,■

::

(• \>.

i:

■4

i! ■

Hi

Marine Biological Laboratory bbrary

Voods Hole, Massachusetts

Presented by the MBL Associates-1971 Gift

methuen's monographs on biological subjects

General Editor: Kenneth mellanby

THE PHYSIOLOGY OF INSECT SENSES

The Physiology
of Insect Senses

V. G. DETHIER

Professor of Zoology and Psychology
University of Pennsylvania

London: METHUEN & CO. LTD
New York: JOHN WILEY & SONS INC

First published September 1963

Reprinted 1964

1.2

© 1963 Vincent G. Dethier

Printed in Great Britain by

Richard Clay (The Chaucer Press), Limited

Bungay, Suffolk

Catalogue No, 12/4125/62

Contents

PREFACE Vii

I INTRODUCTION J

II GENERAL CHARACTERISTICS OF THE SENSORY

SYSTEM 5

III MECHANORECEPTION 26

The Tactile Sense-Sensilla Trichodea - Proprioceptors -
The Role of Mechanoreception in Locomotion -
Responses to Gravity

IV SOUND RECEPTION 77

Tympanic Organs - Hairs as Sound Receptors - The
Johnston's Organ of Culicidae - Perception of Vibra-
tions in Solids

V CHEMORECEPTION 112

The Receptors - Olfaction - Contact Chemoreception

VI RESPONSE TO HUMIDITY 156

VII PHOTORECEPTION 162

The Compound Eye - Colour Vision - Polarized Light
- Form Perception - Ocelli - Stemmata

REFERENCES 214

INDEX 256

Preface

Thirty years have elapsed since the appearance of Eltringham's
The Senses of Insects (first published 1933 in this series). In the inter-
vening time our knowledge of sensory physiology has advanced
further than during any previous period of study. The augmented
and accelerated advance may be attributed to new technological
developments and a notable increase in the number of research
workers interested in insects for their own sake or as material
uniquely suited to the solution of one or another basic biological
problem.

The two most powerful modern tools placed at the disposal of the
sensory physiologist are electronic apparatus for detecting electrical
events in nerve tissue and the electron-microscope. The first has been
responsible for removing from the realm of mystery the function of
so-called Type II neurons in insects, for confirming the facts that
campaniform sensilla are mechano- rather than olfactory receptors,
and for initiating the unraveUing of the multitude of events that occur
as a part of vision. Electric recording from neural tissue has greatly
expanded our rather conservative estimation of sense organs in
general.

Developments in the field of electronics has not, however, ren-
dered the behavioural approach to sensory physiology obsolete.
In fact, experiments in behaviour not infrequently point the direc-
tion that electrophysiological research should follow. Two cases in
point are the studies of the tympanic organ of moths and of the
chemoreceptors of flies. Without behavioural correlates the electro-
physiological findings are of small value.

Just as behavioural studies have engendered electrophysiological
experiments, these in turn have driven the physiologist back to
structure. Whereas in many instances conventional histology has
been extended to its limit, in others it has barely scratched the sur-
face, and large areas of the nervous system are still empty areas
whose only detail is the notation 'unexplored'. Where light micro-
scopy has reached the limit of resolution, the electron-microscope
has opened up vast new areas of investigation and stimulated addi-
tional electrophysiological studies, which in turn have posed new
behavioural questions to bring the endeavour full turn.

vii

Vm PREFACE

We are in the midst, therefore, of a revolution in sensory physio-
logy. There is no better evidence for this than the high element of
controversy that flourishes in many areas indicating that exploration
is still in its early stages. This state of affairs is best illustrated by the
controversies in photoreception. The final verdict is by no means
rendered. In many cases our knowledge is still in a state of flux.
While every effort has been made in this book to present both sides
of questions, a personal bias has been applied. This bias is fashioned
by an appraisal of how well the antagonists have presented their
cases in the literature.

Recent advances in the physiology of the insect sensory system
have proceeded at such a pace that it is impossible to refer to every
published report. In the field of sound reception alone there are over
2,000 articles. For all phases of this book, about 2,000 references in
toto have been examined. Of these, about 700 have been selected for
listing, but the accumulated knowledge of others has been built upon
freely. For the reader interested in the older literature, the works of
Demoll and von Buddenbrock should be consulted. For all purposes,
the works of Wigglesworth and Roeder are invaluable. Where
possible, references have been made to reviews rather than to the
multitudes of individual papers. Where an investigator has published
a series of papers on a given subject, only the most recent has been
quoted unless an eariier paper is germane to some particular point.
During the period of preparation of this book insect physiologists
have not been idle. The most recent papers have probably been
treated somewhat cavalierly, since a running revision is hardly
feasible, but at the very least they are included in the references to
assist the reader in pursuing the matter himself.

The segment of knowledge represented by this book has been
rather narrowly circumscribed, a circumstance dictated by the enor-
mous volume of literature but ameliorated by the fact that other
excellent volumes exist to fill the lacunae. The subject matter has
been 'sensory physiology'; behaviour, as such, has been neglected.
It is well represented, however, by Carthy's fine book, by Roeder,
by Wigglesworth, and by von Buddenbrock. Also, I have not ven-
tured farther into the central nervous system than the lamina gang-
lionaris of the optic lobes, partly because I have been interested
primarily in sensory input and the interaction of sense organs and
their environment and possibly because one naturally hesitates to
venture far into an unexplored jungle. And finally, the temperature
sense has been completely neglected. This, in part, merely reflects the

PREFACE ix

neglect that it has received at the hands of research investigators.
There is considerable Hterature on the behavioural aspects of tem-
perature sensitivity, little of value on the physiological aspects.

The co-operation and courtesy of a number of persons and organi-
zations has been invaluable in the preparation of this book. Paul B.
Hoeber, Inc. has granted permission to reprint sections of Chapter
II which formerly appeared in Roots of Behavior, edited by E. L.
Bliss. Similarly, the Society for Experimental Biology has granted
permission to reprint sections of Chapter V which appeared in
Symposium No. 16of the Society for Experimental Biology. Figs. 17,
18, 21, and 25 were supplied through the courtesy of M. L. Wolbarsht ;
Figs. 40 and 41, through the courtesy of O. Lowenstein and L. H.
Finlayson; Figs. 56, 65, 66, and PI. I, through the courtesy of K. D.
Roeder and A. E. Treat; PI. Ill, through the courtesy of J. R. Adams.
Peter G. Walsh has assisted in the preparation of illustrations. To
all of these I express my sincere gratitude.

V. G. Dethier

E. Blue Hill, Maine
June 1962

CHAPTER I

Introduction

The gross characteristics of the planet earth are the same for all living
things inhabiting it. It has a characteristic gravitational field; it re-
ceives radiant energy from outer space to the extent determined by the
filtering properties of its atmosphere; it is built of specific kinds of
chemical compounds in certain quantitative relationships. The details
of these properties differ from place to place on the planet. Some are
more conducive to animal life than others, and the uneven distribution
of animals in the biosphere is a measure of this and of the adjustments
that various forms of life have made to the details. To regulate their
relations with the earth's environment, animals have evolved sensing
devices for detecting such details as are of direct adaptive value.

Since we *see' the world with our own particular sensing devices, it
is difficult to understand fully how other animals *see' it. What picture
of the details of the world is perceived by the bee that sees in the ultra-
violet, the snake that is deaf to air-borne sounds but detects infra-red,
the shrew that cannot see colour, the electric fish that detects changes
in the electric field about it? To appreciate these points of view, let
alone understand the manner in which all biological sensing devices
work, it is helpful to look first at the world, not at the level of detail
perceived by us but at a level that we cannot perceive directly, the
elementary particles whose interactions and combinations form
the universe.

Of the thirty or so kinds of particles that make up earth or come to
it from outer space, the only stable ones are protons, neutrons,
neutrinos, electrons, and photons (Ruderman and Rosenfeld, 1960).
Some of these (e.g., photons, neutrinos) occur free; others, usually
combined into atoms and molecules. Some, such as electrons and
photons with certain energy, have profound effects on living material ;
others, such as neutrinos, pass through organisms undetected and
without effect. All have different kinds of energy. In the final
analysis it is to these particles that animals must make adjustments.
Organisms living in other parts of the universe would be exposed to
other kinds of particles and would have to evolve differently, each

in its own milieu.

1

2 THE PHYSIOLOGY OF INSECT SENSES

Pertinent details of an animal's environment are detected first
by utilizing the energy of particles to perform work in the biological
system. Thus, light is detected by virtue of an absorber which trans-
forms the energy of photons of certain specific energy. There is no
absorber to transform the energy of photons which, for example,
make up X-rays and gamma-rays. Heat is detected by absorbing the
energy of photons of a different energy level. Electricity is detected by
utiUzing the energy of electrons. Tastes and smells are detected by
utilizing the potential energy existing in the mutual attraction and
repulsion of the particles making up atoms. Sound is detected by
using the energy of moving particles of molecular size.

A sense organ is a part of an organism specialized to receive a small
amount of energy from certain of the sources mentioned above and to
utiUze it to set off a train of events culminating in a nerve impulse.
Many cells and parts of cells do work with energy derived from the
sources mentioned. Many cells use the potential energy in carbo-
hydrates to do work, yet they are not sense organs. Certain of the cells
of green plants absorb photons of certain energies and perform work
much as the pigment rhodopsin of the eye absorbs protons and does
work. A sense cell differs from these in at least three major respects :
first, the work which is done by the sense cell is done at the expense of
its own potential energy, the energy of the environment is merely a
trigger; second, all sense organs, as far as we know, transform their
potential energy into electrical energy; third, they transmit this
electrical energy to another cell. And this energy, in turn, is transmitted
by an element which by itself could not have detected the original
environmental energy.

A sense organ implies another element that will receive the change.
In this sense our attitude towards sense organs is teleological, but
there seems to be no escape from this attitude. As Granit (1955) has
concluded: Tor many purposes, e.g. physiochemical studies of
primary events, we can neglect the teleological aspects of the sensory
message and the general problem of central decoding of the code of
spikes, but I want to emphasize that research into special senses
differs from many other recognized branches of physiology in pre-
supposing and accepting the fact that understanding of biological
purpose is part of its aim, be it movement or perception. To close
one's eyes to this aspect of sensory physiology is to neglect the
biological, psychological, and philosophical implications of a branch
of natural science which actually is capable of giving some meaning
to "meaning'

.'♦ >

INTRODUCTION 3

Receptors change reversibly when energy is apphed to them (i.e.,
they are sensitive, they detect), they do work (i.e., they are responsive)]
they generate a message to be transmitted beyond their boundaries.
Thus, there are three levels at which the physiology of receptors may
be considered.

Considered first as detectors, receptors would not make much
information available to an organism if they were so imperfect as to be
insensitive to any but the greatest energy changes or so perfect as to
be sensitive to every single elementary particle. Furthermore, their
usefulness would be limited if they were so indiscriminate as to detect
equally all kinds of particles and so simple as to detect only the pres-
ence or absence of a stimulus. From the point of view of an animal's
survival, there must be an optimum sensitivity, a capacity for discrim-
inating different kinds of stimuli, and a capacity for measuring not
only on-off, but rate of change, magnitude of change, absolute change,
and direction of change.

It is characteristic of protoplasm that it has moderate sensitivity to
many forms of stimuli and may be sensitive to more than one para-
meter of the stimulus. But receptors have become specialized in that
they possess enhanced sensitivity to some particular form of energy
and to some particular parameter. First, there is development of
special characteristics at the molecular level. The presence of a pig-
ment to absorb radiant energy in some particular wavelength
(rhodopsin in the rods and cones of the vertebrate eye) or the presence
of molecules to uncouple the potential energy in sugar molecules
(sweet taste receptors) are more well-known examples.

