific photokinetic stimulatory organs”. As recently as 1959, Medioni has claimed that the ocelli are stimulatory organs whose principal function is to exercise a general photokinetic action. Bullock and Horridge (1965) in their discussion of “S timulationsorgane” have pointed out that whilst the idea is possible it has not been demonstrated with certainty that any sense organ does act in this way. Those authors who claim that the ocelli act as stimulatory organs do so on the basis of the depression of activity which they have observed in many insects after occlusion of their ocelli. Measurement of depression of activity in this context has chiefly involved comparing either the rate of movement or the latency period of normal and ocellar occluded individuals. Cassier (1 962, 1965) and Dufay (1964) have compared the latency period of normal and ocellar blinded locusts and noctuid moths. The insects were placed in a beam of light and the time elapsing before they responded by walking or flying towards the light source was measured. In Locusta migratoria the latency period shortens slightly up to 50 Lux but above this point it increases again (Fig. 19(a)). When the ocelli are occluded the insects are not active at very low intensities. The latency shortens as the intensity increases and there is no reversal point at high intensities. If the median ocellus only is covered the latency is longer but otherwise follows the pattern of response of the intact insect with a reversal in the region of 50 Lux.

134

LESLEY J. GOODMAN \

\

600

500

,. 400

0

0 L 0

-

6 300 D

.-

a B 200

100 \

0

070.913 1.7 2 27 273 3.3

Log E (0)

/

/

60

005

L

1

L

1

0.1

05

1

5

Log E

(b) Fig. 19. (a) The latency period of adult male locusts before moving up a light beam. (From Cassier, 1965.) Ordinate, latency period in seconds. Absicca, Log. light intensity in Lux. -Ocelli intact. -0-0Median ocellus only covered with colourless varnish. . _ - -Median -ocellus only covered with opaque varnish. -.-.-.All ocelli covered with colourless varnish. -..-All ocelli covered with opaque varnish. (b) The latency period of Noctua pronuba before flying up a light beam. (From Dufay, 1964.) Ordinate, latency period in seconds. Absicca, Log. light intensity in Lux. - All eyes intact. - - - - - Ocelli only covered. Compound eyes only covered.

-..-..

- .- .-.-

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

135

In the Noctuids the latency period is independent of intensity when the compound eyes alone are functioning. When the ocelli are present as well the latency period gradually decreases with intensity but when the ocelli operate alone there is a reversal effect at higher intensities and the latency period lengthens again (Fig. 19(b)). Functioning ocelli clearly shorten the interval before Locusta and Noctua respond at low intensities but at high intensities they appear to have an inhibitory effect. Thus if latency period can be taken as a measure of the responsiveness of the insect then functioning ocelli apparently enhance responsiveness at low intensities and inhibit it at high intensities. From Cassier’s results it would seem that the change in the effect which they exert is triggered by the median ocellus. Many authors have examined the effect of covering the ocelli on the speed of walking or flying insects. The most recent of these results are summarized in Table 111. The effect of light on the speed of walking in intact insects varies. In some, Calliphora, bee, Schistocerca, Noctua, light intensity appears to have no effect on walking speed, in others, Periplaneta, Gryllus, Locusta, the walking speed is reported to increase as the intensity increases. In Calotermes and Drosophila the speed slows as the intensity increases. Goustard (1958) has shown that walking speed increases as light intensity increases in photonegative specimens of Blatella germanica but decreases in photopositive specimens. However, this behaviour is modified by experience and after a number of trials the speed of both photopositive and photonegative insects becomes independent of light intensity. Most authors comparing the walking speeds of intact and ocellar blinded insects have done so only when the insects were behaving in a positively phototactic manner. It is not known therefore whether the phototactic state of an insect can influence the result obtained. The effect of the light intensity upon the speed of flight is not necessarily the same as that as speed of walking for a given insect. Increasing intensity results in a reduction of flight speed in many insects (Richard, 1956). With a few exceptions occluding the ocelli is reported to reduce the speed of walking, climbing and flying. Where measurements have been made over a wide range of intensities occluding the ocelli produces little or no change of speed at low intensities but as the intensity is increased so the reduction in speed becomes larger (Fig. 20) (Goodman, 1968). The reduction in speed at a given intensity depends upon the degree to which the ocelli are occluded. Covering two ocelli produces a greater effect than covering one. Covering the

Table I11 Photokinetic behaviour. A summary of the results of the more recent measurements of the effect of ocellar or compound eye occlusion on the speed of locomotion Intact insects Author

Insect

Dim light

Bright light

Ocelli occluded

Compound eyes occluded

Dim light

Bright light

Dim light

Bright hght

Blind

Cassier (1965)

Locusfa

Walking speed independent of light intensity over three orders of magnitude

Reduced

Reduced

No effect

No effect

-

Jander and Barry (1968)

Gryllus

Run slightly faster at higher intensities

Reduced

Reduced

Reduced

Reduced (but less than in dim Iight)

-

Goustard (1958)

Bhtellu

MBdioni (1959)

Drosophilu

In photopositive state walking Increased speed decreases at higher intensities, in photonegative state it increases. With experience of tests it becomes independent of light intensity Reduced (walking speed)

~~

~

Reduced

Reduced

Increased

-

Richard (1950)

Calotermes

Walking speed varies with age. Speed decreases at higher intensities in older insects

-

Reduced

-

Reduced

-

BayramogluErgene (1964) (1965)

Schisrocerca

Flight speed increases very slightly at higher intensities

Reduced

Reduced

No effect

No effect

Cease flying

-

Climbing speed reduced

-

-

Anacridium aegyptium Mantis religiosa

-

____ ~~~~

Goodman (1968)

Schricker (1965)

Schistocerca

Walking speed independent of light intensity

No effect

Reduced

No effect

No effect

Reduced

Penplaneta

Walking speed increases at higher intensities

No effect

Reduced

No effect

Reduced

Much reduced

Calliphora

Walking speed independent of light intensity

No effect

Reduced

No effect

Reduced

-

Apis

Walking speed independent of light intensity

No effect

Reduced

No effect

No effect

-

-

-

-

Apis

-

Covering ocelli has no effect on speed of foraging flights

138

LESLEY J . GOODMAN

I- 5

--

.... .

=---a

Calliphora erythrocephola

Oal-

- --s--

0

--*,

-_

-a-

0

Intact

- - -sOcelli covered

D

65 .

Periplaneta arnericano

4 3.

2. 1 .

Light intensity log ft. lamberts

Fig. 20. The relationship between the real speed of walking and the light intensity in Calliphora, Apis, Schistocerca and Periphneta. Closed circles, all eyes uncovered; open circles. compound eyes only covered; crosses, ocelli only covered. (Goodman,unpublished.)

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

139

ocelli with black varnish is more effective than using a grey varnish which admits some light (Bayramoglu-Ergene, 1964). The path taken towards the light source by ocellar blinded insects is generally much longer and more tortuous than that of intact insects. Allowance has been made for this in most recent studies by marking the tracts taken and determining the real speed of the insects. Jander and Barry (1968) have recently shown that in bright light the compound eyes and the ocelli can act together to promote a more accurate directional orientation and indeed when the ocelli are covered the insects seem to require stationary “orientation intervals” from time to time as they move towards the light. Unless allowance has been made for these stationary periods misleading figures have been obtained for the speeds of ocellar blinded specimens. Obviously since the ocelli have been shown to be involved in phototactic orientation it is desirable that this method should not be used for comparing the rate of latency periods and movement of intact and ocellar blinded insects. It is interesting that, with one exception, occasions on which no decrease in speed was found when the ocelli were blinded are cases in which the insect’s speed has been measured either in its natural environment or in an experimental situation which has attempted to reproduce such an environment. Goustard ( 1 958) established a test situation resembling the normal environment of Blutella where the insects were free to move between small U-shaped cervices illuminated from above. Under these conditions he found that covering the ocelli increased the speed of walking. He does not indicate that repetitive trials produced any change in the result. Schricker (1965) comments that the speed of foraging bees is not reduced by covering the ocelli. After an initial drop of speed lasting for some minutes the flight speed of free flying specimens of Lucilia sericata is unaffected by ocellar occlusion (Goodman, unpublished observations). Where it has proved possible to compare the speed of movement of insects with the compound eyes blinded with those of intact and ocellar blinded individuals there has either been no change in the speed over a wide intensity range or else a reduction in speed at higher intensities although not usually quite so marked as the reduction in speed produced by covering the ocelli. Light either exerts no effect upon speed of movement via the compound eyes or else they appear to have an excitatory role at high intensities (Fig. 20). Proponents of the “Stimulationsorgane” theory interpret the

140

LESLEY J. GOODMAN

phenomenon of photokinesis as simply an increase in central excitability as a result of increase in light. Goustard (1958) found that the general activity of Blutellu was reduced by 16% if the ocelli were covered, by 8% if the compound eyes were covered and by 21% if both sets of photoreceptors were covered. Bullock and Horridge (1965) point out that the presence of a kinesis does not necessarily argue for a general stimulatory role as opposed to a sensory role for a sense organ. Demonstration that responsiveness to all kinds of environmental stimuli is lower in the dark and higher in the light would strengthen the case. This has not been shown for Blutellu. Evidence that there is a non-specific loss of sensitivity when the ocelli are occluded is necessary to demonstrate that they have a true stimulatory role. This aspect of the supposed stimulatory function of the ocellus has received little attention. M6dioni (1959) has compared the rates of geokinetic climbing in intact and ocellar blinded Drosophilu and found a reduction in the rate of climbing when the ocelli were covered. Mimura et ul. (1 969) have reported depression of spontaneous activity in certain antenna1 units in the brain of Boettcheriscu peregrinu when the ocelli are occluded. In other units however the spontaneous activity is increased (see Section V). The effect of ocellar illumination upon threshold of response to other stimulus modalities is not known. Although loss of tonus and slowed walking speeds have been so frequently reported as a consequence of ocellar blinding most authors have not noted whether there is any difference in the mode of stance or of the pattern of movement of the legs after occlusion. No alteration in pattern of wing movements has been reported. Only one author has described a resulting loss of muscular tone when the ocelli were covered. Schremmer (1950) points out that ocellar blinded bees show tonus disturbances when walking, the third pair of legs being dragged behind. This effect is very marked in the bee but it is also seen when the compound eyes are covered (Goodman, 1968). In fact the entire stance and pattern of leg movement is altered when one set of photoreceptors is not operating. The hind pair of legs are twisted backwards and outwards and the weight of the insect appears to be supported on the first two pairs of legs only. In view of this observation and the fact that ocellar stimulation is known to be able to influence motor activity in the thorax (Goodman, 1968) (see Section V), this aspect of ocellar function might well receive more atten tion. In summary it may be said that although the ocelli have frequently

141

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

been cited as examples of stimulatory organs none of the behavioural observations made have contributed conclusive evidence that these sense organs d o function in this manner. In view of the participation of the ocelli in phototactic orientation, discussed in the next section, they can no longer be regarded as having solely a photokinetic effect, indeed some of the apparent kinetic manifestations are possibly epiphenomena of the mechanism of compound eye and ocellar interaction discussed below. C. THE OCELLAR CONTRIBUTION TO PHOTOTACTIC ORIENTATION

000

It is commonly reported that phototactic orientation by means of the ocelli alone is impossible. However, the ocelli appear to make

CC

180"

180"

180"

Fig. 21. Orientation of a flying locust, S. greguria, suspended in a frame allowing it to rotate about its longitudinal axis, to 30 stimuli from a narrow light beam at position + C . A, Locust with all eyes intact. B, With ocelli only occluded. C, With compound eyes only occluded. (From Goodman, 1965.)

some contribution to positive phototactic orientation mediated by the compound eyes since ocellar occlusion results in a decrease in the accuracy of such orientation (Gotze, 1927-28; Miiller, 193 1 ; Chen and Young, 1943; Cornwell, 1955; Dufay, 1964; Cassier, 1965). The ocelli also contribute to accuracy of orientation in the dorsal light reaction (Mittelstaedt, 1950; Goodman, 1965) (Fig. 21). Cornwell was the first author to notice that the quality of the ocellar contribution to phototactic orientation depended upon the light intensity and the type of phototactic behaviour shown by the insect. In Culliphoru he found that the influence of the ocelli was manifest only under conditions of high light intensity and in Locusta only when the light adapted nymphs were walking directly towards a light source. Cornwell thought that the primary turning reaction was mediated by the compound eyes and directional deviations corrected by the ocelli during movement at high intensities. Cassier (1965),

142

LESLEY J . GOODMAN

when looking at the positive phototactic behaviour of Locustu over a wide range of intensities, found that in normal insects the accuracy of orientation improved gradually until the light intensity rose to 50 Lux. Above this level it decreased again. This falling off in accuracy of orientation at high intensities could be reduced if the median ocellus was partially occluded. It could be abolished entirely by completely occluding the median ocellus or partially occluding all the ocelli. Cassier proposed that the orientation behaviour of the locust over the whole range of intensities was due to an interaction between a gradual improvement in orientation performance as mediated by the compound eyes with rising intensity and the development of a photoinhibitory effect of the ocelli at higher intensities. He suggested that the median ocellus was particularly implicated in the development of ocellar photoinhibition at higher intensities. Jander and Barry ( 1968) have examined the ocellar contribution to the various manifestations of phototactic behaviour exhibited by the locust and the cricket, Gryllus bimaculutus, and have proposed a comprehensive theory of ocellar and compound eye interaction which embraces these earlier observations. Koehler (1950) has pointed out that the concept of “taxis” should not include progressive movements since locomotion and directional orientation are two independently varying processes although the majority of workers have not separated these two factors in their studies. Jander and Barry have measured the “turning tendency” (i.e. the resultant message of the directing mechanism) alone without a progressive component in Locustu and in Gryllus. Jander ( 1963) had previously shown that, in common with adult Orthoptera as a whole, locusts and crickets exhibit protaxis (i.e. they do not show constancy of angular orientation to directional stimuli), their orientation varying with the stimulus intensity. Prophototaxis can have two manifestations in that the strength of the turning tendency can grow or decline with increasing light intensity and the same individual may switch from the proportional to the inversely proportional relationship between stimulus and response. In both species Jander and Barry compared the turning tendency and the menotactic orientation of normal and ocellar occluded individuals showing either proportional or inversely proportional prophototaxis. Ocellar occlusion produced a decrease in the accuracy of orientation and a reversal of phototactic behaviour in those insects showing inversely proportional prophototaxis. The menotactic angle of intact insects showing the inverse relationship is significantly larger in bright light than it is in

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

143

dim light but after ocellar blinding it is significantly larger in dim light than it is in bright light. Loss of the ocelli increases the menotactic angle in dim light (Fig. 22). Orientation has thus become less accurate which implies that the ocelli were acting synergistically with the compound eyes at low intensities. In bright light loss of the ocelli decreases the menotactic angle. Orientation has apparently been improved which implies that the ocelli are acting in opposition to the compound eyes at higher intensities. Turning tendency measurements revealed the same reversal of phototactic \I/

\I/ - 'IO r

-0-

:a

m a t

-3OO

/I\

100 30'

-300

300

-300

30°

-60: a

s It 2 L . 2 O

o s= f 19.80

a

b

-300

30°

\

C

Fig. 22. Measurements of the menotactic angle of the locust, L . migratoriu, with intact (a) and (b) and completely occluded ocelli (c) and (d) (From Jander and Barry, 1968). The insects were walking on a horizontal surface towards a light source. In (a) and (c) the locusts used were behaving in a directly proportional tropotactic manner (see text) when unblinded and in (b) and (d) in an inversely proportional tropotactic manner. The menotactic angle was measured in dim light (filled circles) and in bright light (open circles). S equals the respective deviations of the direction of walking from the direction of the light (0'). Blinding the ocelli reduces the accuracy of phototactic orientation and in the case of the inversely proportional tropotactic insects causes a reversal of behaviour to directly proportional tropotaxis.

144

LESLEY J . GOODMAN

behaviour in ocellar blinded insects showing inversely proportional phophotactic behaviour (Fig. 23). Further information was gained by partial blinding of the ocellar group of ,the crickets whilst they were orientating phototropotactically. The crickets were allowed to orientate o n an horizontal platform between two parallel beams of light of equal intensity which intersected at 90". Intact insects orientated between the lights with an average angular value of 45". Blinding the median ocellus alone, the two lateral ocelli together or all three ocelli produced no significant difference in the angle of orientation.

oo o :

300

-3OO \

+C

I \

-6OO

O0o /

240

600 L 30

o 0

d

Fig. 23. Turning tendency measurements on the locust, L . rnigrutorziz, with the ocelli intact (a) and @) and completely occluded (c) and (d) (From Jander and Barry, 1968). In (a) and (c) the locusts were showing directly proportional prophototaxis, in (b) and (d) inversely proportional prophototaxis (see text). The locusts were placed on an evenly illuminated 30' slope with the light beam coming from the direction + 90'. Measurements were made in dim (filled circles) and bright light (closed circles), each circle representing three measurements. The appropriate m o w s indicate the average angular deviation from the vertical (=Oo). In the case of inversely proportional prophototactic insects ocellar blinding causes a reversal of behaviour to proportional prophototaxis.

However, blinding one of the lateral ocelli alone caused the insects to orientate more towards the light on the blinded side which suggested that the ocellus concerned normally acted in an antagonistic manner to its ipsilateral compound eye. Blinding the median ocellus reversed this effect and the insect orientated more towards the light on the unblinded side (i.e. the ocellus appeared to be acting synergistically with the ipsilateral compound eye). The median ocellus appeared t o have a switching action, reversing the type of interaction occurring

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

145

between the ocelli and the compound eyes. The role of the median ocellus was confirmed by determining the turning tendency of the crickets in a homogeneous white environment. In a white, diffusely lit arena the number of right and left turns made by intact crickets is not significantly different. If the right compound eye is covered, however, they make a significantly greater number of turns towards the right-hand blinded side. If, instead of the right compound eye, the right lateral ocellus is covered the insects turn more often towards the left. With the median ocellus intact the lateral ocellus and its adjacent compound eye are producing turning tendencies in opposite directions. If the median ocellus and right lateral ocellus are covered the cricket once more turns more often to the right-hand side; under these conditions the lateral ocellus appears to be working synergistically with its compound eye. In crickets showing the inversely proportional type of prophototaxis the amount of light falling on the median ocellus appears to control the type of interaction between a compound eye and its adjacent ocellus; in dim light they act synergistically, in bright light antagonistically. In insects showing proportional prophototaxis the ocelli and compound eyes appear to act synergistically at all light intensities. Jander and Barry have suggested that the mode of interaction of ocellar and compound eye inputs whilst the insect is exhibiting inversely proportional prophototaxis is consistent with a negative feedback system. They point out that such a system would have the advantage of countermanding the effects of any large scale changes of intensity and, since the ocelli are generally believed to be more sensitive to light than the compound eyes, the fact that they act synergistically with the compound eyes at low intensities would increase steering ability in the low intensity range. Jander and Barry’s suggested model of the functional connections necessary to achieve this feedback system, based on the observed behaviour of the insect and the electrophysiological evidence of the nature of the ocellar input, is illustrated in Fig. 24. The ocellar afferents run in series with those of the compound eyes. The first ocellar junction they suppose must of necessity be inhibitory because of the electrophysiological evidence. To make possible both a synergistic and an antagonistic coupling with the compound eye tracts the ocellar afferents must branch once and, since the compound eye afferents are known t o cross over, the ipsilateral ocellar branch should produce the antagonistic effect and the contralateral one the synergistic effect. The points of interaction are assumed to be excitatory. Switching is achieved by supposing that the median ocellar afferents add their effects t o those of the

146

LESLEY J GOODMAN

contralateral, synergistic branches. In dim light the weaker synergistic effect of these contralateral branches is strengthened by the activity of the median ocellus. In bright light, when the median ocellar activity is assumed t o be inhibited, the effect of the contralateral tracts is cut down or lost and the effect of the ipsilateral tracts predominates. No sites have been postulated for the points of interaction between signals from the optic centres and the

Fig. 24. Jander and Barry’s proposed model of the functional connections necessary to achieve negative feedback between the ocelli and compound eyes in phototactic orientation. E, excitatory connection points. I, inhibitory connection points. K, compound eyes. The direction of information flow in the feedback system is indicated by arrows. (From Jander and Barry, 1968.)

ocellar centres. A certain amount of evidence from electrophysiological recordings is now available (see Section V). Jander and Barry have observed that all the Polyneoptera with ocelli so far investigated, with the exception of Perla marginata, show an inversely proportional prophototaxis whereas in those without ocelli it is absent. It seems likely therefore that this type of feedback system occurs throughout the Polyneoptera where ocelli are present. They have suggested that this type of coupling might be widespread in insects possessing ocelli and that in this might lie their principal function. Cornwell’s work on Culliphora lends some support to this idea. If it is so then probably the photokinetic and phototactic functions of the ocelli should not be considered in isolation but as particular expressions of the operation of the feedback system A photokinetic feedback effect for example is seen in Drosophila (MCdioni, 1959) where light falling on the ocelli appears to increase the running speed and on the compound eyes, to inhibit it. In

STRUCTURE A N D FUNCTION O F INSECT DORSAL OCELLUS

147

Blatellu the effect is reversed, the ocelli inhibit, and the compound eyes stimulate, running speed (Goustard, 1956). The photokinetic effect of the ocelli which has received such prominence in many studies may well be an epiphenomenon of their phototactic role which until now has not been fully appreciated. Certainly in studies where the stimulatory effect of the ocelli has been estimated during positive phototactic orientation insufficient allowance has been made for the phototactic role of the ocelli. fi is likely that the mode of interaction between ocelli and compound eyes varies in detail between different groups and comparative studies would be of interest, particularly in insects such as cockroaches which lack a median ocellus. D . DETECTION OF THE PLANE OF POLARIZED LIGHT

The only report which suggests that ocelli might be capable of detecting the plane of polarized light is that of Wellington ( 1 953) on the behaviour of the fly, Surcophugu alrichi Parker. He compared the phototactic behaviour of intact flies; flies with only the compound eyes functioning; flies with only the ocelli functioning; completely blind flies and larvae of Neodiprion sertifer (sawflies) whose larval ocelli he considered to be comparable in structure t o the dorsal ocelli of the fly. Indoors his experimental arrangement consisted of a flat surface with an artificial light source with a 6-watt bulb at one side. Intact flies walked in a straight line towards the light source, flies with the compound eyes alone functioning walked more slowly and in a fairly straight line to the light. Flies with only the ocelli functioning behaved like blind flies and did not walk towards the light. The larvae of Neodiprion took a more circuitous route but did reach the light (Fig. 25 A). However, if the insects were allowed to walk on a flat surface out of doors in sunlight all flies, except the completely blinded ones, maintained a constant direction in relation to the sun when they walked (Fig. 25 A). The ocelli appear capable of evoking a motor response and of steering the insect in sunlight but not in artificial light. If a piece of Polaroid is held over flies whilst they are walking, with its axis parallel to the direction of the sun, they change direction and start to walk at approximately 90" to their previous direction (Fig. 25 B). Wellington states that there were only two consistent differences between the out of doors behaviour of flies with functional compound eyes and flies with functional ocelli. Those with functional compound eyes were very easily startled by movements nearby whilst those with only ocelli functional were indifferent to all but large and rapid changes of

148

LESLEY J . GOODMAN

intensity produced by such movements. When the sky was clouded flies with compound eyes intact seldom stopped walking, although they did not proceed in a directed manner, whereas those with only ocelli intact always stopped moving until the sun reappeared. In similar experiments with the Dipteran, Lucilia sericata, flies with ocelli alone operating were not able t o produce directioned movement in sunlight (Goodman, unpublished observations).

(A)

(e)

Fig. 25. A. Phototactic responses of adult Sarcophaga aldrichi and larval Neodiprion sertifer. Tracts A-E show their movements in a darkened room towards a 6-watt lamp, A’-E’ movements of the same individuals on the ground out of doors. Surcophaga with all eyes intact, A; with compound eyes only covered, B; with ocelli only uncovered, C; completely blind, D. A fourth instar larva of Neodiprion, E. Time marks on tracts indicate 10 second intervals. B. Effects of rotation of the axis of a sheet of “Polaroid” on the direction of walking of Sarcophaga and Neodiprion out of doors. Circles show the points at which the sheet of “Polaroid” was placed over a moving insect, or rotated, or withdrawn. Bars indicate the orientation of the axis with respect to the sun and darkened circles indicate that the sky was appreciably darkened when viewed through the “Polaroid” with the axis set as shown. Sarcophuga with all eyes intact, A; with compound eyes alone functioning, B; and with ocelli alone functioning, C. Fourth instar larva of Neodiprion, D. (From Wellington, 1953.)

The anatomy of the ocelli of Sarcophaga is not known in detail but unless it is very different from other Dipteran ocelli it is different to see how, even if the retinular cells of the ocelli were able to discriminate the plane of polarization, this information could be conserved beyond the junction with the second order neurons where there is such a high degree of convergence. E. REGISTRATION OF INTENSITY LEVEL AND OF CHANGES OF INTENSITY

Anatomical evidence suggests that the ocelli may be particularly sensitive to changes of light intensity and electrophysiological evidence has shown that the ocelli can signal information concerning both change and absolute level of intensity (Ruck 1961; Metschl,

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

149

1963; Autrum and Metschl, 1963). It is difficult to assess the level of performance of the ocelli in this respect from behavioural studies since they are seldom solely responsible for initiating motor activity. Their contribution t o photic behaviour which involves intensity discrimination can be determined in some cases. A startle reaction given by flying locusts, in which the ocelli and the compound eyes appear to act synergistically, suggests that the locust ocelli are very sensitive to decreasing intensity. When the light intensity is reduced S, gregaria shows a transient increase in flight speed, the size of which is related to the rate and amount of the decrease when both sets of photoreceptors are functioning (Goodman, 1968). The ocelli do not normally mediate this response but if they are occluded it is very much reduced. If the intensity changes are very small or very slow the locust will only respond if the ocelli are functioning together with the compound eyes (Fig. 26). Schricker (1965), investigating the orientation behaviour of the honey bee, found that in phototactic runs where bees could choose between two lights a significantly greater number of them chose the brighter light. Below 1 Lux only a two-fold difference of intensity is needed for this to occur. Bees with one ocellus blind need a four-fold difference, bees with two ocelli blind need a six-fold difference and bees with all the ocelli blind need an eight-fold intensity difference before a significantly greater number choose the brighter of the two lights. At intensities above 1 Lux the differences of intensity required are smaller in each case. The threshold intensity required for a phototactic run is lower (0.03 Lux) in normal bees than in bees with three blind ocelli (0.1 Lux). In the course of his experiments Schricker found that the ocelli play an important role in determining the onset and cessation of foraging activity in bees. Flight activity in bees is chiefly governed by two factors; the weather and the light intensity. The temperature, the quality and quantity of food available and the light intensity influence the start and end of foraging at dawn and dusk. The first and last of the daily flights is dependent upon the intensity level and Shricker has shown that bees must calculate very exactly the increase and decrease of intensity at dawn and dusk. They are also able to compensate for seasonal variations in the onset of dawn and dusk. If the ocelli are occluded foraging bees behave normally in most respects. Their flight speed, the frequency of their collecting flights during the day and the frequency of their waggle dance is the same as that of normal bees. If food of a lower concentration is A.1.P.-6

150

LESLEY J . GOODMAN

offered, bees with ocelli occluded, like normal bees, cease foraging earlier in the evening. Occlusion of the ocelli however does interfere with the timing of the first and last foraging flights. Bees with one, two or three ocelli occluded start to collect food later in the morning and cease collecting earlier in the evening than normal workers. The light intensity required for the first and last collecting flight is increased by a factor of two if one ocellus is covered, 3.3 if two ocelli are covered and 4.5 if all the ocelli are covered. Schricker concludes that the ocelli are providing the bees with information

a

Fig. 26. S. gregaria shows a transient increase in flight speed in response to a decrease of light intensity. The increase in speed is related to the rate (a) and the amount (b) of the decrease of intensity. The response is reduced by occluding the ocelli (closed circles) and greatly reduced by occluding the compound eyes (open circles). (From Goodman, 1968.)

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

151

concerning either the absolute brightness level or the rate of change of brightness at dawn and dusk. Somewhat unexpectedly there appears to be no connection between the perception of absolute light intensity or change of intensity by the ocelli and the use of light intensity as a trigger for time training. Ocellar blinded bees, trained to feed at a specific time of day, came to the food table with the same temporal deviation as normal bees. Schricker’s observations are of interest in that they reveal the use by the insect of the particular type of information provided by the ocelli in normal behavioural activity. No doubt other examples will be found if the possibility of an ocellar contribution is carefully assessed when the factors controlling specific insect behaviour patterns are being determined. It has been claimed that the ocelli are implicated in the entrainment of circadian rhythms. Cloudsley-Thompson ( 1 953) maintained, although without sufficient supporting evidence, that both the compound eyes and the ocelli are involved. Gautier and Vancassel (personal communication) have recently reported experiments with P. americana in which the factors influencing entrainment and maintenance of rhythm were examined separately both in rhythmic and arrhythmic insects. They claim that normally both the compound eyes and the ocelli together are responsible for entrainment and maintenance of rhythm but that either type of receptor can be sufficient. On the basis of these results it could be supposed that this was yet another example of reinforcing action by the compound eyes and the ocelli. However other authors have obtained conflicting results. Harker (1956), on the basis of experiments in which she severed the optic nerves in P. americana, leaving only the ocellar pathway intact, reported that the ocelli alone were “directly connected with the establishment of the rhythm by the external factors of light and darkness”. Roberts (1965) has presented evidence that the converse is true; that the compound eyes are the principal receptor, certainly sufficient in themselves. Lees ( 1964) has reported that photoperiodic induction can be effected by direct illumination of the protocerebrum in aphids. Since in the opinion of some authors (Bunning, 1936; Pittendrigh and Minis, 1964), photoperiodic induction is mediated by a circadian oscillation, this raises the possibility that the central nervous tissue responsible for circadian rhythmicity might absorb light directly, by-passing the compound eyes and ocelli. If this were so it might account for the discrepancies between the reports of different authors. This possibility has been considered in the careful studies of Nishiitsutsuji-Uwo and Pittendrigh (1 968a, b) on the cockroach,

152

LESLEY J . GOODMAN

Leucophaea madeirae. They locate the circadian clock within the optic lobes and believe that the circadian rhythmicity of locomotory activity is caused by an autonomous self-sustaining oscillation in the output of the optic lobes which causes a circadian periodicity of secretion by the pars intercerebralis. The pars intercerebralis secretion in turn imposes a circadian periodicity on the activity of the thoracic ganglia which control walking. They have failed to demonstrate entrainment of the pars intercerebralis by the direct action of light on the nervous tissues. The coupling of the oscillation in the optic lobes t o the environment is apparently effected by the compound eyes and optic nerves. This was demonstrated by the behaviour in a 24-hr light : dark cycle of insects in which the optic tract was severed on the left side and the optic nerve on the right. Although the left optic lobe was still coupled to the light cycle by an intact optic nerve it was unable t o drive the pars intercerebralis. The right optic lobe was still coupled to the pars intercerebralis and imposed rhythmicity upon it but severance of the optic nerve on the right side prevented entrainment of this clock by the environmental light cycle and the insect displayed a clear free-running circadian rhythm in spite of being in LD 12 : 12. A branch of the ocellar tract on each side of the brain is reported t o run in the posterior optic tract to the medulla of the compound eye in Drosophila (Power, 1943). If such connections are common to all insects possessing ocelli then the possibility that the ocelli might be able to drive a clock situated in the optic lobes exists. However, bilateral sectioning of the optic nerves showed that the presence of an intact ocellar pathway was not sufficient to entrain the rhythm. Surgical removal of both the ocelli in other cockroaches did not interfere with the entrainment of the clock, indicating that the ocellar pathway makes no contribution. At present the sum of the evidence suggests that the ocelli are not implicated in circadian activity, at least not in Leucophaea, but comparative studies would be of interest.

IV. ELECTRICAL ACTIVITY IN THE OCELLUS A. THE ELECTRICAL RESPONSE OF THE VISUAL CELLS

Extracellular recordings from the corneal and retinular cell layers and from the ocellar nerves have been made by several authors (Parry, 1946; Hoyle, 1955; Burtt and Catton, 1958, in Locusta, and

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

153

Metschl, 1 963, in Calliphora erythrocephala). The most comprehensive study has been made by Ruck using a wide range of species including Blaberus craniifer (1 957, 196 1 a), Melanopus bivittatus (1 957), Pachydiplax longipennis and Phormia regina (1 958b), Libellula luctuosa ( 196 1a), Libellula vibrans ( 196 1b), Anax junius, Aeshna sp., and Sympetrum ribicundulum ( 196 1c), Periplaneta americana (1957, 1958) and, in conjunction with Goldsmith, Periplaneta americana and Apis mellifera (1958). The ERG of the ocellus is a complex of components originating in the photoreceptor cells and in the ocellar nerve fibres. Resolution of the components in a single species is difficult but by combining data from several species and in particular from the dragonfly, Libellula luctuosa, and the cockroach, Blaberus craniifer, Ruck has been able to offer a functional interpretation of the ERG. He suggests that the total electrical response of the ocellus to a light stimulus of high intensity consists of at least four distinct components. Component 1 is a slow, monophasic, depolarizing potential, component 2 a transient, cornea positive wave superimposed upon component 1, and component 3 a slower, hyperpolarizing potential not resolvable in all species. These three components have been identified in 25 species. Component 4 is the afferent response of some of the ocellar nerve fibres. Light “on” results in inhibition of the spontaneous dark discharge of these fibres for a period of 5-1 5 s followed by a slow return of spontaneous firing if the light stimulus is prolonged. The first three events together constitute the “on-effect”. When the light stimulus goes off a negative depolarization is recorded together with a high frequency discharge in the ocellar nerve. This depolarization, registered at “off”, which occasionally shows a second, positive component (Metschl, 1963) constitutes the “offeffect” of the ERG. Ruck has assigned components 1 and 2 of the “oneffect” to the retinular cells after comparing the ERG recordings made between a corneal and an indifferent lead with those made between the ocellar nerve and an indifferent lead. The receptor cells lie nearest the cornea and hence components originating there would be expected to appear early in the sequence of events, as they do in fact, and to be larger at the corneal than at the nerve electrodes. Component 3, which he has assigned to the postsynaptic units, together with the afferent discharge 4, would be expected to appear later than components 1 and 2 and to be relatively larger at the nerve electrode. Demonstration of this in ocellar ERG recordings from the dragonfly is complicated by the fact that threshold responses of the receptor cells

154

LESLEY J . GOODMAN

appear adequate to provoke nearly maximal postsynaptic responses, consequently the latter dominate in both corneal and nerve leads at low intensities. With an increase of light intensity components I and 2 increase in size and at high intensities they become large enough in the corneal lead to almost obscure the postsynaptic elements. In the nerve lead component 3 continues to dominate even at high intensities (Fig. 27). In Blaberus craniifer the threshold response is an Electrode on nerve

Electrode on nerve

+---) Y - 7 ’ o - w

&*Id

*

”-Y*

Fig. 27. A comparison of ERG recordings made between a corneal and an indifferent lead and the ocellar nerve and an indifferent lead in the lateral ocellus of Sympetrum rubicundulum indicating those portions of the ERG to which Ruck assigns the terms components 1, 2, 3 and 4. (From Ruck, 1961a.) Component 3 is seen to appear later than components 1 and 2 and to be relatively larger at the nerve electrode. The afferent discharge, component 4, also appears at this electrode. Stimulus duration, 0.25 s; stimulus flickers at 2/s. Stimulus intensity, row (a), Log I = -6.00; row (b) = - 5 . 5 ; row (c) = -4.5; row (d) = -3.3; row (e) = -2.0; row (f) = -1.0. Negativity of either corneal or nerve electrode gives an upward deflection.

“off-effect” which is believed to represent nerve impulses in the largest ocellar nerve fibres. Components 1 and 2 appear above threshold level and increase in amplitude with increasing intensity. Component 3 is rarely discernible in the cockroach recordings made with a corneal lead (Fig. 28). In order t o isolate the response of the photoreceptor layer and determine whether components 1 and 2 do in fact originate there, Ruck recorded ERG’S from different depths in the cockroach ocellus with a fixed reference electrode at the surface of the receptor cell layer. The tip of a microelectrode was moved downwards in stages from the surface of the receptor layer

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

155

until it penetrated the ocellar sheath at the point at which the receptor cell axons meet the transversely running ocellar nerves (Fig. 29(a)). Since the postsynaptic elements are orientated transversely to the ocellar path in the cockroach differences of potential between the synaptic regions and more central regions of the nerve fibres would not be expected to appear on the electrode path and only minor contributions, if any, would be expected from postsynaptic

Fig. 28. ERG recordings made between a corneal and an indifferent lead in Bluberus cranijfer. The threshold response is an “offeffect”. Components 1 and 2 appear above threshold level but component 3 is not seen. Stimulus duration, 0.125 s. Stimulus intensities (a) Log I = -7.0;(b) = -6.5; (c) = -5.5; (d) = -4.0; (e) = -0.5.

elements. The receptor cells on the other hand are contained within the ocellar sheath and orientated parallel to the electrode path and maximal receptor cell responses should thus be recorded as the electrode passes from the base of the cornea through to the ocellar sheath. Components 1 and 2, if they both originate in this layer, should appear together at about the same depth. Figure 29(b) shows that 1 and 2 appeared simultaneously when the micro-electrode was one-third to one-half the way through the receptor cell layer. They increased in amplitude as the electrode penetrated until it passed through the ocellar sheath, when they suddenly decreased. If the reference electrode is moved to an indifferent site so that only the microelectrode lies within the electrical fields established on excitation of the ocellus the electrode tip on its passage through the receptor cell layer should move from one “pole” of a dipolar receptor field to the other whilst remaining in the same pole of a

156

LESLEY J. GOODMAN

dipolar postsynaptic field. Under these circumstances components originating in the receptor fields should reverse polarity whilst those originating postsynaptically should remain unchanged. Recording in this manner from the cockroach Ruck found that components 1 and 2 reversed their polarity near the base of the ocellar cup; the “offeffect” did not (Fig. 29(c)). In a similar experiment with the dragonfly components 1 and 2 were found to reverse polarity near

Fig. 29. (a). Ruck’s preparation of BZubems ocelli used to record ERG’s at successively greater depths. Fixed reference electrode at the left. ERG’s recorded with the tip of the microelectrode at positions 1-9. (b). ERG’s at depths 1-9 with electrodes as in (a). Negativity of microelectrodes gives a downward deflection. Components 1 and 2 appeared simultaneously at position 3 when the microelectrode was about one-third of the way through the receptor layer. They increased in amplitude until the electrode passed through the ocellar sheath when they suddenly decreased (position 9). Contributions from postsynaptic elements are minor under these recording conditions. (c). ERG’S from the ocellus of Blabems when the reference electrode was moved to an indifferent site at the base of an antenna. Under these recording conditions components 1 and 2 reverse polarity near the base of the ocellar cup (position 6). The “offeffect” (a postsynaptic component) shows no such reversal. (a, b and c from Ruck, 1961a.)

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

157

the base of the ocellar cup, but component 3 did not. This suggests an origin deeper than the layer of photoreceptor cells for both component 3 and the “offeffect”. Only component 1 can exist in the absence of any other component. As ocellar preparations age the other components drop out, eventually leaving component 1 alone and when this happens illumination no longer affects the impulse discharge in the ocellar nerve in any way. Ruck equates component 1 with the receptor potential and attributes it to the rhabdomere bearing ends of the visual cells. No intracellular recordings have been reported from ocellar visual cells but a number of studies have been made using visual cells from various arthropod compound eyes and in all cases high intensity illumination provokes a slow depolarization which is generally accepted as originating at some point in the rhabdomere region of the cell (Naka, 196 1 ; Naka and Eguchi, 1962; Autrum and von Zwehl, 1962, Fuortes and Yeandle, 1964; Benolken, 1962; Scholes, 1965). Ruck has questioned the sufficiency of the limited extracellular space between the retinular cells as an ion store for maintaining the prolonged positive ion entry phase presumed necessary to maintain depolarization in this region. He suggests that the additional membrane systems present inside the receptor cell may be implicated in the initial depolarization (see Section I1 for further discussion). Ruck sees the diphasic “on” transient (component 2) as the response of the photoreceptor cell axons and evidence of a relay mechanism between the distal, rhabdomere bearing region of the cell and the synaptic terminals. He argues that in the cockroach ocellus, for example, the high axoplasmic resistance of the small diameter (0.5 p ) , relatively long (300 p ) , retinal cell axon would cause rapid attenuation of the receptor potential and necessitate some form of relay mechanism. The ocellar “on” transient has certain properties suggestive of regenerative action potentials, a transient appearance, a sharp threshold in the honey bee and a refractory period. In Bluberus preparations it can be blocked by a high concentration of potassium in the saline medium. At the same time as it disappears light apparently ceases to have any effect upon the frequency discharge of the ocellar nerve, suggesting that its presence is necessary for the development of the hyperpolarizing postsynaptic potential, component 3 . If the K + is replaced by Na’ the “on” transient reappears within a few minutes and the coupling between receptor cell activation and ocellar nerve inhibition is restored. One objection A.I.P. -6

*

158

LESLEY J . GOODMAN

to assigning component 2 to the receptor axons is the fact that the postsynaptic inhibitory potential, component 3, presumed to be evoked by the action of 2, outlasts it by a considerable time. Ruck however points out that if the potential is actually the envelope formed by the synchronized spike discharges in small nerve fibres it would be visible, whereas desynchronized firing at a lower level would not be seen at all in the presence of component 3. Kennedy ( 1964) has suggested in addition that the inhibitory postsynaptic potential may show strong facilitation and/or have a long time constant. A postsynaptic potential with these properties could far outlast a phasic peak of presynaptic activity since facilitation would make the later impulses “count” more and a long time constant would tend to hold the already developed potential change. The increasing number of intracellular recordings from single visual cells has not solved the problem of the origin of the “on” transient and the mode of transmission between retina and second order neurons. The majority of such records show a fast transient superimposed upon the rising phase of the slow depolarizing wave. Naka and Eguchi (1962) found that the “on7’ transient was graded according to stimulus intensity in the dragonfly and dependent upon the state of adaptation. Fuortes ( 1 958) and Benolken (1962) found that the “on” transient of Limulus was graded in amplitude with stimulus intensity and at high light intensities reversed the membrane potential by 30-40 mV. Baumann (1969) has found a large spike superimposed on the slow wave of the honey bee drone which appears to have regenerative properties. The spike is abolished by tetrodotoxin which selectively blocks the regenerative Na+ influx in many membranes (Narahashi et al., 1964) and by bathing in Na+-free solutions. Though single spikes are common, trains of spikes have rarely been found although they have been recorded from the retinular cells of the locust and the honey bee drone and from the retinular and eccentric cells of Limulus. The spikes recorded from Locustu are small ( 1 mV) and are attenuated below noise level at higher intensities by the conductance changes underlying the receptor potential so that they must be conducted electrotonically from a distant part of the axon or from another cell (Scholes, 1965). Little is known of the relationship between receptor potential waves and the possible generation of action potentials in the receptor cell axons except that single, discrete receptor waves do not appear to evoke spikes. In Limulus trains of spikes have now been recorded from the smaller retinular cell axons but it is not clear whether they

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

159

are in fact generated by illumination of the retinal cell (Borsellino

et al., 1965). Naka and Eguchi (1962) and Kennedy (1964) have suggested that the single transient wave seen in the majority of intracellular records is an artifact resulting from a complete abolition of all spike potentials except the first one in cells which have been damaged by penetration. Baumann (1969) points out that single transients are normally seen in fresh intracellular preparations and in extracellular preparations, whilst trains of spikes tend to be seen in ageing or otherwise abnormal preparations. In extracellular recordings taken at different depths through the eye spikes are first clearly seen at the level of the second synaptic region in the lamina and it is possible that compound eye retinular cell recordings are influenced by inhibitory effects spreading from the lamina (Burtt and Catton, 1959, Baumann, 1969; Horridge, 1968). It is disappointing that intracellular records have not been made from the simpler ocellar receptor in view of this possibility. Certainly all that can be said at present is that the available evidence supports Ruck’s belief that the ocellar “on” transient is a depolarizing response of the retinular cell axons but the nature of the relay mechanism remains unknown. The hyperpolarizing potential, component 3, which appears to be located deeper than the photoreceptor cells, is believed to be a postsynaptic potential exerting direct inhibitory control over the discharge in the ocellar nerve fibres on the following grounds; it is relatively larger when recorded from the ocellar nerve lead than from the corneal lead; it is not recorded if leads are confined to the receptor layer; it shows a positive sign in the synaptic region; and it appears to be clearly linked with the inhibition of ocellar nerve impulses in recordings taken from the dragonfly. Figure 30 shows the inhibition of spontaneous discharge with the appearance of component 3. Spontaneous, miniature, hyperpolarizing potentials also occur in the dragonfly ocellar nerve during darkness and they can be seen to produce a short inhibition of the spontaneous dark discharge, Autrum and Metschl (1963) have raised certain objections to this interpretation. Metschl ( 1963) finds no hyperpolarizing potential in his recordings from Culliphoru and it is rarely seen in Ruck’s recordings from the cockroach; over a very wide range of stimulus conditions Metschl finds no correlation between the size of any ERG component and impulse frequency; no miniature hyperpolarizing potentials producing inhibition of the ocellar nerve discharge are ever seen during darkness as in the dragonfly; although illumination results in inhibition of the discharge in the ocellar nerve

160

LESLEY J . GOODMAN

of Culliphora the frequency of discharge eventually recovers during a long period of light stimulation to reach a level almost equivalent to the steady rate of discharge during darkness. Ruck suggests that the absence of the hyperpolarizing potential in the cockroach recordings may possibly be accounted for by the fact that the synapsing dendrites of the ocellar cup run transversely across the receptor cells

J

Fig. 30. Inhibition of the dark discharge in the ocellar nerve of the dragonfly. Illumination produces a hyperpolarizing postsynaptic potential and consequent inhibition of the dark discharge, (a). Firing returns at a low level in (b) and upon cessation of illumination (c) a high frequency off discharge occurs, probably generated by rebound depolarization. The small, spontaneous inhibitory hyperpolarizing potentials present in darkness are absent immediately after illumination. (From Ruck, 1961c.)

unlike the dragonfly in which the ocellar nerves lie parallel to the receptor cells. If the dipolar electric field associated with component 3 is orientated transversely across the ocellus then a corneal electrode in the cockroach would record little or no sign of component 3. Possibly this could account for Metschl’s failure to record it in Calliphora although recent histological studies suggest that the synaptic elements here lie parallel to the receptor cells as in the dragonfly (Goodman, 1970). There is no direct evidence concerning the mode of synaptic

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

161

transmission between the ocellar retinular cells and the second order neurons but the nature of the electrical events comprising components 2 and 3 suggest that chemical, rather than electrical, transmission occurs. Component 2 appears within the ocellus as a dipolar electrical field, the negative “pole” of which lies in the synaptic region. If the electrical effect of component 2 were the significant factor it might be supposed to exert a depolarizing effect on the postsynaptic membrane. However the postsynaptic response evoked by component 2 is in fact a hyperpolarization. The presence of miniature i.p.s.p.’s in the ocellar nerve during darkness suggests the random release of small packets of inhibitory substance by individual receptor axons. Light stimulation presumably drives the whole population of receptor cells, simultaneously resulting in a massive release of inhibitory transmitter producing a sustained postsynaptic potential. Spontaneous miniature i.p.s.p.’s disappear for a time after illumination as would be expected if some synaptic transmitter had been depleted during activity. The ocellus, with its first order inhibitory synapse, was earlier thought to be a rather specialized photoreceptor but evidence of first order inhibitory synapses is accumulating from other arthropod photoreceptors. Retinular cell arborizations upon each other in Limulus appear to be inhibitory. Insect retinular cells are normally small field “on” units but small field “off” units can be found peripherally in the insect lobe (Hamori and Horridge, 1966). Scholes (1965) has reported the presence of a small, sustained hyperpolarizing potential at high intensities when recording from the proximal region of the retinal layer of Locusta. In the presence of this potential the low frequency dark discharge of action potentials ceases. At light “off’ they are transiently accelerated in a characteristic inhibitory rebound. Scholes suggests that the hyperpolarizing potential is the response of the rudimentary retinular cells found in individual omma tidia. Basal retinular cells with reduced rhabdomeres or eccentric cells have been found in a number of compound eyes but their role is still obscure. No such cells have been reported from ocelli. 0. ELECTRICAL ACTIVITY IN THE SECOND ORDER NEURONS

Although Parry (1947) was unable to detect impulses in the ocellar nerve of Locusta and suggested that “darkening the ocellus causes a depolarization which spreads down the ocellar nerve and depolarizes a ganglion in the brain” subsequent authors have been able to record spikes. Hoyle ( 1955) found a continuous discharge in

162

LESLEY J . GOODMAN

what he believed to be the smaller ocellar fibres during darkness or dim light which ceased immediately the eye was illuminated. At light “on” he recorded a single very large spike which increased in size as intensity was increased but never became repetitive. Hoyle suggested that this response was in fact a single impulse in one of the giant fibres, the larger spike at higher intensities being due to an almost simultaneous firing in a second giant axon. Later authors have never reported a large action potential in the ocellar nerve at light “on” and it is very probable that Hoyle was recording the postsynaptic hyperpolarizing potential described by Ruck which does increase in size with increasing intensity. A sudden decrease of intensity evoked a transient burst of activity in those fibres which were firing spontaneously in darkness which later fell off until the level of the general dark discharge was reached. The experiences of other authors suggest that these fibres are in fact more likely to be the giant axons than the smaller ocellar fibres. Similar recordings are obtained from the dragonfly ocellar nerve (Ruck, 1961b). Activity in the dark adapted state consists of a spontaneous, arrhythmic discharge of impulses of two sizes and hence presumably in two fibres. A light stimulus generates a sustained hyperpolarizing, postsynaptic potential which appears at the nerve electrode as a negative wave and inhibits the nerve impulses (Fig. 30). Both fibres show partial adaptation to a steady stimulus, the smaller spikes adapting faster than the large ones. The most detailed study of the properties of the ocellar nerve has been made by Metschl who has been able to follow the activity in the nerve during and subsequent to illumination for periods of half an hour or more in Calliphora. In darkness he also recorded impulses of two sizes. The smaller ones, having a steady discharge in darkness of between 40-70 impulses/s, appear in all preparations. Larger impulses, having a low frequency of 5-8 impulses/s, appeared in about half the preparations (Fig. 3 1). Stimulation by light was followed by a silent period and then after a variable interval of between 10-180 s, longer generally than in the dragonfly, the smaller impulses reappeared with their frequency rising slowly over 2 or 3 min and then settling down to a steady rate rather lower than the dark rate. The larger impulses reappeared more slowly and at a very low rate. At light “off” a burst of impulses at a high frequency, reaching a peak of 200-300 impulses/s, was recorded in both types of fibre. The rate of discharge fell off rapidly in the first half minute in both fibres and then slowly returned to the previous steady dark frequency over a period of 5-10 min. After the

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

163

initial transition period the impulse frequency during darkness and light remains constant over a period of at least 30min. At near threshold values for the eye the silent period disappears and the steady light discharge increases so that in effect it shows little difference from the steady dark discharge. The discharge peak following light “off’ is also much reduced. Metschl’s observations on Culliphora suggest that in addition to the phasic information about 3001 Irnplsec

0

14

5

6

rnin

Fig. 31. Variation in impulse frequency in the ocellar nerve of Calliphora erythrocephala in relation to light intensity. Ordinate: impulses/sec. Abscissa: time in minutes. Duration of light stimulus 5 min, indicated by open bar. Solid line shows frequency at the highest intensity used, broken lines show frequency at different percentages of this intensity. (From Metschl, 1963.)

intensity changes the ocellus could function as a tonic receptor yielding information about the absolute level of the light intensity over long periods. Not all ocelli appear to function in the same way. Ruck notes that the cockroach ocellus is normally silent in darkness and in light giving only a short burst of impulses at light “off’. Under suitable lighting conditions the dragonfly ocellus can be made to signal phasically only at light “off”. A low background illumination was used which was just sufficient to inhibit completely the spontaneous dark discharge. A bright flash was superimposed upon this and nerve impulses occurred only when the bright light went off, giving a response in a manner similar to the normal

7

164

LESLEY J . GOODMAN

cockroach response. The fact that decreases of light intensity within the range of intensities sufficient, under sustained exposure, to produce total inhibition of the nerve impulses can produce transient discharges increases the informational output of the ocellus at higher levels of illumination. The discharge of impulses in the ocellar nerve is controlled by the level of postsynaptic membrane potential. Ruck ( 1 9 6 1 ~ )suggests that in the dragonfly spontaneous and random activity in the receptor cells in the dark evokes transient, miniature, inhibitory postsynaptic potentials which modulate the inherent, rhythmic activity of the ocellar nerve fibres and results in the characteristic arrhythmic dark discharge. Light stimulation synchronizes excitation in many or all of the receptor cells and results in a large, sustained inhibitory, postsynaptic potential, completely silencing the ocellar nerves. At light "off" the postsynaptic potential declines, generally overshoots, and frequently oscillates briefly before terminating in a sustained after-potential believed to be a depolarizing component of the postsynaptic response. The transient "off" discharge to a decrease of intensity in the illuminated dragonfly eye could arise from this cycle of the "off" oscillation and the sustained after-potential. The cockroach ocellar nerve has the capacity for spontaneous, rhythmic activity since this is encountered in some preparations, particularly ageing ones, but normally the level of postsynaptic membrane potential is such that it is silenced although the mechanism here is not understood. It would be of interest to know the signalling characteristics of other ocellar nerves in this context. Most authors appear to be recording the response of the giant ocellar fibres, possibly the very much smaller fibres present are excited by light but their responses are too small to be seen by the recording techniques used. C. SENSITIVITY, LIGHT AND DARK ADAPTATION AND FLICKER FUSION FREQUENCY

In searching for the role of the ocelli one major point of interest is how the visual information they yield compares with that available from the compound eyes. The high degree of convergence of receptor axons on the ocellar nerve fibres, the multiple synapsing of single receptor axons with ocellar nerve fibres, lack of screening pigments and other arrangements for cutting down the amount of light admitted together with the presence of a white tapetal layer in the majority of cases suggests that the sensitivity of the ocellus might be enhanced compared with the compound eye. It is difficult to

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

165

decide at which point in the visual pathways a meaningful comparison can be made. No comparison of the thresholds of individual retinular cells can be made since no intracellular records are available for the ocellus. In any case part of any difference in sensitivity between the two eyes will be due to differences in the neural connections rather than in the properties of individual retinular cells. Sensitivities determined from ERG recordings have been compared by Ruck in Periplaneta, Pachydiplax, Apis and Phormia. His measurements of threshold levels in these four species have shown that it is not possible to generalize about the relative sensitivities of the ocelli and compound eyes. The ocellus of the dragonfly, Pachydiplax, is a very sensitive receptor giving an electroretinogram when the corneal illumination is as low as lo4 ft. candles (Ruck, 1958). Ocelli of Periplaneta are equally sensitive, and in these two insects the sensitivity of the ocelli appears to be one or two orders of magnitude greater than that of their compound eyes. In Apis and Phormia, however, the thresholds for the ocelli were the highest of the several compound eyes and ocelli which he studied. In these measurements, although corneal illumination was equal for ocelli and compound eyes there was of course no measure of the illumination received at the retinular cells in each case. No comparisons have been made at any other level between the retinal layer and the ventral nerve cord and those made in the cord are very largely reflecting the differences in the wiring connecting the compound eye and ocellar retinular cells with the higher order neurons in the cord. At the level of the thoracic ganglia in S. gregaria the most sensitive response to brightening is that of a compound eye unit, to dimming that of an ocellar unit. The sensitivity of the ocelli to dimming is at least one order of magnitude greater than that of the compound eyes at this level (Goodman, 1968). Measurement of the increment threshold over a range of intensities for the compound eye and ocellar units at mesothoracic ganglion level in S. gregaria shows that the ocelli are able to signal the occurrence of much smaller decreases of light intensity to this level than the compound eyes. The increment threshold for increases of light intensity is higher in both ocellar and compound eye units, the threshold of the compound eye units being lower than that of the ocellar units (Goodman, 1969). Measurements of the flicker fusion frequency of compound eyes made by many workers have revealed two classes of compound eye; “fast” eyes with a flicker fusion frequency in the range

166

LESLEY J . GOODMAN

200-300 stimuli/s, and “slow” eyes with a fusion frequency in the range of 25-60 stimuli/s. Autrum and his co-workers ( 1 950a, b, 195 1, 1957, 1958, 1960) suggested as a result of their experiments with Calliphora, Tachycines, Dixipus, the dragonfly and Apis, that low flicker fusion frequencies were associated with high photic sensitivity and slow rates of dark adaptation whilst high flicker fusion frequencies were associated with rapid adaptation rates and lower sensitivity. Electroretinograms from complete slow eyes yielded a monophasic negative potential at the onset of light which was maintained during illumination and died away slowly afterwards. ERG’S from ‘fast” eyes on the other hand yielded a diphasic potential, a positive “on-effect”, a return to the baseline during illumination, and a negative “off-effect”. Autrum suggested that the retinular cells of both fast and slow eyes were essentially similar in so far as the basic electrical events originating there were concerned, the differences between the two types of eye being due to the underlying structural organization. In fast eyes the lamina ganglionaris, which is the first synaptic region in the afferent pathway from the compound eyes, lies close to the retinular cells. Autrum claims that excitation of the retinular cells initiates depolarization in the retinal layer and that associated current flow evokes a postsynaptic electrical response from cells within the lamina. The postsynaptic response, opposite in electrical sign to the presynaptic one, acts as a “bucking potential” and prevents sustained depolarization of the retinular cells. This action minimizes the desensitizing effect of exposure to a bright light and permits the resolution of a large number of stimuli per second. In the slow eyes, where the lamina lies much further away from the retinular cells, the effects of the “bucking potential” are not seen, depolarization continues as long as illumination lasts and a longer recovery period is needed. The flicker fusion frequency of the ocelli of a number of species has now been measured (Table IV) and in all the species so far examined the flicker fusion frequency lies in the same range as that of the associated compound eye. As Ruck (1958) has pointed out, the fact that the dorsal ocelli of Apix, Pachydiplax and Phormia have high flicker fusion frequencies though lacking the lamina ganglionaris indicates that this layer is not essential for a high temporal resolution. It appears that the characteristics of the retinular cells themselves determine whether or not an eye can resolve high rates of flicker. It is of interest that in two classes, dragonfly and locust (Ruck, 1958; Hoyle, 1955) ocellar nerve records give a much lower

Table IV Flicker fusion frequency. A comparison of measurements of the flicker fusion frequency of the compound eye and the ocellus in several insects ~~

Author

Insect

Hoyle (1955)

Locusta

Ruck (1958b)

Apis

Ocellus

Compound eye

-- CKCi

120

250-265

60

45-60

Dragonfly

190

200+

Phormia

350

220+

Calliphora

Autrum (1950)

Calliphora

Goodman (1968)

Schistocerca

100 for “on-discharge’’ (possibly ERG response) 30 for “off-discharge”

-_.. Higher’ order neuron in VNC 30 for “off-discharge’’ in circumoesophageal commissures

250-265 for the negative wave, the positive wave fuses at a much lower frequency

Periplaneta

Metschl (1963)

Ocellar nerve

ERG

230-240

“Off-discharges’’ of largest fibre in ocellar nerve 20-40

230-240

265 35-50

25-35

20-30 for ocellar “off” unit in VNC

168

LESLEY J . GOODMAN

flicker fusion frequency than ERG records. Ruck suggests that the receptor cell axons are not able to follow the higher rates of fluctuation. In contrast to earlier results for the compound eye Ruck found that there was no general relationship in ocelli between flicker fusion frequency and sensitivity. As noted earlier the ocellus of Periplaneta and Pachydiplax are equally sensitive but the cockroach eye fuses at 50 stimuli/s, the dragonfly’s at 200 stimuli/s. Such measurements as there are reveal no relationship between the rate of dark adaptation and flicker fusion frequency in the ocellus. Figure 32 shows the course of dark adaptation in the dorsal ocelli of four insects and in the corresponding compound eyes. Dark adaptation is complete in periods ranging from 5-30 min. Ruck was not able to obtain sensitivity values earlier than 20-30 s after the start of the dark adaptation period and some of his results and those of Metschl for Calliphora (in which the major part of dark adaptation has occurred within 1-3 s) suggest that in some cases, for example Phormia and Calliphora and probably Pachydiplax, the greater part of the adaptation is complete with the first 100 ms. Eyes with high flicker fusion frequencies must be able to recover the necessary responsiveness in shorter times than eyes with lower fusion frequencies. In Periplaneta the recovery time must be longer than 8.3 ms if the ERG is not to fuse but for the Pachydiplax ocellus the recovery time need only be about 2.3 ms at the same intensity. It is not possible with the high intensities used in order to obtain the best value for flicker fusion frequency to determine sensitivity levels accurately during the first second in electroretinogram recordings. It is likely that if the course of adaptation could be followed during this period a marked difference in rate would be seen between “fast” and “slow” ocelli. The behavioural significance of high temporal resolution in the ocelli remains obscure and this property may simply reflect the rapidity of the process of dark adaptation in those eyes. Studies of sensitivity, rate of dark adaptation and flicker fusion frequency do not give any obvious indication of the role of the ocelli. The compound eyes and ocelli of three diurnal species, Apis mellifera, Pachydiplax longipennis and Phormia regina, have high flicker fusion frequencies; those of two diurnal and one nocturnal species, Locusta, Schistocerca, and Periplaneta have low fusion frequencies. The ocelli of two diurnal and one nocturnal species are more sensitive to photic stimulation than the compound eye whilst the sensitivity of the ocellus of two other diurnal species is less than that of the compound eyes. The presence of these differences in

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

169

Time in the dark

Time in the dark (C)

(d)

Time in the dark

(e)

Fig. 32. Dark adaptation curves of compound eyes and ocelli. (a) P. umericunu, ocellus. @) P. umericunu, compound eye. (c) Apis, ocellus. (d) Apis, compound eye. (e) Puchydiplax

longipennis, ocellus. (From Ruck, 1958.)

170

LESLEY J . GOODMAN

properties between ocelli lends support to the idea that the ocelli are adapted for specific functions in each species; functions which are possibly complementary to those of the compound eyes of that species. D. THE SPECTRAL SENSITIVITY OF OCELLI

The spectral sensitivity of the photoreceptors of the dragonfly and of the honey bee and cockroach have been measured by Ruck ( 1957) and by Ruck and Goldsmith ( I 958). The spectral sensitivity curve of dark adapted ocelli of the dragonfly Libellula luctosa has a peak at about 5 18 mp and a second one in the ultra-violet beyond 380 mp. There is a transitional region in the valley between these two

I’

I

I

I

I

I

Component of erg 0- negotive positive

4-

-

-

0 20

“ 1 “ 1 “ , 1 1 1 ‘ 1 1 1

300

350

450

400

500

550

600

mP Fig. 33. Spectral sensitivity of a single median ocellus of Apis showing two maxima. Sensitivity is here defined as the reciprocal of the relative number of quanta required to produce a constant effect in ERG recordings. (From Goldsmith and Ruck, 1957.)

peaks at 410 mp. At 400 mp the “off-effect” of the ERG wave is broader than at wavelengths longer than 410 mp and with an increase in stimulus intensity the “on-effect” too acquires a component which is not visible at longer wavelengths. The whole ocellus was stimulated and the ERG recorded was of course the summed response from a number of cells. The simultaneous stimulation of two or more classes of photoreceptor does introduce the possibility

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

171

of functional interaction amongst the classes but the response levels used by Ruck to define spectral sensitivity here approached threshold levels where the probability of isolating one receptor class is much higher. It seems likely that the dorsal ocelli of the dragonfly do have two classes of photoreceptor, one with a maximum sensitivity at about 5 18 mp, the other with maximum sensitivity in the ultra-violet at a wavelength shorter than 3 8 0 mp . A similar situation exists in the honey bee where the maxima of the spectral sensitivity curve lie at 490mp and 335-34Omp respectively (Fig. 33). In Periplunetu and Bluberus only one class of photoreceptor with a maximum sensitivity in the region of 500 mp was found. However, Walther and Dodt (1957) have found a second peak in the ultra-violet. It is unlikely that discrimination of wavelength could be preserved beyond the first synaptic layer because of the high degree of convergence. Since the ocelli do not generally initiate motor reactions, tests for behavioural discrimination of wavelength to complement electrophysiological work are not possible.

V . OCELLAR UNITS IN THE BRAIN A N D VENTRAL NERVE CORD A. OCELLAR UNITS IN THE BRAIN

Extracellular recordings from single neurons have been made from the brain of L. migrutoriu by Horridge e t ul. (1965). Recordings made in the region of the ocellar nerves, presumably from the central ends of the second order fibres, showed “on” and “off’ responses similar to those reported by Hoyle ( 1 955). However, no units with a maintained dark discharge were found. Horridge e t ul. suggest that this discrepancy may be due to the fact that Hoyle was recording from the whole nerve whereas they were sampling single units and may have missed the dark discharge ones. Responses from the contralateral ocellus were picked up in the protocerebral lobes. One type of unit found in the lateral protocerebral lobe, presumably a third or higher order neuron, reveals an interaction of the two lateral ocelli; the “off” response to a contralateral light is abolished by a light stimulus to the ipsilateral ocellus at the appropriate time. Although there is some histological evidence of second or third order ocellar fibres running to the optic lobes no units have so far been picked up in recordings made from either the optic lobes or the commissures.

172

LESLEY J . GOODMAN

In extracellular recordings from the brain of Boettcherisca peregrina ocellar illumination appears to influence the level of spontaneous discharge in units which respond either to mechanical or to chemical stimulation of the antennae (Mimura et al., 1969). Both phasic and tonic excitatory and inhibitory units have been encountered, phasic units largely in the protocerebrum and tonic units in the deuto- and tritocerebra and the suboesophageal ganglia. Illumination of the ocelli speeded up the spontaneous discharge of these antennal interneurones in most cases but decreased the rate of firing in some units. When no light was falling on the ocelli the response given to antennal stimulation was reduced in most of the units encountered in the protocerebrum and enhanced in most units in the deuto- and tritocerebra and suboesophageal ganglia. Units whose performance was enhanced by ocellar illumination have been termed light facilitatory units, those whose activity was reduced light occlusive units. Hoyle (1 955) has also noted that the overall activity in the circumoesophageal commissures of the locust is enhanced when the ocelli are in darkness (see later). Ocellar blinding increased the latency of the light facilitatory units and reduced that of the light occlusive units. The light facilitatory units were found mainly in the median region of the brain and the light occlusive units in the dorsal region of the protocerebrum and in the ventral region of the deuto- and tritocerebra and suboesophageal ganglia. The ocellar nerve of Boettcherisca shows a higher rate of discharge in the dark adapted state than in the light adapted state in common with all other ocellar nerves so far examined. Mimura et al. (1969) point out that the response obtained from the light occlusive interneurons could be obtained by simple convergence of antennal and ocellar information but, in view of the apparent properties of the ocellar nerve, to obtain a unit with the properties of the light facilitatory unit requires inhibitory interneurons or presynaptic inhibition. Multimodal interneurons are well-known in the brain and cord of the Crustacea (see Horridge, 1968, for recent summary) and are being increasingly reported from insects. In the locust an auditory stimulus can cause arousal of a compound eye visual unit that has habituated to a visual stimulus (Horridge et al., 1965). Dingle and Caldwell (1967) have reported multimodal in terneurons in the cockroach brain that respond to mechanical stimuli and to light although it is not clear whether the compound eyes or the ocelli or both are involved. In addition to interaction between ocellar units ventral nerve cord

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

173

recordings indicate that there must be interaction between compound eye and ocellar units in the brain (see later). It is, therefore, an oversimplification for Mimura et al. to claim on the basis of their results that the ocellus controls brain excitability and acts as a stimulatory organ. B. OCELLAR UNITS IN THE VENTRAL NERVE CORD

Behaviour experiments have shown that the ocelli participate in some responses initiated or steered by visual stimulation although they are unable to mediate these responses in the absence of the compound eyes. It is also claimed that they assist in regulating muscular tonus. It is of interest therefore to know whether the ocelli can drive interneurons in the ventral nerve cord and if so what motor activity, if any, results from ocellar stimulation. To understand the mechanisms underlying behavioural observations it is also necessary to know the nature of compound eye and ocellar interaction and the level at which it occurs. Only Locusta and Schistocerca have been studied from this viewpoint and the examination is at an early stage. Hoyle ( 1955), recording from the entire circumoesophageal commissures of Locusta, observed three types of response to ocellar stimulation. In darkness he recorded a continuous, irregular discharge of impulses in the commissures, the “dark discharge”, which was greatly reduced during illumination of the ocellus. A brief discharge was recorded following a sudden increase or decrease of illumination, the “on” and “off” responses respectively. The “off” response was distinguishable from the general “dark discharge” which followed it closely. In flickering light he noticed an enormous increase of activity in the circumoesophageal commissures at frequencies up to 30 flashes/s. Burtt and Catton ( 1954), also recording from the whole ventral nerve cord of Locustu between the pro- and meso- and mesoand metathoracic ganglia, reported a very brief “off” discharge of one or two spikes only. N o “on” responses were reported nor was any response recorded when objects were moved through a few degrees in the ocellar visual field. Burtt and Catton did not report either any “dark discharge” or decrease of activity in the connectives in response to illumination. Examination of single units in the ventral connectives of Schistocercu has revealed several types of interneuron that can be driven by ocellar stimulation (Goodman, 1968). As in Locustu, the most readily recorded unit type is that which fires

174

LESLEY J . GOODMAN

briefly with one or two large spikes to a sudden decrease of intensity. Also commonly recorded is a unit which fires in response to brightening and dimming; an “on/off” unit. Small units which fire in a sustained manner in response to illumination with a very slow rate of adaptation have been found in 25% of the preparations examined (Fig. 34 A, B and C). These three types of unit can be found as far down the cord as the metathoracic ganglia but never beyond that

A

B

c

.l.m-.

n -

-

I

1

.-

.

I

.

n

Fig. 34. Three types of ocellar unit recorded extracellularly in a ventral connective of S. gregana between the pro- and mesothoracic ganglia. Light stimulus indicated on bottom trace, trace reads from left to right: A, the large ocellar off unit; B, an ocellar on/off unit; C, unit firing in a sustained manner to stimulation of the median wellus. (From Goodman, 1968.)

point. They cannot be excited by compound eye stimulation. It is possible that at least one interneuron which can be driven by both types of photoreceptor is present in the cord but it has been seen too infrequently to determine its characteristics. 1. The Ocellar “Off” Units The ocellar “off” units fire in response to a sudden decrease of intensity given to one of the three ocelli. One or occasionally two large spikes are normally recorded of the same order of size as the large compound eye “on/off” units found in the cord ( 1 .O-1.5 mV).

175

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

The response of the ocellar “off” units is amongst the largest that can be recorded from the ventral connectives. Transverse sections through the connectives of Schistocercu show numerous comparatively large axons, ranging from 5-15 p in diameter, and it seems probable that these ocellar “off” units are amongst the larger of these neurons. Unlike other visual units recorded from the ventral connectives this “off” unit does not readily habituate with repetitive stimulation. It will fire at “light off” to flashes of light given at a rate Response

1.

r

1 I

I

.

I IO/sec

stimulus

b

.

15

20

.)

30

.

’

35

Fig. 35. The response of an ocellar off unit in the ventral connective to flashes of white light at various frequencies. The unit will follow flickering light up to 20 stimuli/s but begins to fail above this and by 35 stimulils has ceased to respond. (From Goodman, 1970.)

of 5 / s for 1 or 2 hr. It will follow a flickering light up to 20 stimuli/s. At higher frequencies it fails after the first few stimuli and above 25-30 stimuli/s it fails completely (Fig. 35). The latency of the “off” units varies with the duration and intensity of the light stimulation in a manner common to many such systems, being considerably shorter after longer exposure times and higher intensities. After 4 m s exposure the mean latency is 50 ms, after 1 min exposure it is 25 ms (Fig. 36). A similar reduction in the latency time of the “on” response of the large compound eye units in the ventral cord of Locusta has been attributed by Burtt and Catton (1959) t o by-passing of the third synaptic layer of the optic lobes at higher intensities, impulses thus travelling along a tract having fewer

176

LESLEY I. GOODMAN

synaptic junctions. From the little information we have about the ocellar pathways it seems unlikely that there is an alternative, shorter, pathway in this case. The effect is probably due to reduced delay at the photoreceptor level. No figures are available for Schistocerca, but in Calliphora Metschl ( 1 963) has shown that, although the latency of the negative “off-effect” of the ERG appears independent of stimulus intensity and duration, the latency of the slow positive after fluctuation of the ERG and of the first nerve 5550, , *

s .

045.

0

I I

1 1

I

I

1 1 I 1 I I I

1

r

I

I

$ 35,. c

I !

E 40,

_J

’

30.

I ’ ! I I

25,.

;; I 1

I 1 1 I 1

I 1

k l I

1

I I

1

I

I

I

1

Compound eye off unit

1 ‘ 1

I 1

I 1

.

Ocellor giant off unit 0

0

...I...(

Fig. 36. The relationship of the latency of the ocellar off-unit to the duration of the preceding light stimulus. Recording made in the ventral connectives between the pro- and mesothoracic ganglia. The latency of the fEst “off” signal from the compound eyes received at this level is shown at longer durations for comparison. (From Goodman, 1970.)

impulses to appear in the ocellar nerve is related to these two factors. After a stimulus duration of 40 ms, the latency of the first impulses lies between 38-60 ms. It decreases with a longer stimulus duration to a minimum of 8-10 ms, following stimuli lasting 1 min or longer. The latency of the maximum impulse frequency in the ocellar nerve at light “off” also decreases with increased stimulus intensity and duration. The latency of the response produced by dimming in the large compound eye units in the connectives of Schistocerca is very variable depending, amongst other factors, on the area of the eye stimulated, but it is always longer than that of the ocellar “off” units for a given intensity and duration of exposure. The first information on dimming to reach a given point in the ventral cord appears to come via the ocellar “off” units. The unit has a fairly low threshold but it is however more than one order of magnitude greater than the threshold of the ERG of the median ocellar response to a decrease of

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

177

intensity. It is difficult to make a meaningful comparison of threshold values with the compound eye units of the cord. When the compound eye is presented with the same decrease of intensity through a light guide of identical diameter the compound eye units have a variable although markedly higher threshold. However, their performance can be improved by increasing the number of ommatidia stimulated although the threshold level for detecting a decrease of intensity never drops as low as the ocellar “off” units (Goodman, 1970). The ocellar “off” unit may signal a prolonged dimming by a very brief train of spikes but it will only respond to fairly rapid decreases of intensity. These units have a very large receptive field, virtually as large as the field of view of the ocellus. This is perhaps not surprising in view of the very high degree of convergence of visual cells upon the giant axons in the ocellar nerve. The pathways of the ocellar “off” units have not been established anatomically. There is some indication that each ocellus sends several fibres into each ventral connective (see Section 11) although it is not clear whether they are second, third or higher order neurons. Each ocellus appears to send one larger, probably third order, neuron into the connectives in Locusta (Satija, 1957), and this seems also to be the case in Schistocerca. It is suggested that these larger neurons are in fact the ocellar “off” units. Electrophysiological recordings indicate that a decrease of illumination at the median ocellus and at each of the lateral ocelli in turn causes a separate “off” unit to be fired in each case in both the left and right connectives. It has not been shown conclusively that three larger neurons, one from each ocellus, descend into each ventral connective but this seems the most probable arrangement on the present evidence.

a. Interaction between the ocellar “ofr’unitand compound eye input. The ocellar “off” units are not affected by compound eye input. Their firing is never inhibited when compound eye stimulation precedes, is simultaneous with or overlaps ocellar stimulation. The properties of the ocellar “off” units have not been found to be affected by any pattern of compound eye stimulation. The threshold decrease of intensity necessary to fire one of the ocellar units is not affected either by the state of adaptation of the compound eyes or by simultaneous or overlapping compound eye stimulation. Certainly there is no summation of a subthreshold response of the ocellus with a compound eye response to evoke firing of the ocellar “off’ unit. Ocellar “off” units, responding with one or two large spikes at

178

LESLEY J . GOODMAN

“light off”, have been found in the ventral connectives of P. umericunu, extending as far as the metathoracic ganglion, in the cervical connectives of L. sericutu and in the thoracic ganglion of A . mellificu. The units show the same general properties as the ocellar “off” response in Schistocercu (Goodman, 1970). The role of these ocellar “off” units in the normal behaviour of the insects possessing them is unknown. They resemble in many ways the giant fibres involved in various insect startle reactions and since they are the first units to signal a sudden decrease of intensity to the thoracic ganglia they might be expected to be implicated in such startle reactions given in response to dimming. Behaviour studies show that dimming the ocelli alone does normally not initiate such reactions. However, studies of a startle reaction given by flying locusts to dimming show that the reaction is extremely small unless both the ocelli and the compound eyes are operating (see Section 111, Fig. 26). It seems probable that the ocellar “off’ units are involved in some way in this particular startle reaction but although they can drive certain motor units in the thoracic ganglia (see later) they are apparently not able to initiate motor activity.

2. The Ocellur “onloff” Units A unit responding to dimming and very weakly to brightening can be found in many preparations. The fact that the spikes are small (0.5 mV) and the units not so readily found as the ocellar “off” units suggests that this unit is one of the smaller neurons in the cord. The threshold for a response to brightening is high and that for a response to dimming is higher than that of the ocellar “off” unit. At high intensities and relatively long exposure times, 250 ms and upwards, a small number of spikes are given in response to light at frequencies between 15 and 40 impulses/s. The “on” response has a mean latency of 60 ms at a stimulus duration of 1 s. The discharge given at “off” is somewhat variable. Frequently there is a brief train of spikes, at a frequency of 60-70 impulses/s, with a mean latency of 42 ms at exposure times of 1 s and over. On occasion, after a short depressed period, the unit fires again at a lower frequency, 15-20 impulses/s. The threshold for dimming is lower than that for brightening and the “off” response is frequently seen alone. With a high intensity flash the unit will follow flickering light up to 25-30 stimuli/s, when fusion of the “off” effect occurs at ocellar nerve level (Hoyle, 1955; Goodman, 1970). The receptive field of this type of unit although wide is smaller than that of the ocellar

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

179

“off” unit. This type of unit can be found in the ipsi- and contralateral connectives when stimulating any one of the three ocelli but the pathway and the number present in each connective are unknown. The longer latency times and the possibility of central integration (described below) suggest that these may be higher order neurons than the ocellar “off” units. a. Interaction with other units. Inhibition of the “off” effect of an ocellar “on/off” unit in the right connective driven by stimulation of the left ocellus can be achieved by appropriate stimulation of the right ocellus. Similarly, when such a unit in the left connective is driven by the median ocellus the “off” discharge can be inhibited by appropriate stimulation of the right ocellus. A similar interaction has been reported by Horridge et al. (1965) in an “on/off” unit in the lateral protocerebral lobe of Locusta in what they suggest is a third or higher order neuron. The “on” effect of the “on/off” unit can be inhibited or depressed if the ocellar stimulation is preceded or overlapped by high intensity compound eye stimulation (Fig. 37(a), (b), (c) and

(a).

1

1

Fig. 37. The offeffect of an ocellar on/off unit (a) in the right ventral connective, driven by stimulation of the left ocellus, can be inhibited by simultaneous high intensity stimulation of the right ocellus (b). Lower trace indicates duration of stimulation, trace reads from left to right. (b). The on-effect of an on/off unit driven by the median ocellus (C) can be inhibited or depressed (d) by appropriate high intensity stimulation of a compound eye. Middle trace in (d) indicates duration of stimulation of compound eye, lower trace of stimulation of ocellus.

