Advances in Insect Physiology, Volume 08

Advances in

INSECT PHYSIOLOGY

VOLUME 8

This Page Intentionally Left Blank

Advances in

Insect Physiology Edited b y J. W. L . BEAMENT, J. E. TREHERNE and V. B. WIGGLESWORTH

Department of Zoology, The University; Cam bridge, England

VOLUME 8

1971 ACADEMIC PRESS London and New York

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

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

Copyright 0 1971 by Academic Press Inc. (London) Ltd

All rights reserved NO PART O F THIS BOOK MAY BE REPRODUCED IN ANY FORM BY PHOTOSTAT, MICROFILM, OR ANY OTHER MEANS, WITHOUT WRITTEN PERMISSION FROM THE PUBLISHERS

Library of Congress Catalog Card Number: 63-14039 ISBN: 0-12-024208-7

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

List of Contributors toVolume 8 D. DAGAN, Department o f Zoology, Hebrew University, Jerusalem, Israel (D. 9 5 ) C. H. FRASER ROWELL, Department of Zoology, University of California at Berkeley, Berkeley, California (p. 145) S . H. P. MADDRELL, Agricultural Research Unit o f Invertebrate Chemistry and Physiology, Department of Zoology, University o f Cam bridge, England (P. 199) TOSHIO NARAHASHI, Department of Physiology and Pharmacology, Duke University Medical Center, Durham, North Carolina, U.S.A. (p. 1 ) I. PARNAS, Department of Zoology, Hebrew University, Jerusalem, Israel (P. 9 5 )

V

This Page Intentionally Left Blank

Contents LIST O F CONTRIBUTORS TO VOLUME 8 .

. . . . . . . . . .

v

EFFECTS O F INSECTICIDES ON EXCITABLE TISSUES TOSHIO NARAHASHI Introduction . . . . . . . . . . . . . . . . . Process of Insecticidal Action . . . . . . . . . . . . Mechanism of Nerve Excitation . . . . . . . . . . . A. Excitation and Conduction in Nerve Fibers . . . . . B. Synaptic and Neuromuscular Transmission . . . . . IV . Functional Changes Caused by Insecticides in Nerve and Muscle . A . DDT . . . . . . . . . . . . . . . . . . Lindane . . . . . . . . . . . . . . . . . . B. C. Cyclodienes . . . . . . . . . . . . . . . . D. Pyrethroids . . . . . . . . . . . . . . . . E. Rotenone . . . . . . . . . . . . . . . . F. Organophosphates . . . . . . . . . . . . . V . Mechanisms of Functional Changes Caused by Insecticides in Nerve and Muscle . . . . . . . . . . . . . . . . A . DDT . . . . . . . . . . . . . . . . . . Allethrin . . . . . . . . . . . . . . . . . B. VI . Temperature Coefficient of Insecticidal Action . . . . . . A . DDT . . . . . . . . . . . . . . . . . . B. Pyrethroids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Insecticide Resistance . . . . . . . . . A . Nerve Sensitivity t o Insecticides B. GenesControlling theNerve Sensitivity . . . . . . VIII . Structure-activity Relation . . . . . . . . . . . . . A . DDT . . . . . . . . . . . . . . . . . . B. Pyrethroids . . . . . . . . . . . . . . . . C. Rotenone . . . . . . . . . . . . . . . . IX . Road t o the Molecular Mechanisms . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . I. I1. I11.

1 3 5 5 17 21 21 23 24 26 27 27 31 31 45

56 56 61 65 66 70 72 73 75 76 78 80 80

FUNCTIONAL ORGANIZATIONS O F GIANT AXONS IN THE CENTRAL NERVOUS SYSTEMS O F INSECTS: NEW ASPECTS I . PARNAS and D . DAGAN I.

Introduction . . . . . . . . . . A . Definition . . . . . . . . . B. Occurrence and Examples of Function C . Giant Fibre System of the Cockroach vii

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

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

96 96 96 97

viii

CONTENTS

Histological Observations . . . . . . . . . . . . . A. Abdominal Connectives . . . . . . . . . . . . B. Sheaths . . . . . . . . . . . . . . . . . C. Abdominal Ganglia . . . . . . . . . . . . . D. Giant Fibres in Thoracic Ganglia . . . . . . . . . E. Thoracic Connectives . . . . . . . . . . . . F . Degeneration . . . . . . . . . . . . . . . G . Giant Fibre Somata . . . . . . . . . . . . . I11. Membrane Properties . . . . . . . . . . . . . . IV . Through Conduction-"Continuity vs. Contiguity" . . . . . A. Collision Experiments . . . . . . . . . . . . B. Low Safety Factor Zones . . . . . . . . . . . . C. Continuity of Giant Axons in Mole, Cricket and Locust V . Do Giant Fibres Activate LegMotoneurones? . . . . . . VI . Afferent Inputs . . . . . . . . . . . . . . . . A. Cercal Inputs . . . . . . . . . . . . . . . B. Inputsat Abdominaland 1'horacicGanglia . . . . . VII . Giant Fibre Outputs . . . . . . . . . . . . . . . A. Output to Antenna1 Motoneurones . . . . . . . . B. Efferent Activity of Giant Axons in the Metathoracic Ganglion of the Cockroach . . . . . . . . . . . VIII . Giant Axon and Small Fibre Pathway-Timing Relations. . . . IX . Possible Functions of Axons in Integration . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . I1.

100 100 101 102 104 104 106 108 110 110 110 114 121 121 128 129 129 130 130 132 135 136 139 140

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS C. H . FRASER ROWELL I. I1. I11. IV .

V.

VI .

Definitions. Terminology. and Taxonomy . . . . . . . . Introduction-Variable Coloration and the Natural History of Grasshoppers . . . . . . . . . . . . . . . . . . Genetic Factors . . . . . . . . . . . . . . . . . A. Genetic Polymorphism . . . . . . . . . . . . B. Genetic Modification of Phenotypic Polymorphism . . Environmental Factors . . . . . . . . . . . . . . A . The Homochrome Responses to Background: The Orange and Black Pigment Systems . . . . . . . . . . . B. The Green/Brown Polymorphism . . . . . . . . C. Phase Coloration . . . . . . . . . . . . . . Physiological Mechanisms . . . . . . . . . . . . . A. The Green/Brown Polymorphism and the Corpus Allatum . B. The Black Pigment System and the Corpus Cardiacum . . C. Other Endocrine Correlates of Pigmentation . . . . . Pigments . . . . . . . . . . . . . . . . . . . A . The Green Component of the Green/Brown Polymorphism

146 147 152 152 155 156 157 167 175 177 178 186 181 183 184

CONTENTS

ix

B.

The Brown Component of the Green/Brown Polymorphism. and the Black and Orange Pigment Systems . . . . . . Implications of the above for the Green/Brown PolyC. morphism . . . . . . . . . . . . . . . . . D. Pattern . . . . . . . . . . . . . . . . . E . The Phase Coloration of Gregarious Locust Hoppers . . . . . . . . . . . . . . . . . . . . . Acknowledgements References . . . . . . . . . . . . . . . . . . . . .

186

188 189 189 190 190

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

.

S. H . P MADDRELL

I. I1. I11.

Introduction . . . . . . . . . . . . . . . . . Deposit and Storage Excretion . . . . . . . . . . . Less Common Excretory Organs . . . . . . . . . . . A . The Excretory Role of the Pericardial Cells and Nephrocytes B. The Midgut of Larvae of Saturniid Silkmoths . . . . . C. The LabialGlandsof Saturniid Silkmoths . . . . . . D. The Anal Papillae of Mosquito Larvae . . . . . . . IV . The Malpighian Tubules . . . . . . . . . . . . . . A. The Malpighian Tubules of Curuusius . . . . . . . . B. The Malpighian Tubules of Calliphora . . . . . . . C. The Malpighian Tubules of Tipulu . . . . . . . . D. The Malpighian Tubules of Rhodnius . . . . . . . E . The Malpighian Tubules of Culpodes . . . . . . . F. The Ultrastructure of Malpighian Tubules and Its Functional Significance . . . . . . . . . . . . . . G . Formed Bodies . . . . . . . . . . . . . . H. The Handling by Malpighian Tubules of Organic Solutes . V. The Hindgut . . . . . . . . . . . . . . . . . A . The Action of the Hindgut Anterior to the Rectum . . B. The Action of the Rectum . . . . . . . . . . . C. Rectal Absorption of Ions and Water in Schistocercu . . D. Rectal Absorption of Ions and Water in Calliphora . . . E . The Mechanism of Water Absorption by the Rectum . . F. The Mechanism of Ion Absorption by the Rectum . . . G . Rectal Recovery of Amino Acids, Sugars and Other Small Organic Molecules . . . . . . . . . . . . . . H . The Role of the Cuticular Lining of the Rectum . . . . I. Absorption of Water Vapour from Subsaturated Atmospheres by Thermobia . . . . . . . . . . . J. Absorption of Water Vapour from Subsaturated Atmospheres by Tenebrio . . . . . . . . . . . . VI . Concluding Remarks . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .

200 201 204 205 206 209 212 212 213 217 238 238 264 268 276 279 286 287 289 29 1 295 296 303 304 304 307

310 319 324 324

X

CONTENTS

Author Index . . . . . . . . . . . . . . . . . . Subject Index . . . Cumulative List of Authors Cumulative List of Chapter Titles .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

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

333 341 353 355

Effects of Insecticides on Excitable Tissues TOSHIO NARAHASHI Department of Physiology and Pharmacology Duke University Medical Center Durham. North Cizrolina. U.S.A. Introduction . . . . . . . . . . . . . . . . . . Process of Insecticidal Action . . . . . . . . . . . . 111. Mechanism of Nerve Excitation . . . . . . . . . . . . A . Excitation and Conduction in Nerve Fibers . . . . . . B. Synaptic and NeuromuscularTransmission . . . . . . IV . Functional Changes Caused by Insecticides in Nerve and Muscle . A . DDT . . . . . . . . . . . . . . . . . . . B. Lindane . . . . . . . . . . . . . . . . . . C . Cyclodienes . . . . . . . . . . . . . . . . D. Pyrethroids . . . . . . . . . . . . . . . . E. Rotenone . . . . . . . . . . . . . . . . . F. Organophosphates . . . . . . . . . . . . . . V . Mechanisms of Functional Changes Caused by Insecticides in . . . . . . . . . . . . . . . . Nerve and Muscle A. DDT . . . . . . . . . . . . . . . . . . . B . Allethrin . . . . . . . . . . . . . . . . . VI . Temperature Coefficient ot lnsectlcldal Action . . . . . . . A . DDT . . . . . . . . . . . . . . . . . . . B . Pyrethroids . . . . . . . . . . . . . . . . VII . Insecticide Resistance . . . . . . . . . . . . . . . A. Nerve Sensitivity to Insecticides . . . . . . . . . . B. Genes Controlling the Nerve Sensitivity . . . . . . . VIII . Structure-activity Relation . . . . . . . . . . . . . A . DDT . . . . . . . . . . . . . . . . . . . B . Pyrethroids . . . . . . . . . . . . . . . . Rotenone . . . . . . . . . . . . . . . . . C. IX . Road to the Molecular Mechanisms . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . .

I. I1 .

1 3 5

5

17 21 21 23 24 26 27 27 31 31 45 56 56 61 65 66 70 72 73 75 76 78 80 80

I . INTRODUCTION

Mode of action of insecticides has been studied extensively for the past two decades since the development of a variety of synthetic AIP-i

1

2

T. NARAHASHI

insecticides. One of the most remarkable achievements in this field is the study of the metabolism of insecticides which includes activation and degradation. Another contribution worthy of note is the study of the inhibition of cholinesterases (ChE’s) by a number of insecticides, most of which are either organophosphates or carbamates. However, in view of the fact that most insecticides are potent nervous poisons, it is surprising to find that a less amount of effort has been devoted to the study of the effects of various insecticides on the nervous system, especially on its excitable mechanism. As described in this articl6, it was not until the mid-1 960s that the cellular mechanism of action of certain insecticides on the nerve was satisfactorily elucidated. The action of insecticides on the nervous system may be classed into three categories: (1) functional changes in the nervous system as a result of insecticide intoxication; (2) biochemical mechanisms which are responsible for the functional changes; (3) biophysical or physico-chemical mechanisms which are responsible for the functional changes. First of all, the symptoms of poisoning caused by an insecticide must be interpreted in terms of disorders of various tissues. In most cases, the target site is the nervous system. The site of action of the insecticide in the nervous system must be determined, and changes in the nervous function must be observed. Since electric potential change or action potential is the only signal easily observable while the nerve is in the excited state, electrophysiological techniques are the most straightforward way of studying this problem. The biochemical aspects of the mechanism of insecticidal action on the nerve require some comments, because this problem is often misunderstood. As will be described later (Section I11 AS), the excitation and impulse propagation of the nerve fiber are not directly dependent upon the metabolic energy. In other words, the enzyme system is not directly involved in excitation. The only region where the enzyme system plays an immediate role in excitation is synapse and neuromuscular junction. At such junctions, the transmitter substance must be produced in the nerve terminals by enzymatic reactions, and the transmitter substance released from the nerve terminals upon excitation must be destroyed rapidly by the action of enzymes t o regulate the transmitter action on the postsynaptic element. Cholinesterase is the enzyme hydrolyzing the released acetylcholine at the cholinergic junctions. The inhibition of ChE causes severe disturbances of impulse transmission across the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

3

synapses. This is the major mechanism of action of a number of organophosphorus and carbamate insecticides. The mechanisms of ChE inhibition by these insecticides are out of the scope of the present article. Only physiological aspects of synaptic disturbances will be discussed. It is obvious from the explanation given above that the biophysical or physico-chemical aspects are of utmost importance in understanding the mechanism of action of insecticides. Electrophysiological techniques prove to be powerful in this study also. In view of these considerations, the present article covers the following aspects. First of all, the mechanism of the nerve excitation will be briefly described to help readers t a fully understand the subsequent sections. In the second place, changes in the nervous function caused by insecticides are described and discussed. This is the first step of the study of the mode of action. In the third place, the action on the nerve membrane is discussed in detail. This is the mechanism of action at the cellular or membrane level. In the fourth place, the electrophysiological techniques are applied to various problems of the mode of action of insecticides. This includes the mechanism involved in the temperature effect on insecticidal activity, the resistance of insects to insecticides, and the structure-activity relationship. It must be emphasized that the present article is not intended to cover these areas evenly and comprehensively. Most of the insecticides covered here are those in which the author has directly been involved for the past 20 years. For more comprehensive aspects of the mode of action of various insecticides, readers are urged to consult review articles (Brown, 1951, 1960, 1964; Casida, 1963; Colhoun, 1960, 1963; Dahm, 1957; Dahm and Nakatsugawa, 1968; Wilkinson, 1968; Fukuto, 1961; Hayes, 1959; Kearns, 1956; Lipke and Kearns, 1960; March, 1958; Metcalf, 1955, 1967, 1968; Perry, 1960; Roan and Hopkins, 1961 ; Smith, 1962; Spencer and O'Brien, 1957; Terriere, 1968; Winteringham and Lewis, 1959; Gordon, 1961; Hoskins and Gordon, 1956; O'Brien, 1966, 1967; Yamamoto, 1970; Winteringham, 1969). 11. PROCESS OF INSECTICIDAL ACTION

Before discussing the major problems of the present article, it will be appropriate to describe the process of insecticidal action, especially that in insects, because there are some reactions which are

4

T. NARAHASHI

not common for other drugs or other animals. This is also important to visualize the role of each reaction in the whole intoxication process. Figure 1 illustrates a schematic process of the intoxication of insect by an insecticide. The insecticide may enter the insect body through the integument, the mouth, or the stomata. It may be

I1 1 4 DETOXICATION

ACCUMULATION

4 ACTIVATION

7

4

DETOXICATION

I

4 * ACCUMULATION

Lr_'irl EXCITABLE

MEMBRANE

ENZYME

I AUTOTOXIN

DEATH

Fig. 1. Process of toxic action of contact insecticide (Narahashi, 1964a).

insecticidally active as its original form (e.g. DDT) or may have to be converted into an active form t o exert the toxic action. For example, parathion becomes effective in inhibiting ChE's after having been oxidized t o paraoxon. The insecticide may be detoxified (e.g. from DDT to DDE) and excreted, or may be stored in the adipose tissue without exerting any toxic effect. In any case, the insecticide or its activated form finally reaches the site of action, which is in many cases the nervous system. However, there are generally diffusion barriers surrounding the nerve such as the nerve sheath. After having

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

5

penetrated the nerve sheath, the process of activation or detoxication may take place. The insecticide can now exert its toxic action at the real site of action, e.g. at the nerve membrane or at the synaptic junctions. There are at least two ways by which the insecticide works there: (1) the direct physico-chemical action on the nerve membrane; (2) the action through the inhibition of enzymes. Symptoms of poisoning develop, but these do not necessarily lead the poisoned insect to death. In some cases, the hyperactivity of the nerve caused by insecticides liberates a toxin or toxins which in turn stimulate and paralyze the nerve (Sternburg, 1960, 1963; Sternburg and Kearns, 1952; Shankland and Kearns, 1959; Blum and Kearns, 1956; Hawkins and Sternburg, 1964; Sternburg et al., 1959). Death in fact results from the complicated multiple actions of the insecticide, including the exhaustion of energy, the paralysis of nerve and muscle systems, etc. Unlike vertebrate animals, death cannot be caused by a sole disturbance of the vital organ such as the paralysis of the respiratory center or the stoppage of the heart beat, because many functions in insects are not centralized. 111. MECHANISM OF NERVE EXCITATION

It would be appropriate t o briefly describe here the mechanism of nerve excitation, because without having proper knowledge on this problem it will be impossible t o fully understand the mode of action of insecticides on the nerve. This is a very specialized field so that readers in other fields may not be familiar with it. For more detailed information, readers are urged to consult specialty articles or text books (Hodgkin, 1958, 1964; Ruch et al., 1965; Katz, 1962, 1966; Eccles, 1964; Nastuk, 1966; Davson, 1964; Narahashi, 1963a). A. EXCITATION AND CONDUCTION IN NERVE FIBERS

I . Structure of the Nervous System The unit of the nervous system is called “neuron”. A neuron is composed of a nerve cell from which a number of “dendrites” and a long “nerve fiber” or “axon” emerge. Such neurons are synaptically connected with each other, or make synaptic contact with effective organs such as the skeletal muscle or the smooth muscle. At the synapse, there is a gap of a few hundred Angstroms between the presynaptic and postsynaptic membranes. At some junctions,

6

T. NARAHASHI

however, the membranes of both presynaptic and postsynaptic elements are in close contact with each other forming a tight junction. A nerve membrane, which surrounds the axoplasm, is only about 75 A in thickness. This is the site of excitation of the nerve. A number of studies based on electron microscopic observations, X-ray diffraction, etc. have now resulted in the general agreement in that the nerve membrane is composed of a double phospholipid layer sandwiched by two protein layers. However, there have been many arguments regarding the exact arrignment of these macromolecule components in the nerve membrane. Cholesterol is also contained in the membrane, and calcium ions are said to maintain the integrity of the membrane by means of their positive charges. The axoplasm usually contains a large amount of potassium and a small amount of sodium and chloride. In the external medium such as the blood serum, the concentrations of these ions are reversed. Therefore, there are concentration gradients with high potassium inside and high sodium and chloride outside. However, in some insects the concentration gradient is in the opposite direction (see review by Narahashi, 1963a). 2. Resting Membrane Potential

When a glass capillary microelectrode is inserted into a giant nerve fiber (Fig. 2), a steady potential difference is recorded with the Stimulator

Nerve

rnutw \ {Inward

A t D e p o b r i z a t i o n

V

_----_-_-

IHyperpdarizotion

Fig. 2. Diagram of two-microelectrode experiment. Lower part depicts membrane potential changes produced by square pulses of current of various magnitudes in either outward or inwatd direction across the nerve membrane.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

7

inside negative with respect to the outside. This is the resting membrane potential, and is usually of the order of -50 to -100 mV (Fig. 3). Since the concentration of potassium is higher inside than outside, and since the resting membrane is permeable to potassium but is scarcely permeable to sodium or chloride, the membrane behaves more or less like a potassium electrode and the resting mV

50r

- 501

Fig. 3. Action potential recorded by means of intracellular microelectrode from the giant axon of the cockroach. Two tracings are photographiqlly superimposed, one before and the other after inserting the electrade into the axon.

potential approaches the equilibrium potential for potassium (EK) which is given by the Nernst equation:

where R, T and F represent the gas constant, the absolute temperature, and the Faraday constant, respectively, and [ I, and [ I i are the concentrations in the outside and inside of the axon, respectively.

3. Action Potential (a) Initiation of Action Potential: When a brief electric shock is applied to a giant nerve fiber preparation via a pair of wire electrodes, an action potential can be recorded by means of a microelectrode inserted in the fiber (Fig. 3). The action potential recorded from the nerve is usually very brief in duration, lasting only about 1 ms. At the peak of the action potential, the membrane potential is reversed in polarity the inside of the axon becoming positive with respect to the outside. This overshoot amounts to 20-50 mV.

8

T. NARAHASHI

The falling phase of the action potential may simply return t o the original resting potential level in some kinds of nerve fibers, whereas in others the initial quick falling phase is followed by a slow terminal phase which gradually returns to the resting level. This slow repolarization phase is sometimes called “negative after-potential”. It should be noted that during the negative after-potential the potential actually deflects in the positive direction, and that the term “negative” comes from the classical method of external recording of the monophasic action potential whereby the depolarizing direction is recorded as a negative deflection. In some other nerve fibers, the falling phase of the action potential is followed by an undershoot or positive phase which may return t o the resting level gradually or may be in turn followed by a small negative after-potential. Again in this case, the “positive phase” is in fact a negative deflection. Instead of stimulating the nerve fiber by means of a pair of wire electrodes which are in contact with the nerve, one can insert another microelectrode into the axon very close to the recording microelectrode (Fig. 2). When a square current pulse is applied in the inward direction across the nerve membrane, the membrane is slowly hyperpolarized and attains the steady state. Upon cessation of the current pulse, the membrane potential returns slowly to the resting level. The membrane hyperpolarization is increased with increasing the intensity of the inward current pulse. When outward current pulses are applied t o the membrane, the situation is somewhat different. With a weak outward current, the membrane is slowly depolarized and the potential change is a mirror image of that produced by an inward current of the same intensity. With increasing the outward current intensity, a hump may appear during the early phase of depolarization. Upon increasing the current intensity slightly, an action potential is produced from the hump. The latency between the onset of current and the foot of the action potential is shortened as the current intensity is increased, whereas the threshold membrane potential where the action potential is produced remains constant. If one of the microelectrodes shown in Fig. 2 is withdrawn and reinserted at varying distances from the other microelectrode, and similar measurements are repeated, it can be seen that the height of the action potential remains unchanged whereas the steady-state amplitude of the hyperpolarization (anelectrotonic potential) or of the depolarization (catelectrotonic potential) declines with the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

9

distance in an exponential manner. This demonstrates that, although the action potential is propagated along the axon without decrement, the electrotonic potential passively produced by current is not propagated but simply spread. (b) Conduction of Action Potential: The diagram of impulse conduction along a nerve fiber is illustrated in Fig. 4. Since the membrane potential is reversed in polarity at the peak of the action REF

ACT

REST

+_ +_ +_ +_ +-n + - - -n -+++++++ - ++++-------

-

u

w

_ _ - - - - ++++------+ + + + + + ----+++ ++++ IMPULSE

Fig. 4. Diagram of impulse conduction in an axon. REF, refractory; ACT, active; REST, resting state (Narahashi, 1965a).

potential, a potential gradient is established between the activated area and the adjacent areas of the axon membrane. Hence, a local circuit current will flow across the membrane in such a direction as to depolarize the adjacent areas. Under normal conditions, the local circuit current is 3-5 times stronger than the threshold current necessary to produce an action potential. Therefore, an action potential is initiated from the area of the axon ahead of the activated area. No action potential will be produced from the area behind the activated area because the membrane is in a refractory state. Thus, the action potential is propagated along the axon by means of the local circuit current. (c) Ionic Mechanism: The ionic mechanism of action potential production is schematically shown in Fig. 5. As described before, the permeability of the nerve membrane to sodium is very low at resting conditions. Upon depolarizing stimulus, however, the sodium permeability (or sodium conductance, g N a ) rapidly increases so that the membrane becomes almost exclusively permeable to sodium. Therefore, the membrane potential approaches the equilibrium potential for sodium (ENa)defined by the Nernst equation for sodium

Because of the concentration gradient for sodium, sodium ions now

10

T. NARAHASHI

20

0 -20

- 30 - 20

-40

-60 -80

1

1

I

I

1

1

0

I

2

3

4

t~rrssc)

-------------RESTING STATE

K T l M STATE

Fig. 5. Diagram of the mechanism of action potential production. RP, resting potential; AP, action potential; EN,, sodium equilibrium potential; E K , potassium equilibrium potential; ma,membrane sodium conductance; gK , membrane potassium conductance. See text for further explanation (Narahashi, 1965a).

flow across the nerve membrane in inward direction. The increased sodium permeability starts decreasing soon, and the potassium permeability (or potassium conductance, gK ) starts increasing beyond its resting level. These permeability changes make the membrane almost exclusively permeable t o potassium again, thereby bringing the membrane potential back to the resting level. Potassium ions flow outwardly across the membrane according t o the concentration gradient. When the falling phase of the action potential approaches the resting potential level, the potassium permeability may still be maintained at a higher level. This enables the membrane potential to approach the E K closer than at resting conditions producing a positive phase or undershoot of the action potential. Experimental analyses have shown that the negative after-potential

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

11

in squid and cockroach giant axons is produced by the transient depolarization caused by an accumulation of the released potassium in the immediate vicinity of the nerve membrane (Frankenhaeuser and Hodgkin, 1956; Narahashi and Yamasaki, 1960b).

4. Voltage Clamp Experiment (a) Rationale and Membrane Currents: In order to interpret changes in action potential or excitability caused by experimental procedures such as applications of drugs, it is necessary to measure changes in ionic permeabilities of the membrane. This can be achieved most efficiently by means of voltage clamp techniques. Since membrane ionic permeabilities are directly related to membrane conductances to the ions in question, and since conductance is obtained by dividing current by potential, measurements ought to be made on both membrane potential and membrane current. Figure 6 depicts electrical equivalent circuits of an axon. Under normal conditions, the internal resistance (ri) or the axoplasm resistance and the external resistance (r,) or the resistance of the external fluid such as the blood serum or the physiological saline solution are much smaller than the membrane resistance.(r,). There is the membrane capacity (c,) in parallel with the membrane resistance. When a square pulse of electric current is passed across the membrane through a pair of electrodes, one inside the axon and the other outside, the current is spread along the nerve fiber as is shown by arrows in Fig. 6. The longitudinal current inside the axon decreases in intensity with distance, because part of the longitudinal current crosses the membrane at any particular point of the axon. Thus the membrane current is not uniform but declines in intensity along the axon. Moreover, there are two components of the membrane current, one through the membrane resistance (ionic current, i i ) and the other across the membrane capacity (capacitative current, i, ). Under these experimental conditions, it is very difficult to measure the membrane current and potential in any reasonable manner. If, however, a long wire electrode is inserted into the axon longitudinally, and another wire electrode is placed immediately outside of the axon, the longitudinal current in the axon and the membrane current become uniform throughout the entire length of the axon where the wire electrodes are applied (Fig. 6, middle diagram). The situation therefore becomes much simpler, but there are still two components of the membrane current, i.e. i, and ii. This

12

T. NARAHASHI

L

0

C

a i m

Fig. 6. Diagram of current flow in an axon preparation. Top, current is applied to the axon through internal and external microelectrodes. The membrane current and longitudinal current are not uniform along the axon. Middle, current is applied through internal and external wire electrodes. The currents are uniform along the axon (space clamp). Bottom, the ionic current (ii) across the membrane can be measured under voltage clamp conditions. ic, capacitative current; i,, total membrane current; c,, membrane capacity; ro, external resistance; ri, internal resistance; r,, membrane resistance. See text for further explanation.

condition is called “space clamp” and is prerequisite to voltage clamping. In order to eliminate the capacitative current across the membrane, one can make use of the fact that the membrane capacity is generally very small (1 pF/cm2 for squid giant axons) thereby making the duration of the capacitative current short. Instead of applying a square pulse of current and observing the resultant potential change, the membrane potential is changed in a square manner with the aid of an electronic feed-back circuit, and the membrane current necessary t o change the membrane potential is observed. This method is called “voltage clamp”. Because of the small value for the membrane capacity, the capacitative current under this condition ends in a brief period of time. In fact no

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

13

important change in membrane ionic current occurs during this period of time. Therefore, if one ignores the very beginning and the very end of the membrane current under voltage clamp conditions, membrane ionic currents can be observed as a function of time and membrane potential. A family of membrane currents associated with step membrane potential changes under voltage clamp conditions is shown in Fig. 7. When the membrane is hyperpolarized from - 1 15 mV t o - 155 mV, -155mv

-,

L

Fig. 7. Family of membrane currents recorded under voltage clamp conditions from the lobster giant axon. The membrane potential is suddenly changed from -115 mV to the values indicated. The top record represents the membrane current associated with a step hyperpolarization, and the other records the membrane currents associated with various magnitudes of step depolarizations (Narahashi e? al., 1964).

an ionic current flows inwardly across the membrane (top record). This is easy to interpret from the Ohm’s law. If, however, a depolarizing pulse is applied, the membrane ionic current flows in quite a different manner. A large inward ionic current is followed by a steady-state outward ionic current. The current pattern changes with a change in membrane potential; although the steady-state current simply increases in magnitude with increasing the depolarization (for example, compare the record at -45 mV and that at +15 mV in Fig. 7), the transient current, with increasing depolarization, first increases in magnitude (record at -25 mV), decreases again (records at -5 mV and +15 mV), and finally is converted into a transient outward current. (b) Current- Voltage Relations: When the peak value of the transient current and the final value of the steady-state current are plotted as a function of the membrane potential, current-voltage relations are obtained (Fig. 8). Extensive analyses of voltage clamp data have demonstrated that the peak transient current is carried mostly by sodium, whereas the late steady-state current is carried mostly by

14

T. NARAHASHI

potassium (Hodgkin and Huxley, 1952a, b, c, d; Hodgkin et al., 1 952). Therefore, the membrane potential where the transient current reverses its polarity is the sodium equilibrium potential. The potassium equilibrium potential cannot directly be measuted from the current-voltage curves such as those shown in Fig. 8, but separate measurements show that it is of the order of -80 mV.

-10

t

Fig. 8. Current-voltage relations for peak transient sodium current ( 1 ~ and ~ ) for steady-state potassium current ( I K ) in the voltage clamped lobster giant axon. I,, membrane current; Em, membrane potential; Eh, holding membrane potential from which the membrane is depolarized to various membrane potential levels (Narahashi, 1964b).

(c) Membrane Conductances: The membrane conductances t o sodium (gNa) and potassium (gK) are given by the following equations:

where ZN, and ZK are sodium and potassium currents, respectively, and E is the membrane potential. When the logarithms of the membrane conductances are plotted against the membrane potential, curves such as shown in Fig. 35 are obtained. The conductances thus calculated are the “chord conductances”, and are different from the “slope conductances” defined by aZ/aE.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

15

(d) Sodium Inactivation: As can be seen in membrane current records, the sodium current is transient even when the membrane is kept depolarized. Therefore, there are two mechanisms associated with the sodium conductance change, one being the mechanism whereby the sodium conductance is increased upon depolarization and the other the mechanism whereby the increased sodium conductance is decreased during sustained depolarization. The latter is often called the “sodium inactivation”. The time course of the sodium inactivation can be measured by applying two potential pulses. A conditioning pulse with a constant amplitude and varying durations is immediately ’followed by a test pulse of a constant amplitude and duration. The amplitude of the sodium current associated with the test pulse is then plotted as a function of the duration of the conditioning pulse. An exponential curve is obtained showing the time course of the sodium inactivation. The time constant depends on the membrane potential during the conditioning step, and on the temperature. An alternative way of obtaining the sodium inactivation curve is to plot the falling phase of the sodium current. However, the membrane current must be corrected for the potassium current. This may be achieved by the use of a specific inhibitor. For example, tetrodotoxin (TTX) is known to block the sodium current without any effect on the potassium current (Narahashi et al., 1964). If the membrane current recorded in TTX solution is subtracted from that recorded before application of TTX, the sodium current can be obtained. Another example for specific inhibitors is tetraethylammonium (TEA) which blocks the potassium current only when applied inside of the squid giant axon (Armstrong and Binstock, 1965). Therefore, the membrane current recorded from the TEA-treated axon represents the sodium current. Sodium inactivation is also a function of membrane potential. This relationship can be measured by the following two-pulse voltage clamp method. A conditioning pulse with a constant duration (30 ms or longer) and varying amplitudes is immediately followed by a constant test depolarizing pulse, and the amplitude of the transient sodium current associated with the test pulse is measured. The sodium current is plotted against the membrane potential of the conditioning pulse, and a sigmoid curve is obtained (Fig. 9). This is the steady-state sodium inactivation, and represents the availability of the sodium current at each membrane potential. This is one of the very important parameters in connection with drug actions, because

16

T. NARAHASHI

1

a,

I

-100

-50

1

0

Membrane potential (mV)

Fig. 9. Steady-state sodium inactivation curve from a squid giant axon. h-, the peak amplitude of sodium current associated with test step depolarization in a value relative to its maximum value. The abscissa represents the membrane potential of the conditioning pulse preceding the test depolarization (Moore et nl., 1964a).

changes in excitability are often explained in terms of a shift of the steady-state sodium inactivation curve along the potential axis. 5. Role of Metabolism Owing to the concentration gradients for sodium and potassium across the nerve membrane, the axon could gradually gain sodium and lose potassium. However, such a change does not in fact occur in the living tissue in situ, because there is in the axon the metabolic energy that constantly pumps out sodium and retains potassium. As long as the proper concentration gradient is maintained across the nerve membrane, the axon is capable of producing action potentials upon stimulation. This metabolic mechanism is called “sodium pump”. Immediately after excitation, the axon gains a small amount of sodium and loses a small amount of potassium. These changes in the ionic concentrations inside of the axon stimulate the sodium pump, and the concentration gradient is restored to the original value. The changes in internal sodium and potassium concentrations caused by one impulse are indeed very small. In the case of the squid giant axon which is about 5 0 0 p in diameter, the net influx of sodium and the net efflux of potassium are about 4 x lo-’*

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

17

mole/cmZ of the membrane per impulse (Keynes, 195 1 ; Keynes and Lewis, 1951). The sodium and potassium concentrations in the axoplasm are about 50 mM and 400 mM, respectively. Hence, the increase in axoplasm sodium concentration caused by one impulse is of the order of In fact, the squid axon is capable of producing a number of action potentials after its sodium pump has been completely inhibited by treatment with metabolic inhibitors such as iodoacetate and cyanide (Hodgkin and Keynes, 1955). Therefore, it can be said that the excitation is a metabolism independent process and not directly supported by the metabolic energy. B. SYNAPTIC AND NEUROMUSCULAR TRANSMISSION

I . Classification of Junctions At synapses or neuromuscular junctions, impulse propagation is in most cases mediated by a chemical substance called “transmitter”. In other cases, however, the membrane of the presynaptic and postsynaptic elements form a tight junction, and impulses are transmitted by means of a local circuit current in much the same way as in the axon. Examples of the electrical synapse are found in the synapse between the giant axon in the nerve cord and the motoneuron in the crayfish (Furshpan and Potter, 1959), and in the Mauthner cell of the goldfish (Furshpan, 1964; Furukawa and Furshpan, 1963; Furukawa, 1966). Presynaptic impulses could exert either excitatory or inhibitory effects on the postsynaptic element. This is true for both chemical and electrical synapses. In the excitatory synapse or neuromuscular junction, the presynaptic impulses stimulate the postsynaptic cell to cause excitation such as action potential or contracture. In the inhibitory junctions, the presynaptic impulses prevent the postsynaptic cell from being excited by the excitatory presynaptic impulses. These two different kinds of synapses and junctions form the basis of the complicated nerve network. Thus synapses and neuromuscular junctions are classed as follows: Mode of impulse transmission Electrical Chemical Role of junction Excitatory Inhibitory

18

T. NARAHASHI

2. Mechanism of Impulse Transmission (a) Excitatory Junctions: The preparation in which the mechanism of excitatory impulse transmission has been most extensively studied is the neuromuscular junction of the frog or mammal. When an impulse arrives a t the nerve terminal, a large amount of the transmitter substance acetylcholine (ACh) is released, and depolarizes the end-plate membrane of the muscle fiber. Unlike the membrane of the nerve or muscle fiber, the end-plate membrane is highly sensitive to ACh. The depolarization of the end-plate causes a local circuit current to flow across the muscle membrane surrounding the end-plate, so that an action potential is initiated from the muscle membrane. Cholinesterases present in the junction area hydrolyze the released ACh quickly, so that the stimulating action of ACh does not last an unnecessarily long period of time. When the nerve-muscle preparation is treated with d-tubocurarine, the neuromuscular transmission is blocked although the conduction of nerve or muscle is not impaired. Under these conditions, a microelectrode inserted in the end-plate region will record a small and slow depolarizing response upon nerve stimulation. This response is called the “end-plate potential (e.p.p.)”, and represents the depolarization of the end-plate membrane (Fig. 10). Since

Fig. 10. Action potenitial recorded from a sartorius muscle fiber of the frog and end-plate potential recorded from ancither fiber after treatment with d-tubc)CUrarine (Urakawa et QZ., 1960).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

19

d-tubocurarine suppresses the sensitivity of the end-plate membrane to ACh, the e.p.p. does not reach the threshold membrane potential beyond which an action potential of the muscle fiber is produced. In normal muscle preparations, the e.p.p. reaches the threshold, and the action potential almost masks the e.p.p. The e.p.p. is seen to be augmented and prolonged by treatment with anticholinesterases, because the transmitter action lasts longer under these conditions. In non-curarized preparations, anticholinesterases may initiate repetitive afterdischarges of the muscle fiber, and the transmission may eventually be blocked by high concentrations of accumulated ACh. In normal muscle preparations, small depolarizing responses can be observed from the end-plate without any presynaptic stimulation. The amplitude is of the order of 0.5-1 mV, and the duration is the same as that of the e.p.p. The responses occur spontaneously at a frequency of about 1Is, but the frequency is quite variable. They are called the “miniature end-plate potentials (m.e.p.p.s)”, and are produced by spontaneous release of ACh from the nerve terminal. The ACh is in fact released in quanta, and one m.e.p.p. is indicative of the depolarization caused by one ACh quantum which contains several thousand ACh molecules. The frequency and amplitude of the m.e.p.p.s may be differently affected by application of drugs. The change in frequency is a measure of the ability of the nerve terminal to release ACh, whereas the change in amplitude is due either to a change in quanta1 size (the number of ACh molecules in one quantum) or to a change in the sensitivity of the end-plate membrane to ACh, or both. The sensitivity of the end-plate membrane to ACh can directly be measured by recording the depolarization produced by application of ACh. A glass capillary microelectrode filled with ACh is brought close to the end-plate area and a brief positive electric pulse is applied to the electrode. Since ACh is positively charged, a small amount of ACh is ejected from the electrode tip, the amount being calculated from the intensity and duration of the pulse. The depolarization of the end-plate caused by the ACh is recorded by another microelectrode inserted in the end-plate region. At the excitatory synapses, a similar sequence of events occurs during the impulse transmission. The postsynaptic response caused by the transmitter is called the “excitatory postsynaptic potential (e.p.s.p.)”. It should be noted that the transmitter substance at the excitatory synapses or excitatory neuromuscular junctions is not

20

T. NARAHASHI

necessarily ACh. The transmitter is most probably 1-glutamate in crayfish and insect neuromuscular junctions (Faeder, 1968; Takeuchi and Takeuchi, 1964; Kerkut et aZ., 1965; Usherwood and Machili, 1966; Usherwood et al., 1968), whereas it is noradrenaline in adrenergic synapses of mammals. (b) Inhibitory Junctions: There are two types of inhibition, i.e. presynaptic inhibition and postsynaptic inhibition. In the presynaptic inhibition, the inhibitory nerve terminates near the excitatory presynaptic nerve terminals. The excitation of the inhibitory nerve causes a depolarization of the excitatory presynaptic nerve thereby decreasing the magnitude of the action potential there. This in turn causes a decrease in the amount of the transmitter released from the excitatory nerve terminals and results in an inhibition. In the postsynaptic inhibition, the excitation of the inhibitory nerve usually causes a transient hyperpolarization producing an “inhibitory postsynaptic potential (i.p.s.p.)”. The inhibitory impulse arriving at the nerve terminals at about the same time as the excitatory presynaptic impulse causes an inhibition of synaptic transmission through a suppression of the e.p.s.p. Gamma aminobutyric acid is a possible inhibitory transmitter substance in some inhibitory synapses and neuromuscular junctions, but the evidence is not very confirmative (e.g. Takeuchi and Takeuchi, 1965, 1966, 1967, 1969). (c) Ionic Mechanism: The ionic mechanism responsible for the potential change of the postsynaptic membrane is entirely different from that of the axonal membrane. Detailed voltage clamp analyses have been performed with the end-plate membrane of the frog (Takeuchi and Takeuchi, 1959, 1960). These studies are based on the observation of the end-plate currents (e.p.c.s) produced by nerve impulses when the end-plate membrane potential is clamped at various levels. Since the end-plate area is much smaller than the space constant of the muscle fiber, space clamp conditions can be established by a microelectrode inserted in the end-plate area. Another microelectrode is also inserted closely and serves as the potential electrode. ACh causes both sodium and potassium conductances t o increase almost simultaneously. The equilibrium potential for the overall conductance is about - 15 mV. In other excitatory or inhibitory synapses, different conductance changes may be involved. For example, at the inhibitory postsynaptic membrane of the cat motoneurons and snail neurons,

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

21

an increase in chloride conductance is the major factor (Ito et al., 1962; Kerkut and Thomas, 1964). Because the equilibrium potential for chloride is more negative than the normal resting potential, the inhibitory transmitter action causes a transient hyperpolarizing i.p.s.p. The excitatory postsynaptic membrane of the cat mo toneurons becomes permeable to all ions upon transmitter action (Eccles, 1964). The equilibrium potential is therefore near zero membmne potential and the response is a depolarizing e.p.s.p.

IV. FUNCTIONAL CHANGES CAUSED BY INSECTICIDES IN NERVE AND MUSCLE

In this section changes in nerve and muscle functions caused by insecticide intoxication will be described. Observations and experiments in this category are naturally descriptive and superficial, yet they will give a clue t o further exploration of the mechanism of action at the cellular and molecular levels. Experimental materials covered here are mostly limited to lower animals, especially to insects and other crustaceans. The studies described here also give the basis on which some important aspects of insect toxicology can be accounted for as will be described later (Sections VI, VII, andd VIII). A. DDT

1. Symptoms of Poisoning

The observation of the symptoms of poisoning is the first step in the study of the mechanism of action of an insecticide. For example,

if ataxia, hyperactivity or convulsion is observed in the insects poisoned with the insecticide, one can naturally suspect neuromuscular actions of the insecticide. On the other hand, if only paralysis occurs, the major action could either be neuromuscular blockage or metabolic inhibition. Intoxication with DDT results in ataxia and discoordination of insects. Convulsions of appendages and somatic muscle follow and last for a while, the period of which depends on the dosage and the kind of insects. The poisoned insect is eventually paralyzed (Yamasaki and Ishii," 1954a). *Former name of T. Narahashi.

22

T. NARAHASHI

2. Effects on Nervous Functions

One of the most sensitive nervous tissues to the action of DDT is the campaniform sensilla in the trochanter of the cockroach. When injected directly into the leg, DDT is effective in initiating trains of impulses at a concentration of 10-7-10-8 M (Fig. 11) (Becht, 1958; Lalonde and Brown, 1954; Roeder and Weiant, 1946, 1948, 1951; Yamasaki and Ishii, 1954a, b). However, not all sensory cells are

A

B

t .---- -- ----- -- ----- -- ----- -- ---- - - - 100 rnsec

0.2 mV

Fig. 1 1 . Trains of impulses from the sensory cells of the cockroach leg after injection of DDT into the leg. A, before injection of DDT; B and C, after injection (Narahashi, 1966).

equally sensitive to DDT. For example, the sensory cells on the cerci of the cockroach are less sensitive to DDT in producing trains of impulses (Roeder and Weiant, 1948; Eaton and Sternburg, 1967). The chemical sense organs on the tarsus and labellum of the housefly become more sensitive to adequate stimuli such as those by sucrose after intoxication with DDT (Smyth and Roys, 1955; Soliman and Cutkomp, 1963). Although the sense organs in the cockroach legs are highly sensitive t o DDT, repetitive discharges from them do not seem to be the sole factor responsible for the symptoms of poisoning. In the DDT-poisoned insect and other animals, the spontaneous discharge in the central nerve cord is increased in frequency, and synaptic transmission is facilitated (Dresden, 1949; Harlow, 1958; Heslop and

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

23

Ray, 1959; Tobias and Kollros, 1946; Yamasaki and Ishii, 1952b, 1954a, c). The role of these functional changes in producing the symptoms of poisoning can be studied by changing the temperature after intoxication with DDT (Yamasaki and Ishii, 1954a; Eaton and Sternburg, 1964). As will be described later (Section VI), DDT has a large negative temperature coefficient of action. When given an appropriate dose, the symptoms of poisoning appear at low temperature ( 15" -20" C) but reversibly disappear upon raising the temperature to 29"-35°C. At the low temperature, the poisoned cockroach produces ataxia and convulsion, and both the sensory nerve of the leg and the abdominal nerve cord discharge impulses at high frequencies. Upon raising the temperature, the symptoms of poisoning disappear and the impulse discharge in the abdominal nerve cord decreases in frequency, yet the sensory nerve in the leg is still discharging repetitively. Therefore, it is concluded that the sensory repetitive discharge alone cannot produce the symptoms of poisoning. DDT also stimulates nerve fibers or nerve terminals to produce repetitive discharges (Gordon and Welsh, 1948; Narahashi and Yamasaki, 1960c; Shanes, 1949a, by 195 1 ;Welsh and Gordon, 1947; Yamasaki and Ishii, 1952a; Roeder and Weiant, 1948; Harlow, 1958; Bodenstein, 1946; Van den Bercken, 1968). As a result of the hyperactivity of the nervous tissue, an unidentified toxic substance is believed t o be released from the nerve. This substance, sometimes called "autotoxin", is able t o stimulate and then paralyze the nerve (Hawkins and Sternburg, 1964; Shankland and Kearns, 1959; Sternburg, 1960, 1963; Sternburg and Kearns, 1952; Sternburg et al., 1959). In summary, DDT stimulates the nerve t o cause hyperactivity, and the resultant toxic substance eventually paralyzes the nerve. Another important feature of DDT action on the nerve is an increase in negative after-potential. This will be discussed in detail in a later section (V A). B. LINDANE

The symptoms of poisoning in lindane-intoxicated insects are characterized by ataxia, convulsions and eventual paralysis. The convulsions are more severe than those observed in DDT-poisoned insects (Yamasaki and Ishii, 1954d).

24

T. NARAHASHI

In lindane-poisoned insects, the effect on the central nervous system dominates over that on the peripheral nervous system. The frequency of spontaneous discharges in the central nerve cord is increased significantly by treatment with lindane, and the synaptic after-discharge is greatly prolonged (Fig. 12) (Dallemagne and Philippot, 1948; Fritsch, 1952; Fritsch and Krupp, 1952; Harlow,

Fig. 12. Effects of lindane on the synaptic after-discharges recorded from the abdominal nerve cord (postsynaptic) of the cockroach. Single stimuli were applied to the cercal nerve (presynaptic). A, control; B, C and D, after direct application of lindane lo-’ M to the nerve, 1 h 25 min (B), 4 h 10 min (C), and 6 h 10 min (D). Time marker 50 C.P.S. (Yamasaki and Ishii, 1954d).

1958; Vidal-Sivilla and Larralde, 1949; Yamasaki and Ishii, 1954d). No remarkable effect of lindane is observed on the sensory cells, nor on the nerve fibers (Becht, 1958; Lalonde and Brown, 1954; Yamasaki and Ishii, 1952a). There is no paralyzing action of lindane on the nerve. C. CYCLODIENES

Based on the observation of the symptoms of poisoning of dieldrin and aldrin (Fig. 13), it was suggested that the major site of action

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

25

was at the central nervous system (Gianotti et al., 1956). It was in fact found that the frequency of spontaneous discharges in the central nerve cord of the cockroach was increased and the synaptic after-discharge was prolonged by treatment with dieldrin, although the effect was less pronounced than that of lindane (Yamasaki and Ishii, 1958a). However, later experiments with highly purified

t Aldrin ( I )

Dieldrin

(II)

CI’

Aldrin-Transdiol (V)

Fig. 13. Conversion of aldrin and dieldrin to their analogs (Wang etal., 1971).

dieldrin samples showed that dieldrin, when applied directly to the nerve, exerted no effect on the spontaneous discharge and synaptic after-discharge (Narahashi, unpublished observation; J. W. Ray, personal communication). Apparently, the previous finding was due to impurity in the dieldrin sample used. Recent experiments with several derivatives and metabolites of dieldrin strongly point out that dieldrin is converted into active forms before exerting the neural effects (Wang et al., 1971). First of all, observations were made of synaptic transmissions across the last abdominal ganglion and the metathoracic ganglion in the dieldrin-poisoned cockroach. Although the last abdominal ganglion

26

T. NARAHASHI

was only negligibly affected, the synaptic transmission across the me tathoracic ganglion was found to be facilitated. When dieldrin was directly applied to the exposed metathoracic ganglion, it took 45 min or longer for the effect of dieldrin t o become apparent. Aldrin-transdiol (Fig. 13) stimulated the ganglion very quickly, and the synaptic transmission began to be prolonged only 5 min after treatment. When injected into the cockroach leg, dieldrin itself was able to stimulate the sensory cells to initiate repetitive trains of impulses only after a latency of 45 min. A longer latency was reported with topical application of dieldrin on the leg (Lalonde and Brown, 1 954). However, aldrin-transdiol produced repetitive discharges in only 2.5 min. Since aldrin-transdiol is one of the dieldrin metabolites (Matthews and Matsumura, 1969; Klein et al., 1968), it seems reasonable to assume that it is one of the active forms of dieldrin. D. PYRETHROIDS

It has long been known that pyrethrum is a fast acting insecticide, stimulating and paralyzing insects in a brief period of time. The active ingredients, pyrethrins, and the synthetic pyrethroid, allethrin, stimulate the nerve to cause repetitive discharges (Fig. 14) and then paralyze it (Lowenstein, 1942; Narahashi, 1962a; Welsh and Gordon, 1947; Yamasaki and Ishii, 1952a). However, no repetitive trains of

Ab

5 msec

I

P

rnsec Fig. 14. Compound action potentials recorded from the abdominal nerve cord of the cockroach by means of external electrodes. Aa and Ab, control; Ba and Bb, 1 rnin 30 s after M. Temperature 28OC (Narahashi, 1962a). treatment with allethrin 3.3 x 50

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

27

impulses can be observed in the cockroach sensory cells by treatment with pyrethrins (Lalonde and Brown, 1954). As in the case of DDT, pyrethrins were found to produce a toxin in the intoxicated insect (Blum and Kearns, 1956). The toxin is presumably responsible for the paralysis of the poisoned insect together with pyrethrins themselves. E. ROTENONE

Rotenone has been known to cause flaccid paralysis of insects, decrease in oxygen consumption, and decrease in the frequency of the heart beat (Harvey and Brown, 1951; Tischler, 1936; Hatai, 1941; Krijgsman et al., 1950; Orser and Brown, 195 1; Yamasaki and Ishii, 1951). The major mechanism of action is the inhibition of the electron transfer from DPNH to cytochrome b (Fukami, 1961; Fukami and Tomizawa, 1956, 1958a, b; Lindahl and Oberg, 1961; Oberg, 1961). Rotenone blocks the conduction of nerve (Fukami et aL, 1959), and depolarizes the nerve membrane (Yamasaki and Narahashi, 1957~).However, it remains to be seen whether the depolarization is due to the accumulation of potassium around the nerve membrane caused by the inhibition of the metabolic pump or the direct action on the nerve membrane. The muscle is also paralyzed by rotenone, but the effect is brought about later than the nerve paralysis (Fukami, 1954, 1956). It is noteworthy that three effects of rotenone, i.e. the nerve blockage, the metabolic disturbance, and the insecticidal action, go parallel with each other among a number of rotenone derivatives tested (Fukami et al., 1959) (see Section VIII C). F. ORGANOPHOSPHATES

It has well been established that organophosphorus insecticides inhibit ChE’s as their original forms or after having been converted into active forms. The transmitter substances in synapses and other junctions in insects still largely remain to be explored, but there is some evidence in support of the idea that 1-glutamate is the transmitter substance at the neuromuscular junctions of insects (Faeder, 1968; Kerkut et al., 1965; Usherwood and Machili, 1966; Usherwood et al., 1968). It seems also likely that ACh is the transmitter substance at the synapses in the last (sixth) abdominal ganglion of the cockroach (Yamasaki and Narahashi, 1960; Callec

28

T. NARAHASHI

and Boistel, 1967; Kerkut et al., 1969). These synapses connect the cercal sensory nerve fibers with the giant axons in the abdominal nerve cord. The effects of various organophosphorus insecticides on the insect nerve and muscle system are explicable in terms of their antiChE activity.

I . Synaptic After-discharges When the cercal nerve of the cockroach is stimulated by a single shock, the postsynaptic response can be recorded from the abdominal nerve cord. The postsynaptic response is composed of an initial few spikes of large amplitudes followed by an after-discharge of small amplitudes. The after-discharge lasts for about 100 ms (Fig. 15). After treatment with an organophosphorus or other antiChE compound, the synaptic after-discharge is prolonged in duration and increased in amplitude. The effect becomes more pronounced with time, and eventually the postsynaptic neurons produce a burst of high-frequency discharges which is followed by a sudden cessation and paralysis. The synaptic transmission is blocked at that time. However, spontaneous discharges begin to appear soon and the synaptic transmission is restored in the continuous presence of the ChE inhibitor. This block-and-recovery process is repeated many times in the presence of the ChE inhibitor (Narahashi and Yamasaki, 1960a; Roeder, 1948; Roeder and Kennedy, 1955; Roeder et al., 1947; Twarog and Roeder, 1957; Yamasaki and Narahashi, 1958c, 1960). Similar effects are observed in the locust (Harlow, 1958).

2. Postsynaptic Potential and Membrane Potential The mechanism of action of anti-ChE’s on the synaptic transmission was studied in more detail by recording the excitatory postsynaptic potential and membrane potential (Yamasaki and Narahashi, 1958c, 1960). When one external electrode is in contact with the last abdominal ganglion and the other with the abdominal nerve cord (e.g. the connective between the second and third abdominal ganglia), changes in the membrane potential of the last abdominal ganglion produced by drugs or presynaptic stimulation can be recorded together with the action potentials. Examples of such records of the postsynaptic responses are shown in Fig. 15. In normal physiological saline solution, the initial large spikes are followed by a slow depolarization on which small after-discharges are superimposed (far left record in A). After

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

29

M on the synaptic transmission across the last Fig. 15. Effects of eserine 7.7 x abdominal ganghon of the cockroach. A, from left to right, before, 12,19and 29 min after treatment with eserine, respectively. B, as in A, but 56 min after eserine. All records are arranged on the same base line shown on the record far left. B1,from left to right, responses at the time of presynaptic stimulation, 15, 30 and 45 s after stimulation. B2, continuation of B1, from left to right, 60, 75, 90, and 180s after stimulation. C, spontaneous depolarization and repolarization, accompanied by discharges, 81 min after treatment with eserine. The time intervals between successive records are 15, 30, 60, 30, 15,45,and 15 s, respectively. D, postsynaptic responses showing their dependence on the membrane potential, 240 min after treatment with eserine. Voltage calibration in A, 0.5 mV, applied to A, B, and C; voltage calibration in D, 5 mV. Time marker in C2,50c.P.s., applied to A, B, and C; time marker in D, 100 C.P.S. (Yamasaki and Narahashi,1960).

30

T. NARAHASHI

treatment with eserine 7.7 x lo-’ M, the late slow depolarization is increased in magnitude and prolonged in duration with an increasing number of spikes superimposed on it (A). Finally, the late slow depolarization reaches a threshold level beyond which discharges are blocked (B 1 and B2). The depolarizing phase may last as long as 10 s. At this stage, therefore, the prolonged slow depolarization and after-discharges initiated by a single presynaptic stimulus are followed by a cessation of discharges which are in turn followed by a reappearance of discharges as the membrane is slowly repolarized toward the resting level. The repolarizing phase may last as long as 3 min. This process involving depolarization and repolarization appears even spontaneously, and discharges can be seen at a certain depolarized level (C1 and C2). The synaptic transmission is blocked when the membrane is spontaneously depolarized beyond the threshold, and is restored as the membrane is repolarized (D). The mechanism whereby such a spontaneous depolarization-repolarization is produced under the influence of antiChE’s remains to be seen. The slow depolarization observed in the preparation treated with antiChE has been found to be a prolonged large e.p.s.p. Phenobarbital or urethan, when applied at appropriate concentrations, is capable of blocking synaptic transmission without affecting the conduction of the presynaptic nerve fibers. Under these conditions, the e.p.s.p. can be observed without being disturbed by spike discharges superimposed upon it. Examples of records of the e.p.s.p.s are shown in Fig. 16. It is clearly seen that the e.p.s.p.

Fig. 16. Effects of eserine 1.5 x lo-’ M on the excitatory postsynaptic potential recorded from the last abdominal ganglion of the cockroach. The preparation is under the influence of urethan 5.6 x lo-’ M to stop discharges. A, before application of eserine; B, 8 min after application of eserine; D, 12 min after application of eserine. Three short records on the right are 1, 2, and 3 s after the stimulation, respectively. Voltage calibration in C, 0.5 mV, applies to all records. Time marker in C, 50 c.P.s., applies to A and B. Time market in D, 50 C.P.S. (Narahashi, 1965b).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

31

observed under the influence of urethan 0.56 M is greatly augmented in magnitude and prolonged in duration after treatment with eserine 1.5 x 10-5 M.

3. Cholinesterase Inhibition in Nerve The effect of antiChE’s described in the foregoing section has been found t o be related to the inhibition of ChE in the nerve. The ChE activity in the nerve preparations that have started showing prolonged synaptic after-discharges after treatment with parathion, thiol-demeton, thiol-methyldemeton, or thiono-methyldemeton is partially inhibited (Yamasaki and Narahashi, 1960; Narahashi and Yamasaki, 1960a). The exact percentage of inhibition is difiicult to estimate because the ChE activity may be partly restored while the nerve preparation is homogenized and diluted for assay. It is reasonable t o assume that the persistent presence of the transmitter substance in the synaptic area causes a prolongation of the e.p.s.p. thereby producing a prolonged after-discharge.

4. Effects on Other Nerve-Muscle Systems When treated with TEPP, the sensory cells in the cockroach leg produce trains of impulses after a long latency. However, parathion and schradan have no effect on them (Lalonde and Brown, 1954). Although parathion has no effect on the chemo-receptors on the labellum of the housefly, paraxon stimulates it to increase the frequency of discharges (Leski and Cutkomp, 1962; Soliman and Cutkomp, 1963). The neuromuscular transmission of insects is not affected by anti-ChE’s (Colhoun, 1960; Harlow, 1958; Narahashi, unpublished observation) in agreement with the observation that the transmitter substance is likely t o be 1-glutamate.

V. MECHANISMS OF FUNCTIONAL CHANGES CAUSED BY INSECTICIDES IN NERVE AND MUSCLE A. DDT

1. After-potential and Repetitive Discharges

Increase in negative after-potential by treatment with DDT was discovered by Shanes (1 949b) for the first time using crab nerve as material. Shortly after that time, Yamasaki and Ishii (1952b) found that the cockroach nerve fibers underwent a similar change by

32

T. NARAHASHI

intoxication with DDT. These two studies were performed by means of external recording techniques. Detailed analyses of the increased negative after-potential in the DDT-poisoned cockroach giant axon were since undertaken with the aid of both extracellular and intracellular electrodes (see Fig. 20). (a) After-potential and Supernormal Phase: In the cockroach nerve, the negative after-potential is greatly prolonged after treatment with DDT. The effect can be observed with the crural nerve or with the abdominal nerve cord by means of external recording electrodes (Yamasaki and Narahashi, 1957a). Since there was general agreement that the negative after-potential is accompanied by a supernormal phase or a decrease in threshold and the positive after-potential is accompanied by a subnormal phase or an increase in threshold (Gasser, 1941 ;Gasser and Grundfest, 1936; Graham, 1930; Graham and Gasser, 193 1 ; Lehmann, 1937), changes in excitability were examined during the course of the increased negative after-potential in the DDT-poisoned cockroach axon (Y amasaki and Narahashi, 1957a). A supramaximum conditioning stimulus is followed by a test stimulus of just above threshold intensity. The height of the action potential produced by the test stimulus is plotted as a function of the interval between the two stimuli. As is shown in Fig. 17, there is no significant supernormal phase in the control normal nerve. In the DDT-poisoned nerve, however, the absolute refractory period is followed by a marked supernormal phase, its time course running almost parallel with the negative after-potential (Fig. 18). (b) After-potential as Studied by Microelectrodes: There was no doubt that the prolonged falling phase of the externally recorded action potential from the DDT-poisoned cockroach nerve bundle was not due to a summation of repetitively firing small action potentials but due to an increase in negative after-potential in each fiber, because the same effect was observed in action potentials from single fibers in the nerve bundle (Yamasaki and Narahashi, 1957a). However, it is necessary t o confirm this notion by means of intracellular microelectrodes. Moreover, more detailed analyses on the mechanism of action of DDT can be undertaken by this technique. The first observation of the DDT-induced large negative after-potential with intracellular microelectrodes was made by Yamasaki and Narahashi (1957b). More detailed analyses were performed later as described below.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

33

Fig. 17. Recovery process after an action potential i n the cockroach abdominal nerve cord bathed in normal solution. Ordinate, the amplitude of the externally recorded action potential (in percentage) relative to that produced by the supramaximum stimulation. Abscissa, the interval between a maximum conditioning stimulus and a weak submaximum test stimulus. The broken horizontal line represents the amplitude of the action potential produced by the test stimulus alone, and the dots represent the responses by the test stimuli when they are preceded by a conditioning stimulus. The action potential produced by the conditioning stimulus is also drawn with the peak as 100%. Note the absence of the supernormal phase after the conditioning action potential (Yamasaki and Narahashi, 1957a).

Fig. 18. Recovery process after an action potential in the abdominal nerve cord from the DDT-poisoned cockroach. See Fig. 17 for explanation. Note that the increased and prolonged negative after-potential is accompanied by a marked supernormal phase (Yamasaki and Narahashi, 1957a).

34

T. NARAHASHI

(i) After-potential in Normal Cockroach Giant Axons: In the normal cockroach giant axon, the action potential is followed by an undershoot or positive phase of about 5 mV amplitude which in turn is followed by a small negative after-potential of about 1.5 mV amplitude (Fig. 19). The positive phase can be explained in terms of a persistent increase in membrane potassium conductance (Yamasaki and Narahashi, 1959). The negative after-potential has been

Fig. 19. Action potentials recorded intracellularly from the giant axon of the cockroach. Note that the spike is followed by an undershoot or positive phase which is in turn followed by a slight depolarizing phase or negative after-potential (Narahashi, 1965a).

demonstrated to be due to an accumulation of potassium in the immediate vicinity of the nerve membrane (Narahashi and Yamasaki, 1960b). The conclusion is based primarily on the analyses of the time course of the negative after-potentials during repetitive stimuli. (ii) After-potential in DDT-Poisoned Cockroach Giant Axons: After introducing DDT into the nerve chamber at a concentration of M ythe negative after-potential is slowly increased in magnitude and prolonged in duration (Fig. 20). Repetitive after-discharges are usually produced by a single stimulus as the negative after-potential is increased. The repetitive responsiveness disappears as the negative after-potential grows further, and the latter finally reaches about 30 mV or more in magnitude (Fig. 20). The resting potential remains essentially unchanged, and the rising phase and the peak magnitude of the action potential are unaffected (Narahashi and Yamasaki, 1960~). In contrast to the negative after-potential in the normal unpoisoned cockroach giant axon, the DDT-induced large negative after-potentials are not built up upon repetitive stimuli. This behavior is shown in Fig. 21 in which the initial height of the negative after-potential remains almost constant during repetitive stimuli. If the effect of DDT on the negative after-potential were due

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

0.1

msec

35

10 msec

Fig. 20. Changes in intracellularly recorded action potential of the cockroach giant axon M. A, from top to bottom, before 38 min after, and 90 min after treatment with DDT after treatment with DDT. The horizontal lines indicate zero potential level. B, as in A, but with slower sweep (Narahashi and Yamasaki, 1 9 6 0 ~ ) .

to an increase in potassium accumulation around the nerve membrane, the negative after-potentials would be built up during repetitive stimuli as has been observed in the normal axon. Therefore, the effect is possibly due t o changes in conductance parameters responsible for the falling phase of the action potential, i.e. the mechanism whereby the sodium conductance is decreased or the sodium inactivation, or the mechanism whereby the potassium conductance is increased upon stimulation, or both. When a square pulse of current is applied to the nerve membrane, the resultant electrotonic potential rises exponentially and attains a steady-state level. A current-voltage relation for the steady state shows a rectification in the depolarizing direction, increasing the intensity of outward current producing a smaller magnitude of steady depolarization than the corresponding steady hyperpolarization. This is called “delayed rectification”, and can be ascribed to the increase in potassium conductance of the membrane. The delayed rectification has been found to be suppressed by application of DDT (Narahashi and Yamasaki, 1960c). It was therefore suggested that the

36

T. NARAHASHI

Fin. 21. After-uotentials during reuetitive stimuli of varvine freauencies in the no1.mal (A) and DDT-poisoned (B) cockro&h-giant axons. The spikk pkentials are too large to be recorded. The frequencies of stimuli are, from top to bottom in A, 50, 100, 150, and 200 c.P.s., and in B, single stimulus, 50, 100, 200 and 300 C.P.S. (Narahashi and Yamasaki,

1960b, c).

suppression of the potassium conductance increase by DDT was at least partly responsible for the increase in negative after-potential. The large negative after-potential in the DDT-poisoned cockroach giant axon is further augmented in magnitude by removal of potassium from the bathing medium, producing a plateau resembling cardiac action potentials (Narahashi and Yamasaki, 1960d). The DDT-poisoned axon in K-free medium resembles cardiac tissues not only in the shape of the action potential but also in its electrical properties. For example, anodal break response is easily produced, the action potentials are very often produced spontaneously

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

37

(Fig. 22), the plateau phase is abolished by application of anodal current, and the refractory period for the duration of the plateau phase is extremely long. Based on the measurements of membrane conductance during the plateau, it was suggested that the sodium conductance, after having risen to the normal value, declined slowly and the potassium conductance underwent little or no change during

Fig. 22. Action potentials produced by a single stimulus (top record), and spontaneously (middle and bottom records) in the cockroach giant axon bathed in K-free DDT medium. Spike potentials are too large to be recorded (Narahashi and Yamasaki, 1960d).

the plateau phase in the DDT-poisoned cockroach giant axon (Narahashi and Yamasaki, 1960d). This notion has now been subjected t o voltage clamp analyses (Section V A 2 ) . (c) Repetitive Discharge: Welsh and Gordon ( 1947) and Gordon and Welsh (1 948) made interesting observations on the role of calcium in the DDT-induced repetitive discharge in crustacean nerves. An increase in calcium concentration in the bathing medium generally suppresses the repetitive responsiveness induced by DDT. This observation is taken as indicating that DDT somehow disturbs the binding of calcium with the nerve membrane components thereby causing an unstabilizing effect. The increase in negative after-potential during the course of DDT

38

T. NARAHASHI

poisoning is no doubt one of the factors responsible for repetitive responsiveness, because a sustained depolarization works as a stimulant. However, it should be noted that the sustained negative after-potential is not the sole factor responsible for repetitive firing. A prolonged outward current applied to the normal cockroach giant axon does not produce repetitive firing (Yamasaki and Narahashi, 1959). Therefore, the DDT-poisoned axon must undergo changes in such a way as the sustained negative after-potential can initiate repetitive firing. It is also of interest that the repetitive responsiveness caused by DDT has a very high negative temperature coefficient of action as will be described later (Section VI A). 2. Effects on Membrane Ionic Conductances The hypothesis decribed in the foregoing section can be demonstrated by voltage clamp experiments whereby each component of membrane ionic conductances is measured. Detailed analyses have been performed using lobster giant axons as material (Narahashi and Haas, 1967, 1968). (a) Methods: Squid giant axons are most convenient for voltage clamp experiments because of their large diameter (about 5 0 0 ~ ) . However, it was found that they were extremely insensitive to DDT, external or internal application of DDT causing only a small increase in negative after-potential even at a very high concentration of 10-4 M. Therefore, lobster giant axons, which were sensitive to DDT as cockroach giant axons, were used as material for voltage clamp experiments. The lobster giant axon is only about 80 p in diameter on an average, but large enough to do voltage clamp experiments if the sucrose-gap apparatus is used. As described earlier (Section I11 A 4), measurements of membrane currents by voltage clamp techniques require the space clamp conditions in which membrane current and membrane potential are distributed uniformly in a limited area of the nerve preparation. This can be achieved by inserting a wire electrode longitudinally into the giant axon. The squid giant axon is large enough for the wire insertion, but such a technique is almost impracticable in the lobster giant axon. However, if the sucrose-gap insulation originally developed by S t h p l f i (1954) is combined with voltage clamp, one can measure membrane currents. This method called “sucrose-gap voltage clamp” was developed by Julian et al. (1962a, b) for the lobster giant axon, and has since been used not only for the lobster

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

39

giant axon but also for the squid and crayfish giant axons (e.g. Moore et al., 1964a, by 1967; Takata et al., 1966a, b; Narahashi et al., 1967a, b y 1969a, byc; Narahashi and Anderson, 1967; Frazier et al., 1969). A narrow portion (about 100 p wide) of an isolated giant axon, which is in contact with the physiological saline solution, is insulated from both ends of the axon by means of two isotonic sucrose solutions. One end of the axon across this sucrose insulation is in contact with the physiological saline solution, while the other end is in contact with the isotonic KCl solution to depolarize the membrane completely. All of the solutions are flowing continuously. Because the sucrose insulation under these conditions is almost perfect, the absolute value for the membrane potential, without any significant attentuation, can be measured on the central narrow portion called “artificial node” using the KC1 pool as the zero reference potential. The artificial node can be stimulated by application of current pulses between the artificial node and the physiological saline pool. Since the width of the artificial node is much smaller than the length constant of the axon, the space clamp conditions are established in this node. Thus it is possible to make voltage clamp measurements of membrane currents on the artificial node. The sucrose-gap voltage clamp method has several advantages and disadvantages over the conventional axial wire voltage clamp method. Details of the technique will be discussed elsewhere, and will not be described here. The only point worthy of mentioning in connection with the study of DDT action is the fact that the survival time of an artificial node is relatively short (less than 20 min) in the lobster giant axon. Since the action of DDT progresses slowly, it is necessary to measure membrane currents on normal control axons and on DDT-treated axons separately. The survival time of the artificial node under sucrose-gap conditions is longer for larger axons. This is probably due to leakages of internal ions into the sucrose solution across the nerve membrane. (b) Membrane Currents: The top set of records in Fig. 23 represents a family of membrane currents associated with step depolarizations of various magnitudes recorded from a normal lobster giant axon under sucrose-gap voltage clamp conditions. The second set from the top shows a similar family of membrane currents recorded from another lobster giant axon poisoned with 5 x M DDT for a period of 40 min. Two changes brought about by DDT treatment are

40

T. NARAHASHI

4

ma/crn2 10

mv

Normal

-5

5 X 10-4M DDT t 3 X I O-7M T T X - 4 min 0

-

-20

-5

6min

-

8 ,0 -60

-40

20

-__l_m

C

.

.

0

1

2

'

*

k

3

4

5

'

I

-28

\-40 -50

6 7 msec Fig. 23. Families of membrane currents associated with step deuolarizations in a normal lobster giant axon, and another axon treated with DDT 5 x M and with DDT and tetrodotoxin ('ITX) 3 x lo-' M. The third set of records shows changes in membrane current during the course of TTX action. The dotted lines in each set refer to the zero base line (Narahashi and Haas, 1967).

easily recognized: (1) the transient sodium currents, though they rise almost normally, fall more slowly in the DDT-poisoned axon; (2) at certain membrane potentials (-20mV, -40mV, and -50mV in Fig. 23), the transient sodium current is followed by an inward steady-state current in the DDT-poisoned axon. This is never observed in the normal axon. There are two possible explanations for the inward steady-state current flow in the DDT-poisoned axon. One of them is to assume potassium as its carrier. However, this possibility can be excluded by the fact that the resting potential remained essentially unchanged by

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

41

treatment with DDT. The constant resting potential suggests that the internal potassium concentration is not changed drastically. Hence it is not possible for potassium current to flow inwardly at those membrane potentials mentioned above. The other explanation would be that the inward steady-state current represents a residual component of prolonged sodium current. In order t o demonstrate this possibility, tetrodotoxin (TTX) was applied on the DDT-poisoned axon. Tetrodotoxin is the active principle of the puffer fish poison, and has been demonstrated to block the sodium current selectively without any effect on the potassium current (Narahashi et al., 1964). Changes in membrane currents during the application of TTX 3 x lo-' M to the DDT-poisoned axon are shown in the third set of records in Fig. 23. It is seen that the transient sodium current is completely blocked and the inward steady-state current is now converted into a small outward steady-state current 4 min after introduction of TTX. The difference between the membrane current .at 0 min in TTX and that at 4 min in TTX should represent the sodium current flowing in the DDT-poisoned axon. The bottom set of records in Fig. 23 represents a family of membrane currents associated with various step depolarizing pulses in DDT plus TTX. The sodium currents are completely blocked, whereas the potassium currents are suppressed in magnitude compared with those from the normal axon (note that the ordinate scale is different). (c) Current- Voltage Relations: The current-voltage relations for the peak value of the transient current and for the steady-state current are illustrated in Fig. 24. The current-voltage curve for the peak transient current is not appreciably affected by exposure t o DDT (open circles). However, the steady-state current undergoes a considerable change (open triangles). In the normal axon, the steady-state current flows in outward direction in the entire range of membrane potential studied (Fig. 24(a)). In the DDT-poisoned axon, however, the steady-state current is seen to be flowing in inward direction at the membrane potentials ranging from -60mV to -15 mV (Fig. 24(b)). Moreover, the amplitude of the steady-state current is suppressed at more depolarized membrane potential levels. The membrane currents in DDT plus TTX are depicted by closed symbols in Fig. 24(b)). The transient sodium current is almost completely inhibited (closed circles). The steady-state current is now flowing in outward direction in the entire range of membrane

T. NARAHASHI

42 (a)

2-13-67-Al Normal

3-10-67-02

5XIO*M DDT 3XIO-'M TTX

+

Fig. 24. Current-voltage relations for peak transient (sodium) current (Ip) and steady-state (potassium) current (1,J in a normal lobster giant axon, and in another axon treated with DDT 5 x M or with DDT and tetrodotoxin (TI'X) 3 x lo-' M. The broken line shows the residual component of the transient current and was obtained by subtracting I,in DDT plus TTX from I,in DDT (Narahashi and Haas, 1967).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

43

potentials studied (closed triangles), and should represent the potassium current. Therefore, the residual component of the sodium current can be obtained by subtracting the steady-state potassium current in DDT plus TTX (closed triangles) from the steady-state current in DDT (open triangles), and is drawn by a broken line. It is noteworthy that the residual sodium current reverses its polarity at the membrane potential where the peak transient sodium current also reverses its polarity (+40mV). Thus it is clear that in the DDT-poisoned axon the sodium current is greatly prolonged in its falling phase and the potassium current is suppressed. Since these two are the mechanisms that are directly responsible for the falling phase of the action potential, the inhibition of both of them naturally causes a prolongation of the action potential as has actually been observed. (d) Time Course of Na Inactivation: The time course of sodium inactivation can be plotted from the membrane current corrected for the potassium component. The procedure is shown in Fig. 25. The upper set of tracings shows the time course of the sodium current in a normal lobster giant axon. In this case saxitoxin (STX) is used

Fig. 25. Separation of membrane current into sodium current ( 1 ~ and ~ ) potassium current ( I K ) by use of saxitoxin or tetrodotoxin 3 x lo-' M in a normal and in a DDT-treated lobster giant axon. The membrane current in saxitoxin and that in DDT plus tetrodotoxin represents I K . IN^ is obtained by subtraction of IK from the total membrane current (Narahashi and Haas, 1968).

44

T. NARAHASHI

instead of TTX. Saxitoxin is the toxic principle of the poison from the toxic Alaska butter clam, Saxidomas giganteus. It is suggested that STX in the clam originally derives from the dinoflagellates, Gonyaulax catanella (Kao, 1966; Schantz et al., 1966). It has been shown that STX behaves in almost the same way as TTX in selectively blocking the sodium conductance increase (Narahashi et al., 1967b). The net sodium current (INa)is obtained by subtraction of the membrane current in STX from that before STX. The net sodium current in the DDT-poisoned axon is obtained in the same manner by using TTX, and is illustrated in the lower half of Fig. 25. The falling phase of the sodium current is then plotted in Fig. 26 on a semilogarithmic scale. It is clearly seen that the sodium current in the normal axon declines exponentially, whereas in the DDT-poisoned axon it declines slowly in two or more exponential functions. The time constant of the falling phase of the sodium current is estimated as 0.65 ms on an average for the normal axons. The average value for the DDT-poisoned axon is 2.93 ms for the first phase and 1 1.9 ms for the second phase.

0

2

4

6 8 Tinr ( m i r )

10

12

I1

Fig. 26. Semilogarithmic plot of the time course of the falling phase of the peak transient (sodium) current (Ip) at -20 mV in a normal and in a DDT-treated lobster giant axon after correction for the steady-state (potassium) current in the same way as in Fig. 25. The straight lines were drawn by eye (Narahashi and Haas, 1968).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

45

(e) Effects on Other Kinetic Parameters: Other kinetic parameters of membrane currents are also slowed by treatment with DDT, but the effect is much less than that on the time course of sodium inactivation. The time for the transient sodium current to reach its peak is slowed from the normal average value of 0.77-0.99ms at -20 mV membrane potential after exposure to DDT. The time for the steady-state potassium current to reach its half maximum is also slowed from the normal average value of 2.33 ms to 3.51 ms at 0 mV membrane potential by DDT intoxication. Thus, it can be concluded that the kinetics for the on-process of sodium current, the sodium inactivation, and the on-process of potassium current are all slowed by the action of DDT. The slowing of the sodium inactivation is most remarkable. ( f ) Discussion: The results of voltage clamp experiments account for the prolongation of the action potential by treatment with DDT. DDT inhibits the mechanism whereby the sodium conductance is turned off and that whereby the potassium conductance is turned on, and these mechanisms are directly responsible for the falling phase of the action potential. Therefore, the prolongation of the falling phase of the action potential by DDT can be ascribed to these changes in membrane conductances. These effects of DDT on membrane conductances were confirmed with the giant axon of the cockroach (Pichon, 1969a, b). The same effect of DDT on the time course of the sodium inactivation was found with the node of Ranvier of the frog (Hille, 1968). However, the potassium current is not appreciably suppressed by application of DDT. The experiments performed with both lobster and frog nerves show that the so-called “sodium channels” remain open for an unusually long period of time after intoxication with DDT. The sodium channel here simply refers to a conceptual pathway through which sodium ions flow according to the electrochemical potential gradient. It does not necessarily mean any anatomical hole or pore, or any carrier mechanism. It is also suggested that DDT has no effect on the sodium channels that are not open (Hille, 1968). B. ALLETHRIN

1. Effects on Action Potential

Intracellular microelectrode recordings of resting and action potentials from the cockroach giant axons have revealed three effects

46

T. NARAHASHI

of allethrin (Narahashi, 1962a): (1 ) the negative after-potential is increased and prolonged; (2) repetitive afterdischarges are produced by a single stimulus; (3) at a higher concentration of allethrin, the nerve conduction is eventually blocked. (a) After-potential and Repetitive Discharge: Figure 27 shows an example of a series of records of action potentials from the

Ca

Fig. 27. Action potentials of the cockroach giant axon before (Aa-c), 6 min after (Ba-c), M (Narahashi, 24 min after (Ca-c), and 88 min after (Da-c) treatment with allethrin 1962a).

cockroach giant axon before and after application of allethrin at a concentration of M. The rising phase of the action potential is only slightly slowed after application of allethrin, whereas the falling phase is greatly slowed and is followed by a prolonged negative after-potential. The resting potential remains essentially unchanged during the course of the allethrin action (Fig. 28). Repetitive discharges are often superimposed on the large negative afterpotential in the allethrin-poisoned axon. This phenomenon is especially remarkable at high temperature beyond about 27"C, as will be discussed later (Section VI B).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

47

ARP

W AP 12 120

80

20

Fig. 28. Changes in the amplitude of the action potential (AP), the initial amplitude of the negative after-potential (NAP), and the resting potential (RP) in the cockroach giant M (Narahashi, 1962a). axon after treatment with allethrin

A number of experiments have been performed by means of the microelectrode technique in an attempt to elucidate the mechanism of the increase in negative after-potential by allethrin (Narahashi, 1962b). Upon repetitive stimuli the negative after-potentials of the allethrin-poisoned axon are built up to some extent as in the case of the normal axon (Fig. 29). This is in sharp contrast with the situation of the DDT-poisoned axon in which no remarkable addition of the negative after-potentials is observed. This observation was taken as indicating that a depolarizing substance is accumulated outside or inside of the nerve membrane during the course of the repetitive stimuli.