In addition to specialized sensitivity per se, receptors have also
become surrounded with accessory structures whose presence
modifies in various fashions the incident energy. A striking example is
the complicated accessory structures of the auditory labyrinth of the
vertebrate ear. These accessory structures may not only affect the
sensitivity of the combined system but may also determine which
parameter (magnitude, rate of change, etc.) of a stimulus may be used.
Thus, refinement of the receptor from a generalized non-discriminat-
ing sensing element with many imperfections has come about through
evolutionary specialization of the sensing part of the receptor itself
and the structures with which it has become associated.

In addition to looking at the sensitivity of a receptor, one must also
look at its responsiveness. In transforming the energy that comes to
it to do some form of work, the receptor does not maintain a one-to-
one ratio between input and output. A great deal of integration

THE PHYSIOLOGY OF INSECT SENSES

OS
OS

o

m
p
6

.2

c

■«->

o

c

s

o

I

>

u

c

O

</5

a
>^

■4-*

D

H

O

INTRODUCTION 5

(putting parts together into a whole) takes place in the receptor, so it is
probably desirable to anticipate a bit and to present a general picture
of events as they occur in a receptor cell.

Energy impinging upon the cell initiates changes that ultimately
generate a propagated all-or-none nerve impulse. As Bullock (1958,
1959) has pointed out, 'The determination of firing is not merely the
build-up of an adequate stimulus or its transduced and amplified
resultant to a critical level, but rather it is a sequence of labile couplings
between graded events, each occurring in a limited fraction of the
neuron.* A series of separate steps with alternate pathways lead to
firing. Some are reflected in changes in the cell membrane potential
(Fig. 1). Others are excitability cycles and the presence or absence of
after-effects not reflected in the membrane potential. Between the
advent of the physical stimulus from outside the receptor, therefore,
and the final production of a nerve impulse, there is a long complicated
series of events. The importance of these is that the receptor, a highly
complex, physiologically non-uniform cell, not only detects energy
and transduces it but also carries out a considerable amount of
integration.

When it comes to the transmission of information from the receptor,
one begins the long and difficult intellectual trek into the central
nervous system. This is beyond the scope of our endeavour here ;
however, it is worth pointing out that much additional integration is
possible long before the central nervous system is reached. Different
channels (nerves) from the receptor may have very different trans-
mitting characteristics - different gain, time constants, and maximum
output (cf. Bullock, 1953).

Thus, the sensory physiologist is interested in the way in which
accessory structures filter or direct the energy going to a receptor, the
way in which the receptor transduces this energy into another form,
and how there is generated a propagated nerve impulse which is a
code relating to the central nervous system or to special effectors the
nature and parameters of a stimulus. And in seeking answers to these
questions, it is of no import to the sensory physiologist whether the
stimulus originates outside of the animal or from within or whether or
not it is ever perceived at a conscious level.

CHAPTER II

General Characteristics of the
Sensory System

One of the biggest challenges that animals faced in their evolutionary
history occurred when they essayed terrestrial life. By all counts the
sea is a more permissive environment than land. It is above all a
more stable and uniform environment. It exhibits no profound
temperature changes; humidity is no problem; osmotic relations are
constant. Consequently, the need to develop sense organs to detect
changes in these realms is minimal. The physical properties of water
obviate the necessity of differentiating between olfaction and taste
in the sense that terrestrial organisms do. Density precludes distant
vision as well as limiting wavelength discrimination. Since water is a
medium of transport, food procurement does not present all the
problems that confront the terrestrial animal. By the same token
water serves as a medium of dispersal of eggs and sperm and as a
cradle for the young so that the complex sensory discrimination and
behaviour patterns which have evolved among land animals for
reproduction and parental care are largely absent in the sea.

One of the more fundamental problems that confronted animals on
emergence to land was that of support, since air does not lend the
helping hand that water does. Before emerging on land, animals had
already set out on two paths of skeletal development. The arthropods
cast the die for an exoskeleton; the chordates, for an internal
skeleton. The choice of skeleton had profound effects upon the
direction that the development of the nervous system followed. The
skeleton is a major limiting factor; all other organs accommodate to
it. The arthropod exoskeleton determined the method of growth - the
only way to increase size is by moulting - and it also limited the overall
size of the animal. Once free of the support of the sea, the animal was
limited in size by the engineering principles of a frame dwelUng (cf.
Thompson, 1943). This may be one of the reasons why insects in
general are small animals (although during the Carboniferous one
dragonfly attained a wingspread of more than 2 ft.). The largest living
species are somewhat larger than the smallest mammals, while the

6

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 7

smallest are smaller than many protozoa (Folsom and Wardlc, 1943).
The range extends from about 166 mm. (the Venezuelan grasshopper
Tropidacris latreillei and some East Indian walking sticks which arc
even larger) down to a fraction of a millimetre (some springtails,
ceratopogonine midges, and beetles of the family Trichopterygidae).

An exoskeleton and small size are therefore two of the outstanding
characteristics of insects. Clearly they must impose upon the nervous
system certain restrictions which will be reflected in behaviour. For
example, smallness reduces the distance over which conduction of
impulses is required. This would imply, other things being equal, more
rapid response and movement. At the same time, however, size
limitations compel a reduction in the number of neurons possible in
the system. A reduction in number of units implies a reduction in the
informational capacity of the system. Reduction is carried further by
the development of so-called giant fibres. Thus, in the abdominal
nerve cord of the cockroach Periplaneta americana the giant fibres
occupy about 12 per cent of its cross-sectional area (Roeder, 1948).
The largest of these fibres measure 30 microns in diameter, exceeding
in this respect the largest (alpha) fibres in the mammalian system.

Roeder (1959) argues persuasively that the relative merits of a
nervous system composed of a few large units versus one consisting of
many small units can be appreciated if one concluded that detail of
information has been sacrificed for speed. It is noteworthy that large
insects tend to react more slowly than smaller ones and that one at
least, the giant Australian cockroach {Macropanesthia rhinocerus
Suass.), lacks giant fibres (Day, 1950). A large fibre cannot carry as
much information from one point to another as can a number of
smaller fibres because of the on-off or all-or-none nature of the nerve
impulse, but it can transmit its information more rapidly. The giant
fibres are the internuncial units in an alarm reaction. In the detection
of, and escape from, predators, speed has greater survival value than
detailed information. From the point of view of a predator also, speed
is important, since attack must be as rapid as the startle response of a
prey. Here, however, the information required is of a much more
complex nature, but it, too, must be handled by a small nervous system
with relatively (as compared with vertebrates) few units.

Another example of the parsimony of neuronal elements is seen in
the motor system of insects. As Hoyle (1957) has pointed out, there
are functionally important muscles which are microscopically small
and yet move joints with precision and delicacy. In contrast with
vertebrate muscles, which are innervated by hundreds of nerve fibres

B

8 THE PHYSIOLOGY OF INSECT SENSES

under a complex central control, the insect muscle is supplied with a
very small number of motor fibres. Some muscles are supplied by four
or more axons; some are mono-axonic. More commonly, a muscle is
supplied by two axons. Thus, the entire system of nervous control
differs from that in the vertebrates (Hoyle, 1957).

In the sensory system, too, there is a limitation on the number of
neurons. This is related not only to the small size of the body but also
to the fact of its being encased in a rigid non-living cuticle. Only at
certain points do the energies of the external environment filter
through to sense cells. As compared with the integument of an
echinoderm, which may have as many as 4,000 sense cells per square
millimetre of surface, and a mammal, whose skin receptors may run
into the millions, the integument of an insect is relatively barren and
insensitive. The fourth-stage larva of Rhodnius, for example, has only
about 420 receptors on the entire ventral surface of each abdominal
segment (Wigglesworth, 1953); the sensory complement of the entire
leg of a fly is less than 500 (Grabowski and Dethier, 1954; Dethier,
1955 b) ; the total number of stress receptors on the whole body of the
drone honeybee is only about 2,948 (Mclndoo, 1914 b, 1916). It is
only when one counts the cells in the two most highly developed
sensory areas of insects, the eye and the antennae, that the number of
receptors becomes large. And even then they fall short, by several
orders of magnitude, of equalling the number in vertebrates. The
maximum number of receptors in any compound eye is that found in
Odonata and is estimated to be approximately 210,000 (Snodgrass,
1935). The number is more usually a few hundred or thousand. The
antennal sense cells of the honeybee only number between 30,000 and
500,000 (Vogel, 1923 b; Snodgrass, 1956).

Economy of cells is seen further in the remarkable organization of
the nervous system whereby stimulation of a single sense cell may be
adequate to set off a whole chain of behaviour. Stimulation of one
neuron in the labellar taste organs of flies initiates proboscis extension
and the initial steps in the feeding pattern (Grabowski and Dethier,
1954; Dethier, 1955 a; Arab, 1959). Roeder and Treat (1957) have
shown that the acoustic response in noctuid moths is mediated by
only two receptor cells. Similarly, stimulation of one neuron in tactile
receptors can initiate a whole series of behaviour patterns ranging
from simple withdrawal of an appendage to running.

The ultimate in parsimony is achieved by all known receptors being
primary sense cells, that is to say, they are true neurons rather than
modified epithelial cells connected synaptically to a neuron (vom

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 9

Rath, 1888, 1895, 1896; Hanstrom, 1928). Although the sensory
system of vertebrates also contains some primary sense cells (e.g., the
rods, cones, and olfactory receptors), the taste receptors and auditory
receptors are specialized non-neural cells connected with neurons.
Possession of primary sense cells means that one cell does the work of

^

m

ISIJ'!M ©

Q

Fig. 2. Stages in the embryonic development of a hair. Tr, trichogen cell;
To, tormogen cell; N, bipolar neuron. (Redrawn from Krumins,
1952.)

at least two; it performs the multiple function of detecting environ-
mental energy, transducing it, and initiating and transmitting im-
pulses to the appropriate ganglion. From the point of view of the
sensory physiologist, the primacy of the insect receptor is a boon
to investigations, since all events occur in the one cell. It is
generally believed, furthermore, that the primary neurons make their

10 THE PHYSIOLOGY OF INSECT SENSES

connexions directly with the central nervous system. From time to
time peripheral ganglion cells have been described, but these not
infrequently turn out to be non-neural cells. On the other hand, one
of the more recent reports, that of Peters (1961) describing a multi-
polar neuron in the labellum of flies, warrants further investigation.

While it has generally been accepted for some time that sense cells
arise by mitosis from epidermal cells (vom Rath, 1888, 1895, 1896;

Fig. 3. A 'circular nerve' in the integument of fourth-stage larva of
Rhodnius and the cell bodies from which it was derived following an
extensive burn of the cuticle in the third-stage. (Redrawn from
Wigglesworth, 1953.)

Haffer, 1921; Hanstrom, 1928; Snodgrass, 1935) (Fig. 2), proof that
the cells are primary sense cells is of relatively recent origin. Vogel
(1 923 b) claimed that in Hymenoptera the sensory axon grows out from
the central nervous system and joins the sense cell. Newton (1931),
Henke and Ronsch (1951), and Krumins (1952) maintained that the
epidermal sense cells differentiate axons which then grow centripetally
to join the central nervous system. The matter would seem to have
been settled, at least for Rhodnius prolixus, by Wigglesworth (1953),

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 11

who proved by interrupting sensory nerves that axons are regenerated
by epidermal cells and grow centripctally until they encounter a
nerve. If they do not encounter a nerve or if they become trapped
above the basement membrane, which separates the epidermis from
the rest of the body space, they grow in loops and circles, apparently
indefinitely (Fig. 3).