180

LESLEY J. GOODMAN

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

181

3. The Ocellar “On” Units Sustaining units which fire at a low frequency, 20-40 impulses/s, during illumination have occasionally been recorded. These units give small spikes, of the order of 0.5 mV, and have a long latency, 100-200 ms [cf. the long “silent period” (10-180 ms) after illumination in the ocellar nerve of Calliphoru, before the A

I

1

0

Fig. 38. A, An ocellar sustaining unit recorded from the right ventral connective between the pro- and mesothoracic ganglia in S. gregariu driven by illumination of the median ocellus. Top trace indicates duration of illumination, trace reads from left to right. B, Simultaneous stimulation of the left compound eye depresses the rate of fning. Top trace indicates duration of illumination of ocellus and left compound eye. A.I.P.-7

182

LESLEY J. GOODMAN

reappearance of small nerve impulses. Section IV] . No correlation has been observed between rate of discharge and net luminosity. The problem remains of how a unit with these characteristics or those of the preceding unit are derived from the known ocellar input. a. Interaction with other units. Simultaneous stimulation of the left

compound eye and the median ocellus at high intensities has been observed to result in a reduction of the rate of discharge in this type of unit (Fig. 38). Interaction between ocellar inputs has not been examined. C. THE EFFECT OF OCELLAR INPUT ON COMPOUND EYE UNITS

Several compound eye units run in the ventral connectives as far as the metathoracic ganglion. Those most readily recorded are a pair of large axons, one in each connective, which fire in response to dimming and weakly in response to brightening and are descending fibres from the contralateral optic lobes (Burtt and Catton, 1954, in Locusta; Palka, 1967, in Schistocerca). The optic lobes also send fibres into the ipsilateral connectives (Satija, 1968) which give a similar response to changes of intensity. Small units, excited by either the lateral or the contralateral compound eye, and firing in a sustained manner in response to a light stimulus with very slow adaptation, are also found in each connective (Goodman, 1969). It is likely that other small units exist which are capable of being driven by compound eye stimulation but which possess different properties. The mode of interaction of five photoreceptors of two different types over a wide range of stimulus intensities and combinations is difficult to determine. Extensive recording has shown that the ocelli do on occasion have some effect upon the activity of compound eye units in the cord but consistent results have been hard to obtain. When all the ocelli are occluded general activity in the cord in response to a flash of light is of course reduced since ocellar units are removed. However, as well as this effect, the response of the compound eye units is quite often depressed. In some instances no effect is seen and on occasion the opposite effect has been recorded and activity in the compound eye units increased when ocelli are occluded. When the interaction of single units belonging to the compound eyes and the ocelli is examined it is found that the ocelli can exert an inhibitory effect on compound eye units. Suitable stimulation of the median ocellus at a high intensity and the left compound eye can result in depression of the “on” effect in one of

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

183

the large compound eye “on/off” units or inhibition of the “off” effect (Fig. 39). It can also cause reduction in the rate of firing of a sustaining fibre. This type of interaction is not obtained consistently, neither is it possible to relate these effects to the stimulus parameters except t o say that they operate at fairly high intensities. Palka has shown that the response of the large “on/off” compound eye units in the connectives is very much dependent upon the previous history of

A

r”

I . .

1

I..

.

.

.

.

B Fig. 39. The response of a V.N.C. unit firing to brightening and dimming of part of the left compound eye, A, is depressed when the three ocelli are occluded, B. Lower trace indicates duration of light stimulus. Trace reads from left to right.

stimulation of the eye and this is true also of visual interneurons in the connectives of other insects (Chapple, 1967). Techniques in which the collective pattern of activity in the interneurons can be recorded and analysed over a longer period of time may reveal more about the instantaneous integration normally occurring. At this stage certain conclusions can be drawn about the information signalled by the ocelli to the thoracic ganglion level. They contribute information about changes in brightness and about dimming in particular via at least three different types of unit. They have a large fibre pathway t o the thoracic ganglia in which the onset of dimming is signalled rapidly and apparently without modification from the other set of photoreceptors. Dimming is also signalled in another pathway but this information reflects the integration at brain level of information from the respective ocelli and the compound eyes. Increases of brightness must be large t o be signalled and here also the signal can reflect the change registered in both types of photoreceptor. Response t o an increase in brightness presented to the median ocellus, for example, can be inhibited or reduced by presenting an increase t o the contralateral compound

184

LESLEY J GOODMAN

eye. Although the site of interaction of inputs from the various photoreceptors has not yet been established at higher intensities the median ocellus and the contralateral ocellus can be seen to interact in each connective and also the contralateral eye and ocellus. The interactions recorded have so far been all seen at high intensities and all are inhibitory. Ocellar input can also have an inhibitory input on compound eye units. From observations made to date the impression is gained that the compound eyes exert an inhibitory influence mainly on brightening responses in ocellar units whereas the ocelli exert inhibitory influences mainly on dimming responses in other ocellar units and on brightening responses on compound eye units. The ocellar nerve signals both phasic and tonic information to the protocerebrum and phasic information is certainly transmitted down the connectives. However only one ocellar unit signalling tonically in light has been found in the connectives and there is no other evidence that information about net luminosity is passed down to this level, although the recording and analysis techniques employed would probably not have detected the transmission of such in forma tion. D. THE INFLUENCE OF THE OCELLI ON MOTOR ACTIVITY IN THE THORACIC GANGLIA

The only electrophysiological information on the effect of ocellar input on motor output comes also from the locust. In nerves 2 and 3, in the mesothoracic segment, motor units have been found which can be driven by both compound eye and ocellar stimulation. Two types of unit have been found, one which is silent in darkness and fires in response to compound eye and ocellar stimulation and one which fires tonically, having its rate of firing speeded up or slowed down by visual stimulation (Goodman, 1968). Two classes of response have been obtained from “silent” fibres. Some units respond with a large train of impulses to contralateral compound eye stimulation and a smaller train in response to median ocellar stimulation. Simultaneous stimulation of the two photoreceptors results in an apparent summation of their effects. Others have apparently opposite characteristics, responding more strongly to ocellar stimulation than compound eye stimulation. Simultaneous stimulation here results in an apparent averaging of inputs (Fig. 40). It is possible that a single unit is exhibiting different response characteristics in different preparations due to variation in input from the two receptors amongst other factors. However it has not proved possible to alter

185

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

the type of response given by such an individual unit by varying the stimulus parameters. Recordings have only been made whilst stimulating the median ocellus and the contralateral compound eye and other photoreceptor combinations may elicit a different response from the motor unit. A

-------B

Ocellar stim

I

I

1

Compound eye stim

,

Ocellar stim

1

Ocellar sti m

I

I

Compound

I

I

Compound

eye stim

eye stim D Fig. 40. A, A compound eye on/off unit recorded from the right connective between the pro- and mesothoracic ganglia. Bottom trace indicates duration of stimulation of compound eye. Note trace reads from right to left. B, Unit does not fue in response to stimulation of the median ocellus alone. C, If the median ocellus is stimulated just before compound eye stimulation ceases the offeffect is inhibited. D, If median ocellar illumination precedes the compound eye stimulation the oneffect can be depressed and offeffect inhibited.

C

Tonic units firing at a low frequency during darkness and speeded up by changes of intensity at the ocellus and the compound eye have been found. The frequency and duration of the train of impulses resulting from light "off" at the ocellus is related to the duration and intensity of the preceding light stimulus. Firing of the ocellar "off" unit in the connective is normally followed by a large train of impulses in these tonic motor units. Tonic units which can be

186

LESLEY J . G O O D M A N

speeded up by stimulation of one type of photoreceptor and slowed down by stimulation of the other can be found. Exploratory recordings have yielded units in which the rate of firing is increased by stimulation of the median ocellus and slowed by stimulation of a restricted area of the contralateral compound eye and others in which firing has been slowed by stimulating the median ocellus and speeded up by stimulating the contralateral compound eye (Fig. 41 ). No records are available for other combinations of the pho torecep tors. Electrophysiological evidence to date has shown that in the locust

I

D

P

I

Fig. 41. Motor units in nerves 2 and 3 of the mesothoracic ganglion of S. gregaria can be driven by both compound eye and ocellar stimulation. A, Motor unit in nerve 2 driven by stimulation of the left compound eye. Bottom trace indicates duration of compound eye illumination. All traces read from right to left. B, The same unit driven by stimulation of the median ocellus. Middle trace indicates duration of ocellar stimulation. C, The same unit responding to simultaneous stimulation of the median ocellus and left compound eye. D, A second motor unit (nerve 2) stimulated by the left compound eye. Duration of compound eye stimulation shown on bottom trace. Stimulation of the median ocellus (median trace) slows the rate of firing in this case. A long after discharge is seen.

STRUCTURE AND FUNCTION O F INSECT DORSAL OCELLUS

187

units carrying visual information from the compound eyes and others carrying information from the ocelli are present at the level of the thoracic ganglia and are capable of driving thoracic motoneurons. It is not known which muscles the recorded motoneurons supply nor is it known if any locomotory effect is produced by visual stimulation, although in view of behaviour studies some effect seems probable. The fact that simultaneous compound eye and ocellar activity can A

Ocellar stim

I

B

I

Comp eye

I

Fig. 42. A motor unit firing at a low frequency in darkness is inhibited by stimulating the left compound eye, A, and speeded up by stimulating the median ocellus, B. Duration o f stimulation indicated by bottom trace in A, top trace in B. Trace reads from right to left.

have either a synergistic or an antagonistic effect on thoracic motoneurons provides a physiological basis for Jander and Barry’s proposed model of interaction of the two sets of photoreceptor in the locust. Interaction appears to take place at two levels. Although one visual unit at least, the ocellar “off” unit, descends to thoracic

188

LESLEY J. GOODMAN

ganglion level without being influenced by other photoreceptors, the activity of most visual units can reflect inhibitory interaction at cerebral ganglion or in the case of compound eye units possibly at optic lobe level. Integration of activity takes place at thoracic ganglion level where interneurons from both sets of receptors interact with individual motoneurons. Jander and Barry propose a change from synergistic to antagonistic interaction between the compound eyes and ocelli as the intensity increases, the median ocellus providing the switching mechanism. At this stage it is not possible to propose a detailed physiological mechanism for the change-over except to point out that the inhibitory interaction at brain level becomes more pronounced as intensity increases. Several authors have commented upon the loss of tonus when ocelli are occluded but most of them have considered it as a purely central effect (see Section 111). Whilst the ocelli have been shown to influence the activity of interneurons in the brain associated with various modalities, they have also been shown to drive certain interneurons in the thoracic connectives and thoracic motoneurons and may contribute directly to the promotion of normal muscle tonus at thoracic ganglion level. Although the ocelli can make some contribution to motor output they are evidently not normally able to activate command interneurons capable of initiating the complex patterns of motor output required for locomotory activity. They cannot initiate motor behaviour but can contribute to fine control as exercised for example in steering or in rate of movement. VI. CONCLUSION

It is now fairly clear how the ocellus functions as a photoreceptor; it signals phasic and tonic information about intensity levels to the brain. Ocellar structure, with the dioptric system forming an image outside the retinal layer and a high degree of convergence on to second order neurons, suggested such a function to earlier workers and physiological information has brought confirmation. So far all ocelli have been found to signal in the same way with light modulating a dark discharge in the ocellar nerve via a first order inhibitory junction. Light “on” is signalled phasically by sudden silence and “off” by a brief high frequency discharge; the steady rate of discharge maintained over long periods for a given intensity yields

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

189

information about absolute intensity level. Recordings made from the ocellar nerve have probably all been from the larger fibres and the properties of the small fibres present in some cases are unknown. The electrical activity of the retinular cells and their axons and the inhibitory junction with the second order neurons are of intrinsic interest, especially since evidence of first order inhibitory synapses is accumulating from other arthropod photoreceptors. Studies made on this eye are uncomplicated by possible effects of underlying neuropile areas but for various reasons the material is more difficult to work with than the compound eye. The nature of the input once made it a promising candidate for the entrainment and maintenance of circadian rhythms. This role has now been strongly disputed but it is perhaps too soon to say that the ocellus has no neurosecretory function at all and this possibility requires further investigation. Unfortunately we know little of how and where the ocellar input is integrated within the central nervous system and in consequence can still say little in physiological terms about the role of the ocellus in insect behaviour. Until recently behavioural observations have not given much impetus to physiological studies. Only the growing realization that the ocelli were implicated in phototactic orientation and must therefore interact at some level with the compound eye input to modify steering has provoked the formulation of functional models and directed a search for the underlying physiological mechanisms. The ocelli do not appear to initiate motor responses yet they obviously have some effect on locomotory activity which has led some workers to the view that they act solely as stirnulatory organs influencing the general level of excitation centrally at some unstated level. Since ocellar input declines with rising intensity whilst locomo tory responsiveness increases the simplest idea is to assume that ocellar illumination causes a release of inhibition at some level. The input of the darkened ocellus might contribute to the inhibition exercised over the motor centres by certain areas of the protocerebrum. Control of the time of the first and last foraging flights in bees might be explained in these terms, increasing light causing a gradual reduction in the inhibition of the excitatory motor centres. Such evidence as we have of ocellar activity in the brain suggests that the true situation is not nearly so simple. Ocellar involvement in phototactic orientation presupposes a specific interaction between individual ocellar and compound eye inputs and electrophysiological records have shown that this does occur. The A.I.P.-I*

190

LESLEY J . GOODMAN

ocelli can drive interneurons found in the ventral nerve cord down to the level of the thoracic motor centres. The compound eyes drive other interneurons present in the cord and recordings from these two types of unit show that interaction between compound eye and ocellar input must take place at supra- or suboesophageal ganglion level since illumination of a compound eye can under certain circumstances depress or inhibit the response of a unit to ocellar stimulation and vice versa. Interneurons which can be driven by both types of photoreceptor may also be present. Motoneurons that can be driven either synergistically or antagonistically by the two types of photoreceptor have been found in the thoracic segments and some integration of compound eye and ocellar input apparently takes place at thoracic motor centre level. The function of the ocellar giant "off" unit found in the VNC is of interest in view of the inability of the ocelli to initiate motor activity. It has many of the properties of other insect giant axons, is unaffected by compound eye stimulation and appears able to initiate trains of impulses in motoneurons. Presumably it is involved in startle reactions to dimming possibly by lowering thresholds to subsequent stimuli from the compound eyes travelling via other interneurons. The ocelli can certainly no longer be regarded as exercising solely a general stimulatory effect and it is possible that this is not their major role at all. Recent observations by Mimura et al. (1969) show that ocellar illumination can raise or depress spontaneous activity in antennal units and increase or inhibit responses to antennal stimulation but this does not make the ocelli unique since other sense organs can produce similar effects in multimodal units. It is of interest to know whether threshold levels of other modalities are affected in such units by ocellar stimulation, something which has not so far been demonstrated. The possibility of an ocellar contribution has been somewhat neglected by electrophysiologists looking at the properties of interneurons in the brain and ventral nerve cord in the past. It is to be hoped that this may be remedied in the future. REFERENCES Autrum, H. J. (1950). Die Belichtungspotentiale und das Sehen der Insekten. 2. vergl. Physiol 32, 176-227. Autrum, H. J. (1958). Electrophysiological analysis of the visual systems in insects. Expl. Cell. Res. Suppl. 5 , 426439. Autrum, H. J. and Stoecker, M. (1950). Die Verschmelzungsfrequenzen des Bienenauges. Z. naturf. 56, 38-43. Autrum, H . J. and Gallwitz, U. (1951). Zur Analyse der Belichtungspotentiale der Insectenaugen. 2. vergl. Physiol. 33,407.

STRUCTURE A N D FUNCTION O F INSECT DORSAL OCELLUS

191

Autrum, H. J. and Hoffman, E. (1957). Die Wirkung von Pikrotoxin und Nikotin auf das Retinogramm von Insekten. Z. naturf: 126, 752-757. Autrum, H. J. and Hoffmann, C. (1960). Diphasic and monophasic responses in the compound eye of Calliphora. J. ins. Physiol. 4, 122-127. Autrum, H. J. and Metschl, N. (1963). Die Arbeitsweise der Ocellen der Insekten. Z. vergl. Physiol. 47, 256-273. Autrum, H. J. and von Zwehl, V. (1962). Die Sehzellen der Insekten als Analysatoren fur polarisiertes Licht. Z. vergl. Physiol. 46, 1-7. Baccetti, B. and Bedini, C. (1964). Research on the structure and physiology of the eyes of a Lycosid spider. Arch’s. ital. Biol. 102, 97-122. Baumann, F. (1969). Slow and spike potentials recorded from retinula cells of the honey-bee drone in response to light. J. gen. Physiol. 52, 855-875. Bayramoglu-Ergene, S. ( 1964). Untersuchungen iiber den Einfluss der Ocellen auf die Fluggeschwindigkeit der Wanderheuschrecke Schistocerca gregaria Z. vergl Physiol. 49,461480. Benolken, R. M. (1962). Effects of light and dark adaptation processes on the generator potential of the Limulus eye. Vision Res. 2, 103. Bernstein, M. (196 1). In “Structure of the Eye” (G. K. Smelser, ed.). Academic Press, New York. Borsellino, A., Fuortes, M. G. F. and Smith, T. E. (1 965). Visual responses in Limulus. Cold Spring Harb. S y m p . quant. Biol. 30,429-443. Bozler, E. (1926). Experimentelle Untersuchungen uber die Funktion der Stirnaugen der Insekten. Z. vergl. Physiol. 3, 145-182. Bullock, T. H. and Horridge, G. A. (1965). “Structure and Function in the Nervous Systems of Invertebrates”. Vols I and 11. W. H. Freeman & Co., San Francisco and London. Bunning, E. (1936). Die endogene Tagesrhythmik als Grundlage der photoperiodischen Reaktion. Ber. d t . bot. Ges. 54, 590-607. Bursell, E. (1967). The excretion of nitrogen in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne, and V. B. Wigglesworth, eds). Vol. 4, pp. 33-69. Burtt, E. T. and Catton, W. T. (1954). Visual perception of movement in the locust. J. Physiol. 125, 566-580. Burtt, E. T. and Catton, W. T. (1956). Electrical responses to visual stimulation in the optic lobes of the locust and certain other insects. J. Physiol. 133, 68-88. Burtt, E. T. and Catton, W. T. (1959). Transmission of visual responses in the nervous system of the locust. J. Physiol. 146,492-5 15. Caesar, C. J. (1913). Die Stirnaugen der Ameisen. Zool. Jb. (Anat.) 35. 161-242. Cassier, P. (1960). Le phototropisme du criquet migrateur. C.r. SOC.zool. Fr. 85, 165. Cassier, P. (1 962). R61e des ocelles frontaux chez Locusta migratoria migratoriodes (R. et F.). Insectes SOC.9 , 213. Cassier P. (1965). Contribution h 1’6tude des reactions photomenotactique du criquet migrateur: Locusta migratoria migratoriodes (R. et F.) leur determinisme sensoriel. Insectes SOC. 12, 365-38 1. Chapple W. D. (1966). Motoneuron responses to visual stimuli in Oncopeltus fasciatus Dallas. J. exp. Biol. 45,401-410. Chen, S . H. and Young, B. (1943). Vision and flight: an experimental study on the cicada, Cryptotympana pustulata. Sinensia, Shanghai 14, 55-60.

192

LESLEY J . GOODMAN

Cloudsley-Thompson, J. L. (1953). Studies in diurnal rhythms. 111. Photoperiodism in the cockroach, Periplaneta americana. A n n . Mag. nat. Hist. 6, 705-712. Cornwell, P. B. (1955). The functions of the ocelli of Calliphora (Diptera) and Locusta (Orthoptera). J. exp. Biol. 32, 2 17-237. Curtis, D. J. (1969). The fine structure of photoreceptors in Mitopus morio (Phalangida). J. Cell. Sci. 4, 327-35 1. Demoll, R. and Scheuring, L. (1912). Die Bedeutung der Ocellen der Insekten. 2001.Jb. (Abt. Zool. Physiol.) 31, 519-628. Dingle, H. and Caldwell, R. L. (1967). Multimodal interneurons in Cockroach cerebrum. Nature Lond. 215.63-64. Dufay, C. (1964). Contribution h l’btude du phototropisme des lbpidoptkres Noctuides. Annls Sci. nat. Zool. Ser. MI. 6, 28 1-406. Fernandez-Morbn, H. (1958). Fine structure of the light receptors in the compound eyes of insects. Expl. Cell. Res. Suppl., 5, 586-644. Fuortes, M. G. F. (1958). Electric activity of cells in the eye of Limulus. A m . J. Ophthal. 46, 210-223. Fuortes, M. G. F. and Yeandle, S. (1964). Probability of occurrence of discrete potential waves in the eye of Limulus. J. gen. Physiol. 47,443-463. Goldsmith, T. H. and Ruck, P. (1958). The spectral sensitivities of the dorsal ocelli of cockroaches and bees-an electrophysiological study. J. gen. Physiol. 41, 1171-1 185. Goodman, L. J. (1960). The landing responses of insects. I. The landing response of the fly, Lucilia sericata, and other Calliphorinae. J. exp. Biol. 37, 854-878. Goodman, L. J. (1965). The role of certain optomotor reactions in regulating stability in the rolling plane during flight in the desert locust Schistocerca gregaria. J. exp. Biol 42, 385407. Goodman, L. J. (1968). Interaction between ocellar and compound eye visual units in Schistocerca gregaria. Proc. 13th Int. Congr. Entomology, Moscow. Goodman, L.J. (1969) (in preparation). Goodman, L. J. and Kirkham, B. (1970) (in preparation). Gotze, G. (1927). Untersuchungen an Hymenopteren uber das Vorkommen und die Bedeutung der Stirnaugen. 2001.Jb. (Abt. Zool.) 44, 21 1-268. Goustard, M. ( 1958). R6action phototropique et rbgulation du comportement chez l’image male de Blatella germanica. Bull. biol. Suppl. 45, 1-1 12. Hbmori, J. and Horridge, G. A. (1966). The lobster optic lamina. 11. Types of synapses. J. Cell. Sci. 1,257-270. Harker, J. E. (1956). Factors controlling the diurnal rhythm of activity in Periplaneta americana. J. exp. Biol. 33, 224-234. Hertweck, H. (193 1). Anatomie und Variabilitat des Nervensystems und der Sinnesorgane von Drosophila melanogaster. 2. wiss. Zool. 139, 559-663. Hesse, R. (1908). Untersuchungen uber die Organe der Lichtempfindung bei niederen Thieren. VII von den Arthropoden Augen. Z . wiss. 2001.70, 347-473. Homann, H. (1924). Zum Problem der Ocellenfunktion bei den Insekten. 2. vergl. Physiol. 1, 541-578. Horridge, G. A. (1968). “Interneurons”. W.H. Freeman & Co., London and San Francisco.

STRUCTURE AND FUNCTION OF INSECT DORSAL OCELLUS

193

Horridge, G. A. and Barnard, P. B. T. (1965). Movement of palisade in locust retinula cells when illuminated. Q. Jl microsc. Sci. 106, 13 1-135. Horridge, G. A., Scholes, J. H., Shaw, S. and Tunstall, J. (1965). Extracellular recordings from single neurons in the optic lobe and brain of the locust. I n “The Physiology of the Insect Nervous System” (J. E. Treherne and J. W. L. Beament, eds). Academic Press, London and New York. Hoyle, G. (1955). Functioning of the insect ocellar nerve. J. exp. Biol. 32, 397-407. Jander, R. (1963). Insect orientation. A . Rev. Ent. 8. Jander, R. and Barry, C. K. (1968). Die phototaktische Gegenkopplung von Stirnocellen und Facettenaugen in der Phototropotaxis der Heuschrecken und Grillen (Saltatoptera: Locusta migratoria und Cryllus bimaculatus). Z. vergl. Physiol. 57,432-458. Jawlowski, H. (1958). Nerve tracts in the bee (Apis mellifica) running from the sight and antenna1 organs to the brain. Annls Univ. Mariae CurieSklodowska (C), 12, 307-323. Kalmus, H. (1945). Correlations between flight and vision, and particularly between wings and ocelli in insects. Proc. R . ent. SOC.Lond. A 20,84-96. Kennedy, D. (1964). The photoreceptor process in lower animals. In “Photophysiology” (A. C. Giese, ed.). Academic Press, New York, pp. 79-1 2 1. Koehler, 0. ( 1950). Die Analyse der Taxisanteile instinkartigen verhaltens. Syrnp. SOC.exp. B i d . 4, 361-384. Kolbe, H. J. (1893). Insektenkunde, Berlin. Lammert, A. (1925). Pigment movements in ocelli. 2. vergl. Physiol. 3, 225-278. Lees, A. D. (1964). The location of the photoperiodic receptors in the aphid, Megoura vici5e Buckton. J. exp. B i d . 41, 119-133. Link, E. (1908). Uber die Stirnaugen der Orthopteren. Verh. d t . zool. Ges., 161-167. Link, E. (1909a). Uber die Stirnaugen der hemimetabolen Insekten. Zool. Jb. (Abt. Anat Ontog.) 27, 281-376. Link, E. (1909b). Uber die Stirnaugen der Neuropteren und Lepidopteren. 2001. Jb. 27,213. Lowne, B. (1870). “The Anatomy and Physiology of the Blow Fly”. London, pp. 121. Lubbock, J. (1889). “The Senses, Instincts and Intelligence of Animals with Special Reference to Insects”. London, pp. 292. MBdioni, J. (1959). Les organes de stimulation photique chez Drosophila rnelanogaster. Non-spBcifitB de cette stimulation rklativement aux reactions a la lumikre. C.r. Seanc. SOC.Biol. 153, 1845. Metschl, N. ( 1963). Electrophysiologische untersuchungen an den ocellen von Calliphora. 2. vergl. Physiol. 47, 230-255. Mittelstaedt, H. ( 1950). Physiologie des Gleichgeswichtsinnes bei fleigenden Libellen. 2. vergl. Physiol. 32, 422-463. Mimura, K., Tateda, H., Morita, H. and Kuwabara, M. (1969). Regulation of insect brain excitability by ocellus. Z. vergl. Physiol. 62, 382-394. Muller, E. (193 1). Experimentelle Untersuchungen au Bienen und Ameisen iiber die Funktionsweise der Stirnocellen. Z. vergl. Physiol. 14, 348-384. Muller, J. ( 1826). “Zur vergleichenden Physiologie des Gesichtssinnes der Menchen und der Tiere”. Leipzig, p. 462.

194

LESLEY J . GOODMAN

Naka, K. (1961). Recording of retinal action potentials from single cells in the insect compound eye. J. gen. Physiol. 44, 571-584. Naka, K. and Eguchi, E. (1962). Spike potentials recorded from the insect photoreceptor. J. gen. Physiol. 45,663-680. Narahashi, T., Moore, J. W. and Scott, W. R. (1964). Tetrodotoxin blockage of sodium conductance increase in lobster giant axons. J. gen. Physiol. 47, 965. Nishiitsutsuji-Uwo, J. and Pittendrigh, C. S. (1968a). Central nervous system control of circadian rhythmicity in the cockroach. 11. The pathway of light signals that entrain the rhythm. 2.vergl. Physiol. 58, 1-14. Nishiitsutsuji-Uwo, J. and Pittendrigh, C. S. (1968b). Central nervous system control of circadian rhythmicity in the cockroach. 111. The Optic Lobes, Locus of the Driving Oscillation? 2.vergl. Physiol. 58, 14-46. Palka, J. (1967). An inhibitory process influencing visual responses in a fibre of the ventral nerve cord of locusts. J. Insect Physiol. 13,235-248. Parry, D. A. (1947). The function of the insect ocellus. J. exp. Biol. 24, 2 11-219. Power, M. E. (1943). The brain of Drosophila melanogaster. J. Morph. 72, 5 17-559. Power, M. E. (1948). The thoracicoabdominal nervous system of an adult insect, Drosophila melanogaster. J. cornp. Neurol. 88,347-409. Priesner, H. ( 1928). “Die Thysanopteren Europas”. Wien. Pittendrigh, C. S. and Minis, D. H. (1964). The entrainment of circadian oscillations by light and their role as photoperiodic clocks. A m . Nut. 98, 26 1-294. Redikorzew, W. (1900). Untersuchungen iiber den Bau der Ocellen der Insekten. Z. wiss. Zool. 68,581-625. Richard, G. (1956). Contribution i l’Etude du Phototropisme des Insectes au cours de leur vol. Proc. 10th Int. Congr.Entomology. Vol2. 1956-58. Roberts, S. K. (1965). Significance of endocrines and central nervous system in circadian rhythms. In “Circadian Clocks” (J. Aschoff, ed.). North Holland Publishing Co., Amsterdam, pp. 198-213. Ruck, P. (1957). The electrical responses of dorsal ocelli in cockroaches and grasshoppers. J. Insect Physiol. 1, 109-123. Ruck, P. (1958a). Dark adaptation of the ocellus in Periplaneta americana: a study of the electrical response t o illumination. J. Insect Physiol. 2, 189-198. Ruck, P. (1958b). A comparison of the electrical responses of compound eyes and dorsal ocelli in four insect species. J. Insect Physiol. 2,261-274. Ruck, P. ( 1 9 5 8 ~ ) . Post synaptic giant fibre response in ocellar nerves of dragonflies. Anat. Rec. 132,499. Ruck, P. (1961a). Electrophysiology of the insect dorsal ocellus. I. Origin of the components of the electroretinogram. J. gen. Physiol. 44, 605-627. Ruck, P. (196 lb). Electrophysiology of the insect dorsal ocellus. 11. Mechanisms of generation and inhibition of impulses in the ocellar nerve of dragonflies. J. gen. Physiol. 44, 629-639. Ruck, P. (196 lc). Electrophysiology of the insect dorsal ocellus. 111. Responses to flickering light of the dragonfly ocellus. J. gen. Physiol. 44, 641-657. Ruck, P. and Edwards, G. A. (1964). The structure of the insect dorsal ocellus. I. General organisation of the ocellus in dragonflies. J. Morph. 115, 1-26.

STRUCTURE A N D FUNCTION OF INSECT DORSAL OCELLUS

195

Ruck, P. and Goldsmith, T. H. (1958). Spectral sensitivities of dorsal ocelli in cockroaches and honeybees. Fedn. Proc. Fedn. A m . SOCSexp. Biol. 17, 137. Rutherford, D. J. and Horridge, G . A. (1965). The rhabdom of the lobster eye. Q. Jl microsc. Sci. 106, 119-130. Satija, R. C. (1957). Studies on the course and functions of the giant nerve fibres from the eye and the tegumentary nerve to the ventral nerve cord in Locusta migratoria. Res. Bull. Panjab. Univ. Sci. 132, 51 1-519. Satija, R. C. (1958). A histological and experimental study of nervous pathways in the brain and thoracic nerve cord of Locusta migratoria migratoriodes (R.& F.). Res. Bull. Panjab. Univ. Sci. 137, 13-32. Scholes, J. H. (1965). Discontinuity of the excitation process in locust visual cells. Cold Spring Harb. Symp. quant. Biol. 30, 5 17-527. Schremmer, F . ( 1950). Bemerkungen zur Ocellenfunktion bei Hummeln. 6 s t . ZOO^. Z. 2,242-274. Schricker, B. (1965). Die Orientierung der Honigbiene in der Dammerung. Zugleich ein Beitrag zur Frage der Ocellenfunktion bei Bienen. Z . vergl. Physiol. 4 9 , 4 2 0 4 5 8 . Trujillo-Cenoz, 0. (1 965). Some aspects of the structural organization of the arthropod eye. Cold Spring Harb. Symp. quant. Biol. 30, 371-382. Uchizono, K. (1965). Characteristics of excitatory and inhibitory synapses in the central nervous system of the cat. Nature, Lond. 207, 642. Waddington, C. H. and Perry, M. M . (1961). The ultrastmcture of the developing eye of Drosophila. Proc. R . SOC.B. 153, 155-178. Walther, J. B. and Dodt, E. (1957). Electrophysiologische Untersuchungen iiber die Ultraviolettempfindlichkeir von Insektenaugen. Experientia 13, 133. Wellington, W. G. (1953). Motor responses evoked by the dorsal ocelli of Sarcophaga aldrichi Parker, and the orientation of the fly to plane polarised light. Nature, Lond. 172, 11 77-1 179. Wolsky, A . (1930). Optische Untersuchungen iiber die Bedeutung und Funktion der Insektenocellen. Z . vergl. Physiol 12, 783-787. Wolsky, A . (193 1). Weitere Beitrage zum Ocellenproblem. Die optischen Verhaltnisse der Ocellen der Honigbiene. Z . vergl. Physiol. 14, 385-391. Wolsky, A . (1933). Stimulationsorgane. Biol. Rev. 8, 370.

This Page Intentionally Left Blank

Polarity and Patterns in the Postembryonic Development of Insects PETER A . LAWRENCE Department of Genetics. Cambridge. England

*

I. I1.

Introduction . . . . . . . . . . . . . . . . . . Cell Polarity . . . . . . . . . . . . . . . . . . A. Insect Segmental Gradients . . . . . . . . . . . B. Origin of the Segmental Gradient . . . . . . . . . C. Polarized Transmission of Information During Growth and . . . . . . . . . . . . . . . . Regeneration D. Growth and Gradients . . . . . . . . . . . . . E. Gradients and Patterns . . . . . . . . . . . . . F. Insect Segmental Gradient: a Summary of the Working Hypo thesis . . . . . . . . . . . . . . . . . G . Gradient Phenomena in Other Organisms: a Comparison . . I11. Pattern Formation . . . . . . . . . . . . . . . . A. The Development of Spaced Bristles and Hairs inRhodniusand Oncopeltus . . . . . . . . . . . . . . . . . B . Genetic Mosaics . . . . . . . . . . . . . . . IV . Determination and Regulation . . . . . . . . . . . . A. Introduction . . . . . . . . . . . . . . . . B. Experiments with Cut Discs -Regulation . . . . . . . C. Dissociation Experiments . . . . . . . . . . . . D . Changes in the Determined State . . . . . . . . . . V . Cellular Differentiation . . . . . . . . . . . . . . . VI . Outlook . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .

197 198 199 206 207 212 216 220 221 224 224 231 235 235 237 247 254 257 259 260 260

I . INTRODUCTION

Developmental biology is still a largely uncharted subject: analysis of developing embryos is difficult mostly because so much is happening at any one moment that it becomes impossible to observe * Present address: MRC Laboratory of Molecular Biology. Cambridge. England . 197

198

PETER A. LAWRENCE

individual components of the morphogenetic machinery. In all organisms development continues beyond the embryonic stage, where it often takes a simpler form. In insects this postembryonic development usually occurs as growth and metamorphosis, particularly of the epidermis, and here it is a help that any changes in the epidermis are permanently registered in the cuticle. I hope t o show in this chapter that basic embryological phenomena do recur at insect metamorphosis, and how work with insects has produced some central information. There are dangers in the assumption that embryonic and later development depend on similar mechanisms, but in spite of these dangers, knowledge in depth should be the first aim. Developing cells are usually polarized in relation to the main axes of symmetry of the organism. During epigenesis, t o the continued accompaniment of growth, the possible destinies of individual clones of cells become progressively defined. Once the developmental identity of a cell is established, it is said to be determined and the process which orders requisite determinations in space is called pattern formation. Often the mechanisms of pattern formation are plastic so that they adapt to interference or loss of parts so as to reconstruct the whole pattern from the available material. This process is termed regulation. After a cell reaches its final determined status it constructs specialized organelles and synthesizes specific biochemical products-it differentiates. Although these five features of developing systems: polarity, determination, pattern formation, regulation and differentiation, are each components of one process, they can, to a certain extent, be treated independently. In this essay they will be discussed in relation to development of postembryonic insects. I have concentrated only on telling examples and have made no attempt to include all references in any of the areas considered. 11. CELL POLARITY

Individual cells are asymmetric in structure and in function. Insect epidermal cells sit on a basement membrane and secrete a cuticle apically. However, there is another, more subtle orientation: the first layer of. the exocuticle contains chitin fibrils that are laid down in a particular direction-usually the antero-posterior axis of the body, or the disto-proximal axis of the appendage (Neville, 1967; Neville and Luke, 1969). This might indicate some extracellular stress which aligns the fibrils, but as we shall see, experiments point to the conclusion that the orientation of these fibrils is a reflection

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

199

of an intrinsic polarity in the underlying cells. This polarity in the epidermis also has direct expression in the orientation of asymmetric structures, such as scales or bristles, which commonly grow out antero-posteriorly . While the basic cue t o asymmetry may be in the egg itself, the complex polarities of the highly folded and structured epithelia which make up the mature insect originate as part of the epigenetic process. It is often not possible to analyse these events in the egg, but it is highly probable that the mechanisms which act during the genesis of cell-polarity there, also act to maintain it during growth and regeneration. A. INSECT SEGMENTAL GRADIENTS

After wounding insect epidermal cells migrate individually from the surrounding area into the damaged parts (Wigglesworth, 1937). It seems unlikely that they retain their polarity during this amoeboid movement, and indeed when bristles are regenerated by the

Fig. 1. Two stages in the acquisition of a preferred orientation in the differentiating hair cells in Oncopeltus. The alignment of the three-cell groups progresses from the early (open circles) to the late (closed circles) stages. (From Lawrence, 1966a.)

200

PETER A. LAWRENCE

reconstituted epidermis they are often poorly orientated (Wigglesworth, 1940a, 1959). At subsequent moults the bristles are reorientated to their normal alignment, and after large wounds this proper orientation seems to spread in centripetally from moult to moult (Lawrence, 1966a). It would seem that cellular polarity is retrieved from the undisturbed peripheral cells. During the development of scales, hairs and bristles in a moth (Piepho and Marcus, 1957), a bug (Lawrence, 1966a) and in Drosophilu (Stern and Hannah, 1950; Tokunaga and Stern, 1969) the formative cells acquire an orientation from the cells around them (Fig. 1); an observation which implies that polarity can be communicated from cell to cell.

am Fig. 2. An abdominal segment of adult Culleriu after reversal of a square of integument in the last stage larva. Arrows mark the orientation of scales. Note the central area of reversed polarity. am = anterior margin, pm = posterior margin of intersegmented membrane. (From Piepho, 1955a.)

What, therefore, would happen to a cell placed between two masses of cells of mutually opposing polarity?Piepho (1955a, b) investigated this question by rotating a square of Galleria larval integument and grafting the piece back (Fig. 2). The scales of the adult were orientated by polarizing influences emanating from both the rotated piece and the area around it; Piepho described these influences in terms of forces, and noted how the scales oriented in the direction taken by the resultant of these opposing forces. Locke (1959, 1967) also found that rotation of pieces of integument in

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

5th- stage larva 10 8

20 1

Adults

transplant

0 10

0

IW

Fig. 3. Locke’s experiments on Rhodnius. The figures mark the h e a t of the segmental gradient in the host segment and in the transplanted pieces. The adult ripple patterns are shown, in stylized form, on the right. At the bottom of the diagram the serially repeated segmental gradient is depicted; the thick lines represent intersegmental membranes. (After Locke, 1959.)

202

PETER A. LAWRENCE

Rhodnius disturbed the orientation of epidermal structures (in this case folds or “ripples” in the adult cuticle), whereas removal and replacement of squares of unrotated integument were without effect (Fig. 3). Transplantation up and down the segment in the proper orientation also produced alteration of the adult ripple pattern and the degree of pattern disturbance was related to the amount of displacement of the transplant. Since the polarity of the cells in the host and transplant was not altered by this experiment, it was clear that there was a gradient of some property within the segment, and it was interaction between cells of different gradient level which caused the disturbance. As transplantations between equivalent levels of adjacent segments had no effect on the ripple pattern, this gradient must be serially repeated in each segment. Lateral transplantations at any particular level had no effect on the adult pattern; this suggested that the gradient was the same along the segment. The segmental repetition of the gradient implies that the intersegmental membrane must intervene between the high point of one gradient and the low in the adjacent (Fig. 3), and indeed

Fig. 4. Hairs near a discontinuity in the intersegmental membrane (im) of an adult Oncopelfus. Note the reorientation of the hairs in the centre of the picture. (From Lawrence, 1966a.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

203

Rhodnius which have an interrupted intersegmental membrane also show a considerably altered orientation of cuticular structures (Locke, 1960). In Oncopeltus the hairs are orientated normally in the antero-posterior axis, but in insects bearing a natural discontinuity of the intersegmental membrane there is a complex but precise pattern of hairs, which are oriented in relation to it (Fig. 4). Insects such as these suggested a model of the insect segmental gradient in which I proposed (Lawrence, 1966a) that the gradient shared properties with a sand gradient, in that it has a maximum stable slope resulting from an equilibrium between two forces (gravity and friction between the sand grains). When a steeper slope is created experimentally, flow of sand is initiated and continues until the maximum stable slope is reconstituted. In this view the gap in the intersegmental membrane would result in an unstable situation, the sand then flowing from the high to the low to produce a new stable gradient topography (Figs 5 and 6). This model

Fig. 5 . The sand model-1. Glass plates separate two sand gradients. (From Lawrence, 1966a.) Fig. 6 . The sand model-2. The glass plates have been opened and the sand has flowed to ~ of the hairs in Fig. 4 and the set up a new, stable landscape. Note that i i orientation direction of the sand gradients are identical. (From Lawrence, 1966a.)

204

PETER A. LAWRENCE

provided a common explanation to Piepho’s and Locke’s results: their transplantation experiments set up unstable gradient situations (“sand precipices”) which resulted in flow of the gradient itself until new gradients were set up at the maximal steepness that was stable. In Galleria the scales point down (say) the gradients in the new landscape, and in Rhodnius the ripples run along the contours. A similar model of a concentration gradient of diffusible substance was proposed independently by Stumpf ( 1965a, b) and Lawrence (1966a). In this model, like the sand model, experiment produces an unstable disposition of the gradient itself, subsequent flow and the formation of a new landscape. It is not known how the gradient originates, or how it is maintained, but several models have been suggested. The intersegmental membrane of Galleria has two margins, one forming the anterior and one the posterior boundary of adjacent segments (Fig. 17). Marcus (1962) implanted small pieces of segment margin into different regions of the segment surface in Galleria; and these preparations were studied by Stumpf (1967a). At a distance from the implant the orientation of individual scales was the resultant of two influences: one emanating from the extant segment margins, and the other from the implanted piece. It was possible to deduce the relative contributions of these two influences from the alignment of the scales in relation to the distance from the transplant. There were many sources of experimental error, but Stumpf concluded that the orienting force attributable to the transplanted piece of margin declined more or less linearly with increasing distance from the transplant; whereas that attributable to the undisturbed segment margins was approximately constant all over the segment. This implied a linear gradient, and Stumpf therefore proposed that the anterior and posterior margins of the intersegmental membrane maintain two different concentrations of a substance, which is diffusing constantly from one margin to another. If the passage of material from anterior to posterior (say) were rapid, local disturbances resulting from transplantation would soon be overcome. As this does not happen (disturbed ripple patterns can persist for three moults in Rhodnius-several weeks) a substance with a very low rate of diffusion was postulated. An objection t o this model arises from an elegant experiment of Piepho’s in which he generated a patch of segment surface totally surrounded by just one segment margin. This resulted in a field of centrally pointing scales (Fig. 7) (Piepho, 1955b; Lawrence, 1970) and showed that a gradient of

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

donor

Presumptive adult posterior margin

205

host

t

Moults to adult (Fig. 7 b ) Fig. 7(a). Piepho's method of constructing a piece of Galleria segment surface totally surrounded by presumptive adult posterior segment margin (toothed Line). The dotted lines mark the cuts made. The three pieces (A, B and C) were transferred to a host as shown. (After Piepho, 1955b.)

Fig. 7@). The scale pattern after the operation. Note the even surround of intersegmental membrane, and the centripetal alignment of the scales. (From Piepho, 1955b.)

some sort does not depend on the presence of both margins. However, the additional postulate that the substance spontaneously breaks down can accommodate Piepho's result to Stumpf s model. I propose that the gradient is generated and maintained by the cells themselves, by active transport of a substance against the gradient. Such pumping by the cells sets up a stable situation, in which the forces of pumping and diffusion are in equilibrium. If the direction of the gradient slope is altered, the cells respond

206

PETER A. LAWRENCE

antagonistically and transport the substance against the new concentration gradient and thereby maintain it in its new orientation. This model suggests that the intersegmental membranes might act simply as passive impermeable barriers. It seems that a combination of both models fits the facts best: on the one hand, transplantation of pieces of anterior or posterior intersegmental margins into the segment surface produces profound alterations in the orientation of scales or ripples which suggests positive participation by two margins of opposite effect (Piepho, 1955b); on the other hand, deformation in the gradient landscape resulting from transplantation within the segment does point to the participation of the epidermal cells themselves. B. ORIGIN OF THE SEGMENTAL GRADIENT

Natural malformations cast some light on the origin of the segmental gradient system. The abdominal stemites of Oncopeltus can conveniently be described in terms of two lateral halves, for although typically these two halves fuse medially to give one entire sternite, occasionally one half fails to appear, possibly because it merges with an adjacent segment half (Fig. 8). Any defect in a sternite is not carried over to the equivalent tergite, and their development can therefore be regarded as independent. Similar malformations occur occasionally in many insects and frequently in stocks of Drosophila carrying certain mutations, (e.g. abnormal abdomen, Zimmermann, 1954) or can be elicited by heat shocks (Lobbecke, 1958). A detailed study of a strain of Drosophila showing a high frequency of segmental defects was undertaken by Sobels (1952) who found, in agreement with Maas ( 1 948) that the adult segment pattern descended directly from the larva with very few alterations. The most interesting type of defect, which was found only rarely, is the genesis of a segment in part bounded by two “anterior” or “posterior” margins. In the latter case the bristles orient towards both margins, with a divide in the middle. A similar defect in Oncopeltus is shown in Fig. 9. These malformed segments point to the intersegmental membrane as being vital to the genesis and organization of polarity of the segment surface; they suggest that the proper relations of the two segmental margins are an essential prerequisite for the normal development of each sternite and tergite. Indeed, from study of many kinds of segmental defects, Sobels (1 952) concluded “both the anterior and the posterior margin of a segmental border seem to be involved in the determination of the

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

-

clrnrn.

207

---

Fig. 8. Diagrammatic sketch of a segmental defect in the sternites of an individual Oncopeltus adult. On the right side of the diagram the normal sequence of segments is present. Note that segments 2 and 3 bear paired groups of large bristles centrally, whereas segments 4 and 5 each have two large bristles by the dorso-ventral muscle insertions (black bars). On the left side of the Figure the intersegmental membrane which separates segments 3 and 4 is missing and the combined segment, which is only as wide as one normal stemite, bears large bristles appropriate to both segments 3 and 4. The shading marks pigmentation.

outgrowth of the tergite anlage. The anterior margin mainly governs the differentiation of the anterior part of the tergite, whereas the posterior margin seems to be responsible for the differentiation of the caudal pigment band.” C. POLARIZED TRANSMISSION OF INFORMATION DURING

GROWTH AND REGENERATION

The balanced growth of a complex organ is easily taken for granted and yet t o achieve it, information exchange must take place

208

PETER A. LAWRENCE

Fig. 9. The fifth (5) sternite of an Oncopeltus adult. Over a restricted region the orientation of hairs and bristles is outwards from the centre. Note that pigmentation and the length of the bristles suggest that in this region the segment is bounded by two “posterior” margins. am = anterior margin: pm = posterior margin of normal region of segment.

between the various parts. One such organ has been analysed by Locke (1958). In the ramifications of the tracheal system of Rhodnius the minimal total area of air tubes at any particular level limits the air-flow in the whole system, and in fact measurements of tracheae showed that the total cross-sectional area of the branches remained constant right through the system, whether measured at the base or at the finely branched tips. When the tracheae branch distally during growth the proximal supply tracheae show a corresponding increase in diameter. If connections between these supply tracheae and the distal branches were cut, they survived normally and the distal system continued to branch. New cuticle was successfully formed by all parts of the system but the supply tracheae did not expand in diameter. This experiment suggested that growth stimuli normally pass down the tracheal epithelium from the tissues to the basal trunks. By means of similar experiments Locke

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

209

discovered there were two sources of growth stimuli: the distal branches in the tissues and the nodes (sites of cuticular breaks at ecdysis). Growth information could pass only in one direction along the tracheae. Here there was no evidence for a gradient as such, only for polarity. Locke (1 964) did however suggest that the nodes might

1 ..._.

Fig. 10. The effect of cutting a V-shaped piece from the tibia of Leucophaea. The next instar bears a lateral regenerate. (From Bohn, 1965b.)

210

PETER A. LAWRENCE

be equivalent t o the intersegmental membranes in the part they play in the polarized transmission of growth stimuli (p. 2 16). When an insect limb is -cut off, regeneration of a new limb from a wound blastema of cells may ensue. A simple cut in the limb, although like amputation in inducing cell multiplication and cellular migration, does not lead to regeneration. It would seem t o be the presence of the distal limb itself which effectively inhibits expression of the latent powers of the wounded limb cells to regenerate. If the cut is made in a Leucophaea limb as a deep V , the proximal portion of cut epithelium apparently receives no information from the extant limb for it makes a regenerate (Fig. 10) (Bohn, 1965b). This experiment illustrates the inability of these messages to by-pass gaps by making a lateral detour, and suggests, as with the tracheal experiments, that information is transmitted only in a particular direction. Confirmation comes from another experiment by Bohn: amputation, followed by grafting of the distal portion back on to the stump, effectively stops regeneration by the stump blastema (Bohn, 1965a). Thus the epithelium has grown together. If instead of the appropriate limb, a leg from the other side of the animal is grafted on to the stump (Fig. 1 1 ) information does not effectively pass to the

/

i

'3

... I

!

16 * '

a

P

P

Fig. 1 1 . Transplantation of a distal portion of the limb from side to side in Leucophaea. After the operation the antero-posterior axes (a, p ) are compatible in host and transplant, whereas the dorso-ventral axes (d, v ) are not. (From Bohn, 1965b.)

POLARITY AND PATTERNS IN DEVELOPMENT O F INSECTS

21 1

stump, and at the two sites of axial incompatibility two supernumerary limbs regenerate from the proximal wound tissue (Bodenstein, 1937; Bart, 1965a, b; Bohn, 1965b). Transplantation from left to right with rotation through 180" (Fig. 12) results in axial incornpatability in the other axis of the limb and here also two supernumerary limbs grow out (Bohn, 1965b). One may conclude

Fig. 12. Transplantation of a distal portion of the limb from side to side, with 180' rotation in Leucophuea. This experiment creates incompatibility in the antero-posterior axis, but not in the dorso-ventral. Note that the regenerated limbs are in a different plane than in Fig. 1 1 . (From Bohn, 1965b.)

that information which inhibits regeneration can only move between cells whose symmetry is equivalent. This conclusion is complicated by some other experiments by Bohn in which he rotated the amputated portion of the limb through 90" or 180" and then grafted it back. This altered the relative situation of both axes and one might predict the regeneration of four extra limbs. However, the topological quandary was resolved by the gradual rotation of the distal portion of the limb until the proper

212

PETER A. LAWRENCE

\

a

I

!I

Q

Fig. 13.- 90° rotation of a distal portion o f the limb in Leucophaeu. Correct axial alignment is regained by compensatory rotation of the transplanted portion. No regeneration occurs. (From Bohn, 1965b.)

relationship between stump and transplant was re-established (Fig. 13). There is no explanation for this simple solution by the insect. D. GROWTH AND GRADIENTS

There is other evidence that segmental gradients are connected with the organization of growth; it would seem that the tissue has an

n

213

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

i

I I

1 Y

stu

-t

Reg

A Fig. 14. Transplantation between two tibia in Leucophaea. (a). Note after transplantation there is a gradient discrepancy at the junction. Intercalary regeneration occurs, and the spines o n the regenerated portion Reg, unlike those on the rest of the limb, grow out disto-proximally. The final tibia is twice the length of a normal tibia.

n

+- ' \

.......... .-...."

8' B

Fig. 14(b). Diagram of the gradient situation. (From Bohn, 1967.) A.I.P.-8

v !

\

214

PETER A. LAWRENCE

inherent steepness of gradient and growth continues until that steepness is reached-the gradient being determinant therefore of the size of the structure. For the clearest example of this we must turn again to the insect leg. In experiments homologous to his studies on the abdomen, Locke (1966) showed that the insect leg segments bear serially repeating gradients which are concerned in the orientation of bristles. Bohn ( 1967) found that individual segments of Leucophaea could regenerate excised tissue. Regeneration proceeded only until the lost tissue had been reconstituted, and he showed, moreover, that it was not the length of the tibia per se that the system was reconstructing but the steepness of the gradient (Fig. 14). By transplantation between the short mid tibia and the long hind tibia Bohn also demonstrated that the differences between the two

0)

'{

. ..j....A'.:

......I

t t

/

\$ \ \

*

.._.,_. . .... ... ... :: :. ! ,,,.;

\ \

\

~

b)

8.

--------* 4

$,,/

.. ...... ; .:.. :.

c)

.

i. .; ........ .. ...

Fig. 1 5 . Transplantation between the short mid-tibia and the longer hind tibia. The numbers represent the percentage of the total length of the tibiae. These experiments show that regeneration only occurs when there is a gradient discrepancy, and that the gradients are of different steepness, but not extent, in the tibiae of the two legs. (From Bohn, 1967.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

215

tibia reside not in the total extent of the gradient, but also in the declivity (Fig. 15). The gradients are homologous in the two limb segments for they will fuse into one, but nevertheless are of different steepness. Needless to say these observations raise further problems: How is the steepness of the gradient system varied in the different limbs? How is growth controlled quantitatively from moult to moult? This latter problem is highlighted by a segmental defect found occasionally in cultures of Oncopeltus (Lawrence, 1966a) (Fig. 16).

:.:.%::Ti ..:.,..:.:.o.

Fig. 16. Sketch of Oncopeltus sternum, showing central fusion of sternites 2 and 3. Note that sternites in the fused area are only together as wide as one normal segment.

In such insects the sternites 2 and 3 are fused medially, and together are only about the width of one normal sternite. Consequently the ventral abdominal cuticle is considerably shorter than the dorsal, and the whole abdomen is uncomfortably hunched. Earlier studies (Wigglesworth, 1942, 1964) had demonstrated that growth of the epidermis is a homeostatic response to cuticular expansion due to feeding. In the defective insects the lateral parts of sternites 2 and 3 and the tergites are the proper width, and witness that normal distension has occurred during feeding. If growth were

216

PETER A. LAWRENCE

completely dependent on stretching one would have expected the medial parts of segments 2 and 3 to grow as wide as two segments. They do not do so, showing that in addition to the effects of distension on epidermal cell division, there is also some kind of growth control stemming from the segment margins. Locke (1959) reported that integumental grafts only grew properly when they were oriented correctly, in relation to the intersegmental membranes. I found (Lawrence, 1965) that such grafts did grow regardless of their orientation, providing that the graft included a piece of intersegmental membrane, which again points to the segment margins as sources of growth control-as suggested by Locke (p. 209). E. GRADIENTS AND PATTERNS

The insect segment is an excellent model system of pattern formation. When, in the egg, a presumptive segment is generated, it may consist of fewer cells than the number of qualitatively different regions that will be present in the mature segment. Thus during growth, there is diversification into different regions. Formally this could either occur by means of unequal cell divisions, by which the different parts of a single cell’s cytoplasm are partitioned amongst the daughters (mosaic development) or alternatively by some supracellular determinative process which occurs much later (regulative development). An experimental investigation of the developing segment can allow a choice between these two hypotheses. The adult segment of Galleria bears strips of qualitatively different integument, but the larval cuticle is more or less homogeneous. Marcus (1962, 1963) showed that if small pieces of larval segment from different parts are cultured through metamorphosis in the abdominal cavity of a mature caterpillar, they develop into cuticle according to their prospective fate. This experiment indicates that the different regions of the segment are already determined prior to metamorphosis. However, transplantation of small pieces of cuticle in the larva, from presumptive region 0 into presumptive region 2 (Fig. 17) resulted in the reorientation of scales in accordance with the segmental gradient system, and moreover the two tissues interacted to produce a region of integument type 1 between them (Fig. 18). I have argued that region 0 was the posterior margin of the intersegmental membrane (Lawrence, 1970), which maintained the lower gradient position; this resulted in flow and the lowering of the surrounding level 2 cells to level 1. Marcus (1962) showed that none

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

217

A A A A A A d I I anterior

5 4

................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... 3 ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... ................................................................... 2 ................................................................... ................................................................... ................................................................... ................................................................ .........-~- ...................... .Y ................................................................... .................................................... L

Crmh..-

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

h

Fig. 17. Diagram of the adult Galleria abdominal segment. The thick bar is a ridge of hard cuticle (region 4). The dotted areas (3, 2 and 1) represent regions bearing scales of three different sizes. The intersegmental membrane has a posterior margin (0) and an anterior margin (7). The arrow points anteriorally. (After Marcus, 1962; interpretation from Lawrence, 1970.)

of the type 1 cells originate from cells of the transplant (Fig. 18); they were all presumptively type 2 cells from the segmental surface and effectively therefore the posterior margin imposed its gradient position on the surrounding cells. This result confirms that the intersegmental membranes do have a special part to play in the maintenance of the gradient, for if their gradient position were as labile as the cells of the segment surface after transplantation, some of their cells would also have made type 1 structures. The actual formation of the adult integument was elicited by general metamorphosis of the host, but the nature of the product depended on interaction between the cells within the segment. Thus the determination of cells in the larval segment was only “provisional”. These elegant experiments (Marcus, 1962, 1963) showed that orientation of scales and the sequence of cuticular types in the segment had a common gradient basis, the level of the segmental

218

PETER A. LAWRENCE

0 Fig. 18. Transplantation of presumptive region 0 into presumptive region 2 in Galleria 2, the shaded nuclei mark scales of type 1. The host nuclei (presumptively type 2) are each marked with a central dot, the donor nuclei are undotted. Note that some host nuclei have become cells of type 1, and others have become nuclei of type 0 (small nuclei in centre). (From Marcus, 1962.) (see Fig. 17). The large peripheral nuclei mark scales of type

gradient being determinant of the type of integument formed. The proper development of isolated pieces of segment suggest that the gradient level can be autonomously maintained for some time, again pointing to active participation by the epidermal cells in the maintenance of the segmental gradient, rather than only the intersegmental membranes. Such experiments argue against mosaic development and demonstrate that the interaction of growing cells is responsible for the increasing complexity of the epidermis during growth. They further indicate two central points: one, that cellular polarity and pattern formation are causally interrelated and two, that pattern formation, at least in this example, can be analysed as two systems-a basic underlying gradient system and the response of the

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

219

cells to it. Both could vary independently. These two points will be illustrated further in the next example, and in the section on pattern formation. In Galleria pupae and adults there is a ridge of cuticle (region 4 in Fig. 17) which is present on segment 6 but not on segment 4. In a series of transplantation experiments Stumpf (1966, 1967b) showed that ridge cuticle developed wherever the appropriate level was generated by the process of gradient flow. Rotation of a square of cuticle with a slightly eccentric presumptive ridge gave the predicted pattern of an isolated ellipse and deflected line of ridge cuticle (Fig. 19). Here many cells which were not normally destined to make

_ - _ _ _ ............................................. - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _a.m. _ _ ___ _ _ _ .......................................... ________________-----------0 L

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

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

(c)

Fig. 19. (a) Rotation of a square piece of Galleria larval cuticle containing eccentrically placed presumptive ridge cuticle (thick line). (b) Situation immediately after rotation. (c) Predicted pattern after flow; the contour for level 6 is marked with a thick line. a.m. =anterior margin; p.m. =posterior margin. (From Stumpf, 1968.)

220

PETER A. LAWRENCE

ridge cuticle now did so. Stumpf (1968) then turned her attention to the difference between segment 6 and segment 4. She transplanted pieces of tissue between these two segments and the shape of the ridge, which only developed from sixth segment cells, showed that the segmental gradients were equivalent and that they interacted as before. The difference between the segments was totally due to varied cellular responses to identical gradient systems. F. INSECT SEGMENTAL GRADIENT: A SUMMARY OF THE WORKING HYPOTHESIS

It has been established that the scales and bristles of Galleria and Oncopeltus mark the direction of slope of a linear gradient, which is set up between the anterior and posterior margins of the two bordering intersegmental membranes and behaves as the sand gradient model. The direction of slope can be altered predictably by experiment. The maintenance of the gradient seems to depend on some active participation by the cells themselves between the limits set by the organizing margins of the intersegmental membrane. A chemical model which would fit the facts would consist of the following points: (i) that the gradient is of a diffusible substance; (ii) that one margin produce, and the other absorb this substance; (iii) that the substance is actively transported by the cells against the concentration gradient in which they find themselves; (iv) that the substance is labile or broken down by the epidermal cells. Some kind of gradient would be set up by points (i) and (ii) alone or point (iii) alone. However, the experiments of Piepho (1955b) and Marcus (1962) which show that both membranes are opposite in effect and that they act on the adjacent epidermal cells, and those of Locke (p. 20 1 ) which demonstrate that the orientation of the cells themselves is not totally dependent on the intersegmental membranes, together demand the bipartite hypothesis. Points (i), (ii) and (iii) still do not predict the result when a segment is bordered by two margins of identical effect, which occurs naturally (Fig. 8) and has been created experimentally (Fig. 7): a further premise (iv) is required. This working hypothesis is little more than a description of the experimental results, and must be regarded as very provisional. Indeed, Goodwin and Cohen (1969) have shown one way in which a

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

221

gradient with the required properties could be generated without a unique substance. G. GRADIENT PHENOMENA IN OTHER ORGANISMS: A COMPARISON

The insect offers a very clear advantage over other systems in the elucidation of the morphogenetic gradient: there are markers such as bristles, which point to the direction of the gradient in any small locality, and provide an accurate register of gradient changes after experiment. No other system possesses such local polarity indicators. Moreover, as we have seen, there is convincing evidence that at least in the insect segment, the gradient itself is the basis for qualitative pattern development, particular levels in the gradient being determinate of particular integumental types. While there is no compelling reason to believe that morphogenetic phenomena in different groups have a common basis, it is often instructive to see how far information on one group is in accord with evidence from other groups. In this case the experiments on the insect segment have produced an incomplete, but moderately demanding hypothesis; how far can this hypothesis be applied to these other groups where morphogenetic gradients have been mooted? We shall restrict ourselves chiefly to studies of Hydra and Tubuluriu although most gradient phenomena have also been found in planarians (Wolff, 1962; Lender, 1965; Hay, 1966). These hydroids possess the ability to regenerate new distal and proximal ends, and to reorganize the tissues between, after section. There is a suggestion that all cells are normally restrained from developing into distal structures by inhibitory influences emanating from extant distal structures (Rand et ul., 1926; Webster, 1966). Webster and Wolpert (1 966) have demonstrated that in addition to this inhibitory influence there is a gradient of hypostome forming propensity, and that the most anterior portion of the cut piece, now released from inhibition, will develop fastest to the state where it can induce a secondary axis when grafted to an intact hydra. A similar gradient of head forming propensity has been postulated for planarians (Brqhdsted, 1955). Rose (1957a) has argued the case for specific inhibitors; he believes that each organ of the regenerating hydroid produces an inhibitor which prevents other parts from developing into that structure. Under any circumstances the cells of Tubuluriu will transform into the most “efficient” state not already occupied. This A.I.P.4’

222

PETER A. LAWRENCE

hypothesis proposes that the sequence of structures along the disto-proximal axis is stored in the cells’ developmental repertoire as a series of states of different “efficiency” and that each of these differentiated structures, when formed, will inhibit the development of like structures. After removal of the hypostome, the most distal remnant tissue will transform into new hypostome, and the specific sequence of structures will spread proximally from the new hypostome until it includes the whole reconstituted animal. Rose (1 963) has moreover gathered an impressive body of evidence which speaks for the existence of specific inhibitory substances. Pieces of Tubularia will normally regenerate new distal tentacles, but homogenates of distal structures incorporated into agar could be made to inhibit this regeneration. The agar was inserted into a piece of hydranth stem and the preparation subjected to an electrophoretic current. In this situation distal regeneration was always inhibited if the distal end of the test piece faced the cathode, but not if it faced the anode. If clean agar, or agar impregnated with a homogenate of proximal material, was inserted into the hydranth hem, the current was without effect on regeneration. Quantitatively the results from these experiments are impressive. Further experiments (Rose, 1963) suggested that it was possible to influence the movement of inhibitory substances in situ. These observations justify the notion that an electric field can interfere with the disto-proximal passage of something which blocks regeneration of distal structures behind pre-existing distal structures. Specific inhibitory effects have also been demonstrated in nemerteans (Tucker, 1959), polychaetes (Smith, 1963) and planarians (Lender, 1956). So far the insect segmental gradient and the morphogenetic gradient of hydroids seem to have little in common. One phenomenon, which brings the two systems together, is that of the direction of transmission of inhibitory information. It appears that inhibitory information from an extant or nascent hypostome can only pass disto-proximally. If two distal portions of Tubularia were grafted together in the same polarity one was overcome by the other and they merged (Rose, 1957b) (Fig. 20(b)). But when two distal portions were grafted together in opposite polarity they did not inhibit each other and maintained their independent status (Fig. 20(c)). Rose argued that the inhibitory information could not pass up the stem proximodistally, so that the two distal ends were effectively insulated from each other.

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

1 0 9

8

a

7

6

10 910 9

b

8

7

223

6

C

Fig. 20. Experiments on Tubukzria. Above: appearance of hydrbid. Below: gradient situation. Transplanted tissue is shown in black, host tissue in white. (a) Control. (b) Grafted distal portion on to complete presumptive hydranth in same orientation. (c) Distal portion grafted in opposite orientation to host, which leads to development of twin-headed hydranth. Note how part of the host becomes changed in polarity. This change can be explained in terms of the kind of flow which happens in the sand model. (After Rose, 1957b.)

The hydroid hypothesis now stands as follows: (i) there is a gradient of propensity to transform into distal structures; (ii) each organ makes a specific inhibitory product which inhibits the development of the same organ elsewhere; (iii) there is polarized transmission of these inhibitory agents. This hypothesis offers no explanation for polarity reversal. When two pieces of hydroid of different length are grafted together some of the stem of one becomes reversed in polarity (Rose, 1957a, b) (Fig. 20(c)). The Rose hypothesis does in no way predict this reversal but the sand-model type of gradient does (Fig. 20(c)). Is it possible that a gradient of the sand-rnodel type is set up below the hypostome in hydroids? Are there any advantages to such a hypothesis? The neogenesis of a hypostome after section is not predicted by

2 24

PETER A. LAWRENCE

the sand-model hypothesis, but once one is formed, because of the plastic nature of the gradient, depending on active transport of a diffusible substance between the two new limits (basal disc and hypostome in hydra) a new gradient would be reconstituted. If two pieces of different gradient level are grafted together there should be “flow” from high to low, and over the area affected by flow a newly oriented gradient covering intermediate gradient levels would be established. After these gradient alterations the sequence of structures would develop because each cell responds appropriate to its level in the gradient. This hypothesis requires no specific inhibitory substances, but demands that the cells respond to the concentration of one gradient substance. In this respect it is simpler than Rose’s, but if one were to adopt the sand-model hypothesis for hydroids, and perhaps planarians, one needs to explain away the demonstrations of specific inhibitory substances (p. 222, Tardent, 1963). Of course, one can find difficulties: the substance ought to interfere with the growth of the producing organ itself, particularly as it can allegedly cause the regression of extant structures of the same type as the producing organ (Rose 1957b). One might argue that the substances discovered by Rose are internal growth regulators rather than morphogenetic messengers. One must conclude, unsatisfactorily, that a choice between these two incomplete hypotheses is premature now and probably will be until the basic mechanisms of gradient phenomena are uncovered. 111. PATTERN FORMATION

Embryonic development depends on the spatial organization of cellular differentiation. One apparently simple, and two-dimensional, example of this phenomenon is the generation of spaced bristle patterns in insects: there have been two main approaches to analysis of this system. One is a cytological study of the development of bristles in two genera of bugs, and the other utilizes Drosophila which are mosaic for tissues of two genotypes that have different effects on bristle pattern. These are both reviewed below. k THE DEVELOPMENT OF SPACED BRISTLES AND HAIRS IN RHODNIUS AND ONCOPELTUS

The abdominal cuticles of larval Rhodnius and Oncopeltus are covered with evenlyspaced bristles. Bristles are tactile sensilla, and

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

225

their constituent cells develop from single epidermal cells which transform into bristle mother cells and then undergo special differentiative divisions (Wigglesworth, 1953; Lawrence, 1966b). As a result of cell divisions during each moult cycle the number of epidermal cells intervening between extant bristles increases, and in the subsequent moult cycle some of these cells are transformed into bristle mother cells. In this way the bristle density is kept more or less constant from instar to instar. It has been experimentally demonstrated that the number of new bristles is related to the number of cells rather than to the distance intervening between extant bristles (Wigglesworth , 1940a; Lawrence, 1966b). Wigglesworth noted that new bristles were likely to appear in any large space which existed in the old bristle pattern, and that they never formed very close to existing bristles; it was clear that particularly situated cells were being selected as bristle mother cells. To explain his observations Wigglesworth ( 1940a) postulated that extant bristles effectively inhibited the appearance of new bristles nearby. During the last moult cycle in Oncopeltus, but not in Rhodnius, new structures develop, so that in addition to the bristles the adult abdominal sternites are covered with dense hairs. These hairs also develop as a result of unusual divisions by transformed epidermal cells, but unlike the bristles, lack innervation (Lawrence, 1966b, 1968). They form a completely integrated pattern with the bristles (Fig. 21). As these hairs are added to the pattern the density rises from about 15O/sq mm for the larval bristles alone to 2000/sq mm for the bristles and hairs; these hairs are so precisely spaced that the uniformity of distribution* of the pattern rises from R = 1.44 f 0.03 in fifth-stage larvae to R = 1.70 f 0.02 in adults. Analysis and understanding of this process requires detailed knowledge of cellular events during metamorphosis. We shall consider some of the relevant information: (i) Cellular metamorphosis: Juvenile hormone is present in the earlier instars and maintains the larval state; in its absence in the fifthstage larva the moult cycle is prolonged and metamorphosis occurs. During metamorphosis the epidermal cells change in many ways and synthesize a qualitatively different cuticle (Lawrence, 1969). If juvenile hormone is injected at the beginning of the fifth

* The uniformity of distribution is defined as a scalar quantity, R, which is equal to the product of twice the square root of the density and the average distance from each unit in the pattern to its nearest neighbour (Clark and Evans, 1954; Claxton, 1964). For a random distribution R = 1.00, and for perfect hexagonal packing R = 2.15.

226

PETER A. LAWRENCE

Fig. 21. Distribution of hairs (closed circles) and bristles (open circles) on the sternite of adult Oncopelrus. Note hairs and bristles are distributed in the same way. (~360.)

larval stage, metamorphosis is inhibited and no hairs develop (Lawrence, 1969). Two experimental results show that development of hairs does not result from the altered moult cycle but is a local phenomenon dependent on the state of the epidermal cells themselves; firstly if the hormone is applied topically to the insect, local patches of hairless larval cuticle are formed-even though the length of the moult cycle is not affected and the remainder of the insect is a perfect adult (Lawrence, 1969) and secondly, adult tissue transplanted back on to a larva will develop new hairs even during a fourth- t o fifth-stage moult of the host, in the continuous presence of the juvenile hormone (Lawrence, 1 9 6 6 ~ ) . (ii) The timing of determination: Uniformly sized fifth-stage larvae, which normally develop into adults of a certain size, will moult into very small adults if they are partially starved (Lawrence, 1966b). These small adults have approximately the normal epidermal cell and hair density and accordingly the number of hairs in the particular region studied may be as low as half the normal. As was earlier shown by Wigglesworth (1940a) for Rhodnius, the density of epidermal cells is homeostatically controlled and their number therefore related

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

227

to the amount the cuticle expands during each moult cycle. In the starved Oncopeltus the number of hairs has responded to the lower number of cells, and consequently hair determination must have occurred after at least the majority of cell divisions. We may note that in contrast bristle determination in larvae does not respond to starvation in any one moult cycle, and by the same argument must therefore have occurred prior to the majority of cell divisions in that moult (Lawrence, 1966b) (Fig. 22).

2nd

3rd

4th

/*/ *cell

divisions

hair determination

5th

starve

1

bristle determination

Fig. 22. The timing of bristle and hair determination. Starvation in the thud larval stage reduces cell divisions. Bristle determination responds to this lower cell number in the fourth larval stage, so that the reduced bristle number can be seen on the fifth instar. By contrast, starvation in the fifth larval stage, because hair determination follows cell divisions, directly reduces hair determination in that moult.

(iii) The process of hair spacing: Cell divisions in the fourth-stage larva effectively separate the bristles and uncover competent areas, which are now outside the bristles’ “inhibitory influence”. These areas are then populated with some new bristles during the determinative phase which begins before the majority of cell divisions in the fifth-stage larva. Cell divisions in the fifth-stage larva will again separate old and nascent bristles, and uncover areas free from inhibition. These divisions are followed by hair determination. So many hairs are formed that in addition to those sponsored by divisions in the fifth-stage larva, there must be a considerable reduction in the extent of inhibition by the bristles so as to increase massively the total competent area. This drop in the diameter of the inhibitory circle which surrounds each bristle could either occur suddenly prior to the initiation of hairs, or could be continuous with

228

PETER A. LAWRENCE

the process. If one knows the spatial order in which hairs are added one can distinguish between these possibilities; in the first instance large competent areas are defined and within them hairs with their inhibitory circles are introduced at random until the area is saturated, and in the second instance the initial hairs are confined to areas relatively distant from extant bristles, and as the inhibitory

eO

3

0

.

O

.O 0.

. e

0

0

0

0

0 0

0 0 ,0

‘ e

0

0

0

Fig. 23. Identical patterns of bristles with (a) large and (b) small inhibitory regions. The first new hairs irr (a) improve the uniformity of distribution, and in (b) spoil it.

circles shrink, the hairs are sited closer and closer to each other and to the bristles. The uniformity of distribution will be very different during the early phases of pattern development in the two cases (Fig. 23). With the aid of a computer we (Lawrence and Hayward, 1970) have been simulating hair pattern development of various models of this type. We have found that patterns which are almost identical to the natural hair pattern can be generated using the simple rules developed by Claxton (1964). The starting point is a typical larval bristle pattern, and each bristle is surrounded by an inhibitory field

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

229

within which new structures are forbidden. Points are generated at random, and added (with an appropriate inhibitory circle) if they fall within competent regions. During the process the computer periodically pauses and measures parameters of the pattern under construction. The process continues until no competent areas remain. We found that the adult pattern cannot be generated if at any one time the inhibitory circles are all exactly the same size (this gives a probability surface around the average hair or bristle shaped

I,

c %

.-.-

ti

r

l,

&O C.

Distance-

a

c

Distance

b

-.