A

B

C

Fig. 29. After-potentials produced by repetitive stimuli of varying frequencies in the cockroach giant axon. Only the positive phase and the negative after-potential are seen; the spike phase is too large to be recorded. A, in normal saline solution, 50 c.P.s.; B, 10 min M, 50 c.P.s.; C, 11 min, 10 C.P.S. (Narahashi, 1962b). after treatment with allethrin

48

T. NARAHASHI

However, the following experiment illustrated in Fig. 30 excludes the possibility that the depolarizing substance is potassium ion. When the potassium concentration is raised from the normal value of 3.1 mM to 30 mM, the after-potential of the allethrin-poisoned axon undergoes a considerable change in shape (Fig. 30, Ab). The after-potential associated with the second action potential elicited during the course of the after-potential of the first action potential

Aa

Ba

J

111

r-pF

Ab

J

5mv

50 msec 50 msec

I

Fig. 30. Effects of K-rich solution and conditioning action potentials on after-potentials of the allethrin-treated cockroach giant axon. Aa, in 3.3 x lo-' M allethrin; Ab, after treatment with 30 mM K; Ac, after washing with 3.3 x lo-' M allethrin. Bas, the test action potentials are produced at various moments during the course of the negative after-potential associated with the conditioning impulse (Narahashi, 1962b).

also undergoes a change, but in an entirely different way from the change brought about by high potassium (Fig. 30, Ba, Bb, Bc). Therefore, the large negative after-potential in the allethrin-poisoned axon cannot be ascribed to the accumulation of a large amount of potassium. Later experiments with voltage clamp techniques have demonstrated that this effect of allethrin is due to changes in sodium inactivation and potassium conductance increase mechanism (Section V B 2). (b) Conduction Block: At a high concentration of 3.3 x M, allethrin eventually blocks the action potential of the cockroach giant axon. The membrane is slightly and progressively depolarized during the course of this allethrin action, but the depolarization is not enough to account for the conduction block. An example of

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

49

such experiments is illustrated in Fig. 3 1, in which the maximum rate of rise of the action potential is plotted as a function of membrane potential. The maximum rate of rise of the action potential is proportional to the inward ionic current (sodium current) at that moment, and therefore can be used as a good measure of excitability. Before application of allethrin (open circles) the maximum rate of rise of the action potential is increased by anodal hyperpolarization

*

a

-

a

0

-

-

e

w

-

a

0 Control. Fblorizotion

-

A Allethrin I m l . Course of block 0 Allethrin ldptnl. Wrizotion I

I

Membrane potential (mV) Fig. 31. The maximum rate of rise of the action potential plotted as a function of membrane potential before and after application of allethrin (steady-state sodium inactivation curve). Cockroach giant axon. Open circles represent the measurements while the membrane potential is displaced from the resting potential (arrow) by polarizing currents. Filled triangles are the measurements during the course of allethrin action. Filled circles are similar measurements as the control after the membrane is depolarized by allethrin to the level shown by arrow (Narahashi, 1965a).

of the membrane and finally attains a steady value. It is decreased by cathodal depolarization and is finally blocked. When allethrin is applied, the maximum rate of rise of the action potential starts decreasing without any appreciable change in resting potential as shown by filled triangles. The resting potential then starts decreasing, and the excitability is completely blocked. However, the maximum rate of rise of the action potential is partially restored by anodal hyperpolarization (filled circles). This experiment strongly suggests that the mechanism by which the sodium conductance is increased AIP-3

50

T. NARAHASHI

upon stimulation is inhibited by allethrin. The notion was later demonstrated by the voltage clamp experiment (Section V B 2). Another point worthy of note in the experiment shown in Fig. 31 is that after treatment with allethrin the curve relating the maximum rate of rise of the action potential to the membrane potential (sodium inactivation curve) is shifted along the potential axis in the direction of hyperpolarization. This shift can be estimated from the membrane potentials where the respective curves attain 50% maximum value, or more directly if the two curves are normalized. Recent voltage clamp experiments with crayfish giant axons have demonstrated the shift of the sodium inactivation curve by poisoning with allethrin (unpublished observation). Thus this shift also contributes to the suppression of the action potential by allethrin. 2. Effects on Membrane Ionic Conductances Voltage clamp experiments on allethrin action have been performed with the giant axon of the squid (Narahashi and Anderson, 1967). Unlike DDT, allethrin exerts similar actions both on the cockroach giant axon and on the squid giant axon. It should be emphasized that the squid axon is much superior to the cockroach or lobster axon for voltage clamp analyses, because measurements on membrane currents can be made more accurately on the squid axon than on the other axons owing to its larger diameter and longer survival time under sucrose-gap conditions. (a) Methods: The sucrose-gap voltage clamp technique used for the squid giant axon is essentially the same as that for the lobster giant axon (Section V A 2a). Both external and internal applications of allethrin were attempted. It is somewhat surprising to find that allethrin exerts an additional effect when applied internally in view of its high lipophilic property. Two methods of internal perfusion of the squid giant axon were developed by different groups. Baker et al. (1961) developed a technique whereby the axoplasm was squeezed out by a small roller. When the crushed axon preparation was inflated by perfusion of internal media, the normal sized resting and action potentials were recorded and lasted for a few hours. Oikawa et al. ( 1961) developed a cannulation method. Two glass capillaries, one large (about 300 p in diameter) and the other small (about 150 p), were inserted longitudinally from both ends of the axon. After the capillaries met at the middle of the preparation, they were slowly withdrawn while internal media were introduced from the small capillary. The

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

51

axoplasm was gradually washed out from the large capillary and finally a continuous internal perfusion was established. In our experiments, a modified squeezing method was employed exclusively. The internal media generally contain potassium in the form of fluoride or glutamate salt. At the early stage of internal perfusion, potassium sulfate was often used (e.g. Narahashi, 1963b; Baker et al., 1962a, b), but it was later found that fluoride and glutamate were among the best anions for this purpose (Tasaki et al., 1965). In the present experiment on allethrin, the following two kinds of internal media were used: Solution I contained 400 mMK', 50 mM Na', 420 mM F-, 15 mM HzPO;, and 250 mM sucrose, and the pH was adjusted to 7.3. Solution I1 contained 350 mM K ' , 50 mM Na', 320 mM glutamate-, 50 mM F-, 15 mM H, PO,, and 333 mM sucrose and the pH was adjusted to 7.3. Both solutions gave essentially the same result. Before performing voltage clamp experiments, it was confirmed that the squid giant axon responds t o externally applied allethrin in the same manner as the cockroach giant axon, i.e. the negative after-potential is increased and prolonged, repetitive after-discharges

P-

............... -30pM

Allethrin Internally 10 min

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

Control

100mv

Fig. 32. Prolongation of the action potential of squid giant axons by internal perfusion of allethrin under sucrose+p conditions (Narahashi and Anderson, 1967).

are superimposed on the negative after-potential, and the action potential is eventually blocked. However, it should be noted that the effect on the negative after-potential is much more conspicuous when allethrin is applied internally than externally. The spike phase of the action potential is followed by a large and prolonged falling phase forming a plateau (Fig. 32). (b) Effects o n Membrane Currents b y External Application: Both peak transient sodium current and steady-state potassium current are

52

T. NARAHASHI

inhibited by application of allethrin at a concentration of lo-' M (Fig. 3 3 ) . When the peak current and the steady-state current are plotted as a function of the membrane potential, a current voltage relation can be obtained (Fig. 34). Control ......IOOmv

I------

-_

......

10 JJMAllethrin Externally 2.5 min

..... ......

........

8-5-65

..100mv

......

.......... 0

:-s0 Fig. 33. Families of membrane currents associated with step depolarizations in a voltage clamped squid giant axon before and during treatment with allethrin externally (Narahashi and Anderson, 1967). '-5 0

2 mnc

-

Control

8-4-65-C

Fig. 34. Current-voltage relations for the peak transient (sodium) current (Ip) and for the steady-state (potassium) current (lsJ in a voltage clamped squid giant axon before and during exposure t o allethrin externally (Narahashi and Anderson, 1967).

The membrane chord conductance during the peak sodium current can be calculated from the equation:

where g , refers to the peak conductance, I p the peak current, E the membrane potential, and E , the equilibrium potential for I,. The membrane slope conductance during the steady-state potassium

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

53

current is calculated instead of the chord conductance because of difficulty in estimating the potassium equilibrium potential. The slope conductance (g), for the steady-state current (Iss) is given by 4

s

=dEThe peak sodium conductance is suppressed t o an average of 53% normal control value by lo-’ M allethrin applied externally (seven experiments), whereas the steady potassium conductance is suppressed to an average of 67% (seven experiments). In Fig. 35, the logarithm of g , is plotted against the membrane 0-4-65-C

i

gP

300 (rnrnho/crnz)

I00

30

-

Control

10

I -60

1

-40

I

-20

1’ I

0

_ _ _IOpM - Allethrin

I

I

1

20

40

60

Externally 5 min

I

00 E (mv)

I

100

Fig. 35. The membrane conductance (g,) at the peak transient (sodium) current plotted

as a function of membrane potential in a voltage clamped squid giant axon before and during exposure to allethrin externally (Narahashi and Anderson, 1967).

potential before and during application of allethrin externally. The gp-E curve is shifted downward indicating the suppression of g , . (c) Effects on Membrane Currents b y Internal Application: Both the peak sodium current and the steady-state potassium current are suppressed by internal application of allethrin in a concentration of 10- -1 0-4 M. However, an additional effect has been found as might be expected from the extremely large negative after-potential when allethrin is applied internally (see Fig. 32). An example of membrane currents before and after application of allethrin internally is shown in Fig. 36. It is clearly seen that the peak sodium current, which is partially suppressed by allethrin, is followed by a steady-state inward

’

54

T. NARAHASHI

current at certain membrane potentials. This steady-state inward current cannot be due t o potassium ions, because the concentration gradient for potassium is maintained at a constant value by a continuous perfusion in both external and internal phases thereby eliminating the possibility of an inward flow of potassium at any membrane potentials. It is most likely that the peak sodium current 30 JIM Allethrin Internally 12 min

Control

........ 80mv ........ 60

...........8Omv

......... 40

........... 60

...........40

.........

..........20

.......

2 maec

Fig. 36. Families of membrane currents associated with step depolarizations in an internally perfused giant axon of the squid before and during exposme to allethrin internally. The dashed line on the right of each family represents the zero membrane current (Narahashi and Anderson, 1967).

is not terminated as quickly as normal but maintained for a while. This possibility is analyzed by drawing the current-voltage relatjon. Figure 37 shows that the peak sodium current is partially suppressed (open and filled circles). The steady-state current flows in inward direction at the membrane potentials ranging between -50 mV and -5 mV (filled triangles). Since the steady-state outward potassium current before treatment with allethrin starts flowing when the membrane is depolarized beyond -25 mV (filled triangles), and also since the steady-state inward current after treatment with allethrin reaches its maximum at the same membrane potential (filled triangles), it is reasonable to assume that the steady-state potassium current in the allethrin-poisoned axon also starts flowing at -25 mV membrane potential. The peak sodium current in the allethrinpoisoned axon reverses its polarity at +40 mV (filled circles) so that there should be no sodium component in the steady-state current at that membrane potential. Therefore, the potassium component in the steady-state current in the allethrin-poisoned axon can be approximated by a straight line connecting zero current at -25 mV and the steady-state current at +40 mV. This is shown by a dotted line in Fig. 37; the potassium current is now seen to be suppressed by

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

\

O\

55

30pM Allethrin Internally

....._ --- Ccfrected I,.

l2 rn'n

Fig. 37. Current-voltage relations for the peak transient current (Ip) and for the steady-state current (Id in an internally perfused giant axon of the squid before and during exposure to allethrin internally. The dotted line represents the potassium component in I,, corrected for the residual sodium current as described in the text (Narahashi and Anderson, 1967).

allethrin. Thus the sodium component in the steady-state current is the difference between the total steady-state current (filled triangles) and the potassium current (dotted line). Our recent voltage clamp experiments with crayfish giant axons have demonstrated the validity of this assumption. As in the case of DDT experiments (Section V A 2), TTX was used to eliminate the sodium component from the total membrane current recorded from the allethrin-poisoned axon. Detailed results will be reported elsewhere. (d) Effects on Kinetics of Conductance Change: In contrast to the marked prolongation of the kinetics of the sodium inactivation, the time for the sodium current to reach its peak is only slightly prolonged. The average prolongation amounts t o 18% and 12% for external and internal applications of allethrin, respectively. The time for the steady-state potassium current to reach 50% maximum is not affected by internal application of allethrin. (e) Discussion: The changes in conductance parameters by allethrin can adequately account for the changes in action potential. The suppression of the sodium conductance increase is directly responsible for the suppression or blockage of the action potential.

56

T. NARAHASHI

The suppression of the steady-state potassium current and the slowing of the sodium inactivation by internal application of allethrin are responsible for the slowing of the falling phase of the action potential. The fact that the slowing of the sodium inactivation is rather negligible when allethrin is applied externally reflects the small increase in negative after-potential. It is of interest t o see the differential effect of allethrin from both sides of the squid nerve membrane despite the fact that allethrin is highly lipid soluble. This might suggest that the mechanism whereby the sodium conductance is inactivated is located near the internal surface of the nerve membrane. VI. TEMPERATURE COEFFICIENT OF INSECTICIDAL ACTION A. DDT

It has long been known that the insecticidal action of DDT is stronger at low temperature than at high temperature (Barker, 1957; Dustan, 1947; Fullmer and Hoskins, 195 1; Guthrie, 1950; Efliger, 1948; Hoffman and Lindquist, 1949; Hoffman et al,, 1949; Kaeser, 1948; Lindquist et al., 1945, 1946; Menn et al., 1957; Nagasawa and Hoskins, 1962; Potter and Gillham, 1946; Pradhan, 1949; Rhoades and Brett, 1948; Richards and Cutkomp, 1946; Tahori and Hoskins, 1953; Tomaszewski and Gruner, 1951; Vinson and Kearns, 1952; Yamasaki and Ishii, 1953, 1954b; Yates, 1950). Because the rate of chemical reactions in general has a positive temperature coefficient, the negative temperature coefficient of the insecticidal action of DDT is of great interest from the viewpoint of the mode of action. At least four major factors must be taken into consideration to account for this phenomenon, i.e. (1) penetration of DDT through the integument, (2) detoxication of DDT, (3) accumulation of DDT in the adipose tissue, and (4) sensitivity of the target site or receptor to DDT. 1. Penetration of DDT Through the Integument When applied as a suspension, DDT was found to exert a stronger insecticidal action against the mosquito larvae at low temperature than at high temperature. However, when injected into the larvae, DDT was stronger at high temperature than at low temperature. Based on these observations, it was suggested that DDT was adsorbed more at low temperature than at high temperature (Fan et al., 1948).

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

57

However, the actual measurements o n the DDT penetration through the integument revealed that the rate of penetration was in fact faster at high temperature than at low temperature (Barker, 1957; Vinson and Kearns, 1952) (Table I). Therefore it can be concluded that the factor of penetration plays a role antagonistic to the negative temperature coefficient of the insecticidal action of DDT. Table I Qlo values of various actions of DDT in the cockroach (From Yamasaki and Ishii, 1954b)

Action of DDT

Temperature ("C)

-

Qio

Penetrability through the cuticle

15

Detoxication

15

1/LD50 by injection

15

-

35

1/LD50 by topical application

15

-

35

Potency in developing poisoning symptoms

15-35

<0.258

16-30

0.117-0.316

Nerve sensitivity

16

-

35 35

30

Source of data

1.414- 1.581 Vinson and Kearns (1952) 1.024 1.906 Vinson and Kearns (1952) 0.277 0.377 Vinson and Kearns (1952)

-

0.223

-

0.365 Vinson and Kearns (1952)

0.181

Vinson and Kearns (1952) YamasakiandIshii (1954b) Yamasaki and Ishii (1954b)

2. Detoxication of DDT DDT is detoxified into DDE, DDA or kelthane by the action of enzymes such as DDT dehydrochlorinase (e.g. Abedi et a/., 1963; Agosin et al., 1961 ;Bull and Adkisson, 1963; Miller and Perry, 1964; Perry, 1960; Perry et al., 1963; Tsukamoto, 1959, 1960, 1961). Because the action of enzymes has in general a positive temperature coefficient, it seems reasonable t o assume that the detoxication by enzymes plays a n important role in the negative temperature coefficient of the insecticidal action of DDT. This has been demonstrated to be the case (Barker, 1957; Menn et al., 1957; Vinson and Kearns, 1962) (Table I).

58

T. NARAHASHI

However, the detoxication factor is not sufficient t o account for the whole phenomenon in question. For example, when the amount of undetoxified DDT in the insect body is measured, a larger amount can be found in the survived individuals at high temperature than in the dead individuals at low temperature (Vinson and Kearns, 1952). In other words, the insect can tolerate a larger amount of DDT at high temperature than at low temperature. Another line of evidence in support of the idea that the detoxication is not the sole factor comes from the observation that the symptoms of DDT poisoning are reversible upon changing the temperature. When treated with an appropriate dose, the DDT-poisoned insect exhibits the symptoms of poisoning at a low temperature (1 5"C), but the symptoms disappear upon raising the temperature to 30°C. This process can be repeated several times by changing the temperature between the low and high values. If this reversibility is t o be explained in terms of detoxication, one must assume the reversible detoxication of DDT with respect to the temperature, but no DDT has been discovered in the insect injected with DDE o r DDA (Sternburg and Kearns, 1950). Therefore, it is not possible to account for the reversibility of the DDT symptoms in terms of the detoxication alone. 3. Accumulation of DDT in the Adipose Tissue Owing t o high lipid solubility, DDT is accumulated in the adipose tissue. The DDT molecules stored in the adipose tissue may not exert any toxic effect. Therefore, if DDT is discharged from the fat upon lowering the temperature, this might account for the stronger insecticidal action at the low temperature. In fact, the fat of the cockroach kept at low temperature for a while has a higher solubility and holding capacity of DDT than that kept at high temperature (Hurst, 1949; Munson, 1953; Munson et al., 1954). However, since no further quantitative data are available, this hypothesis remains speculative.

4. Sensitivity of Nerve to DDT From the foregoing discussion, it is naturally predicted that the sensitivity of the target site to DDT may change upon changing the temperature. This problem has been studied using the sensory neurons in the cockroach leg as material (Yamasaki and Ishii, 1953, 1954b). DDT was more effective in inducing trains of impulses in the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

59

sensory cells at a low temperature (1 6°C) than at a high temperature (30°C). In Fig. 38, the percentage of the individuals that produce the trains after injection of DDT into the leg is plotted as a function of the concentration of DDT. The values for the effective dose fifty (ED50) are estimated as 0.95 x M M at 16°C and 10.41 x at 30°C. The value of Qlo is calculated t o be 0.181 (Table I). However, it should be noted that the frequency of appearance of

-

I00

80

60 40

20 0

lo

lo-

10-6

Mol.conc. of DDT

Fig. 38. Percentage of the individuals that produce trains of impulses in the sensory nerve of the cockroach leg after injection of various concentrations of DDT into the leg at 16°C (open circles) and at 30°C (filled circles) (Yamasaki and Ishii, 1954b).

trains is higher at 30" C than at 16°C when comparison is made at an equivalent concentration of DDT, for example at the ED50 for each temperature. This reflects the fact that the symptoms of poisoning are more intense and easier t o observe at 30°C than at 16°C. The effect of DDT on the trains of impulses was found t o be reversible upon changing the temperature. Figure 39 depicts changes in the frequency of appearance of trains when the temperature is altered. After injection of DDT at a concentration of 3 x M, the trains appear as the temperature is lowered from 30°C to 12"C, disappear as the temperature is raised to 30"C, and this process can be repeated. The .dose of DDT necessary to develop the initial symptoms of poisoning was compared at 16" C and 30" C by injection method. The concentration of DDT to produce the symptoms is lower at 16°C than at 30"C, the difference being 5-20 times. This reflects the difference in nerve sensitivity t o DDT, because under these conditions the effects of other factors such as detoxication can be

60

T. NARAHASHI

Time (midafter injection

Fig. 39. Reversibility of appearance of trains of impulses in the sensory nerve of the cockroach leg injected with DDT when temperature is altered. DDT is injected once at zero time in a concentration of 3 x lo-' M (Yamasaki and Ishii, 1954b).

effectively eliminated. The value of Qro is calculated as 0.3 16 0.1 17 (Table I). Table I summarizes the values of Qlo for various factors involved in the DDT action. It can be concluded that, although the detoxication of DDT contributes t o the negative temperature coefficient of the insecticidal action of DDT, the major factor appears to be the sensitivity of the nerve to DDT. The nerve sensitivity not only has a large negative temperature coefficient but also can adequately account for the reversible symptoms of DDT poisoning with respect t o the temperature.

-

5. Discussion Eaton and Sternburg (1964, 1967) also studied the effect of temperature on the trains of impulses induced by DDT. Part of their results seems to contradict t o our results described above. However, careful examinations of the experimental conditions have revealed that both data agree in essence. Eaton and Sternburg (1964) stated that the trains of impulses from the sensory cells of the DDT-poisoned cockroach leg show a positive temperature coefficient, increasing the temperature increasing the frequency of appearance of trains, whereas the trains from the abdominal nerve cord show a negative temperature coefficient decreasing the temperature increasing the frequency of appearance of trains (their Fig. 2 and Table I). However, the threshold concentration of DDT to induce the trains of impulses is much higher for the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

61

abdominal nerve cord than for the sensory cells. Therefore, the dose of DDT used is well above the threshold for inducing the trains in the sensory nerve at both high and low temperatures. Thus, changes in temperature simply result in changes in the frequency of appearance of trains in the sensory nerve, increasing the temperature increasing the frequency. This agrees with the observation by Yamasaki and Ishii (1954b). For the abdominal nerve cord, however, the dose used is near the threshold to induce trains at high temperature. Therefore, the trains are not observed at the high temperature, and appear as the temperature is lowered. This also agrees with the observation by Yamasaki and Ishii ( 1954b) who applied the near-threshold concentration of DDT t o the sensory nerve. Eaton and Sternburg (1967) later performed similar experiments using the cercal nerve of the cockroach. The number of trains of impulses is markedly reduced as the temperature is lowered. Because of low threshold concentrations of DDT in the sensory nerve, this result is to be expected from the previous observations by Yamasaki and Ishii (1 954b) and by Eaton and Sternburg ( 1964). Thus, if the concentration of DDT is carefully chosen, the potency of DDT t o induce trains of impulses 'has a negative temperature coefficient both in the sensory nerve and in the abdominal nerve cord. In other words, the threshold concentration of DDT to produce trains becomes low as the temperature is lowered. Therefore, the DDT-poisoned insect shows reversible symptoms of poisoning upon changing the temperature if an appropriate dose is chosen. In the poisoned cockroach, the impulse trains in the central nervous system probably play an important role in exhibiting the symptoms of poisoning, because the trains disappear upon increasing the temperature in parallel with the symptoms of poisoning of insect, whereas the trains from the sensory nerve can still be observed at high temperature in the absence of the apparent symptoms of poisoning (Yamasaki and Ishii, 1954a; Eaton and Sternburg, 1964). B. PYRETHROIDS

The insecticidal activity of pyrethroids increases as the temperature is lowered (Blum and Kearns, 1956; Harries et al., 1945; Chamberlain, 1950; Guthrie, 1950; Hartzell and Wilcoxon, 1932). The effects of temperature on various aspects of allethrin action on the cockroach nerve have been studied. As described in an earlier section (IV D), allethrin causes the nerve to produce repetitive

62

T. NARAHASHI

discharges at relatively weak concentrations and blocks the conduction at higher concentrations.

I . Repetitive Discharge The repetitive discharge in the allethrin-poisoned axon shows a positive temperature coefficient. Figure 40 illustrates a series of records from the allethrin-treated cockroach giant axon when the

Fig. 40. Effects of temperature on the action potentials recorded from the allethrinpoisoned giant axon of the cockroach. Temperature 33°C (A), 28°C (B), 26.5"C (C), and 26°C (D) (Narahashi, 1962a).

temperature is gradually lowered from 33°C in record A (Narahashi, 1962a). As the temperature is lowered from 26.5"C in record C to 26OC in record D, the axon stops firing repetitively, leaving small oscillations of potential. Repetitive firing at high temperatures can be attributed, at least in part, t o the increase in negative after-potential upon raising the temperature (Narahashi, 1963a). The effectiveness of allethrin in producing repetitive discharges in the isolated abdominal nerve cord of the cockroach was studied at high and low temperatures (Y.Suzuki, personal communication). The allethrin concentrations to cause this effect in 50% of the individuals are estimated as 2.07 x M at 30°C and 1.1 x lod6 M at 15°C.

2. Nerve-blocking Action In contrast t o the positive temperature coefficient of the action of allethrin in producing repetitive discharges, the blocking action of

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

63

allethrin shows a negative temperature coefficient. Figure 4 1 depicts an example of experimental data in which the temperature and the amplitude of the action potential recorded externally from the cockroach nerve cord are plotted as a function of time (Narahashi, unpublished observation). Separate control experiments show that the action potential of the normal nerve cord increases in amplitude

0

5

10

15 20 Time after

30 35 - 4 0 g/ml Allelhrin (min.)

25

45

50

55

Fig. 41. Effects of temperature on the amplitude of the action potential recorded externally from the allethrin-poisoned abdominal nerve cord of the cockroach.

by only about 20% with lowering the temperature from 29°C to 15°C. In the nerve cord poisoned with allethrin 3.3 x M, however, the action potential is reduced in amplitude upon lowering the temperature from 29"C, and eventually blocked when the temperature reaches 12°C. This process can be repeated as is shown in Fig. 4 1. Y. Suzuki (personal communication) also found that M allethrin blocked the action potential of the cockroach nerve cord in a few minutes at 15" C, whereas it took more than 50 min to block at 30°C. To block the conduction at 30°C in a few minutes, the concentration of allethrin had to be raised to M. These observations clearly demonstrate that the nerve-blocking action of allethrin has a negative temperature coefficient. The mechanism involved in the negative temperature coefficient of the blocking action of allethrin has been studied by means of intracellular microelectrodes using the cockroach giant axon as material (Narahashi, unpublished observation). As described in an earlier section (V B l ) , the curve relating the maximum rate of rise of the action potential to the membrane potential (sodium inactivation curve) is shifted along the potential axis in the direction of inside

64

T. NARAHASHI

more negative membrane potential by application of allethrin (Fig. 3 1). In the allethrin-poisoned axon, the sodium inactivation curve is further shifted in the same direclion upon lowering the temperature from 26.5"C t o 14°C (Fig. 42). The absolute magnitude of the maximum rate of rise of the action potential is also decreased by lowering the temperature. Also plotted in Fig. 42 are the

:

:

: 0

I 1000

-10

AP T p -20 (mV)

-30

.........o.........

v.m.9.

0%.

4..

-

-40-..& -50

-

p...,

-m

->--*-

d...

,,*.-;*4

......A,......... .......... 0

% , . doto b

00

@ ."'

%

--ocfl --o--ai-a-

:j

,

,('

..

p.&.4..

,

800

6oo dV/dt

4oo (V/reC)

200

'4?&..n

Fig. 42. The amplitude (AP) and the maximum rate of rise (dV/dt) of the action potential and the threshold membrane potential (TP) plotted as a function of membrane potential displaced from the resting potentials (arrows at the top) before and during exposure to allethrin 3.3 x M. AP, before (0) and during (4 allethrin. dV/dt, before (0) and during (A) allethrin. Arrows on the curves show the membrane potentials where dV/dt is half maximum. TP, before (0) and during (m)allethrin.

amplitude of the action potential and the threshold membrane potential where the action potential arises as a function of membrane potential. The threshold membrane potential becomes inside less negative upon lowering the temperature. The resting membrane potential is decreased by lowering the temperature as shown by arrows with symbols in Fig. 42. All of these changes tend to suppress the conduction of the action potential. The amplitude of the action potential is increased upon lowering the temperature, so that once the action potential is produced it can reach a higher level at low temperature. In summary, the resting membrane potential is decreased, the threshold membrane potential is also decreased (inside less negative), the sodium inactivation curve is shifted along the potential axis in the direction of hyperpolarization, and the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

65

maximum rate of rise of the action potential is suppressed by a lowering of the temperature, and all of these changes are responsible for the stronger blockage of the conduction of the action potential at low temperature.

3. Discussion Although the study on the mechanism underlying the negative temperature coefficient of the insecticidal action of pyrethroids is less extensive than that for DDT, it is no doubt that the sensitivity of nerve to allethrin plays an important role in this phenomenon. The nerve-blocking action of allethrin is probably the major factor, because the cycle of blockage-recovery of the nerve conduction can be observed by changing the temperature in parallel with the reversible symptoms of poisoning in insects. From the positive temperature coefficient of the allethrin action in producing repetitive firing, it is predicted that the cockroach poisoned with a low dose of allethrin will become more hyperactive at high temperature than at low temperature. Careful observations of the symptoms of poisoning as a function of dose, time and temperature are necessary to evaluate the validity of this notion. Since pyrethroids are known t o be metabolized in insects as well as in mammals (see review by Yamamoto, 1970), it is reasonable to assume that the metabolic degradation of allethrin is accelerated by a rise in the temperature. If this is the case, then the metabolism of allethrin will play an additional role in the negative temperature coefficient of the insectidical action of allethrin. VII. INSECTICIDE RESISTANCE

A number of studies have been performed in an attempt to elucidate the mechanism of resistance to insecticides (see reviews by O’Brien, 1966, 1967; Brown, 1960, 1961, 1964). Although different mechanisms are in fact involved in different strains of resistant insects and in different insecticides, four major factors are easily recognized from Fig. 1, i.e. (1) penetration of insecticides through the integument, (2) detoxication and excretion of insecticides, (3) store of insecticides in non-target tissues, and (4) sensitivity of nerve to insecticides. In addition, there is so-called “behavioral resistance”, in which insects develop the ability to avoid contact with insecticides. This is outside the scope of the present article, and will not be described here. AIP-4

66

T. NARAHASHI

Importance of the nerve sensitivity to insecticides in insecticide resistance can easily be seen in the observation in which the amount of undertoxified insecticide is compared between susceptible and resistant strains of insects. The survived resistant insects in many cases contain a larger amount of undetoxified insecticides than the dead susceptible insects (Babers and Pratt, 1953; Perry and Hoskins, 1951; Sternburg et al., 1950; Tahori and Hoskins, 1963). This indicates that the resistant insects can tolerate a larger amount of insecticides than the susceptible insects without showing any sign of intoxication. For this reason, factors other than the cuticule penetration and detoxication are suspected to play an important role in insecticide resistance, although the detoxication factor has been demonstrated, in a number of resistant strains of insects, to be one of the key factors for the resistance. There should be some defense mechanisms whereby the target site in the resistant insects is protected from the toxic action of insecticides.

A. NERVE SENSITIVITY TO INSECTICIDES

Low sensitivity of the nerve to insecticides is one of the most probable mechanisms whereby the resistant insects can tolerate a large amount of the insecticide present in the body. Earlier studies indicate that the nerves from the resistant strains are less sensitive to insecticides than those from the susceptible strains (Pratt and Babers, 1953; Smyth and Roys, 1955; Weiant, 1955). Detailed studies were performed using a variety of insecticide-resistant strains of houseflies (Yamasaki and Narahashi, 1958b, 1962; Narahashi, 1964a; Tsukamoto et al., 1965). The test solution containing insecticide is applied to the exposed thoracic ganglia and the impulse discharge is recorded by means of external silver wire electrodes inserted in the femur of the metathoracic leg. An example of such records is shown in Fig. 43. The top record (A) shows a burst of discharges produced by stimulating the normal housefly with an air puff. Spontaneous discharges are low in both amplitude and frequency in the normal M to the housefly (record B). Direct application of DDT 2.8 x exposed thoracic ganglia induces bursts of discharges which increase in intensity with time (records C, D, and E). Therefore, in the experiments described in the following sections, the increase in the frequency of discharges is taken as a measure of response.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

67

Fig. 43. Discharges of motoneurons originating from the thoracic ganglia and innervating the femur muscle of the housefly (DDT-susceptible NAIDM strain). Recordings are made externally from the femur muscle by means of silver wire electrodes. A, burst of impulses induced by an air puff applied to the housefly. B, spontaneous discharges in another normal M to the housefly. C, burst of discharges 7 min after direct application of DDT 2.8 x exposed thoracic ganglia. D, 13 min after DDT. E, 26 min after DDT (Yamasaki and Narahashi, 1962).

1. DDT

The value of the effective dose fifty (ED50) in stimulating the nerve to increase the discharge frequency in 50% individuals is estimated as 3.4 x lo-' M for the strain susceptible to DDT (NAIDM), 2.1 x 10-6 M for the strain moderately resistant to DDT (CSMA), and 2.6 x M for the strain highly resistant to DDT (DKM). Thus the ED50 ratio DKM/NAIDM is 7.6 as against the LD50 ratio of 217 (Yamasaki and Narahashi, 1962). Another DDT-resistant strain of houseflies exhibits much lower nerve sensitivity to DDT. The ED50 ratio of the resistant strain to the susceptible strain (Lab) is estimated to be 100 (Tsukamoto et al., 1965) (Fig. 44). Recently, the labellar taste receptors of DDT-resistant houseflies were found to be also less sensitive t o DDT than those of DDT-susceptible houseflies (Browne and Kerr, 1967).

68

T. NARAHASHI x

104

Io4

DDT cu-centraticn (MI

o3

I

Fig. 44. Dose-response relations for DDT in inducing high frequency discharges from the thoracic ganglia of the housefly. Ordinate represents the percentage of the individuals that respond to various concentrations of DDT. Lab, oDDT-susceptible strain. R (bwb : ocra : ar : ac), DDT-resistant strain. F, , F, hybrid (R + x Labs d) (Tsukamoto ef a/., 1965).

2. Lindane The lindaneresistant strains of houseflies also exhibit low nerve sensitivity to lindane. Lindane exerts a similar effect as DDT on the thoracic ganglia of the housefly. The ED50 ratio of the resistant HR(2356) strain to the susceptible Lab strain is estimated t o be 124-162 as against the LD50 ratio of 59,000 (Narahashi, 1964a). 3. Dieldrin In the dieldrin-resistant strain (Hikone) of the houseflies, the effect of directly applied dieldrin in increasing the discharge frequency of the thoracic ganglia can be observed after a longer latency than in the susceptible strain (Takatsuki), the difference being 1.5 times (Yamasaki and Narahashi, 1958b). However, since impurity in the dieldrin sample used is suspected (Section IV C), it% possible that this difference in the latency does not represent the difference in the nerve sensitivity to dieldrin. Alternatively, this difference may reflect the difference in the ability of the nerve to convert dieldrin into an active compound for which aldrin-transdiol is a possible candidate (Wang et al., 1971, also see Section IV C). No

EFFECTS O F INSECTICIDES ON EXCITABLE TISSUES

69

significant difference has been found between susceptible and resistant strains of insects in the detoxication of dieldrin and in the penetration of dieldrin through the integument (Khan and Brown, 1966; Perry et al., 1964; Ray, 1963; Winteringham and Harrison, 1959). In view of these considerations, it is urged t o study the sensitivity of the nerve t o the activated dieldrin metabolites such as aldrin-transdiol. Matsumura and Hayashi (1966a, 1969) studied the binding of dieldrin with various components of the German cockroach nerve. The nerve from the dieldrin-resistant strain binds a less amount of dieldrin than that from the susceptible strain. However, it remains to be seen whether this factor is causally related to the resistance of the cockroach to dieldrin. 4. Diazinoti The nerve sensitivity t o insecticides plays a minor role in the diazinon-resistant strain of houseflies (Narahashi, 1964a). Since diazinon is converted into an active form diazoxon in insects, the sensitivity of the nerve was studied both for diazinon and diazoxon. The results are shown in Fig. 45, in which the nerve sensitivity to diazinon and diazoxon is only slightly lower in the resistant strain

90

- 70 ?i 50 0

5z

W

30

ti 10 I

10-6

I

I

10'~ CONCENTRATION

I 0-4

(M)

Fig. 45. Dose-response relations for diazinon and diazoxon in inducing high frequency discharges from the thoracic ganglia of the housefly. Ordinate represents the percentage of the individuals that respond to various concentrations of the insecticides. NAIDM, diazinon-susceptible strain. L-S-5,diazinon-resistant strain.

70

T. NARAHASHI

than in the susceptible strain. It is clear that the nerves from both strains are more sensitive to diazoxon than to diazinon. This is to be expected because diazoxon is more potent than diazinon in inhibiting ChE’s in vitro and also because the inhibition of ChE’s is directly responsible for the multiple discharges produced by organophosphorus insecticides (Section IV F). The results of experiments described above are consistent with the observation that ChE’s from both resistant and susceptible strains of houseflies are equally inhibited by diazoxon (T. Shono, personal communication).

B. GENES CONTROLLING THE NERVE SENSITIVITY

The apparent low nerve sensitivity in DDT- and lindane-resistant strains of houseflies described in the preceding section does not exclude the possibility that hese insecticides are detoxified inside the nerve thereby making the nerve less sensitive. In fact, it has been shown that the activity of DDT dehydrochlorinase is higher in DDT-resistant houseflies than in susceptible houseflies (Miyake et al., 1957). There are at least two genes controlling DDT resistance in the housefly, one being located on the second chromosome and the other on the fifth chromosome. A single recessive gene pair on the second chromosome is known to control the inheritance of the so-called knockdown resistance to DDT (Harrison, 1951 ; Milani, 1954; Milani and Travaglino, 1957). A dominant resistance gene on the fifth chromosome controls dehydrochlorination of DDT (Tsukamoto and Suzuki, 1964). Since knockdown of DDT-poisoned houseflies is caused by the action of DDT on the nerve, it is possible that the knockdown resistance gene on the second chromosome controls the low nerve sensitivity to DDT. On the other hand, it is also possible for the dominant dehydrochlorination gene on the fifth chromosome plays a major role in the low nerve sensitivity to DDT, because the nerve can detoxify DDT (Miyake et al., 1957). The nerve sensitivity to insecticides was analyzed using multichromosomally marked resistant strain of houseflies, R (bwb : ocra : ar : ac), and a susceptible strain, Lab (Tsukamoto et al., 1965). As is shown in Fig. 44, the nerve from the resistant strain was much less sensitive to the directly applied DDT, the difference between the susceptible and resistant strains being about 100. The F1 hybrid between the females of the resistant strain and the males

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

71

of the susceptible strain showed an intermediate nerve sensitivity to DDT. In order t o analyze the genetic factor responsible for the nerve sensitivity to DDT, the males of the F1 hybrid were backcrossed to the females of the resistant strain. Since each autosome except for the fourth chromosome was labeled with a visible mutant marker, it was possible t o determine to which linkage group the recessive nerve sensitivity character belonged. Eight out of 16 phenotypes were examined for their nerve sensitivity to DDT, because the data with these eight phenotypes were sufficient to make the proposed analyses. Table I1

M Response of the nerve of the housefly t o 1.7 x DDT in different genetic make-ups of chromosomes Phenotype (2 : 3 : 5 : 6) +:

+:

+:+

+: +:ar:ac + : ocra: + : ac + : ocra : ar : + bwb: +: +:ac bwb: +:ar:+ bwb : ocra : + : + bwb : ocra : ar : ac

Exp. 1

Exp.2

87.5 81.2 87.5 87.5 18.7 25.0 37.5 13.3

75 .O 58.3 79.1 45.8 12.5 8.3 16.6 12.5

Data are given in percentages of the houseflies that respond to DDT by an increase in discharge frequency from the motoneurons innervating the femur. The houseflies are obtained by backcross, R (bwb : ocra : ar : ac) x F1 [R(bwb : m a : ar : ac) ? x Lab dl d. (From Tsukamoto et al., 1965.)

Table I1 gives the results of experiments with these eight phenotypes. The data are expressed as the values in the percentage of the houseflies whose nerves are stimulated in response t o 1.7 x M DDT. These percentage values were transformed into the arc-sine unit, and the homozygous effect of each chromosomal factor o n inheritance of low nerve sensitivity was calculated by the partial factorial analysis. The result clearly shows that the recessive gene on the second chromosome is responsible for the low nerve sensitivity to DDT in the resistant strain, and the contribution of the fifth chromosomal factor to the nerve sensitivity is very small.

72

T. NARAHASHI

In view of the evidence that the nerve of the housefly can detoxify DDT (Miyake et aZ., 1957), it is tempting t o ascribe the fifth chromosomal factor described above to DDT detoxication in the nerve. However, it should be noted that we are here dealing with recessive genes. The gene on the fifth chromosome that controls DDT dehydrochlorination is a dominant one (Tsukamoto and Suzuki, 1964). Therefore, the recessive gene on the fifth chromosome is controlling the nerve sensitivity through some other mechanism or through the detoxication of DDT other than dehydrochlorination. The recessive gene on the second chromosome controls the nerve sensitivity to DDT. There are at least two possible mechanisms whereby the nerve exhibits low sensitivity to DDT, i.e. (1) low permeability of DDT through the nerve sheath, and (2) low sensitivity of the nerve excitable membrane to DDT. The present experiment does not distinguish these two possibilities, and this problem remains to be explored. Preliminary experiments with lindane-resistant houseflies show that the gene controlling low nerve sensitivity t o lindane is located neither on the second nor on the fifth chromosome (Tsukamoto et al., 1965). VIII. STRUCTURE-ACTIVITY RELATIONSHIP

A number of experiments have been carried out in an attempt to find out the structure-activity relationship of various insecticides. Most of the experiments were based on the observation of insecticidal activities using a wide variety of derivatives and analogs of any particular parent compound. Much progress has indeed been made in terms of creations of new compounds of potential use. Many insecticides currently developed emerged as a result of such broad searches of compounds. However, it should be emphasized that our knowledge on the structure-activity relationship remains poor despite an enormous amount of efforts so far made. This is at least in part due to the fact that the insecticidal activity is a v e complex ~ chain of various reactions. Figure 1 clearly shows the situation. Because we are dealing with the insecticide molecule on the one hand, it is almost impossible to relate the chemical structure to the complicated chain of reactions in the insect body. It is absolutely necessary to dissociate the whole reaction that leads to the death of the poisoned insect into each component such as the penetration through the

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

73

cuticule, the activation, the detoxication, the action on the nerve, etc. To elucidate the structure-activity relation for the primary toxic action for many of the insecticides, one would naturally be forced to compare the potency of action of various derivatives and analogs on the nervous function both qualitatively and quantitatively. In case the target site is known to be an enzyme system, the study will be relatively easy at least from a technical point of view, because one can perform in vitro experiments on that particular enzyme. This is true for ChE’s which ,are the target site of a number of organophosphorus and carbamate insecticides. For some other insecticides such as chlorinated hydrocarbons and pyrethroids, the study of structure-activity relationship involves a very time-consuming comparison of relative effectiveness of various derivatives on the nervous function. For this reason, not much progress has so far been made along this line of approaches. A. DDT

Preliminary electrophysiological experiments were performed using the sensory nerve of the cockroach leg and the giant axon of the crayfish. Since initiation of repetitive discharges and increase in negative after-potential are two major actions of DDT on the nervous function, the relative potency of a number of derivatives and metabolites of p,p’-DDT (I) in exerting these two effects was compared (Yamada and Narahashi, 1968).

(I)

Detailed results will be described elsewhere, and only a few points will be mentioned here. Amino substitute (11) is not effective on the nerve in agreement with the absence of insecticidal activity (Metcalf and Fukuto, 1968). However, nitro substitute (111) is effective in initiating trains of impulses in the sensory nerve of the cockroach leg and in increasing the negative after-potential in the crayfish giant axon, despite the lack of insecticidal activity (Metcalf and Fukuto, 1968; Holan,

74

T. NARAHASHI

Methyl substitute (IV) has an insecticidal activity, especially for mosquitoes (Metcalf and Fukuto, 1968). It is effective on the sensory nerve in producing trains of impulses, but has no effect on the negative after-potential of the crayfish axon.

(Is9

Substitution of chlorines at para positions by methoxy (-OCH3) (compound V) or ethoxy (-OC2 H, ) (VI) group still maintains the effectiveness on the sensory nerve.

However, although methoxy substitute is capable of increasing the negative after-potential of the crayfish axon, ethoxy substitute lacks this action. Both substitutes are insecticidally active (Metcalf and Fukuto, 1968). When the p,p’-substituents are increased in size (e.g. -OC4H,), the compound becomes inert on both types of nerve preparations. Metabolities of p,p’-DDT show an interesting spectrum of action on the nerve. Although o,p‘-DDT (VII) and p,p’-DDD (VIII) are effective in producing trains of impulses, they are ineffective in augmenting the negative after-potential. However, o,p’-DDD (IX) is somewhat effective in both respects.

(mI)

(nm)

(Ix)

The absence of the effect of p,p’-DDD on the negative after-potential was confirmed by Van den Bercken (1 969) using single nodes of Ranvier of Xenopus Zuevis. He also found that DDD suppressed the action potential. This action was never observed with p,p’-DDT. Dehydrochlorinated metabolite p,p’-DDE (X) and acid form metabolite p,p’-DDA (XI) have no effect on both nerve preparations.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

75

This agrees with the general observation that they lack the insecticidal activity. CI O

f

O

C

J

CI

OF-@ COOH

CCI,

(XI)

(XI

These observations on the effectiveness on the nervous function may be interpreted in terms of (1) the profile of formal charges on the chloroform group, (2) the steric factor, and (3) the hydrophobicity. Definite conclusion about the structure-activity relation awaits further experimental analyses. One important point emerging from these preliminary observations is that when the direct effect of insecticides on the nerve is examined as a measure of activity, there are qualitative as well as quantitative differences among derivatives or metabolites having very similar structures. These differences may not be elucidated if the insecticidal activity is simply compared. B. PYRETHROIDS

Several pyrethroid derivatives of a new type were compared for their insecticidal activity, mammalian toxicity, and ability to affect the nervous function (Berteau et al., 1968). Allethrin (XII) and four other pyrethroid-like compounds were used (Table 111). Compound XI11 is the keton analog of allethrin and lacks the ester function. In compound XIV, chrysanthemum-monocarboxylic acid of allethrin is replaced by tetramethylcyclopropanecarboxylicacid, and allethrolone by 5-benzyl-3-furylmethanol. In compound XV, tetrame thylcyclopropanecarboxylic acid of compound XIV is further replaced by a carbamate, tetramethylaziridine carboxylic acid. In compound XVI, the cyclopropane of compound XIV is substituted by N,N-diisopropylcarbamicacid. Some of the data on insecticidal activity and nerve activity are given in Table 111. It is clear that both activities run parallel with each other, and that all of the compounds tested do not loose the activities by changes in chemical structure described above. It is noteworthy that all of the compounds exert very similar actions on the crayfish giant axon, i.e. (1) slight and progressive depolarization, (2) increase in negative after-potential, and (3) repetitive afterdischarges by a single stimulus. Our recent voltage clamp

76

T. NARAHASHI

Table I11 Chemical structures and biological activities of allethrin and four structurally related compounds Compound

Structure

XI1 Allethrin

21

XI11

171

XIV

xv XVI

Toxicity to housefly LD5 0 (mg/kg)

0.9

Jm

Potency on nerve ED50 (PM 1

2.6 17 1.6

228

24

750

130

Nerve potency: micromolar level to decrease the maximum rate of rise of the action potential of crayfish giant axons to 50%normal. (Berteau etul., 1968.)

experiments with the crayfish giant axon show that compound XI11 exerts the same effects on membrane conductances as allethrin (compound XII), i.e. the sodium and potassium conductance increases are suppressed, and the sodium inactivation is greatly slowed. These results are in a way contradictory to the classical concept concerning the structure-activity relation of pyrethroids. It has been believed that the cyclopropane ring and the ester function are essential for the insecticidal activity. The results described here rather suggest that the configuration of the molecule, relative to appropriate size and shape to interact with the receptor of the nerve membrane, appears to be of critical importance in exerting the nerve action. The receptor for pyrethroids can be visualized as specific group(s) of macromolecules in the nerve membrane such as proteins and phospholipids which control the gate mechanism involved in conductance changes. Further experimental analyses, especially those by means of voltage clamp techniques, are necessary to explore the structure-activity relationship of pyrethroid-like compounds. C. ROTENONE

Rotenone (XVII) inhibits the electron transfer from DPNH to

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

77

cytochrome b and blocks the nerve conduction as described earlier (Section IV E). cH2%-C cH3’

-CHI

A% OCH,

@33

mm)

Thirty-four derivatives of rotenone were examined for their insecticidal activity, and their potency in inhibiting the glutamic dehydrogenase activity. Some of them were also examined for their potency to block the nerve conduction (Fukami et al., 1959). In short, the three activities go parallel with each other for most of the derivatives, and those which have a strong insecticidal activity are, without exception, capable of inhibiting the enzyme activity and blocking the nerve conduction. Some of them, however, inhibit the enzyme activity and block the nerve conduction effectively, yet lack strong insecticidal activity. For example, rotenone hydrochloride (XVIII) is almost equipotent to rotenone for the enzyme inhibition and only slightly less active on the nerve, but it possesses only 20% insecticidal activity of rotenone. This may be due to rapid degradation of rotenone hydrochloride in the insect.

~xszllt)

See XVII for the rest of structure,

Detailed results will not be repeated here. The results confirm the earlier suggestion by Martin (1 946) that the asymmetric carbons at positions 7 and 8 are of critical importance in maintaining the activities. The presence of the chromanochromanone ring in the molecule is not essential, and the chromanochromanol ring can substitute for it. This is shown by experiments with rotenol (XIX), dihydro-rotenol (XX), and acetylrotenone (XXI), all of which are potent in exerting the three actions.

78

T. NARAHASHI

(XIXI

(xx)

(XXI)

See XVlI for the rest of structure.

IX. ROAD TO THE MOLECULAR MECHANISMS

Little has been known concerning the molecular mechanisms of action of insecticides. The fact that most insecticides interact with the nerve membrane makes the direct in vitro study of this problem very difficult. For example, it is first of all difficult to isolate the pure nerve membrane component without being contaminated by the components of other membranes such as those of Schwann cells and connective tissues. Even when this is accomplished satisfactorily, one will have to demonstrate that the interaction of insecticides with the isolated nerve membrane component is the same as that occumng in the nerve membrane in situ. It is also absolutely necessary to demonstrate that the interaction between the insecticides and the nerve membrane is directly related to the toxic action. In view of these considerations, there will be no single approach whereby one can obtain a clean-cut answer to this problem. The approach will have t o be multidisciplinary in nature. Classical electrophysiological techniques such as those by voltage clamp and microelectrodes will continue to be very useful and powerful in measuring the nerve activity in terms of membrane ionic conductances and membrane potential changes. These parameters, especially conductance changes, will provide us with the basis t o explore the molecular mechanisms involved. Attempts were made to isolate receptors for the insecticide action on the nerve membrane. This will give us the chemical basis of interpretation of the mode of action. DDT and dieldrin bind with various components of the nerve (Matsumura and Hayashi, 1966a, b, 1969; Hayashi and Matsumura, 1967; O’Brien and Matsumura, 1964; Matsumura and O’Brien, 1966a, b; Hatanaka et al., 1967; Brunnert and Matsumura, 1969). However, the role of such bindings in the toxic action of insecticides on the nerve remains to be explored.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

79

Artificial membrane systems, either the monomolecular film spread over the aqueous phase or the bimolecular membrane formed between two aqueous phases, will also be highly useful for elucidating the mechanism of action of insecticides at the molecular level. One of the advantages in this type' of experiment is that the artificial membrane is chemically defined. Experiments along this line have already been started. The potassium conductance of the lecithindecane membrane is increased by application of valinomycin, and DDT partially antagonizes the valinomycin action and decreases the potassium conductance (Hilton and O'Brien, 1970). The effect of DDT on the membrane potassium conductance is, on the surface, at least in the same direction as that found in the lobster axon membrane (Narahashi and Haas, 1967, 1968). However, the potassium conductance of the natural nerve membrane and that produced by valinomycin in the artificial membrane are not necessarily the same in every respect. Despite this, the experiments with artificial membranes will provide us with a clue to approach the molecular mechanism of action of insecticides. In addition to the mechanism of action of insecticides at the membrane or molecular level, electrophysiological techniques will be extremely useful for studies of other aspects as has been described in this article. Among many possible applications is the study of the structure-activity relationship of insecticides. It should be emphasized that the knowledge of the structure-activity relation for any particular type of insecticides is useful not only for creation of new insecticides but also for interpretation of insecticide-receptor interactions at the molecular level. In this connection, the hypotheses put forward by Mullins (1 954, 1955, 1956) and by Holan (1968) in an attempt to explain the structure-activity relationship of DDT and its analogs are worth while to note. Mullins (1954, 1955, 1956) proposed a model in which the molecules of DDT and its analogs must fit into an interspace formed by membrane macromolecules. For example, iodo-DDT in which two chlorine atoms on the phenyl rings are substituted by two iodine atoms does not fit because the p,p'-substituents are too large, and in fact it is ineffective as the insecticide. In DDE, the tetrahedral bond angle is changed and causes non-fit. Holan (1969) modified the Mullins' original hypothesis to explain the structure-activity relationship of new DDT analogs, 1,-l-di(p-chlorophenyl)-2,2-dichlorocyclopropane and its derivatives. Part of the insecticide molecule containing the phenyl rings locks itself into the overlaying protein

80

T. NARAHASHI

layer in the nerve membrane by forming a molecular complex with it. An attempt is made t o explain the prolongation of sodium current while the whole insecticide molecule is locked in the membrane. Projection of the van der Waals outline of the active insecticides can adequately explain the fit of all of the active compounds into the membrane. ACKNOWLEDGEMENTS

Part of the results described in the present article was supported by a grant from the National Institute of Health (NS 068SS), and by a contract with the National Institute of Environmental Health Sciences (PH-43-68-73). I wish t o thank Mrs. R. M. Crutchfield and Mrs. C. A. Munday for their secretarial assistance.

REFERENCES Abedi, Z. H., Duffy, J. R. and Brown, A. W. A. (1963). Dehydrochlorination and DDT-resistance in Aedes aegypti. J. e c o n Ent. 56, 51 1-517. Agosin, M., Michaeli, D., Miskus, R., Nagasawa, S. and Hoskins, W. M. (1961). A new DDT-metabolizing enzyme in the German cockroach. J. econ. Ent. 54,340-342. Armstrong, C. M. and Binstock, L. (1965). Anomalous rectification in the squid giant axon injected with tetraethylammonium chloride. J. gen. Physiol. 48, 859-872. Babers, F. H. and Pratt, J. J., Jr. (1953). Resistance of insects to insecticides: the metabolism of injected DDT. J. econ. Ent. 46, 977-982. Baker, P. F., Hodgkin, A. L. and Shaw, T. I. (1961). Replacement of the protoplasm of a giant nerve fibre with artificial solutions. Nature, Lond. 190,885-887. Baker, P. F., Hodgkin, A. L. and Shaw, T. I. (1962a). Replacement of the axoplasm of giant nerve fibres with artificial solutions. J. Physiol. 164, 330-354. Baker, P. F., Hodgkin, A. L. and Shaw, T. I. (1962b). The effects of changes in internal ionic concentrations on the electrical properties of perfused giant axons. J. Physiol. 164, 355-374. Barker, R. J. (1957). Some effects of temperature on adult house flies treated with DDT. J. e c o n Ent. 50,446-450. Becht, G . (1958). Influence of DDT and lindane on chordotonal organs in the cockroach. Nature, Lond. 181, 777-779. Berteau, P. E., Casida, J. E. and Narahashi, T. (1968). Pyrethroid-like biological activity of compounds lacking cyclopropane and ester groupings. Science, N . Y . 161, 1151-1153.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

81

Blum, M. S. and Kearns, C. W. (1956). Temperature and the action of pyrethrum in the American cockroach. J. econ. Ent. 49, 862-865. Bodenstein, D. (1946). Investigation on the locus of action of DDT in flies (Drosophila).Biol. Bull. 90, 148-157. Brown, A. W. A. (1951). “Insect Control by Chemicals”, 817 pp. John Wiley & Sons, Inc., New York. Brown, A. W. A. (1960). Mechanisms of resistance against insecticides. A . Rev. Ent. 5, 301-326. Brown, A. W. A. (1961). The challenge of insecticide resistance. Bull. ent. SOC. A m . 7, 6-19. Brown, A. W. A. (1964). Animals in toxic environments: resistance of insects to insecticides. In “Handbook of Physiology, Section 4, Adaptation t o the Environment” (D. B. Dill, E. F. Adolph and C. G. Wilber, eds), pp. 773-793. Am. Physiol. SOC.,Washington, D.C. Browne, L. B. and Kerr, R. W. (1967). The response of the labellar taste receptors of DDT-resistant and non-resistant houseflies (Musca domestica). Ent. Exp. Appl. 10,337-346. Brunnert, H. and Matsumura, F. (1 969). Binding of 1,l,l-trichlore2,2-dip-chlorophenylethane (DDT) with subcellular fractions of rat brain. Biochem. J. 114, 135-139. Bull, D. L. and Adkisson, P. L. (1963). Absorption and metabolism of C 14-labeled DDT by DDT-susceptible and DDT-resistant pink bollworm adults. J. econ. Ent. 56, 641-643. Callec, J. and Boistel, J. (1967). Les effets de l’acktylcholine aux niveaux synaptique et somatique dans le cas du dernier ganglion abdominal de la Blatte Periplaneta americana. C.r. Sbanc. SOC.Biol. 161,442-446. Casida, J. E. (1963). Mode of action of carbamates. A . Rev. Ent. 8, 39-58. Chamberlain, R. W. (1950). An investigation on the action of piperonyl butoxide with pyrethrum. A m . J. Hyg. 52, 153-183. Colhoun, E. H. (1960). Approaches to mechanisms of insecticidal action. J. agric. Fd Chem. 8, 252-257. Colhoun, E. H. ( 1963). The physiological significance of acetylcholine in insects and observations upon other pharmacologically active substances. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 1, pp. 1-46. Academic Press, London and New York. Dahm, P. A. (1957). The mode of action of insecticides exclusive of organic phosphorus compounds. A . Rev. Ent. 2, 247-260. Dahm, P. A. and Nakatsugawa, T. (1968). Detoxication of pesticides and the mechanism of synergism. In “Enzymatic Oxidations of Toxicants” (E. Hodgson, ed.), pp. 89-1 12. North Carolina State University at Raleigh. Dallemagne, M. J. and Philippot, E. (1948). Recherches sur la toxicitk de l’hexachlorocyclohexane. Arch. int. Pharmacodyn. Thbr. 76,274-296. Davson, H. (1964). “A Textbook of General Physiology”. Little, Brown and Company, Boston. Dresden, D. (1 949). “Physiological Investigations into the Action of DDT”, 114 pp. Drukkerij en Uitgeverij G. W. van der Wiel and Co. Arnheim, Netherlands. Dustan, G. G. (1947). Effect of temperature on toxicity of DDT. Can Ent. 79, 1-4. .

__

82

T. NARAHASHI

Eaton, J. L. and Sternburg, J. (1964). Temperature and the action of DDT on the nervous system of Periplaneta americana (L.). J. Insect Physiol. 10, 4 7 1-485. Eaton, J. L. and Sternburg, J. G. (1967). Temperature effects on nerve activity in DDT-treated American cockroaches. J. econ. Ent. 60, 1358-1364. Eccles, J. C. (1964). “The Physiology of Synapses”, 328 pp. Springer-Verlag, Berlin, Gottingen, Heidelberg. Faeder, I. R. (1968). “Neuromuscular Transmission in Insects”. Thesis, Cornell University, Ithaca, N. Y. Fan, H. Y., Cheng, T. H. and Richards, A. G. (1948). The temperature coefficients of DDT action in insects. Physiol. Zool. 21, 48-59. Frankenhaeuser, B. and Hodgkin, A. L. (1956). The aftereffects of impulses in the giant nerve fibres of Loligo. J. Physiol. 131, 341-376. Frazier, D. T., Narahashi, T. and Moore, J. W. (1969). Hemicholinium-3: Non-cholinergic effects on squid axons. Science, N. Y. 163,820-82 1. Fritsch, H. (1952). Versuche zur Analyse des Angriffspunktes kontactinsektizider Stoffe. Biol. Zbl. 71, 5 12-528. Fritsch, H. and Krupp, H. (1952). Wirkung von Insektiziden auf ein isoliertes Ganglien-Muskel-Praparat von Dytiscus marginalis (Gelbrand-kafer). Arch. exp. Path. Pharmak. 214,227-241. Fukami, J. (1954). Effects of rotenone on succinoxidase system in the muscle of the cockroach. Jap. J. appl. Zool. 19, 29-37. Fukami, J. (1956). Effects of some insecticides on the respiration of insect organs, with special reference to the effects of rotenone. Botyu-Kagaku 21, 122-128. Fukami, J. (1961). Effect of rotenone on respiratory enzyme system of insect muscle. Bull. Nut. Znst. Agric. Sci. (Ser. C) No. 13, 33-45. Fukami, J. and Tomizawa, C. (1956). Effects of rotenone on the 1-glutamic oxidase system in the insect. Botyu-Kaguku 21, 129-133. Fukami, J. and Tomizawa, C. (1958a). Effect of rotenone and its derivatives on the glutamic dehydrogenase in insects. Botyu-Kagaku 23, 1-4. Fukami, J. and Tomizawa, C. (1958b). The effects of rotenone and its derivatives on the respiration of brain in guinea pig. Botyu-Kaguku 23, 205-208. Fukami, J., Nakatsugawa, T. and Narahashi, T. (1959). The relation between chemical structure and toxicity in rotenone derivatives. Jap. J. appl. Ent. ZOO^. 3,259-265. Fukuto, T. R. (1961). The chemistry of organic insecticides. A . Rev. Ent. 6, 3 13-332. Fullmer, 0 . H. and Hoskins, W. M. (1951). Effects of DDT upon the respiration of susceptible and resistant house flies. J. econ. Ent. 44,858-870. Furshpan, E. J. (1 964). “Electrical transmission” at an excitatory synapse in a vertebrate brain. Science, N. Y. 144, 878-880. Furshpan, E. J. and Potter, D. D. (1959). Transmission at the giant motor synapses of the crayfish. J. Physiol. 145, 289-325. Furukawa, T. (1966). Synaptic interaction at the Mauthner cell of goldfish. In “Progress in Brain Research” (T. Tokizane and J. P. SchadB, eds), Vol. 2 1A, pp. 44-70. Elsevier Publishing Co., Amsterdam. Furukawa, T. and Furshpan, E. J. (1963). Two inhibitory mechanisms in the Mauthner neurons of goldfish. J. Neurophysiol. 26, 140-176.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

83

Gasser, H. S. (1941). The classification of nerve fibers. Ohio J. Sci. 41, 145-159. Gasser, H. S. and Grundfest, H. (1 936). Action and excitability in mammalian A fibers. A m . J. Physiol. 117, 113-133. Gianotti, O., Metcalf, R. L. and March, R. B. (1956). The mode of action of aldrin and dieldrin in Periplaneta americana (L.). Ann. ent. SOC.A m . 49, 5 88-5 92. Gordon, H. T. (1961). Nutritional factors in insect resistance to chemicals. A . Rev. Ent. 6, 27-54. Gordon, H. T. and Welsh, J. H. (1 948). The role of ions in axon surface reactions to toxic organic compounds. J. cell. comp. Physiol. 31, 395-419. Graham, H. T. (1930). Modification of the response of nerve by veratrine and by narcotics. J. Pharmac. exp. Ther. 39,268-269. Graham, H. T. and Gasser, H. S. (1931). Modification of nerve response by veratrine, protoveratrine and aconitine. J. Pharmac. exp. Ther. 43, 163-185. Guthrie, F. E. (1950). Effect of temperature on toxicity of certain organic insecticides. J. econ. Ent. 43, 559-560. Hafliger, E. (1948). Der Einfluss der Temperatur auf die Giftwirkung des DDT bei Honigbienen (Apis mellifica L.) Experientia 4, 223-225. Harlow, P. A. (1958). The action of drugs on the nervous system of the locust (Locusta migratoria). Ann. appl. Biol. 46, 55-73. Harries, F. H., DeCoursey, J. D. and Hofmaster, R. N. (1 945). Some factors affecting the insecticidal action of pyrethrum extracts on the beet leafhopper. J. agric. Res. 71, 553-565. Harrison, C. M. (1 95 1). Inheritance of resistance to DDT in the housefly, Musca domestica L. Nature, Lond. 167, 855-856. Hartzell, A. and Wilcoxon, F. (1932). Some factors affecting the efficiency of contact insecticides. 11. Chemical and toxicological studies of pyrethrum. Contr. Boyce Thompson Znst. PI. Res. 4, 107-1 17. Harvey, G. T. and Brown, A. W. A. (195 1). The effect of insecticides on the rate of oxygen consumption in Blattella. Can J. Zool. 29,42-53. Hatai, N. (1961). Effects of pyrethrin on the heart preparation from the silkworm larva. J. Coll. Agric. T o k y o Imp. Univ. 15, 69-81. Hatanaka, A., Hilton, B. D. and O’Brien, R. D. (1967). The apparent binding of DDT t o tissue components. J. Agric. Fd Chem. 15, 854-857. Hawkins, W. B. and Sternburg, J. (1964). Some chemical characteristics of a DDT-induced neuroactive substance from cockroaches and crayfish. J. econ. Ent. 57, 241-247. Hayashi, M. and Matsumura, F. (1967). Interactions of DDT with the nervous system of the resistant and susceptible German cockroaches. Nature, Lond. 215,1510-1512. Hayes, W. J., Jr. (1959). Pharmacology and toxicology of DDT. In “DDT. The Insecticide Dichlorodiphenyltrichloroethane and its Significance” (P. Muller, ed.), Vol. 11, pp. 11-247. Birkhauser Verlag, Basel. Heslop, J. P. and Ray, J. W. (1959). The reaction. of the cockroachPeriplaneta americana L. t o bodily stress and DDT. J. Insect Physiol. 3, 395-401. Hille, B. (1968). Pharmacological modifications of the sodium channels of frog nerve.J. gen. Physiol. 51, 199-219. Hilton, B. D. and O’Brien, R. D. (1970). Antagonism by DDT of the effect of valinomycin on a synthetic membrane. Science, N . Y. 168, 841-843.

84

T. NARAHASHI

Hodgkin, A. L. (1958). Ionic movements and electrical activity in giant nerve fibres. Proc. R . SOC. B 148, 1-37. Hodgkin, A. L. (1964). “The Conduction of the Nervous Impulse”. Liverpool Univ. Press, Liverpool. Hodgkin, A. L. and Huxley, A. F. (1952a). Currents carried by sodium and potassium ions through the membrane of the giant axon of Loligo. J. Physiol. 116,449-472. Hodgkin, A. L. and Huxley, A. F. (1952b). The components of membrane conductance in the giant axon of Loligo. J. Physiol. 116, 473-496. Hodgkin, A. L. and Huxley, A. F. ( 1952~ ) .The dual effect of membrane potential on sodium conductance in the giant axon of Loligo. J. Physiol. 116,497-506. Hodgkin, A. L. and Huxley, A. F. (1952d). A quantitative description of membrane current and its application t o conduction and excitation in nerve. J. Physiol. 117, 500-544. Hodgkin, A. L. and Keynes, R. D. (1955). Active transport of cations in giant axons from Sepia and Loligo. J . Physiol. 128, 28-60. Hodgkin, A. L., Huxley, A. F. and Katz, B. (1952). Measurement of current-voltage relations in the membrane of the giant axon of Loligo. J. Physiol. 116,424-448. Hoffman, R. A. and Lindquist, A. W. (1949). Effect of temperature on knockdown and mortality of house flies exposed to residues of several chlorinated hydrocarbon insecticides. J. econ. Ent. 42, 89 1-893. Hoffman, R. A., Roth, A. R. and Lindquist, A. W. (1949). Effect of air temperature on the insecticidal action of some compounds on the sheep tick and on migration of sheep tick on the animal. J. econ. Ent. 42, 893-896. Holan, G. (1969). New halocyclopropane insecticides and the mode of action of DDT. Nature, Lond. 221, 1025-1029. Hoskins, W. M. and Gordon, H. T. ( 1956). Arthropod resistance to chemicals. A. Rev. Ent. 1, 89-122. Hurst, H. (1949). Reversible action of DDT. Nature, Lond. 163, 286287. Ito, M., Kostyuk, P. G. and Oshima, T. (1962). Further study on anion permeability of inhibitory post-synaptic membrane of cat motoneurones. J. Physiol. 164, 150-156. Julian, F. J., Moore, J. W. and Goldman, D. E. (1962a). Membrane potentials of the lobster giant axon obtained by use of the sucrose-gap technique. J. gen. Physiol. 45, 1195-12 16. Julian, F. J., Moore, J. W. and Goldman, D. E. (1962b). Current-voltage relations in the lobster giant axon membrane under voltage clamp conditions. J. Gen. Physiol. 45, 1217-1238. Kaeser, W. ( 1948). Zur Frage einer temperaturebedingten Wiederstandsftihigkeit der Honigbiene (Apis mellifica L.) gegeniiber dem Kontaktinsektizid DDT (Gesarol). Anz. Schudlingsk. 21, 129-132. Kao, C. Y.(1966). Tetrodotoxin, saxitoxin and their significance in the study of excitation phenomena. Pharmacol. Rev. 18, 997- 1049. Katz, B. (1962). The tranmission of impulses from nerve t o muscle, and the subcellular unit of synaptic action. Proc. R . SOC. B 155, 455-477. Katz, B. (1966). “Nerve, Muscle and Synapses”. McGraw Hill, New York.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

85

Kearns, C. W. (1956). The mode of action of insecticides. A Rev. Ent. 1, 123-148. Kerkut, G. A. and Thomas, R. C. (1964). The effect of anion injection and changes in the external potassium and chloride concentration on the reversal potentials of the IPSP and acetylcholine. Comp. Biochem. Physiol. 11, 199-213. Kerkut, G. A., Leake, L. D., Shapira, A., Cowan, S. and Walker, R. J. (1965). The presence of glutamate in nerve-muscle perfusates of Helix, Carcinus and Periplaneta. Comp. Biochem. Physiol. 15, 485-502. 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. Keynes, R. D. ( 195 1). The ionic movements during nervous activity. J. Physiol. 114, 119-150. Keynes, R. D. and Lewis, P. R. (1951). The sodium and potassium content of cephalopod nerve fibres. J. Physiol. 114, 151-182. Khan, M. A. Q. and Brown, A. W. A. (1966). Lipids and dieldrin resistance in Aedesaegypti. J. econ. Ent. 59, 1512-1514. Klein, A. K., Link, J. D. and Ives, N. F. (1968). Isolation and purification of metabolites in the urine of male rats fed aldrin and dieldrin. J. Ass. off: Analyt. C h e m 51, 895-898. Krijgsman, B. J., Dresden, D. and Berger, N. E. (1 950). The action of rotenone and tetraethyl pyrophosphate on the isolated heart of the cockroach. Bull enf. Res. 41, 141-15 1. Lalonde, D. I. V. and Brown, A. W. A. (1 954). The effect of insecticides on the action potentials of insect nerve. Can. J. Zool. 32, 74-81. Lehmann, J. E. (1937). The effect of changes in pH on the action of mammalian A nerve fibers. A m . J. Physiol. 118, 600-612. Leski, R. A. and Cutkomp, L. K. (1962). The influence of parathion and para-oxon on sensory hairs of flies. J. econ. Ent. 55, 281-285. Lindahl, P. E. and Oberg, K. E. (1961). The effect of rotenone on respiration and its point of attack. Expl. Cell Res. 23, 228-237. Lindquist, A. W., Madden, A. H. and Schroeder, H. 0. (1946). Effect of temperature on knockdown and kill of mosquitoes and bedbugs exposed to DDT. Kans. Ent. Soc. J. 19, 13-15. Lindquist, A. W., Wilson, H. G., Schroeder, H. 0. and Madden, A. H. (1945). Effect of temperature on knockdown and kill of houseflies exposed to DDT. J. econ. Ent. 38,261-264. Lipke, H. and Kearns, C. W. (1960). DDT-dehydrochlorinase. In “Advances in Pest Control Research” (R. L. Metcalf, ed.), Vol. 111, pp. 253-287. Interscience Publ., Inc., New York and London. Lowenstein, 0. (1942). A method of physiological assay of pyrethrum extract. Nature, Lond. 150, 760-762. March, R. B. (1958). The chemistry and action of acaricides. A . Rev. Ent. 3, 355-376. Martin, H. ( 1946). Insecticides: Chemical constitution and toxicity. J. SOC. Chem. Znd., Lond. 65, 402-405. Matsumura, F. and Hayashi, M. (1966a). Dieldrin: Interaction with nerve components of cockroaches. Science, N. Y. 153, 757-759.

86

T. NARAHASHI

Matsumura, F. and Hayashi, M. (1966b). Interaction of dieldrin with the subcellular components of both resistant and susceptible strains of Aedes aegypti L. Mosq. News 26, 19G194. Matsumura, F. and Hayashi, M. (1969). Dieldrin resistance. Biochemical mechanisms in the German cockroach. J. Agric. Fd Chem. 17,231-235. Matsumura, F. and O’Brien, R. D. (1966a). Absorption and binding of DDT by the central nervous system of the American cockroach. J. Agric. Fd Chem. 14, 36-39. Matsumura, F. and O’Brien, R. D. (1966b). Interactions of DDT with components of American cockroach nerve. J. Agric. Fd Chem. 14,39-43. Matthews, H. B. and Matsumura, F. (1969). Metabolic fate of dieldrin in the rat. J. Agric. Fd Chem. 17,845-852. Menn, J. J., Benjamini, E. and Hoskins, W. M. (1957). The effects of temperature and stage of life cycle upon the toxicity and metabolism of DDT in the house fly. J. econ. Ent. 50, 67-74. Metcalf, R. L. (1955). “Organic Insecticides. Their Chemistry and Mode of Action”. Interscience Publ. Ltd., New York and London. Metcalf, R. L. (1967). Mode of action of insecticide synergists. A . Rev. Ent. 12, 229-256. Metcalf, R. L. (1968). The role of oxidative reactions in the mode of action of insecticides. In “Enzymatic Oxidations of Toxicants” (E. Hodgson, ed.), pp. 15 1-173. North Carolina State University at Raleigh. Metcalf, R. L. and Fukuto, T. R. (1 968). The comparative toxicity of DDT and analogues to susceptible and resistant houseflies and mosquitoes. Bull. Wd Hlth Org. 38, 633-647. Milani, R. (1954). The genetics of the house fly. Preliminary note. Caryologia (SUPPI.),pp. 791-796. Milani, R. and Travaglino, A. (1957). Ricerche genetiche sull a resistenza al DDT in MUSCQdomestica concatenazione del gene kdr (knockdown-resistance) con due mutanti morfologigi. Riv. Parassitol. 18, 199-202. Miller, S. and Perry, A. S. (1964). Separation and purification of DDT-degrading enzymes from the human body louse. J. Agric. Fd Chem. 12, 167-169. Miyake, S. S., Kearns, C. W. and Lipke, H. (1957). Distribution of DDT-dehydrochlorinase in various tissues of DDT-resistant house flies. J. econ. Ent. 50, 359-360. Moore, J. W., Narahashi, T. and Ulbricht, W. (1964a). Sodium conductance shift in an axon internally perfused with a sucrose and low potassium solution. J. Physiol. 172, 163-173. Moore, J. W., Ulbricht, W. and Takata, M. (1964b). Effect of ethanol on the sodium and potassium conductances of the squid axon membrane. J. gen. Physiol. 48, 279-295. Moore, J. W., Blaustein, M. P., Anderson, N. C. and Narahashi, T. (1967). Basis of tetrodotoxin’s selectivity in blockage of squid axons. J. gen. Physiol. 50, 1401-141 1. Mullins, L. J. (1954). Some physical mechanisms in narcosis. Chem. Rev. 54, 2 89-32 3. Mullins, L. J. ( 1955). Structure-toxicity in hexachlorocyclohexane isomers. Science, N . Y. 122, 118-1 19.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

87

Mullins, L. J. (1956). The structure of nerve cell membranes. Zn “Molecular Structure and Functional Activity of Nerve Cells”. Publ. No. 1, pp. 123-124. Am. Inst. Biol. Sci., Washington, D.C. Munson, S. C. (1953). Some effects of storage at different temperatures on the lipids of the American roach and on the resistance to this insect to heat. J. econ. Ent. 46, 657-666. Munson, S. C., Padilla, G. M. and Weissmann, M. L. (1954). Insect lipids and insecticidal action. J. econ. Ent. 47, 578-587. Nagasawa, S. and Hoskins] W. M. (1962). The relation between temperature and lethal action of p,p -DDT on adults of the housefly, Musca domestica L. Ent. Exp. Appl. 5, 139-146. Narahashi, T. ( 1962a). Effect of the insecticide allethrin on membrane potentials of cockroach giant axons. J. cell. comp. Physiol. 59, 61-65. Narahashi, T. (1962b). Nature of the negative after-potential increased by the insecticide allethrin in cockroach giant axons. J. cell. comp. Physiol. 59, 67-76. Narahashi, T. (1963a). The properties of insect axons. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. I, pp. 175-256. Academic Press, London and New York. Narahashi, T. (1963b). Dependence of resting and action potentials on internal potassium in perfused squid giant axons. J. Physiol. 169,91-115. Narahashi, T. ( 1964a). Insecticide resistance and nerve sensitivity. Jap. J. med. Sci. Biol. 17, 46-53. Narahashi, T. ( 1964b). Excitable membrane and calcium. Seibutsudutsuri (Biophysics) 4, 101-1 14. Narahashi, T. (1965a). The physiology of insect axons. Zn “The Physiology of the Insect Central Nervous System” (J. E. Treherne and J. W. L. Beament, eds), pp. 1-20. Academic Press, London and New York. Narahashi, T. (1965b). Mode of action of insecticides. In “Shin Noyaku Soseiho (Creation of New Insecticides)” (R. Yamamoto and T. Noguchi, eds), pp. 169-209. Nankodo, Tokyo. Narahashi, T. (1966). Mode of action of insecticides. Kagaku To Seibutsu (Chemistry and Biology) 4, 134- 140. Narahashi, T. and Anderson, N. C. (1967). Mechanism of excitation block by the insecticide allethrin applied externally and internally to squid giant axons. Toxic. appl. Pharmac. 10.529-547. Narahashi, T.;-Anderson, N. C. and Moore, J. W. (1967a). Comparison of tetrodotoxin and procaine in internally perfused squid giant axons. J. gen. Physiol. 50, 1413-1428. Narahashi, T. and Haas, H. G. (1967). DDT: Interaction with nerve membrane conductance changes. Science, N. Y. 157, 1438- 1440. Narahashi, T. and Haas, H. G. (1968). Interaction of DDT with the components of lobster nerve membrane conductance. J. gen. Physiol. 51, 177-198. Narahashi, T., Haas, H. G. and Therrien, E. F. (1967b). Saxitoxin and tetrodotoxin: Comparison of nerve blocking mechanism. Science, N. Y. 157, 1441-1442. Narahashi, T., Moore, J. W. and Shapiro, B. I. (1969a). Condylactis toxin:

88

T. NARAHASHI

Interaction with nerve membrane ionic conductances. Science, N . Y. 163, 680-68 1. Narahashi, T., Moore, J. W. and Poston, R. N. (1969b). Anesthetic blocking of nerve membrane conductances by internal and external applications. J. Neurobiol. 1, 3-22. Narahashi, T., Moore, J. W. and Frazier, D. T. ( 1 9 6 9 ~ ) .Dependence of tetrodotoxin blockage of nerve membrane conductance on external pH. J. Pharmacol. exp. Ther. 169, 224-228. Narahashi, T., Moore, J. W. and Scott, W. R. (1964). Tetrodotoxin blockage of sodium conductance increase in lobster giant axons. J. gen. Physiol. 47, 965-974. Narahashi, T. and Yamasaki, T. (1960a). Nervous and cholinesterase activities in the cockroach as affected by demeton and methyldemeton. Studies on the mechanism of action of insecticides (XVIII).Jap. J. appl. Ent. 2001. 4, 64-69. Narahashi, T. and Yamasaki, T. (1960b). Mechanism of the after-potential production in the giant axons of the cockroach. J. Physiol. 151, 75-88. Narahashi, T. and Yamasaki, T. ( 1 9 6 0 ~ ) .Mechanism of increase in negative after-potential by dicophanum (DDT) in the giant axons of the cockroach. J. Physiol. 152, 122-140. Narahashi, T. and Y amasaki, T. (1 960d). Behaviors of membrane potential in the cockroach giant axons poisoned by DDT. J. cell. comp. Physiol. 55, 131-142. Nastuk, W.L. ( 1966). Fundamental aspects of neuromuscular transmission. Ann. N . Y . Acad. Sci. 135, 110-135. Oberg, K. E. (1961). The site of the action of rotenone on the respiratory chain. Expl. Cell Res. 24, 163-164. O’Brien, R. D. (1966). Mode of action of insecticides. A . Rev. Ent. 11, 369402. O’Brien, R. D. ( 1967). “Insecticides. Action and Metabolism”. Academic Press, New York and London. O’Brien, R. D. and Matsumura, F. (1964). DDT: A new hypothesis of its mode of action. Science, N. Y . 146, 657-658. Oikawa, T., Spyropoulos, C. S., Tasaki, I. and Teorell, T. (1961). Methods for perfusing the giant axon of Loligo pealii. Acta Physiol. Scand. 52, 195-196. Orser, W. B. and Brown, A. W. A. (1951). The effect of insecticides on the heartbeat of Periplaneta. Can. J. 2001.29, 54-64. Perry, A. S. (1960). Metabolism of insecticides by various insect species. J. Agric. Fd Chem. 8, 266-272. Perry, A. S. and Hoskins, W.M. (195 1). Detoxification of DDT as a factor in the resistance of house flies. J, econ. Ent. 44, 850-857. Perry, A. S., Miller, S. and Buckner, A. J. (1963). Insect metabolism of insecticides. The enzymatic in vitro degradation of DDT by susceptible and DDT-resistant body lice. J. Agn’c. Fd Chem. 11,457-462. Perry, A. S., Pearce, G. W. and Buckner, A. J. (1964). The absorption, distribution, and fate of C14-aldrin and C’4-dieldrin by suscept’ibie and resistant house flies. J. econ. ent. 57, 867-872. Pichon, Y. (1969a). Effets du D.D.T. sur la fibre nerveuse isoEe d’insecte. l h d e en courant et en voltage imposks. J. Physiol., Paris 61, (Suppl. 11, 162-163.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

89

Pichon, Y. (1969b). “Aspects electriques et Ioniques du Fonctionnement Nerveux chez les Insectes. Cas Particulier de la Chaine Nerveuse Abdominale d’une Blatte Periplaneta americana L”. TMse, Univ. Rennes. Potter, C. and Gillham, E. M. (1946). Effects of atmospheric environment, before and after treatment, on the toxicity to insects of contact poisons. I. Ann. appl. Biol. 33, 142-159. Pradhan, S. (1949). Studies on the toxicity of insecticide films. Part 11. Effect of temperature on the toxicity of DDT films. Bull. ent. Res. 40,239-265. Pratt, J. J., Jr. and Babers, F. H. (1953). Sensitivity t o DDT of nerve ganglia of susceptible and resistant house flies. J. econ. Ent. 46,700-702. Ray, J. W. (1963). Insecticide absorbed by the central nervous system of susceptible and resistant cockroaches exposed to dieldrin. Nature, Lond. 197, 1226-1227. Rhoades, W. C. and Brett, C. H. (1948). The relation of temperature in the susceptibility of grasshoppers to synthetic insecticide dusts. J. Kansas Ent. SOC.21,6670. Richards, A. G. and Cutkomp, L. K. (1946). Correlation between the possession of a chitinous cuticle and sensitivity t o DDT. Biol. Bull. 90, 97-108. Roan, C. C. and Hopkins, T. L. (1961). Mode of action of insecticides. A . Rev. Ent. 6,333-346. Roeder, K. D. (1948). Organization of the ascending giant fiber system in the cockroach (Periplaneta americana). J. expl. Zool. 108,243-261. Roeder, K. D. and Kennedy, N. K. (1955). The effect of certain tri-substituted phosphine oxides on synaptic conduction in the roach. J. Pharmac. exp. Ther. 114, 2 11-220. Roeder, K. D. and Weiant, E. A. (1 946). The site of action of DDT in the cockroach. Science, N . Y. 103, 304-306. Roeder, K. D. and Weiant, E. A. (1948). The effect of DDT on sensory and motor structures in the cockroach leg. J. cell. comp. Physiol. 32, 175-186. Roeder, K. D. and Weiant, E. A. (1951). The effect of concentration, temperature, and washing on the time of appearance of DDT-induced trains in sensory fibers of the cockroach. Ann. ent. SOC.A m . 44,372-380. Roeder, K. D., Kennedy, N. K. and Samson, E. A. (1947). Synaptic conduction t o giant fibres of the cockroach and the action of anticholinesterases. J. Neurophysiol. 10, 1-10. Ruch, T. C., Patton, H. D., Woodbury, J. W. and Towe, A. L. (1965). “Neurophysiology”. W. B. Saunders Co., Philadelphia and London. Schantz, E. J., Lynch, J. M., Vayvada, G., Matsumoto, K. and Rapoport, H. (1966). The purification and characterization of the poison produced by Gonyaulax catenella in axenic culture. Biochemistry 5, 1191-1195. Shanes, A. M. (1949a). Electrical phenomena in nerve. I. Squid giant axons. J. gen. Physiol. 33, 57-73. Shanes, A. M. (1949b). Electrical phenomena in nerve. 11. Crab nerve. J. gen. Physiol. 33, 75-102. Shanes, A. M. (1951). Electrical phenomena in nerve. 111. Frog sciatic nerve. J. cell. comp. Physiol. 38, 11-40. Shankland, D. L. and Kearns, C. W. (1959). Characteristics of blood toxins in DDT-poisoned cockroaches. Ann. ent. SOC.A m . 52, 386-394. Smith, J. N. (1962). Detoxication mechanisms. A . Rev. Enr. 7,465-480.