Nc Oc

Fig. 4. A. Stages in the moulting of sensilla in a fourth-stage larvae of
Rhodnius. B. Campaniform organ (haematoxylin). C. Tactile hair
(haematoxylin). D. Campaniform organ (Romanes' method). E.
Tactile hair (Romanes' method). N, nerve; S, sense cell; Dp, distal
process; A, axon; Tr, trichogen ; To, tormogen; Nc, new campaniform
organ; Oc, old campaniform organ; Ns, new seta; Os, old seta; Ni,
neurilemma cell; Nso, new socket; E, extension of distal process from
new to old seta. (Redrawn from Wigglesworth, 1953.)

12 THE PHYSIOLOGY OF INSECT SENSES

At each moult, when new receptors arise to service the increased
body size, they, just like regenerating receptors, send axons centri-
petally to the central nervous system (Fig. 4). Whether or not the
axons fuse when they come together is still in doubt. Pringle (1938 a)
was of the opinion that fusion did occur in the cockroach, but
Wigglesworth (1953) felt that the matter was in doubt in Rhodnius.

Fig. 5. A small area of the integument of a fourth-stage larva of Rhodnius
showing hairs and neurons. The number of axons present in each nerve
as calculated from the number of sensilla is indicated by the figures. The
inset shows the distribution of epidermal cell nuclei. (Redrawn from
Wigglesworth, 1953.)

He showed in Rhodnius (Wigglesworth, 1959) by making a rough
count of the numbers and types of sensilla on the antenna and a count
of the axons in the antennal nerve that there must be a fusion of at
least fifteen sense cells to one axon. Similar counts in the leg indicated
that similar fusion of the tactile fibres here does not occur. He pointed
out that a lack of fusion is understandable in view of the need of

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 13

accurate touch localization. On the other hand, fusion of axons within
groups of campaniform sensilla, hair plates, and some chordotonal
sensilla does occur in the leg of the cockroach (Pringle, 1938 a;
Nijenhuis and Dresden, 1952).

A B

Fig. 6. Diagrammatic representation of two types of ommatidia. A. Apposi-
tional type of diurnal Lepidoptera. B. Superpositional type of nocturn-
al Lepidoptera. Co, cornea; Cp, corneal process; Cn, crystalline cone;
CnN, crystalline cone nuclei; R, retinal cell; Rh, rhabdom; T, tracheal
tapetum. (Redrawn from Snodgrass, 1935 after Nowikoff.)

The presence or absence of fusion is a matter of considerable
importance to the electrophysiologist, who must try to interpret
recordings from afferent nerves. Another unsettled question of
importance is whether or not the large cells frequently observed at
the junction of uniting nerves or axons are multipolar nerve cells.
If they are, it means that a synapse is interpolated between the
peripheral sensory neuron and the central nervous system. Otherwise,

14

THE PHYSIOLOGY OF INSECT SENSES

Fig. 7. Examples of Type II neurons. A and B. Subepidermal nerve plexus
in the larva of Melolontha. (Redrawn from Zawarzin, 1912.)

the path from the primary neuron to the central nervous system is a
direct one. According to Snodgrass (1935) there are no cells in
sensory nerves, and Wigglesworth (1953) is of the opinion that the
cells seen there are actually neurilemma cells.

Morphologically there are two broad categories of sense cells:
those whose dendrites are nearly always associated with the cuticle or
its invaginations (apodemes, tracheae, and cuticle of preoral and oral
cavities) (Type I) (Fig. 5) ; multipolar neurons never associated with

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 15

cuticular processes but lying instead on the inner face of the body wall,
the walls of the alimentary canal, muscles, and connective tissue
(Type II). Only the cells of photoreceptors lack an obvious distal
process or dendrite, although even in these cells the distal end is

B • ^

Fig. 8. Type II neurons. A. The hypopharynx of termites. (Redrawn from
Richard, 1951.) B. The muscles of the mid-gut of Melolontha larvae.
C and D. The muscles of the oesophagus of Oryctes larvae. (Redrawn
fromOrlov, 1924.)

structurally modified (Fig. 6). Depending on the position of the
neurocyte with respect to the epidermis. Type I neurons are termed
intraepidermal or sub-epidermal, but there are certainly all gradations
between the two positions.

The ontogenetic origin of Type II neurons from the ectoderm has
not yet been determined (Snodgrass, 1935). Numerically they are less

16 THE PHYSIOLOGY OF INSECT SENSES

common than Type I and are most abundant in soft-skinned larvae.
Their branching distal processes may form an elaborate subepidermal
net, as in the larva of Melolontha vulgaris (Fig. 7) (Zawarzin, 1912 b).
They also occur in the walls, muscles, and connective tissue of the

B

Fig. 9, Types of campaniform sensilla. A. From the haltere of Calliphora
erythrocephala. B. From the cercus of Gryllus domes ticiis. C. From the
cercus of Blatta orientalis. (Redrawn from Hsii, 1938.)

alimentary tract of some insects {Periplaneta americana [Zawarzin,
1912 a] and larvae of scarabaeid beetles [Orlov, 1924] [Fig. 8]).
Richard (1951) has described them in the labium, labrum, and
hypopharynx of termites (Fig. 9). Until Finlayson and Lowenstein
(1955, 1958) studied some Type II receptors electrophysiologically,

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 17

C

Fig. 10. A campaniform sense organ from the antenna of a grasshopper. N
bipolar neuron; S, scolopoid sheath; A, axon; D, dendrite; C, cuticular
dome. (Redrawn from Slifer et aL, 1959.)

there was no certain knowledge of the function of any of them.
Electrophysiological studies of those associated with muscle and
connective tissue have proved them to be proprioceptors. Their
structure and function will be described in Chapter III.
Type I neurons and the cells giving rise to the cuticular structures

18 THE PHYSIOLOGY OF INSECT SENSES

with which they are usually associated originate from the same parent
epidermal cell. The combination of the neuron (or neurons), the
immediate cuticular area, when present, and its generative cells is
termed a sensillum. A sensillum is thus a sense organ ; the neuron is
the receptor. With the possible exception of the photoreceptors, all
sensilla are believed to be homologous and to have been derived from
setae. Classically, because of the ease of examining the cuticular
portion, sensilla were classified on the basis of external form. They

Fig. 1 1 . Sensillum placodeum
from the antenna of the
honeybee. N, bipolar
neuron; A, axon; D,
dendrites; P, plate; F,
attachment of fibres. (Re-
drawn from Snodgrass,
1935.)

include setiform varieties (sensilla trichodea), bristles (sensilla
chaetica), scales (sensilla squamiformia), pegs and cones (sensilla
basiconica), pegs, cones, or bristles sunk in shallow depressions
(sensilla coelonconica) or in deep pits (sensilla ampullacea) (Schenk,
1903).

Other sensilla whose external cuticular structures are not seta-like
are undoubtedly derived from setae (Snodgrass, 1935; Lees, 1942).
The campaniform organs (sensilla campaniformia, sense pores, etc.)
are visible externally as minute pits in the cuticle, but in section are
seen to possess a bell-shaped cuticular cap (Figs. 10, 30). Each is
innervated by a single neuron. Their structure will be discussed in

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 19

detail in Chapter III. The plate organs (sensilla placodea) are marked
externally by an oval or elliptical plate surrounded by a narrow
membranous ring. Each is innervated by a number of neurons (Fig. 1 1).
One type of sensillum (sensillum scolopophorum, scolopoid
sensillum, peg organs, chordotonal organs, stiftfUhrcnder Sinnesor-

FiG. 12. Different forms of scolopoid sensilla. A. Terminal peg from a
grasshopper sensillum. (From Eggers.) B. Terminal peg from the
tracheal organ of Gryllus. (From Schwabe.) C. Terminal peg from the
femoral chordotonal organ oiPediculus. (From Graber.) D. Terminal
peg from a chordotonal organ in the abdomen of a cerambycid larva.
(From Hess.) E. A simple chordotonal organ with one sensillum.
(From Eggers.) F. A. chordotonal sensillum from the haltere of a
muscid fly. (From Pflugstaedt.) G and H. Scolopoid sensilla from the
tympanic organ of Cicadetta coriaria S- (From Vogel.) I. Sensillum
from the tibial organ of Decticus. (From Schwabe.) J. Sensillum from
the subgenual organ of Decticus. (From Schwabe.)

gane) does not fit conveniently into any category. The sense cell is a
bipolar neuron, but its dendrite is not always associated with a
particular cuticular structure. These sensilla usually do not occur
singly. They are generally gathered together in bundles having a
common point of attachment to an undifferentiated part of the body
wall. Sometimes a small external pit, thickened disk, or nodule

20 THE PHYSIOLOGY OF INSECT SENSES

marks the point of attachment. In special cases they are associated
with a tympanic membrane. They all possess in the region of the
dendrite a cuticle-like sheath, usually ribbed, and in the majority of
cases capped with a prominent refractile body. Because of the

Fig. 13. a sensillum basiconicum from the antenna of a grasshopper. N,
bipolar neuron; S, scolopoid sheath ; A, axons ; D, dendrites. (Redrawn
from SHfer^/^/., 1959.)

resemblance of the combined sheath and cap to a peg-shaped rod,
the structure has been termed the scolops (Stift). A terminal strand
may or may not extend from the cap distally to the point of attachment
on the body wall.
There is no completely satisfactory name to apply to these sensilla.

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 21

They were called chordotonal organs by Graber (1882 a) because
many of them stretched like taut cords from point to point and were
believed to be auditory organs. Those with a terminal strand were
termed amphinematic because they appeared to be stretched between
two threads; the others were termed mononematic (Graber, 1882 a).
It soon became clear that these sensilla were neither exclusively audi-

FiG. 14. A thick-walled sensillum basiconicum from the antenna of a
grasshopper. N, bipolar neuron; S, scolopoid sheath; A, axons;
D, dendrites; Tr, trichogen cell; To, tormogen cell. (Redrawn from
Slifer et al., 1957.)

tory nor always cord-like in appearance. In recognition of the
universal existence of a peg-like rod within them, Eggers (1923, 1924,
1928) proposed that they be called scolopophorous or scolopal organs.
With this terminology one sensillum would be called a scolopidium
(by analogy with the ommatidium of the eye) and a group of scolopidia
comprising a unit would be termed a scoloparium. Even Eggers
admitted, however, that other kinds of sensilla possessed peg-like

22 THE PHYSIOLOGY OF INSECT SENSES

rods, and it now appears that such structures are a constant feature of
insect sensilla. In recognition of this fact Snodgrass (1926) originally
preferred the term chordotonal in deference to custom ; however, he
later (1935) switched to the term scolopophorous sensillum. Neither

Fig. 15. Sensillum coeloconicum from the antenna of a grasshopper. N,
bipolar neuron; S, scolopoid sheath; A, axons; D, dendrites; Tr,
trichogen cell; To, tormogen cell; M, dried residue from moulting
fluid. (Redrawn from Slifer et al, 1959.)

term seems singularly appropriate, so the choice at the moment is a
matter of taste (for a complete discussion of this matter see Eggers,
1923, 1924, 1928; Snodgrass, 1926, 1935).

A typical sensillum consists of a minimum of four cells (Fig. 16):
(1) the trichogen or hair-forming cell; (2) the tormogen or socket-

B
Fig. 16. A. Tactile hair from the larvae of Vanessa urticae. B. Scolopoid
body in the tactile hair of the larva of Fieri s rapae. S, scolopoid sheath;
Sb, scolopoid peg; D, dendrite; N, bipolar neuron; A, axon; Tr,
trichogen cell ; To, tormogen cell. (Redrawn from Hsii, 1 938.)
C

24 THE PHYSIOLOGY OF INSECT SENSES

forming cell; (3) the bipolar neuron; (4) the reniform neurilemma cell
(Haffer, 1921; Wigglesworth, 1933; Hsu, 1938). This group of cells
may be surrounded on all sides by the epithelium of the integument
or may lie sunken in a subepidermal position. In any case, the base-
ment membrane of the cells is continuous with the basement
membrane of the epidermis, so that the whole unit is walled off from
the body cavity except where the neuron enters.