Fig. 24. The probability of a new hair developing in relation to distance from an extant bristle. (a) With all inhibitory circles the same size and (b) with normal variation in the size of the inhibitory circles.

as in Fig 4(a) and a final distribution that is too uniform) and that some normally distributed variation of these circles is required (Fig. 24(b)). The simplest model supposes that the reduction in circle size, which is a prerequisite for adding a large population of hairs, occurs at the very beginning; the first and all subsequent hairs are therefore added in random order. When this is done the first hairs lower the uniformity of distribution, and it only recovers later (Lawrence, 1969). If however the reduction in circle size occurs later, or gradually during the process, then the first hairs formed will be a highly dispersed sample. There is direct evidence that this latter system obtains: when accurate drawings were made of the differentiating epidermis in the late fifth-stage larva the most advanced stages, that is the e a r k s t formed hairs, were so precisely dispersed that the uniformity of distribution for the fifth-stage bristles and these stages together, was higher than for the bristles alone. Further evidence that shrinkage of the circle size is occurring while the hairs are being added comes from insects in which the process has been curtailed, either by an inhibitor of DNA synthesis, hydroxyurea, or by means of the insect moulting hormone, ecdysone

230

PETER A. LAWRENCE

[which causes premature deposition of the cuticle (Williams, 1968)]. When injected late in the moult both these agents specifically curtail hair development, and do not inhibit cuticle formation or cellular metamorphosis. Measurements of the partially completed hair patterns of such insects are completely at variance with computer simulations in which the diameters of inhibitory circles are fixed at the outset, but much closer to simulations incorporating a gradual shrinkage of the inhibitory circle size (Lawrence and Hayward, 1970). (iv) Models to explain spatial inhibition: In order to simulate the development of hair and bristle patterns, we require only that extant and nascent bristles should effectively inhibit the transformation of nearby epidermal cells into new structures. The extent of this inhibitory effect is decreased during hair determination. We need t o know how this inhibition is achieved. As a model Wigglesworth ( 1940a) proposed that extant bristles continuously absorb some diffusible substance which is made by the epidermal cells. Only epidermal cells in an area of sufficiently high concentration of this substance can transform into bristle mother cells and as soon as the transformation has occurred in an individual cell, this cell begins to absorb the substance, reduces the concentration around it, and thereby precludes nearby determinations. The longer this process takes the greater would be the chance of two simultaneous determinations occurring in neighbouring cells. Such determinations almost never occur, and this makes literal application of the model a bit difficult. Another problem is that the inhibitory influence which is presumed to emanate from a hair or bristle must be maintained independent of the cytological state of the organ itself: old bristles, whose cells have secreted several previous cuticular structures; nascent bristles whose cells have not yet differentiated; new hairs in the various phases of cell division and hair mother cells only barely distinguishable from epidermal cells. All clearly have equivalent inhibitory effects which result in an even, integrated pattern. Each of these organs, in their divers states has to, according to the Wigglesworth model, absorb a substance at the same rate, and it is difficult to imagine how this could be done. The essential factor that they all have in common is the surrounding epidermal cells and this suggests to me that it is the epidermal cells rather than the bristle cells which are the prime mover. This hypothesis proposes that a small group of epidermal cells sponsor the appearance of a new organ in their centre and in this view the developmental state of the bristle

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

231

itself becomes irrelevant to the pattern forming process. This hypothesis is not without disadvantages, is more complex than the Wigglesworth model, but does bring the bristle patterns more obviously into line with other field systems in embryology, where a population of cells is clearly capable of subdividing itself into two or more equivalent independent field systems (e.g. “Homonomous arealization”, p. 244). This means also that they adopt the same mantle of mystery which surrounds embryological fields generally. There is no doubt that the epidermal cells surrounding bristles are sometimes different from other cells. In Rhodnius (Wigglesworth, 1933) these cells secrete a plaque of smooth cuticle; and as a result of delicate manipulation of hormone timing (Wigglesworth, 1940b) can be made to d o so even when the bristle itself has degenerated. In Drosophila some bristles are always accompanied by a small cuticular process from a neighbouring cell termed a bract (Peyer and Hadorn, 1966). After dissociation and reaggregation of leg disc cells the orientation and site of this bract is related to the orientation of the bristle itself (Garcia-Bellido, 1966a). These examples have been considered to result from induction by the bristle cells, but there is no evidence against their determination occurring synchronously with the bristle. From this viewpoint it becomes likely that in some mutants the epidermal cells could become organized into a local bristlesupporting field, but because of some genetic alteration their central cells might not be able to form a bristle. This realization separates the inducing from the responding systems, ideas that have been developed by Stern and his school, reviewed in the next section. B. GENETIC MOSAICS

Drosophila, homozygous for the recessive gene achaete (ac), lack two large bristles on the thorax. Flies were bred which carried ac and the gene yellow (y) [which alters the colour of the cuticle of each cell bearing it (Hannah, 1953)l on a normal X chromosome, and ac + y+, on a ring X chromosome. The ring X chromosome is eliminated with fairly high frequency during development to give clones of male cells which are hemizygous for ac y , and consequently the extent of the ac tissue is independently marked by yellow cuticle (Hannah, 1953). Analysis of such mosaics (Stern, 1954, 1968) showed that regardless of the amount of ac and ac/ac+ tissue the expression of ac was almost always autonomous to the yellow patch of tissue. When yellow

232

PETER A. LAWRENCE

cuticle covered the bristle site the bristle was missing; when wild-type cuticle covered it, the bristle was present (Fig. 25). There was no gross interaction between the two tissues; it was as if the only difference between the ac and the ac+ epidermis was the presence or absence of the bristles themselves, rather than in the pattern forming process. Since bristles are, as we have seen, particular epidermal cells selected because of their position in the whole, this lack of interaction was surprising until the dual nature of the pattern

Fig. 25. Some mosaic half-thoraces of Drosophila. Black = QC tissue, white = QC+ tissue. Wherever the QC+ tissue covers the site of the bristle, the bristle develops normally. (From Stern, 1954.)

forming process was perceived. In both ac and ac+ tissue, Stern argued, there are identical co-operative pattern forming systems (prepatterns) but in ac tissue the cell selected by the prepattern is unable to respond. The genes involved in the pattern forming process which locates the bristle, and those concerned in the development of the bristle itself, were different. These lines of thought converge with those expressed above when I suggested that it was the epidermal cells not the bristle cells which are the prime mover in pattern formation; and it should be possible, as apparently in achaete tissue, to have a bristle supporting field without a bristle. However the situation is more complicated, because occasionally when ac tissue only just covers a bristle site, a slightly displaced y + bristle will sometimes form in wild-type tissue near by. Here we have evidence that the prepattern does not define a bristle site absolutely, and that some regulation is possible (compare Schubiger’s experiments with Edgebristle, p. 254). How then in wild-type tissue is it ensured that only one bristle is formed? There is no answer as yet. Some mutants of Drosophila bear adventitious structures: male flies homozygous for the gene engrailed ( e n ) have a secondary sex-comb. Somatic cross-overs produced, in stocks of suitable genotype, mosaics of en/en and en+/en tissue (Tokunaga, 1961). The homozygous tissue was marked with yellow. It was found that regardless of the relative amount of en/en and en+/en epidermis that

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

233

were present, the development of a secondary sex-comb was always autonomous to the patch of yellow tissue. In some cases a little yellow patch behaved always as if it were part of a limb which was all mutant; the large areas of wild-type tissue likewise developed as if the limb were all wild-type. In other cases the limb would be mostly of en/en tissue and again the expression of pattern was quite autonomous to the patches of different genotype. Tokunaga concluded that in the wild-type leg there is normally an underlying facility to make a secondary sex-comb, and that en+ cells have lost the ability to respond to it. In fact nearly all the pattern genes tested by means of mosaics turned out to be alterations in the mutant cells’ reaction to an invariant prepattern (Arnheim, 1967). However, an example of a prepattern mutant has been recently described (Stem and Tokunaga, 1967). Male insects heterozygous for the mutant eyelessdominant bear extra rows of sexcomb bristles. Mosaics of eyeless-dominant and wild-type tissue behaved non-autonomously : patches of wild-type tissue in a primarily eyelessdominant background conformed with the general pattern of the eyelessdominant tissue and formed sexcomb teeth in the extra rows not found in exclusively wild-type insects. It was noticed that the eyelessdominant gene interferes with the proper segmentation of the leg, which confirms the importance of the segmental gradient, and the intersegmental membranes in the development of prepattern. We have seen how the orientation of hairs is an indicator of the underlying cellular polarity (p. 203). Mosaics of two alleles of aristaless (ul), which have differing effects on bristle orientation of the scutellum, have been studied by Tokunaga and Stern (1 969). They found that bristle orientation was not expressed autonomously, but was related to the proportion of mutant and wild-type tissue on that side of the scutellum. The posterior part of the scutellum grew less in mutant than in wild-type tissue, as did patches of mutant tissue in a mosaic. This growth, autonomously expressed by mutant tissue, controlled the orientation of the bristles near by, regardless of the genotype of the bristles themselves. These observations closely related the fundamental pattern forming system of the organ to the orientation of the bristles. Homeotic mutants cause the development of an inappropriately situated appendage ; one example is aristapedia where tarsal segments develop in place of the antenna1 arista. A suitable genotype will give flies mosaic for aristapedia (marked with yellow) and wild-type

234

PETER A. LAWRENCE

tissues, after somatic crossing over. Roberts (1 964) found that patches of mutant tissue were autonomously expressed as tarsal tissue in amongst wild-type aristal filaments. These observations imply that the underlying prepattern is equivalent in the two appendages and the essential variable, which has been so strikingly altered in the mutant tissue, is the competence to respond t o it. A similar point was beautifully illustrated much earlier by Bodenstein (1935) who took advantage of the very different structure of the fore- and hind-legs of adult Vanessa urticae; the fore-leg being exceptionally short and covered with long hair-like scales. The adult leg develops initially from a region in the second segment of the caterpillar limb and can first be cytologically distinguished at the beginning of the fifth and last larval stage. When different amounts of forelimb were transplanted on to part of the hindlimb in the third-stage larva, harmoniously formed limbs resulted, each with only one femur, tibia and the normal number of tarsi. The limbs were chimaeric and made up of different amounts of tissue of the two limb types: when second, third and fourth segments of the forelimb were transplanted on t o the first segment of the hindlimb the adult leg was almost pure donor, but when only segments 3 and 4 were transplanted the adult limb was about half and half fore- and hindlimb. The contribution of the transplanted limb to the adult structure depended also on the amount of material left as the stump on the donor. If segments 4 and 3 of the forelimb were transplanted on to segments 1 and 2 of the hindlimb the grafted tissue only formed tarsal material, but if the same tissue was transplanted on to segment 1 of the hindlimb, it formed both tarsi and some of the tibia. Clearly segmentation of the limbs was not established at the time of the operation. But the chimaeric nature of the limb showed that determination to limb type had been completed, although nevertheless these differently determined structures were responsive to one organizing system. Again, Kroeger (1959, 1960) tied fore- and hind-wing discs Ephestiu together and implanted the combinate into a mature caterpillar These two alien discs combined t o form single integrated wing hinges in which, although the characteristic type of structure was formed by each tissue, the components fitted together neatly. Both different discs clearly conformed to a single combining pattern influence, but the precise way they reacted depended on the established identity of the disc cells themselves. These studies build up a picture of the pattern forming system as a

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

235

series of independent gradient systems which determine the shape and size of an organ. In different organs such systems are much more similar in construction than are the final organs themselves-the difference being primarily due to myriad variety in the competence of the cells. It is possible that these cells respond to a level in the gradient system, as occurs in the segmental gradient of Galleria (p. 218), and that gradient systems are based on the same mechanisms, wherever morphogenesis is to be found. Similar ideas are also being developed by Wolpert (1 970). Differences between competences of cells are, therefore, mostly responsible for the structural heterogeneity. This pattern of responses by the cells to a gradient system depends on the earlier acquisition by those cells of an identity (that they are determined as fore-wing cells, hind-wing cells and so on); this identity-the determined state-will be considered in the next section. IV. DETERMINATION AND REGULATION A. INTRODUCTION

Although “Determination is among those indispensable terms in the vocabulary of embryologists and geneticists which are most difficult to define” (Hadorn, 1965), it is obviously important to know when, in a growing and diversifying population of cells, the developmental identity of an individual cell becomes established. Determination is this choice of a developmental route, and Waddington (1956) argues that when “determined” the cell is irrevocably committed to develop along that pathway. However, as he points out, what is irrevocable may depend on experimental circumstances: Marcus has shown (p. 216) that isolated pieces of the Galleria larval segment will develop according to their prospective fate, and most would argue that this shows that the cells are determined; and yet if placed adjacent to cells of different gradient position, they develop differently. In that case, moreover, determination to form a particular type of scale is provisionally a property of a local population of cells. Still later individual cells from that population are further determined as scale mother cells, leaving the others to make cuticle (Lawrence, 1970). In this case determination is progressive; as growth proceeds diversification continues, to result in a narrowing developmental repertoire of the cells. If the development of an embryo is completely and irrevocably

236

PETER A. LAWRENCE

mapped out, parts dissected from it would always develop, as far as conditions allow, towards their prospective fate. If small sections of an embryo were extirpated the embryo would mature with these parts wanting. Often, however, there is plasticity in the developmental process and isolated pieces, or incomplete embryos, overreach their prospective fate so that a whole embryo may be constructed from some of its parts. This process is termed regulation. In theory, regulation could only occur in the absence of determination, but in fact these two cellular properties are not as mutually exclusive as abstract definitions demand. We shall discuss these matters in connection with experiments on imaginal discs. Many of the properties of the determined state have been elucidated by a brilliant series of experiments on imaginal discs of Drosophilu: “one of the great stories in modern biology” (Williams, see Hadorn, 1965). Diptera are almost completely remade in the pupal stage: this is not done by “demolition and reconstruction” but by “progressive substitution” (Wigglesworth, 1965) as many of the larval tissues break down and the adult develops from small clusters of cells which have been sequestered during larval growth. These clusters of cells, termed imaginal discs, appear in the egg and grow during larval life. At metamorphosis the cells of the disc differentiate and form adult organs, including much of the integument. Each disc has independent generative powers, for if removed from a fully-grown larva and implanted into the abdomen of another, it will develop almost completely normally, although many structures fail to evaginate (Bodenstein, 1941 ; Vogt, 1946). Younger and smaller discs will attempt to make their own structures if forced to metamorphose prematurely by implanting them into a fully grown larva (Bodenstein, 1939). These experiments show that the discs are determined as a whole to make their appropriate organ but they d o not tell how precise determination is at the cell level. Vogt (1944, 1946) investigated the determined state of the eye-antennal disc of Drosophilu. She transplanted bisected discs into late third-stage larvae, and found, unlike Bodenstein (1 94 l), that the development of half discs quantitatively overreached their prospective fate and, taken together, the two half discs made more than the normal number of ocelli and aristae. Hadorn and Gloor (1946) investigated the determined state of parts of the female genital disc of Drosophila. By cutting the mature larval disc into several pieces and transplanting those pieces directly into another full grown larva, they found that different parts of the

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

237

disc developed independently, into organs of the genital system and associated structures, according to a pattern. They concluded that the disc was divided into areas which were committed t o form the different organs of the genital system and constructed a determination map of the disc. In addition they noted that the disc pieces often made complete supernumerary structures. As each portion of the cut region reacted to the interference by regenerating a complete structure, this process was termed regulation. It seemed that the determined regions of the disc possessed independent powers of regeneration, and Hadorn and Gloor (1946) therefore proposed that the disc was a mosaic of separate field systems or anlagen. In an examination of this disc, Hadorn and Chen (1956) studied regulation of the spermatheca and found that regulation occurred most often when the presumptive spermathecal region was bisected. But if only small portions of this area were isolated by the cut they had a reduced chance of forming a spermatheca. These studies raised two interrelated questions: first, what is the mechanism of regulation? Second, how precisely determined are the disc cells prior to metamorphosis? The f m t question has been investigated primarily by studies of the discs’ performance after section and different periods of time in culture, and the second by dissociation and reaggregation of disc cells. I shall review these approaches separately. B. EXPERIMENTS WITH CUT DISCS-REGULATION

The male genital disc was examined in great detail by Hadorn, Bertani and Gallera ( 1949) and Ursprung ( 1957, 1959). This disc will develop into structures illustrated in Fig. 26 if implanted into a mature larva. Some structures are paired (anal and lateral plates, claspers, paragonia, vasa) others unpaired (ductus, sperm pump and penis plate). If the male disc was cut into two sagittal halves (Fig. 27) and both pieces implanted into a full-grown larva, regulation usually occurred only of the unpaired parts (Fig. 28); so that each half disc developed into a genital system with complete ductus, sperm pump and penis plate, but only one of each of the paired structures. Sometimes, however, the paired structures regulated so that each part of the disc made two of these organs. Hadorn et al. (1949) noted that regulation was more likely in younger hosts, and Ursprung (1959) refined these delicate methods still further and found that variability in the results could be reduced if the age of the host was carefully controlIed. He moreover found that a host of only 55 hr

238

PETER A. LAWRENCE

Fig. 25. The male genital system of Drosophifa. P = paragonia, V = vasa, D = ductus, L = lateral plate, A = anal plate, M = claspers, S = sperm pump, T = penis plate. (From Hadorn el al., 1949.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

239

old was capable of supporting complete regulation by both halves of the disc. Even half a disc of a mature (96 hr) larva was capable of regulation to a complete male genital system when implanted into a 55-hr-old larva. It seemed that regulation was dependent not on the age of the disc itself but on the time available prior to metamorphosis, and this was confirmed later (Ursprung, 1962). Hadorn et al. (1949) noted also that when the disc was cut paramedially (Fig. 27) the unpaired

Fig. 27. The genital disc showing approximate position of the anlagen in relation to a sagittal cut (thick bar) and paramedial cuts (dotted lines). P=paragonia, V=vasa, C = claspers, A = anal plates, S = sperm pump, T = penis plate. (From Hadorn et ol., 1949.)

central structures usually did not duplicate while the paired ones did. This drew further attention to the observation of Hadorn and Chen (1956) that cut areas regulate particularly well. UV damage does not act in the same way as a cut, for Ursprung (1957, 1959) found that local irradiation often deleted parts of the genital system without inducing regulation. Hadorn et al. (1949) had noted that cutting induced growth and proposed that “cell proliferation and regulation go hand in hand”. The size of one of the paired plates can be gauged by the number of bristles it bears, and because Hadorn et al. (1949) found no difference in the mean bristle number of anal plates or claspers, whether two were formed by each of the three pieces resulting from paramedial cuts, or whether there were only two plates in all, they argued that regulation is an all-or-none phenomenon. However, soon Hadorn (1953) reported cases of partial regulation in D. skgguyi. The anal plates of this species bear four different kinds of bristles, and if disc transplantation was performed some time before pupation, two anal plates of different size were found in the

240

PETER A. LAWRENCE

Fig. 28. The genital disc of Drosophila from a larva of 96 hr is bisected and the half discs implanted into another larva of 96 hr. After metamorphosis of the host each half has a complete set of unpaired structures, but no duplication of paired structures. A = anal plate, D = ductus, Da = hindgut, L = lateral plate, La = lappetts of sperm pump, M = claspers, P = paragonia. PB = peripheral bristles, Sk = sperm canal, S p = sperm pump, V = vasa. (From Ursprung, 1959.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

241

adult, one was normal in size and shape and the other was very much smaller (Fig. 29). Hadorn argued that the small disc was regenerating at. the time when metamorphosis of the host commanded the cells to express their current status. Other very carefully timed experiments caught “regulation actually under way” (Ursprung, 1959). If a sagittal half of a disc were implanted into a third-stage larva of 72 hr,

Fig. 29. Section and implantationof the genital disc of D. skguyi. After metamorphosis the anal plates show incomplete duplication. Note that all four types of bristles are present on each plate, although their numbers may be much reduced. (From Hadorn, 1953.)

each half gave rise to a complete set of unpaired structures. However, the anal plates and claspers were asymmetric; one of each pair on the same side bore the normal number of bristles, whereas on the other side (the cut side) the anal plate and clasper bore a reduced number of bristles (Fig. 30). This allows an insight into the mechanism of regulation: as

242

PETER A. LAWRENCE

Fig. 30. Compare with Fig. 28. Transplantation of bisected male genital disc of Drosophilu performed at 72 hr with host also aged 72 hr. Note partial duplication of the lateral plates, anal plates and claspers on the cut side (arrow). The paragonia, vasa and ductus remain unduplicated. For abbreviations see Fig. 28. (From Ursprung, 1959.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

243

pointed out by Ursprung, regulation could either result from growth of the field of determined cells, followed by its subdivision into two equal parts at metamorphosis (this would give rise always to two equally sized organs, even if they might be abnormally small), or by growth of an independent field system from the wound blastema caused by the cut. The original half of the field would develop normally. Ursprung’s observations support the second hypothesis. In a further analysis of this question Ursprung (1962) made use of Bodenstein’s discovery (Bodenstein, 1943) that discs could be cultured in vivo in the abdominal cavity of adult flies. He removed the implanted half of the disc after various periods of time in culture, and at seven days illustrate a small blastema which could be clearly seen growing out from the half disc. Similarly Kroeger (1958) had observed that during duplication of an implanted wing disc of Ephestiu, mitosis was mainly restricted to the outgrowing blastema. Luond ( 1961) transplanted male genital discs of DrosophiZu siguyi taken from fully mature larvae into hosts of different age. In mature hosts, as before, only the unpaired structures regulated, each half disc only formed one anal plate and one clasper although occasionally (6%) two such structures were formed by regulation. The percentage of regulation, as in earlier studies, was increased when younger hosts were employed. Luond demonstrated that there is hierarchy in regulative ability; certain structures invariably regulated before others, and this suggested that the regulative process proceeded continuously in the cut disc, at differential rates, until metamorphosis. During initial regulation of the anal plate the proportion of bristles of the various types is different from the final proportion (a feature also noted by Ursprung, 1959) but the proper proportions are reestablished by the time the bristle number is half the normal. In these very small anal plates it is the medial teeth-shaped bristles which are relatively in excess, and Luond argues that this implies a mediolateral outgrowth of the anal plate primordium. Moreover, intermediate bristles, with sockets like the teeth-shaped bristles and more typical shafts, were found between regions bearing the pure bristle types. Since the cuticular parts of bristles are formed by only two cells, which have certainly descended from one mother cell (Lees and Waddington, 1942; Lawrence, 1966b) this points to the plastic nature of this morphogenesis. Luond postulated a mediolateral gradient of something whose concentration in any one place determines the type of bristle that will be formed there.

244

PETER A. LAWRENCE

Unlike the genital disc the eye-antenna disc is mapped out asymmetrically so that section separates two qualitatively different pieces (Gehring, 1966). For instance, when the cut is made at a particular level and the pieces implanted into a mature larva the ability to make palpus is almost exclusively restricted t o one half. We know that half of a genital disc, given time, can regulate to form a complete genital system, but since the disc and the genital system itself is symmetrical this only implies duplication of parts already present. Gehring asked if during regulative growth the half discs could surpass their prospective fate and make quaZitatively different structures. The anterior half of the antenna1 part of the eye disc invariably made palpus when implanted into a mature larva, and the posterior half never made palpus. Gehring therefore implanted these halves into young (72-80 hr) hosts t o allow time for growth and regulation. After metamorphosis of all 30 implants of posterior half, the palpus was lacking. In the anterior halves palpi were formed and often duplication occurred. After more prolonged culture in adult hosts, followed by retransplantation into larvae, duplication of the palpus in the anterior half was more frequent and in 3/23 cases palpus was formed by the posterior half. Gehring then isolated the most posterior third of four discs, fused them together and implanted them into adults as before. Of 12 combined implants of this type, six formed palps, the other six were completely free of palp. Gehring argued that since even after extensive growth of these combined implants half were without palp, this must imply that basically the posterior third of the disc contains no prospective palp cells, and that its appearance in the remainder must have been spontaneous (regenerative Neubildung). Such a change of determined state from one structure t o another normally found in the same disc is called regional specific transdetermination (p. 25 5 ) . Gehring then turned his attention to the duplication of the palpus. Because the bristle numbers are always equal on each side, the duplication of palpus seemed to occur by growth of the palpus anlage and its subsequent subdivision into two equal parts: he called this “homonomous arealization”. This is a different method to that described for the genital disc by Ursprung ( 1962) and Luond ( 196 1). The leg disc is also asymmetric and the results of similar studies to Gehring’s allowed Nothiger and Schubiger ( 1966) to conclude that regulation of discs depended on proliferation of existing cells, which handed down the determined state to their daughters. It did not

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

245

seem possible for cells belonging to one anlage to reliably replace missing parts normally made by cells from another. In order to test this hypothesis further Nothiger and Schubiger made the genital disc asymmetric by irradiating one half of the disc in approximately the region of the claspers. The irradiated disc was then cut into an irradiated and a non-irradiated half and these were implanted into young larvae ( 5 5 hr), or for a time in an adult, to allow regulation to proceed.

Fig. 31. Diagram to illustrate proliferative regulation. W irradiation destroys the clasper anlage (C) on the right side of the disc. On this implanted half, even though other organs duplicate, the claspers are not replaced. On the left half all organs duplicate. G = hindgut, Pe = penis. For other abbreviations see Fig. 28. (From Nothiger and Schubiger, 1966.) A.I.P.4

246

PETER A. LAWRENCE

The non-irradiated half often regulated completely and almost invariably contained at least one of the paired structures; however, about 60% of the irradiated halves developed into genital systems which completely lacked one structure, even though other structures were duplicated. Here even after extensive growth, a particular anlage could not be replaced by other parts (Fig. 31). At first sight this study contradicts Gehring’s observations on transdetermination to palpus by cells of the posterior third of the eye-antenna1 disc. This occurred in only one half of Gehring’s compound implants, however, and such changes if they occurred only rarely in the irradiated disc, could not be individually detected. These two studies allow the conclusion that reliable replacement of completely lost parts does not occur and further illustrates the mosaic nature of the disc. Regulation of this kind, which is not a complete restructuring of the available material into an entire system, but involves cell division and results only in a duplication of parts already present, was termed proliferative regulation by Nothiger and Schubiger. Proliferative regulation could be achieved by: (i) cell division in the different determined anlagen of the disc followed by migration to the new sites; (ii) migration of determined cells from the different anlagen to the cut and their subsequent division and spatial reorganization; (iii) a combination of (i) and (ii). The experiments of Nothiger and Schubiger do seem to rule out the formation of a “dedifferentiated” blastema, which develops into a new half-genital system autonomously. Some further observations relevant to this problem have come from studies on the proboscis disc. The proboscis of Drosophila is constructed from two equal halves, each made by one separate labial disc. When a single disc is implanted into a 72-hr-old larva it may produce an entire proboscis (Wildermuth and Hadorn, 1965) and the frequency of such duplication can be increased by culturing the disc for several days in the adult (Wildermuth, 1968a). It would seem that the disc in situ is normally under some developmental control, and indeed some supernumerary bristles are formed by a single disc even when it is implanted into a larva of the same age as the donor. The development of these extra bristles could result from the shock of the operation itself [ Loosli (1969) has suggested that the formation of similar adventitious bristles in the implanted haltere disc is due to such a shock], or be due to the absence of contact with other tissues

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

247

(Gehring has demonstrated that palpus duplication in the eye-antenna disc is inhibited when the eye part of the disc is in cwtact with the palpus area: in this case, however, there may be less wounding as the cut is located far from the palpus). It is noteworthy that duplication of the proboscis disc can occur without section [although the operation itself would provide some wounding stimulus (Nothiger, personal communication)] ,and from a study of partially duplicated discs Wildermuth concluded that duplication resulted from growth and the subsequent reorganization of the cellular material into two parts. He regarded the process as an example of proliferative regulation. This hypothesis demands that the cells of the new structure should descend from the different determined regions of the regulating disc and not from a local blastema. Kroeger (1958) reported that in duplication of the wing disc of Ephestiu, mitoses were concentrated along the outgrowing new wing. However an autoradiographic study of the duplicating labial disc (Wildermuth, 1968b) did not show such a concentration. Because of the short life of the thymidine in the fly, a pulse of label was taken up by all of the cells undergoing DNA synthesis within only about 1 hr of injection. Labelled cells were found evenly distributed over the disc when uptake of thymidine was assayed immediately after implantation into an adult. In a disc labelled at the beginning of regulation and left to go through metamorphosis prior to fixation, the label was localized only on the old side of the now complete proboscis. This result could well be due to dilution of the label on the new side by repeated divisions, as has been shown to occur during the larval-pupal moult of silkmoths (Krishnakumaran e t ul. !967), but there are other possible explanations. Even in the case of the genital disc, where regulation of a half occurs as a result of growth of a new blastema from the cut, the source of new cells is not established. Choice between the basic models for proliferative regulation is not yet possible, but the autoradiographic method pioneered by Wildermuth should, if the whole process of regulation is followed in detail, give the answer. C. DISSOCIATION EXPERIMENTS

The studies on regulation have suggested that the imaginal disc is a mosaic of independent populations of cells determined to make particular structures. During superfluous or regulatory growth, the cells pass on their determined state to their offspring. We do not yet

248

PETER A. LAWRENCE

know the precision of cellular commitment within each anlage of the mature larval disc. It is conceivable that each cell is rigidly determined to make each component, say a particular bristle, of the pattern, and that pattern reconstruction after section and regulative growth, or after excessive growth leading to duplication, occurs by the ordered assortment of daughter cells. Alternatively, the determination of cells could only be a commitment to form a particular region of the disc such as clasper, and deployment of the available clasper material could depend on a supercellular organization process occurring at the onset of metamorphosis itself. Hadorn e t a1. (1959) began to investigate this matter by means of genetically marked discs that were partially dissociated and then mixed up together. They mixed wing discs marked with yellow 0.1) with others marked with ebony (e). These chimaeric discs were then implanted, allowed to pass through metamorphosis, and the cuticle formed by the combinates examined. Many areas were genetically mosaic, and they found that the cells from both discs combined to make integrated patterns (Fig. 32). The authors argued that at the

Fig 32. Rows of bristles like those normally found along the wingedge but containing cells of two genotypes. Note that the pattern includes bristles regardless of genotype. Arrows mark sockets without shafts. (From Hadorn et ul., 1959.)

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

249

time of dissociation the pattern was “not anchored in single cells’’ and that some pattern forming process integrated their differentiation. Sometimes the patterns were imperfect; Ursprung and Hadorn (1962) found lines of partially ordered bristles which continued across genetically different areas. These authors argued that such a pattern, which did not have all the components of the pattern found in situ, and yet clearly indicated an integrating influence that extended across the genetic boundaries, could only have come from epigenetic development of pattern, which had not been completed in time for metamorphosis. Nothiger (1964) extended these studies and mixed pieces of genetically marked discs from both sexes, two species and from different discs. He found that partially dissociated discs from genetically marked males readily formed chimaeric but mostly normal structures, and that sometimes even single bristles from one donor could be found isolated in a homologous structure: such bristles were integrated into the entire pattern (Fig. 33). Male

Fig. 33. A male anal plate containing one yellow bristle in an ebony background. (From Nothiger, 1964 .)

250

PETER A. LAWRENCE

Fig. 34. Autonomous development of long female bristles in a predominantly male anal plate. (From Nothiger, 1964.)

genetical discs from two species of Drosophilu (0. melunoguster and D.skguyil also formed chimaeric structures. Here the cells from each donor could only form organs characteristic for that species, although cells of both species combined to form integrated patterns. Likewise, male and female cells also combined to form integrated structures, but only in the anal plates which occur in both sexes (Fig. 34). That mosaics of claspers or vaginal plates were not formed

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

251

suggested that these cells separated from each other. Indeed Nothiger discovered that wing and genital disc cells segregated completely. The sorting of disc cells has moreover been observed in vitro (Lesseps, 1965) and can be accomplished without growth (Garcia-Bellido, 1967). These experiments illustrated that, at the time of the operation, the disc cells were already determined to form particular structures, and while cells which were determined identically associated, those determined differently segregated.

Fig. 35. An aberrant pattern: a single clasper tooth, in a lateral plate. (From Nothiger, 1964.)

Occasionally some isolated cells belonging to, say, clasper, were found inappropriately placed, in a lateral plate (Fig. 35): this result is of importance because it shows that such errant cells are not sensitive to the nature of the surrounding cells and develop autonomously. Tobler (1966) did some rather similar experiments on the leg disc and in particular partially dissociated and combined the central part

252

PETER

A.

LAWRENCE

of one leg disc (corresponding to tarsus and tibia) with the periphery (corresponding to coxa, femur and tarsus) of another differently marked disc. If alterations of prospective fate were to occur within the leg disc one would expect all regions of the legs to be chimaeric, but if determination was already so established that different regions of the disc sorted by migration one would only expect to find mosiac patterns in those regions actually near the cut where tissues of one type are likely to be donated from both halves. Tobler found that 64% of the implants were mosaic for tarsus, whereas none were mosaic for femur, which confirmed that cells sort out from the different regions of the leg disc, and do not change their determined status. Garcia-Bellido ( 1966a) refined these methods further, and succeeded in dissociating the disc t o about 90% single cells with the remainder as small clumps consisting of 2-20 cells. These cells were then mixed and reassociated by centrifugation, and were then implanted into an adult for two days because “reconstruction of integrated structures proved to be more efficient” before final transplantation into a larva. Using the wing disc (Hadorn and Buck, 1962) he confirmed that cells from different regions of the disc would only form chimaeras with cells from the same regions and not from others. Generally each part of the wing disc only formed structures appropriate to their origin, but there was one case where cells of the anterior part of the disc (marked with y ) combined with those from a posterior half (marked with e) to form a row of integrated bristles normally found along the wingedge. At the overlap two bristles of the posterior type were formed by anterior cells. Garcia-Bellido proposed that this resulted from growth and “restitutive regeneration” of the anterior cells. Restitutive regeneration must clearly mean, in this case, a change in the prospective fate of the cells, possibly dependent on the interaction between cell types at the junction. The patterns formed by completely dissociated leg disc cells were not usually as finely integrated as those formed by partially dissociated discs. Often single sex-comb teeth or claw elements would appear in an alien environment. Garcia-Bellido argued that this implies that their formative cells were precisely determined prior to dissociation, and that the basis for reconstruction of pattern in these mixed cells is cellular migration. Moreover, since occasionally bristles in a row were quite inappropriately oriented to the others, he argued that cellular

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

253

polarity was also a persistent feature of these cells, and was not epigenetically determined after reaggregation. These hypotheses would demand that the cells not only normally migrate to their precise site in a developing pattern but also rotate until their polarity is appropriate to the other cells. He also combined fore- and middle-leg discs which did not sort out from each other. In the male fore-leg there are neat transverse rows of closely packed bristles that are lacking in the middle-leg. He found that only fore-leg cells made transverse rows; they were never chimaeric. Garcia-Bellido (1 966a) argues that this observation implies that “the determined cells carry not only information which enables them to join in regions, but also more specific information which enables them to reconstruct the patterns they come from”. Garcia-Bellido’s argument for highly specific determination of single cells is based on the differentiation of errant cells in the abnormal patterns he gets after complete dissociation. If a sex-comb tooth is formed in isolation, this, he argues, implies that the sex-comb mother cell was determined as such prior to dissociation. But even if the determination were less precise, what would one predict an isolated cell determined as, say, “tarsus” would do when commanded to differentiate and metamorphose? It must make something, and as such a cell could make anything on the tarsus there seems no reason why it should not spontaneously develop into a sex-comb tooth. In my opinion the reconstruction of pattern reported previously is not likely to result from the migration of each cell to its precise place in the developing pattern, and that at the time of dissociation, determination had not reached down to the single cell. This is clearly so for the cells which will make the little bracts which form nearby some kinds of leg bristles and are never found in isolation. After dissociation and reaggregation it is common to find a bristle of one genotype whereas its associated bract is of another. This conclusion is underwritten by the incisive observation of Nothiger (1964) who considered the earlier work on regulation in the genital disc by Hadorn et al. (1 949) (p. 237). On each of the claspers there is one characteristic long bristle. These authors cut the disc into three sections and sometimes each section made two claspers, each with one such bristle. The cells which prospectively were to make only one pair of bristles cannot have been located in all three pieces and therefore in this case at least two bristles must have been made by cells not predetermined to make them. Furthermore Schubiger (1968) examined development of the leg and constructed a fate map ALP. -9.

254

PETER A . LAWRENCE

of exceptional detail. On the trochanter in particular there is invariably one large “edgebristle” of characteristic form. The effect of cutting the disc into four pieces and reimplanting the pieces into a mature larva was to reduce the appearance of this edgebristle to only 60% of the discs. This seemed to be due to damaging the presumptive bristle cells, for if the cut was oriented even closer to the presumptive bristle the frequency of appearance dropped to 15%. When portions of the disc were reared in young larvae or kept for some days in an adult the frequency of appearance of the edgebristle increased and sometimes two were formed, one in each half: even in the cut portions there must be cells which can recover the ability to make edgebristle if they have time. Halves of younger discs are more likely to make two edgebristles than fragments of older discs; this implies that the area competent t o form edgebristle contracts with the age of the disc. These experiments make clear that the capacity to form an edgebristle is not restricted to a single cell until the process of metamorphosis itself; and may not be so restricted even then. In Oncopeltus, even after metamorphosis and development of hairs, the remnant epidermal cells have not lost their ability to make hairs, and under experimental conditions will do so again (Lawrence, 1966c). D. CHANGES IN THE DETERMINED STATE

Hadorn (1963) investigated the ability of the imaginal disc cell to maintain the same determined state through many cell generations. A half disc was cultured in an adult female abdomen, and after about two weeks growth was removed, bisected and one piece was implanted into another adult, while the other was implanted into mature larva to see what structures it could make during metamorphosis. After culture in several successive adults the proliferative rate of the cells increased rapidly and from then on several implants could be made, and new sublines initiated. It is a consequence of the continued section of the growing disc that later test implants descend from smaller and smaller regions of the original disc. If determination in the original half disc were to the single cell level, one would expect eventually to have cultures producing only one cell type. There is in fact some narrowing of the repertoire of those sublines producing genital structures over many transfer generations (Hadorn, 1963, 1965, 1966), which confirms that the genital disc is subdivided into districts of different determination. However this specialization does not continue indefinitely, and even

POLARITY AND PATTERNS IN DEVELOPMENT O F INSECTS

255

after many generations the narrowest of sublines still produced, for instance, anal plates (with characteristic bristles and cuticle) and hindgut. Hadorn (1966) noted that the test implants from lines bearing anal plates and hindgut nearly always contained either both organs, or only anal plates, but very rarely (6/966) hindgut alone. This and studies of the lineage suggested that the capacity to make hindgut was often lost from the line, but was repeatedly being regenerated from presumptive anal plate cells. Such a “regulative” change is reminiscent of that reported by Gehring for the neogenesis of palpus (p. 244)and is termed regional specific transdetermination. During the first five transfer generations genital disc cells gave only parts of the genital system in test implants, but later perfectly formed structures appropriate to other discs made a sudden appearance (Hadorn, 1963). Use of differently marked hosts and implants (Gehring, 1966) showed that these cells had not come from the host and resulted instead from a change in the determined state of the genital disc cells. The relevance of such transdetermination to normal development is not yet clear but there are certain features which are of particular interest. We may note: (i) the change to the new state is abrupt and complete, and intermediate cell types are not formed, or at least not recognized ; (ii) the cell affinities of the transdetermined cells become appropriate to their new identity so that sorting occurs (Garcia-Bellido, 1966b); (iii) the change is probably to a wide repertoire such as “leg” or “wing” and not just to “trochanter” or “wing-fringe”. Gehring (1 966) has shown that in one case of transdetermination from palpus to antenna, at first an antenna1 third segment is made, which later develops into an arista and a second segment. Hadorn (1966) examined a number of test implants in his genital disc lines in which transdetermination had only recently occurred to antenna. These showed no predominant bias to be one particular segment of the antenna; (iv) the probability of transdetermination occurring is directly related to the proliferative rate (Tobler, 1966; Wildermuth, 1968a); (v) some transdeterminations occur with greater frequency than others, and transdetermination may be both reversible and irreversible (Hadorn, 1966);

256

PETER A . LAWRENCE

(vi) transdetermination is a communal act by a small group of cells. This key fact was shown by irradiating larvae of the genotype y+sn+/ysn, mwh'lmwh with X-rays so as to induce somatic cross-over and mutations (Gehring, 1967, 1968). Such changes happen only with a low probability in a single cell and from this a clone of cells marked with, say, yellow and singed integument will grow. Gehering found, as was to be expected, transdetermined elements within a clone, but he also found regions of transdetermined tissue

Fig. 36. Gehring's experiments on transdetermination. Note that the two tissue patches (wing and occiput) and the patches of two genotypes (rnwh'sn' and mwh sn) are spatially unrelated. (After Gehring, 1967.)

that were genetically mosaic. This could imply either that transdetermination occurred spontaneously in a group of adjacent cells; or that transdetermination occurred independently in regions of two genotypes and that migration and sorting had brought these two groups of cells together. A detailed examination of such a mosaic showed, however, that both the genetic boundary and the tissue boundary were continuous and definite, and independent from each other (Fig. 36). Yet, if the area of wing tissue had been formed by two populations of cells migrating together there would have been a random assortment of the two genotypes in the wing tissue. The observed pattern could best be explained if transdetermination was a co-operative event and happened synchronously in two or more cells. As transdetermination is a co-operative event, could the normal determined state also depend on cellular interaction?Attempts to

POLARITY AND PATTERNS IN DEVELOPMENT O F INSECTS

257

start imaginal disc cell lines with single cells have not been successful (Gateff, personal communication), but in vertebrate cell lines the determined state can be transmitted via a single cell (Coon, 1966). In summary, these experiments on imaginal discs have confirmed Hadorn and Gloor’s (1 946) original description of the mature larval disc as being divided into fields or anlagen. Each anlage contains cells that are determined for a particular organ, and through growth can compensate for loss of part of the anlage. Occasional spontaneous changes of the determined state do occur, but normally the cells of one anlage cannot replace lost cells from another. Mixed cells from different anlagen will normally sort themselves out, but if a few cells are trapped in an alien environment they will retain their determined state and develop autonomously. In my opinion the formation of integrated patterns after dissociation and reaggregation depends on some supercellular organization process, rather than migration of previously determined cells to their appropriate site in the pattern under reconstruction. V . CELLULAR DIFFERENTIATION

Cellular differentiation is a constant feature of postembryonic life; even in adult insects there is continuous regeneration of intestinal cells and haemocytes (Krishnakumaran et al., 1967). Most typical examples concern cases where proliferating cells cease dividing, acquire characteristic organelles and make specialized products. Of course dividing cells are also synthesizing many characteristic products and assembling them into conspicuous organelles but cell differentiation begins essentially with the transition from this set of syntheses to another. Often this will result in the creation of a mature, stable cell type, and the idea has therefore grown up that differentiation involves a loss of developmental plasticity (Grobstein, 1966), but so many cases of “dedifferentiation” have been established that this part of the definition could well be dropped and the attention turned to the acquisition of particular syntheses (Hay, 1966, 1968). The transformation from larval to adult syntheses which occurs in the epidermal cells of many insects at metamorphosis (Wigglesworth, 1934, 1940b; Lawrence, 1966c, 1969) is an example of differentiation. One lesson to be culled from such insects is the real distinction between determination and differentiation, for here these processes are often widely separated in time. The adult pattern of

258

PETER A. LAWRENCE

Rhodnius is determined very early-although it can be evoked in the first-stage larva if it is parabiosed to a fifth stage (Wigglesworth, 1934) it normally does -not appear until metamorphosis, many cell divisions later. During this growth the particular characteristics of the latent adult pattern are passed on to daughter cells. If wounding results in excessive divisions of cells determined to form a pigment spot, or a particular bristle type (Wigglesworth, 1940b), then a t metamorphosis a larger pigment spot or more bristles of that type are formed in the adult. The same system is responsible for the maintenance of determination in discs, and such latent but propagable determined states are probably usual in developing systems (Baker, 1967). The controlling factor in cellular metamorphosis is juvenile hormone; in its presence the cells make larval cuticle and in its absence adult cuticle. If, in Oncopeltus, juvenile hormone is applied to cells undergoing cellular metamorphosis the cells will make cuticle which contains both adult and larval components. The synthesis of melanin changes from the larval to the adult pattern and becomes insensitive to the juvenile hormone early in the moult; in some areas this change involves the initiation, and in others the cessation, of melanin production. The surface sculpturing of the cuticle remains sensitive to the hormone for much longer, and all intergradations of cuticle between larval and adult can be made (Lawrence, 1969). This production of intermediate cuticle shows that the transformation can be broken down into components, which illustrates that the synthetic pathways may be autonomous, and may differ in sensitivity to the juvenile hormone, or in their moment of commitment. Little is known about the intercellular switches which control the synthetic pathways involved in differentiation in any organism. One feature of insects, which is certain to help, is puffing in polytene chromosomes (Ashburner, 1967, 1970),but the simplest systems for studying the biochemistry of differentiation employ cells which make a great deal of one product. Kafatos has explored one such system in an insect. During metamorphosis of the pupa of Antherueu particular epidermal cells of the galea differentiate and diversify in t o three types. This diversification stems from unusually oriented mitoses (vertical division illustrated in Fig. 1(b), Kafatos and Feder, 1968) which probably can be regarded as differentiative divisions like those found in bristle or dermal gland development (Lawrence, 1966b). One cell becomes highly polyploid and manufactures a large

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

259

quantity of a pure enzyme (cocoonase) which is deposited as crystals of active substance on the outside surface of the pupal galeae (Kafatos and Feder, 1968) and later is dissolved in a buffer solution secreted by the labial glands (Kafatos and Williams, 1964; Kafatos, 1968). The' active enzyme then digests the hard proteinaceous cocoon to allow escape of the emergent moth. Kafatos and Feder (1968) have shown that DNA synthesis associated with increasing polyploidy of the secretory cell nucleus occurs concurrently with rapid synthesis of cocoonase. This is of particular interest because of the rather widely held view that DNA synthesis and differentiated cell function are mutually exclusive (Holtzer, Abbott, Lash and Holtzer, 1960). As the cocoonase is synthesized it is sequestered in a large vacuole in the cytoplasm, and pulse chase studies with explanted galeae have shown that about 70% of labelled protein enters the vacuole within 2 hr of labelling (Kafatos and Reich, 1968). The remaining 30% of synthesized protein stays in the cytoplasm. In the presence of 60pg/ml of Actinomycin D in vitro, uptake of labelled uridine into nuclear RNA drops to only about 1% of its value in paired control galeae. After 12 hr incubation in Actinomycin D rate of uptake of amino acids into cocoonase is about 80% of controls, whereas the uptake into cytoplasmic protein has dropped to about 10% of its former level. Kafatos and Reich interpret these interesting results as evidence that the messenger RNA for cocoonase, in contrast to that specifying the indefinite proteins of the cytoplasm, is long-lived. More will, no doubt, come from this system. VI. OUTLOOK

If one is interested in biological phenomena rather than in a particular species it is well to choose one's experimental organism carefully. There is a case for using insects to study all branches of developmental biology, but it is in the study of pattern formation that they offer outstanding advantages. The two-dimensional nature of the cuticle and the precisely local differentiation that occurs, as well as the orientation of cuticular structures themselves, has already offered a tantalizing glimpse at methods of intercellular communication, and may well take us further than the traditional systems have done. In studying intercellular communication we need to know what chemical molecules and ionic interchange occur between the cells of a developing tissue and how information is encoded. By this

260

PETER A. LAWRENCE

route it may be possible to escape from hazy fields and gradients to a more mechanistic understanding.

ACKNOWLEDGEMENTS

I am grateful to Professor J. M. Thoday for the hospitality of his department; to the Agricultural Research Council for financial support; to F. H. C. Crick, P. Hayward, G. Webster and L. Wolpert for useful discussion; to M. Ashburner and J . Willis for critically reading the manuscript; and to H. Bohn, E. Hadorn, W. Marcus, R. Nothiger, H. Piepho, C. Stern, the late H. Stumpf and H. Ursprung for permission to use figures. I am particularly indebted to R. Nothiger for his most careful criticism of the imaginal disc section.

REFERENCES Arnheim, N. (1967). The regional effects of two mutants in Drosophila analysed by means of mosaics. Genetics, Princeton 56, 253-263. Ashburner, M. (1967). Gene activity dependent on chromosome synapsis in the polytene chromosomes of Drosophila melanogaster. Nature, Lond. 2 14, 1159-1 160. Ashburner, M. (1970). Behaviour of polytene chromosomes in insect development. (This Volume.) Baker, W. K. (1967). A clonal system of differential gene activity in Drosophila. Dev. Biol. 16, 1-17. Bart, A. (1965a). Sur l’origine des formations surnumeraires au cours de la regeneration des pattes chez Carausius motosus Br. C.r. hebd. Sianc. Acad. Sci., Paris 261, 1901-1903. Bart, A. (1 965b). Induction experimentale d’une morphogenk accompagnant ou non la regeneration de la patte de Carausius morosus Br. C.r. hebd. Seanc. Acad. Sci., Paris 261,5619-5622. Bodenstein, D. ( 1935). Beintransplantationen an Lepidopterenraupen. 111. Zur Analyse der Entwicklungspotenzen der Schmetterlingsbeine. Arch. EntwMech. Org. 133, 156-192. Bodenstein, D. ( 1937). Beintransplantationen an Lepidopteranraupen. IV. Zur Analyse experimentell erzeugter Bein-Mehrfachbildungen. Arch. EntwMech. Org. 136,745-785. Bodenstein, D. (1939). Investigations on the problem of metamorphosis. V. Some factors determining the facet number in the Drosophila mutant Bar. Genetics, Princeton 24,494-508. Bodenstein, D. ( 1941). Investigations on the problem of metamorphosis. VIII. Studies on leg determination in insects. J. exp. 2001.8 7 , 31-53. Bodenstein, D. (1943). Hormones and tissue competence in the development of Drosophila. Biol. Bull. Mar. biol. Lab., Woods Hole 84, 34-58.

POLARITY AND PATTERNS I N DEVELOPMENT O F INSECTS

261

Bohn, H. ( 1965a). Analyse der Regenerationsfahigkeit der Insektenextremitat durch Amputations und Transplantationsversuche an Larven der afrikanischen Schabe Leucophaea maderae. Fabr. (Blattaria) I. Mitteilung Regenerationspotenzen. Arch. EntwMech. Org. 156, 49-74. Bohn, H. ( 1965b). Analyse der Regenerationsfaigkeit der Insektenextremitat durch Amputations und Transplantationsversuche an Larven der afrikanischen Schabe (Leucophaea maderae Fabr.) 11. Mitteilung. Achsendetermination. Arch. EntwMech. Org. 156, 449-503. Bohn, H. ( 1967). Transplantationsexperimente mit interkalarer Regeneration zum Nachweis eines sich segmental wiederholenden Gradienten im Bein von Leucophaea (Blattaria). 2001.Anz. 30, (Suppl.) 499-508. Brqhdsted, H. V. (1955). Planarian regeneration. Biol. Rev. 30, 65-126. Child, C. M. (1941). “Patterns and Problems of Development”. 811 pp. University of Chicago Press, Chicago. Clark, P. J. and Evans, F. C. (1954). Distance to nearest neighbour as a measure of spatial relationships in populations. Ecology 35, 445-453. Claxton, J. H. (1964). The determination of patterns with special reference to that of the central primary skin follicles in sheep. J. theor. Biol. 7 , 302-317. Coon, H. G. (1966). Clonal stability and phenotypic expression of chick’ cartilage cells in vitro. Proc. natn. Acad. Sci. U.S.A. 55,66-73. Garcia-Bellido, A. ( 1966a). Pattern reconstruction by dissociated imaginal disc cells of Drosophila melanogaster. Devl. Biol. 14, 278-306. Garcia-Bellido, A. (1966b). Changes in selective affinity following transdetermination in imaginal disc cells of Drosophila melanogaster. Expl. Cell Res. 44, 382-392. Garcia-Bellido, A. ( 1967). Histotypic reaggregation of dissociated imaginal disc cells of Drosophila melanogaster cultured in vivo. Arch. EntwMech. Org. 158,212-217. Gehring, W. (1966). Ubertragung und Anderung der Determinationsqualitaten in Antennenscheiben-Kulturen von Drosophila melanogaster. J. Embryol. exp. Morph. 15,77-111. Gehring, W. (1967). Clonal analysis of determination dynamics in cultures of imaginal discs in Drosophila melanogaster. Devl. Biol. 1 6 , 4 3 8 4 6 . Gehring, W. (1968). The stability of the determined state in cultures of imaginal discs in Drosophila. In “The Stability of the Differentiated State” (H. Ursprung, ed.), pp. 136-154. Springer-Verlag, Berlin. Goodwin, B. C. and Cohen, M. H. (1969). A phase-shift model for the spatial and temporal organization of living systems. J. theor. Biol. 25,49-107. Grobstein, C. (1 966). Differentiation of vertebrate cells. In “The Cell” (J. Brachet and A. E. Mirsky, eds), Vol. I, pp. 437-496. Academic Press, New York. Hadorn, E. (1953). Regulation and differentiation within field districts in imaginal discs of Drosophila. J. Embryol. exp. Morph. 1 , 2 13-2 16. Hadorn, E. ( 1963). Differenzierungsleistungen wiederholt fragmentierter Teilstiicke mannlicher Genitalscheiben von Drosophila melanogaster nach Kultur in vivo. Devl. Biol. 7 , 617429. Hadorn, E. ( 1965). Problems of determination and transdetermination. Brookhaven Symp. Biol. 18, 148-161. Hadorn, E. (1966). Konstanz, Wechsel and Typus der Determination und

2 62

PETER A. LAWRENCE

Differenzierung in Zellen aus msjlnlichen Genitalanlagen von Drosophila melanogaster nach Dauerkultur in vivo. Devl. Biol. 13,424-509. Hadorn, E., Anders, G. and Ursprung, H. (1959). Kombinate aus teilweise dissoziierten Imaginalscheiben verschiedener Mutanten und Arten von Drosophila. J. exp. Zool. 142, 159-175. Hadorn, E., Bertani, G. and Gallera, J. (1949). Regulationsfiihigkeit und Feldorganisation der mannlichen Genital-Imaginalscheibe von Drosophila melanogaster. Arch. EntwMeGh. Org. 144, 3 1-10. Hadorn, E. and Buck, D.( 1962). Uber Entwicklungsleistungen transplantierter Teilstucke von Fliigel-Imaginalscheiben von Drosophila melanogaster. Rev. suisse Zool. 69, 302-310. Hadorn, E. and Chen, P. S . (1956). Die Feldorganisation der Spermathekenanlage bei Drosophila melanogaster. Rev. suisse Zool. 63, 268-211. Hadorn, E. and Gloor, H. (1946). Transplantationen zur Bestimmung des Anlagemusters in der weiblichen Genital-Imaginalscheibe von Drosophila. Rev. suisse Zool. 53,495-501. Hannah, A. (1953). Non-autonomy of yellow in gynandromorphs of Drosophila melanogaster. J. exp. Zool. 123, 523-560. Hay, E. D. (1966). “Regeneration”. Holt, Rinehart and Wilson, New York. Hay, E. D. (1968). Dedifferentiation and metaplasia in vertebrate and invertebrate regeneration. In “The Stability of the Differentiated State” (H. Ursprung, ed.), pp. 85-108. Springer-Verlag, Berlin. Holtzer, H., Abbott, J., Lash, J. and Holtzer, S. (1960). The loss of phenotypic trait by differentiated cells in vitro. I. Dedifferentiation of cartilage cells. Proc. natn. Acad. Sci. U.S.A. 46, 1533-1542. Kafatos, F. C. (1968). The labial gland: a salt-secreting organ of saturniid moths. J. exp. Biol. 48,435-453. Kafatos, F. C. and Feder, N. (1968). Cytodifferentiation during insect metamorphosis: the galea of silkmoths. Science, N . Y. 161,470-472. Kafatos, F. C. and Reich, J. (1968). Stability of differentiation-specific and nonspecific messenger RNA in insect cells. Proc. natn. Acad. Sci. U.S.A. 60, 1458-1465. Kafatos, F. C. and Williams, C. M. (1964). Enzymatic mechanism for the escape of certain moths from their cocoons. Science, N . Y. 146,538-540. Krishnakumaran, A., Berry, S . J., Oberlander, H. and Schneiderman, H. A. (1967). Nucleic acid synthesis during insect development. 11. Control of DNA synthesis in the cecropia silkworm and other saturniid moths. J. Insect Physiol. 13, 1-57. Kroeger, H. (1958). Uber Doppelbildungen in die Liebeshohle verpflanzter Fliigelimaginalscheiben von Ephestia kuhniella Z. Arch. En twMech. Org. 150,401-424. Kroeger, H. ( 1959). Determinationsmosaike aus kombiniert implantierten Imaginalscheiben von Ephestia kuhniella Zeller. Arch. EntwMech. Org. 151, 113-135. Kroeger, H. (1960). Die Entstehung von Form im morphogenetischen Feld. Naturwissenschaften 47, 148-153. Lawrence, P. A. (1965). The Determination and Development of Hairs and Bristles in the Milkweed Bug, Oncopeltus fasciatus, Dall. Thesis, University of Cambridge.

POLARITY AND PATTERNS IN DEVELOPMENT O F INSECTS

263

Lawrence, P. A. (1966a). Gradients in the insect segment: the orientation of hairs in the milkweed bug, Oncopeltus fasciatus. J. exp. Biol. 44, 607-620. Lawrence, P. A. (1966b). Development and determination of hairs and bristles in the milkweed bug, Oncopeltus fasciatus (Lygaeidae, Hemiptera). J. Cell. Sci. 1 , 4 7 5 4 9 8 . Lawrence, P. A. ( 1 9 6 6 ~ ) The . hormonal control of the development of hairs and bristles in the milkweed bug, Oncopeltus fasciatus, Dan. J. exp. Biol. 44, 5 07-5 22. Lawrence, P. A. (1968). Mitosis and the cell cycle in the metamorphic moult of the milkweed bug, Oncopeltus fasciatus. A radioautographic study. J. Cell. Sci. 3, 391-404. Lawrence, P. A. (1969). Cellular differentiation and pattern formation during metamorphosis of the milkweed bug Oncopeltus. Devl. Biol. 19, 12-40. Lawrence, P. A. (1970). The development of spatial patterns in the integument of insects. In “Developmental Systems-Insects” ( S . J. Counce, ed.), Academic Press, New York. (In press.) Lawrence, P. A. and Hayward, P. (1970). (In preparation.) Lees, A. D. and Waddington, C. H. (1942). The development of the bristles in normal and in some mutant types of Drosophila rnelanogaster. Proc. R . SOC.B131,87-110. Lender, T. (1956). L’inhibition de la rkgknkration du cerveau des Planaires Polycelis nigra (Ebb.) et Dugisia lugubris (0. Schm.) en presence des broyats de tQtesou de queues. Bull. SOC.zool. Fr. 81, 192-199. Lender, T. (1965). La rkgknkration des planaires. In “Regeneration in Animals and Related Problems” (V. Kiortsis and H. A. L. Trampusch, eds), pp. 95-1 11. North Holland Publishing Co., Amsterdam. Lesseps, R. J. ( 1965). Culture of dissociated Drosophila embryos: aggregated cells differentiate and sort out. Science, N. Y. 148, 502-503. Lobbecke, E. A. (1958). Uber die Entwicklung der imaginal Epidermis des Abdomens von Drosophila ihre Segmentierung und die Determination der Tergite. Biol. Zbl. 77, 209-237. Locke, M. (1958). The co-ordination of growth in the tracheal system of insects. Q. Jl microsc. Sci. 99, 373-391. Locke, M. (1959). The cuticular pattern in an insect, Rhodnius prolixus Stal. J. exp. Biol. 36 459477. Locke, M. (1960). The cuticular pattern in an insect-the intersegmental membranes. J. exp. Biol. 37,398-406. Locke, M . (1964). The structure and formation of the integument in insects. In “Physiology of the Insecta” (M. Rockstein, ed.), pp. 379470. Academic Press, New York. Locke, M. (1966). The cuticular pattern in an insect: the behaviour of grafts in segmented appendages. J. Insect Physiol. 12, 397402. Locke, M. (1967). The development of patterns in the integument of insects. Adv. Morphogen. 6 , 33-88. Loosli, R. ( 1959). Vergleich von Entwicklungspotenzen in normalen, transplantierten und mutierten Halteren-Imaginalscheiben von Drosophila rnelanogaster. Devl. Biol, 1,24-64. Luond, H. ( 196 1). Untersuchungen zur Mustergliederung in fragmentierten Primordien des mannlichen Geschlechtsapparates von Drosophila seguyi. Devl. Biol. 3 , 615-656.

2 64

PETER A. LAWRENCE

Maas, A. H. ( 1948). Uber die Auslosbarkeit von Temperaturmodifikationen w a r e n d der Embryonalentwicklung von Drosophila melanogaster Meigen. Arch. EntwMech. Org. 143, 5 15-572. Marcus, W. (1962). Untersuchungen uber die Polaritat der Rumpfhaut von Schmetterlingen. Arch. EntwMech. Org. 154, 56-102. Marcus, W. ( 1963). Beitrage zur Entwicklungsphysiologie der Rumpfhaut von Galleria mellonella L. Zool. A n z . (Suppl.) 26, 168-180. Neville, A. C. (1967). Daily growth layers in animals and plants. Biol. Rev. 42, 42 1-44 1. Neville, A. C. and Luke, B. M. (1969). Molecular architecture of adult locust cuticle at the electron microscope level. Tissue and Cell 1, 355-366. Nothiger, R. ( 1964). Differenzierungsleistungen in kombinaten hergestellt aus Imaginalscheiben. Verschiedenes Arten, Geschlechter und Korpersegmente von Drosophila. Arch. EntwMech. Org. 155,269-301. Nothiger, R. and Schubiger, G. (1966). Developmental behaviour of fragments of symmetrical and asymmetrical imaginal discs of Drosophila melanogaster (Diptera). J. Embryol. exp. Morph. 16, 355-368. Peyer, B. and Hadorn, E. (1966). Zum Manifestationsmuster der Mutante “multiple wing hairs” (mwh) von Drosophila melanogaster. Arch. Julius Klaus-Stiff. VererbForsch. 40, 19-26. Piepho, H. (1955a). Uber die Ausrichtung der Schuppenbalge und Schuppen am Schmetterlingsrumpf. Natunuissenschaften, 42,22. Piepho, H. (1955b). Uber die polare Orientierung der Balge und Schuppen auf dem Schmetterlingsrumpf. Biol. Zbl. 74,467-474. Piepho, H. and Marcus, W. (1957). Wirkungen richtender Faktoren bei der Bildung der Schuppen und Balge des Schmetterlingsrumpfes. Biol. Zbl. 76, 23-27. Rand, H. W., Bovard, J. F. and Minnich, D. E. (1926). Localization of formative agencies in Hydra. Proc. natn. Acad. Sci. U.S.A. 12, 565-570. Roberts, P. (1964). Mosaics involving aristapedia, a homeotic mutant of Drosophila melanogaster. Genetics, Princeton 49, 593-598. Rose, S. M. (1957a). Cellular interaction during differentiation. Biol. Rev. 32, 35 1-382. Rose. S . M. (1957b). Polarized inhibitory effects during regeneration in Tubularia. J. Morph. 100, 187-205. Rose, S. M. (1963). Polarized control of regional structure in Tubularia. Devl. Biol. 7,488-501. Schubiger, G. ( 1968). Anlageplan, Determinationszustand und Transdeterminationsleistungen der miinnlichen Vorderbeinscheibe von Drosophila melanogaster. Arch. EnhvMech. Org. 160, 9-40. Smith, S. D. (1963). Specific inhibition of regeneration in Clymenella torquata. Biol. Bull. Mar. biol. Lab., Woods Hole 125, 542-555. Sobels, F. H. (1952). Genetics and morphology of the genotype “asymmetric” with special reference to its “abnormal abdomen” character. Genetica, 26, 117-279. Stern, C. (1954). Two or three bristles. A m . Sci. 42, 213-247. Stern, C. (1968). “Genetic Mosaics and Other Essays”. 185 pp. Harvard University Press, Cambridge, Mass. Stern, C. and Hannah, A. (1950). The sex combs in gynanders of Drosophila melanogaster. Port. Acta biol. Ser. A 798-8 12.

POLARITY AND PATTERNS IN DEVELOPMENT OF INSECTS

265

Stem, C. and Tokunaga, C. (1967). Nonautonomy in differentiation of patterndetermining genes in Drosophila. I. The sex comb of Eyeless-dominant. Proc. natn. Acad. Sci. U.S.A. 57, 658-664. Stumpf, H. (1965a). Die Deutung der Riefenmuster bei Rhodnius prolixus auf Grund eines Konzentrationsgefalles. Naturwissenschaften 52, 500-50 1. Stumpf, H. (1965b). Deutung der Richtungsmuster der Schuppen von Galleria mellonella auf Grund eines Konzentrationsgefalles. Naturwissenschaften 52, 522. Stumpf, H. (1 966). Mechanism by which cells estimate their location within the body. Nature, Lond. 2 12,430-43 1. Stumpf, H. ( 1967a). Uber den Verlauf eines Schuppenorientierenden Gefalles bei Galleria mellonella. Arch. EntwMech. Org. 158, 3 15-330. Stumpf H. ( 1967b). Uber die Lagebestimmung der Kutikularzonen innerhalb eines Segmentes von Galleria mellonella. Devl. Biol. 16, 144-167. Stumpf, H. ( 1968). Further studies on gradientdependent diversification in the pupal cuticle of Galleria mellonella. J. exp. Biol. 49, 49-60. Tardent, P. (1963). Regeneration in the Hydrozoa. Biol. Rev. 38,293-333. Tobler, H. ( 1966). Zellspezifische Determination und Beziehung zwischen Proliferation und Transdetermination in Bein- und Flugelprimordien von Drosophila melanogaster. J. Embryol. exp. Morph. 16, 609-633. Tokunaga, C. (1961). Differentiation of a secondary sex comb under the influence of the gene engrailed in Drosophila rnelanogaster. Genetics, Princeton 46, 158-176. Tokunaga, C. and Stern, C. (1969). Determination of bristle direction in Drosophila. Devl. Biol. 20,411-425. Tucker M. (1959). Inhibitory control of regeneration in nemertean worms. J. Morph. 105, 569-600. Ursprung H. ( 1957). Untersuchungen zum Anlagemuster der weiblichen Genitalscheibe von Drosophila melanogaster durch UV-Strahlenstich. Rev. suisse Zool. 64, 303-3 16. Ursprung, H. ( 1959). Fragmentierungs- und Bestrahlungsversuche zur Bestimmung von Determinationszustand und Anlageplan der Genitalscheiben von Drosophila melanogaster. Arch. EntwMech. Org. 15 1, 504-558. Ursprung, H. ( 1962). Einfluss des Wirtsalters auf die Entwicklungsleistung sagittalhalften mannlicher Genitalscheiben von Drosophila melanogaster. Devl. Biol. 4, 22-39. Ursprung, H. and Hadorn, E. (1962). Weitere Untersuchungen uber Musterbildung in Kombinaten aus teilweise dissoziierten FlugelImaginalscheiben von Drosophila melanogaster. Devl. Biol. 4,40-66. Vogt, M. ( 1944). Beitrag zur Determination der Imaginalscheiben bei Drosophila. Naturwissenschaften 32, 39-40. Vogt, M. (1946). Zur labilen Determination der Imaginalscheiben von Drosophila. I. Verhalten verschiedenaltriger Imaginalanlagen bei operativer Defektsetzung. Biol. Zbl. 65, 223-238. Waddington, C. H. (1956). “Principles of Embryology”. 5 10 pp. Allen and Unwin Ltd., London. Webster, G. (1966). Studies on pattern regulation in hydra. 11. Factors controlling hypostome formation. J. Embryol. exp. Morph. 16, 105-122. Webster, G. and Wolpert, L. (1966). Studies on pattern regulation in hydra. I.

266

PETER A. LAWRENCE

Regional differences in time required for hypostome determination. J. Embryol. exp. Morph. 16,91-104. Wildermuth, H. ( 1968a). Differenzierungsleistungen, Mustergliederung und Transdeterminationsmechanismenin hetero- und homoplastischen Transplantaten der Riisselprimordien von Drosophila. Arch. EntwMech. Org. 160,41-75. Wildermuth, H. ( 1968b). Autoradiographische Untersuchungen zum V e r mehrungsmuster der Zellen in proliferenden Russelprimordien von Drosophila melanogaster. Devl. Biol. 18, 1-13. Wildermuth, H. and Hadorn, E. ( 1965). Differenzierungsleistung der Labial-Imaginalscheibe von Drosophila melanogaster. Rev. suisse Zool. 72, 686-694. Wigglesworth, V. B. (1933). The physiology of the cuticle and of ecdysis in Rhodnius prolixus (Triatomidae, Hemiptera); with special reference to the function of the oenocytes and of the dermal glands. Q. Jl microsc. Sci. 76,269-318. Wigglesworth, V. B. (1934). The physiology of ecdysis in Rhodnius prolixus (Hemiptera). 11. Factors controlling moulting and “metamorphosis”. Q. Jl microsc. Sci. 77, 191-222. Wigglesworth, V. B. (1937). Wound healing in an insect (Rhodnius prolixus Hemiptera). J . exp. Biol. 14,364-381. Wigglesworth, V. B. (1940a). Local and general factors in the development of “pattern” in Rhodnius prolixus (Hemiptera). J. exp. Biol. 17, 180-200. Wigglesworth, V. B. (1940b). The determination of characters at metamorphosis in Rhodnius prolixus (Hemiptera). J. exp. Biol. 17,201-222. Wigglesworth, V. B. (1942). The significance of “chromatic droplets” in the growth of insects. Q. Jl microsc. Sci. 83, 141-152. Wigglesworth, V. B. (1953). The origin of sensory neurones in an insect, Rhodnius prolixus (Hemiptera). Q. Jl microsc. Sci. 94,93-112. Wigglesworth, V. B. (1959). “The Control of Growth and Form: A Study of the Epidermal Cell in an Insect”. 140pp. Cornell University Press, Ithaca, N.Y. Wigglesworth, V. B. (1964). Homeostasis in insect growth. Symp. SOC.exp. Biol. 18,265-28 1. Wigglesworth, V. B. (1965). “The Principles of Insect Physiology”. 6th edition, 741 pp. Methuen, London. Williams, C. M. (1968). Ecdysone and ecdysone-analogues: their assay and action on diapausing pupae of the Cynthia silkworm. Biol. Bull. Mar. biol. Lab., Woods Hole 134,344-355. Wolff, E. (1962). Recent researches on the regeneration of Planaria. In “Regeneration” (D. Rudnick, ed.), pp. 53-84. Ronald Press, New York. Wolpert, L. (1969). Positional information and the spatial pattern of cellular differentiation. J. theor. Biol. 2 5 , 1-47. Zimmermann, W. (1954). Uber genetische und modifikatorisch bedingte Storungen der Segmentierung bei Drosophila melanogaster. Z . VererbLehre 86, 327-372.

Regulation of Intermediary Metabolism. with Special Reference to the Control Mechanisms in Insect Flight Muscle BERTRAM SACKTOR Gerontology Research Center National Institute of Child Health and Human Development National Institutes of Health. Baltimore. Maryland. U.S.A. I . Introduction . . . . . . . . . . . . . . . . . . 268 I1. Physiological Properties and Structural Organization of Insect Flight Muscle . . . . . . . . . . . . . . . . . . . . 269 A . Utilization of Oxygen during Flight . . . . . . . . . 269 Supply of Oxygen and Fuel to the Flight Muscle B. . . . . 269 C. Nature of the Substrate Consumed during Flight . . . . 271 D . Properties of the Contractile Proteins . . . . . . . . 271 E . Morphological Organization of Flight Muscle . . . . . . 275 F. Structural-Functional Correlates . . . . . . . . . . 281 111. Regulation of Carbohydrate Metabolism . . . . . . . . . 281 A . Glycogenolysis . . . . . . . . . . . . . . . 283 B. Phosphorylase b Kinase . . . . . . . . . . . . . 295 C. Glycogen Synthetase . . . . . . . . . . . . . . 295 D . Trehalase . . . . . . . . . . . . . . . . . 296 E. Biosynthesis of Trehalose . . . . . . . . . . . . 300 F. Glycolysis . . . . . . . . . . . . . . . . 303 G . Identification of Other Loci of Control of Metabolism . . 310 IV . Regulation of Fat Metabolism . . . . . . . . . . . . . 312 A . Fatty Acid Catabolism . . . . . . . . . . . . . 313 B. The Role of Carnitine . . . . . . . . . . . . . 314 C. Biosynthesis of Fat . . . . . . . . . . . . . . 316 D . Mobilization and Transport of Fat . . . . . . . . . 319 V . Regulation of Mitochondria1 Metabolism . . . . . . . . . 322 A . The Respiratory Chain and Oxidative Phosphorylation . . 323 B. Control of Pyruvate Oxidation . . . . . . . . . . 325 C. Control of Proline Oxidation . . . . . . . . . . . 330 D . Control of a-Glycero-P Oxidation . . . . . . . . . 332 E. The Energy-Dependent Accumulation of CaZ+and Pi . . . 333 F . Interactions of Metabolic Effectors with the Respiratory Chain 3 34 VI . Conclusions . . . . . . . . . . . . . . . . . . . 336 References . . . . . . . . . . . . . . . . . . . . . 338 261

268

B. SACKTOR

I . INTRODUCTION

During the rapid transition of a tissue from a “resting” state to one carrying on intense physiological work, the metabolic rate of the tissue increases many-fold. This indicates that the tissue has the full capacity to carry out glycolytic and oxidative reactions at the higher rate, in fact, classical determinations ’of activities of individual enzymes and concentrations of substrates in the tissue may suggest that metabolism should occur at maximal rates at all times. This is not the case, however, because biochemical pathways in living organisms are regulated or controlled. Studies of the regulation of metabolism, especially of systems in vivo, were initiated, therefore, with the following question in mind. What are the mechanisms by which energy-yielding biochemical pathways become activated instantaneously; or, perhaps more to the point, what are the mechanisms for keeping energy reserves from being used until needed? Previously, studies on the regulatory mechanisms of glycolysis and oxidative metabolism, in vivo, were limited to viable yeast or ascites cells or, in vitro, to a wide variety of preparations, ranging from reconstructed systems with purified enzymes to the more physiological isolated perfused heart or brain made anoxic by decapitation. The approach in our laboratory has been to seek regulatory processes in intact animals that are amenable to this kind of experimental inquiry. A unique system to determine the control mechanisms in a living animal during the transition from rest to activity is found in the flying insect. Initiation of flight induces an instantaneous increase in rate of energy transformation which is far in excess of that for any other biological system. In addition, muscular activity of the insect is readily controllable without use of traumatic experimental techniques. Moreover, as will be shown subsequently, kinetic measurements, in vitro, can be used successfully to predict the factors which control activity, in vivo. The energy used by insect muscle in performance of work which, at times, can total more than one million successive wingbeats with rates as great as lOOO/s, is ultimately derived from chemical reactions going on within the muscle. The nature and regulation of these chemical interconversions in the muscle and in other tissues, such as in fat body, that have significant and direct effects on the metabolism of the muscle, are the subjects of this review.

REGULATION OF INTERMEDIARY METABOLISM

269

11. PHYSIOLOGICAL PROPERTIES A N D STRUCTURAL ORGANIZATION

OF INSECT FLIGHT MUSCLE A. UTILIZATION OF OXYGEN DURING FLIGHT

The over-all level of metabolism, or biochemical interconversions, in the working muscle may be estimated from either the respiratory exchange or the depletion of the animals’ depots of fuel. In terms of cal/g of muscle/hr, values as high as 2400 for the bee during prolonged periods of continuous flight have been reported (Weis-Fogh, 1952). The cost of flight in insects can also be measured by comparing the rate of oxygen uptake during flight with that of the same insect at rest. Increases as great as 50-100 times the resting values have been recorded in a variety of insect species (see Sacktor, 1965). For example, Davis and Fraenkel (1940) reported that the resting respiratory rate of the blowfly, Luciliu, is 33-50 pl/g/min and this is increased 30-50 times during flight. Some individuals have oxygen consumptions of about 3000 pl/g/min during flight, thus elevating their resting rates approximately 100-fold. Such large increases in respiration upon initiation of flight are not restricted to Diptera and Hymenoptera, which have the asynchronous or fibrillar type of muscle striation and are characterized by a high frequency of movement of their wings. Essentially identical increases in oxygen uptake between individuals at rest and during flight have been observed in Orthoptera and Lepidoptera, which have the synchronous or close-packed type of muscle morphology and, in general, have relatively slow rates of wingbeat. For instance, in a variety of moth species, Zebe (1954) reported oxygen uptakes of from 7-1 2 pl/g/min at rest. These increased to values of 700-1 660 pl during flight, an increment of over 100-times in some cases. The 50to 100-fold increase in respiratory rate upon initiation of flight indicates that there is a large degree of control of respiration in muscle, in vivo. The mechanism of respiratory control will be discussed in a later section. B. SUPPLY OF OXYGEN AND FUEL TO THE FLIGHT MUSCLE As pointed out, oxygen uptake upon initiation of flight may reach a rate of 3000 pl/g/min. In spite of these enhanced respiratory rates, the observations that Drosophilu (Chadwick, 1953) and locusts (Krogh and Weis-Fogh, 195 1) can maintain flight for hours while accruing no, or only a small, oxygen debt indicate that the metabolic processes are not limited by the availability of oxygen. In insect

270

B. SACKTOR

flight muscle, myoglobin and hemoglobin are absent and air is conveyed directly to muscle through an elaborate conduit of tracheae. In most muscles the rich tracheolar system invades the fibers. Electronmicrographs, which will be described in detail later, show tracheoles to be in close opposition with mitochondria in a “mitochondrion-tracheole continuum” (Edwards and Ruska, 1955). The minute distances between tracheoles and oxygen-consuming elements of muscle suggest that diffusion suffices to transport at least part of the extra oxygen utilized. In fact, Weis-Fogh (1964, 1967) calculated that in small insects, including the flies, Drosophilu, Muscu, and Culliphoru, diffusion of respiratory gases is sufficient to account for the entire transport between the spiracles and the end of the tracheoles even at the highest rate of metabolism. In large insects, such as in dragon flies, locusts and wasps, the primary tracheole supply must be strongly ventilated while diffusion is sufficient in the remaining part of the air tubes. The ventilating mechanisms may be of two types: abdominal movements which result in abdominal pumping of air and hemolymph (in Vespu) and movements of the thoracic walls which are caused by the wing movements themselves and which result in thoracic pumping at wing-beat frequency (in Aeshnu). Weis-Fogh (1964) suggested that pumping of air and blood due to shortening of the fibers of the flight muscle is of little importance for the exchange of gases but of major importance for the supply of metabolic fuel. In many insects the flight muscles are so large that diffusion alone is quite inadequate t o transport substrate from the surface t o the interior and that exchange is achieved by a combination of diffusion and muscular pumping. He claimed that because of the inefficiency of the tidal process of moving blood into and out of intramuscular spaces as the muscle relaxes and contracts that the concentration of sugar in the blood must be very high (of the order of 0.5-1%) to provide gradients that can supply substrate to the catabolic enzymes in the interior of the fibers at the rates required for the exceptionally high activity of this tissue. The dependence of the rate of wingbeat on concentration of blood sugar in the blowfly (Clegg and Evans, 1961) is relevant to this view. Although muscular pumping may be important, the organization of the transverse tubular system (the Tsystem) and specific facilitated transport mechanisms may also have significant functions in transporting substrates. These latter concepts will be discussed more fully in later sections.

REGULATION OF INTERMEDIARY METABOLISM

27 1

C. NATURE OF THE SUBSTRATE CONSUMED DURING FLIGHT

Metabolic energy in working muscle is generated by oxidation of foodstuffs by atmospheric 0 2 ,with concomitant production of C02 . Measurements of the volumes of C 0 2 liberated and of O2 consumed, with calculation of the ratio of these volumes, the RQ, have been of value by virtue of inferences they permit in regard to the kind of substance undergoing oxidation. A compilation and full discussion of the data on the RQ during flight as well as the depletion of the insects’ reserves after flight was made previously (Sacktor, 1965). In summary, with insects having the asynchronous fibrillar type muscle, such as Diptera and Hymenoptera, the RQ is equal to unity and carbohydrates are the main, if not the exclusive, substrate. In those insects having the synchronous type of muscle, including Lepidoptera and Orthoptera, RQ values of 0.73 are found and fats are depleted, even though some species (moths) were gorged with glucose (Zebe, 1954). Locusts, roaches and aphids may use both carbohydrates and fat. Glycogen and trehalose are used during the initial period of flight; however, as flight continues, the RQ decreases and fat becomes the principal fuel and is able to sustain flight for hours. Also, flight muscle homogenates of the cecropia moth oxidize sugars and glycogen, and a moderately active trehalase was reported (Stevenson, 1968; Gussin and Wyatt, 1965). The role of amino acids as substrates for flight was considered earlier (Sacktor, 1961, 1965). The rapid utilization of proline on initiation of flight of the blowfly (Sacktor and Wormser-Shavit, 1966) to prime Krebs cycle activity (Sacktor and Childress, 1967) will be dealt with later. A unique example of the utilization of proline as an energy-furnishing reserve is found in the Tsetse-fly (Bursell, 1963, 1966). D. PROPERTIES OF THE CONTRACTILE PROTEINS

An understanding of the biochemical control mechanisms in insect flight muscle demands a knowledge of many aspects of insect organization, especially the structural and mechanical characteristics of the muscle as well as the physiological and biochemical properties of the contractile elements. These subjects have been the object of much experimental inquiry and a full description of them is clearly beyond the scope of the present discussion. Fortunately, excellent reviews on the properties of the contractile proteins have been published recently, including those by Maruyama (1 9 6 9 , Pringle (1967a, b) and Ruegg (1968). However, since the significances of various metabolic regulatory mechanisms become evident only when

272

B. SACKTOR

related to physiological events, a minimum of pertinent and essential features of the contractile processes in insect muscle must be noted. 1. Excitation-Contraction Coupling There are two fundamentally different ways in which the wing-beat frequency of insect flight muscle is determined. In one way, found in synchronous or afibrillar type muscles (Roeder, 1951), wingbeat frequency is determined by the activity of the central nervous system; each cycle of mechanical activity is triggered by one or a short burst of motor nerve impulse. This classical kind of excitation-contraction coupling is found in moths, dragonflies, and locusts and the flight of these insects is characterized by a rate of wingbeat of about 5-30/s. In the other way, found in asynchronous or fibrillar type muscles (Pringle, 1949), contraction is not triggered by nerve impulses and by action potentials in synchrony with the contraction cycle, rather mechanical activity is maintained by a self-oscillatory mechanism and is merely initiated and terminated by motor nerve control. Flies, bees, beetles and most bugs have this kind of flight muscle and these species have rapid rates of wingbeat, over 1OOO/s in some instances. Pringle and collaborators (see Pringle, 1967a) have shown that fibrillar muscle oscillates because of a unique property of the myofibrils themselves, in which there is a delay between changes of length and changes of tension in a mechanically resonant system. ATP is the source of energy and Ca and Mg ions are necessary activators. The threshold concentration for Ca2+ is approximately l o p 7 to lO-'M (Jewel1 and Ruegg, 1966); sensitivity of the muscle preparation to Ca2+depends on the concentration of MgZ+,which has a weak antagonistic effect on the action of Ca2+,and on mechanical conditions. The effect of muscle length is of particular importance, as stretching the fiber seems to be equivalent to raising the concentration of Ca2+.

2. Myofibrillar A TPase A Ca-activated myofibrillar ATPase in asynchronous flight muscle was reported some time ago (Sacktor, 1953a). More recently, vom Brocke (1966) noted that low levels of CaZ+( M) are needed for activation and that the degree of CaZ+ stimulation is less in oscillatory than in non-oscillatory muscles. This latter observation was further explored by Maruyama et al. (1 968), who showed that in synchronous muscle ATPase activity is low in the absence of Ca2+

REGULATION OF INTERMEDIARY METABOLISM

273

and that activity increases greatly over a narrow range of concentration of Ca2+. On the other hand, asynchronous muscle has a greater ATPase in the absence of Ca2+ but shows a much smaller increase over a wide range of concentration. The low Ca*+sensitivity of fibrillar flight muscle can be of marked physiological importance in that the small increment of Castimulation makes possible further increases in ATPase activity during stretch and oscillation (Ruegg and Tregear, 1966; Chaplain, 1967) and that the fibrillar muscle is less sensitive to small changes in concentration of Ca2+ over the physiological range. 3. Relaxing Factor Activity Tsukamoto e t al. (1966) described a granular fraction from locust flight muscle that inhibits myofibrillar ATPase, retards superprecipitation of actomyosin, and, in the presence of ATP, removes Ca2+ from myofibrils because of their strong Ca2+-binding capacity. These observations suggest that the locust granules are capable of acting as a relaxing factor. Indeed, the Ca2+-binding capacity of this insect preparation is comparable to that of rabbit skeletal muscle. Further, Smith (1966) showed, as will be illustrated in the next section, that locusts, having the synchronous-type muscle, have a well-developed sarcoplasmic reticulum, the source of the Ca2+ secluding vesicles. On the other hand, the sarcoplasmic reticulum is drastically reduced in asynchronous muscle (Smith, 1966). Since adequate Ca2+ cannot be released rapidly by the remnants of the reticulum, it is unlikely that the rapid rise and fall of tension occurring during oscillatory activity of fibrillar muscle involves the sarcoplasmic reticulum. In fact, Jewel1 and Ruegg (1966) and Chaplain (1967) demonstrated that glycerinated fibers carry out oscillatory contraction in the presence of ATP, Mg2+, and Ca2+, and that the frequency is independent of changes in concentration of Ca2+.Thus, fluctuations in the level of Ca2+ probably are not involved in the actual mechanism of oscillatory contraction, although Ca2+ is necessary to activate this system. Significantly, actomyosin ATPase is inhibited by excess MgATP and the allosteric inhibition is counteracted by physiological levels of Ca2+ (Chaplain, 1966). The possible role of mitochondria in maintaining Ca2+fluxes must not be overlooked in view of recent experiments by Carafoli e t al. (1969) showing energydependent rapid uptake and release of Ca2+ by mitochondria from blowfly flight muscle.