90

T. NARAHASHI

Smyth, T., Jr. and Roys, C. C. (1955). Chemoreception in insects and the action of DDT. Biol. Bull. 108, 66-76. Soliman, S. A. and Cutkomp, L. K. (1963). A comparison of chemoreceptor and whole-fly responses to DDT and parathion. J. econ. Enf. 56,492-494. Spencer, E. Y. and O’Brien, R. D. (1957). Chemistry and mode of action of organophosphorus insecticides. A. Rev. Ent. 2, 261-278. Stampfli, R. (1 954). A new method for measuring membrane potentials with external electrodes. Experientia 10, 508-509. Sternburg, J. ( 1960). Effect of insecticides on neurophysiological activity in insects. J. Agric. Fd Chem. 8, 257-261. Sternburg, J. (1963). Autointoxication and some stress phenomena. A. Rev. Enf. 8, 19-38. Sternburg, J. and Kearns, C. W. (1950). Degradation of DDT by resistant and susceptible strains of house flies. Ann. e n f . SOC.A m . 43, 444-458. Sternburg, J. and Kearns, C. W. (1952). The presence of toxins other than DDT in the blood of DDT-poisoned roaches. Science. N . Y. 116. 144-147. Sternburg, J., Kearns, C. W. and Bruce, W. N. (1950). Absorption and metabolism of DDT by resistant and susceptible house flies. J. econ. Ent. 43, 214-219. Sternburg, J., Chang, S. C. and Kearns, C. W. (1959). The release of a neuroactive agent by the American cockroach after exposure t o DDT or electrical stimulation. J. econ. Ent. 52, 1070-1076. Tahori, A. S . and Hoskins, W. M. (1953). The absorption, distribution, and metabolism of DDT in DDT-resistant houseflies. J. econ. Enf. 46, 302-306 and 829-837. Takata, M., Moore, J. W., Kao, C. Y. and Fuhrman, F. A. (1966a). Blockage of sodium conductance increase in lobster giant axon by tarichatoxin (tetrodotoxin). J. gen. Physiol. 49, 977-988. Takata, M., Pickard, W. F., Lettvin, J. Y. and Moore, J. W. (1966b). Ionic conductance changes in lobster axon membrane when lanthanum is substituted for calcium. J. gen. Physiol. 50,461-471. Takeuchi, A. and Takeuchi, N. (1959). Active phase of frog’s end-plate potential. J. Neurophysiol. 22, 395-4 1 1. Takeuchi, A. and Takeuchi, N. (1960). On the permeability of end-plate membrane during the action of transmitter. J. Physiol. 154, 52-67. Takeuchi, A. and Takeuchi, N. (1964). The effect on crayfish muscle of iontophoretically applied glutamate. J. Physiol. 170, 296-3 17. Takeuchi, A. and Takeuchi, N. ( 1965). Localized action of gamma-aminobutyric acid on the crayfish muscle. J. Physiol. 177, 225-238. Takeuchi, A. and Takeuchi, N. (1966). A study of the inhibitory action of y-aminobutyric acid on neuromuscular transmission in the crayfish. J. Physiol. 183,418-432. Takeuchi, A. and Takeuchi, N. (1967). Anion permeability of the inhibitory post-synaptic membrane of the crayfish neuromuscular junction. J. Physiol. 191, 575-590. Takeuchi, A. and Takeuchi, N. (1969). A study of the action of picrotoxin on the inhibitory neuromuscular junction of the crayfish. J. Physiol. 205, 377-39 1. Tasaki, I., Singer, I. and Takenaka, T. (1965). Effects of internal and external

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

91

ionic environment on excitability of squid giant axon. A macromolecular approach. J. gen. Physiol. 48, 1095-1 123. Terriere, L. C. (1968). The oxidation of pesticides: The comparative approach. In “Enzymatic Oxidations of Toxicants” (E. Hodgson, ed.), pp. 175-196. North Carolina State University at Raleigh. Tischler, N. (1936). Studies on how derris kills insects. J. econ. Ent. 28, 2 15-220. Tobias, J. M. and Kollros, J. J. (1946). Loci of action of DDT in the cockroach (Periplaneta americana). Biol. Bull. 91, 247-255. Tomaszewski, W. and Gruner, H. E. (1951). Die Temperaturabhangigkeit der Wirkung synthetischer Insektizide (DDT, Hexa) auf den Kornkafer (Calandra granaria L.). Beitr. Ent. 1, 105-109. Tsukamoto, M. (1959). Metabolic fate of DDT in Drosophila melanogaster. I. Identification of a non-DDE metabolite. Botyu-Kagaku 24, 141-15 1. Tsukamoto, M. (1960). Metabolic fate of DDT in Drosophila melanogaster. 11. DDT-resistance and Kelthane-production. Botyu-Kagaku 25, 156-162. Tsukamoto, M. (1961). Metabolic fate of DDT in Drosophila melanogaster. 111. Comparative studies. Botyu-Kagaku 26, 74-87. Tsukamoto, M. and Suzuki, R. (1964). Genetic analysis of DDT-resistance in two strains of the house fly Musca domestica L. Botyu-Kagaku 29, 76-89. Tsukamoto, M., Narahashi, T. and Yamasaki, I. (1965). Genetic control of low nerve sensitivity t o DDT in insecticide-resistant houseflies. Botyu-Kagaku 30, 128-1 32. Twarog, B. M. and Roeder, K. D. (1957). Pharmacological observations on the desheathed last abdominal ganglion of the cockroach. Ann. ent. SOC.A m . 50, 231-237. Urakawa, N., Narahashi, T., Deguchi, T. and Ohkubo, Y. (1960). Dual action of maltoxin, a neuromuscular blocking agent from malt rootlet. A m . J. Physiol. 198, 939-942. Usherwood, P. N. R. and Machili, P. (1966). Chemical transmission at the insect excitatory neuromuscular synpase. Nature, Lond. 21 0 , 634-636. Usherwood, P. N. R., Machili, P. and Leaf, G. (1968). I-Glutamate at insect excitatory nerve-muscle synapses. Nature, Lond. 219, 1169-1 172. Van den Bercken, J. (1 968). The action of DDT and dieldrin on nerves and muscles of Xenopus laevis. Meded. R ijksfak. Landbouw-Wetenschappen Gent 33, 1241-1251. Van den Bercken, J. (1969). The effect of DDD on single Ranvier nodes of Xenopus laevis. Eur. J . Pharmac. 9, 146-148. Vidal-Sivilla, S. and Larralde, J. (1949). Effect of hexachlorocyclohexane on the electroencephalogram. Rev. espafi. fisiol. 5, 299-303. Vinson, E. B. and Kearns, C. W. (1952). Temperature and the action of DDT on the American roach. J. econ. Ent. 45,484-496. Wang, C. M., Narahashi, T. and Yamada, M. (1971). The neurotoxic action of dieldrin and its derivatives in the cockroach. Pesticide Biochem. Physiol. 1, 84-9 1. Weiant, E. A. (1955). Electrophysiological and behavioural studies on DDT-sensitive and DDT-resistant house flies. Ann. ent. SOC. A m . 48, 48 9 4 9 2. Welsh, J. H. and Gordon, H. T. (1947). The mode of action of certain

92

T. NARAHASHI

insecticides on the arthropod nerve axon. J. cell. comp. Physiol. 30, 147- 172. Wilkinson, C. F. (1968). Detoxication of pesticides and the mechanism of synergism. In “Enzymatic Oxidations of Toxicants” (E. Hodgson, ed.), pp. 113-149. North Carolina State University at Raleigh. Winteringham, F. P. W. (1 969). Mechanisms of selective insecticidal action. A . Rev. Ent. 14,409-442. Winteringham, F. P. W. and Harrison, A. (1959). Mechanisms of resistance of adult houseflies to the insecticide dieldrin. Nature, Lond. 184, 608-61 0. Winteringham, F. P. W. and Lewis, S. E. (1959). On the mode of action of insecticides. A . Rev. Ent. 4, 303-3 18. Yamada, M. and Narahashi, T. (1 968). Electrophysiological analyses of the structure-activity relationship in DDT derivatives. Report at Annual Meeting Entomol. SOC. Am., Dallas, Texas. Bull. ent. SOC.A m . 14, 208. Yamamoto, I. (1970). Mode of action of pyrethroids, nicotinoids, and rotenoids. A . Rev. Ent. 15,257-272. Yamasaki, T. and Ishii, T.* (1951). Studies on the mechanism of insectides (111). The actions of BHC and rotenone on the heart of the larva of Allomyrinu dichotomus L. Oyo-Kontyu 7 , 31-34. Yamasaki, T. and Ishii, T.* (1952a). Studies on the mechanism of action of insecticides (IV). The effects of insecticides on the nerve conduction of insect. Oyo-Konfyu 7 , 157-164. Yamasaki, T. and Ishii, T.* (1952b). Studies on the mechanism of action of insecticides (V). The effects of DDT on the synaptic transmission in the cockroach. Oyo-Kontyu 8, 11 1-1 18. Yamasaki, T. and Ishii, T.* (1953). Studies on the mechanism of action of insecticides (VI). DDT symptoms in the cockroach, with special reference to the effect of temperature. Oyo-Kontyu 9, 87-96. English translation (1957). In “Japanese Contributions t o the Study of the InsecticideResistance Problem,’, pp. 13G139. Publ. by Kyoto Univ. for W.H.O. Yamasaki, T. and Ishii, T.* (1954a). Studies on the mechanism of action of insecticides (VII). Activity of neuron soma as a factor of development of DDT symptoms in the cockroach. Botyu-Kuguku 19, 1-14. English translation (1957). In “Japanese Contributions t o the Study of the Insecticide-Resistance Problem” pp. 140-154. Publ. by Kyoto Univ. for W.H.O. Yamasaki, T. and Ishii, T.* (1954b). Studies on the mechanism of action of insecticides (VIII). Effects of temperature on the nerve susceptibility to DDT in the cockroach. Bofyu-Kuguku 19, 39-46. English translation (1957). In “Japanese Contributions to the Study of the InsecticideResistance Problem”, pp. 155-162. Publ. by Kyoto Univ. for W.H.O. Yamasaki, T. and Ishii, T.* ( 1954~ ) .Studies on the mechanism of action of insecticides (IX). Repetitive excitation of the insect neuron soma by direct current stimulation and effects of DDT. Jup. J. upp2. Zool. 19, 16-28. English translation (1 957). In “Japanese Contributions to the Study of the Insecticide-Resistance Problem”, pp. 163-175. Publ. by Kyoto Univ. for W.H.O.

* Former name of T.Narahashi.

EFFECTS OF INSECTICIDES ON EXCITABLE TISSUES

93

Yamasaki, T. and Ishii, T.* (1954d). Studies on the mechanism of action of insecticides (X). Nervous activity as a factor of development of y-BHC symptom in the cockroach. Botyu-Kuguku 19, 106-1 12. English translation (1957). In “Japanese Contributions to the Study of the Insecticide-Resistance Problem”, pp. 176-183. Publ. by Kyoto Univ. for

W.H.O.

Yamasaki, T. and Narahashi, T. (1957a). Increase in the negative after-potential of insect nerve by DDT. Studies on the mechanism of action of insecticides (XIII). Botyu-Kuguku 22, 296-304. Yamasaki, T. and Narahashi, T. ( 1957b). Intracellular microelectrode recordings of resting and action potentials from the insect axon and the effects of DDT on the action potential. Studies on the mechanism of action of insecticides (XIV). Botyu-Kuguku 22, 305-31 3. Yamasaki, T. and Narahashi, T. ( 1 9 5 7 ~ ) .Effects of metabolic inhibitors, potassium ions and DDT on some electrical properties of insect nerve. Studies on the mechanism of action of insecticides (XV). Botyu-Kuguku 22,354367. Yamasaki, T. and Narahashi, T. (1958a). Nervous activity as a factor of development of dieldrin symptoms in the cockroach. Studies on the mechanism of action of insecticides (XVI). Botyu-Kuguku 23,47-54. Yamasaki, T. and Narahashi, T. (1958b). Resistance to house flies to insecticides and the susceptibility of nerve to insecticides. Studies on the mechanism of action of insecticides (XVII). Botyu-Kuguku 23, 146-157. Yamasaki, T. and Narahashi, T. ( 1 9 5 8 ~ )Synaptic . transmission in the cockroach. Nature, Lond. 182, 1805-1806. Yamasaki, T. and Narahashi, T. (1959). Electrical properties of the cockroach giant axon. J . Insect Physiol. 3, 230-242. Yamasaki, T. and Narahashi, T. (1960). Synaptic transmission in the last abdominal ganglion of the cockroach. J. Insect Physiol. 4, 1-13. Yamasaki, T. and Narahashi, T. (1962). Nerve sensitivity and resistance to DDT in houseflies. Jup. J. uppl. Ent. Zool. 6 , 293-297. Yates, W. W. (1950). Effect of temperature on the insecticidal action of mosquito larvicides. Mosq. News 10, 202-204.

This Page Intentionally Left Blank

Functional Organizations of Giant Axons in the Central Nervous Systems of Insects: New Aspects I . PARNAS and D . DAGAN Department of Zoology. Hebrew University. Jerusalem. Israel

Introduction . . . . . . . . . . . . . . . . . . 96 A. Definition . . . . . . . . . . . . . . . . 96 B. Occurrence and Examples of Function . . . . . . . 96 C. Giant Fibre System of the Cockroach . . . . . . . . 97 I1. Histological Observations . . . . . . . . . . . . . . 100 A. Abdominal Connectives . . . . . . . . . . . . 100 B. Sheaths . . . . . . . . . . . . . . . . . . 101 C . Abdominal Ganglia . . . . . . . . . . . . . . 102 D. Giant Fibres in Thoracic Ganglia . . . . . . . . . 104 E. Thoracic Connectives . . . . . . . . . . . . . 104 F . Degeneration . . . . . . . . . . . . . . . . 106 G. Giant Fibre Somata . . . . . . . . . . . . . . 108 111. Membrane Properties . . . . . . . . . . . . . . . 110 IV . Through Conduction-“Continuity vs. Contiguity” . . . . . 110 . . . . . . . . . . . . 110 A. Collision Experiments B . Low Safety Factor Zones . . . . . . . . . . . 114 C. Continuity of Giant Axons in Mole Cricket and Locust . . 121 V . Do Giant Fibres Activate Leg Motoneurones? . . . . . . . 121 VI . Afferent Inputs . . . . . . . . . . . . . . . . . 128 A. Cercal Inputs . . . . . . . . . . . . . . . . 129 B. Inputs at Abdominal and Thoracic Ganglia . . . . . . 130 VII . Giant Fibre Outputs . . . . . . . . . . . . . . . . 130 A. Output t o Antennal Motoneurones . . . . . . . . . 130 B . Efferent Activity of Giant Axons in the Metathoracic Ganglion of the Cockroach . . . . . . . . . . . 132 VIII . Giant Axon and Small Fibre Pathway-Timing Relations . . . . 135 IX. Possible Function of Axons in Integration . . . . . . . . 136 Acknowledgements . . . . . . . . . . . . . . . . . . . 139 References . . . . . . . . . . . . . . . . . . . . . . 140

I.

95

96

I. PARNAS AND D. DAGAN

I. INTRODUCTION A. DEFINITION

“Nerve cells or fibres that are notably larger than most others in any given animal are called giant” (Bullock and Horridge, 1965). According to this definition, fibres having a diameter as small as 4 p, could and have been termed giants (in Drosophila, Power, 1948). This definition, which is based only on relative size, implies that the role of the giant fibres is simply to increase conduction velocity, providing rapid, uninterrupted conduction over long distances. However, we hope to demonstrate in this chapter that the importance of giant fibres in insects goes beyond mere increased conduction velocity; namely that they function as integratory neurons. Furthermore, we present some new suggestions and experimental interpretations which have not appeared so far in previous reviews dealing with various aspects of insect giant axons (Hughes, 1965; Huber, 1965; Bullock and Horridge, 1965; Roeder, 1967). B. OCCURRENCE AND EXAMPLES OF FUNCTION

There are many phylogenetic groups in which one can find giant fibres, e.g. flatworms, nemertineans, polychetes, annelids, arthropods, molluscs, phoronids, enteropneusts and vertebrates. For detailed references the reader should consult the book of Bullock and Horridge (1 965). Among arthropods giant fibres are found in arachnids (scorpions: Saint-RCmy, 1866a, b; spiders: Hanstrom, 1923) and crustaceans (Wiersma, 1947). In insects it has been suggested that Aeschnu nymphs (Odonata) employ a longitudinal system of abdominal muscles activated by giant axons and eject water through the anus during escape (Fielden, 1960; Hughes, 1958; Mill, 1963). In other insects the presence of giant fibres was shown histologically without correlated physiological studies. Thus, in Drosophila, sudden take off evoked by visual stimuli is attributed to giant axons (Power, 1948). The prime importance of giant axons was attributed to their fast conduction velocities and thus it was suggested that giant axons serve as premotor interneurons to activate in the fastest way motoneurons involved in a stereotyped escape response. However, we would like to show in this chapter that this hypothesis is not generally correct and that still based on fast conduction velocity the giant axons serve as

FUNCTIONAL ORGANIZATION OF GIANT AXONS

97

interneurons to rapidly activate or inhibit pathways not directly connected to the motoneurons involved in escape (e.g. leg mo toneurons). Recent findings in the cockroach giant fibre system, which is probably the most thoroughly studied, are used to illustrate this point. We will also show that these findings apply to Locusta and Gryllotalpa as well. C. GIANT FIBRE SYSTEM OF THE COCKROACH

It is a well-known phenomenon that tactile and vibratory stimuli to the cerci induce a rapid forward escape response in the cockroach (Roeder, 1948). In their studies on the organization of the nervous system of the cockroach, Periplaneta americana, Pumphrey and Rawdon-Smith (1 937) reported the presence of ascending giant fibres and concluded that some of these originate in the cerci and ascend the ventral nerve cord to the suboesophageal ganglion without any synapse. If these pathways were indeed responsible for the evasive response (this is not the case, vide infra) it would appear to be the most efficient way to ensure the shortest startle response time. It has been shown by Roeder (1948) and Callec and Boistel (1965) that all of the fibres that originate in the cerci synapse in the sixth abdominal ganglion. Furthermore, Pumphrey and RawdonSmith (1937) and Roeder (1948) showed that abdominal giant axons of each connective are innervated ipsilaterally and contralaterally from both cercal nerves. Roeder (1948) also suggested that all of the ascending giant fibres traverse the ventral cord without synapses in the abdominal ganglia and terminate at the metathoracic ganglion. The scheme of the oganization of the giant axons as summarized by Bullock and Horridge (1965) is shown in Fig. 1 and is based on extracellularly recorded action potentials (Roeder, 1948), a noticeable delay in conduction observed across each of the thoracic ganglia, block of conduction across the thoracic ganglia by nicotine, and Hess’s (1 958, 1960) degeneration experiments where the cord was transected between the ganglia A5-A6, and degeneration was seen to proceed only up to T, . Hess (1 958) also showed that two of the giant axons did not degenerate, and suggested that these two conduct descending information. Bullock and Horridge (1 965) also suggested the presence of continuous descending inhibitory fibres and continuous ascending giant axons. AIP-6

98

I. PARNAS AND D. DAGAN long dou. i. \ short dou. I.\

JU. I X .

Fig. 1. Diagram of pathways in the cockroach central nervous system involved in evasion, according to Bullock and Horridge (1965). Typical action potentials recorded at different levels of the cord to ipsi- and contra-lateral cercal stimulation are shown (Roeder, 1948). The scheme includes in addition to the ascending giant axons, continuous ascending axon (asc. ax.), long descending inhibitors (long desc. I) and short descending inhibitory axons. ce.n., Cercal nerves. cr.n., Crural nerve. Compare this scheme with final diagram in Fig. 29.

On the basis of the findings until 1965, Hughes inferred the existence of a hypothetical ancestral type in which a continuous giant fibre system carried information in both directions. He postulated subsequent evolution of the two giant systems generally accepted to exist in the cockroach and dragonfly (Fig. 2). Some difficulties in explaining known integratory mechanisms, however, result from the use of this scheme. For example, one might ask what

FUNCTIONAL ORGANIZATION OF GIANT AXONS

I*

99

\

I‘

...........

1

........

...........

.........

..........

.........

-7

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

-7

1

/

.........

/

Cockroach

/

-7

........ -I

.......

........

.......

........ -\

......*../L -\

.........+

i

Aerchnid

Fig. 2. Diagrams illustrating neuronal pathways involved in the giant fibre system. (After Hughes, 1965.) Note that the scheme for the cockroach includes descending giant axons. In the cockroach and dragonfly larva, sensory inputs (dashed line) to the ascending giant axons are found only from the caudal end. In the hypothetical ancestral form, both sensory inputs and motor outputs (dots) are found along the axon.

would happen if information should reach the giant axons at the same time from two widely separate inputs. Thus (Fig. 2c) in the hypothetical ancestor, simultaneous inputs from two sources will result in occlusion of giant axon spikes, interfering with co-ordinated movement. Recent findings (Spira et al., 1969a, b; Parnas et al., 1969; Farley and Milburn, 1969) show that the organization of giant fibres in the cockroach is, in fact, different from the one described and appears to offer a satisfactory explanation for known behaviour including the question of occlusion. We shall now summarize the recent findings concerning giant

100

I. PARNAS AND D. DAGAN

fibres in the cockroach and other insects. At the end of the chapter some suggestions and speculations as to the integratory function of the giant axons will be made. I1 HISTOLOGICAL OBSERVATIONS A. ABDOMINAL CONNECTIVES

In each of the cockroach abdominal connectives, giant fibres are arranged in two conspicuous groups: a ventral group of four axons of 25-60 p diameter and a dorsal group of 4-5 axons of 20-35 p diameter (Fig. 3(a)). Guthrie and Tindall (1968), report measwements from freshly thawed sections which generally agree with the earlier findings of Roeder (1 948), Pipa et al. (1 959) and Hess (1958). An arrangement into two groups is also present in Locusta (Cook, 195 1 and Fig.,4(a)). A single median fibre or possibly two (Satija, 1958a) of an approximate average diameter of 13 p and a peripheral laterodorsal group of three fibres: two of 12 p and a third of 8.5 p

Fig. 3. Cross-sections of cockroach nerve cord: (a) abdominal connectives at A,-A, ;(b) abdominal ganglion A,: (c) caudal part of metathoracic ganglion; (d) central portion of metathoracic ganglion. Note pronounced differences in diameter of giant axons in connectives and ganglion. (From Spira et aL, 1969b.)

FUNCTIONAL ORGANIZATION OF GIANT AXONS

101

Fig. 4. Cross-sectionof locust nerve cord: (a) abdominal connectives at the caudal end of ganglion A,; (b) abdominal ganglion A, ; (c) caudal end of the metathoracic ganglion; (d) thoracic connectives at T, -T, .

in diameter were described. In dragonfly nymphs (Anax imperator) (Anisoptera), Hughes ( 1953) demonstrated, seven giant fibres of 12-16 p in diameter. A similar situation exists in the Dragonfly nymph of Aeschna (Zawarzin, 1924). The general picture found in Orthoptera is evident also in the mole-cricket Gryllotalpa and the giant fibres here appear also in a dorsal group of three fibres and a ventral group of four fibres all about 30-40 p in diameter. Surrounding these giant fibres are about 20 other large fibres of 18-23 p i n diameter (Fig. 5). The same arrangement exists at all levels of abdominal connectives in the different species. E.SHEATHS

Giant fibres are found to be ensheathed with extraganglionically nucleated sheaths probably serving as a barrier t o diffusion, and possibly increase the conduction velocity by decreasing the electrical

102

I. PARNAS AND D. DAGAN

Fig. 5 . Cross-section of mole cricket nerve cord: (a) abdominal connectives at A,-A, ;@) abdominal ganglion A, ; (c) thoracic ganglion T, ; (d) thoracic connectives at T,-T, Note that the arrangement of giant axons in the abdominal cord resembles that of the cockroach.

.

shunting of the action current. These sheaths can be seen surrounding giant fibres of Periplaneta americana. Anax and Dytiscus (Hughes, 1953, Treherne, 1967). C. ABDOMINAL GANGLIA

In the cockroach, the giant fibres narrow considerably, the diameter decreasing to 15-30 p (Fig. 3(b)) (Roeder, 1948; Hess, 1958; Guthrie and Tindall, 1968; Spira et al., 1969b) and even less in the locust (Fig. 4(b), and Cook, 1951) while passing through an abdominal ganglion. In Gryllotalpa, the fibres narrow to 10-18 p (Fig. 5(b)). These changes in diameter are more or less constant throughout the abdominal ganglia and are superimposed on a gradual posterior-anterior tapering of giant fibres. While no lateral branches

FUNCTIONAL ORGANIZATION OF GIANT AXONS

103

could be traced in the abdominal ganglia from the giant fibres in the cockroach (Roeder, 1948), Mill (1963) has shown that the giant fibres in Aeschnid nymphs synapse at each abdominal ganglion with mo toneurons to the abdominal muscles. These changes in diameter should have been reflected in conduction velocity and cause a delay across the ganglion. The length of an abdominal ganglion is about 0.5 mm and the space constant for the giant fibres in the connectives was found to be 0.86mm (Yamasaki and Narahashi, 1959). If we assume homogeneity of the giant axon so that its specific membrane and axoplasmic resistances do not change along the fibre we can calculate the space constant of the narrow region while passing through an abdominal ganglion. Since the space constant is given by

(Hodgkin and Rushton, 1946) where a = radius of axon in p Rm = specific membrane resistance C2 cm2 Ri = specific axoplasm resistance C2 cm then the correction factor for the narrower parts’ space constant can be shown to be

where a, and a,, are the radii of the large and narrow parts respectively. Thus,

In other words the space constant of the narrow part is 0.6 mm, i.e. longer than its total length. Intracellular recordings by Pichon (1969) show that the spike magnitude is 96 mV, and the threshold depolarization is 18.8 mV, thus the spike could propagate decrementally 1.5 space constants and still produce enough depolarization to initiate a spike. It is therefore not surprising that a change in conduction velocity due to changes in diameter is not exhibited during passage through the ganglion. Furthermore one might expect these narrowings to serve as

104

I. PARNAS AND D. DAGAN

zones with a low safety factor on the basis of current density considerations. Namely the current density of a spike emerging from the tunnel into the wider part of the axon may not be sufficient to generate the next spike. Although this apparently is the case in the thoracic ganglia (see p. 120) the narrowing of the giant axons in the abdominal ganglia is neither abrupt nor long enough to endanger propagation of the spike. Indeed conduction in the abdominal cord does not fail even with prolonged stimulation at frequencies of up to 200/s in either direction. D. GIANT FIBRES IN THORACIC GANGLIA

Pipa et al. (1 959) claim that the abdominal giants are continuous throughout the thoracic ganglia. The two groups of abdominal giant fibres are distinctly separated in the methathoracic ganglion (Fig. 3(d)) and a sharp posterio-anterior narrowing is evident. Central and anterior sections of the same ganglion no longer show the conspicuous separation into two groups. Spira et al. (1969b) and Farley and Milburn (1969) both describe a sharp narrowing of the dorsal giant fibre group in the metathoracic ganglion without a widening at the anterior part. Satija (1958a) describes three groups of giant fibres in the thoracic ganglia of the locust: dorsal, medial and ventral bundles (Fig. 4(c). In Aeschna nymphs, Satija (1958b) reports that five bundles of giant fibres widen on entering the thoracic ganglia from 7-8 p in the connective to 10-11 p in the posterior end of the prothoracic ganglion. However, this may be interpreted as similar to the Orthopteroid situation of a general caudorostral tapering with narrowings in the ganglia. In Gryllotalpa, longitudinal sections show giant fibres from the first abdominal ganglion, which is in close proximity t o the- metathoracic ganglion, entering the thoracic ganglion and tapering down to 16 p in the caudal third of the ganglion. In contrast to the abdominal region of the cockroach where giant fibres apparently do not branch, Farley and Milburn (1969) have shown metathoracic lateral branches from the ventral giant fibres. They propose that these branches fit Roeder's (1948) and Pipa et al. 's (1959) suggestion that the giant fibres trigger efferent activity in the metathoracic c r u d nerves. E. THORACIC CONNECTIVES

Cross-sections of thoracic connectives show distinctly the ventral

FUNCTIONAL ORGANIZATION OF GIANT AXONS

105

giant fibre group (Pipa et al., 1959; Spira et al., 1969b, and Fig. 6a) while it is difficult t o distinguish the abdominal dorsal group since there are now many more dorsal giant fibres present than in the abdominal connectives. In Gryllotalpa the situation is rather similar and the largest giant axons here are about 27 p (Fig. 5). These additional giant fibres may well serve to co-ordinate leg movements. Further studies are needed t o answer some of the contradictions

Fig. 6. Crowsections at thoracic connectives of cockroach: a, T, -T, ;b, anterior portion of T,-T,; c, T,-T,. Note the clear arrangement of the ventral goup of abdominal giant fibres (marked by rectangles), and the progressive posterio-anterior decrease in their diameter. (After Spira et al., 1969b.)

106

I. PARNAS AND D. DAGAN

between Pipa et al.’s and Roeder’s findings: namely the issue of “continuity vs. contiguity”* of giant fibres in the thoracic region. F. DEGENERATION

In order to provide the exact localization of somata and to rule out the possibility of septate giant fibres in the abdominal ganglia, degeneration experiments were performed. These experiments are based on the assumption that an axon will degenerate if it is no longer continuous with its soma while the proximal part of the axon should undergo less drastic histological changes (Hess, 1958, 1960). The absence of transneuronal degeneration is also assumed in such experiments. Athough some cases of transneuronal degeneration have been reported (e.g. sensory cells of the lateral line of the newt degenerated following sectioning of all nerve supply, Bennet, 1970, personal communication and Jones et al., 1970). Thus axons can be sectioned in various connectives and their somata can be localized to specific ganglia. Hess (1 958) reported that following transection of the abdominal cord at various levels he could show degeneration of all giant fibres except for two in the ventral group and localized the somata of the giant fibres as being in the last abdominal ganglion. This latter finding was confirmed by Farley and Milburn (1969). Hess (1958) considered the two “nondegenerating” giant fibres to be descending. The most significant finding of Hess was his inability to cause degeneration of giant fibres in the thoracic connectives, thus confirming a barrier (synapse or septum) in T3. To summarize our conclusions from Hess’s findings: 1 . abdominal giant fibre somata are located in A, ; 2. giant fibres that degenerated have no other somata in the abdomen or if present they are not sufficient to prevent degeneration; 3. two giant fibres are different; 4. all giant fibres that degenerated terminate in T3, supporting Roeder’s scheme.

Farley and Milburn (1969) and Spira et al. (1969b) have used unilateral connective transections, leaving the other connective intact to serve as a control, and reported signs of degeneration in thoracic

* “Continuity” denotes a continuous axoplasm all along the axon, while “contiguity” implies a continuous pathway, but its cells are separated by synapses or septa.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

107

connectives 30 days following transection at As -A5.Comparing this with the unoperated connective they conclude that the ventral giant fibres ascend continuously from the last abdominal ganglion to the head, in accordance with Pipa et al. ' s (1 95 9) observations. Farley and Milburn (1969) also argued that the dorsal giant fibre group terminates in the mesothoracic ganglion. Spira et al. (1 969) in similar experiments, however, reported degeneration also of the dorsal giant fibre group up to the premesothoracic level (Fig. 7a, b).

Fig. 7(a). Cross-sections at abdominal connectives of the cockroach: A, A,-A, normal animal; B, A, -A, eight days after transection at A, -A6 ;C, A, -A, 10 days after transection; D, A,-A, 10 days after transection; E, A, -A6 m d F, T,-A, 17 days after cutting the left A,-A, connective; G, A,-A, and H, T,-A, 30 days after cutting the right A,-A, connective. Calibration 100 ~.(After Farley and Milburn, 1969.)

108

1. PARNAS AND D. DAGAN

Fig. 7(b). Degeneration of abdominal giant axons after unilateral transection of one connective at A,-A,, in cockroach: (a) cross-section at T,-T,; (b) cross-sections at T,-T,. Arrows mark degenerated axons in right connectives. Note degeneration of both ventral and dorsalabdominal giants both at T, -T2 and T,-T,. (After Spira el af., 1969b.)

It should be borne in mind that these degeneration experiments do not prove direction of conduction since there are axo-axonal synapses in insects which may be remote from the somata (Bullock and Horridge, 1965). Furthermore, an axon having a cluster of somata (see next section) may possibly degenerate when disconnected from this cluster while having another soma distal to this cluster which may not be sufficient to prevent degeneration. Satija’s (1 958a) degeneration experiments in the locust were inconclusive and in our experiments, 30 days after transection of the nerve cord between A4-A5 in locust, no signs of degeneration could be observed. Boulton (1 969) and Boulton and Rowel1 (1969) report that locust axons show very little degeneration while any part of them is juxtaposed to a non-sectioned nerve and imply strong trophic influences from neighbouring axons. But it is also possible that in the locust somata of the giant axons are located in ganglia anterior to the cut (e.g. thoracic ganglia). G. GIANT FIBRE SOMATA

Roeder (1 948) attempted to trace giant axons to their somata in the last abdominal ganglion, and suggested that the giant fibres might

FUNCTIONAL ORGANIZATION OF GIANT AXONS

109

originate from clusters of somata. Hess (1958) and Wigglesworth (1960), indicated that the location of giant fibre cell bodies is indeed in the last abdominal ganglion. Further degeneration experiments (see previous paragraph) ascertained the origin in A6 of giant axons. Following transection of the cord at A5-A6 , Farley and Milburn (1969) traced giant fibres into the last abdominal ganglion using both light and electron microscopy. They identified the injured giant

cN

/

Fig. 8. Organization of giant axons and their somata in the last abdominal ganglion of the cockroach. D, Dorsal giant neurons. V, Ventral giant neurons. CN, Cercal nerves. (After Farley and Milburn, 1969.)

fibres by appearance of myelin globules (which they assume to be secondary lysosomes), disarrangement of neurotubules, increase in number of mitochondria, and the presence of rough endoplasmatic reticulum. Processes of these giant fibres led to several groups of somata 30-50 p in diameter. A schematic drawing summarizing the giant fibre arrangement in the last abdominal ganglion according to Farley and Milburn (1 965) is given in Fig. 8. In the locust, Cook (195 1) traced one giant fibre to a single large cell body in the last abdominal ganglion and concluded that this axon is unicellular. Three other axons were seen to disappear in the

110

I. PARNAS AND D. DAGAN

region of large cell bodies in the same ganglion but connections to these cells could not be demonstrated. Thus the somata of some of the locust’s giant axons may possibly be located in thoracic ganglia. 111. MEMBRANE PROPERTIES

Since this review is primarily aimed at integratory mechanisms and functional organization of giant fibres, we shall not elaborate in detail on the ionic and membrane properties. Excellent reviews on insect giant fibre membrane properties have been written by Narahashi (1 963, 1965)’ Treherne (1 967) and Guthrie and Tindall (1968) and the reader should consult the recent papers of Boistel and Pichon ( 1969) and Pichon and Boistel (1 967a, by 1968). Some of the recent findings on membrane properties of giant axon are summarized in Table I. Table I Summary of cockroach giant axon membrane properties (after Pichon, 1969) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Temperature Diameter of axon Input resistance Time constant Specific membrane resistance Resting potential Threshold depolariz ation Action potential Rate of depolarization Rate of depolarization Positive after-potential Negative after-potential Specific membrane capacitance

25.loC 46 I.( 598.5 KS2 0.96 ms 293 a c m 2 in situ -58.4 mV 18.8 mV 96.6 mV in situ in situ 1370 V/s 440 Vls in situ -1.2 mV +0.96 mV 3.33 pF/crn2

IV. THROUGH CONDUCTION-“CONTINUITY

-43.0 mV 105.5 mV 1161 V/s 386 Vls

vs. CONTIGUITY”*

A. COLLISION EXPERIMENTS

The functional organization of the giant axons in the nerve cord of Periplaneta americana has been recently re-examined by at least two groups. Spira et al. (1969a, b) and Parnas et al. (1969) concentrated on the “continuity vs. contiguity” aspect of the giant fibres in the * See footnote on p. 106.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

111

thoracic region while Farley and Milburn (1969) focused more on the structural and ultrastructural aspects. Spira et al. (1969a) used spike collision experiments (Fig. 9) to show that the giant pathway is continuous from the last abdominal ganglion to the prothoracic ganglion. In these experiments ascending and descending impulses given simultaneously, result in mutual extinction along the nerve cord, the point of collision depending on the timing of the ascending and descending pulses. A quantitative analysis of such collision experiments (Fig. 9A) showed that ascending and descending spikes could collide all along the nerve cord, and anywhere along the thoracic connectives, implying a common route for ascending and descending spikes from “head” to “tail”. Descending spikes initiated at the sub-oesophageal ganglionTI connectives, showed similar features to the ascending ones evoked at the cercal nerves when both were recorded at any common place along the connectives. Furthermore no crossconnections of giant axons were found between left and right connectives. Although these results suggest at first glance the existence of giant axons as through fibres without any synapses, two other experimental results namely blocking of conduction in the thoracic ganglia by 5 y/ml of nicotine and a ganglionic delay of 0.6 ms (Roeder, 1948; Spira et al., 1969a) have to be explained. There are three possible models (Fig. 10) that may explain these results: A. The giant axons are continuous from A6 to the suboesophageal ganglion, a model which does not agree with the model presented in Fig. 1. B. The giant axons form at each of the thoracic ganglion axo-axonal mutually excitatory synapses, enabling bidirectional conduction. Such a model agrees with the results of Roeder ( 1948) and Hess’s (195 8) degeneration experiments. C. The common pathway is interrupted in each of the thoracic ganglia by a septum with tight or gapjunctions or other low safety factor elements arranged in series. In all of these three models, the anatomical fact of a general tapering is superimposed rostrally. The smaller diameter of the ascending giant axons in the thoracic connectives probably accounts for the smaller compound action potentials recorded extracellularly by Roeder (1948) and which led him to conclude that only a fraction of the abdominal giant axons reach the head.

112

I. PARNAS AND D. DAGAN

Fig. 9.

113

FUNCTIONAL ORGANIZATION OF GIANT AXONS

B I

g

I

I I I

I

I

I I

g -?

I

I

I

I

‘

I

I

I

1

I

I

Fig. 10. Schematic representation of three models suggested by the experimental findings. A, The ascending pathway is continuous from A, to the sub-oesophagealganglion, tapering towards the head with narrowings at each ganglion (“continuity” model). B, Ascendingdescending pathway with mutually exciting synapses at each thoracic ganglion “contiguity model”. C, The pathway is interrupted by septa “contiguity model”. (After Spira er al., 1969a).

As already mentioned, both the work of Farley and Milburn (1 969) and Spira et al. (1969b) exclude the axo-axonal synaptic and septa1 models. Their degeneration experiments imply a continuous pathway, assuming that significant transneuronal degeneration does not take place. It is difficult to exclude transneuronal degeneration (see p. 106), transection of the cercal nerves, however, did not cause degeneration of the giant axons in the abdomen (Farley and Milburn, 1969; Hess, 1958), nor was any degeneration found in metathoracic ~~~

~

Fig. 9. Ascendingdescending interaction in the CNS of the cockroach. Inset shows experimental set-up, each electrode served for either stimulation or recording through a switch box. A, Quantitative analysis of collision point along the nerve cord. Solid line 0-Brepresents the interval between the stimulus and the first large ascending spike, recorded at different electrodes, and evoked at the cercal nerves (El). Corresponding values for descending impulses, elicited above the prothoracic ganglion (E2),are plotted on the broken line F-F, . The intersection of the two curves at F , , denotes site of collision after simultaneous excitation at El and E, . Intersection points beyween broken lines and solid line, show the point of collision when the descending impulses preceeded (K, I, H, C) or followed (D, C, B, A) the ascending pulses. B, Collision of ascending and descending impulses at different delays; activity was recorded at E, and E, : 1. Control, ascending response evoked at CN (El 1; 2. Control descending responses evoked at E, . Delays between the ascending and descending stimuli in ms were: a, +lo;b, +9; C, +5; d, +2; f, o;g, -1; h, -3; i, -5; k, -6. Single arrow indicates a descending spike recorded by E,. Double arrow indicates a descending spike recorded by E,. Dotted double arrow shows position of “missing” descending spike. Full dot shows ascending spike recorded by E,. Circle shows position of “missing” ascending spike recorded by E, Scale 10 ms,0.25 mV. (After Spira el ul., 1969a.)

.

114

I. PARNAS AND D. DAGAN

leg nerves following giant fibre degeneration (Hess, 1958). Thus neighbouring synapses in the same preparation do not exhibit transneuronal degeneration. Injection by current or pressure of dyes such as procion yellow (Stretton and Kravitz, 1968) into the giant axons at the base of ganglion T3 , might give a more direct answer to this problem. Such experiments are now in progress in our laboratory. The appearance of dye in the giant axon rostra1 to the ganglion would be strong evidence of continuity. It would seem at this stage, therefore, that the first of the three models suggested, namely a continuous axon from A6 to the sub-oesophageal ganglion, is the most probable (see note added in proof, p. 143). The narrowing of the giant fibres while traversing a thoracic ganglion explains the ganglionic delay if the narrowings were sul’ficiently small and long enough (see p. 103). The susceptibility of these regions to nicotine would appear to be also an attribute of the narrowness of thoracic giant axons in the ganglionic mass. It has been shown that finer axons are blocked by nicotine more readily than larger ones and that nicotine blocked thinner axons in the abdominal region as well (Spira et al., 1969a).

B. LOW SAFETY FACTOR ZONES

Microelectrode studies (Parnas et al., 1969) in the region of the isthmuses in the thoracic ganglia were useful in deciding on the validity of the various models and suggested that morphological nuances are capable of explaining mechanisms of complexity hitherto unthought of. The following experiments demonstrate that the “narrowings” in the thoracic ganglion behave as regions of low safety factor for axonal conduction. A microelectrode was inserted into a giant axon at the level of A3-A4 and the connectives stimulated at So-T, , T I -T,, T,-T3 , or As-A6. When twin pulses of varying interval were given to the connectives above ganglion T, the response to the second stimulus was blocked abruptly (Fig. 11 AB), after a delay of 4 ms. On the other hand when the same axon was stimulated below ganglion T, the usual relative refractory period of 2 ms was observed (Fig. 11 C, D). In the same axon application of nicotine blocked only the responses induced above ganglion T, (Fig. 12) and when the axon was stimulated at a rate of 50/s, again the responses initiated above T, were almost immediately blocked (Fig. 13 A, B). It is clear that in this case a zone of low safety factor

115

FUNCTIONAL ORGANIZATION OF GIANT AXONS

UuumL I B

A

D

C

E

Fig. 11. Twin pulse experiment. Microelectrode recordings at A, -A, connective. Twin pulses with varying intervals were applied to So-T, . A, T, -T2 ; B, T, -T, ; C, A, -A6; D and cercal nerves E. Note in A and B where the pulses were given above ganglion T, , the one step drop of the second response starting at a delay of 4 ms. In C and D where the pulses were initiated below ganglion T, , there is a slight decrease in size of the second response starting at a delay of 4 ms and a block at a delay of 2 ms. E, One step block of second response at a delay of 10 ms. Scale 5 ms, 50 mV. (After P ~ M etS al., 1969.)

existed in ganglion T,. In other experiments similar zones of low safety factor were identified also in ganglia T, and T3. Intracellular recordings from a giant fibre at the caudal base of the metathoracic ganglion showed spikes in a one to one fashion to single stimuli at all levels of the cord. On the other hand stimulation at 50/s at TI-T, (Fig. 14), elicited a spike which slowly increased in latency. A more slowly rising pre-potential became evident and the spike was now accompanied by depolarizing after-potential of long duration (subsequent traces are seen to rise from a more depolarized level). Finally, the spike failed to invade the region of recording, and only a brief depolarization of 12 mV remained (Fig. 14L). This depolarizing response showed neither facilitation nor further fatigue and thus differed from the usual synaptic potential. This finding strongly supports the histological finding of continuity of the giant axons through the me tathoracic ganglion. Furthermore, it appears that the “narrowing” in the thoracic ganglion behaves like a zone of a low safety factor and that the current spread from the narrow part to the large axon is not sufficient to reach the threshold for spike initiation after prolonged stimulation at high frequencies. The depolarization seen may be either causally or only concomitantly related to the

I. PARNAS AND D. DAGAN

116

CL

D

-

Fig. 12. Effect of nicotine on conduction. The same axons as in Fig. 11. Superimposed sweeps in all cases recording at A, -A,. A, B, Stimulation at So-T, . Block occurred in less than 30 s; after application of 10 rg/ml nicotine to thorax only. C, D, Stimulation at T, -T,, responses also blocked in 30 s. Again, here stimulation was given above ganglion T, E, F, Stimulation at T,-T, and A,-A, respectively. No block occurred even after nicotine was applied for several minutes only to thorax. G, Control response to cercal stimulation. H-I, Block of the response to cercal stimulation after 10pg.rglml nicotine were applied to the whole bath. K-L,Responses to T, -T, and A, -A, were not blocked. Scale 5 ms, 50 mV. (After Parnas et al., 1969.)

.

117

FUNCTIONAL ORGANIZATION OF GIANT AXONS

A l-e ’ 1 A

C

L

Fig. 13. Repetitive stimulation. The same axon as in Figs 11 and 12. Conduction failure induced by repetitive stimulation superimposed sweeps in all cases, recording at A , -A,. A, 5O/s stimulation at So-TI. B, Block of the responses from So-T, took place after stimulation continued for 30 s. C, Stimulation at T, -T,. after block at So-T, . Note that in both cases stimulation was given above ganglion T,. D, Responses to stimulation below ganglion T, , at T, -T, . E, 5O/s stimulus at A, -A, . Note that there was no block of the spike response even after few minutes of stimulation. (From Parnas e l al., 1969.)

failure of propagation. This last problem is both suitable and worthy of further investigation. In another series of experiments, a microelectrode was inserted at the connective T2-T, (rostrally to ganglion T3). Thus the microelectrode was between two narrowings: one in T2 , conducting descending information, and the second at T3, carrying ascending information. When the connectives below ganglion T3 were stimulated, responses were recorded above T3, both to twin pulses and high frequencies and this narrowing did not behave as an area of low safety factor (Fig. 15 D,E). In these experiments, however, stimulation above ganglion T2 showed that for down-going information the “neck” behaved as an area of low safety factor (Fig. 15 G-I), and was rapidly blocked by high frequency stimulation. This result shows that conduction in both directions in such a “narrowing” is not symmetrical ,and that this area prefers to transfer high frequency information in the up-going direction. Furthermore, these zones tend to behave like low-passfilters, passing information only up to a certain frequency. The non-symmetrical changes in diameter are probably responsible for this asymmetric behaviour. Since the giant axons are

118

I. PARNAS AND D. DAGAN

.L E

.

-

*

F

I

!

@k. G

H

K

L

Fig. 14. Multiple sites of low safety factor. Responses to stimulation at So-T, (A-D) and T,-T, (E-H) recorded at the caudal base of ganglion T,. Note that A, E and F are at lower gain. A, Control response to stimulation at So-T,. B-D, Progressive changes during repetitive stimulation at 5O/s. E, Stimulation at T, -T2. F, After repetitive stimulation at 5O/s, note slowing of initial rise time with spike appearing at a threshold of 12 mV (arrow). G-H,Superimposed sweeps at higher maghification during repetitive stimulation of T, -T2. I-L, 5O/s repetitive stimulation of T,-T, . Scale: 5 ms; A, E, F, I, 25 mV; others 12.5 mV. (From Parnas ef al., 1969.)

continuously tapered towards the head with the “necks” located in the thoracic ganglia superimposed on the general tapering, the change in diameters in the upgoing direction is always less pronounced, than in the descending direction. During invasion of the giant axon from the neck in the ascending direction, the safety factor is not critically reduced and current spread is apparently sufficient to induce spikes at all frequencies. It is also possible that the neck is more abruptly formed at the caudal ends of the ganglion. This has not been investigated to date. Failure of invasion of spikes in the transition of sudden narrow axons into regions of greater diameter is a well-known phenomenon: the lack of antidromic invasion into a soma through the axon hillock is one example seen both in vertebrates (Coombs ef al., 1957) and invertebrates (Eyzaguirre and Kuffler, 1955). A more closely related example may be that of Lumbricus giant axons, whele Bullock and Turner (1950) also reported non-symmetrical

FUNCTIONAL ORGANIZATION OF GIANT AXONS

119

A

t

Fig. 15. Mid-thoracic recording. Intracellular recording a t connective T, -T, of spikes elicited from the cercal nerves (A€), from A,-A, D-E and from T, -T,(F-I). A, Twin pulses delivered to the cercal nerves; block occurred at a delay of 5 ms. B C , Repetitive stimulation of the cercal nerves produces a n increase in latency (B) and block (C). D,Responses elicited by twin pulses at A,-A,. E, Repetitive stimulation at 50/s of A,-A,. F-I, Repetitive stimulation (SO/s) at T, -T, caused slowing in rise time (G) and block of the spike leaving only a local response ( G I ) . Scale: A-G, 10 ms; H-I, 5 ms; A G , 33 mV; H-I, 20 mV. (From Parnas e l 41.. 1969.)

conduction: at higher stimulation frequencies a unidirectional block occurred but this finding was not correlated with changes in axon diameter. On p. 103 we explained the absence of significant changes in conduction velocity during passage through the neck on the basis of calculations of the length constants. Since the length constant of even the narrowed giant axon in an abdominal ganglion was longer

120

I. PARNAS AND D. DAGAN

than the length of its narrower part in the ganglion, no significant changes in conduction velocity would be expected and none were recorded. Similar calculations for the metathoracic ganglion show that the space constant at the narrow zone in T3, is 0.15-0.2 mm, assuming that the 6 0 p diameter axon at the base of T, narrows to 2 p . Since the length of the ganglion is 2 m m and the narrowing appears to be present throughout most of the interganglionic passage of the giant axon, the magnitude of the spike is greatly attenuated and a significant decrease in conduction velocity would be expected. A minimal length of 0.6-0.7mm of axon of 2 p diameter is sufficient to account for the recorded conduction delays of 0.6-0.7 ms, assuming a conduction velocity of 1.0-1.2 m/s at the narrow zone. Furthermore, the short space constant and the relatively long length of the narrowing mitigate strongly against electrotonic conduction through the ganglion and for active propagation throughout the length of the ganglion, in contrast to the suggestion presented by Parnas e l al. (1969). Moreover, the asymmetry of the high frequency conduction block indicates that the narrow regions themselves, despite their brief length constant cannot be held responsible for the block. Again we must revert to the postulate that the larger change in diameter, perhaps a more abrupt change as well, in the caudal direction is the responsible factor in this finding. Despite our delineation of the underlying structural basis for conduction block at high frequencies, we have not shown the specific changes that cause the region of low safety factor to exhibit failure of conduction. It is quite possible, that failure of conduction is associated with meaningful changes in the external ionic medium locally bathing the axons during their passage through the ganglia. A local increase in external potassium (Baylor and Nicholls, 1969) or activation or inactivation of an ionic pump mechanism by high frequency stimulation (Nakajima and Takahashi, 1966) would explain the prolonged depolarization seen in response to high frequency stimulation. The depolarization, in turn, would produce sodium inactivation and/or increased potassium conductance, factors which in themselves could account for the failure of conduction in a particularly susceptible region, in this case the transition from a neck of 2 1.1 to a giant axon up t o 30 times larger in diameter. Three questions arise: first of all, is there any functional significance of these unidirectional low pass filters and do they relate to the escape behaviour of the cockroach? Second, if so-what and how? The third question is whether this situation is limited to the

FUNCTIONAL ORGANIZATION OF GIANT AXONS

121

cockroach or whether the phenomenon of “neck filters” is more widespread. We will treat the third question first. C. CONTINUITY OF GIANT AXONS IN MOLE CRICKET AND LOCUST

In the cockroach, continuity of the abdominal giant axons through the thoracic nervous system was established after a variety of experiments, including: spike collision using extracellular recording, histological studies of the normal and degenerated giant axon population and intracellular recordings close to the area of contention. In previous sections the organization of abdominal giant axons in locust and mole cricket has been described on the basis of histological studies (Figs 4 and 5). The locust is readily available and several studies of the CNS have already been made (Rowell, 1965; Cook, 195 1). Notably the degeneration experiments of Satija (1958a) as well as our own (see p. 108) are inconclusive, but suggesting that the locust giant fibre system might be somewhat different than that of the cockroach. The mole cricket Gryllotalpa sp. lives in narrow tunnels where it can advance only forward or backwards. It was therefore speculated that two sets of giant fibres might control escape-one group for escape in each direction. Although all of the experiments done with cockroaches were not repeated, enough experiments were performed to provide quantitative analysis of collision both in locust and mole cricket. In both cases it was shown that all giant fibre spikes initiated at the lower part of the abdomen and at So-T, connectives could occlude each other all along the nerve cord (Figs 16 and 17). As a corollary the point of collision could be varied by appropriate timing of the ascending and descending volleys. Only the giant spikes were occluded by this procedure and responses of small axons could be recorded at any interval between the two volleys. We can thus conclude, that in the Orthoptera: cockroach, locust and mole cricket, the arrangement of giant axons is similar. V. DO GIANT FIBRES ACTIVATE LEG MOTONEURONES?

We have shown that giant fibres run without synapses to the head. Furthermore, the giant axons narrow considerably while passing through a thoracic ganglion and such “narrowings” can behave as unidirectional low-pass Titers.

122

I. PARNAS AND D. DAGAN

(C)

(d 1

Fig. 16. Ascendingdescending interaction in locust. Recordings from abdominal cord at A, -A2. First spikes in each trace evoked at A, -A4 (ascending). Second volley at So-T, (descending). (a)-(c) Reduction of interval between ascending and descending stimuli. Note occlusion of descending spikes in giant axons in (d). Cal.: 5 ms.

Fig. 17. Ascendingdescending interaction in mole cricket giant axons. Activity recorded in abdominal cord at A,-A, (superimposed sweeps). First action potential evoked by stimulation at A, -A4 (ascending). The descending impulses were evoked at So-T,.The time interval between the ascending and descending impulses was shortened until all descending spikes occluded. Cal.: 5 ms.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

123

The assumption that these abdominal giant axons are directly involved in the activation of leg motoneurones was widespread for many years (Roeder, 1948, 1967; Hughes, 1965; Guthrie and Tindall, 1968). The basis for this assumption is first comparative: crayfish, squid, and Mauthner fibres are examples drawn from two different invertebrate phyla as well as two different vertebrate orders. The second reason for this assumption is one of economics: the cost in volume and integrative capacity of the nervous system involved in doubling the velocity must be paid back by use in a vital rapid function. What better example in these slow feeders than escape? However, despite such compelling arguments, no unequivocal evidence is available to warrant such an assumption. If indeed the giant fibres are responsible for the passage of information from the sixth abdominal ganglion to the metathoracic ganglion, and not smaller axons, the overall conduction time in the abdomen is 2.8 ms (Roeder, 1967). However, this time represents only a small fraction of the total response time to startle, which was found to vary between 28-90ms. Roeder, in his book “Nerve Cells and Insect Behaviour” (1967), stresses that any saving of time is crucial for escape and survival, and emphasizes the “high cost” in space taken by the giant fibres. In his words: “The information handling capacity of 100 small axons operating in various numbers and combinations is astronomically greater than that of a single axon.” Indeed, if the evasive response operates through a single system which activates only leg movements, saving of any fraction of time is important. However, the insect escape response is complicated and it involves more than mere activation of leg motoneurones, a complex of stereotyped activities designed to bring the animal in a state of alert, including antenna1 movements and the cessation of on-going activity. With a complicated response of this magnitude it is possible that the role of the giant axons is to rapidly co-ordinate and simultaneously activate several functional systerns. A further cautionary note appeared when several workers (Roeder, 1967; Hughes, 1965) indicated the lability of the connection between giant axons and motoneurones: more than one volley is usually needed to induce firing of the leg nerve. It is strange to find the necessity of temporal summation in a system supposedly evolved to transmit information to leg motoneurones in the shortest time. These considerations led us to reinvestigate the mode of activation of the leg motoneurones during escape. Stimulation of the cerci with air puffs initiates activity both in the

124

I. PARNAS AND D. DAGAN

giant axons and also in an unknown number of small axons (Roeder, 1948). We have recently demonstrated that these small axons transmit information to leg motoneurones and that the giant axons do not have excitatory effect on leg motoneurones (Dagan and Parnas, 1970).

Fig. 18. Set-up for experiments dedcribed in Figs 20, 21, 22,24 and 26. Thoracic ganglia marked T , -T, ,abdominal gangliaA , -A6.S t , -Sf,,extracellularhook stimulation electrodes. R hook electrode for extracellular recording. R , suction electrode for recording. Sf,-R, microelectrode for recording and stimulation, and R , pointed stainless steel electrode to record activity from antenna1 muscles. (From Dagan and Parnas, 1970.)

This conclusion is backed by the following experimental results done with the experimental set-up described in Fig. 18. When the nerve cord was stimulated at low intensities, the giant axons were activated but no response was observed in leg nerve N, (Fig. 19a). Only when stimulus intensity was increased to such a degree that small axons in the abdomen were activated, a synchronized response appeared in N, (Fig. 19 b, c). To evaluate the conduction velocity of the activating pathway, stimulating electrodes were placed at A, -A6 and A, -T3 while the responses were recorded at Al -T3 and from the leg nerve N, (Fig. 20). From the difference between latencies of the N, response the conduction velocity was calculated to be 1.5-3.5 m/s, which is half the conduction velocity of giant fibres. Further evidence to show that N, activation is induced by a pathway other than that of the giant fibres was obtained by repeating the abovedescribed experiment on cockroaches whose giant axons were degenerated previously (see p. 106). In these animals evoked activity of the c r u d nerve was indistinguishable from that of normal cockroaches (Fig. 2 1). Another result further confirming the “slow

FUNCTIONAL ORGANIZATION OF GIANT AXONS

a

125

b

L

C

d

Fig. 19. Evoked potentials recorded from abdominal cord (lower trace) and leg nerve N, (upper trace). a, At low strength of stimulation, note giant fibre responses in abdomen but no responses in leg nerve. b-d, Gradual increase of stimulus-strength; an evoked response is observed in N, together with potentials of small fibres in abdomen. In d, a spontaneous response is observed before the evoked one in N, . Cal.: 0.4 mV,4 ms. (From Dagan and Pamas, 1970.)

pathway hypothesis” was obtained after application of 5/ml nicotine on the abdominal cord. After this treatment conduction of small abdominal fibres is blocked (Spira et al., 1969a) and consequently induced activity in N5 is eliminated while the giant axon response is unaffected (Fig. 22). It is not clear whether or not the slow axons are blocked at synaptic sites in the abdominal ganglion, but the relatively fast conduction velocity of the slow system responsible for activation of motoneurones mitigates this possibility. Despite this, for the sake of generality, and recognizing the incompleteness of the evidence,

126

1. PARNAS AND D. DAGAN

b

Fig. 20. Conduction velocity of the abdominal pathway inducing the response in N, . Upper trace response from N,, lower trace response from T,-A, . a, Responses to stimulation at A, -A,. b, Responses to stimulation at T,-A, . Distance between the two stimulating electrodes 12 mm, difference in delays for response in N, , 6 ms, conduction velnritv nf ahdnminal n a t h w a v indiirinu t h e rpcnnncp in N

3 m/c Cnl

. 0 3 mV

10 ms

1

Fig. 21. Evoked potentials recorded in N, (upper trace) and abdominal cord flower trace) in normal (a) and cords with degenerated giant fibres (b). Note responses in N, even when giant fibre responses are completely absent in (b). Cross-sections of control and degenerated cords from which the recordings were made are shown in (c) and (d). Cal.: 0.2 mV, (a), 10 ms, (c), (d), 100 p . (From Dagan and Parnas, 1970.)

FUNCTIONAL ORGANIZATION OF GIANT AXONS

127

b

d

C

e

L

I

f

-+----

Fig. 22. Evoked potentials recorded from N, (upper trace) and abdomen (lower trace) to stimulation at A,-A, (left column) and T , - A , (right column) in normal and nicotine treated cords. Nicotine lo-’ mg/ml w a s topically applied to ganglia A,-A, between stimulating electrodes, dotted area in scheme. a, b, Control: note diferences in delay of N, response in a and right b. c, d, Three minutes after lo-’ mg/ml nicotine; note blockage of small axon responses in abdominal recording and lack of an N, response to stimulation caudally to the blocked region. e, f, Recovery after washings. Cal.: 1 mV, 10 ms.

128

I. PARNAS AND D. DAGAN

synapses are included in this pathway in our final scheme (Fig. 29). Finally a negative result supports these previous findings: activation of single giant axons by an intracellular microelectrode failed to induce leg movements neither to single nor to repetitive stimuli at rates of up to 200/s for durations up to 1 s, while rapid leg movements were observed following a single extracellular stimulus of sufficient intensity. This result shows that activity of a single giant axon is insufficient t o evoke leg movements, but, the results discussed previously rule out also convergence of giant fibres on leg motoneurones. It is thus possible to conclude that activation of leg motoneurones is not mediated by the abdominal giant fibres but rather by a pathway made up of small axons with a conduction velocity of up to 3.5 m/s. In conformity with this result, Cook (195 1) holds that escape in locust is not associated with activation of abdominal giant fibres. His results are based on the finding that cercal stimulation evoked giant axon activity without causing escape. Since the giant axons apparently never activate leg motoneurones in the cockroach, the possibility of a dual system of activation either by giant fibres or by small axons would appear to be absent. It should be noted, however, that such a situation apparently exists in the escape reflex or tail-flip of the crayfish. Rapid and powerful tail flexion can be produced in crayfish without mediation by either medial or lateral giant fibres (Krasne and Wine, personal communication). Krasne and his collaborators found that very sudden tactile stimuli, but only such stimuli excite the giant fibres. Rostra1 stimuli excite the medial, while the lateral giants are stimulated by caudal stimuli. Visual stimuli and non-abrupt somatic stimuli cause identical tail-flip escape responses which are not mediated by giant fibres. An analogous situation apparently also holds in the escape reflex of the polychaete Branchioma vesiculosum (Krasne, 1965). VI. AFFERENT INPUTS

We now return to the problem raised of the possibility of physiological bidirectional conduction by the giant axons. In the scheme suggested by Roeder (1 948), the input synapse in A, and the postulated output synapses in the thoracic ganglia are in accordance with the classic concept of unidirectional axonal conduction. The postulated thoracic synapses serve as valves to assure unidirectional conduction of ascending information. The bidirectional continuous

FUNCTIONAL ORGANIZATION OF GIANT AXONS

129

giant axons postulated by Hughes (1 965) for a hypothetical ancestral form are illustrated in Fig. 2c. It is clear that with such an organization, information entering a giant axon, along its course will spread in both directions. Simultaneous inputs at one end and any other point will cause extinction of the two volleys. This possibility was already discussed by Hughes in his 1965 review. Our experimental findings in Orthroptera suggest a similar situation if indeed several sensory inputs feed into the giant axons. However, it is quite possible that the narrowings that the giant axons form while passing through a thoracic ganglion serve as valves or filters to direct information coming from thoracic inputs at high frequencies in the up-going direction and at low frequencies in both directions. An alternative solution could be a one-point entrance to the giant axons, e.g. the caudal end. In this case the giant axons will serve to transfer only ascending information. A. CERCAL INPUTS

About 150 axons converge from the cercal nerves to synapse on to the giant fibres (Roeder, 1967). These synapses have been studied in detail by Yamasaki and Narahashi (1960) and Boistel and his collaborators. In recent reviews, Boistel (1 968) summarized in detail the topic of synaptic transmission in insects. Activation of the cercal nerves causes an excitatory post-synaptic potential (EPSP) recorded probably from a giant fibre (Callec and Boistel, 1965). The EPSP’s show temporal and spatial summation and a spike is induced at a threshold depolarization. The ipsilateral and contralateral inputs from the two cercal nerves show similar properties. The transmitter substance at this synapse has not been fully elucidated (see the 1966 review of R. Werman for criteria for identification of transmitter substance in the central nervous system), but, acetylcholine is the best candidate for this transmitter. Thus, the synapse is rapidly blocked by nicotine (Roeder, 1948; Spira ef aL, 1969a). Eserine and other anticholinesterases and atropine block conduction through this synapse. B. INPUTS AT ABDOMINAL AND THORACIC GANGLIA

In cockroach, the only inputs t o abdominal giant fibres found so far, are from the cerci. Although sought for, no inputs could be found at the abdominal or thoracic ganglia. Activity was recorded at AIP-7

130

I. PARNAS AND D. DAGAN

the abdominal cord and all of the sensory nerves leading into the metathoracic ganglion were stimulated with supramaximal shocks, using both a single pulse or trains (to provide for the possible requirement of temporal or spatial summation). In no case was a giant fibre response recorded in the abdomen. This lack of activation cannot be attributed to the filtration characteristic of the neck as it was shown (p. 115) that it acts as a low pass filter. The only descending responses recorded were those of small fibres with a conduction velocity of 3.5 m/s (Fig. 23(b)). On the other hand giant

I

L

Fig. 23. (a) Evoked potentials recorded at abdominal connectives (lower trace) and thoracic connectives (upper trace) to crural nerve activation. Note that giant axon responses were recorded only from the thoracic connectives. (b) Determination of conduction velocity of the pathways activated by N, stimuli in abdominal connective. Evoked potentials were recorded at two points on the abdomen 11 mm apart. From the time interval between the fust responses in each trace (3.0 ms) conduction velocity has been calculated to be 3.6 m/s.

fibre ascending volleys are recorded from the thoracic connectives (Fig. 23(a)). Since the thoracic connectives contain many giant axons, it is probable that stimulation of the sensory axons activated only thoracic giant axons. In this connection the work of Maynard (1 956) should be recalled, where stimulation of the antennae activated only thoracic giant axons. VII. GIANT FIBRE OUTPUTS

Since neither abdominal nor thoracic motoneurones are activated by abdominal giant axons, there are two logical places to look for outputs of giant fibres: one in the head where the abdominal giant axons terminate, the second in interneurones of the thoracic ganglia where the giant axons give off collaterals (Farley and Milburn, 1969). A. OUTPUT TO ANTENNAL MOTONEURONES

Behaviouml experiments with cockroaches during evasion showed that cercal stimulation is associated not only with leg movements,

FUNCTIONAL ORGANIZATION OF GIANT AXONS

131

but also with a forward thrust of the antennae. Furthermore the normal cockroach avoids obstacles in flight from noxious stimuli. It is quite possible that this forward thrust of antennae is a response to leg movements, which in that case should be activated first. However, cockroaches, whose nerve cord was cut between ganglia A5-A6,and the giant fibres given sufficient time t o degenerate, run when stimulated by touch at the abdomen, without the forward thrust of the antennae and continually bump into obstacles. It seems therefore that activation of the antennal muscles is not dependent solely on leg movements and is somehow connected to the function of abdominal giant fibres. Indeed when the giant fibres were stimulated alone, the antennal muscles were activated while the crural nerves remained silent (Dagan and Parnas, 1970). When both giant and small axons in the abdomen were activated, the N, nerve fired as well (Fig. 24). If the forward thrust of the antennae is associated with the escape response, it should occur prior to, o r together with, the leg movements in order to be most effective. As Fig. 24 demonstrates,

Fig. 24. Evoked potentials recorded at the base of an antenna (upper trace) and at N, (lower trace) to stimulation at A, A , . First response from antennae is marked by a dot. Note that the responses at N, and the antennae appear with the same delay. Cal.: 1 mV, 10 ms. (From Dagan and Parbas, 1970.)

132

I. PARNAS AND D. DAGAN

when activity was recorded at N5 and antennal muscles, although the distance from the stimulating electrode at A5-A, to the antennae is twice as long, as that to the crural nerve, activity appears at both places at the same time. It should be noted that while activity at the base of antennae was recorded from the muscles, in the leg activity was recorded from the nerve. If we allow 5-8 ms for conduction in N5 and synaptic delay at the neuromuscular junction (Roeder, 1948; Hughes, 1965) the forward thrust of the antennae will occur prior to leg movements well into the pre-evasive period. We do not know if the giant axons synapse directly on to the motoneurons of the antennal muscles or whether they activate the muscles through interneurons. In our final scheme (Fig. 29), we therefore show the giant axons as activating the antennal musculature via a black box. B. EFFERENT ACTIVITY OF GIANT AXONS IN THE METATHORACIC

GANGLION OF THE COCKROACH

The role of the giant fibre branches demonstrated histologically by Farley and Milburn (1969) remains to be elucidated. We have shown that these neither provide excitatory inputs to leg motoneurons (as previously thought) nor are they a sink for sensory information from leg sensory organs. Werman ( 1968, personal communication) suggested that all other on-going activity that might interfere with the escape must be inhibited just prior to initiation of the complicated behaviour called the escape response (Roberts, 1968; Parnas et al., 1969). It was speculated that the giant fibres might be assigned this “clear-all-stations” function. The high velocity would serve to inhibit all ongoing activity just prior to the slower signals which signal the location and possibly the nature of the noxious stimulus. Two approaches were taken in studying this problem. First, intracellular recordings were made from neurons in the metathoracic ganglion. These neurons were classified either as motor or interneurones on the basis of antidromic stimulation of leg nerves. Since both sensory and motor nerves were stimulated, the classification was based on delay and size of intracellular recofd responses, and responses to membrane potential setting via a bridge connected to the microelectrode. The experimental set-up is shown in Fig. 25. At times the neurons were stimulated through the recording electrode and responses recorded in the leg nerves. None of

FUNCTIONAL ORGANIZATION OF GIANT AXONS

133

B ! I

!

A Fig. 25. Schematic set-up for microelectroderecording at T,. St, extracellular electrode for stimulating abdominal connectives. St, suction electrode to stimulate N, . R , and R, hook electrodes to record activity at abdominal connectives and N, respectively. ME, Microelectrode connected to a bridge for recording and current passing.

the motoneurones thus far examined responded to abdominal giant fibre single pulse stimulation or after trains of impulses, but action potentials could be elicited in them upon activating the slower conducting abdominal pathways. Since the synaptic region may be remote from the cell body or recording site, it is possible that some minor synaptic effects, too small to detect, were elicited. But these must either be inhibitory or if excitatory, far below threshold, since summation to firing levels were never seen. On the other hand, several interneurones that also receive synaptic inputs from the crural

134

I. PARNAS AND D. DAGAN

nerve stimulation-probably sensory in origin-responsed to giant fibre stimulation with a spike (Fig. 26). Many details must still be clarified before one is able to draw a map of exactly where the giant fibres act and to adequately describe the relationships between identified inter-motoneurone pools (Cohen and Jacklet, 1967) and specific giant fibres. This functional-anatomical mapping is currently under way in both our laboratory and at the University of Oregon (Rowe et aZ., 1969 and personal communication).

Fig. 26. Evoked potentials recorded from the abdomen (middle trace), N, -upper trace, and intracellular recording from an interneuron in T, , lower trace. A gradual increase in stimulus intensity to the abdominal cord, resulted in synaptic and spike potentials at the interneuron (A-B). Note that at this stage, no activity is recorded from N, . Sensory axom in N, also induced synaptic and spike potentials in the same interneuron. Cal.: 10 mV, 1 ms for lower trace.

Another approach undertaken to study the possible inhibitory function of giant fibres is based on external recordings of the crural nerve activity as evoked by descending information and its modification by ascending giant fibre activity. By computing the mean evoked responses in the crural nerve, alternately with and without concurrent giant fibre stimulation, it could be shown that giant fibre activity diminished the size of the response evoked in N,. Furthermore, this inhibitory effect disappears when the experiment w/v is repeated after bathing the metathoracic ganglion with picrotoxin (Fig. 27). Following sectioning of the thoracic cord between T3 and T2, cercal grooming activity can be monitored from the crural nerve of T3 (Eaton and Farley, 1969). During this activity which seems to be nonevasive, the firing of a certain unit can be inhibited completely by a train of giant axon spikes (Fig. 28). Further experiments are being made on this system t o help elucidate the inhibitory mechanism of the giant axons.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

135

Fig. 27. Effect of abdominal giant axon’s stimulation on compound potentials evoked at thoracic connectives and recorded from leg nerve, N, Compound action potentials recorded from leg nerve N, . Each trace is composed of 64 evoked potentials summed by Fabri Tek 1072 computer. A, Control responses to stimulation of the So-T, contralateral connective. B, When both So-T, connective and abdominal giant axon were stimulated, the response recorded from N, is reduced in magnitude. (It should be remembered that stimulation of abdominal giant axons by themselves, does not evoke any activity at leg nerve N, J C-D, Responses after application of w/v picrotoxin. C, Control, responses to stimulation of so-T,connectives. D, Responses t o So-T, and abdominal giant axon activation. Note that response in (D) is not reduced in comparison with (C). (Stimuli with and without abdominal giant axons were given alternately to overcome fatigue effects.) Cal.: 5 ms.

.

.

Fig. 28. Inhibition of unit activity in leg nerve N, Upper trace activity recorded from leg nerve N, following section of the cord a t T,-T,. Lower trace recording of giant axon activity at A, -A4. Note that when giant axons were stimulated at A, -A, ,cessation of fuing of unit in N, occurred. Cal.: 50 ms.

VIII. GIANT AXON AND SMALL FIBRE PATHWAY-TIMING RELATIONS

To summarize, it might be valuable to look at the timing mechanism of the giant axons. A noxious signal triggers cercal activity in the cockroach which evokes activity in giant and small axons. The information then sweeps up the abdominal cord in these two parallel channels. A difference, however, of up to 5 ms may

136

1. PARNAS AND D. DAGAN

occur between the arrival of the giant axon in the metathoracic ganglion where it “prepares” by exerting an inhibitory effect on on-going activity, and the excitatory signal that arrives via the slower pathway. In the meantime, however, the information continues to ascend the cord via the giant axons and switches on a general alarm system in the head ganglia. This attention arousal occurs almost simultaneously and even precedes slightly the actual motor escape activity of the metathoracic legs, perhaps providing directional signals for movement and other stereotyped activity associated with flight. In conclusion we would like to suggest the following scheme (Fig. 29) for the organization of giant axons and systems involved in escape in the cockroach. IX. POSSIBLE FUNCTION OF AXONS IN INTEGRATION

The structural organization of cockroach giant axons together with their physiological characteristics might have wider implications in the elucidation of mechanisms involved in the spread of information through several terminals of one axon. Figure 30 shows a scheme in which terminals of one axon innervate several post-synaptic cells. Such cases are known to occur in invertebrates and are the rule in the CNS of vertebrates (Bullock and Horridge, 1965; Eccles, 1964). An important example of this phenomenon occurs in the neuromuscular systems of crustacea, where a single axon may provide all of the excitatory input to more than one muscle. For example, the opener and stretcher muscles in the claw of crayfish are innervated by a single excitatory axon (Wiersma, 1961) and the deep abdominal extensor muscles in crayfish and lobster, share common excitatory and inhibitory innervation (Parnas and Atwood, 1966; Atwood, 1967). In spite of this, the opener and stretcher muscles, or in general the post-synaptic cells which are innervated by a single axon, can be activated separately, even without the superimposed activity of inhibitory nerves. Since an axon conducts an all-or-none spike, it is usually assumed that the information reaching the different presynaptic terminals is the same, i.e. that the same number and pattern of spikes reach all terminals. Thus, if different information should reach each of the post-synaptic cells, the integratory mechanisms must be ascribed to the synaptic regions and/or to the properties of the post-synaptic

FUNCTIONAL ORGANIZATION OF GIANT AXONS

9c t

n

I37

P

Fig. 29. Organization of abdominal giant axons and small pathways involved in evasion, in the CNS of the cockroach. (a) General scheme. The giant axons slightly narrow while passing through the abdominal ganglia, and isthmuses are formed at thoracic ganglia. The dorsal giant axons narrow in one step at ganglion T, and they do not form isthmuses at T,-TI. Note the asymmetrical arrangement of the isthmuses in thoracic ganglia, at the caudal end of each thoracic ganglion, the change in diameter of the axon is abrupt. In the head, the giant axons terminate on a black box, activating the antenna1 muscles. The small pathway responsible for the activation of leg motoneurones, is shown to form synapses at abdominal ganglia. It is not known if the same small axon pathway activates all three ganglia, or if separate pathways reach each thoracic ganglion. (b) Connections of giant axons and small axon pathway in the metathoracic ganglion. A “black box” denotes the complex of neurones connected to the leg motoneurones. Note that both ascending small axons and descending axons terminate with excitatory synapses on the “black box”. The giant axon sends an excitatory co-lateral to interneurones which has inhibitory action on the ‘black box” (black triangle). The same inhibitory interneurone is activated by sensory axons (broken line) from the leg.

138

1. PARNAS AND D. DAGAN

Fig. 30. Mode of control of several post-synaptic cells innervated by a single axon. Terminal a, releases large amounts of transmitter per impulse, and cell A is “sensitive”. Terminal b, releases small amounts of transmitter per pulse, and cell B is %on-sensitive”. Terminals a and b, each could represent a cluster of terminals on the post-synaptic cells, and a, and b, can be seen as main branches. Note the “narrowing” in branch a, which blocks conduction in line a-A at high frequencies. For further detail, see text.

cells. However, such integration would only explain differences in responses of the cells, not independency of operation. It is known that cells which are innervated by a common axon can show different cable properties (Atwood, 1963, 1965, 1967). Such post-synaptic cells, therefore, will be activated at different firing frequencies of the pre-synaptic axon, depending on temporal summation and respective thresholds for contraction or spike initiation. Thus, marked differences could be exhibited by combination of post-synaptic factors (effective integratory surface and threshold for spike initiation) and properties of the terminals (different degrees of facilitation, richness of innervation). Different terminals of a given axon are preferentially invaded by spikes (Bittner, 1968) and the amount of transmitter released per-impulse can differ (Atwood and Parnas, 1968; Bittner and Harrison, 1970). However, both Atwood and Parnas (1968) and Bittner and Harrison (1970), assumed that the same number and patterns of spikes pass the main axon and main branches, therefore, the above-mentioned

FUNCTIONAL ORGANIZATION OF GIANT AXONS

139

mechanisms are not sufficient to provide satisfactory explanation to enable each of the post-synaptic cells to operate independently. To show how this ,might be accomplished we will analyse the scheme in Fig. 30. Here terminals (a) and (b) innervate cells (A) and (B) respectively ((a) and (b) can be looked at also as a group of terminals on each post-synaptic cell). Terminal (a) releases large quantities of transmitter per-impulse, the post-synaptic cell (A) will be termed “sensitive” because it reaches threshold at low frequencies (either because of low threshold or because of high input impedance). Terminal (b) releases small amounts of transmitter per-impulse and cell (B) is “non-sensitive”. At low frequencies of the pre-synaptic axon, only line a-A will operate. At higher frequencies it is quite possible that line a-A will respond maximally while line b-B will only begin to respond. Still, higher frequencies will now induce line b-B to fire maximally, while line a-A is already firing at its maximal rate, and such prolonged activity at high frequencies may make this line susceptible to fatigue. Note that, thus far, only a-A, could operate independently, while line b-B will always fire together with line a-A. However, to enable line b-B to fire alone, we have only to introduce a low puss filter into line a-A. In this case line a-A will operate alone at low frequencies of firing at the pre-synaptic cell, at intermediate frequencies both lines will operate and at high frequencies line b-B will fire alone. It is possible that terminals of the same axon, or even its main branches have different diameters or show pronounced “narrowings” in their courses. Such “narrowings” could serve as filters and at different frequencies at the main axon, some branches or terminals could pass information while others would not. The relative change in diameter along an axon branch will determine the effective range of that particular filter. It is therefore possible that integration in its widest sense takes place not only at the level of synapses, but along the axons as well. ACKNOWLEDGEMENTS

The authors wish to thank Professor R. Werman for his critical reading of the manuscript, and for the many helpful discussions throughout the writing of this review. We would also like to thank Dr. I. McCance for correcting the manuscript, Mrs. Z. Shapira and Miss I. Harrari for technical assistance in histology, and Mr. A. Gilai for drawing the schemes.