The different types of sensilla depart from this general picture in
details, and considerable variation exists among any single type from
one insect to another or from one part of an insect to another. The
most significant variations occur with respect to the number of
neurons and their relations with other parts of the sensillum. The
number of neurons associated with a sensillum may range from one
to more than fifty (Fig. 13) (Slifer, Prestage, and Beams, 1959). The
distal process, or dendrite, of the neuron may terminate at the
base of a seta or extend variable distances up the shaft. The exact
relations between the dendrites of the neurons and the cuticular
portions of the sensillum is imperfectly known despite extensive
studies (Snodgrass, 1926; Hsli, 1938; Vogel, 1923; Sihler, 1924;
Eggers, 1928; Debauche, 1935; Slifer, 1961). In almost all cases the
dendrite enters a sheath (not of neural origin) whose optical and
staining properties resemble in many respects that of cuticle. The
sheath, best described as a cuticula-like tube, has been variously
termed sense rod, scolopale, Stift, Stiftkorperchen, corps scolopoid,
Hiille, Chitineurium, etc. The exact relation of the dendrite to this
tube differs in the various sensilla. In chemoreceptors especially the
top (distal) end of the tube may open to the shaft of the hair or peg
so that the dendrites can extend the length of the tube and into the
lumen of the external process (Fig. 14). Or it may open to the outside,
in which case the dendrites pass through openings in the side of the
tube and continue in the lumen of the hair or peg (Fig. 13).

In many mechanoreceptors the top of the tube is capped with
a dark-staining apical body. In some cases the dendrite terminates
within or just beneath this cap. This is the case with many tactile hairs
(Fig. 16) and campaniform sensilla (Figs. 9, 10, 30) and with most
chordotonal sensilla (Figs. 12, 48). In some cases, especially in
chordotonal organs, a long filament may extend from the cap to the
cuticle (Fig. 12). There are, however, chordotonal organs in which the
terminal filament is absent (Figs. 47, 48).

Recent studies with the electron-microscope (Gray and Pumphrey,
1958; Gray, 1960; Slifer, 1961; Adams, 1961; Larsen, 1962; Dethier,

GENERAL CHARACTERISTICS OF THE SENSORY SYSTEM 25

Larsen, and Adams, 1963) have revealed a most extraordinary
complexity in the structure of the dendrites and scolopoid bodies of
sensilla. Since details appear to differ so much from one sensillum to
another, specific cases will be dealt with in later chapters.

While certain generalizations can be made about the function of
some types of sensilla (e.g., sensilla campaniformia are mcchano-
receptors sensitive to deformation of the cuticle, and chordotonal
sensilla are mechanoreceptors sensitive to sound, vibration, and
stretching of parts of the body), it is dangerous as well as unprofitable
to attempt to link a particular structure with a particular function.
This is so because a type of sensillum which may subserve one function
in one species may subserve a different function in another. Further-
more, a single sensillum may consist of more than one type of receptor
(e.g., the labellar hairs of Phormia, which have two chemoreceptors
and a mechanoreceptor associated with them). Each is a case unto
itself.

CHAPTER III

Mechanoreception

Every insect maintains its body in a particular attitude with respect to
gravity. This attitude, its primary orientation, means for most
ventral-side down. There are, however, many that habitually live
upside down (the backswimmers, praying mantids, many lepi-
dopterous larvae) and a few, such as those living in vertical burrows,
whose normal attitude is posterior-end down. For aquatic insects
and flying insects, both of whom are freely suspended in a homo-
geneous medium, there must be sense organs that can give information
about the direction of the force of gravity. Vision is of some help in
this connexion, especially in flying insects, since the sky is invariably
up and the ground down, but that is not the prime service of vision.
Insects in contact with a substrate derive some information from
sense organs touching the substrate, but this information tells them
only which surface of the body is in contact, not what the attitude of
their body is with respect to gravity.

In addition to maintaining a primary orientation all insects have
characteristic postural relations. To maintain these, to have infor-
mation concerning the position of one part of the body with respect to
another, requires special sense organs. The same sort of information is
needed to be able to move one part of the articulated body precisely
with respect to another in order to walk, swim, fly, spin cocoons, dig
burrows, court, copulate, and feed. Finally, for the proper working of
the internal organs there must be information concerning the presence
of faeces in the rectum, eggs in the ootheca, and food in the alimentary
canal.

Serving all of these needs is a whole spectrum of sense organs
responding to energy derived, in the final analysis, from the surface
gravitation of the earth. They are organs sensitive to stretch, com-
pression, or torque imparted to cuticle, connective tissue, or muscles
by the weight (which is a measure of the force with which the earth
attracts it) of parts of the body, the relative movement of parts, the
gyroscopic effects of moving parts, and impingement of the substrate
or surrounding media. Since there is no neuron directly sensitive to
gravitational force, all of the receptors concerned with primary

26

MECHANORECEPTION 27

orientation, postural relations, and touch are associated with compli-
cated and highly specific accessory structures whose function is to
transduce the energy of the stimulus into mechanical deformation of
the protoplasm of the neuron. Many cells are sensitive to mechanical
deformation, but the mechanoreceptors are highly specialized in this
respect. The different accessory structures and the different response
characteristics (e.g., rate of adaptation) of the receptor determine the
sensitivity of the sense organ and to which parameter of the stimulus
it will respond.

THE TACTILE SENSE-SENSILLA TRICHODEA

A non-living, tough, and most often rigid cuticle effectively insulates
tissues from all but the grossest mechanical disturbances of the
external environment. To provide the surface sensitivity denied by the
exocuticle there are numerous thin, hollow extensions, the sensilla
trichodea, on all surfaces of the body. They are most numerous on
those areas, such as legs, which come most frequently into contact
with the substrate, those, such as the antennae and mouth-parts, which
are employed in palpating or manipulating the environment, and on
such extended portions of the body as the antennae and cerci which
represent perimeter guards of the body. They also occur abundantly
on surfaces between joints, segments, and other appressed areas of the
body where their function is proprioceptive. Their shape, mechanical
properties, and the physiological properties of the neurons associated
with them vary considerably, depending upon the service they are
designed to perform. Those at the joints are short and slow adapting;
those on the cerci, extraordinarily long and delicate, and hence
sensitive to the most gossamer disturbance ; the thick tibial leg spines,
gross and slow adapting.

The sensilla trichodea are set in membraneous sockets. The shaft is
so rigidly constructed that any force appUed to the structure is trans-
mitted to the socket. It is here that movement takes place and, because
of the leverage, is greatly ampHfied. The simplest sensilla trichodea are
those that are innervated by one neuron. These are simple mechano-
receptors. There are others, however, that are innervated by more
than one neuron, and these are compound sense organs. In the labellar
hairs of the blowfly, for example, one neuron is the mechanoreceptor
responding when the hair is moved in its socket ; the remaining neurons
are chemoreceptors responding when certain chemicals touch the tip
of the hair (cf. Chapter V).

The exact manner in which the dendrite of the neuron is associated

28 THE PHYSIOLOGY OF INSECT SENSES

with mechanoreceptive hairs is not always clear (Fig. 16). It is reported
as being inserted on that part of the hair shaft that extends below the
socket, on the socket itself, or at some point part way up the shaft of the
hair. It is enclosed in a scolopoid sheath, but may or may not be
capped with an apical body. The sheath may be ribbed and possess
hickened zones. Just below the cap may be seen with the light micro-
scope some darkly staining spots, the sense rods (Snodgrass, 1926) or
points sensoriels (Hsu, 1938). Further observation with the electron-
microscope may reveal more complicated structural relationships
similar to those seen in tympanal chordotonal sensilla (Gray, 1960).

Richard (1952) showed clearly in the case of sensilla trichodea of
termites that at moulting the sheath and the neuron process extending

.«''sw>'»«*».^n«<JujW

1.0
MV

1 1

MMMiffMMI

Fig. 17. Response to repetitive mechanical stimulation of a chemosensory
hair on the wing ofSarcophaga. The arrow indicates the onset of a rapid
one-directional displacement, and the direction of the displacement
with reference to the preceding stimulus. Positive potential at recording
electrode is down. Time marks recur at 0-2-second intervals. (Courtesy
of M. L. Wolbarsht.)

through and beyond it are lost. A new sheath and filament are regener-
ated to supply the new hair. During the period when the new cuticle is
forming the distal fibre remains attached to the old hair to provide
tactile sensitivity until the last moment. Shortly before the moult the
distal fibre ruptures, and at this time there is a loss in tactile sensitivity
(Richard, 1952; Wiggles worth, 1953).

From the geometry of these sensilla it is likely that bending of the
hair in its socket deforms the terminal region of the distal process of
the neuron. The response to angular deflexion varies with direction
relative to the long axis of the joint (Pumphrey, 1936; Pringle, 1938 a).
Stimulation is followed by an electrical response that can be divided
into two components. The work of Wolbarsht (1960) has shown that
there is a graded slow potential, the receptor potential, that can be
recorded from the distal process. It occurs prior to any impulses, varies

MECHANORECEPTION 29

smoothly, and must attain some critical level before any impulses are
initiated.

All hairs have a steady resting potential until a mechanical stimulus
is applied. The potential may be either positive or negative, depending
upon the type of hair. When the hair is stimulated an increase in
negativity, which varies smoothly as the stimulus is increased or
decreased, is seen at the recording electrode. Two types of receptor
potentials have been observed. In one kind of hair the potential per-
sists only during the motion of the hair (Fig. 17); in another kind, it
lasts as long as the hair is deformed (Fig. 18). In the former the return
of the hair to the unstrained position initiates another receptor

-wi^/^^^wftifmri^^

Fig. 18. Response of mechanosensory hair on outer clasper of male
Phormia to mechanical stimuli. A, B, and C are successive records taken
from the same hair, but the stimuli were not the same. Arrows indicate
onset and cessation of deformation. The deformation of the hair was
approximately proportional to the amplitude of the response for each
record. Positive at recording electrode is down. Time marks recur at
0-2-second intervals. (Courtesy of M. L. Wolbarsht.)

potential. This kind of hair adapts very rapidly. The other type adapts
very slowly. Sometimes adaptation is still incomplete at the end of
twenty minutes.

The impulses always have an initial positive phase in contrast to the
negative-going receptor potential. They are usually monophasic.
Impulses occur only after some threshold of receptor potential has
been reached. The threshold receptor potential for two types of fast-
adapting hairs which is accompanied by an impulse is plotted for a
series of test stimuli in Figs. 19 and 20. There is also a relation between
the frequency of discharge and the magnitude of the receptor potential.
Bernhard, Granit, and Skogland (1942) had shown that frequency was
determined by the refractory period of the spike-generating mech-
anism and the magnitude of the generator potential. Up to the point
where the increase in generator potential causes no further increase in

30 THE PHYSIOLOGY OF INSECT SENSES

frequency because the interval between impulses is equal to the re-
fractory period of the neuron there is a direct relation between fre-
quency and the amplitude of the generator potential. These relations
hold for the insect mechanoreceptor (Figs. 21, 22).

The most striking feature of insect mechanoreceptors is the change
in size of the impulse as the receptor potential changes (Figs. 23, 24,

o
o

o

Threshold Receptor Potential for Appearance of impulses MV

Fig. 19. The threshold receptor potential of a mechanoreceptor on the
grasshopper claw which is accompanied by an impulse is plotted for a
series of test stimuli. In the largest number of trials the threshold was
2-5 MV with the spread as shown. (Redrawn from Wolbarsht, 1960.)