REGULATION OF INTERMEDIARY METABOLISM

275

E. MORPHOLOGICAL ORGANIZATION O F FLIGHT MUSCLE

The complexities and interactions of the biochemical systems in flight muscle become fully appreciated in light of structuralfunctional studies. The morphological organization of the tissue provides a basis for the biochemical similarities and differences between synchronous and asynchronous muscles, for the distinctive aggregation of enzymes into intracellular compartments, and for the regulatory mechanisms of the various metabolic pathways. Our knowledge of the structure of flight muscle stems largely from the elegant studies of David S . Smith (Smith, 1961a,b, 1963, 1965, 1966a, b). The reader is urged to examine these papers in detail since only the barest essentials consistent with the present discussion will be given here. The general organization of insect flight muscle was described in the pioneering investigations of Von Siebold (1848), Kolliker ( 1 857), and Watanabe and Williams (195 1). Among the distinguishing characteristics of the tissue is the large size of the fibrils, their arrangement into giant fibers, with diameters of up to several hundred microns, that are massed in turn, to fill the greatest part of the insect thorax and the short sarcomeres formed by restriction of the I-band. In addition, elongate columns of closely-packed mitochondria lie between the fibrils, the fibers are invaded by a rich tracheolar system and nuclei are few in number and often located peripherally. All the components of the muscle fiber are enveloped in an exceedingly thin sarcolemma. Structural unity results, in part, from the network of tracheae. Electron microscopic studies have elaborated the general morphological features of the tissue and have pointed out the complex, fine structure of insect flight muscle. A representative of the synchronous or close-packed, type of musclc is described in Fig. 1. This longitudinal section of the dragonfly muscle shows cylindrical fibrils approximately 2.3 p in length and 0.6 p in diameter. The sarcomeres are delineated by the prominent Z-bands. The I-bands are much-reduced. Transverse sections (Figs 2 and 3) indicate that the Fig. 1 . Longitudinal section of a synchronous type flight muscle of the dragonfly, Sympetmm, showing satcomere striations indicated by the prominent Z-bands. The mitochondria are arranged precisely opposite the sarcomeres. Cisternae of the sarcoplasmic reticulum lie in a bead-like fashion between the fibrils and mitochondria. Transversely sectioned profiles of the T-system tubules lie midway between the 2-band and center of the sarcomere. (Unpublished photo, kindly furnished by David S. Smith, similar to Fig. 3 in Smith, 1966b) ~35,000.

276

B. SACKTOR

actin filaments lie opposite to and between pairs of myosin filaments (Smith, 1966b). In the dragonfly, the mitochondria are slab-like and very large, and are interposed between the myofibrils and arrayed alongside each fibril precisely opposite the sarcomeres defined by successive Z-bands. In internal structure the mitochondria contain an enormous number of doubly lamellate cristae arranged in whirled and subparallel arrays. The large number of cristae is indicative of the high metabolic activity of these mitochondria. As first described by Smith (1966a), the fibrils of the muscle fibers are invested with two distinct series of membrane-limited components: a longitudinal system of cisternae, the sarcoplasmic reticulum; and a transversely oriented tubular system, the T-system. As seen in Fig. 1, membranes of the sarcoplasmic reticulum lie between the fibrils and the mitochondria in a continuous sheet bearing fenestrations which afford a beaded appearance. In synchronous muscle, the T-system tubules transverse the fiber midway between the Z-line and the middle of the sarcomere, lying in an indentation in the surface of the adjoining mitochondrion (Fig. 1). The two membrane systems enter into intimate association, in a dyad configuration. In the region of the dyad the fenestrations of the sarcoplasmic reticulum are absent; electron opaque material is present in the reticulum cisternae while the interior of the T-system appears to be entirely devoid of structural content. The origin of the T-system is apparent from Fig. 2. In the dragonfly the tubules are seen interposed between cisternae of the sarcoplasmic reticulum and the mitochondria and are derived as open invaginations of the plasma membrane at the surface of the fiber. Tubules opening directly to the interfiber space are in evidence. Figure 3 shows a fiber soaked in ferritin prior to fixation. The electron-dense particles of ferritin are clearly seen in the T-system space, between fibrils as well as between the fibrils and mitochondria, demonstrating the deep infusion of the muscle fiber by the extracellular milieu. The fine structure of flight muscle of the asynchronous or fibrillar type is illustrated in Figs 4 and 5. The cylindrical fibrils are very large, approximately 2 p in diameter and 3 p in length. The I-band is very narrow. As in other fibrillar muscles (Smith, 1961a and 1963; Shafiq, 1963; Ashhurst, 1967; Gregory et al., 1968), the Fig. 2. Transverse section of Sympetmm flight muscle showing origin of the transverse T-system as direct invaginations of the plasma membrane. (Unpublished photo, kindly furnished by David S. Smith, similar to Fig. 6 in Smith, 1966b) ~40,000.

A.1.P.-10

REGULATION OF INTERMEDIARY METABOLISM

279

mitochondria of the blowfly, Phormiu, are ovoid and irregular in shape, up to 4 p in length, and are not precisely aligned with respect to the myofibrillar striations. The extensive sarcoplasmic reticulum that surrounds the fibrils in synchronous muscle is markedly reduced in the asynchronous muscle (Fig. 4). The reticulum is represented only by small flattened vesicles, closely applied to the T-system tubules in a dyad configuration. On the other hand, the tubular plasma membrane system, derived from the circumtracheolar sheaths drawn into the fiber around the invaginated tracheoles, is profusely distributed throughout the fibers and plasma membranes lie in close proximity to the mitochondria (Fig. 5). Ferritin injected into the blowfly in vivo is distributed throughout, and exclusively in, the T-system (Smith and Sacktor, 1970). The numerous and large extracellular spaces between the myofibrils and delineated by the tracheolar plasma membrane system (Fig. 5) have been described erroneously by some recent authors as degenerative mitochondria, lysosomes, and storage vesicles. The blowfly flight muscle mitochondrion, as is true for other mitochondria, has a distinct outer limiting membrane and cristae which are continuous with the inner membrane of the mitochondrion. The cristae are arrayed as parallel plates, 30-35 cristae per micron (Smith, 1963; Gregory et al., 1967). The cristae are fenestrated; the fenestrations of successive cristae are aligned to form cylindrical channels within the mitochondrial matrix (Figs 4 and 5, and Smith, 1963). The membranes of the cristae are about 70 A in thickness, and the intervening spaces, the matrix of the mitochondrion and the intracristal spaces, are each approximately 100 A wide. The osmotically active matrix space was reported to be about 2.5 pl/mg mitochondrial protein (Hansford, 1968). Cristae, derived from osmotically disrupted mitochondria and negatively stained, are displayed as membraneous ribbons flanked on the side by spherical particles, each 80-95 A in diameter (Fig. 6 and Smith, 1966). Each particle is attached to the membrane by a stalk. Smith (1963) calculated that each cubic micron of the mitochondrion contains 200,000 particles, representing 10% of the total mitochondrial volume. As will be discussed subsequently, these particles are the sites of ATP synthesis, at least in mammalian mitochondria (Kagawa and Racker, 1966). Fig. 3. Transverse section of Symperncrn flight muscle soaked in ferritin solution prior to fixation in glutaraldehyde. (From Smith, 1966b,reproduced with kind permission of David S. Smith and J. Cell Biol.) ~105,000.

REGULATION OF INTERMEDIARY METABOLISM

28 1

F. STRUCTURAL-FUNCTIONAL CORRELATES

The physiological role of the T-system in excitation-contraction coupling was suggested from the observations of Hill (1948, 1949) that the transition from rest to activity following depolarization is so rapid in large fibers as to preclude the possibility that the activation is mediated by diffusion into the sarcoplasm of a substance released inside the surface cell membrane by membrane depolarization. The current view maintains that in muscles with a well-developed sarcoplasmic reticulum and T-system, surface depolarization of the plasma membrane is spread internally along the tubules of the T-system, which is morphologically and physiologically continuous with the external plasma membrane. Excitation is transferred from Tsystem to sarcoplasmic reticulum at the dyad. Ca2+is released from the cisternae of the sarcoplasmic reticulum, which diffuses to the myofibrils and activates the myofibrillar ATPase concomitant with contraction of the muscle. Following contraction, Ca2+ is actively reaccumulated by the sarcoplasmic reticulum and MgATP dissociates the actin myosin filaments, permitting relaxation. The resting potential across the surface plasma membrane and across the junction between the T-system and sarcoplasmic reticulum is then re-established. In fibrillar flight muscle, which has a well developed T-system but a reduced sarcoplasmic reticulum, the situation is less clear. It is possible that while excitation-contraction coupling at the onset of activity may be initiated and subsequently maintained by a mechanism similar to that occurring in other striated fibers, the mechanical activity of the muscle in the oscillatory state is controlled by an intrinsic property of the myofibrillar elements. The marked reduction in sarcoplasmic reticulum seems to rule out the possibility that sarcomere length is controlled by movement of Ca2+into and out of the reticulum. However, other evidence indicates that Ca2+is vitally important, not only in initiating the oscillatory process and in activating the myofibrillar ATPase but, as will be seen later, in participating in the control of the critical energy-furnishing metabolic enzymes in this muscle. 111: REGULATION OF CARBOHYDRATE METABOLISM

The chemical sources of muscular energy in an insect, such as the Fig. 4. Transverse section of asynchronous type flight muscle of the blowfly, Phormiu. Note the large diameter of the fibrils and mitochondria as well as the absence of an extensive sarcoplasmic reticulum around the fibrils. Remnants of the sarcoplasmic reticulum are in juxtaposition with the T-system tubules and appear as dyads. (From Smith and Sacktor, 1970) ~ 2 8 , 0 0 0 .

REGULATION OF INTERMEDIARY METABOLISM

283

blowfly, include glycogen depots in flight muscle and fat body, and trehalose and other sugars in blood and gastrointestinal tract. The depletion of these reserves during exercise and the general description of the glycolytic and oxidative pathways in flight muscle have been discussed previously (Sacktor, 1965). The present review represents a continuation of these findings, but with special emphasis on the biochemical control system in this muscle, which on initiation of flight undergoes a transition from a state of minimal to that of maximal metabolic flux. By measuring coincident and sequential changes in concentrations of carbohydrates, glycolytic and citric acid cycle intermediates, amino acids and nucleotides at the start of flight and during the “steady state” of a continuous flight, the loci of metabolic control have become evident (Sacktor and WormserShavit, 1966; Sacktor and Hurlbut, 1966). A. GLYCOGENOLYSIS

1. Utilization of Glycogen

Figure 7 describes the changes in concentrations of carbohydrates in flight muscle of the blowfly during flight. Changes in concentration are evident 5 s after induction of flight. Although the concentration of glycogen is not measurably decreased during this brief initial period, after about 2 min of flight glycogen in flight muscle does serve as a major energy reserve, supplying approximately 2.5 pmoles (as equivalents of glucose) each minute until it is depleted after about 10 min. This rapid utilization of glycogen in muscle, commencing shortly after flight is induced, indicates control of glycogenolysis. The concentration of glycogen in the fat body does not significantly decrease during the first 5 min of flight. However, the depletion of this depot is large after about 15 min and becomes more marked with flights of longer durations. Thus, during flight, glycogen is mobilized from its two principal storage loci at independent rates; glycogen in muscle being used before the polysaccharide in fat body.

2. Characterization of Flight Muscle Glycogen Until recently, virtually nothing was known of the physical or biochemical properties of glycogen from insects. Since Bueding and Orrell (1964) have shown that the molecular properties of the Fig. 5 . Longitudinal section of Phomia flight muscle including dyads and tracheoles surrounded by invaginated plasma membrane of muscle fiber. (From Smith and Sacktor, 1970) ~ 2 5 , 0 0 0 .

285

REGULATION OF INTERMEDIARY METABOLISM 8

32

24

G

9

-+ "

1 6 ' 0

+ W 15

2

GLYCOGEN

T R E HA L 0 S E

2 -

1

I

I

30

45

J

60

0

I

I

I

15

30

45

60

cs \

v, Y

2

0

za

.05t

GLU-6-P , ,

':kl 0

0

15

30

45

60

0

15

30

45

1

2

60

MINUTES Fig. 7. The concentrations of carbohydrates in thoraces of blowflies during 1 hr of flight. Insets with expanded time scale show alterations at the initiation of flight. (From Sacktor and Wormser-Shavit, 1966.)

polysaccharide may vary with the tissue of origin and with the metabolic state of the tissue, a characterization of glycogen from flight muscle is essential to understanding how glycogenolysis is activated. Examination of glycogen in situ (Fig. 8), reveals that it is located mostly in the interfibrillar sarcoplasm with scattered deposits on the myofibrils (Childress et al., 1970).The glycogen is in the form of alpha particles or rosettes which vary in size with diameters of up to 0.25 micron. The individual components comprising the rosettes, the beta particles, are not clearly delineated. However, the polydisperse nature of native glycogen is apparent from electron micrographs of isolated glycogen (Fig. 8(d)). Incubation of the flight muscle tissue with amylase following fixation results in disappearance of the glycogen rosettes, leaving only an amorphous network of electron dense material in this region. Sections of the Fig. 6. A portion of the cristae from disrupted mitochondria of the blowfly, Culliphoru, negatively stained with phosphotungstate. The laterally placed particles (ATPase) are attached by stalks to the axial structure. (From Smith, 1963, reproduced with kind permission of David S . Smith andJ. CellBioZ.) ~550,000. A.1.P.-10*

286

B. SACKTOR

Fig. 8. Localization of blowfly flight muscle glycogen in situ. Glycogen rosettes are located between mitochondria (Mi) and myofibrils (M). The area within the square is shown . Glycogen rosettes in higher magnification in Fig. 8(b) ~ 5 , 8 2 0 ,bar represents 1 . 0 ~ (b) (arrows) are prominent. ~ 7 5 , 0 0 0 bar , represents 0 . 2 ~(c) . Section similar to (b) but treated with a-amylase. Only amorphous electrondense material can be seen adjacent to the . Isolated rosettes of negatively stained mitochondria. ~ 7 5 , 0 0 0 ,bar represents 0 . 2 ~ (d) glycogen. ~ 7 6 , 5 9 0bar , represents 0 . 2 ~(From . Childress el al., 1970.)

muscle examined by light microscopy following this treatment are no longer PAS positive. Native glycogen, isolated from blowfly flight muscle by a mild aqueous procedure, is a white amorphous powder containing less than 0.1% protein (Childress et al., 1970). On incubation with phosphorylase b, amylo-1, 6-glucosidase, AMP and Pi, the polysaccharide is degraded completely to glucose-1-P (90.5%) and glucose 9.5%). Sedimentation analysis of the pure glycogen also

REGULATION OF INTERMEDIARY METABOLISM

287

shows that it is polydisperse (Fig. 9), with particles having molecular weights as high as 50-100 million. Enzymatic degradation of the glycogen reveals that the outer chains are quite short, only about 25% of the total glucosyl residues being released by phosphorylase alone. Treatment of native glycogen with hot alkali, typical conditions for the extraction of glycogen from tissues, increases the release of the glucosyl residues to 3176, reflecting the availability of

,P

WATM W G H l NUXU UVCWEN

i !y YI

>

I so0 1

I I I I ' i 500 1000 1500 2000 2500 3Ooo

SEDIMENTATION COEFFICIENT

1000 1500 ZOO0 2500 xwx) SEDIMENTATION COmaENT

Fig. 9. The sedimentation coefficient distributions of native and alkali-treated blowfly flight muscle glycogens. (From Childress er al., 1970.)

additional terminal residues which previously were resistant to enzymatic attack, perhaps because of steric hindrance. Chemical alteration of native glycogen by the harsher alkali treatment is also demonstrated by a marked lowering of the sedimentation coefficient (Fig. 9), indicating a five- to 1 0-fold decrease in molecular weight. The significance of using native glycogen in kinetic studies of glycogenolysis is evident from the data in Fig. 10, showing that flight muscle phosphorylase a has a lower affinity for native flight muscle glycogen than for the same substrate treated with alkali or for preparations of glycogen from other species. The apparent K, values are 0.09 and 0.29 mM for the alkali-treated and native glycogens, respectively, in the presence of saturating levels of Pi and AMP. A much greater difference between K, values for the two substrates is observed at low AMP levels. Values for maximum velocity with the native substrate are approximately 50% of the values obtained with alkali-treated or commercial glycogens. In addition, phosphorylase a

288

B. SACKTOR

has differing affinities for Pi depending on which glycogen is present as cosubstrate. In the presence of saturating levels of AMP and glycogen, the apparent Km values for Pi are 4.5 and 9.5 mM, respectively, for alkali-treated and native glycogens. In view of these glycogendependent differences in kinetic properties of the phosphorylases, it is advisable to use native glycogen, extracted from flight muscle tissue, as substrate in experiments on the regulation of glycogenolysis in flight muscle. I

I

I

I

I

I

I

KOH glyc -AMP

04

08

12

16

I

I

I

20

24

28

[GLYCOGEN] mM END GROUP Fig. 10. A comparison of the effects of concentration of native and alkaliextracted glycogens on the initial velocity of flight muscle phosphorylase a. The concentration of Pi was 80 mM and of AMP, when present, was 1 mM. (From Childress et al., 1970.)

3. Glycogen Phosphorylase The glycogen phosphorylases, phosphorylase a and b, from flight muscle of the blowfly have been characterized by the recent studies

REGULATION OF INTERMEDIARY METABOLISM

289

of Childress and Sacktor ( 1970). Glycogen phosphorylase catalyses the reaction: Glycogen,

+ Pi * Glycogen, -, + Glucose- 1-P

It has been calculated that the phosphorylases comprise approximately 1.5% of the total muscle protein and have a potential activity of 9.6 pmoles xmin-' x g-' wet wt of thorax, a value more than adequate to account for the rate of glycogenolysis during flight. Studies of sedimentation velocity and electrophoresis on polyacrylamide gel demonstrate a high degree of homogeneity of the purified enzyme preparations. The SZO value for flight muscle phosphorylase b is 7.4. Estimations of the molecular weight of purified phosphorylase b by sedimentation equilibrium, by gel filtration, and by calculation from sedimentation and diffusion coefficients indicate a value of approximately 100,000. Flight muscle phosphorylase a, purified as the a form or converted from b to a with phosphorylase b kinase and ATP just before sedimentation analysis, also has a SZO of 7.4, demonstrating that flight muscle phosphorylase a has the same molecular weight as does the b form of the enzyme. This is in marked contrast to the rabbit muscle enzyme. Other studies on the purified blowfly enzyme show that phosphorylase b is not dissociated in to smaller molecular species by p-chloromercuribenzoate and that pyridoxal phosphate is the prosthetic group. Table I shows an amino acid analysis of flight muscle phosphorylase b. For comparison, data are also presented on the composition of human, rabbit, and frog muscle phosphorylases. Although the overall compositions are similar, striking differences between the flight muscle enzyme and the vertebrate muscle enzymes are found in the contents of half-cystine, arginine, and lysine residues. The kinetic properties of Phormia flight muscle phosphorylase a and b have been examined in vitro and during flight (Childress and Sacktor, 1960,1970). The interactions between cosubstrates, glycogen and Pi with phosphorylase a are shown in Fig. 11. These double reciprocal plots give a series of straight lines which intersect in the fourth quadrant, suggesting a bimolecular, sequential mechanism in which increasing levels of either substrate enhance the binding of the other. The values for apparent K, for Pi range from 47 mM at 0.15 mM glycogen to 7.3 mM at 1.7 mM glycogen; the values for apparent K, for glycogen range from 6.4 mM at 1.6 mM Pi to 0.72 mM at 40.3 mM Pi. In contrast to phosphorylase a, double

290

B. SACKTOR

TABLE I Amino acid

composition

Amino Acid Aspartic acid Threonine Serine Glutamic acid Proline Alanine Me thionine Isoleucine Leucine Tyrosine Valine Glycine Lysine Histidine Arginine Phenvlalanine Try piophane Half-cystine

of

rabbit, human, phosphoelases

frog, and flight

Rabbita

Human"

Frogb

Phormia

100 35 26 99 37 65 22 51 83 37 63 50 54 25 70 41 13 9

100 34 26 94 35 66 23 49 80 35 60 49 50 25 68 42 12 9

100 37 37 92 47 60 22 47 76 38 63 51 46 21 57 38 14 12

100 35 32 95 38 66 24 53 79 40 52 57 66 16 35 32 10 60

muscle

Compositions are expressed in percentiles relative to the values of Aspartate which were set arbitrarily at 100. Therefore, the differences in amino acid compositions of the four enzymes which are found by this method of calculation are not due to differences in molecular weight or absorbence indices. a Data of Appleman et ul. (1963); data of Metzger et ul. (1968). (From Childress and Sacktor, 1970.)

reciprocal plots of velocity vs. substrate concentration for phosphorylase b are non-linear. However, as in the case of phosphorylase a, increasing levels of one substrate enhance the binding of the other. Although not required for activity, low levels of AMP stimulate phosphorylase a two- to three-fold at saturating levels of substrates and ten-fold or higher at very low substrate levels. The apparent K, for AMP is 0.6 pM at saturating levels of Pi and glycogen. Lowering the level of either substrate decreases the affinity of the enzyme for the activator. Moreover, AMP increases the affinity of phosphorylase a for both substrates. The apparent K, for glycogen is decreased by one-third and that for Pi is lowered from 100 mM to 9 mM in the presence of 10.w and high concentrations of AMP, respectively. Unlike the a form of the enzyme, phosphorylase b has an absolute requirement for AMP. Furthermore, levels of AMP 100-fold greater than those which stimulate phosphorylase a are needed to stimulate

29 1

REGULATION OF INTERMEDIARY METABOLISM

I

1

.-c

2.4

I

I

I

I

I

1,

I

aJ

c

E

2.0

m

E

1

.-=

1

E

1.6

1 -0 43 L E

1.2

0 LL

4 -

0.8

c3 v)

% 0.4 E a

0.0

I

0

I

200

I

400

I

600

Fig. 1 1 . Initial velocity of phosphorylase “a” as a function of substrate concentration at several fixed levels of cosubstrate. The AMP concentration is 1.6 mM. The apparent K, for Pi are, respectively, 47.6, 23.7, 16.3, 11.2 and 7.3 mM in order from the lowest to the , for glycogen are, respectively, 6.4, 4.5, highest glycogen concentration. The apparent K 2.6, 1.7 and 0.7 mM in order from the lowest to the highest Pi concentration. (From Childress and Sacktor, 1970.)

the b form of the enzyme. As with phosphorylase a, however, increasing amounts of AMP lower the apparent K, values for Pi and glycogen, and increasing levels of either substrate lowers the affinity of phosphorylase b for AMP. Childress and Sacktor (1970) have found that ATP is a potent inhibitor of phosphorylase b but not of phosphorylase a. The Ki value for ATP is approximately 2 mM. At infinite concentration of AMP there is little change in Vmax, suggesting competitive inhibition with respect to AMP. Consistent with this view is the fact that ATP causes increases in the apparent K, values for both glycogen and Pi. Other nucleoside triphosphates are less potent as inhibitors while

292

B. SACKTOR

various metabolites, including trehalose, glucose-6-P and arginine-P, at concentrations found in flight muscle, are non-inhibitory.

4. Flight Muscle Phosphorylase a and b in Rest and Flight These detailed kinetic data for the flight muscle phosphorylases coupled with knowledge of the concentrations in resting and working muscle of various metabolites that affect phosphorylase activity T A B L E I1 values for flight muscle Comparison of metabolite levelsU a n d a p p a r e n t K, phosphorylases u n d e r in vivo conditions

Metabolite

in vivo in vi vo Level Level a t rest d u r in g flight

Apparent K, valueb b

a

Potential activity at simulated conditions o f Rest Flight b

a

a

b .__

mM

mM

Glycogen

6-9

Pi

14 0.2 14

6+O.7Sc 15 0.6

AMP

ATP

13

mM 0.6 10 0.5

2(K;)

mM

%

%

%

%

3

64

9

65

1.7

8 0.001 -

Metabolite levels, expressed by Sacktor and Wormser-Shavit (1966) and Sacktor and Hurlbut (1966) as pmoles x g-l wet wt of thorax, were doubled with the assumption that muscle represents half the thorax and the remainder of the thorax does not contribute to the metabolite concentration. The apparent K, values were determined from the kinetic data at the metabolite concentrations which exist in vivo under resting conditions. The level of glycogen in vivo decreased steadily during flight to a value of 0.75 mM after 10 min of flight where it remained steady. A glycogen concentration of 1% corresponds to 5.9 mM end groups. Glycogen data are given in mM end groups. (From Childress and Sacktor, 1970.)

(Sacktor and Wormser-Shavit, 1966; Sacktor and Hurlbut, 1966), enable estimation of the state of the phosphorylases, in vivo. The calculated apparent K, values for cosubstrates and activator of phosphorylase a and b are shown in Table 11. The data indicate that the level of glycogen present is sufficient to saturate the enzymes except after 10 min or more of flight when the muscle glycogen reserve is near depletion. The level of Pi under resting conditions is 14 mM, increasing slightly to 15 mM during flight. In this value, concentration range, which is slightly above the K, phosphorylase activity is quite dependent upon changes in the Pi level. The levels of AMP in vivo are 100-fold greater than the K,

REGULATION OF INTERMEDIARY METABOLISM

293

value for phosphorylase a and it seems reasonable to assume that the enzyme is saturated with AMP at all times. This level of AMP is near the K, value for phosphorylase b and thus the activity of the b form of the enzyme is responsive to changes in AMP level from 0.2 mM at rest to 0.6mM during flight. However, the level of ATP, which is strongly inhibitory to phosphorylase b, is very high in this tissue and decreases only slightly, from 14 mM to 13 mM, upon initiation of flight. Table I1 also shows the predicted relative activities of phosphorylase a and b under conditions of rest and flight. The strong inhibitory effect of ATP on phosphorylase b is most critical. In flight, phosphorylase b retains only 9% of its potential activity, while the a form of the enzyme, unaffected by ATP, retains 65%. Even though the changes in metabolite concentrations which occur during the transition from rest to flight cause a three-fold stimulation of phosphorylase b, the maximal rate of the b form is still far too low to account for the rate of glycogenolysis during flight. Based on the specific activity of 9.6 units of activity x g-’ wet wt of thorax, when assayed with native glycogen at 30°C and the fact that the potential activity of phosphorylase is the same whether it exists in the a or the b form, it can be estimated from the kinetic data that at least 50% of the total phosphorylase must be present in the a form in order to account for the rate of glycogen breakdown that occurs during flight (a minimum of 2.5 pmoles glucosyl residues x min-’ x g-’ wet wt of thorax). Using special precautions t o prevent alterations in the relative TABLE 111 Relative amounts of phosphorylases “a” and “b” in vivo at rest and during flight State

% of total phosphorylase in the “a” form

Resting-Unmounted 17.8f 3.6 (17) Mounted-Rested 2 hr 3 4 . 3 f 8.5 (27) Flown 5 s 63.9 12.3 (16) Flown 15 s 69.1 f 13.4 (16) Flown 30 s 72.2 f 12.3 (10) Flown 6 0 s 71.5 f 14.2 (10) Flown 10 min 72.1 13.1 (10) The figures shown represent mean values and their standard deviations. The number of extracts, each consisting of five thoraces, is shown in parentheses. (From Childress and Sacktor,

* *

1970.)

294

B. SACKTOR

proportions of phosphorylase a and b, Childress and Sacktor (1970) have measured the levels of both forms of the enzyme in flight muscle during the transition from rest to activity (Table 111). An increase in the relative amount of phosphorylase a upon initiation of flight is demonstrated. The level of phosphorylase a reaches its maximum of about 70% at 15 s of flight and remains there during a flight of 10 min. The total amount of phosphorylase present (a plus b ) does not change during flight. This level of phosphorylase a is adequate t o account for the observed rate of glycogenolysis during flight. From the kinetic data obtained under in vitro conditions, a 70% level of phosphorylase a can catalyze a rate of glycogenolysis of 4.3 pmoles glucosyl residues x min-' x g-' wet wt of thorax as compared to a minimum rate of 2.5 actually measured (Sacktor and Wormser-Shavit, 1966).

5. Glycogenolysis in Fat Body As noted previously, glycogen stored in the fat body is used as a major energy depot during flights of long durations (also see review by Sacktor, 1965). In the blowfly, glycogen from the abdomen, presumably from the fat body, is consumed primarily after 10 min of flight and this reserve may support activity for at least an hour of continuous flight (Sacktor and Wormser-Shavit, 1966; Clegg and Evans, 1961). In general, essentially nothing is known of the mechanisms that initiate glycogenolysis in the fat body and knowledge of the biochemical properties of the phosphorylases in this tissue is fragmentary. The phosphorylases in the fat body from various stages of Cynthia and cecropia silkmoths have been partially characterized by Stevenson and Wyatt (1964) and Wiens and Gilbert (1967a). Activation by AMP is found and K, values for glucose-1-P and glycogen are given. It is obvious that, in view of the complicated kinetics of flight muscle phosphorylases, the values reported for fat body phosphorylases are to be interpreted with caution. Nevertheless, it is clear that when fat body glycogen is degraded, glucose does not leave the tissue. Instead, sugars are mobilized into blood as the disaccharide trehalose (Clegg and Evans, 1961). This trehalose, in turn, pools with blood trehalose and is transported to muscle, where it supplies energy for flight. There is reasonable evidence that hormones may enhance glycogenolysis in fat body. Besides changes in phosphorylase activity during development (Wyatt, 1967; Wiens and Gilbert, 1967b), Steele (1961, 1963) and others (McCarthy and Ralph, 1962; Ralph, 1962;

295

REGULATION OF INTERMEDIARY METABOLISM

Bowers and Friedman, 1963) have shown that corpus cardiacum extract stimulates the release of trehalose from isolated fat body and that the chief precursor of the released disaccharide is glycogen. However, the slow response of fat body phosphorylase to the hormone, 2 hr in some cases, makes the role of this hormonal mechanism in flight metabolism debatable. B. PHOSPHORYLASE b KINASE

Childress and Sacktor (1970) have shown that the rapid glycogenolysis induced at the initiation of blowfly flight requires the conversion of phosphorylase b to a. This conversion is catalyzed by the enzyme, phosphorylase b kinase. The mechanisms by which phosphorylase b kinase, itself, becomes activated to transform phosphorylase from the b to the a form, are, therefore, of significance to the regulation of flight muscle metabolism. The kinase is localized with phosphorylase a phosphatase, and phosphorylases a and b in the post-mitochondrial supernatant of the muscle, and can be readily isolated. In some recent experiments with Phormia flight muscle, Hansford and Sacktor (1970b) have found that phosphorylase b kinase is stimulated by Ca2+ at with maximal enhancement at about concentrations > g atoms/litre. These concentrations of Ca2+ are within the physiological range of the cation. An additional increase in phosphorylase b kinase activity is obtained by high concentrations of Pi. Both Pi and Ca2+seemingly function by increasing the Vmax of the enzyme. C. GLYCOGEN SYNTHETASE

Despite the importance of glycogen and its metabolism in flight muscle, little is known of the mechanisms by which the polysaccharide is synthesized. There is evidence that in insect muscle, as in mammalian tissues, the synthesis of the a-l,4-glucosidic linkage in glycogen is a function of UDP-glucose-glycogen transglycosylase (glycogen synthetase): UDP-glucose + Glycogen,

f,

UDP + Glycogen,,

,

The UDP-glucose is generated by UDP-glucose pyrophosphorylase by the reaction: Glucose-1-P + UTP * UDP-glucose + PPi Trivelloni ( 1960) has shown incorporation of 14C-glucose from UDP-glucose into glycogen and release of UDP by extracts of

296

B. SACKTOR

thoracic muscles of locusts. On the other hand, the histochemical findings of Hess and Pearse (1961) are not in full agreement with these data. These authors claim that little glycogen is deposited by the transglycosylase reaction but much glycogen is formed by a reversal of the phosphorylase reaction. However, as pointed out previously (Sacktor, 1 9 6 3 , since the histochemical method is dependent on the visualization of synthesized glycogen, the apparent low activity of the UDP-glucose system may have resulted from the presence of catabolically active phosphorylase which prevents the accumulation of glycogen from UDP-glucose by rapidly breaking down the newly formed polysaccharide. Hess and Pearse (1961) do report low phosphorylase and high UDPglucose-glycogen transglycosylase activities in the leg muscle of the locust, however. The synthesis of glycogen by the glycogen synthetase pathway in fat body of insects is well documented (see Wyatt, 1967). The synthetase is bound t o the particulate glycogen (Murphy and Wyatt, 1965; Vardanis 1967); in fact, with preparations from bee larvae it is not necessary to add glycogen as a primer. Glucose-6-P activates the synthetase, as in mammalian tissues. In cecropia larval fat body (Murphy and Wyatt, 1965), the K, for glucose-6-P is 0.6 mM and the K, for UDP-glucose is 1.6 mM. Glucose-6-P activates without significantly changing the K, for UDPglucose, an effect which resembles that found with the D form of the enzyme in mammalian tissues (Rosell-Perez and Larner, 1964). In the bee (Vardanis, 1967), the K, is slightly less than that in the silkmoth, but glucose-6-P increases approximately 10-fold the affinity of the synthetase for UDP-glucose. Thus, the bee has properties of both the D and I forms of the mammalian enzyme. Vardanis has suggested that the main factor limiting incorporation of glucose into glycogen is the length of outer branch chains in the primer. When the limit is reached, incorporation stops. Addition of glucose-6-P causes incorporation to resume until a higher limit is attained. He reasons that either glucose-6-P changes the specificity of the enzyme for primer outer chains, or that glucose-6-P activates a glucose-6-P dependent form of the enzyme that can utilize longer outer chains as effective primer units. The possible hormonal regulation of glycogen synthesis has been suggested (Van Handel and Lea, 1965; Wyatt, 1967). D. TREHALASE

1. Utilization of Trehalose In addition to glycogen, it is now well established that the

REGULATION OF INTERMEDIARY METABOLISM

297

disaccharide trehalose ( 1-a-D-glucopyranosyl-u-D -glucopyranoside) also supports flight muscle activity. The enzyme trehalase is in flight muscle (Sacktor, 1955). Trehalose, the principal blood sugar in many species of insects (see Wyatt, 1967), is found in muscle, and is reduced in concentration within these loci after flight (Bucher and Klingenberg, 1958; Clegg and Evans, 1961; Sacktor and WormserShavit 1966). A correlation has been observed between the titer of trehalose in the blood and the frequency of wingbeat in the blowfly (Clegg and Evans 1961). Figure 7 describes sequential changes in the concentration of trehalose on initiation of flight and during the steady-state of a continued flight (Sacktor and Wormser-Shavit, 1966). The concentration of disaccharide falls precipitously, approximately 1 pmole x g-I wet wt during the first 5 s, and continues to decrease rapidly for about 30 s. The level decreases progressively but at a lesser rate during the remainder of flight. This rapid initial utilization of trehalose coupled with the relatively slow rate of glycogenolysis on induction of flight suggests that during the first few seconds of flight the catabolism of trehalose probably serves as the major energydonor reaction. It has been pointed out earlier (Sacktor and Wormser-Shavit, 1966), that two kinetically different pools of trehalose are indicated from the discontinuity in the rate at which trehalose is utilized (Fig. 7). One is metabolized at a great rate and becomes exhausted within 30s; the other is depleted gradually during sustained flight. The kinetics of the latter pool resemble closely those of the decrease in the concentration of trehalose in the blood of the thorax during flight. The former pool has been designated as “muscle trehalose”. However, it may represent the trehalose in the extracellular fluids of the tubular system which invaginates the muscle profusely (Fig. 5 ) , and is compartmented from the sugar in the blood. If this interpretation is found to be valid, then all the trehalose is, in fact, extracellular and the enzyme trehalase, which hydrolyzes the disaccharide to its two glucose moieties, functions in the transport of sugar across the cell membrane. The marked decrease in the concentration of trehalose at the onset of exercise is coincident with the rapid and marked increase in the level of glucose. As shown in Fig. 7, the concentration of glucose increases by over 2 pmoles x g-I (wet wt) during the first few seconds of flight. These opposite changes, occurring at a time when there has been considerable enhancement of glycolysis, clearly indicates that the cleavage of trehalose to glucose by trehalase has

298

B. SACKTOR

been facilitated greatly and is controlled. The elevation of the concentration of glucose is transient, however; within 30 s its concentration has returned fo the original level. The changes in the intramuscular concentration of glucose reveal suggestions of overshoots and periodic fluctuations. This indicates that the steady-state levels of metabolites are not reached monotonically but in an oscillatory manner. Such oscillations are indicative of fine adjustments in the regulatory mechanisms of glycolysis in the muscle during exercise.

2. Hydrolysis of Trehalose by Trehalase The enzyme trehalase hydrolytically splits trehalose into two glucose moieties: Trehalose + H, 0 + 2 Glucose Reports on the activity of trehalase in a host of insect species and various tissues are common and these have been ably tabulated by Wyatt (1967). In general, two different trehalases, specific for trehalose, have been characterized and both may be found in the same insect, silkmoth (Gussin and Wyatt, 1965), roach (Gilby et al., 1967), and blowfly (Sacktor, unpublished). One type is represented by the enzyme from intestine; the other has been described from muscle. The gut trehalase is soluble, has a pH optimum in the range 5.0-5.7 and has a K, mostly less than 1 mM.The enzyme from muscle is largely associated with particles, has a pH optimum less acid and has a K, value indicating less affinity for substrate. The particulate thoracic muscle trehalase from roaches, locusts and cecropia silkmoths can be activated several-fold by various physical and chemical treatments, i.e. freezing and thawing, detergents, phospholipase A, that tend to disrupt lipoprotein structure (Zebe and McShan, 1959; Gussin and Wyatt, 1965; Gilby et al.. 1967; Stevenson, 1968). After activation, the pH optimum of the muscle trehalase remains the same as that of the original particulate preparation but the K, is lowered to approximately one-half. It has been suggested (Gussin and Wyatt, 1965) that this activation phenomenon may be related to the biological regulation of the muscle trehalase but further studies from the same laboratory cast doubt on this hypothesis (Gilby et al., 1967). They point out that in the housefly and two species of blowflies, insects which utilize carbohydrate for flight energy, the muscle trehalase is not activated by freezing and thawing. However, their values of trehalase activity

REGULATION OF INTERMEDIARY METABOLISM

299

in Dipteran muscle preparations are already high. In fact, no further increase should have been expected since they are probably measuring the uncontrolled rate. It can be calculated that for Phormia, at least, the activity of trehalase in the muscle homogenate is sufficient to account for the rate of trehalose utilization by the blowfly on initiation of flight (Sacktor and Wormser-Shavit, 1966). The mechanism of regulation of the membrane-bound trehalase in insect muscle remains unknown despite much experimental probing. An important question is the precise cellular localization of the enzyme and to date the fragmentary observations are seemingly inconsistent. The first report of trehalose catabolism in flight muscle (Sacktor, 1955) has established a requirement for both the cytosol and particulate fractions in the complete oxidation of the disaccharide. Since it is now known that some of the glycolytic enzymes are localized exclusively in the cytosol whereas the respiratory chain is in the mitochondria, and that the complete catabolism of trehalose to C 0 2 and H,O involves both these metabolic pathways, the need for the two fractions is understandable and, therefore, these original observations cannot be used as an argument for any specific localization of trehalase. Approximately 25% of the total trehalase is found in the 100,000 x g for 20 or 30 min supernatant fraction of flight muscle from Phormia blowfly (Hansen, 1966), cecropia silkmoth (Gussin and Wyatt, 1965), and Blaberus roach (Gilby et al., 1967). In both cecropia and Blaberus, the post-mitochondria1 “microsomal” fraction contains about half the total enzyme with relatively high specific activities. Appreciable activity remains with a low-speed fraction that contains myofibrils and nuclei, but mitochondria show very low activity. In contrast, Zebe and McShan (1959) and Hansen (1966) claim that mitochondria are the chief locus of muscle trehalase in the roach Leucophaea and blowfly, respectively. However, cellular fractions, other than the myofibrils of Phormia, apparently have not been tested. Wyatt (1967) correctly points out that “microsomes” is an operational term rather than a specific morphologic‘al entity. By the procedures used in his experiments, the “microsomal” fraction contains sarcoplasmic reticulum, ribosomes, glycogen, and perhaps elements of the plasma membrane. It is clear from the electron micrographs shown in Figs 2 and 5 that the plasma membrane, as indicated by the Tsystem, deeply invaginates the muscle. Thus, this membrane makes intimate contact with other cellular components and, in view of the tendency of lipoprotein membranes to adhere to

300

B. SACKTOR

neighbouring particulates, claims as to the localization of enzymes by cosedimentation alone without appropriate controls with marker enzymes and electron microscopic examination of the pellets may be somewhat rash. In preliminary studies with flight muscle of Phormia (Sacktor and Reed, unpublished), substantial trehalase activity is found cosedimenting with mitochondria. However, electronmicrographs of the preparations frequently show what appears to be a third mitochondrial membrane, which suggests the adherence of an extra mitochondrial component. Other studies with marker enzymes clearly show that trehalase is located on or exterior to the outer mitochondrial membrane. If trehalase is, in fact, part of the plasma membrane, the seemingly discrepant observations of Hansen ( 1966) with blowflies and Gussin and Wyatt (1 965) and Gilby et al. (1967) with silkmoths and roaches, respectively, may be reconciled. In the latter insects, which have the synchronous type muscle with a welldeveloped sarcoplasmic reticulum, fragments of the plasma membrane may adhere to the reticulum and appear in the “microsomal” fraction. On the other hand, in the asynchronous muscle of the blowfly the sarcoplasmic reticulum is markedly reduced while the T-system and mitochondria are well-developed. Disruption of cell structure by homogenization, however mild, may cause the plasma membranes to collapse on t o the mitochondria, and trehalase activity will be evident in isolated mitochondria. Such hypothesis for the cellular localization of trehalase is consistent with the proposed function of the enzyme in transporting the sugar into the cell. This transport process will be activated, in a manner yet to be discovered, upon neuroelectrical stimulation of the muscle, depolarization of the plasma membrane and initiation of contraction. E. BIOSYNTHESIS OF TREHALOSE

Clegg and Evans (1961) have suggested that the intensity of flight in Phormia is determined largely by the interaction of two processes: the rate of trehalose utilization by the flight muscle and the rate of trehalose synthesis. Exhaustion results when trehalose cannot be supplied to the muscle at the necessary rate. While this view may be an oversimplification of the mechanisms regulating flight muscle, it is perfectly clear that the synthesis of trehalose is an important component of the overall process. A number of investigators (Treherne, 1958; Winteringham, 1959; Clegg and Evans, 1961; Wyatt, 1967) have shown that 14C-glucose injected into insects is converted to trehalose. Clegg and Evans (1961) introduced labeled

REGULATION OF INTERMEDIARY METABOLISM

301

glucose into blood of adult Phormia, previously starved for 24 hr, and measured the rate of trehalose formation in uivo. Within 30 s, radioactivity is measurable in trehalose of blood. The percentage of radioactivity appearing in blood trehalose increases very sharply in time, reaching about 50% within 2 min and about 90% at 10 min following the injection. Since over 96% of the total injected radioactivity is recovered as either trehalose or glucose, it is apparent that little glycogen is formed under these experimental conditions. Glucose, fructose and mannose are converted to trehalose with equal rapidity, while incorporation of galactose is appreciably slower. This pattern of conversion is similar to that for the rates of oxidation of these monosaccharides by fly flight muscle homogenates (Sacktor, 1955). Candy and Kilby (1 959) and Clements (1959) have shown that biosynthesis of trehalose takes place rapidly in the locust fat body whereas blood, leg muscle, and gut tissues are largely inactive in this respect. The synthesis in fat body has been collaborated by Clegg and Evans (1961) for the blowfly and the woodroach, and by other workers with a variety of insect species. Clegg and Evans (1961) also confirm that I4C-glucose is not incorporated into trehalose by mid-gut or blood of the blowfly, but a limited quantity of disaccharide is synthesized by flight muscle, although the possible contamination of the muscle by the diffuse fat body in the thorax is not ruled out. However, Trivelloni (1960) and Hines and Smith (1963) have detected trehalose synthesis in locust muscle. Clegg and Evans ( 1961) have demonstrated that the trehalose that is rapidly synthesized by isolated fat body is immediately released to the incubation medium. One may conclude from these data that the primary site of trehalose synthesis is the fat body and that the rate of disaccharide formation and liberation by this tissue is sufficiently rapid for it to be the source of most of the trehalose that accumulates in blood and is utilized in flight. The enzymes catalyzing the conversion of glucose to trehalose in the fat body have been described by Candy and Kilby ( 196 1 ) and are identical with those leading to the formation of the disaccharide in yeast, as previously discovered by Cabib and Leloir (1958). The pathway from glucose to UDP-glucose is common to the synthesis of both trehalose and glycogen. In the biosynthesis of trehalose, UDP-glucose serves as donor of one of the glucose moieties. The other hexose moiety stems from glucose-6-P, which, upon condensation with the nucleotide in a reaction catalyzed by the

302

B. SACKTOR

enzyme, trehalose-6-P synthetase, forms trehalose-6-P: UDPglucose

+ Glucose-6-P -,Trehalose-6-P + UDP

This is followed by the dephosphorylation of trehalose-6-P by a specific phosphatase (Friedman, 1960): Trehalose-6-P + H, 0 + Trehalose + Pi Murphy and Wyatt (1965) have studied the kinetics of the synthesis of trehalose in fat body of cecropia larvae. The K, for UDPglucose is 0.3 mM. Glucose-6-P shows more complex kinetics, a plot of glucose-6-P concentration against velocity yields a sigmoid curve with half-maximal velocity at about 5 mM. High concentrations of glucose-6-P are inhibitory. Mgz+ enhances the binding of glucose-6-P. These allosteric properties (Monod et al., 1963, 1965) are lost when the enzyme is partially purified or mildly maltreated. After such treatment the enzyme has classical Michaelis-Menton kinetics towards glucose-6-P with a K, of 5 mM. Significantly, trehalose-6-P synthetase is strongly inhibited by trehalose. The extent of inhibition varies not only with the trehalose concentration but also with the levels of glucose-6-P and Mgz+.Trehalose decreases the affinity of the enzyme for glucose-6-P and is non-competitive with respect to UDP-glucose. Treatments that cause loss of sigmoid kinetics towards glucose-6-P also reduce the sensitivity of the enzyme to trehalose. This feedback inhibition of trehalose-6-P synthetase by trehalose may be important to the mechanism of regulation of blood trehalose. Murphy and Wyatt (1965) suggest that elevation in glucose (by ingestion or other processes) causes a rise in glucose-6-P, and this will activate both glycogen and trehalose-6-P synthetase, the former reaching saturation (K, = 0.6 mM) before the latter (K, = 5 mM). However, trehalose-6-P synthetase has a greater affinity for UDP-glucose, K, = 0.3 mM as compared to 1.6 mM for glycogen synthetase. This will enable the preferential synthesis of trehalose when UDPglucose levels are low. When trehalose accumulates sufficiently to inhibit trehalose-6-P synthetase, the UDP-glucose level rises and allows increased synthesis of glycogen. This mechanism can provide for rapid production of both trehalose and glycogen when glucose is increased, followed by readjustments to homeostasis in the titer of blood trehalose. The scheme also provides a rational for the increased synthesis of trehalose after starvation (Clegg and Evans, 1961), when the blood trehalose titer is low; and the decreased

REGULATION OF INTERMEDIARY METABOLISM

303

synthesis of disaccharide induced by corpus cardiacum hormone in well-fed flies (Friedman, 1967), when a high level of blood trehalose is expected. On the other hand, Friedman (1968) has suggested an alternative mechanism which may partially explain the action of trehalose on trehalose-6-P synthetase. In addition to inhibiting the synthetase by decreasing the affinity of the enzyme for glucose-6-P (Murphy and Wyatt, 1965), Friedman has found that the rate of GLYCOGEN

G LY C 0 L Y T l C EN0-PRODUCTS Fig. 12. A simplified schematic representation of the interactions of hexoses, trehalose and glycogen in the regulation of carbohydrate metabolism. A = Hexokinase. Inhibited by glucose-6-P. B = Glucose-6-P phosphatase. Activated by trehalose. C = Phosphoglutomutase. D = UDP-glucose pyrophosphorylase. E. = Glycogen synthetase. Activated by glucose-6-P. F = Glycogen phosphorylase. Inhibited by ATP. G = Trehalose-6-P synthetase. Inhibited by trehalose. H = Trehalose-6-P phosphatase. I = Trehalase. Control mechanism as. yet unknown. J = Glycolysis. Control at phosphofructokinase.

glucose-6-P phosphatase is increased greatly and specifically by trehalose. Thus, the accumulation of trehalose may result in feedback inhibition by removing a necessary substrate for trehalose-6-P synthetase without affecting the synthesis of glycogen. It is evident that hexoses, trehalose and glycogen mutually influence one another and that their cross-reactions are both numerous and complex. A simplified schematic representation of some of these interactions is illustrated in Fig. 12. F. GLYCOLYSIS

I . General Considerations The principal pathway for catabolism of carbohydrate in muscle

304

B. SACKTOR

of insects is unquestionably a variant of the classical EmbdenMeyerhof glycolytic scheme. An alternative route, designated by different authors as the ‘tpentose-P pathway”, “hexose monophosphate oxidation shunt”, or the “Warburg-Dickens pathway”, although prominent in some tissues during the life-cycle of the insect and present in muscle (Sacktor, 1965; Chefurka, 1965), probably contributes little to the energetics of the muscle. VogeTl et al. (1959) have reported that the activities of glucose-6-P dehydrogenase in muscle from wing and leg of locusts are only 0.1% of those of enzymes in glycolysis and Agosin et al. (1961) have noted that in thoracic extracts of the bug, Triatoma, 6-phosphogluconate dehydrogenase is even less active than is glucose-6-P dehydrogenase. Additional facets pertaining to the significance of the hexose monophosphate oxidative shunt in muscle as well as a description of the reaction sequences are to be found in previous reviews (Sacktor, 1965, Chefurka, 1965). The early studies of glycolysis in insects and details of the individual enzymatic reactions have also been thoroughly discussed (Sacktor, 1965; Chefurka, 1965) and no attempt will be made here to characterize each step in the pathway. Rather, the overall system and its regulation in muscle will be considered. It is now quite clear that flight muscle of a variety of insect species have a very high activity of a-glycero-P dehydrogenase and an extremely low activity of lactate dehydrogenase. Thus carbohydrates, whether originating as trehalose, glycogen on monosaccharides, yield stoichiometrically pyruvate and a-glycero-P as the products of glycolysis (Fig. 13). Comparative studies by several investigators (see Table I of Sacktor, 1965; Brosemer, 1967; Brosemer and Marquardt, 1966) on a-glycero-P and lactate dehydrogenase activities in different kinds of muscle reveal a reciprocal correlation between these activities. The highest ratio of a-glycero-P to lactate dehydrogenase activities is found in flight muscle of Dipterous and Hymenopterous insects. In this respect, the contrast between flight and leg muscle of a given species, between flight muscle of insects that are active flyers and the same muscle of related species that have lost during evolutionary development the ability to fly, and between some insect and mammalian muscle, is particularly striking. The significance of the formation of pyruvate and a-glycero-P in glycolysis becomes evident in light of the a-glycero-P cycle (Bucher and Klingenberg, 1958; Estabrook and Sacktor, 1958a; Sacktor and Dick, 1962). The cycle consists of two reactions:

REGULATION OF INTERMEDIARY METABOLISM

3 05

NADH(DPNH) + H + + Dihydroxyacetone-P + NAD+(DPN+) + a-glycero-P a-glycero-P + 1 / 2 O2+ Dihydroxyacetone-P + H20 NADH(DPNH) + H + + 1/ 2 O2-+ NAD+(DPN') + H20 Sum :

As shown in Fig. 13, the a-glycero-P cycle provides a mechanism whereby NADH, which is formed extramitochondrially in the glyceraldehyde-3-P dehydrogenase reaction, becomes oxidized. With the virtual absence of both lactate dehydrogenase and a direct mitochondrial oxidation of extramitochondrial NADH, because mitochondria are impermeable to NADH, glycolytically formed NADH is reoxidized, concomitant with the reduction of dihydroxyacetone-P and formation of a-glycero-P, by the extremely active soluble a-glycero-P dehydrogenase of the cystosol. The a-glycero-P is oxidized, in turn, by the mitochondrial a-glycero-P dehydrogenase, a flavoprotein, thereby regenerating additional GLUCOSE

ATp4

AD P GLU-6-P

3.

FRU-6-P ADP FRU-1,6-diP

1GLYC-3-P

4

NADH

* OHAP

NAO'

f _ aGP

a GP CYCLE

1

MlTOCH 0 N 0 RIAL

t

CITRIC ACID CYCLE

D ( u GP OEHYDROGENASE)

Fig. 13. Schematic diagram of glycolysis and the aglycero-P cycle in flight muscle.

306

B. SACKTOR

dihydroxyacetone-P. This dihydroxyacetone-P is then available for further oxidation of extramitochondrial NADH. Accordingly, the cycle is a shuttle system, in which NAD-linked substrates, in reduced and oxidized states, respectively enter and leave the mitochondria. In this way, hydrogen or reducing equivalents from the extramitochondrial pool of NADH pass the cytosol-mitochondria1 barrier and are oxidized by the mitochondria1 respiratory chain. Further, the cyclic process is self-generating in that only a catalytic quantity of dihydroxyacetone-P is needed to oxidize the NADH being continuously formed (Sacktor and Dick, 1962). This suggests that most of the dihydroxyacetone-P that is produced by the aldolase reaction can be isomerized to glyceraldehyde-3-P and that essentially all of the carbon of the carbohydrate metabolized in a prolonged flight is convertible to pyruvate and is available for further oxidation via the Krebs citric acid cycle. Thus the system is remarkably efficient, in that end-products of glycolysis, such as lacate, need not accumulate wastefully as it does in exercising vertebrate muscle. The oxidation of the two end-products of glycolysis, pyruvate and aglycero-P, by flight muscle mitochondria will be examined in a subsequent section. In considering glycolysis in muscle, the usable chemical energy derivable from the reactions is central to the discussion. The net yield of productive chemical energy from glycolysis may be computed from a balance of the moles of ATP consumed and regenerated per mole of carbohydrate degraded to end-products. As illustrated in Figs 12 and 13, when the substrate is glucose, which becomes available to muscle either as the result of the action of trehalase on trehalose or as free hexose, 1 mole of ATP is consumed at each of two kinase steps, the phosphorylations of glucose and of fructose-6-P. Since in anaerobic glycolysis in insect flight muscle the chief end-products are 1 mole each of pyruvate and a-glycero-P from each mole of glycosyl residue, the highenergy phosphate transferred to ADP to generate ATP is different from that typically described for mammalian tissues. The anaerobic formation of a-glycero-P is not concomitant with synthesis of ATP. In the formation of pyruvate, one equivalent of ATP results from the conversion of 1,3-diPglycerate to 3-Pglycerate and another is generated when P-enolpyruvate is transformed to pyruvate. Thus, per mole of glucose glycolyzed, 2 moles of ATP are used and 2 moles are regenerated, with no net gain of ATP. When starting with glycogen as the substrate instead of glucose, the synthesis of glucose-6-P is achieved

307

REGULATION OF INTERMEDIARY METABOLISM

without involvement of ATP. Hence, per mole of glucose residue glycolyzed, only 1 mole of ATP is utilized (the fructose-6-P kinase reaction), and the net gain in ATP is 1 mole. Thus, anaerobic glycolysis in insect muscle is extremely inefficient with respect to conservation of energy in the form of ATP and the need for an essentially completely aerobic metabolism is now fully appreciated. Indeed, the modifications, both anatomical and biochemical, that have evolved in the rapidly flying insect are largely those which provide for an efficient aerobic metabolism.

2.Control of Glycolysis From concurrent measurements of the concentrations of glycolytic intermediates in flight muscle of the blowfly at rest and after periods of induced flight, the enzymatic reaction regulating glycolysis has been identified (Sacktor and Wormser-Shavit, 1966). On initiation of flight, the concentration of glucose-6-P decreases and that of fructose-6-P remains essentially constant. In contrast, the 4.

-

.30

.

.

L

3 + Y 3

J

I

2.

.

.

1-

I5

DHAP

I

I

30

45

,

-k

1 I

I

.

”

,o

’ 1

2

w 2

0

H

a

‘2

d-GP

8

4

L 0

15

60

0

I

I

15

30

45

60

MINUTES Fig. 14. The concentrations of fructose-1, 6 d i P (FDP), dihydroxyacetoneP (DHAP), aglycero-P (aGP), and pyruvate (Pyr) in thoraces of blowflies during flight. (From Sacktor and Wormser-Shavit, 1966.)

308

B. SACKTOR

concentration of fructose-l,6diP increases during the first 5 s of the start of contraction (Fig. 14). The level of fructose-l,6-diP in the muscle reaches a peak at 15 s. The concentration of the hexose diphosphate returns to the resting level by the second minute of flight and from 10 to 60 min of the flight is maintained at a value slightly less than that measured initially. The slight decrease in the concentration of the hexose monophosphates coincident with a rapid accumulation of fructose- 1,6diP under conditions of maximal glycolytic flux brought about by initiation of active contraction identifies an additional crossover or control point, i.e. the phosphofructokinase reaction. The mechanism whereby the phosphofructokinase reaction is facilitated in the transition from a slowly metabolizing resting muscle to an intensely active working muscle becomes evident from an examination of changes in concentrations of the adenine nucleotides and Pi in flight muscle at the initiation of flight (Sacktor and Hurlbut, 1966). In earlier in vitro studies with mammalian enzymes, Lardy and Parks (1956) have discovered that, although ATP is a substrate for the reaction, excess ATP is inhibitory to phosphofructokinase. Passonneau and Lowry ( 1962) have found that the inhibition by ATP may be overcome by either ADP, AMP, Pi, 3’,5kyclic AMP, fructose-l,6-diP or, more effectively, by a combination of these activators. Lowry et al. (1964) postulate that whenever formation of ATP does not keep up with use of ATP, however slight, then Pi, ADP, and, particularly, AMP will increase and that this combination enhances phosphofructokinase activity autocatalytically. As shown in Fig. 15, on initiation of flight the concentration of ATP in blowfly flight muscle decreases whereas the levels of Pi, ADP and, especially, AMP increase. These changes, as flight begins, are in complete accord with their theory and extend the hypothesis to a working muscle in vivo. Recent studies by Grasso and Miglioro Natalizi (1968) with roach leg muscle confirm the effects of ATP and other phosphates on the phosphofructokinase in an insect tissue. Thus, their findings are consistent with the previously suggested mechanism for regulation of phosphofructokinase in insects. The pattern of changes in the concentrations of the other glycolytic intermediates in flight muscle during flight does not reveal any other locus of control of glycolysis (Sacktor and WormserShavit, 1966). The concentrations of dihydroxyacetone-P and 3-Pglycerate reflect the changes in the concentration of

REGULATION OF INTERMEDIARY METABOLISM

309

7 .

t

6 -

ATP

.3

. I.

5 . 4.

2 -

*

AMP

3 -

2.r-

..

ADP

1 -

I c

8-

Pi

1-

6 -

6 -

5-

5-

4 -

4 -

3 -

3

2-

2

1 -

1 -

A RG I NI NE-P

LL

Fig. 15. The sequential changes in the concentrations of ATP, ADP, AMP and Pi in flight muscle ofPhomioduring a 1 hr flight. (From Sacktor and Hurlbut, 1966.)

fructose-l,6diP although neither increases as much as that of the hexose diphosphate. The concentrations of glyceraldehyde3-P, 2-P-glycerate and 2-Penolpyruvate are low initially and do not change significantly during prolonged muscular work (Sacktor and Wormser-Shavit, 1966). A purified hexokinase from honeybee flight muscle is inhibited by its products, glucose-6-P and ADP (Ruiz-Amil, 1962), but the significance of this inhibition in regulation of glycolysis in the working muscle is uncertain. Interestingly, the concentration of glucose-6-P in blowfly flight muscle at rest greatly exceeds the value observed after exercise (Fig. 7) and the ratio of glucose-6-P to fructose-6-P approaches the expected value of 3 : 1 during continuous flight, suggesting that the isomerase reaction is not at equilibrium in the resting muscle. A.1.P.-11

310

B. SACKTOR

G . IDENTIFICATION OF OTHER LOCI OF CONTROL OF METABOLISM

Sequential analyses of metabolites in flight muscle during flight have uncovered three additional points of regulation, the oxidations of a-glycero-P, pyruvate and proline. As noted above, aglycero-P and pyruvate are the end-products of glycolysis and they, as well as proline, are metabolized further by mitochondrial oxidations. The evidence on the identification and significance of these control sites will be presented in this section, whereas the mechanisms of the regulation will be discussed later in conjunction with a general exposition of regulation of oxidative activity in mitochondria.

1 .a-Glycero-P Oxidation Flight muscle of Phormia contains an exceptionally high concentration of aglycero-P, about 1.5 pmoles x g- 1 wet wt of thorax, a value 10-100 times those of other phosphorylated glycolytic intermediates (Fig. 14, Sacktor and Wormser-Shavit, 1966) and approximating the apparent K, of the mitochondrial dehydrogenase (Estabrook and Sacktor, 1958a). Significantly, the concentration of a-glycero-P does not change during flight. This is in contrast with the findings of a large accumulation of a-glycero-P during anaerobic glycolysis in insect muscle in vitro (Kubista, 1957; Chefurka, 1958; Heslop et al., 1963). It indicates that glycolysis in insect flight muscle during prolonged and continuous work is completely aerobic, the a-glycero-P produced concomitant with oxidation of glycolytically formed NADH being immediately oxidized by the extraordinarily active mitochondrial a-glycero-P dehydrogenase. Since, as pointed out previously, only a catalytic quantity of the regenerated dihydroxyacetone-P is needed in the a-glycero-P cycle, the excess triosephosphate is metabolized to pyruvate. Thus, these in vivo data demonstrate that essentially two equivalents of pyruvate are formed from each mole of hexose, verifying the concept and function of the a-glycero-P cycle in the intact organism. 2. Pyruvate Oxidation The concentration of pyruvate increases strikingly on initiation of flight (Fig. 14), doubling within the first few seconds and reaching a maximal value in 15 s. Thereafter, this concentration decreases rapidly so that by the second minute a steady-state level only slightly higher than that seen at rest is achieved. Coincident with the onset of active contraction and with the substantial rise in the concentration

REGULATION OF INTERMEDIARY METABOLISM

311

of pyruvate, there is an enormous accumulation of alanine in the flight muscle (Fig. 16, Sacktor and Wormser-Shavit, 1966). About 1 pmole/g (wet wt of thorax) is formed within 5 s. The concentration of alanine continues to increase after 15 s, the time at which the concentration of pyruvate reaches its maximum. A maximal concentration of alanine is measured after 5 min of flight. Polecek and Kubista (1960) have also noted an increased concentration of pyruvate within the muscle of the roach after a flight of 1-3 min and Kirsten et al. (1 963) have found elevated concentrations of pyruvate and alanine after a 20 s flight of the locust. In addition to the formation of alanine after the onset of flight, pyruvate is also converted to acetyl carnitine (Childress et al., 1967). A four-fold increase in acetyl carnitine parallels the four-fold increase in pyruvate. After about 1 min of flight the levels of both acetyl carnitine and pyruvate decrease, with acetyl carnitine attaining a steady-state concentration of about twice that in the muscle at rest. The increase in concentration of pyruvate and accumulations of both alanine and acetyl carnitine in the transition from rest to flight suggest that, on the initiation of flight, pyruvate is not metabolized in the citric acid cycle as fast as it is formed by glycolysis. This indicates that there is a limitation in the oxidation of pyruvate that is relieved shortly after the onset of flight. The nature of this control will be considered subsequently.

3.Pro line Oxidation The large accumulation of alanine with the induction of contraction has prompted a search for potential sources of the amino group (Sacktor and Wormser-Shavit, 1966). As shown in Fig. 16, the increase in alanine approximates the decrease in proline, which is present in resting muscle at a remarkable concentration of over 6 pmoles x g-1 wet wt of thorax. A slight decrease is found in the concentration of glutamate, while no consistent or significant changes are observed in the concentrations of aspartate, free NH;, total amide, glutamine, or asparagine within the muscle on initiation of active contraction or during prolonged flight. The coincident and stoichiometric relationship between the formation of alanine and the utilization of proline suggests that the amino moiety of alanine is derived mostly from the large store of free proline. Bursell (1963) has also reported a decrease in the concentration of proline during short flights of the tsetse fly and has suggested that the amino acid is of significance as an energy reserve in this pest because of the

3 12

B. SACKTOR

50

25

-

c

ALANINE

3 + Y

S

J

1

1

-

2t

:

25

0

2

4

Y

- 21

3

2

a

2

.

-%0

,

I

1

2

30

.

.6

* .3

-

9-

MALATE

.6 0

15

30

45

60

45

..

GLUTAMATE

5

15

60

1

1

I

0

15

30

,o

45

1

2 60

MINUTES Fig. 16. The sequential changes in the concentrations of alanine, proline, glutamate and malate in flight muscle of Phormiu during a 1 hr f l a t . (From Sacktor and Wormser-Shavit, 1966.)

blood-sucking habit of the latter as well as its lack of carbohydrate stores. The triggering of the rapid decrease in the concentration of proline in flight muscle of the blowfly on initiation of flight demonstrates the sixth enzymatic site of regulation in the rest-to-flight transition. IV. REGULATION OF FAT METABOLISM

As in the present discussion of carbohydrate metabolism, this examination of the regulation of fat metabolism will not include all the details of the biosynthetic and degradative pathways nor the multitude of data on various aspects of the biochemistry of lipids. These topics have been dealt with in earlier reports by Gilby (1 965), Tietz (1965), and Sacktor (1965) and more recently in the comprehensive review by Gilbert ( 1967a). The present discussion is limited, in part, by the author’s interests and, therefore, will treat topics related to flight muscle metabolism more extensively than subjects not necessarily of less importance.

REGULATION OF INTERMEDIARY METABOLISM

313

A. FATTY ACID CATABOLISM

1. Utilization of Fatty Acids As described in a previous section, the fact that some insects deplete their reserves of fat during sustained flight indicates that fat can serve as a metabolic fuel for muscular contraction. In these cases, the importance of lipid should not be underestimated. Lipids are the most concentrated source of energy, yielding per gram over twice as many calories as d o carbohydrates or proteins. However, it should be reemphasized, as pointed out earlier (Sacktor, 1965), that locusts and roaches, species which utilize fats in a sustained flight, will first consume their carbohydrates (Bucher and Klingenberg, 1965 ; Hofmanova et al., 1966). Furthermore, the conclusion that moths exclusively utilize fats for flight (Zebe, 1954; Domroese and Gilbert, 1964) is not indubitable. The RQ of a prolonged flight is not indicative of the metabolic events at the beginning of flight; indeed, the RQ of flying moths fed glucose, although low, is higher than those of individuals not fed (Zebe, 1954). Moreover, the observations that added glucose fails to increase the endogenous rate of 0, uptake in muscle homogenates and that the relative rates of 14C02 produced from labeled glucose and pyruvate are low (Domroese and Gilbert, 1954) may not indicate the full potential of the muscle because there is no indication that the conditions are optimal for these substrates in these experiments. In fact, Stevenson (1968a) has shown that flight muscle homogenates of the Southern armyworm moth, Prodenia, can completely and rapidly oxidize glucose, trehalose and glycogen and that mitochondria isolated from this muscle metabolize pyruvate plus malate at a rate comparable to that measured with mitochondria from flies. The glycogen content of Prodenia is relatively low, but it is sufficient to support flight for about 8 min. On the other hand, the ability of fly mitochondria to oxidize fatty acids is extremely limited (Sacktor, 1955; Childress et al., 1967), although Gregory et al. (1968) claim that in the development of the flight muscle in the pupa of the blowfly, Lucilia, fatty acids are utilized significantly for the synthesis of ATP. Thus, the distinction between insects that supposedly use only fats and those that utilize only carbohydrates is even more equivocal than has been suggested previously (Sacktor, 1965). 2. Oxidation of Fatty Acids in Muscle Based on known values of 0, uptake in locusts during flight, Beenakkers ( 1965) has calculated that the flight muscle will consume

314

B. SACKTOR

fatty acids at a rate of 2 3 mg x g-' wet wt of muscle x hr-'. In vitro studies by Meyer et al. (1960), Domroese and Gilbert (1964) and Bode and Klingenberg (1-964) show the utilization of fatty acids by flight muscle at a rate only a fraction of this calculated in vivo rate. However, with the work of Stevenson (1 966, 1968b) using isolated flight muscle mitochondria from Prodenia, rates of respiration greater than 700 pl O2 x mg-' mitochondrial protein x hr-' are reported, and these values approach the required in vivo rates. The requirement for ATP, MgZ+,CoA and a member of the Krebs cycle (for priming) for maximal rates of oxidation of fatty acid by flight muscle preparations (Meyer et al., 1960; Domroese and Gilbert, 1964), suggests that in insects the activation of the fatty acid to its acyl CoA derivative and, in turn, its catabolism via the poxidation pathway of fatty acids is the same as that established for vertebrate systems. Although the overall process of P-oxidation has yet to be demonstrated in insects, the findings of 0-ketoacylthiolase and P-hydroxyacyl dehydrogenase in locust flight and leg muscle (Zebe, 1960; Beenakkers, 1963a, b) as well as fatty acyl-CoA synthetase in moths (Domroese and Gilbert, 1964; Stevenson, 1968b), strongly support the presence of the entire sequence of reactions in muscle of insects. The enzymes are localized largely in the mitochondria of the flight muscle (Beenakkers, 1963a; Stevenson, 1968b). Successive repetition of the P-oxidation cycle results in the complete degradation of even-numbered fatty acids to acetyl CoA. The acetyl CoA generated by degradation of fatty acids pools with acetyl CoA arising from the oxidative decarboxylation of pyruvate, derived largely from glycolysis. The fate of acetyl CoA, upon its entry into the citric acid cycle, will be considered later. B. THE ROLE OF CARNITINE

In pioneering experiments, Friedman and Fraenkel (1 955) have shown that extracts of mammalian tissues mediate the reversible acyl transfer between CoA and carnitine. The physiological significance of this reaction becomes apparent with the finding of Fritz (1 955) that carnitine stimulates the rate of oxidation of fatty acids. This and other studies have led to the hypothesis (Fritz and Marquis, 1965) that fatty acyl CoA thioesters d o not readily penetrate mitochondrial membranes, whereas fatty acyl carnitine esters do, and that the formation of carnitine esters by acyl transferases effects the translocation of fatty acyl groups to the site of fatty acid oxidation. In accord with this view, Beenakkers (1963b) and Bode and

REGULATION OF INTERMEDIARY METABOLISM

315

Klingenberg (1964) have found that added carnitine markedly stimulates the oxidation of fatty acids in locust flight muscle and that fatty acids supplied as acyl carnitine esters are metabolized at even greater rates. The requirement for carnitine for oxidation of fatty acids is strongly correlated with the presence in the muscle of active carnitine-acetyl and -palmityl transferases (Beenakkers and Klingenberg, 1964; Beenakkers et al., 1967). Particularly striking are the differences in transacetylase activity between the flight muscle of two insects, those of the locust, which oxidize fatty acids, and those of the bee, which utilize only carbohydrates in flight. The enzyme is absent from flight muscle of the bee, whereas it is very active in locust flight muscle. Conflicting with this scheme, however, are the surprising observations that flight muscle of two species of moths oxidize palmitate vigorously without added carnitine and that carnitine palmityl transferase cannot be detected in the muscles (Stevenson, 1966, 1968b). Interestingly, the flight muscle of the blowfly, which like the bee is deficient in fatty acid oxidase and has only a negligible capacity to oxidize palmityl carnitine, has a high content of carnitine and a very active acetyl carnitine transferase (Childress et al., 1967). It has been found in the fly, but not in the bee, that carnitine affects carbohydrate utilization, via a role in pyruvate metabolism. The acetyl carnitine transferase in mitochondria from flight muscle of Phormia catalyzes the synthesis of acetyl carnitine from carnitine and acetyl CoA, derived from pyruvate. Formation of acetyl carnitine has been demonstrated both in vitro and in vivo; on initiation of flight its concentration in flight muscle increases four-fold, paralleling the increase in pyruvate (Childress et al., 1967). Approximately 90% of the acetyl carnitine transferase in flight muscle of Phormia is found in the mitochondria (Childress et al., 1967). Exogenous acetyl CoA, in the presence of carnitine, is not oxidized by mitochondria, although acetyl carnitine is oxidized with a QO, of over 300. This indicates that the blowfly mitochondrial inner membrane is not permeable to the thioester and that the mitochondrial carnitine acetyl transferase does not transfer acetyl groups from extramitochondrial acetyl CoA to carnitine and, thus, into the mitochondrial matrix. Instead, the evidence suggests that the mitochondrial enzyme mediates the transfer of acetyl groups out of the mitochondria. On the other hand, the apparent presence of about 10% of the carnitine acetyl transferase activity in the extramitochondrial fraction of the muscle may permit the extramitochondrial

316

B. SACKTOR

acetylation of carnitine, with subsequent transport of the acetyl carnitine into the mitochondria. In contrast to the situation in blowfly mitochondria, in both locust (Beenakkers and Henderson, 1967) and moth (Stevenson, 1968b) flight muscle, exogenous acetyl CoA plus carnitine as well as acetyl carnitine, but not acetyl CoA alone, are oxidized at appreciable rates. From this, it has been inferred that there are two pools of acetyl carnitine transferase; one between the outer and inner membrane, the other within the cristae space. The formation of acetyl carnitine in the blowfly at the start of flight raises the question as to its physiological significance. It has been shown, as discussed in a previous section, that during the initial phase of flight pyruvate is generated faster than it is utilized via the Krebs cycle. Childress et al. (1967) have offered several possible ways in which the formation of acetyl carnitine during this critical period can be of advantage t o the blowfly. These possibilities include: (1) lowers the acetyl CoA : CoA ratio, alleviating inhibition of pyruvate decarboxylase; (2) makes free CoA available for the oxidation of a-ketoglutarate t o succinate, providing oxaloacetate for citrate synthesis; (3) increases the energy production from glycolysis from 6 to 12 moles of ATP per mole of glucose; and (4) provides a readily oxidizable substrate (acetyl carnitine) rather than alanine during the period prior t o the maximal activation of pyruvate oxidation. A more comprehensive discussion of these alternative mechanisms has been presented earlier (Childress et al., 1967). C. BIOSYNTHESIS OF FAT

Data on the mechanisms for biosynthesis of fats in insects are exceedingly limited. In general, the existing knowledge tends to suggest that insects synthesize fat by pathways similar to those which operate in microbial and mammalian systems. Acetate, injected into insects, is readily transformed into fats (Robbins et al., 1960; Louloudes et al., 196 1;Sedee, 1961 ; Sridhara and Bhat, 1964; and Bade, 1964). Labeled acetate is incorporated primarily into C I 6 C l 8 fatty acids, both saturated and unsaturated. Strong (1963) has reported the conversion of U-l"C-glucose to fats in aphids; the pattern of fatty acids formed from glucose is not significantly different from that produced with acetate as precursor. Experiments with specifically labeled glucose indicate synthesis of fatty acid via pyruvate, decarboxylation of pyruvate and incorporation of 2-carbon units into lipids (Horie et al., 1968).

REGULATION OF INTERMEDIARY METABOLISM

317

Interestingly, Van Handel and Lum (1 96 1) have found that female mosquitoes, but not houseflies nor male mosquitoes, can synthesize fatty acids from glucose. Locust fat body, in vitro, converts the carbon moiety of amino acids into fat (Clements, 1959). For the most part, the fatty acids synthesized from acetate and glucose are found combined as triglycerides and phospholipids (Tietz, 196 1). Clements (1959) and Zebe and Mcshan (1959b) have demonstrated the in vitro biosynthesis of fatty acids using isolated, intact fat body and cell-free homogenates of fat body, respectively. Primarily, saturated long-chain fatty acids are formed, although Bade and Clayton (1963) have shown that stearate is rapidly desaturated to oleate. The cell-free system requires or is stimulated by ATP, Mg2+,glutathione, bicarbonate, malonate, CoA, an intermediate of the Krebs cycle, and NADP (Zebe and McShan, 1959b; Tietz, 1961). Under optimal conditions, the particle-free supernant (20,000 x g for 20 min) is as active as the whole homogenate in incorporating acetate into fatty acids. Preincubation with avidin completely inhibits fatty acid synthesis. The inhibition is reversed by biotin. The stimulatory effect of NADP on production of fatty acids strongly suggests a role for NADPH in reductive synthesis. An active hexose monophosphate pathway, which yields NADPH in the oxidations of glucose-6-P and 6-P-gluconate, has been discovered in fat body of silkworms (Horie et ul., 1968). Moreover, the rate of NADPH oxidation in this tissue is remarkably dependent on both malonyl CoA and acetyl CoA (Horie, 1968). These results, although fragmentary at best, do support the view that insects synthesize fatty acids in a manner similar to those of mammals and microorganisms. A generalized scheme for fatty acid biosynthesis has been described in the recent review of Gilbert (1 967a). Zebe and McShan (1959b) have demonstrated that, although fat body is a major site of fat formation, synthesis of fatty acids can take place in muscle. Glucose is incorporated into fatty acids in muscle at a rate approximately 20% that in fat body. Mitochondria from Drosophilu larvae are also capable of de novo synthesis of fatty acids (Goldin and Keith, 1968). Acetate is incorporated into even-number, saturated and monoenoic, fatty acids from 12-18 carbons in chain length. Antimycin A, an inhibitor of the mitochondrial respiratory chain, lowers the rate of incorporation of acetate but the rate can be partially restored by ATP. Labeled acetate appears in free fatty acids, neutral lipids, as well as phospholipids. It is not known whether this mitochondria1 system for fatty acid A.1.P.-1 1

318

B. SACKTOR

biosynthesis differs from that found in the soluble fraction of the fat body (Tietz, 1961). The mechanism by which fatty acids are incorporated into glycerides in insects is essentially unknown. Isolated fat body from cecropia, roach and locust incorporate 14C-palmitate into diglyceride and triglyceride (Chino and Gilbert, 1965; Tietz, 1962, 1967). Incorporation into phospholipid is negligible. The total amount of triglyceride is approximately 5 0 times greater than that of the diglyceride, so that most of the label appears in the triglyceride fraction. However, the specific activity of the diglyceride is at least 50 times greater than that of the triglyceride (Tietz, 1967). Preliminary studies from Tietz's laboratory indicate that the biosynthesis of glycerides proceeds by the pathway previously shown for mammalian tissues, namely : a-Glycero-P + 2 Acyl-CoA +. Phosphatidic acid Phosphatidic acid + H20+. DqP-Diglyceride + Pi D-a,P-Diglyceride + Acyl-CoA +.Triglyceride

Tietz (1967) reports that the relative amounts of palmitate incorporated into diglyceride and triglyceride will depend on the relative concentrations of a-glycero-P and diglyceride available as acceptors for palmityl CoA. Using fat body of resting locusts, which have a relatively high level of tissue a-glycero-P, biosynthesis of diglyceride predominates. In fat body from flown locusts, which is depleted of glycogen and the concentration of a-glycero-P is, therefore, low, the incorporation of palmitate into triglyceride will prevail. Addition of glucose to the incubation medium containing fat body from flown locusts will initiate a-glycero-P formation, and the biosynthesis of diglyceride will proceed at a higher rate. There are suggestions that lipid synthesis in fat body is under hormonal regulation. Bodenstein (1953) has reported that in the roach, Periplunetu, removal of the corpora allata results in a quantitative increase in the fat deposited in the fat body. This effect of the corpus allatum has been confirmed in the blowfly (Orr, 1964). Vroman et ul. (1965) have shown that incorporation of acetate into the triglyceride fraction is more than doubled in the allatectomized insects, whereas incorporation into phospholipid is not appreciably affected. Allatectomy also noticeably slows the turnover of both triglyceride and phospholipids and it has been suggested that the corpora allatum hormone increases triglyceride by regulating the mechanisms responsible for the utilization of lipids, i.e. ovarian

REGULATION OF INTERMEDIARY METABOLISM

319

development (also, see Gilbert, 1967b). In the female mosquito, the medial neurosecretory cells restrict the synthesis of glycogen and stimulate triglyceride synthesis (Van Handel and Lea, 1965). Removal of these cells greatly increases the storage capacity for glycogen at the expense of triglyceride storage. However, these authors caution against ascribing the effects directly to a neurosecretory cell hormone since bilateral sectioning of the combined nervus corporis allati and esophagi, which emanate from these cells, has the same metabolic effect as removal of the neurosecretory cells. D. MOBILIZATION AND TRANSPORT OF FAT

It has been calculated that a locust possessing about 180 mg of flight muscle consumes fatty acids at a rate of 4.1 mg x hr-1. The fatty acid content of the muscle is about 3 mg, while the fat body has more than 18 mg (Beenakkers, 1965). Since the locust flies continuously for 7-8 hr and its reserve of carbohydrate can last for only 1-2 hr, it is obvious that during flight fat is mobilized in the fat body and is transported to the flight muscle, presumably by way of the blood. 1. Release of Fatty Acids from Fat Body

I’ietz (1962) has found that when fat body of locusts, previously prelabeled with 14C-palmitate, is incubated in hernolymph, glycerides are released from the tissue into the medium. The effect of blood is specific; little glyceride is released in phosphate-saline, bovine serum, or buffered solutions of albumins. The amount of glyceride released is proportioaal to the amount of hemolymph that is added. Effectiveness of the blood is not affected by dialysis, but is destroyed by heating, and is inhibited by fluoride and cyanide. Confirmatory observations have been reported by Chino and Gilbert (1965) for three additional species. Further, they have noted that the release of glyceride is inhibited by azide and dinitrophenol. Chino and Gilbert (1965) have concluded that in pupal and adult cecropia, as well as in the roach and grasshopper, that the glyceride released by prelabeled fat body is in the form of diglyceride. Little triglyceride is liberated. Free fatty acids are found in the incubation medium; however, this release is not specific for insect hemolymph but is also enhanced by serum albumin. The metabolic inhibitors that block the release of diglyceride stimulate the yield of free fatty acids. The apparent accelerated rate of release of free fatty acids by the inhibitors is attributed t o their interference with the incorporation of

3 20

B. SACKTOR

free fatty acid into neutral lipid. According to Gilbert (1967a), the release of free fatty acids is a passive process which follows a concentration gradient and depends on the level of free fatty acid in the fat body. Recently, the possible presence of two different glyceride pools in fat body, only one of which directly releases fatty acids to the blood, has been suggested (Beenakkers and Gilbert, 1968). As noted previously, when isolated fat body is incubated with IT-palmitate, the fatty acid is incorporated into both di- and triglycerides. The specific activity of the diglyceride in the tissue is 50 times that of triglyceride, although the amount of triglyceride in the fat body is far greater (Chino and Gilbert, 1965; Tietz, 1967). When the prelabeled fat body is transferred to a medium containing hernolymph, the quantity of diglyceride that is released is greater than its concentration in the fat body, although the amount in the tissue does not change. However, its specific activity is markedly reduced. Consequently, the average specific activity of the glycerides in the blood is much greater than the average specific activity of the glycerides in the fat body because the diglyceride with an extremely high specific activity is continuously being released into hemolymph, while the low specific activity triglyceride largely remains in the fat body. When the fat body is prelabeled in vivo and the specific activities of the di- and triglyceride are identical, the specific activity of the diglyceride in the fat body is not reduced during diglyceride release (Tietz, 1967). The general conclusion that diglyceride is specifically released from the fat body of insects (Gilbert, 1967a), may now need revision. Wlodawer and co-workers (1966, 1967) have suggested that, in the larvae of the waxmoth, Galleria, prelabeled fat body releases free fatty acids into the hemolymph and these are subsequently incorporated into triglyceride by an active lipase in the blood. Very little radioactivity is found in diglyceride although the diglyceride represents the largest fraction of the glycerides. The findings of Cook and Eddington ( 1967) with Periplunetu are also inconsistent with the views of Gilbert and Tietz. Analytical as well as isotopic analyses show that free fatty acids and triglyceride are the major lipids released from fat body. Diglyceride appears to be the principal lipid in the hernolymph, however, as shown previously by Chino and Gilbert (1965). The contrast between the data of Cook and Eddington and those of Chino and Gilbert (1965) on the identification of the fatty acid that is released may be due to the

REGULATION OF INTERMEDIARY METABOLISM

321

latter authors’ use of counts rather than analyses, of levels of hemolymph which are well below the optimal for triglyceride release, and of not considering fully the consequences of diffusion. During flight the concentration of diglyceride in the hemolymph of bcusts increases several-fold (Beenakkers, 1965; Tietz, 1967). The levels of triglyceride and phospholipids in the blood do not change. Whether there is a change in the free fatty acid content of the blood of flown locusts is still inconclusive (Beenakkers, 1965 ;Tietz, 1967). The fat body is the source of the increased diglyceride in the blood and it has been suggested that the excess fatty acids are transported to the muscle for utilization in flight. However, until the turnover rates of the hernolymph lipids during the flight are known, it is difficult to assess the relative contributions of the different forms of lipid to flight metabolism.