140

I. PARNAS AND D. DAGAN

REFERENCES Atwood, H. L. (1963). Differences in muscle fibre properties as a factor in “fast” and “slow” contraction in Carcinus. Comp. Biochem. Physiol. 10, 17-3 2. Atwood, H. L. (1965). Characteristics of fibres in extensor muscle of a crab. Comp. Biochem. Physiol. 14,205-207. Atwood, H. L. (1967). Crustacean neuromuscular mechanisms. A m . 2001. 7 , 527-55 1. Atwood, H. L. and Parnas, I. (1968). Synaptic transmission in crustacean muscles with dual motor innervation. Comp. Biochem. Physiol., 27, 381-404. Baylor, D. A. and Nicholls, J. F. (1969). Changes in extracellular potassium concentration produced by neuronal activity in the central nervous system of the leech. J. Physiol. 203,555-569. Bittner, G. D. (1968). Differentiation of nerve terminals in the crayfish opener muscle and its functional significance. J. gen. Physiol. 5 1, 73 1-758. Bittner, G. D. and Harrison, J. (1970). A reconsideration of the Poisson hypothesis for transmitter release at the crayfish neuromuscular junction. J. Physiol. 206, 1-23. Boulton, P. S. (1969). Degeneration and regeneration in the insect central nervous system, I. 2. Zellforsch. 101,98-118. Boulton, P. S. and Rowell, C. H. F. (1969). Degeneration and regeneration in the insect central nervous system, 11. 2 . Zellforsch. 101, 119-134. Boistel, J. and Pichon, Y. (1969). Aspects ioniques et electriques du fonctionnement des fibres nerveuses d’un insecte, la Blatte, Periplaneta americana. Actual. neurophysiol. (In press.) Boistel, J. (1968). The synaptic transmission and related phenomena in insects. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V.B. Wigglesworth, eds), pp. 1-57. Academic Press, London and New York. Bullock T. H. and Horridge, G. A. (1965). “Structure and Function in the Nervous System of Invertebrates”. W. H. Freeman, San Francisco and London. Bullock, T. H. (1944). The giant nerve fibre system in Balanoglossids. J. Comp. Neurol. 80,355-367. Bullock, T. H. and Turner, R. S. (1950). Events associated with conduction failure in nerve fibres. J. cell comp. Physiol. 36, 59-82. 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, London and New York. Cohen M. J. and Jacklet, J. W. (1967). The functional organization of motor neurons in an insect ganglion. Phil. Trans. R. SOC.B252,263-571. Cook, P. M. (1951). Observations on giant fibers of the nervous system of Locusta migratoria. Q. Jl Microsc. Sci 92, 297-305. Coombs, J. S., Curtis, D. R. and Eccles, J. (1957). The interpretation of spike potentials of motoneurons. J. Physiol., Lond. 139, 193-23 1.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

141

Dagan D. and Parnas, I. (1970). Giant fibre and small nerve pathways involved in the evasive response of the cockroach. Periplaneta americana. J. exp. Biol. 52, 313-324. Eaton, R. C. and Farley, R. D. (1969). The neural control of cercal grooming behavior in the cockroach Periplaneta americana. J. Insect Physiol. 15, 1047-1065. Eccles, J. C. (1964). “The Physiology of Synapses”. 3 16pp. Springer-Verlag, Berlin. Eyzaguirre, C. and Kuffler, S. W. (1955). Further study of soma, dendrite and axon excitation in single neurons. J. gen. Physiol. 39, 121-153. Farley, R. D. and Milburn, N. S. (1969). Structure and function of the giant fibre system in the cockroach, Periplaneta americana. J. Insect Physiol. 15, 457. Fielden A. (1960). Transmission through the last abdominal ganglion of the dragonfly nymph, Anax imperator. J. exp. Biol. 37,4, 832-844. Guthrie, D. M. and Tindall, A. R. (1968). “The Biology of the Cockroach”. Edward Arnold, London. Hanstrom, B. (1 923). Further notes on the central nervous system of arachnids: scorpions, phalangids and trap door spiders. J. comp. Neurol. 35, 249-274. Hess, A. (1958). Experimental anatomical studies of pathways in the severed central nerve cord of the cockroach. J. Morph. 103,479-499. 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. Hodgkin, A. L. and Rushton, W. A. H. (1946). The electrical constants of a crustacean nerve fibre. Proc. R. SOC.B133, 444. Huber, F. (1965). 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. Hughes, G. M. (1953). Giant fibres in dragonfly nymphs. Nature, Lond. 171, 87-88. Hughes, G. M. (1958). The co-ordination of insect movements. 111. Swimming in Dytiscus, Hydrophilus and a dragonfly nymph. J. exp. BioL 35, 567-583. Hughes, G. M. (1965). 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. Jones, D. P. and Singer, M. (1970). Neurotrophic dependence of the lateral newt Triturus viridescens. J. exp. 2001.171,433-442. Krasne, F. B. (1965). Escape from recurring tactile stimulation in Branchiomma vesiculosum J. exp. Biol. 42, 307-322. Maynard, D. M. (1956). Electrical activity in the cockroach cerebrum. Nature, Lond. 177,529-530. Mill, P. J. (1 963). Neural activity in the abdominal nervous system of Aeschnid Nymphs. Comp. Biochem. Physiol. 8,83-98. Nakajima, S. and Takahashi, E. (1966). Post-tetanic hyperpolarization and electrogenic Na pump in stretch receptor neurone of crayfish. J. gen. Physiol. 49, 61 3-628.

142

I. PARNAS AND D. DAGAN

Narahashi, T. (1963). The properties of insect axons. In “Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 1, pp. 175-256. Academic Press, London and New York. Narahashi, T. (1965). The physiology of insect axons. In “The Physiology of the Insect Central Nervous System” (J. W. L. Beament and J. E. Treherne, eds). Academic Press, London and New York. Parnas, I., Spira, M. E., Werman, R. and Bergmann, F. (1969). Non-homogeneous conduction in giant axons of the nerve cord of Periplaneta americana. J. exp. BioL 50, 635-649. Parnas, I. and Atwood, H. L. (1966). Phasic and tonic neuromuscular systems in the abdominal extensor muscles of the crayfish and rock lobster. Comp. Biochem. Physiol. 18, 701-723. Pichon, Y. (1969). Aspects electriques et ioniques du fonctionnement nerveux chez les insectes. Cas particulier de la chaine nerveuse abdominale d’une blatte Periplaneta americana L. PbD. Thesis submitted to University of Rennes. Pichon, Y. and Boistel, J. (1967a). Current-voltage relations in the isolated giant axons of the cockroach under voltage clamp conditions. J. exp. Biol. 47, 343. Pichon, Y. and Boistel, J. (1967b). Microelectrode study of the resting and action potentials of the cockroach giant axon with special reference to the r61e played by the nerve sheath. J. exp. Biol. 47, 357 Pichon, Y. and Boistel, J. (1968). Ionic composition of haemolymph and nervous function in the cockroach, Periplaneta americana L. J. exp. Biol. 48,31. Pipa. R. L., Cook, E. F. and Richards, A. G. (1959). Studies on the hexapod nervous system. 11. The histology of the thoracic ganglia of the adult cockroach, Periplaneta americana (L.). J. comp. Neurol. 113,401-433. Pumphrey, R. J. and Rawdon-Smith, A. F. (1937). Synaptic transmission of nervous impulses through the last abdominal ganglion of the cockroach. Proc. R. SOC.B122, 106-118. Power, M. E. (1948), The thoracico-abdominal nervous system of an adult insect. Drosophila melanogaster. J. comp. Neurol. 88, 347-4 10. Roberts, A. (1 968). Recurrent inhibition in the giant-fibre system of the crayfish and its effect on the excitability of the escape response. J. exp. Biol. 48, 5 45-567. Roeder, K. D. (1948). Organization of the ascending giant fiber system in the cockroach (Periplaneta americana). J. exp. Zool. 108, 243-26 1. Roeder, K. D. (1967). “Nerve Cells and Insect Behavior”. Harvard University Press, Cambridge, Massachusetts. Rowe, E. C. (1969). Morphology of branches of functionally-identified motoneurones in cockroach neuropile. Am. Zool. 9 , Abst. 247. Rowell, C. H. F. (1965). The control of reflex responsiveness and the integration of behaviour. In “The Physiology of the Insect Central Nervous System” (J. W. L. Beament and J. E. Treherne, eds). Academic Press, London and New York. Satija, R. C. (1958a). A histological and experimental study of nervous pathways in the brain and thoracic nerve cord of Locusta migrutoria migrutoriodes (R.C.F.).Res. Bull. Panjab Univ. 137, 13-32.

FUNCTIONAL ORGANIZATION OF GIANT AXONS

143

Satija, R. C. (1958b). A histological study of the brain and thoracic nerve cord of Aeschna nymph with special reference to the descending nervous pathways. Res. Bull. Panjab Univ. 138, 33-48 Saint-Rbmy, G. (1866a). Recherches sur la structure du cerveau du scorpion. C.r. hebd. Sianc. Acad. S c i , Paris, 102, 1492-1494. Saint-Rbmy, 0.(1 866b). Sur la structure des centres nerveux chez le scorpion. Bull. Skanc. SOC.Sci. Nancy, (2) 8 (20), xxix. Spira, M. E., Parnas, I. and Bergmann, F. (1969a). Organization of the giant axons of the cockroach Periplaneta arnericana. J. exp. Biol. 50, 615-627. Spira, M. E., Parnas, I. and Bergmann, F. (1969b). Histological and electrophysiological studies on the giant axons of the cockroach Periplaneta americana. J. exp. Biol. 50, 629-634. 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. Treherne, J. E. (1967). Axonal function and ionic regulation in insect central nervous tissues. In “Insects and Physiology” (J. W. L. Beament and J. E. Treherne, eds), pp. 175-188. Oliver and Boyd, Edinburgh and London. Werman, R. (1966). A review-criteria for identification of a central nervous system transmitter. Cornp. Biochem. Physiol. 18, 745-766. Wiersma. C. A. G. (1947). Giant nerve fiber .system of the crayfish. A contribution to ‘comparative physiology of synapse. Neurophysiol. 10, 23-38. Wiersma, C. A. G. (1 96 1). The neuromuscular system. In “The Physiology of Crustacea” (T. H. Waterman, ed.), Vol. 11, pp. 191-240. Academic P r e s New Y ork and London. Wigglesworth, V. B. (1960). Axon structure and the dictyosomes (golgi bodies) in the neurones of the cockroach, Periplaneta arnericana. Q. J. Microsc. Sci. 101, 381-388. Yamasaki, T. and Narahashi, T. (1959). Electrical properties of the cockroach giant axon. J. Insect Physiol. 3,230-242. 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). Zun Morphologie der Nervenzentren. Des Bouchmark der Insecten. Ein Beitrag zur vergleichenden Histologie. 2. Wiss. Zool. 122, 323424.

NOTE ADDED IN PROOF Since this review was submitted for publication, several findings have been reported which make some of the data given obsolete and add new information. It was thought that the giant axons are supported by a cluster of cell bodies as shown in Fig. 8. Recent studies by Milburn (1971, personal communication) and Smyth (1971, personal communication), after injection of Procion yellow into cockroach giant axons, indicate that each giant fiber has only one cell body and no syncitial giant fibers were observed. The Procion yellow injections also show (Smyth, 197 1, personal communication) that giant axons of the cockroach do branch in the abdominal ganglia,

144

I. PARNAS AND D. DAGAN

however we have n o knowledge as to the role of these branches. These findings with those of Cook (1951) and Seabrook (1970) showing that in the desert locust giant axons have only one cell body throw doubt on the syncitial origin of giant axons in insects. An additional possibility as to integration mechanisms of cockroach giant axons has been found recently (Dagan, 197 1). It was found that following an ascending pulse in the giant fibers, sub-threshold stimuli in the head became effective and evoked descending giant axon responses. This phenomenon was shown not to be a result of synaptic loop activation, but rather that a state of increased excitability exists after an ascending spike. This period of increased excitability lasts up to 120 ms. Similar periods of hyperexcitability attributed to negative after potentials were reported in mammalian C fibers (Gasser, 1941) and crayfish interneurons (Kennedy and Mellon, 1964). Since n o inputs to the giant axons were described so far in the head, this phenomenon might be without physiological implication. However, we found similarly that antidromic activation of the giant axons lowers markedly the stimulus intensity required to evoke a spike in the giant axons, when the last abdominal ganglion A, is stimulated either directly, or through one of the cercal nerves. As previously mentioned giant axons of the cockroach are innervated both ipsi- and contralaterally by the cercal nerves. The excitation of the giant axons by the contralateral cercal nerve requires temporal and spatial summation. It is thus possible to determine stimulation conditions whereby a single pulse to the contralateral cercal nerve will not evoke a spike in the giant fibers. However, if this is preceded by stimulation of the ipsilateral cercal nerve the “ineffective” contralateral stimulus becomes effective. Again the period of hyperexcitability following the ipsilaterally evoked spike lasts about 80-100 ms. This finding might have physiological implications and it is possible that an input through one cercus may modulate the input from the other.

REFERENCES Dagan, D. (197 1). The neural basis for the escape response in the cockroachPeriplaneta americana. Ph.D. Thesis submitted to the Hebrew University of Jerusalem (in Hebrew). Gasser, H. S. (1941). Properties of mammalian C fibers. Ohio J. Sci. 41, 145-159. Kennedy, D. and De Forest Mellon, Jr. (1964). Synaptic activation and receptive fields in cray-fish interneurons. Comp. Biochem. Physiol. 13, 275-300.

The Variable Coloration of the Acridoid Grasshoppers C. H. FRASER ROWELL Department of Zoology, University of California at Berkeley, Berkeley, California Definitions, Terminology, and Taxonomy . . . . . . . . 146 Introduction-Variable Coloration and the Natural History of Grasshoppers . . . . . . . . . . . . . . . . . . 147 . . . . . . . . . . 152 111. Genetic Factors . . . . . . A. Genetic Polymorphism . . . . . . . . . . . . . 152 B. Genetic Modification of Phenotypic Polymorphism . . . 155 . . . . . . . 156 IV. Environmental Factors, . . . . . . . A. The Homochrome Respbnses to Background: The Orange and Black Pigment Systems . . . . . . . . . . . 157 B. The Green/Brown Polymorphism . . . . . . . . . 167 C. Phase Coloration . . . . . . . . . . . . . . . 175 V. Physiological Mechanisms . . . . . . . . . . . . . . 177 A. The Green/Brown Polymorphism and the Corpus Allatum . 178 The Black Pigment System and the Corpus Cardiacum . . 1 8 0 B. C. Other Endocrine Correlates of Pigmentation . . . . . . 181 VI. Pigments . . . . . . . . . . . . . . . . . . . 183 A. The Green Component of the Green/Brown Polymorphism . 184 B. The Brown Component of the Green/Brown Polymorphism, and the Black and Orange Pigment Systems . . . . . ., 186 C. Implications of the above for the GreedBrown Polymorphism . . . . . . . . . . . . . . . . . 188 D. Pattern . . . . . . . . . . . . . . . . . 189 E. The Phase Coloration of Gregarious Locust Hoppers . . . 189 Acknowledgements . . . . . . . . . . . . . . . . . . . 190 References . . . . . . . . . . . . . . . . . . . . . . 190 1. 11.

.

As he smoked, his legs stretched out in front of him, he noticed a grasshopper walk along the ground and up on to his woollen sock. The grasshopper was black. As he had walked along the road, climbing, he had started many grasshoppers from the dust. They were all black. They were not the big grasshoppers with yellow and black or red and black wings whirring out from their black wing sheathing as they fly up. These were just ordinary hoppers, but all a sooty black in colour. Nick had wondered about them as he walked, without really thinking about them. Now, as he watched the black hopper that was nibbling at the wool of his sock with its fourway lip, he realized that they had all turned black from living in the burned-over land. He AIP-8

145

146

C. H. FRASER ROWELL

realized that the f i e must have come the year before, but the grasshoppers were all black now. He wondered how long they would stay that way. Carefully he reached his hand down and took hold of the hopper by the wings. He turned him up, all his legs walking in the air, and looked at his jointed belly. Yes, it was black too, iridescent where the back and head were dusty. He tossed the grasshopper into the air and watched him sail away t o a charcoal stump across the road.

Big Twehearted River (Ernest Heminpay, 1963, with permission of the publishers)

I. DEFINITIONS, TERMINOLOGY, AND TAXONOMY

Coloration includes both colour, which describes the reflectance of the pigments of a given area, and pattern, which describes the distribution of pigments. Variable coloration in this review is understood to describe the possibility of different colorations existing in different individuals of the species who are otherwise similar in age, sex, and maturational state. It excludes differences associated with regular sexual dimorphism, sexual maturation, or changes in coloration characteristic of different stadia, but invariable within those stadia. For example Phyteumus purpurascens (Pyrgomorphidae) has a typical and different coloration for almost every one of its seven stadia, but all individuals of the same age and sex are similar; this is not an example of variable coloration. In other words, only simultaneous polymorphism is here considered. The polymorphism may be due to differences in the genotype or to phenotypic differences derived by the interaction of a given genotype with varying environments. In the grasshoppers, variation of this sort is usually long-term, there being almost no short-term colour change. The exception to this is the temperature-dependent colour change described by Key and Day (1954a, b) from Kosciuskola (Catantopinae) and a few other genera, and this is excluded from the discussion below. The Acridoidea are here understood and classified as by Uvarov (1966, p. 397). Most of the specific examples cited are European or African forms, and authors for generic or specific names are not given, as these can be obtained from either Uvarov (1 966) or Dirsch (1964). The names of families and subfamilies given after cited genera are abbreviated after the following scheme: Pyrg. Hemiacr. Trop.

Pyrgomorphidae Hemiacridinae Tropidopolinae

Euryph. EY Pr. Catan t.

Euryphyminae Ey prepocneminae Cat an topinae

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

ox. Copt. Cyrt. Callipt.

147

Acrid. Acridinae Oxyinae Oed. Oedipodinae Coptacridinae Tru x a linae Cyrtacanthacridinae Trux. Gomph. Gomphocerinae Calliptaminae thus, Heterucris vinuceus (Eypr.).

The review of the literature on which this article is based was completed in December, 1969. 11. INTRODUCTION-VARIABLE COLORATION AND THE NATURAL

HISTORY OF GRASSHOPPERS

It is common knowledge, among naturalists and entomologists in temperate regions, that the coloration of acridoid grasshoppers is very variable. This usually reflects the realization that coloration is not an adequate guide to the taxonomy of temperate grasshoppers, in the way that it is for temperate Lepidoptera. To some extent this prevalent view reflects the preponderance in the North Temperate of Oedipodine and Gomphocerine genera, which are particularly variable. In all subfamilies there are, however, species which show little or no variation, and in many divisions, particularly among the tropical fauna, there is a preponderance of invariant species. The Pyrgomorphidae, Lentulidae, Hemiacridinae, Coptacridinae and Eyprepocnemidae of Africa afford examples of such mainly invariant groups. The majority of grasshopper species pass through at least two different colorations in their life-history, often changing from one to the other at the final moult, and the larval and adult colorations, as would be expected from evolutionary theory, often show divergent specializations. Such divergence includes variation as here understood; for example, most Cyrtacanthacridinae have larvae which show one or more sorts of simultaneous polymorphism, especially green/brown polymorphism and occasionally phase polymorphism; the adults are much more invariable, losing the green/brown polymorphism completely in most cases and showing phase effects on coloration to a much reduced extent. The selection pressures which act in coloration are unendingly diverse, and the balance between variable and invariable coloration in any given life-stage of grasshopper species must be dictated by their resultant. It is, however, possible to list common associations between variant and invariant coloration and various other

148

C. H. FRASER ROWELL

specializations which the different species show. These associations are observable empirically, but would also be expected on theoretical grounds. Variable coloration is associated with (i) Geophilous habit (ii) Grassland habitats (iii) Temperate or alpine habitats (iv) Phase transformation (v) Acoustic epigamic displays. The first two of these result in environments in which the prevailing colour of the background is subject to seasonal change or local variation, and the variation found in’the acridoid inhabitants is such as to make them predominantly cryptic. Grasslands change seasonally not only from green to brown with the wet and dry seasons, but are also subject to grass fires which transform the environment to predominantly black, while the substrate colour available to a geophilous species changes sharply with local geology and humidity. Relatively few species of grasshopper inhabit temperate and alpine zones; one presumes that interspecific competition is less and each is found in a greater diversity of habitats than in the lowland tropics. The capacity for variation is accordingly of greater selective value. The colour variation brought about by phase transformation differs sharply from the first three categories, in that it seems to make the individual less rather than more cryptically coloured. The selective advantage of this is obscure, but it seems possible that conspicuous coloration of gregarious hoppers facilitates visual responses which make possible aggregation into bands and their subsequent maintenance. Acoustical interspecific communication reduces the importance of species-specific coloration and of associated visual displays, and thus the selection pressure against variable coloration. Invariant coloration is associated with a complex of factors which complement the above. These are: (i) aseasonal or otherwise colour-stable habitats, including especially tropical wet forest and swamp, and excluding most other grasslands; (ii) intraspecific visual signals, including epigamic displays and patterns promoting social cohesion in obligately social forms; (iii) interspecific visual communication, including aposematic and possibly mimetic coloration, “flash” coloration of rear wings, etc.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

149

It is obvious that many of these factors are either not strictly independent, or if they are, are not exclusive. Most grasshoppers have coloration elements which can be attributed, in an evolutionary sense, to several of the factors listed above, and particular combinations are especially common in certain habitats. This is perhaps best illustrated by example, a number of which are given below. a. Cryptic grassland forms. These genera are usually elongate slender animals, often sluggish or slow moving, sometimes with reduced or absent wings. They show more or less striking resemblances to the plants on which they live. They usually have a well-developed green/brown polymorphism and a rapid and effective homochrome response to black backgrounds, but lack the orange pigment homochrome systems which are associated with a geophilous habitat. Some (e.g., Acrida. Gznnula) have pattern polymorphism, probably allelomorphic, which results in either longitudinal striping (disruptive coloration), or symmetric or asymmetric blotching with yellow pigment in the black homochrome form, which gives a striking resemblance to charred grass fragments. Most have well-developed acoustical signals, either stridulatory as in the relevant Truxalines, Gomphocerines and Hemiacridines, or. by mandible clicking, as in Acrida. Where there are also visual displays, the associated specialized colour areas are confined to areas which are normally hidden in the resting position of the animal, such as the blue and red patches on the inner surface of the metathoracic femora of Mesopsis laticornis (Gomph.). In species which fly, the hind wings often provide a flash colour as in Acrida (Acr.) and Truxalis (Trux.). Other representative genera are Amphicremna, Machaeridia, Cannula (Acr.); Mesopsera (Catant.). As exceptions to the above generalizations it may be noted that the Hemiacridines Leptacris and Acanthoxia apparently lack the green morph. This group makes an interesting contrast with the small number of genera typical of a very different grassland habitat, that of tropical swamp. These species are more active and less morphologically specialized for a cryptic appearance. Their environment is almost completely stable in colour. They themselves show very little variation. They are green, without the brown morph; show no homochrome response to background; and often have an invariant linear pattern. Examples are seen in Oxya hyla (Ox.), and Paracinema tricolor (Oed.).

150

C. H. FRASER ROWELL

b. Gramnivorous geophilous forms. These show no extreme morphological specializations, but are perhaps the most prone to coloration polymorphism. They show the full range of homochromic responses: black and orange pigment systems, a well-developed green/brown polymorphism, and often considerable allelic polymorphism, involving both specific features such a metathoracic tibia1 colour and the whole patterning of the individual. They are often extremely cryptic. Associated with this they have well-developed stridulatory and other acoustic mechanisms; where these are supplemented by visual displays at close quarters, the displays are either purely dynamic, or involve colours not normally visible, such as the back wings. The group is most typically represented by the Oedipodinae; the genera Chortoicetes, Gastrimargus, Locusta, Locustam, Oedipoda, and Oedaleus account for most of the experimental work on environmentally controlled colour variation. Many of the Gomphocerinae could also be included, though the group is a much more diverse one and less specifically geophilic in its colour adaptations. The best known members of this group are the temperate, rather unspecialized, genera such as Chorthippus, where variation is predominantly genetic. The extreme colour variation of Afroalpine forms, such as the related Dnopherula werneriana, or of Coryphosima amplificata and Uganda kilimanjarica (Acrid.), all of which are dominant species with few competitors over large areas of montane grassland, is probably of the same type. The variation seen in tropical lowland species of the same genera is less. c. Geophilous forms eating broad-leaved plants. This is not a large category, but it makes an interesting contrast with the above. Examples are Chrotogonus (Pyrg.) and Trilophidium (Oed.) and possibly Gemeneta (Catant.). These animals are almost never seen on any substrate other than bare earth; they feed on the edges of low-growing herbs, while still remaining on the earth themselves. They give excellent homochrome response to background with both the black and orange pigment systems, but to my knowledge lack completely any green morph. d. Wingless high alpine forms. A number of genera of different families living in the Afroalpine habitat have lost or greatly reduced their wings. The larvae fulfil the expectation that they will be of variable coloration, in view of their relatively uncolonized habitat. Thus the larvae of Occidentosphem and Parasphena (Pyrg.) and of Pezocatantops (Catant.) are green/brown polymorphic; in Parasphena they are additionally able to make homochromic response to

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

15 1

background with black, orange and possibly purple pigment systems, while Pezocatantops sometimes also assumes the quite different adult coloration in the mid-larval stadia. As adults, however, they are invariable with bold patterns and quite bright colours. Presumably visual signals are important in mating, as none stridulate. The brachypterous subalpine and high montane species of Phymeurus (Euryphyminae) are similarly silent and use visual epigamic displays, though in this case associated coloration is confined to the inner surface of the metathoracic femora. They differ from the preceding genera in several ways. All stages are dully coloured and obscurely patterned in grey and black and relatively invariable, though undoubtedly cryptic. e. Wingless tropical forest forms. These are a strikingly uniform group, derived by convergence from many different subfamilies. They tend to be brachypterous or apterous, forb-feeding, scuttlers rather than jumpers, brilliantly and invariantly coloured, and having visual, not acoustic, epigamic displays. They are perhaps at their best in wet lowland secondary forest in Africa; a selection of flightless forms from there would include Pterotiltus (Ox.); Cyphocerastis, Paracoptacra, Ruwenzoracris (Copt .); Heteracris spp. (Eypr.); Serpusia, Auloserpusia and related genera (Catant.); and Odontomelus (Acrid.). Similarly invariable brightly coloured forms from the same forests, but having functional wings, extend both the genus and family lists; typical examples are Heteracris vinaceus (Eypr.); Parapropacris rhodopterus, Orbillus coeruleus (Catant.); and Pachynotacris amethystinus (Cyrt.). The edges of montane forests also contain populations of flightless species in many ways intermediate between this group and the last, such as Eyprepocnemis mo ntigena (Eypr .), Kinangopa jea nn eli, Ixalidium haema toscelis, (Catant.), Coryphosima sp. A (Moroto) (Jago, in prep.) (Gomph.). f. Aposematic coloration of distasteful species, and other invariant patterns with a non-sexual signal function. Some grasshoppers have patterns and colour combinations which are either glaringly conspicuous or are conventionally associated with a mimetic assemblage, such as black and yellow stripes or bands. Some of the brightly coloured forest species of the preceding group may possibly be distasteful, but there is no evidence of this. The only grasshoppers known to be habitually poisonous or distasteful are some members of the Pyrgomorphidae and Romalinae, and those that have been investigated (e.g. Poecilocerus, Phymateus and related genera) derive their active constituents from their food plants

152

C. H. FRASER ROWELL

(Rothschild and Parsons, 1962; van Euw et al., 1967; Reichstein, personal communication and 1967). Not all of these are brightly coloured, and the provenly distasteful Phymateus group are not conspicuously coloured or patterned at all as adults, though when molested they display flash coloration; Dictyophorus has brilliant orange and black stripes as a larva but is totally obscure, apart from its pink back wings, as an adult; and some of the most brilliantly coloured are not known to be poisonous at all. In several of these cases it seems that the coloration may serve signal functions other than or additional to the aposematic, such as group cohesion; see Rowell ( 1967a) for a discussion of these.

111. GENETIC FACTORS

Much of the variable coloration of the Acridoidea is at least partially determined by environmental factors, and shows phenotypic lability more or less independently of genotype. The majority of the experimental work concerns this type of coloration, and this review is primarily devoted to it. However, it is obvious that the limits of this phenotypic lability are set by the genotype; further, some examples of simultaneous polymorphism are purely or largely genetic in character, and the environment influences them little or not at all. Examples of both of these genetic influences are given below. A. GENETIC POLYMORPHISM

The relatively small amount of experimental work which has been performed indicates considerable differences in the extent to which the phenotype is modifiable by environmental factors, and it seems that this condition is commonest among the Acridoidea. In the other groups the indications are that genetic polymorphism is the more important. The extreme case of a polymorphism with an exact correspondence between genotype and phenotype appears among the Tettrigidae. Nabours (1 929) bred Paratettix texanus and Apotettix eurycephalus for over 20 years and more than 60 generations. The species both show an extraordinary range of colour and pattern composed of defined forms, and which does not form a graded series. In Paratettix at least 18 factors were isolated influencing colour and

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

153

pattern. Fisher (1930, 1939) showed from these data that the diversity of the genotype was maintained by selection against the homozygous dominant; the homozygous recessive was not so affected. Further, in wild though not in caged populations, there was strong selection pressure against the pairing of more than one heterozygote in one individual. No environmental effects were found influencing coloration: “neither excessive humidity, temperature, aridity, acidity, salinity, sunlight through glass or direct, darkness, color of soil, food, excreta, starvation, fungus disease, parasitism, nor any other observable feature of the environment has ever changed color pattern to any appreciable extent” (Nabours, 1929). Comparable situations exist in other Tetrigid genera (see, e.g., Good, 194 1 , Tettigidea parvipennis) and in the Eumastacid Morabine grasshoppers (e.g. Lewontin and White, 1960). They have been important in the development of the concept of the super-gene. Relatively little is known of the control of colour polymorphism in the Grylloidea or Tettigonioidea. Within the latter a number of phytophilous families (e.g. the Phaneropteridae, Tettigoniidae and Conocephalidae) include genera in which there is well defined green/brown polymorphism, the basis of which is largely unknown. Verdier (1 958) produced evidence that Barbistites fischeri, Ephippiger provincialis, and Orphania denticauda and 0. scutata reared in captivity were green in isolation, but brown when crowded; this is especially unexpected in E. provincialis, as a green form is unknown in the wild. The constancy of the green and brown proportions in the E. African races of Homorocoryphus nitidulus (Owen, 1965; Karuhize, 1968) argues either a genetically controlled polymorphism or a remarkably consistent environment ; individuals of this species, reared in high humidity, show no colour differences associated with density (Rowel1 and Mukwaya, unpublished). In addition, many of these polymorphic tettigoniid species show occasional purple coloration, affecting either morph. In at least two species this is controlled by a single dominant allele (Amblycorypha oblongifolia, Phaneropteridae, Hancock, 1916; Homorocoryphus nitidulus, Conocephalidae, Rowell and Mukwaya, unpublished). Data on genetic polymorphism of the Acridoidea themselves are disparate and uncommon. It is clear that such polymorphisms are frequent, if not well known. For example, many genera ((e.g. Truxalis (Trux.) and Morphacris (Oed.)) are polymorphic in the colour of their hind wings, and the different geographical populations vary markedly in the relative frequency of these forms.

154

C.

H. FRASER ROWELL

In such widely distributed species, colour and pattern may vary considerably over the range, and on occasion it has been noted that these differences persist for generations under standard culture conditions, indicating genotypic differences (Locusta pardalina, Nolte, 1962; Gastrimargus africanus, Rowell and Hunter-Jones, unpublished). King and Slifer (1 955), investigating the variation in colour of the hind tibia of Melanoplus sanguinipes (Catant.) showed that red colour is due to a dominant autosomal gene, and blue to its recessive allele, though other genes modify these effects, mostly in dilution of the colour. Brett (1947) had previously demonstrated that the proportions of these colours in a population and their intensities are also modifiable phenotypically by food plant, and slightly by temperature and humidity. Many genera include a form (=forma porphyrica of Rubtzov, 1935) in which part of the animal, mostly the dorsal areas of vertex, pronotum and elytra, is purple; European examples are given by Vorontskovskii (1 929), Rubtzov (1935), and Ramme (1951) from the Gomphocerinae, and other subfamilies showing this trait include the Hemiacridinae (Spathosternum), Oedipodinae (Aiolopus) and Acridinae (Duronia, Roduniella, Gymnobothris, etc.) in Africa. In Aiolopus thalassinus this form appears to be controlled by a single dominant allele (Rowell, unpublished observation) as in the tettigoniids noted above, and this may well be the general case. Collectors are familiar with a number of very rare but reproduceable colour variants, which are probably due to recurrent mutation. Thus in the related genera Humbe, Gastrimargus and Heteropternis (Oed.) the normal pronotal coloration may be replaced by a uniform pale yellow, pink, or white; in Acanthacris ruficornis (Cyrt.) the dorsal pronotwn is occasionally dark green instead of sienna; Cyphocerastis sp. A (Jago, in preparation) (Copt.) normally has a clear greenish-yellow pronotum, but occasionally wild individuals with this colour replaced by pink are found. The incidence of these forms has not been properly determined, but in all these examples is certainly well below 0.1%. Substantiated cases of such mu tation are the albino forms reported for Schistocerca gregaria (Cyrt.), Melanoplus sanguinipes (Catant.) and Locusta migratoria (Oed.) (Hunter-Jones, 1957; Putnam 1958; Verdier, 1965, Nolte, 1968), and the melanic form of S. gregaria (Volkonsky, 1938). At a greater level of complexity, Byrne (1962, 1967a, b) showed that nine taxonomically recognized forms (Key, 1954) of Chortoicetes terminifera (Oed.) corresponded to the possible

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

155

combinations of four alleles (i.e. four homozygotes and six heterozygotes), two of which gave an indistinguishable phenotype. These factors affected only the melanic pattern and part of the brown pigment system, probably the ommochrome fraction. In high humidity the animals developed green pigment, and this masked or excluded the other pigments. White (1 968) describes a bilateral gynandromorph of Valanga irregularis (Cyrt.) which expressed two of the six pattern polymorphs of this species. Together with the naturally occurring proportions of the different forms in different localities, this suggests genetic control of colour and pattern in this species. Finally, it seems likely that the best known Gomphocerine genera, such as Chorthippus, may derive their bewildering variation almost entirely from the genotype, after the manner of the tettrigids. There is strictly only one genetic experiment. In Chorthippus parallelus Sansome and La Coeur (1935) were able to separate ten factors affecting colour and/or pattern. Six of these inhibit the expression of at least two nonallelomorphs, and in addition at least two others interfere with the expression of several other genes. The authors point out that one phenotype may thus correspond to a comparatively large number of genotypes, and that by this epistasy the species is able to react readily to different habitats (i.e. through selection) and yet remain fairly uniform in one habitat. The authors do not attach much importance to phenotypic variation. Rubtzov (1 935) confirmed this belief with field observations and laboratory experiments on related Gomphocerine species, including C albomarginatus; with the possible exception of social stimuli derived from crowding, environmental stimuli appeared unimportant. Similar views were expressed by Richards and Waloff (1 954) after a field study of C brunneus and other British Gomphocerines; variation appeared to reflect a stable genetic polymorphism, different in each population, and to be relatively unaffected by environmental fluctuations other than via long-term fecundity or differential survival of some forms. It is, however, likely that Chorthippus (Burton, 1960) does have a homochrome response to low albedo (see Section IIIA). B. GENETIC MODIFICATION OF PHENOTYPIC POLYMORPHISM

Of the various phenotypic responses to environmental stimuli, the green/brown polymorphism is known also to be affected in a number of species by the genotype of the individual. Thus in the

156

C. H. FRASER ROWELL

Oedipodinae green forms are always commoner among males than females in some genera, regardless of experimental conditions (e.g. Humb e tenuicornis, Walter, 1965 ; Gastrimargus africanus, Rowell, 1967b, 1970; Heteropternis couloniana, Rowell, unpublished), while in other genera the opposite is the case (Chortoicetes terminifera, Byrne, 1962; Ailopus thalassinus, Walter, 1965; Locustana pardalina, Nolte, 1963). The probable basis for such distributions is provided by the work of Stehr (1959) on the similar, but environmentally insensitive, polymorphism of haemolymph colour in the lepidopteran genus Qzoristoneura. Both protagonistic and suppressor loci for the green colour were found, and the latter was situated on the sex chromosome. It was noted above that in Chortoicetes the environmental response to high humidity is a green pigmentation which obscures the genetically determined patterning. However, there is also the opposite interaction. Not all the genetic variants are equally likely so to respond to high humidity, and Key (1954) considers that the majority of wild green forms represent the form nigrovirgata, which is the most susceptible t o the green morph. The example of Chortoicetes, in which a coloration element which is primarily controlled environmentally is also influenced genetically, is presumably typical. Other known cases are seen in the locusts. Locustana pardalina responds to humid environments when solitary with the green form of the green/brown polymorphism. The ability to respond to the environment in these ways is selectable, and strains of different capacity can be separated (Nel, 1968). Similarly, Hunter-Jones (1958), Nel (1967a, b) and Fuzeau-Braesch (1 96 1) have shown that the propensity for responding to high density with gregarious phase coloration can be readily selected for in Schistocerca gregaria, Locusta and Locustana, and in the Gryllid G. bimaculatus. Ultimately such genetic differences in the ability to respond differentially to environmental stimuli distinguish most grasshoppers from the locusts and other species which respond to the social environment by change, not only in colour, but often in behaviour and other aspects of their physiology as well. IV. ENVIRONMENTAL FACTORS

The variable coloration of the acridids that are phenotypically polymorphic consists of the interplay of a number of basically independent pigment systems. These are here defined phenomeno-

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

157

logically in terms of the colours they produce, and the stimuli they respond to, rather than as metabolic processes. A specific pigment metabolism may be influenced by more than one of these systems and a system may include more than one pigment; this is treated further in Section VI. The evidence for these separate pigment systems derives from the observation that they can vary independently in the population, from experimental work on the adequate environmental factors and the physiological mechanisms which they excite, and from pigment chemistry. The pigment systems here recognized are as follows: (a) The black pigment system. (b) The orange pigment system, which may itself be a heterogeneous category, and mediates colours of yellow to reddish orange. Together (a) and (b) mediate the homochrome responses to background colour. (c) The green/brown polymorphism and the pigments which mediate it. (d) Phase coloration, in those forms which have a coloration characteristic of gregarious phase. Though basically independent, these systems do interact, and these interactions will be returned t o below. A. THE HOMOCHROME RESPONSE TO BACKGROUND: THE BLACK AND ORANGE SYSTEMS

1. Extent and Occurrences A majority of geophilous and many gramnivorous grasshopper species resemble the general coloration of their background. Further, this resemblance holds when the species extends over a variety of differently coloured habitats, and also when the habitat is prone to seasonal or irregular colour change. The best known case of the latter is the observation that grasshopper populations of recently burnt vegetation are predominantly black. The explanation of these observations must lie among the following: (i) the population is variable in coloration. Matching is achieved by differential predation; (ii) the population is variable in coloration, and individuals select habitats matching their own colour;

158

C. H. FRASER ROWELL

(iii) individuals change colour to match the background: i.e. they make a homochrome response. The above hypotheses are not mutually exclusive. Indeed some overlap is logically necessary; in many environments, a homochromic response would have little or no selective value unless the individual also thereafter selected its habitat for background colour. The first proposition is immediately acceptable: differential predation will always tend t o eliminate the less well-matched individuals, regardless of whether the variation among individuals is produced genotypically or phenotypically. Confirmatory data on green and brown Acrida (Acr.) and grey and brown Oedipoda (Oed.) during predation by chameleons and storks has been obtained by Ergene (1 95 1, 1 9 5 3 ~ )Grasshopper . populations differ from those of many insects in that the second and third propositions can also be demonstrated in a variety of species. The evidence for this is presented below. While it is usual to suppose that the cryptic coloration which results from homochromic change is the major selective advantage achieved by the mechanism, it is clear that colour change will have other effects. To take an obvious example, dark coloured grasshoppers heat up more swiftly to a higher equilibrium temperature in radiant heat, especially when measurements are conducted on dead, anaesthetized or restrained insects; but freely moving animals, at least under conditions where air temperature as well as radiation is high, show very little difference in body temperature between opposite extremes of coloration (see Uvarov, 1966, pp. 207-224, for a review of the evidence). As it is clear that the blacker animals will always absorb more heat, the behavioural and physiological thermoregulation of individuals of the two extreme colorations must be quite different. Although a number of earlier observers assumed that the homochromy of grasshoppers on very recently burnt ground was the result of an active colour change, rather than of differential predation, migration or habitat selection (e.g. Poulton, 1926), the first successful experimental work corroborating this was performed by Faure (1 932) on Locusta and Locustana (Oedip.), establishing many of the basic facts, though in qualitative form. Solitary larvae of these species were found to be able to make homochromic responses to white, grey, black, yellow, orange and brown backgrounds, but not to red or pink or green. These homochromic responses were made at the moult following several days’ exposure to the

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

159

background and only occurred in individuals which were not stimulated by the appropriate factors to be green morphs or to enter gregarious phase (see B and C below). All or part of these results have since been experimentally confirmed in a considerable number of genera by several different authors: Pyrgomorphidae: Chrotogonus, Parasphena (larva only), (Rowell, unpublished). Eyprepocnemidae: Tylotropidius (Burtt, 196 1). Catantopinae: Melanoplus (Faure, 1933). Oedipodinae: Locusta Hertz and Imms, 1937; de Wilde and Staal, 1955; Albrecht, 1965; Fuzeau-Braesch, 1966; Nicolas and Fuzeau-Braesch, 1968), Oedipoda (Ergene, 1952c et seq. ; Levita, 1966a), Oedaleus (Ergene, 1955b), Gastrimargus (Rowell, 1970), Humbe, Heteropternis (Rowell, unpublished). Acridinae: Acrida (Ergene, 1950 et seq.), Cannula (Rowell, unpublished). Gomphocerinae: Phorenula (Burtt, 195 1 ), Mesopsis (Rowell, unpublished). Of the above, only black pigment homochromic change was found in the Gomphocerines, and in Cannula and Tylotropidius. The orange pigment change appears to be confined to the more geophilous forms, except in the case of Acrida. In the Cyrtacanthacridines Acanthacris and Ornithacris (Rowell, unpublished) no homochromic response has been found, though specifically looked for under experimental conditions. In a number of these genera, the change to black can take place in the adult, i.e. without a moult, though this requires longer exposure to appropriate conditions (Phorenula, Tylotropidius, Burtt, 195 1; Acrida, Oedipoda and Oedaleus, Ergene, 1953b, 1954a; Locusta, Albrecht, 1967; Nicolas and Fuzeau-Braesch, 1968). Once made, this change is not reversible, though larvae can assume a pale coloration again following a moult (Burtt, 195 1; Ergene, 1953b; Albrecht, 1964; Rowell, 1970). There has been no report of a similar adult change involving the orange pigment. Subsequent to Faure’s work, it has been found that homochromy to blue is impossible, at least among the Oedipodinae. Ergene (1952a, 1955b) has claimed homochromy to green and violet backgrounds in Oedaleus and Acrida, but this needs substantiation, especially as Acrida is sometimes found with purple coloration that appears to be identical with that which is genetically controlled in other species (see Section IIIA). True reds are not matched, but a

160

C. H. FRASER ROWELL

close approximation is achieved up to about 600 mp; in general the range from yellow through orange to red or purplish brown can be matched with great accuracy throughout. Homochrome adaptation is never total in any large population; even under the most favourable conditions virtually the full range of possible colorations can be found, though the homochrome form greatly predominates and some types, such as completely black animals, are almost never found except on the appropriate background colour (Faure, 1932; Hertz and Imms, 1937; Rowell, 1970). It may be noted that this is also true of those Gomphocerine genera where the polymorphism is probably entirely genetical (see Section 111) (Rubtzov, 1935; Peterson and Treherne, 1949; Richards and Waloff, 1954; N. Elsner, personal communication).

2. Light and Mediation All authors, with the partial exception of JovanCiC (1 963), agree that the most important environmental factor governing homochromic change is the light t o which the insect is subject. Hertz and Imms (1937), Grayson (1942), Levita (1966b), and Rowell (1 970) found that homochromy failed in three different Oedipodine genera and Melanoplus (Catant.) if kept in the dark. Ergene (1950, 1952b, c, 1953b, 1954a, 1955a, 1956) claimed that larvae and adults of Acrida, Oedipoda and Oedaleus would make homochromic change even though the eyes were covered with an opaque lacquer; further (1 954), that areas of the integument which were similarly lacquered did not make the change while the rest of the exposed insect did. The conclusion drawn was that the epidermal cells are directly responsive to light, and that the eyes do not play an essential role. Levita (1966b), in contrast, found that Oedipoda larvae with lacquered eyes made no homochromic response to background at all. Rowell (1970) repeated both of Ergene’s experiments with Gastrimargus. Lacquered eye animals kept on black backgrounds did show a slight darkening relative to controls on white backgrounds, but no more than a further control group with normal vision which was exposed t o black backgrounds for 12 h after each moult. It was concluded that the change seen in the blinded animals was caused by the visual experience during and subsequent to each ecdysis or during temporary damage to the opaque lacquer used, but that this change was slight compared to that seen in normal animals kept continuously on the same dark background. The experiments with

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

16 1

lacquered areas of the cuticle did not confirm Ergene’s findings. Acridu is a member of a different subfamily and has not been tested for a homochrome reaction by any other worker; but with this caveat, it will be assumed below that the visual response to backgrounds is mediated via the eyes. Faure and many subsequent workers have found that homochrome change proceeds better under strong illumination than weak. Some claims have been made that homochromy is impossible under weak light, but this is too extreme a view: a significant proportion of Gustrimurgus larvae made homochromic change to a variety of background colours including both black and orange under maximum illumination of only 23 pW/cm*, about one-thousandth of the maximum intensity recorded during control experiments carried out in sunlight. The change took much longer under the weaker light, none being seen before two larval instars had elapsed, but by the final larval instar there was relatively little difference between the two populations (Rowell, 1970). Hertz and Imms (1937) were the first to point out that the entire range of homochrome coloration in Locustu could be described in terms of variation of three components, black, orange, and yellow; this was achieved by comparison of the insect colours with the Ridgeway standards. Levita (1968) has confirmed these results using microspectrophotometric measurements of the reflected light from homochrome adapted Oedipodu. Her measurements show that excellent subjective matching can be achieved with backgrounds with dominant reflectivity in the range 574-600 mp, by altering saturation and lightness in a smaller range of pigments (576-588 mp). This fits well with the finding (Levita, 1966; Passama-Vuillaume and Levita, 1966) that the epidermal cells of these insects contain orange and yellow granules of tetrapyrrolic character and brown-black granules of ommochrome (see also Section VI). The orange and yellow mixtures were found by Hertz and Imms to be evoked only on backgrounds which were predominantly of these colours, whereas black pigment was evoked in varying proportion by all backgrounds, being least on white or on yellow-green, and greater on black and far red. Blue background colour elicited grey. From these observations, the authors concluded that the two systems were separately controlled. (In Section VI it is shown that they also use chemically different pigments.) The black pigment was thought to be evoked by the intensity difference between incident and reflected light, and the responses to blue and red backgrounds were taken to

162

C. H. FRASER ROWELL

mean that these wavelengths appeared respectively light and dark to the insect eye. Orange pigment was elicited only on backgrounds of that wavelength, and the authors explained its absence in insects raised on white backgrounds in white light by postulating that the short-wave components of less than 550 m p not only did not elicit but actually inhibited the production of this pigment. The findings of Hertz and Imms on the control of black pigmentation have been confirmed and extended with Gastrimargus (Rowell, 1970), a genus which is very close to Locusta but shows little or no phase behaviour. Populations of insects raised in white light on coloured backgrounds were scored for black pigmentation, and the resulting rank order shown t o correspond with one exception to the apparent reflectance of the backgrounds, the latter being calculated from the spectral response curves obtained by Bennett et ul. from Locustu retinula cells. The exception was the population reared on orange. The retinula cells are relatively insensitive to light of longer wavelength than 550 mp, and orange backgrounds would be expected to give marked black pigmentation, as do red. In practice, little black pigment is elicited, and the animals turn orange instead. The most probable explanation for this was thought t o be an inhibitory relationship between the two pigment systems, the black and the orange, both of which are evoked under these circumstances. The independence of the black homochrome response with respect to wavelength of light was confirmed by raising groups on white or black backgrounds under a variety of approximately monochromatic lights. In all these experiments, high reflectance backgrounds inhibited the production of black pigment, and absorbant ones facilitated it. The term albedo response was suggested for the control of black pigmentation by these factors. It is not known whether the differential illumination of the retina, which these findings indicate as the trigger stimulus of the albedo response, is topographically fixed, as has been claimed in the isopod Ligiu. If so, then species with characteristically different positions relative to the horizontal (e.g., grass-living Acridines) would be expected to differ in this topography, or alternatively space-constant visual interneurones (Wiersma and Yamaguchi, 1967) could mediate the response. Experiments on Gustrimurgus led t o a new hypothesis of the factors influencing the orange pigment system. While orange, brown or yellow backgrounds caused the appropriate homochrome change when illuminated with white light, animals illuminated with

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

163

monochrome yellow, red or orange light on white backgrounds made no such change, confirming an earlier result of Grayson (1 942) with Melanoplus. Backgrounds of other colours under this illumination induced only black pigment response, in proportion to their albedo in the incident light used. The factor present in the first situation which is missing from the second appears to be light of short wavelengths. This is confirmed by the finding that animals simultaneously illuminated with both orange and blue light from discrete sources do make a homochromic colour change to orange, though not to either light singly. The response is not affected by reversing the lights, so that the blue source is ventral and the orange dorsal; the response therefore does not depend upon specific areas of the retina, but only demands that some areas receive predominantly short-wave illumination and others, simultaneously, receive predominantly long. This hypothesis is compatible with the results of Hertz and Imms (1 937). The electrophysiology of the acridid eye is not so far advanced as to give direct evidence for or against the hypothesis, but the available information is compatible. The retinula cells respond over the range of about 300-600 mp, and their peak sensitivities form a spectrum so that some are very much more sensitive in the long wavelengths than others (Bennett et al., 1967). Interneurones, which were fed with opposite sign from two such cells of opposite extremes, would effectively function as either long- or short-wave detectors, with little or no response in the green; such units are not yet known from acridid nervous systems but have been described (Swihart, 1969) from lepidopterans. Using elements with these characteristics, a neuronal model can be constructed which will represent the environmental stimuli and the command output to the orange pigment system (Rowell, 1970). The evidence presented in Section VI suggests that this command would ultimately affect the oxidative state of epidermal biliverdin. It should be stressed that the “orange” pigment system here referred to may not be unitary. But yellow and orange components are present (see discussion above), and though Levita (1 968) stresses that these cover only a small range of wavelengths and implies that the distinction may not be real, her published datum points from red-brown insects form a population which differs significantly from those obtained from insects adapted t o yellow backgrounds. Subjectively, homochrome insects from these backgrounds appear very different. Retinula cells do not fall into two distinct categories of frequency sensitivities, as would be the minimal requirement to

164

C. H. FRASER ROWELL

drive one pigment system in the manner proposed; they form instead a graded population between two extremes, and appropriate combinations of these could give detector systems for several closely spaced wavelengths.

3. Other Environmental Factors In comparison with other elements of variable coloration, the homochrome responses are relatively unaffected by environmental stimuli other than light. Temperature is known t o influence the development of black pigment in a variety of grasshoppers and locusts, but the relation of this to the albedo response is almost unknown. A number of grasshopper species are generally darker in the colder parts of their range than in the warmer, and in Chortioicetes (Oed.) (Key, 1954) this relationship has been experimentally shown to be causal. Grayson (1 942) found that low temperatures increased both areas of black pattern and the general level of black pigmentation in Melonoplus bivattatus (Catant.) and Okay (1956) found that Acridu (Acr.) reared at sublethal cold temperatures (c. 16°C) were darker than normal (32°C) controls, while those reared at high temperatures (40"C+) were much lighter than normal. Duck (1944) had obtained similar variation in both green and brown forms of Schistocerca obscura (Cyrt.). rearing them at 21", 30" and 32°C. Similar effects are known with gregarious phase locusts (see Section IV C.). Nicolas and Fuzeau-Braesch (1 968) showed that if gregarious Locusta (Oed.) are prevented from developing dark pigmentation by cold treatment, they are thereafter, at normal temperature, rendered able to make a homo chrome response to black background, which is normally impossible with individuals of this phase. In Poekilocerus hieroglyphicus (Pyrg.) cold temperatures (1 7" C) caused a retention of melanic pigment in total darkness, which at normal temperatures would have caused it t o disappear (Abushama, 1969). This homogeneous body of results apparently shows that the Acrididae differ markedly from the Phasmid Curansius morosus, in which melanin and ommochrome formation is inhibited by low temperature (Dustmann, 1964). Crowding, other than in those species in which it causes phase change or influences the green/brown polymorphism (Sections B and C below) has no apparent effect on homochromic change; humidity, too, affects it only in so far as it biases the green/brown polymorphism. A variety of agents other than temperature are

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

165

known to influence black pigmentation, but their effect is probably a general one acting through the responsible endocrine system (see Section V B) and has no special relevance to homochromic change.

4. Interaction of the Homochrome Responses with Other Pigment Systems The apparent inhibitory relations between the orange and the black pigment systems has been noted above. Both systems are also inhibited to a greater or lesser degree when either green morphs or (in locust species) gregarious forms are produced. It is usually stated that homochrome responses are absent from green morph grasshoppers. This is not strictly true. In some species the entire body surface is subject to green/brown polymorphism, but in many, especially among the Oedipodinae, there remain some areas of epidermis which retain in a characteristic pattern the brown ground colour, even though the rest of the animal is green. These areas remain capable of giving a homochrome response with either the orange or the black pigment system. However, only very rarely does one see an insect in which it appears that both green and brown pigments are present either in the same or in closely adjacent cells. There is undoubtedly an inhibitory relation between green and black pigmentation. In experiments with Gastrimargus, all circumstances which altered the probability of black pigmentation also altered inversely the probability of green; this held to include the case of animals raised on orange backgrounds. As explained above, the orange pigment response which is thus elicited appears to inhibit the black pigment which is expected on the basis of the apparent albedo of orange backgrounds; and these animals were not only less black than expected, but had an unusually high proportion of green morphs (Rowell, 1970). A similar relation between green and black pigmentation was found in Locusta by Nicolas (1 969), and Nicolas, Cassier, and Fain-Maurel (1969), when the activity of the black system was manipulated with C 0 2 concentration. However, the green and black pigment systems are not completely exclusive. Both in the wild and in experimental situations one finds a small proportion of green morph animals with heavy black pigmentation superimposed on the green areas; this is especially frequent in Oedipodine grasshoppers raised on black backgrounds in high humidity (Albrecht, 1967; Rowell, 1970), or captured in corresponding habitats, and is also seen in other groups, e.g., the elongate gramnivorous genus Mesopsis (Gomph.) and in

166

C. H. FRASER ROWELL

Amphicremna (Acrid.). Nolte (1963) noted that black and green pigments could vary independently in his experiments with Schistocerca gregaria, Locusta and Locustana. Homochromic responses were found by Faure (1 932) to be absent in Locusta and Locustana hoppers which were kept crowded and which tended to enter the gregarious phase. This has been confirmed by all subsequent investigation and is now accepted for the locust species (Uvarov, 1966), though crowding does not affect the homochrome responses of related species, such as Gastrimargus, which show no locust traits (Rowell, 1970). This is not to deny that the gregarious phase coloration which is found in such locust populations is itself in part influenced by .conditions of the physical environment (see Section IV C), but these influences do not result in homochromy. It has recently been shown (Fuzeau-Braesch, 1968; Nicolas, 1969; Nicolas and Fuzeau-Braesch, 1968; Nicolas et al., 1969) that this and several other aspects of the transformation to gregarious phase can be inhibited or reversed by short daily exposures to high concentrations of C 0 2 . This has itself a small positive effect on black pigmentation (see above), but a much larger one in reducing the number of gregarious phase animals and thus increasing the number of solitarious forms capable of making homochromic responses to black. Only the responses to black backgrounds has been tested in these experiments; whether the orange pigment system is affected in any way is not yet known. 5. Active Selection of Background In view of the importance of this behaviour for a full explanation of homochromy , remarkably little experimental evidence is available. Ergene (1951a, 1952d, 1953a, 1957) performed a series of experiments using wild-caught Oedipoda and Acrotylus (Oed.) from a variety of habitats, in which the animals were differently coloured; black grasshoppers from burnt grassland, yellow from clay, grey from chalk, reddish from a terra rosa soil, and speckled yellow from sand. When liberated into a cage containing the different substrates, some 80% of the animals which settled on one of these substrates chose the correct one; similar discrimination was shown by larvae, but it was abolished in all animals when the eyes were lacquered. Experiments performed on Gustrimargus adults which had been reared since hatching on white or black backgrounds confirm these experiments, and similar observations have been made in the wild on the same species (Rowell, unpublished observation). In principle, a preference for a particular background colour could

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

167

be built into the animal’s genotype. This appears to be the case for other aspects of substrate selection in a wide variety of marine animals, and Meadows (1967) has shown that in the amphipod Corophium previous experience of atypical substrates does not alter the preference for typical ones. Such a fixed preference may exist in some acridids, presumably ones with invariant coloration, though none have been specifically identified. It would not, however, suffice for the present examples; all the genera cited above are known to perform homochromic colour change, and the preference for a background colour must therefore be acquired simultaneously with the homochromic coloration. There seem to be three possible ways in which this could be done. In many species, the homochromic coloration extends to the eye, and might be thought to act as a filter over the retinal apparatus. Homochrome backgrounds would therefore tend to be relatively brighter than non-homochrome ones. This seems an improbable explanation, if only because the animal could not discriminate between homochrome and white backgrounds, and the system would clearly not work for black backgrounds. Secondly, the animal can see various parts of its anatomy, and might conceivably match its background with itself by visual comparison. This again seems unlikely; the most affected areas, the dorsal surfaces, are not visible to the animal. The third possibility derives from the fact that appreciable homochromic change requires five or more days’ exposure to the background colour. It seems likely that the animals could simultaneously learn this colour, and select it preferentially thereafter, resulting in a sort of visual homeostasis. This hypothesis is supported by an important experiment (Ergene, 1957) in which it was shown that the normal preference of wildcaught grey Oedipodu from a grey habitat could be significantly diminished if they were kept for 5- 10 days on a black background. A surprising additional result was that the change to the new preference for black was less if the animals were kept in dimmer light inside than if kept in daylight outside; this would have adaptive significance, for the former conditions would produce less pigmentation in the population, but the mechanism by which the effect is achieved is hard to imagine. B. THE GREENIBROWN POLYMORPHISM

1. Distribution and Occurrence A large number of acridid species (and also many other insect groups, including the tettigonioids, mantoids, phasmids, cicadids, and

168

C. H. FRASER ROWELL

lepidopterans, t o mention some of the best known) have the capacity for a green/brown polymorphism of the pigmentation of the epidermal cells over much or part of the body surface. Closely related to this polymorphism is an associated polymorphism of haemolymph colour; the two share identical or very closely related pigments, and probably functionally linked (see Section VI). Of a sample of East African acridoids, comprising 180 species and 107 genera of the Acridoidea, a green/brown polymorphism was known to occur in 85 and 43 of these respectively (Rowell, unpublished). The polymorphism may be even more generally distributed than such figures as this suggest. In some species one or the other form is excessively rare; Richards and Waloff (1 954) obtained two green forms of Chorthippus brunneus (Gomph.), a sample of 2300, and Wise (1966) recorded the fifth known green form of this species from Britain; only two brown forms of the larva of Acanthacris rujicornis (Cyrtacanth.) were found in a sample of several thousand wild specimens in Uganda, though the brown form is common under culture conditions (Rowell, 1967); only one green adult of Catantops kissenjianus (Catant.), an abundant species, was seen in six years of collecting in Uganda. Mesopsis laticornis (Gomph.) was uniformly brown in the Rukwa Valley, Tanzania, in 1956 (Chapman, 1962), but uniformly green in the Murchison Falls National Park, Uganda, in 1967. One, but only one, green adult form has been seen in cultures of Schistocerca vaga (Cyrt.) (G. B. Staal, personal communication). Thus in many species one of the colour forms may have been overlooked. This is supported by hormone experiments (see Section V) in which it is possible t o obtain morphs unknown in the wild, such as green adult Acanthacris ruficornis (Rowell, 1967b), demonstrating that the epidermal cells may retain the capacity for the polymorphism even though the appropriate hormonal climate rarely or never occurs. It has been frequently reported in those species where both forms are reasonably common that the green form is more abundant in wetter parts of the habitat, or during and after a rainy season (see, e.g., Golding, 1934). As green vegetation is also more probable in the same circumstances, the resultant cryptic coloration is presumed and has occasionally been shown (Ergene, see p. 158) to confer a selective advantage. The correlation of green colour in background and insect does not, however, imply that cryptic coloration is the only aspect of this polymorphism on which selection could act. Albrecht (1964, 1965) has for example shown that the different

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

169

morphs in Locusta tolerate starvation differently; green forms are more stress-resistant in damp atmospheres, and brown in dry atmospheres (though not so markedly as the gregaria form). Similarly, Richards and Waloff (1 952) found that population figures for some British Gomphocerines (Omocestus viridulus, Chorthippus parallelus and Stenobothrus lineatus) suggested a differential longevity correlating with different colour forms, including the green; though it should be borne in mind that the polymorphism of these species is probably largely genetic, and such differences are perhaps more expected in a polygenic situation. Logically, the same interacting causal hypotheses as shown (p. 157) valid for the homochrome response apply to the green/brown polymorphism. Differential predation has already been acknowledged. There are no data which show that grasshoppers select green or brown environments according to their own morph, but the experiments of Albrecht (above) and others make it probable that they would so select an appropriate humidity; as this is apparently the dominant environmental factor regulating the polymorphism, as shown below, it would probably have the same effect. The issue appears to be less acute than in the homochrome situation, however, as the environment is likely to change more gradually in its overall vegetational state than in the colour of its earths and rocks. Genetic factors undoubtedly play a part in determining the green/brown polymorphism in many species (see Section III), but most of the experimental evidence concerns the role of environmental factors. As in other aspects of grasshopper coloration, relatively few subfamilies have been investigated, and there seem to be three main groups of experimental animals giving rise to conflicting beliefs, and which may possibly show real differences, at least with regard as to which of the various factors is the most important in determining the polymorphism of the population. These three are the Oedipodinae, the Gomphocerinae, and Acrida. The evidence relating to the first two groups is internally consistent, and seems to indicate a predominantly environmentally determined polymorphism in the former (though with genetic modifiers), and a predominantly genetic polymorphism in the latter. Enough data have been obtained for the Oedipodine species to make it sure that genetic variation is of secondary importance. It should, however, be noted that there is a virtual absence of data on the effects of different environments o n experimentally maintained populations of Gomphocerines, and the evidence for a predominantly genetic

170

C. H. FRASER ROWELL

polymorphism comes either from direct genetic experimentation or by deduction from measurements of wild populations and observation of their stability under varying seasonal conditions. This evidence, while certainly valid, does not exclude environmentally determined phenotypic variation as a further possibility. The only experimental evidence against this possibility is that of Rubtzov (1935), who found that in Chorthippus and a number of other Gomphocerine genera no experimental manipulation of background, density or humidity ever produced a transition from green to brown or vice versa. Finally, in the case of Acrida, there is considerable disagreement between the various workers who have used it. 2. Light and Radiation It was noted in the previous section that the probability of the green morph is reduced if lighting conditions are such as to favour black pigmentation. Except for this indirect effect, there is no unequivocal link between reflected light and green/brown polymorphism. The green/brown polymorphism of the cyrtacanthacridine genera is not influenced by background colour, and the oedipodine genera are unable to make a homochrome response to green backgrounds, in spite of their ability so to react to albedo and to backgrounds reflecting longer wavelengths. This homochrome response to green has, however, been vigorously claimed for Acrida bicolor (Ergene, 1950, 1952a, 1954b, 1955a) and as energetically denied (JovanCiC, 1953, 1960, 1963; Okay, 1954, 1956). Okay’s experiments were performed in total darkness; many grasshopper species do not eat under these conditions and soon die, but Acrida and a few Locusta lived and changed from green to brown or vice versa with other experimental factors. A . turrita is green/brown polymorphic in E. Africa; a laboratory culture derived from the wild stock remained uniformly green for some generations under conditions of background colour, lighting, humidity and density which would have induced a large proportion of brown forms in cyrtacanthacridine or oedipodine genera (Rowell, unpublished). It therefore seems likely that if these environmental factors are important to Acrida, they have thresholds differing from those of the other subfamilies. It can be speculated why a visual response to environmental colour is apparently suitable to control black and orange pigmentation but not green. One possible reason derives from the mechanism proposed (pp. 162-3 above and more fully in Rowell, 1970) for the colour

THE VARIABLE COLORATION OF THE A C R l W l D GRASSHOPPERS

17 1

discrimination leading to the homochrome responses. The spectral response curves of the different receptors are most different in the long and short ranges, and the comparator cells which are postulated would have difficulty in distinguishing green from yellow. Such a confusion would lead to undesirable interaction of green and orange pigmentation: and in Gastrinaargus more orange forms than expected are indeed produced on green backgrounds, though these do not affect the proportion of green forms. The quantity and quality of incident radiation, however, may well play a part in determining the green/brown polymorphism. JovanCiC (1963) has suggested that strong incident light will tend to produce brown, rather than green forms, with the implication that the latter provides a better shield against excessive radiation. There is no direct evidence that this is the function of the green/brown polymorphism, but a correlation between low light intensities and the green morph has been found in other orthopteroid insects. Willig (1 969) found that Gzrausius tended to be green if reared .in total darkness, but brown in normal lighting: this was associated with a change in the quantity of biliverdin in the epidermal cells (see Section VI) and was not due merely to masking by other pigments. Passama-Vuillaume (1964, 1965a) has adduced evidence that the labile bile-pigment chromoprotein, which is either solely or partly responsible for the green coloration (see Section VI), is directly affected by incident radiation. In Mantis religiosa and Sphodromantis it was found that the extracted water-soluble pigment tended to oxidize to brown in far red and infra-red light, and t o colourless in blue or ultraviolet light. Normal white light of low intensity left the green pigment unchanged. The same changes were produced in living animals under similar conditions of illumination, and it was suggested that the direct response of the pigment to radiation explained these changes. A similar pigment was found in Locusta (Passama-Vuillaume, 1965b), but was much more resistant t o colour change under similar conditions of radiation, which was attributed t o differences in the bound protein fraction of the pigment. However, the possibility clearly exists that the green pigment of acridids is directly responsive to illumination under some circumstances, and the expected direction of such a response would be that high intensities of white light or of infra-red radiation would favour the brown form. In view of the conflicting reports on Acrida, this would be a likely place to look for such an effect. A correlation of this sort was found by Rowel1 (1970) in the

172

C. H. FRASER ROWELL

oedipodine Gastrimargus ; otherwise identical cultures illuminated by tungsten light sources produced very significantly fewer green forms when bulbs of higher wattage were used. Air temperature was not appreciably altered by the larger source, and though it is possible that the greater intensity of radiation could produce a higher internal temperature in the insects, this is unlikely in view of the extent to which grasshoppers actively regulate their temperature. A similar effect is suspected in Cyrtacanthacrid larvae. Under culture conditions with tungsten illumination larvae of Cyrtacanthacris, Acanthacris, and Schistocerca vaga show a pronounced greenlbrown polymorphism; the population is all green in the first instar, but shows a progressive change to a majority of brown forms in the last larval instar. This effect is accentuated by stronger wattage tungsten bulbs, and by longer illumination cycles, and is inhibited if fluorescent light sources, which are notably deficient in longer wavelengths, are substituted (Rowell, unpublished; W. Loher, personal communication).

3. Humidity There is general agreement that humidity is the most important single factor in predisposing experimental populations in' favour of the green morph (Locusta and Locustana, Faure, 1932; Pyrgomorpha cognata (Pyrg.), Golding, 1934; Melanoplus sanguinipes (Catant.), Faure, 1933; Locusta, Hertz and Imms, 1937; Albrecht, 1967; Chortoicetes (Oed.), Byrne, 1967a, b; Gastrimargus, Rowell, 1970; Acrida, Okay, 1956; JovanCiC, 1953, 1960; Schistocerca vaga (Cyrt.), Rowell and Cannis, unpublished; for a dissenting view, see Ergene, loc. cit supra). A number of reports linking green forms with fresh as opposed t o wilted foodplants (Schistocerca paranensis, S. gregaria, Gastrimargus africanus, Hunter-Jones, personal communication and 1962; Hunter-Jones and Ward, 1960; Locusta and Acrida, Okay, 1953) are probably picking up the same effect. The ecological and physiological implications of this sensitivity to humidity are of interest. Humid atmospheres will be associated with a larger probability of green backgrounds than will dry ones, but in order that full use may be made of this correlation it would seem necessary to distinguish between (a) the humidity of the microclimate surrounding the insect, which because of its feeding habits and other behaviour will tend t o the moist parts of the environment, and (b) the humidity of the atmosphere as a whole, which is more likely t o be of significance to the vegetation.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

173

Cutaneous hygroreceptors of grasshoppers (see, e.g., Slifer, 1955; Aziz, 1958; Riegert, 1960) will sample only the former. It has been suggested (Rowell, 1970) that the apparent sensitivity of the green pigment to infra-red radiation may contribute to an assessment of atmospheric humidity, as radiation a t the earth’s surface varies with this parameter by up to 60% in certain wavelength bands in the far red or near IR (Gates, 1966), changes much greater than those found to influence the green/brown polymorphism under culture conditions. But whatever the possible role of hygrometry by IR detection in the wild, the experimental data indicate pronounced effects of humidity on green/brown polymorphism under conditions where it could play no part, so that some other detector system, presumably neural, must play a major role.

4. Crowding In locust species, crowded rearing conditions turn experimental populations of larvae into the gregarious phase, which has its own characteristic coloration which excludes both the green/brown polymorphism and the homochrome responses to background colour and albedo. Relatively little work has been carried out on the effects of crowding on the coloration of non-locust species. It is clear that crowding does not invariably influence coloration in grasshopper species, for the normally social species of Phyteumas and Phymateus (Pyrg.) retain all their normal complex series of coloration changes if isolated experimentally in the first instar (Rowell, 1967a). In Gastrimargus (Oed.) it has been demonstrated that crowding is significant in influencing the green/brown polymorphism in most sex and age categories (Rowell, 1970). The exceptions are that under the experimental conditions used the density effect was overruled by humidity effects in two populations; adult males in high humidity gave only green forms (males have a genetic predisposition to the green form), and female larvae in low humidity produced a uniformly brown population. In the remaining six experimental categories the crowded population had fewer green forms. The nature of the effective stimulus which is derived from other individuals, visual, tactile, or chemical, is not known. It is unlikely that it reflected any profound change in metabolism for there was no difference in the length of the larval development between the crowded and solitary populations, unlike locusts or social pyrgomorphs. The selective advantage of the sensitivity t o this factor is also obscure.

174

C. H. FRASER ROWELL

In Cyrtacanthacridines the situation is confused. Johnson (1 929, 1932) claimed that crowding changed nymphs of Cyrtacanthacris tartarica and Acanthacris ruficornis from green to brown. This experiment is not fully reported, but it involved the experimental crowding of wild-caught green larvae. Duck (1944) found that solitary larvae of Schistocerca obscura (Cyrt.) were green, but when reared two or more to a cage were brown with more or less dense black markings, depending on temperature. The author attributes this to a phase polymorphism, but it is not clear from his description whether the crowded form is analogous to the gregarious phase of S. gregaria or the brown forms of S. vaga, Acanthacris ruficornis or Cyrtacanthacris tartarica; the latter alternative appears more probable. Cultures of these three species develop increasing numbers of brown forms with age, resulting in a majority by the fourth and fifth instar, although in the wild the brown form is very rare (Rowell, unpublished and 1967b). In S. vagu crowding increases the probability of brown larval forms in either wet or dry environments by 10% relative to isolated controls, but the effect is less that that due to humidity (Rowell and Cannis, unpublished). Experimental isolation of brown Acanthacris did not produce a reversion to green in succeeding instars, but the isolation was not enough to prevent, e.g., pheromonal communication. Some Cyrtacanthacridinae do not respond to culture with a brown form (e.g. Ornithacris turbida, Rowell, unpublished) but in many of this subfamily it seems that crowding increases the probability of the brown form. 5. Temperature Okay (1956), keeping brown Acrida (Acr.) larvae in saturated air and total darkness (both being conditions conducive to the green morph), found that at 33°C all larvae eventually became green, whereas as the temperature approached the upper and lower lethal limits at 16" and 46°C the proportion dropped to 50% or less. This result may be interpreted as reflecting more a depression of vitality than a specific effect on the polymorphism. JovanEiC (1963 and preceding works) has also consistently stressed the role of temperature in determining the green/brown polymorphism of Acrida and of Mantis. The evidence is complicated by the apparent need to distinguish between temperature and long-wave radiation (see Section 1 above) and by and large this has not been done. Passama-Vuillaume (1 965a) found that the responses to radiation of Mantis pigment in vitro and in the living animal were constant over

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

175

the range 29-30°C: but according t o JovanCit (1 960) it is only over 3OoC that mantises turn brown in response to high temperature. It certainly seems possible that an oxidative change such as Passama-Vuillaume describes could be facilitated by higher temperature. On the other hand, the green forms of Schistocerca obscura (Cyt.) remain green with temperatures up to 34" C, but show a progressive lightening of colour from dark green to a pale greenish white, an effect presumably due to inhibition of melanin and black ommochrome (see p. 164); there was no indication of a change to brown (Duck, 1964). In Chrausius (Phasmidae) the amount of biliverdin in the epidermis is reduced at 28" C, relative to 18"C, and this increases the probability of brown morphs-(Willig, 1969). High temperatures inhibit black pigmentation (both melanin and ommochrome fractions) in both solitary grasshoppers and gregarious phase locusts (pp. 164, 177). In view of the reciprocal relation between black pigmentation and green morph found in the homochrome and other responses of Oedipodines (pp. 165-166) one would expect that under some conditions green pigmentation would be facilitated by high temperatures, not inhibited, but no report of this is available. C. PHASE COLORATION

The vast literature on phase polymorphism of locusts has been frequently reviewed, two of the most recent and extensive treatments being those of Uvarov (1 966) and Albrecht (1 967), where further detail and bibliography can be obtained. Here only brief attention will be given to the coloration of the gregarious phase in locust species, and to the interrelation of this and other components of variable coloration. The characteristic coloration of gregarious locust hoppers is basically a bold alternation of areas of black and yellow or orange; gregarious phase hatchlings are usually black or at least very dark. Such a coloration is well known in, e.g., Locusta and Locustana (Oed.), and in Schistocerca gregaria, Nomadacris septemfasciatum, and Anacridium aegyptium (Cyrtacanth.). A similar patterning has also been found in the wild in exceptionally dense populations of grasshoppers which are not normally considered liable to phase change, including Pyrgodera armuta (Oed. ; Popov, 1952), Faureia milanjica (Gomph. ; Sjostedt, 1929), and Chorthippus albomarginatus (Gomph.; Rubtzov, 1 9 3 9 , though only in C. albomarginatus was the

176

C. H. FRASER ROWELL

effect reproduced experimentally. In other species of economically important grasshoppers, the high density populations may have colour patterns which differ at least statistically from the normal, and may approach the classic orange and black type, as in, e.g., Melanoplus sanguinipes (Catant.) (Faure, 1933; Grayson, 1942) and Austriocetes males (Oed.) (Key, 1954). There are also species in which the only correlate of density is a shift in the proportions of the colour variants which occur under all environmental conditions (e.g., Dociostaums maroccanus adults (Gomph.), Uvarov et al., 195 1;Gastrimargus musicus (Oed.), Common, 1948). At least three different trends can be discerned in the behaviour of the pigment systems during the assumption of gregarious phase coloration. These are: (i) increased black pigmentation; (ii) inhibition of the pigments responsible for both the green and brown ground colours, and thus also of the green/brown polymorphism; (iii) inhibition of the visually mediated homochrome responses to albedo or to orange backgrounds. On the evidence of pigment chemistry, it is possible (Section VI) that these trends reflect only two major changes in pigment metabolism. The loss of the green/brown polymorphism and the loss of the orange homochrome response may be two aspects of a single fundamental change in the metabolism of bile pigments; the increase in black pigmentation involves the same metabolic processes as the homochrome response to albedo, and presumably this system is uncoupled from the normal triggering mechanism during gregarious phase. Most of the literature on the complex of environmental factors that promote gregarious phase deals with phase criteria other than colour (e.g. morphometrics or behaviour). There are some data on the environmental conditions which influence at least the black pigment component. Stimuli derived from crowding appear to facilitate black patterning (as opposed to general melanization) in a variety of grasshoppers as well as the locusts. Instances are recorded for Anacridium moestum and Cyrtacanthacris tartarica (Cyrt.; Johnson, 1932); Aiolopus tergestinus (Oed.; Plotnikov, 1926); Chorthippus spp. and other Gomphocerines (Rubtzov, 1935); Schistocerca obscura (Cyrt.; Duck, 1944); Gastrimargus africanus (Oed.; Rowell, 1970). This response is, however, not entirely general,

THE VARIABLE COLORATION OF THE ACRlDOID GRASSHOPPERS

177

for it is lacking in many species closely related to the above (Uvarov, 1966; personal observation). Nolte ( 1963) claimed a pheromonal effect in Schistocerca gregaria, Locusta and Locustana facilitating melanization; recent reports of agents with at least some similar properties in Schistocerca gregaria and Chortoicetes terminifera (Oed.; Anon., 1969) tend to support this suggestion. The nonspecific stress effects of crowding, due to constant disturbance, irritation, tactile stimulation, optomotor input, and so on, must not be overlooked, in view of the way in which such input is likely to influence release of corpus cardiacum hormone (see Section V). Rubtzov published as early as 1935 the observation that grasshoppers infested with mites were invariably darker, and assumed that the increased coloration was a result of irritation. Enforced activity and increased COz concentrations both promote black pigmentation of S. gregaria hoppers (Husain and Mathur, 1936a, b). Temperature is known to influence the development of black pigment in gregarious Locusta and Schistocerca gregaria hoppers. High temperatures (c. 40°C) inhibit the pigment and low temperatures (c. 2OoC) promote it (Husain and Ahmad, 1936; Stower, 1959; Dudley, 1964; Uvarov, 1966, review; Nicolas and Fuzeau-Braesch, 1968). This seems to be an effect identical with that seen in solitarious locusts and grasshoppers (p. 164), so that is probably not strictly relevant to phase coloration specifically. The effect of the phase transformation upon the systems responsible for the green/brown polymorphism and for the orange homochrome response is entirely obscure. If all these colours are in fact mainly due to variation of the bile pigment fraction, then it is possible that the various colour states, green, brown, orange and colourless (which last would represent the gregaria state), represent different oxidation levels (see Section VI). There are suggestions that some Cyrtacanthacridine larvae may respond to crowding by a switch from the green to the brown morph (see Section IV B), and if this were confirmed, would tend to support the idea that crowding causes a variety of changes in bile-pigment metabolism, and not merely its total absence in the gregarious phase. V. PHYSIOLOGICAL MECHANISMS

The environmental stimuli reviewed in the preceding section, with the exception of temperature and the possible direct effect of IR radiation on the green pigment, act only on the sensory receptors of AIP-9

178

C. H. FRASER ROWELL

the animal, and are integrated by the central nervous system. (Indeed, control of variable coloration and associated choice of homochrome backgrounds are virtually the only functions which can so far be ascribed to the colour vision of acridoids, which is otherwise certain only from electrophysiology.) All the colour changes are mediated by the epidermal cells. The mechanisms whereby the CNS communicates with the epidermis is the subject of this section. Of the classical mechanisms whereby such communication is achieved, there is no evidence to suggest a direct nerve supply to the epidermis which could be used for this function. Such innervation has been described from the Hemiptera (Maddrell, 1966); there is no comparable morphological account from the Orthoptera. Neurosecretion, perhaps additionally supplemented by non-neural endocrine organs, appears to be the only alternative, and t o date only two of the various pigment changes seem to have any endocrine correlate; more are presumably to be expected.