25) (Wolbarsht and Dethier, 1958; Wolbarsht, 1960). Wolbarsht
(1960), after analysing the electrical events occurring in the mechano-
receptor, concluded that: the mechanical stimulus effects a change
in the membrane resistance at the receptor site; the receptor potential
(which is undoubtedly the generator potential) is the difference
between the potential across the membrane at the receptor site and the
general polarization of the cell; the impulses are generated proximad

MECHANORECEPTION

31

30

—

to

C 20

._

O

-ij

<XJ

c-

<u

-Q

O

«-*-

o

1^

a>

_o w

'~~

E

3

^

1 1 i 1

Threshold Receptor Potential for Appearance of Impulses Ml/

Fig. 20. The threshold receptor potential of the mechanosensory neuron of
a chemosensory hair on the wing of Sarcophaga which is accompanied
by an impulse is plotted for a series of mechanical stimuli. In the
largest number of trials the threshold was 0-38 MV with the spread as
shown. (Redrawn from Wolbarsht, 1960.)

of the receptor site and do not invade it ; changes in impulse size are due
to changes in the membrane resistance of the receptor site ; there is a
high resistance between the fluid-filled centre of the hair and the
general body fluid ; the value of this resistance is effectively determined
by the close anatomical conjunction of the receptor membrane and
the lumen of the hair cavity. In most respects the electrical response of
the insect mechanoreceptor is similar to that of the Pacinian corpuscle
of vertebrates.

-r*

_U^

f{rm'^''f^^rfmrf0m^^,

t

^tf^.

O^^'^^f^v*

10
MV

Fig. 21. Response of mechanosensory hair on the anal plate of a female
Pliormia to changing mechanical stimulation. The left-hand arrow
indicates the approximate beginning of the deformation; the middle
arrow, the minimum; and the right-hand arrow, the return to the
undeformed position. Positive at the recording electrode is down.
Time marks recur at 0-2-second intervals. (Courtesy of M. L.
Wolbarsht.)

32 THE PHYSIOLOGY OF INSECT SENSES

While it is convenient to speak of all sensilla trichodea as a class of
sensilla responding to deflexion, the variations in structure, hence in
physiological characteristics, impart to the different ones quite
different functions in the economy of the insect. One of the salient
differences encountered is the rate of adaptation. Generally, the
smaller, more delicate hairs adapt rapidly during prolonged deflexion
while the stouter spines adapt more slowly and incompletely. The
nature of the process is still unclear.

160

—

140

-

•

120

—

oioo

-

•

,

■-60

•

•

60

—

•

•

•

40

~"

•

n

1

1

1

1

1

6

Receptor Potential MV

Fig. 22. Receptor potential amplitude plotted against frequency of
impulses for a mechanosensory hair on the anal plate of a female
Phormia. Each point is the average of 2 seconds. (Redrawn from
Wolbarsht, 1960.)

Evidence has been presented to show that the generation of impulses
follows depolarization of the neuron membrane. The rate of adapta-
tion appears to be closely related to the rate of decay of the depolariz-
ation. The generator potential of the rapidly adapting Pacinian
corpuscle of the cat (Alvarez-Buylla and de Arellano, 1953 ; Gray and
Sato, 1953), the fast adaptory crustacean muscle receptor (Eyzaguirre
and Kuffler, 1955), and the sensory hairs on the wings of flies (Fig. 17)
(Wolbarsht, 1960) fall rapidly below threshold. The slowly adapting
frog-muscle spindle (Katz, 1950), the slow muscle receptor of Crust-

MECHANORECEPTION

33

— 6

—6

O"— 2

TlH

^ nl— ?U-

c •

}5

T

• 1.}

B« - • 5

T O

.?'

I

A •

Receptor Potential M V

Fig. 23. Impulse size plotted against receptor potential for a mechano-
sensory hair on the outer clasper of a male Phormia. Points are average
values; bars denote extreme values. (Redrawn from Wolbarsht, 1960.)

acea, and the mechanosensory hair on the claspers of male flies
(Fig. 18) (Wolbarsht, 1960) exhibit slow potentials that are main-
tained above threshold for a considerable period of time. It has been
assumed, at least in some cases, that depolarization is a consequence
of the unfolding of the cell membrane (Katz, 1950; Gray and Sato,
1953). Lowenstein (1956) has suggested that the degree to which the

34 THE PHYSIOLOGY OF INSECT SENSES

membrane can unfold imparts a mechanical component to adap-
tation. As he pointed out, this concept, while not explaining rates of
adaptation, brings it to another level of understanding.
All types of mechano -hairs studied (Wolbarsht and Dethier, 1958;

7 —

• } ■•

Receptor Potential M V

Fig. 24. Impulse size plotted against receptor potential for a mechano-
sensory hair on the second antennal joint of Melanopliis. All points
except lowest receptor potential are averages of at least twenty impulses.
Lowest value is a single impulse. Bars denote extreme values. (Redrawn
from Wolbarsht, 1960.)

Wolbarsht, 1960; Pumphrey, 1936) appear to fall into two classes:
velocity sensitive and pressure sensitive. The former fire only while the
stimulus is changing. Some, such as the hairs on the leading edge of the
wings of flies and other insects (Wolbarsht and Dethier, 1958), may
fire at a rate of 600 or more impulses per second. Pressure-sensitive
hairs show a repetitive discharge during a static deformation.

MECHANORECEPTION 35

Tactile receptors are usually of the velocity-sensitive type and are
most common on those portions of the body that encounter the en-
vironment as the animal progresses or on those appendages with which
the animal explores its environment. Pressure-sensitive receptors are

.hf^H^^^*^^llk0iifm0^F^'^i ( I i ! 1 1 I 1 I

M

4.0
MV

I

i4iYiTfrrf^T>^^

Fig. 25. Response of a mechanosensory hair on the second antennal joint
of Melanoplus to increasing mechanical stimulation. Approximate
onset of stimulation is indicated by arrow. The stimulus continues until
the end of the record. A and B are responses to successive stimuli of
approximately the same size. Positive at the recording electrode is
down. Time marks recur at 0-2-second intervals. (Courtesy of M. L.
Wolbarsht.)

most common on those areas where positioning is important. They are
found, for example, in the genital regions and between joints. They
are proprioceptors.

PROPRIOCEPTORS

Proprioceptors may be defined as sense organs capable of continuous
response to deformations (changes in length) and stresses (tensions
and compressions) in the body (Lissmann, 1950). They provide some
of the information necessary for the animal to maintain certain re-
lations of one part of the body with another and of the body as a whole
with respect to gravity. In this capacity they are assisted by other recep-
tors, as photo- and tactile receptors, for example, whose primary
function may lie in another realm. In other words, the insect maintains
position by assessing much information from a multitude of receptors.

36 THE PHYSIOLOGY OF INSECT SENSES

Five kinds of sensory structures are known to take direct part in
proprioception: hair plates, campaniform sensilla, stretch receptors,
chordotonal organs, and statocyst-like organs.

Hair Plates (Position Receptors)

Concentrations of minute sensilla trichodea similar in general
structure to tactile hairs were first described by Lowne (1890) on the
anterior part of the thorax of the blowfly (Fig. 26). Similar structures,
termed hairplates by Pringle (1938 c), are distributed in the joints of the
legs and palpi of cockroaches (Fig. 27). Others occur in the dragonfly
and mantis (Fig. 27) (Mittelstaedt, 1950, 1952, 1957) and in the honey-

FiG. 26. The prosternal organ of Calliphora. B, basal part of head; C, neck
sclerite; Co, connecting membrane; N, nerve; P, prothorax. (Redrawn
from Lowne, 1890.)

bee (Lindauer and Nedel, 1959). Single hairs of varying size and shape
were found in the body articulations of the cockroach by Diakonoff
(1936). Hair plates are undoubtedly a common feature among insects.
In all cases the sensilla are stimulated by folds of the intersegmental
membrane or contact with adjoining surfaces as the joints are moved.

The hair plates of the cockroach have been studied in some detail
(Markl, 1962; Peters, 1962). The discharge of these receptors is
proportional to the degree of bending of the hairs as the joint is
flexed. The initial frequency may be 800 or more per second (Fig.
29). The rate of adaptation is very slow. Thus, the hair plates act as
'position' organs and behave independently of muscular tension.

The proprioceptive function of hair plates is strikingly illustrated
by Mittelstaedt's (1952, 1957) analysis of prey capture by mantids.
Successful prey capture is a problem of absolute optic localization.

MECHANORECEPTION 37

Mantids lying in ambush detect tlieir prey visually. The eyes being
immovable, the head is turned so that the prey is always faced.
Although the mantis may then turn its body to line it up with the prey,
it can capture a prey which sits at a considerable lateral deviation
from the median plane of the thorax. If the prey is close enough it is
seized by a rapid (10-30 milliseconds) stroke of the prothoracic legs.
The accuracy of hitting in normal mantids is about 85 per cent. Since

Fig. 27. A. Proprioceptors of the neck region of the mantis. S, sterno-
cervical hair plate; N, tergocervical plate. (Redrawn from Mittel-
staedt, 1957.) B. Proprioceptors of the second leg of Periplaneta.
C, coxa; I, inner coxal hair plate; O, outer coxal hair plate; T, troch-
anteral hair plate. (Redrawn from Pringle, 1 938 c.)

the speed of the stroke is so great, its direction is not controlled by
watching the difference between its direction and that of the prey. Its
direction is predetermined by information concerning the position of
the prey relative to the plane of the prothorax. Thus, the message
steering the stroke must contain information about the direction of the
prey relative to the head and of the position of the head relative to the
prothorax. The first information is provided by the compound eyes ;
the second, by hair plates in the neck region. Two pairs of plates are
involved ; the sternocervical plate (Pringle, 1938 c) and the tergocervical

38 THE PHYSIOLOGY OF INSECT SENSES

plate (Mittelstaedt, 1952) (Fig. 27). The role of these sensory systems
was demonstrated by the following series of experiments. If the inner-
vation from all hair plates is cut so that no information is forthcoming
from these organs the accuracy of hitting drops to 20-30 per cent.
With deafferentation of the left side only, missing increases, and there
is a tendency to strike to the right. In other words, proprioceptive
information coming only from the right misleads the animal into
believing that its head is turned to the right whereas it is not. If the
head is held in a fixed position relative to the prothorax by a balsa-

FiG. 28. Ventral view of the campaniform sensilla in the trochanteral region
of the third right leg of Pe rip lane ta. C, coxo-trochanteral condyle;
E, extensor trochanteris and its accessory apodemes ; F, flexor troch-
anteris; H, trochantero-femoral hinge joint; 2, 3, 4, groups of campani-
form sensilla. (Redrawn from Pringle, 1938c.)

wood bridge so that the neck region is not touched accuracy of hitting
decreases to 25 per cent if the head deviates from the body axis by
10-30 degrees. The prey is missed to the left if the head has been turned
to the right, and vice versa.

If head fastening and unilateral extirpation of proprioceptors are
combined the effects of both are superimposed, the loss of one-half of
the neck receptors being equivalent to a deviation of the head of less
than 20 degrees. If the free head is loaded with an extraneous force
performance remains normal until the load surpasses twice the head
weight at twice its diameter. In short, direction of stroke depends on
feedback processes which control the position of the head. Fixation
movements of the head are steered by the difference between the
optic-centre message (a function of the angle between prey and

MECHANORECEPTION

39

700-

500

400

c
o

a.

C 300
U; 200

100

\«. • " • '

♦ -•— • • — « .

5 6

Seconds
A

B

Fig. 29. A. Adaptation of a nerve-ending in the inner coxal hair plate of
Periplaneta to a stimulus of a constant deflexion of the hairs. Two
records from the same preparation. B. Diagram to illustrate the
method of excitation of the inner coxal hair plate by a fold of the inter-
segmental membrane. C, coxa; H, hair plate; P, pleuron. (Redrawn
from Pringle, 1938c.)

fixation-line) and the proprioceptive-centre message (a function of
the angle of the head with respect to the body axis). If fixation move-
ments have come to rest the direction of stroke is determined by the
optic and, to a lesser extent, by the proprioceptive-centre messages,
which then both contain the required information (Mittelstaedt,
1957).