2. Transport of Fat in Hemolymph The specific requirement for hemolymph in the liberation of lipids from the fat body suggests the involvement of a blood protein in the release and/or transport mechanism. Indeed, Tietz (1962) has shown that a lipoprotein component of locust blood becomes radioactive after incubation with prelabeled fat body and Chino and Gilbert ( 1955) have identified a diglyceride-hemolymph protein conjugate. In cecropia, four distinct protein bands are resolved electrophoretically, three of which are lipoproteins. Almost all the radioactive glyceride is concentrated with a single fraction (Chino and Gilbert, 1965). The three classes of lipoproteins have been isolated by ultracentrifugal techniques (Thomas and Gilbert, 1968). A high density lipoprotein contains about 75% of the total lipid. Electrophoresis demonstrates that each class is composed of several lipoprotein species. Previously, a diglyceride-bound lipoprotein from Cynthia has been isolated by chromatography and is described as a globulin-like protein (Chino et al., 1967). Acrylamide gel electrophoresis of the isolated lipoprotein reveals two protein bands, indicating also that the fraction obtained is not yet pure. Mayer and Candy (1967) have examined changes in hemolymph lipoprotein during flight of the locust. Electrophoresis of blood from resting locusts shows eight protein components, only two (Group A) of which contain lipid. After a 2 hr flight, the total lipid content (mainly diglyceride) of the hemolymph increases three to four times the resting value. Part of this increase, 70%, is accounted for by an increase in the lipid content of the A group lipoproteins. The

322

B. SACKTOR

remaining 30% is found in a second pair of proteins (Group B), which previously have been devoid of lipid. Group A lipoprotein contains both tri- and diglycerides, whereas Group B lipoprotein contains chiefly diglyceride. Three hours after flight has stopped the hemolymph pattern has returned to the resting state.

3. Uptake of Fat by Muscles The mechanisms by which the glycerides from the lipoproteins in the blood are transported into the muscle are essentially unknown. Lipase activities have been demonstrated in leg and flight muscles (George and Bhakthan, 1960a, b , 196 1). Considerably greater activities of the enzyme are found in flight muscle of locusts and dragonflies than in that of bumblebees. This difference between species seems to be correlated with locusts’ and dragonflies’ dependence on fat as a metabolic fuel during prolonged flights, whereas bumblebees, if like the honeybee, rely exclusively on carbohydrate for the energy for flight. Significantly, the lipase in flight muscle of cecropia hydrolyzes diglyceride at a rate of five times that of triglyceride (Gilbert et al., 1965). The lipase is not Ca*+-activated, As a consequence of lipase activity, free fatty acids and glycerol are formed. The role of cartinine in the transport of fatty acids has been considered previously. Glycerol is glycogenic and follows the pathway of carbohydrate in metabolism. Flight muscles of flies oxidize glycerol at a very slow rate (Sacktor, 1955). Presumably, glycerol is first converted to a-glycero-P at the expense of ATP in a reaction catalyzed by glycerol kinase. The diminutive rate of oxidation of glycerol relative to that of a-glycero-P suggests that the kinase reaction is rate-limiting.

V. REGULATION OF MITOCHONDRIAL METABOLISM

The large increase in rate of oxygen uptake upon initiation of flight indicates that there is an exceptionally high degree of respiratory control in flight muscle, in vivo. Since cellular respiration is attributed almost exclusively to mitochondria (Watanabe and Williams, 195 1 ; Sacktor, 1953b), an examination of mitochondrial metabolism is crucial to the understanding of the regulatory mechanisms at the three mitochondrial loci of control in flight muscle; namely, the oxidations of pyruvate, proline and a-glycero-P.

REGULATION OF INTERMEDIARY METABOLISM

3 23

A. THE RESPIRATORY CHAIN AND OXIDATIVE PHOSPHORY LATION The primary energy-conserving reaction in mitochondria is the formation of ATP coupled to the exergonic passage of protons and electrons from substrate to molecular oxygen via the respiratory chain. This vital process, oxidative phosphorylation, accounts for over 90% of the ATP generated. The components of the respiratory chain (NAD, flavoproteins, quinone, and the cytochomes), their spectral properties, concentrations and kinetic parameters, have been examined in detail in mitochondria of flight muscle of flies (Chance and Sacktor, 1958; Estabrook and Sacktor, 1958b) and locusts (Klingenberg and Bucher, 1959, 196 1 ). These fundamental studies, and some others which are largely confirmatory in nature, have been reviewed in depth (Sacktor, 1965). Since then, recent advances on the respiratory enzymes in insects include the isolation, crystallization and determination of the amino acid composition of cytochrome b from houseflies (Ohnishi, 1966a, b) and the establishment of the primary structure of cytochrome c from flight muscle of Cynthia (Chan and Margoliash, 1966). A general discussion of the current views of the reaction mechanisms of oxidative phosphorylation can be found elsewhere. The status of oxidative phosphorylation in insects has been summarized by Sacktor (1 965) and Harvey and Haskell (1966). Recently it has become clear that the respiratory pigments and phosphorylating enzymes are essential constituents of the mitochondria1 inner membrane (Figs 4 and 5). If analogous to the situation as found in mammalian mitochondria (Kawaga and Racker, 1966), the respiratory chain of insects is an integral part of the cristae and the phosphorylating machinery is localized in the stalked spherical particles, shown in Fig. 6. Although these knobs have not been isolated from insect preparations, a Mg*+- and DNP-activated (Sacktor, 1953a; Sacktor and Cochran, 1957), cold-labile (Mills and Cochran, 1967), oligomycin-sensitive (Hansford and Sacktor, 1970a) ATPase has been described. A fairly typical experiment illustrating classical respiratory control by ADP is shown in Fig. 17. Mitochondria exhibit high rates of respiration in the presence of ADP and Pi (State 3) but respire only slowly in the absence of phosphate acceptor (State 4). The respiratory control ratio, defined as the ratio of the rate of oxygen uptake in the presence of added ADP to the rate of respiration after the added ADP has been completely utilized, is approximately 6 in this experiment (Childress and Sacktor, 1966). The ADP : 0 ratio is

324

B. SACKTOR

defined as the ratio of the pmoles of ADP added t o the patoms of oxygen utilized, induced by the addition of ADP. A value of about 2.7 for the oxidation - of pyruvate approaches the theoretical maximum of 3.0 for this substrate. These ratios are in general agreement with those obtained manometrically by Gregg e t al. (1960) and Van den Bergh and Slater (1962) with fly mitochondria

F l

RESPIRATORY CONTROL DURING PYRUVATE OXIDATION

2.67

5.95

RATES IN MATOMS 0 2 / M I N I

c . 2 #MOLE ADP

Fig. 17. Respiratory control of pyruvate oxidation by blowfly mitochondria. (From Qlildress and Sacktor, 1966.)

and potentiometrically by Klingenberg and Bucher (1 959) with locust mitochondria. Although the stimulation of the rate of pyruvate oxidation by ADP is apparently indicative of tightly coupled mitochondria, respiratory control values with aglycero-P as substrate are relatively small when measured with the same insect preparations. Thus, R.C. ratios ranging from no stimulation to a high of about 3 have been reported for isolated mitochondria (Sacktor, 1954; Sacktor and Cochran, 1958; Gregg et al., 1960;Van den Bergh and Slater 1962; Birt, 1961; Klingenberg and Bucher, 1959; Hansford, 1968), and from 3-5 with the use of teased muscle preparations (Sacktor and Packer, 196 1). As emphasized previously

REGULATION OF INTERMEDIARY METABOLISM

3 25

(Sacktor, 1965), it is most important to point out that these observed stimulations of pyruvate and a-glycero-P oxidation by ADP, although of considerable significance, are much too small to account, alone, for the physiological control of respiration in the insect initiating flight. 9. CONTROL OF PYRUVATE OXIDATION As discussed in a previous section, the increase in concentration of pyruvate and accumulations of alanine and acetyl carnitine in the transition from rest to flight demonstrates that, on initiation of flight, pyruvate is not metabolized in the Krebs cycle as fast as it is formed by glycolysis. This indicates that there is a control on the oxidation of pyruvate that is released shortly after the onset of flight. The mechanisms of this regulation will be considered at this time. 1. Respiratory Substrates and Permeability of Mitochondria

Although historically there has been much dispute as to the relative abilities of fly flight muscle mitochondria to oxidize a-glycero-P, pyruvate and Krebs cycle intermediates (see Sacktor, 1965), it is now clear, as shown in Table IV, that intact mitochondria oxidize, at appreciable rates, only exogenous a-glycero-P, pyruvate, TABLE IV Respiratory activities of mitochondria from blowfly flight muscle Substrate a! -Glycero-P

Pyruvate Acetyl carnitine Proline Citrate a-Ketoglutarate Succinate Fumarate Malate Glutamate Asuarate

QOz 1400 65 5 330 130

10 45 60 35

30 25 10

QO, = rlOz x mg-I mitochondria1 protein x h r l . Data compiled from Sacktor and Childress (1967), Childress e l al. (1967), Childress and Sacktor (1966) and Chance and Sacktor (1958).

326

B. SACKTOR

acetyl carnitine and, to a lesser extent, proline. Other experiments (Van den Bergh and Slater, 1962; Hansford, 1968) show approximately equal rates of oxidation with a-glycero-P and pyruvate. Indeed, in some instances pyruvate is oxidized at twice the rate for a-glycero-P (Hansford, 1968). In contrast, Krebs cycle intermediates, such as citrate, a-ketoglutarate, succinate and malate, the amino acids glutamate and aspartate, and NADH, added to isolated mitochondria are not effective respiratory substrates (Chance and Sacktor, 1958; Van den Bergh and Slater, 1962; Childress and Sacktor, 1966; Childress e t al., 1967; Sacktor and Childress, 1967). An explanation for the low rates of oxidation of these substrates has been suggested by Van den Bergh and Slater (1962), who have discovered the unusual phenomenon that the mitochondria are not readily permeable to these compounds. Subjecting mitochondria to sonic disintegration or freezing-thawing, procedures that damage the mitochondrial membranes, increase the respiratory rates with these substrates many-fold (Van den Bergh and Slater, 1962; Sacktor and Childress, 1967). On the other hand, the oxidations of a-glycero-P and pyruvate are not stimulated by the disruptive techniques, provided cofactors are added back to the reaction. Van den Bergh (1967) can find no evidence in fly flight muscle mitochondria for the presence of specific exchange-diffusion carriers for Krebs cycle substrate anions.

2. The Function of Proline in Pyruvate Oxidation Isolated mitochondria rapidly lose the capacity to oxidize pyruvate (Childress and Sacktor, 1966; Sacktor and Childress, 1967). This loss can be reversed by proline but not by Krebs cycle intermediates nor glutamate (Fig. 18). Since blowfly flight muscle mitochondria are permeable to proline, but not t o other amino acids nor metabolites of the citric acid cycle, these findings suggest that proline enhances the rate of pyruvate metabolism by penetrating the mitochondrial barrier, forming intramitochondrial precursors of oxaloacetate, enabling the synthesis of citrate, and effecting the complete oxidation of pyruvate via the Krebs cycle at a maximal rate. In support of this view, it has been found that proline is metabolized by flight muscle with formation of A'-pyrroline-5carboxylate and glutamate (Bursell, 1963; Brosemer and Veeradhadrappa, 1965; Sacktor and Childress, 1967). Transamination of glutamate with pyruvate gives rise to alanine and a-ketoglutarate. The intramitochondrial a-ketoglutarate is further metabolized to

REGULATION OF INTERMEDIARY METABOLISM

3 27

dicarboxylic acids, forming oxaloacetate and C 0 2 (Sacktor and Childress, 1967; Bursell, 1967). Alternatively, the bicarbonate liberated by this sequence of reactions can be used to provide oxaloacetate by the enzyme pymvate carboxylase (Hansford and Chappell, 1968). The carboxylase has been reported in flight muscle (Lewis and Price, 1956; Pette et al., 1962; Bursell, 1965). RESTORATION OF PVRUVATE OXIDATION I V PROLINE

A-ASSAYED IMMEDIATELV B,C,L D ASSAYED AFTER 4 0 MIN EXPOSURE TO MEDIUM

0 . 9 4 n A T O M 02

\

M W

CVRUVATE IN MEDIUM INITIALLY

1.0 m M PYRUVATE IN MEDIUM INITIALLY

I.

\\

5.0 m M PROLINE IN MEDIUM INITIALLY

5.0 mM PROLINE

10mM GLUTAMAT

.._ .nM PVRUVATE

5.0 n M PROLINE

A R A T E S I N IA A T O M S

Fig. 18. Restoration of pyruvate oxidation by proline. (From Sacktor and Childress, 1967.)

The control on pymvate oxidation at the initiation of flight suggests that, in vivo, flight muscle mitochondria may be deficient in Krebs cycle intermediates and that these are generated from proline. By this mechanism, one of the limitations in the oxidation of pyruvate during flight is relieved.

3. The Requirement for Piand ADP for Activation at the Dehydrogenase Level During a study of kinetic factors controlling Krebs cycle activity in mitochondria from flight muscle of the blowfly, Calliphora, Hansford and Chappell (1968) have discovered that the rate of pyruvate oxidation is markedly influenced by Pi. As shown in Fig. 19, more than 25 mM Pi is required for a maximal State 3 rate. It is

328

B. SACKTOR

unlikely that this requirement has anything to do with respiratory chain phosphorylation as only 2 mMPi is needed for the ADP-stimulated oxidation of a-glycero-P. In contrast, the State 4, or controlled, rate of pymvate oxidation is unaffected by Pi. Thus, in the presence of ADP, the rate of oxygen uptake with pymvate increases by a factor of about 10 as the concentration of Pi is raised from 1.3 to 25 mM.

Concn. of Phosphate mM Fig. 19. The dependence upon Pi concentration of the rate of pyruvate oxidation by blowfly mitochondria. Incubation mixture contained pyruvate, bicarbonate, ATP and an excess of ADP. (From Hansford, 1968.)

This finding implies that the enzyme catalyzing the rate-limiting step of the Krebs cycle requires a very high level of Pi. Likely candidates are the substrate level phosphorylation concomitant with sketoglutarate oxidation and the NAD-linked isocitrate dehydrogenase. Substrate level phosphorylation has been largely ruled out, since at 1.2 mM Pi gramicidin-treated mitochondria oxidize a-ketoglutarate at a rate well in excess of that needed to be

329

REGULATION OF INTERMEDIARY METABOLISM

consistent with the very low rate of pyruvate oxidation at this level of Pi (Hansford, 1968). On the other hand, the NADdependent isocitric dehydrogenase of blowfly ritochondria has an absolute dependence on Pi (Fig. 20). Moreover, the requirement for Pi is very high; at 5 mM only 5% of the maximal activity is expressed, activity is 80%of maximum at 30 mM Pi (Hansford and Chappell, 1968).

. In al

d

0

E 1

E

z

0 bV 3

n

W LY

n

a

z

0

0

20

4Q

60

80

100

mM PHOSPHATE Fig. 20. The effect of Pi on blowfly NAD-isocitric dehydrogenase activity. (From Hansford and Chappell, 1968.)

Further support for the view that isocitrate dehydrogenase is subject to tight regulation and may limit pyruvate oxidation in the Krebs cycle has come from the finding that the enzyme is activated by ADP, as shown in Fig. 2 1. The mechanism of action of ADP is to lower the K, for isocitrate (Hansfordland Chappell, 1968). In the presence of a concentration of isocitrate approximating that found in the mitochondrion, there is a 20-fold increase in enzyme activity on adding ADP. It has also been found that the dehydrogenase is strongly inhibited by ATP (Hansford and Chappell, 1968). Thus,

330

B. SACKTOR

although the stimulation by ADP and inhibition by ATP are dependent on the concentrations of the effectors, the activity of isocitric dehydrogenase is determined by the relative proportions of the two nucleotides in a mixture of a fixed total concentration of adenine nucleotide. A concentration of adenine nucleotide in the mitochondria of about 6 mM has been estimated (Price and Lewis,

0 1 2 3 4 Concn. of ADP mM Fig. 21. The dependence of isocitric dehydrogenase activity upon the level of ADP. (From Hansford, 1968.)

1959; Hansford, 1968), and, if one assumes that in resting flight muscle most of the adenine nucleotide is ATP (Sacktor and Hurlbut, 1966), then the control of isocitrate dehydrogenase and, in turn, pyruvate oxidation by this mechanism must be quite rigorous. C. CONTROL OF PROLINE OXIDATION

As described in Fig. 16, the level of proline in flight muscle drops abruptly on the initiation of flght. It has been suggested that the mitochondria1 oxidation of proline is facilitated by the rest-to-flight

33 1

REGULATION OF INTERMEDIARY METABOLISM

transition and that this oxidation is crucial in providing the Krebs cycle intermediates necessary for the rapid and complete oxidation of pyruvate (Sacktor and Childress, 1967). In view of this role, the mechanism for the regulation of proline oxidation is important. Hansford and Sacktor (1970a) have found that the oxidation of proline by flight muscle mitochondria from Phormiu is stimulated by

0 12

-

0.10

f

s ' 0.08 N

0 v3

s 0.06

m

=l

w

k 0.04 a 0.02

2

1

3

mM ADP Fig. 22. The effect of ADP on proline oxidation. The uncoupling agent, FCCP, is in the reaction mixture. (From Hansford and Sacktor, 1970a.l

ADP in the presence of uncoupling agents and oligomycin (Fig. 22). The stimulation is enhanced further by high levels of Pi. These findings indicate that the site of action of the nucleotide is proline dehydrogenase, rather than on the respiratory chain. The mode of action of ADP is to lower the apparent K, for proline (Hansford and Sacktor, 1970a). Significantly, the K, in the presence of ADP approximates the concentration of proline found in

332

B. SACKTOR

the muscle, at rest (Sacktor and Wormser-Shavit, 1966). In a manner analogous to that for isocitric dehydrogenase, ATP inhibits proline dehydrogenase. Therefore, the activity of the enzyme is dependent, in part, on the composition of the adenine nucleotide mixture. As the nucleotide in the resting muscle is predominantly ATP, the small changes in ADP and ATP levels which are found when blowflies begin to fly (Sacktor and Hurlbut, 1966) will lead to a considerable enhancement in the rate of proline oxidation. D. CONTROL OF a-GLYCERO-P OXIDATION

For both pyruvate and proline oxidations, the key dehydrogenase is sensitive to a signal of the metabolic state of the muscle. It is also evident that for the third mitochondria1 site of regulation, u-glycero-P oxidation, control is manifested at the dehydrogenase level. Estabrook and Sacktor (1958a) have found that EDTA blocks 0

.-c € 0.7

H 0-6 0

-as 0.5 1

Y

b

0.4

3

0.3 2 !!

6

0.2

rc

0

Q,

0.1

. c .

b

'0

4 8 12 16 20 24 Concn. of D L Glycerol-3-P mM

Fig. 23. The effect of aglycero-P concentration on the rate of oxidation by blowfly flight muscle mitochondria, in the presence and absence of free calcium. Reaction mixture contains FCCP. (From Hansford and Chappell, 1967.)

REGULATION OF INTERMEDIARY METABOLISM

333

a-glycero-P oxidation and that this inhibition is reversed by the divalent cations, Ca2+ and Mg2+, and by additional substrate. The locus of inhibition has been observed to be at the dehydrogenase level. Based on these findings, a hypothesis has been suggested that regulation of a-glycero-P oxidation is achieved by reversal of the inhibited state by either the accumulation of substrate or, more likely, by release of divalent ions coincident with nervous stimulation of the muscle at the initiation of flight. This suggestion has received extensive confirmation through the work of Hansford and Chappell (1967). They have found that the metal ion involved is Ca2+,rather than Mg2+, and activation is maximal at very low levels of Ca2+.The level of free calcium that is required is about 5 x 10-7 g-ions/litre. As shown in Fig. 23, Ca2+has been seen to act by lowering the K, of the a-glycero-P dehydrogenase for its substrate (Hansford and Chappell, 1967). A plot of enzyme activity versus concentration of a-glycero-P reveals allosteric kinetics. At about 2 mM a-glycero-P, which is the physiological level in flight muscle (Sacktor and Wormser-Shavit, 1966), a 10-fold increase in rate is obtained on adding Ca2+. Increasing Ca2+ from 10-8 to 10-6 g-ion/litre causes a progressive increase in the ADPstimulated (State 3) rate, without an increase in the resting (State 4) rate. E. THE ENERGY-DEPENDENT ACCUMULATION OF Ca2+ AND Pi

It has been shown that Pi markedly stimulates the rate of pyruvate and proline oxidations and that Ca2+ enhances the oxidation of a-glycero-P. The need for very high levels of Pi has suggested that blowfly mitochondria may be capable of accumulating Pi against a concentration gradient. Direct support for this has been obtained by Hansford and Chappell (1968), who have found in Calliphora an energydependent uptake of Pi to about 60 mM. The accumulation of Pi is supported by both pyruvate and a-glycero-P and is inhibited and reversed by FCCP. Recently, Carafoli et al. (1969) have demonstrated that Phormia flight muscle mitochondria accumulate Ca2+ to a level as high as 400 nmoles xmg-1 protein (Fig. 24). The uptake requires respiratory energy derivable from either a-glycero-P or pyruvate. Uncoupling agents block and reverse, in part, the accumulation. Ca2+ uptake is not accompanied by ejection of H+. Instead, the simultaneous uptake of the Pi anion is required; inhibition of Pi uptake by mersalyl inhibits Ca2+binding. Acetate or chloride anions do not substitute for the Pi anion.

334

1

B. SACKTOR

4oo.

0 LL a

PYRUVATE PROLINE

300

++ 200 m 0

cn W

0

100

2 c

+ACETATE

0 0

3 6 MINUTES

Fig. 24. 'Ihe accumulation of (3% by blowfly flight muscle mitochondria. (From Carafoli et al., 1969.) F. INTERACTIONS OF METABOLIC EFFECTORS WITH THE RESPIRATORY CHAIN

Changes in the steady state oxidation-reduction level of the components of the respiratory chain in flight muscle mitochondria during the transition from a controlled to an active metabolic state have been examined by Chance and Sacktor (1958), Sacktor and Packer (1961), in flies, and Klingenberg and Bucher (1961), in locusts. In general, it has been seen that the respiratory components are largely oxidized in the presence of oxygen and absence of substrate. They become considerably reduced upon addition of a-glycero-P or pyruvate. For the most part, the extent of reduction is graded along the chain, the percentage decreasing from dehydrogenase to oxygen. (With a-glycero-P, reduction of NAD is due to reverse electron transport.) In most of these experiments a-glycero-P has been the substrate, and it has been found that addition of ADP

335

REGULATION OF INTERMEDIARY METABOLISM

or uncoupling agents cause the respiratory components to become more oxidized as the rate of oxygen uptake increases. When the added ADP has become phosphorylated to ATP, the respiratory rate decreases and the steady state levels of the cytochromes become more reduced. A second addition of ADP again initiates a transition to an increased respiratory rate (see Fig. 17), and decreased levels of reduction of cytochromes. When dissolved oxygen becomes exhausted and respiration ceases, the respiratory pigments go completely reduced.

TI

H

1 minute

Reduct ion

o*=0

r

(A1

glycerolphosphate I

02=0 I

I

I

Ca**and ADP

ADP

ADP

NADH Fig. 25. The effect of the simultaneous addition of ADP and CaB on the steady state levels of the respiratory components. oGlycero-P is the substrate. (From Hansford, 1968.)

Similar measurements of the redox states of cytochrome c and NAD have been made by Hansford (1968, 1969). The classical picture is seen when ADP is added to blowfly mitochondria respiring with a-glycero-P (Fig. 25). Surprisingly, however, during pyruvate oxidation, ADP addition results in an increased reduction even though, as described previously, the rate of respiration increases many-fold. Added Pi also causes a reduction, especially in the presence of oligomycin. These findings strongly support the concept that ADP and Pi activate a rate-limiting step in pyruvate oxidation (isocitrate dehydrogenase) even more than the respiratory chain. Hansford has also found that, when u-glycero-P is the substrate, added Ca*+increases reduction of the respiratory carriers at the same time that respiration is increased. Figure 25 shows an attempt to simuiate flight when both ADP and Ca2+ are added simultaneously,

336

B. SACKTOR

as may happen at the initiation of flight. It is evident that both cytochrome c and NAD become more reduced. These observations give strong support to the hypothesis that respiratory control in flight muscle may be due to control at the substrate level as well as to control in the respiratory chain (Chance and Sacktor, 1958; Estabrook and Sacktor, 1958a; Hansford and Sacktor, 1970a). Hansford’s elegant experiment also provides the experimental explanation for Keilin’s (1925) observations that the reduced cytochrome bands appear in the flight muscle, in situ, when the waxmoth starts to flap its wings. VI. CONCLUSIONS

This review has attempted to analyze the metabolic events that are fundamental to the transition of a muscle from a controlled resting state to one so active that energy transformations are taking place at a rate far in excess of that for any other biological process. The approach has been to identify the enzymatic reactions that are facilitated during the transition, to determine the mechanisms of activation at each locus of control, and then to formulate a working hypothesis that will unite the experimental findings into an overall scheme for the metabolic regulation of this tissue. This summarization should be premised with the knowledge that many of the experimental observations have been made with isolated enzymes and subcellular organelles. Thus, the in vivo environment can never be adequately reconstructed. Moreover, knowledge of the metabolite levels permitting simulation of in vivo conditions assumes uniform distribution of these substances and excludes compartmentation. These reservations made, the sites of regulation in fly flight muscle have been identified by measuring coincident and sequential changes in the concentrations of the metabolic intermediates. On the basis of the crossover theorem and initiation of utilization of stored reserves, these loci are: phosphorylase and trehalase, which determine the entrance of carbohydrate into the catabolic pathway; phosphofructokinase, which is rate-limiting for glycolysis; the mitochondria1 oxidations of a-glycero-P and pyruvate, the two end-products of glycolysis; and the oxidation of proline, which presumably provides intermediates for the initiation of the Krebs cycle. It is implicit to this discussion of the mechanisms of activation that, when the muscle is at rest and the rate of metabolism is low,

REGULATION OF INTERMEDIARY METABOLISM

337

the regulatory enzymes are in an inhibited state. Conversely, initiation of contraction sets off a chain of biochemical events that deinhibit or activate the enzymes, allowing for the accelerated rate of metabolism. With this in mind, it is noteworthy that high levels of ATP (or MgATP), a condition likely to be found in the muscle at rest, inhibit actomyosin ATPase, phosphorylase b, phosphofructokinase, isocitrate dehydrogenase and proline dehydrogenase. On the other hand, it is now known that the products of ATP breakdown, ADP, AMP and Pi, which are formed when the muscle contracts, counteract the effects of ATP or activate phosphorylase b, phosphofructokinase, isocitrate dehydrogenase and proline dehydrogenase. Significantly, and perhaps more than fortuitously, the low level of Ca2+,about 5 x 1O-' g-ions/litre, that activates myofibrillar ATPase, is the same concentration that activates phosphorylase b kinase and a-glycero-P dehydrogenase. Using these facts, as well as others cited previously in the review, it is tempting to hypothesize a scheme for the metabolic regulation of blowfly flight muscle. With the arrival of an electrical impulse and depolarization of the sarcolemma, the CaZ+which is sequested in the remnants of the sarcoplasmic reticulum is liberated. The free Ca2+ activates myofibrillar ATPase and phosphorylase b kinase. Phosphorylase b kinase converts the inhibited phosphorylase b to an active phosphorylase a, resulting in glycogenolysis. With the dephosphorylation of ATP by actomyosin and formation of ADP, AMP and Pi, phosphofructokinase becomes activated autocatalytically, enabling glycolysis to proceed at a maximum rate. The decrease in the concentration of extramitochondrial ATP by the myofibrillar ATPase, with the resultant increase in extramitochondrial ADP, also initiates a concentration-dependent exchange with the intramitochondrial adenine nucleotides; ATP leaves and ADP enters the mitochondria. The mitochondria will also accumulate Pi and Ca2+, now available. The decrease in mitochondrial ATP coupled with the increase in mitochondrial ADP, Pi and Ca2+ deinhibits the dehydrogenases. Ca2+, by lowering the K, of the u-glycero-P dehydrogenase, increases the oxidation of u-glycero-P. A 10-fold increase in oxidation is to be expected, at the level of substrate found in the tissue. A shift in the proportion of ATP and ADP initiates proline oxidation, also by reducing the K, of the enzyme for its substrate. With the limitation to pyruvate oxidation relieved by ample Krebs cycle intermediates derived from proline,

338

B. SACKTOR

ADP and Pi affects a 20-fold increase in isocitrate dehydrogenase activity, thus facilitating pyruvate oxidation. With controls at the dehydrogenase levels removed, the rates of oxidation of pyruvate and a-glycero-P via the respiratory chain, activated, in turn, several-fold by ADP in the classical manner, become sufficiently rapid to account for the oxygen uptake of the blowfly in flight. Concomitant with this respiration is the synthesis of the ATP for continued contraction.

REFERENCES Agosin, M., Scaramelli, N. and Neghme, A. (1961). Intermediary carbohydrate metabolism of Triatoma infestans (Insecta; Hemiptera). I. Glycolytic and pentose phosphate pathway enzymes and the effect of DDT. Comp. Biochem. Physiol. 2, 143-159. Appleman, M. H., Yunis, A. A., Krebs, E. G. and Fischer, E. H. (1963). Comparative studies on glycogen phosphorylase-the amino acid composition of rabbit and human skeletal muscle phosphorylase. J. biol. Chem. 238, 1358-1361. Ashhurst, D. E. (1967). The fibrillar flight muscle of giant water-bugs: an electronmicroscope study. J. Cell Sci. 2,425-444. Bade, M. L. (1964). Biosynthesis of fatty acids in the cockroach, Eurycotis jloridana. J. Insect Physiol. 10,333-34 1. Bade, M. L. and Clayton, R. B. (1963). Cholesterol esters of the cockroach, Eurycotis floridana. Nature, Lond. 197, 77-79. Beenakkers, A. M. Th. (1963a). Enzyme der fettsaure-oxydation in den flugmuskeln von Locusta migratoria wiihrend firer entwicklung. Biochem. Z. 337,436-439. Beenakkers, A. M. Th. (1963b). Fatty acid oxidation in insect muscles. Acta physiol. pharmacol. neerl. 12,332-335. Beenakkers, A. M. Th. (1965). Transport of fatty acids in Locusta migratoria during sustained flight. J. Insect Physiol. 11,879-888. Beenakkers, A. M. Th., DeWaide, J. H., Henderson, P. T. and Lutgerhorst, A. (1967). Fatty acid oxidation and some participating enzymes in animal organs. Comp. Biochem. Physiol. 22, 675-682. Beenakkers, A. M. Th. and Gilbert, L. I. (1968). The fatty acid composition of fat body and haemolymph lipids in Hyalophora cecropia and its relation to lipid release. J. Insect Physiol. 14,481-494. Beenakkers, A. M. Th. and Henderson, P. T. (1967). The localization and function of camitine acetyltransferase in the flight muscles of the locust. Eur. J. Biochem. 1, 187-192. Beenakkers, A. M. Th. and Klingenberg, M. (1964). Carnitine-coenzyme A transacetylase in mitochondria from various organs. Biochim. biophys. Acta 84,205-207. Birt, L. M. (1961). Flight muscle mitochondria of Lucilia cuprina and Musca dornestica. Estimation of the pyridine nucleotide content and of the

REGULATION OF INTERMEDIARY METABOLISM

339

response of respiration to adenosine diphosphate. Biochem. J. 80, 623-63 1. Bode, C. and Klingenberg, M. (1964). Carnitine and fatty acid oxidation in mitochondria of various organs. Biochim. biophys. Acta 84, 93-95. Bodenstein, D. (1953). Studies on the humoral mechanisms in growth and metamorphosis of the cockroach, Periplaneta americana. 111. Humoral effects on metabolism. J. exp. 2001.124, 105-1 16. Bowers, W. S. and Friedman, S. (1963). Mobilization of fat body glycogen by an extract of corpus cardiacum. Nature, Lond. 198,685. Brosemer, R. W. ( 1967). The levels of extramitochondrial glycerophosphate dehydrogenase in wing muscles of a flightless grasshopper. J. Insect Physiol. 13,685-690. Brosemer, R. W. and Marquardt, R. R. (1966). Insect extramitochondrial glycerophosphate dehydrogenase. 11. Enzymic properties and amino acid composition of the enzyme from honeybee (Apis mellifera) thoraces. Biochim. biophys. Acta 128,464-473. Brosemer, R. W. and Veerabhadrappa, P. S. (1965). Pathway of proline oxidation in insect flight muscle. Biochim. biophys. Acta 110, 102-112. Bucher, T. and Klingenberg, M. (1958). Wege des wasserstoffs in der lebendigen organisation. Angew. Chem. 70,552-570. Bueding, E. and Orell, S. A. (1964). A mild procedure for the isolation of polydisperse glycogen from animal tissues. J. biol. Chem. 239,4018-4020. Bursell, E. (1963). Aspects of the metabolism of amino acids in the tsetse fly, Glossina (Diptera). J. Insect Physiol. 9,439-452. Bursell, E. (1965). Oxaloacetic carboxylase in flight musculature of the tsetse fly. Comp. Biochem. Physiol. 16,259-266. Bursell, E. (1966). Aspects of the flight metabolism of tsetse flies (Glossina). Comp. Biochem. Physiol. 19,809-818. Bursell, E. (1967). The conversion of glutamate to alanine in the tsetse fly (Glossinamoristans). Comp. Biochem. Physiol. 23, 825-829. Cabib, E. and Leloir, L. F. (1958). The biosynthesis of trehalose phosphate. J. biol. Chem. 231, 259-275. Candy, D. J. and Kilby, B. A. (1959). Site and mode of trehalose biosynthesis in the locust. Nature, Lond. 183, 1594-1595. Candy, D. J. and Kilby, B. A. (1961). The biosynthesis of trehalose in the locust fat body. Biochem. J. 78,531-536. Carafoli, E., Sacktor, B. and Lehninger, A. L. (1969). Manuscript in preparation. Chadwick, L. E. (1953). Aerodynamics and flight metabolism. In “Insect Physiology” (K. D. Roeder, ed.), pp. 615-636. John Wiley and Sons, New York. Chan, S. K. and Margoliash, E. (1966). Properties and primary structure of the cytochrome c from the flight muscles of the moth, Samia Cynthia. J. biol. Chem. 241,335-348. Chance, B. and Sacktor, B. (1958). Respiratory metabolism of insect fight muscle. 11. Kinetics of respiratory enzymes in flight muscle sarcosomes. Archs Biochem. Biophys. 76, 509-531. Chaplain, R. A. (1966). The allosteric nature of substrate inhibition of insect actomyosin ATPase in presence of magnesium. Biochem. biophys. Res. Commun. 22,248-253.

340

B. SACKTOR

Chaplain, R. A. (1967). The effect of Ca2+and fibre elongation on the activation of the contractile mechanism of insect fibrillar flight muscle. Biochim. biophys. Acta 131, 385-392. Chefurka, W. ( 1958). On the importance of a-glycerophosphate dehydrogenase in glycolysing insect muscle. Biochim. biophys. Acta 28,660-66 1. Chefurka, W. (1965). Intermediary metabolism of carbohydrates in insects. In “The Physiology of Insecta” (M. Rockstein, ed.), Vol. 11, pp. 58 1-768. Academic Press, New York and London. Childress, C. C. and Sacktor, B. (1966). Pyruvate oxidation and the permeability of mitochondria from blowfly flight muscle. Science, N . Y. 154,268-270. Childress, C. C. and Sacktor, B. (1969). Regulation of glycogenolysis in insect flight muscle in vitro and in vivo. Fedn Proc. Am. SOCSexp. Biol. 28,412. Childress, C. C. and Sacktor, B. (1970). Regulation of glycogen metabolism in flight muscle. I. Purification and properties of phosphorylases. 11. Kinetic properties and control of phosphorylase in vivo. J. biol. Chem. in press. Childress, C. C., Sacktor, B., Groseman, I. W.and Bueding, E. (1970). Isolation, ultrastructure, and biochemical characterization of glycogen in insect flight muscle. J. Cell Biol. 45, in press. Childress, C. C., Sacktor, B. and Traynor, D. R. (1967). Function of carnitine in the fatty acid oxidasedeficient insect flight muscle. J. biol. Chem. 242, 754-760. Chino, H. and Gilbert, L. I. (1965). Lipid release and transport in insects. Biochim. biophys. Acta 98, 94-1 10. Chino, H., Sudo, A. and Harashima, K. (1967). Isolation of diglyceride-bound lipoprotein from insect hemolymph. Biochim. biophys. Acta 144, 177-179. Clegg, J . S. and Evans, D. R. (1961). The physiology of blood trehalose and its function during flight in the blowfly. J. exp. Biol. 38,77 1-792. Clements, A. N. (1959). Studies of the metabolism of locust fat body. J. exp. Biol. 36,665475. Cook, B. J. and Eddington, L. C. (1967). The release of triglycerides and free fatty acids from the fat body of the cockroach, Periplaneta americana. J. Insect Physiol. 13, 1361-1372. Davis, R. A. and Fraenkel, G. (1940). The oxygen consumption of flies during flight. J. exp. Biol. 17,402407. Domroese, K. A. and Gilbert, L. I. (1964). The role of lipid in adult development and flight-muscle metabolism in Hyalophora cecropia. J. exp. Biol. 41,573-590. Edwards, G. A. and Ruska, H. (1955). The function and metabolism of certain insect muscles in relation to their structure. Q. JI microsc. Sci. 96, 15 1-159. Estabrook, R. W. and Sacktor, B. (1958a). a-Glycerophosphate oxidase of flight muscle mitochondria. J. biol. Chem. 233, 1014-1019. Estabrook, R. W. and Sacktor, B. (1958b). The respiratory metabolism of insect flight muscle. 111. Low-temperature spectra of the cytochromes of flight muscle sarcosomes. Archs Biochem. Biophys. 76, 532-545. Friedman, S. (1960). Occurrence of trehalose-6-phosphatae in Phormiu regina Meig. Archs Biochem. Biophys. 88,339-343. Friedman, S . (1967). The control of trehalose synthesis in the blowfly,Phormia regina Meig. J. Insect Physiol. 13,397-405.

REGULATION OF INTERMEDIARY METABOLISM

34 1

Friedman, S. ( 1968). Trehalose regulation of glucose-6-phosphate hydrolysis in blowfly extracts. Science, N.Y. 159, 110-1 1,l. Friedman, S. and Fraenkel, G. (1955). Reversible enzymatic acetylation of camitine. Archs Biochem. Biophys. 59,491-501. Fritz, I. B. (1955). The effects of muscle extracts on the oxidation of palmitidic acid by liver slices and homogenates. Acta pkysiol. scand. 34, 367-385. Fritz, I. B. and Marquis, N. R. (1965). The role of acylcamitine esters and carnitine palmityltransferase in the transport of fatty acyl groups across mitochondria1 membranes. Proc. natn. Acad. Sci. U.S.A. 54, 1226-1233. George, J. C. and Bhakthan, N. M. G. (1960a). A study on the fibre diameter and certain enzyme concentrations in the flight muscle of some butterflies. J. exp. Biol. 37, 308-315. George, J. C. and Bhakthan, N. M. G. (1960b). The fibre diameter and certain enzyme concentrations in the flight muscles of some moths. J. Anim. Morph. Physiol. 7, 141-149. George, J. C. and Bhakthan, N. M. G. (1961). Lipase activity in the slow and fast contracting leg muscles of the cockroach. Nature, Lond. 192,356. Gilbert, L. I. (1967). Lipid metabolism and function in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 4, pp. 69-2 11. Gilbert, L. I. (1967b). Changes in lipid content during the reproductive cycle of Leucophaea maderae and effects of the juvenile hormone on lipid metabolism in vitro. Comp. Biochem. Physiol. 21, 237-257. Gilbert, L. I., Chino, H. and Domroese, K. (1965). Lipolytic activity of insect tissues and its significance in lipid transport. J. Insect Physiol. 11, 1057-1070. Gilby, A. R. (1965). Lipids and their metabolism in insects. A . Rev. Ent. 10, 14 1-160. Gilby, A. R., Wyatt, S. S. and Wyatt, G. R. (1967). Trehalases from the cockroach, Blaberus discoidalis: activation, solubilization and properties of the muscle enzyme and some properties of the intestinal enzyme. Acta biochim. pol. 14,83-100. Goldin, H. H. and Keith, A. D. (1968). Fatty acid biosynthesis by isolated mitochondria from Drosophila melanogaster. J. Insect Physiol. 14, 887-899. Grasso, A. and Migliori Natalizi, G. (1968). Studies on insect (Periplaneta americana, L.) phosphofructokinase. Comp. Biochem. Physiol. 26, 979-984. Gregg, C. T., Heisler, C. R. and Remmert, L. F. (1960). Oxidative phosphorylation and respiratory control in housefly mitochondria. Biochim. biophys. Acta 45, 561-570. Gregory, D. W., Lennie, R. W. and Birt, L. M. (1968). An electron-microscopic study of flight muscle development in the blowfly Lucilia cuprina. JI R . microsc. SOC.88, 15 1-175. Gussin, A. E. S. and Wyatt, G. R. (1965). Membrane-bound trehalase from cecropia silkmoth muscle. Archs Biochem. Biophys. 112,626-634. Hansen, K. (1966). Zur cytologischen lokalisation der trehalase in der indirekten flugmuskulatur der insekten. Biochem. 2. 344, 15-25. Hansford, R. G. (1968). The oxidative metabolism of fly flight muscle. University of Bristol, Ph.D. Thesis, July 1968. A.1.P.-12

342

B. SACKTOR

Hansford, R. G. (1969). The oxidative metabolism of fly flight-muscle. In “Advances in the Study of Metabolic Control” (J. B. Chappell and P. B. Garland, eds), Vol. I. John Wiley, London. (In press.) Hansford, R. G. and Chappell, J. B. (1967). The effect of Ca2* on the oxidation of glycerol phosphate by blowfly flight-muscle mitochondria. Biochem. biophys. Res. Commun. 27,686-692. Hansford, R. G. and Chappell, J. B. (1968). The energy dependent accumulation of phosphate by blowfly mitochondria and its effect on the rate of pyruvate oxidation. Biochem. biophys. Res. Commun. 30,643448. Hansford, R. G. and Sacktor, B. (1970a). The control of the oxidation of proline by isolated flight muscle mitochondria. J. biol. Chem. 245,991-994. Hansford, R. G. and Sacktor, B. (1970b). Regulation of glycogen metabolism in insect flight muscle. Activation of phosphorylase b kinase by calcium and inorganic phosphate. FEBS Letters, in press. Harvey, W. R. and Haskell, J. A. (1966). Metabolic control mechanisms in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 3, pp. 133-205. Academic Press, London and New York. Heslop, J. P., Price, G. M. and Ray, J. W. (1963). Anaerobic metabolism in the housefly, Musca domestica L. Biochem. J. 87,35-38. Hess, R. and Pearse, A. G. E. (1961). Localization of dehydrogenases and glycogen metabolizing enzymes in muscle tissue of the desert locust (Schistocercagregaria). Enzymol. Biol. Clin. 1, 15-33. Hill, A. V. (1948). On the time required for diffusion and its relation to processes in muscle. Proc. R . SOC.B135,446453. Hill, A. V. (1949). The abrupt transition from rest to activity in muscle. Proc. R . SOC.B136,399-419. Hines, W. J. W. and Smith, M. J. H. (1963). Some aspects of intermediary metabolism in the desert locust (Schistocerca gregaria Forskal). J. Insect Physiol. 9,463-468. Hofmanova, O., Cerkasovova, A., Foustka, M. and Kubista, V. (1966). Metabolism of thoracic musculature of Periplaneta americana during flight. Acta Univ. Carol. Biol. 1966, 183-189. Horie, Y. (1968). The oxidation of NADPH by the soluble fraction of the fat body of the silkworm, Bombyx mori L. J. Insect Physiol. 14,417-424. Hone, Y.,Nakasone, S. and Ito, T. (1968). The conversion of 14C-carbohydrates into C02 and lipid by the silkworm, Bombyx mori. J. ZnsectPhysiol. 14, 97 1-98 1. Jewell, B. R. and Ruegg, J. C. (1966). Oscillatory contraction of insect fibrillar muscle after glycerol extraction. Proc. R . SOC.B164.428-459. Kagawa, Y. and Racker, E. (1966). Partial resolution of the enzymes catalyzing oxidative phosphorylation. X. Correlation of morphology and function in submitochondrial particles. J. biol. Chem. 24 1, 2475-2482. Keilin, D. (1925). On cytochrome, a respiratory pigment common to animals, yeast and higher plants. Proc. R . SOC.B98,312-339. Klingenberg, M. and Biicher, Th. (1 959). Flugmuskelmitochondrien aus Locusta migratoria mit atmungskontrolle. (Aufban und zusammen setzung der atmungskette.) Biochem. Z. 33 1 , 312-333. Klingenberg, M. and Biicher, Th. (1961). Glycerin-1-P und flugmuskel mitochondrien. Biochem. Z. 334, 1-17.

REGULATION OF INTERMEDIARY METABOLISM

343

Kolliker, A. (1857).Z. wiss. Zool. 8, 311. Krogh, A. and Weis-Fogh, T. (1951). The respiratory exchange of the desert locust (Schistocerca gregaria) before, during, and after flight. J. exp. Biol. 28,344-357. Kubista, V. (1957). Accumulation of a stable phosphorus compound in glycolysing insect muscle. Nature, Lond. 180,549. Lardy, H. A. and Parks, R. E., Jr. (1956). Influence of ATP concentration on rates of some phosphorylation reactions. Zn “Enzymes: Units of Biological Structure and Function” (0. H. Gaebler, ed.), pp. 584-587. Academic Press Inc., New York. Lewis, S. E. and Price, G. M. (1956). Malic enzyme activity in blowfly muscle. Nature, Lond. 177, 842-843. Louloudes, S. J., Kaplanis J. N., Robbins, W. E. and Monroe, R. E. (1961). Lipogenesis from (?‘-acetate by the American cockroach. Ann. ent. SOC. A m . 54, 99-103. Lowry, 0. H., Passonneau, J. V., Hasselberger, F. X. and Schulz, D. W. ( 1964). Effect of ischemia on known substrates and cofactors of the glycolytic pathway in brain. J. biol. Chem. 239, 18-30. Maruyama, K. (1965). The biochemistry of the contractile elements of insect muscle. In “Physiology of Insecta” (M. Rockstein, ed.), Vol. 11, pp. 45 1482. Academic Press Inc., New York. Maruyama, K., Pringle, J. W. S. and Tregear, R. T. (1968). The calcium sensitivity of ATPase activity of myofibrils and actomyosins from insect flight and leg muscles. Proc. R. SOC.B169,229-240. Mayer, R. J. and Candy, D. J. (1967). Changes in haemolymph lipoproteins during locust flight. Nature, Lond. 215,987-990. McCarthy, R. and Ralph, C. L. (1962). The effects of corpora allata and cardiaca extracts on hemolymph sugars of the cockroach. Am. Zool. 2,429. Metzger, B. E., Glaser, L. and Helmreich, E. (1968). Purification and properties of frog skeletal muscle phosphorylase. Biochem. 7, 202 1-2036. Meyer, H.,Preiss, B. and Bayer, Sh. (1960). The oxidation of fatty acids by a particulate fraction from desert locust (Schistocerca gregaria) thorax tissue. Biochem. J. 76,27-35. Mills, R. R. and Cochran, D. G. (1967). Adenosinetriphosphates from thoracic muscle mitochondria of the American cockroach. Comp. Biochem. Physiol. 20,919-923. Monod, J., Changeux, J. P. and Jacob, F. (1963). Allosteric proteins and cellular control systems. J. molec. Biol. 6,306-326. Monod, J., Wyman, J. and Changeux, J. P. (1965). On the nature of allosteric transitions: a plausible model. J. molec. Biol. 12, 88-1 18. Murphy, T. A. and Wyatt, G. R. (1965). The enzymes of glycogen and trehalose synthesis in silkmoth fat body. J. biol. Chem. 240, 150@1508. Ohnishi, K. (1966a). Studies on cytochrome b. 11. Crystallization and some properties of cytochrome b from larvae of the housefly, Musca domestica L. J. Biochem. Tokyo, 59, 9-16. Ohnishi, K. (1 966b). Studies on cytochrome b. 111. Comparison of cytochrome b’s from beef heart muscle and larvae of the housefly. J. Biochem. Tokyo, 59, 17-23. Orr, C. W. R. (1964). The influence of nutritional and hormonal factors on the chemistry of the fat body, blood, and ovaries of the blowfly, Phormia regina Meig. J. ZnsectPhysiol. 10, 103-1 19.

344

B. SACKTOR

Passonneau, J. V. and Lowry, 0. H. (1962). Phosphofructokinase and the Pasteur effect. Biochem. biophys. Res. Comniun. 7, 10-15. Pette, D., Klingenberg, M. and Biicher, Th. (1962). Comparable and specific proportions in the mitochondrial enzyme activity pattern. Biochem. biophys. Res. Commun. 7,425-429. Price, G. M . and Lewis, S . E. (1959). Distribution of phosphorus compounds in blowfly thoracic muscle. Biochem. J. 71, 176-185. Pringle, J. W. S. (1949). The excitation and contraction of the flight muscle of insects. J. Physiol. 108, 226-232. Pringle, J. W.S. (1967a). The contractile mechanism of insect fibrillar muscle. In “Progress in Biophysics and Molecular Biology” (J. A. Butler and H. E. Huxley, eds), Vol. 17, pp. 1-60. Pergamon Press, Oxford and New York. Pringle, J. W. S. (1967b). Evidence from insect fibrillar muscle about the elementary contractile process. J. gen. Physiol. 50, 139-156. Ralph, C. L. (1962). Action of extracts of cockroach nervous system on fat bodies in vitro. Am. Zool. 2 , 5 5 0 . Robbins, W. E., Kaplanis, J. N., Louloudes, S. J. and Monroe, R. E. (1960). Utilization of I€14-acetate in lipid synthesis by adult houseflies. Ann. ent. SOC.Am. 53, 128-129. Roeder, K. D. (195 1). Movements of the thorax and potential changes in the thoracic muscles of insects during flight. Biol. Bull. 100, 95-106. Rosell-Perez, M. and Lamer, J. (1964). Studies on UDPG-cu-glucan transglucolyase. IV. Purification and characterization of two forms from rabbit skeletal muscle. Biochem. 3 , 7 5 4 1 . Ruegg, J. C. (1968). Oscillatory mechanism in fibrillar insect flight muscle. Experientia 24, 529-536. Ruegg, J. C. and Tregear, R. T. (1966). Mechanical factors affecting the ATPase activity of glycerolextracted insect fibrillar flight muscle. Proc. R . SOC. B165,497-512. Ruiz-Amil, M. (1962). The hexokinase of the honey bee. J. Insect Physiol. 8, 259-265. Sacktor, B. (1953a). Investigations on the mitochondria of the housefly, Musca domestica L. J. gen. Physiol. 36, 37 1-387. Sacktor, B. ( 1953b). Investigations on the mitochondria of the housefly, Musca domestica L. 11. Oxidative enzymes with special reference to malic oxidase. Archs Biochem. Biophys. 45,349-365. Sacktor, B. (1954). Investigations on the mitochondria of the housefly, Musca domestica L. 111. Requirements for oxidative phosphorylation. J. gen. Physiol. 37,343-359. Sacktor, B. (1955). Cell structure and the metabolism of insect flight muscle. J . biophys. biochem. Cytol. 1 , 2 9 4 6 . Sacktor, B. (1961). The role of mitochondria in respiratory metabolism of flight muscle. A. Rev. Ens. 6 , 103-130. Sacktor, B. (1965). Energetics and respiratory metabolism of muscular contraction. In “The Physiology of Insecta” (M. Rockstein, ed.), Vol. 2, pp. 483-580. Academic Press, New York. Sacktor, B. and Childress, C. C. (1967). Metabolism of proline in insect flight muscle and its significance in stimulating the oxidation of pyruvate. Archs Biochem. Biophys. 120,583-588.

REGULATION OF INTERMEDIARY METABOLISM

345

Sacktor, B. and Cochran, D. G. (1957). Dephosphorylation of nucleotides by insect flight muscle. J. biol. Chem. 226,241-254. Sacktor, B. and Cochran, D. G. (1958). The respiratory metabolism of insect flight muscle. I. Manometric studies of oxidation and concomitant phosphorylation with sarcosomes. Archs Biochem. Biophys. 74, 266-276. Sacktor, B. and Dick, A. (1962). Pathways of hydrogen transport in the oxidation of extramitochondrial reduced diphosphopyridine nucleotide in flight muscle. J. biol. Chem. 237,3259-3263. Sacktor, B. and Hurlbut, E. C. (1966). Regulation of metabolism in working muscle in vivo. 11. Concentrations of adenine nucleotides, arginine phosphate, and inorganic phosphate in insect flight muscle during flight. J. biol. Chem. 241,632-634. Sacktor, B. and Packer, L. (1961). The stimulation of a-glycerolphosphate oxidation by adenosine diphosphate in teased flight muscle. Biochim. biophys. Acta 49,402404. Sacktor, B. and Wormser-Shavit, E. (1966). Regulation of metabolism in working muscle in vivo. I. Concentrations of some glycolytic, tricarboxylic acid cycle, and amino acid intermediates in insect flight muscle during flight. J. biol. Chem. 24 1,624-63 1. Sedee, D. J. W. (1961). Intermediary metabolism in aseptically reared blowfly larvae, Calliphora erythrocephala (Meig). 11. Biosynthesis of fatty acids and amino acids. Arch. Int. Physiol. Biochim. 69, 295-309. Shafiq, S. A. (1963). Electron microscopic studies on the indirect flight muscle of Drosophila melanogaster. I. Structure of the myofibrils. J. Cell Biol. 17, 35 1-362. Smith, D. S. (1961a). The structure of insect fibrillar flight muscle. A study made with special reference to the membrane systems of the fiber. J. biophys. biochem. Cytol. 10, 123-158. Smith, D. S. (1961b). The organization of the flight muscle in a dragonfly, Aeshna sp. (Odonata). J. biophys. biochem. Cytol. 11, 119-145. Smith, D. S. (1963). The structure of flight muscle sarcosomes in the blowfly Calliphora erythrocephala (Diptera).J . Cell Biol. 19, 1 15-138. Smith, D. S. (1965). The organization of flight muscle in an aphid, Megoura viciae (Homoptera). With a discussion on the structure of synchronous and asynchronous striated muscle fibers. J. Cell Biol. 27, 379-393. Smith, D. S. (1966a). The organization and function of the sarcoplasmic reticulum and T-system of muscle cells. In “Progress in Biophysics and Molecular Biology” (J. A. Butler and H. E. Huxley, eds), Vol. 16, pp. 107-142. Pergamon Press, Oxford and New York. Smith, D. S. (1966b). The organization of flight muscle fibers in the Odonata. J. Cell Biol. 28, 109-126. Smith, D. S. and Sacktor, B. (1970). Disposition of membranes and the entry of haemolymph-borne ferritin in flight muscle fibers of the fly Phormia regina. Tissue and Cell. in press. Sridhara, S. and Bhat, J. V. (1964). Incorporation of [ 1-14C]-acetate into the lipids of the silkworm, Bombyx mori L. Biochem. J. 91, 12@123. Steele, J. E. (1961). Occurrence of a hyperglycemic factor in the corpus cardiacum of an insect. Nature, Lond. 192, 680-68 1. Steele, J. E. (1963). The site of action of insect hyperglycemic hormone. Gen. Comp. Endocr. 3,46-52.

346

B. SACKTOR

Stevenson, E. (1966). Rapid oxidation of palmitate with concomitant phosphorylation of adenosine 5’-diphosphate by moth flight-muscle mitochondria. Biochim. biophys. Acta 1.28,29-33. Stevenson, E. (1968a). Carbohydrate metabolism in the flight muscle of the southern armyworm moth, Prodenia eridania. J. Insect Physiol. 14, 179-198. Stevenson, E. (1 968b). The carnitine-independent oxidation of palmitate plus malate by moth flight-muscle mitochondria. Biochem. J. 110, 105-1 10. Stevenson, E. and Wyatt, G. R. (1964). Glycogen phosphorylase and its activation in silkmoth fat body. Archs Biochem. Biophys. 108, 420-429. Strong, F. E. (1963). Fatty acids: in vivo synthesis by the green peach aphid, Myzus persicae (Sulzer). Science, N . Y. 140, 983-984. Thomas, K. K. and Gilbert, L. I. (1968). Isolation and characterization of the hemolymph lipoproteins of the American silkmoth, Hyalophora cecropia. Archs Biochem. Biophys. 127, 5 12-521. Tietz, A. (1961). Fat synthesis in cell-free preparations of the locust fat body. J. Lipid Res. 2, 182-187. Tietz, A. (1962). Fat transport in the locust. J. Lipid Res. 3,421426. Tietz, A. (1965). Metabolic pathways in the insect fat body. In “Handbook of Physiology” (A. E. Renald and G. F. Cahill,Jr., eds), Section 5, pp. 45-54. Williams and Wilkins, Baltimore. Tietz, A. (1967). Fat transport in the locust: the role of diglycerides. Eur. J. Biochem. 2,236-242. Treherne, J. E. (1958). The absorption and metabolism of some sugars in the locust, Schistocerca gregaria (Forsk). J. exp. Biol. 35,611-625. Treherne, J. E. (1960). The nutrition of the central nervous system in the cockroach, Periplaneta americana L. J. exp. Biol. 3 7 , s 13-533. Trivelloni, J. C. (1960). Biosynthesis of glucosides and glycogen in the locust. Archs Biochem. Biophys. 89, 149-150. Tsukamoto, M., Nagai, Y.,Maruyama, K. and Akita, Y.(1966). The occurrence of relaxing granules in the muscle of the locust, Locusta migratoria. Comp. Biochem. Physiol. 17, 569-581. Van Handel, E. and Lea, A. 0. (1965). Medial neurosecretory cells as regulators of glycogen and triglyceride synthesis. Science, N. Y. 149, 298-300. Van Handel, E. and Lum, P. T.’M. (1961). Sex as a regulator of triglyceride metabolism in the mosquito. Science, N. Y . 134, 1979-1980. Van den Bergh, S. G. and Slater, E. C. (1962). The respiratory activity and permeability of housefly sarcosomes. Biochem. J. 82, 362-37 1. Van den Bergh, S. G. (1967). Permeability towards substrate anions of mitochondria from rat liver and housefly and locust flight muscle. In “Mitochondrial Structure and Compartmentation” (E. Quagliariello, S. Papa, E. C. Slater and J. M.Tager, eds). Bari, 203-206. Vardanis, A. (1967). Glycogen synthetase of bee larvae. Utilization of native primer and the effects of glucose-6-phosphate. J. biol. Chem. 242, 2306-23 11. VogeU, W., Bischai, F. R., Biicher, Th. and Klingenberg, M. (1959). Uber strukturelle und enzymatische muster in muskeln von Locusta migratoria. Biochem. Z . 332,8 1-1 17. Vom Brocke, H. H. (1966). The activating effects of calcium ions on the

REGULATION OF INTERMEDIARY METABOLISM

347

contractile systems of insect fibrillar flight muscle. Pflugers Arch. ges. Physiol. 290,70-79. von Siebold, C. T. E. (1848). Lehrbuch der vergleichende. Anatomie der werbellosen Tiere. Veit and Co., Berlin, p. 561. Vroman, H. E., Kaplanis, J. N. and Robbins, W. E. (1 965). Effect of allatectomy on lipid biosynthesis and turnover in the female American cockroach, Periplaneta americana (L). J. Insect Physiol. 11,897-904. Watanabe, M. I. and Williams, C. M . (195 1). Mitochondria in the flight muscle of insects. I. Chemical composition and enzymatic content. J. gen. Physiol. 34, 675-689. Weis-Fogh, T. (1952). Fat combustion and metabolic rate of flying locusts (Schistocerca gregaria Forskal). Phil. Trans. R. SOC.B237, 1-36. Weis-Fogh, T. (1 964). Diffusion in insect wing muscle, the most active tissue known. J. exp. Biol. 4 1 , 229-256. Weis-Fogh, T. (1967). Respiration and tracheal ventilation in locusts and other flying insects. J. exp. Biol. 47,561-587. Weins, A. W. and Gilbert, L. I. (1967a). The phosphorylase system of the silkmoth, Hyalophora cecropia. Comp. Biochem. Physiol. 21, 145-159. Wiens, A. W. and Gilbert, L. I. (1967b). Regulation of carbohydrate mobilization and utilization in Leucophaea maderae. J. Insect Physiol. 13, 779-794. Winteringham, F. P. W. (1959). Comparative aspects of insect biochemistry with particular reference to insecticidal action. IT.‘ Inter. Cong. Biochem., Vienna, 1958. Vol. 12, 201-215. Wlodawer, P. and Lagwinska, E. (1967). Uptake and release of lipids by the isolated fat body of the waxmoth larva. J. Insect Physiol. 13, 3 19-33 1. Wlodawer, P., Lagwinska, E. and Baranska, J. (1966). Esterification of fatty acids in the waxmoth haemolymph and its possible role in lipid transport. J. Insect physiol. 12, 547-560. Wyatt, G. R. (1967). The biochemistry of sugars and polysaccharides in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 4 , pp. 237-360. Academic Press, London and New York. Zebe, E. (1954). Uber den stoffwechsel der Lepidopteren. 2. vergl. Physiol. 36, 290-3 17. Zebe, E. (1 960). Condensing enzyme und 0-ketoacylthiolase in verschiedenen muskeln. Biochem. 2. 332, 328-332. Zebe, E. C. and McShan, W. H. (1959a). Trehalase in the thoracic muscle of the woodroach, Leucophaea maderae. J. cell. comp. Physiol. 5 3 , 2 1-29. Zebe, E. C. and McShan, W. H . (1959b). Incorporation of 14C-acetate into long chain fatty acids by the fat body of Prodenia eridania. Biochim. biophys. Acta 31,513-518.

This Page Intentionally Left Blank

Cellular Mechanisms Underlying Behavior* -Neuroethology GRAHAM HOYLE Department of Biology. University of Oregon Eugene. Oregon. U.S.A. I.

Introduction . . . . . . . . . . . . . . . . . 349 A . Insects and Ethology . . . . . . . . . . . . . . 349 B. Behavior Amenable t o Analysis: Defining Neuroethology . . 353 I1. Neural Architecture and Physiology . . . . . . . . . . . 356 A . Anatomy . . . . . . . . . . . . . . . . . 356 B . CellularPhysiology of Motorneurons . . . . . . . . 361 C . The Electrical Activity of Neuropil . . . . . . . . . 375 D . Mechanism of Ijabituation . . . . . . . . . . . . 387 E. Memory and “Learning” Machinery . . . . . . . . . 392 I11. Motor Mechanisms . . . . . . . . . . . . . . . . 398 A . General Aspects . . . . . . . . . . . . . . . 398 B . Respiration . . . . . . . . . . . . . . . . . 401 C. Perambulatory Locomotion . . . . . . . . . . . 403 D . Flight . . . . . . . . . . . . . . . . . . . 408 E. Insect Song: Crickets . . . . . . . . . . . . . 412 F . Courtship Behavior . . . . . . . . . . . . . . 417 IV . Models of Neural Activity and Terminology . . . . . . . . 420 V . Discussion . . . . . . . . . . . . . . . . . . . 425 Acknowledgements . . . . . . . . . . . . . . . . . . 430 References . . . . . . . . . . . . . . . . . . . . . 431 Note Added in Proof . . . . . . . . . . . . . . . . . . 439 f

I . INTRODUCTION A . INSECTS AND ETHOLOGY

There are so many attractions in principle to the use of insects in the study of neural mechanisms underlying behavior-availability, simplicity of neural organization. small numbers of nerve cells. short life cycle. etc.-that the subject has been surprisingly slow to get * Unpublished research by G . Hoyle reported in this article was supported by research grant GB 7413 from NSF . A.1.P.- 12’

34 9

350

G . HOYLE

under way. The reference list will bear witness to a recent rapid expansion and, although it is much shorter than might have been hoped for, some prosperous years must surely lie ahead. Because insects are largely terrestrial and have invaded a very wide variety of environments, they have also evolved many complex behavior patterns. Overt insect behavior has been thoroughly reviewed by Baerends ( 1959) and by Mark1 and Lindauer ( 1965) and has been the topic of a symposium (Haskell, 1966). Its study has produced several classic examples of behavior, such as the marvellous work of von Frisch (1 967) and Lindauer (1 961) on bees and of Baerends (1941), Evans (1957, 1966) and others on the digger wasps Ammophile, Bembix, etc. Unfortunately, these examples are not likely to be helpful to neurophysiologists, but of the greatest importance in general is the fact that modern generalizers in the field of ethology (Tinbergen, 195 1 ; Thorpe, 1963; Manning, 1967) include insects quite naturally with other animals all the way up the evolutionary scale to man, in their consideration of principles, as if there is a true continuum in the underlying neurological machinery. Younger investigators who have been brought up on these liberal views may raise the question “Is any other view reasonable?”, but the possibility of giving serious attention to the insect nervous system, in the search for general principles of neurobiology, is associated only with the last decade, and is still denied by most cat-brain physiologists. Instinctive behavior is not at all well understood in higher animals, nor is the mechanism of learning, and they must eventually be understood at the cellular level before real progress in neurophysiological understanding of behavior can be achieved. Both have been amply demonstrated to occur in a number of insects and therefore insects might be able to play a significant role in fundamental studies on them, just as did Drosophila in genetics. But for an adequate scale of experimental attack the logical continuity must be generally accepted, as it was for genetics. Bullock (1 966) has called for clever “strategy” in relation to the approach. I believe that what we need most is a massive and above all a concerted attack. Furthermore, the direct approach is sadly neglected. Whilst Mittelstaedt (1 962), Hassenstein (1958), Thorson (1 964), McCann (Bishop et al., 1968) and others make highly sophisticated whole-system cybernetic-model studies on bits of reflexive behavior, there is still not a single published study of the excitability of an insect neuron soma. Whilst a profusion of electron micrographs of neuropil appear, maps of neural

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

35 1

pathways, which would be of much greater value, especially if pertaining to insects which are popular for experimental work, have progressed but little beyond that given for an Aeschna larva ganglion by Zawarzin in 1924. Judging by the frequency with which the relevant picture of Zawarzin’s is referred to, one might suppose that young investigators anxious for quick and widespread recognition would be attracted to neuroanatomy; but unfortunately this has not yet been the case. The problem is that research in insect neurobiology proceeds by happenstance, not plan; the whims of investigators, current fashions, and the influence of immediate “key” figures, guide an army consisting entirely of captains-no men, n o generals. This is partly because no organism has selected itself as the “Drosophila” of insect neurobiology. Under the circumstances it is clearly necessary for senior investigators to lead the way, and to focus attention. The most-investigated insect, both neurophysiologically and otherwise, at the present time, is the locust Schistocerca gregaria. The efforts of the British Anti-Locust Research Centre, in making this insect available to laboratories all over the world, are largely responsible for its popularity, and it is now freely available commercially. It is a tough, reliable experimental animal, and of course as more information accumulates about it the more valuable it becomes. But the desert locust unfortunately has many shortcomings. The subtlety of its behavioral repertoire leaves much t o be desired. Nevertheless, it is the only insect about which a strong case can be made for a concerted effort to fill in the gaps in knowledge, at the present time. The locust is closely followed, as first choice, by the cockroach Periplaneta americana, which is even more readily and widely available. Of special advantage is the ease with which this insect accepts supernumerary limbs, or regenerates extirpated ones. The cockroach is readily available throughout the world, but is fragile, suffers from blood which clots readily, and like the locust has a limited behavioral repertoire. We are still searching for an ideal subject ! Notable advances have been made in the study of genetics of overt behavior mechanisms. For example Manning ( 196 1) showed that the speed of mating may be improved by selection in Drosophila. Of tremendous interest and significance is the demonstration by Rothenbuhler (1964a, b) that small bits of a behavior sequence are subject to Mendelian laws. In this case the removal of dead larval honey bees from the comb by certain strains of bees, termed

352

G. HOYLE

“hygienic”, is the behavioral act, which includes two major sections: uncapping of the cell, and removal of the dead larva. When the normal bees were crossed with a strain not showing this behavior, termed “unhygienic”, the progeny were all unhygienic. But a back-cross of the hybrids to the recessive, hygienic strain led t o a simple segregation into four equal groups: (1) bees which uncapped only and did not remove the corpses; (2) bees which would not uncap but would remove dead larvae if the cap was first removed by the investigator; (3) ones which did both, i.e. were hygienic; (4) bees which were unhygienic. A general review of the genetics of insect behavior has been given by Ewing and Manning (196’7). The possibilities of combining gene tic with neurophysiological ahalysis in furthering our understanding of gene tics, developmental biology and neurophysiology are clearly profound. But is it really technically feasible to explore the insect’s neural machinery, bearing in mind that we need to be able to associate electrical activity in single cells with actual behavioral acts? This combination has been achieved to a limited extent with two vertebrates-monkey and cat, but astronomical numbers of nerve cells (Ca 10’0) make a full understanding virtually impossible. Insects have something of the order of lo5 or lo6 neurons in their entire nervous system! These are manageable numbers for complete model-building. Also, since about 80%are concerned with immediate integration of sensory input-a great deal of which is redundant information-a remarkably small number deal with the determination and integrative co-ordination of motor activity. Quite small numbers of neurons must also be concerned with storing the information required to execute the acts of simple and complex instinctive behavior and such learning as the insect engages in. These facts should greatly encourage anatomists and physiologists to work on insects. So far the only organism in which significant behavioral acts have been obtained and correlated with events in single neurons of the nervous system is the large nudibranch mollusk Tritonia gilberti (Willows, 1967). An aquatic organism can be maintained in an apparently normal, translational environment by having the water move by it, whilst it is otherwise stationary. In the Tritonia preparation, it is effectively suspended by the brain whilst microelectrodes are placed in specific brain cells according to a simple map. Comparable experiments in the case of terrestrial insects are being attempted in the study of locomotion, both of flight and

353

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

walking, and of singing in crickets. The extent to which these experiments are being successful will be considered below. B. BEHAVIOR AMENABLE TO ANALYSIS: DEFINING NEUROETHOLOGY

It will simplify our approach to neuroethology if we can categorize behaviors from a neurophysiological point of view. This is most readily done on the basis of a hierarchy of complexity. A physiologist will equate behavior with movement, and movements permit a preliminary classification depending upon the number of muscles involved, the number of sequences of contraction, and the variety of the sequences. Thus the simplest behavior results from a single contraction of a single muscle. A more complex act will involve several muscles, part synchronously and part sequentially. There are different degrees of significance though, at each level, and these can only be described ethologically. A simple movement may have, or be capable of acquiring, significance of a subtle character. For example, if the antenna1 levator contracts this may represent only a spontaneous random increase in frequency of its motor neuron and have zero behavioral significance. Alternatively, it may be significant in the context of searching for a scent signal, which is perhaps more closely related t o sensory physiology than to ethology, although it may then be regarded as behaviorally significant. But it can also serve as a communication signal, or a releaser of behavior as part of a courtship pattern. When the latter occurs, the simple act is associated with a high level of ethological significance. At each stage the final neural program is identical, but the relations 1p to the nervous system and to other motor acts becomes in'treasingly complex. Understanding the motor machinery of antenna leyation is primarily the concern of physiology and it may be an essential prerequisite to the study of the movement in relation t o the total sequence. But its study alone is not enough. Neuroethology really begins when the act is demonstrated to be used in a larger context and when the neural machinery of the relationship to the larger act is the concern of a neurophysiological investigation. A second level of motor acts, non-repetitive movements which involve the co-ordinated activities of more than one muscle and which have adaptive value may, in themselves, be recognized as acts of behavior, not merely muscular contractions. Examples are: jumping, seizing food, clasping, defensive kicking. Stereotyped sub-systems of behavior compounded of cyclically-repeated

t

354

G. HOYLE

sequences of co-ordinated contractions may be recognized as the third step in complexity, e.g. waving, waggling, raising and lowering, scissoring of wings, nodding, rocking, grooming. These are not without behavioral interest, as the work of Blest (1960) on Lepidoptera and Bastock and Manning ( 1955) on Drosophila courtship amply bring out. A sustained level of central nervous excitation is required for their production, unlike the former categories. In a class by itself comes the perpetual rhythmical cycle of the movements of respiration. From the point of view of neuroethology this may be relegated to the level of pure physiology, but there are probably some common features of the underlying neural machinery. The simplicity and the persistence of respiratory phenomena permit extensive neural analysis which the neuroethologist will wish to be acquainted with. Also in this category are the more complex cycles comprising the locomotory movements of walking, running, flying, swimming, in which there is a complete overlap with the interests of the physiologist. The kinds of activities which will come strictly within the purveyance of the neuroethologist are the great classes of instinctive behavior: escape, defense, aggression (display and attack), sexual display, courtship, mating, egg-laying, nest-building, bodily care, exploration, food-search, feeding, feeding mate or offspring, nest care, etc. Each of these contains one or more elemental units of motor action, many of which are amenable to analysis by neurophysiological methods. A sixth level is that associated with various forms of learning, from simple habituation through avoidance conditioning to the learning of the geography of the environment. The phenomena involved in learning require a different type of classification, for they must be categorized in a way which is related only to the quality of underlying neurophysiological processes and not to the complexity of associated motor activity. Finally, we have the multiple complex acts of which only a few insects are capable, such as “foraging” in the bee, which involve the receipt, interpretation and communication of complex information, memorization, orientation, flight, food collection and distribution. Neuroethology must be concerned with the machinery of the nervous system which programs these various acts and it may be argued that so complex an activity as the foraging of the bee is not amenable to analysis. But at least some components are, and it is possible to visualize a tethered bee with microelectrodes implanted in its nervous system being informed by a waggle dance of the distance and direction of food. We cannot at the present time suggest where

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

355

in the bee’s brain the information is sorted and stored, i.e. where to place electrodes, for far too little is known about the bee’s brain, but this is a foreseeable experiment. The physiologist has three possible ways of approaching the nervous system: he may proceed from the sensory apparatus inwards, from the central nervous system outwards, or from the muscles upwards. The writer is predisposed to prefer the latter route and the present review leans heavily in this direction. The reason for the preference is that if the output is known there is a good chance that anything we learn about the central activity which is correlated with it can be interpreted in an exact and perhaps quantitative manner. This is true in insects because so few motor neurons innervate each muscle that the output consists of but a few trains of impulses. It often consists of a single train. By contrast, if we start with the input or directly with the ganglia, there is no guarantee that any behavior will result or that it will be correlated with the input, or central activity. Also, much of what is done at the input end relates strictly to sensory transduction and data reduction. It triggers behavior or guides it but does not participate in the actual programing of output. The potentialities of the subject are clearly vast. The present article will attempt to review some of the recent studies which are relevant to neuroethology and will also present some results of unpublished exploratory work being carried out in our laboratory, by various persons. Much of this has so far been done on locusts, and also on the cockroach P.americana. It is a paradox of the subject that the investigator suffers simultaneously from a dearth of information-especially about anatomical localizations and pathways and the properties of the neurons-but a surfeit of recordings. In one month an active laboratory can accumulate a mile of paper, every inch of which contains neural recordings which require analysis; yet merely to glance through them takes several hours. But unless such results can be distilled down to a single significant statement they cannot be communicated to other scientists. It is an unfortunate aspect of this subject that it not only suffers from the usual interdisciplinary problem of several sets of research literature, but also of a remarkably diverse publication loyalty even of the contributors to the various branches of the subject. For example, on one relevant contributory subject alone, that of insect central nervous synaptic transmission, the most recent review I saw covered 129 references, a modest number by modern standards, but these were drawn from no fewer than 46 different journals or books.