A. THE GREEN/BROWN POLYMORPHISM AND THE CORPUS ALLATUM

Pfeiffer (1945) showed a correlation between naturally or artificially high titres of corpus allatum secretion and green coloration of the haemolymph in Melanoplus sanguinipes (Catant .). This pigment has many resemblances to that found in the epidermis of green morphs, and may be identical (Section VI); it also appears in the haemolymph prior to an epidermal change to green (Ergene, 1954c, and subsequent authors). Joly (1951, 1952) showed that implantation of additional corpora allata into brown individuals of Locusta and Acridu bicolor increased the number of green morphs in the next instar. This effect has been exhaustively confirmed in Locusta (Joly e t al., 1956; Joly, 1960; Staal, 1961 ; Joly, 1962). For effective change to green, the implantation must be made late in the preceding instar, for while earlier implantation produces a variety of effects, such as metathetaley, on the resultant animal in the following instar, it has no effect on coloration in a majority of cases. It was therefore deduced that the pigment system is sensitive to hormone titre only at or immediately around the moult, and that the hormone increase given by the implant is transitory and decreases soon after the implantation. However, Novak and Ellis (1 967) found that the sensitive period in gregarious larvae of S. greguria was during

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

179

the first third of the instar; whether this reflects differences in animal, phase status or technique is not clear. Implantation of even a single additional corpus allatum into larvae of Humbe tenuicornis, Gastrimargus africanus (Oed.) and Acanthacris ruficornis (Cyrt.), species which show the greenlbrown polymorphism but none of the phase characteristics of the locust species to which they are closely related, induces green forms in subsequent instars. Indeed, several green adult Acanthacris developed from implanted larvae, although such a coloration has never been reported in the wild, which demonstrates the continuing competence of the epidermis to respond t o the hormonal climate (Rowell, 1967b). The green colour frequently lasted for several stadia, which argues against a merely transitory survival of the implanted organ. It may be that in previous work the large number of implanted glands induced a retrogressive change in those of the host, and thus a long-term reduction in hormone titre. Injection of synthetic juvenile hormone into larvae of Schistocerca vaga (Cyrt.) increased the probability of the green morph in subsequent larval instars, while allatectomy resulted in a brown coloration in the next instar, and adultoid coloration and morphogy in the subsequent one (G. B. Staal, personal communication). Allatectomy of green Syrbilla fuscovittata (Oed.) induces brown coloration, but neither allatectomy nor implantation affects the green/brown polymorphism of Gomphocerus mfis (Gomph.) (W. Loher, personal communication). This not only supports the view that the polymorphism of the Gomphocerinae is genetically rather than environmentally controlled, but also suggests that there may be real differences between the control of the pigment systems of the epidermal cells in this group as compared with Acridines, Oedipodines and Cyrtacanthacridines. There is no information available on the relation between juvenile hormone and the green/brown polymorphism in the Pyrgomorphidae (Pyrgomorpha itself well exemplifies this polymorphism) or in the remaining subfamilies. Cautery of the A and B cells of the pars intercerebralis in Locusta led to green larval pigmentation, and other effects associated with hyperactivity of the corpora allata; these persisted if the nervous connections of the corpora allata were destroyed, but not if the corpora themselves were first removed from the animal (Girardie, 1967). The implication of these and similar results is that the A and B cell regions inhibit by their products the secretion of the corpora allata while the C cells facilitate it; this would form a control system

180

C. H. FRASER ROWELL

of the green/brown polymorphism which could be readily accessible to neural input from the periphery (Girardie and Joly, 1968). B. THE BLACK PIGMENT SYSTEM A N D THE CORPUS CARDIACUM

The first indication of a humoral factor controlling black pigmentation was obtained by Nickerson (1 954, 1956). Injections of haemolymph from gregarious phase Schistocerca gregaria hoppers into transiens or solitarious phase recipients resulted in an increase in black pattern; the reverse transfer had no such effects. The active agent was found to be ether-soluble, pH stable, non-proteinaceous and slowly degraded by boiling, which led Nickerson t o suggest a sterol; but a peptide could have similar properties and would be a more likely neurosecretory product. Staal (1961) found that implants of corpora cardiaca into Locusta larvae increased their black pigmentation. This was confirmed by Girardie and Cazal (1 965), who further showed that ablation of the corpus cardiacum led to transitory loss of pigment. If the pars intercerebralis was also cauterized, cardiectomy resulted in a permanent loss of pigment. Implantation of partes intercerebrales into cardiectomied animals also increased pigmentation; Girardie (1967) showed by microcautery that the active component of the pars is derived from Type C cells. The evidence thus suggests that the black patterning of gregarious larvae is regulated by the secretion of these cells, via the corpus cardiacum, which acts merely as a store. Highnam ( 1961) and subsequent workers have demonstrated a release of stored neurosecretory material from the corpus cardiacum after a variety of nonspecific stimuli, including flight, presence of mature males, low frequency electric shock, or tumbling in a rotating jar. Tumbling procedures also produced a significant increase in the black pigmentation of Schistocerca hoppers (Husain and Mathur, 1936a), and it is probable that action of this sort underlies the darkening properties of such factors as enforced locomotor activity, and crowding (pp. 164-1 65). Clarke (1 966) demonstrated differences in the histological appearance of at least Type A neurosecretory material in the nervi corporis cardiaci I of Locusta under different temperature regimes; it will be recalled that this also influences the amount of black pigmentation in Schistocerca and Locusta larvae (p. 177). However, the relevance of this system in solitarious locusts or grasshoppers during albedo responses to non-reflectant backgrounds

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

18 1

(Section I V A ) is quite unknown. In this response the releasing stimulus is not at all non-specific, but comprises a highly defined optical input. One way in which the gregarious transformation could inhibit the albedo response would be by blocking the pathway from the appropriate visual integration system to the Type C neuro secretory cells, but many other alternative hypotheses are tenable in the present absence of data. There are some complicating data available; Nicolas and Fuzeau-Braesch (1 968) confirm that gregarious adult Locusta isolated on a dark background will make very little response to the albedo, but if they have been previously raised at high temperatures which inhibit black pigmentation, their response to the dark background approaches that of isolated solitary animals, and they darken very well. This suggests a post-inhibitory excitatory process, though with very long time relations. C. OTHER ENDOCRINE CORRELATES OF PIGMENTATION

Phase transformation has a number of endocrine correlates, which are discussed in the reviews cited previously. The changes in the black pigment system and in the ground colour associated with phase transformation are described in Section IVC, and the hormonal control of these systems indicated in A and B above could clearly be integrated into the hormonal basis of phase transformation without conceptual difficulty. A number of additional lines of evidence are, however, less easily accommodated. Nickerson (1 956) studied the changes in pigmentation which occurred when portions of epidermis from solitarious or gregarious Schistocerca gregaria larvae were grafted into hosts of the same or opposite phase status. The results were complicated by degenerative changes, but are none the less suggestive. Grafts into hosts of the same phase produced no changes in either graft or host epidermis until degeneration set in. Solitarious green epidermis grafted into gregarious hosts lost the green colour, as would be expected from the dependence of green colour on high juvenile hormone levels. Grafts of yellow gregarious epidermis into solitarious hosts did not however become green, as might have been anticipated; instead, the yellow colour was maintained, and spread to surrounding host epidermis, in some cases extending over the whole animal uniformly. If these effects were not artifacts of the operational procedure, they may indicate that the response of the epidermis to the determination of gregarious phase is more complex and more persistent than usually thought.

182

C. H. FRASER ROWELL

A further interesting but very curious finding is the demonstration by Fuzeau-Braesch ( 1968), Nicolas ( 1969), and Nicolas et al., ( 1969) that short daily anaesthesia with CO, reverses or inhibits many of the characteristics of gregarious phase in crowded Locusta larvae, including the gregarious coloration and the inability to make a homochrome response to albedo. CO, is known t o have a variety of effects on neural activity, as seen in its anaesthetic action; the present finding suggests that it may also have specific and long-term effects on neurosecretion, or alternatively on the integrative areas which process gregarizing stimuli. Ellis and Carlisle (1961) suggested that the prothoracic glands might influence pigmentation in solitarious Schistocerca larvae. Removal of about three-quarters of the gland from green fourth instar hoppers caused a change t o yellow at the next moult. However, these effects were not found in Locusta by Staal. The COz treatment referred to above caused among other effects an unusual degree of retention of apparently functional prothoracic glands in adult Locusta, which might support Ellis and Carlisle’s contention, but the interactions of the retrocerebral complex and the prothoracic gland are likely t o be so complex in the experiments described that it is difficult to draw any useful conclusion from these data. The remaining evidence all suggests that this particular aspect of coloration is primarily a function of corpus allatum activity. Repeated implantation of supernumary corpora allata into female gregarious Locusta resulted in a variety of solitarious characters in the eggs and resultant offspring produced by that female, including pale coloured hatchlings; the dark colour characteristic of gregarious hatchlings was absent (Cassier, 1964). (In Locustana pardalina there is the further complication that the phase status of the female determines not only the colour of the hatchlings but also the diapause status of the eggs (Matthee, 1950), which is also a hormonally regulated character.) Together with the evidence cited above, this suggests that the corpus cardiacum and the corpus allatum may act reciprocally in their determination of coloration in solitarious and gregarious coloration, the latter being characterized by low levels of juvenile hormone and high activity of the corpus cardiacum system, resulting in an absence of biliverdin pigments and heavy black patterning, and the solitarious larvae having opposite characteristics. However, this simple apposition of the main endocrine glands is clearly an oversimplification even if coloration alone is considered, and is certainly not a valid general statement of the endocrine basis of phase (see also discussion in Staal, 1961).

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

183

The few data on the endocrine correlates of coloration in other orthopteran groups do not illuminate the situation in the Acridoidea. For example, Roussel (1 967) and Fuzeau-Braesch ( 1968) found that alletectomy resulted in black Gryllus bimaculatus, while even a fragment of corpus allatum or injection of synthetic juvenile hormone led to orange pigmentation; the effect was thought to be a direct one of hormone on integument. It is difficult to see where this finding relates to the Acridoid data, where melanization and black ommochrome synthesis appears to be effectively independent of corpus allatum function, unless it too indicates an inhibitory interaction between the corpora allata and cardiaca. The differences between the two groups are emphasized by the fact that the coloration induced by higher juvenile hormone levels is characteristic of crickets bred at high density, while in the acridoid locusts the opposite is true. VI. PIGMENTS

Despite a considerable body of work, the available information on the pigment chemistry of variable coloration is not easy to reconcile, and it is clear that further data are required. Most of the experiments so far have been performed in vitro on extracted material; it seems that some of the ambiguities will only be settled by microhistochemistry which will give information both on the unextracted pigment and also on its distribution in the cells. There is broad agreement on the range of pigment types found, though it should be noted that the analytic work has been virtually confined t o Schistocerca gregaria and to Locusta migratoria and other Oedipodine species. These pigments are melanins, ommochromes, carotenoids, and bile pigments and other pyrrole derivatives. Other pigments, such as flavins and pterines, are present in minute amounts. The confusion lies in the role that these pigments play in visible coloration. The following points should be borne in mind when assessing the evidence: i. Some of the pigments belong to chemical families which are still poorly known or have only recently been elucidated, such as the ommochromes. ii. Pigments exist both in the free form or as the prosthetic group of a chromoprotein. The same prosthetic group can be combined with a range of proteins, and thus acquire different properties, including different colours. iii. All the important pigments, with the exception of melanin, are

184

C. H. FRASER ROWELL

labile compounds which change colour readily with a change in redox potential. Thus biliverdin at different oxidation levels can be colourless, yellow, red or violet, blue or green; and mixtures of these can also give grey, brown, green and orange colours (Lemberg and Legge, 1949). The oxidative state of a pigment extracted into a particular solvent is likely to differ from that of a pigment granule in the cytoplasm, perhaps dramatically. iv. The colour of light reflected by a pigment depends on its concentration and its dispersal; thus melanin can produce orange, brown or black colour. Finally, pigments are distributed in determined ways in the integument, the epidermal cells, and the haemolymph, and the pigment in lower layers is often masked by those in the upper. General accounts of pigmentation of acridids or of specific species have been given in recent years by Fuzeau-Braesch (1963, 1965), Nolte (1 965), and Uvarov (1 966). This account will merely correlate the different views held on the pigmentary basis of variable coloration, and consider the discrepancies between them. A. THE GREEN COMPONENT OF THE GREENIBROWN POLYMORPHISM

Views on this pigment have passed through several historical stages. Prior to experimental work it was widely assumed that green pigmentation in phytophagous insects was derived from ingested chlorophyll, a view which became untenable in the late 'twenties. Przibram and Lederer (1933) considered that the green pigment of Dixippu haemolymph was derived from a complex of blue and yellow pigments. Further work, reviewed by Goodwin (1 952), led to the view that the green pigments of solitary Locusta and Schistocerca were of this type; the two components were identified as chromoproteins, the blue having as its prosthetic group mesobiliverdin and the yellow having either or both @-carotene or astaxanthine, both carotenoids. This view is accepted by most of the recent reviewers, and supported by recent experimental work on other groups; thus Willig (1969) isolated one biliverdin and four carotenoids, the most prevalent of which was isozeaxanthin derived from @-carotene,from the epidermis of Carausius, and four different biliverdins and three carotinoids, principally @-carotene, from Tettigonia. In a recent series of papers Passama-Vuillaume (1964, 1965a, b, 1966) and Passama-Vuillaume and Levita (1966) have advanced a different view: that the green pigment is not a complex

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

185

of blue and yellow components, but instead is a single water-soluble chromoprotein, having as its prosthetic group IXa-biliverdin. The studies were initially on mantids, but have been extended to both Locustu and Oedipodu ; slight differences between the Acridoids and Mantoids, as, e.g., in resistance of the pigment t o oxidation when irradiated with light, are attributed to differences in the protein fraction. It is at first sight difficult t o reconcile these two views. Many of the biochemical determinations which support the two-pigment hypothesis have been made on haemolymph rather than epidermis (for an acridid example see Dadd, 1961), but this is not the case with the results of Goodwin and Skrisukh (1951) who investigated them separately in both Locustu and Schistocerca. Further, Blackith and Blackith (1969) found that in Morabine Eumastacids and in Atructomorphu (Pyrg.) electrophoresis of haemolymph proteins produce bands which are colourless, yellow, or green, not blue. Dadd (1961) and Nayar (1964) have shown that diets deficient in carotene produce blue, rather than green, larvae in Locustu, Schistocerca and Melanoplus (Catant.), which supports the two-pigment hypothesis. Similar results have been obtained with the lepidopteran caterpillar Munducu (Dahlman, 1969) where the pigment is contained in the haemolymph. However, the effects of carotene deprivation were complex, causing the blue biliverdin pigment to be synthesized in all three Acridoid species in circumstances in which it would normally be absent (e.g., in the haemolymph of gregarious locust hoppers) and also disrupting normal melanin and ommochrome pigmentation. The explanation of these effects is not known; no metabolic disorder other than in pigmentation was observed. It is possible that retinene production was sufficiently disrupted to cause malfunction of the visual system, or it is conceivable that the various pigment systems within the epidermal cells are linked by feedback, such that dietary deficiency affects the other pigments merely by removing the normal car0tene pigmentation. It may also be that the discrepancy between the two theories is less than appears. The presence of carotene is not denied by any worker; the argument turns on the colour of the mesobiliverdinprotein complex. It is quite possible that this compound is sufficiently labile t o be subject to colour change between green and blue depending on conditions of extraction and analysis. Perhaps both forms exist in the epidermal cells, for there is appreciable variation in the colour of green assumed by a population of green larvae.

186

C. H. FRASER ROWELL

B. THE BROWN COMPONENT OF THE GREEN/BROWN POLYMORPHISM, AND THE BLACK AND ORANGE PIGMENT SYSTEMS

The coloration of the brown morph is known (see Section IV) to include components which respond differently under different environmental conditions; thus the Oedipodine genera possess a black pigment system; an “orange” pigment system which produces coloration varying from yellow to reddish or purple-brown, and which may itself be heterogenous; and possibly a separate ground-colour element, which obtains in the absence of any homochromic stimulation of these two systems. There is also the element of pattern, which is dealt with separately below. Unfortunately, few of the chemical investigations have been made on insects whose status with respect to these different components of coloration had been determined. It is perhaps not surprising that a variety of pigments have been implicated in the “brown” coloration. Of the major pigments, melanin is agreed to be confined to the integument; the remaining ommochromes, bile pigments and carotenoids have all been held to be involved in the coloration of the brown morphs. Redox-sensitive pigments soluble in acidic alcohol can be isolated from all morphs and stages of Schistocerca, and of Locusta and various other oedipodines, with the exception of extreme green morphs with no areas of brown coloration (Goodwin, 1952). A similar situation pertains in Mantis (Susec-Michieli, 1965). Such pigments have been isolated from many other insects. At least part of this fraction consists of ommochrome (phenoxazone) pigments. The simplest of these, xanthommatine, has been isolated by FuzeauBraesch (1 960, 1968) from Gryllus and from Locusta and Oedipoda, and found identical with the synthesized chemical. There seems little doubt that the black homochrome response t o non-reflectant backgrounds is mediated primarily by epidermal ommochrome, together with some cuticular melanin. Ommochrome is present in large amounts in individuals of Locusta, Gastrimargus and Heteropternis (Oed.) and Coryphosima (=Paracomacris, Gomph.) which have been experimentally darkened by rearing on a black background, or wildcaught on burnt grassland (Fuzeau-Braesch, 1965, 1966). Melanic patterning of the cuticle overlies more extensive black ommochrome pigmentation of the epidermis in both gregarious and solitarious colorations (Nickerson, 1956; FuzeauBraesch, 1965, 1966). This empirical association of the two pigments

THE VARIABLE COLORATION OF THE ACRIWID GRASSHOPPERS

187

may be due to a functional coupling between the synthetic processes which derive melanin from tyrosin and the ommochromes from tryptophane (Fuzeau-Braesch, 1963b). This hypothesis is supported by the finding that mutant albino Locusta are devoid not only of melanin but also of xanthommatine in the larval instars (Fuzeau-Braesch, 1 968). The role of ommochrome pigments in other variable components of the coloration is much less certain. Certainly the range of colours which ommochromes are capable of under appropriate redox conditions would provide the entire observed range of “brown” pigments. Goodwin (1952) concluded that they were responsible for the brown ground colour of adult Schistocerca and Locusta (though not the additional yellow of mature males) and of solitary brown-morph larvae, and this has been followed by several subsequent authors. However, it is not clear that the experimental evidence goes beyond showing that the brown coloration was not due to carotenoids and that ommochromes were present in the animal; in fact, the only analytic data Goodwin presents on the acidic-alcohol-soluble pigments is that they gave rise to pyrrole degradation products, which seems incompatible with ommochromes. Passama-Vuillaume (loc. cit.) contends that the brown ground colour of Locusta, Oedipoda, and Mantis is due not to an ommochrome but to the same biliverdin protein complex to which she attributes the green colour, but in a higher oxidative state. Certainly the correlation between the response of the biliverdin compound in vitro and of living Mantis religiosa to far red and far blue light seems compelling. The red and yellow granules found in epidermal cells of Oedipoda which has made a homochromic response to orange or red-brown backgrounds are also considered to be tetrapyrroles derived from biliverdin by oxidation (PassamaVuillaume and Levita, 1966). This interpretation is in accord with Goodwin’s finding of pyrrole products in the acidic-alcohol extract, rather than with his own ommochrome identification; a similar point was made by Okay (1 953). However, Passama-Vuillaume’s analysis applies t o the water-soluble brown pigments. Goodwin (1952), Susec-Michieli ( 1 9 6 3 , and Fuzeau-Braesch ( 1969) all confirm the presence of ommochromes in brown morphs in considerable amounts, and the former two authors and many previous workers have also found carotenoids. It seems probable that all are involved; it is tempting to suppose that the neutral yellow or yellow-grey

188

C. H. FRASER ROWELL

ground colour of brown morph insects which have made no homochromic response (Hertz and Imms, 1937; Levita, 1966; Rowell, 1970) consists of oxidized biliverdin chromoproteins and 0-carotene, with ommochromes providing the non-melanic patterning; that the response to low reflectance backgrounds causes synthesis of black ommochrome and some extra melanin; and that orange and red backgrounds result in further oxidation of biliverdin to give pyrrolic pigments of these colours. A special difficulty is raised by species (e.g., Gastrimargus (Oed.), or Chro togonus, Parasphena (Pyrg.)) which become almost pure white when raised on a white background. No analyses have been made of the pigmentation of such individuals, but it seems certain that either they contain an as yet unidentified white pigment masking the remainder, or that they have lost their brown pigments. This last presents no especial problem in the case of the bile pigments, for these can be oxidized to a colourless form; but this is not known to be possible for the carotenoids or the ommochromes, and these would have t o be actively removed. As ommochromes are not found in the haemolymph (Goodwin, 1952), this would imply their intracellular breakdown. C . IMPLICATIONS O F THE ABOVE FOR THE GREEN/BROWN POLYMORPHISM

The attraction of Passama-Vuillaume’s interpretation of pigmentation, in which the green and many of the brown colours are derived from basically the same biliverdin pigment, is that it allows the almost ubiquitous green/brown polymorphism to be correlated with a simple redox shift at the cellular level. However, further experimental verification is required before this view can be accepted without reservation. Even if this is the primary mechanism, it still requires to be supplemented by a further command sequence which will inhibit the production of ommochromes in the green morph, and allow their synthesis in the brown morph. Even this does not suffice t o explain all observations. For example, while oedipodines are usually either green or “brown” in a given area, which is compatible with the hypothesis, some green individuals are capable of making a homochromic response to a black background without concomitantly losing their green colour (pp. 165-6). It is clear that at least under these circumstances the inhibition of ommochrome synthesis (which appears to be epidermal, Goodwin, 1952) is not complete.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

189

D. PATTERN

Pattern, as used by Fuzeau-Braesch (1 965) or Rowel1 (1 970), is a heterogenous category. It includes both exocuticular melanin and epidermal ommochrome. In the case of the black markings which form part of the pattern of, e.g., Locusta and Schistocerca gregarious larvae there is a close correspondence between the two pigments, as noted in Section I1 above. However, pattern can also describe variation in colour or density of the epidermal pigments only, as in many larvae of solitarious Locusta or of Gastrimargus. In these animals the pattern is composed of denser areas of brown or grey pigment, and the moulted exocuticle has no corresponding melanic markings. It is presumed that this pattern is ommochromal, but the remarks in Section B above will show that this does not have to be the case. The melanic pattern of the exocuticle is perhaps the only element of grasshopper coloration in which there seems no remaining ambiquity with regard to the responsible pigments; the detailed confwmation of this has been largely due to Fuzeau-Braesch (1 963a, 1965, 1966). E. THE PHASE COLORATION OF GREGARIOUS LOCUST HOPPERS

In many species gregarious larvae are predominantly black and orange or yellow, in a bold pattern. The black patterning differs from that of the solitary form in amount and distribution, but has the same pigment chemistry. The yellow/orange ground colour is more debatable. Grayson and Tauber (1 943) thought that carotene was largely responsible for the differences between solitarious and gregarious coloration of Melanoplus sanguinipes (Catant .). Goodwin (1952) concluded that while carotenoids and ommochromes were both present in gregarious Schistocerca and Locusta, the yellow ground of the former was due solely to carotenoids, while that of Locusta in contrast was derived from a partial melanization of the cuticle, and not from underlying pigment. The evidence for this role of carotene in Schistocerca is the loss of yellow colour caused by extraction in acetone, although Goodwin’s plate and text indicate that only a small fraction of the yellow colour is in fact removed by this treatment. In Locusta larvae, acetone treatment produces no obvious change in the whole animal, and it was concluded that the orange colour was due not t o carotene, but to some other pigment, probably cuticular melanin. Dadd (1 963) raised gregarious

I90

C.

H. FRASER ROWELL

Schistocerca and Locusta larvae on diets deficient in carotenoids, and obtained individuals with less than the normal yellow coloration, which supports the view that carotene is the yellow epidermal pigment in both species. Histochemistry of frozen sections of epidermis and cuticle is required t o elucidate this point. The evidence seems to suggest that the typical gregarious phase coloration is derived by an overall simplification of the pigment systems in which at least the bile pigment complex is not manufactured. The absence of this pigment, if the appropriate hypotheses as to pigmentation are selected from those reviewed above, could simultaneously explain why gregarious hoppers have a yellow colour, rather than a brown; why they are never green, despite otherwise favourable environmental conditions; and why they are unable to make homochromic response to orange or red backgrounds. Added to this deficiency, gregarious hoppers must be presumed to have modified systems for the synthesis of ommochrome, which produce more black area than in the solitarious forms, but which are less responsive to environmental factors, especially background reflectance. ACKNOWLEDGEMENTS

I am grateful to Drs R. H. Dadd, S. Fuzeau-Braesch, and D. S. King for their critical reading of sections of this review, and especially to those who have kindly allowed me to cite unpublished work. The original work reported here was supported in part by Grant No. 259 from the Research Fund of Makerere University College, Uganda, and I thank Mr J. Ngirumwe for his untiring technical assistance in that work.

REFERENCES Abushama, F. T. ( 1 969). The effect of light in the dispersal of “yellow” melanin pigments in the integument of the grasshopperPoekilocerus hieroglyphicus (Klug). Experientia 2 5 , 326. Anon. (1969). Report of the Anti-Locust Research Centre, London, for 1967. Albrecht, F. 0. ( 1 964). Etat hygromdtrique, coloration et rdsistance chez l’imago de Locusta migratoria migratorioiiies R. and F. Experientia 20, 97. Albrecht, F. 0. (1965). Influence du groupement, de l’dtat hygromdtrique et de la photop6riode sur la r6sistance au jefine de Locusta migratoria migratorioTdes R. et F. (Orthoptsre acridien). Fanlac, Perigeux.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

19 1

Albrecht, F. 0. (1967). “Polymorphisme Phasaire et Biologie des Acridiens Migrateurs”. Masson et Cie, Paris. Aziz, S. A. (1958). Probable hygrocepters in the desert locust, Schistocerca gregaria Forskkl (Orthoptera: Acrididae). Indian J. Ent. 19,164-170. Bennett, R . R., Tunstall, J. and Horridge, G. A. (1967). Spectral sensitivities of single retinula cells of the locust. Z. vergl. Physiol. 5 5 , 195-206. Blackith, R. E. and Blackith, R. M. (1969). The anatomy and physiology of the morabine grasshoppers. IV. Proteins and pigments in solution. Aust. J. ZOO^. 17, 687-696. Brett, C. H. (1947). Interrelated effects of food, temperature, and humidity on the development of the lesser migratory grasshopper, Melanoplus mexicanus mexicanus Sauss. (Orthopt.). Tech. Bull. Okla. agric. Exp. Stn T26:50. Burton, J . G. (1960). Notes on Orthoptera in S.E. England in 1959. Entom. Rec. 72,68-7 1. Burtt, E. (195 1). The ability of adult grasshoppers to change colour on burnt ground. Proc. R . SOC. A . 26,4548. Byrne, 0. W. (1962). Colour pattern polymorphism in the Australian plague locust, Chortoiocetes terminifera (Orthoptera: Acrididae). Ph.D. Thesis, Adelaide 1962. Byrne, 0. W. (1967a). Polymorphism in the Australian Acrididae: I. Inheritance of colour patterns in the plague locust Chortoiocetes terminifera. Heredity, Lond. 22, 561-568. Byrne, 0. W. (1967b). Polymorphism in the Australian Acrididae: 11. Changes in colour pattern gene frequencies in the plague locust, Chortoiocetes terminifera. Heredity, Lond. 22, 569-589. Cassier, P. (1964). Effets immidiats et transmis des implantations de corps allates sur la ficonditC et la descendance des femelles de Locusta migratoria migratorioiaes ( R . et F.) phase gregaria (Insecte OrthoptCroYde). C.r. hebd. Seanc. Acad. Sci., Paris 259,2706-2708. Chapman, R. F. (1962). The ecology and distribution of grasshoppers in Ghana. Proc. zool, SOC.Lond. 139, 1-66. Clarke, K . U. (1966). Histological changes in the endocrine system of Locusta migratoria L. associated with the growth of the adult under different temperature regimes. J . Insect Physiol. 12, 163-170. Common, I. F. B. (1948). The yellow-winged locust, Gastrimargus musicus Fabr., in central Queensland. Qd. J. agric. Sci. 5 , 153-219. Cromartie, R . I . T. (1959). Insect Pigments. A . Rev. Ent. 4, 59-76. Dadd, R. H. (1 96 1). Observations on the effects of carotene on the growth and pigmentation of locusts. Bull. ent. Res. 52, 63-81. Dahlman, D. L. (1969). Cuticular pigments of tobacco hornworm (Manduca sexta) larvae: effects of diet and genetic differences. J. Insect Physiol. 15, 807-814. Dirsch, V. M. (1964). “The African Genera of Acridoidea”. Cambridge University Press, London. Duck, L. D. (1944). The bionomics of Schistocerca obscura (Fabr.). J. Kansas ent. SOC.17, 105-1 19. Dudley, B. (1964). The effects of temperature and humidity upon certain morphometric and colour characters of the desert locust (Schistocerca

192

C. H. FRASER ROWELL

gregaria Forskil) reared under controlled conditions. Trans. R. ent. SOC. Lond.116, 115-129. Dustmann, J . H. (1964). Die Redoxpigmente von Carausius morosus und ihre Bedeutung fur den morphologischen Farbwechsel. Z . vergl. Physiol. 49, 28-57. Ellis, P. E. and Carlisle, D. B. (1961). The prothoracic gland and colour change in locusts. Nature, Lond. 190, 368-369. Ergene, S. ( 1950). Untersuchungen iiber Farbanpassung und Farbwechsel bei Acrida turrita. Z . vergl. Physiol. 32, 530-55 1. Ergene, S . (1 95 la). Wahlen Heuschrecken ein homochromes Milieu? Dt. zool. Z. 1, 123-133. Ergene, S. (1951b). Hat homochrome Farbung Schutzwert? Dt. zool. Z . 1, 187-195. Ergene, S . (1952a). Farbanpassung entsprechend der jeweiligen Substratfarbung bei Acrida turrita. Z . vergl. Physiol. 34, 69-74. Ergene, S . (1952b). Spielt das Auge beim homochromen Farbwechsel von Acrida turrita eine Rolle? Z . vergl. Physiol. 34, 159-162. Ergene, S. ( 1952c). Homochrome Farbanpassung bei Oedipoda-larven. Z. vergl. Physiol. 34, 163-165. Ergene, S. (1952d). Erkennen homochrome Heuschrecken die Farbe oder die Graustufe ihrer Umgebung? Rev. Fac. Sci. Univ. Istanbul B17, 305-314. Ergene, S. (1 953a). Bevorzugung homochromen Milieus durch Oedipoda larven. Rev. Fac. Sci. Univ. Istanbul. B18, 63-66. Ergene, S. ( 1953b). Homochromer Farbwechsel ohne Hautung.bei Heuschrecken auf schwarzen Untergrund. Zool. Jb. 81,437-624. Ergene, S. ( 1 9 5 3 ~ )Weitere . Untersuchungen uber die biologische Bedeutung der SchutzfZrbung. Mitt. zool. Mus. Berl. 29, 127-133. Ergene, S. (1954a). Weitere Untersuchungen uber Schwarzanpassung bei Imagines von Heuschrecken. Mitt. zool. Mus. Berl. 30, 107-1 10. Ergene, S. (1954b). Uber den angeblichen Einfluss von frischem griinem Futter auf den Farbwechsel von Acrida turrita. Z . vergl. Physiol. 36, 235-237. Ergene, S. ( 1 9 5 4 ~ ) . Das Verhalten der Hamolymphe in Verlauf des Farbwechsels. Z . vergl. Physiol. 36,238-240. Ergene, S. (1 954d). Homochromer Farbwechsel bei Oedaleus-imagines. Zool. Anz. 153, 110-1 13. Ergene, S. (1955a). Uber die Faktoren, die grunfarbung bei Acrida bedingen. Z . vergl. Physiol. 37,221-225. Ergene, S. (1955b). Weitere Untersuchugen uber Farbanpassung bei Oedaleus decorus. Z . vergl. Physiol. 37,226-229. Ergene, S. ( 1 9 5 5 ~ ) .Die imaginale homochrome Farbwechsel bei Orthopteren. Rev. Fac. Sci. Univ. Istanbul. B20, 113-116. Ergene, S. ( 1956). Homochromer Farbwechsel bei geblendeten Oedaleus decorus Z . vergl. Physiol. 38, 3 11-3 14. Ergene, S. (1 957). Homochromie und Dressbarkeit nach Versuchen mit Oedipoda coerulescens-Imagines. Zool. Anz. 158, 38-44. Euw, J. V., Fishelson, L., Parson, J. A., Reichstein, T. and Rothschild, M. (1967). Cardenolides (heart poisons) in a grasshopper feeding on milkweeds. Nature, Lond. 214, 35-39.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

193

Fisher, R. A. (1930). The evolution of dominance in certain polymorphic species. A m . Nat. 64, 385-406. Fisher, R. A. (1 939). Selective forces in wild populations of Paratettix texanus. A n n . Eugen. 9 , 109-122. Faure, J. C. (1932). The phases of locusts in South Africa. Bull. ent. Res. 23, 293424. Faure, J . C. (1933). The phases of the Rocky Mountain locust Melanoplus mexicanus. J . econ. Ent. 26, 706-718. Fuzeau-Braesch, S. (1 960). Etude biologique et biochimique de la pigmentation d’un insecte, GryNus bimaculatus de Geer (gryllide, orthopt6re). Bull. biol. Fr. Belg. 94, 525-627. Fuzeau-Braesch, S. (1 963a). Etude de la pigmentation tkgumentaire des insectes A l’aide de radioklkments. In “Radiation and Radioisotopes Applied to Insects of Agricultural Importance”. International Atomic Energy Agency, Vienna. Fuzeau-Braesch, S. (1963b). Un carrefour de pigments: le tegument des insectes et ses curieuses adaptations. Science progrbs la Nature. No. 3336. Fuzeau-Braesch, S. (1 965). Etude du pigment cuticulaire du mutant albinos de Locusta rnigratoria (Orthop.) Bull. SOC.zool. Fr. 9 0 , 477-486. Fuzeau-Braesch, S. (1 966a). Etude du noircissement cuticulaire non-exuvial chez Locusta migratoria L. J. Insect Physiol. 12, 1363-1368. Fuzeau-Braesch, S. (1966b). Contribution A l’ktude de l’homochromie chez Locusta migratoria. Colloques int. Cent. natn. Rech. scient. 173,163-171. Fuzeau-Braesch, S . (1968). Action de l’hormone juvknile: synthbse dans la morphogknbse et la pigmentogknkse de Cryllus bimaculatus. C. r. SOC.Biol. Paris. 162, 1086. Fuzeau-Braesch, S. (1969). Recherche de la xanthommatine chez deux Insectes: Locusta migratoria L. et Oedipoda coerulescens L. C.r. hebd. Stanc. Acad. Sci., Paris. 267, 2030-2033. Girardie, A. (1967). Contr6le neuro-hormonal de la mktamorphose et de la pigmentation chez Locusta migratoria cinerascens. (Orthoptkre). Bull. Biol. Fr. Belg. 107, 79-1 14. Girardie, A. and Cazal, M. (1965). Rble de la pars interckrkbralis et des corpora cardiaca sur la mklanisation chez Locusta migratoria. C.r. hebd. Skanc. Acad. Sci., Paris. 261,4525-4527. Girardie, A. and Joly, P. (1968). Mecanism physiologique de l’effet de groupe chez les Acridiens. Colloques int. Cent. natn. Rech. scient. 173, 127-145. Golding, F. D. (1934). On the ecology of Acrididae near Lake Chad. Bull. ent. Res. 25,263-303. Good, C. M. (1941). The genetics of the grouse locust Tettigidea parvipennis Harris. Trans. Kans. Acad. Sci. 44, 235. Goodwin, T. W. (1952). The biochemistry of locust pigmentation. BiolRev. 27, 439-460. Grayson, J. M. and Tauber, 0. E. (1943). Carotin-the principle pigment responsible for variations in coloration of the adult grasshopper, Melanoplus bivattatus Say. Iowa St. Coll. J . Sci. 17, 191-196. Grayson, J. M. (1942). Studies of some factors influencing coloration of the grasshopper Melanoplus bivittatus Say. Iowa St. Coll. J. Sci 17, 69-70.

194

C. H. FRASER ROWELL

Hancock, J. L. (1916). Pink katydids and the inheritance of pink coloration. Ent. News. 27,70-82. Hemingway, E. (1963). “The Snows of Kilimanjaro and Other Stories”. Charles Scribners Sons, New York and Jonathan Cape, London. Hertz, M. and Imms, A. D. (1937). On the responses of the African Migratory Locust to different types of backgrounds. Proc. R . SOC.B122,281-297. Highnam, K. C. (1961). Induced changes in the amounts of material in the neurosecretory system of the desert locust. Nature, Lond. 191, 199-200. Hunter-Jones, P. (1957). An albino strain of the desert locust. Nature, Lond. 180, 236-237. Hunter-Jones, P. (1958). Laboratory studies on the inheritance of phase characters in locusts. Anti-Locust Bull. No. 29. Hunter-Jones, P. and Ward, V. K. (1960). The life-history of Gastrimargus africanus Saussure (Orth. Acrididae) in the laboratory. Entomologist’s mon. Mag. 9 5 , 169-172. Hunter-Jones, P. (1962). Coloration of the desert locust (Schistocerca gregaria Forskil) reared in isolation. Entomologist’s mon. Mag. 98, 89-92. Husain, M. A. and Mathur, C. B. (1936). Studies on Schistocerca gregaria Forsk. V. Pigmentation and physical exertion. Indian J. Agric. Sci. 6 , 591-623. Husain, M. A. and Ahmad, T. (1936). Studies on Schistocerca gregaria Forsk. VI. Influence of temperature o n the intensity and extent of black pattern in Desert Locust hoppers bred crowded. Indian J. Agric. Sci. 6 , 624-664. Husain, M. A. and Mathur, C. B. (1936). Studies on Schistocerca gregaria Forsk. VIII. Influence of carbon dioxide on development of black pigmentation in Schistocerca gregaria Forsk. Indian J. Agric. Sci. 6 , 1005-1030. Johnson, H. B. (1930). Report of the Government Entomologist of the Sudan Government for the year 1929. Johnston, H. B. (1932). Notes on 2 locusts of minor economic importance in the Sudan. Bull. ent. Res. 23,49-64. Joly, L. (1960) “Fonctions des Corpora Allata Chez Locusta migratoria (L.)”. Imprimerie des Dernihres Nouvelles de Strasbourg. Joly, P. (1951). Dkterminisme endocrine de la pigmentation chez Locusta migratoria L. C.r. Sianc. SOC.Biol. 145, 1362-1366. Joly, P. (1952). Dkterminisme de la pigmentation chez Acrida turrita. C.r. hebd. Sianc. Acad. Sci., Paris. 235, 1054-1056. Joly, P. (1962). RBle jouk par les corpora allata dans la rkalisation du polymorphisme de phase chez Locusta migratoria L. Colloques int. Cent. natn. Rech. Scient. 114, 77-88. Joly, P., Joly, L. and Halbwachs, M. (1956). Contr6le humoral du dkvelopment chez Locusta migratoria. Ann. Sci. Nat. Zool. 11, 257-261. JovanEiE, L. (1960). G6nkse des pigments tkgumentaires et leur r61e physiologique chez la mante religieuse et chez d’autres espbces animales. Posebna Izd. Prir. Muz. Belgrade. 29, 1-1 14. JovanEiE, L. (1963). Les caracthres mimktiques au service de la protection contre les radiations solaires nocives chez certains insectes et chez d’autres espkces animales. 16th Int. Congr. Zool., Washington, 1963. 1, 241. Karuhize, G. R. (1 968). The biology of Homorocoryphus (Orthoptera, Tettigoniidae) in Uganda. M.Sc. Thesis, Univ. of E. Africa.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

195

Key, K. H. L. (1954). “The Taxonomy, Phases and Distribution of the Genera Chortoiocetes Brunn. and A ustroiocetes Uv. (Orthoptera: Acrididae)” Canberra. Key, K. H. L. and Day, M. F. (1 954a). A temperature-controlled colour change in the grasshopper Kosciuskola tristis Sjost. (Orthopt., Acrididae). Aust. J. ZOO^. 2, 309-339. Key, K. H. L. and Day, M. F. (1954b). The physiological mechanism of colour change in the grasshopper Kosciuskola tristis Sjost. (Orthopt., Acrididae). Aust. J . ZOO^. 2 , 340-363. King, R. L. and Slifer, E. H. (1955). The inheritance of red and blue hind tibiae in the lesser migratory grasshopper. J. Hered. 46, 302-304. Lemberg, R. and Legge, J. W. (1949). “Haematin Compounds and Bile Pigments”. Interscience Publishers, Inc. New York. Levita, B. (1 966a). MCcanisme pigmentaire de l’homochromie chez Oedipoda coerulescens L. (Acridien, Orthoptgre). C.r. hebd. Skanc. Acad. Sci., Paris. 262, 1763-1 765. Levita, B. (1 966b). Pigmentation et photosensibilitC chez Oedipoda coerulescens L. (Acridien, Orthoptke). C,r. hebd, S6anc. Acad. S c i , Paris. 262, 2496-2497. Levita, B. (1969). New method of spectrophotometric and colorimetric investigations on insect integuments in vivo. Zeiss Information 68,44-50. Lewontin, R. C. and White, M. J. D. (1960). Interaction between inversion polymorphisms of two chromosome pairs in the grasshopper Moraba scurra. Evolution 14, 116-129. Linzen, B. (1 959). Ueber neue Ommochrome aus Heuschrecken, Verh. XVII int. Kong. reine u. angew. Chemie, Munchen. Matthee, J. J. (1950). The egg of the brown locust. Fmg. S. A fr. 25,255-257. Meadows, P. S. ( 1967). Discrimination, previous experience, and substrate selection by the amphipod Corophium. J. exp. Biol. 47, 553-560. Nabours, R. K. (1 929). “The Genetics of the Tettigidae”. Verlag M. Nijhoff, den Haag. Nayar, J. K. (1964). The nutritional requirements of grasshoppers. I. Rearing of the grasshopper Melanoplus bivittatus (Say) on a completely defined synthetic diet and some effects of different concentrations of B-vitamin mixture, linoleic acid, and B-carotene. Can. J. Zool. 42, 11-22. Nel, M. D. (1967). Selection of phase types based on hopper coloration in Locusta migratoria migratorioides (R. and F.). S. Afr. J. Agric. Sci. 10, 4 6 1469. Nel, M. D. (1967). Selection of phase types based on hopper coloration in Locustana pardalina (Walker). S. A fr. J. Agric. 10, 823-830. Nel, M. D. (1968). Selection at a high humidity for green and brown solitaria hopper coloration in Locustana pardalina (Walker). S. A fr. J. Agric. Sci. 11, 163-172. Nickerson, B. ( 1 954). A possible endocrine mechanism controlling locust pigmentation. Nature, Lond. 174, 357-358. Nickerson, B. ( 1956). Pigmentation of hoppers of the desert locust (Schistocerca gregaria Forskil) in relation to phase coloration. Anti-Locust Bull. 24, 34 PP.

196

C . H. FRASER ROWELL

Nicolas, G. (1969). Action du gaz carbonique chez Locusta migratoria L. (Orthoptkre-Acrididae). Doctoral Thesis, UniversitC de Rennes. Nicolas, G., Cassier, P. and Fain-Maurel, M. A. (1969). Evolution vers le phenotype solitaire et persistance des glandes de mue chez Locusta rnigratoria cinerascens, phase grkgaire, sous l’influence du gaz carbonique. Etude exphimentale et infrastructurale. C.r. hebd. Stanc. Acad. Sci., Paris 268, 1532-1534. Nicolas, G. and Fuzeau-Braesch, S. (1968). Etude de quelques facteurs controlant l’homo chromie chez L ocusta m igratoria rnigratorioides. (Orthoptkre). C.r. Stanc. SOC.Biol. 162, 1091. Nolte, D. J. (1 962). Strain differentiation in locusts. Proc. 2nd. Congr. S. Afr. Genet. SOC.Pretoria, 1962, 96-100. Nolte, D. J. (1963). A pheromone for melanisation of locusts. Nature, Lond. 200,600-66 1. Nolte, D. J . (1965). The pigmentation of locusts. S. Afr. J. S c i 61, 173-178. Novak, V. J. A. and Ellis, P. E. (1967). The metamorphosis hormones and the phase dimorphism in Schistocerca gregaria. 11. Implantations of the glands into hoppers reared in crowded conditions (Abstract). Gen. comp. Endocr. 9,477478. Okay, S. (1953). Formation of green pigment and colour changes in Orthoptera. Bull. ent. Res. 44, 299-3 15. Okay, S. (1956). The effect of temperature and humidity on the formation of the green pigment in Acrida bicolor (Thunb.). Archs int. Physiol. Biochim. 64,80-9 1. Owen, D. F. (1956). Swarming and polymorphism in the African edible grasshopper Hornorocoryphus nitidulus (Tettigonoidea, Conocephalidae). Acta Trop. 22,55-61. Passama-Vuillaume, M. (1964). Sur la pigmentation verte de Mantis religiosa ( L ) . C.r. hebd. SPanc. Acad. Sci., Paris (0) 258,6549-6552. Passama-Vuillaume, M. (1 965a). Etude de l’irradiation lumineuse, facteur essentiel du brunissement de Mantis religiosa ( L ) . C.r. hebd. Stanc. Acad. Sci., Paris (0) 261, 3683-3685. Passama-Vuillaume, M. (1965b). Etude du pigment vert chez Locusta migratoria L . normal et albinos. Bull. SOC.zool. Fr. 9 0 , 4 8 5 4 9 2 . Passama-Vuillaume, M. ( 1966). Sur la pigmentation brune de trois Orthoptkres: Mantis reZigiosa(L), Sphodromantis viridis (F), et Locusta migratoria (L). C.r. hebd. Stanc. Acad. Sci., Paris (0) 262, 1597-1599. Passama-Vuillaume, M. and Levita, B. ( 1966). Sur les pigments tetrapyrroliques d’oedipoda coerulescens. C.r. hebd. Seanc. Acad. Sci., Paris (0)263, 1001-1003. Peterson, B. and Treherne, J. E. (1949). On the distribution of colour forms in Scandinavian Omocestus viridulus L. Oikos 1, 175- 183. Pfeiffer, I. W. (1945). Effect of corpora allata in the metabolism of adult female grasshoppers. J. exp. Biol. 99, 183-233. Plotnikov, V. 1. (1926). “Insects Injurious to Agricultural Plants in Central Asia” (in Russian). Tashkent. Popov, G. (1952). Apparent tendency t o phase variation in an Iranian grasshopper, Pygodera armata (F.W.) (Orthoptera, Acrididae). Eos, Madr. 28,277-283.

THE VARIABLE COLORATION OF THE ACRIDOID GRASSHOPPERS

197

Poulton, E. B. (1926). Protective resemblances borne by certain African insects to the blackened areas caused by grass fires. Proc. 3rd Int. Congress Ent. Zurich, July 1925. 2,433. Putnam, L. G. (1 958). Albinism in the migratory grasshopper Melanoplus bilituratus (Wlk.). Nature, Lond. 182, 1529. Ramme, W. (1 95 1). Die parallele Farbungsvariation (“Parallelochromie) des Acrididae; ihre genetische and phylogenetische Bedeutung. (Orthopt.). Eos, Madr. 1950,249-266. Reichstein, T. ( 1967). Cardenolide (herzwirksame Glykoside) als Abswehrstoffe bei Insekten. Natunv. Rdsch., Stuttg. 20,499-51 1. Riegert, P. W. (1960). The humidity reactions of Melanoplus bivittatus (Say) (Orthoptera, Acrididae): antenna1 sensilla and hygroreception. Can. Ent. 92, 8 . Richards, 0. W. and Waloff, Z. (1954). Studies on the biology and population dynamics of British grasshoppers. Anti-Locust Bull. 17, 182 pp. Rothschild, M. and Parsons, J. (1 962). Pharmacology of the poison gland of the locust Poekilocerus bufonius Klug. Proc. R . Ent. SOC. Lond. (C). 27, 21-22. Roussel, J. P. (1967). Fonctions des corpora allata et contrble de la pigmentation chez Gryllus bimaculatus de Geer. J. Insect Physiol. 13, 1 13-130. Rowell, C. H. F. (1967a). Experiments on aggregations of Phymateus pupurascens (Orthoptera, Acrididae, Pyrgomorphinae). J. Zool., Lond. 152, 179-193. Rowell, C. H. F. (1967b). Corpus allatum .implantation and green/brown polymorphism in three African grasshoppers. J. Insect Physiol. 13, 1401-1412. Rowell, C. H. F. (1970). Environmental control of coloration in Gastrimargus africanus (Acridinae, Orthoptera). Anti-Locust Bull. 4 7 , 4 8 pp. Rowell, C. H. F. and Cannis, T. L. Environmental factors affecting green/brown polymorphism in Schistocerca vaga. (In preparation.) Rubtzov, I. A. (1935). Phase variation in non-swarming grasshoppers. Bull. ent. Res. 26,499-574. Sansome, F. W. and La Coeur, L. (1935). The genetics of grasshoppers: Chorthippus parallelus. J. Genet. 30,415422. Sjostedt, Y . (1929). Acridoidea aus Zentralafrika gesammelt von R. Grauer wiihrend seiner Expedition 1909-191 1. Ark. 2001.A. 20, 1-41. Slifer, E. H. (1955). The detection of odors and water vapor by grasshoppers (Orthoptera, Acrididae) and some evidence concerning the sense organs which may be involved. J. exp. Zool. 130, 307-3 18. Staal, G. B. (196 1). Studies on the physiology of phase induction in Locusta migratoria migratorioides. Publs Fonds. Landb. Export. Bureau 1916-1918, 40. 125 pp. Stower, W. J. (1959). The colour patterns of hoppers of the Desert Locust (Schistocerca gregaria Forskil). Anti-Locust Bull. 32, 75 pp. Stehr, G. (1959). Haemolymph polymorphism in a moth and the nature of sex-controlled inheritance. Evolution 13, 537-560. Susec-Michieli, S. (1965). Biochemical investigations of the morphological colour changes in Mantis religiosa (Dictyoptera). Proc. XII Int. Congr. Ent. (Lond.) 1964, 208-209.

198

C.

H. FRASER ROWELL

Swihart, S. L. (1969). Colour vision and the physiology of the superposition eye of a butterfly (Hesperiidae). J. Insect Physiol. 1 5 , 1347-1365. Uvarov. B. P. (1966). “Grasshoppers and Locusts”, Vol. I. Cambridge University Press, London. Uvarov, B. P., Chapman, E., Waloff, N. and Waterston, A. R. (1951). Observations on the Moroccan Locust (Dociostaurus maroccanus Thunberg) in Cyprus, 1950. Anti-Locust Bull. 10, 5 5 pp. Verdier, M. (1958). Modifications pigmentaires likes rl la densit6 chez les tettigonides. Bull. SOC.Zool. Fr. Paris. 83, 252-253. Verdier, M. (1965). La mutation albinos de Locusta migratoria. Origine et description. Bull. Soc. 2001. Fr. 9 0 , 4 9 3 . Volkonsky, M. A. (1 938). Une mutation mklanique de Schistocerca gregaria Forsk. obtenue en klevage. C.r. Stanc. SOC.Biol. 127, 254-256. Vorontzovskii, P. A. (‘1938). On the question of homologous ranges of colour variation in Acrididae. Izv. Orenburg Stants Zaslieh. Rast. 1 , 27-39. (In Russian .) Walter, C. (1 965). Colour phases in Uganda grasshoppers. MSc. thesis, Univ. of E. Africa. White, M. J. D. (1968). A gynandromorphic grasshopper produced by double fertilisation. Aust. J. 2001.16, 101-109. Willig, A. (1969). Die Carotinoide und der Gallenfarbstoff der Stabheuschrecke Carausius morosus, und ihre Beteiligung in der Entstehung der Farbmodifikation. J. Insect Physiol. 15, 1907-1928. Wise, A. J. (1966). A rare colour variety of Chorthippus brunneus Thunberg (Orthoptera: Acrididae) from Wiltshire. Entomologist 99, 294. de Wilde, J. and Staal, G. B. (1955). Kleuranpassungsvermogen bij Springhanen (Orthopt.). Ent. Ber., Arnst. 1 5 , 4 9 7 4 9 8 . Wiersma, C. A. G. and Yamaguchi, T. (1967). Integration of visual stimuli by the crayfish central nervous system. J. exp. Biol. 4 7 , 4 0 9 4 3 2 .

The Mechanisms of Insect Excretory Systems S . H . P . MADDRELL Agricultural Research Council Unit of Invertebrate Chemistry and Physiology. Department of Zoology University of ambridge. England Introduction . . . . . . . . . . . . . . . . . . Deposit and Storage Excretion . . . . . . . . . . . . 111. Less Common Excretory Organs . . . . . . . . . . . . A. The Excretory Role of the Pericardial Cells and Nephrocytes B . TheMidgut of Larvae of Saturniid Silkmoths . . . . . C. TheLabialGlandsof Saturniid Silkmoths . . . . . . D. The Anal Papillae of Mosquito Larvae . . . . . . . . IV. The Malpighian Tubules A . The Malpighian Tubules of Carausius . . . . . . . . B . The Malpighian Tubules of Calliphora . . . . . . . . C. The Malpighian Tubules of Tipula . . . . . . . . . D. The Malpighian Tubules of Rhodnius . . . . . . . . E . The Malpighian Tubules of Calpodes . . . . . . . . F. The Ultrastructure of Malpighian Tubules and Its Functional Significance . . . . . . . . . . . . . . . . G. Formed Bodies . . . . . . . . . . . . . . . H . The Handling by Malpighian Tubules of Organic Solutes . . V. TheHindgut A. The Action of the Hind-gut Anterior to the Rectum . . . . . . . . . . . . . . B. The Action of the Rectum Rectal Absorption of Ions and Water in Schistocerca . . . C. D. Rectal Absorption of Ions and Water in Calliphora . . . . The Mechanism of Water Absorption by the Rectum . . . E. The Mechanism of Ion Absorption by the Rectum . . . . F. G . Rectal Recovery of Amino Acids, Sugars and Other Small Organic Molecules . . . . . . . . . . . . . . H . The Role of the Cuticular Lining of the Rectum . . . . I. Absorption of Water Vapour from Subsaturated Atmospheres by Thermobia . . . . . . . . . . . . . J. Absorption of Water Vapour from Subsaturated Atmospheres by Tenebrio . . . . . . . . . . . . . . VI . Concluding Remarks . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . I. I1 .

199

200 201

204 205 206 209 212 213 217 238 238 264 268 276 279 287 289 291 295 296 303 304 304 307 310 319 324 324

200

S. H. P. MADDRELL

I. INTRODUCTION

It is usual under the heading of insect excretion to consider almost exclusively the activities of the Malpighian tubules and the rectum; the tubules producing what is essentially an ultrafiltrate of the haemolymph from which the rectum selectively reabsorbs or adds substances according to the current needs of the insect. Such a consideration, while it undoubtedly focuses attention on the most important aspect of the functioning of the excretory system, does not include other activities which might reasonably and importantly Secretion Controlled uptake Unavoidable uptake

* +

Metabalic pod

*

EXCRETION

Unavoidable loss

Fig. 1 . Flows of materials into and out from the metabolic pool of an organism.

be considered as excretion. At this stage it is, therefore, necessary to define the term excretion. A functionally satisfactory definition is that excretion covers those processes which lead t o the effective removal of substances from the metabolic pool of the organism in order to avoid the harmful effects their continued presence would have on metabolism. Figure 1 shows the many ways in which substances enter and leave the metabolically active parts of an organism. The differences between the processes whereby substances leave the metabolic pool may be explained in terms of the concept of purpose. Broadly and teleologically speaking, molecules are secreted from the metabolic pool to achieve some purpose which is positively useful to the organism outside the metabolic pool. Excretion may be thought of as those processes which remove substances from the metabolic pool for the negative reason that they interfere with ordered metabolism. In addition, some substances may leave by neither of these processes. This is exemplified by the loss of water by evaporation-a process which can be defined as unavoidable loss. As with so many biological criteria it is not always possible to achieve a tight definition and decide whether a substance has been secreted or excreted. For example, the incorporation of pteridines and uric acid

T H E MECHANISMS OF INSECT EXCRETORY SYSTEMS

201

into the wings of Pieris brassicae serves the dual purposes of removing potentially injurious nitrogenous wastes and of usefully pigmenting the wings (Harmsen, 1966). A definition of excretion like that advanced in the above paragraph allows us t o focus attention on several different methods and organs of excretion. One large area which will not be exhaustively considered in this chapter is detoxification. By this is meant those processes in which metabolically deranging substances are altered so that their poisonous effects are very much reduced or abolished. This might be termed biochemical excretion. Important cases of this are the conversions of ammonia t o uric acid or urea and the detoxification of extrinsic poisons such as the conversion of DDT to the relatively harmless DDE by DDT-resistant insects (Brown, 1958). Those processes which are examined more fully here are related to three separate aspects of excretion. The first is the excretion of molecules which are undesirable, perhaps even poisonous, at all except exceedingly low concentrations. The second is the excretion of molecules which are ordinarily useful or, indeed, essential to L metabolism but which are present t o excess. Under this second heading comes excretion of water and ions and this process is of course the outwardly directed part of osmoregulation. The third process is the excretion of substances that are not toxic but merely useless. Their excretion is worth while because if allowed to accumulate they would become obstructive. 11. DEPOSIT AND STORAGE EXCRETION

Another fundamental division of the subject depends on whether the excreted material is removed from the body of the animal or merely isolated within it by confinement in a particular tissue or in a special physico-chemical state. These latter cases are usually covered by the term storage excretion, but since this implies the later usefulness of the stored material it ought properly to be reserved for such processes as the removal of osmotically embarrassing concentrations of substances such as glucose by deposition of glycogen in the fat body. The term deposit excretion would be better employed for those cases of isolation of poisonous or useless molecules, as for example, with deposits-of uric acid found in many insects. Considering storage excretion first, one can ask if insects, which as

202

S . H. P. MADDRELL

small truly terrestrial animals face the problem of water conservation, might store water from periods when it is available to excess. Such vertebrate animals as the toad and frog do exactly this (Jprgensen, 1950). Surprisingly, it appears so far that among insects there is only one case where this has come to light-in Dysdercus fasciatus (Berridge, 1965a). During the last larval stage of this insect, which lasts approximately eight days, the animal only feeds during the first five days. In this period the rate of excretion of urine from the body is high (Fig. 2), but little is retained in the rectum. Just at

\

0

I

I

1

I

I

I

I

I

2

3

4

5

6

7

Days

F i g . 2. Rate of excretion (open circles) and accumulation of fluid in the rectum (solid circles) by fifth instar Dysdercus (data from Berridge, 1965a).

the end of the feeding period, however, there is a dramatic increase in the size of the rectum (Fig. 2) which depending on the humidity may swell with watery fluid until it weighs more than 10% of the body weight. Thereafter practically no fluid is voided from the rectum. Its size, however, decreases and the osmotic pressure of its contents concomitantly increases as water is resorbed into the haemolymph to replace that lost by evaporation. The rate of this change depends o n the rate of evaporated loss. Berridge has calculated that the insect is potentially capable of replacing more than 40% of the volume of its haemolymph from this water store in the rectum. Measurements of the osmotic pressure of the haemolymph and rectal contents showed that the insect could keep

203

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

its haemolymph at a constant osmotic pressure while that of the rectal contents was rising sharply (Fig. 3). It is clear that in this case the insect can use the excretory system to store water gained during feeding to tide it over a period when feeding activity ceases. It would be interesting to look for further examples of such storage excretion among other insects which feed in a similar discontinuous way.

-!- 160-, zs

2

140-

Haemolymph

‘“r1 e

+.

00

60

8 40

8

20 OO

I

2

3

4

5

6

7

0

Days

Fig. 3. Osmotic concentrations of the haemolymph (solid circles) and the contents of the rectum (open circles) of fifth instar Dysdercus (redrawn from Berridge, 1965a).

We turn now from the storage of materials which, though surplus to requirements, may later be useful, to cases where useless or poisonous materials are deposited in special areas of the body. This is properly called deposit excretion. The best known cases of deposit excretion concern the fate of uric acid. Uric acid is the characteristic end-product of nitrogen metabolism in terrestrial insects (though by no means the only one, Bursell, 1967). In most groups of insects it is eliminated through the Malpighian tubules to appear in the urine. In a few insects this process for some reason does not occur (or does not proceed fast enough), so that some other site is needed in which uric acid can be deposited. Particularly is this the case in cockroaches, where uric acid is deposited in special urate cells in the fat body (Kilby, 1963). However a number of other insects also adopt the same solution, notably endoparasites of insects such as the Hymenopteran Nemeritis canescens (Corbet and Rotheram, 1965), which do not eliminate

204

S. H. P. MADDRELL

uric acid from the body, presumably so that they do not interfere more than they need to with the metabolism of the host. There is also a report of uric acid deposition in the fat body of a species of beetle (Gupta and Sinha, 1960). With high protein diets, the increased deposition of uric acid in the fat body may become a problem. In Periplaneta americam for example, the fat body becomes greatly swollen by deposits of uric acid (Haydak, 1953) and this may perhaps be the cause of reduced longevity observed in such insects (Martignoni, 1964). In some insects excess uric acid may find its way into the epidermis and, because of its white colour in the crystalline state, perform a useful function there. The wings of Pieris brassicae owe their white colour partly to deposits of uric asid (Harmsen, 1966), while the white bands on the abdomen of Dysdercus are attributable to uric acid granules in the epidermal cells (Berridge, 1965b). A bizarre development is the ability of male cockroaches of several species to accumulate large amounts of uric acid in the utriculi majores of the accessory sex glands (Roth and Dateo, 1964). These glands may contain up to 88% by weight of uric acid and the uric acid may comprise 5% of the weight of the insect (Roth and Dateo, 1965). In Blattellu germanica, the utriculi majores of the recently emerged male contain no uric acid at first but they fill with white granules of uric acid in 1-2 days. If the male does not have access to females the glands become very enlarged and fill up a large part of the abdominal cavity. At mating, however, the glands empty almost completely and cover the spermatophore and parts of both male and female with a chalky white deposit. The significance of this development is unclear-the more so when it is found that closely related species may not have such uric acid filled glands (Roth and Dateo, 1965).

111. LESS COMMON EXCRETORY ORGANS

From the definition of excretion advanced on p. 200, several organs other than the Malpighian tubules and rectum may be implicated in activities which are highly important parts of excretion. As is discussed on p. 306 the Malpighian tubules and rectum do not seem suited to deal with the excretion of substances of high molecular weight. The excretion of such substances is the function of other organs. Among these are the pericardial cells and nephrocytes.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

205

A. THE EXCRETORY ROLE OF THE PERICARDIAL CELLS AND NEPHROCYTES

The pericardial cells as their name suggests are to be found arranged in a dense array around and along the length of the heart and aorta. Their function has recently been reviewed by Wigglesworth (1 970). The most characteristic property of these cells is their ability to take up colloidal particles from the circulating blood. For example, haemoglobin, chlorophyll, egg white and dyes of high molecular weight such as trypan blue and ammonia carmine when injected into the haemolymph all appear as inclusions in the pericardial cells (Hollande, 1922; Lison, 1937). Various other cells scattered around the body perform the same function (Wigglesworth, 1970) and so th$y have been termed nephrocytes. The ultrastructure of these cells show them t o be very active pinocytotically and they contain many coated vesicles. These characteristics are found in tissues engaged in the selective uptake of protein. It may well be then that the pericardial cells are particularly active in ingesting proteins from the haemolymph. The cytoplasm contains large numbers of lysosomes so that it is probable that the cells degrade the ingested proteins and return them to the haemolymph as amino acids. For example, from Wigglesworth’s work on the fate of haemoglobin after ingestion by Rhodnius (Wigglesworth, 1943), it is clear that trace amounts of haemoglobin are absorbed without being digested and that this circulates in the haemolymph. Most of this is taken up by the pericardial cells and is there converted to biliverdin. Wigglesworth suggests that the digested products of the protein moiety are returned into the circulating haemolymph. Whether or not foreign or waste proteins are toxic or merely useless and thus obstructive, their removal can clearly be classed as excretion in the sense outlined on p. 200. Larger particles than are dealt with by the pericardial cells are engulfed by phagocytic haemocytes as are invading bacteria and even metazoan parasites. The ways in which the excretory system handles the range of sizes and materials to be excreted is discussed on p. 306. It seems very likely that the pericardial cells and the nephrocytes are largely responsible for excretion of materials of the size of colloidal particles, particularly proteins. Not only do parts of the body other than the Malpighian tubules and rectum handle excretion of large substances, they often play an important role in the excretion of smaller molecules. We have already seen (p. 203) that uric acid may be deposited in the fat body and

206

S. H. P. MADDRELL

epidermis or may be eliminated from the body by the utriculi majores of male cockroaches (p. 204). Several other organs contribute to the excretion of ions and water. B . THE MIDGUT OF LARVAE OF SATURNIID SILKMOTHS

Silkmoth larvae feed voraciously on leaves of such plants as the mulberry, willow, holly oak, etc. Such a diet contains very large amounts of potassium and relatively small amounts of sodium. For example, leaves of Viburnum notatum on which larvae of Hyulophoru cecropia feed were found to contain 153 mE . kg" of potassium and only 4 mE . kg-' of sodium (Harvey, and Nedergaard, 1964). This diet is reflected in the composition of the midgut contents which has 208 mmol . 1-' potassium and 0.7 mmol . 1-' sodium. The potassium concentration in the haemolymph (27 mmol . 1 - I ) is strikingly lower than that of the gut contents while the sodium concentration in the haemolymph (6.0 mmol . 1 - I ) is close to that of the leaves. Thus the various tissues are protected from what is a potentially hostile ionic environment in the gut. It turns out that this protection is very largely due to a highly developed ability of the midgut of this animal to pump potassium at a very high rate into the midgut even in the face of extremely steep electrochemical potential gradients (Harvey and Nedergaard, 1964). The basis of this potassium movement has been the subject of a good deal of interest, with much research still in progress. So far it has been shown that: (i) Potassium is pumped by a sodium-independent potassium pump (Harvey and Nedergaard, 1964). (ii) The pump is specific for potassium-it will not normally transport sodium, lithium or choline; rubidium ions which closely resemble potassium ions chemically can however substitute fully for potassium (Nedergaard and Harvey, 1968; Wood, 1971). (iii) The potassium pump is insensitive t o ouabain, to the vertebrate hormones vasopressin and oxytocin and t o drugs such as eserine, atropine, and adrenaline. (iv) The pump is almost certainly situated on or in the apical plasma membrane of the epithelial cells where they face the lumen of the gut (Wood, Farrand and Harvey, 1969). (v) The short-circuit current developed by isolated midguts is wholly attributable t o the potassium transport (Wood, 1971). (vi) The specific radioactivity of potassium pumped into the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

207

lumen approaches that of the bathing solution on the blood side only after a lag of about 35 min (at a potassium concentration in the bathing solutions of 35 mmol . 1 - I ) (Wood, 1971). This shows that the transported potassium mixes with a significant proportion of the potassium within the tissue before it emerges into the lumen. One possibility is that potassium transport is performed by only one of the two types of cells present in the epithelium. The goblet cells, from their ultrastructural appearance (Anderson and Harvey, 1966), b seem a likely choice. (vii) The potassium content of the tissues of isolated midguts falls from 90 mmol . kg-' of tissue to 65 mmol . kg-' in 30 min, when bathed in 32 mmol . 1-' K+ solutions (Harvey and Zerahn, 1969; Wood, 197 1). During this same interval the short-circuit current developed by the midgut falls by about 30% (Wood, 1971). After this time both the potassium content and the short-circuit current fall away very much more slowly. It could well be that these facts are related and that the pump draws on intracellular potassium and that its rate of pumping depends on the concentration of potassium available to it. If the pump is on the apical plasma membrane (see (iv) above) then this is as would be expected. Figure 4 shows in diagrammatic form a summary of these facts and proposed mechanisms for the silkworm midgut. Now that attention has been focused on the potassium pump, future research might well concern itself with such topics as the energy supply of the pump. Another interesting question which arises is the extent to which large-scale movements of potassium give rise to concomitant movements of water. If, as indicated in the diagram, potassium enters the cells largely from the lumen (because of the very much larger difference in electrochemical potential for potassium across the plasma membrane on this side in vivo) to be pumped actively back into the lumen, on the face of it there is little reason to suppose that water movements may occur. However, if potassium ions enter through the membranes readily accessible to the lumen contents and are pumped back via the confined space of the goblet cavity, one would expect water to be osmotically dragged into the goblet cavity from the goblet cell cytoplasm. Whether this water is replaced from the haemolymph, in which case a large net trans-wall water movement would occur, or from the lumen contents would depend on the relative areas and permeabilities of the membranes facing the haemolymph and lumen contents. In one series of experiments, Nedergaard and Harvey (1 968) showed an increase in

Fig. 4. The midgut of larvae of Saturniid silkmoths drawn to show both the ultrastructure of the tissue and the inferred route of potassium movements in uiuo. Broken lines represent passive movements and the continuous line active transport.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

209

the concentration of potassium in the solution bathing the lumen side of an isolated short-circuited midgut. This increase was of a size which would be expected if there had been no net trans-wall water movement. Presumably then water movement, if any, must involve absorption from the lumen and its subsequent return. This might involve flow in through the columnar cells and out through the goblet cells. If so the inward flow would lead t o an elevated concentration of solutes between the microvilli of the columnar cells similar to the “sweeping-in” ‘effect described by Berridge and Oschman (1969) for Malpighian’tubules. It is conceivable that this might speed absorption of food materials by the gut. C. THE LABIAL GLANDS OF SATURNIID SILKMOTHS

The labial glands of adult saturniid silkmoths are a pair of long convoluted tubules which lie largely in the ventral part of the thorax. The anterior ends of the two glands unite and open t o the exterior of the animal through a medial pore on the labium. They produce a copious fluid secretion which appears largely to consist of an isotonic solution of potassium bicarbonate (Kafatos, 1968). The compositions of the secretion and that of the haemolymph bathing the tubules are shown in Table I. These tubules are of concern t o us here because they may well act as regulators of the haemolymph volume, which in a large flying insect, needs to be kept low (Edwards, 1964). It is clear that an important function they have in most saturniids is TABLE I The composition of the secretion produced by the labial glands of Antheraeu pernyi compared with the haemolymph composition (concentrations in mmol . 1-’ ). Figures from Kafatos (1 968)

a

AIP- 10

I on

Haemolymph

K+ Na’ Ca2+ Mg2+ HCO; c1Phosphate

38 3 8 50

Estimated.

5a

2oa 40

Secreted fluid 190 0.4

0.2 0.2 175 19 <0.03

210

S. H . P. MADDRELL

to provide a solvent of appropriate pH for the enzyme cocoonase secreted at ecdysis. This enzyme is used by these adult saturniid silkmoths to digest the sericin component of the cocoon prior to their emergence from it (Kafatos and Williams, 1964). The moth Hyalophora cecropia does not secrete cocoonase though it too has labial glands which are very active at the time of ecdysis. It is found that these glands are capable of secreting fluid for several days after ecdysis (Edwards, 1964). It may well be that here the glands ?re primarily excretory and serve to jettison surplus fluid. It is worth recalling here that adult Gzlliphora also excrete large volumes of fluid after emergence prior to their first flight (Cottrell, 1962) and that the blood volume of adult locusts declines dramatically after the moult from the last larval instar (Lee, 1961). Probably the labial glands (at least in Antheruea pernyi) do not regulate the composition of the haemolymph apart from lowering the potassium concentration, but they do decrease the blood volume by about 15%(Kafatos, 1968). Any regulation that is necessary is very likely achieved by the Malpighian tubulelrectum complex. The mechanism underlying the secretion of the fluid is of interest because of its rather unusual composition. Kafatos (1 968) has been able to get surprisingly large amounts of valuable information from in uivo experiments. Based on these results he proposes a model in which blood K + ions are exchanged for H + ions at the basal side of the cells. At the apical surface the reverse occurs, K + ions leaving the cells and H + ions entering them. The alkaline secretion than traps carbon dioxide by converting it t o HCO,. These points are emphasized in Fig. 5. Kafatos was unable to use in uitro preparations of the labial glands because injury to the moth caused the glands to stop secreting. If, as seems probable, the glands are controlled by a hormone released into the blood or perhaps less likely are made to secrete by direct nervous stimulation, dissection might well lead to a failure of such mechanisms. However, one can suggest that if tubules could be isolated either into haemolymph from ecdysing adults or into a Ringers solution containing cyclic AMP, the tubules might well continue to secrete. Cyclic AMP acts as an intracellular mediator in the control of many secretory tissues (Robison et al., 1968). The labial glands are rather larger than the other insect fluid secreting tubules so that research on in uitro preparations might be most rewarding. It will also be interesting to investigate moths which spin no

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

21 1

HaerocOel

Fig. 5 . Diagrammatic representation of the ultrastructure of the cells of the labial gland of Antheraea to show the postulated ion movements underlying fluid secretion (after Kafatos, 1968).

cocoons at all to see whether they use their labial glands at ecdysis. If they do, one could argue that the primary role of the labial glands is one of weight reduction and that their co-operative action with cocoonase is a secondary addition. Those organs which are important to the excretion of small molecules and which we have so far discussed, namely the utriculi majores of cockroaches, the midgut of the larvae and the labial glands of the adults of saturniid silkmoths, have in common that their excretory role is limited to one aspect. They are all concerned with a gross removal from the body of one substance-the utriculi majores remove uric acid, the midgut excretes potassium (mostly as chloride) and the labial glands excrete water (as isotonic potassium bicarbonate solution). In each case they are to be thought of as accessory excretory organs, for their activity alone would certainly not suffice to control the composition of the internal environment. The crucial control is still the responsibility of the Malpighian tubules and rectum. The activity of the accessory organs seems to be

212

S. H. P. MADDRELL

to relieve the Malpighian tubules and rectum of a part of their work in cases where a large amount of one substance needs to be excreted. We may note in passing that another example of this occurs in Arachnids where the cattle tick, Boophilus microplus, faced with the necessity for excreting a lot of fluid as it feeds, secretes it back into its host, through the activity of its salivary glands (Tatchell, 1967). Bloodsucking insects, however, are able to use the Malpighian tubules and rectum to solve the same problem (Wigglesworth, 193 la; Lester and Loyd, 1928). D. THE ANAL PAPILLAE OF MOSQUITO LARVAE

Some recent work has opened up the possibility that anal papillae of mosquito larvae which are known t o be able t o take up ions may also be able to excrete them. Phillips and Meredith (1969) have recently described their work on the anal papillae of the larvae of the salt water mosquito Aedes campestris and established that these organs can take up ions from hypo-osmotic media. In this work they also did an experiment which showed that larvae with damaged papillae took up chloride ions from a hyper-osmotic sodium chloride solution faster than did control insects. One explanation is that the ions may normally be transported from the insect into the surrounding medium. If this proves t o be the case, then anal papillae can be counted as accessory excretory organs. As Phillips and Meredith point out, the gills of euryhaline fish can similarly pump ions either into or out of the animal. IV. THE MALPIGHIAN TUBULES

We turn now to consider the recent work that has been done with the most important excretory organs of the insect, the Malpighian tubules and the hind-gut-in particular the rectum. As was indicated earlier the role of the Malpighian tubules is in general to supply the hindgut with a fluid containing a good many of the smaller sized constituents of the haemolymph more or less in proportion to the concentrations at which they occur in the haemolymph. It is the function of the hind-gut to reabsorb those constituents which the insect requires and t o reject the others so that the composition and volume of the haemolymph is kept relatively constant or adjusted to suit the needs of the insect. This simple picture needs to be modified somewhat to include the ability of Malpighlan tubules actively to

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

213

secrete other substances as is indicated by their secretion of dyes (see p. 280). As is argued on p. 284 it is very likely a function of Malpighian tubules to secrete a range of aromatic substances produced in metabolism. It is also possible, at least, that the hind-gut may add substances to the fluid passed to it from the Malpighian tubules. The rectum, for example, is certainly able to acidify the contents of the lumen, and it is conceivable that other additive changes may occur, either in the rectum or in the rest of the hind-gut. A further modificatioh may soon be needed for more attention has been paid to the activity of the rectum than that of the rest of the hind-gut. There are however, clear indications in the literature that the hind-gut anterior to the rectum may well be active in water absorption and possibly therefore may absorb other substances as well. This point is considered further on p. 287. A. THE MALPIGHIAN TUBULES OF CARA USIUS

We shall consider first the recent work on insect Malpighian tubules. As a prologue to this section it will be helpful first of all to summarize the early classical work of Ramsay who, in a series of papers on the Malpighian tubules of Gzrausius (then Dixippus) morosus, largely laid the foundations for more recent work on a wider variety of tubules. The roots of this work were established in an early paper (Ramsay, 1953). In this paper Ramsay observed that the fluid* collected from Malpighian tubules of both Rhodnius and larvae of Aedes contains considerably more potassium and less sodium than is found in the haemolymph. This might be expected in Rhodnius, where the diet is rich in potassium. It seemed pointless, however, in Aedes where the tubules continue to elaborate a fluid containing high concentrations of potassium even when the larva is kept in a solution of pure NaC1. This was the stimulus to look at the fluid secreted by Malpighian tubules of eight insects carefully chosen to cover as wide a range of orders, habitats and diets as possible. In each case it was found that the concentration of potassium in the fluid produced by the tubules was higher than in the haemolymph. From measurements of the trans-wall potential made at the same * The fluid produced by Malpighian tubules is often termed urine (see for example Ramsay, 1958; Berridge, 1968). The view taken in this paper is that the term urine is only properly applied to the fluid which is excreted after having passed through the hindgut. The fluid produced by the Malpighian tubules is homologous with the glomerular filtrate of vertebrates

.

214

S. H. P. MADDRELL

time, it was apparent that potassium movements into the lumen are thermodynamically uphill-that is potassium entry is active. With this as a background Ramsay then went on t o look in detail at the operation of the Malpighian tubules of the stick insect, Carausius. He first (Ramsay, 1954) developed a technique of isolating single tubules in a drop of serum under liquid paraffin (Fig. 6). It is the Hiemolymph Air bubble

Urdne

Fig. 6 . The experimental arrangement used by Ramsay to investigate secretion by isolated Malpighian tubules of Curuusius (from Ramsay, 1954).

exploitation of this technique which has allowed the rapid advance of our knowledge of the way in which Malpighian tubules work and has recently been extended to insect salivary glands (Berridge and Patel, 1968). The main conclusions of a series of papers which followed (Ramsay, 1954, 1955b, 1956) were: (1954) The fluid secreted by isolated tubules is nearly iso-osmotic with the haemolymph over a wide range of osmotic pressure. In the majority of cases, however, the fluid is significantly hypo-osmotic to the haemolymph. (1955b) The rate of fluid secretion is crucially dependent on the potassium concentration of the haemolymph, but it is scarcely affected by the concentration of sodium (Fig. 7). Sodium, like potassium, can be actively transported against an electrochemical gradient and does not appear to compete with potassium in the secretory mechanism. The rates of secretion of sodium and potassium both vary in direct proportion to the respective concentrations of these ions in the medium, but potassium is secreted more than ten times faster than sodium.

21 5

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

0

0

10

I

1

I

I

I

I

1

I

20

30

40

50

60

70

80

90

1 100

Concentration of ion in bathing fluid (rnrnol.I-I) Fig. 7. Rate of fluid secretion by isolated Malpi@an tubules of Carausius. as a function of the potassium concentration (open circles, continuous line), the sodium concentration being held constant at 16-17 mmol . I-' , and as a function of the sodium concentration (solid circles, dotted line), the potassium concentration being held constant at 15-16 mmol. I-' (data from Ramsay, 1955).

Calcium, magnesium and chloride ions are always at a lower concentration in the secreted fluid than in the bathing medium. The concentration of phosphate is always higher in the secretion than in the medium. As the external phosphate concentration is increased, so is the rate of secretion. The secreted fluid is always more alkaline than the serum. Based on these results Ramsay (1956) concluded that the active transport of potassium across the tubule wall was fundamental to secretion. Because the secreted fluid was consistently hypo-osmotic to the bathing medium he rejected the idea that the secretion of potassium (together with an anion) could establish an osmotic pressure which would in turn produce an inward osmotic influx of water. This conclusion left unclear why potassium movements were basic t o secretion. Ramsay had also to suppose that water movements were somehow adjusted to the movements of solutes so as always to produce a fluid more or less iso-osmotic with the haemo lymp h.

216

S . H. P. MADDRELL

These then, very briefly, are Ramsay’s results and conclusions on the movements of ions and water in the mechanism of secretion by the Malpighian tubules of Chrausius. We shall defer until p. 279 consideration of his very important paper (Ramsay, 1958) which deals with the movements of organic molecules and provides a rationale for the functioning of the Malpighian tubules and rectum together. We go on now t o consider the recent experiments on isolated tubules of Calliphora, Rhodnius, Calpodes and Chrausius in the light of Ramsay’s conclusions and use these results as a basis for a general statement of the mechanism of secretion by Malpighian tubules. A great drawback encountered by Ramsay in his work on isolated Malpighian tubules was that the tubules would not function well in an artificial medium-they needed at least a proportion of serum in the bathing medium. A most useful advance since then has been the discovery by Berridge (1966b) that Malpighian tubules of the blowfly, Calliphora, can be made to function for extended periods, up to three days, provided that the bathing solution contains an energy source. A wide range of compounds such as maltose, glucose, pyruvate and glutamine will support prolonged secretion. Without such an energy source secretion rapidly comes to a halt (Berridge, 1966b). Based on this discovery, it has now been shown that Chrausius tubules can also be made to secrete in an artificial medium containing added glucose (Pilcher, 1969, 1970a). Their ability to do this depends rather on the season of the year; in late spring and summer these tubules require little more than an energy source added to a basic solution of ions, but that tubules isolated from stick insects later in the year require some extra ingredient from the haemolymph. The situation with Rhodnius tubules is that in a Ringer’s solution containing only ions, secretion fails in 10-30 min (Maddrell, 1969). Secretion persists for 8-10 h in the presence of glucose. In haemolymph alone secretion also persists for about 8-10 h, but if glucose is added t o the haemolymph, secretion may go on for as long as 2 4 h . Clearly there is some as yet unidentified ingredient of haemolymph which allows secretion to persist longer than in its absence. This may be similar to the factor required for maintained secretion by tubules of Carausius (Ramsay, 1953). The requirement for this material is not as marked with Rhodnius tubules as it is with Carausius tubules. The Malpighian tubules of the skipper Calpodes ethlius (Lepidoptera) secrete for long periods in solutions containing maltose (Irvine, 1969, 1970).