D

40

THE PHYSIOLOGY OF INSECT SENSES

Campaniform Sensilla (Compression and Stretch Receptors)

The term sensilla campaniformia was proposed by Berlese (1909) to
apply to all sensilla similar to those originally described by Hicks
(1857). These organs consist of a canal in the cuticle covered by a
domed cap and innervated by a single bipolar neuron (Fig. 11, 30)
(Newton, 1931; Vogel, 1911; Sihler, 1924; Pflugstaedt, 1912;
Mclndoo, 1914 a, 1914 b ; Snodgrass, 1935 ; Hsu, 1938 ; Slifer, Prestage,
and Beams, 1959). The cap is lined with a substance that has different
staining properties from the rest of the cuticle. The dome may extend
above the general cuticular surface, be flush with it, or be recessed. The
distal neuronal process extends within the scolopoid sheath (Slifer

Fig. 30. Structure of the cuticular parts of various types of campaniform
sensilla. A, outer lamella of dome; B, inner lamella of dome; C, cuti-
cular connexion of dendrite of sense cell. (Redrawn from Snodgrass,
1935.)

et al, 1959) and is inserted on the underside of the cap by a highly
refractile body.

These sensilla occur on practically all parts of the body. One of the
earliest suggestions that they are proprioceptive organs came from
Demoll (191 7), many of whose speculations concerning the mechanism
of action were given substance by the work of Pringle (1938 b). It is
interesting that there is no known case of these sensilla occurring on a
part of the cuticle free from strain. They are concentrated especially
where stresses are set up by muscular contractions (e.g., leg, wing,
haltere, ovipositor, and mandibular joints). A model has been
constructed by Pringle (1938 b) to explain their probable action
(Fig. 31).

A thin, flat surface such as the cuticle can be subjected to bending

MECHANORECEPTION 41

by forces at right angles to its plane or shearing in its plane, as illus-
trated by a sheet of rubber stretched over a rectangular frame (Fig. 3 1 ).
When the surface is distorted by changing the rectangle to a parallel-
ogram there is tension in the direction AC and compression in the
direction BD.

Any force applied to a hollow cylinder sets up shearing forces (see
Fig. 31). This is the situation as it occurs in a homogeneous surface to
which forces are applied evenly. Insect limbs do not conform to this
ideal, but, as Pringle pointed out, the important point is that on a
hollow structure such as the insect exoskeleton all stresses can be ex-

FiG. 31. Diagram to illustrate the resolution of shear forces into com-
pression and extension components. A, a flat surface; B and C, a
hollow cylinder subjected to twist and bend. Continuous lines, ex-
tension. Broken lines, compression. D, model representing actual
sensillum; E, model constructed in a circular framework for measure-
ment of the effect of compression in different directions. (Redrawn
from Pringle, 1938 b.)

pressed as shearing, which can be resolved into compression and
extension components.

Based upon these ideas, the probable mode of action of a campani-
form sensillum can be explained in terms of a second model (Fig. 31).
In a sheet of rubber stretched over a frame a circular or oval hole
corresponding to the cuticular canal of the sensillum is cut. A domed
strip of paper is fastened across the long diameter of the oval aperture.
This corresponds to the longitudinal thickening frequently seen in the
dome of the sensillum where the nerve fibre is attached. The rest of
the cap, assumed to be more elastic, is omitted from the model. The
assumption of differential elasticity of the two parts is essential to the
theory.

Distortion of the model from the rectangular shape causes the hole

42

THE PHYSIOLOGY OF INSECT SENSES

150

50 92 100

Resistance (arbitrary units)

Fig. 32. Graph obtained from the model in Fig. 3 IE showing the relation
between the angle of the applied compression force and the resistance
of the electrolytic cell for various values of the compression force.
(Redrawn from Pringle, 1938 b.)

to be lengthened in one direction and shortened in another. As a result
of this action, the paper strip humps up in the middle. If the end of the
neuron is imagined as being inserted on the underside of the middle of
the strip it will be stretched by movement in one direction and com-
pressed by that in the other. The magnification of movement is very
large.
In order to provide some measurement of the force involved,

MECHANORECEPTION 43

Pringle constructed another model in which there were two strips of
paper, one on each surface (Fig. 31). The movement measured
between the two and the relation between the angle of the applied
compression force and the resistance of the detecting electrolytic cell
is shown in Fig. 32.

If it be assumed excitation occurs only with increased doming, then
the neuron will be excited when the compression component of the
shear force makes an angle of less than a certain critical value with the

^^f ..fl^

/i)

<tBB^

Fig. 33. Details of the orientation of the groups of campaniform sensilla on
the third leg oiPeriplaneta. (Redrawn from Pringle, 1938 b.)

direction of thickening of the cap membrane. The sensitivity of the
sensillum will depend, among other things, on the absolute length of
the long diameter of the cap. Sensilla with parallel orientation will
respond to the same type of shear force. Each group will act as a unit
and, if innervated from the same fibre, will extend the range of these
fibres quantitatively but not quaHtatively (Fig. 33). This condition
occurs in the sensilla of the palps (Pringle, 1957).

Although there is yet no direct evidence that the sensillum has
undirectional sensitivity, considerable circumstantial evidence sup-
ports the idea. In many instances groups of sensilla are orientated in
the same direction. The sensilla on the tarsus have their long diameters

44 THE PHYSIOLOGY OF INSECT SENSES

parallel to the length of the leg (similarly in the palps) and when the
tarsus is pressed against the ground the compression lines in the cuticle
of the upper surface, where these sensilla are located, will also be
longitudinal. It must be the compression component which is effective.
Electrical recordings have shown clearly that these sensilla are most
sensitive to pressure on the cuticle (Pringle, 1938 b). Adaptation is
slow and incomplete.

Stretch Receptors

Certain of the Type II receptors for which no function had been known
have recently been shown by Finlayson and Lowenstein (1955, 1958)

Fig. 34. Stretch receptor in the right side of the fourth abdominal segment
of Periplaneta. The broken lines indicate the posterior ends of the
fourth and fifth muscle bands, counting from the dorsal mid-line.
(Redrawn from Finlayson and Lowenstein, 1958.)

to be stretch receptors analogous to the stretch receptors of Crustacea
discovered by Alexandrowicz (1951, 1952 a, 1952 b, 1954, 1956) and to
the muscle spindles of vertebrates. Firstdescribedby Zawarzin(1912 a)
and Rogosina (1928) in dragonfly larvae, they have now been found in

MECHANORECEPTION 45

Orthoptera, Hymenoptera, and Lepidoptera by Slifer and Finlayson
(1956) and Finlayson and Lowenstein (1955, 1958).

Three types of stretch receptors are known : those associated with
connective tissue, those associated with muscle, and those consisting
of a specialized muscle fibre. They are represented by receptors in
dragonfly larvae, bees, and cockroaches, by those in Orthoptera, and
by those in moths respectively.

Each connective tissue receptor is a multipolar neuron embedded in
a strand of connective tissue which stretches either from an inter-

FiG. 35. Right half of the sixth abdominal segment of a full-grown larva of
Aeschna juncea with various muscles cut to reveal the three stretch
receptors. Dva, dorso-ventral anterior muscle; Dvm, dorso-ventral
median muscle; Dvp, dorso-ventral posterior muscle; Sx, sextic
longitudinal tergal muscle; Obi, oblique receptor; Lo, longitudinal
receptor; Ve, vertical receptor. (Redrawn from Finlayson and
Lowenstein, 1958.)

segmental fold or from a nerve to a point on the body wall (Fig. 34).
In the cockroach Periplaneta americana a pair has been found in the
dorsal region of abdominal segments two to seven. In dragonfly larvae
{Aeschna juncea) three pairs of receptors have been found in each
abdominal segment one to eight (Fig. 35). In the honeybee there is a
pleural pair in each abdominal segment three to six. In all cases the
neuron is encapsulated in connective tissue. The capsule is multi-
nucleated. Detailed studies with light- and electron microscopy of
the abdominal stretch receptor of the cockroach Blaberus (Osborne
and Finlayson, 1962; Osborne, 1963) have shown that the terminal

46 THE PHYSIOLOGY OF INSECT SENSES

regions of the dendrites are naked, that is, not invested by Schwann
cells, and embedded in the connective tissue matrix. There are no
connexions with the collagen-like connective tissue fibrils. They
resemble very closely the sensory terminations of the vertebrate
muscle spindle and Pacinian corpuscle.

In Orthoptern (Acrididae) a pair of muscle receptor organs is present
in each of abdominal segments one to ten in the region of the mid-
dorsal line. The receptor organ is usually attached to the medial edge
of one of the longitudinal muscle bands, although it may be nearer the

Fig. 36. Dorsal longitudinal muscles of the third and fourth abdominal
segments of a pupa of Antheraea pernyi, with a part of the band in the
fourth segment removed to show the receptor. Motor and sensory
innervation of the receptor are shown. (Redrawn from Finlayson and
Lowenstein, 1958.)

middle. Each consists of a large multipolar neuron surrounded by a
thick nucleated capsule and a modified muscle fibre. There are also
many neurilemma cells. The muscle fibre originates near the anterior
end of the segment, and is inserted on the anterior end of the following
segment. The fibre possesses striations. Attached to it are numerous
thick connective tissue fibres which also extend to adjacent muscle
fibres. The axon of the receptor joins the ventral branch of each
segmental tergal nerve.

The most complex receptors are those in the larvae, pupae, and
adults of Lepidoptera. In the larvae there is a pair in the meso- and
meta-thorax and abdominal segments one to nine (Fig. 36). They are

MECHANORECEPTION 47

present in the abdominal segments of pupae and adults but have not
been found in the thorax. The receptor consists of a multipolar re-
ceptor cell, a modified muscle fibre, and motor innervation to the fibre
(Fig. 37a). The neuron is enclosed in the usual connective-tissue sheath .

Fig. 37. A. Diagram of the structure and innervation of a lepidopteran
stretch receptor, based on a preparation from the right side of the fifth
abdominal segment of a pupa of Antheraeapernyi. The receptor is short
because no tension is applied. B. Diagram showing the relative sizes
and positions of the two giant nuclei, the neuron sheath, and the fibre
tract of a lepidopteran stretch receptor. (Redrawn from Finlayson and
Lowenstein, 1958.)

Its dendrites sometimes run on finger-like extensions of the sheath,
which is continuous with a tubular sheath extending along the edge of
the receptor (Fig. 37b). The muscle which constitutes the bulk of the
receptor has fainter striations than normal muscles. At the centre there
is a clear swelling with a giant nucleus (100-360 ijl). Adjacent to this
is another large nucleus. This muscle has its own separate motor

48 THE PHYSIOLOGY OF INSECT SENSES

innervation. The resemblance of this receptor to the vertebrate muscle
spindle is striking. Both have rich motor innervation, an absence of
striations in the central region, and much nuclear material (nucleus in
the insect; nuclear bag in the vertebrate). In Crustacea and other
insects there is no giant nucleus, but there is, as already described, a
richly nucleated sheath.

Despite morphological differences, the mode of action of all of these
receptors is nearly identical. Under minimum tension, a condition

Time fsec.)

Fig. 38. A series of curves showing the discharge frequencies from the
stretch receptor of the larva of A. pemyi in response to stretch of
different intensities plotted against time. Arrows indicate beginning and
end of stretch. Figures below the curves denote the stretch in mm. The
broken line indicates response to 'over-stretch'. (Redrawn from
Finlayson and Lowenstein, 1958.)

approaching the normal situation, the receptor discharges impulses at
the rate of 5-10 a second. Under zero tension it is silent. The basal
rate is maintained for hours. When the receptor is stretched there is a
large increase in the frequency of discharge (Fig. 38). This drops very
rapidly to a new level which is maintained for a long period. Upon
release there is a post-stimulatory reduction in discharge. The rela-
tively large drop in initial discharge is probably due both to neuronal
adaptation and a mechanical accommodation to stretch. Over what
may be judged a normal range of stretch there is a linear relation be-
tween the intensity of stretch and the frequency of impulses (Fig. 39).
Beyond a certain point this relationship breaks down, probably

MECHANORECEPTION 49

because of *over stretch', i.e., extension beyond the physiological
range. The slowly adapting receptors clearly serve a static function.