356

G. HOYLE

Too much time is wasted in searching so diverse a set of sources as these represent, bearing in mind that this is one fragment only. One may wish that eventually there will be a Journal of Neuroethology. In the meantime, it is to be hoped that contributors to what must become a rapidly growing field will favor a few of the more widely circulating journals in which this work is acceptable, such as J. Exp. Biol., J. Insect Physiol., J. Neurophysiol. and Z. Vergl. Physiol., and the newly-formed J. Neurobiology. The account which follows is an attempt to summarize relevant recent additions to knowledge, especially of the basic physiology of insect neurons, and of the operation of specific neural circuits controlling motor activity, with special reference to the new field of neuroethology. I have included a number of unpublished results of my own, and also a few others which were kindly supplied by Drs Bentley*, Runion* and Delcomyn. The results, except those of Dr Bentley, fall short of what is desirable in regard to the aims of neuroethology, but they are a necessary preliminary to larger-scale investigations. 11. NEURAL ARCHITECTURE AND PHYSIOLOGY A.

ANATOMY

I . General Anatomy of Insect Ganglia Pursuance of the ultimate goals outlined will require a detailed knowledge of the anatomy of the relevant nervous systems. The combined use of light microscopic serial sections and low-power electron microscopy are already adequate to permit a complete threedimensional model of a part of or even the whole insect nervous system to be made. The magnitude of the task should, however, not be underestimated. It will take a few investigators, each with adequate technical assistance, several years of continuous work to accomplish the task, in any one case. What is needed is not so much strategy as organization and effort. No satisfactory detailed account of ganglion structure yet exists for any ganglion of any insect. There are but a few papers on ultrastructure (e.g. Hess, 1958; Trujillo-CCnoz, 1962; Smith and Treherne, 1963); the subject has been reviewed by Bullock and Horridge ( 1965). Cohen and Jacklet ( 1967) have made an important methodological contribution in demonstrating that in a cockroach ganglion the nuclei of motor neurons whose axons have been cut, * Since published-see Note Added in Roof.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

357

within a few hours of the cut being made come to be surrounded by a sphere of newly-formed RNA. Since this can be stained histochemically it is possible in principle t o locate within the ganglion the cell bodies associated with a particular motor nerve fibre. A rough map of the location of some of the motor neurons has already been made by severing whole nerve trunks (Fig. l), but it has yet to be

Fig. 1. Threedimensional drawing of the locations of some of the principal motorneurons in the metathoracic ganglion of the cockroach Periplunetu umericuna. The numbered cell bodies were identified in relation to the nerve trunk from which their motor axons emerge by cutting the nerve trunks and later staining the ganglion for RNA. A few of them have now been related to a particular muscle. A ring of new RNA occurs round the nuclei of neurons whose axons have been cut. (From Cohen and Jacklet, 1967.)

shown that the method can be used to locate the soma of individual specified (identified) neurons. Several matters of interest emerge from this study. Firstly, the cell bodies are largely on the same side of the ganglion as their motor axons, though there are exceptions for particular nerves. Secondly, the suspected symmetry of the two halves of the ganglion has been confirmed. Thirdly, neurons in ganglia of different individuals occur in the same locality within the ganglia, as has been shown for molluskan ganglia possessing giant neurons (Coggeshall, 1967 ; Willows, 1967). One hopes that Cohen’s (1 967) conclusion that “the ability to identify specific individual neurons within a cockroach central ganglion allows concentration of a variety of anatomical, chemical and physical approaches on

358

G. HOYLE

individual cells known to be engaged in specific integrative activity” will in fact be borne out, though it was certainly premature to make the suggestion in the present state of knowledge. It was also irresponsible of Eisenstein and Krasilovsky ( 1967) to state that they are “investigating . . . cellular changes associated with learning in the prothoracic segment of the cockroach”. There is not yet the possibility of such an approach in insects, for physiological phenomena are confined to bits of dendritic material embedded in the neuropil where they are not amenable to analysis by currently available methods. The dendrites must have physiological features, of

B ,

loop

,

Fig. 2. A. General sketch of the nervous system of the pond skater Gerris, showing the location of the giant internuncial neuron (4th). B. Giant internuncial neuron having dendritic connections in several segments. The fourth giant internuncial of the pond skater Gerris. (From Guthrie, 1961.)

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

359

extreme significance for neuroethology , which are not necessarily reflected in the chemical and biophysical properties of the soma. It is perhaps possible that there are chemical reflections of the significant phenomena in the neuron soma, but we are a long way from establishing that this is the case, and this is absolutely necessary before such claims can even in principle be justified. An anatomical study made by Guthrie (1961) on the swimming bug Gerris was very successful, and it is hoped that correlated physiological studies will be made. Silver impregnation by the Golgi-Cox method revealed giant interneurons half a millimeter long, having dendritic branches in at least three ganglionic neuropils. These are also thick enough to be penetrable by intracellular microelectrodes, and long enough to permit insertion of two or three (Fig. 2). Locations and pathways of numbers of motorneurons and other interneurons were also given. Gross maps of the brain and ganglia are also urgently required, and electrophysiologists are finding it necessary to make their own, as Maynard (1967) has done for cockroach brain. Mill (1963) made me taI microelectrode recordings from Anax larval last abdominal ganglia, and has backed these up by histological studies (1 964). Although this is an excellent approach, the results are so far disappointing, at least in the present context. The possibility of there being some kind of anatomical division of the axon groups within the nerves emerging from ganglia of insects, as occurs in the vertebrates, where dorsal roots of the spinal cord are sensory and ventral roots motor, has been examined for the last abdominal ganglion of a dragonfly nymph by Fielden (1 963). She split the nerve trunk and both stimulated and recorded from dorsal and ventral portions. The results showed that motor elements are dorsal and sensory ones ventral, the reverse of the vertebrate situation. This is somewhat surprising in view of the (presumed) ventral location of the motor nerve cell bodies. This is, however, an isolated example and is not borne out in general in my own experience. Some nerves are either wholly motor or wholly sensory. Further comparative studies are needed in regard to this point. Electron microscopists have been attracted t o the neuropil (Hess, 1960; Smith, 1965, 1967; Trujillo-CCnoz, 1962; Smith and Treherne, 1963; Osbome, 1966) and their studies are providing valuable information, but unfortunately nothing of direct help yet to the neurophysiologists. The picture of the neuropil remains one of utter confusion. Only the sort of actual or semi-serial sectioning, combined

3 60

G . HOYLE

with inspired guesswork such as has been so successfully carried out by Dowling and Boycott (1967) on the vertebrate retina is likely to be really helpful in interpreting the neuropil from the point of view of function, as indeed the authors quoted above point out. A dramatic break-through in invertebrate neuroanatomy has recently been achieved by Stretton and Kravitz (1968). These ,authors, and also Remler et al. (1 968) used glass capillary microelectrodes filled with a saturated solution of the fluorescent dye Procion yellow. The electrode was inserted into a soma and depolarizing current passed through it. If the cell was motor it could be identified by observing the ensuing contraction. The current was then reversed, causing the electrically charged dye t o pass by iontophoresis out of the electrode into the cell. After some hours the dye passes down the axon and into the dentritic branches of the neuron. The neuron may then be observed in a whole mount and in sections, with the aid of fluorescence microscopy. Its topography within the ganglion and its associations with other neurons can then be mapped. Dr David Bentley has shown (in unpublished studies) that the method can be applied to some of the flight motorneurons in the mesothoracic ganglion of Schistocerca gregaria, and it is to be anticipated that the method will eventually be widely used in insects. Walker and Pitman (1969, see Note Added in Proof) have now succeeded in injecting Procion yellow into soma of the cockroach 6th abdominal ganglia. It seems unlikely that neuropil will in fact turn out to be a disorderly uninterpretable tangle, or even that it works in some way by local electric fields as Hughes (1952) has suggested, or in some extremely complex-and therefore virtually un-analyzable-manner, as Horridge (1968) implies. But it is certainly complex and difficult and the problem will be to persuade a new generation of anatomists to tackle it. If physiologists were agreed as to which bit of which ganglion they would most like to see worked out the anatomists would be more likely to make the great efforts needed on their part. If the former will call the tune I feel sure the latter will gladly co-opera te.

2. Location of Synapses There has yet to be discovered a single clear example in insects of synapses occurring on the neuron soma surface as occurs in the vertebrate CNS. They were claimed some time ago by Leghissa

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

361

(1942) for Carausius, but TrujilloCCnoz (1962) did not find any on Pholus nerve cell bodies, and none of the electronmicroscopists nor any other light microscopists who have examined insect CNS have been willing to claim that they exist. All the central synapses which have been clearly established occur within the neuropil on dendritic branches and we may reasonably assume that these are the major, or even the only, ones. The cell bodies are very tightly enveloped in glial cells (Smith and Treherne, 1963). There are two levels of “explanation” of the difference between insects and vertebrates, an immediate structural and/or chemical one, and a teleological one. Considering the latter, in comparison with the vertebrate, only matters associated with a different form of processing of neural information, i.e. a different brand of computer appropriate to large-scale data handling by a small amount of software, could be relevant. Of more direct interest is the possibility that the physiology of the cells is significantly different. For synapses on the neuron surface to be functional they would need to be capable of initiating junctional potentials large enough to influence impulse-generating sites on the axon. B. CELLULAR PHYSIOLOGY OF MOTORNEURONS

I . Excitability -Intracellular Recording All the available evidence-most of it unpublished because it is of the “negative” kind, but passed on by private communication among investigators-suggests that the insect neuron soma is universally, or nearly universally, electrically inexcitable. Furthermore, complete depolarization of the soma leads to such a small effect in the axon that the resulting electrotonic potential there may not be sufficient to initiate a spike. Very large depolarizing currents nevertheless initiate impulses. The attachment of the soma to its axon is very thin in insects (e.g. Wigglesworth, 1959), and presumably has a high electrical resistance. Conversely, therefore, spikes in the axon fail to give more than a minute electrotonic potential in the soma. I have personally penetrated most of the larger cell bodies in the metathoracic ganglia of Schistocerca and Periplaneta with glass microelectrodes filled with potassium chloride, citrate or acetate. A total of about 800 penetrations and tests of excitability have been made. So far, all the cell bodies have proved to be electrically inexcitable. A similar experience has been stated for cricket ganglion cells (Bentley, 1969a, b).

362

G. HOYLE

Only one positive assertion of success existed in the insect literature for 10 years following successful recording from mammalian spinal cord motorneurons, ,that due to Hagiwara and Watanabe ( 1956) on the motorneuron of the sound-producing muscle of the tymbal muscle of a cicada. Using a KC1-filled glass capillary microelectrode they obtained a 60-mV resting potential in A

B

Fig. 3. Intracellularly-recorded synaptic and action potentials from the motorneuron of the tymbal muscle of the cicada. A. Responses to antidroniic stimulation after penetration of a neuron. B. Same as A, with the electrode first outside the cell (upper three traces), then inside the cell (lower three traces). C. Repetitive discharge recorded intracellularly in the motorneuron, which resemble the natural output causing a sound, evoked by a single afferent volley in the auditory nerve. (From Hagiwara and Watanabe, 1956.)

the ganglion, in the region where they expected the cell body to be. When the nerve which evokes the sound was stimulated they obtained large excitatory postsynaptic potentials (EPSP’s). The largest of these elicited all-or-nothing, overshooting, action potentials (70mV). The interval between the potentials was similar to that occurring in the sound-producing muscle during orthodromic reflex action giving rise to sound. Antidromic stimulation also gave overshooting spikes. These results (Fig. 3) were very similar to those

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

363

which had been obtained by Eccles from motorneurons of the spinal cord of the cat and gave great promise for insect neurophysiology. Hagiwara and Watanabe determined the conduction times for sensory and motor nerves in their cicada preparation very carefully and were able to obtain the central reflex time. This had the long duration of 30-40 ms, which is sufficiently great to include two or more synaptic delays. A single electric stimulus applied directly to the caudal sensory nerve elicited up to seven equally-spaced EPSP’s and six or seven impulses in the motorneuron, the number depending upon the stimulus strength. The number of impulses in a reflex discharge was not influenced by an antidromic impulse. The motorneuron is not itself the source of the rhythm; one or more interneurons are. Hagiwara and Watanabe were not able to record intracellularly from relevant interneurons. In crayfish, Preston and Kennedy (1 960a, b; Kennedy and Preston, 1960) and others, obtained many results using conventional intracellular electrodes which even more clearly resemble the classical vertebrate motorneuron results. The recordings were all, however, made “blind”, i.e. the tips of the electrodes could not be seen in the microscope during the penetration. When the cell bodies were definitely penetrated, comparable activity was not recorded and Kennedy ( 1966) was convinced that the successful results were obtained from axons, not cell bodies. However, it has now been established by several persons (e.g. Roberts, 1968) that some crayfish and lobster cell bodies reflect both synaptic activity (up to a few mV) and spikes (up to 30 mV) occurring in neuropilar dendrites. Professor Hagiwara tells me that he did not see the electrode tip penetrate the soma of the cicada and that therefore it is possible that in his case the recordings were made from a dendrite in the neuropil. Earlier attempts were all made with the aid of KC1-filled electrodes, however, and it is possible that chloride diffusing from the electrode tip depressed the excitability of the cell. Recent penetration of cockroach cell bodies with electrodes filled instead with citrate are giving more encouraging results. Kerkut et al. ( 1968) removed the 5th and 6th abdominal ganglia of male cockroaches and placed the tissue in an insect saline. After dissecting away a small piece of sheath from the dorsal surface they found it “fairly easy to insert a microelectrode into a superficial specific neuron”. Resting potentials of 55-60 mV were obtained with citrate-filled electrodes, and about 15% of the cells penetrated showed spontaneously-

364

G . HOYLE

occurring action potentials having a 25-30 mV overshoot. Activity continued up to more than 3 hr. KC1-filled electrodes caused 5-10 mV hyperpolarization after penetration. The authors think they were “certainly recording from the cell body of the neurone” and that any synaptic activity is taking place some distance along the axon away from the cell. E. C. Rowe’s Ph.D. thesis (1963) was exclusively devoted to exploring activity in the cockroach me tathoracic ganglion. One hundred and seventeen cells were recorded from with intracellular electrodes but only three gave overshooting action potentials and these were all nerve branches. No action potentials were found in soma. Of 18 sites showing synaptic potentials, only two were deemed possible cell body ones, the rest all being in the neuropil. Many in tracellular recordings were made from the mesothoracic ganglion of locusts by Kendig ( 1968) but she failed completely t o record resting potentials. Spikes as large as 50 mV were, however, obtained from neuropil dendrites, or axons coursing through it, by the microelectrode serving as a very efficient extracellular lead.

2. Synaptic Potentials Synaptic potentials of conventional appearance were recorded extracellularly by Yamasiki and Narahashi ( 1960) and used as a basis for pharmacological studies. EPSP’s and IPSP’s were recorded intracellularly by Rowe ( 1960, 1963) in Peripluneta fuliginosu, by stimulating the leg nerves in the tibiae, and ganglion. Units were found which did not at first give responses, even at stimulation rates of 30/s, but eventually started to give rise to EPSP’s at progressively decreasing intervals. On cessation of the stimuli, the unit did not immediately cease t o follow, but slowly declined over several seconds. Inhibitory impulses could be obtained by applying single stimuli to certain other nerves. These were evoked in trains by single shocks. Spontaneously-firing units were also found, and when the inhibitory response was evoked, these spontaneous IPSP’s were arrested temporarily. The frequency of their discharge was extremely high, in the region of 150/s. Results of this kind are extremely promising but, of course, would be much more valuable if both input stimuli and neurons recorded from could be prescribed acutely. Callec and Boistel (1965, 1966) quite recently were able to record good resting potentials intracellularly from the cockroach last

365

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

abdominal ganglion, but from unidentified cells. Stimulation of the cercal nerves gave large EPSP's in the cells, and action potentials if these were large enough. Evidence of IPSP's was also obtained, and a spontaneous action potential, at intervals of about 1 s, each preceded by a very slow depolarization, but its total height was only some 35 mV. In a later paper (1967) the same authors report that they obtained 60 mV action potentials. In the cricket mesothoracic ganglion Bentley (1969a, b) obtained

I

A

B

Y

-

I"'

"

I

'

'

Y

6

Fig. 4. Electrical activity recorded from the neuropil of the cricket Gryllus carnpestris with acetate or critrate-filled capillary microelectrodes. A. Unitary EPSP's, and one small spike (inset). B. Two IPSP's, and a burst of IPSP's evoked by cercal stimulation. C. Neuron firing closely-spaced pairs of impulses at regular intervals. D. Unit giving bursts similar to the pulse timing of calling stridulation. E. Unit fiing in long bursts a t uniform intervals. F. Unit giving parabolic-type bursts of impulses (burst expanded in inset). G. May be recorded from a sensory neuron. H. Mixed IPSP's and EPSP's. 1. Continuation of H, showing initiation of two spikes. Potential calibration: 5 mV (A-D, H, I); 20 mV (E, G). Time calibration: 200 ms (A, B); 100 ms (C); 500 ms (D, F (Insert)); 1000 ms (E); 2000 ms (F). (From Bentley, 1969a.)

3 66

G. HOYLE

I

B

C

I

I

Fig. 5 . Events during the evasion response to cercal stimulation in the cricket. Upper beam: extracellular recording from a nerve. Lower beam: intracellular recording from the neuropil. A. Upper beam: axon to second basalar motor unit. Lower beam: recording from neuropilar segment of same motor axon. B. Upper beam: 1st promotor nerve. Lower beam: recording from possible inhibitory intemeuron to unit recorded from on upper beam. C.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

367

stable resting potentials from neuron somata, but neither synaptic nor action potentials occurred in them. By contrast, in the neuropil, where he was also able to record some fairly stable resting potentials in the region of 50-70mV, he obtained EPSP’s, IPSP’s and action potentials (Fig. 4). In some cases these were from dendrites which could be identified as belonging to specific motorneurons (Fig. 5A). These results are the most promising yet obtained for permitting a direct examination of neural mechanisms underlying a behavioral act. Kendig (1968), although having less success than Bentley, was nevertheless able to use some of her results on the locust mesothoracic ganglion as an aid in choosing between some possible alternative hypothetical schemes for neural control mechanisms related to flight, thereby directly relating her results to interests of behavior. For the single excitatory motorneuron supplying the locust anterior coxal adductor muscle (see p. 233), I have on two occasions been able to record from the motorneuron in the neuropil with a citrate-filled intracellular electrode. I have also attempted to locate the relevant cell body, and probably have indeed recorded from it without recognition, because insufficient spread of the electrical events into the soma occurs. The action potential in the motor axon is already attenuated when recorded from the neuropil segment (Fig. 6 ) and prolonged in duration as a result of decremental electrotonic spread. Spontaneous activity persisted at about normal rates in preparations consisting of only about onequarter of the ganglion, provided this included the neuropil closest to the origin of the 3rd nerve. To make such a preparation, the ganglion was desheathed, the soma layer cut away, whilst the motorneuron discharge was continually monitored. Then, the neuropil itself was whittled away. All of the contralateral side of the ganglion as well as the ipsilateral posterior lobe and extreme anterior portion were removed without causing permanent cessation of the discharge. Following each cut there was usually a brief acceleration or an inhibition, after which normal firing was resumed. This means that the pacemaker is in the ipsilateral neuropil, not Upper beam: from second basalar and 1st promotor nerve. Lower beam: unidentified unit showing spontaneous spikes followed by prolonged IPSP’s. The change was brought about by cercal stimulation, which led to firing in the nerve. D. Upper beam: mainly 2nd basalar nerve. Lower beam: unidentified neuron showing at first IPSP’s, then EPSP’s. Note that as total depolarization changes, firing rate of motor unit (upper beam) increases. (From Bentley, 1969a.)

368

G . HOYLE

A

B

C

Fig. 6. Intracellular recordings from Schistocerca greguria anterior coxal adductor muscle (upper traces) and from neuropilar segment of motorneuron (lower traces). Note small EPSP’s and IPSP’s. Calibration: vertical, 10 mV; horizontal, 5 ms (A and B); 20 ms (C). Note that spike interval (C) is directly related to mean membrane potential. (From Hoyle, unpublished.)

too far from the point of emergence of the cell. When trying to find a site of origin for this discharge in terms of conventional EPSP’s I have so far drawn a complete blank, however. It may be postulated that the pacemaker is a part of the final motorneuron itself and that this region is not receiving synaptic bombardment directly, but only a number of events of small amplitude and various durations. These

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

369

may be electrotonic (including ephaptic), in addition to being conventionally chemically synaptically mediated, and may include components due to activation of the membrane electrogenic ion pumps. A review of various aspects of synaptic transmission has been given by Boistel ( 1968).

3. Ephaptic Excitation The implication behind Hughes’ ( 1952) interpretation of neuropil functioning was that ephaptic transmission occurred and caused gross excitatory or inhibitory effects. Synchronous activity in many dendrites would lead to unequally distributed current sinks (i.e. “field” effects) which in turn would influence the firing of many motorneurons. My work with 25 p or 50 p leads implanted into locust and cockroach ganglia has shown that d.c. potential differences of up t o 6 mV (see p. 233) occur in neuropil either spontaneously or as a result of sensory stimulation. The larger of these shifts are followed by motorneuron firing. Most of the “slow” motorneurons fire erratically, and some fire continually, with a regular rhythm, so the mean membrane potential of these cells at the impulse initiation zone must be within a few millivolts of the threshold for firing. It is thus quite possible that current flow associated with the neuropilar potentials significantly influences the membrane potential. Nerve impulses in some of the larger axons passing through the ganglion, such as those cercal giants which course right through the me tathoracic ganglion, could also influence impulse initiation within the ganglion. Some interneurons may make electrotonic (tight) junctions with the motorneuron, and fine intercellular bridges could occur between neurons, though none has yet been identified in an insect. Waldron ( 1968) and Bentley ( 1969a) have both presented evidence for possible ephaptic influences between motorneurons which normally act synergistically. Bentley ( 1969a) found that when a given cricket motorneuron fired, an electrode recording intracellularly from one of its synergists registered a sharp depolarization of a few millivolts (Fig. 7). 4. The ‘%eneral” Insect Motorneuron It seems worth while to attempt to summarize the available data on the insect motorneuron and whilst allowing for a probable variety in certain details, to try to distinguish features which are present in a majority of those so far examined and thereby to outline a “general”

3 70

G. HOYLE

I

I

I

Fig. 7. Demonstration of possible electrical coupling between motor axons innervating different groups of muscle fibers in the same muscle. Upper beam: 2nd basalar nerve. Lower beam: neuropilar segment of one of the two units frring in nerve. Spikes in one unit are accompanied by a simultaneous short depolarization in the second unit (indicated by arrows). Neuropilar electrode placed successively in each of the two units in A and B. (From Bentley, 1969a.)

insect motorneuron (Fig. 8). It is safe to conclude that the soma of the “general” motorneuron is electrically inexcitable, and that synaptic terminals d o not occur upon it. Instead, synapses occur on an element of the neuron which runs in the neuropil, and on side branches from this element. The connection to the soma is relatively long and thin so that electrical events occurring in the neuropilar segment are highly attentuated by the time they reach the soma, even to the point of being below the noise level of the recording system. From the major neuropilar element the single motor axon emerges. The junction between the two is an element which is narrower than either the motor axon or the principal neuropilar segment. Somewhere on this narrower section is the impulse initiation zone. This conclusion is reached for several reasons. If the motor impulse is initiated distally and traced antidromically it can easily be followed into the ganglion but then it fades rapidly at about the point where the axon narrows. The spike can be recorded from the neuropil but is

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

371

very small and also has a longer time-course. Intracellular recordings from the spike initiation zone are rarely obtained and extremely unstable, even with a 40 MA2 electrode having a 0.2 p tip. The resting potential starts to decline immediately following penetration and thereafter orthodromic activity ceases.

Fig. 8. Diagram of “general” insect motorneuron. The peripheral axon narrows as it enters the neuropil, then widens into a neuropdur segment (wider segment). This commonly runs dorsoventrally; from it dendritic branches are given off. Excitatory synapses (-) and inhibitory synapses (+) occur on these branches. Three other influences occur: electric currents associated with depolarization in clusters of nearby dendrites; specific ephaptic activity from neighboring axons and neuropil; possible fine cytoplasmic bridges from neighboring cells. The soma is connected to the neuropilar segment by a long, narrow neural filament; all three portions are electrically inexcitable. The impulse-initiation region lies on the narrow portion of the cell which connects the axon with the neuropilar segment.

The damage is associated with the small diameter of the cell at this point. Stable recordings, although difficult to obtain, are relatively easily obtained from the same cell deeper in the neuropil. In this thicker segment of the (“general”) neuron, however, overshooting action potentials are not found. Even when resting potentials of more than 50 mV are obtained from the region an overshooting spike is not seen when an action potential occurs in the motor axon. The size of the response in the segment may be as small as 12 mV. The conclusion to be drawn from this is that the neuropilar segment, like the soma, is electrically inexcitable. Furthermore, owing to the geometry of the neuron, with a widening in the neuropil, rapid attentuation of the action potential, which is spreading antidromically and electrotonically, occurs. The further attenuation along the narrow connection to the soma is apparently

3 72

G. HOYLE

sufficient to reduce the size in the soma to a fraction of a millivolt only, in smaller neurons. 5. Background Activity Most insect motorneurons are silent, in preparations, unless reflexly activated or unless there is a “spontaneous” escape struggle. However, many slow axons and also inhibitory axons are often active, and some are continually active. Anne Knights ( 1965) split segmental nerves of dragonfly nymphs and was able to record the electrical activity in several different motor units. Although some remained silent for long periods many had a constant background discharge in the absence of any specific stimulation applied to the preparation. The frequency did not remain fixed, for a particular unit, but showed accelerations and depressions, apparently as central fluctuations occurred. This may be characteristic for “slow” motorneurons concerned with maintaining postural tone. One of these, a specific motorneuron, the single motor axon which supplies the metathoracic anterior coxal adductor muscle of the locust Schistocerca gregaria, has been recorded from over periods of several hours in preparations ranging all the way from the intact, free-moving insect, with implanted, trailing leads, to the fully isolated ganglion-nerve muscle preparation (Hoyle, 1965, 1966a, b). This neuron has a more-or-less fixed basic rate of firing, at about 8 per s which occurs continually, though it is subject to rises and falls or even brief cessation, dependent on changing input or spontaneous central fluctuation. About 13 years ago I started making recordings-sporadically, for I find it a tedious business-from various muscles of intact locusts and grasshoppers which were free to move or “behave” in a fairly large cage (Hoyle, 1957, 1964). The simplicity of the motor innervation permits a quantitative comparison of the output patterns associated with movements so that a number of significant questions can be asked about the underlying mechanisms. The same approach has enabled recordings of motor activity t o be made during enactment of three different kinds of song in crickets: calling, aggression and courtship (Ewing and Hoyle, 1965; Bentley and Kutsch, 1966). Such recording provides an important “window” into the nervous system of an intact organism. Simplified bits of tieddown organism are important in studying neural organization and function, but they are no substitute for knowing what happens in the intact animal-if this can be found out. The technique has recently been developed t o a

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

373

remarkable extent by Howell Runion (Runion and Usherwood, 1966; Galloway et al., 1966), who has hooked up both nerves and muscles to a micro-miniaturized solid-state amplifier which the locust carries around on its back (Fig. 9). This is enabling a study of the role of the inhibitory axons in locomotion which could be achieved in no other way (Runion and Usherwood, 1969). In correlated, independent work on more conventional “preparations”, Runion and Usherwood (1 968) found that sense organs on the tarsal spines mediate a reflex excitation of the metathoracic extensor tibiae muscle via the “slow” axon provided at least four spines are moved simultaneously, and provided the general level of excitability is high due to other afferent inputs. In addition, the inhibitory axon innervating the same muscle is excited at the same time by bending the spines. The excitatory action promotes long tonic discharges, with a brief phasic burst if sufficiently vigorous, whilst the inhibitory consists of only two impulses with rather specific timing which in this case is late in relation t o the burst so that its net effect will be to promote more rapid relaxation. A study of the role of spontaneous and evoked discharges of the inhibitory axon supplying the me tathoracic anterior coxal adductor muscle of locusts (Hoyle, 1966b) suggested that its major functional role is to aid rapid relaxation of a muscle rather than to modulate tension or actually inhibit its development. In spite of spending many hundreds of hours watching locusts behave and at the same time watching the traces of their electromyograms, I d o not feel the observations enable me to contribute, except in the most general terms, to the analysis of underlying neural mechanisms. One very quickly obtains an enormous amount of recorded data, as pointed out above, and this takes a very long time to examine critically. It might be possible to build up nice “stories” by carefully selecting from the data, but these must surely be prejudiced by ideas about control mechanisms. Reflexes have turned out to be extremely difficult t o detect in spontaneous behavior of the freely-moving insect and I am at present of the opinion that their role is much less important than was once believed. This largely subjective impression is nevertheless in line with the experience of others, as has been pointed out, and will be further emphasized below. Reflexes may be concerned, apart from simple evasive and defensive acts, in which they are paramount, to perfect movements initiated by the central machinery, and to maintain the nervous system in a state of excitation adequate to the continuance of motor A.I.P.-13

374

G . HOYLE

Fig. 9. Methodology for recording simultaneously nerve activity, electromyograms and movement of a joint from an intact, free-walking insect. Upper: stippled area on grasshopper’s back denotes multichannel preamplifier carried by insect. A and C show general positions of recording leads and movement monitor (coils attached to limbs). B shows sites for placing electrodes on exposed nerves. (From Runion and Usherwood, 1966.)

activity. There are, however, also central mechanisms for achieving the latter. A hungry locust will explore and “march” for an hour or so, without sensory reinforcement other than that which results from its own movements, not simply because its gut is empty, but after a stimulus which may be brief, and as simple as a small environmental temperature drop, which need not be maintained (Hoyle, 1964). The

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

375

reaction has a counterpart in the wild and may have adaptive significance. A well-fed locust will not move at all when subjected to the same treatment, so that there is no comparable proprioception, but there is still a raised frequency of discharge in tonic motorneurons which follows a similar time-course, suggesting that a long-lasting central nervous “driving” activity has been initiated by the stimulus.

C. THE ELECTRICAL ACTIVITY OF NEUROPIL 1. Extracellular Unit Recording Penetration of ganglia with microelectrodes capable of giving intracellular recordings has yielded significant new results as a result not of intracellular, but of extracellular pick-up. Maynard ( 1956, 1967) obtained spike potentials of surprisingly large size (up to 30 mV) from the cockroach cerebrum. Potentials of such size and speed (total duration about 5 ms) could only have occurred as a result of a large propagated disturbance occurring synchronously in a large number of nerve cells. The optic ganglia of locusts yield single unit activity associated with both second- and third-order synapses (Burtt and Catton, 1960; Suga and Katsuki, 1961). These units fire in response to rather specific changes in the visual field. Burtt and Catton have now been able to trace these responses into the thoracic ganglia and should eventually be able to assess their motor effects. In the cricket brain, microelectrode recording during light flash stimulation (Dingle and Fox, 1966a) revealed four different kinds of units. Some fire continually in the light at a frequency related to light intensity, whilst others are their inverse, i.e. they fire at higher frequencies in lower light intensities. There are also “on” and “on-off” phasic units. They also found units in the brain which fired during stimulation of the cerci (Dingle and Fox, 1966b). Again, it would help if the roles of any of these units in behavior could be determined. Analyses of the motion detection mechanism in the optic lobes and brain of flies have been made by Bishop and Keehn (1 967) and by Bishop e t al. (1968). These are important in regard to understanding sensory information processing, which in turn must result in commands being passed on to motor centers. Vowles (1964) recorded the activity from several units in the bumblebee brain. Some corpora pedunculata cells have activity

376

G . HOYLE

which is closely associated with movements of the antenna or mouthparts. Some very interesting cells fired when eyes were illuminated or antennae moved, but with a variable latency as long as 2 min. This latency was reduced by repetitive stimulation, but the cells then gave an afterdischarge. Attempts to record from the neuropil were paqially successful, but were not correlated with movements. These results are typical of the state of affairs in the field. They indicate that successful experiments are possible with the most highlyevolved, sophisticatedly-behaving insect there is, and yet they do not give us a single positive bit of information which helps to understand the brain machinery. Of direct value in the interpretation of behavioral acts is the recent work of Roeder and Payne (1 966), who have been able to locate a second-order interneuron relevant to the input from the auditory organ of a moth. This gives a single spike or sometimes a close pair in response to behaviorally significant bat sounds. It is not only clearly an integrator cell, for multiple discharges from a few units converge upon it and call forth only this simple output, but also perhaps a “command” unit. That is, it may call forth specific behavior, in this case avoidance action, on the part of the moth. It is extremely desirable that recordings be made from single ganglionic neurons during cyclically repeated behavior, such as flight movements, and Wilson’s collaborators have made a number of attempts, regrettably largely unsuccessful, to d o this during tethered flight of locusts. Iwasaki and Wilson ( 1966) did obtain an EPSP from the neuropil with a glass electrode when they stimulated the sensory nerve from the wing. The activity of this nerve is not directly concerned with establishing patterned output to the wing muscle, but it can modify the central “oscillator” which does, and so its effects are of considerable interest. Input from the wing-hinge stretch receptor, which is a single unit, summated with it. When the summed EPSP’s were large enough they gave rise to a spike. What kind of nerve the spike arose in and what its motor effects were, if any, was not disclosed. The activity appears to have been picked up extracellularly, from a site close to the synapse. Maynard’s recent work (Maynard, 1967) shows one of the many ways in which it is desirable that future investigations should proceed. Microelectrodes were passed through the cockroach brain at several stereotactically-measured points and responses obtained at various depths, during the passage of a single afferent volley in the antenna1 nerve. Electrical potential contours were plotted for various

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

377

time intervals following the stimulation, permitting a study of the passage of the excitation wave through the brain. Early, brief components with only 10-20 ms latency were followed by more prolonged, more slowly-conducted components with 30-40 ms latency. The latter produced activity in the glomeruli of the antenna1 lobe (Fig. 10). Subsequently, activity spread throughout the ganglion. The optic tubercle, calyx, stalk, and corpora pedunculata all have characteristic spikes. Probably the wave originates in the calyces and travels down the peduncle to terminate in the a and /3 lobes. Although it may fragment, the wave commonly behaves in an all-or-nothing manner. Either there is a very synchronous discharge in a large number of axons, or there must be large axons. The latter seems unlikely on the ground that the conduction velocity is only 0.2 m/s. The initiation of such a synchronized burst requires either an equally synchronous input, or perhaps an input from a single fibre. The globular cells must interact with each other subsequent to their excitation, as suggested by Maynard in the diagram shown in Fig. 11. The existence of such synchronous activity by parts of the corpora pedunculata which have been thought to be intimately involved in the most complex behavior of which an insect is capable, nevertheless must seem paradoxical. This work, taken in conjunction with other studies, such as that of Roeder on the moth evasion response, and ones to be described below, point the way towards a general understanding of one of the simplest types of innate behavior, the escape response. This seems to be dependent on certain rapidly reacting cells in specific pathways which are all caused to be stimulated when a single key cell, sometimes called a “command” cell, which also must be the convergence unit for specific input, fires what may be a single impulse, in some cases, or a brief, highfrequency burst in others. Tungsten microelectrodes have been inserted into neuropil of the locust tritocerebrum by Rowel1 and Horn (1967). Specific single units responding to complex visual inputs were found. For example, a disk moving upwards gave a response in one unit, but not the same disk moving downwards. The protocerebrum of Lepidoptera has been explored with extracellular microelectrodes by Blest and Collett (1965), during stimulation of the visual field. They found certain units which only fired when specific, fairly complex input was given. Such units are now known for cat (Hubel and Wiesel, 1959) and frog (Maturana et al., 1960) visual systems. It is a pity that so little interest seems to be paid by visual sensory physiologists to the motor

378

G . HOYLE

Fig. 10. Spread of electrical activity in cockroach cerebral ganglion following a single afferent volley in the antennal nerve. Recordings made in the middle saggital plane. The numbers above the records correspond to the paths made by the electrodes as indicated in the diagram below. a, alpha lobe of corpora pedunculata; AG, antennal glomeruli; GC, globular cell bodies; LC,lateral calyx; MC, middle calyx; 0, optic commisure; P,peduncle of corpora pedunculata. (From Maynard, 1967.)

Fig. 11. Diagram of corpora pedunculata illustrating possible sites of synaptic electrotonic interaction. Cells shown are globule cells located in the cup of the calyx, sending axonal processes down the peduncle to terminate in the Q and p lobes, and dendritic processes receiving afferent intemeurons (arrows) in the calyx. Depolarization originating there summates with local currents in neighboring globule cell axons, resulting in a synchronized spike. (From Maynard, 1967.)

380

G . HOYLE

consequences of successful sensory input. That motor activity can be obtained by visual signals in an insect preparation other than the familiar optomotor response has been shown by Chapple (1966) in studies on the milkweed bug Oncopeltus. In his experiments leg motorneurons which were normally silent in the preparation gave a continuous “tonic” discharge when the light intensity was changed sufficiently.

2. Recording from the Neuropil of Intact Insects The key to breaking open the problems outlined above is clearly to develop methods of recording from specific neuropilar sites. If this can be done with extracellular electrodes, so much the better, for then it should be possible to record from a relatively unrestrained, intact insect which can be expected to show some of its subtler and more interesting behavioral repertoire. I have made some exploratory studies in this direction, the results of which were reported at the symposium on comparative neurophysiology held in Tokyo following the 23rd International Congress of Physiology. Since neither these, nor more recent studies have yet been published, I am taking the liberty of including some of the results here. Initially, a muscle was chosen which could be recorded from relatively satisfactorily in the intact, free-moving situation with trailing leads (Hoyle, 1964). Its motor nerve was then located in the muscle and stimulated electrically. The antidromic pathway was traced into the neuropil by probing with a capillary glass microelectrode. Originally this method was used in an attempt to locate the cell body of the relevant motorneuron in the manner of Eccles. Usually, a point was found in the ipsilateral neuropil where a good extracellular record could be obtained of an antidromically-invading spike. This point was noted, and was approximately the same in different preparations. Now a fine factory-insulated 25 p platinum wire was passed through one of the minor tracheae entering the ganglion just anterior to the neuropil site in such a manner as to come to rest with its tip as close as possible to this site. A small platinum plate with a wire attached was inserted into the thoracic cavity, t o form an indifferent electrode. The wires were then cemented into the cuticle and the thoracic cavity, opened to permit insertion, sealed up. Pairs of insulated fine copper wires were placed through the cuticle overlying the muscle, so as to just touch the muscle. In some experiments a few such pairs were placed in joints of different legs. The leads were collected together and the insect either

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

38 1

held on its back but with all its legs free, or released completely but restrained by the wires. Such experiments have been carried out on Schistocerca gregaria and Periplaneta americana, with the objective to simply record electrical activity in selected parts of the neuropil. But in addition, when a promising site had been located, motor activity was evoked either by natural stimuli or by electrical stimulation of the nerve cord after implanting additional leads through the abdominal cuticle. The experiments require great patience, partly because obtaining any results at all is a lottery with odds of about 4-1 against. Such odds would be marvellous for a lottery but are not satisfying to an experimental biologist. Those were the odds against getting any result for a day’s work. The odds against obtaining similar results in a different insect, with the neuropil electrode, were very much higher. I shall refer to the electrical activity obtained from the neuropil as an electroneuropilogram, or ENG. Although the aim was to obtain localized rather than overall electrical activity, the precise source of the recorded activity could not be established. Possibly activity is recorded, using this method, not only from the specific site, but to some extent from the rest of the ganglion. This is mainly below amplifier noise level, however, and lost. Neither the electrocardio. gram nor the myograms of thoracic muscles were picked up in most experiments. Figure 12 shows a typical record of electrical activity obtained from the metathoracic ganglion of P. americana from a neuropilar site into which antidromic potentials from the “fast” axon supplying the metathoracic extensor tibiae could just be traced. It lay just dorsal to the mid-line and just anterior to the center of the ganglion on the ipsilateral side. All these discharges were “spontaneous”. The largest spikes in the record from the muscle (EMG) were due to firing of the “fast” axon supplying the extensor muscle. The second largest spikes were due to cross-talk from a “fast” axon response of a unit of the flexor tibiae. The small potentials occurring, if at all, in long trains, were due to the “slow” axon supplying the extensor. Small oscillations occur in the ENG record preceding each “fast” axon response in the muscle and these are presumed to be due to the “fast” motor axon itself, but there are also a lot of other small bumps in the record. In addition to these potentials, a slow negative wave was found t o occur at some active neuropilar site preceding and overlapping each burst of muscle potentials. The slow wave was never completely smooth, but only when its mean value exceeded a A.1.P.-

13.

382

G . HOYLE

emg

Fig. 12. Electromyograms (EMG) obtained from the left metathoracic extensor tibia (LMET) of the cockroach P.arnericana at the same time as focal activity in the neuropil-electroneuropilogram (ENG). The insect was intact and only partially restrained. All discharges shown occurred spontaneously. Largest action potentials in EMG occurred when single extensor “fast” axon to LMET fired. Second and third largest were due to flexor “fast” axons, smallest to “slow” extensor axon. Several extensor-flexor stepping sequences occurred in the fust five and the last discharges. Note similarity of neural “programs” in €ourth and fifth discharges. Otherwise there are marked differences. (From Hoyle, unpublished.)

I

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

383

characteristic height did any activity ensue in the muscle. The most probable interpretation of this form of ENG is that it is due to summed small EPSP’s (and IPSP’s) and perhaps other sources of depolarization , such as ephaptically transmitted current. On a few occasions the character of the potentials occurring at the neuropilar electrode was quite different from the above. In the most interesting of these, potentials occurred spontaneously which were of nearly constant height and shape and which were probably EPSP’s. Their size was up to 4 mV, so the electrode must have been quite close to their site of origin. They appeared sporadically, during silence in the muscle, but from time to time appeared in little bursts of four or five in quick succession, effectively at about 20-3O/s. Each EPSP had a duration of almost 200 ms, so at these frequencies they also summated sharply. The summated responses of two of them were always followed by one or two closely-spaced action potentials in the muscle due to the “fast” extensor axon. Longer bursts were associated with more extensor potentials. Sometimes they appeared in a long series of brief bursts (Fig. 12). The decay rates were faster than can be accounted for on the basis of the passive decay constant of single EPSP’s. There seems little doubt that these resulted from the presence of inhibitory postsynaptic potentials (IPSP’s). Following such a burst, as in the example shown, the EPSP’s were about twice their normal size. Thereafter they slowly and progressively declined, both in height and in frequency, taking two or three minutes to return to normal. The increased magnitude is due to a facilitation process which decays remarkably slowly. It is of great interest that these EPSP’s did not give rise to action potentials within range of the site giving rise to them. Spikes of small amplitude occurred superimposed upon them, and correlated with them in time, but were never apparently initiated by them. The probable occurrence of IPSP’s is reinforced by occasionally finding a site at which not only negative-going, but also positivegoing potentials of similar shape, but slightly slower time course, were obtained (Fig. 13). A perfect matching of the shapes of EPSP’s and IPSP’s, even allowing for a possible slower time-course of the latter, was not obtained. This suggested that the sites of their formation are spatially separated. This is what must be expected in neuropil, where excitatory synapses and inhibitory synapses occur on a thin process rather than on a nearly spherical unit as in the spinal cord. At the neuropilar site of Fig. 13, although one EPSP site dominated the summed depolarizing response which occurred

3 84

G . HOYLE

Fig. 13. Spontaneous and evoked ENG activity related to EMG activity in MET of P. americunu. Time in seconds. (a) Spontaneous leg kicks in an insect held upsidedown but with its legs free. Small presumed EPSP's started to occur in the ENG record and abruptly increased in frequency, summating to give a peaked record, but with no evidence of spikes. The peak was followed by a brief period of silence, then a series of bursts. The decay of the summed ENG potential in these records was faster than the natural decay time, an effect which suggests the presence of IPSP's. Each burst is associated with EMG activity. Note that the EPSP's facilitate markedly (about 60%) and that the facilitation decays very slowly. (b) Another preparation, in which the giant axons of the ventral nerve cord were stimulated electrically. Note the burst of EPSP's, outlasting the stimuli, and the slow decline in the afterdischarge of these EPSP's. EMG activity also outlasted the stimuli. (c) Spontaneouslyoccurring presumed EPSP's and large junctional potentials of opposite sign in a preparation similar to (a). A natural stimulis in the form of a very brief puff of air was given at the arrow. Note the summated ENG record and associated burst of activity in the leg. All records read from left to nght. (From Hoyle, unpublished.)

following a natural excitatory stimulus, in this case an air puff applied to the anal cerci, other, smaller contributory units could be recognized. These were both EPSP's and small IPSP's. The latter were ineffectual. The cell giving the big IPSP was apparently itself inhibited during the excitatory burst, but returned to its sporadic firing immediately afterwards. The basic EPSP rate was only about 1 per 15 s before the puff. It rose to an estimated maximum of 1 l/s as it

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

385

summated in the response period, and then declined to its background rate in 12 s. The output in the metathoracic extensor tibiae “fast” axon (METFA) was well correlated with the summed EPSP’s in a general way, but there was no precise relationship. It might be that the ENG is purely incidental to the true synaptic events occurring in the reflex arc mediating the extensor thrust of the escape response, but this seems rather unlikely. Nevertheless, it is evident that we are not dealing with a chain of neurons, each exciting the other in turn with a simple input : output relationship. Rather, what we have is a situation resembling that occurring in a sense organ such as the crayfish stretch receptor, where multiple inputs summate to give a generator potential that forms a platform of depolarization such as is caused by a long rectangular pulse of current. When this exceeds the threshold for firing impulses in the impulse initiation region of the motor axon, action potentials are fired at a rate depending upon the motor axons’ intrinsic properties and the magnitude of the current. The large EPSP’s in the example under consideration appear to be developed in an interneuron of some kind, not the final motor cell, and this interneuron does not give spikes. The summed EPSP’s could be depolarizing the final motor cell by ephaptic current. In some recent experiments on locusts I have found that the firing of impulses to the metathoracic anterior coxal adductor (MACAM-see p. 241) continues whilst the ganglion is desheathed and whittled away until only about onequarter of the anterior ipsilateral neuropil alone remains. In these studies the surface layer of cell bodies was progressively removed, and although this drastic operation caused temporary cessation, the regular discharge eventually returned. Exploration of the neuropilar mass with intracellular electrodes is also being attempted, and has revealed the probable site of origin of MACAM impulses. This is a part of the dendritic apparatus of the motor axon itself. Motor antecedent activity has not been found. The rate of firing is influenced by steady shifts of membrane potential rather than discrete EPSP’s and IPSP’s. A number of similarities may be noted between aspects of these results and those obtained with glass “intracellular” electrodes by Bentley ( 1 969a, b). The most significant is the absence of stable, high-frequency impulse activity in the neuropil which was to be expected on the neurally-driven reciprocating oscillator model. Rise-times and decays of all forms of change of potential are rather slow, the fastest activity being that which occurs in the motor-

386

G. HOYLE

neurons. Neuropilar dendrites d o give high-frequency trains, but only when damaged by the microelectrode tip. It was hoped that ENG records might be obtained representing activity in the “driver” neurons exciting motorneurons and thereby providing clues to the origin of patterned outputs, and it is worth while examining the possibility that those obtained and presented in Figs 1 1 and 12 actually represent “driver” or “command” activity. They are certainly occurring in an interneuron, and one which does not apparently give spikes. If this is so, the driving is produced not by an interneuron firing at a high frequency and giving synaptic potentials on the motorneuron, but from a depolarizing “bias” voltage. The EPSP’s recorded are not only just as well “coupled” to the flexor tibiae motorneurons as to the extensor, they are also equally intimately associated in time with impulses in other muscles of the same leg, and with those of the contralateral leg and even of legs of other segments. In another experiment (Fig. 13(c)) a site showing EPSP’s was obtained and a pair of stimulating electrodes was placed in contact with the ventral nerve cord. Weak electrical stimuli applied to the cord gave rise to impulses only in the giant through-conducting axons which elicit the startle response initiated in life by air puffs on the anal cerci (Roeder, 1948). The ENG potentials obtained were femarkably dimilbr to those given in response to the natural stimulus. Two similar excitations gave almost identical responses at all levels, including similar ENG’s. Further repetitions gave the familiar habituation phenomenon, which will be discussed below. The potential shifts associated with the activity in the neuropil, together with the absence of synaptic activity neatly correlated with motor output, recall the predictions of Hughes (1 952) arising out of experiments with d.c. applied to cockroach ganglia. Their presence does not necessarily support his view that local, or widespread potential fields are important in the neuropil, but it helps to explain why d.c. fields have such interesting effects in causing reciprocallycoupled movements of antagonistic sets of muscles. The latter could, perhaps, be driven normally by summed, slow EPSP’s occurring in interneurons with which they synapse and be electrically rather than chemically synaptically mediated, by a set of specific electrotonic junctions rather than some kind of diffuse “field” effect. These junctions would be excited by extrinsic current flow. A decade ago Roeder (1959) measured the delay between an air puff and the onset of movement in the startle response, obtaining a

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

387

range of from 28-90 ms. The lowest value is what interests us in attempting to prescribe the probable number of synapses interposed in the metathoracic ganglion between the giant fibers and the motorneurons. The delay in the cercal nerve and in the synapses between cercal afferents and giant fibers is about 1.5 ms. Conduction up the giants to the metathoracic ganglion takes 2.8 ms. Conduction from the ganglion to the muscle takes a further 2.5 ms. Thereafter neuromuscular synaptic delay occurs, followed by excitation contraction coupling. The latter is much slower via the “slow” axon than via the “fast”. Normally, the startle response involves the firing, initially, of the slow axon, followed by a brief burst in the fast axon. It can be initiated by fast axon impulses occurring at the same time as slow, but this is very rare. When it does occur, there is still an 8-10 ms latency before detectable tension develops. The sum total of minimum, irreducible delays, excluding the synaptic events in the metathoracic ganglion, is thus 17 ms. Metathoracic events therefore occupy about 1 1 ms. Roeder concluded that it may be 8-10 ms. This is a long time from the point of view of an animal interested in survival, and it should be borne in mind that these are the minimum values. Maximum times are up to 10 times longer. It demonstrates conclusively that the synaptic integrative events in the ganglion are complex. The time is sufficiently great that as many as three synapses-two interneurons, must be postulated. Roeder has on several occasions (e.g. 1953) illustrated the abdominal giant to motorneuron synapses as being monosynaptic, but it cannot possibly be of this kind. The minimum central synaptic delays occurring in insect ganglia are, in general, surprisingly long. There is, in fact, no established instance of a monosynaptic reflex arc; the shortest known latency is 10 ms, for the proprioceptive reflex arc driving the Periplunetu extensor tibiae (Wilson, 1956a). When one considers the number of different actions in which each single motorneuron is called upon to act with precision, the complexity and lability of reflex responses is not so surprising. D. MECHANISM OF HABITUATION

The startle reaction is one which has attracted attention in a specifically behavioral context as an example which permits the study of habituation to a constant stimulus in a quantitative manner (Roeder, 1963). Habituation is usually classified as the simplest form

388

G . HOYLE

of learning (Thorpe, 1963). Very few attempts have been made to investigate the physiological processes which underly the habituation phenomenon and n o satisfactory explanation of it has been offered. Fatigue of sensory processes can be ruled out very simply because the habituated response is immediately restored by the simple expedient of applying a different “neutral” form of subthreshold sensory input at the same time as the one which has become habituated. This is termed the dishabituating stimulus. There is some evidence that at least part of the process occurs in the last abdominal ganglion in the case of the cockroach startle reaction (Hughes, 1956b), but the metathoracic ganglion must be the immediate site of the major process, perhaps with the intervention of head ganglia. The preparations utilized in the present work permitted a preliminary examination of the physiology of habituation. If, as has been suggested, the summed EPSP’s constituting the ENG directly excite the motorneurons, it follows that habituation could be a result of their attenuation. Habituation occurs readily to stimulation of the cerci by air puffs. This is illustrated by the sequence shown in Fig. 14. The heights of individual EPSP’s in these sequences remained unchanged, but the frequency and durations of the bursts evoked by successive, identical air puffs repeated once every 30 s, declined

L

I

I

r .

Fig. 14. Habituation of the response to an air puff. Four exactly similar, brief air puffs were applied at the arrows, at intervals of 1 min. Note the marked decline in both the EMG, recorded from the right metathoracic femur, and the ENG, recorded from a site giving a perceptible response to antidromic firing of the “fast” axon supplying RMET.Calibration: 0.5 mV. (From Hoyle, unpublished.)

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

3 89

markedly after the first two responses, resulting in smaller total depolarizations and reduced motor output. These experiments suggest that habituation of the startle response results from a reduced frequency of firing of the interneuron which is responsible for the ENG, but they d o not give any indication as to the reason for this; it could be due to sensory fatigue or to events in the last abdominal ganglion. Both sources can be ruled out by stimulating the giant fibers in the cord directly and this was achieved by implanting leads connected to an electronic stimulator through the abdomen so as t o touch the nerve cord. Experiments of this kind showed that when bursts of impulses were directly initiated in the giant fibers, equal in numbers of spikes and mean spike interval to those occurring in response to the natural stimulus of a puff of air applied to the cerci, they were subject to habituation of a similar kind. There was also an associated similar decline in ENG. Habituation is associated with a marked reduction in the number of the interneuron discharges. In the experiments shown in Fig. 15 a large ENG was obtained by stimulating the cord at just above threshold strength, for 1 s. Both ENG and the evoked discharge in the leg muscles were large for the first seven excitations, but declined rapidly at the eighth and ninth. Then abruptly, the tenth was missing completely. A reduced ENG such as might have been expected to occur on the tenth did appear, but 1 s late. The eleventh to thirteenth stimuli likewise failed to cause responses. In an experiment in which only single shocks were given (Fig. 16) at intervals of only ?4 s, brief evoked ENG’s were obtained continually without any trace of reduction in size. Each was associated with a brief discharge of impulses in leg muscles. The activity of the “fast” axon supplying the MET is conspicuous in these records. In an air-puff-induced response, or one evoked by bursts of stimuli, the “fast” axon response quickly disappears but this is not the case for single shocks. It seems that simple fatigue processes can be ruled out as causes of habituation at all levels: sense organ, abdominal synapses, giant axon, ENG formation, excitation of motorneurons. Nor is there any sign of a collateral inhibitory action affecting the same sites as those which give the ENG. It is the neuron which is responsible for the ENG recorded which is influenced, and this is perhaps itself affected by an inhibitory interneuron also excited by the giant fibers, but activated to a smaller extent, so that it is ordinarily ineffective against the excitatory one. It may be postulated that when the inhibitory in terneuron is stimulated by repeated, prolonged bursts it becomes

390

G. HOYLE

ENG

A

ENG

Fig. 15. Habituation of the response to burst of electrical activity applied to the ventral nerve cord. Upper trace: monitor record of stimuli. Second trace: ENG recorded from a site giving a perceptible response to antidromic fUing of the “fast” axon supplying RMET. A. With reducing intervals between stimulus bursts. B. Equal intervals between stimulus bursts. (From Hoyle, unpublished.)

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

391

* -ENG

1sec

J ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ " " " " " " " " ' " " " " " " " ' " ' ' ' ~

Fig. 16. Non-habituation of the response to single shocks applied to the ventral nerve cord. The shocks were applied at just threshold strength for stimulating the giant axons, and the electrical response monitored on an event-marker channel of the oscillograph (upper trace). Some stimuli just failed to evoke a response, when there is no resulting ENG (second trace) or EMG (third trace). Large spikes in EMG are due to "fast" axons supplying RMET. Time in seconds. (From Hoyle, unpublished.)

effective in suppressing the excitatory interneuron giving rise to the ENG. With a mechanism of this kind disinhibition might work either by increasing the total excitatory input to the interneuron which gives rise to the ENG or by suppressing the neuron which inhibits it. When the inhibitory interneuron circuitry is strongly excited, it can continue to suppress the habituated response for several minutes. To account for this it is necessary for it either to be capable of prolonged discharge naturally, or to form part of a reverberatory circuit. A tentative scheme by which the experimental findings on the ENG associated with motor activity could be understood along these lines is given in Fig. 17. The principal feature of the model is an inhibitory collateral pathway, and it should be noted that a somewhat similar model has been proposed by Horn (1967) to explain the general features of habituation in vertebrate nervous systems. In the present model it is postulated that on the pathway of excitation from the abdominal giants to the leg motorneurons there is interposed at least one large interneuron, which is drawn stippled in the figure. An interneuron is required to explain both the present findings and the older observations on the minimum central reflex time. An inhibitory collateral hypothesis was proposed to explain habituation of the escape response of the crayfish and subjected to testing by the use of the crayfish inhibitory blocking agent picro toxin (Krasne and Roberts, 1967). The drug did not affect the process so the authors

3 92

G. HOYLE

Fig. 17. Diagram of one possible interneuron circuit which might give the kind of results obtained in the conduction of excitation from the giant fibers to the leg muscles, and which allows for habituation and dishabituation. Filled triangles represent inhibitory synapses; open triangles represent excitatory synapses. “Tongues” from extensor and flexor represent branches making lephaptic contact with the large interneuron (shaded) responsible for the ENG recorded.

concluded that some other mechanism besides collateral inhibition should be sought. Habituation was detected by Horridge et al. (1965) whilst recording from single units in the locust brain in response to both optic and auditory input. But their results do not permit any consideration of mechanism and indeed, these investigators end their report by issuing a general counsel of despair about ever analyzing the neuropil. E. MEMORY AND “LEARNING” MACHINERY

The more beautiful examples of learning in insects, maze-running, memorization of visual clues related t o the nest, etc. d o not readily lend themselves to neurophysiological analysis. In an effort to provide something simpler Horridge (1962), with the aid of his students, set out to see if insect ganglia can be trained by simple avoidance-conditioning procedures. Headless cockroaches and locusts were suspended above a fluid-filled dish connected to a stimulator with a return lead to the body of the insect, so that if they extended a specified foot into the dish they received a shock through the tarsus. Precisely what kind of stimulation the insect received was not

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

393

determined, nor is it perhaps important. The result, as amply confirmed by a number of amateur and professional scientists since, is that in many examples-though there are some very stubborn exceptions-after only a few “errors” punished by shocks the headless insect holds its tarsus well away from the dish for quite a long time, even for many hours. Control headless insects connected to the test animals in parallel, and therefore receiving similar numbers of similar shocks but not correlated with leg position, failed to “learn” to hold their leg up and took just as long as untrained animals to learn to do so when switched to a coupled program. Eisenstein and Cohen ( 1965) whittled the cockroach preparation down to the level of a single ganglion and the training procedure still worked. They have even mixed the blood of trained animals with untrained ones to see if the former contain any “factor” which is transmissible. The results were inconclusive. I was attracted to the possibilities the preparation seemed to offer for a biophysical study of modifiable behavior. The leg may be extended (this is the “error” which leads to its receiving a shock) for one of two reasons, physiologically speaking. Either the firing rate of relevant flexor motorneurons decreases, or that of extensor motorneurons increases. By implanting leads in the leg muscles of cockroaches which had been trained to hold the leg up an increased frequency in some levators was detected, but the results were variable, and several muscles seemed to be involved. A simpler result was obtained in locusts, where one muscle in particular, the anterior coxal adductor (ACAM), always had a relatively high frequency of firing after training. This muscle lifts the whole leg upwards by rotation around the coxal joint. It is not necessarily the only muscle involved, but it did present a possible test object for making a learning preparation. It is innervated by a single excitatory axon which fires continually at about 8/s, maintaining a tonic contraction in the untrained animal. After training it fires at higher frequencies, up to 30/s, and maintains the muscle in a state of strong contraction. The necessity for proprioception in the learning process was eliminated by fixing the leg as a whole, severing the ACAM apodeme, and holding the cut stump in a force transducer device. Under these conditions, as in the intact insect, the single excitatory axon supplying the muscle fires continually at about 8/s. These impulses were monitored, not in the nerve, which also contains an inhibitory axon that fires sporadically, but in the muscle itself, with the aid of an intracellular electrode. An electronic interval measuring and

394

G. HOYLE

averaging device (in the latest procedure) follows the discharge. If the pulse interval falls below a pre-set value, say 166 ms (6/s) for three or four pulses the device trips a relay switch which causes a shock to be delivered to the leg. The pre-set value was termed the “demand” level. The behavioral argument used is that if a fall in the frequency of firing of this magnitude had occurred in a locust suspended over a dish, the contraction in the ACAM would not have been sufficient to counteract the combined gravitational and abductor force and so the leg would have lowered, thereby resulting in a shock being received in the Horridge-type experiment. The shock is followed, not surprisingly, by a little burst of impulses to the ACAM. Thereafter, the frequency returns to the previous mean but again fluctuates and sometimes falls below 6/s. It then gets another shock. After only a few such experiences the mean rate is found to have risen, usually to about 12/s, and shocks are not received for quite a long time, if even at all, and rarely within the life of the preparation (Fig. 18). Sometimes a headless insect preparation will hold its leg up for several hours after only a few training shocks. The interval monitor in the above experiment may now be reset to a different demand level of, say, 80 ms interval, resulting in several shocks being received by the preparation. Again the results follow the same course, but now the mean frequency settles down at the much higher value of about 20/s. In some preparations the mean frequency has been sextupled by progressively raising the demand level. This may be done very slowly, taking a few hours, or inside a

0

I

1

I

I

1

1

10

20

30

40

50

I

I

I

1

1

60

70

80

90

I00

min Fig. 18. Progressive rise in the firing frequency of the motor axon supplying the anterior coxal adductor of Schistocerca gregaria in response to electric shocks applied to the leg which were correlated with falls in the mean frequency of firing below a predetermined “demand” level. The “demand” level was raised a total of three times. (From Hoyle, 1965.)

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

.

,

3 95

-- _-- _. - - aclwlty after

'learning '

Fig. 19. An experiment in which shocks were given each time a maintained fall in frequency of fuing of the S. greguria ACAM occurred. The mean frequency of firing was caused to rise from 6 per s to over 20 per s in only 4 min. (From Hoyle, 1965.) Shocks were given to the leg at arrows.

matter of about 10 min, depending on how quickly the demand level is re-set. An example of a rapid rise is shown in the experiment of Fig. 19. Equally interesting experiments have now been done in the reverse direction, giving shocks when the frequency rises, i.e. with demand levels of progressively shorter intervals. The net result then is a maintained decrease in frequency. The preparation trained to fire continually at a high frequency can be brought back to a normal rate of firing or even to zero firing rate, by this reversal of the training procedure. Of course the results are not always as pretty as this description perhaps somewhat falsely conveys; sudden inhibitions occurring apparently at random ruin many tests. Some preparations d o not respond at all, but the general trends are usually there. The same phenomena can be repeated, though with more uncertainty of success, in a preparation in which only the isolated ganglion is used (Hoyle, 1966a). The ACAM may be retained and recorded from to monitor the activity of the excitor motorneuron, and conditioning stimuli applied to the nerve stump. In most preparations the neuron continues to fire at about 8-12/s. This

396

G . HOYLE

frequency can be raised by conditioning to above 40/s. (The sheath can be removed from the ganglion without altering the activity.) Of course it would be of considerable interest t o be able to locate the cellular site of the relevant pacemaker. The activity can be picked up in a highly localized region of the ipsilateral neuropil, anterior to the point of emergence of the axon, and this site appears to be the pacemaker region. Intracellular recordings have been made from near this region (Fig. 6) and most recently from the cell body (see addendum). The biological implications of the results are quite simple, for the reaction has adaptive significance. The excitor neuron fires spontaneously, driven by a pacemaker which is probably within the same cell. The pacemaker is coupled to sensory input from the leg. Should a strong sensory input from an electric shock or pinch occur as a result of a change in the pacemaker frequency, the latter itself changes in a direction such as to oppose the likelihood of recurrence. The phenomenon has value in relation to maintaining appropriate posture, and one may imagine a continual resetting of the postural pacemakers as loading or other demands change. If such mechanisms had been known it would have been possible to predict the outcome of the Horridge experiments. But the biophysical implications are very complex. How does the pacemaker “recognize” that its earlier inaccuracy resulted in the sensory input? There must be some kind of “memory” of the events which led to interruption or acceleration of the pacemaker; that is, a slowlydecaying chemical remainder. This question calls to mind the “memory” element of the famous principle recognized by the German control mechanisms school known as “re-afference” in which the equivalent of a carbon copy of instructions is supposed to be kept by the nervous system for later use in comparing with actual afference. It is some time, however, since we have heard about this principle. Other questions must also be raised: why is a pacemaker change only instigated after a repetition of errors, rarely after a single one? What molecular changes occur to ensure that it is long-lasting? In all the experiments it takes about 8 min at 22°C to achieve a stable change. This is also the time taken by a locust to learn to maintain its leg in the raised position. It is probably just sufficient time for a change in protein composition of the pacemaker membrane to occur. The phenomenon could be regarded as a special kind of “physiological adaptation” and some psychologists apparently prefer to think of it as such. Whether or not it is called “learning” depends

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

397

upon the definition adopted. It fits most. The kind of plasticity involved is akin to that expected of various kinds of learning. It is encouraging to find total acceptance of the phenomenon by two experimental psychologists (Kandel and Spencer, 1968), even to the extent of using it as an example of avoidance-conditioning. Further studies are being carried out in an attempt to record from neuropilar sites where the underlying phenomena occur, and the preliminary results have been referred to above (p. 216), but progress is hampered as always by the absence of neuroanatomical information. I hope that investigators do not continue t o interpret Bullock’s remark that a “maximum of electicism is needed” as a license to scatter their effort thinly, otherwise we may wait a long time for the gaps to be filled in. It should be worth while for an anatomist to work out the locations and connectivity of this experimentally manipulatable neuron. In the present context, the relevant finding is that the rate of firing is proportional to the mean membrane potential near the pacemaker site (Fig. 6C). This is lower after “training” in one direction, and higher after training in the other. The lowering of membrane potential may be due to increased excitatory synaptic bombardment or to other causes. The information currently available does not permit a resolution of the problem. Luco and Aranda ( 1964) have also been investigating physiological phenomena associated with a form of learning in insects. They find that although cockroaches normally clean their antennae (see work of Rowel1 on “grooming” reflex referred to on p. 247) only with the forelegs, if these are amputated, they eventually begin to clean them with the middle pair. The actions are crude at first, but get progessively better co-ordinated. It is hard to avoid thinking of the improvements as progressively learning new co-ordinations, and this may in fact be apt. At the same time as they studied the behavioral changes, they examined a number of physiological parameters. When a connective between the subesophageal and prothoracic ganglia is stimulated electrically an output of impulses follows in the large 5th nerve of the metathoracic ganglion. There is a delay in the normal preparation of 4.42 f 0.37 ms. But after the operation, there is a stage at which this reduces by about half, to only 1.93 ? 0.07 ms. This has always been reached by the 12th day following the operation. The probable explanation is that a central synapse associated with the larger delay present in the normal ganglion has been by-passed as a result of physical or physiological changes. These are initiated by leg extirpation.

398

G. HOYLE

Other evidence for memory processes has come from studies of optomotor responses. If a locust is shown a pattern of moving contrasting stripes it turns its head in the directior! of movement. Horridge ( 1966) was able to show that a movement as little as 0.1" is an adequate stimulus and that if plunged into the dark for a time, the locust then reacted in such a way as to indicate that it retained a visual impression of the previous positions of the stripes for several seconds or even minutes. It thus demonstrates an optokinetic memory. 111. MOTOR MECHANISMS A. GENERAL ASPECTS

A few aspects of insect motor activity have been subjected to preliminary analyses in varying degrees of intensity. These are: respiration, especially with reference to spiracle control; perambulatory locomotion; flight; song-of crickets; and mating behavior. Each has something to contribute to our knowledge of how the insect neural machinery operates, and each will be reviewed briefly. Only such work as includes an actual or theoretical cellular aspect will be included. Much of the early literature on central nervous function did not attempt to reduce the problems to the level of single neurons. Studies in which insect cerebral (supraesophogeal) and other ganglia were progressively severed from the rest of the nervous system led to the postulate of a hierarchical set of inhibitors of general activity. Excitatory control is then achieved in part by disinhibition. Clearly this was an attempt to bring the insects into line with the view of the vertebrate nervous system propounded by Sechenov, but which is a rather crude generalization and only partially true. Roeder, who has supported this view following his pioneering researches on the praying mantis (Roeder, 1937), has elegantly summarized his work (1 963). The mantis is very inactive after the removal of the cerebral ganglia, but hyperactive after removal of the subesophageal. The same are true of the cockroach, but only for a short while after the operation. Locusts are not very active after removing the cerebral ganglia, but occasionally they will walk, jump or even fly spontaneously. After removing also the subesophageal ganglia they remain immobile, with a high extensor tone, but they will, rarely, walk, jump or even make flight movements.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

399

Most authors have been impressed by the capabilities of the thoracic ganglia to co-ordinate and control locomotion, sound production, flight, etc. without intervention from higher centers. Even the prodigious capabilities of the last abdominal ganglion, in regard to sexual movements and egg-laying (Roeder et ul., 1960) are impressive, giving an additional reason for the increasing interest in insect ganglia. Views on “general” controlling functions must adjust to the recent results of Rowell (1964) on the grooming activity of Schistocercu greguriu and Acunthucris ruficornis. He refers, perhaps a little tentatively, to a “grooming reflex”. Rowell must at first have thought of grooming as a possible example of “cen trally-programed” behavior, but his experience with the movement soon changed that. Only very occasionally in the intact insect did Rowell obtain grooming behavior by tactile stimuli applied to a few well-defined sensilla. If the legs were isolated from the substratum the chances of obtaining the response increased, but were still slight. Since the insect cannot be expected to take to the air to groom itself, it must be supposed that the tarsal inhibitory effect is incidental, part of a generally depressing action on neural activity. But if connectives to the other thoracic ganglia were severed, the probability of obtaining a response increased dramatically, becoming 96% with an isolated prothoracic ganglion. Of greatest interest was the fact that Rowell noted that before the operation severing a single connective the animal used its left and right sides equally, whilst afterwards the leg of the same side as the cut was used about five times as often as its partner. The legs were used equally often also in the case of cutting both connectives. Rowell explains these results by postulating mutual inhibition between the two halves of the prothoracic ganglion. He suggests that in the normal state inhibitory signals primarily inhibit their own halves, but after sectioning of one connective this side is disinhibited, sending stronger signals to the other half and therefore adding still further t o its own disinhibition. He suggested an alternative explanation, which I favor because it fits better with my own experiences in studying inhibitory effects within the ganglia. In this other interpretation a single inhibitory neuron is postulated which synapses with a majority of elements in the nervous system -including those associated with the grooming movement. In comparing the alternatives, he rejects the second on the grounds that if the common inhibitory neuron were replicated sufficiently for all comparable behavior, this would be wasteful of

400

G . HOYLE

neural elements. I am not convinced by this argument. For one thing the number of bits of behavior for which such a neuron would be appropriate is rather limited, so the number of inhibitory cells would not be prohibitively large, perhaps no more than 20-30. An inhibitory cell of this type could be shared by more than one behavioral unit. Related behavioral units are presumably linked in just this way, so that disinhibiting one automatically disinhibits the others. This may be one way in which excitation spreads among the various units during elicitation of a complex set of acts. Once it is accepted that such general inhibiting effects are widespread, the search for disinhibition as a device to disclose excitatory pathways-reflexes if you prefer-will become a common practice. Rowell’s findings should serve to emphasize how unrealistic may be results which are obtained from preparations which are less than the whole insect. A number of ablation studies have led to results which suggest the existence of opposed general excitatory and general inhibitory influences. These could serve as indicators of preparations which it would be worth while following up neurophysiologically . For example, Huber ( 1 959) found that removal of the dorsal parts of the mushroom bodies of a cricket and a grasshopper caused a large increase in general locomotory activity which persisted for several days. This sort of observation is a step ahead of the whole ganglion type of experiments, but a satisfactory understanding will not be achieved until the influence of individual units has been determined. The locust has been brought into the whole animal study field not only at the output end, but at the level of the brain also. Rowell ( 1963) chronically implanted stimulating electrodes into the brains of locusts which were free t o move. He obtained antenna1 movements, feeding, and locomotion in response to stimulation, and also noted that ongoing sexual activity could be inhibited. The feeding, and perhaps even a foraging activity, occurred only if the food stimulus was also present at the time of brain stimulation. Rowell thus considered that the effect was due to a disinhibition being applied to sites which are active in suppressing reflexes which might otherwise have been going on. Here in Rowell’s work we see the influence of something an investigator has convinced himself of, affecting his subsequent studies and disclosing further examples of the same “principle”-in this case that of disinhibition at the level of a specific act of behavior.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

40 1

B. RESPIRATION

An excellent and thorough review of present knowledge about the neural mechanisms of respiratory movements in insects has been presented by Miller (1966). He has also recently contributed additional experimental data on the control mechanisms of locust spiracles (Miller, 1967a). In the context of the present review we need consider only the generalizations about the underlying neural events which may have significance for behavior. Two separate pacemakers occur, one for ventilation, the other for the spiracles. Both are essentially endogenous. That is, the rhythm of action can be generated by cellular events which do not require feedback from responding elements to ensure a return action. They are nevertheless sensitive to a variety of internal and external environmental changes. The ventilatory center is influenced by the COz concentration locally, and by input from abdominal proprioceptors. Both “modulate” the basic rhythm which they cannot directly determine. The spiracular control center is influenced by the COz concentration, but very little by proprioception. The spiracular and ventilation centers are synchronized by interneurons which run from the latter, which, in locusts, is located in the metathoracic ganglion, to the interneurons immediately antecedent to the spiracle motorneurons lying in the same segment as the relevant spiracle. Abdominal ganglia can initiate activity independently in their own segments, but an overriding mechanism co-ordinates the activity. This influence is carried by interneurons which extend down the cord. The same neurons inhibit inspiratory motorneurons located in the abdominal segments. Such co-ordinated activity is advantageous where a complex organ such as the abdomen, comprising many segments, must act as a whole to subserve functional needs. Miller ( 1967a and see also 1967b) gives a summary diagram (Fig. 20) which contains a number of interesting points. The long, intersegmental neurons he termed “command” fibers after the usage of Wiersma and Ikeda (1 964), developed from a study of crayfish interneurons. Each motorneuron in Miller’s scheme has a large number of separate specific sites of action along its length within the neuropil. These include an intrinsic pacemaker, a nearby frequency adjuster which influences the pacemaker, various inhibitory sites and a spike-initiation zone. It will be generally agreed that any insect interneuron which is a pacemaker follows this pattern, and that it is similar to that postulated above for insect motorneurons in general

402

G. HOYLE

~

Spikeinitiating zone

??‘.

Spike-initiating zone

,.‘ -‘.-,- I‘

To other spiracle 1 motoneurone

Pacemaker -

Frequency adjuster -b

4

Pattern 2a interncurone

k

Pattern 1 interncurone

/

Inhibitory interneu rone from head

Inhibitory interncurone from G 111

Fig. 20. Diagram of neurons involved in the control of the spiracles of the locust. (From Miller, 1967a.)

so that when driven by a constant depolarizing input they fire repetitively. We should nevertheless ask a question of the model. Do such neurons have an intrinsically unstable membrane potential at their mean “resting” level? A similar general control system, based upon a master pacemaker located in the metathoracic ganglion, but with other local pacemakers occasionally taking over, has been shown t o occur also in Periplaneia by Farley e i al. ( 1967). Internuncials propagate impulses, occurring in bursts, t o the abdominal expiratory centers at 3 m/s. Expiratory bursts of impulses (Mill and Hughes, 1966; Hughes and Mill, 1966), which lead to the expulsion of jets of water occur rhythmically in dragonfly larvae. Single neural units were recorded from and showed bursts which increased in frequency as the burst proceeded. Again the rhythm is intrinsic, but can be modified by the afference. B y electrically stimulating the first segmental nerve they

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

403

were able to interpose an extra burst and this led to a resetting of the rhythm. C. PERAMBULATORY LOCOMOTION

There are several recent contributions to the study of walking in insects, all of which permit some examination of possible underlying mechanisms. Indeed, investigators are increasingly oriented in principle towards the cellular level in neural control problems and direct their experiments accordingly. The method of recording electromyograms from the intact insect with trailing leads seems destined t o play a major role in elucidating the detailed operation of the locomotor apparatus. It also may be helpful in evolutionary and genetic studies. For example, Ewing and Manning (1966) made an interesting comparative study of the underlying control mechanisms of walking in three cockroach species which are representative respectively of high activity and great speed of movement (Periplaneta americana), medium speed (Nauphoeta cinerea) and very slow speed (Gromphadorhina portentosa). The basic patterns of motor output were found t o be virtually identical in all of them. The same kinds of changes in recruitment of motor axons and reduced reciprocity with increasing speed occurred in each species. Similar neural machinery may thus be postulated for each, with speed differences caused by different rates of response of the muscles. It would be worth while studying the muscles by direct physiological methods and by electron microscopy. The results of most new studies on locomotion are enigmatic in that they emphasize the importance of proprioceptive reflexes in determining the details of timing, but at the same time also show that centrallydetermined programs can function independently of proprioception. A preparation in which recordings are made of the neural activity t o an antagonistic pair, levator and depressor muscles of the coxa of a leg of P. americana, has been developed by Pearson ( 1969). The mo torneurons give reciprocal bursts when efforts have been made to exclude proprioceptive feedback. He found that levator burst durations are relatively constant, whilst those in the depressor are variable. This is what earlier behavioral (filmed) studies suggested, but Delcomyn’s ( 1969) more intensive analysis shows that protraction times (levator action) in fact decrease greatly with increased speed of movement. Pearson (in preparation) has found that a negative coupling occurs between the levators in the metathoracic

404

G. HOYLE

segment, and also, in stronger form, between ipsilateral meso- and metathoracic sets. The significance of proprioceptors is, by contrast, clearly brought out in a series of experiments made on stick insects by Wendler (1961, 1964, 1966). These demonstrated the relevance of the hair plates to locomotion. When they were removed the movements were awkward, and the amplitude exaggerated. When present they form part of a negative feedback loop which stabilizes the joints against bending forces, assuring normal movements. At the same time, however, the experiments stressed the non-importance of proprioceptors as agents in the determination of the sequences of movement. When various amputations were performed, there was an alteration of the co-ordination of the walking movements, and also a reduction in the speed. Wendler deduced from these results that neither the changed equilibrium nor the changed velocity accounted directly for changed co-ordination. The explanation probably lies in the lack of phasic input from receptors of the amputated limb, which must therefore influence the timing of the patterns. Following the kind of approach being developed by Wilson, Wendler concluded that there are two ways in which sensory influences might work, neither of which involves a direct reflexive action of the classical kind. Firstly, they might influence the motor pattern generators (= oscillators) of the neighboring legs. Secondly, they might influence the motor pattern generator (= center for control of local perambulatory movement) of the same leg, leaving phase determination entirely to mutual connections between other centers. Elimination of afference from a particular leg could reduce the amplitude of oscillation of that leg to such an extent that it would no longer play a role in phase determination. The new co-ordination would result from a smaller number of active neural pattern generators (NPG’s). When both middle legs are amputated, for example, the insect walks with a CNPG co-ordination: the middle leg stumps move, but with a small amplitude. Simply by adding little extensions to the stumps the amplitude is returned to normal and at the same time the co-ordination pattern returns to that of the normal 6 NPG’s. These extensions touch the ground and lead to proprioceptors in the stumps being stimulated with their normal time relations. Wilson in 1965b and more recently ( 1 966, 1967) has commented on the possible origins of perambulatory locomotion, based both upon earlier and upon his original investigations. He has defined a

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

405

tonic reflex as one which shows phase independence between input and output. The problem with this definition, based upon my own studies on locomotion, is that the degree of phase relative to any given afference, for a particular motor act, is never fixed (in insects) in a simple manner. Thus a movement may show a marked phase dependence for a while and then revert to complete independence and vice versa. In locomotion the driving of a movement may be for a time dependent mainly on fluctuations in frequency of the motor output to the extensors, but later change to dependence on flexor fluctuation. Nevertheless, the afference can produce adaptive adjustment to the locomotory control when averaged over several cycles of movement. The minimum reflex time in cockroaches is at least 10 ms, compounded of: activation of sense organ, centripetal conduction, ganglion transmission, centrifugal conduction, neuromuscular delay and contraction time. The latter depends upon the existing state in the muscle. Tension fluctuations occurring against a background of tonic tension, in which the muscle fibres are continually activated, occur much more rapidly than contraction starting from rest, and this is the way they are occurring in the intact animal. Estimates of delays in reflex pathways should take this into account. The normal stepping speed for a running cockroach has an upper limit of about 20 steps/s. The reflex mechanisms can be driven at speeds up to twice this value, but only for a short time. It seems that the central nervous mechanisms quickly fail and the output becomes more or less continuous, with sporadic fluctuations. Very recently Delcomyn ( 1969) has analyzed the co-ordination of cockroach legs during walking in great detail. He filmed the walking animal with a high-speed movie camera, and made a frame-by-frame analysis of the movements of each leg and the phase relations between legs in the intact insect and after leg amputation. These results were supplemented by electrical recording from relevant muscles during walking and by analysis of reflex actions to forced movements of leg joints. Several of the results are relevant to our understanding of the detailed means by which the motor output is determined. Firstly, phase relations between legs were found not to change as the speed of walking changed, at speeds from 5 cm/s to 80 cm/s (they did change at slow speeds from 5 cm/s down to below 1 cm/s), but the reflexes in a given leg shift with the frequency changes. Following any amputation the phase relations between legs changed. Separate A.1.P.- 14

406

G. HOYLE

mechanisms are therefore required for the co-ordination of legs and the generation of a step by a given leg. The results stress the importance of the reflex arcs in co-ordination in cockroach locomotion. Pringle ( 1940) many years ago investigated cockroach reflexes and failed to find a neural coupling between segments, concluding that it was instead a mechanical one dependent on stresses physically transferred and actuated by local proprioception. Experiments can readily be performed which rule out the latter and which nevertheless show a coupling-sometimes. We are faced with an apparent paradox. The explanation is, I think, quite clear and again demonstrates the lack of dependence on reflexes in the simple sense. When the insect is neurally ready to locomote, pathways between segments are opened up, permitting cross influences that are not normally detectable. But again, these influences are subtle ones. There is no simple, forceful automatic turning on of a relay. Wilson has been forced t o refer to the influences as a form of “loose coupling” which can have different degrees of intensity. A “tonic reflex” represents one extreme in a continuum of possible relationships in one pathway, where the other extreme of complete phase dependence can also occur. This generalization was very similar to that made by Weis-Fogh (1964) in respect mainly to the flight control system, though meant to be generally applicable. Co-ordinated movements were said t o comprise an intricate command system connected with many sense organs by means of changing, facultative coupling! For absolute co-ordination of legs in different segments during walking the coupling must be strong enough that there is no frequency difference between segments. Each segment has its own motor mechanism which is self-operating, but the 3rd, or fastest, drives the 2nd and the 2nd drives the lst, each with a phase lag. This determines the metachronal sequence from hind to front. The term “drive” is meant to imply the condition of extreme intensity of coupling; the other extreme, very weak coupling, is to imply complete independence or “ n ~ n d r i v e ’ ~Two . elements are present in the coupling, a fixed lag due to conduction times, and a variable lag dependent on the intensity of the coupling at the time. Wilson (1 967) has given a general diagram of the cross influences between segments (Fig. 2 1). As the frequency of stepping increases the lags are shorter, although the phase remains constant. Thus the apparent pattern change with increasing speed is illusory. The pattern is in fact

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

407

_ J V '\ /

:5I

f I

I'

Fig. 21. Diagram illustrating the pathways coupling neural pattern generators (here giving bursts and indicated as -) between adjacent segments. (From Wilson, 1967.)