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

B. THE MALPIGHIAN

-

217

TUBULES OF CALLIPHORA

Secretion by the Malpighian tubules of Calliphora (Berridge, 1968, 1969) has a number of similarities with the equivalent process in Curausius tubules. It is, for example, extremely sensitive to the external potassium concentration-particularly when the medium contains sodium ions. In the absence of potassium (i.e. when the major cationic component is sodium) fluid secretion occurs at a very

Concentmtim of potassium in the batting ~ d u t i(rnmol.l-') i

Fig. 8. Effect of the potassium concentration in the bathing solution on the rate of secretion by isolated Malpighian tubules of Culliphoru. The osmotic concentration of the bathing solution was kept constant with sodium (closed circles) or with sucrose (open circles) (from Berridge, 1968).

low rate. The addition of a small amount of potassium has a disproportionately large effect-replacement of 8 mmol .1-' of the sodium by potassium causes a ten-fold increase in the rate (Fig. 8). Subsequent increases in the potassium concentration cause a linear increase in the rate so that a t 1 4 0 m m o l . 1-' potassium and 0 mmol . I-' sodium the rate is at its maximum. The preference of the tubules for potassium over sodium is also clearly shown from measurements of the concentrations of these two ions in the secreted fluid (Fig. 9). In a potassium-free medium, sodium is secreted at a high concentration-but at a very slow rate. Replacement of a small quantity of sodium with an equivalent amount of potassium produces a large change in the potassium and sodium concentrations in the secreted fluid; potassium almost

218

S. H. P. MADDRELL

Fig. 9. The concentrations of sodium and potassium in the fluid secreted by isolated Malpighian tubules of Calliphora as a function of their concentrations in the bathing solutions ( o d , potassium concentration when sucrose replaced sodium in the bathing solution) (from Berridge, 1968).

completely replaces sodium. Sod& is in fact secreted no slower but the presence of potassium stimulates a very large increase in the rate of fluid secretion so that sodium appears in the secreted fluid at a much lower concentration. Potassium is always more concentrated in the secreted fluid than it is in the fluid bathing the tubules, particularly so at low values of the latter. Sodium does not compete with potassium-the substitution of sodium for a small amount of potassium in a previously potassium-rich sodium-free medium scarcely alters the rate of potassium secretion. Ramsay reached similar conclusions from his experiments with the tubules of Carausius. Berridge (1968) found that rubidium could be substituted for potassium with little slowing in rate but that caesium and sodium were less effective, the rates being 18% and 3% respectively of the rate in a potassium based solution. Ammonium, lithium and choline ions were completely ineffective. Although sodium does not compete with potassium for transport it has a marked stimulatory effect on the rate of potassium secretion and fluid secretion. Plotted in Figs 8 and 9 are values of

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

219

potassium concentrations and rates of fluid formation determined in experiments where the osmotic pressure of the bathing fluid was kept constant with sucrose instead of sodium. By comparing these values with those where sodium was present it is obvious that sodium markedly promotes secretion by tubules secreting in solutions containing low concentrations of potassium. Berridge (1969) investigated the ability of a large number of anions to support fluid secretion by Culliphoru tubules. The results are best summarized as a graph of the rate of fluid secretion plotted against hydrated radius (Fig. 10). Two points emerge. One is that for all anions except phosphate, the ability to support secretion appears to depend critically on its size. Anions larger than acetate will not support secretion at measurable rates. Smaller anions will permit secretion and the smaller they are the faster the flow. As is discussed on p. 245, permeabilities of monovalent ions are not often related purely to size as in this case, but are determined by the free energy decrease associated with the ion's move from water into the membrane. Only in cases where the membrane is relatively little

10

-

I OI

I

I

,SOLH:iCiH&P-

,%m*

,

2

3

6

7

4

nydroted mdiur

5

(A)

Fig. 10. The ability of anions with different hydrated radii to support fluid secretion by isolated Malpighian tubules of CaZliphora (from Berridge, 1969).

220

S . H. P. MADDRELL

charged will this result in ions penetrating in order of size. The second point to emerge is that phosphate ions do not fit this story. They behave as if there were a separate mechanism to handle them. Berridge suggests that there may be a carrier-mediated transport system for phosphate. This is consistent with the observation that phosphate may be concentrated in the secreted fluid (Fig. 11). It 2624

4

t

2 z

22

-

0

’ 0

2 0 -

0

.

0

18-

k

1 6 -

e

1 4 -

0

.

12

0

*

0

o 0

-

0

10

08

I

I

I

I

I

I

I

I

I’lio\pli.stc CunCcntr.it1on In mcdium ( p g . l’O,//fI.)

Fig. 1 1 . The relation between phosphate U/P and phosphate concentration of the bathing solution for isolated Malpighian tubules of CdiphOrQ. The anionic concentration was maintained either with sulphate (closed circles) or chloride (open circles) (from Berridge, 1969).

may also explain an increased rate of fluid secretion observed when phosphate is included in a chloride-based medium (Fig. 12). That there are two separate pathways for anion transport was neatly shown in Berridge’s experiments where one pathway was blocked with cupric ions and the other blocked with arsenate ions. When cupric ions were added to tubules secreting in solutions based on chloride or nitrate, secretion stopped; secretion by tubules bathed in a phosphate-containing solution was unaffected (Fig. 13). Arsenate ions by contrast inhibit phosphate transport but not chloride transport. Indeed when arsenate (1 x M/1) was added to tubules bathed in solutions containing phosphate and chloride ions, the chloride concentration in the secreted fluid was more than doubled. Because the rate of fluid secretion decreased sharply, this does not imply a faster rate of chloride secretion. Ramsay in his earlier work with Gzrausius tubules also found that phosphate ions were very effective in allowing secretion and that phosphate was concentrated by the tubules. Just how similar the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

221

t td' t

:$0 I

OoYo=

slo

Id0

o1;

Chloride concentrotion in the bathing solution (mm0l.l -'I

Fig. 12. Rate of fluid secretion by isolated Malpighian tubules of Culliphoru showing the effects of including phosphate ions (open circles) or sulphate ions (closed circles) in a chloride-based solution (from Berridge, 1969).

tubules of Culliphora and Curausius are in this respect is brought out in Figs 14 and 15. The osmotic pressure of the bathing medium very much affects the rate of secretion of Culliphora tubules. The rate is inversely proportional to the external osmotic pressure, whether changes in osmotic pressure are achieved by adding potassium chloride or sucrose (Fig. 16), though at the highest concentrations of sucrose the rate of secretion appears to be somewhat depressed by comparison with that in concentrated solutions based on potassium chloride. The osmotic pressure of the secreted fluid was always slightly hyper-osmotic to the bathing medium in a range of solutions with freezing point depressions of 0.3-2.1 "C (Fig. 17). The addition of sucrose to the bathing medium produces an increase in the concentration of potassium in the secreted fluid very close to that required exactly to balance the increased osmotic pressure of the bathing medium due t o the addition of sucrose (Fig. i 8).

222

S. H. P. MADDRELL

Tc

-c T

I1

-

10

-

98

-

7 -

c

.-

t

H

6-

0

5 -

-

r

B 4P 32I -

-, -\

0:

I

; 5

T i m (hr)

mol . 1-' ) on the rate of fluid secretion by Fig. 13. Effect of copper ions (1 x isolated Malpighian tubules of Calliphora. Twelve tubules were set up in solutions containing the anion to be studied. After 2 h copper ions were added to six tubules and the remainder served as controls. Phosphate control ( 0 ) ; phosphate with copper (0);nitrate control (A); nitrate with copper (A); chloride control (m); chloride with copper (0) (from Berridge, 1969).

As in Caruusius, secretion by isolated tubules of Gzlliphora is affected by changes in the external concentration of calcium and magnesium (Berridge, 1968). The results are shown in Fig. 19. The optimum concentration of either ion is about 10 mmol . I-' and higher concentrations depress secretion, particularly so in the case of calcium. pH affects secretion by Gzlliphora tubules scarcely at all (Fig. 20). This conclusion must be accepted with some reservation because pH was controlled by altering the concentrations of H2PO: - and from Fig. 10 it is possible that H,PO, supports a higher rate of secretion than does HPO: -. At acid pHs then one might expect an acceleration of secretion due to an increase in the concentration of H2PO;.That such an acceleration does not occur suggests that the secretory mechanism itself may be depressed by the low pH but this is

2 23

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS Chloride

Phosphote

I

80-

- 70 L

'c)

260-

Lo-

E

540L

2-

.c

K

20-

u 100 0 -

i 20

40

60

80

100

Concentration of ion in bathing

fluid

Fig. 14. Concentrations of chloride and phosphate in fluid secreted by isolated Malpighian tubules of Calliphora and Carausius as a function of the concentrations of these ions in the medium. In each case the concentrations of the ions in the medium are complementary, i.e. a low chloride concentration is balanced by a high phosphate concentration and vice versa (based on figures from Berridge, 1969 and Ramsay, 1956).

10-

= 9;1

F

5

t

P

654-

2 32g l-

Caroustus

0

8 - -~_ - _ _ _ - _ _ _ _

- 7-

2.5-

C a l f i ~

0

2.0 -

\o \

9,

, , 0,

\

\ O

1.5 '\ \

..

*.

..

, 0

h

o

b' \

1.0 -

-,,o

.

0.5 -

Fig. 15. Rate of fluid secretion by isolated Malpighian tubules of Calliphora and Camusius as a function of the concentration of chloride and phosphate in the medium. In each case, as the chloride concentration increases, the phosphate concentration decreases. For Calliphora, each point represents the average of six determinations; for Carausius, each point represents the average of four determinations (figures for Calliphora from Berridge, 1969; figures for Carausius from Ramsay, 1956).

224

S. H. P. MADDRELL

.5

c

c

osmotic concentrotion of the bathing solution (AOC)

Fig. 16. The rate of secretion of isolated Malpighian tubules of Culliphoru as a function of the osmotic concentration of the medium. Higher osmotic concentrations were produced either by adding potassium chloride (open circles) or sucrose (solid circles) (based on data of Berridge, 1968).

0

2

%-

0

,Sh e-

c

t 0

Osmotic concentration of bathing solution (AOC)

Fig. 17. The osmotic concentration of the fluid secreted by isolated Malpighian tubules of Culliphorn as a function of the osmotic concentration of the bathing solution. The line is that of an iso-osmotic relationship (from Berridge, 1968).

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

225

310 290 -

250 270

50

I00

150 200

250

300

Sucrose concentrotion in the bathing solution (mmol. I - ' )

Fig. 18. The increase in the potassium concentration of fluid secreted by isolated Malpighian tubules of Calliphora as sucrose is progressively added to the bathing solution. 0 , values expected if extra potassium chloride in the secreted fluid is osmotically to balance the increased osmotic concentration of the bathing solution; 0 , observed values (from Berridge, 1968).

Concentration in bathing solution (mmolkl-')

Fig. 19. Effects of magnesium (closed circles) and calcium (open circles) on the rate of fluid secretion by isolated Malpighian tubules of Calliphom (from Berridge, 1968).

226

S.

01 5.0

1

60

H. P. MADDRELL

I

7.0

I

I

80

90

pH of bathing solution

Fig. 20. Effect of pH on the rate of fluid secretion by isolated Malpighian tubules of Culliphoru (from Berridge, 1968).

counteracted by an increase in the concentration of H,PO,. Such an argument does not affect the fact that the tubules will secrete fast at pH 9.2. By contrast the tubules of Gzruusius secrete poorly outside the range of pHs from 5.2 to 7.5. Finally it should be added that Culliphoru tubules are not affected by ouabain, acetazolamide (Diamox) and sulphanilamide. It is probable therefore that neither a sodium/potassium exchange pump of the kind found in many tissues (Glynn, 1964) nor carbonic anhydrase is essentially involved in the transport mechanism. What emerges from this consideration of Berridge’s work on Culliphoru tubules is that they behave in a very similar fashion to those of Curuusius. This encourages one t o attempt to use the information from both sources t o produce a working hypothesis as to how they might operate. But before doing this it will be appropriate to consider some additional information on Cizruusius tubules which comes from Pilcher’s recent work (Pilcher, 1969, 1970b). This work is largely concerned with the hormonal control of these tubules but does provide a good deal of valuable information about the way in which the tubules work. Her work is the first to use Curuusius tubules isolated into artificial media. From experiments using solutions containing only one monovalent cation, Pilcher found that sodium would support a slow rate of secretion about 20% of that observed in solutions containing

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

227

potassium only. Ammonium ions would scarcely support any secretion. This was not due to any toxic effect because tubules in a potassium and ammonium solution secreted at a rate comparable to that observed in control normal saline. At relatively low concentrations of potassium ions the presence or absence of sodium ions had a pronounced effect on the rate of secretion (Fig. 21). That this effect was not obvious from Ramsay's

20

40

60

80

Concentration of potassium in t h bathing mdium (mmol.l-')

Fig. 21. The dependence of the rate of fluid secretion by isolated Malpighian tubules of Curausius on the potassium concentration in the medium in the presence of sodium (open circles) or in its absence (closed circles) (from Pilcher 1970b).

earlier experiments (Ramsay, 1955b) is because Ramsay had to work with solutions containing at least some serum and therefore a minimum of 3 mmol . 1-' sodium. Pilcher has now shown that the addition of even 2 mmol. 1-I of sodium is enough to cause a drama tic increase in the rate of secretion (Fig. 2 1). Further increases have a relatively much smaller effect. Using solutions containing only one of a range of anions, Pilcher (1969) was able to show that anions larger than chloride supported only very reduced rates of secretion, but that phosphate, as in Cizlliphoru, supported a higher rate than does chloride. Her results are plotted in Fig. 22 together with those from Gzlliphoru for comparison. However she found that cupric ions not only slowed secretion by tubules secreting in chloride containing solutions as in Culliphoru, but also slowed secretion by those bathed in phosphate-containing Ringer's solution. This latter finding is exactly

S. H. P. MADDRELL 150

,

125

ONO,

I I I I Br-0 \ CI-O\

100

\

\ OCIO, 1OCIOi

'p-

75

\ \

\ \ 50

\

\

25 \

\

CH3COO-of/ 0

I

I

I

2

I 3

SG-

o&o C H 3 C H @ O 4 5

Radius of hydrated anion

I

6

(A)

Fig. 22. The ability of anions with different hydrated radii to support fluid secretion by isolated Malpighian tubules of Calliphoru (open circles) and Caruusius (solid circles) (data for Culliphom from Berridge, 1969 and for Cumusius from Pilcher, 1969).

opposite to that found in Calliphoru and is one of the few points of difference between the behaviour of the two tubules. As in Calliphoru, ouabain has no effect on secretion by Caruusius tubules. By cannulating the tubules and perfusing them with a fluid exactly similar to that of the bathing medium, Pilcher has shown from the transmural potential difference, that the lumen is usually positive with respect to the bathing medium, especially when the tubules are stimulated by their diuretic hormone. Ramsay also found a transmural potential difference of the order of 20 mV, lumen positive. In this case the tubule was not cannulated and so the tubule

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

229

wall had fluids of different composition on its two sides. Berridge (unpublished results) has found that the lumen of uncannulated tubules of Chlliphora is typically positive with respect to the bathing medium. Again using cannulated tubules, Pilcher showed that the trans-wall potential was very sensitive t o changes in the external potassium concentration (the sodium concentration being altered in the opposite sense so that the sum of the sodium and potassium concentrations was constant). A plot of log [Ko]/[Ki] against potential gave lines with a slope close t o 58 mV for a ten-fold change in [KO]. While not the only interpretation of these results, this probably indicates that the basal cell membrane is much more permeable to potassium than to sodium. If this is so it may help to explain why the tubules transport potassium so much faster than sodium. It could well be that access to pumps on the luminal side of the cell is controlled by the permeability characteristics and the pumps on the basal cell wall. If the basal cell wall is permeable to potassium then this would allow potassium in the bathing medium easy access to the pumps on the luminal side but deny it to sodium. When Chrausius tubules are bathed in a solution whose only anions are bicarbonate and sulphate, the trans-wall potential changes so that the lumen is a good deal more positive with respect to the bathing solution. The tubules are relatively impermeable to these ions and the increase in potential can be likened to the similar increase in potential seen in isolated preparations of frog skin bathed in sulphate solutions (Ussing, 1960). With these new findings for the tubules of Carausius we are now in a position to suggest a model of how the tubules of Gzrausius and Calliphora might operate. Any such model which attempts to account for the mode of operation of these tubules must incorporate the following points: (1) The concentration of ions in the secreted fluid is different from that in the haemolymph-in particular the potassium and phosphate levels are elevated and the sodium, calcium, magnesium and chloride concentrations are depressed. (2) Fluid is secreted at a rate which is dependent on the potassium concentration of the bathing fluid and is faster still in the presence of small concentrations of sodium ions (as low as 2 mmol . I-' ) especially at lower potassium concentrations. (3) The fluid produced by the tubules is nearly iso-osmotic with the bathing fluid over a wide range of concentrations of the latter. Fluid is secreted at a rate inversely proportional t o the osmotic pressure of the bathing medium. It follows that solute transport is

230

S. H. P. MADDRELL

not affected by changes in osmotic pressure of the bathing medium. (4) The lumen is usually positive with respect to the bathing solution and becomes more so when secretion is accelerated by the diuretic hormone or when secretion is slowed in a sulphatel bicarbonate solution. (5) A barrier accessible from the basal side is much more permeable to K + than to Na+. (6) Fluid secretion is insensitive t o ouabain even when stimulation by sodium (see (2) above) is maximal. (7) Potassium ions move against their electrochemical gradient and under some circumstances so do sodium ions. (8) Sodium and potassium each seem not t o detract from the transport of the other. (9) For prolonged secretion, Malpighian tubules require an energy source in the medium. (10) Anions (except phosphate) support secretion at a rate which is related to their hydrated size; the smaller the anion, the faster the rate of secretion supported. Phosphate ions support a higher rate of secretion than does any other anion. Taking these points into consideration a possible model for the operation of these tubules is shown in Fig. 23. This model while it is consistent with the known results is of course not the only model which could have been put forward. It must strongly be stressed that the value o f such a model lies in the experiments it suggests; it certainly must not be taken as an unambiguous conclusion derived from the results. To emphasize this point, it is important to stress the unsupported, though it is hoped, reasonable, assumptions which underlie the model. It is assumed that transport of ions and water across the layer of cells involves passing through the cell cytoplasm after crossing the basal cell membrane and before leaving through the apical cell membrane. This may well be what happens but, for example, a recent model for sodium transport across frog skin has sodium moving only in the plane of cell membranes and not penetrating into the cell interior (Cereijido and Rotunno, 1968). The model also assumes that the major barriers to transport are the cell membranes. Based on the work of neurophysiologists (see Katz, 1966) it is clear that for nerve axons the cell membrane is the major permeability barrier. The work of Loewenstein et al. (1965) showed that the external cell membranes of Malpighian tubule cells of Chironomus had a very high electrical resistance as compared with the apposed lateral cell membranes and the cytoplasm. If transport

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS Bathing solution

Cell

c1--+ cu2'

t

23 1

Lumen

---b

cI-

-

Fig. 23. Hypothetical scheme for operation of Malpighian tubules of Curuusius and Culliphora. Active movements are indicated by unbroken lines and passive movements by broken lines.

goes through the cells then it is clear that the apical and basal cell membranes will be the major permeability barriers. The main features of the model are that active ion transport occurs both at the basal and apical cell membranes and that water movements are osmotically linked to ion movements. The necessity for the first of these is that for example potassium transport can occur from bathing solutions of low potassium concentrations. It is difficult to envisage net passive movement of potassium into the cell from an external concentration of, say, 5 mmol .1-' when it is known that most cells have a intracellular concentration of potassium of between 50 and 200 mmol . 1-' (see for example Robertson, 1957) even taking into account that it is likely that the interior of the cell is at a potential which is negative with respect to the external medium. It seems very likely then that potassium is actively transported into the cell by a pump situated on the basal side of the cell. This arrangement has the drawback that it is probable that the basal side of the cell is permeable to potassium

232

S. H . P. MADDRELL

(point (5) above) so that transported potassium might tend to leak back-especially if the pump were electrogenic as this would tend to produce both a concentration and an electrical gradient favouring such back diffusion. A better arrangement would be an electrically neutral pump such as a linked potassium and chloride pump-in this case if the membrane were not highly permeable to chloride extensive back diffusion of potassium would be prevented by the relative inability of chloride to follow. The problem of back diffusion at both the basal and apical cell membranes is much reduced if, as argued below, water movements are osmotically coupled t o net ion movements. As a result high concentrations of transported ions d o not build up as readily and the effectiveness of back diffusion is reduced by the rapid movements of water proceeding in the same direction as ion movements. To account for the stimulation of transport by sodium ions at low potassium concentrations the model proposes that it is the entry of potassium into the cell which is sodium stimulated. This is more likely than the alternative which is that sodium ions directly stimulate the rate at which potassium ions leave the cell at the apical membrane. The stimulant action is pronounced even at low concentrations of sodium and in view of the probably low permeability of the basal cell membrane t o sodium ions it seems unlikely that a low external concentration could have much effect on events occurring on the other side of an unfavourable permeability barrier. The sodium stimulation of a basal potassium transporting system might be brought about by sodium diffusing into the cell and being pumped out in exchange for potassium. This is the arrangement suggested by Berridge (1967, 1968). It has the advantage of being electrically neutral so that the tendency of potassium ions t o move back out of the cell is reduced. The objections to this arrangement are the same as those against an action of sodium on the apical side of the cell. They are that, at least in Caruusius, a pronounced sodium stimulation is produced by as little as 2 mmol . 1-' of sodium in the external solution and that the basal cell membrane is probably rather impermeable to sodium. Under these conditions, it is difficult t o see that sodium could enter the cell fast enough t o support a high rate of sodium/potassium exchange. Against this objection it should be pointed out that in a solution containing low amounts of sodium a slightly higher concentration might accumulate in the basal infoldings (see Fig. 49) swept in there by the steady influx of fluid on the one hand and pumped there

THE MECHANISMS O F INSECT EXCRETORY SYSTEMS

233

from the cells on the other. Nonetheless it may be said of this type of pump that it does not produce a net movement of ions into the cell and so does nothing t o produce an osmotic gradient down which water movements could occur. It must also be added that secretion is unaffected by ouabain (Berridge, 1 968; Pilcher, 1969, 1970b), although this drug inhibits all the known sodium/potassium exchange pumps. In view of these considerations it is suggested that it is more likely that sodium stimulates the potassium-transporting system from the outside of the cell. In the presence of potassium as the only monovalent cation, secretion is very slow at low potassium concentrations. This would stem from the reduced potassium transport across the basal cell wall in the absence of sodium. This would starve the apical potassium pump because potassium ions could reach it only slowly by passive entry into the cell. As a further consequence fluid would enter the cell only slowly because the development of an osmotic gradient across the basal side of the cell would largely be established indirectly by the action of the apical pumps and not directly by basal ion transport. A further factor to be taken into account is that secretion is slow in solutions of low chloride concentration (Fig. 24). This would be expected in those cases where chloride movements are passive but not in cases where chloride movements are active. This then is an objection to entry of potassium into the cell by involvement in a

-

'0

20

40

60

00

100

120

Concentration of chloride ions in the bathing solution (rnmobl-')

Fig. 24. Fluid secretion by Malpighian tubules of Calliphora as a function of the chloride concentration in phosphate-free solutions (redrawn from Berridge, 1969).

234

S. H. P. MADDRELL

linked potassium/chloride pump. Clearly each of the possible ways in which potassium is transported into the cells faces objections and it is not prudent at the moment t o decide between them as to which is the most likely. To summarize the position on this point the possible modes of entry of potassium are listed in Table I1 together with the objections against and the points in favour of each. Ion transport into the lumen is satisfactorily accounted for by an electrogenic potassium pump on the apical cell membrane. This would provide an electrical gradient to drag anions in behind potassium ions and together these ions would provide the osmotic concentration which would force water to follow. It is envisaged that sodium ions also gain access to the lumen via an electrogenic pump for this ion, or possibly they follow phosphate ions transported into the lumen. Phosphate ions cross the tubule wall, according to the model, by involvement with a carrier at both apical and plasma membranes. There is some evidence that the carrier at the apical membrane may be the alkaline phosphatase which has been localized there (Berridge, 1967). If this step is an active one, that is it acts as a pump, then this may provide an electrical gradient which could promote the entry of sodium into the lumen. In this case there would be no need to involve an electrogenic sodium pump on the apical membrane. As tests of this idea one can suggest measuring the trans-wall potential in a potassium-free solution containing sodium and phosphate, where one would expect the lumen to become negatively charged if the sodium ions follow active movements of phosphate ions. It would also be worth while discovering if secretion is possible in a sodium-based solution containing no potassium or phosphate. In this case secretion would be expected t o fail. To account for the fact that in Curuusius cupric ions slow phosphate transport, it may be that these ions interfere with the basal side carrier or that phosphate, though it may be pumped into the lumen from the cell, enters the basal side passively and so is vulnerable there t o the action of copper ions (which are thought selectively t o block the passive movements of anions through membranes, Berridge, 1969). Neither of these suggestions is particularly satisfactory. The other feature of the model which needs discussion here is that water movements are osmotically linked to ion movements. Berridge (1 967, 1968) and Berridge and Oschman (1 969) have fully discussed the possible mechanisms whereby this might be brought about.

TABLE I1 Ways in which potassium ions may gain access to Malpighian tubule cells from the haemolymph and the points in favour of and objections to these suggestions Potassium transport system at basal side of the cell

Objections

Points in favour (1) Net transport of ions provides a direct driving force for water entry into the cells. (2) Chloride ions enter the cell passively so that, as observed, low external concentration of chloride causes a reduction in the rate of fluid transport.

(a) Sodium-stimulated electrogenic potassium Pump

( 1 ) Inefficient because it creates an

(b) Sodium/potassium exchange pump

(1) Requires entry into the cell from low external concentrations across a membrane thought to be relatively impermeable to this ion. ( 2 ) N o net transport of solute produced. (3) Requires that this Na/K pump is unusual in being insensitive to ouabain.

(1) Electrically neutral so that back

(c) Sodium-stimulated linked potassium chloride pump.

( 1 ) Lowered external chloride concentrations would not be expected t o slow secretion, but this is what is observed.

(1) Electrically neutral so that back diffusion of potassium is reduced. (2) Net transport of ions provides a direct driving force for water entry into the cells.

electrical gradient encouraging back diffusion of potassium across a membrane thought to be permeable to this ion. (2) Cell interior would be at a positive potential with respect to the bathing medium-not a usual finding.

diffusion of potassium is reduced. (2) Chloride ions enter the cell passively so that, as observed, low external concentrations of chloride cause a reduction in the rate of fluid transport. (3) Sodium stimulation is an integral part of the system.

236

S. H. P. MADDRELL

Basically the system of long infoldings of the basal cell wall and the elongated microvillar projections of the apical cell wall (see Fig. 49) may be suited to a coupling of solute and water transport in iso-osmotic proportions by means of standing osmotic gradients (Diamond and Tormey, 1966; Diamond and Bossert, 1967, 1968). Active movements of ions across both basal and apical cell membranes are a feature of the model and they provide a force which could drag water across both surfaces of the cell. The difficulty with this aspect of the model is that it does not explain Ramsay’s results with Gzrausius tubules that the secreted fluid is hypo-osmotic to the bathing fluid. One can suggest, as indeed Ramsay did (1954), that the explanation might be that an iso-osmotic fluid may be produced by most of the cells of the tubules but that solute resorption in parallel by a few cells could easily reduce the osmotic pressure so that the fluid produced would be slightly hypo-osmotic. At the time of Ramsay’s experiments he pointed out “that there is no direct or circumstantial evidence in favour of the argument”. There is still no trulv direct evidence in favour of the explanation advanced above but several pieces of circumstantial evidence do make it more likely. They are: (1) The epithelial cells of the Malpighian tubules of Gzlliphora turn out to be a heterogeneous population-there is a small number of stellate cells among a much larger number of more typical cells (Berridge and Oschman, 1969). Were there such a heterogeneous population in the tubules of Carausius this would provide at least a structural basis for a secreting system with a small resorbing element in parallel with it. It is known for example, that the tubules of Gzrausius are regionally different in appearance (Ramsay, 1955a). (2) The upper or distal portions of the Malpighian tubules of Rhodnius consist only of one type of cell, and bathed in solutions of a wide range of osmotic concentrations, they consistently produce fluid which is very slightly hyper-osmotic (Maddrell, 1969; p. 247). (3) The tubules of Gzlliphora (Berridge, 1968) produce fluid which is slightly hyper-osmotic and the tubules of Dysdercus fluid which is very close to being iso-osmotic with the bathing fluid (Berridge, 1966a). A recent preliminary study using isolated tubules of the larvae of Tipula paludosa shows that here too the fluid secreted is iso-osmotic (Coast, 1969). Clearly it is characteristic of many Malpighian tubules that they produce fluid which is iso-osmotic with the bathing medium. (4) Fluid secretion by the Malpighian tubules of Calliphora

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

237

(Berridge, 1968) and Rhodnius (Maddrell, 1969) occurs at a rate which is inversely proportional to the osmotic pressure of the bathing medium. It follows that the rate of solute transport is unaffected by osmotic pressure, but that fluid movements are obligatorily circumscribed by it. (5) The addition of sucrose to the medium bathing the tubules of Culliphoru results in the secretion of an osmotically exactly compensating amount of potassium. Again it is not the rate of secretion of potassium which alters, but the rate of water movement. (6) The osmotic pressure of the fluid produced by isolated tubules of Dysdercus is unchanged when the rate of fluid secretion increases more than six times under the influence of this insect’s diuretic hormone (Berridge, 1966a). The standing osmotic gradient model of Diamond and Bossert (1967) predicts just such an independence. One’s overriding impression from (2), (3), (4), (5) and (6) is that the rate of fluid movement in the Malpighian tubules of several species of insects is determined by the rate of solute secretion and the osmotic pressure of the bathing medium over a wide range of conditions. The alternative is to suppose that water transport is adjusted to the movements of solutes and to the osmotic pressure of the bathing medium by some means other than an osmotic coupling. What form such an adjustment might take is difficult to envisage. Given the possibility of a small number of cells behaving differently from the majority (point (1) above) it is easy to see how small deviations in osmotic concentration might be produced in a fluid whose osmotic concentration would otherwise be iso-osmotic. The great advantage of the suggestion that water movements follow solute secretion is that it accounts very neatly for the otherwise meaningless secretion of potassium ions (and accompanying anions). In view of the new circumstantial evidence adduced in the preceding account, it now seems reasonable to accept that solute secretion is the driving force for movements of fluid. Why it should largely be potassium rather than sodium that is secreted by Malpighian tubules remains, as it was (Ramsay, 1955b), a matter for speculation. However, as we shall shortly see, the Malpighian tubules of Rhodnius secrete sodium very nearly as fast as they secrete potassium and they will secrete fluid at an unchanged rate in a medium containing no potassium at all. Parts of the tubules of Chlpodes may also secrete sodium at high rates (Irvine, 1969). Such an arrangement seems uncharacteristic but it should be pointed

238

S. H. P. MADDRELL

out that very few Malpighian tubules have been investigated in detail so far. Sodium secretion may turn out to be a characteristic of wider occurrence than now seems the case. C. THE MALPIGHIAN TUBULES OF TIPULA

It is worth including here some preliminary results from work on isolated tubules of larvae of Tipula paludosa (Coast, 1969). The tubules isolated into serum secreted fluid which was iso-osmotic with the serum (over a three-fold range in concentration of the latter). The concentration of potassium in the'tubule fluid was invariably higher and that of sodium always lower than that in the serum. Chloride concentrations were similar in both fluids. The lumen of the tubule was positive (on average 32 mV) with respect to the bathing medium. The rate of secretion of fluid was proportional, presumably inversely so, to the osmotic concentration of the serum when the potassium concentration was held constant. With the osmotic concentration of the serum constant, the rate of fluid secretion was related to the concentration of potassium. Without further details it is not possible to say whether the model set up for tubules of Gzrausius and Calliphora will in detail fit the operation of the tubules of Tipula, but it is certainly compatible with the results. D. THE MALPIGHIAN TUBULES OF RHODNIUS

Rhodnius' Malpighian tubules are of particular interest because they are specialized for very fast secretion of fluid. The insect spends long periods of starving between blood meals when its excretory rate is very low because of the need to conserve water. After a blood meal which is usually ten times as large as the insect itself it excretes during the next 2 h or so a weight of fluid equal to its previous body weight every half an hour. This excretion is largely a result of the very great acceleration in operation of the Malpighian tubules brought about by the release into the haemolymph of a very potent diuretic hormone (Maddrell, 1963). The secretory rate of the tubules increases promptly by a factor of about 1000 times to a rate as high as 3.3 pl . min-' . cm-*. The Malpighian tubules of Rhodnius are divided into two structurally and functionally distinct regions (Wigglesworth, 1931a, b, c; Ramsay, 1952). The upper or distal

239

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

length secretes fluid into the lumen, while the lower or proximal region is thought to modify fluid passing through it, possibly by resorption. It is known that isolated preparations of the lower lengths do not secrete fluid (Maddrell, 1969). The work described was therefore all done on the upper lengths. In view of the very high rate of secretion by the Malpighian tubules it is perhaps not surprising that the basis of secretion differs

"E 10

10 -

o!

210

40

60

eo

Id0

o ;l

Id0

$0

Concentration of potassium ions in the bathing solution (mnd,l-')

Fig. 25. Fluid secretion by the Malpighian tubules of Rhodnius, Carausius and Calliphora as a function of the potassium concentration of the bathing solution, the cationic balance being maintained with sodium ions.

in many respects from those of other tubules so far examined. The first of these differences is that fast secretion does not require potassium ions in the bathing medium. Not only can tubules secrete in potassium-free solutions but they secrete just as fast as do tubules bathed in standard Ringer's solution or in haemolymph. This is quite different from the situation in Curuusius and Culliphora as is brought out in Fig. 25. In potassium-based solutions containing no sodium ions, secretion is only about 40% as fast as in standard Ringer's solution. This effect is similar to that observed in Curuusius and Culliphoru except that it is more marked. Rhodnius' tubules are greatly affected by lack of sodium ions even in the presence of 150 mmol . I-' of potassium ions. In Curuusius the absence of sodium ions has an effect only when the external potassium concentration falls below about 30 mmol .1-' (F'ilcher, 1969, 1970b). In Culliphoru the absence of

240

S. H. P. MADDRELL

sodium ions has an effect at all potassium concentrations in the bathing medium lower than 140 mmol . 1-' . These effects are illustrated in Fig. 26. The addition of as little as 1-1.5 mmol . 1-' of sodium ions to Rhodnius tubules secreting in a potassium-rich solution causes an abrupt acceleration of secretion to near the normal level-though the effect soon dies away because sodium is rapidly depleted from the bathing solution (Maddrell, 1969). By

c

E

K

E a

-I 150 Concentration of potassium

n)m

in the bathing solution (rnmol I-')

Fig. 26. Fluid secretion by the Malpighian tubules of Camusius. Calliphora and Rhodnius as a function of the potassium concentration of the bathing solution when no sodium ions are present.

analogy with the situation in Gzrausius and Gdliphora where a similar acceleration is produced, it may well be the rate of potassium access which is sodium dependent. The ability of Malpighian tubules of Rhodnius to secrete in a potassium-free medium is almost certainly due to their ability to transport sodium ions nearly as fast as they can pump potassium ions. This ability is one of the points revealed by analyses of fluid secreted from solutions whose sodium and potassium concentrations were altered concomitantly so that the sum of their concentratio'ns was kept constant at 150 mmol .1-'. The results of these analyses are shown in Fig. 27. Several points emerge. The first is that the sum of the concentrations of sodium and potassium in the secreted fluid is constant over the whole range of concentrations. This can be explained by the facts that the osmolarity of the fluid produced is always very near to that of the bathing fluid and that sodium and

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

0

24 1

10 20 30 40 50 60 70 80 90 100110120130140150

W+l

150140130120110100 90 80 70 60 50 40 30 20 10 0 "a+] Wentrntion in the taiting fluid (mmol.l-')

Fig. 27. The concentrations of sodium and potassium ions in the fluid secreted by isolated Malpighian tubules of Rhodnius as a function of their concentrations in the bathing solution (from Maddrell, 1969).

potassium are virtually the only cations secreted from these solutions. Secondly, the tubules' preferential handling of potassium is clear-potassium appears at a higher concentration in the secreted fluid than does sodium at all concentrations of potassium in the bathing fluid higher than about 15 mmol . 1-' ; at equal concentrations of sodium and potassium in the bathing fluid, potassium is about 30% more concentrated in the secreted fluid than is sodium; AIP- 1 1

242

S. H.

P. MADDRELL

and finally potassium is more concentrated in the secreted fluid than in the bathing fluid over the range 0-1 10 mmol . 1-' , while this is true for sodium over the smaller range 0-85 mmol . 1-' . Nevertheless, it is plain that sodium ions move through the tubule wall relatively little slower than do potassium ions, and this is a major difference from the behaviour of tubules of Calliphora and Carausius. This point is apparent from Fig. 28 which shows the similar figures for

120 100.

80

5 e

c

f

0

Na'

*-

60

4020

Calliphara

"

.'

40

.' ,

I

,

,

I

#

-

I

I

I

I

.

1

-

and Fig. 28. A comparison of the ways in which isolated Malpighian tubules of Rhodnius and Calliphora handle sodium and potassium ions (from Maddrell, 1969).

Calliphora. For example, in Rhodnius, when sodium is present at low concentrations in the bathing fluid it appears at a substantial concentration in the secreted fluid, and this is far from being the case in Calliphora. In Rhodnius the rate of fluid secretion is constant over the range K = 0, Na = 150 to K = 149, Na = 1 mmol . 1-' . It follows that at low concentrations of potassium (below about 10 mmol .1-' ) the rate of secretion of sodium increases rapidly as the potassium concentration is reduced to nothing. In a similar fashion, the rate of secretion of potassium rapidly increases as the sodium concentration decreases below 10 mmol . 1-' (though as described above, below about 1 mmol. 1-' of sodium, the rate of secretion of potassium drops to less than half its previous value). Figure 29 illustrates these points. These facts are explicable if one supposes there to be separate

THE MECHANISMS O F INSECT EXCRETORY SYSTEMS

0

2 43

10 20 30 40 50 60 70 80 90 100110120130140150 Concentration in bathing solution (cationic balance made up in each ccse by the other ion) (mmol.l-'l

Fig. 29. Rates of secretion of sodium and potassium ions by isolated Malpighian tubules of Rhodnius as a function of their concentrations in the bathing solution (from MaddreU, 1969).

transporting systems for sodium and potassium. At low concentrations of potassium either the potassium-transporting system moves sodium ions or the sodium system is very much stimulated. In an analogous fashion at low concentrations of sodium, the potassium secretion system is stimulated or the sodium-transporting system moves potassium. A further series of experiments shows that in the presence of a single monovalent cation the rate of secretion depends on the concentration of the cation, as in Fig. 30. Of course, as we have seen already, in sodium-free potassium solutions the rate of secretion does not reach a high rate even at a high potassium

244

S. H. P. MADDRELL

concentration. With sodium as the only cation, secretion reaches the normal maximum rate at about 125 mmol . 1-' ; below this concentration, the rate is concentration-dependent. The addition of small amounts of one cation to tubules secreting slowly in a low concentration of the other cation causes a disproportionately large acceleration of secretion. Analysis shows that the secretion of both cations is increased.

80-

70 -

6050 40 -

/*

30 -

Concentration of icm in the bathing solution (rnmol.l-')

Fig. 30. Rate of fluid secretion by the Malpighian tubules of Rhudnius as a function of the concentration in the bathing solution either of sodium or of potassium. In each case the cationic balance was made up with choline ions.

It is worth pointing out at this stage that Rhodnius tubules unlike those of Carausius qnd Calliphora will secrete ammonium ions as if they were unable tb distinguish them from potassium ions, at least for periods up to 1 h. Hydrated ammonium ions are closely similar in size to potassium ions (Using, 1960) though parameters other than charge and size may well be important in determining the rate of penetration of cell membranes by ammonium ions (Diamond and Wright, 1969).

Anions Again in contrast to the situation in Carausius and Gzlliphora, the Malpighian tubules of Rhodnius seem to secrete at high rates only in solutions containing chloride ions as the main anionic component. Phosphate ions support a rate of secretion only about 1 % as fast as

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

245

do chloride ions, though in Carausius and Gzlliphora secretion is considerably faster in phosphate based solutions than those based on chloride. It certainly seems that in Rhodnius no special transport system exists for phosphate ions. Apart from this difference, Rhodnius tubules also perform less well than do those of Carausius and Calliphora in solutions of anions other than chloride. The rates of secretion are plotted in Fig. 31 together with those for the two

c

e

I2,t

Radius of hydrated anion

(A)

Fig. 31. Dependence of the rates of fluid secretion by isolated Malpighian tubules of Rhodnius, Calliphora and Carausius on the hydrated size of the anion present in the bathing solution.

other insects for comparison. It is however possible from these results to suppose that the dependence of the tubules on the presence of chloride ions reflects the greater permeability of the tubule walls to hydrated chloride ions than the other anions, and in this case this is associated with the smaller size of the hydrated chloride ion. It is important to realize however that passive ion permeabilities cannot often be predicted from the hydrated ion size. As Diamond and Wright (1969) have recently pointed out in a signal paper it is the size of the decrease in the free energy of an ion as it

246

S. H. P. MADDRELL

becomes associated with a membrane which determines the permeability of the membrane t o the ion. Only in cases where the membrane has a very low electric field strength (due perhaps to very few fixed charges) will ions permeate at a rate determined by their hydrated size. The poor performance by Rhodnius tubules secreting in solutions based on larger anions than chloride suggest that this may be the case in Rhodnius and because the tubules of other insects do better in such cases suggests that the Malpighian tubules of Rhodnius are less permeable t o anions than are the other tubules. If that were the case one would expect secretion to be slowed even more in low concentrations of chloride than is the case in Culliphoru where the cell membranes behave as if they were quite permeable to anions larger than chloride. In fact, secretion by Rhodnius tubules is only slowed when the external chloride concentration drops below 3 0 mmol . 1-' . The behaviour of the tubules of Rhodnius and Culliphoru in this respect are compared in Fig. 32. So although the tubules of Rhodnius secrete considerably faster than do those of Culliphoru and behave as if they are less permeable to the larger anions, secretion is considerably less effected by lowered concentrations of chloride. This represents a prima facie case for a chloride transport system being involved. This case is greatly strengthened from measurements of the trans-wall potential made from cannulated tubules perfused with a solution identical with the bathing medium. At 24"C, for example, the

;520

<'B '

L=L--

OO

25

50

75

100

125

150

Concentration of chlwide ms m the batting sdution (mmol.l-')

Fig. 32. Fluid secretion by isolated Malpighian tubules of Rhodnius and Cdiphoru as a function of the concentration of chloride ions in the bathing solution.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

247

potential is on average 30 mV with the lumen negative with respect to the bathing medium (see p. 254 and Maddrell et al., 1969). It is worth pointing out that secretion in solutions containing both sodium and potassium is slowed when the sum of the concentrations of sodium and potassium is lower than about 60 mmol . 1-' , whereas as mentioned above secretion slows in low chloride solutions only when the concentration of chloride ions is less than 30 mmol . I-' . This suggests, at least, a more powerful chloride transporting system than cation transporting systems.

Water Movements Rhodnius' tubules produce a fluid which is marginally but consistently hyper-osmotic to the bathing fluid (by about 0.01 00.015"C (Fig. 33) which is equivalent to 2.7-4 mmol . 1-' NaCl). The rate of fluid secretion is close to being inversely related to the osmolarity of the bathing fluid (Fig. 34). From these facts it follows that solute secretion does not vary with changes in osmotic 1.50

-

Q

-

1.25 -

0

= U

g

e

1.00-

P E

0.75-

Osmotic concentration of the bathing fluid ( A T )

Fig. 33. The osmotic concentration of the fluid secreted by isolated Malpighian tubules of Rhodnius as a function of the osmotic concentration of the bathing solution (from .Maddrell,1969).

248

S. H. P. MADDRELL

B

40

,

30

'

\

a '

b.,

al c

B

.,*'

',.

20

10

:

I

l

l

l

l

l

l

l

l

l

l

l

l

Fig. 34. The rate of fluid secretion by isolated Malpighian tubules of Rhodnius as a function of the osmotic concentration of the bathing solution (from Maddrell, 1969).

concentration of the bathing medium, but that the rate of water movement is inversely proportional to the osmolarity of the bathing solution. These facts are directly comparable to those obtained in Gzrausius and Calliphora. From them it seems very likely that water movements are osmotically linked to the movements of solutes. It is worth mentioning that the osmotic concentration of the secreted fluid does not depend on the length of the tubule portion taken. The fluid produced by very short lengths is the same as that produced by longer lengths. This argues that the fluid emerging into the lumen after crossing the layer of cells is already iso-osmotic and shows that the standing osmotic gradient is not, as might have been thought

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

249

possible, along the length of the lumen of the tubule with the distal contents more concentrated than the fluid further down the tubule. It has also been discovered recently that the osmotic pressure of the fluid produced by unstimulated Rhodnius tubules secreting at a rate of about 0.1 nl . min-’ does not differ appreciably from that of fluid produced by stimulated tubules secreting some 600 times faster. As was mentioned on p. 237 this is predictable from Diamond and Bossert’s standing osmotic gradient model. These two new points further strengthen the case that water movements across the cells of these Malpighian tubules are osmotically linked to the simultaneous movements of solutes. We are now in a position to consider marshalling the results from these experiments on Rhodnius tubules and to use them as a basis for a possible model to explain the way in which the tubule operates. Again it must be stressed that the point o f doing this is to suggest further experiments and that many o f the features of the model, while they are reasonable suggestions, are speculative and are certainly not to be taken as conclusions which are unavoidable from the results. The model is shown in Fig. 35. Its main features are Bothng solution

Cell

h4--

K*

‘7 Lumen

CI-

Basal cell membrane

Apical cell membrane

Fig. 35. Hypothetical scheme for operation of Malpighian tubules of Rhodnius. Active movementsare indicated by unbroken lines and passive movementsby broken lines.

250

S. H. P. MADDRELL

similar to those described earlier for tubules of Gzruusius and Gzlliphoru with the addition of a sodium pump on both basal and apical sides and a chloride pump on the apical side. The necessity for the basal sodium transporting system is to account for the ability of the tubules to transport sodium ions at a high rate even when the tubules are bathed in solutions of low sodium content. As in Curuusius, a plot of the effect of changing external potassium on the trans-wall potential in cannulated and perfused tubules shows (Fig. 36) a slope close to 5 0 m V for a ten-fold increase in the external potassium concentration. Again the most likely explanation is that *30-

*20 '

r> *I0

'

s

.c

w

2

0.

X s

c

8

-10

e! .f 3

-a

-20

c

P

8

-30

-40

-4 6 'I Q Q

I

I

1

20

30

40

1

I

I

I

l

J

50 60708090100 Concentration of potassium In h t b q fluid (mmol I-') Ib

Fig. 36. The trans-wall potential difference of isolated cannulated Rhodnius Malpighian tubules as a function of the potassium concentration in the bathing medium. The continuous lines join determinations made on a single tubule. The dashed line is that of a 5 8 m V change in potential difference for a ten-fold change in external potassium concentration.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

25 I

the basal cell membrane is more permeable to potassium than sodium. If this is the case, then clearly a basal sodium pump seems essential to provide the apical pump with sodium ions-for without it, the apical pump would quickly be starved because of the impermeability of the basal cell wall to sodium ions. The necessity for a powerful chloride pump on the apical cell membrane follows from the negative trans-wall potential and from the ability of the tubules to transport fluid at the maximum rate in solutions containing only 30 mmol . 1-' of chloride ions-in spite of there being evidence to suggest a relatively slow rate of penetration by other anions (see p. 246). In the absence of sodium ions the model proposes that the entry of potassium ions and of chloride ions would fall to a low level. The apical pumps would then be starved of these ions. That secretion is markedly slowed by such an event in Rhodnius, but less so in Calliphora and less again in Carausius (see Fig. 26) may reflect the fact that Rhodnius tubules secrete faster than do Calliphora tubules which in turn secrete faster than do Carausius tubules. In solutions containing say 50 mmol . 1-' of potassium and no sodium the same rate of passive entry of ions may satigfy the apical pumps of Carausius, partially starve those of Gdliphora and allow only a very low rate of secretion by those of Rhodnius (see Fig. 26). The subsequent addition of sodium ion in Rhodnius has such a dramatic effect on secretion because it leads to the supply of potassium, sodium and chloride ions to the apical pumps, potassium ions by stimulating the basal potassium pump, sodium by providing ions for the basal sodium pump to transport, and chloride ions by allowing them freer entry. In the absence of potassium on the other hand the model proposes that sodium ions are still passed to the apical pumps at a high rate because the basal pump while it is slowed by the absence of potassium ions is still able to pump fast enough to maintain the rate of secretion. The ability of the rate of secretion to stay high in a solution with very little sodium or in a solution with no potassium suggests that the basal pumps can supply ions faster than the apical pumps can deal with them. In the absence of potassium for example, the total sodium transport increases so that the rate of fluid secretion stays the same. The model proposes that under these conditions the apical potassium pump now pumps sodium or that the apical sodium pump accelerates. This increased output of sodium through the apical side could not happen if the rate of entry of sodium ions through the

252

S. H. P. MADDRELL

basal cell membrane were not sufficiently high to allow it. Similarly in external solutions containing say 3 mmol . 1-' of sodium and 147 mmol .1-' of potassium, the rate of movement of potassium across the apical cell wall could not increase to the high level that it does unless the potassium entry across the basal side of the cell were high enough to satisfy it. The fact that secretion occurs at a constant rate in solutions whose composition varies from Na = 150, K = 0 to Na = 1 , K = 149 mmol . 1-' is most satisfactorily explained by supposing that the apical cation pumps together are the rate limiting step. Variations in cation composition would not affect the rate at which sodium and potassium taken together are transported because any unused capacity of one pump would merely pump more of the other cation. An alternative explanation is that the energy supply available limits apical transport so that as the supply of, say, sodium decreases because of reduced entry in a low sodium solution, more energy becomes available for the apical potassium pump which therefore pumps potassium faster. The permeability of the basal cell membrane to chloride ions must be high enough so that the apical chloride pump is not starved of ions except when the external concentration of chloride ions is below 3 0 m m o l . 1-'. The alternative would be to have a basal chloride pump which would supply chloride ions fast enough for the apical pump at concentrations higher than 30 mmol . 1-' . This is less likely because it would then not be clear why copper ions should slow secretioncopper ions are thought only to decrease passive permeability to anions (Bemidge, 1969) so that, for example, they do not prevent the carrier-mediated transport of phosphate ions in Calliphora (p. 220). The necessity of supposing that the basal sodium pump is sensitive to lowered concentrations of potassium follows from the facts that in the absence of potassium, sodium transport from low concentrations is very much slower than it is when potassium is present. This can be deduced from the dependence of the rate of sodium transport on the sodium concentration in tubules bathed in solutions containing more than 140 mmol .1-' of potassium and in solutions containing no potassium. The available results are plotted in Fig. 37. The shift of the curve to the right in zero potassium reminds one very much of the Bohr effect of COz on the oxygen carrying capacity of haemoglobin, which is also an inhibition. That the basal sodium pump is not completely stopped in the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

253

'r

0 0

10

20

30

40

50

60

70

Concentration of sodium in bathing solution (mmol,l-')

Fig. 37. Sodium transport by the Malpighian tubules of Rhodnius as a function of the sodium concentration of the bathing medium, either in the presence of relatively large amounts of potassium or in its absence (data from Maddrell, 1969).

absence of potassium ions follows from the fact that sufficient sodium still reaches the apical pumps to allow the maximum rate of secretion in solutions containing more than 130 mmol . 1-' of sodium. This is unlikely to happen passively because the basal cell wall is thought to be much more permeable to potassium than to sodium ions (p. 251) and yet it does not allow passive potassium entry in the absence of sodium to go on fast enough to allow more than 40% of the maximum rate of secretion (p. 239). This section ought properly to be ended by emphasizing again the speculative nature of its proposals. It is assumed throughout, for example, that very large changes in ion concentrations affect only ion pumps and not say water permeability or the availability of energy to the pumps. While these may be reasonable assumptions, there is little proof of their validity. What is needed now is more information from experiments designed to test predictions based on the model. There is some information on the electrical events involved in secretion by Rhodnius Malpighian tubules (Maddrell, 1971) and this throws some light on the subject. As has been mentioned already the trans-wall potential in Rhodnius tubules after stimulation with the diuretic hormone is such that the lumen is negative t o the

254

S. H. P. MADDRELL

haemolymph by about 30 mV (in cannulated tubules perfused with a solution identical t o the bathing medium). The course of the potential as the diuretic hormone stimulates the tubules follows a spike-like series of changes (Fig. 38). The potential first rapidly goes more negative from an unstimulated level (which averages -1 3 mV) and then more slowly climbs until the lumen is positive with respect t40

c

I

I

0

I

10

I

20

I

30

I

40

I

50

I

60

I 70

Time after addition of hormone(min) Fig. 38. The changes in trans-wall potential difference of isolated cannulated Rhodnius Malpighian tubules after diuretic hormone is added to the bathing solution. The figure is based on the measured responses of 75 tubules.

to the haemolymph (by an average 38 mV) and then it reverses still more slowly t o reach a steady level at -30mV. This steady negativity is maintained while the tubule is still stimulated but the potential climbs as the hormone is used up. Thus the negativity is only maintained for a short time after stimulation by a small dose of the hormone but for considerably longer by a large dose. It is clear from experiments with uncannulated tubules that negativity and rate of secretion go hand in hand (Fig. 39). As has been pointed out, the steady potential produced after stimulation is sensitive t o changes in

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

255

c

I

U

i2 -

20-1 I I

10-

I

I

I ",W \ I T W I ,

Fig. 39. Secretion and trans-wall potential difference in isolated Malpighian tububs of Rhodnius. The potential difference (broken line) has here been plotted as luminal negativity so that it can be compared directly with the changes in rate of secretion (unbroken line).

external potassium concentration. The potential before stimulation is rather similarly affected by changes in potassium concentration. It seems therefore that the hormone does not much affect the relatively much higher permeability of the tubule wall to potassium than to sodium. In contrast the steady potentials before and after stimulation by the diuretic hormone are very differently affected by changes in chloride concentration as Fig. 40 shows. Before stimulation, a ten-fold increase in the external chloride concentration increases the lumen negativity by 11 mV, whereas after stimulation the same change in chloride concentration produces a change in potential of 52 mV. This strongly suggests that one action of the

256

S. H. P. MADDRELL

+

40

> E

: *30 .Q =

L

H .-p +20 f

z

=P

+I0

t

o

0

c

0

I

I

I

I

I

I

I

I

I

1 0

20

30

40

50

€0

70

80

90

Time(rnin)

Fig. 40. Changes in the trans-wall potential difference in an isolated cannulated Malpighian tubule of Rhodnius induced by changes in the concentration of chloride in the bathing medium. Note that a doubling of the chloride concentration before diuretic hormone is added has a much smaller effect than after the hormone has had its effect.

diuretic hormone is to increase the permeability of the tubule wall (most likely the basal cell wall) to chloride ions. Alterations in the ionic composition of the fluid perfusing cannulated tubules has surprisingly little effect on the trans-wall potential. This could mean that the apical cell wall is not passively permeable to these ions, and this one might expect if this wall contains sodium, potassium and chloride pumps; they would have a greater effect if passive back diffusion is kept low. Alternatively, it might be that fluid in the lumen is different from that in contact with most of the luminal cell wall-that is the fluid between the microvilli. If, as has been speculated (Berridge, 1968; Berridge and Oschman, 1969; Maddrell, 1969) transport occurs into the spaces between the microvilli and that the ability of this system to produce an iso-osmotic fluid depends on the increased time for which the transported ions stay near the cell membrane, then it would not be surprising if the composition of this fluid were less affected by changes in the solution out in the lumen proper, especially as the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

257

fluid between the microvilli moves rapidly towards the lumen. However, in the similar situation in the roach gall-bladder where iso-osmotic fluid transport is also achieved by water movements following ion transport into restricted extracellular spaces, in this case between the cells opening on t o the external surface, the potential difference across the wall is very sensitive tochanges in the total concentration of monovalent cations in the external solution, towards which fluid transport occurs (Diamond, 1962). It is very much less sensitive to changes in chloride concentrations. In this tissue there are large areas of cell membrane on the external surface between the intracellular gaps thought to be responsible for fluid transport. These areas of membrane would be easily accessible to the external solution and it seems possible that the potential changes recorded by Diamond were developed across them and not across the transporting lateral cell membranes. It seems then that in Rhodnius Malpighian tubules changes in the luminal solution do not affect the trans-wall potential difference as much as they might, probably because the apical cell membrane has reduced ion permeabilities and because a large part of the microvillar surface is of decreased accessibility from the lumen because of its geometry and because of the fluid flow away from, it. By varying the external composition of the bathing solution and from a knowledge of the composition of the secreted fluid one can, from the steady trans-wall potential difference developed by non-cannulated tubules, estimate the electrochemical gradient against which any ion is moving. Some of the most extreme gradients recorded against which transport can occur are shown in Table 111. It is obviously possible to create conditions such that transport of each of the main monovalent ions is thermodynamically very much uphill. It should clearly be understood that this does not establish that each of these ions is actively transported, though this may well be the case. It is probable, and indeed in a very active fluid and ion transporting system like the tubules of Rhodnius almost certain, that some apparently uphill transport is achieved by frictional forces between the molecules-this includes both solvent drag (Ussing, 1960) and solute drag (Franz et al., 1968; Biber and Curran, 1968). At first sight it is unlikely that such frictional interactions could account for transport against such large gradients as are shown in Table 111 and so it is likely that active transport of each of the ions involved does occur, but proof of this will require examination of cases where only one ion is moving (short-circuited preparations) and

258

S. H. P. MADDRELL

TABLE 111 The most extreme electrochemical gradients recorded against which transport of each major ion can occur in Malpighian tubules of Rhodnius Composition of Composition of bathing solution secreted fluid (mmoles . 1-I) (mmoles . I - ' ) Na K C1 Na K C1 a Sodium

9

b (i) Potassium

C

o

Trans-wall potential difference (lumen with respect t o bathingsolution)

141 155

85

100 180

+4 5

150 155

0

185 180

+ I 58

(ii)

141

9

26

95

90

(i)

141

9 155

95

90 180

-105

(ii)

150

0

170

-20

170

+3 0

Chloride

30 185

0

fluid movement stopped, or cases where from the fluxes of other permeant molecules the extent of solvent-solute and solute-solute interactions can be gauged. The ability of the tubules to transport ions and fluid at various trans-wall potential differences may well reflect the existence of active transport systems for both cations and anions. The sign of the trans-wall potential would depend on whether cations or anions were being transported the faster. As Diamond (1962) has pointed out a similar situation arises in three other preparations: small intestine and in uiuo frog skin both transport sodium and chloride and stomach transports hydrogen ions and chloride ions. In each case it has been possible to observe a potential difference of the sign expected for one pump by eliminating or slowing the other. To some extent this is also possible in Rhodnius Malpighian tubules. The clearest case involves the use of sulphate-based solutions or copper-containing solutions to stop chloride transport. In both cases a high positive potential results (Fig. 41)-as one would expect if lumen-directed cation pumps persisted in their action and the potential difference produced by their action were not shunted out by active (or passive) anion movements. Substitution of cations by tetra-ethylammonium ions produces a negative trans-wall potential difference as expected but it is not larger than that produced in

259

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

1

0

1

10

1

20

1

30

I

40

I

50

1

60

I

70

I

80

Time(rmn1

Fig. 41. The change in trans-wall potential difference of isolated cannulated Rhodnius Malpighian tubules after diuretic hormone is added to the bathing solution which contains, as its only anions, sulphate, bicarbonate and phosphate.

normal solutions. This might be explained on the basis of the model (shown on p. 249) by supposing that chloride entry into the cells is reduced in the absence of cations and so the apical chloride pump is starved of chloride ions and does not generate a large potential. What is needed is an inhibitor which will selectively block cation pumping. It will be worth measuring the potential produced by Malpighian tubules when they are bathed in solutions containing such substances as tetrodotoxin or procaine which tend to block passive cation movements, to see if thereby the potential developed by the chloride pump becomes more apparent. Malpighian tubules of Rhodnius bathed in solutions containing no sodium ions develop a trans-wall potential difference which averages 8 4 m V , lumen positive. To some extent this is expected from the high potassium concentration. However even if the potential is affected similarly by potassium concentrations over 100 mmol . 1-' as it is by lower concentrations one would predict a trans-wall potential difference of only about 30 mV. If, as has been suggested earlier (on p. 251) and above, chloride entry is decreased in the

260

S. H. P. MADDRELL

absence of sodium ions, one would expect the apical chloride pump to be starved and so a positive potential to be developed by the activity of the apical cation pumps. Starvation of chloride achieved in chloride-free solutions leads to a trans-wall potential difference of an average about +55 mV. The basal side effect of high potassium would be expected to add to this and together they would yield a trans-wall potential difference of +85 mV, and this is very similar to that observed. An analogous situation is the effect on the trans-wall potential difference of a potassium sulphate solution-that is a chloride- and sodium-free solution. Again the potential is high (average +75 mV) suggesting the addition of the same two effects mentioned . The results of these experiments on steady potentials suggest: (a) that the basal side cell wall is much more permeable to potassium than sodium; (b) chloride ions are normally secreted faster than are cations; (c) after treatment with the diuretic hormone the wall of the tubule becomes much more permeable to chloride ions; (d) transport of each of the major'ions can occur against steep electrochemical potential gradients; (e) the passage of chloride ions across the wall of the tubule is slowed in the absence of sodium ions. Secretion is much affected by temperature (Maddrell, 1964b) and the trans-wall potential is no exception. At low temperatures the spike potential change induced by the diuretic hormone occurs more slowly and the potential levels out with the lumen positive to the bathing solution. At high temperatures the spike is very fast and the potential level out with the lumen at a potential negative to that of the bathing solution. Typical responses are shown in Fig. 42. The interpretation of this effect is not easy but it does show again that the net trans-wall potential can change even in sign. Such a change in sign would not be expected if the transport of one ion only were the driving force for secretion; probably, as suggested, both anions and cations are actively transported. The spike potential changes induced by the hormone (Fig. 38) can be altered by changes in the external solution. The initial drop is abolished in sulphate-based, chloride-free solutions and is much increased in size in double strength solution where the chloride concentration is 3 10 mmol .1-'. It may be, then, that the initial drop is a result of an increase in chloride permeability on the blood

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

26 1

/S'C J

I

0

I

I

10

20

1

30

1

I

40

50

1

60

1

70

1

80

Tim (inin)

Fig. 42. Spike-like changes induced in the transwall potential difference of isolated cannulated Rhodnius Malpighian tubules by the diuretic hormone at two different temperatures. The bathing solution in each case w a s the standard solution.

side; the increased tendency of chloride ions to enter the cell would make the trans-wall potential difference become more negative. The rise in potential which starts about 1 min after stimulation may be the consequence of cation pumps beginning to pump faster. In a few experiments using tetraethyl ammonium ions to replace sodium and potassium, the potential changes show no rising phase (Fig. 43). In solutions containing no chloride, on the other hand, the rising phase is pronounced and the potential rises to a high level (Fig. 41). The final negative-going change in potential can reasonably be attributed to the activity of the apical chloride pump. If chloride-containing solutions are added to chloride-free ones, the

262

H. P. MADDRELL

S.

P

c

a 0 v1

I ime iminl

Fig. 43. Changes induced in the trans-wall potential difference of an isolated cannulated Malpighian tubule of Rhodnius when diuretic hormone was added to the bathing solution which contained tetraethyl ammonium ions in place of sodium and potassium ions.

0

I

I

I

I

I

I

I

I

20

I

10

30

40

50

60

70

80

90

Time (mn)

Fig. 44. Changes in the trans-wall potential difference of an isolated cannulated Malpighian tubule of Rhodnius in response to the sequential addition of diuretic hormone and then chloride ions to an initially chloride-free, sulphate-based bathing solution.

2 63

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

potential at first rapidly and then more slowly becomes much more negative (Fig. 44). It has also been possible with tracer ions to show that in a normal saline solution a dramatically increased chloride influx into the lumen occurs at the time when the potential starts to go negative. It is also at this time that in tubules bathed in normal saline, fluid secretion starts. These suggestions are summarized in Fig. 45. The ideas contained Flow of flud beqim

*40-

Maintained lumen negatmty during fast fluid secretnn 1

0

I

I

1

I

I

I

I

I

1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0

I

DO

Time (mid

Fig. 45. The spike potential change of hormone-stimulated Rhodnius Malpighian tubules and speculations as to the underlying mechanism.

in it can best be tested by further experiments involving intracellular microelectrodes so that events occurring at the two sides of the cells can be disassociated. Secretion by the upper lengths of the Malpighian tubules of Rhodnius has some similarities with that performed by tubules of Calliphora and Carausius. All these tubules secrete a nearly iso-osmotic fluid and are capable of transporting potassium ions at high rates. The distinctive differences can largely be attributed to the additional abilities of Rhodnius tubules to pump sodium and chloride ions at speed. This in turn, as has been suggested (Maddrell, 1969), may underlie the ability of these tubules to secrete so much faster than the other two tubules considered.

2 64

S. H. P. MADDRELL

E. THE MALPIGHIAN TUBULES OF CALPODES

We turn now to consider the recent work on the Malpighian tubules of Calpodes (Irvine, 1969, 1970). These tubules are structurally and functionally separated into a number of regions which are, taking them in a progressively downstream order: a complex associated with the rectum in a cryptonephridial arrangement; two clear regions, the rectal lead and iliac plexus; and two pigmented lengths, the yellow and white regions which discharge into a small pulsatile bladder which passes fluid into the hind-gut. Malplghrn tubulas

White region

Yellow regm

Iliac plexus

Rectal lead Perinephric membrane

I

/

/

Bladder

MID- GUT

ILEUM

RECTUM

Fig. 46. The various regions of the Malpighian tubules of Calpodes and their relationship with the gut (redrawn from lrvine, 1969). The direction of fluid flow along the Malpighian tubules is shown by the small arrows.

These regions are shown in Fig. 46. Most of Irvine’s work has been concerned with the upstream transparent lengths which can secrete fluid at a high rate. The downstream pigmented regions also are capable of secreting fluid though of a very different composition. This is most neatly shown from one of Irvine’s experiments where he ligated a tubule between the transparent and yellow regions and isolated the tubule into a drop of saline. By puncturing both ends of the tubule, fluid secreted by each of the two regions was collected separately and the sodium and potassium content measured. The results are set out below in Table IV. It seems likely from these results and the fact that the osmotic pressure of the medium was 230mosmol . 1-I that the upper,

T H E MECHANISMS OF INSECT EXCRETORY SYSTEMS

265

TABLE IV Concentrations of sodium and potassium in the fluids secreted by different lengths of Malpighian tubules of Calpodes

"a+]

[K+l

in mmol . 1-'

Transparent region

Bathing medium

Yellow region

12 130

12

50

30 40

transparent region secretes an approximately iso-osmotic fluid and the yellow region a markedly hypo-osmotic fluid. In their function the upstream transparent regions display several similarities with the tubules of Culliphora and Curausius. They secrete an approximately iso-osmotic fluid at a rate which is dependent on the bathing medium potassium concentration; the fluid produced contains an elevated concentration of potassium and a lower concentration of sodium relative to the bathing medium; and the trans-wall potential difference is such that the lumen is positive (by about 25 mV) to the bathing medium. The main differences are that secretion at low potassium concentrations is not slowed if sodium ions are omitted and that there is a more effective sodium transporting system. As a result of the latter, in the absence of potassium a high sodium concentration can support a rate of secretion equivalent t o 30% of that supported by an equal concentration of potassium. The model shown on p. 231 will also fit secretion by the upper transparent lengths of Gzlpodes if it is modified so that the basal potassium pump is not sodium-sensitive, if it includes a more active apical pump for sodium and if it includes a basal element which excludes sodium when its external concentration is below about 60 mmol . 1-' , and in the absence of external potassium. This basal sodium element, which might perhaps be a potassium-sensitive sodium extruding pump, is necessary to explain why sodium is secreted at such slow rates from potassium-free external solutions containing less than about 6 0 mmol .1-' of sodium. If sodium crossed the basal side purely passively one would expect its rate of secretion by an apical pump to increase linearly with the external concentration, and this does happen in the presence of potassium. The potential activity of the apical sodium pump is shown by the

266

S. H . P. MADDRELL

rapid increase in rate of sodium transport as the external sodium concentration is increased beyond 60 mmol . I-' in potassium-free solutions. The handling of sodium and potassium ions by these upper lengths of tubules are summarized in Fig. 47. One surprising fact is that even in a solution containing no sodium or potassium a slow rate of fluid secretion is maintained. Irvine suggests that choline ions may b

0

C

c

$0 5b40 3b 20

Ib 0

N3+

Concentration of ions in battmng fluid (mmol I-')

Concentration of ms in bathing flud (mmd I-')

Fig. 47. Secretion of sodium and potassium ions by the rectal lead and iliac plexus of Malpighian tubules of Calpudes (figures calculated from those of Irvine, 1969). (a) The concentrations of ions in the secreted fluid as a function of their concentrations in the bathing fluid. (b) The rate of secretion of potassium ions as a function of their concentration in the bathing fluid, in the presence of sodium ions (solid circles) or in their absence (open circles). (c) The rate of secretion of sodium ions as a function of their concentration in the bathing fluid, in the presence of potassium ions (left-hand curve) or in their absence (right-hand curve).

be transported by the tubules at low or zero concentrations of sodium ions in the absence of potassium ions. The downstream pigmented regions of the Malpighian tubules of Calpodes will secrete fluid as mentioned earlier. The basis of this secretion is very different from secretion by the upper lengths of the same tubules and from secretion by tubules of Rhodnius, Calliphora, Tipula and Chrausius. When these lower lengths of tubules are bathed in a 230-mosmol medium containing 5 5 mmol . 1-' of potassium and 13 mmol . 1-' of sodium the fluid they produce contains 30 mmol . 1-' of potassium and 60 mmol . 1-' of sodium. They concentrate sodium rather than potassium, and unless calcium and magnesium are transported the fluid secreted must be considerably hypo-osmotic (by up to 60 mosmol . 1-' ). Without further details it is difficult to estimate what mechanisms might account for this secretion. From Irvine's detailed figures for the performance of

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

267

upper and lower lengths of tubules together and for the upper transparent lengths alone, one can deduce that the lower lengths of tubule resorb potassium ions from the lumen at a rate which depends on the concentration gradient across the tubule wall. This movement does not appear to be accompanied by large movements of chloride ions and it is probable that cations from the bathing medium are exchanged for potassium to some extent. As there is little net movement of ions involved in this potassium absorption it is unlikely that much water is absorbed from the lumen. At the same time

K' I

60

I

I

I

50

40

30

I

20

I

10

A

0

No*

Concentrotion of ions in bothing fluid(mmol.l-')

Fig. 48. Net fluid movements produced by the activity of the lower regions (yellow and white regions) of the Malpighian tubules of CaZpodes as a function of the concentrations of sodium and potassium in the bathing fluid. The whole tubule was isolated in each case and the activity of the lower regions deduced from the activity of the whole tubules and that of the upper regions determined later after cutting away the lower regions. The vertical lines attached to the points represent twice the standard error. It is clear that large volume changes are produced only when the sodium concentration is high and the potassium concentration concomitantly low (calculated from data of Irvine, 1970).

268

S. H. P. MADDRELL

sodium ions are transported into the lumen of the tubule at a rate which depends o n the external concentration. This movement appears to be more successful in producing a water movement in the same direction. As a result these lower lengths of tubule add to the volume of fluid passing along the lumen when bathed in solutions containing a good deal of sodium but do not alter it much when the external solution has a low sodium concentration. These points are brought out in Fig. 48. The differences in mechanism, whatever they are, are perhaps to be expected from the fact that these lower lengths of tubule appear not to be primarily concerned with secreting a presumptive urine but rather with a regulation of excretion of sodium and potassium (Irvine, 1970). It is worth recalling that the lower (downstream) lengths of the Malpighian tubules of Rhodnius behave in a totally different fashion from the upper lengths of the same tubule. They d o not secrete fluid and they may well exchange luminal potassium partly for sodium from the bathing medium (Maddrell, 1969). The tubules of Culpodes gradually change in character along their length from a “typical” tubule producing an iso-osmotic primary excretory fluid rich in potassium, to a tubule primarily concerned with ion regulation by modifying fluid passing through it.

F. THE ULTRASTRUCTURE OF MALPIGHIAN TUBULES AND ITS FUNCTIONAL SIGNIFICANCE

Many ultrastructural studies have been made of Malpighian tubules (see Berridge and Oschman, 1969 for references) and generally speaking they all agree in revealing a common plan which qualitatively is the same for all Malpighian tubules examined. Figure 49 is a diagram of a cross section through the wall of a Malpighian tubule to show those features of its organization which appear, at the moment, t o be significant. Taking them in the order from the basal or haemolymph side of the cell, they are: (1) basement membrane; (2) extensive tracheal and tracheolar supply; (3) deep and tortuous basal inpushings of the basal cell membrane; (4) relatively little cytoplasm t o cross before reaching ( 5 ) long closely packed microvilli, a proportion of which have a long mitochondrion running along the length of the microvillus;

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

269

Fig. 49. Diagrammatic representation of a cross section of the wall of a Malpighian tubule emphasizing those features which are, at the moment, thought to be significant for its function.

(6) the cells are joined to one another laterally by very extensive septate desmosomes. The basement membrane appears in some Malpighian tubules t o deny very large molecules access t o the cells of the tubule. Evidence for this comes from the work of Locke and Collins (1967, 1968) on Culpodes. They showed that while the basement membrane of the fat body allowed blood proteins and injected peroxidase to penetrate freely, that of the Malpighian tubules did not. In an epithelial tissue which transports fluid at high rates this makes good sense. If it did not occur, one can imagine that the basal infoldings might very soon be clogged by large molecules swept in by fluid flowing through the tubule wall. However this apparently desirable feature is not found in

270

S. H .

P. MADDRELL

all Malpighian tubules. The basement membrane of the tubules of the dragonfly, Libellulu, is permeable to horseradish peroxidase (MW ca 40,000) and this substance is found in all the basal infoldings after being injected into the haemocoel (Kessel, 1970). As Kessel points out, it will be interesting to see whether or not such tubules with permeable basement membranes can continue t o secrete fluid when the basal unfoldings contain large amounts of protein. The extensive tracheal and tracheolar supply typical of many tubules indicates that the tubules are, or can be at times, very active. It has been suggested that isolated Rhodnius tubules, whose tracheal supply is cut off, are limited in their rate of secretion by the availability of oxygen (Maddrell, 1969). Similarly, it is known that in vitro preparations of the midgut of Antheruea pernyi require a constant circulation of the bathing medium and its saturation with 70% oxygen before oxygen supply is no longer limiting (Wood, 1971). That Malpighian tubules are not as sensitive as this probably follows partly from their less frenetic activity, partly from the fact that fluid surrounding them is constantly moving towards them carrying oxygen with it (Maddrell, 1969) and partly from their favourable surface area/volume ratio. The long basal infoldings of Malpighian tubules are very likely specializations to increase the rate of fluid transport across this side of the wall; they may also help to promote the passage into the tubule fluid of small molecules occurring in the haemolymph. The increased rate of fluid transport follows largely from the fact that the area of the basal side of the cell is increased many times by the infoldings so that ion transport across this side is faster. Also the dimensions of the basal channels are such as t o allow the development of local standing osmotic gradients (Diamond and Bossert, 1967) though because their absolute length (5-10 pm) is a lot shorter than the 50-100 pm channels examined by Diamond and Bossert, the effect will not be a large one. Infoldings which face the medium from which fluid is being absorbed are thought t o be capable of osmotically linking water movements to those of ions (Diamond and Bossert, 1968). The fact that the infoldings are close together means that the folds of cytoplasm between them also have dimensions which would slow down the rapid mixing of transported ions with the main body of the cytoplasm, as Berridge and Oschman (1969) have pointed out. It could well be that this effect together with the concomitant reduced ability of the hypo-osmotic fluid within the basal extracellular channels t o equilibrate as fast as

27 1

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

otherwise with the bathing medium, means that fluid movements are even more efficiently coupled with ion movements. To put it another way, ion movements into the cell not only tend to leave behind a hypo4smotic fluid but they establish a hyper-osmotic one on the other side of the wall. Since neither fluid can equilibrate as quickly with the bathing medium or with the main body of the cytoplasm as otherwise would be the case, conditions are established for an accelerated flow of water into the cell. The osmotic gradient created in such a situation may very well be steeper than in the situations examined by Diamond and Bossert for which they assumed that only on one side of the cell wall were conditions appropriate to restrict diffusion. This steeper gradient may avoid the limitation on rate of transport in “backwards” facing channels discussed by Diamond and Bossert (1 968), by effectively speeding the rate of water movements. It may also counteract the tendency for gradients in short channels t o be of small size and so of reduced effectiveness. It appears at first sight that a further important result of having narrow channels present on both sides of the cell wall may be that effectively most ions are moved across the wall at the base of the system. This is because the concentration of ions in the fluid from which absorption is occurring tend to fall away as the fluid passes deeper into the system; ions are not able to enter this compartment by diffusion as rapidly as they could if the channel were very short. Figure 50 gives an impression of this process. It is worth noting that

. ..

.

.

:

Fig. 50. Possible events in the coupling of ion and water movements at a cell membrane thrown into a series of close-packed folds-as in the basal infoldings and apical microvilli of Malpighian tubules. It should be noted that, for clarity, the concentration gradients occurring have been exaggerated; in the corresponding structures of Malpighian tubules the gradients will be very small because the channels between folds of cell membrane are short (5-10pm). Active ion movements are indicated by the smaller continuous arrows and passive water movements by the larger broken arrows.