The corresponding organs in Crustacea possess two neurons, a
*fast' and a *slow' one; that is, a phasic and a static receptor. It turned
out when the insect organs were examined with phasic stimulation
that the single neurons were in fact dual-purpose receptors (Lowen-
stein and Finlayson, 1960). At rest the discharge activity is, within
certain limits, a linear function of absolute length or total displace-

so

ii 40

3

mm.

Fig. 39. Peak discharge frequencies plotted against intensity of stretch.
A, larva of A. pernyi; b, longitudinal receptor of larva of A. juncea.
(Redrawn from Finlayson and Lowenstein, 1958.)

ment. With plastic alternation of stretch and relaxation the impulse
activity is a combined function of displacement and velocity. At low
stimulus frequencies the phase relation of maximum response is
approximately constant, and maximum activity coincides with maxi-
mum slope (= max velocity) (Fig. 40). Impulses begin to drop out as
higher stimulus frequencies (1-5-5 c/s) occur (Fig. 41). Above 5 c/s
the organ fails to signal phasic stimuli accurately, although the res-
ponse is still in phase with the stimulus. Up to this point the stimulus-
response relationship is linear (Fig. 42).

This receptor is thus seen to respond to displacement at low stretch
velocities, but from 3 to 4 c/s onward there is no response activity at
maximum displacement and during relaxation. Lowenstein and Fin-
layson (1960) were convinced that the decline of activity after the
velocity peak of the stretch is passed is accentuated by post-excitatory

50

THE PHYSIOLOGY OF INSECT SENSES

I f I I I ( I r ( c I I I r t f :i ( I c. ( r ( ( f r (. t. f I I I r f r rr r- t r.c r r 1- 1 r I t I I 1 I f I I i; r I I- 1 t.,

a'i,i| iiilillil li i i ilii i lllli lillll t lll l lll l|l l llill l ll| ll l ll l l l l ll i li i U'

1 second i __ -»___^___-^________^__.

( I (• f ( 1 t I ( I ( I r 1. I I . I I r c r I r I f r [ c. c c I I. r ( 1 ( ( ( r ( t f 1 I ( I i. I ( f t c ( r

I 1 I. I I 1

t (• III r I I r I ( I I (.1 I I. i'l I 'r ri '( t c t f c I c i. i i i"( i- c,- 1 l i c. c i i. ( ( r, l, u l t c r r c i; c'l. c l

llii i liii i iii i i i i i lii l iiilli l liiilil li iiiliiiil l llli i il ll l l li' l^^ ^

1 I . t • r I ( I r I I c

r r I I r r r r (. I r. r. l r i rmTTTTTm"^rrT c f i r- r f r f i l- r r l r. c- r r t. r

UiMllliilllllllilllilllllll

t.t t I c 1 1 I < r I I I ( L 1 .1 r c t.i I r c C fe I t t 1^1 I I I ( I. f r ( i I I ( ( I, I ( ( I ^ I I ; I I 1 J ( i t 1
•t^i^cTT r' I CI I I .f't I r t.i r. ui t.t i- 1- 1. i-.cL*cin*t t>f i r l i r i i i i i i i. i i i i i" i 1. 1 i i r i i i'

I.I I « I I I I I I I I

Fig. 40. Response to phasic stimulation during a continuous experiment
with a range of frequencies of alternation between stretch and re-
laxation, a, resting discharge ; b-e, frequency of stimulus and velocity
of stretch respectively, 0-25 c/s, 4-3 mm./sec; 0-5 c/s, 7-3 mm./sec;
1 -5 c/s, 22-2 mm./sec. ; 2 c/s, 33 -6 mm./sec. (Redrawn from Lowenstein
and Finlayson, 1960.)

il l W^'iV itl i V 'W ' l jl ' i 'i i' ' k 'iii Mi i i\ "i'i\"

mAAAAAAAAAAAAAAAAAAA

f) J ■. I .: J I 'i 1 1 1 .1 1 :) I- I I 1 I T 1 I I 1 1 1 1 I I I 1 1 1 1 1 ) I J J I I I o 1 i I 1 1 I 1 I 1 I 1 I 1 1 I I I

<) J 1 I I .1 I I I 1 J .1 .1 J I .1 I 1 I I I I I ' I I I I 1 I 1 I I I J ] I I I 1 I I I I I I I I ) I I I .1 ) I I I I I I I I

'i . . , , , I .■,,,,,,,,,,,,,,-•,,.,.■.,..... . •-.,.,,,...• . .....•..!

. ] 1 1 I i I I ) T 1 1 Jl 1 I 1 1 1 i I 1 I I I I I I 1_ I I T_1J )■) 1 I J ) I •) II ) I ) ) I 1 1 .1 1. I I I I 1 f 1 I 1 I 1;
H 1 J D 1 ■; J .1 J J :j > i J ! J J .) .1 J ) .) .1 J '1 .1 •.! .1 r J .1 .1 I.I II II I I I J I 1 I i r I I ) I ) I 1 II A I I 1 i'

m^^^ w^.^^^ ■ » I p^^ ^fi^A^ y^^^^mmm^m^tf^^m

i 1-i 1 J T 1 I 1 •] 1 1 1 n n a 1 J J ) I I I 1 1 I i I J I I • ,.j i i .■ i i i i i 4 i i i i i i i i ■ i i i ■ i i • i

Fig. 41. Continuation of experiment in Fig. 40. a-e, 4 c/s, 67-2 mm./sec;
5 c/s, 71 mm./sec; 15-5 c/s, 89 mm./sec; 18-5 c/s, 59 mm./sec;
recordings of responses in a and b respectively at ten times speed.
(Redrawn from Lowenstein and Finlayson, 1960.)

MECHANORECEPTION 51

inhibition. Thus, the gradual disappearance of response to displace-
ment and the transition from a static to a dynamic response is related
to the post-excitatory silencing.

The range of accurate frequency monitoring is more than adequate
to keep pace with such rhythmic activity as respiratory movements. In
their static capacities these stretch receptors are able to provide
information as to the relations of one part of the body with respect to

3^100

Q~-

1

1

1 1 1

1 lo y^ 1

/ • o

• /

y^ *

i

1 1

1 1 1

2

Stimulus Frequency cjs (•)

Fig. 42. Stimulus-response relationship. Frequency of phasic stimulus
(solid circles) (abscissa). Velocity of stretch (open circles) (abscissa).
Impulse frequency (ordinate). (Redrawn from Lowenstein and
Finlayson, 1960.)

another. In the dragonfly larvae, for example, the three pairs of
receptors are well adapted to supply information for analyses of body
movements. As Fig. 43 illustrates for dragonfly larvae, the oblique
receptor will be stretched during respiration (A) and relaxed during
expiration (B). In longitudinal abdominal movement, as, for example,
jet propulsion locomotion, the vertical and longitudinal receptors
will act antagonistically.

52

THE PHYSIOLOGY OF INSECT SENSES

Chordotonal Sensilla

Cord-like sensilla devoid of a specialized exocuticular component
and stretching from one point of the body wall or its derivatives to
another are peculiar to insects and have been found in every species
in which they have been sought (Siebold, 1844; Leydig, 1851 ; Graber,
1882a, 1882b; Eggers, 1924, 1928; Snodgrass, 1926; Debaisieux,

Fig. 43. Diagrams to illustrate probable alterations in tension of the three
stretch receptors of Aeschna larvae. A (inspiration) and B (expiration)
show the movements of the sternal and pleural regions of the body
wall during ventilation. Upper figures, dorso-ventral muscles. Lower
figure, oblique receptor of right side. The oblique receptor will be
stretched during A and relaxed during B. C and D illustrate probable
effects of longitudinal movement on longitudinal and vertical re-
ceptors. (Redrawn from Finlayson and Lowenstein, 1958.)

1935, 1938 ; Debauche, 1935; and others). They are widely distributed
in the body. In all insects they occur in the appendages of the mouth
and in the legs. They are universally present at the bases of the wings
(Fig. 44) and halteres (Fig. 45) and in the antennae (Graber, 1882 a,
1882 b; Schon, 1911; Vogel, 1912; Lehr, 1914; Eggers, 1928). They
occur segmentally in the abdomen. In many species simple chordo-
tonal organs are associated with the tracheal system (Demoll, 1917;
Larsen, 1955). They are especially numerous in larval insects, where
they stretch from one point on the body wall to another and criss-

MECHANORECEPTION

53

^^^^^=^^^i^^=:.^-^^g^^

Fig. 44. A. Right wing oiEmpis. B. Enlarged viewof insectshowinglocation
and orientation of campaniform sensilla. Ri, R2, R3, radial groups;
Sci, Scg, Scg, subcostal groups with arrows indicating the orientation of
the long axes of the sensilla. C. Right wing of Panorpa communis
showing the location of chordotonal sensilla. 1, 2, 3, 4, ante-alar,
radial, medial, and cubital sensilla respectively; Co, costa; Sc, sub-
costa; R, radius; M, media; Cu, cubitus; A, anal. (Redrawn from
Pringle, 1957 after Zacwilichowski.)

54 THE PHYSIOLOGY OF INSECT SENSES

cross in many directions (Fig. 49) (Graber, 1882 a, 1882 b; Radl,
1905; Hess, 1917).

Typically the sensilla are stretched between two undifferentiated
points of the body wall by a ligament at the proximal (i.e., axonal) end
and by the cap cell or accessory cells at the distal (dendritic) end.
Consequently, the organs have the appearance of thin strands
(Fig. 46) or sheets (Fig. 47).

Cue

Fig. 45. Dorsal view of the base of the left haltere of Tipula paludosa
showing the location and orientation of the group of campaniform
and chordotonal sensilla. Re, Cue, aac, radial cubital, and ante-alar
chordotonal organs respectively; Ri, Rg, R3, Scg, radial and subcostal
groups of campaniform sensilla with arrows indicating long axes of
sensilla. (Redrawn from Pringle, 1957 after Zacwilichowski.)

The structure of chordotonal sensilla is very complex and has yet
given no clue as to their exact mode of action. Each sensillum consists
of one bipolar neuron and a minimum of two companion cells which
are probably homologous with the trichogen and tormogen of other
sensilla (Schwabe, 1906; Schon, 1911 ; Vogel, 1912; Snodgrass, 1926;
Eggers, 1928). The three cells are arranged in overlapping linear
fashion. One, the enveloping cell, surrounds the proximal portion of
the dendrite ; the other, the cap cell, surrounds the remaining length

MECHANORECEPTION 55

of the dendrite. As in the case of other sensilla, the dendrite of the
neuron is enclosed terminally in a ribbed scolopoid sheath. This is
invariably capped with a darkly staining, peg-shaped apical body
(Fig. 48). In some instances there is a terminal strand extending from
the apical body to the distal point of attachment of the sensillum.

Although they were originally thought to be exclusively audio-
receptors, it is now known that they are not organs of constant
function. A proprioceptive function has been demonstrated for many.