408

G. HOYLE

constant. In saying that the 3rd segment drives the 2nd drives the lst, Wilson is not taking into account the independence of the two sides. There are, as referred to in considering Wendler’s work, six motor mechanisms, one for each leg, in each half ganglion. There is a remarkable degree of independence of those located in each half of a ganglion. In normal locomotion there is a reciprocal or negative coupling between the two sides which leads to phase locking at 0.5, or antiphase, but this can be switched to positive coupling for such acts as the locust hop or jump, or the swimming strokes of aquatic insects as Wilson points out. The capability of this switching, which is complete when it occurs, demonstrates an intrinsic pitfall in looking for fixed reflex pathways in insects. Their central nervous systems are much too subtle for that, though their “flexibility” can also be their undoing. I have seen locusts fry to death for want of the central part of their jump “reflex” mechanism to permit impulses to pass through, or flap their wings to exhaustion after falling on their backs-a reflex caused by loss of tarsal contact. D. FLIGHT

Several reviews have covered the remarkably elegant work of Wilson and Weis-Fogh ( 1962) and Wilson and his school (e.g. Wilson, 1961, 1964; Wilson and Wyman, 1965) on the neural mechanisms underlying the flight motor output in locusts (see review by Wilson, 1968). These studies have resulted in the best analysis we have about any single act of behavior of an insect. Wilson’s very significant contribution has been to show that a simple pattern of neural output to the flight muscles, quite adequate for functional flight and closely resembling that which can be shown by recording from the muscles actually to occur, at least in tethered flight, is produced by the appropriately active nervous system in the complete absence of sensory input from the wing proprioceptors. The control patterns occur in each of a small, still unknown, number of motorneurons for each muscle. They are substantially similar, though independently produced, in each. According to Neville (1963) five “fast” motor axons supply each of five motor units in the flight muscle of locusts. The output consists of spaced single or close pairs, rarely triplets or quadruplets of impulses. In the gaps between these, similar singles or pairs occur in the antagonistic neurons. The timing of the second impulse in a pair is considered to be due to a long, 30-50 ms, refractoriness of the relevant axons, but it is of

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

409

significance in determining the power of the ensuing contraction. A single action potential causes the muscle to develop only a small proportion of the maximum force of which it is capable, whilst a pair or triplet, closely-spaced as they are, gives much more force. The number of spikes increases from 1 to 4 in proportion as the interval between successive outputs decreases. The nature of the electrogenic events causing the 1-4 output is not known, but from general principles of excitation of cells it may be deduced that as the total excitatory level increases, a depolarizing event occurs at progressively briefer intervals. At the Same time, by a process of facilitation the individual depolarizations become more steeply-rising. In the absence of any secondary response they would simply lead to higher plateaux. Because of their depolarizing action and the electric excitability of the motorneuron, impulses are fired, increasing in number as the facilitation proceeds. The initial depolarization could be either a pacemaker within the same neuron, or synaptic bombardment. The latter would suggest that the firing cell is driven from an interneuron operating at the interval frequency. Neither interpretation is considered by Wilson, who has especially considered schemes based on the reciprocating oscillator model developed by Librascope personnel, notably Reiss ( 1962), in which continuous trains of impulses at high frequency (the primary oscillator) drive secondary oscillators. The first secondary oscillator fires exactly 90" out of phase with the second or antagonist secondary oscillator. This means that these two oscillators may be reciprocally coupled, so that when one fires it inhibits the other. In locusts, on this model, switching on of the primary oscillator is achieved by a combination of loss of contact with the substratum by the tarsi, which removes an inhibitory influence, and air pressure on special sense organs on the head (Weis-Fogh, 1956), which is excitatory. Since no evidence exists which can directly support any of these models, the reader may take his pick. Ann Kammer has recently examined the warm-up phase of moth wing movements (Kammer, 1967) and concluded that they were substantially identical in underlying neural pattern to those of overt flight. In detail, there may be considerable differences, but a close inspection of all the records suggests a continuum, so that modification of the single basic flight mechanism provides an adequate explanation. Of particular interest is that antagonists may contract against each other, proving that even the simplest type of reciprocating system is not determined in its action by simplistic A.I.P.-14*

410

G . HOYLE

connections with only a limited functional mode. In a study of the motor pattern to a flying fly Wilson (1965b) found that most of the phase locking which is so characteristic of locust flight control “has been lost”. This led him to regard the fly neural mechanism as a “degenerate locust pattern”. In both insects frequency modulation in pulse trains to flight-controlling muscles accomplishes changes in both power output and spatial orientation. In the locust, frequency modulation is super-imposed on a pattern of oscillatory interaction between reciprocating antagonists, to use the Wilson terminology. He considers the control system of the fly to approach that of an ideal pulse frequency modulating code, i.e. in which the timing of individual pulses has no importance. In the locust, patterned output to the muscles controlling wing movement can be obtained not only by natural stimuli but also by stimuli applied to the nerve cord. Even when these are random, a flight pattern output occurs (Wilson and Wyman, 1965). The power output is increased not only by increased numbers of impulses per burst, as outlined above, but also by recruitment of additional motor units (Wilson and Weis-Fogh, 1962). Some of the wing muscles, the anterior and posterior tergocoxals and the 2nd basalar and the subalar, are also leg promotors or remotors. Wilson (1 962) was able to show that during flight they operate in nearly perfectly repeating fixed patterns, co-ordinated with other flight muscles. However, when they function in leg movements they become completely independent of each other, and give a wide variety of detailed activity. Thus we have here a situation in which switching of motor control occurs from one mode which causes a rigidly fixed motor pattern, probably independent of sensory input, to another which involves a lot of reflex action. Waldron (1967a, b) has also examined details of the locust flight control pattern, with special reference to the influence of sensory input, which Wilson has shown to be of only secondary significance. She found that elevator activity is depressed by wing sensory input, whilst the same input tends to cause rhythmic bursts to continue in depressors. The individual mo torneurons undergo slow changes in average excitation, as Wilson had found. In examining interaction between ganglia, no evidence was found for greater co-ordination within ganglia than between them. Other forms of sensory input which were found to influence the flight pattern were flashing lights. If flashes are rhythmical, at a period not too different from the wingbeat frequency, the latter come to be synchronized with the flashes.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

41 1

Although perhaps not directly within the expected scope of the field of neuroethology, studies on detailed aspects of the motor control of flight muscles during flight is of dual interest to us. Firstly, it provides details in an experimentally accessible preparation of a neural program, as briefly outlined above. Secondly, it gives an indication as to how complex can be the interrelations of individual muscles in causing movements. The 2nd basalar and the 2nd subalar muscles both contract synergically during flight of the dragonfly, aiding wing depression. But they each give a different angle of attack, so that the actual angle of attack, and therefore important aspects of the lift pattern, is determined by the balance of excitation to the two muscles (Neville, 1960). The 3rd subalar muscle, and also the coxoalar, are brought in for attitude and directional control. Many acts of behavior are likely to involve many muscles, some of them multi-functional ones. A direct experimental study of these postulates has been achieved by extracellular recording combined with flash photography by Dugard (1 967). She found that the 1st and 2nd basalar muscles cause turning by acting synergistically to increase the pronation of the wing. Extracellular microelectrode recordings from single units in the 1st and 2nd segmental nerves of Locustu migrutoriu, and possibly associated with the flight pattern, were varied. Some had a rhythmic spontaneous pattern and these were inhibited by almost any afferent input. Others were silent until continued afferent input occurred, when they gave a patterned output. Unfortunately, as with nearly all of this kind of work, it is not yet possible to assess the significance of the results in relation to behavior. Some very interesting anatomical studies on neuron localizations of cells associated with the flight control apparatus of Schistocercu greguriu were presented by Guthrie (1 964). Motorneurons to the longitudinal indirect muscles were located and found to be large. Some of them have now been located in the living ganglion and their threedimensional geometry revealed by injection of Procion yellow by Bentley (1 969c). A valiant attempt to record intracellularly from specific neurons associated with the generation of impulses to muscles controlling flight has recently been made by Joan Kendig (1 968). Her aim was not so much to understand the physiology of these neurons as to see if she could throw some light on the mechanism of generation of the impulse patterns. She did succeed in obtaining preparations in which functional activity remained after removing the sheath from the

412

G . HOYLE

ventral surface of the ventrallyexposed mesothoracic ganglion. Electrical activity was recorded from a specific muscle at the same time as a glass capillary microelectrode was slowly probed into the ganglion. Successful penetrations of the somas or even large dendrites surprisingly was rare, and stable resting potentials were not obtained. She obtained long trains of injury discharges, all of which eventually ceased, presumably a result of damage caused by the electrode. But impulses as great as 50-60 mV were sometimes recorded, indicating a very close proximity of the electrode tip to a conducting axon. The extracellular records permitted a study of the influence of impulses in contralateral neurons or specific neurons. Potentials as large as 2 mV occurred with almost no latency, and although uncertainty regarding the exact recording conditions prohibits assessment with any degree of confidence, a low resistance pathway, permitting weak electrotonic coupling between motorneurons, is a possible explanation. Synergically acting cells could be synchronized by this simple means. Kendig also recognized a larger, later (5-6 ms latency) influence which was probably synaptically mediated. These findings are not compatible with one-layered, simple models of neural interactions as the basis of the flight control system, but instead lead to the postulation of a multilayered, hierarchical arrangement with patterns generated independently of the relevant mo torneurons.

E. INSECT SONG: CRICKETS

Male crickets spontaneously emit a characteristic calling song when mature. If another male cricket comes within visual range, the song changes to one of aggression, but if the intruder is a female it is more likely t o change to the soft, continuous courtship song. Clearly it would be of interest to know how these different songs are determined by underlying neural events, and how they are “selected” by the different sensory clues. The output causing the sounds can serve as a starting point for investigations and this we attempted t o do by implanting leads in various wing muscles in the manner mentioned for locust locomotory muscles. It turned out that the presence of the leads did not prevent the insects from singing normally (Ewing and Hoyle, 1965). Bentley and Kutsch (1966) were also successful with similar experiments. To determine the nature of the sound produced in relation to wing

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

413

movements is itself a major undertaking requiring high-speed films of the movements. Davis (1 968), in our laboratory, concluded from a preliminary study of this kind that opening during stridulation is homologous with downstroke of flight, so that sound is produced during the closing stroke. A functional anatomy of the muscles responsible for the wing movements has unfortunately not yet been worked out satisfactorily. Dorsal longitudinal muscles, for example, are expected t o give only wing depression when the wings are in the flight position. What do they do in the song position? Bentley and Kutsch found that one bundle gives action potentials on the opening stroke and the other on the closing, and they claimed that they must therefore be antagonistic. This in itself is a remarkable and unexpected finding, subject to substantiation by direct experiments. The larger action potentials obtained by Ewing from the dorsal longitudinals during the courtship song were always followed by a loud sound within a few milliseconds. The wing depression stroke does not close the wings, it opens them, and it seemed possible that this sound was made by the wings being pressed together during opening rather than during closing, which is the stroke which is considered to give the sound (Pierce, 1948). Kutsch (personal communication) has tested the statement of Pierce concerning the closing stroke as sound producer by dissecting out a small central piece of the file and sticking it back in place in the reversed position. This leads to a softer sound in the middle of a pulse, and this would be expected to occur only if the sound is produced by closing, in view of the slight angle to the vertical made by the teeth which results in a less stiff contact with the scraper after reversal, but only in the closing direction. An intensive study of song in a grasshopper has been presented by Elsner (1968). He states that he was able to insert the remarkable number of 15 flexible wire electrodes made from 30 pm/steel wire into muscles concerned with singing and still have the insect emit a normal-sounding song. He has obtained detailed information concerning the relative timing of activity in different muscles from the electrical records. He found that some motor units are continuously active during song, whilst others are only active in association with louder sounds, in all the muscle groups involved. The former units lead the latter by from 0.2-2 ms. Elsner also studied co-ordination of muscles on opposite sides, and found that in the first sequence they are always synchronous but that in the second and third they either fire alternately or else a fixed delay is shown

414

G . HOYLE

between one side and the other. At all events there is a marked correlation between them. A great deal of progress had already been made by Hiiber (1959, 1962, 1965a, b) in his classical studies which were the first to show that behavior could be obtained by local electrical stimulation of the brain of an insect. By combining such studies with local lesion and coagulation experiments in crickets, he concluded that parts of the mushroom bodies and of the central body neuropil of the brain determine the sound patterns. Sounds can be produced by local stimulation of thoracic ganglia of insects from which the brain has been removed, but they are crude and unlike any natural sounds. Some more promising sounds were, however, obtained whilst passing direct current through the thoracic ganglia. In principle, as pointed out in the paper by Ewing and myself, a very simple form of neural machinery not unlike that postulated by the Librascope group for determining reciprocal activity in antagonists, could cause the appropriate sounds. The significant point is that quite small numbers of nerve cells might be adequate on the motor side. Many cells may be needed for the process of selecting which song to sing. Huber may have been stimulating neurons in the selection circuitry, or perhaps the “command” interneurons on to which sensory input converges, when he obtained singing during brain stimulation. His electrodes need not have been anywhere near the actual motor machinery of the song determination. They did stimulate complex additional machinery, however, for there was a decrease in the intensity of stimulus required to just evoke a motor response threshold, and other aftereffects lasting 30 min or more. Another interest of the cricket preparation is that the muscles which are responsible for sound production may also be used in flight. Studies on such a muscle provide a way into a motor pattern generating system which must involve some form of central switching process, from “flight” channel to “song” channel; this permits, in principle, the study of a motor tape selection mechanism. An example of such a study is that by Elsner of a cricket subalar muscle. In flight, the left and right subalars contract synchronously, but in the courtship song of the grasshopper Gomphocerippus rufus the legs alternate. When courtship song contractions start in the subalars, they are at first synchronous, as in flight, but gradually shift to the alternating pat tern. Huber (1962) proposed that there are three separate song centers and channels. I have offered (in Ewing and Hoyle, 1965) a simpler

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

415

alternative, in the form of a single continuously active major “oscillator”, normally inhibited, which is disinhibited in slightly different but characteristic ways, depending on the specific sensory input. The various wing muscles would be driven by the oscillator

M

1

1

M

-

Fig. 22. Recordings from a cricket ganglion during singing. Motor nerve activity (upper traces-M) was extracellularly recorded from the 2nd basalar and 1st promoter nerves during the period (marked by a bar) between the onset of the opener motorneuron burst and the onset of the closer motorneuron burst during a chirp, correlated with activity in mesothoracic ganglionic neuropilar segments of unspecified neurons (lower traces-(;). Note the small bursts of EPSP’s (upper neuropilar trace) and IPSP’s (lower neuropilar trace) which also shows some spikes, during the song burst. Calibration: horizontal, 100 ms; vertical, 10 mV. (From Bentley, 1969b.)

416

G. HOYLE

automatically. The purpose of such models is to focus our attention on neural mechanisms and to encourage a neurophysiological search, and perhaps to aid in the recognition of such units as may be found during the search. A preliminary search for the underlying neural machinery has now been made by Bentley (l969b). He succeeded in getting a single cricket to sing continually for 10 hr following lesions made with the tip of a sharpened heated steel needle in dorsal areas of the mushroom bodies of the supraesophageal ganglia. The song was mainly of the calling kind, but with some transitions t o both aggression and courtship calls. Whilst it was singing, he carefully exposed the mesothoracic ganglion and made exploratory excursions with a glass microelectrode. His most significant finding was that gaps sometimes occur in the calling song, and the duration of the gaps is an integral multiple of interchirp intervals and chirp durations. This shows that there is an independent chirp rhythm generator which can be active even when it is not actually triggering chirps. This rhythm is not reset when an out-of-phase chirp occurs in a two-chirp gap. Bentley found EPSP’s underlying each opener neuron burst and evidence for electrical coupling between various neurons active in association with opening. IPSP’s followed each burst of EPSP’s (Fig. 22). By contrast, IPSP’s occurred in motorneurons associated with closing after the onset of opener firing, to be replaced by EPSP’s at the termination of the burst (Fig. 23). There were no intracellular events during the interchirp interval, but some interneurons fired action potentials during the opener and closer inhibitions. In the dying phase both chirp rhythms and the timing of pulses within the chirp were greatly disorganized, but unpatterned firing of neither closers nor openers occurred. It may thus be deduced that the antagonistic groups are obligatorily linked together, either by direct synapses or by a single interneuron layer. This was also a part of the model I proposed. Bentley’s proposed model is one in which a slowly firing excitatory interneuron drives the openers, which are supposed to have an intrinsic pacemaking activity. The latter inhibit the closers reciprocally, and the opener burst is followed by a counter-burst of opener inhibition and closer excitation. 1. Cicadas There are close parallels between some aspects of cricket song and the song of cicadas and of grasshoppers. The former has already been

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

417

Fig. 23. Further recordings from a cricket ganglion during singing. Motor nerve activity (upper traces) extracellularly recorded from 2nd basalar and 1st promoter nerves, correlated with activity recorded with an intracellular electrode placed in the neuropil in or close to neuropilar segments of neurons provoking closing of wings. Filled arrows indicate onset of PSP’s; open arrows indicate impulse in motor nerve which follows large depolarization in the neuron. Calibration: horizontal, 25 ms (A); 50 ms (B). Vertical, 10 mV. (From Bentley, 1969b.)

shown to be amenable to electrophysiological study. The cicada should be the simplest to examine since it involves simple, paired muscles not used for any other purpose, unlike the leg and wing muscles used for cricket and grasshopper song. The songs have been well analyzed (e.g. Pringle, 1954), and are surprisingly complex. A single burst of impulses in the auditory nerve leads to a high frequency (about 200/s) burst in a central ganglion cell, according to Hagiwara and Watanabe (1956), but the tymbal muscles fire in alternating bursts at about 1OO/s each. Presumably, then, the patterning is determined by central interneuron properties and by the timing phenomena associated with a reciprocally-coupled output. F. COURTSHIP BEHAVIOR

An excellent account has recently been published by Loher and Huber (1966) of a bit of courtship behavior of the grasshopper Gomphocerus which must delight all potential neuroethologists. It is exactly the kind of behavioral study on an insect which, though small, may be “practical” for electrophysiology , which investigators will need. This is doubtless the objective which the authors had in mind-and likewise may explain the clarity of their account. It stresses the advantage of collaboration of ethologist and physiologist when the study of neuroethology is intended; only the anatomist is A.I.P.-14*-*

418

G . HOYLE

needed in addition. The behavior observed could be broken down into linked units, each of which is either the contraction of a single muscle, or a small group of muscles giving a co-ordinated contraction. In all, Loher and Huber were able to recognize about 5 0 units and separated these into three main divisions during each of which units are put together in a predictable way. The male orients visually in front of the female before beginning his calling song. The first division of behavior starts with the male taking up a position diagonally in front of the female. He then begins to shake his head for 1-2 s, slowly at first. The palps swing at the same rhythm but with a small and constant delay. Both antennae are held forwards and upwards. The hind femora now start to glide up and down at a small amplitude. This movemement of the femora, which carry the stridulating organs, produces a soft sound. Of interest to the neuroethologist is that the leg movements are synchronized with the head-shaking movements. The first division is terminated by an increase in frequency of both head and leg movements, simultaneously, to 4-6/s. Also at the same time, the palps begin to vibrate. A second period now starts, with the antennae quickly flung back and downwards, accompanied by a sudden raising of the anterior part of the body. The antennae return to their original positions and are swung backwards again. Meanwhile, the hind legs are lifted and one or two sharp sounds are produced by quick backward kicks. The third and final period starts with a pause of about half a second and leads into prolonged .calling, the legs rubbing together as the erected antennae swing slowly at a low amplitude, alternating with each other. The whole behavior takes 8-9 s at 28°C. It is delightfully subtle and I venture t o suggest of much greater interest than cricket song because the primary behavior, singing, is accompanied by other movements and both are directly related to the member of the opposite sex. Extremists will doubtless want to term this complex of activities the Gornphocerus courtship reflex. Indeed it looks like a series of chain reflexes in some respects but it is difficult in principle to see how each preceding movement automatically results in each later one solely because specific sense organs are stimulated. This traditional explanation has already been almost completely ruled out by experiments (see below) and it will hardly be surprising if further work does not substantiate it. It looks like a superb example of a “motor tape” being switched on and played into the local command

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

419

apparatus of the nervous system. The illusion is reinforced by the fact that the moment the sequence is complete the whole cycle repeats itself exactly, i.e. this “tape” is a loop which does not require to be wound back! Various ablation, fixing and gross nervous system damage experiments have already been carried out by Loher and Huber, and their effects on behavior determined. There is a complete absence of contralateral influences. Following unilateral cuts between the brain and the 3rd thoracic ganglion there is at first a slight delay in the onset of movements by the ipsilateral leg. But after 24 hr this has disappeared. Thus “commands” for one side must pass to the other, in the metathoracic ganglion. Of far-reaching significance was the fact that fixing, or even ablating an organ (one antenna, one hind femur, or one palp) did not affect the sequence. Subsequent performance by another part occurred with perfectly normal timing and amplitude. Now these operations presumably prevented, or at least greatly distorted, the normal sensory feedback associated with proprioceptors from the relevant organ. The first step in the process of disposing of the chain-reflex explanation of this behavioral sequence has thus been taken, with a positive result. In another experiment, the head was fixed to the whole thorax by a balsa bridge so that it could not shake. This part of the courtship sequence was then replaced by a rhythmical flexing and extending of the forelegs: shades of displacement activity? Behavior of insects can usefully be studied in the context of neuroethology even without the possibility or intent of doing electrophysiological experiments, provided the observations on behavior are made in a precise, quantitative manner. This is amply brought out by a remarkable series of experiments made mainly by Manning and collaborators on Drosophila. In this work the behavior was broken down into unitary components and the intensity of each was measured, as numbers of a movement per unit time, total number in the sequence or amplitude of movement. The work has additional interest in that it has further opened up the possibility of a study of the genetics of behavior mechanisms. The various elements of courtship in Drosophila: orienting, “tapping”, scissoring, vibrating, licking, attempted copulation, can all appear in the normal behavior sequence in the presence of an etherized and therefore completely unreactive female, provided the male is “sufficiently excited” (Manning, 1967). This observation shows that the sub-behavioral components which together constitute

420

G . HOYLE

courtship are not only each programed in “motor tape” fashion, they are linked together in much the same manner as are the movements of a symphony on a - r e a l tape. In actual courtship, however, scissoring may be omitted (Manning, 1959). The playing back of a part of the “tape” further enhances the level of excitation for the next part, and if extremely high a stage may be by-passed. The reaching of a more advanced stage must initiate measures which inhibit the preceding sections. IV. MODELS OF NEURAL ACTIVITY AND TERMINOLOGY

It seems desirable to now consider the various analogical terms which are being used by the various contributors to the field. Some sort of general language terms are essential whilst we are groping for actual mechanisms, in order to focus attention on these. At the moment the situation is anarchic, which can only mean that a satisfactory set of terms has not yet been devised. My own use of the word “tape” was meant to convey a picture which is easily grasped, though in terms of recent computer construction the relevant information stores are more likely to be in the “core program” than tape or disk. For the purposes of functional description the nervous system can be considered to have three major functional divisions. In one of these the enormous mass of incoming sensory input is reduced to a small series of sets of information which can serve to drive motor output. The second “selects” from the sets the one which shall at a given time be operative, and inhibits others. The third division contains neural pattern generators whose purpose is to produce sequences of motor nerve impulses. The sensory input serves a dual role: it stimulates the nervous system as a whole and thereby increases the probability that some motor activity will occur, spontaneously or reflexly, and specific channels or modalities Serve to cause defined reflex action. The neural pattern generators operate in unknown ways, and may exist in different forms. Some utilize, or themselves constitute, inherited motor neural program stores (“tapes”). There may also be sensory program stores which can be coupled to them. Reflexes are undoubtedly of great significance in determining behavior but in addition, as described above, a role is played by neural activity whose programming is achieved by central nervous cells to a large extent independently of the detailed sensory input.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

42 1

The totality of nerve cells which achieves a particular motor pattern independently of the detailed ongoing sensory input are termed collectively neural pattern generators. A simple form of neural pattern generator is one which produces a regular alternation of contractions of antagonistic muscles of a particular joint independently of proprioception. The fact that proprioceptors can further emphasize the alternation in the normal situation in the intact animal should not detract from the importance of the fact that the alternation can be generated independently of it. A problem in understanding the operation of a particular neural pattern generator is to determine if the output is programmed to control force, displacement, or their time derivatives. Usually it seems to be the former alternative. The locust programs I have studied, which were classified as “motor tape” programs, give similar patterns in a variety of loading conditions, so they are programming a force development pattern in the expectation that it will result in the desired displacement at appropriate rates. Wendler’s ( 1966) experiments on Carausius showed that in that instance of locomotion control, the basic program is a generator of a force program which is somewhat more powerful than that actually required. Here, co-operation with the local reflex action is also in effect built into the plan. The reflex trims the motor output down by central inhibition to a more appropriate size. In the recent experiments by Runion and Usherwood it was shown that in the locust step a similar inhibition may be produced peripherally. The step is produced centrally, but it is modified by reflex action. Nervous systems cause these kinds of movements to be produced by output which unfolds sequentially with a real time base, as must be the case with a taped control of inanimate machinery. In practise the detailed output to the muscle is rendered complex in most instances by a compounding of reflex influences with the central program, and there must then be a guidance system leading to the execution of correct movements. This can only be the afference associated with that movement. In other words, the nervous system must have a record, for certain sequences of movements at least, of the sensory input associated with execution of that movement, to which it can refer in order to determine motor output aimed at achieving a repetition of the movement. It is quite clear that the human nervous system can do this. Such a store of information also has a very important time element, and I termed it a sensory tape (Fig. 24). A “sensory” tape program is in principle much more efficient than

422

G . HOYLE

MOTOR TAPES

SENSORY TAPES

I *oCOMPARE

a w

a w 1

a

0

L

-

COMPUTE FROM DIFFERENCE

a

(3

A

B

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

423

a “motor” program, since it permits continuous adjustment of power output to meet varying loads. A force program to the contractile portions of a stretch receptor operating a servo-loop would be similarly effective. It is perhaps surprising that stretch receptors are not more common in insect limbs. It has been tacitly assumed that the higher vertebrates program for displacement, with an extensive use of proprioceptive feedback for control. However, in a recent dramatic test in monkeys, this was not found to be the case. Monkeys were trained to receive a reward of fruit juice for a displacement activity, flexing the wrist. Detailed microelectrode studies were then made of the neural programs, including those of pyramidal tract cells, and the loading conditions suddenly changed. It was made clear that the learned neural program which achieved the reward was not displacement of the hand but a neural program for force development in the wrist muscles (Evarts, 1968). The insect is supposed to have a small repertoire of both motor tapes and sensory tapes, the former and perhaps even also the latter being largely inherited, at least in a rather crude form. Learning may be supposed to improve their performance during the life of the animal and to result in the addition of new, non-innate tapes. Wilson (1968) has proposed the use of the term “template” instead of tape. I d o not find this an improvement. The template analogy possibly derives from the information coding device used by DNA. Essential features of the neural “tape” are the way they repeatedly play out their instructions from beginning t o end, the extremely important time relations and the way several different patterns operate a command output (read-out) machine, all of which can more readily be visualized by the tape analogy than by a template. In considering alternatives which are less committed t o hardware models, the term “neural program store” is the most general I can currently concoct. _ _ ~ _ _ ~ _ Fig. 24. Models to illustrate the two possible alternative major forms of central nervous control of behavioral acts determined by complex sequences of movements. A. Control by “motor tape”. This involves centrallydetermined patterns with appropriate time sequences generated by neural aggregates, which occur without reference to peripheral feedback, i.e. “open loop”. Various impulse sequences are available (different tapes-causing different behaviors), and then played into a common motor apparatus. B. Control by “sensory tape”. Here centrally-determined sequences of impulses serve as a source of reference with which specific ongoing sensory input may be compared so as to permit the nervous system to compute (from the difference) motor commands on a continuous basis. This is a closed-loop control system. It operates through the same effector machinery as the “motor tapes”. (From Hoyle, 1964.)

424

G . HOYLE

Wilson has also suggested that my earlier models, which involve a hierarchical layering of impulse initiators, may be unnecessarily complex. It is true that the individual motorneuron is itself a complex, effectively multilayered unit, with separate pacemaker and integrator sections. However, this does not, by itself, remove the need for consideration of hierarchical “driving” elements. There are three reasons for holding the latter view: firstly, the extremely high degree of convergence on a small number of final motorneurons whose detailed impulse patterns control behavior; secondly, the long delays occurring between specific input and output in motorneurons argues against direct connections with motorneurons; thirdly, a variety of activities is obtained from interneurons during reflex action, and this cannot be related in a simple, direct manner to the firing pattern of the motorneuron. It has become customary in recent years to regard the various groups of reciprocally-acting antagonistic muscles as being driven by loosely-coupled but basically independent “oscillators”. The oscillation might be due to their being driven from a pair of neurons which are themselves crossconnected reciprocally. A single neuron would drive the pair by simultaneous excitation. The modern version of this basically 50-year-old scheme has originated in a rather unexpected way, from a group of scientists with a physics and electronics orientation rather than a physiological or biological one. They were employed by a firm named Librascope, which manufactures computers. Believing that neurobiology, which deals with the computers evolved naturally by the evolutionary process, might be able t o contribute new ideas to computer technology, the company set up a group (Hamilton, Reiss, Lewis and others) to keep abreast of neurobiological progress, at the same time giving them freedom to make models of neural processes as they thought fit. The group quickly became aware of the reciprocal, oscillatory nature of many of the events determined by neural machinery-and of significance in many behavioral acts-and sought to explain them in simple terms. They showed by both theoretical and model (neuromime) treatment that a reciprocallyconnected pair of neurons driven from a common oscillator of variable frequency can produce many of the patterns which are familiar to neurophysiologists studying nerve cell impulses, and to students of muscular activity in behavior-the reverse of the original intention (see Wilson, 1964, Fig. 1O-locust flight; Huber, 1967, Fig. 5 -cicada singing; Hoyle, 1964, Fig. 13-stepping elements of antagonistic pairs of muscle in

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

425

locust locomotion; Ewing and Hoyle, 1965, Fig. 10-movements controlling wing opening and closing in cricket singing). Oscillation need not be a characteristic of only special interneurons. Motorneurons also are capable of firing repetitively when depolarized continually, thereby generating their own oscillation. By cross-connecting two such neurons by excitatory links one may drive the other, or block it via inhibitory links. V. DISCUSSION

The overwhelming impression which readers must have obtained from the foregoing account is that here is a subject which is as ripe with promise for exploitation as almost any in biology today yet which is not quite off the mark. It would be nice to be able to predict a period of rapid progress, but none can be guaranteed. The path ahead is strewn with difficulties of many kinds. Insect neurophysiology has only recently acquired its first professionals. Only a single research generation ago it was difficult enough for insect physiologists even to pry loose an oscilloscope-some had to make their own, a situation which is happily no longer with us, but the intellectual climate which will reward a man adequately for success in the field has still to materialize. Most investigators referred to in this account owe their allegiance to what they term “general principles of neural integration” and nobody would be surprised to find them working on a fat sea slug instead of an insect. There are both hopeful and discouraging aspects in the recent results. Hopeful is the progress made in recording from the intact animal and in its continuing to tolerate the electrodes and giving normal behavior. Discouraging are the difficulty in recording electrical activity from specific neurons and the variability of detailed neural patterns. Some of the latter may be due to the difficulty in getting a clean record lacking pick-up from adjacent dendrites by cross-talk. But it may be genuine, and whilst it foretells a very interesting data-handling system in the nervous system it makes the task of analysis by the method of looking into the nervous system from the output more difficult than it might otherwise have been. All recent investigators seem to be following Bullock’s ( 196 1) lead in trying to draw attention to the intrinsic capabilities of the groups of neurons as generators of “programmed output” not related to input. It has become almost a virtue to follow the new “party line”

426

G . HOYLE

in condemning the narrowness of the older apostles of the chain reflex, from Pavlov and Sherrington to Skinner. Actually, the timing of this change of emphasis should be traced t o the about-face of von Holst ( 1948), but Bullock’s influence on the new generation of comparative neurophysiologists has been paramount. In my own case, I was brain-washed most by G. P. Wells, who was greatly influenced by his findings on the complexities shown by isolated extroverts of the lugworm, Arenicolu, and the way these were nicely correlated with actual behavior of the whole worm (Wells, 1950). This change can be traced in part also t o the success of the views of ethologists Tinbergen and Lorenz concerning instinctive behavior, which the animal does not need to learn. Instead the behavioral act unfolds in response t o a total situation, rather than a specific stimulus, and therefore owes much to Gestalt psychology. The act can appear spontaneously or, rarely, in an unusual situation (displacement activity). The change in attitude did not arise, as might have been expected, from the researchers or neurophysiologists, although many of the latter did seek to test the reflexological dogmas by critical experiments. The problem was, and still is, that it is not possible satisfactorily t o separate the organism from its sensory input. Even when a specific sensory input is not directly concerned in a given act of behavior it may so greatly contribute t o the total state of excitation in the central nervous system that the unrelated action cannot occur without it. Nevertheless, the new results have largely served to show the extent to which behavior can be programmed by central neural program stores and much further research will be directed towards elucidating the modes of storage of this information and of its “read-out” by the motor elements. The controversy as t o whether behavior is the result of reflexes or central patterns is clearly a sterile one, for it is already clear that there is every shade of intermingling of the extremes as well as shifts from one t o the other during a specific activity. However, if any side is going t o have the “last say” it is bound to be the reflex dogmatists because every complex action involves pathways to which contact is made by sensory input, and this is bound to influence it. It is hardly surprising that strong stimulation of sense organs often results in behavior. It is, however, a big leap from such results t o the suggestion that behavior is all caused to occur in this manner, and it is unfortunate that the suggestion was ever supported as strongly as it was. It is surprising, perhaps, that even those most intimately concerned with these matters are

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

427

sometimes not given to making concessions. Loher and Hiiber were quite convinced that their ablation and immobilization experiments on Gomphocerus prove that the command sequences are not dependent on peripheral feedback. They compare it with walking, which they regard as definitely due to chains of reflexes. In fact, most modern students of insect locomotion will probably dispute the latter. The central nervous system is probably just as capable of producing the functional patterns of motor output for walking in the absence of proprioceptors as it is for courtship, but if strong sensory input occurs, as must the case during ordinary walking, the nervous system must be influenced by it and may even be dominated by it in a given sequence. This is not t o say that it is always so dominated or that the sensory input directly causes each succeeding movement. It would be interesting to know if the courtship pattern of Gomphocerus can be generated in the total absence of proprioception from the units involved in the courtship behavior. Furthermore, it would be of interest to know how long after a particular sensory deprivation the pattern sequence retains its “perfect timing”. The reason for raising this question is that the interplay between sense organs and motor program stores may be a continual one, with a “learning” process influencing the central program. The central neurons may produce, by chance, too low or too high a frequency of discharge in a particular situation, too early or too late an onset, producing either a weak or an exaggerated movement or an ill-timed one. Any of these imperfections will be reflected in the sensory feedback. As was shown experimentally in the case of the postural motorneuron supplying the locust anterior coxal adductor, this results in a long-term adjustment which can be equated with a learning process. A basic frequency of firing of the tonic motorneuron is probably due to biophysical features whi’ch are inherited. This is the simplest possible kind of central program. But even this is not fixed absolutely like the grooves in a gramophone record. The program can be “erased” and “re-recorded” within a few minutes. The analogy with magnetic tape is thus very good indeed. The tape as such is the inherited neural machinery and it comes with a tune impressed on it, but if a part of the “tune” is inadequate it can be modified to some extent so as to d o a better overall job. Doubtless the extent of this flexibility varies with the activity, but variability must be associated with all behavior, both for immediate adaptation and for long-term evolution to be possible. It is inevitable that during the course of evolution organisms shall

428

G . HOYLE

tend towards progressive improvement of one or another aspect of their neural as well as their skeletal machinery. Where complex actions are involved, i t would seem that progressive loss of dependence of the central programming machinery upon the vicissitudes of multifarious sensory inflow has been favored. Reliability and reproducibility of actions are here the keynote. With very short life-spans there was no other way. But there may be a modicum of plasticity in even the more rigorous behaviors, which in turn are regulated by neural program stores. The motorneural patterns, which are perhaps generated in a different set of neurons, driven by the neural program stores, can possibly be re-programmed, at least to a limited extent, within the life-time of the animal, by a learning process. This takes care of problems arising out of different stresses associated with growth, at one extreme, or permits even the use of different limbs in an act such as grooming, should damage or ablation occur. Wilson (1968) has very recently suggested that there may be a continuum of control mechanisms, rather than any simple categorization of the kind I proposed, but there will be no possibility of directly resolving the issue until neuroethology has made some giant strides forward. Unless one wishes t o recognize fine divisions of control, only three fundamentally different mechanisms are possible. These are reflexes, motor neural program store control and sensory neural program store control. The latter two must both feed information into an effector system which is subject to sensory input capable of causing reflex action-unless it can be fully inhibited locally. Certain aspects of a complex behavior may be “commanded” overall from a motor program store, with other aspects determined by a sensory program store and yet other, perhaps individual muscular movements, dependent upon reflex actions. A neural program store could serve as a motor program for one read-out channel and a sensory program for a different read-out channel acting simultaneously. At the same time it could act as a general excitor for a third channel, a general inhibitor for a fourth, and even as an inhibitor of inhibition for a fifth. The ingenious reader will doubtless think of several other combinations. It may be seen from the above that it will not help our understanding of neuroethology for either the reflex school or the central programs school to continue t o pursue the study of exclusive aspects of behavior and exaggerate their importance. The emphasis must be switched away from the examination of either actions which can be reflexly evoked or those which can be observed t o occur

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

429

spontaneously, to the study of the functioning of nerve cells as producers of motor neural patterns. We may ask the question: to what extent have the detailed aspects of the neural machinery of insects evolved in parallel with those of vertebrates or even other major invertebrate groups, notably the mollusks? Have they diverged also and invented unique mechanisms? The small numbers of neurons available to insects must have set some limits to their capabilities and restricted the modes of information processing. What these differences might be, however, we still cannot state, since knowledge of central mechanisms in all groups is so limited. There seems to be commonality in certain aspects of the processing of certain kinds of sensory input, especially that related to the special senses: visual, olfactory and auditory. These are at least in part concerned to detect a few specific clues, patterns of movement, combinations of odours or sounds which relate to the life of the animal, which signal for example food to be procured, a predator to be avoided, a partner to mate with. These clues lead to convergence upon what present data from all investigated phyla suggests are single nerve cells which are excited by this input, and by it alone. The cells then cause appropriate co-ordinated behavior to occur automatically, presumably on account of their connectivity and the reactivity of the cells they excite. In turn, the cells which they excite generate sequences of motor impulses, perhaps of great complexity, lasting from a few milliseconds to many minutes, the executive sequences of which are to a considerable extent independent of proprioception. The extent of dependence is, however, very different in different actions. Collectively, the cells generate a neural program. A lot of recent research suggests that two classes of cell are of importance in influencing insect behavior. These are general inhibitors and general excitors. These certainly have their counterpart in the vertebrate brain and the upper part of the spinal cord. One difference is that in insects their total number may be quite small. All motor activity requires that either general excitors be active or that inhibitory activity be suppressed, or that both occur simultaneously, in addition to the activation of specific excitatory of disinhibitory pathways. These general excitatory and inhibitory cells have ramifying connections which permeate the ganglia. They raise or lower the excitability of both interneurons and motorneurons alike. Their detailed mode of action is not yet known, but it is possible that they act electrotonically rather than by chemically transmitting synapses. It has not been the intention of those who have emphasized central programs to belittle the importance of sensory input in

430

G . HOYLE

general, or reflexes in particular. These are obviously needed both t o maintain the nervous system in a sufficiently excited state to permit any kind of action, as well as for orientation, posture, etc. Their purpose was to redress a balance which was upset by overzealous supporters of reflexological theories of behavior. In this they have been successful, and any further discussion of the subject until more information exists concerning unitary activity during behavior seems futile. The worst offenders are, in my opinion, those who belittle in advance the knowledge to be obtained by microelectrode recording. These methods are admittedly intrinsically limited, but the limits are also equated with the extent to which understanding of the nervous system is possible. What we cannot find out by these methods cannot be found out at all in terms which are worthy of serious consideration. If a final projection may be permitted, it is that single neurons will be found to be of much greater importance than has hitherto been believed. The simplest motor neural program store or neural pattern generator will probably turn out to be a single nerve cell. The information content of the discharge pattern of a single neuron which fires several impulses when itself excited by a unitary step, together with amplification of information which may be achieved by “read-out” mechanism at follower synapses, is quite sufficient for programming behavior. Complex stores may turn out to be built up from hierarchical “cascades” of lesser stores, each a single cell, rather than complex networks for which the whole is greater than the sum of its parts. At least, I hope so, for otherwise it may not be possible fully to comprehend the functioning of even the insect nervous system. Readers who have not lost their interest after reading this article, and who have made some excursions into the original literature, will wish to read the witty, scholarly and thoughtful article by J . S. Kennedy (1967). They will find there much food for thought, but few spurs to action. My advice to students has been and will continue to be: record first, think later-then record some more. In the long run we can never have too much data.

ACKNOWLEDGEMENTS

The original research reported in this article was supported by research grant GB 3 160 from the U.S. National Science Foundation. It was written during the tenure of a Guggenheim Foundation

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

431

Fellowship whilst on sabbatical leave at the Department of Physiology, University College, London, and at the Station Zoologique, Villefranche (University of Paris). I am indebted to Dr Malcolm Burrows, Dr Fred Delcomyn, Dr D. Bentley and Dr D. M. Wilson for helpful comments on the manuscript. Especial thanks are due to Dr David Bentley for permission to include material which was in press at the time of writing this article. REFERENCES Baerends, G. P. ( 1941). Fortpflanzungsverhalten und Orientierung der Grabwespe Ammophile campestris, Jr. Tijd schr. Ent. 84,68-275. Baerends, G. P. (1959). Ethological studies of insect behaviour. A . Rev. Ent. 4, 207-234. Bastock, M. and Manning, A. (1 955). The courtship of Drosophila melanogaster. Behaviour, 8,85-111. Bentley, D. R. (1969a). Intracellular activity in cricket neurons during the generation of behaviour patterns. J. Insect Physiol. 15, 677-699. Bentley, D. R. (1969b). lntracellular activity in cricket neurons during generation of song patterns. Z . vergl. Physiol. 62, 267-283. Bentley, D. R. ( 1 9 6 9 ~ ) .Topological relationships of the locust flight motor neurons. A m . Zool. 9 , 1106-1 107(A). Bentley, D. R. and Kutsch, W. (1966). The Neuromuscular mechanism of stridulation in crickets (Orthoptera: Grillidae). J. exp. Biol. 45, 15 1-164. Bishop, L. G. and Keehn, D. G. (1967). Neural correlates of the optomotor response in the fly. Kybernetik 3,288-295. Bishop, L. G., Keehn, D. G. and McCann, G. D. (1968). Motion detection by interneurons of optic lobes and brain of the flies Calliphora phaenicia and Musca domestica. J. Neurophysiol. 31,509-525. Blest, A. D. (1960). The evolution, ontogeny and quantitative control of the settling movements of some new world saturnid moths, with some comments on distance communication by honey-bees. Behavior 16, 188-253. Blest, A. D. and Collett, T. S. (1965). Microelectrode studies of the medial protocerebrum of some Lepidoptera. I. Responses to simple, binocular visual stimulation. J. Insect Physiol. 11, 1079-1 102. Boistel, J. (1968). The synaptic transmission and related phenomena in insects. In “Recent Advances in Insect Physiology” (J. E. Treherne and J. W. L. Beament, eds), Vol. 5, pp. 1-64. Academic Press, New York. Bullock, T. H. (1961). The origins of patterned nervous discharge. Behaviour 17, 48-59. Bullock, T. H. (1966). Strategies for blind physiologists with elephantine problems. Symp. SOC.exp. Biol. 20, 1-10. Bullock, T. H. and Horridge, A. (1965). “Structure and Function in the Nervous Systems of Invertebrates”. 2 vols., 1719 pp. Freeman, San Francisco. Burtt, E. T. and Catton, W.T. (1960). The properties of single unit discharges in the optic lobe of the locust. J. Physiol. 154,479-490.

432

G. HOYLE

Callec, J. J. and Boistel, J. (1965). Analysis with microelectrodes of the synaptic transmission at the level of the sixth abdominal ganglion of a cockroach, Periplaneta americana. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 59-65. Academic Press, New York. Callec, J. J. and Boistel, J. (1966). PhCnomknes d’excitation et d’inhibition au niveau du dernier ganglion abdominal de la Blatte, Periplaneta americana, L . C o m p t . Rendus 160,2418-2420. Callec, J. J. and Boistel, J. (1967). Les effets de I’acCtylcholine aux niveaux somatique et synaptique dans le cas du 6e’me ganglion abdominal de Periplaneta arnericana. Compt. Rendus 161,442-446. Chapple, W. D. (1966). Motorneuron responses to visual stimuli in Oncopeltus fasciafus Dallas. J. exp. Biol. 45, 4 0 1-410. Coggeshall, R. E. (1967). A light and electron microscope study of the abdominal ganglion of Aplysia californica. J. Neurophysiol. 30, 12 63- 1 287. Cohen, M. J. (1967). Correlation between structure, function and RNA metabolism in central neurons of insects. In “Invertebrate Nervous Systems” (C. A. G. Wiersma, ed.), pp. 65-78. Chicago University Press. 370 pp. Cohen, M. J. and Jacklet, J. W. (1967). The functional organization of motorneurons in an insect ganglion. Phil. Trans. R . SOC. B 2 5 2 , 561-569. Cohen, M. J., Jacklet, J. W. and Eisenstein, E. M. (1968). Neural regeneration and behavior in insects. In “Physiological and Biochemical Aspects of Nervous Integration” (F. D. Carlson, ed.), pp. 267-306. Prentice Hall. 391 pp. Davies, W. J. (1968). Cricket wing movements during stridulation. A n . Behav. 16.72-73. Delcomyn, F. (1969). Reflexes and locomotion in the American cockroach. Ph.D. Thesis, University of Oregon. Dingle, H. and Fox, S. S. (1966a). Microelectrode analysis of light responses in the brain of the cricket (Gryllus campestris). J. Cell. comp. Physiol. 68, 45-60. Dingle, J. and Fox, S. S. (1966b). Responses t o cercal stimulation recorded in the cricket brain. Nature, Lond. 210, 1950. Dowling, J. E. and Boycott, B. B. (1967). Organisation of the primate retina: electron microscopy. Proc. R. SOC. B 166,80-111. Dugard,.J. J. (1967). Directional change in flying locusts. J. Insect Physiol. 13, 1055-1063. Eisenstein, E. M. and Cohen, M. J. (1965). Learning in an isolated prothoracic insect ganglion. A nim. Behav. 13, 104-108. Eisenstein, E. M. and Krasilovsky, G. H. (1967). Studies of learning in isolated insect ganglia. In “Invertebrate Nervous Systems” (C. A. G. Wiersma, ed.), pp. 329-332. Chicago University Press. 370 pp. Elsner, N. ( 1968). Die neuromuskularen Grundlagen des Werbeverhaltens der Roten Keulenheuschrecke Gomphocerippus rufus (L). Z . vergl. Physiol. 60,308-350. Evans, H. E. (1957). “Comparative Ethology of Digger Wasps of the Genus Bembix”. Cornell University Press, New York.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

433

Evans, H. E. (1966). The behavior patterns of solitary wasps. A . Rev. Ent. 11, 123-154. Evarts, E. V. (1968). Relation of pyramidal tract activity to force exerted during voluntary movement. J. Neurophysiol. 31, 14-27. Ewing, A. and Hoyle, G. (1965). Neuronal mechanisms underlying control of sound production in a cricket: Acheta domesticus. J. exp. Biol. 43, 139-153. Ewing, A. and Manning, A. (1966). Some aspects of the efferent control of walking in three cockroach species. J. Insect Physiol. 12, 11 15-1 118. Ewing, A. and Manning, A. (1967). The evolution and genetics of insect behaviour. A . Rev. Ent. 12,471-494. Farley, R. D., Case, J. F. and Roeder, K. D. (1967). Pacemaker for tracheal ventilation in the cockroach, Periplaneta americana (L.). J. Insect Physiol. 13, 1713-1728. Fielden, A. (1963). Properties of interneurones in the abdominal nerve cord of a dragonfly nymph. J. exp. Biol. 40,441-452. Frisch, K. von. (1967). “The Dance Language and Orientation of Bees”. Harvard University Press. 566 pp. Galloway, N. S., Runion, H. I. and Usherwood, P. N. R. (1966). An electrophysiological approach to the study of insect behaviour. J. Physiol. 187.9-10 P. Guthrie, D. M. (1961). The anatomy of the nervous system in the genus Gerris (Hermiptera-Heteroptera). Phil. Trans. R . SOC.B 244. 65-102. Guthrie, D; M. (1964). Observations on the nervous system of the flight apparatus in the locust Schistocerca gregaria. Q. Jl. microsc. Sci. 105, 183-201. Hagiwara, S . and Watanabe, A. (1956). Discharges in motorneurons of cicada. J . Cell. comp. Physiol. 4 7 , 4 15-428. Haskell, P. T. (Ed.). (1966). “Insect Behaviour”. Symp. R o y . Ent. SOC.Lond. No. 3. Hassenstein, B. (1958). Uber die Wahrnehrnung der Bewegung von Figuren und unregelmbsigen Helligkeitsmustern. 2. vergl. Physiol. 4 0 , s 56-592. Hess. A. (1958). The fine structure of nerve cells and fibers, neuroglia and sheaths of the ganglion chain in the cockroach. J. Morph. 103,479-502. Hess, A. (1960). The fine structure of degenerating nerve fibres, their sheaths and their terminations in the central nerve cord of the cockroach (Periplaneta americana). J. biophys. biochem. Cytol. 7,339-344. Holst, E. von. (1948). Von der Mathematik der nervosen Ordnungsleistung. Experientia, 4, 374-38 1. Horn, G. (1967). Neural mechanisms of habituation. Nature, Lond. 215, 707-7 1 1. Horridge, G. A. (1962). Learning leg position by the ventral nerve cord in headless insects. Proc. R . SOC.B 157, 33-52. Horridge, G. A. (1966). Optokinetic memory in the locust. J. exp. Biol. 44, 255-26 1. Horridge, G. A. (1968). “Interneurons”. Wiley & Sons, New York. 436 pp. Horridge, G. A., Scholes, J. H., Shaw, S. and Turnstall, J. (1965). Extracellular recordings from single neurones in the optic lobe and brain of the locust. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne

434

G . HOYLE

and J. W. L. Beament, eds), pp 165-202. Academic Press, London and New York. Hoyle, G. (1957). Nervous control of insect muscles. In “Recent Advances in Invertebrate Physiology” (B. T. Scheef, ed.), pp. 73-98. University of Oregon Press. 304 pp. Hoyle, G. ( 1964). Exploration of neuronal mechanisms underlying behaviour in insects. In “Neural Theory and Modelling” (R. F. Reiss, ed.), pp. 346-376. Stanford University Press. Hoyle, G. ( 1965). Neurophysiological studies on “learning” in headless insects. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 203-232. Academic Press, London and New York. Hoyle, G. ( 1966a). An isolated insect ganglion-nerve-muscle preparation. J. exp. B i d . 44, 413-427. Hoyle, G. ( 1966b). Functioning of the inhibitory-conditioning axon innervating insect muscles. J. exp. Biol. 44, 429-453. Hubel, D. H. and Wiesel, T. N. (1959). Receptive fields of single neurons in the cat’s striate cortex. J. Physiol. 148, 574-591. Hiiber, F. ( 1959). Auslosung von Bewegungsmustern durch electrische Reizung des Oberschlundgangions bei Orthopteren (Saltatoria: Gryllidae, Acridae). 2001.A ~ z (Suppl. . 23), 248-269. HGber, F. (1962). Central nervous control of sound production in crickets and some speculations on its evolution. Evolution, Lancaster, Pa. 16,429-442. Huber, F. (1965a). Akruelle Probleme in der Physiologie des Nervensystems der Insekten. Naturw. Rdsch. Braunschw. 18, 143-156. Hiiber, F. (1965b). Brain controlled behaviour in Orthopterans. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 233-246. Academic Press, London and New York. Hiiber, F. (1967). Central control of movements and behaviour of invertebrates. In “Invertebrate Central Nervous Systems” (C. A. G. Wiersma, ed.), pp. 333-35 1. Chicago University Press. 370 pp. Hughes, G. M . (1952). Differential effects of direct current on insect ganglia. J. exp. Biol. 29,387-402. Hughes, G. M. ( 1965a). Locomotiog: terrestrial. In “The Physiology of Insecta” (M. Rockstein, ed.), Vol. 11, pp. 227-254. Academic Press, New York. Hughes, G. M. (1965b). Neuronal pathways in the insect central nervous system. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J . W. L. Beament, eds), pp. 79-1 12. Academic Press, London and New York. 277 pp. Hughes, G. M. and Mill, P. J. (1966). Patterns of ventilation in dragonfly larvae. J. e x p . Biol. 44, 317-333. Iwasaki, S. and Wilson, D. M. (1966). Neuropile potentials following wing sensory nerve stimulation in locusts. C o m p . Biochem. Physiol. 17, 693-695. Kammer, A. (1967). Muscle activity during flight in some large Lepidoptera. J. exp. Biol. 47,277-295. Kandel, E. R. and Spencer, W. A. (1968). Cellular neurophysiological approaches in the study of learning. Physiol. Rev. 48, 65-134.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

435

Kendig, J. J. (1968). Motor neurone coupling in locust flight. J. exp. Biol. 48, 389-404. Kennedy, D. (1966). The comparative physiology of invertebrate central neurons. A d v . comp. Physiol. 2, 117-184. Kennedy, D. and Preston, J. B. (1960). Activity patterns of interneurons in the caudal ganglion of the crayfish. J. gen. Physiol. 43,655-670. Kennedy, J. S. (1967). Behavior as physiology. In “Insects and Physiology” (J. W. L. Beament and J. E. Treherne, eds), pp. 249-265. Oliver and Boyd, Edinburgh and London. 278 pp. Kerkut, G. A., Pitman, R. M. and Walker, R. J. (1968). Electrical activity in insect nerve cell bodies. Life Sciences 7 , 605-608. Knights, A. (1965). The activity of single motor fibres in arthropods. The dragonfly nymph. J. exp. Biol. 42,447-461. Krasne, F. B. and Roberts, A. (1967). Habituation of the crayfish escape response during release from inhibition induced by pirotoxin. Nature, Lond. 215,769-770. Leghissa, L. ( 1942). Peculiari connessioni nervose nella catena ganglionare di Carausius morosus. Arch. zool. ital. 30, 289-310. Loher, W. and Hiber, F. (1966). Nervous and endocrine control of behaviour in a grasshopper (Gomphocerus rufus L., Acrididae). S y m p . SOC.exp. Biol. 20, 38 1-400. Lindauer, M. ( 1961). “Communication Among Social Bees”. Harvard University Press, Massachusetts. Luco, J. V. and Aranda, L. C. (1964). An electrical correlate to the process of learning. A cta physiol. Latinoam. 14, 274-288. Manning, A. (1959). The sexual behaviour of two sibling Drosophila species. Behaviour 15, 123-145. Manning, A. (1961). The effects of artificial selection for mating speed in Drosophila melanogaster. Anim. Behav. 9,82-92. Manning, A. (1967). “An Introduction to Animal Behaviour”. Arnold, London. 208 pp. Maturana, H. R., Lettvin, J. Y., McCulloch, W. S. and Pitts, W. H. (1960). Anatomy and physiology of vision in the frog (Rana pipiens). J. gen. Physiol. 43, 129-175. Markl, H. and Lindauer, M. (1965). Physiology of insect behaviour. In “The Physiology of Insecta” (M. Rockstein, ed.), pp. 3-122. Academic Press, New York. Maynard, D. M. (1956). Electrical activity in the cockroach cerebrum. Nature, Lond. 177,529-530. Maynard, D. M . (1962). Organization of neuropile. A m . Zool. 2,79-96. Maynard, D. M. (1967). Organization of central ganglia. In “Invertebrate Nervous Systems” (C. A. G. Wiersma, ed.), pp. 231-258. Chicago University Press. 370 pp. Mill, P. J. (1963). Neural activity in the abdominal nervous system of Aeschna nymphs. C o m p . Biochem. Physiol. 8,83-98. Mill, P. J. (1964). The structure of the abdominal nervous system of aeschnid nymphs. J. comp. Neurol. 122, 157-171. Mill, P. J. and Hughes, P. J. (1966). The nervous control of ventilation in dragonfly larvae. J. exp. Biol. 44,297-31 6.

436

G . HOYLE

Miller, P. L. (1966). The regulation of breathing in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), pp. 279-354. Academic Press, London and New York. Miller, P. L. (1967a). The derivation of the motor command to the spiracles of the locust. J. exp. Biol. 46, 349-37 1. Miller, P. L. (1967b). The origins of motor acts in insects. In “Insects and Physiology” (J. W. L. Beament and J. E. Treherne, eds), pp. 267-299. Oliver and Boyd, Edinburgh and London. 378 pp. Mittelstaedt, H. (1962). Control systems of orientation in insects. A. Rev. Ent. 7, 177-198. Neville, A. C. (1960). Aspects of flight mechanics in anisopterous dragonflies. J. exp. Biol. 37,631-656. Neville, A. C. (1963). Motor unit distribution of the dorsal longitudinal flight muscles in locusts. J. exp. Biol. 40, 123-136. Osborne, M. P. (1966). The fine structure of synapses and tight junctions in the central nervous system of the blowfly larva. J. Insect Physiol. 12, 1503-1512. Pearson, K. G. and Bergman, S. J. (1 969). Common inhibitory motoneurones in insects. J. Exp. Biol. 50, 4 4 5 4 7 1. Pierce, G. W. (1948). “The Songs of Insects”. Harvard University Press, Massachusetts. 329 pp. Preston, J. B. and Kennedy, D. (1960a). Activity patterns of interneurones in the caudal ganglion of the crayfish. J. gen. Physiol. 43,655-670. Preston, J. B. and Kennedy, D. (1960b). Integrative synaptic mechanisms in the caudal ganglion of the crayfish. J. gen. Physiol. 43,67 1-68 1. Pringle, J. W. S. (1940). The reflex mechanism of the insect leg. J. exp. Biol. 17, 3-17. Pringle, J. W. S. (1954). The mechanism of the myogenic rhythm of certain insect striated muscles. J. Physiol. 124, 269-29 1. Remler, M., Selverston, A. and Kennedy, D. (1968). Lateral giant fibers of crayfish: location of somata by dye injection. Science, N. Y . 162,28 1-283. Reiss, R. F. (1962). A theory and simulation of rhythmic behaviour due to reciprocal inhibition in small nerve nets. A m . Fed. In$ Proc. 21, 171-194. Roberts, A. (1968). Some features of the central co-ordination of a fast movement in the crayfish. J. exp. Biol. 49,645-656. Roeder, K . D. (1937). The control of tonus and locomotor activity in the praying mantis (Mantis religiosa L.). J . exp. Zool. 76, 353-374. Roeder, K. D. (1948). Organization of the ascending giant fiber system in the cockroach (Periplaneta americana). J. exp. Zool. 108,243-261. Roeder, K. D. (1953). “Insect Physiology”. Wiley, New York. Roeder, K. D. (1959). A physiological approach to the relation between prey and predatory. Smithson. misc. Collns 137,287-306. Roeder, K. D. (1963). “Nerve Cells and Insect Behavior”. Harvard University Press. Massachusetts. 238 pp. Roeder, K . D. (1966). Interneurons of the thoracic nerve cord activated by tympanic nerve fibres in noctuid moths. J. Insect Physiol. 12, 1227-1244. Roeder, K. D. and Payne, R. S. (1966). Acoustic orientation of a moth in flight by means of two sense cells. S y m p . SOC.exp. Biol. 20,25 1-272. Roeder, K. D., Tozian, L. and Weiant, E. A. (1960). Endogenous nerve activity and behaviour in the mantis and cockroach. J. Insect Physiol. 4,45-62.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

437

Rothenbuhler, W. C. (1964a). Behaviour genetics of the nest cleaning in honey-bees. I. Responses of four inbred lines to disease-killed brood. Anim. Behav. 12,578-583. Rothenbuhler, W. C. (1964b). Behaviour genetics of nest cleaning in honey-bees. IV. Responses of F1 and backcross generations to disease-killed brood. Am. ZOO^. 4, 111-123. Rowe, E. C. (1960). Activity of single nerve cells in an insect thoracic ganglion. Anat. Rec. 137,389. (Abstr.) Rowe, E. C. (1963). Microelectrode records from an insect thoracic ganglion. Ph.D. Thesis, University of Michigan. Rowell, C. H. F. (1963). A method for implanting chronic stimulating electrodes in the brains of locusts, and some results of stimulation. J. exp. Biol. 40, 27 1-284. Rowell, C. H. F. (1964). Central control of an insect segmental reflex. I. Inhibition by different parts of the central nervous system. J. exp. Biol. 41,559-572. Rowell, C. H. F. and Horn, G. (1967). Response characteristics of neurones in an insect brain. Nature, Lond. 216, 702-703. Runion, H. I. and Usherwood, P. N. R. (1966). A new approach to neuromuscular analysis in the intact free-walking insect preparation. J. Insect Physiol. 12, 1255-1263. Runion, H. I. and Usherwood, P. N. R. (1968). Tarsal receptors and leg reflexes in the locust and grasshopper. J. exp. Biol. 49,42 1-436. Runion, H. I. and Usherwood, P. N. R. (1969). (In preparation.) Smith, D. S. (1965). Synapses in the insect nervous system. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 39-57. Academic Press, London and New York. Smith, D. S . (1967). The organization of the insect neuropile. In “Invertebrate Nervous Systems” (C. A. G. Wiersma, ed.), pp. 79-85. Chicago University Press. 370 pp. Smith, D. S. and Treherne, J. E. (1963). Functional aspects of the organization of the insect nervous system. In “Advances in Insect Physiology. 1” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), pp. 401-484. Academic Press, London and New York. Stretton, A. 0. W. and Kravitz, E. A. (1968). Neuronal geometry: determination with a technique of intracellular dye injection. Science, N . Y . 162, 132-134. Suga, N. and Katsuki, Y. (1961). Pharmacological studies on the auditory synapses in a grasshopper. J. exp. Biol. 38,759-770. Thorpe, W. H. (1963). “Learning and Instinct in Animals”. 2nd Edition, Methuen, London. Thorson, J. (1964). Dynamics of motion perception in the desert locust. Science, N . Y . 145,69-7 1. Tinbergen, N. (195 1). “The Study of Instinct”. Oxford University Press. TrujilloCCnoz, 0. (1962). Some aspects of the structural organization of the arthropod ganglia. Z . Zellforsch. mikrosk. Anat. 56,649-682. Vowles, D. M. (1964). Models and the insect brain. In “Neural Theory and Modelling” (R. F. Reiss, ed.), pp. 349-373. Stanford University Press. Waldron, I. (1967a). Mechanisms for the production of the major output pattern in flying locusts. J. exp. Biol. 47,201-2 12.

438

G . HOYLE

Waldron, I. (1967b). Neural mechanism by which controlling inputs influence motor output in the flying locust. J. exp. Biol. 47, 213-228, Waldron, I. (1968). The mechanisms of coupling of the locust flight oscillator to oscillatory input. 2.vergl. physiol. 57, 331-347. Weevers, R. de G. (1965). Proprioceptive reflexes and the co-ordination of locomotion in the caterpillar of Antheraen pernyi (Lepidoptera). In “The Physiology of the Insect Nervous System” (J. E. Treherne and J. W. L. Beament, ed.), pp. 113-124. Academic Press, London and New York. Weis-Fogh, T. (1956). Biology and physics of locust flight. IV. Notes on sensory mechanisms in locust flight. Phil. Trans. R . SOC.B239,553-584. Weis-Fogh, T. (1964). Control of basic movements in flying insects. S y m p . SOC. exp. Biol. 18, 343-363. Wells, G. P. (1950). Spontaneous activity cycles in polychaete worms. S y m p . SOC.exp. Biol. 4, 127-142. Wendler, G. (1961). Die Regelung der Korperhaltung bei Stabheuschrecken (carausius rnorosus). Natunvissenschaften 48, 667-677. Wendler, G. (1964). Laufen und Stehen der Stabheuschrecke Carausius morosus: Sinnesborstenfelder in den Beingelenken als Glieder von Regelkreisen. 2. vergl. Physiol. 48, 198-250. Wendler, G. (1966). The co-ordination of walking movements in arthropods. S y m p . SOC.exp. Biol. 20, 229-249. Wiersma, C. A. G. and Ikeda, K . (1964). Interneurones commanding swimmeret movements in the crayfish, Procambarus clarkii (Girard). C o m p . Biochem. Physiol. 12,509-525. Wigglesworth, V. B. (1959). The histology of the nervous system of an insect, Rhodnius probiscus (Hemiptera). 11. The central ganglia. Q. JI. microsc. Sci. 100,299-313. Willows, A. 0. D. (1967). Behavioral acts elicited by stimulation of single identifiable brain cells. Science, N . Y. 157, 570-574. Wilson, D. M . (1961). The central nervous control of flight in a locust. J. e x p . Biol. 38,47 1-490. Wilson, D. M. (1962). Bifunctional muscles in the thorax of grasshoppers. J. e x p . Biol. 39,669-677. Wilson, D. M. (1964). Relative refractoriness and patterned discharge of locust flight motorneur0ns.J. exp. Biol. 41, 191-205. Wilson, D. M. (1965a). Proprioceptive leg reflexes in cockroaches. J. exp. Biol. 43,397-409. Wilson, D. M. (1965b). The nervous co-ordination of insect locomotion. In “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 125-140. Academic Press, London and New York. Wilson, D. M. (1966). Insect walking. A. R e v . Ent. 11, 103-122. Wilson, D. M . (1967). An approach to the problem of control of rhythmic behavior. In “Invertebrate Nervous Systems” (C. A. G. Wiersma, ed.), pp. 2 19-229. Chicago University Press. Wilson, D. M. (1968). The nervous control of insect flight and related behavior. In “Recent Advances in Insect Physiology” (J. E. Treherne and J. W. L. Beament, eds). Academic Press, New York.

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

439

Wilson, D. M . and Weis-Fogh, T. (1962). Patterned activity of co-ordinated motor units, studies in flying locusts. J . e x p . Biol. 39, 643-667. Wilson, D. M . and Wyman, R . J . (1965). Motor output patterns during random and rhythmic stimulation of locust thoracic ganglia. Biophys. J . 5 , 12 1-143. Yamasaki, T. and Narahashi, T. (1960). Synaptic transmission in the last abdominal ganglion of the cockroach. J . Insect Physiol. 4, 1-13. Zawarzin, A. ( 1924). Uber die histologische Beschaffenheit des unpaaren ventralen Nervs der Insekten (Histologische Studien iiber Insekten V). Z . wiss. Zool. 122,97-115.

NOTE ADDED IN PROOF

Whilst this article was being edited and typeset several important papers appeared. In addition, a most significant development has occurred following the use of improved techniques for impaling motorneurons in our laboratory, and even the preliminary results necessitate a revised statement. The publishers have kindly permitted the inclusion of this short addendum to cover these developments. A nice behavioral study giving insight into aspects of the neural control mechanisms was published by Elsner and Huber (1 969). The subject of their investigation was the courtship sequence shown by male grasshoppers of the species Gomphocerippus rufus, from which electromyograms were recorded in the freely-moving insect with implanted, trailing leads. The head, antennae and hind legs all give discrete sequential movements during courtship, which is evoked by either optical, acoustic, or tactile stimuli. The movement patterns evoked do not seem to be simply reflexive, however. Head movements have two phases, a slow one and a fast one. Visual stimuli given during the former can interrupt the sequence, but when the same stimuli are given during the latter there is no response. The complex temporal sequences of movements were not changed by amputation of, or fixing, the hind legs but individual motorneuron discharges could decrease in mean frequency after the operation. Loading the hind legs did not alter the discharge patterns. The investigators connected the hind legs together mechanically and recorded from both simultaneously. The motor impulses were still as much out of phase as if the legs could move independently. In other experiments the head was fixed to the prothorax but electrical activity in the head-moving muscles was unchanged during elicitation of the

440

G . HOYLE

courtship behaviour, apart from a slight lowering of the frequency of major bursts (equals head oscillations) at the beginning of a courtship sequence. One simple operation, that of fixing the pterothorax to the head, did have a dramatic influence: song was then abolished. Either one of the head/thorax connectives could be severed without altering the behaviour. These results provide further evidence for the existence of neural pattern generators or “tapes”, which are in overall control of complex behaviour, but in which a small degree of modification occurs in each sequence as a result of local reflex action when the sequential movements occur. Additional light on the means by which locomotion is determined has been provided by Pearson and Iles (1970) for the cockroach. They recorded extracellularly from nerves supplying metathoracic leg levator and depressor muscles whilst evoking walking movements. There was a marked reciprocity, with very little overlap. Next they cut off all sources of sensory feed-back from all legs, to find that the alternating rhythm persisted. Unfortunately, all the movements described were slow, more than 10 times as slow as occur in a fast-running cockroach, but they nevertheless serve conclusively to demonstrate that a central locomotory rhythm generator exists in a cockroach ganglion. This paper draws attention to the need for more rapid relaxation in some of the leg muscles than occurs naturally and suggests that peripheral inhibition may be an important aid in achieving this. The problem has actually been tackled experimentally in extremely well-conceived direct tests using freely-walking locusts with implanted leads, by Usherwood and Runion (1 970). To overcome the intrinsic limitations upon analysis, for example of determining forces developed, in the free-walking preparation, they used the impulses actually recorded from nerves to excite, following amplification, exactly similar patterns in simple preparations. In this way they were able to demonstrate an actual useful utilization of a peripheral inhibitor in the intact insect. At the same time, their results showed that reflexes are very important in locust locomotion. The earlier overall conclusion is thereby reinforced, that in insect Fig. 1. Metathoracic ganglion of adult male Schistocercu greugu~tuviewed from ventral surface and photographed in ultraviolet light after fust filling both the “fast” extensor tibiae and “slow” extensor tibiae motorneurons with Rocion yellow. The larger, more posterior cell is the “fast” one. These cells are always found near these same positions. Experiment by G . Hoyle and M. Burrows, with technical assistance from B. Moberly. Photograph by H. Howard.

This Page Intentionally Left Blank

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

44 1

walking, unlike flight, the programming is the result of a combination of a central motor “tape” with reflex activity. Young (1969), who served as a technician for the project resulting in the cockroach metathoracic ganglion map (Fig. 1, above), has published a comparable survey of the mesothoracic ganglion. A study of electrical activity in the cockroach metathoracic ganglion using microelectrodes, part of which was referred to above, has been published (Rowe, 1969). Recordings indicative of synaptic activity were obtained, but were from unidentified sites. Bentley (1969) has published a brief note asserting his ability, referred to above, to penetrate mesothoracic motorneurons which drive phasic muscles involved in flight of locusts and to fill them with Procion yellow dye. The cells were identified by direct intracellular excitation, when the innervated muscles contracted. Subsequent serial sectioning is enabling him to produce a three-dimensional topographical model of the neurons, which occur in three clusters, corresponding to the three different nerve roots through which the axons emerge. Somata of neurons supplying a given muscle were found to be adjacent, those of anteriorly-located muscles being anterior in the ganglion. A remarkable success was obtained by Ikeda and Kaplan (1969), who were able to obtain a rhythmic discharge from a motorneuron of the thoracic ganglion of a Drosophilu. The specimens carried the sex-linked, single gene mutant HKIp, which is associated with rhythmic leg shaking under etherization. Two kinds of burst occur, a slow one, with 10 sec intervals, and a fast one, with 0.2 sec intervals. The latter is not affected by removing the head, but the former is thereby caused to stop. The cell from which the discharge was obtained could not be identified, but the spikes were about 50 mV. Since even the largest cell bodies are less than lop across, it seems probable that the electrode was inside one and that it was well invaded by the spike. This may be possible because, the insect being so small, the conduction distance is much less than in the larger insects referred to above in which action potentials invade the soma to only a small extent. Results of studies referred to above by Walker and Pitman have been published (Kerkut et ul., 1969). They were able to penetrate unidentified neurons in the cockroach metathoracic and sixth abdominal ganglion. In each case they removed the sheath. Resting potentials averaging 54 mV were recorded, with overshooting spikes of 8 1 mV. The discharges were spontaneous, but were accelerated by

442

G. HOYLE

acetylcholine in a concentration of only 1.31 x 1013mole, or hyperpolarized and inhibited by GABA in the even lower concentration of 1.05 x 10-I3mole. A step towards cell identification in the cockroach metathoracic ganglion was made by Rowe et ul. (1969), who were able to cause leg extension or flexion after impaling cell bodies and depolarizing them. In my own laboratory, I had been frustrated, as mentioned above, in several attempts spanning a 16-year period, to record from and excite locust neurons. Dr Dennis Willows, in our laboratory, has made a significant step in correlating activity in identified ganglion cells with behaviour by a preparation in which the whole, intact animal is free to move about whilst microelectrodes impale identified brain cells (Willows, 1967; Willows and Hoyle, 1969). In this preparation the glass microelectrodes used are shaped so as to render them relatively stiff, for Willows had found that the brain cells deteriorated when the sheath was removed; so it is left in place and electrodes driven through it by a vibration caused by tapping the micromanipulator arm holding them. The sheath surrounding the locust nervous system is tough, and it serves as an ionic pump (Hoyle, 1952, 1953). Hence it seemed worth while to try to adapt the Tritoniu technique to locust ganglia. These experiments have been carried out by Dr Malcolm Burrows and myself and they have been dramatically successful. We have been able to hold microelectrodes in one or two motorneurons or interneurons at a time, for up to a few hours after driving them across the sheath (Hoyle and Burrows, 1970). Almost all the motorneurons supplying metathoracic leg muscles have been located and identified by direct stimulation. Each cell was filled with Procion yellow dye and serial sections have been made after filling, enabling three-dimensional maps to be drawn. The probability of penetrating a selected neuron has risen to about 0.5 in the case of the larger cells, as we have gained experience. The neurons are in almost exactly the same location in every ganglion. Homologous neurons in the two halves of the ganglion are in similar locations. All the neurons penetrated were found to have electrically inexcitable soma membrane, as previously, but each was invaded electrotonically by its action potential so that it could be determined when the cell fired. Resting potentials were consistently from 50 to 56 mV and action potential heights were from 3 to 25 mV. Spike heights recorded from the somas of desheathed ganglia were much smaller than these or even absent, resulting in the belief that there is

CELLULAR MECHANISMS UNDERLYING BEHAVIOR

443

tension

membrane potent ia I myogram current

SLOW

EXTENSOR

tens ion

membrane myogram current

FAST

EXTENSOR

Fig. 2. Results of depolarization applied through intracellular microelectrodes to “slow” (A, B) and “fast” (C, D) metathoracic extensor tibiae motorneurons shown in Fig. 1. Traces in each case are from top down: tension recorded from tibia; intracellular potential; extracellular myogram from extensor muscle; current applied to soma. Calibrations: A, B-vertical 10 mV or 0.05 g; C, D-vertical 10 mV or 5.0 g; current-lx 10-8amp (vertical); time-200 msec in all.

no invasion of the cell body. The sheath potential was surprisingly large, at about 58 mV, inside negative. The cell bodies of the left metathoracic fast extensor tibiae and its partner the slow, filled together with Procion yellow, are illustrated in Fig. 1. The fast extensor soma is a flattened cell, 90-1 10 p across by 40 p deep, and is the largest cell body in the ganglion. The slow cell body is almost spherical, 3540 p across, and lies considerably anterior to the fast cell body. The smallest mo torneuron we have recorded from and filled was a flexor tibiae soma 30 p across. EPSP’s and IPSP’s of various sizes have been recorded with the electrode in the soma of both motor- and interneurons. In interneurons there is a continual barrage of junctional potentials in a large range of sizes, occurring at rather high rates.

444

G . HOYLE

Upon passing depolarizing current of sufficient intensity across the membrane of a motorneuron each has been made to fire action potentials, causing contraction (Fig. 2). In almost every case the cell had to be driven to a positive value several times the resting potential. By intracellular, extracellular and mechanical recording from the muscle or limb, each axon has been positively identified. A three-dimensional topographical map of leg motorneurons is being prepared.