272

S. H. P. MADDRELL

exactly the same arguments apply whichever of sides 1 or 2 is the extracellular fluid. That is both “backwards” and “forwards” transporting membranes to use Diamond and Bossert’s terminology. One of the assumptions that Diamond and Bossert have had to make in their theoretical treatment of fluid transport into a single channel (Diamond and Bossert, 1967) is that in order for fluid transport to be iso-osmotic most of the ion pumps have to be at the base of the channel into (or from which) secretion is occurring. In a situation with many backward and forward facing channels in an alternate close-packed parallel array, there may be, as shown in Fig. 50, a decrease in ion concentration in the fluid from which transport is occurring. If, as is possible, the rate of pumping of ions depends on the concentration available, transport will tend to be faster at the base of the channel-not because of any restriction on the sites of ion pumps but by reduced amounts of substrate for them. If this is what happens, it will neatly fulfil the condition that for iso-osmotic transport, ion transport needs to be concentrated in the base of the channels. It should be emphasized that these consequences of having an alternate array of backward and forward facing channels are based on assumptions, which while it is thought they are reasonable, require further support. In particular it must be stressed again that the depth of the basal infoldings at 5-1 0 pm is considerably less than the 100 pm which in Diamond and Bossert’s analysis is needed in widely separated channels t o produce a virtually complete osmotic coupling when such parameters as the solute transport rate, water permeability, channel radius and solute diffusion coefficient are set at values roughly averaged for transporting tissues. It may be that a parallel array of alternatively forwards- and backwards-facing channels is a more efficient way of coupling solute and water transport than widely separated channels and so the close-packed channels can be shorter than would otherwise be the case. What is now needed is a mathematical analysis of the situation illustrated in Fig. 50, using the approach established by Diamond and Bossert, to see if the predictions made here are correct. It is clear from their structural similarities that fluid transport across basal infoldings may be qualitatively similar to that across the apical microvilli. In both systems restricted channels on either side of the cell membrane are arranged alternately and parallel with one another. The apical microvilli are often longer than the basal infoldings. From the point of view of the need to transport fluid, the tubule wall enjoys the advantage of two similar systems in series.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

273

Probably fluid transport across both cell walls is linked with ion transport and so fluid transport is not slowed because of restrictions imposed by passive flow across one cell wall. It seems that in Malpighian tubules, Diamond and Bossert’s idea of “backward” and “forward” systems facing respectively the fluid from which and to which secretion is occurring, breaks down. Both sides of Malpighian tubules contain both backwards- and forwards-facing channels. There is a further advantage peculiar t o excretory systems in having “backwards” facing channels on both sides of the cell. Primary excretory fluids in most animals consist of a more or less indiscriminate ultrafiltrate of the intercellular fluid. This is relatively easily produced by a positive hydrodynamic pressure acting across a leaky extracellular membrane as in the kidney glomerulus. It is of course not so easily produced by excretory systems which, because of the need to secrete a fluid by a process led by ion transport consist of a relatively “tight” layer of epithelial cells. Malpighian tubule cells are joined together by extensive septate desmosomes which are thought very much to reduce extracellular movement of material between the cells. As is discussed by Diamond and Bossert (1968) and in particular for Malpighian tubules by. Berridge and Oschman (1 969) it is a consequence of “backwards” facing channels that non-transported solutes are swept into them in the flow of fluid entering. They accumulate in the channel and reach a concentration which is determined by the rate of fluid flow, the dimensions of the channel, the rate of back diffusion out of the channel and the net rate of penetration through the cell wall. For molecules able t o diffuse through the cell wall, this higher concentration will ensure a faster penetration of these molecules than in the absence of such a concentrating mechanism. This mechanism will not only apply at the basal side of the Malpighian tubule as discussed by Berridge and Oschman (1969) but will also apply at the luminal side. This is important because it allows molecules to cross both sides of the cell faster than otherwise. If such a system operated on only one side the final concentration of permeant non-transported molecules in the lumen would be lower and more fluid would have t o be produced effectively to ultrafilter the same amount of haemolymph. This is of course important for excretory systems where, as is argued on p. 280, the ability to filter permeant molecules from the extracellular fluid is crucial for their function. It is noticeable that several nonexcretory fluid secreting systems have either no basal infoldings or few and short ones. As an example, the gall bladder of the rabbit AIP- 12

274

S . H. P. MADDRELL

(Kaye et al., 1966; Tormey and Diamond, 1967) has facing the luminal solution (from which absorption occurs) short microvilli fairly widely separated. It seems unlikely that the short relatively wide spaces between the microvilli could act as do basal infoldings in Malpighian tubules. The tendency of non-transported molecules to accumulate at this surface of the cell would readily be counteracted by the ease with which they could diffuse away. Again, what is now needed is a mathematical analysis of this sweeping-in effect so that its effectiveness can be gauged for basal infoldings of various dimensions and for various solutes. A particular calculation to this end has been most kindly done for me by Dr. R. B. Moreton. The result is rather surprising in that it shows that the effect is almost negligible for small molecules. For example, in a 5 pm infolding of diameter 60 nm into which water is flowing at 2 pm . min-' , alanine at 100 mmol . 1-' in the incoming fluid is concentrated only to 100.1 mmol . 1-' in the deepest part of the infolding when the U/P ratio is 0.3. It follows that the tubule wall in this case must be very permeable to alanine. The figures used in the calculation are taken from Ramsay's and Pilcher's work on stick insect tubules. That the sweeping-in effect is not bigger really follows from the fact that the half-time for diffusion out of the infolding is for alanine only 13.5 ms which is much smaller than the transit time of 150 s for the fluid entering. There is one factor which may increase the effectiveness of sweeping in. This is that water in channels as narrow as the basal infoldings of Malpighian tubules may have properties significantly different from those of bulk water (Diamond and Bossert, 1967). Diffusion coefficients of solutes in narrow channels may well be less than in bulk water, so that their tendency to accumulate in the infoldings would be increased. A similar argument applies of course, to osmotic coupling of ion and water movements in basal infoldings and in the luminal microvilli. Until more is known of this possibility it will be prudent not t o overestimate the importance of the sweeping-in effect for diffusible molecules. It is clear that only for very slow-diffusing molecules in deeper infoldings with water entering faster will there be a tendency for them to become concentrated to significant extents. Since it is possible to control the rate of fluid secretion of Malpighian tubules by the osmotic concentration, by the addition of hormones or stimulant substances and in most cases by the potassium concentration also, it is open to experiment to show how important sweeping-in is in

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

275

promoting trans-wall movements by following such movements of a range of molecules at a range of fluid secretion rates. Recent work has shown that both water and solute movements through biological membranes can entrain other solute molecules by the frictional effects known respectively as solvent and solute drags (Ussing, 1960; Franz et al., 1968; Biber and Curran, 1968). These effects may play a large part in securing an accelerated movement of small solute molecules into the lumen of Malpighian tubules. It may turn out to be that it is these effects, rather than sweeping-in, which explain the surprisingly high concentrations of organic solutes found in the fluid secreted by Malpighian tubules (Ramsay, 1958). The presence of mitochondria within the basal infoldings and especially within the apical microvilli, where they occupy a large proportion of the cross sectional area (see Fig. 49), may indicate the high energy demands of these areas of the cells. It is not clear why there should be a higher concentration of them towards the luminal side of most tubules. It may be that the apical pumps are more active than the basal ones. Their presence within the folds of cytoplasm on both sides of the cell must also help to increase the surface area/volume ratio of these parts and so ,contribute to osmotic equilibration across the cell walls. It is noticeable that the proportion of microvilli containing mitochondria varies widely in different species of insect. The significance of this is not yet certain, but at the moment it appears that the faster-secreting tubules have a greater density of mitochondria in the microvilli. In summary, the ultrastructure of the Malpighian tubules suggests several ways in which the structure of the cells may promote the movement of molecules across the tubule wall. In particulzr, it suggests a special role for the convoluted cell membranes on the basal and apical sides of the cells. These specializations probably act to ensure that active lumen-directed movements of ions through these membranes are osmotically coupled with movements of water in the same direction. There will also be a tendency for substances which are not actively transported to penetrate the tubule wall faster because of the lumen-directed flow of secreted fluid. This effect seems more likely, a t the moment, to be attributable t o frictional entrainment between such substances and the constituents of the secreted fluid than to accumulation at elevated concentrations in the backwards-facing channels, which now seems unlikely to occur to significant extents. Again it should be stressed that the arguments

276

S. H. P. MADDRELL

used in supporting these ideas are largely qualitative. What is now needed is both a further mathematical analysis of the situation and experimental investigations so that the importance of these effects can be gauged. Three more points can be made from the ultrastructure of Malpighian tubules. They are: (a) the demonstration of magnesium activated ATPases on both basal and apical plasma membranes, and of an alkaline phosphatase on the apical cell membrane of tubule cells of Calliphora (Berridge, 1967); (b) the microvilli on the apical side of the tubule cells of Calliphora have on the cytoplasmic side of the plasma membrane numerous 100-Aparticles (Berridge and Oschman, 1969); (c) within the cytoplasm of the cells of the Malpighian tubules of Carausius, Drosophila and Schistocerca is found a system of tubules which are structurally similar t o the smooth endoplasmic reticulum (Pilcher, 1969; Wessing and Eichelberg, 1969; Joyner, 1970). These small tubules have a preferred orientation at right-angles to the tubule wall, that is in the direction of fluid flow. They run into the basal cytoplasmic folds and into the apical microvilli. All these findings may in the future throw light on the way in which tubules work. Already the demonstration of ATPases on the cell membranes of Calliphora tubules ties in with the suggestion of active pumps on both sides of the cell (p. 23 1). The apically situated alkaline phosphatase may be concerned with phosphate transport there (p. 234). 1 0 0 4 cytoplasmic particles similar to those found on the apical cell membranes of Calliphora tubules have also been seen in several other insect-transporting tissues (see Berridge and Oschman, 1969 for references). Their function is as yet unknown. The intracytoplasmic tubules found in the cells of Malpighian tubules may perhaps have some role in directing or channelling fluid movement in these cells. G. FORMED BODIES

One aspect of ultrastructural studies of insect Malpighian tubules which deserves special mention is the presence within the lumen and between the microvilli of numbers of small and a fair number of larger membrane bound vesicles. These vesicles often appear to contain within themselves further small vesicles packed together or arranged concentrically. It is very likely that these are the “formed

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

277

bodies” which have been described by Riegel (1 966a) as typical of the fluid produced by excretory systems of several animals as diverse as the frog, the crayfish and the stick insect. One’s first reaction to the discovery with the electron microscope of such bodies within Malpighian tubules is to question whether they might be artefacts. It is well known, for example, that Malpighian tubules undergo prominent cytological changes under unsuitable in vitro conditions (Wigglesworth, 1965). It might be thought, therefore, that such structures might arise during fixation especially as they occur on the luminal side of a structure fixed from the basal or haemolymph side. That this is unlikely comes from several lines of evidence. Micropuncture samples taken from living in situ excretory organs of frogs, crayfishes and stick insects agree in that they all contain formed bodies (Riegel, 1966a). The fluid produced by isolated Malpighian tubules of Rhodnius which are secreting at a high rate contains formed bodies (unpublished results of J. A. Riegel and the writer). More convincing is the demonstration of formed bodies in the urine freely produced by intact Rhodnius after a blood meal (unpublished results of J. A. Riegel and the writer). Clearly the formed bodies cannot be artefacts-though it still might be argued that they are the products of ageing cells being replaced or renovated. The view taken by the writer is that formed bodies are a consistent and natural feature of the fluid produced by many excretory systems. What might be the function of formed bodies? In his earlier work, Riegel (1966a) suggested that in glomerular nephrons they may accumulate and release into the lumen such substances as albumen and peroxidase. They may in the lumen of the coelomosac of the antenna1 gland of the crayfish accumulate congo red previously injected into’ the haemolymph or lipochrome, and orange pigment found in the haemolymph (Riegel, 1966b). On this basis, Riegel (1966b) proposes that formed bodies are the vehicle for the transport into the lumen of materials which are too large or are otherwise unsuitable for entry by other methods. As a speculation, Riegel (1 966a) suggests that formed bodies will be found to underly a very large number of secretory processes. He suggests that fluid secretion by aglomerular kidneys and Malpighian tubules is initiated by the secretion of formed bodies. The evidence outlined in this paper leads to the conclusion that fluid secretion in Malpighian tubules is driven by ion secretion, in particular by movements of potassium, chloride and to a lesser extent sodium ions (and also

278

S. H. P. MADDRELL

phosphate ions in Gdliphora and Gzmusius). The two points of view might be reconciled if it could be shown that formed bodies in Malpighian tubules were very rich sources of these ions. There has been no suggestion yet that this is the case. Formed bodies from crayfish antenna1 glands contain only small amounts of potassium (2-3 mM/l) and less sodium (Riegel, 1968). Recent work on crayfish formed bodies have shown however that they are capable of absorbing very large quantities of fluid from saline solutions (Riegel, 1970). On this basis one cannot then entirely exclude the possibility that formed bodies do play some role in the formation of fluid secretion. It is the view of the writer that it is unlikely that formed bodies are essential to fluid secretion for the following reasons. When the tubules of Rhodnius are secreting fast, a volume of fluid equal to the volume of the cells passes through the wall every 16 s (Maddrell, 1969). Yet, tubules rapidly fixed while secreting at this maximum rate have only a few formed bodies visible in the lumen and among the microvilli. One interpretation of this might be that during fast secretion nearly all the formed bodies swell and burst. If this were so, then there ought to be visible large numbers of fragments of membranes but such is not the case. In the writer’s view, the reduced number of formed bodies found in tubules fixed while secreting rapidly and in the fluid produced by such fast secreting tubules is best explained by supposing that the processes of fluid secretion and formed body formation are independent of each other. So when tubules are secreting fast, formed body formation would be relatively slow compared with fluid secretion and a sample of fluid or a length of tubule will have only a low concentration of formed bodies. At a low rate of fluid secretion one finds larger numbers of formed bodies in the secretion (unpublisGed results.of J. A. Riegel and the writer). Again this can be explained if formed body formation did not alter in rate so that fluid collecting in the lumen at a low rate would contain a higher concentration of formed bodies. The alternative explanation is that if fluid secretion is a result of formed bodies swelling and bursting, a slow rate of fluid secretion might be correlated with a much higher survival of intact formed bodies. Further relevant evidence comes from an electron micrographic study of the salivary glands of Culliphoru (Oschman and Berridge, 1970). This tissue also secretes fluid at a high rate (Berridge and Patel, 1968). Electron micrographs of this tissue show no structures which resemble formed bodies. Although no formed bodies can be

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

279

seen in the salivary glands, they are a common feature of Malpighian tubules of the same insect (Berridge and Oschman, 1969). It may be argued that formed bodies of the salivary glands of Culliphora are short-lived because of the high rate of secretion. Ultrastructural examination of this tissue by the freeze-etching technique which allows a much more rapid fixation than do chemical fixatives, should show whether formed bodies are present or not. In summary, it seems t o the writer that formed bodies may be an important feature of many secretory tissues and that their role may well be, as Riegel first suggested, the transport into the secreted fluid of materials not able to enter by other methods. There is clearly scope for more research into these intriguing structures. H. THE HANDLING BY MALPIGHIAN TUBULES OF ORGANIC SOLUTES

1. Organic Solutes of Low Molecular Weight

This article has so far in very large part concentrated attention on the handling by the excretory system of ions and water. This is because of course nearly all recent research has been concerned with this side of the subject. However, as Ramsay (1 958) showed, various organic solutes of low molecular weight when added to the fluid bathing isolated Malpighian tubules of Curuusius rather surprisingly appear at substantial concentrations in the secreted fluid. In these experiments, Ramsay used the following substances: DL-alanine, L -arginine, glycine, L -lysine, L-proline, DL-valine, D -glucose, ~ - f r u c t o s e ,sucrose and urea. He found that all of these substances were easily able to cross the wall of the tubule-Table V shows figures for the ratio of the concentrations in the secreted fluid and medium (U/P ratio) for these substances. What is surprising at first sight is that the concentrations should be so high in the secreted fluid. It is possible that frictional effects between the fluid moving into the lumen and these substances (p. 278) may contribute to this, but it is clear that the tubule wall must be very permeable to these substances. Indeed if the tubule is to perform its function of forming a fluid containing most small molecules in a relatively unselective fashion (see below) then such a permeability is to be expected. Ramsay’s evidence supports the proposition that the substances investigated entered the secreted fluid by diffusion. In particular (1) the U/P ratio is never greater that 1; (2) the U/P ratio for each substance is largely independent of P; (3) the effects of rate of flow upon the U/P ratio are those predicted assuming diffusion to occur

280

S.

H. P. MADDRELL

TABLE V Concentrations of various organic substances in the fluid secreted by isolated Malpighian tubules of Carausius (figures from Ramsay, 1958) Substance Alanine Arginine Gly cine Proline Valine Glucose Fructose Sucrose Urea

Concentration in bathing solution (mmol . I-'

U/P ratio (Mean f S.D.)

56 53 43 52

0.25 f 0.05 0.35 f 0.09 0.17 f 0.05 0.59 f 0.04 0 . 2 4 f 0.04

53 53 53 56

0.86 f 0.18 0.61 k 0.08 0 . 5 3 f 0.10-

59

0.96 f 0.06

and (4)there is no interference between different substances. The crucial finding from this work is that metabolically useful molecules are to be found in the fluid leaving the Malpighian tubules. It is not immediately obvious what advantage there might be from such an arrangement. Indeed, these molecules have to be actively resorbed in the hind-gut (see p. 286) which process consumes energy. Ramsay's elegant explanation of this apparently useless manoeuvre is that it reflects a principle of the design of the insect excretory system. Given a circulation of water through the excretory system, unwanted substances may either be secreted into the stream or all substances useful and useless may be allowed t o enter the stream passively and useful substances are later reabsorbed. The fact of the commital of useful substances to the stream suggests that insects have adopted the second scheme. The major advantage of such a scheme is that it ensures automatic excretion of any unwanted substance simply by not providing a specific mechanism for its reabsorption. This idea is discussed further on p. 306. 2. The Excretion by Malpighian Tubules of Dyes and Other Complex Organic Molecules We now touch on the problem of the possible excretion by the Malpighian tubules of the useless or toxic substances of a size slightly larger than considered in (1) above and which may be formed in metabolism or which are absorbed from the diet or external medium.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

28 1

By analogy with studies on the vertebrate kidney (see Smith, 195 1) one might expect the excretion of such substances to occur in insects. Unfortunately, very few studies have been made to investigate these sort of problems in insects. The literature on insect excretion however does contain sporadic references t o the excretion of dye molecules such as phenol red and indigo carmine. Here again research has to a large extent merely scratched the surface of the problem. The work of Lison (1937, 1938) and Palm (1952) have been the only worthwhile attempts to investigate the phenomenon in a systematic way. As will be argued, the subject is a very important one so that no excuse needs to be made for referring fully to this early work. Lison took as his experimental material individuals of Periplaneta orientalis and Forficula auricularia and injected into their haemolymph small amounts of dyes made up as 0.5% solutions. He then by subsequent dissection and microscopic examination followed the course of their excretion. It is useful to record here a summary of his main conclusions. He first of all (Lison, 1937) systematically worked through some 62 acid dyes. He found them to be divisible into four groups according to the way they were handled by the Malpighian tubules. Members of the first group were continuously concentrated by the tubules, the second group appeared to permeate the tubule freely but were not concentrated, the third group appeared to stain the cytoplasm of the tubule cells very intensely-often leading to their disintegration. As a result it was difficult to estimate the depth of colour in the lumen and further experiments were not done with these dyes. The fourth group were not able to penetrate the basal side of the tubule cells and so were not excreted at all. Examples of the structures of some of these dyes are shown in Table VI. Whether or not a dye molecule was concentrated by the tubules did not depend on the effective size of the dye molecules as determined by measurements of their diffusion coefficients. Nor was it related to the speed with which the dyes were able to penetrate gelatine. It appears from Table VI that taken with the results of his work on basic dyes (Lison, 1938) only those molecules with strongly acid groups and no basic groups such as -NH2 are actively transported by the tubules. Molecules with both acid and basic groups appear not to be able t o cross the basal membranes, while substances with weakly acid groups which can cross the basal membrane tend not to be concentrated in the lumen, though they may lead t o vital staining of the cytoplasm.

TABLE VI Structures of various dyes tested by Lison ( 1 937, 193 8) on Malpighian tubules of Periplunefu and Forficulu Acid dyes concentrated by the Malpighian tubules

a

m

*dxL& 2R

Acid dyes which permeated the tubules and stained the cytoplasm very intensely

nhmk r*M

Acid dyes not able to penetrate the basal side of

m “ild red

Basic dyes which permeated the tubules but were not concentrated

rcn,~,~ nwtml md

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

283

Palm, by contrast with Lison, confined his attention to a small number of dyes but examined their excretion or deposition in the body in more than 125 different species of insects. His conclusions support those of Lison in showing that acid dyes are swiftly excreted. More basic dyes are excreted a good deal more slowly. He found that trypan blue, which in Lison’s experiments on Forficula and Periplaneta was not excreted at all, can be excreted by Malpighian tubules of some Coleoptera, Lepidoptera and Heteroptera. He was also able to show that the rate of excretion of a dye is proportional to its concentration in the blood. The most valuable part of Palm’s work comes from his observations on the distribution of dyes in the insect as a whole rather than just their handling by the excretory system. He found that, in general, there was a correlation between the ease with which a dye is excreted by the Malpighian tubules and the extent to which it is taken up by the pericardial cells and other tissue. “Dye injected is taken up and excreted by Malpighian tubules if the substance in question can be excreted by these organs. If it cannot, it is taken up by the pericardial tissues and by the blood cells as far as they are able to cope with the dye. But if this resorption is limited, and if a dose exceeding the capacity of these tissues is injected, the remainder stays in the blood, possibly staining other organs. Substances which are insoluble as dispersed free carmine or indian ink (carbon particles) will be phagocytosed by special cells, or coated with free haemocytes.” Further than this he noted that when he injected dyes which were only slowly excreted by the Malpiaan tubules, they were first taken up by the pericardial cells and slowly released as excretion by the Malpighian tubules lowered the concentration in the haemolymph. Palm’s observations on the handling of particles of large size fit well with those of other workers. The ways in which the excretory system is able to handle a range of size of substances which need to be excreted are discussed on p. 306. Although the work of Palm and Lison provides the data on which to base a theory for predicting which of a range of dyes will actively be excreted in various insects, it does not provide an explanation of why such molecules are excreted at all. This has been a problem which seems not to have aroused comment until fairly recently, when Ramsay (1958) observed “it is not easy to see why natural selection should have operated to provide this one class of substances (dye molecules) with such an efficient mechanism of elimination or how this observation can be assimilated to any theory of the normal

284

S . H. P. MADDRELL

operation of the excretory system. It may be that the ability to concentrate dyes is an incidental property of some other feature of the excretory mechanism”. Two points arise from these remarks. The first is to ask whether dye secretion and fluid secretion are linked phenomena or whether they are separate activities of Malpighian tubules. This point is considered below. The second is why dyes are so actively secreted. Mammalian physiologists have already provided an answer to this question in relation to the similar activities of the kidney. There it appears that dye molecules are concentrated by a mechanism normally used to excrete various non-metabolizable aromatic residues such as hippuric acid whose effective excretion is important (Smith, 1951). It seems very likely that a similar explanation will be found to underlie the apparently gratuitous ability of the insect to excrete various dyes. Dyes are not excreted because they are dyes but because they structurally resemble other uncoloured excreted molecules. It is clear that here is a whole field of research remaining to be worked. However we must return to the question asked above-are the abilities of Malpighian tubules to secrete dye and to secrete fluid coupled activities or not? Some recent work of Pilcher (1969) and of the author and S. E. Reynolds (unpublished results) shows rather clearly that for the Malpighian tubules of Gzruusius and Rhodnius, at least, they are not. For these experiments, tubules were isolated into drops of Ringer’s solution containing the dye indigo carmine (molecular weight 446). A burst of fast fluid secretion was then initiated by adding a small amount of a diuretic hormone preparation or a little 5-hydroxytryptamine-containingsolution (Maddrell et al.. 1969). Drops of fluid secreted at a variety of rates were collected. The dye content of the drops was measured by spotting them on to strips of chromatography paper and scanning the dried strips with a spectrophotometer. Figure 51 and Fig. 52 show how dye clearance varies with speed of fluid secretion. It is clear that dye clearance (rate of dye secretion) is very little affected even by large changes in the rate of fluid secretion. At low rates of fluid secretion, the dye is concentrated by the tubules but at higher rates, dye secretion does not speed up and as a result the fluid produced contains only low concentrations of the dye. Further than this, it was found that for the Malpighian tubules of Rhodnius dye clearance was enhanced by an increase in pH from 6.7 to 7.5. Such a change in pH has no effect on fluid secretion (see p. 226). This suggests, of course, that it is the

285

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

"0

I

2

3

4

5

Rate of fluid secretion (n1.rnin-I)

Fig. 51. The rate of dye secretion by isolated Malpighian tubules of Curuusius as a function of the rate of fluid secretion. Each line represents the linear regression calculated for the tubules from a single insect. The continuous lines represent results from experiments with indigo carmine and the dashed lines the results of experiments using neutral red (calculated from data of Pilcher, 1969).

Rate of fluid secretion (nl-mn-')

Fig. 52. The rate of secretion of indigo carmine by isolated Malpighian tubules of Rhodnius as a function of the rate of fluid secretion. The line represents the linear regression of the results of 26 determinations on the tubules of nine insects. The pH of the dye solution was 6.7.

anionic form of the dye which is transported and that at lower pH's this dissociation is suppressed. This demonstration that for Rhodnius and Chusius at least, dye secretion and fluid secretion are separate activities of Malpiaan tubules, means that dye secretion cannot be used as a marker for following fluid secretion in these two species. Some workers however

286

S.

H. P. MADDRELL

(see Gersch, 1967; Unger, 1965) have used dye movements as markers for fluid movement in other species. Their interpretation of their results must now come under suspicion until it can be shown whether or not, in the particular species used, fluid movements are correlated with dye movements. As a summary of this section we can say that the ability of the Malpighian tubules of many insects to concentrate acidic dyes is well established. It seems very likely that this ability is merely a visualization of an important ongoing excretion of various non-metabolizable organic acids. In two species of insects at least, Rhodnius and Curuusius, the mechanisms of dye excretion and fluid excretion seem not to be coupled and the two processes t o be separate activities of Malpighian tubules.

V. THE HINDGUT

The preceding sections have described in some detail how the Malpighian tubules function. In general, it may be said that they produce a fluid more or less iso-osmotic with the haemolymph, containing an elevated concentration of potassium ions, considerable amounts of chloride and phosphate ions, and a reduced concentration of sodium ions. The fluid also contains in solution, substances of low molecular weight at a concentration related to that in the haemolymph and t o the ease with which they cross the cell membranes. As Ramsay (1956) has pointed out the excretion from the insect of such a fluid would be so inappropriate as to result rapidly in death. It is very largely the function of the hind-gut t o act on the primary excretory fluid from the Malpighlan tubules in such a way that useful substances are conserved and that toxic substances and substances present t o excess are eliminated. We have already seen that in Rhodnius and Calpodes the lower lengths of the Malpighian tubules themselves undertake some modification of and resorption from the primary excretory fluid produced further upstream in the tubules. In most cases, however, these functions are the preserve of the hind-gut. The hind-gut typically consists of a narrow anterior tube, the ileum, leading posteriorly into a much thicker walled rectum (Fig. 53 and see Wigglesworth, 1965). In some insects the hind-gut preceding the rectum can be divided into an anterior ileum and a posterior colon. All the hind-gut is lined with cuticle which as we shall see has important consequences (p. 304).

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

287

Mdpiohion

tubule Roventriculus

Fig. 53. Diagram to show, for a "typical" insect, the relationship of the Malpighian tubules with the rest of the alimentary canal. Note that fluid from the Malpighian tubules passes through the ileum before reaching the rectum. A. THE ACTION OF THE HIND-GUT ANTERIOR TO THE RECTUM

So far more attention has been paid by research workers to the activity of the rectum in controlling what is finally to be eliminated. However it is probable that the ileum or colon also modify the fluid passing through them. In several insects, the material passing along the hind-gut becomes progressively drier as it passes through in turn the ileum, colon and rectum (Wigglesworth, 1965). Phillips (1964a) observed in the locust a concentration of the dye amaranth in the hind-gut anterior to the rectum. Ramsay (1955a) showed for the stick insect, Carausius, that fluid could be absorbed by the hind-gut anterior to the rectum, but that such resorption was not regulative in that the fluid remaining was more or less unchanged in composition. More recently, Wall (1970), has shown that while the Malpighian tubules of PeripZaneta produce a slightly hyper-osmotic fluid, the fluid entering the rectum is usually hypo-osmotic. This suggests that the anterior lengths of the hind-gut may absorb solutes from the contents. Finally, it is worth pointing out that in Thermobia the rectum must almost always be filled with air as it is the site of uptake of water vapour from sub-saturated air (see p. 308). It follows that resorption of water and ions must occur in the hind-gut. A preliminary ultrastructural examination of the ileum and colon in this insect kindly carried out for me by Miss Y. Carter has shown that the hind-gut anterior to the rectum has a structure (Fig. 54) compatible with the idea that it may remove ions and water from its lumen. It may well be therefore that in many insects the hind-gut anterior to the rectum serves to reduce the volume of fluid finally delivered to the rectum. Its function would then in this respect be analagous to that of the proximal and distal convoluted tubules of the vertebrate

Fig. 54. Diagram to show the ultrastructure of the hind-gut anterior to the rectum in Thermobia.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

289

nephron. In some insects, Rhodnius among them, the tubules discharge straight into the rectum. In the case of Rhodnius, at least, a considerable amount of resorption still goes on before the fluid reaches the rectum, but here such resorption occurs at the lower ends of the Malpighian tubules (Wigglesworth, 1931; Ramsay, 1952; Maddrell, 1969). B. THE ACTION OF THE RECTUM

It is clear from the foregoing remarks that the study of the function of the hind-gut other than the rectum would be very rewarding-particularly i f it were done on an insect like the locust, blowfly or cockroach where the resorptive abilities of the rectum are already known. The functions of the rectum can be divided into the process of selective resorption-so that the fluid eliminated is appropriate for the insect’s needs-and the process of acidification. Rather little work has yet been done on acidification yet this function is an important one. As we have seen from the ability of Malpighian tubules to concentrate acid dyes the fluid leaving the tubules probably contains soluble organic anions at appreciable concentrations and it also contains uric acid also in anionic form. The acidification of the rectal contents will lead to precipitation of uric acid as a result of increase in the concentration of the un-ionized form. Other organic anions will associate with hydrogen ions and may also be precipitated. It is not yet known how rectal acidification is brought about, nor indeed how effective it is in quantitative terms in precipitating uric acid and thus reducing its osmotic effect. The function of the rectum in selective reabsorption is better understood. Even here, however, rather little is known of the recovery of amino acids, sugars and other useful compounds from the rectal contents. Still less is known of the mechanism of their uptake. Research has so far been concentrated on the reabsorption of ions and water to which we now turn our attention. A good deal of early work has suggested that the rectum of several terrestrial and brackish water insects is able to produce a final excretory fluid which is considerably hyper-osmotic to the haemolymph (Wigglesworth, 193la; Ramsay, 1950, 1952, 1955; Sutcliffe, 1960). The fluid which enters it from the hindgut or Malpighian tubules is usually at or near the osmotic concentration of the haemolymph. The elimination from the insect of a hyper-osmotic

290

S.

H. P. MADDRELL

final fluid then is certainly due to the activities of the rectum. However, this might be achieved either by secretion of solutes into the rectal fluid or by absorption of water from the rectal fluid or, of course, by a combination of these two processes. That a hyper-osmotic fluid is largely achieved by an absorption of water has been directly shown in two insects, the locust, Schistocercu greguria and the blowfly, Culliphoru erythrocephulu, by Phillips (1 961 , 1964a, b yc, 1969). Phillips’ work (1961 , 1969) has also shown that the blowfly can, under conditions of excess of water, produce a very hypo-osmotic fluid. The cockroach too is capable of producing a hypo-osmotic urine (Wall and Oschman, 1970). Phillips’ classic work has provided the basis and the stimulus for the work which has been done since on the ultrastructure of the rectal cells. In turn knowledge of the ultrastructure of these cells has allowed new hypotheses as to the mechanism of their action. The most recent research has provided evidence which has allowed us to discriminate between these hypotheses. We should begin our survey, however, by considering Phillips’ research on the function of the rectum in the locust and blowfly. The locust, unlike the blowfly, is not capable of producing a hypo-osmotic urine. Even when fed on tap water for 4-6 days the rectal fluid is still twice as concentrated as the haemolymph (Phillips, 1 9 6 4 ~ )the ~ concentrations of the main ions however being spectacularly reduced (to as little as 1 mmol . 1-I ). By contrast, after being fed on hyper-osmotic saline (1000 mosmol . l-’), the rectal fluid is nearly four times as concentrated as is the haemolymph, and is rich in the main ions. The details of Phillips’ analyses are shown in Table VII. The problem was to discover how these results are brought about by the rectum. Phillips developed the useful technique of isolating the rectum in the insect from the hind-gut by tying a ligature between them.* He then was able t o introduce fluids of various composition into the rectum and observe changes in the volume and composition of the fluid without contamination by material entering

* This technique with the rectum isolated from the rest of the hind-gut, yet retained within the living insect allows the rectum to retain its tracheal supply and to be bathed with haemolymph. These points may be very important in the success of the experiments. It is worth recalling that the midgut of Antheraea pernyi, another active insect ion-transporting tissue, requires fast circulation of oxygen-enriched saline over both luminal and haemolymph surfaces before it will operate at the maximum rate (Wood, 1971). In vitro preparations of very active insect tissues must be suspect unless it can be shown that oxygen supply is not limiting.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

29 1

TABLE VII The osmotic concentration and concentrations of ions in the haemolymph, hind-gut and rectal fluid of the locust under extreme conditions (figures from Phillips, 1964c) Osmotic

Ion concentration concentrations as melting ( mmoles . 1- 1 ) point depression (A") Na K c1 Insects fed on tap water

Haemolymph

0.74

108

11

115

Hind-gut

0.78

20

139

93

Rectal fluid

1.52

1

22

5

0.96

158

19

163

-

67

186

192

405

241

569

Insects fed hyper-osmo tic Haemolymph saline (NaCl 300 mmoles . 1-' Hind-gut KCl 150 mmoles . 1-' Mg(CH3COO)z Rectal fluid 30 mmoles .1-' Ca(N03)2 3 0 mmoles . l - ' )

3.47

the rectum from the hind-gut. Changes in volume were followed by including radio-iodinated ( 1 1 3 l) serum albumin in the solutions placed in the rectum, and measuring the number of counts/unit volume in fluid taken from the rectum. C. RECTAL ABSORPTION OF IONS AND WATER IN SCHISTOCERCA

An early key experiment was to show (Phillips, 1964a) that water was rapidly absorbed from the rectum when it contained xylose solution iso-osmotic to the haemolymph. The system is in equilibrium when the rectal fluid reaches an osmotic concentration between two and three times that of the haemolymph (Fig. 5 5 ) . The movement of fluid is not attributable to any hydrostatic pressure gradient (Phillips, 1964a). The hyper-osmotic rectal fluid is achieved almost entirely by absorption of water. This was shown by similar experiments using trehalose which is not absorbed from the lumen; it was found that very few ions were secreted into the lumen, the total

292

S. H. P. MADDRELL

40 r

I

Initial hyperosrnolarity of rectal contents (rnosrnol I-')

Fig. 55. The rate of water absorption from xylose solutions injected into the lumen of the locust rectum, as a function of the initial osmotic concentration difference across the rectal wall (redrawn from Phillips, 1970).

rectal concentration of monovalent ions not reaching more than S m m o l . 1-'. Plainly, absorption of water from the lumen can proceed in the absence of net solute absorption from the lumen and of large-scale influx of ions from the haemocoel. The rectal wall behaves as if there were an active movement of water from the lumen to the haemolymph in parallel with a passive leak, whose magnitude and direction depends, of course, on the osmotic gradient across the wall (Phillips, 1964a). Phillips suggests (1 965) that regulation of water reabsorption in response to the requirements of the insect might be achieved by altering the passive permeability of the rectal wall to water, possibly under the influence of a hormone (for which there is some evidence in other insects, see p. 304), rather than by altering the active movements of water. The mechanism of the active absorption of water was discussed by Phillips (1 964a) and he suggested several possibilities, among them water movements led by active ion transport within the rectal wall and water movements produced by a pump specific for water. Both were thermodynamically possible. New research has thrown a considerable amount of light on this subject and is dealt with on p. 296.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

293

Table VII suggests that, in the locust, ions can be absorbed from the rectum against very steep concentration gradients. Phillips confirmed this by following water and ion absorption from saline solutions made hyper-osmo tic with high concentrations of xylose. Under these conditions water entered the rectum and ions left it. The trans-wall potential difference was 20 mV with the haemolymph negative to the lumen. Therefore chloride ions move against their electrochemical potential gradient. Sodium and potassium ions must also do so from the fact that the concentration gradient against which they move is much steeper than can be supported by the potential gradient. Solvent drag can clearly be excluded because these ion movements occur in the face of water movement in the opposite direction. Since there is still the possibility of frictional interaction between solute molecules as they move out of the rectum (solute drag), one cannot say that all three ions are actively transported, though this is certainly a possibility. The main difference between water-fed and saline-fed insects is the ability of recta of the former to absorb ions at high rates from high concentrations in the lumen. The recta of saline-fed insects absorb ions at a much lower rate under the same conditions. This appears to be due to a difference in the capacity of the uptake mechanism under the different conditions. Figure 56 shows the relevant figures for the main monovalent cations. Clearly the absorptive ability of the recta of saline-fed locusts becomes saturated at luminal concentrations of higher than about 125 mmol .1-' of sodium or chloride ions and about 20 mmol . 1-' of potassium ions. Water-fed insects exhibit no such saturation of the absorbing mechanism at the concentrations used. There was some evidence from Phillips' experiments that the additional absorption of chloride ions, at least, in water-fed animals, can be attributed to passive diffusion (Phillips, 1964b). However this is very difficult to equate with the very low concentrations of ions found in the recta of water-fed insects (Phillips, 1 9 6 4 ~because )~ with passive diffusion increased, one would expect ions to invade the rectal lumen from the haemolymph or the rectal cells. It may well be then that it is the active transport systems which become saturated in saline-fed insects, but they have a higher capacity in water-fed insects. Finally, it should be noted that the rectum can absorb potassium ions about ten times faster than it can sodium or chloride ions. This neatly compensates for the relatively higher concentrations of potassium ions in the fluid derived from the Malpighian tubules. It is

294

S. H. P. MADDRELL

K

lmects fed tap water

Insects fed hypertmic saline

4

I

100

200

0.9-

1

300

I

400

I 500

0.5 -

04 -

03No I

200

I

300

I

400

5bO

Concentration in the luminal solution (mr0l.l")

Fig. 56. Rate of absorption of ions from recta of locusts previously fed either tap water or hypertonic saline. Note that in each case the uptake is non-saturable in insects fed tap water but is saturated at relatively low concentrations in insects fed hypertonic saline. The similarity in the uptake of sodium and chloride ions is not coincidental; not unnaturally, chloride ions are taken up at the same rate as the cation in the luminal solution. Thus with a potassium chloride solution in the lumen both potassium and chloride ions are taken up at the higher rates characteristic for potassium ion (redrawn from Phillips, 1964b).

not so well matched with the high concentrations of chloride ions it.1 the fluid secreted by the tubules. However absorption rates for chloride were determined from injected sodium chloride solutions. Chloride ions are absorbed several times faster from potassium chloride solutions (Phillips, 1964b). The explanation may well be that with sodium chloride in the lumen, chloride ions cannot be removed faster than can sodium ions, for reasons of maintenance of electrical neutrality in the transported material. Since potassium ions

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

295

can be removed faster than sodium ions and chloride ions must keep pace with this, their absorption must o c c u much faster in this case. Phillips showed that the trans-wall potential difference was sensitive to changes in the luminal potassium concentration but not to similar changes involving sodium. It is likely then that the lumen-facing cell membrane is more permeable to potassium than sodium (cf. the similar situation in Malpighian tubules where the membrane facing the side from which transport occurs is also potassium selective, p. 229). This may be the basis for the preferential handling of potassium ions by the rectum. D . RECTAL ABSORPTION OF IONS AND WATER IN CALLIPHORA

The technique used by Phillips for his investigation of absorption by the rectum of the locust, he has also used to study the function of the blowfly rectum. As was pointed out on p.290, the main difference between the two is the ability of the blowfly to produce a final urine which is strongly hypo-osmotic. However, recta isolated from the rest of the hindgut in water-fed flies still absorbed water from fluids introduced into the rectum so that after 18 h, the fluid in the lumen was more than 40% more concentrated than the haemolymph. Perhaps then the rectum plays little part in the production of hypo-osmotic urine. However, as Phillips (1 969) points out, the technique for isolating the rectum from the rest of the gut involves relatively extensive incisions of the body wall; it would not be surprising if the fly were to respond to such damage by at once attempting to conserve water. These factors make it impossible to say how a hypo-osmotic urine is achieved. Flies deprived of water for two days could be used to show a rapid concentration of initially iso-osmotic saline injected into their recta (which as before were isolated from the length of the hind-gut anterior to it). As with the locust, Phillips was able to show that this concentration was achieved by water absorption against an increasing osmotic gradient rather than by solute secretion into the lumen. It was also shown (Phillips, 1969) that water absorption could occur from sugar solutions. These studies taken with those from the locust agree in showing that water absorption from the rectal lumen can occur against increasing osmotic gradients. It is clear also that secretion of solutes into the lumen is not an important factor in the formation of hyper-osmotic urine. Finally, since water absorption can occur from

296

S. H. P. MADDRELL

initially pure sugar solutions with scarcely any ion movements into the lumen, it is apparent that this absorption does not require simultaneous solute absorption. And in fact the maximum osmotic gradient developed across the rectal wall is not significantly smaller when the lumen contains sugar solution than it is when the lumen contains a saline solution (Phillips, 1964a). It is with the explanation of these central facts about water absorption that recent research has been concerned. E. THE MECHANISM OF WATER ABSORPTION BY THE RECTUM

Phillips’ own suggestions as to how water absorption might be brought about were made before the ultrastructure was known. He suggested (Phillips, 1965) that water movement might be a result of “active solute transport and back diffusion across individual membranes so that net flux of solute across the whole rectum is not involved”. He also pointed out that a carrier that specifically combines with and transports water molecules could not be excluded on thermodynamic grounds. He concluded (Phillips, 1965) that “the nature of water transport remains a matter for speculation and experimentation”. What followed were investigations of the ultrastructure of the rectum. The first of these was an examination of the rectal papillae of the blowfly (Gupta and Berridge, 1966a, b; Berridge and Gupta, 1967, 1968). This study revealed the rectal papillae as very much more complex than a single layer of simple cells. In fact, as Gupta and Berridge point out, the structural complexity of these papillae had been recognized very much earlier in a painstaking and accurate light microscope study by Graham-Smith (1 934). The main details of the structure of the rectal papilla are illustrated in Fig. 57. The main points to be noted are as follows: (i) The cell membrane facing the lumen is infolded somewhat so as to increase the surface area by about 10-20 times. Such infoldings, as with similar ones in Malpighian tubules (p. 270) would tend to link any solute movements across this membrane with a Fig. 57. Diagram to show the ultrastructure of the rectal papilla of CulZiphoru and the postulated ion movements underlying the absorption of water from the lumen. The solid continuous arrows represent active movements of solute, the solid broken arrows passive movements of water and the open continuous arrows the flow of fluid in the extracellular spaces (based on Gupta and Berridge, 1966b).

8

4

298

S.

H. P. MADDRELL

water flow in the same direction. The cytoplasmic side of this membrane is adorned with a dense regular array of particles. (ii) Between the large cortical cells is a wide ranging system of intercellular spaces and sinuses which open out to the haemolymph through a one-way valve as shown by Graham-Smith (1934). (iii) The lateral walls of the cortical cells are developed and extensively folded to form large numbers of stacks of membranes closely associated with large mitochondria in the adjacent cytoplasm. (iv) The intercellular sinuses lead into infundibular channels which run in contact with the basal membranes of the cortical rectal cells before debouching into the haemocoel. The basal cell membranes are not infolded. (v) The cortical cells at both luminal and basal edges are joined by pronounced tight junctions and septate desmosomes. This must isolate the lumen, intercellular spaces and the infundibular space from direct communication with each other between the rectal cells. The next valuable piece of information came from the work of Berridge and Gupta (1967) who examined the ultrastructure of the epithelial (cortical) cells of the rectal papillae of the blowfly under various physiological conditions. Basing their work on Phillips’ findings, they injected hypo-osmotic fluid into the recta of starved flies to produce maximum water absorption. They also examined the recta of similar flies injected with iso-osmotic and hyper-osmotic fluids. Also based on Berridge’s ( 1966b) findings that the excretory system of female flies varies in activity during the cycle of development of the ovaries, they were able to examine the recta of flies in various natural physiological states. Very briefly their results can be summarized as showing that both in injected and untreated insects the intercellular spaces are dilated to an extent directly related to the inferred rate of water absorption. They are grossly distended in starved flies injected with hypo-osmotic media, are somewhat less dilated under natural absorbing conditions but are completely collapsed in naturally fasting or artificially starved flies. It is a strong point in favour of the consistency of these results that they rest on the examination of 76 papillae each taken from a different fly, each randomly selected from a group treated in one of the ways outlined above. More than 1000 micrographs were examined. From these ultrastructural findings it is reasonable to suppose that during absorption of water from the lumen of the rectum, it passes

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

299

from the cells into the intercellular spaces, then through the infundibular spaces into the haemocoel. The examination of the locust rectum by Irvine (1 966-quoted in Phillips, 1970) showed that the locust rectum was very similarly organized to that of the blowfly. There are infolded luminal cell membranes, extensively folded lateral cell membranes and associated mitochondria, and a complex of intercellular spaces leading to a subepithelial sinus draining through a limited number of channels into the haemocoel. Further ultrastructural studies of insect recta have been undertaken by Noirot and Noirot-Timothke (1 966), Hopkins (1 967) and Oschman and Wall (1969) on a termite, the mosquito, Aedes aegypti, and the cockroach, Periplaneta americana, respectively. They reveal a structure essentially similar to that described for the blowfly, though of course there are differences in detail. It is now feasible to examine the suggested mechanisms for water absorption from the rectum in these insects. Essentially the argument is to decide whether or not it is necessary to invoke a pump specific for water or not. From the facts that water absorption can occur against increasing osmotic gradients without a net absorption of solutes and with very little recruitment of ions from the haemolymph (Phillips, 1969; Stobbart, 1968), the idea of a water pump has attractions. However it has been the lesson of history that, as new facts have emerged, many proposals for water pumpsprimary water transport (Kedem, 1965)-have had to be exchanged for schemes in which water follows solute movement-secondary transport of water (Kedem, 1965). It therefore seems prudent to attempt to explain rectal water absorption by a scheme involving primary solute transport, bearing in mind the strictures outlined above. The structure of the rectal cells suggests a prima facie case for primary solute transport leading to water absorption. If this were not the case, the effective isolation of the intercellular spaces from the haemolymph and lumen contents would have no obvious function. A scheme in which water absorption is primarily produced by solute transport is shown in Fig. 58. It proposes that solute transport occurs across the folded regions of the lateral cell wall. Because of the geometry and dimensions of this region, water follows this transport in iso-osmotic proportions. The fluid once in the large intercellular spaces or the subepithelial spaces, is subjected to ion reabsorption. In this case, water tends not to follow because extreme local osmotic gradients cannot develop as the membranes facing

300

S.

H. P. MADDRELL

lntercellulai channel

Fig. 5 8 . A hypothetical scheme to explain water absorption from the rectum. The drawing is based on that of Oschman and Wall (1969) of Periplaneta, but the general scheme would apply as well to Calliphora and Schistocerca. The solid continuous arrows refer to active movements of solutes, the solid broken arrows to passive movements of water and the open continuous arrows to fluid movements in the extracellular spaces.

these spaces are relatively plane and the spaces have a comparatively low surface area/volume ratio. It follows that fluid in the subepithelial spaces is very hypo-osmotic to the lumen contents and that in the intercellular spaces is either hyper-osmotic to the lumen or less concentrated depending on the extent to which ion reabsorption occurs here. Various other mechanisms are possible, but as emerges from a careful examination of them by Phillips (1970), apart from one involving primary water transport, they can be excluded or they suffer from the necessity of involving ions from the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

301

lumen or the haemolymph which as we have seen are not likely to play a major role. If the mechanism is that shown in Fig. 58, then it follows that: (1) the osmotic pressure of the cells must be greater than that of the lumen contents so that water enters the cell by osmosis from the lumen; (2) the fluid in the intercellular spaces may well also be hyper-osmotic to the lumen contents but will certainly be much more concentrated than the fluid in the subepithelial space; (3) the fluid in the subepithelial space must be hypo-osmotic to the lumen contents. (1) and (3) would also follow from a system in which primary water transport was responsible for water absorption. It is the combination of (2) and (3) which points to water absorption being secondary to solute movements. Work on the cockroach by Wall and Oschman has provided the information necessary to check some of these points. The first discovery was that the fluid in the subepithelial space is hypo-osmotic to the fluid in the lumen in dehydrated cockroaches (Wall and Oschman, 1970). It was also found that in cockroaches provided with water, but no food, this subepithelial fluid was hyper-osmotic to the luminal fluid, indicating a faster net absorption of solute than water. These facts can indeed be explained in terms of a mechanism involving solute transport first into and then out from the fluid initially produced in the lateral intercellular spaces. However, they do not of themselves exclude the alternative of primary water transport, with ion transport either failing to keep pace (dehydrated animals), or occurring at a higher rate (hydrated animals). The crucial evidence that enables one confidently to assert that water is secondary to solute transport comes from the work of Wall et al. (1 970). These workers were able by micropuncture to collect fluid from the intercellular spaces in the rectal pads of the cockroach during water-uptake. The animals used were adult male cockroaches which had been kept without water for 4-7 days. The animals were operated upon to expose the rectum and allow the rectal contents to be flushed through into the colon by means of a syringe and cannula. The fluid injected was either iso-osmotic saline or a hyper-osmotic solution of trehalose which contained in addition about 60 mmol . 1-' of sodium chloride. The junction between colon and

302

S. H. P. MADDRELL

rectum was then ligated. By including Nigrosine dye in the injected fluid, the intercellular spaces could be seen clearly against the dark background of the contents of the lumen. 0.025-0.1 nl samples were then removed from the intercellular spaces by micropuncture (using capillaries with bevelled tips of 4-5 j,un outer diameter). All such samples proved to have osmotic concentrations higher, some of them considerably so, than that of the contents of the rectal lumen (Fig. 59). It is this observation which is difficult to reconcile with 0

/

Osmotic concentration of rectal contents ( mosml. I-'

Fig. 5 9 . The osmotic concentrations of samples of fluid removed from the intercellular spaces of the rectal pads of Periplaneta compared with the osmotic concentrations of the rectal contents (data from Wall e f al., 1970). The line is that of an iso-osmotic relationship between the two fluids.

primary water transport, especially since it is known that the subepithelial fluid during such uptake is hypo-osmotic. Solutes clearly are present in sufficient quantity to make the intercellular fluid hyper-osmotic to the contents of the rectum, but during its passage through the subepithelial spaces the concentration becomes very much reduced. While it is still possible that this reduction in concentration is achieved by water transport into the fluid, it seems more likely that it is brought about by ion resorption. * This new work on the cockroach rectum now makes it reasonable

* Finally to establish that ions move out of this fluid would require, ironically enough, a repetition in miniature of Phillips' pioneering work on the rectum as a whole. To follow volume changes as the extracellular fluid becomes more dilute requires the introduction into the intercellular spaces of some volume marker such as radio-iodinated serum albumin. The difficulties of this make remote the prospect of this experiment being done.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

303

to assume that water moves from the rectal cells into the intercellular spaces in a purely passive manner following solute movements in the same direction, the solute being resorbed in another part of the cell. Wall et al. (1970) were also able to show by direct observation that the intercellular spaces of the rectum distend rapidly at the onset of water absorption and remain so during the time that water-uptake occurs. This is in spite of their inability to find such a correlation from examining fixed and sectioned material. Their direct observation is a convincing proof of what was shown to be very probable by Berridge and Gupta (1967), who, as we have seen, were able to find in fixed material from blowflies a relation between the size of the intercellular spaces and the rate of water absorption prior to fixation. One can conclude that the route and the mechanism whereby water leaves the rectal cells and reaches the haemolymph is known. However there is still~theproblem of how water enters the rectal cells from the luminal contents, The model scheme shown in Fig. 58 proposes that the osmotic concentration of the cells is higher than that of the contents of the lumen so that water is removed from the lumen passively by osmosis. The difficulty here is that the only measurements made of concentrations of monovalent ions in the rectal tissue are for the locust and they show that the concentration of these ions is lower in the rectal tissue than in the haemolymph (Phillips, 1964b; Stobbart, 1968). If the model scheme is to survive then it follows that there must be high concentrations of other solutes in the cells or a reduction in the activity of water within the cell. Clearly there is scope for research into this problem. F. THE MECHANISM OF ION ABSORPTION BY THE RECTUM

By contrast with the active research on water movements, rectal resorption of ions has not been followed beyond Phillips’ work on the locust and blowfly. This work showed that ions could be removed from the lumen against steep gradients and that in the locust, at least, the rate of ion uptake was reduced in insects fed on hyper-osmotic saline. Very little is known of how this change is brought about or of the sites of the various ion pumps which must be involved. It is a distinct possibility that the reabsorptive activities of the rectum are regulated by hormones released from neurosecretory axon endings. Such endings are found in profusion in the blowfly in the medulla near the rectal cells (Gupta and Berridge, 1966b) and

3 04

S. H. P. MADDRELL

similar nerve endings have also been found in the rectal pads of the cockroach, Periplunetu (Oschman and Wall, 1969). In the cockroach, these nerves come from the last abdominal ganglion, extracts of which can increase water-uptake by isolated recta (Wall, 1967). G. RECTAL RECOVERY OF AMINO ACIDS, SUGARS AND OTHER SMALL ORGANIC MOLECULES

It is a consequence of the functioning of the Malpighian tubules that many substances of low molecular weight appear at appreciable concentrations in the fluid passed into the hind-gut. Since such substances as amino acids and sugars are virtually absent from the matter which is finally eliminated from the insect, they must presumably be reabsorbed by the hind-gut. Only very recently have any attempts been made to investigate this process. Wall and Oschman (1 970) in their work on the operation of the rectum in the cockroach, Periplunetu, took the opportunity of measuring the concentration of free amino acids in the colon, rectal lumen, subepithelial sinus and haemolymph. From their measurements, it was clear that the concentration of amino acids in the rectal lumen was lower than that to be found in the hind-gut anterior to the rectum, and lower than that in the haemolymph. Significantly the fluid in the subepithelial space was rich in amino acids. It seems very likely that amino acids are actively transported out of the rectal lumen and that they follow the same route as resorbed water via the intercellular and subepithelial spaces into the haemolymph. A good deal of research remains to be done in this field to follow up Wall and Oschman’s pioneering investigation. H. THE ROLE OF THE CUTICULAR LINING OF THE RECTUM

The hindgut is lined with cuticle of appreciable thickness. Since this cuticle lies between the contents of the lumen and the epithelial cells which modify the contents, its permeability properties are of great interest. Apart from this, the rectal cuticle may well also act to prevent mechanical damage t o the epithelial cells from abrasive materials in the luminal contents. In a recent paper, Phillips and Dockrill ( 1 968) describe their work on the isolated cuticular lining of the locust rectum. They prepared the lining as a sac by destroying the rectal cells by immersion in tap water for some hours. As a result, the tissue surrounding the

305

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

cuticular lining could easily be pulled off with forceps. Such sacs of cuticle could be perfused with a saline resembling rectal fluid in its ionic composition and containing the radioactive material under test. From the rate of appearance of isotope in the bathing solution, the one-way flux of the test substance could be calculated. The flux figures in turn were used to calculate the permeability coefficient of the cuticle to the test molecule. The results show that the cuticle selectively restricts permeation; only small molecules penetrate with any ease. The lining is more or less impermeable t o molecules with a radius of more than 5-6 e.g., glucose with an estimated hydrated radius of 4.2 A penetrates readily but trehalose with an estimated hydrated radius of 5.2 has a permeability coefficient more than 30 times smaller. What is the function of this permeability barrier? The rectum as a whole selectively resorbs many of the small molecules which are contained in the fluid derived from the Malpighian tubules. Plainly it is important that the cuticle lining the rectum should allow these molecules easy access to the active rectal cells. The Malpihan tubules however also actively secrete dye molecules which, as was suggested on p. 284, probably reflects their important ability to remove from the haemolymph those toxic organic acids which are by-products of metabolism. The concentration of these substances in the Malpighian tubules, while appreciable, may not be high enough to have toxic effects there. However in the rectal lumen, as Phillips’ work has shown, water is absorbed at a high rate and any toxins there present are very rapidly concentrated. Now the well being of the insect depends vitally on the ability of the rectal cells to reabsorb useful molecules from the rectal contents. If a high concentration of toxic substances was allowed to come into contact with the rectal cells, they would certainly not be able to continue operating. Therefore, the ability of the rectal lining to prevent penetration by molecules of the size actively secreted by the Malpighian tubules may well be its most important role. It is worth pointing out that it is important that the permeability properties of the Malpighian tubules and of the cuticle lining the hind-gut are well matched. If the tubules were to allow larger molecules to diffuse into the secreted fluid than can cross the lining of the hind-gut, then reabsorption could not take place and these molecules would be lost from the insect. This is probably why Malpighian tubules have developed secretory mechanisms for the larger toxic molecules rather than allow such large molecules to enter by diffusion. If this is the case then one

a,

AIP-13

306

S. H. P. MADDRELL

must modify the attractive idea, first put forward by Ramsay (1958), that excretion in insects is basically an unselective collecting of substances into a primary excretory fluid followed by reabsorption of useful molecules. This may be true for substances of low molecular weight but clearly cannot be true for larger molecules. It is not the case then that all toxic molecules are automatically excreted,

a

1

Large particles Bacteria Protozoan EL metazoan parasites EL parasitcids

Haemocytes

Pencardial cells and Nephrocytes

Large molecules Colloidal particles

Medium-sized molecules Small molecules

- k-

c

Malpighian tubule

+ - / my l l

1

Selective reatsorption of smoll molecules

?

-

Midgut

P

I

? Selective excretion of small molecules

Fig. 60. A scheme to show the different ways in which undesirable materials are removed from the haemolymph of insects by the activity of the excretory system.

for if they are larger than can enter the Malpighian tubules by diffusion, they will only be removed if there exists a specific mechanism for their secretion into the lumen. Figure 60 summarizes these ideas by showing the fate of molecules of various sizes which the insect wishes t o excrete. Only small toxic molecules are eliminated simply by not being reabsorbed. Medium-sized toxic molecules are eliminated by active secretion into the lumen of the Malpighian tubules; when they reach the rectum they cannot be returned to the haemolymph nor affect the rectal cells because they cannot penetrate the cuticular lining. Very large molecules and colloidal particles may be too big even to cross the basement membrane of the Malpighian tubules (see p. 269) and are removed from circulation by the activities of the pericardial cells (p. 205). Gross particles in the haemolymph are engulfed by haemocytes.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

307

While it is important for the design of the excretory system that the permeability properties of the tubules and rectal lining are matched, there has as yet been no investigation of both these parameters in one insect. The permeability of stick insect tubules to some organic substances of low molecular weight is known (p. 280) and as we have just seen the permeability of the locust rectal lining has been explored. These two investigations do not tally very well for the Malpighian tubules of Gzrausius are freely permeable to sucrose to which substance the rectal lining of Schistocerca is almost impermeable. Perhaps the cuticular lining of the rectum in Gzrausius is more permeable and the Malpighian tubules of Schistocercu less permeable. There is however a further possibility. So far only the cuticular lining of the rectum has been investigated. It is possible that the cuticular lining of the hind-gut anterior to the rectum is more freely permeable and that slightly larger useful molecules can be resorbed there. It is clear that here is a field which would profit from further research. It is also time that the permeability properties of Malpighian tubules was further explored. In Ramsay's original experiments (1 958), only substances up to the size of sucrose were examined. All these substances were able to cross the tubule wall with some ease. It would be instructive to find out how permeable the tubule wall is to a whole range of substances of larger size so that an accurate idea of what sorts of substances are likely to appear and at what concentrations in the fluid passed by the Malpiaan tubules to the hind-gut.

I. ABSORPTION OF WATER VAPOUR FROM SUBSATURATED ATMOSPHERES BY THERMOBIA

Thermo bia domestica (also known as Lepismodes inquilinus) (Watson, 1967) has the remarkable ability to take up water from subyturated atmospheres down to 45% relative humidity at 21°C (Beament et al., 1964). Several other insects also have this ability but to a less marked extent (see Noble-Nesbitt, 1969 for references). Noble-Nesbitt (1969) has shown that desiccated Thermobia weighmg 30 mg can take up water at a rate of about 1-5 pg . min-' depending on the conditions. The insect, once it recovers its initial level of body-water content, no longer absorbs water so that the process is a truly regulative one. Noble-Nesbitt (1 970) has now reported the astonishing finding AIP-13.

308

S. H. P. MADDRELL

that this impressive uptake of water can be entirely prevented by occlusion of the anus. He suggests that this may well mean that the rectum is the site of water-uptake. As support for this suggestion he cites the work of Grimstone et al. (1 968) who showed that Tenebrio can equilibrate air in the rectum to 88% relative humidity and that this humidity is the lowest from which the animal can take up water. As might be expected from these findings, Noble-Nesbitt's preliminary results indicate that blocking the anus in Tenebrio prevents water uptake. As is shown below, the rate of water uptake in Thermobia is so fast that it could scarcely occur at sites deeper within the animal than the rectum. From these considerations it seems probable that the rectum is indeed the site of uptake. One suggestion as to how the system might work is to suppose that air might be taken in by dilation of the hind-gut and then after closure of the anus, pressure might be applied by the muscles of the hind-gut and abdomen. This would reduce the volume of the air and so increase its relative humidity which would make the absorption of water easier. To make the point, suppose it were possible to halve the volume of the contained air and that initially it was at 50% relative humidity. After compression it would be saturated with water vapour and water would start t o condense on the surrounding walls. Of course it is likely in fact that the pressure which could be exerted is much less than this but any increase in pressure would help, as would removal of oxygen by the respiratory activity of the surrounding tissues. Unfortunately such an idea although it is an attractive one, does not survive when one considers how often such a cycle of events would have to occur. Thermobia can take up water at a rate of 5 pg . min-I. Even saturated air at 20°C only contains 18 pg . cm-3 of water so that supposing the insect can remove half of this it must be able to absorb the water vapour from 0.55 cm3 . min-' . The rectum can contain about 100 nl of air so that if it is the site of uptake, it must fill and empty more than 80 times/s. Even if more of the gut is involved in absorption it is clear that any such cyclical process must occur at a rate which is too high to be envisaged without supposing the hind-gut musculature t o operate at similar frequencies t o and over a much greater change of length than that of the flight muscles of other insects! How then can air be supplied to the hind-gut fast enough for the uptake of water? Since it is clear that mechanical ventilation at any

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

309

ordinary rate is not sufficient, it does not immediately strike one that diffusion might be fast enough. In fact calculation shows that diffusion is capable of ensuring that water vapour enters the hind-gut at a high enough rate. This is another case where provided the distances involved are not too large, diffusion is capable of very high rates of transfer. I am most grateful to Professor T. Weis-Fogh who did the calculation for me. In essence the calculation shows that diffusion can supply water vapour at a sufficiently high rate provided that the absorbing surfaces are not deeper within the animal than about 1 mm. Since the rectum in Thermobiu is 1.5 mm long, it is clear that the rest of the hind-gut, for example, is too deep within the animal to be supplied by diffusion at a high enough rate. It is on this basis that one concludes that the rectum is the site of uptake of water vapour. In a sense it is logical that the rectum, which as we have seen earlier is able to reabsorb water against steep osmotic gradients, should be the site of water-uptake from the air. However in this latter case, the gradient against which water uptake occurs is extraordinarily steep. This will be clear from the fact that, for example, 7 M NH4C1 solution is in equilibrium with 79.5% relative humidity at 20" C. It is difficult at the moment to make any definite suggestion as to how the rectal cells are able to create conditions within themselves so that water in the air at 45% relative humidity can move thermodynamically downhill into them. To do this they must reduce the activity of the water within themselves by more than half. If this is done by some substance in solution then it can be predicted that this substance must be extremely soluble in water and have as well in all probability a large proportion of sites at which hydrogen bonding can occur and reduce the thermodynamic activity of water. It is doubtful if any non-toxic, naturally occurring inorganic molecule would have a sufficient effect. It seems likely that some organic molecule might be involved in which case one can suggest that it should have a low melting point, be not easily crystallized and have a high proportion of hydrophilic groups to give it a high affinity and capacity for water. A further possibility is that the luminal membrane of the rectal cells might contain hydrophilic macromolecules which when hydrated might undergo a conformational change to bring in the adherent water which might then be removed because of the different intracellular conditions (perhaps an ionic or pH difference).

310

S. H. P. MADDRELL

Such a system would of course be energy-requiring and would constitute a primary water pump. The energy-consumption of the water-uptake is of interest to enable one t o form an idea of how metabolically expensive the process is. Water is moved from air at say 50% relative humidity to a cell in equilibrium with air at close to 100% relative humidity-that is against a concentration gradient for water of 2. The minimum work involved is given by the formula W = R T log, !.! a2

where W = work/mole of water moved, R =gas constant, T is the temperature in OK, and a , and u2 are the activities of water in the cells and rectal lumen respectively. From this it follows that approximately 410 calories are required to secure the uptake of 1 mole of water. Since 1-5 pg . min-' of water can be absorbed, the minimum work required for this can be calculated t o be 0.0013-0.0066 cal . h-' . If this energy is derived from glucose and the process is, say, 10% efficient, then glucose would be consumed at no more than 18 pg . h-' or less than 0.07% of the body weight per hour which is much less than is consumed by flying insects, which depending on the substrate, consume fuel to the extent of from 0.6 to 20% of the body weight per hour (Weis-Fogh, 1967). It is also interesting to calculate the energy consumption of the rectal tissue. The rectum is a cylindrical structure, 1.5 mm long and 0.4 mm in diameter. Even assuming that this is 100% tissue, its weight can be no more than 190 pg. If this amount of tissue consumes 0.00 13-0.0066 cal . h- l , this is equal to 7-35 cal . g-' . h-' , even assuming 100% efficiency. The actual figure must be several times higher than this which makes the rectum of Thermobiu one of the most active tissues known other than very active flight muscles (Weis-Fogh, 1967). J. ABSORPTION OF WATER VAPOUR FROM SUBSATURATED ATMOSPHERE BY TENEBRIO

We have so far been concerned with the Malpighian tubules and rectum of insects separately, and this is justifiable in most cases because they act sequentially on the excretory fluid. However, a few insects (some Coleopteru and the larvae of some Lepidopteru) possess

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

311

a most unusual arrangement of the excretory system, with the upper lengths of the Malpighian tubules held closely to the surface of the rectum by an enveloping membrane (for a review, see Saini, 1964). In these cases then, the action of the Malpighian tubules and rectum may affect each other in a much more direct fashion. So far in only one case has the action of such a system been examined in detail; this is the cryptonephridial system of the mealworm, Tenebrio molitor, which is a pest of stored food products. This insect is able to take up water from subsaturated atmospheres down to 88%relative humidity (Mellanby, 1932). Coupled with this is its ability to remove so much water from the faecal pellets that they then can take up water vapour from air at an average of about 90% relative humidity with occasional pellets so dehydrated that they equilibrate with air at 75% relative humidity. These facts together with the results of Noble-Nesbitt’s work on water uptake by Therrnobia and his preliminary experiments on Tenebrio, strongly suggest that the ability of Tenebrio t o dry its faecal pellets is a reflection of water-uptake from the atmosphere through its rectum. This being the case, Ramsay’s results (Ramsay, 1964; Grimstone et al., 1968) are of even more interest because the system not only keeps loss of water via the excretory system to a very low level, but it is the means whereby the insect gains water from air of humidities of 88%relative humidity and above. The structure of the cryptonephridial system of Tenebrio is shown in Fig. 6 1 (based on the work of Grimstone et al., 1968). The values for concentration of potassium, sodium and chloride ions and for osmotic concentration as freezing-point depression are indicated for an insect taken from a culture maintained under dry conditions and which was presumably, therefore, attempting to absorb water in the rectum at a high rate (Ramsay, 1964, Table 111, serial 6). The points of structure which appear significant are as follows: (1) The perinephric membrane which invests the Malpighian tubule/rectum complex is, over most of its area, very thick and composed of many (up t o 40) layers of cells so extremely flattened that each layer consists of little more than two plasma membranes separated by a layer of cytoplasm about 100 A thick. (2) At regular points, called leptophragmata, the perinephric membrane is completely perforated so that the underlying Malpighian tubules are separated from the haemolymph only by two extracellular basement membranes which are probably very permeable .

312

S. H. P. MADDRELL

Cut ends of Malpighian tubules

Perinephric membrane

~

Rectum

t.

1' 1

Na 84 K- 78

1070 3doC

134

Haemolymph

3.O0C Na 2 3 K 132 I . 5 T CI 73

I.4'C

JF] K

Perirectal space

-

Rectal lumen

4

2

<

9.0"C

I

<

IO.5OC

-

Fig. 61. The cryptonephridial arrangement of Malpighian tubules and rectum in larvae of Tenebrio (fQures based on those of Grimstone et a l , 1963). (a) The appearance of the cryptonephridial system as a whole. For clarity the tubules are shown less convoluted and with fewer boursouflures (swellings) than in life. (b) Diagrammatic longitudinal section through the cryptonephridial system to show the functionally important compartments and the concentrations of ions (in mmol . I-') and melting-point depressions of fluids in them. (c) Transverse section of the cryptonephridial system through the posterior region. (d) Details of the ultrastructure of a leptophragma. Note the relationship of the leptophragma cell with the inner part of the perinephric membrane and with the more typical tubule cells.

(3) Just at the points where the perinephric membrane is lacking, the Malpighian tubule wall becomes extremely thin. A specialized cell, the leptophragma cell, comprises the wall in this area and it is attached laterally to both the perinephric membrane and the surrounding cells of the Malpiaan tubule wall. The leptophragma cell has a plane surface facing the haemocoel and long, thin, widely-separated microvilli radiating into the lumen of the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

Fig. 6 1-cont.

313

314

S. H. P. MADDRELL

Malpighian tubule. The cytoplasm of this cell contains many mitochondria. (4) The rectum, which lies within the structures described in (l), (2) and (3) above, is approximately hexagonal in cross-section. The rectal epithelium is made up of a single layer of cells which varies in thickness from about 10 pm at the angles up to about 70 pm between them. The cells contain large numbers of microtubules many of which run parallel to the lateral cell walls. There are many mitochondria and large numbers of ribosomes distributed throughout the cytoplasm. The apical and basal surfaces of the rectal cells are thrown into irregular folds as are portions of the lateral cell walls. Nowhere however, are there visible the regular arrays of mitochondria and folded lateral cell walls so characteristic of the recta of the blowfly, locust and cockroach (see p. 299). Any mechanism which attempts to account for the operation of the cryptonephridial system must take into account the concentrations of ions and osmotic concentrations in the various compartments, as indicated in Fig. 61. It must also account for the fact that what enters the rectum is a mush of undigested material suspended in fluid from the Malpighian tubules which is approximately iso-osmotic to the haemolymph. What leaves the rectum when the cryptonephridial system is active, is dried faecal material and, through the Malpiaan tubules, a solution of potassium and sodium chloride hyper-osmotic to the haemolymph. Clearly, extra potassium chloride ions without iso-osmotic amounts of water must enter the system to account for the fact that the fluid leaving the system is not hypo-osmotic as it is in say the rectum of the cockroach (see p. 301). The work of Ramsay (1964) and Grimstone et al. (1 968) strongly suggests that the potassium chloride is added to the system by the action of the leptophragma cells which probably possess pumps capable of actively transporting potassium from the haemolymph into the lumen of the Malpighian tubules so that chloride ions follow, but iso-osmotic amounts of water do not. In effect then, the system falls into two concentrating mechanisms in series. The rectal epithelium transports water from the rectal contents into the perinephric space against a large osmotic gradient by a mechanism which is at present unknown. The Malpighian tubules then transport water out of the perinephric space by using the osmotic capacity of potassium chloride derived from the haemolymph by the leptophragma cells. The evidence for transport of water by the rectal epithelium

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

315

(whether it be primary or secondary transport-see p. 299) lies in the osmotic gradient across it and the fact of removal of water from the lumen. The evidence for transport of potassium chloride into the tubules by the leptophragma cells is more extensive and comes from several lines of experimentation as outlined in the following paragraphs. Ramsay isolated the rectal complex from larvae of Tenebrio, placing it on a frame which could be rotated under liquid paraffin saturated with oxygen so as t o supply at least some oxygen to the tissues. A small measured volume of haemolymph or saline was then applied t o the rectal complex after any adherent haemolymph had been sucked away. Changes in the freezing-point depression and in the concentrations of sodium, potassium and chloride ions of the medium added were then followed. In brief, what was discovered was that the potassium concentration decreased and in several crucial experiments the osmotic concentration also decreased. If the external medium was a chloride-free nitrate-based solution then the chloride concentration rose dramatically but the potassium concentration fell. In the presence of poisons, both chloride and potassium concentrations rose. It can be concluded that the rectal complex has the ability to take up potassium and probably has a relatively high permeability t o chloride ions. One can now use the results of Lison's method for staining leptophragmata with externally applied silver nitrate, to show where the passive chloride movements are localized. Prolonged treatment with silver nitrate results in a dense precipitate of silver chloride being deposited on the outside of the leptophragmata, showing them to be permeable to chloride ions from within the system. Since the structure of the perinephric membrane, which occurs everywhere except over the leptophragmata, suggests that it is very impermeable, it seems very likely that the movements of potassium and chloride ions occur through the leptophragmata-that is into the lumen of the Malpighian tubules. If this is accepted, one can now ask whether the movements of potassium and chloride are active or passive. The potential difference between the lumen of the tubules and the haemolymph, Ramsay was able to measure as about SOmV, the lumen positive with respect t o the haemolymph. Since the potassium concentration within the tubules averages 500 mmol . I-' and can reach 2 mol . I - ' (Ramsay, 1964) it is clear that potassium movements are thermodynamically very much uphill and so are almost certainly active. Conversely, the chloride concentrations in

316

S. H. P. MADDRELL

the tubule which, although they are similar to the potassium concentrations, constitute a concentration gradient which is not as steep as the potential gradient in the opposite direction (Grimstone et al., 1968). Chloride movements are therefore downhill and unless future measurements show that these movements are too fast to be accounted for by the electrochemical gradient, it is probable that chloride ions move passively through the leptophragmata into the tubule. These findings make it reasonable to suppose that potassium chloride is transported into the tubules by the leptophragma cells. Since the leptophragma cells are Malpighian tubule cells, albeit highly unusual ones, this is not perhaps surprising. What makes the leptophragma cells so different from other Malpighian tubule cells is that the inward transport of potassium chloride is not accompanied by an osmotically compensating amount of water. It is this ability which is crucial to the operation of the system; without it the tubules would not be able to remove water from the perinephric chamber. The structure of the leptophragma cells is very different from that of the other Malpighian tubule cells (see Fig. 61). If the presence of deep basal infoldings and close-packed microvilli are structural indications of an ability to transport iso-osmotic fluid (see p. 271), then leptophragma cells with a plane basal cell membrane and widely separated apical microvilli are structurally suited to secrete ions without water following in iso-osmotic proportions. The evidence for this ability of the leptophragma cells lies in two types of findings. The first of these concerns the effects on the secretion by the Malpighian tubules of alterations in the external bathing medium-that is the fluid outside the perinephric membrane. If distilled water is added to this bathing medium, no obvious change results in the rate of secretion of the tubules and there is no rapid change in the osmotic concentration of the secreted fluid. Such a change would have extremely dramatic effects on other Malpighian tubules, of course, and significantly, if distilled water is injected under the perinephric membrane, the tubules of Tenebrio respond by secreting very fast for a short while. If, on the other hand, the potassium concentration of the external medium is increased, this does accelerate secretion by the Malpighian tubules in the cryptonephridial system. Both these findings are explicable if secretion is induced by potassium secretion from the external medium and the water to accompany it comes from that in the

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

317

perirectal space. On the other hand, although 3 mol . 1-' sucrose added to the bathing medium causes no large fall in the rate of secretion, the osmotic concentration of the secreted fluid is raised considerably. It may be that the leptophragma cells are not completely impermeable to water so that as they transport potassium and chloride ions into the tubules, a little water follows, though not in iso-osmotic amounts. The second finding supporting this last idea that the inward transport of potassium chloride is not accompanied by much water, comes from experiments in which the potassium and osmotic concentrations of fluid bathing isolated rectal complexes were followed. In some of these experiments, there was a substantial fall in the osmotic concentration of the bathing fluid accompanied by a fall in the potassium concentration. This finding of course supports the suggestion that the role of the leptophragma cells is that outlined above. That only in some experiments was such a result achieved, is probably due to the fact that isolated rectal complexes lack their tracheal supply and so may be expected to deteriorate rapidly. Even in successful experiments, the fall in osmotic concentration is not sustained for longer than 10-20 min. As pointed out earlier (p. 290), experiments with active tissues deprived of their tracheal supply are likely to perform much less actively than they do in vivo. Before we go on to consider what might be the point of the cryptonephridial system, there is one finding which is worth discussing. This is Ramsay's observation that the fluid bathing the Malpighian tubules under the perinephric membrane contains a substance which promotes the ability of the fluid to supercool-that is the fluid melts at a higher temperature than that at which it freezes. Just such a behaviour would be expected for a substance with many hydrophilic groups. It was predicted on p. 309 that the rectal cells of Thermobia may reduce the activity of water within themselves by using high concentrations of an organic substance with many hydrophilic groups so as to form hydrogen bonds with water molecules and so reduce their activity. The finding of just such a molecule in contact with the rectal epithelium of Tenebrio, which is known to be able to take up water vapour from air at 75% relative humidity, strongly suggests that this compound is centrally involved in water-uptake. The compound involved has a molecular weight of 10,000-12,000 (Grimstone et al., 1968). That the compound is a large one is perhaps not surprising from the design point of view if it is involved in water-uptake. It would be definitely advantageous for

318

S.

H. P. MADDRELL

the insect if a water absorbing compound once produced by the rectal cells could not easily diffuse away from the cells into the rectal lumen to be carried away in the faeces. Since the cuticular lining of the rectum is probably very impermeable to such large molecules (p. 305), it may well act to restrict to a thin layer under the cuticle any water absorbing substances the rectal cells produce. To carry the speculation one stage further, it could be that the apparent lack of the usual complex rectal cells found in the blowfly, cockroach, locust etc. (see pp. 296-300) is correlated with a different type of water-uptake. Similarly, the rectal cells of Thermobia although they are more complex than are those in Tenebrio, do not have the specialized foldings of the lateral cell walls which are found in say the blowfly. It is the apical cell walls in Thermobia which are highly folded and associated with mitochondria (Noble-Nesbitt, unpublished results). It could be, therefore, that those insects which take up water vapour from subsaturated air via the rectum, use a different mechanism than other insects and that this difference is reflected in the different ultrastructural appearance of the rectal cells. What is the point of the cryptonephridial system? One obvious suggestion is that the movement of water across the rectal epithelium and Malpighian tubule wall in series allows the concentration gradient against which each works to be smaller than otherwise. A further suggestion advanced by Ramsay (1964) is that the Malpighian tubules by removing ions from the perinephric space allow the rectal epithelium to have an ionic environment on this side of the cells which is similar to that of the haemolymph-otherwise the cell would have very “unnatural” media on both sides. An additional suggestion stems from Ramsay’s finding that the perinephric space contains high concentrations of a substance which it is here suggested may be active in water-uptake. If, as suggested above, a function of the cuticular lining of the rectum is to prevent the loss of this substance into the lumen of the rectum, then it may well be that the reason why the water crossing the rectal wall is passed to the haemolymph only via the Malpighian tubules is because in that way the water absorbing substance is not lost into the haemolymph, but is maintained in the region of the rectal cells. All these suggestions may contain some truth. By contrast, however, the rectal epithelium of Thermobia which can remove water from air at 45% relative humidity has no cryptonephridial

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

319

system of Malpighian tubules to back it up. We are left with the uneasy conclusion that while the cryptonephridial system may be one way of producing a very effective water absorbing organ, in other insects capable of even more impressive performance, such a complex system appears not to be needed. VI. CONCLUDING REMARKS

The things that can be usefully said in these concluding remarks include both a summary of the more important recent discoveries in the field and suggestions as to what areas might most profitably now receive attention. The storage of water in the rectum by larvae of Dysdercus towards the end of the feeding cycle in an instar (p. 202) is surprisingly the only case known of storage excretion in insects. Stores of water make such good sense for small terrestrial animals like insects, that one expects that further cases could be found. Deposit excretion (p. 203), in contrast, seems to be widespread. It appears that some insects have difficulty in eliminating uric acid from the body through the Malpighian tubules and rectum complex. This is particularly noticeable in cockroaches, and as a result they usually store uric acid in urate cells in the fat body, while some male cockroaches eliminate uric acid through their accessory glands. In other insects such as adult Pieris and all stages of Dysdercus, uric acid is used as a pigment. Parasitoids also retain uric acid in the fat body, presumably so as not to inconvenience the host more than necessary. It is one of the themes of this article that excretion involves tissues other than the Malpighian tubules and hind-gut. In particular, it seems that this system is not well suited for the excretion of substances of high molecular weight. The pericardial cells and nephrocytes (p. 205) play a substantial role in excretion of these larger molecules up to the size of the colloidal particles. They trap material probably by pinocytosis and after destructive digestion return the smaller products to the haemolymph. Particles which are larger still are engulfed by haemocytes, as are invading bacteria and even metazoan parasites. Several organs play a subsidiary role in the excretion of ions and water. The midgut of larvae of saturniid moths, by means of a powerful potassium pump, actively prevents the high concentrations

3 20

S.

H. P. MADDRELL

of potassium ions in the lumen from crossing the gut wall into the haemolymph (p. 206). The labial glands of adults of the same insects produce a large volume of alkaline fluid at emergence (p. 209). This fluid undoubtedly has a part in the enzymic softening of the cocoon by providing an appropriate pH for the action of the enzyme cocoonase, but it is quite possible that this is a secondary development. Its primary role may be that of jettisoning excess fluid which would hinder flight. Research should now be centred on those moths which spin no cocoons, to see if they also eliminate fluid through the labial glands at the beginning of adult life. The anal papillae of a salt-water mosquito may act to extrude ions when the animal is in a hyper-osmotic medium (p. 212). The evidence for this statement is not conclusive and it will be interesting to see if further research confirms the suggestion. Research on the secretory performance of Malpighian tubules has now broadened to include tubules of several different species and it is now possible to make generalizations about the way in which the “typical” tubule secretes fluid. Reduced to essentials, the primary fluid-secreting parts of typical Malpighian tubules actively transport potassium ions into the lumen. They behave as if they had in series a sodium dependent potassium entry on the basal side and on the luminal side, an electrogenic potassium pump (p. 23 1). Anions follow passively, the path of phosphate ions being very much smoothed by facilitated entry, probably passive in nature. Rhodnius Malpighian tubules, which secrete much faster than other tubules when hormonally stimulated but much slower otherwise, have a somewhat different ionic basis for secretion (p. 249). In this case, the transport of potassium ions although it is fast is not essential; two new elements, transport of sodium and chloride ions, occur so fast that the rate of fluid secretion does not alter in the absence of potassium ions. In both sets of tubules, the transport of ions gives rise to a flow of fluid which is explicable in terms of water passively following ion movements. The basal infoldings and apical microvilli are very probably the sites of standing osmotic gradients which link water and ion movements. More work is needed on the mathematical validity of this concept. In particular, the infoldings and microvilli are short, so any gradient will be small and could only function to couple ion and water movements if the osmotic permeability of the cell walls is very high. It is suggested that the close-packed array of backwards- and

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

321

forwards-facing membranes on both sides of the tubule cells may act to offset this difficulty by effectively increasing the gradients across the individual membranes. What is now needed is research on Malpighian tubules from other species to see if the generalizations which apply to the tubules already examined, are firmly based or not. Further than this, the function of lengths of Malpighian tubules, other than those where primary secretion goes on, is still largely obscure, although their function, one may suppose, is important. The role of the structures known as formed bodies in the operation of Malpighian tubules is controversial (p. 277). It is suggested here that they probably act to secure the excretion of some large molecules, but it is not impossible that they are more centrally involved in the mechanisms underlying the secretion of fluid. One difficulty is that they do not appear in electron micrographs of some secretory tissues, such as insect salivary glands. Since it is possible that the formed bodies here have a very short life, this tissue should be re-examined by the freeze-etching technique. Malpighian tubules secrete fluid containing many organic molecules, some of which enter the lumen passively, while others appear actively to be secreted there. The small molecules that cross the tubule wall passively (p. 280) occur at high concentrations in the fluid and this reflects a high permeability of the wall to these compounds. Probably the process is aided by frictional interactions of these substances with the other constitutents of the secreted fluid. This idea is open to test, since there are now known several different ways of controlling the rate at which tubules secrete fluid. Also, too little is known of the permeability of tubules to substances of molecular weight larger than about 500. Dye molecules are actively secreted by Malpighian tubules (p. 281). It is suggested that only anionic dye molecules are secreted and that this is merely a visualization of an important ongoing excretion of non-metabolizable aromatic acids. It will be interesting to see if competition exists between the various dyes and the similar non-coloured acids which may be excreted. The fluid produced by Malpiaan tubules is thought to be selectively acted upon in the hind-gut, particularly the rectum. In the concentration of research on the very important role of the rectum, the activities of the rest of the hind-gut in excretion has been neglected. One suspects, however, that the rest of the hind-gut may be important in resorption from and selective addition to the

3 22

S. H. P. MADDRELL

primary excretory fluid coming from the Malpighian tubules. Work on rectal water absorption from fluids in the lumen has reached an advanced stage. It seems clear that the active transport of water across the rectal epithelium is secondary in nature and is correlated with a cycling of solutes within the tissue (p. 303). It is probable that ions are secreted at high concentration into the fine ramifications of the lateral intercellular spaces and that water follows osmotically. As the hyper-osmotic fluid produced travels towards the haemolymph in the larger intercellular channels, ions are now reabsorbed so that the fluid which finally emerges into the haemo lymph is considerably hypo-osmotic. The way in which the rectum resorbs ions from the lumen is not yet so well understood. It is clear, however, that ions can be reabsorbed against steep electrochemical gradients (p. 291). It is possible that both ion and water absorption is under hormonal control. Research might well now be directed both towards a greater understanding of ion reabsorption and the ways in which this and water movements can be modified by hormonal action. Little is known about the activity of the rectum in recovering amino acids and other useful organic molecules from the rectal fluid. It seems probable that amino acids, at least, follow the same route as water; they are to be found in high concentrations in the fluid emerging from the intercellular spaces (p. 304). Nothing is yet known of whether the process involves active systems specific for these substances. The layer of cuticle lining the rectum seems to have the function of protecting the rectal epithelium both from mechanical damage by potentially abrasive faecal material and from disruption by toxic molecules which reach high levels of concentration in the rectal fluid (p. 305). The cuticular lining appears to deny such toxic molecules access to the rectal cells by acting as a molecular sieve; molecules of molecular weight higher than 300-500 penetrate the cuticle very slowly indeed. This research raises two problems. If medium-sized molecules cannot cross the rectal lining, they cannot be reabsorbed. As a result, one would expect that if useful molecules of this size are not automatically to be excreted by the excretory system, either they must be excluded from the fluid secreted by the Malpighian tubules or they must be reabsorbed in the rest of the hind-gut or in the lower lengths of the tubules themselves. If it turns out that they do not appear in the fluid produced by Malpighmn tubules, then the idea that excretion in insects effectively consists of an unselective

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

323

gathering of molecules from the haemolymph and the subsequent resorption of useful substances must be abandoned save for the smallest molecules. If on the other hand such substances are present in the fluid secreted by Malpighian tubules then their resorption must go on in the tubules or hind-gut other than the rectum. There is clearly scope for research into this field. Two remarkable and extreme cases of water-uptake in the rectum are known. Both involve the removal of water vapour from subsaturated atmospheres. It is clear that concentrations of inorganic ions cannot easily provide a basis for such a system and it is suggested that organic solutes are used to lower the water activity in or on the surface of the rectal cells so that water can be taken up from subsaturated air. The first of the two cases (p. 307) concerns the firebrat, Thermobia, which has the remarkable ability to absorb water from air at 45% RH (at 21°C). This absorption appears t o take place through the rectum. Absorption is so fast that water is removed from a volume of air equivalent to that of the rectal lumen every 10-15 ms. Diffusion is capable of supplying air to the rectum at this rate. A calculation of the work involved in this uptake of water reveals that the rectal epithelium must be one of the more active tissues known. Nothing is yet known of the way in which the water-uptake is achieved; the mechanism involved will obviously be of great interest and its elucidation poses a most intriguing challenge. A rather less dramatic, though still remarkable, case of water-uptake occurs in the cryptonephric arrangement of Malpighian tubules and rectum of Tenebrio. This insect commonly eliminates pellets so dehydrated that they equilibrate with air at 88% RH and some pellets equilibrate with air at 75% RH (p. 3 1 1). Here the system involves two mechanisms in series. The first is a concentrating mechanism which removes water from the rectal lumen into a perirectal space contained outside the rectum. How this water movement is achieved is no better known than it is in the similar case in Thermobia. Water is removed from the perirectal space by Malpighian tubules contained in the space. The tubules essentially operate by importing potassium chloride into the rectal complex from the haemolymph through modified Malpighian tubule cells-the leptophragma cells. Unlike the situation in more typical tubule cells, this ion movement is not accompanied by osmotically compensating amounts of water from the haemolymph. Osmotically compensating amounts of water do however enter from the perirectal space

324

S. H. P. MADDRELL

through the more typical tubular cells. As a result the tubules emerging from the rectal complex cany water out of the system. How the water is removed from the tubules into the haemolymph is not known. It is clear that there is some most interesting research to be done on several aspects of the operation of the cryptonephric system. It is noticeable that in these two insects where water-vapour can be absorbed against large gradients, the ultrastructure of the rectal cells differs fundamentally from those of insects involved in achieving what by comparison is a more modest concentration of the rectal contents. It may be that this reflects two different basic mechanisms. Several other insects are known to absorb water-vapour from subsaturated atmospheres and it will be interesting to see if this ability is also due to the activity of the rectal epithelium and if so, whether the rectal epithelium is ultrastructurally of the same type as that of Tenebrio and Thermobia. In small terrestrial animals like insects, control of water and salt content is at a premium. A greater knowledge of the ways in which this control is achieved will be of great interest, both for its own sake and for the possibilities it may reveal for developing effective and specific ways of controlling insect pests. ACKNOWLEDGEMENTS

I should like to thank Dr. M. J. Berridge, R. Bridges, B. 0. C. Gardiner, Dr. B. L. Gupta, Dr. R. C. Joyner, Dr. R. B. Moreton, Dr. J. L. Oschman, Dr. W. T. Prince, Dr. J. A. Riegel, Professor Torkel Weis-Fogh, Dr. B. J. Wall and J. L. Wood for many helpful discussions, and J. W. R. Rodford for doing several of the drawings in this paper. REFERENCES Anderson, E. A. and Harvey, W. R. (1966). Active transport by the cecropia midgut. 11. Fine structure of the midgut epithelium. J. Cell Biol. 31, 107-1 34. Beament, J . W. L., Noble-Nesbitt, J. and Watson, J. A. L. (1964). The water-proofing mechanism of arthropods. 111. Cuticular permeability in the firebrat, Thermobia dornestica (Packard). J . exp. Biol. 41, 323-330. Berridge, M. J. (1965a). The physiology of excretion in the cotton stainer, Dysdercus fasciatus Signoret. I. Anatomy, water regulation, osmoregulation. J. exp. Biol. 43, 5 11-521.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

325

Berridge, M. J. (1965b). The physiology of excretion in the cotton stainer, Dysdercus fasciutus Signoret. 111. Nitrogen excretion and excretory metabolism. J. exp. Biol. 43,535-552. Berridge, M. J. (1966a). The physiology of excretion in the cotton stainer, Dysdercus fasciatus Signoret. IV. Hormonal control of excretion. J. exp. Biol. 44,553-566. Berridge, M. J. (1966b). Metabolic pathways of isolated Malpighian tubules of the blowfly functioning in an artificial medium. J. Insect Physiol. 12, 1523-1538. Berridge, M. J. (1967). Ion and water transport across epithelia. In “Insects and Physiology” (J. W. L. Beament and J. E. Treherne, eds). Oliver and Boyd, Edinburgh. Berridge, M. J. (1968). Urine formation by the Malpighian tubules of Calliphora. I. Cations. J. exp. Biol. 48, 159-174. Berridge, M. J. (1969). Urine formation by the Malpighian tubules of Calliphora. 11. Anions. J. exp. Biol. 50, 15-28. Berridge, M. J. and Gupta, B. L. (1967). Fine structural changes in relation to ion and water transport in the rectal papillae of the blowfly, Calliphora. J. Cell Sci. 2, 89-1 12. Berridge, M. J. and Gupta, B. L. (1968). Fine structural localization of adenosine triphosphatase in the rectum of Calliphora. J. Cell Sci. 3, 17-32. Berridge, M. J. and Oschman, J. L. (1969). A structural basis for fluid secretion by Malpighian tubules. Tissue and Cell 1, 247-272. Berridge, M. J. and Patel, N. G. (1 968). Insect salivary glands: ,sti,mulation of fluid secretion by 5-hydroxytryptamine and adenosine-3 5 -monophosphate. Science, N. Y. 162,462-463. Biber, T. U. L. and Curran, P. F. (1968). Coupled solute fluxes in toad skin. J. gen. Physiol. 5 1,606-620. Brown, A. W. A. (1958). “Insecticide Resistance in Arthropods”. World Health Organisation, Geneva. 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. 33-67. Academic Press, London and New York. Cereijido, M. and Rotunno, C. A. (1968). Fluxes and distribution of sodium in frog skin.J. gen. Physiol. 51, 280s-289s. Coast, G. M. (1969). Formation of urinary fluid by Malpighian tubules of an insect. J. Physiol. Lond. 202, 102P-103P. Corbet, S. A. and Rotheram, S. (1965). The life history of the ichneumonid Nemeritis (Devorgilla) canescens (Gravenhorst) as a parasite of the Mediterranean flour moth, Ephestia (A nagasta) kuehniella Zeller, under laboratory conditions. Proc. R. ent. SOC. Lond. A40,67-72. Cottrell, C. B. (1962). The imaginal ecdysis of blowflies. Observations on the hydrostatic mechanisms involved in digging and expansion. J. exp. Biol. 39,431448. Diamond, J. M. (1962). The mechanism of solute transport by the gallbladder. J. Physiol. Lond. 161,474-502. Diamond, J. M. and Bossert, W. H. (1967). Standing-gradient osmotic flow. A mechanism for coupling of water and solute transport in epithelia. J. gen. Physiol. 50,206 1-2083. AIP-14

326

S,

H. P. MADDRELL

Diamond, J. M. and Bossert, W. H. (1968). Functional consequences of ultrastructural geometry in u backward^" fluid-transporting epithelia. J. Cell Biol. 37,694-702. Diamond, J. M. and Tormey, J. M. (1966). Role of long extracellular channels in fluid transport across epithelia. Nature, Lond. 210, 8 17-820. Diamond, J. M. and Wright, E. M. (1969). Biological membranes: the physical basis of ion and non-electrolyte selectivity. A . Rev. Physiol. 31, 581-646. Edwards, J. S. (1964). Diuretic function of the labial glands in adult giant silk moths, Hyalophora cecropia. Nature, Lond. 203,668-669. Franz, T. J., Galey, W. R. and Van Bruggen, J. T. (1968). Further observations on asymmetrical solute movements across membranes. J. gen. Physiol. 51, 1-12. Gersch, M. ( 1967). Experimental examinations on the hormonal control of the water balance and excretion of the larva of Corethra (Chaoborus). Gen. comp. Endocr. 9,453. Glynn, I. M. (1964). The action of cardiac glycosides on ion movements. Pharmac Rev. 16,381-407. Graham-Smith, G. S. (1934). The alimentary canal of Calliphora erythrocephala L., with special reference to its musculature, proventriculus and its rectal valve and papillae. Parasitology. 26, 176-248. Grimstone, A. V., Mullinger, A. M. and Ramsay, J. A. (1968). Further studies on the rectal complex of the mealworm, Tenebrio molitor L. (Coleoptera, Tenebrionidae). Phil. Trans. R . SOC.B253, 343-382. Gupta, B. L. and Berridge, M. J. (1966a). A coat of repeating subunits on the cytoplasmic surface of the plasma membrane in the rectal papillae of the blowfly, Calliphora erythrocephala (Meig.), studied in situ by electron microscopy. J. Cell Biol. 29,376-382. Gupta, B. L. and Berridge, M. J. (1966b). Fine structural organization of the rectum in the blowfly, Calliphora erythrocephala (Meig.) with special reference to connective tissue, trachea and neurosecretory innervation in the rectal papillae. J. Morph. 120, 23-81. Gupta, P. D. and Sinha, R. N. (1960). Excretion and its products in some stored-grain-infesting beetles. Ann. ent. SOC. A m . 53, 632-638. Harmsen, R. (1966). The excretory role of pteridines in insects. J. exp. Biol. 45, 1-13. Harvey, W. R. and Nedergaard, S. (1964). Sodium-independent active transport of potassium in the isolated midgut of the cecropia silkworm. Proc. natn. Acad. Sci. U.S.A. 51, 757-765. Harvey, W. R. and Zerahn, K. (1969). Kinetics and route of active K-transport in the isolated midgut of Hyalophora cecropia. J. exp. Biol. 50,297-306. Haydak, M. H. (1953). Influence of the protein level of the diet on the longevity of cockroaches. Ann. ent. SOC.A m . 46,547-560. Hollande, A . C. (1922). La cellule p6ricardiale des insectes. Archs. Anal. microsc. 18, 85-307. Hopkins, C. R. (1967). The fine-structural changes observed in the rectal papillae of the mosquito Aedes aegypti, L. and their relation to the epithelial transport of water and inorganic ions. J1 R o y . microsc. SOC.86,235-252. Irvine, H. B. (1969). Sodium and potassium secretion by isolated insect Malpighian tubules. A m . J. Physiol. 217, 1520-1527.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

327

Irvine, H. B. (1970). Malpighian tubule secretion in Calpodes ethlius Stoll. Ph.D. Thesis, Case Western Reserve University. JQrgensen, C. B. (1950). The amphibian water economy, with special regard to the effect of neurohypophyseal-extracts. A eta. physiol. scand. 22 (suppl.), 1-79. Joyner, R. C. (1970). Some aspects of the fine structure and function of Malpighian tubules in the desert locust Schistocerca gregaria. Ph.D. Thesis, University of Bristol. Kafatos, F. C. (1968). The labial gland: a salt-secreting organ of saturniid moths. J. exp. Biol. 48. 435453. 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. Katz, B. (1966). “Nerve, Muscle and Synapse”. McGraw-Hill, New York. Kaye, G. I., Wheeler, H. O., Whitlock, R. T. and Lane, N. (1966). Fluid transport in the rabbit gall bladder. J. Cell Biol. 30,237-268. Kedem, 0 . (1965). Water flow in the presence of active transport. Symp. SOC. exp. Biol. 19, 61-74. Kessel, R. G. (1970). The permeability of dragonfly Malpighian tubule cells to protein using horseradish peroxidase as a tracer. J. Cell Biol. 47,299-303. Kilby, B. A. (1963). The biochemistry of the insect fat body. In Advances in Insect Physiology” (J. W. L. Beament, J. E. Treherne and V. B. Wigglesworth, eds), Vol. 1, pp. 11 1-174. Academic Press, London and New York. Lee, R. M. (1961). The variation of blood volume with age in the desert locust (Schistocerca gregaria Forsk.). J. Znsect Physiol. 6, 36-57. Lester, H. M. 0. and Lloyd, L. (1928). Notes on the process of digestion in tsetse flies. Bull. ent. Res. 19, 39-60. Lison, L. (1937). htudes histophysiologiques sur les tubes de Malpighi des Insectes. I. Elimination des colorants acides chez les Orthoptdres. Archs Biol. Paris, 48,321-360. Lison, L. (1938). ktudes histophysiologiques sur les tubes de Malpighi des Insectes. 111. L’klimination des colorants basiques chez les Orthoptkres. 2. Zellforsch. Mikrosk, Anat. 28, 179-209. Locke, M. and Collins, J. V. (1967). Protein uptake in multivesicular bodies in the molt-intermolt cycle of an insect. Science, N. Y. 155,467-469. Locke, M. and Collins, J. V. (1968). Protein uptake into multivesicular bodies and storage granules in the fat body of an insect. J. Cell Biol. 36,453-483. Loewenstein, W. R., Socolar, S. J., Higashino, S., Kanno, Y. and Davidson, N. ( 1965). Intercellular communication: renal, urinary bladder, sensory and salivary gland cells. Science, N. Y. 149, 295-298. Maddrell, S. H. P. (1963). Excretion in the blood-sucking bug, Rhodnius prolixus S t a . I. The control of diuresis. J. exp. Biol. 40,247-256. Maddrell, S . H. P. (1964). Excretion in the blood-sucking bug, Rhodnius prolixus Stil. 11. The normal course of diuresis and the effect of temperature. J. exp. Biol. 41, 163-176. Maddrell, S. H.P. (1969). Secretion by the Malpighian tubules of Rhodnius. The movements of ions and water. J. exp. Biol. 51, 71-97. Maddrell, S. H. P. (1971). Secretion by the Malpighian tubules of Rhodnius. 11. Electrical events. (In preparation.)

328

S. H. P. MADDRELL

Maddrell, S. H. P., Pilcher, D. E. M. and Gardiner, B. 0. C. (1969). Stimulatory effect of 5-hydroxytryptamine (serotonin) on secretion by Malpighian tubules of insects. Nature, Lond. 222, 784-785. Maddrell, S. H. P. and Reynolds, S. E. (1971). The secretion of dyes and organic anions by the Malpighian tubules of insects. (In preparation.) Martignoni, M. E. (1964). Pathophysiology in the insect. A . Rev. Ent. 9 , 179-206. Mellanby, K. (1932). The effect of atmospheric humidity on the metabolism of the fasting mealworm (Tenebrio molitor L., Coleoptera). Proc. R . SOC. B111,376-390. Mordue, W. ( 1969). Hormonal control of Malpighian tubule and rectal function in the desert locust, Schistocerca gregaria. J. Insect Physiol. 15, 273-285. Nedergaard, S. and Harvey, W. R. (1968). Active transport by the cecropia midgut. IV. Specificity of the transport mechanism for potassium. J. exp. Biol. 48, 13-24. Noble-Nesbitt, I. (1969). Water balance in the firebrat, Thermobia dornestica (Packard). Exchanges of water with the atmosphere. J. exp. Biol. 50, 745-769. Noble-Nesbitt, J. ( 1970). Water balance in the firebrat, Thermobia domestica (Packard). The site of uptake of water from the atmosphere. J. exp. B i d . 52, 193-200. Noirot, C. H. and Noirot-Timothie, C. (1966). Revitement de la membrane cytoplasmique et absorption des ions dans les papilles rectales d’un termite (Insecta, Isoptera). C.r. hebd. Skanc. Acad. S c i , Paris. 263, 1099-1 102. Oschman, J. L. and Berridge, M, J. (1970). Structural and functional aspects of salivary fluid secretion in Calliphora. Tissue and Cell 2, 28 1-370. Oschman, J. L. and Wall, B. 1. (1969). The structure of the rectal pads of Periplaneta americana L. with regard to fluid transport. J. Morph. 127, 475-5 10. Palm, N-B. (1952). Storage and excretion of vital dyes in insects. Ark. Zool. 3, 195 -272. Phillips, J. E. (1961). Rectal absorption of water and salts in the locust and blowfly. Ph.D. Thesis, University of Cambridge. Phillips, J. E. (1964a). Rectal absorption in the desert locust, Schistocerca gregaria Forskil. I. Water. J. exp. Biol. 41, 15-38. Phillips, J. E. (1964b). Rectal absorption in the desert locust, Schistocerca gregaria Forskil. 11. Sodium, potassium and chloride. J. exp. Biol. 41, 39-67. Phillips, J. E. ( 1 9 6 4 ~ ) .Rectal absorption in the desert locust, Schistocerca gregaria Forskdl. 111. The nature of the excretory process. J. exp. Biol. 41, 68-80. Phillips, J. E. (1965). Rectal absorption and renal function in insects. Trans. R . SOC.Canada 3,231-254. Phillips, J. E. (1969). Osmotic regulation and rectal absorption in the blowfly, Calliphora erythrocephala. Can. J. Zool. 47,85 1-863. Phillips, J. E. (1970). Apparent transport of water by insect excretory systems. Am. Zool. 10,413-436. Phillips, J. E. and Dockrill, A. A. (1968). Molecular sieving of hydrophilic

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

329

molecules by the rectal intima of the desert locust (Schistocerca gregaria). J. exp. Biol. 48,52 1-532. Phillips, J. E. and Meredith, J. (1969). Active sodium and chloride transport by anal papillae of a salt water mosquito larva (A edes campestris). Nature, Lond. 222, 168-169. Pilcher, D. E. M. (1969). Hormonal control of the Malpighian tubules of the stick insect, Carausius morosus. Ph.D. thesis, University of Cambridge. Pilcher, D. E. M. (1970a). Hormonal control of the Malpighian tubules of the stick insect, Carausius morosus. J. exp. Biol. 52, 653-665. Pilcher, D. E. M. (1970b). The influence of the diuretic hormone on the process of urine secretion by the Malpighian tubules of Carausius morosus. J. exp. Biol 53, 465-484. Ramsay, J. A. (1950). Osmotic regulation in mosquito larvae. J. exp. Bio2. 27, 145-157. Ramsay, J. A. (1952). The excretion of sodium and potassium by the Malpighian tubules of Rhodnius. J. exp. Biol. 29, 110-126. Ramsay, J. A. (1953). Active transport of potassium by the Malpighian tubules of insects. J. exp. Biol. 30, 358-369. Ramsay, J. A. (1954). Active transport of water by the Malpighian tubules of the stick insect, Dixippus morosus (Orthoptera, Phasmidae). J. exp. Biol. 31, 104-1 13. Ramsay, J. A. (1955a). The excretory system of the stick insect, Dixippus morosus (Orthoptera, Phasmidae). J. exp. Biol. 32, 183-199. Ramsay, J. A. (1955b). The excretion of sodium, potassium and water by the Malpighian tubules of the stick insect, Dixippus morosus (Orthoptera, Phasmidae). J. exp. Biol. 32, 200-2 16. Ramsay, J. A. (1956). Excretion by the Malpighian tubules of the stick insect, Dixippus morosus (Orthoptera, Phasmidae): calcium, magnesium, chloride, phosphate and hydrogen ions. J. exp. Biol. 33,697-709. Ramsay, J. A. (1958). Excretion by the Malpighian tubules of the stick insect, Dixippus morosus (Orthoptera, Phasmidae): amino acids, sugars and urea. J. exp. Biol. 35,871-891. Ramsay. J. A. (1964). The rectal complex of the mealworm, Tenebrio molitor L. (Coleoptera, Tenebrionidae). Phil. Trans. R . SOC.B248,279-3 14. Riegel, J. A. (1966a). Micropuncture studies of formed body secretion by the excretory organs of the crayfish, frog and stick insect. J. exp. Biol. 44, 379-385. Riegel, J. A. (1966b). Analysis of formed bodies in urine removed from the crayfish antenna1 gland by micropuncture. J. exp. Biol. 44, 387-395. Riegel, J. A. (1968). Analysis of the distribution of sodium, potassium and osmotic pressure in the urine of crayfishes. J. exp. Biol. 48,587-596. Riegel, 5. A. (1970). In vitro studies of fluid and ion movements due t o the swelling of formed bodies. Comp. Biochem. Physiol. 35,843-856. Robertson, J. D. (1957). Osmotic and ionic regulation in aquatic invertebrates. In “Recent Advances in Invertebrate Physiology”. University of Oregon Publications, Eugene ,Oregon. Robison, G. A,, Butcher, R. W. and Sutherland, E. W. (1968). Cyclic AMP. A. Rev. Biochem. 37, 149-171.

330

S . H. P. MADDRELL

Roth, L. M. and Dateo, G. P. (1964). Uric acid in the reproductive system of males of the cockroach, Blattelh germanica. Science, N. Y. 146, 782-784. Roth, L. M. and Dateo, G. P. (1965). Uric acid storage and excretion by accessory sex glands of male cockroaches. J. Insect Physiol. 11, 1023-1029. Saini, R. S. (1964). Histology and physiology of the cryptonephridial system in insects. Trans. R . ent. SOC.Lond. 116, 347-392. Smith, H.W. (195 1). “The Kidney”. Oxford University Press, New York. Stobbart, R. H. (1968). Ion movement and water transport in the rectum of the locust, Schistocerca gregaria. J. Insect Physiol. 14,269-275. Sutcliffe, D. W. (1960). Osmotic and ionic regulation in the larvae of some euryhaline Diptera. Nature, Lond. 187, 33 1-332. Tatchell, R. J. (1967). Salivary secretion in the cattle tick as a means of water elimination. Nature, Lond. 213,940-941. Tormey, J. M. and Diamond, J. M. (1967). The ultrastructural route of fluid transport in rabbit gall bladder. J. gen. Physiol. 50, 203 1-2060. Unger, H. (1965). Der Einfluss der Neurohormone C und D auf die Farbstoffabsorptions-fihigkeit der Malpighische Gefasse (und des Darmes) der Stabheuschrecke Carausius morosus in vitro. Zool. Jb. (Physiol.) 71, 7 10-716. Ussing, H. H. (1960). The alkali metal ions in isolated systems and tissues. In “Alkali Metal Ions in Biology”. Springer-Verlag, Berlin. Wall, B. J. (1967). Evidence for antidiuretic control of rectal water absorption in the cockroach Periplaneta americana L. J. Insect Physiol. 13, 565-578. Wall, B. J. (1970). Effects of dehydration and rehydration on Periplaneta americana. J. Insect Physiol. 16, 1027-1042. Wall, B. J. and Oschman, J. L. (1970). Water and solute uptake by the rectal pads of Periplaneta americana. A m . J. Physiol. 218, 1208-1215. Wall, B. J., Oschman, J. L. and Schmidt-Nielsen, B. (1970). Fluid transport: concentration of the intercellular compartment. Science, N. Y. 167, 1497-1498. Watson, J. A. L. (1967). Reproduction, feeding activity and growth in the adult firebrat, Lepismodes inquilinus Newman (Thysanura, Lepismatidae). J. Insect Physiol 13, 1689-1698. Weis-Fogh, T. (1967). Metabolism and weight economy in migrating animals, particularly birds and insects. In “Insects and Physiology” (J. W. L. Beament and J. E. Treherne, eds). Oliver and Boyd, Edinburgh. Wessing, A. and Eichelberg, D. ( 1969). Elektronenoptische’ Untersuchungen an der Nierentubuli (Malpighische Gefasse) von Drosophila melanogaster. I. Regionale Gliederung der Tubuli. Z. Zellforsch. Mikrosk. Anat. 101, 2 8 5-32 2. Wigglesworth, V. B. (1931a). The physiology of excretion in a blood-sucking insect, R hodnius prolixus (Hemiptera, Reduviidae). I. The composition of the urine. J. exp. Biol. 8, 41 1-427. Wigglesworth, V. B. (1 93 1b). The physiology of excretion in a blood-sucking insect, R hodnius prolixus (Hemiptera, Reduviidae). 11. Anatomy and histology of the excretory system. J. exp. Biol. 8,428-442. Wigglesworth, V. B. (193 1c). The physiology of excretion in a blood-sucking insect, Rhodnius prolixus (Hemiptera, Reduviidae). 111. The mechanism of uric acid excretion. J. exp. Biol. 8,443-45 1.

THE MECHANISMS OF INSECT EXCRETORY SYSTEMS

33 1

Wigglesworth, V. B. (1943). The fate of haemoglobin in R hodnius prolixus (Hemiptera) and other blood-sucking arthropods. Proc. R. SOC. B131, 3 13-339. Wigglesworth, V. B. (1965). “The Principles of Insect Physiology” (6th ed.). Methuen, London. Wigglesworth, V. B. (1970). The pericardial cells of insects: analogue of the reticuloendothelial system. J. Reticuloendothelial SOC. 7 , 208-2 16. Wood, J. L. (1971). Potassium transport by the midgut of silkworms. Ph.D. Thesis, University of Cambridge. Wood, J. L., Farrand, P. S. and Harvey, W. R. (1969). Active transport of potassium by the cecropia midgut. VI. Microelectrode potential profile. J. exp. Biol. 50, 169-178.

NOTE ADDED IN PROOF: THE MIDGUT OF LARVAE OF SATURNIID SILKMOTHS Isolated midguts of Saturniid larvae will only normally transport potassium ions (see p. 206). However, a recent paper (Harvey and Zerahn, 197 1) has shown that if calcium and magnesium ions are left out of the bathing solution and if no potassium ions are available, isolated midguts of larvae of Hyalophora cecropia can be made to transport sodium or lithium ions. It is possible that omission of calcium ions makes the midgut cells more permeable to sodium and lithium ions and they then can gain access to the pump. This is compatible with the idea advanced on p. 207 that the pump draws ions for transport from an intracellular pool. REFERENCE Harvey, W. R. and Zerahn, K. (1971). Active transport of sodium by the isolated midgut of Hyalophora cecropia. J. exp. Biol. 54, 269-274.

This Page Intentionally Left Blank

Author Index Numbers with an asterisk refer to pages on which references are listed at the end of the paper A Abedi, Z. H., 57, 80* Abushama, F. T., 164, 190* Adkisson, P. L., 57, 8 1 * Agosin, M., 57, 80* Ahmad, T., 177, 194* Albrecht, F. O., 159, 165, 168, 172, 175, 190*, 191* Anderson, E. A., 207, 324* Anderson, N. C., 39, 50, 51, 52, 53, 54, 55,86*, 87* Armstrong, C. M., 15,80* Atwood, H. L., 136, 138, 140*, 142 Aziz, S. A., 173, 191* B Babers, F. H., 66, 80*, 89* Baker, P. F., 5 1, 80* Barker, R. J., 56, 57, 80* Baylor, D. A., 120, 140* Beament, J. W. L., 307, 324* Becht, G., 22, 24, 80* Benjamini, E., 56, 57, 86* Bennett, R. R., 163, 191* Berger, N. E., 27, 85* Bergmann, F., 99, 100, 102, 104, 105, 106, 107, 108, 110, 111, 113, 114, 115, 116, 117, 118, 119, 120, 125, 129,132,142*, 143* Berridge, M. J., 202, 203, 204, 209, 213, 214, 216, 217, 218, 219, 220, 221, 222, 224, 225, 226, 228, 232, 233, 234, 236, 237, 252, 256, 268, 270, 273, 276, 278, 279, 296, 298, 303, 324*, 325*, 328* Berteau, P. E., 75, 76, 80* Bittner, G. D., 138, 140* Biber, T. U. L., 257, 275, 325* 333

Binstock, W. M., 15, 80* Blackith, R. E., 185, 191* Blackith, R. M., 185, 191* Blaustein, M. P., 39, 86* Blum,M. S., 5, 27, 61, 81* Bodenstein, D., 23, 81* Boistel, J., 27, 28, 97, 110, 129, 81*, 140*, 142* Bossert, W. H., 236, 237, 270, 271, 272, 273, 274, 325*, 326* Boulton, P. S., 108, 140* Brett, C, H., 56, 154, 89*, 191 Brown, A. W. A., 3, 22, 24, 26, 27, 31, 57, 65, 69, 80*, 81*, 83*, 85*, 88*, 201, 325* Browne, L. B., 67, 81* Bruce, W.N., 66,90* Brunnert, H., 78, 81* Buckner, A. J., 57,69, 88* Bull, D. L., 57, 8 1 * Bullock, T. H., 96, 97, 98, 108, 118, 136, 140* Bursell, E., 203, 325* Burton, J. G., 155, 191* Burtt, E., 159, 191* Butcher, R. W., 210, 329* Byrne,D. W., 154, 156, 172, 191* C Callec, J., 27, 28, 8 1 * Callec, J. J., 97, 129, 140* Cannis, T. L., 174, 197* Carlisle, D. B., 182, 192* Casida, J. E., 3, 75, 76, 80*, 81* Cassier, P., 165, 182, 191*, 196* Cazal, M., 180, 193* Cereijido, M., 230, 325*

334

AUTHOR INDEX

Chamberlain, R. W., 61, 81* Chang, S. C., 5 , 23, 90* Chapman, E., 176, 198* Chapman, R. F., 169, 19 1* Cheng, T. H., 57, 82* Clarke, K. U., 180, 191* Coast, G. M., 236, 238, 325* Cohen, M. J., 134, 140* Colhoun, E. H., 3, 3 1, 8 1* Collins, J. V., 269, 327* Common, I. F. B., 176, 19 1* Cook, E. F., 100, 102, 104, 105, 107, 109, 142* Cook, P. M., 121, 128, 140* Coombs, J. S., 118, 140* Corbet, S. A., 203, 325* Cottrell, C. B., 210, 325* Cowan, S., 20, 85* Cromartie, R. I. T., 19 1 * Curran, P. F., 257, 275, 325* Curtis, D. R., 118, 140* Cutkomp, L. K., 22, 31, 56, 85*, 89*, 90*

D Dadd, R. H., 185, 189, 191* Dagan, D., 124, 125, 126, 131, 133, 141* Dahlman, D. L., 185, 191* Dahm, P. A., 3, 8 1* Dallemagne, M. J., 24, 8 1 * Dateo, G. P., 204, 330* Davidson, N., 230, 327* Davson, H., 5, 81* Day, M. F., 146, 195* De Coursey, J. D., 61, 83* Degucai, T., 19, 91* De Wilde, J., 159, 198* Diamond, J. M., 236, 237, 244, 245, 257, 258, 270, 271, 272, 273, 274, 325*, 326*, 330* Dirsch, V. M., 146, 191* Dockrill, A. A., 304, 329* Dresden, D., 22, 27, 8 1*, 85* Duck, L. D., 164, 174, 175, 176, 191* Dudley, B., 177, 191* Duffy J. R., 57, 80*

Dustan, G. G., 56, 8 1* Dustmann, J. H., 164, 192* E Eaton, J. L., 22, 23, 60, 61, 82* Eaton, R. C., 134, 141* Eccles, J., 118, 136, 140*, 141* Eccles, J. C., 5 , 21, 82* Edwards, J. S., 209, 326* Eichelberg, G., 276, 330* Ellis,P.E., 178, 182, 192*, 196 Ergene, S., 158, 159, 160, 166, 167, 170, 178, 192* Euw, J. V., 152, 192* Eyzaguirre, C., 118, 141* F Faeder, I. R., 20, 27, 82* Fain-Maurel, M. A., 165, 182, 196* Fan, H. Y., 56,82* Farley, R. D., 99, 104, 106, 107, 109, 11 1, 113, 131, 132, 134, 141* Farrand, P. S., 206, 33 1* Faure, J. C., 158, 159, 160, 166, 172, 176, 193* Fielden, A., 96, 141* Fishelson, L., 152, 192* Fisher, R. A., 153, 193* Frankenhaeuser, B., 11, 82* Franz, T. J., 257,275,326* Frazier, D. T., 39, 82*, 88* Fritsch, H., 24, 82* Fuhrman, F. A., 39, 90* Fukami, J., 27, 77, 82* Fukuto, T. R., 3, 73, 74,82*, 86* Fullmer, 0. H., 56, 82* Furshpan, E. J., 17, 82* Furjkawa, T., 17, 82* Fuzeau-Braesch, S., 156, 159, 164, 167, 177, 181, 182, 183, 184, 186, 187, 189, 193*, 196*

G Galey, W. R., 257, 275,326* Gardiner, B. 0. C., 247, 284, 328*

AUTHOR INDEX

Gasser, H. S., 32, 83* Gates, 173 Gersch, M., 286, 326* Gianotti, O., 25, 83* Gillham, E. M., 56, 89* Girardie, A., 179, 180, 193* Glynn, I. M., 226, 326* Golding, F. D., 168, 172, 193* Goldman, D. E., 38,84* Good, C. M., 153, 193* Goodwin, T. W., 184, 185, 186, 187, 188, 189, 193* Gordon, H. T., 3, 23, 26, 37, 83*, 84*, 92* Graham, H. T., 32,83* Graham-Smith G. S., 296, 298, 326* Grayson, J. M., 160, 163, 164, 176, 189, 193*, 194* Grimstone, A. V., 308, 31 1, 312, 314, 3 1 6 , 3 17,326* Grundfest H., 32, 83* Gruner, H. E., 56, 91* Gupta, B. L., 296, 298, 303, 325*, 32,6* Gupta, P. D., 204, 326* Guthrie, D. M., 100, 102, 110, 123, 141* Guthrie, F. E., 56, 61, 83* H Haas, H. G., 38, 39, 40, 41, 42, 43, 44, 79,87* Hafliger, E., 56, 83* Halbwachs, M., 178, 194* Hancock, J. L., 153, 194* Hanstrom, B., 96, 141* Harlow, P. A., 22, 23, 24, 28, 31, 83* Harries, F. H., 61, 83* Harmsen, R., 201,204, 326* Harrison, A., 69, 92* Harrison, C. M., 7 1, 83* Harrison, J., 138, 140* Hartzell, A., 61, 83* Harvey, G. T., 27,83* Harvey, W. R., 206, 207, 324*, 326*, 328* Hatai, N., 27,83* Hatanaka, A., 78, 83*

335

Hawkins, W. B., 5, 23, 83* Hayashi, M., 69, 78, 85*, 86* Haydak, M. H., 204, 326* Hayes, W. J., Jr., 3, 83* Heslop, J. P., 22, 23, 83* Hemingway, E., 194* Hertz, M., 159, 160, 161, 163, 172, 188,194* Hess, A., 97, 100, 102, 106, 108, 111, 113. 114, 141* Higashino, S., 230, 327* Highnam, K. C., 180, 194* Hille, B., 45, 83* Hilton, B. D., 78, 79, 83* Hodgkin, A. L., 5, 11, 14, 17, 50, 51, 80*, 82*, 84*, 103, 141* Hoffman, R. A., 56, 84* Hofmaster, R. N., 61, 83* Holan, G., 73, 79, 84* Hollande, A. C., 205,326* Hopkins, C. R., 299, 326* Hopkins, T. L., 3, 89* Horridge, G. A., 96, 97, 98, 108, 136, 140*, 163, 191* Hoskins, W. M., 3, 56, 57, 66, 80*, 82*, 84*, 86*, 87*, 88*, 90* Huber, F., 96, 141* Hughes, G. M., 96, 99, 101, 102, 123, 129, 132, 141* Hunter-Jones, P., 154, 156, 172, 194* Hurst, H., 58, 84* Husain, M. A., 277, 180, 194* Huxley, A. F., 14, 84*

I Imms, A. D., 159, 160, 161, 163, 172, 188, 194* Irvine, H. B., 216, 237, 264, 266, 267, 268,299,326*, 327* Ishii, T., 21, 22, 23, 24, 25, 26, 27, 31, 56, 57, 58, 59, 60, 61, 92*, 93* Ito, M., 21, 84* Ives, N. F., 26, 85* J Jacklet, J. W.,134, 140* Johnson, H. B., 174, 176, 194*

336

AUTHOR INDEX

Joly, L., 178, 194* Joly, P., 178, 180, 193*, 194* Jones, D. P., 106, 141* 'Jqhgensen, C. B., 202, 327* JovanCiE, L., 160, 170, 171, 172, 174, 175, 194* Joyner, R. C., 276,327* Julian, F. J., 38, 84*

Larralde, J., 24, 91' Leaf, G., 20, 27, 9 1 * Leake, L. D., 20,85* Lee, R. M., 210, 327* Legge, J. W., 184, 195* Lehmann, J. E., 32, 85* Lemberg, R., 184, 195* Leski, R. A., 31, 85* Lester, H. M. O., 212, 327* Lettvin, J. Y., 39, 90* Letvita, B., 159, 160, 161, 163, 184, 187, 188, 195*, 196* Lewis, P. R., 17, 85* Lewis, S. E., 3, 92* Lewontin, R. C., 153, 195* Lindahl, P. E., 27,85* Lindquist, A. W., 56, 84*, 85* Link, J. D., 85* Linzen, B., 195* Lipke, H., 3, 26, 70, 72, 85*, 86* Lison, L., 205, 281, 327* Lloyd, L., 212, 327* Locke, M., 269, 327* Lowenstein, O., 26, 85* Loewenstein, W. R., 230, 327* Lynch, J. M., 44,89*

K Kaeser, W., 56, 84* Kafatos, F. C., 209, 210, 211, 327* Kanno, Y., 230, 327* Kao, C. Y., 39, 44, 84*, 90* Karuhize, G. R., 153, 195* Katz, B., 5, 84*, 230, 327* Kaye, G. I., 274, 327* Kearns, C. W., 3, 23, 27, 56, 57, 58, 61, 66, 70, 72,81*, 85*, 86*,89*, 90*, 91* Kedem, O., 299,327* Kennedy, N. K., 28,89* Kerkut, G. A., 20, 21, 27, 28, 85* Kerr, R. W., 67, 81* Kessel, R. G., 270, 327* Key, K. H. L., 146, 154. 156, 164, 176,195* Keynes, R. D., 17,84*, 85* M Khan, M. A. Q., 69,85* Machili, P., 20, 27, 9 1 * Kilby, B. A., 203, 327* Madden, A. H., 56,85* King, R. L., 154, 195* Maddrell, S. H. P., 216, 236, 237, 238, Klein, A. K., 26, 85* 239, 240, 241, 242, 243, 247, 248, Kolltos, J. J., 23, 91* 253, 256, 260, 263, 268, 270, 278, Kostyuk, P. G., 2 1, 84* 284,289,327,328* Krasne, F. B., 128, 141* R. B., 3, 25, 83*, 85* March, Kravitz, E. A., 114, 143* Martignoni, M. E., 204, 328* Krijgsman, B. J., 27, 85* Martin, H., 77, 85* Krupp, H., 24, 82* Mathur, C. B., 177, 180, 194* Kuffler, S. W., 118, 141* Matsumara, F., 26, 69, 78, 81*, 85*, 86*, 88* Matsumoto, K., 44,89* Matthee, J. J., 182, 195* L Matthews, H. B., 26, 86* La Coeur, L., 155, 197* Lalonde, D. I. V., 22, 24, 26, 27, 31, Maynard, D. M.,130, 141* Meadows, P. S., 167, 195* 85* Mellanby, K., 31 1, 328* Lane, N., 274, 327*

337

AUTHOR INDEX

Menn, J. J., 56, 57, 86* Meredith, J., 2 12, 329* Metcalf, R. L., 3, 25, 73, 74, 83*, 86* Michaeli, D., 57, 80* Milani, R., 70, 86* Milburn, N. S., 99, 104, 106, 107, 109, 111, 113, 131, 132, 141* Mill, P. J., 96, 103, 141* Miller, S., 57, 86*, 88* Miskus, R., 57, 80* Miyake, S. S., 70, 72, 86* Moore, J. W., 13, 14, 15, 16, 38, 39, 41, 82*, 84*, 86*, 87*, 88*, 90* Morove, W., 328* Mullinger, A. M., 308, 311, 312, 314, 316. 317.326* Mullins', L. J., 79, 86*, 87* Munson, S. C., 58,87*

0 Oberg, K. E., 27,85*, 88* O'Brien, R. D., 3, 65, 78, 79, 83*, 86*, 88*, 90* Ohkubo, Y., 19,91* Oikawa, T., 50,88* Orser, W. B., 27, 88* Okay, S., 164, 170, 172, 174, 187, 196* Oschman, J. L., 209, 234, 236, 256, 268, 270, 273, 276,278, 279, 290, 299, 301, 302, 303, 304, 325*, 328*, 330* Oshima, T., 2 1, 84* Owen, D. F., 153, 196*

P Padilla, G. M., 58, 87* N Palm, N. B., 281,328* Parnas, I., 99, 100, 102, 104, 105, Nabours, R. K., 152, 153, 195* 106, 107, 108, 110, 111, 113, 114, Nagasawa, S., 56, 57, 77, 80*, 87* 115, 116, 117, 118, 119, 120, 124, Nakajima, S., 120, 141* 125, 126, 129, 131, 132, 133, 136, Nakatsugawa, T., 3, 27, 81 *, 82* 138, 140*, 141*, 142*, 143* Narahashi, T., 4, 5, 6, 10, 11, 13, 14, 15, 16, 18, 22, 23, 25, 26, 27, 28, Puson, J. A., 152, 192* 29, 30, 31, 32, 33, 34, 35, 36, 3 7 , , Pasons, J., 152, 197* 38, 39, 40, 41, 42, 43,44, 46, 47, Passama-Vuillaume, M., 161, 171, 174, 184, 187, 196* 48, 49, 50, 51, 52, 53, 54, 55, 6 2 , / 63, 66, 67, 68, 69, 70, 71, 72, 73, Patel, N. G., 214, 278,325* 75, 76, 77, 79,80*, 82*, 86*, 87*, Patton, H. D., 5,89* 88*, 91*, 92*, 93*, 103, 110, 129, Pearce, G. W., 69,88* Perry, A. S., 3, 57, 66, 69,86*, 88* 142*, 143* Peterson, B., 160, 196* Nastuk, W. L., 5,88* Pfeiffer, I. W., 178, 196* Nayar, J. K., 185, 195* Nedergaard, S., 206, 207, 326*, 328* Philippot, E., 24, 81* Nel, M. D., 156, 195* Phillips, J. E., 212, 287, 290, 291, Nicholls, J. F., 120, 140* 292, 293, 294, 295, 296, 299, 300, Nickerson, B., 180, 181, 186, 195* 303,304,328*, 329* Nicolas, G., 159, 165, 166, 177, 181, Pichon, Y., 45, 88*, 89*, 103, 110, 182, 195* 140*, 142* Noble-Nesbitt, J., 307, 324* 328* Pitman, R. M., 27,28,85* Noirot, C. H.,299, 328* Pickard, W. F., 39, 90* Noirot-Timothbe, C., 299, 328* Pilcher, D. E. M., 216, 226, 227, 228, Nolte, D. J., 154, 156, 166, 177, 184, 233, 239, 247, 276, 284, 285, 196* 328*, 329* Novak, V. J. A., 178, 196* Pipa, R. L., 100, 104, 105, 107, 142*

338

AUTHOR INDEX

Plotnikov V.I., 176, 196* Popov, G., 175, 196* Poston, R. N., 39, 88* Potter, C., 56, 89* Potter, D. D., 17, 56, 82* Poulton, E. B., 158, 197* Power, M. E., 96, 142* Pradhan, S., 56, 89* Pratt, J. J., Jnr., 66, 80*, 89* Pumphrey, R. J., 97, 142* Putnam, L. G., 154, 197*

R Ramme, W., 154, 197* Ramsay, J. A., 213, 214, 215, 216, 223, 227, 236, 237, 238, 275,279, 280, 283, 286, 287, 289, 306, 307, 308, 311, 314, 315, 316, 317,318, 326*, 329* Rapoport, H., 44,89* Rawdon-Smith, A. F., 97, 142* Ray, J. W., 22, 23, 69, 89* Reichstein, T., 152, 192*, 197* Reynolds, S. E., 328* Rhoades, W. C., 56, 89* Richards, A. G., 56, 82*, 89*, 100, 104, 105, 107, 142* Richards, 0. W., 155, 160, 168, 169, 197* Riegel, J. A., 277, 278, 329* Riegert, P. W., 173, 197* Roan, C. C., 3,89* Roberts, A., 132, 142* Robertson, J. D., 231, 329* Robison, G. A., 2 10,329* Roeder, K. D., 22, 23, 28, 89*, 91*, 96, 97, 98, 100, 102, 103, 104, 108, 111, 123, 124, 128, 129, 132, 142* Roth, A. R., 56, 84* Roth, L. M., 204, 330* Rotheram, S., 203, 325* Rothschild, M., 152, 192*, 197* Rotunno, C. A., 230, 325* Roussel, J. P.., 183, 197* Rowe, E. C., 134, 142*

Rowell, C. H. F., 108, 121, 140*, 142*, 152, 156, 159, 160, 161, 162, 163, 165, 166, 168, 170, 171, 172, 173, 174, 176, 179, 188, 189, 197* Roys, C. C., 22, 66, 90* Ruch, T. C., 5, 89* Rubtzov, I. A., 154, 155, 160, 170, 176,197* Rushton, W. A. H., 103, 141*

S Saini, R. S., 31 1,330* Saint-Rkmy, G., 96, 143* Samson, E. A., 28, 89* Sansome, F. W., 197* Satija, R. C., 100, 104, 108, 121, 142*, 143* Schantz, E. J., 44, 89* Schroeder, H. O., 56, 85* Schmidt-Nielsen, B., 301, 302, 303, 330* Scott, W. R., 13, 14, 15,41, 88* Shanes, A. M., 31,89* Shankland, D. L., 5, 23, 89* Shapira, A., 20,85* Shapiro, B. I., 39,87* Shaw, T. I., 50, 5 1, 80* Singer, I., 5 1, 9 1 * Singer, M., 106, 141* Sinha, R. N., 326* Sjostedt, Y..175, 197* Slifer, E. H., 154, 173, 195*, 197* Smith, H. W., 281,284, 330* Smith, J. N., 3, 89* Smyth, T., Jnr., 22, 66, 90* Socolar, S. J., 230, 327* Soliman, S. A., 22, 31,90* Spencer, E. Y.,3, 90* Spira, M. E., 99, 100, 102, 104, 105, 106, 107, 108, 110, 111, 113, 114, 115, 116, 117, 118, 119, 120, 125, 129, 132, 142*, 143* Spyropoulos, C. S., 50, 88* Staa1,G. B., 159, 178, 180, 182, 197*, 198*

AUTHOR INDEX

Stampfli, R., 38,90* Stehr, G., 156, 197* Sternberg, J., 5, 22,23, 58, 60, 61, 66, 82*, 90* Stower, W. J., 177, 197* Stobbart, R. H., 299, 303, 330* Stretton, A. 0. W., 114, 143* Susec-Michieli, S., 186, 187, 197* Sutcliffe, D. W., 289, 330* Sutherland, E. W.,210, 329* Suzuki, R., 70, 72, 9 1* Swilhart, S. L., 163, 198* T Tahori, S. S., 56, 66, 90* Takahashi, E., 120, 141* Takata, M., 39, 86*, 90* Takenaka, T., 5 1 , 9 1* Takeuchi, A., 20,90* Takeuchi, N., 20,90* Tasaki, I., 50, 51,88*, 91* Tatchell, R. J., 212, 330* Tauber, 0. E., 189, 193* Teorell, T., 50, 88* Terriere, L. C., 3, 9 1* Therrien, E. F., 39, 44, 87* Thomas, R. C., 21,85* Tindall, A. R., 100, 102, 110, 123, 141* Tischler, N., 27, 91* Tobias, J. M., 23, 91* Tomaszewski, W., 56, 91* Tomizawa, C., 27,82* Tormey, J. M., 236, 374, 326*, 330* Towe, A. L., 5, 89* Travaglino, A,, 70, 86* Treherne, J. E., 102, 110, 143*, 160, 196* Tsukamoto, M., 57, 66, 67, 68,70, 71, 72,91* Tunstall, J., 163, 191* Turner, R. S., 118, 140* Twarog, B. M., 28, 91* U Ulbricht, W., 16, 39, 86* Unger, H., 286, 330* Urakawa, N., 18,91*

339

Usherwood, P. N. R., 20,27,91* Ussing, H. H., 229, 244, 257, 275, 330* Uvarov, B. P., 146, 158, 166, 175, 176, 177,184, 198* V Van Bruggen, J. T., 257, 275, 326* Van den Bercken, J., 23,74,9 1* Vayvada, G., 44,89* Verdier, M., 153, 154, 198* Vidal-Sivilla, S., 24, 9 1 * Vinson,E. B., 56, 57, 58,91* Volkonsky, M. A., 154, 198* Vorontzovskii, P. A., 154, 198* W Walker, R. J., 20, 27, 28, 85* Wall, B. J., 287, 290, 299, 301, 302, 303, 304,328*, 330* Waloff, N., 176, 198* Waloff, Z., 155, 160, 168, 169, 176, 197* Walter, C., 156, 198* Wang, C. M., 25, 68,91* Ward, V. K., 172, 194* Waterston, A. R., 176, 198* Watson, J. A. L., 307, 324*, 330* Weiant, E. A., 22, 23, 66, 89*, 91* Weis-Fogh, T., 3 10, 330* Weissmann, M. L., 58, 8 1 * Welsh, J. H., 23, 26, 37, 83*, 92* Werman, R., 99, 110, 114, 115, 116, 117, 118, 119, 120, 129, 132, 142*, 143* Wessing, A., 276, 330* Wheeler, H. O., 274, 327* White, M. J. D., 153, 155, 195*, 198* Whitlock, R. T., 274, 327* Wiersma,C. A. G.,96, 136, 143*, 162, 198* Wigglesworth, V. B., 109. 143*, 205, 212, 238, 277, 286, 287, 289, 330*, 331* Wilcoxon, F., 61, 83* Wilkinson, C. F., 3,92* Williams, C. M., 210, 327*

340

AUTHOR INDEX

Willig, A., 171, 175, 184, 198* Wilson, H.G., 56, 85* Winteringham, F. P. W., 3, 69, 92* Wise, A. J., 168, 198* Wood, J. L., 206,207,270,290,33 1 * Woodbury, J. W., 5,89* Wright, E. M., 244, 245, 326* Y Yamada, M., 25, 68, 73, 91*, 92* Yamaguchi, T., 162, 198* Yamamoto, I., 3, 65, 92*

Yamasaki, I., 88*, 91 * Yamasaki, T., 11, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 38, 56, 57, 58, 59, 60, 61, 66, 67, 68, 70, 71, 72, 92*, 103, 129,143* Yates, W. W., 56, 93* Z Zawarzin, A., 101, 143* Zerahn, K., 207, 326*

Subject Index A Abdominal connectives, giant fibres, 100-1 Acanthacris ruficornis, coloration, 154, 159, 172, 174, 179 Acanthoxia, coloration, 149 Acetazolamide (Diamox), and Malpighian tubules, 226 Acetylcholine excitatory junctions, 18-19 giant fibres, 129 Acrida, coloration, 149, 158, 159, 160, 161, 164, 169, 170, 171, 172,174 A . bicolor, 178 A . turrita, 170 Acridinae, coloration, 149, 150, 15 1, 153, 158, 159, 162, 166, 179, 184 Acridoid grasshoppers, variable coloration 145-198, see Grasshoppers Acrotylus, coloration, 166 Adipose tissue, DDT accumulation, 58 Aedes, Malpighian tubules, 2 13 Aedes aegypti, rectum, 299 Aedes campestris, anal papillae, 2 12 Aeschna, giant fibres, 96, 101 Aeschnid nymphs, giant fibres, 103 Ailopus, coloration, 154 A . tergestinus, 176 A . thalassinus, 154, 156 Alanine, and Malpighian tubules, 279, 280 Aldrin, nerve and muscle changes, 24-6 Allethrin nerve and muscle changes, 45-56 action potential, 45-50 membrane ionic conductances, 50-6 structure - activity relationships, 75-6

Allethrin-con t. temperature coefficient, 6 1-5 Afferent inputs, giant fibres, 128-30 Amblycorypha o blongifolia, coloration, 153 Amino acids, and rectum, 304, 322 Ammonium ions, and Malpighian tubules, 244 Amphicremna, coloration, 149, 166 Anacridium, coloration A . aegyptium, 175 A . moestum. 176 Anal papillae mosquito larvae, 212 salt-water mosquito, 320 Anax imperator, giant fibres, 101, 102 Animals other than insects annelids,giant fibres, 96 arachnids excretion, 2 1 giant fibres, 96 arthropods, giant fibres, 96 Brunchioma vesiculosum, escape, 128 cat, motoneurones, 2 1 Corophium, coloration, 167 crab, DDT, 31 crayfish abdominal extensor muscles, 136 allethrin, 55, 75-6 claw, innervation, 136 DDT, 73-4 electrical synapse, 17 formed bodies, 276, 278 giant axons, 123, 128 1-glutamate, 20 voltage clamp experiments, 39 Crustacea DDT and calcium, 37

341

342

SUBJECT INDEX

Animals other than insects-cont. Antheraea pernyi labial gland excretion, 209, 210 Crustacea-cont. giant fibres, 9 6 midgut, oxygen, 270 Apotettix eurycephalus, coloration, muscle input, 136 157 Fish, euryhaline, ions, 2 12 Arginine, and Malpighian tubules, 279, flatworms, giant fibres, 96 280 frog Astaxanthine, and grasshopper coloraformed bodies, 277 tion, 184 ion transport, 258 Atractamorpha, coloration, 185 nerve, and DDT, 45 Auloserpusia, coloration, 15 1 neuromuscular junction, 18 Austriocetes, coloration, 176 sodium transport, 230 sulphate solutions, 229 water storage, 202 B goldfish, electrical synapse, 17 Gonyaulax, catanella, saxitoxin, 4 4 Barbistes fischeri, coloration, 153 Ligia, albedo response, 162 Beetle, uric acid, 204 lobster Bile pigments, and grasshopper abdominal extensor muscles, coloration, 171, 176, 177, 183, 136 186, 188 DDT, 38, 39,40, 4 3 , 4 4 Biliverdin, and grasshopper coloration, membrane currents, 13, 14 171, 175, 184, 187, 188 Lum bricus giant axons, 1 18 mesobiliverdin, 184 mammal Blattella germanica, uric acid, 204 kidney and dyes, 284 Boophilus microplus, excretion, 2 12 neuromuscular junction, 18, 20 pyrethroids, 65 C molluscs, giant fibres, 9 6 Calcium ions nemertineans, giant fibres, 96 and DDT, 37 polychaetes, giant fibres, 96 and Malpighian tubules, 215, 222, rabbit, gall bladder, 273 225 scorpion, giant fibres, 9 6 Calliphora erythrocephala Saxidomas giganteus, 44 spider, giant fibres, 96 labial glands, 2 10 Malpighian tubules, 2 16, 2 17-38, squid, giant axons, 17, 123 allethrin, 50-1 239-40, 242, 244-6, 248, DDT, 38 250, 251, 252, 263, 265, toad, water storage, 202 26, 276, 278-9 vertebrates Calpodes, Malpighian tubules, 237, CNS, axons, 136 2648,269 giant fibres, 96 C. ethlius, 216 kidney, 281 Cannula, coloration, 149, 159 Xenopus laevis, DDT, 74 Carausius m orosus Anions, Malpighian tubules, 320 coloration, 164, 171, 175 Calliphora, 2 19-20 cuticular lining, rectum, 307 Carausius, 228, 229, 230 hindgut, 287 Malpighian tubules, 2 13-6, 226-3 1, Rhodnius, 244-7 234, 236, 239-40, 242, Antennae and giant fibres, 130-2

SUBJECT INDEX

Carausius morosus-cont. Malpighian tubules-cont. 244-5, 248, 250, 251, 263, 265, 266, 276, 278, 284-6 Carbon dioxide, and grasshopper coloration, 182 Carotene, and grasshopper coloration, 185, 189, 190 /3-carotene, 184 Carotenoids, grasshopper coloration, 183, 184, 187, 188, 189, 190 Catantopinae, coloration, 146, 149, 150, 151, 154, 159, 160, 164, 158, 172, 176, 178, 185, 189 Catantops kissenjianus, coloration, 168 Cerci and giant fibres, 124, 129 Chironomus, Malpighian tubules, 230 Chloride ions Malpighian tubules, 2 15, 277 Calliphora, 2 2 3, 2 32 Calpodes, 266-7 Rhodnius, 246, 249-52, 255-63 Tipula, 2 38 rectum, Schistocerca, 293-5 water absorption, Tenebrio, 3 1 1, 314-7 Cholinesterases and insecticides, 2, 27, 31 Choristoneura, coloration, 156 Chorthippus, coloration, 150, 154, 170, 176 C. albomarginatus, 154, 175 C. brunneus, 154, 168 C. parallelus, 154, 169 Chortoicefes, coloration, 150, 164, 172 C. terminifera, 154, 156, 177 Chromoprotein and light, 171 Chrotogonus, coloration, 150, 159, 188 Cicadids, coloration, 167 Cockroach action potential, 7 allethrin, 45-9, 5 1 cyclodienes, 25, 26 DDT, 2 2 , 2 3 after potential, 31-6

343

Cockroach-cont. DDT-con t. repetitive discharge, 38 structure - activity, 73 temperature coefficient, 57, 58-61 dieldrin, 69 giant fibres afferent inputs, 129 histology, 100, 102-3, 105, 107-8, 109 integration, 136-9 leg motoneurones, 124, 128 membrane properties, 1 10 Outputs, 130-5 system, 97-100 through conduction, 1 10-2 1 timing relations, 135-6 organophosphates, 27-8, 29-3 1 pyre t hroids, 2 6-7 uric acid, 204, 206, 319 Cocoonase, 2 10 Coleoptera excretory system, 310 Malpighian tubules, 283 Coloration, variable, Acridoid grasshoppers, 145-98 environmental factors, 156-77 genetic factors, 152-6 natural history, 147-52 physiological mechanisms, 177-83 pigments, 183-90 terminology, 146-7 Conduction, giant fibres, 1 10-21 collision experiments, 1 10-4 low safety factor zones, 1 1 4 2 1 mole cricket, locust, 121 Connectives, giant fibres abdominal, 100-1 thoracic, 104-6 Conocephalidae, coloration, 153 Copper Malpighian tubules, 258 phosphate transport, 234 Coptacridinae, coloration, 147, 15 1, 154 Corpus allatum, and grasshopper coloration, 178-80

344

SUBJECT INDEX

Corpus cardiacum, and grasshopper coloration, 180-1, 182 Coryphosima (=Paracomacris), coloration, 151, 186 C. amplificata, 150 Crowding, and grasshopper coloration, 164-6, 173-4, 180, 182 Cryptonephridial system, Tenebrio, 310-9, 323-4 Cupric ions, and phosphate transport, 234 Cuticular lining, rectum, 304-7, 318, 322 Cyclodienes, nerve and muscle changes, 24-6 Cyphocerastis, coloration, 15 1, 154 Cyrtacanthacridinae, coloration, 147, 151, 154, 155, 159, 164, 168, 172, 174, 175, 176, 177, 179 Cyrtacanthacris, coloration, 172 C. tartarica, 174, 176 D DDT action, 4 artificial membranes, 7 9 detoxication, 57-8, 201 nerve and muscle changes, 21-3 mechanisms, 31-45 nerve, sensitivity, 58-60 receptors, 78 resistance, 67-8, 201 structure - activity relationships, 73-5, 79-80 temperature coefficient, 56-61 Degeneration, giant fibres, 106-8 Deposit and storage excretion, 201-4, 319 Detoxication, DDT, 57-8, 201 Diazinon, resistance, 69-70 Dictyophorus, coloration, 152 Dieldrin nerve and muscle changes, 24-6 receptors, 78 resistance, 68-9 Dixippus, coloration, 184 Dociostaunrs maroccanus, coloration, 176

Dragonfly, giant fibres, 98, 99, 101 Dropherula werneriana, coloration, 150 Drosophila giant axons, 96 Malpighian tubules, 276 DPNH, and rotenone, 76-7 Duronia, coloration, 154 Dyes, and Malpighian tubules, 280-6, 32 1 Dysdercus Malpighian tubules, 236, 237 storage excretion, 3 19 uric acid, 204, 3 19 water storage, 202-3 Dytiscus, giant fibres, 102 E Ecdysis, and labial glands, 2 10-1 Endocrine control, grasshopper coloration corpus allatum, 178-180, 182-3 corpus cardiacum, 180-1, 182 juvenile hormone, 179, 181, 183 others, 181-3 Ephippiger provincialis, coloration, 153 Ephidermis and CNS, 178 Escape behaviour and giant fibres, 123-8 Aeschna, 96 Branchioma, 128 cockroach, 97, 120, 130-9 crayfish, 128. Drosophila, 96 Gryllotalpa, 12 1 locust, 128 Eserine, and e.p.s.p., 30 Eumastacid Morabine grasshoppers, coloration, 153, 185 Euryphyminae, coloration, 151 Excretory systems, mechanisms, 199-331 anal papillae, 2 12 deposit and storage excretion, 201-4 hindgut, 286-319 labial glands, 209-12

SUBJECT INDEX

3 45

Excretory systems-cont. Giant axons-cont. Malpighian tubules, 2 12-86 through conduction, 1 10-2 1 midgut, 206-9 timing relations, 135-6 pericardial cells and nephrocytes, Glucose, and Malpighian tubules, 279, 280 205-6 Eyprepocnemidae, coloration, 147, 1-glutamate, as transmitter substance, 20, 27, 31 151, 159 Eyprepocnem is m on tigena, coloration, Glycine, and Malpighian tubules, 279, 280 15 1 Gomphocerinae, coloration, 147, 149, 150, 151, 154, 155, 159, 160, F 165, 168, 169, 170, 175, 176, 179, 186 Faureia m ilanjica, coloration, 175 Gomphocerus rufus, coloration, 179 Flavins, and grasshopper coloration, Grasshoppers, Acridoid, variable 183 coloration, 145-98 Forficula auricularia, Malpighian environmental factors, 156-77 tubules, 281-3 genetic factors, 152-6 Formed bodies, Malpighian tubules, 276-9, 321 natural history, 147-52 Fructose, and Malpighian tubules, 279, physiological mechanisms, 77-83 280 pigments, 183-90 terminology, 146-7 Gregarious locust hoppers, coloration, 166, 173-4, 175-7, 180-2, 185, G 186, 189-90 Gamma aminobutyric acid, and Grylloidea, coloration, 153 impulse transmission, 20 Gryllotalpa, giant fibres, 101, 102, Ganglia, and giant fibres 104, 105, 121 abdominal, 102-4 Gryllus bimaculatus, coloration, 156, thoracic, 104 183, 186 Gastrimargus, coloration, 150, 154, Gymnobothris, coloration, 154 159, 160, 161, 162, 165, 166, 171, 172, 173, 186, 188, 189 G. africanus, 154, 156, 172, 176, 179 H G. musicus, 176 Gemeneta, coloration, 150 Hemiacridinae, coloration, 147, 149, Genes 154 grasshopper coloration, 152-6 Hemiptera, CNS and epidermis, 178 insecticide resistance, 70-2 Heteracris vinaceus, coloration, 15 1 super-gene coloration, 153 Heteroptera, Malpighian tubules, 283 Giant axons, functional organisation, Heteropternis, coloration, 154, 159, 95-143 186 afferent inputs, 128-30 H. couloniana, 156 giant fibre outputs, 130-5 Hindgut, excretion, 286-319, 321 histology, 100-10 amino acids, sugars, 304 leg motoneurone activation, 12 1-8 anterior to rectum, 287-9 membrane properties, 1 10 Calliphora, 295-6 cuticular lining, 304-7 role in integration, 136-9

3 46

SUBJECT INDEX

Hindgut-cont. ion absorption, 303-4 rectum, action, 289-9 1 Schistocerca, 291-5 Tene brio, 3 10-9 Thermobia, 307-10 water absorbtion, 296-303 Homorocoryphus nitidulus, coloration, 153 Hormones corpus cardiacum, and crowding, 177 diuretic, Malpighian tubules, 238, 254, 260-2 grasshopper coloration corpus allatum, 178-80, 182-3 corpus cardiacum, 180-1, 182 juvenile, 179, 181, 183 others, 181-3 rectal ion absorption, 303-4 Housefly DDT, 22 insecticide resistance, 66-72 parathion, 31 Humbe, coloration, 154, 159 H. tenuicornis, 156, 179 Humidity, and grasshopper coloration, 169, 172-3 Hyalophora cecropia, excretion labial glands, 2 10 midgut, 206 Hygroreceptors, grasshoppers, 173 I Inputs, afferent, giant fibres, 128-30 Insecticides and excitable tissues, 1-93 changes in nerve and muscle, 2 1-31 mechanisms, 3 1-56 insecticidal action, 3-5 molecular mechanisms, 78-80 nerve excitation, 5-2 1 resistance, 65-72 structure - activity relation, 72-8 temperature coefficient, 56-65 Integument, penetration by DDT, 56-7 Intestine, small, ion transport, 258 Ions absorption, rectum, 322

Ions-cont. absorption- cont. Calliphora, 295-6 Schistocerca, 291-5 mechanism, 303-4 anions, Malpighian tubules Calliphora, 2 19-20 Carausius, 228, 229, 230 Rhodnius, 244-7 ionic conductances membrane, DDT, 38-45 ionic mechanisms, neuromuscular junctions, 20-1 ionic properties, giant fibres, 110 ionic pump, giant fibres, 120 Ixalidium haernatoscelis, coloration, 151

K Kingangopa jeanneli. coloration, 15 1 Kosciushola, coloration, 146 1

Labial glands, silkmoths, 209-12, 320 Leg motoneurones, and giant fibres, 121-8 Lentulidae, coloration, 147 Lepidoptera coloration, 147, 163, 168, 185 excretory system, 3 10 Malpighian tubules, 283 Leptacris, coloration, 149 Libellula, Malpighian tubules, 270 Light, and grasshopper coloration, 160-4, 170-2 Lindane nerve and muscle changes, 23-4 resistance, 68 Locust giant fibres continuity, 12 1 continuity, 12 1 histology, 100, 102, 104, 108, 109 and leg motoneurones, 122, 128 labial gland excretion, 2 10 organophosphates, 28

347

SUBJECT INDEX

Locusta coloration, 150, 158, 159, 161, 162, 164, 165, 166, 169, 170, 171, 172, 175, 177, 178, 179, 180-1,182 L. migratoria, 154, 183, 184, 185, 186, 187, 189, 190 L. pardalina, 154 giant fibres, 100 Locustana, coloration, 150, 158, 166, 172, 175, 177 L. pardalina, 156, 182

M Machaeridia, coloration, 149 Magnesium ions, and Malpighian tubules Calliphora, 222, 225 Carausius, 2 15 Malpighian tubules, 2 12-86 Calliphora, 2 16-38 Calpodes, 264-8 Carausius, 2 13-6 formed bodies, 276-9 organic solutes, 279-86 Rhodnius, 238-64 Tipula, 238 ultrastructure, 268-76 Manduca, coloration, 185 Mantis, coloration, 174, 186, 187 M. religiosa, 171, 187 Mantoids, coloration, 167, 185 Mauthner fibres, 123 Melanins and grasshopper coloration, 183-9 Melanoplus, coloration, 159, 160, 163, 185 M. bivattatus, 164 M. sanguinipes, 154, 172, 176, 189 Membranes artificial, and insecticides, 79 ionic conductances and allethrin, 50-6 and DDT,38-45 properties, giant fibres, 1 10 Mesopsera, coloration, 149 Mesopsis. coloration, 159, 165 M. laticornis, 149, 168

Metabolism, and nerve excitation, 16-17 Midgut, silkworm, excretion, 206-9 Mole cricket, giant fibres continuity, 121 histology, 101, 102 leg motoneurones, 122 Molecular mechanisms, insecticides, 78-80 Morphacris, coloration, 15 3 Mosquito DDT, 74 larvae, and papillae, 212 salt-water, anal papillae, 320 Moth, saturniid labial glands, 209-12, 320 Potassium, 319 Motoneurones, leg, and giant fibres, 121-8 Muscle changes, insecticides, 2 1-31 mechanisms, 31-56 N

Nemeritis canescens, uric acid, 203 Nephrocytes and pericardial cells, excretion, 205-6 Nerves changes, insecticides, 21-31 mechanisms, 3 1-56 excitation, 5-2 1 sensitivity to insecticides, 66-70 Nervous system, structure, 5-6 Neuromuscular j unctions classification, 17 excitatory, 18-20 inhibitory, 2 0 ionic mechanisms, 20-1 Nomadacris septemfasciatum, coloration, 175 Noradrenaline, 20

0 Occidentosphena, coloration, 150 Odonata, giant axons, 96 Odontomelus, coloration, 15 1 Oedalus, coloration, 150, 159, 160

348

SUBJECT INDEX

Oedipoda, coloration, 150, 158, 159, Parathion action, 4 160, 161, 166, 167, 185, 186, and nerves, 3 1 187 Oedipodinae, coloration, 147, 149, Pattern, grasshopper, coloration, 150, 153, 154, 156, 158, 159, 189-90 160, 164, 165, 166, 169, 172, Pericardial cells, excretion, 205-6, 283 175, 176, 177, 179, 183, 186, Periplaneta americana 188 excretion Ommochromes, and grasshopper hindgut, 287 coloration, 161, 183, 185, 186, rectum, amino acids and sugars, 187, 188, 189, 304 Omocestes viridulus, coloration, 169 rectum, water absorption, 299, Orbillus coeruleus, coloration, 15 1 300-2, 304, 3 14, 3 18 Organic molecules, size uric acid, 204 and hindgut, 304 giant fibres and Malpighian tubules, 32 1 and cerci, 97 and rectum, 322 histology, 102 Organic solutes, and Malpighian through conduction, 1 10-2 1 tubules, 279-86 Periplaneta orientalis, Malpighian complex, 280-6 tubules, 28 1-3 low MW, 279-80 Peroxidase, horseradish, and MalOrganophosphates pighian tubules, 270 Malpighian tubules, 247-9 Pezocarantops, coloration, 150, 15 1 nerve and muscle changes, 27-3 1 pH, and Malpighian tubules, 222, 226 Ornithacris, coloration, 159 Phaneropteridae, coloration, 153 0. turbida, 174 Phase coloration, grasshoppers, 175-7 Orphania denticauda, coloration, 153 Phasmids, coloration, 167 Orphania scutata, coloration, 153 Phoremula, coloration, 159 Orthoptera, giant fibres Phoromids, giant fibres, 96 afferent inputs, 129 Phosphate ions and Malpighian histology, 101, 104 tubules, 320 through conduction, 12 1 Calliphora, 220-1. 223, 234, 252, Ouabain, and Malpighian tubules, 226, 278 Carausius, 215, 234, 278 228,233 Phyrnateus, coloration, 15 1, 152, 173 Outputs, giant fibres, 30-5 Oxya hyla, coloration, 149 Phymeurus, coloration, 150 Oxyinae, coloration, 149, 151 Phy teum as purpurascens, coloration, 146, 173 Pieris brassicae, nitrogenous wastes, 20 1 P Poecilocerus, coloration, 15 1 P. hieroglyphicus, 164 Pachynotacris amethystinus, coloraPolymorphism, grasshopper coloration tion, 151 genetic, 152-5 Paracinema tricolor, coloration, 149 Paracoptacra, coloration, 15 1 greenlbrown, 167-75 Parapropacris rhodopterus, coloration, brown component, 186-8 151 and corpus allatum, 178-80 Parasphena, coloration, 150, 159, 188 green component, 184-5 Paratettix texanus, coloration, 152 implications, 188

349

SUBJECT INDEX

Polymorphism-con t. phenotypic, genetic modification, 155-6 Potassium ions and allethrin, 5 1-6, 76 and DDT, 40-1, 45, 79 and excretion, midgut, 206-9, 21 1 giant fibres, 120 Malpighian tubules, 277, 320 Calliphora, 2 17-8, 225, 226, 227, 229, 230-5, 237, 239-40, 263 Calpodes, 264-8 Carausius. 2 13-5. 239-40. 263 Rhodnius, 239-44, 247,’249-53, 255-63 Tipula, 238 nerve excitation, 7-1 1, 14-6 rectum, 293-5 saturniid moths, 3 19-20 synaptic transmission, 20-1 water absorption, Tenebrio, 3 1 1, 3 14-7 Potentials action, and allethrin, 45-50, 63-5 action, mechanism, 7-1 1 after-, and allethrin, 46-8 after-, and DDT, 31-7 e.p.s.p., mechanism, 19 i.p.s.p., mechanism, 20 resting, mechanism, 6-7 transwall, Malpighian tubules, 253-63 Proline, and Malpighian tubules, 279, 280 Prothoracic glands, and grasshopper coloration, 182 Pteridines, Pieris, 201 Pterines, and grasshopper coloration, 183 Pterotiltus, coloration, 15 1 Pyrethroids nerve and muscle changes, 26-7 structure - activity, 75-6 temperature coefficient, 6 1-5 Orgodera armata, coloration, 175 Pyrgomorpha, coloration, 1 79 P. cognata, 172

Pyrgomorphidae, coloration, 146, 147, 150, 151, 159, 164, 173, 179, 185,188 R Radiation, and grasshopper coloration, 170-2, 173 Rectum, excretion, 289-91, 321-2 Repetitive discharge and allethrin, 46-8, 62 and DDT, 37-8 Resistance t o insecticides, 65-72 Retinula cells, and grasshopper coloration, 163-4 Rhodnius, excretion hindgut, 289 Malpighian tubules, 213, 21 6, 236, 237, 238-63, 266, 268, 284-6, 320 anions, 244-7 formed bodies, 277, 278 ultrastructure, 270 water movements, 247-63 pericardial cells, 205 Roach, iso-osmotic fluid transport, 257 Roduniella, coloration, 154 Romalinae, coloration, 151 Rotenone nerve and muscle changes, 27 structure - activity, 76-8 Ruwenzoracris, coloration, 15 1 S

Salivary glands, formed bodies, 32 1 Saxitoxin, 4 3 , 4 4 Schistocerca gregaria coloration, 154, 156, 166, 172, 175, 177, 178, 180, 181, 182, 183, 184, 185, 186, 187, 189,190 Malpighian tubules, 276 rectum cuticular lining, 304-5, 307 ion and water absorption, 290-5, 299, 300, 303, 314, 318

350

SUBJECT INDEX

T Schistocerca obscura, coloration, 164, 174.175.176 Taste receptors, and DDT, 67 Schistocerca paranensis, coloration, Temperature 172 and grasshopper coloration, 164, Schradan, and nerves, 3 1 174-5 Serpusia, coloration, 151 and Malpighian tubules, 260 Sheath, giant fibres, 101-2 Temperature coefficient, insecticides Silkmoth, saturniid, excretion DDT, 56-61 labial glands, 209-1 2 pyrethroids, 61-5 midgut, 206-9 Tenebrio molitor, water-uptake, 308, Sodium ions 310-9, 323-4 and allethrin, 49-56, 76 Tetra-ethyl ammonium ions and DDT, 34-7,41-5,80 Malpighian tubules, 25, 261-2 giant fibres, 120 and potassium current, 15 Malpighian tubules, 277, 320 Calliphora, 217-8, 226, 227, Tetrodotoxin (TTX) and DDT, 15, 41-3 229, 236, 232, 2345, 237-8, Tettigidea parvipennis, coloration, 1 5 1 239-40, 263 Tettigonia, pigments, 184 Calpodes, 2 6 4 8 Tettigoniidae, coloration, 153 Carausius, 214-5, 239-40, 263 Tettigonioidea, coloration, 153, 167 Rhodnius, 239-44, 247, 249-53, Tettrigidae, coloration, 152 256-63 Thermobia Tipula, 2 38 hindgut, excretion, 287 midgut excretion, 206 water uptake, 323, 324 nerve excitation, 7-1 1, 14 Thermobia domestica (Lepismodes inactivation, 14-6 inquilinus), water absorption, pump, 16-7 307-1 0, 3 1 1 rectum, excretion, 293-5 Tipula paludosa, Malpighian tubules, synaptic transmission, 20-1 234,238,266 and water absorption, 3 1 1, 3 14 Transmitter substances, giant fibres, Somata, giant fibres, 108-10 129, 138-9 Spathosternum, coloration, 154 Trilophidium, coloration, 150 Sphodromantis, coloration, 17 1 Stenobothrus linealus, coloration, 169 Truxalinae acoustical signals, 149 Stick insect, Malpighian tubules, 274, coloration, 149, 153 276 Truxalis, coloration, 149, 153 Storage excretion. 201-4. 319 d-tubocurarine, and excitatory juncStructure - activity relationships tions, 18 DDT. 73-5 Tylotropidius, coloration, 159 pyrethroids, 75-6 rotenone, 76-8 Sucrose, and Malpighian tubules, 279, U 280 Uganda kilimanjarica, coloration, 150 Sugars, rectal recovery, 304 Sulphanilamide, and Malpighian Urea, and Malpighian tubules, 279, tubules, 226 280 Super-gene, grasshopper coloration, Uric acid, excretion, 201, 203-4, 206, 153 211,289,319 Syrbilla fuscovittata, coloration., 179 . Uriculi majores, 204, 206, 21 1 I

r

-

35 1

SUBJECT INDEX

V Valanga irregularis, coloration, 154 Valine, and Malpighian tubules, 279, 280 Voltage clamp experiments, 11-16 allethrin, 50, 75-6 DDT, 38-45

Water-cont. absorption-con t. Tenebrio, 3 10-9 Thermobia, 307-10 movements, Malpighian tubules, 247-63, 270-5 storage, 202-3 Wingless- . grasshoppers, coloration,

W

Wings, Pieris, nitrogenous wastes, 2 10

Water absorption, rectum, 322 Calliphora, 295-6 mechanism, 296-303 Schistocerca, 29 1-5 absorption, unsaturated atmospheres, 323

150-1

X Xanthommatine, and grasshopper coloration, 186, 187

This Page Intentionally Left Blank

Cumulative List of Authors Numbers in bold face indicate the volume number o f 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, 13 1 Bursell, E., 4, 33 Burtt, E. T., 3, 1 Carlson, A. D., 6,51 Catton, W. T., 3, 1 Chen, P. S., 3, 53 Colhoun, E. H., 1, 1 Cottrell, C. B., 2, 175 Dadd, R. H., 1 , 4 7 Dagan, D., 8, 96 Davey, K. G., 2, 219 Edwards, John S.,6,97 Fraser Rowell, C. H., 8,146 Gilbert, Lawrence I., 4 , 6 9 Goodman, Lesley, 7 , 9 7 Harmsen, Rudolf, 6,139 Harvey, W. R., 3, 133 Haskell, 3. A., 3, 133

Hinton, H. E., 5 , 6 5 Hoyle, Graham, 7,349 Kilby, B. A., 1, 11 1 Lawrence, Peter A., 7, 197 Lees, A. D., 3, 207 Maddrell, S. H. P., 8, 200 Miller, P. L., 3, 279 Narahashi, Toshio, 1, 175, 8, 1 Neville, A. C., 4,213 Parnas, I., 8, 96 Pringle, J. W. S., 5, 163 Rudall, K. M., 1,257 Sacktor, Bertram, 7, 268 Shaw, J., 1 , 315 Smith, D. S., 1 , 4 0 1 Stobbart, R. H., 1 , 3 1 5 Treherne, J. E., 1,401 Usherwood, P. N. R., 6,205 Waldbauer, G. P., 5, 229 Weis-Fogh, Torkel, 2, 1 Wigglesworth, V. B., 2 , 2 4 7 Wilson, Donald M., 5, 289 Wyatt, G. R., 4, 287 Ziegler, Irmgard, 6, 139

353

This Page Intentionally Left Blank

Cumulative List of Chapter Titles Numbers in bold face indicate the volume number o f the series

Active Transport and Passive Movement of Water in Insects, 2,67 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, 1 1 1 Biology of Pteridines in Insects, 6, 139 Cellular Mechanisms Underlying Behaviour-Neuroethology, 7, 349 Chitin Orientation in Cuticle and its Control, 4, 213 Chitin/Protein Complexes of Insect Cuticles, 1, 257 Colour Discrimination in Insects, 2, 13 1 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, 219 Effects of Insecticides on Excitable Tissues, 8, 1 Electrochemistry of Insect Muscle, 6, 205 Excitation of Insect Skeletal Muscles, 4, 1 Excretion of Nitrogen in Insects, 4 , 3 3 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, 7, 1 Functional Aspects of the Organization of the Insect Nervous System, 1,401 Functional Organizations of Giant Axons in the Central Nervous Systems of Insects: New Aspects, 8 , 9 6 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 Mechanisms of Insect Excretory Systems, 8, 200 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, 175 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 Variable Coloration of the Acridoid Grasshoppers, 8, 146 355

This Page Intentionally Left Blank