REFERENCES Bentley, D. (1969). Topological relationships of the locust flight motorneurons. A m . Zool. 9, 1106-1 107(A). Elsner, N. and Huber, F. (1969). Die Organisation des Werbegesanges der Heuschrecke Gomphocerippus rufus in Abhangigkeit von zentralen und periphere Bedingungen. Z. vergl. Physiol. 65,389-423. Hoyle, G. (1952). High blood potassium in insects in relation to nerve conduction. Nature, Lond. 169, 281-282. Hoyle, G. (1953). Potassium ions and insect nerve muscle. J. exp. Biol. 30, 12 1-235. Hoyle, G . and Burrows, M. (1970). Intracellular studies on identified neurons of insects. Fedn Proc. Fedn A m . SOCSexp. Biol. 29. Ikeda, K. and Kaplan, W. D. (1969). Neural mechanism for specific leg movements of a mutant Drosophila. A m . Zool. 9,584(A). Kerkut, G. A., Pitman, R. M. and Walker, R. J. (1969). Iontophoretic application of acetylcholine and GABA onto insect central neurones. Comp. Biochem. Physiol. 31, 61 1-633. Pearson, K. G. and Iles, J. F. (1970). Discharge patterns of coxal levator and depressor motorneurones of the cockroach, Periplaneta americana. J. exp. Biol. 5 2 , 139-165. Rowe, E. (1969). Microelectrode records from a cockroach thoracic ganglion: synaptic potentials and temporal patterns of spike activity. Comp. Biochem. Physiol. 30,529-539. Rowe, E. C., Moberly, B. J., Howard, H. M. and Cohen, M. J. (1969). Morphology of branches of functionally-identified motorneurons in cockroach neuropile. A m . Zool. 9,1107(A). Usherwood, P. N. R. and Runion, H. I. (1970). Analysis of the mechanical responses of metathoracic extensor tibiae muscles of free-walking locusts. J. exp. Biol. 5 2 , 3 9 4 8 . Willows, A. 0. D. (1968). Behavioral acts elicited by stimulation of single identifiable nerve cells. Zn “Physiological and Biochemical Aspects of Nervous Integration” (F. D. Carlson, ed.), pp. 217-243. Willows, A. 0. D. and Hoyle, G. (1969). Neural network triggering a fixed action pattern.Science, N.Y. 166, 1549-1551. Young, D. (1969). The Motorneurons of the mesothoracic ganglion of Periplaneta americana. J. ZnsectPhysiol. 15, 1175-1179.

Author Index Numbers marked with an asterisk are the pages on which the references are listed A Abbott, J., 259,262* Abramyan, K. S., 17, 18, 71 * Adams, B. A., 18,91* Agosin, M., 304, 338* Akita, Y.,273, 346* Alanso, P., 9, 7 1 * Alfert, M., 3, 4, 71* Allfrey, V. G., 20, 7 1 * Amabis, J. M., 7, 87* Ammermann, D., 9 , 7 1* Anders, G., 248, 262* Anderson, R., 50, 75* Ansley, H. R., 19,20,75* Appleman, M. H., 290,338* Aranda, L. C., 397,435* Arends, S., 70, 88* Arnheim, N., 233,260* Arnold, G., 15, 71* Ashburner, M., 5 , 6, 24, 25, 26, 35, 36, 37, 46, 50, 57, 68, 70, 71*, 72*, 258,260* Ashhurst, D. E., 276,338* Autrum, H. J., 149, 157, 159, 166, 167, 190*, 191*

B Baccetti, B., 121, 191* Bade, M. L., 316,317,338* Baerends, G. P., 350,431* Bahr, G. F., 17, 18,72*, 73* Baker, W.K.,258,260* Balbiani, E. G., 2,72* Baranska, J., 320, 347* Barigozzi, C., 7 , 7 2 * Barnard, P. B. T., 117, 193* A.I.P.-lS

Barry, C. K., 131, 136, 139, 142, 143, 144, 146, 190, 193* Bart, A., 21 1, 260* Basile, R., 51, 52, 53, 86*, 93,94* Bastock, M., 354,43 1 * Baudisch, W.,29, 59, 65,72* Bauer, H., 2 , 3 , 7 , 8 , 7 2 * , 8 0 * Baumann, F., 158, 159, 191* Baumhover, A. H., 93,95* Bayer, Sh., 314, 343* Bayramoglu-Ergene, S., 137, 139, 191* Becker, H. J., 24, 30, 32, 37, 42, 46, 49,54,72*, 73* Bedini, C., 121,191* Beenakkers, A. M. Th., 313, 314, 315, 316,319,320,321,338* Beermann, W.,2, 3, 4, 10, 11, 12, 14, 15, 16, 17, 18, 19, 29, 31, 33, 41, 47, 64, 65, 70,72*, 73*, 74*, 77*, 78* Bencosme, S. A., 17,81* Benjamin, W.,20,73* Benolken,R.M., 157, 158, 191* Bentley, D. R., 361, 365, 367, 369, 370, 372, 385, 411,412,415,416, 417,431*, 441,444* Berendes, H. D., 2, 11, 14, 17, 18, 19, 25, 30, 31, 35, 36, 37, 38, 40, 44, 50, 54, 60, 66, 70, 73*, 74*, 75* Berger, C. A., 4,74* Bergman, S. J., 403,436* Bernstein, M., 123, 191* Berry, S. J., 17, 18, 74*, 247, 257, 262* Bertani, G., 237, 238, 239, 253,262* Bhakthan, N. M. G., 322, 341*

44 5

446

AUTHOR INDEX

Bhat, J. V., 316, 345* Bier, K., 8, 23, 54, 55, 56, 58, 74*, 75*, 88* Birnsteil, M. L., 94, 95* Birt, L. M., 276, 279, 313, 324, 338*, 341* Bischai, F. R., 304, 346* Bishop, L. G., 350, 375,431* Black, M. M., 1 9 , 2 0 , 7 5 * Blest, A . D., 354, 377,431* Block, W.,7 , 7 5 * Bode, C., 3 7 , 8 1*, 3 14,339* Bodenstein, D., 8, 55, 68, 75*, 82*, 211, 234, 236, 243, 260*, 318, 339* Bohn, H., 209, 210, 211, 212, 213, 214,261* Boistel, J., 364, 365, 369,431*, 432* Bonner, J., 20, 75 * Borsellino, A., 159, 191* Bovard, J. F., 221,264* Bowers, W.S., 295, 339* Boycott, B. B., 360, 432* Boyd, H., 59, 63, 69,70, 75* Boyd, J. B., 35, 59, 63, 69, 70, 74*, 75 * Bozler, E., 132, 191* Brachet, J., 22, 78* Breuer, M. E., 4 , 9 , 12,22,28, 30, 31, 32,75*, 78*, 86* Bridges, C. B., 10, 75* Brendsted, H. V., 22 1 , 2 6 1* Brookes, V. J., 59, 79* Brosemer, R. W.,304, 326, 339* Bucher, Th., 297, 304, 313, 323, 324, 327, 334, 339*, 342*, 343*, 346* Buck, D., 252,262* Buder, E., 4, 92* Bueding, E., 283, 339* Buetti, E., 32, 75* Bullock, T. H., 123, 133, 140, 191*, 350,356,425,43 1* Bultmann, H., 40, 57,77*, 94, 94* Bunning, E., 151, 191* Burdette, W.J., 50, 75* Burrows, M., 442,444* Bursell, E., 127, 191*, 271, 311,326, 327.339*

Burtt, E. T., 152, 159, 173, 175, 182, 191*, 375,431* C Cabib, E., 301,339* Caesar, C. J., 131, 191* Caldwell, R. L., 172, 192* Callan, H. G., 11, 55,75* Callec, J. J., 364, 365, 432* Candy, D. J., 301,321,339*, 343* Cannon, G. B., 38,75* Carafoli, E., 273, 333, 334, 339* Case, J. F., 402, 433* Cassagnau, P., 9, 75*, 76* Cassier, P., 133, 134, 136, 141, 191* Catton, W. T., 152, 159, 173, 175, 182, 191*, 375,431* Cave, M. D., 19, 76* Cerkasovova, A., 3 13, 342* Chadwick, L. E., 269,339* Chan, S. K., 323,339* Chance, B., 323, 325, 326, 334, 336, 339* Changeux, J. P., 302,343* Chaplain, R. A., 273, 339*, 340* Chappell, J. B., 327, 329, 332, 333, 342 * Chapple, W. D., 183, 191*, 380,432* Chefurka, W.,304, 310, 340* Chen, P. S., 237,239,262* Chen, S. H., 141, 191* Child, C. M., 261* Childress, C. C., 271, 285, 286, 287, 288, 289, 290, 291, 292,233, 294, 295, 311, 313, 315, 316,323,324, 325,326,327,331,340*, 344* Childress, D., 8, 76* Chino, H., 318, 319, 320, 321, 322, 340*, 341* Clark, P. J., 225, 261* Claxton, J. H., 225, 228,261* Clayton, R. B., 317,338* Clegg, J. S., 270, 294, 297, 300, 301, 302,340* Clements, A. N., 3 0 1 , 3 17, 340* Clever, U., 2, 14, 15, 18, 19, 20, 25, 27, 31, 32, 33, 34, 35, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 49, 57,

447

AUTHOR INDEX

Clever, U.-cont. 61, 62, 63, 64, 66, 67, 76*, 77*, 89*, 94,94* Cloudsley-Thompson, J. L., 151, 192* Cochran, D. G., 323, 324, 343*, 344* Coggeshall, R. E., 357,432* Cohen, L. H., 2 0 , 7 0 , 7 7 * Cohen, M. H., 220,261* Cook, B. J., 320,340* Cohen, M. J., 356, 357, 393, 432*, 442,444 Collett, T. S., 377, 431* Coon, H. G., 257,261* Corlette, S. L., 22, 89* Cornwell, P. B., 131, 141, 192* Crouse, H. V., 22, 23, 30, 38,46, 77* Curtis, D. J., 110, 121, 192* D Dackerman, M. E., 61, 86* da Cunha, A. B., 22, 23, 51, 52, 77*, 86*, 94,94* Daneholt, B., 10, 11, 16,77*, 78*, 93, 94*, 95* Darrow, J. M., 40,57, 77* Davies, W. J., 413,432* Davis, R. A., 269, 340* Dawid, I. B., 11,92* de Bruyn, W. C., 17, 18,74* de Duve, C., 63,77* Delcomyn, F., 403,405,432* Demoll, R., 132, 192* de Reuck, A. V. S., 20,77* Desai, L., 21,77* Dessen, E. M., 70,87* DeWaide, J. H., 315, 338* Diaz, M., 5 1 , 7 8 * Dick, A., 304, 306, 345* Dietz, W., 17, 18,74* Dingle, H., 172, 192*, 375,432* Doane, W., 59,78* Dodt, E., 171, 195* Domroese, K . A., 313, 314, 322, 340*, 341* Dowling, J. E., 360,432* Doyle, D., 15, 66,78* Dreyfus, A., 4, 22,78* Dufay, C., 133, 134, 141, 192*

Dugard, J. J., 411, 432* Dunbar, R. W., 7 , 8 9 * E Eddington, L. C., 320, 340* Edstrom, J.-E., 10, 11, 15, 16, 77*, 78*, 93,94*, 95* Edwards, G. A , , 108, 114, 124, 127, 194*, 270,340* Eguchi, E., 157, 158, 159, 194* Egyhizi, E., 93, 94, 95* Eisenstein, E. M., 358, 393,432* Ellgaard, E. G., 20, 50, 78* Elmer, N., 413,432*, 439,444* Emmerich, H., 4 1,78* Estabrook, R. W., 304. 310, 323,332, 336,340* Evans, D. R., 270,294,297,300,301, 302,340* Evans, F. C., 225,261* Evans, H. E., 350, 432*, 433* Evarts, E. V., 432,433* Ewing, A., 352, 372, 403, 412, 414, 425,433*

F Farley, R. D., 402,433* Feder, N., 258,259,262* Federoff, N., 48,78* Fernindez-Morin, H., 114, 192* Ficq, A., 22,18* Fielden, A., 359, 433* Filatova, I. T., 27,48, 82* Finlayson, L. H., 53, 79* Fischer, F. M., 53,79* Fisher, E. H., 290, 338* Foustka, M., 3 13, 342* Fox, S. S., 375,432* Fraenkel, G., 32, 59, 79*, 269, 314, 340*, 341* Franca, Z. M., 52,77* Friedman, S., 295, 302, 303, 314, 339*, 340*, 341* Frisch, K., von, 350,433* Fritz, I. B., 314,341* Frizzi, G., 7 , 7 9 * Fujita, S., 14, 79*

448

AUTHOR INDEX

Fuortes, M. C. F., 159, 19 1 * Fuortes, M.G. F., 157, 158, 192*

G Gabrusewycz-Garcia, N., 12, 22, 23, 28, 30, 79*, 93, 95* Galbraith, M. N., 93, 95* Gall, J. G., 23, 79*, 94, 95* Gallera, J., 237, 238, 239, 253, 262* Galloway, N. S., 373,433* Gallwitz, U.,191 * Garcia-Bellido, A., 54, 79*, 231, 251, 252,253,255,261* Garrido, M. C., 23, 52,77*, 94,94* Gaudecker, B. von, 14,79* Gawlik, S., 63, 69, 87* Gay, H., 15, 18.60.79* Gehkng, W:, 54, 80*, 244, 255, 256, 261* Geiger, H. R.,68,79* Geitler, L., 4, 79* Gellhorn, A., 20,73* George, J. C., 322, 341* Gershey, E. C., 20, 7 1 * Gilbert, E. F., 45, 79* Gilbert, L. I.,294, 312, 313, 314, 317, 318, 319, 320, 321, 322, 338*, 340*, 341*, 346*, 347* Gilby, A. R., 298, 299, 300, 312, 341* Glaser, L., 290, 343* Gloor, H., 236,237,257,262* Goidl, J., 45, 79* Goldin, H. H., 317, 341* Goldsmith, M., 66,83* Goldsmith, T. H., 153, 170, 192*, 195* Golikowa, M. N., 9,79* Gonqalves, A. M. A., 52, 77* Goodman, L. J., 104, 105, 106, 107, 108, 110, 112, 113, 114, 117, 118, 121, 123, 124, 127, 129, 130, 135, 137, 140, 141, 149, 150, 160, 165, 167, 174, 175, 176, 177, 178, 182, 184, 190, 192* Goodman, R. M., 20, 45, 62, 73*, 79*, 80*, 91*

Goodwin, B., 220,261* Gorovsky, M. A., 20,80* Gotchel, B. V., 20,70, 77* Gottschalk, A., 70, 80* Gotze, G., 101, 103, 131, 132, 141, 192* Goustard, M., 135, 136, 139, 140, 147, 192* Grasso, A., 308, 341* Green, M. M., 10,88* Greenberg, J. R., 16, 80* Greenwood, H., 93, 95* Gregg, C. T., 324, 341 * Gregory, D. W.,276, 279, 313, 341* Grell, E. H., 10,84* Griffen, A. B., 7,86* Grobstein, C., 257, 261* Gross, J. D., 12,80* Grossbach, U., 29, 64, 70, 80*, 94, 95 * Grossman, I. W.,285,286, 340* Gussin, A. E. S., 271, 298, 299, 300, 341* Guthrie, D. M., 358, 359, 411, 433*

H Hadorn, E., 37, 54, 80*, 89*, 231, 235, 236, 237, 238, 239, 246, 248, 249, 252, 253, 254, 255, 257, 261*, 262*, 264*, 265*, 266* Hagele, K., 23, 81 * Hagiwara, S., 362,417,433* Himori, J., 161, 192* Hannah, A., 200,231,262*, 264* Hansen, K., 299,300,341* Hanser, G., 37, 81 * Hansford, R. G., 279, 323, 324, 326, 327, 328, 329, 330, 331,332,333, 335,336,341*, 342* Harashima, K., 32 1, 340* Harker, J. E., 15 1, 192* Hams, H., 17,80* Harvey, W.R., 323,342* Haskell, J. A., 323, 342* Haskell, P. T., 350,433* Hasselberger, F. X.,308, 343* Hassenstein, B., 350, 433*

449

AUTHOR INDEX

Hay, E. D., 22 1,257,262* Hayward, P., 228, 230, 263* Heisler, C. R., 324, 341* Heitz, E., 2, 7,80* Helmreich, E., 290, 343* Henderson, P. A., 62,64, 66, 89* Henderson, P. T., 315, 316,338* Henderson, S. A., 5, 6, 23, 72*, 80* Hertweck, H., 110, 192* Heslop, J. P., 310, 342* Hess, A., 356, 359,433* Hess, R., 296,342* Hesse, R., 108, 131, 132, 192* Hill, A. V., 28 1,342* Hines, W.J. W.,301, 342* Hitelman, A., 52, 77* Hoffman, E., 191* Hofmanova, O., 313,342* Holst, E. von, 426,433* Holt, Th. K. H., 44, 47, 50, 54, 74*, 83* Holtzer, H., 259, 262* Holtzer, S., 259,262* Homann, H., 131, 132, 192* Horie, Y., 316, 317, 342* Horn, D. H. S., 93,95* Horn, E. C., 19,80* Horn, G., 377, 391,433*, 437* Horridge, A., 356,431* Horridge, G. A., 117, 121, 123, 133, 140, 159, 161, 171, 172, 179, 191*, 192*, 193*, 195*, 360,392, 398,433* Howard, A., 21,87* Howard, H. M., 442,444* Hoyle, G., 127, 152, 161, 166, 167, 171, 172, 173, 178, 193*, 372, 373, 374, 380, 394, 395,412,414, 423, 424, 425, 433*, 434*, 442, 444* Hsu, W.S., 59, 60, 80* Hubel, D. H., 377,434* Huber, F., 400, 414, 417, 424, 434*, 435*, 439,444* Hughes, G. M., 360, 369, 386, 388, 402,434* Hughes, P. J., 402,435* Hurlbut, E. C., 283, 292, 308, 309, 330,332,345*

I Ikeda, K., 401,438*, 441,444* Iles, J. F., 440,444* Ito, S., 44, 80* Ito, T., 316,317,342* Iwasaki, S., 376,434* J Jacklet, J. W.,356, 357,432* Jacob, F., 41,80*, 302,343* Jacob, J., 9, 17, 18, 61, 63, 80*, 81* Jander, R., 131, 136, 139, 142, 143, 144, 146, 190, 193* Jawlowski, H., 129, 193* Jewell, B. R., 272,273, 342* John, H. A., 94,95* Jones, K . W.,94,95* Jurand, A., 52, 61, 63, 80*, 8 1 * K

Kafatos, F. C., 258,259,262* Kagawa, Y., 279, 323,342* Kalmus, H., 99, 100, 193* Kalnins, V. I., 17, 8 1* Kammer, A., 409,434* Kandel, E. R., 397,434* Kanno, Y ., 4 4 , 8 1*, 84* Kaplan, 44 1,444* Kaplanis, J. N., 93, 95*, 316, 318, 343*, 344*, 346* Karlson, P., 32, 33, 34, 36, 37, 38, 40, 41, 59, 70, 77*, 8 1 * , 90*,93,95* Kato, K.-I., 60, 81 * Katsuki, Y.,375,437* Keehn, D. G., 350,375,43 1* Keilin, D., 336,342* Keith, A. D., 3 17,341 * Kendig, J. J., 364, 367,411,435* Kennedy, D., 158, 159, 193*, 360, 363, 435*, 436* Kennedy, J. S., 430,435* Kerkut, G. A., 363, 435*, 441, 444* Keyl, H.G., 11, 21, 22, 23, 52, 77*, 81*

Kidston, M. E., 23, 79* Kiknadze, I. I., 7, 14, 21, 27, 48, 49, 50,82*, 90*

450

AUTHOR INDEX

Kilby, B. A., 301, 339* Kimball, R. F., 52,53,89* Kimoto, Y., 18,82* King, D. S., 38,82* King, R. C., 8 , 5 4 , 5 5 , 8 2 * Kinnear, J. F., 94,95* Kirkham, B., 104, 105, 106, 107, 108, 110, 112, 113, 114,117,118,121, 123, 124, 127, 129, 130, 192* Kitzmiller, J. B., 7 , 8 2 * Kleinfeld, R. G., 22, 79* Kleinsmith, L. J., 20,71* Klingenberg, M., 297, 304, 313, 314, 315, 323, 324, 327, 334, 338*, 339*, 342*, 343*, 346* Kloetzel, J. A., 61, 66, 67, 82* Klug, W. S., 8, 55, 82* Knight, J., 20, 77* Knights, A., 372, 435* Kobel, H. R., 53,82* Kodani, M., 60, 82* Koehler, O., 142, 193* Koenig, P. B., 46,90* Kolbe, H. J., 132, 193* Kolliker, A., 275, 342* Koltzoff, N., 3,82* Kopac, M . J., 53, 89* Kosswig, C., 4, 82* Kraczkiewicz, 6, 82* Krasilovsky, G. H., 358,432* Krasne, F. B., 391, 435* Kravitz, E. A., 360, 437* Krebs, E. G., 290, 338* Krimbas, C. B., 7 , 8 2 * Krishnakumaran, A,, 247, 257, 262* Kroeger, H., 2, 20, 25, 27, 3 5 , 4 0 , 4 1 , 42, 43, 44, 45,46, 47,49, 51,82*, 83*, 84*, 89*, 234,243,247,262* Kubista, V., 310, 311, 313, 342*, 343 * Kunz, W., 23,55,74*, 75*, 83* Kutsch, W., 372,412,431* Kuwabara, M., 140, 172, 190, 193* L Lagwinska, E., 320, 347* Lambert, B., 93, 94*, 95* Lammert, A., 108, 193*

Lamprecht, J., 7, 83* Lardy, H. A., 308, 343* Larner, J., 296, 344* Lash, J., 259, 262* Laufer, H., 2, 14, 15, 47, 61, 62, 66, 67, 68,78*, 82*, 83*, 84*, 93,95* Lawrence, P. A., 200, 202, 203, 204, 215, 216, 217, 225,226, 227,228, 229, 230, 235, 243, 254,257,258, 262*, 263* Lea, A. O., 296,319,346* Lees, A. D., 8, 84*, 151, 193*, 243, 263* Leghissa, L., 360, 435* Lehninger, A. L., 273, 333, 334, 339* Leloir, L. F., 301, 339* Lender, T., 221,222,263* Lennie, R. W.,276, 279, 313, 341* Lesher, S., 59, 60,84* Lesseps, R. J., 25 1, 263* Leth, C. B., 8 , 5 5 , 8 2 * Lettvin, J. Y.,377,435* Lewis, A . L., 46,48, 87* Lewis, S. E., 327,330,343*, 344* Lezzi, M., 2, 4, 18, 19,27,42,43, 47, 5 1,83*, 84* Lindauer, M., 193*, 350,435* Lindsley, D. L., 10, 84* Link, E., 108, 131, 132, 193* Littau, V. C., 20,7 1 * Lobbecke, E. A., 206, 263* Locke, M., 200, 201, 203, 208, 209, 2 1 4 , 2 16,263* Lockschin, R. A., 63,84* Loewenstein, W. R., 44, 45, 80*,8 1*, 84*, 87*, 92* Loffler, 70, 8 1* Loher, W., 417,435* Loosli, R., 246,263* Louloudes, S. J., 316, 343*, 344* Lowne, B., 132, 193* Lowry, 0. H., 308,343* Lubbock, J., 132, 193* Lucchesi, J. C., 46,90* Luco, J. V., 397,435* Luke, B. M., 198,264* Lum, P. T. M., 3 17,346* Luond, H., 243,244, 263* Lutgerhorst, A., 315, 338*

45 1

AUTHOR INDEX

M Maas, A. H., 206,264* MacGregor, H. C., 23, 61, 67, 79*, 84* MacInnes, J. W., 3,84* Mackie, J. B., 61, 67,84* Mainx, F., 7, 84* Makino, S., 2 , 7 , 8 4 * Manning, A., 350, 351, 352, 354,403, 419,420,431*, 433*, 435* Marcus, W., 200, 204, 216, 217, 218, 220, 264* Margles, S., 93,95* Margoliash, E., 323, 339* Markl, H., 350,435* Marquardt, R. R., 304, 339* Marquis, N. R., 314, 341* Martin, M. D., 94, 95* Maruyama, K., 271, 272, 273, 343*, 346* Mateyko, G. M., 3, 86* Mattingly, E., 22, 23, 28,85* Maturana, H. R., 377,435* Matuszewski, B., 5, 6, 52, 82*,85* Maurer, R.,41,81* Mayer, R. J., 321, 343* Maynard, D. M., 359, 375, 376, 378, 379,435* McCann, G. D., 350,375,43 1* McCarthy, R., 294, 343* McCulloch, W. S.,377, 435* McMaster-Kaye, R., 30,85* McShan, W. H., 298, 299, 317, 347* Mechelke, F., 12, 27,29,30, 85* Midioni, J., 133, 136, 140, 146, 193* Melland, A . M . , 4 , 7 , 2 8 , 8 5 * Metschl, N., 148, 149, 153, 159, 163, 167, 176, 191*, 193* Metz, C. W.,7, 12, 21, 85*, 87* Metzger, B. E., 290, 343* Meyer, H., 314,343* Middleton, E. J., 93, 95* Midelfort, M. E., 46,85* Migliori NataLizi, G., 308, 341 * Milkman, R. D., 37, 39, 41, 48, 78*, 85 * Mill, P. J., 359, 402, 434*, 435* Miller, P. L., 401,402, 436*

Mills, R. R.,323, 343* Mimura, K., 140, 172, 190, 193* Minis, D. H., 151, 194* Minnich, D. E., 22 1,264' Mirsky, A. E., 20, 71 * Misch, D. W., 63,85* Mitchell, H. K.,68, 79*, 85* Mittelstaedt, H., 193*, 350,436* Moberly, B. J., 442,444* Monod, J., 41,80*, 302, 343* Monroe, R. E., 3 16, 343*, 344* Moore, J. W., 158, 194* Morgante, J. S., 23, 52, 77*, 94, 94* Morita, H., 140, 172, 190, 193* Mukherjee, A. S., 48, 8 5 * Miiller, E., 103, 129, 132, 141, 193* Miiller, J., 132, 193* Murphy, T. A., 296,302,303,343*

N Nagai, Y.,273, 346* Nagl, W., 9 , 8 5 * Naka, K., 157, 158, 159, 194* Nakase, Y., 14, 15, 62, 66, 83*, 84* Nakasone, S., 316, 317, 343* Narahashi, T., 158, 194*, 364, 439* Naylor, J. M., 21, 91* Neghme, A., 304,338* Neville, A. C., 198, 264*, 408, 41 1, 436* Nishiitsutsuji-Uwo, J., 15 1, 194* Nonato, E., 4, 22,78* Nothiger, R., 244, 245, 249, 250, 251,253, 264* Nur, U., 4, 85 *

0 Oberlander, H., 247,257,262* Ohnishi, K., 323 Ohtaki, T., 37, 39,85* Orell, S. A., 283, 339* Orr, C. W. R., 318, 343* Osborne, M. P., 359,436* Osinchack, J., 63,85* P Packer, L., 324, 334,345*

45 2

AUTHOR INDEX

Painter, T. S., 2 , 7 , 8 , 10, 59,86* Paka, J., 182, 194* Panitz, R., 10, 27, 35,48,49, 65, 72*, 86*, 90* Pardue, M. L., 94,95* Parker, C., 22, 23, 28,85* Parks, R. E., Jr., 308, 343* Parry,D.A., 131, 152, 161, 194* Passonneau, J. V., 308,343* Patterson, E. K., 61,86* Paul, J. S., 3 , 4 8 , 5 0 , 8 6 * Pavan, C., 4, 7, 9, 12, 21, 22, 23, 28, 30, 31, 32, 51, 52, 53, 61, 75*, 77*, 78*, 81*, 86*, 87*, 89*, 94, 94*, 95* Payne, R. S., 376,436* Pearse, A. G. E., 296,342* Pearson, K. G., 403, 436*, 440, 444* Pelc, S.,21,87* Pelling, C., 3, 11, 12, 13, 14, 16, 19, 35,44,73*, 87* Pdrez-Silva, J., 9, 71 * Perkowska, E., 60,81*, 87* Perondini, A. L. P., 7, 52,70,87*, 94, 95* Perry, M. M., 121, 195* Pette, D., 327, 343* Pettit, B. J., 19,23,87*, 88* Peyer, B., 23 1,264* Philip, U . , 7,87* Philips, D. M., 61,87* Picard, T., 94,95* Piepho, H., 200, 204, 205, 206, 220, 264* Pierce, G. W.,413,436* Pistey, W. R., 45,79* Pitman, R. M., 360, 363, 435*, 441, 444* Pittendrigh, C. S.,15 1, 194* Pitts, W.H., 377, 435* Pogo, A. O., 2 0 , 7 1* Pogo, B. G. T., 20,71* Politoff, A., 44,87* Poulson, D., 12, 21, 87* Power, M. E., 129, 152, 194* Prabhoo, N. R., 9 , 8 7 * Preiss, B., 314, 343* Preston, J. B., 363,435*, 436* Price, G . M., 310, 327, 330, 342*, 343*, 344*

Priesner, H., 99, 194* Pringle, J. W. S., 271, 272, 343*, 344*, 406,417,436* Prostakova, T. M., 48,87* Pulitzer, J. F., 14, 19, 48,50,88* R Racker, E., 279,323,342* Radzikowski, S., 9, 87* Ralph, C. L., 294, 343*, 344* Rand, H. W.,221,264* Rao, B., 62,83*, 84* Rasch, E. M., 22, 23, 46, 48, 63, 69, 87*, 88* Rasch, R. W.,19,46,85*, 87* Ray, J. W.,310, 342* Rayle, R. E., 10, 70,88* Redikorzew ,W ., 1 10,194* Reich, J., 259, 262* Reindorp, E. C., 8,86* Reingol’d, W.N., 17, 18, 71* Reiss, R. F., 409,436* Rememsberger, P., 7 , 8 3 * Remler, M., 360,436* Remmert, L. F., 324,341 * Ribbert, D., 8, 23, 55,56, 58, 59,74*, 75*, 88* Richard, G., 135, 137, 194* Richart, R. M., 45,79* Ringborg, U., 93, 94*, 95* Ristow, H. J., 11,70,88*, 92* Ritossa, F., 14, 15, 17, 1 9 , 2 1 , 4 8 , 5 0 , 51,62,88* Rizki, T. M., 60, 88* Robbins, W.E.,41,92*, 93,95*, 316, 318,343*, 344*, 346* Roberts, A., 363, 391,435*, 436* Roberts, B.,24, 56, 57,58,88* Roberts, M., 20,88*, 89* Roberts, P., 234, 264* Roberts, P. A., 52, 53, 89* Roberts, S. K., 151, 194* Rodems, A. E., 62,64,66,89* Rodman, T. C., 4 , 5 3 , 8 9 * Roeder, K. D., 272, 344*, 376, 386, 387,398,399,402,433*, 436* Romball,C. G., 14, 15, 19, 39,77* Rose, S. M., 221,222,223,224,264* Rosell-Perez, M., 296, 344*

AUTHOR INDEX

Rothenbuhler, W. C., 35 1,437* Rothfels, K . M., 7, 89* Rowe, E. C., 364, 437*, 441, 442, 444* Rowell, C. H. F., 377,399,400,437* Ruch, F., 54,80* Ruck, P., 103, 108, 110, 114, 124, 127, 148, 153, 154, 156, 160, 162, 164, 165, 166, 167, 169, 170, 192*, 194*, 195* Rudkin, G. T., 10, 14, 18, 19,21,22, 89*, 90* Ruegg, J. C., 271, 272, 273, 342*, 344* Ruhm, W.,9,91* Ruiz-Amil, M., 309, 344* Runion, H. I., 373, 374, 433*, 437*, 440,444* Ruska, H., 270, 340* Rutherford, D. J., 121, 195* S

Sacktor, B., 269, 271, 272, 273, 279, 281, 283, 285, 286,287,288,289, 290, 291, 292, 293, 294,295,296, 297, 299, 301, 304,306,307,308, 309, 310, 311, 312,313,315,316, 322, 323, 324, 325, 326, 327,330, 331, 332, 333, 334, 336, 339*, 340*, 342*, 344*, 345* Sanborn, R. C., 53,79* Sang, J. H., 8 , 4 6 , 5 5 , 8 2 * , 89* Satija, R. C., 177, 182, 195* Scaramelli, N., 304, 338* Schalike, W.,4,92* Scharrer, B., 37,89* Scheuring, L., 132, 192* Schin, K. S., 61, 63,64,67,89* Schneiderman, H. A., 247, 257, 262* Scholes, J. H., 157, 158, 161, 171, 172, 179, 193*, 195*, 392, 433* Schor, N. A., 70,90* Schremmer, F., 140, 195* Schricker, B., 137, 139, 149, 193*, 195* Schubiger, G., 244, 245,253,264* Schultz, J.,8, 21, 55,61, 86*,90* Schulz, D. W., 308, 343* Schurin, M.F., 17, 18,21,90* A.1.P.-

*

45 3

Scott, W. R., 158, 194* Sebeleva, T. E., 21,90* Sedee, D. J. W., 345* Sekeris, C. E., 41,81* Selverston, A., 360,436* Semenza, L., 7, 72* Sengun, A., 4,82*, 90* Serfling, E., 49,90* Shaaya, E., 36,59,90* Shafiq, S. A., 276, 345* Shaw, S., 171, 172, 179, 193*, 392, 433* Sherudilo, A. I., 21,90* Siddall, J. B., 38,82*, 93, 95* Simoes, L. C., 52,61,81* Sirlin, J. L., 9 , 14, 17, 18, 19, 60,70, 80*, 81*, 90* Slater, E. C., 324,326,346* Slizynski, B. M., 4,90* Smith, D. S., 273,275, 276,279,281, 283, 285, 345*, 356, 359, 361, 437* Smith, M. J. H., 301, 342* Smith, P. D., 46,90* Smith, S. D., 222,264* Smith, T. E., 159, 191* Smith, W.J., 17,48,90* Snodgrass, R. E., 24,90* Sobels, F. H,., 206,264* Socolar, S. J., 44,87* Sorsa, M., 4, 17,90* Sorsa, V.,4, 90* Spencer, W.A., 397,434* Spiegelman, S., 17, 88* Spirin, A. S., 18,90* Spiro, D., 45, 62,80*, 9 1*, 92* Springhetti, A., 7,79* Sridhara, S., 316,345* Stalker, H. D., 9,90* Staub, M.,49,54,80*, 90* Stedman, E., 20,90* Steele, J. E., 294, 345* Steffenson, D. M.,4,90* Stern, C., 200, 231, 232, 233, 264*, 265* Stevens,B. J., 17, 18,90*,91* Stevenson, E., 271, 294, 298, 313, 314,315,316,345*, 346* Stich,H. F., 17, 21,81*, 91* Storbeck, I., 15,77*

454

AUTHOR INDEX

Stretton, A. 0. W., 360,437* Strong, F. E., 3 16, 346* Stumpf, H., 204,219,220,265* Sudo, A., 321,340* Suga, N., 375,437* Sutton, E., 7, 10,91* Swift, H., 2, 4, 17, 18, 19, 20, 22, 48, 60, 61, 87*, 88*, 91* T Takemoto, K., 14, 79* Tanzer, E., 2,91* Tardent, P., 224, 265* Tateda, H.,140, 172, 190, 193* Taylor, J. H., 12,91* Tencer, R., 21,77* Terner, J. Y.,62,80*, 91* Thomas, K. K., 321,346* Thompson, M. J., 93,95* Thomson, J. A., 8,91*, 93,94,95* Thorpe, W. H., 350, 388, 437* Thorson, J., 350,437* Tietz, A., 312, 317, 318, 319, 320, 321, 346* Tinbergen, N., 350,437* Tobler, H., 25 1,255,265* Tokunaga, C., 200,232,233,265* Tozian, L., 399,436* Trager, W., 4,91* Traynor, D. R., 311, 313, 315, 316, 325, 326,340* Tregear, R. T., 272, 273, 343*, 344* Treherne, J . E., 300, 346*, 356, 359, 361,437* Trivelloni, J. C., 295,301, 346* TrujiiloCenbz, O., 121, 124, 195* TrujilloCinoz, O., 356,359, 361,437* Tsukamoto, M.,273,346* Tucker, M., 222,265* Tunstall, J., 171, 172, 179, 193* Turnstall, J., 392,433* U Uchizono, K., 129, 195* Uehawa, M.,60,93* Uretz, R. B., 3,84*

Ursprung, H., 237, 239, 240, 241, 242, 243, 244, 248, 249, 262*, 265* Usherwood, P. N. R., 373, 374,433*, 437*, 440,444* V van Breugel, F. M. A., 44, 50,53,74*, 75*, 82* Van den Bergh, S. G., 324, 326,346* Vanderberg, J., 14, 15, 62, 83* Van Handel, E., 296, 317, 319, 346* Varnadis, A., 296, 346* Veerabhadrappa, P. S., 326, 339* Vogell, W., 304,346* Vogt, M., 236,265* Vam Brocke, H.H.,272,346* von Borstel, R. C., 19,48,88* von Siebold, C. T. E., 275,346* von Zwehl, V., 157, 191* Vowles, D. M., 375,437* Vroman, H. E., 3 18,346* W Wachtler, K., 9,91* Waddington, C. H., 8,84*, 121, 195*, 235,243,263*, 265* Waldner-Stiefelmeier, R. D., 59, 91* Waldron, I., 369,410,437*, 438* Walker, R. J., 360, 363, 435*, 441, 444* Walsche, B. M., 67,91* Walters, V. A., 53,79* Walther, J. B., 171, 195* Ward, C. L., 19,28,80*,91* Watanabe, A., 362,417,433* Watanabe, M. I., 275, 322,347* Webster, G., 221, 265* Weevers, R. de G., 438* Weiant, E. A., 399,436* Weirich, G., 93,95* Weis-Fogh, T., 269, 270, 347*, 406, 408,409,410,438*, 439* Wellington, W. G., 147, 148, 195* Wells, G. P., 426,438* Welsch, U., 9,91* Wendler, G., 404,421,438* Wetzel, R., 3,92* Wharton, L. T., 28,92*

45 5

AUTHOR INDEX

White, M . J . D., 4, 5, 6, 7 , 9 2 * Whitten, J. M., 8, 9, 24, 56, 57, 58, 69,92* Wiener, J., 45,92* Wiens, A. W., 294, 347* Wiersma, C. A. G., 401,438* Wiesel, T. N., 377, 434* Wigglesworth, V. B., 199, 200, 215, 225, 226, 230, 231, 236, 257,258, 266*, 361,438* Wildermuth, H., 246, 247, 255, 266* Williams, C. M., 37, 39, 41, 63, 84*, 85*, 92*, 230, 259, 262*, 266*, 275,322,347* Willmer, E. N., 41, 92* Willows, A. 0. D., 352, 357, 438*, 442,444* Wilson, D. M., 376, 387, 404, 406, 407, 408, 410, 423, 424, 428, 434*, 438*,439* Winteringham, F. P. W.,300,347* Wlodawer, P., 320, 347* Wobels, U., 49, 90* Wobus, U., 67,92* Wolff, E., 221,266* Wolpert, L., 221,235,265*, 266* Wolsky,A., 131, 132, 133, 195* Wolstenholme, D. R., 4, 1 1,92*

Woods, P. S., 14, 21,89* Woodward, J., 20,80* Wormser-Shavit, E., 271, 283, 285, 292, 294, 297, 299,307, 308,309, 310,311,312,332,333,345* Wyatt, G. R., 271,294,296,297,298, 299, 300, 302, 303, 341*, 343*, 346*, 347* Wyatt, S. S., 298, 299, 300, 341* Wyman, J., 302,343* Wyman, R. J., 408,410,439* Y Yamasaki, T., 364,439* Yeandle, S., 157, 192* Yoshimatsu, H.,60, 61, 93* Young, B., 141,191* Young, D., 441,444* Yunis, A. A., 290, 338* Z Zawarzin, A., 35 1,439* Zebe, E.,269,271,313,314,347* Zebe, E. C., 298,299,317,347* Zimmermann, W., 206,266* Zirwer, D., 4, 92*

This Page Intentionally Left Blank

Subject Index A Acalyptratae, polytene chromosomes, 7 Acanthacris ruficornis, grooming activity, 399 Acetylcholinesterase, in ocelli, 121 Acid phosphatase Malpighian tubules, 69 and salivary gland histolysis, 64 Acricotopus lucidus, chromosome puffing during development, 27 and ecdysone, 35, 38 and gibberellin A, 48 salivary gland, 29, 30, 59, 65 Acridines, effect on chromosome puffing, 49 Actinpmycin D, effect on chromosome puffing, 49 Adaptation, ocellus, 164-170 Aeshna neural pathway map, 35 1 ocellus, 153 oxygen supply, 270 Agrion, ocellus, 108, 131 Agromyzidae, polytene chromosomes, 7,9 Albumin, serum, free diffusion, 44 Amino acids muscle phosphorylases, 290 puparium glue, 60 Arnrnophile, behaviour, 376 Anacridium aegyptiurn, ocellus, 1 37 Anal gill, polytene chromosomes, 7 Anatopynia varius, polytene chromosomes,4 Anax abdominal ganglia, 359 ocellus, 114 A . junius, 153 457

Animals other than insects amphibia, nucleolar DNA, 2 3 Arenicola, behaviour, 426 ciliates, polytene chromosomes, 9 crayfish habituation of escape response, 392 interneurons, 401 motor neurons, 363 stretch receptor, 385 frog phosphorylases, 289 visual system, 377 human, phosphorylases, 289 Hydra, morphogenetic gradients, 22 1 Lirnulus, eye, 158, 161 lobster, motor neurons, 363 mammal ATP synthesis, 279 glycogen synthetase, 296 mollusc, giant neurons, 357 monkey, displacement activity, 423 nemerteans, morphogenetic gradients, 222 Noserna, effect on juvenile hormone, 5 3 planarian, morphogenetic gradients, 221-222,222-224 protozoa infection, and chromosome puffing, 52 rabbit, phosphorylases, 289 spider, eye, 1 10 Thelophania, effect on salivary gland, 52 Tritonia gilberti, neuroethology, 352 tubularia, morphogenetic gradients, 22 1-224 vertebrates cell lines, 257

458

SUBJECT INDEX

Animals other than insects-cont. Bee vertebrates-cont. flight muscle metabolism, 269, 272, corticosteroids and chromosome 296,315 puffing, 4 5 foraging, 354 displacement activities, 423 ocellus, 132, 135, 139, 140, lampbrush chromosomes, 55 149-151, 189 Bee, bumbleAnoplura, ocellus, 99 brain, electrical activity, 375 Antenna1 units, and ocellus, 140, 172, 190 carbohydrate, and flight, 322 Antheraea, cellular differentiation, Bee, honey258 carbohydrate, and flight, 322 Aphaniptera, ocellus, 99 glycolysis, 309 Aphid Mendelian laws, behaviour, 352 fat biosynthesis, 316 ocellus, 149-151, 157, 158, 170, flight muscle metabolism, 27 1 171 photoperiodicity, 15 1 Beetle Aphiochaeta xanthina, polyteny and contractile protein, 272 endopolyploidy, 6-7 ocellus, 99 Aphrophora spumaria, ocellus, 102 Behaviour Apis, ocellus cellular mechanisms, 349-444 dark adaptation, 169 anatomy, 356-361 development, 102 courtship, 41 7 4 2 0 spectral sensitivity, 170 flight, 408-412,467 structure, 114 habituation, 389-392 Apis mellifera, ocellus locomotion, 403-408, 4 6 5 4 7 0 electrical response, 153 m e m or y and ‘‘learning”, flicker fusion frequency, 166, 167, 392-398 168 models, neural activity, 420-425 sensitivity, 165 motor neurons, physiology, Apis mellifica, ocellus 361-375 as stimulatory organ, 137, 138 neuropil, electrical activity, units, thoracic ganglion, 178 3 75-387 Aschiza, polytene chromosomes, 7 respiration, 4 0 1 4 0 3 Astarta, ocellus development, 102 song, crickets, 41 2-41 7 ATPase, myofibrillar, flight muscle, role of ocellus, 132-152 light intensity, 148-152 2 72-27 3 phototactic orientation, 141-147 polarized light, 147-148 as stimulatory organ, 133-141 Bembix, behaviour, 350 B Bibio hortulanus, Malpighian tubules, 2 Balbiani rings, 23, 9 4 Bibionidae, polytene chromosomes, 7 in different tissues, 3 1 Bilobella massoudi, polytene chromoduring development, 27, 29-30 somes, 9 and ecdysone, 35 Blaberus and juvenile hormone, 47 ocellus, 114, 171 nurse cells, 55 and RNA synthesis, 13-16 B. craniifer. 108, 153-157 and RNA transport, 17 trehalase, 299-303 and salivary gland function, 64-65, Blatella germanica, ocellus, 135, 136, 68 139, 140, 147

SUBJECT INDEX

Blattoidea, ocellus, 98, 101 Blowfly, flight muscle metabolism carnitine, 315, 316 contractile protein, 223 fat biosynthesis, 3 18 fatty acid catabolism, 31 3 glycogenolysis, 283-295 glycolysis, 307, 308, 309 mitochondria1 metabolism, 326329,332-335 organization, 279 proline oxidation, 3 12 sugar supply, 270 trehalase, 297, 298, 300 trehalose, 301 Boettcherisca peregrina, ocelli, 140, 172 Boophthora erythrocephala, polytene chromosomes, 9 Brachycera, polytene chromosomes, 7 Braconids, ocellus, 101 Bradysia mycorum, polytene chromosomes, 6 1, 94 Brain maps, 359 ocelluar units, 171-173 polytene chromosomes, 7 Bristles and hairs. develoDment cell polarity,’ 199-269, 215, 220, 22 1 pattern formation, 224-233 and imaginal disc development, 339-354 chromosome puffing, 8 , 58 Bug, contractile protein, 272

C Calcium ions flight muscle, 272-273 metabolism, mi t o chondrial 333-334,335 and permeability, 44 Calliphora erythrocephala flight muscle, metabolism mitochondria, 327-329, 333 organization, 284 oxygen supply, 270 ocellus dark adaptation, 168

459

Calliphora erythrocephala- cont. ocellus-cont. electrical activity, 153, 159, 160,162,163 eye, structure, 124 flicker fusion frequency, 166, 167 form perception, 13 1 ocellar units, VNC, 176, 18 1 phototactic orientation, 141 as stimulatory organ, 135, 137, 138 visual field, 1 31 polytene chromosomes ecdysone, 38,41 interspecific transplant, 35 nurse cells, 8, 5 5 trichogen cells, 58 Calliphora stygia, protein synthesis, 94 Calliphoridae, polytene chromosomes, 8 Calotermes, ocellus, 135, 137 Camptomyia, polytene chromosomes, 4 Carausius locomotion control, 42 1 synapses, 36 1 Carbohydrate metabolism, flight muscle, 281-312 glycogen synthetase, 295-296 glycogenolysis, 283-295 glycolysis, 303-309 other loci of control, 3 10-312 phosphorylase b kinase, 295 trehalase, 296-300 trehalose, biosynthesis, 300-303 Carnitine, and fat metabolism, 3 14-3 16 Cecidomyiidae, polytene chromosomes, 7 behaviour, 52 nurse cells, 9 polyteny and endopolyploidy, 5-6, 8 salivary gland function, 60 Cecidomyiinae, polyteny and endopolyploidy, 6 Cecropia, flight muscle metabolism fat biosynthesis, 318 fat mobilization and transport, 3 19, 322

460

SUBJECT INDEX

Cecropia-cont. glycogen synthetase, 296 glycogenolysis, 294 substrate, 271 trehalase, 298, 299 trehalose, 302 Cell polarity, post embryonic develop ment, 198-224 Cellular mechanisms of behaviour, see Behaviour Central nervous system, ocellar input, 189 Ceratopsyllus canis, ocellus, 108 Cerci, habituation, 388-389 Chironomidae, polytene chromosomes, 2,7,25-27, 28-29, 60 Chironomus, polytene chromosomes, 21.49.70 C. dohali;, 18, 27,48, 50, 61 C. melanotus, 23, 52 C. palliduvitatus cellular junctions, 45 during development, 28, 29 ecdysone, 35 RNA synthesis, 15 salivary gland function, 64-65 secretory protein synthesis, 94 C. plumosus, 48 C. tentans biochemistry, 10-11, 12, 13-16, 17, 1 9 , 9 3 , 9 4 d e v e l o p m e n t a1 physiology, 25-27,28,31 endocrine control, 33, 37, 39, 40,42,43,44 experimental modifications, 47, 49 significance, 61, 63, 64, 66, 67 C. thummi biochemistry, 15, 18, 20-21, 27 development, 32 endocrine control, 35, 42, 43, 47,93 oxytetracycline, 49 significance, 61-62, 66, 67 Chromosomes, polytene, during development, 1-93 and endopolyploidy, 4-7

Chromosomes-cont. epidermal cells, 56-59 infection, 5 1-54 nurse cells, 54-56 occurrence, 7-9 puffing biochemistry, 10-24 endocrine control, 32-47 e x pe rim en t a1 modification, 47-5 1 physiology, 24-32 significance, 59-69 Chromomere, unit hypothesis, 10-1 1 Cicada, song, neuroethology, 362-363, 416-417,424 Cicada concinna, form perception, 131 Circadian rhythms, and ocellus, 151-152, 189 Cloeon, ocellus, 102, 03,108 Cockroach neuroe t hology brain map, 359 ephaptic excital m, 369 ganglion structure, 356-359 habituation, 388 learning, 393, 397 motor neurons, 363, 364 nervous system, 375, 377, 378, 379 neuropil, 386 removal of ganglia, 398 walking, 403, 405-406, 466-468 ocellus circadian rhythms, 15 1-152 electrical activity, 153-156, 157, 159,160,163,164 flicker fusion frequency, 168 ocellar units, brain, 172 spectral sensitivity, 170 Cocoon formation Antherea, 258-259 salivary gland function, 60, 61 Coelopidae, polytene chromosomes, 7 Cold, effect on chromosome puffing, 47,50 Coleoptera, ocelli, 99 Collembola ocellus, 99, 101

SUBJECT INDEX

46 1

Collembola-cont. Dicyandiamide, and chromosome puffpolytene chromosomes, 9 ing, 48 Corpora allata D i f f e ren t ia t i o n, p o s t embryonic hormone, and fat biosynthesis, 3 18 development, 257-259 implantation, and chromosome Digestion, and salivary gland function, puffing, 47 59 Corpus cardiacum, and trehalose, 295, Diptera 303 glycolysis, 304 Corticosteroids, and chromosome ocellus, 99, 101, 127 puffing, 45 polytene chromosomes, 1-93, see Cortisone, and chromosome puffing, Chromosomes 45,62 larva, 7-8 Courtship, neural control, 4 1 7 4 18, pupa and adult, 8-9 465466 respiration, 269,271 Cricket trehalase, 299 brain, electrical activity, 375 Discs, imaginal, development, 236-257 ephaptic excitation, 369 changes in determined state, motor neurons, 361, 365,366, 372 254-25 7 phototactic orientation, 142, 144, dissociation experiments, 247 -2 54 145 eye-antenna, 244 removal of mushroom body, 400 genital, 2 3 7-243, 245-246, song, 412417,425 250-25 1, 253-254 Cryptochironomus, salivary gland, 29 leg, 244,251-253 Culicidae, polytene chromosomes, 7 proboscis, 246-247 Cuticle, development of foot pads, regulation, 237-247 56-57 wing, 252 Cycloheximide and chromosome puff- Dissociation experiments, imaginal ing, 3 4 , 3 9 , 4 0 , 4 9 , 63 disc, 247-254 Cynthia, flight muscle metabolism, Dissosteira, ocellus, 114, 117, 118 294,32 1 Dixippus, flicker fusion frequency, 166 D DNA and cocoonase synthesis, 259 Dasyneura affinis, endopolyploidy and “DNA puffs”, 2 1-24 polyteny, 6 and cortisone, 4 5 4 6 Dasyneura crataegi, polytene chromoand ecdysone, 38 somes, 6, 23 Sciaridae, 27-28 Dasyneura urticae, polytene chromoSciarinae, 11,21-24 somes, 5, 6, 52 and polytene chromosomes, 3-4, 5, Dermaptera, ocelli, 99 10-11, 14, 16, 17, 19, 20, 41, Development 51-52,53,54,58,59 polytene chromosomes, 1-93, see in vitro formation from RNA, 70 Chromosomes DNAase postembryonic, 197-266 patterns, Drosophila, 69 cell polarity, 198-224 salivary glands, 62-63 cellular differentiation, 2 5 7-2 59 determination and regulation, Dolichopodidae, polytene chromo235-25 7 somes, 9 pattern formation, 224-235 Dorylus, ocelli, 101

462

SUBJECT INDEX

Dragonfly flight metabolism contractile protein, 272, 274 fat mobilization, 322 organization, 276 oxygen supply, 270 neuroethology, 359, 372, 402, 41 I ocellus electrical activity, 153-156, 158, 159,160, 162, 163, 164 flicker fusion frequency, 166, 167, 168 spectral sensitivity, 170, 171 Drosophila development, postembryonic, 200, 206,231-234 imaginal discs, 236-257 flight muscle metabolism, 269,270, 317 neuroethology courtship, 354, 4 19-420 locomotion, 467 speed of mating, 35 1 ocellus and behaviour, 132 and circadian rhythms, 152 and photokinetic effect, 146 as stimulatory organ, 135, 136, 140 structure, 110, 121, 129 polytene chromosomes biochemistry, 12,20, 21 ecdysone, 41 endopolyploid nuclei, 7 genetic analysis, 70 nurse cells, 55 RNA hybridization, 94 significance, 59-60, 68, 69 Drosophila anannassae, chromosome puffing, 25 Drosophila busckii, chromosome puffing, 2 5 , 4 8 , 4 9 , 5 0 Drosophila hydei, chromosome puffing ecdysone, 35-37, 38,44, 93 and heat, 50 hypertrophy, 53, 54 physiology, 25, 30, 3 1 and RNA transport, 18 significance, 63, 66

Drosophila lebanonensis, chromosome puffing, 28 Drosophila melanogaster chromosome puffing biochemistry, 10, 14, 17, 21 ecdysone, 32-37, 38, 3 9 , 4 2 , 4 5 , 46 epidermal cells, 57 experimental modification, 48, 49,50 hypertrophy, 53, 5 4 juvenile hormone, 46 mutants, 8, 10, 37, 53-54, 55 nurse cells, 5 5 occurrence, 7, 8-9 physiology, 24-26, 30 significance, 60, 61, 68 structure, 2 4 imaginal disc development, 250 Drosophila siguyi, imaginal disc, 239, 241,243,250 Drosophila simulans, chromosome puffing, 25 Drosophila takahashii, chromosome puffing, 25 Drosophila victoria, chromosome puffing, 28 Drosophila virilis, chromosome puffing, 7, 17, 1 8 , 4 6 , 4 8 Drosophila willistoni, chromosome puffing, 52 Drosophilidae, polytene chromosomes, 7 E Ecdysone and chromosome puffing, 32-40, 54,56,59,93 mechanism of action, 40-46 related compounds, 38 and cuticle deposition, 229 Electrical activity, ocellus, 152-1 7 1 second order neurons, 161-1 64 sensitivity and flicker fusion frequency, 164-170 spectral sensitivity, 170-17 1 visual cells, 152-16 1 Embioptera, ocelli, 99 Empydidae, polytene chromosomes, 9

463

SUBJECT INDEX

Endomitosis, 4, 8, 55 Endoplasmic reticulum, after infection, 52 Ehdopolyploidy and polyteny, 4-7 Enzymes, salivary gland, 59,61-63, 66 Ephaptic excitation, 369 E phemeropt era, ocelli, 99 Epidermis cell polarity, 198-209,215-219 pattern formation, 231-232 polytene chromosomes, 56-59 Ephestia, wing development, 234, 243, 247 Eristalis, ocellus, 1 10 Eye, compound -antenna, imaginal disc, 244 and circadian rhythms, 15 1 dark adaptation, 168-169 electrical responses, 157, 158, 159 flicker fusion frequency, 165-168 and light intensity, 149-15 1 and ocellar units, interaction, 172-173, 177-178, 180, 182-184 and ocellus, interaction, 173, 189-190 and phototactic orientation, 141-147 and polarized light, 147-148 sensitivity, 164, 165 and speed of locomotion, 136-140 thoracic ganglia, 184-1 88 F Fat body glycogen synthesis, 296 glycogenolysis, 294-29 5 polytene chromosomes, 4 , 7 Fat metabolism, regulation, 312-322 biosynthesis, 3 16-3 19 carnitine, 3 14-3 16 fatty acid catabolism, 313-314 mobilization and transport, 3 19-322 Flicker fusion frequency, ocellus, 164-170 Flight control, 467 electrical activity, 376 motor mechanisms, 408-412, 424

Flight-conf. muscle, intermediary metabolism, see Muscle Fly flight, 272,410 intermediary metabolism fat transport, 322 fatty acid catabolism, 313 mitochondria1 metabolism, 323, 324,325-326,334 trehalose, 301,303

G 0-galactosidase, salivary gland, 62 Galleria cell polarity, 200, 204, 205, 2 16-220, 235 fat transport, 320 Ganglia general anatomy, 356-360 posterior, polytene chromosomes, 7 thoracic, and xelli, 184-188 Gastric, caecae, polytene chromosomes, 7 Genes, control activities, 45 and histones, 20 rejuvenation, 42 Genetic mosaics, 23 1-234 Genetics of behaviour, 351-352 Genital imaginal disc, 237-243, 245-246, 250-25 1, 253-254 Gerris, giant internuncial neuron, 358, 359 Gibberellins, and chromosome puffing, 48,65 Glucose-6-phosphatase, salivary gland, 62 a-glycero-P oxidation, 3 10, 332-333 Glycogen synt he tase , flight muscle, 295-296 Glycogenolysis, flight muscle, 283-295 Glycolysis, flight muscle, 303-309 Glypotendipes, chromosome puffing, 21 Gomphocerippus rufus courtship, 465-466 song, 4 14

464

SUBJECT INDEX

Gomphocercus, courtship, 41 7 4 1 8, 427 Grasshopper courtship, 41 7 4 1 8 motor mechanisms, 400 motor neurons, 372,374 song, 4 16 Gromphadorhina portentosa, walking, control, 403 Grooming activity, locust, 399 Gryllus, ocelli, 135, 136 G. bimaculatus, 142 Gryllus campestris, synaptic potentials, 365, 366 Gut, polytene chromosomes, 7, 31, 3 7 ,5 2 ,5 3

H Habituation, mechanism, 387-392 Haemocytes, and transfer of materials, 69 Hairs, and bristles, development, see Bristles Heart, polytene chromosomes, 9 Heat shock, effect on puffing, 50, 51 Helophilus, ocellus, 108, 1 10, 13 1 Hemiptera, ocellus, 99 Heterochromatin proliferation, 23-24 Histolysis, and chromosome puffing, 62-64, 68 Histones, and chromosome puffing, 19-21, 22,45 Hormones corpus cardiacum, and trehalose biosynthesis, 303 corpora allata, and fat biosynthesis, 318 ecdysone, and chromosome puffing, 3 2 4 0 and glycogen synthesis, 296 and glycogenolysis, 294-295 juvenile hormone and chromosome puffing, 4 6 4 7 , 53,54,93 and differentiation, 258 and pattern formation, 225-226 Housefly cytochrome b, 323

Housefly-cont. fat biochemistry, 317 trehalase, 298 Humeral imaginal disc, polytene chromosomes, 7 Hyaluronidase, salivary gland, 62 Hybosciara fragilis, “DNA puffs”, 23 Hymenoptera glycolysis, 304 ocellus, 99, 101, 103, 131 respiration, 269,271 Hypertrophy, polytene chromosomes, 52, 53 I

Ichneumonids, ocelli, 101 Imaginal discs, development, 236-257, see Discs Infection, modification of polytene chromosomes, 5 1-54 Integument, development, 200-209, 2 15-2 18 Ions and ecdysone, polytene chromosomes, 4 1 , 4 2 4 5 and rejuvenation, 4 2 4 5 Magnesium, flight muscle, 272 Isoptera, ocelli, 99 L Larva, Diptera, polytene chromosomes, 7-8 “Learning” and memory, 388, 392-398 Leg imaginal disc, 244,25 1-253 regeneration, 209-2 14 Lepidoptera behaviour, 354 ocellus, 99,101, 103 protocerebrum, 377 respiration, 269,271 spinning gland nuclei, 5 Leptidae, polytene chromosomes, 7 Lestodiplosis, endopolyploid polytene nuclei, 5 L. pisi, 5

465

SUBJECT INDEX

Leucophaea limb regeneration, 209-2 14 trehalase, 299 L e u cophaea madeirae, circadian rhythms, 152 Libellula, ocellus, 110, 114, 123, 124, 127 L. luctosa, 153, 170 L. vibrans, 153 Limnobiidae, polytene chromosomes, 7 Locomotion effect of ocelli, 132-141, 189 neural control, 354,465470 perambulatory, motor mechanisms, 403408 Locust intermediary metabolism, flight muscle carnitine, 3 15 contractile proteins, 272, 273 fat biosynthesis, 3 18 fat transport, 3 19, 32 1, 322 fatty acid catabolism, 313, 314 glycolysis, 304 glycogen synthetase, 296 mitochondria1 metabolism, 323, 324,334 oxygen supply, 270 oxygen utilization, 269 substrate, 271 trehalase, 298 trehalose, 301 neuroe thology ephaptic excitation, 369 flight, control, 467 flight, electrical activity, 376 flight, motor mechanisms, 408-410 habituation, 392 learning, 393-395 locomotion, 408, 425, 468470 motor neurons, 372, 373,374 neuropil, 377,380-385 optic ganglia, 375 spiracle, 4 0 1 4 02 synaptic potentials, 367 ocellus electrical response, 158

Locust-cont. ocellus-cont. flicker fusion frequency, 166 light intensity, 149 phototactic orientation, 142, 144 startle reaction, 178 as stimulatory organ, 133, 134 thoracic ganglia, 184-187 units, brain, 171 Locusta, ocellus and eye electrical response, 152, 158, 161 flicker fusion frequency, 167, 168 interaction, 182 ocellar units, VNC, 173-177, 179 “onion bodies”, 123 structure, 117, 118, 123 L. migratoria, 127, 131, 133, 135, 136, 141-144, 171 Lucilia fatty acid catabolism, 313 flight, oxygen, 269 polytene chromosomes, 8 Lucilia sericata, ocellus, 103, 139, 148,178 Lysosomes, and tissue histolysis, 63

M Magnesium ions and chromosome puffing, 43 flight muscle, 272 Malate dehydrogenase, salivary gland, 62 Malpighian tubules, polytene chromosomes and actinomycin D, 14 DNAase activity, 63 and ecdysone, 36,38 hypertrophy, 53 occurrence, 7 , 9 puffing patterns, 69 tissue specificity, 3 1-32 Mantis religiosa, ocellus, 137 Mapping brain, 359 chromosomes, 10-1 1 Mecoptera, ocelli, 99 Melanopus bivittatus, ocellus, 153

466

SUBJECT INDEX

Membranes, cellular, and ecdysone, 4 1-44 Memory and “learning”, 392-398 Mendelian laws and behaviour, 351-352 Metabolism, intermediary, flight muscle, 267-347 carbohydrate metabolism, 281-3 12 fatty acid metabolism, 3 12-322 mitochondria, 322-336 muscle properties, 269-28 1 Metriocnemus hygropetricus, chromosome puffing, 12 Microsporidia, effect on chromosome puffing, 51-52,53 Mikiola fagi, polyteny and endopolyploidy, 6 Mitochondria changes after infection, 52 metabolism Ca2+and Pi, 333-334 a-glycero-P oxidation, 332-333 metabolic effectors, 334-336 oxidative p hosphorylation, 323-325 proline oxidation, 330-332 pyruvate oxidation, 325-330 Mitomycin, and chromosome function, 14 Mitopus morio, ocellus, 1 2 1 Models, neural activity, 420-425 Monedula, ocellus, 102 Mosaics, genetic, 231-234 Mosquito, fat biosynthesis, 3 17, 3 19 Moth auditory organ, 376 cell polarity, 200 evasion response, 377 flight, metabolism, 269, 271, 272, 313,314 flight, motor mechanisms, 409 ocellus, 133, 135 Mot or mechanisms, behaviour courtship, 41 7-420 flight, 408412 general, 398-400 locomotion, 403-408 respiration, 4 0 1-403 song, crickets, 412-416

Moulting, and chromosome puffing, 68 ecdysone, 32-46 juvenile hormone, 46-47 specificity, 24-28 Mucoproteins, in puparium glue, 60 Musca ocellus, 114, 117, 118 oxygen supply, 270 Muscle flight, intermediary metabolism, 267-347 carbohydrate, 281-312 contractile proteins, 27 1-274 fatty acids, 3 12-322 mitochondria, 322-336 organization, 275-28 1 oxygen, 269-271 structural functional correlates, 28 1 substrate, 27 1 polytene chromosomes, 7 Mycetophilidae, polytene chromosomes, 7, 60

N

Nauphoeta cinerea, control of walking, 403 Nematocera, polytene chromosomes, 7, 68 Neodiprion sertifer, ocellus, 147, 148 Neonura, polytene chromosomes, 9 Neotony, ocellus and wings, 101 Nerve cells, polytene chromosomes, 7 Nerve cord, ventral, ocellar units, 173-182 Neuroethology, see Behaviour Neurons motor, 361-375 background activity, 372-375 ephaptic excitation, 369 “general” neuron, 369-372 intracellular recording, 361-364 second order, ocellus, 16 1-164 Neuropil, electrical activity anatomy, 359

467

SUBJECT INDEX

Neuropil-cont. ext racellular recording, 3 75-380 intact insects, 380-387 resting potentials, 367 Neuroptera, ocellus, 99 Noctua pronuba, ocellus, 134, 135 Nucleoli, microformation, 22-23 Sciarids, 50 Nurse cells, polytene chromosomes, 7, 8-9,54-56

P

Pachydiplax longipennis, ocellus, 153, 165,166,168,169 Panorpa communis, occelus, 131 Parasitism, and endopolyploidy and polyteny, 5 Pattern formation, post-embryonic development, 224-235 Pediculus, eye, 98 Pemphigus fraxini, ocellus, 131 Periplan e ta fat metabolism, 318,320 motor neurons, 36 1 respiration, 402 Periplaneta americana 0 motor neurons, 357 Ocellus, dorsal, 97-195 n e u r o pi1 , electrical activity, behavioural studies, 132-152 381-387 brain and VNC,171-188 ocellus distribution and structure, 99-1 32 and circadian rhythms, 151 electrical activity, 152-17 1 dark adaptation, 169 Odonata, ocellus, 99 development, 102 Oesophagus, polytene chromosomes, 7 electrical response, 153 flicker fusion frequency, 167, Oncopeltus bristles and hairs, 254 168 nerve, 127 cell polarity, 199-209 ocellar units, VNC, 178 differentiation, 25 7-25 8 sensitivity, 165, 171 pattern formation, 224-231 as stimulatory organ, 135, 137, leg motor neurons and light, 380 138 Orientation, phototactic, and ocellus, structure, 103 132-133, 141-147, 189 walking, control, 403 Orthoptera Periplaneta fuliginosa, synaptic potenocellus, 99 tials, 364 respiration, 269,271 Ovary, polytene chromosomes, 8, Perla abdominalis, oceUus, 146 Perla marginata. ocellus, 1 31 54-56 Permeability, membrane, salivary Oxidation gland, 44 a-glycero-P, 310,332-333 Perrisia ulmariae, polyteny and endoproline, 31 1-312,330-332 polyploidy, 6 pyruvate, 3 10-31 1, 325-330 Phenol oxidase, S protein, 68-69 Oxidative phosphorylation Pholus, location of synapses, 36 1 and chromosome puffing, 4 4 , 5 1 Phoridae, polytene chromosomes, 6, 7 and respiratory chain, 323-325 Phormia Oxygen flight muscle metabolism anoxia, and chromosome puffing, carnitine, 3 15 50 a-glycero-P oxidation, 3 10 flight muscle, 269-271 glycogenolysis, 289 Oxytetracycline, and chromosome puffing, 49 mitochondria, 331, 333

468

SUBJECT INDEX

Phorrnia -con t. organization, 279, 280,282 phosphorylase b kinase, 295 trehalase, 299, 300 trehalose, 300-301 ocellus, 165, 166, 167, 168 P. regina, 153 Phosphorylase b kinase, flight muscle, 295 Phototactic orientation, and ocelli, 132-133, 141-147, 189 Plecoptera, ocelli, 9 8 , 9 9 Polarized light, and ocellus, 147-148 Polarity, cell, 198-224 Polyneoptera, ocellus, 146 Polytene chromosomes, see Chromosomes Pond skater, giant internuncial neuron, 358,359 Potassium ions, and ecdysone, 41, 42-44 Praying mantis, removal of ganglia, 398 Proboscis, imaginal disc, 246-247 Prodenia, fatty acid catabolism, 31 3, 3 14 Prodiamesia olivacea, chromosome puffing, 28 Proline hydroxylation, 65 oxidation, 3 11-3 12, 330-332 Protanurini, polytene chromosomes, 9 Protease, salivary gland, 62, 64 Proteins and chromosome puffing, 18-19, 34,39,50 contractile, 271-274 S protein, phenol oxidase, 68-69 salivary gland secretion, 66-67 Protura, ocelli, 99 Proventriculus, polytene chromosomes, 7 Pseudachorutini, endopolyploidy, 9 Psocoptera, ocellus, 99 Psychodidae, polytene chromosomes, 7 Pterygota, ocelli, 101 Ptychopteridae, polytene chromosomes, 7

Puffing of chromosomes biochemistry, 10-24 ecdysone, 32-46 epidermal cells, 8 experimental modification, 47-5 1 juvenile hormone, 4 6 4 7 nurse cells, 55 physiology, 24-32 significance, 59-64 Pulvilli of pretarsus, development, 56 Putoniella marsupialis, polyteny and endopolyploidy, 6 Puparium, gluing to substrate, 59-60 Puromycin, and chromosome puffing, 39,49,62 Pyruvate oxidation, 3 1@311, 325-330 R Rectal gland, polytene chromosomes, 7 Rectum, polytene chromosomes, 7, 38 Regeneration, cell polarity, 209-2 14 Rejuvenation, chromosome puffing, 42 Respiration motor mechanisms, 401-403 and oxidative phosphorylation, 323-325 Rhabdophaga saliciperda, polyteny and endopolyploidy, 6 Rhodnius, development cell polarity, 201, 202, 203, 204, 208 differentiation, 258 pattern formation, 224-231 Rhyncosciara, “DNA puffs”, 23, 31 R hyncosciara angelae, polytene chromosomes “DNA puffs”, 12, 21, 22, 32 development, 27-28, 9 3 infection, 51, 5 2 , 5 3 nurse cells, 9 salivary gland, 30 R hyncosciara rnilleri, chromosome puffing, 2 1, 28 Ribonuclease, and chromosome puffing, 48 Ribosomes, after infection, 52

SUBJECT INDEX

Ringer-type solutions, and chromosome puffing, 49 RNA, and polytene chromosomes and cortisone, 45 -DNA hybridization, 70, 94 and ecdysone, 3 9 , 4 0 experimental modification, 47, 50 nurse cells, 55, 56 in puffs, 1 1 synthesis, and puffing, 12-17, 19, 2 1, 22-23, 6 6 , 7 0 transport, 17-18 RNAase, salivary gland, o L Roach, flight muscle metabolism fat, 318, 319 fatty acids, 3 13 glycolysis, 308 pyruvate oxidation, 3 1 1 substrate, 27 1 trehalase, 298,299, 300 S Salivary gland polytene chromosomes, 1-93, see Chromosomes Sarcophaga heterochromatin proliferation, 24 lysosome activity, 6 3 Sarcophaga aldrichi, ocellus, 147, 148 Sarcophaga bullata, polytene chromosomes, 9, 56, 94 Sawfly, ocellus, 147 Schistocerca gregaria neuroethology flight, control, 41 1 flight, motor neurons, 360 grooming, 399 learning, 394-395 locomotion, control, 466 motor neurons, 36 1,372 neuropil, 381 synaptic potentials, 368 ocellus flicker fusion frequency, 167, 168 input, and eye units, 182 light intensity, 149, 150 ocellar units, VNC, 173-177, 178,181 phototactic orientation, 141

469

Sch istocerca gregaria -con t. ocellus-cont. sensitivity, 165 as stimulatory organ, 135, 137, 138 structure, 103-13 1 thoracic ganglia, 186 Sciara, polytene chromosomes “DNA puffs”, 23 ecdysone, 9 3 genetic analysis, 70 histones, 20 Sciara coprophila, polytene chromosomes cortisone, 4 5 4 6 development, 28 “DNA puffs”, 22 ecdysone, 38 Gibberellin A, 48 lysosomes, 63 Malpighian tubules, 69 salivary gland, 30, 61, 62 Sciara ocellaris, polytene chromosomes development, 28 “DNA puffs”, 12, 2 1 infection, 5 2 , 9 3 salivary gland, 6 1 Sciaridae, chromosome puffing, 27-28 Sciarinae, polytene chromosomes, 7, 11,21-24,38 Scales, development, 200-204, 220 Scatopsidae, polytene chromosomes, 7 Scenopinidae, polytene chromosomes, 7 Silk moth larval-pupal moult, 247 trehalase, 298, 300 Silk worm, fat biosynthesis, 3 17 Simuliidae, polytene chromosomes, 7, 9,52,60 Simulium niditifrons, salivary gland, 61 Smittia parthenogenetica lysosome activity, 63 salivary gland, 6 1 Sodium ions, and ecdysone, 4 2 4 3 Song, crickets, motor mechanisms, 412-417,425

470

SUBJECT INDEX

Spectral sensitivity, ocellus, 170-17 1 Sphingids, ocellus, 101 Spinning gland, nuclei, 5 Spiracle, control, 401-402 Startle reaction, 387 Steniola, ocellus, 102 Stick insect, locomotion control, 404 Stomach, chromosome puffing, 31 Strepsiptera, ocellus, 99 Sugars, in puparium glue, 60 Syconastes marginatus, ocellus, 108 Sympetrum flight muscle, 274, 277, 278 ocellus, 108, 127 S. rubicundulum, 153, 154 Synapses, location, 360-361 Synaptic potentials, motor neurons, 364-369 Syrphus, ocellus, 1 10 T Tabanids, ocellus, 101 Tachinidae, polytene chromosomes, 9 Tachycines, flicker fusion frequency, 166 Tachysphex pectinipes, ocellus, 102 Tachytes europaea, ocellus, 102 Thioacetamide, and chromosome puffing, 48 Thysanura, ocellus, 99 Thysanoptera, ocellus, 99, 100 Tipulidae, endopolyploidy, 7 Tormogen cells, polytene chromosomes, 8 , 5 8 Trachea cell polarity, 208 polytene chromosomes, 7 Transport organ, salivary gland as, 66-67 Trehalase flight muscle, 296-300 salivary gland, 62 Trehalose, flight muscle, 296-300 biosynthesis, 300-303 Triatoma, glycolysis, 304

Trichocladius vitripennis, salivary gland, 29 Trichogen cells, polytene chromosomes, 8 , 5 8 Trichoptera ocellus, 99 spinning gland nuclei, 5 Trichosia, chromosome puffing, 52 Trypetidae, polytene chromosomes, 7 Trypsin, effect on chromosome histones, 20-2 1 Tryptophan, and chromosome puffing, 48 Tsetse fly, flight muscle metabolism, 271,311 Tube building, Chironomidae, 60 V Vanessa urticae, leg development, 234 Ventral nerve cord, ocellar units, 173-182, 190 Vesiculae seminales, polytene chromosomes, 9 Vespa, oxygen supply, 270 Vespa crabro, ocellus, 103 Virus, effect on chromosome puffing, 51 Visual cells, ocelli, electrical response, 152-161, 189

W Wasp digger, be haviour, 35 0 oxygen supply, 270 Waxmoth, mitochondria1 metabolism, 336 Wing imaginal disc, 252 and ocelli, correlation, 99-1 01

2 Zavrelia, salivary gland, 29 Zygaena, ocellus, 103, 108, 131

47 1

Cumulative List of Authors Numbers in bold face indicate the volume number of the series Aidley, D. J., 4, 1 Andersen, Sven Olav, 2, 1 Asahina, E., 6, 1 Ashburner, Michael, 7, 1 Beament, J. W. L., 2 . 6 7 Boistel, J., 5, 1 Burkhardt, Dietrich, 2, 1 3 1 Bursell, E., 4, 33 Burtt, E. T., 3, 1 Carlson, A. D., 6, 5 1 Catton, W. T., 3, 1 @hen, P. S., 3, 5 3 Colhoun, E. H., 1, 1 Cottrell, C. B., 2, 175 Dadd, R. H., 1 , 4 7 Davey, K. G., 2, 219 Edwards, John S., 6, 9 7 Gilbert, Lawrence I., 4, 6 9 Goodman, Lesley, 7 , 9 7 Harmsen, Rudolf, 6, 139 Harvey, W. R., 3 , 133 Haskell, J. A., 3 , 133 Hinton, H. E., 5, 6 5 Hoyle, Graham, 7 , 3 4 9 Kilby, B. A., 1 , 11 1 Lawrence, Peter A., 7, 197 Lees, A. D., 3 , 2 0 7 Miller, P. L., 3, 279 Narahashi, Toshio, 1 , 175 Neville, A. C., 4, 213 Pringle, J. W. S., 5, 1 6 3 Rudall, K. M., 1 , 2 5 7 Sacktor, Bertram, 7 , 2 6 8 Shaw, J., 1 , 3 1 5 Smith, D. S., 1 , 401 Stobbart, R. H., 1 , 3 15 Treherne, J. E., 1 , 4 0 1 Usherwood, P. N. R., 6, 205 Waldbauer, G. P., 5, 229 Weis-Fogh, Torkel, 2, 1 Wigglesworth, V. B., 2, 247 Wilson, Donald M., 5, 289 Wyatt, G. R., 4, 287 Ziegler, Irmgard, 6, 139

This Page Intentionally Left Blank

473

Cumulative List of Chapter Titles Numbers in bold face indicate the volume number of the series

Active Transport and Passive Movement of Water in Insects, 2 , 6 7 Amino Acid and Protein Metabolism in Insect Development, 3, 53 Biochemistry of Sugars and Polysaccharides in Insects, 4, 287 Biochemistry of the Insect Fat Body, 1, 11 1 Biology of Pteridines in Insects, 6, 139 Cellular Mechanisms Underlying Behaviour-Neuroethology, 7,349 Chitin Orientation in Cuticle and its Control, 4, 2 13 Chitin/Protein Complexes of Insect Cuticles, 1, 257 Colour Discrimination in Insects, 2,131 Comparative Physiology of the Flight Motor, 5 , 163 Consumption and Utilization of Food by Insects, 5, 229 Control of Polymorphism in Aphids, 3, 207 Control of Visceral Muscles in Insects, 2, 21 9 Electrochemistry of Insect Muscle, 6, 205 Excitation of Insect Skeletal Muscles, 4, 1 Excretion of Nitrogen in Insects, 4, 33 Feeding Behaviour and Nutrition in Grasshoppers and Locusts, 1,47 Frost Resistance in Insects, 6, 1 Function and Structure of Polytene Chromosomes During Insect Development, 791 Functional Aspects of the Organization of the Insect Nervous System, 1,401 Hormonal Regulation of Growth and Reproduction in Insects, 2, 247 Image Formation and Sensory Transmission in the Compound Eye, 3, 1 Insect Ecdysis with Particular Emphasis on Cuticular Hardening and Darkening, 2,175 Lipid Metabolism and Function in Insects, 4 , 6 9 Metabolic Control Mechanisms in Insects, 3, 133 Nervous Control of Insect Flight and Related Behaviour, 5,289 Neural Control of Firefly Luminescence, 6, 5 1 Osmotic and Ionic Regulation in Insects, 1 , 3 15 Physiological Significance of Acetylcholine in Insects and Observations upon other Pharmacologically Active Substances, 1, 1 Polarity and Patterns in the Postembryonic Development of Insects, 7, 197 Postembryonic Development and Regeneration of the Insect Nervous System, 6,97 Properties of Insect Axons, 1, 1 75 Regulation of Breathing in Insects, 3,279 Regulation of Intermediary Metabolism, with Special Reference to the Control Mechanisms in Insect Flight Muscle, 7, 268 Resilin. A Rubberlike Protein in Arthropod Cuticle, 2, 1 Spiracular Gills, 5,65 Structure and Function of the Insect Dorsal Ocellus, 7 , 9 7 Synaptic Transmission and Related Phenomena in Insects, 5, 1

This Page Intentionally Left Blank