Fig. 46. Longitudinal vertical section of a
pleural tubercle of Monohammus con-
fusor showing a simple chordotonal
organ. (Redrawn from Hess, 1917.)

and it is likely that all not associated with tympanic membranes or
grouped to form sub-genual and Johnston's organs (which have a
mixed proprioceptive and exteroceptive function and in Culicidae and
Chironomidae are auditory) will eventually be proved to be proprio-
ceptive. Their association with skeletal articulations, tracheae, pulsa-
tile organs, and blood cavities has led to hypotheses that they are con-
cerned with position, passive body movements, active muscle move-
ments, blood pressure, tracheal air pressure, and vibrations (Radl,
1905; Demoll, 1917; Hertwick, 1931; Debaisieux, 1938). It can
actually be seen by direct observation in many cases that the organs
change in length as the insect moves (Radl, 1905; Larsen, 1955).

E

56 THE PHYSIOLOGY OF INSECT SENSES

Proprioceptive chordotonal organs are usually much simpler than
those concerned with sound reception, although there are some note-
worthy exceptions (e.g., the tympanic organ of noctuid moths).
Whereas the auditory organs may consist of tens to thousands of
chordotonal sensilla, organs subserving a proprioceptive function
consist of very few sensilla, sometimes only one.

Fig. 47. Subgenual organ of Formica sangidnea. A, accessory cell ; C, cap
cell ; E, enveloping cell ; S, sense cell ; N, nerve. (Redrawn from Schon,
1911.)

In the water bug Aphelocheirus, Larsen (1955) discovered a single
chordotonal sensillum attached to each of the large tracheal sacs of
adults and larvae. He suggested that the sensilla are concerned with
respiratory behaviour. When air pressure within the animal falls, as it
does according to Larsen, and when an animal in poorly oxygenated
water has used up its air supply, there is a decrease in the volume of the
air sacs. The decrease triggers the chordotonal sensilla, and the animal

MECHANORECEPTION 57

can attempt to move to a more favourable environment before acute
respiratory distress ensues.

It has been suggested that other chordotonal sensilla in the abdomen
of insects may act as rhythmometers in respiration (Eggers, 1928;

Fig. 48. Diagrammatic repre-
sentation of a chordo-
tonal sensillum. C, cap
cell; E, enveloping cell;
S, sense cell; Sb, scolo-
poid body; D, dendrite;
A, axon. (Redrawn from
Schwabe, 1906.)

Hughes, 1952). In Dytiscus and Locusta comphcated sensory activity
has been recorded in afferent segmental nerves of the abdomen
(Hughes, 1952). Three patterns of impulses have been recorded, indi-
cating that there are some end-organs that discharge during inspira-
tion, others that discharge during expiration, and some whose

58 THE PHYSIOLOGY OF INSECT SENSES

discharge is inhibited during inspiration. Some end-organs with fibres
in the segmental nerves respond to whistUng, ground tapping, and
respiration, others to sound only, others to respiratory movements
only. Those responding to sound are sharply tuned to about 100 c/s,
and the only parameter of the stimulus to which their frequency of
discharge is related is intensity. Since the segmental nerves from which
these signals were recorded are known to contain fibres from seg-

FiG. 49. A segment of the
larva of Corethra
plumicornis showing
the location of one of
the segmental chordo-
tonal organs (C). (Re-
drawn from Graber,
1882.)

mental chordotonal organs, it was suggested that they are the sensilla
involved in these activities (Hughes, 1952). Pumphrey (1940) also
suggested that the activity recorded from the segmental nerves of
locusts arises in chordotonal sensilla and not in hairs as formerly
believed (Pumphrey and Rawdon-Smith, 1936 c). It should be re-
membered, however, that the Type II neurons of stretch receptors are
also sensitive to respiratory movements and to periodic motion, so
that further investigation of chordotonal responses is needed.

MECHANORECEPTION 59

Johnston's Organ

In 1855 Johnston discovered in the antennae of cuhcine mosquitoes
the compound chordotonal organ that now bears his name. Its
structure was not investigated in any detail until nearly forty years
later (Child, 1894). Since then it has been found to be almost a con-
stant feature of insect antennae, and much intensive study has been
given to its structure (e.g., Lehr, 1914; Eggers, 1923, 1924, 1928;
Debauche, 1936; Richard, 1956, 1957; Urvoy, 1958). In its classical
form it is basically a hollow cylinder or truncated sphere in the second
antennal segment (pedicel), where it stretches from the base of the
segment distally to the synovial membrane of segment three. The
sensilla comprising the organ vary in number from tens to hundreds in
the different species. Although the sensilla differ from those in other
chordotonal organs in that they lack caps or apical bodies, both
Eggers (1923, 1928) and Snodgrass (1926, 1935) agreed that they are
chordotonal sensilla.

Each sensillum consists of a bipolar neuron, a scolopoid sheath, a
cap cell, and an enveloping cell (Eggers, 1928). As seen with the hght
microscope, the scolopoid sheath, which is ribbed like most, becomes
attenuated distally into a bundle of thin fibres. The fibres from a
number of sensilla are grouped together for common insertion into a
pore, cleft, or some other attachment on the synovial membrane. In
the simpler Johnston's organs, where the total number of sensilla is
small, each bundle with a common point of attachment is clearly
separate from every other bundle. This arrangement suggests that the
Johnston's organ is really a compound structure composed of many
chordotonal organs. The closed cylinders of the complex Johnston's
organs are a consequence of the extraordinarily large number of
sensilla that must terminate in the limited space of the pedicle
(Eggers, 1928).

The organ is enormously developed in Culicidae and Chironomidae.
So many sensilla are crowded into the pedicel that the neurons lie in
several layers, and the cap and enveloping cells are so forced out of
position that they were originally described as supporting rods (Child,
1894). Furthermore, there is a corresponding modification of the
synovial membrane to accommodate the large number of sensilla. The
base of the third segment in male culicine mosquitos, for example,
flares to form a reinforced circular plate from which rigid spines
extend radially far into the pedicel. The sensilla are attached to these
spinous processes. In these and other species where the organ is a

60 THE PHYSIOLOGY OF INSECT SENSES

hollow cylinder both its inner and outer surfaces are covered with
epithelium that is a continuation of the hypodermis. In one case
(Hymenoptera) every sensillum is individually ensheathed in epi-
thelium. In addition to Johnston's organ there are smaller more typical
chordotonal organs present in the scape, pedicel, and flagellum of
many insects.

Everyone who has ever looked at Johnston's organs is impressed
by the fact that they are ideally situated to respond to movements of
the third antennal segment with respect to the second. This means,
essentially, movement of the entire flagellum of the antennae with
respect to the base, since all antennal muscles are located in the scape
and insert on the pedicel.

The stimuli for motion of the antennae may be many. Johnston
(1885) surmised that the organs of Culicidae were sensitive to air-
borne sounds, and Mayer (1874) demonstrated that the long hairs on
the shaft of the flagellum could be set in motion when a tuning fork
was struck near by. Child (1894) decided that in Culicidae and
Chironomidae the organ has an auditory function, but otherwise is a
tactile organ. The legend has grown up that Johnston's organs in
general are auditory organs, but Demoll (1917) and Eggers (1928)
went to great pains to point out that the only evidence for an auditory
function to date, and that not conclusive, was the observation of
Mayer (1874) for Culicidae. They therefore considered Johnston's
organs to be organs concerned with movements of the antennae used
actively as tactile appendages and with passive movements induced by
air currents.

As might have been suspected from the fact that the Johnston's
organ is a compound and complicated organ, the kinds of mechanical
stimuli to which it can react are varied. A study of the action potentials
produced in the antennae of the blowfly {Calliphora erythrocephald)
revealed that the majority of sensilla house phasic receptors, but that
some of these respond to torsional movements independent of direc-
tion, while others are sensitive only to movements in one of the two
possible directions (Burkhardt, 1960). When wind blows on the an-
tenna it causes the arista to act as a lever arm which rotates the
funiculus outwards around its long axis (Burkhardt and Schneider,
1957). A typical electrical response from a single rotation consists of a
large (+5 mV) action potential (on-wave) foflowed by a series of
decreasingly small action potentials during the course of stimulation
and ending with a large off'-wave at the termination of stimulation.
The small potentials are released by oscillations superimposed at the

MECHANORECEPTION 61

time of rotation. The amplitude of the on-wave increases with stimulus
intensity. It seems that the on-wave represents the summation of many
synchronously discharging neurons. The change in amplitude with
change in stimulus intensity is a reflection of the dilTcrcnccs in thres-
hold of the contributing units.

The time-course of the off-wave differs slightly. If the direction of
rotation of the funiculus is reversed the on-wave now resembles the
off-wave at the end of the stimulus for the original direction of ro-
tation, and vice versa. Burkhardt (1960), by way of explanation,
suggested that some of the receptors respond to turning in one direc-
tion only, while others respond to turning in any direction. The
geometrical arrangement of sensilla is such that mechanical stress in a
particular direction serves as a stimulus; however, different sensilla
are oriented differently with respect to the axis of rotation. Directional
sensitivity of this sort had already been described for Locusta by
Uchiyama and Katsuki (1956) on the basis of action potentials
recorded by means of a microelectrode. There is no support for the
suggestion of Kuwabara (1952 b) that the Johnston's organ is stimu-
lated by changes in the pressure of body fluid in the pedicel caused by
its moment of inertia during antennal movement.

The pattern of discharge following rhythmic stimuli of short dura-
tion changes as the frequency of stimulation changes. As the intervals
between stimuli are shortened below 20-30 msec, the off-wave of one
stimulus approaches more closely the on-wave of the succeeding stim-
ulus. When intervals are double the duration of stimuli the on- and
off-waves present a picture of regularity resembling the pattern
elicited by air-borne sound (Burkhardt and Schneider, 1957). The
response to sound frequencies up to 500 c/s reported by Burkhardt and
Schneider (1957) probably arises from rigidly synchronized discharges
at the beginning and end of each torsional vibration stimulus (Burk-
hardt, 1960). As the duration of stimulus is decreased the on- and off-
waves approach one another, and eventually the off-wave disappears.
It appears to be inhibited by the on-wave rather than to sum with
it. Thus, minute changes in stimulus pattern elicit marked altera-
tions in the spatio-temporal excitation pattern in Johnston's organ
(Burkhardt, 1960).

Other patterns of excitation can also be detected in antennae, but
the identity of the sensilla in which they originate is not known. There
are a few unknown tonic receptors (Burkhardt and Schneider, 1957)
(also in Locusta, Uchiyama and Katsuki, 1956), by means of which a
constant deflexion of the antenna can be signalled. When subjected to

62 THE PHYSIOLOGY OF INSECT SENSES

Steady air-flow, however, the antennae appear to tremble. Under these
circumstances the phasic receptors would register on- and off-waves
and thus could also register a mean deflexion (Burkhardt, 1960). In
short, Johnston's organ can provide the central nervous system with
rather accurate information about the strength and time course of
antennal deflexion.

THE ROLE OF MECHANORECEPTION

IN LOCOMOTION
Walking

Locomotion requires the co-ordination of moving parts. Although it
is possible that movements associated with locomotion can arise from
an innate central pattern, especially in the case of flying, as Wilson
(1961) has shown for the desert locust {Schistocerca gregaria Forskal),
a steady stream of information about the position of parts relative to
one another and about the magnitude of forces acting upon them is
undeniably necessary. In so far as flight is concerned, Weis-Fogh
(1956) and Pringle (1957) favour a complete reflex explanation of
movements. The consensus now is that walking, too, is reflexly con-
trolled (Pringle, 1940; Hughes, 1957, 1958).

In addition to feedback loops from the moving parts, locomotion
also requires information about the orientation of the insect as a whole
in the gravitational field. The source of all of this information is pre-
dominantly the mechanoreceptors. All types of mechanoreceptors
provide the service, and one of the marvels of the central nervous sys-
tem is its ability to integrate all of the incoming signals so that the
necessary adjustments in the performance of the effectors can be made
quickly and accurately.

In walking, for example, the reflex action of the leg muscles is evoked
by stimulation of mechanoreceptors. Most of the muscles of the insect
leg are innervated by two nerve fibres only, a *quick' and a 'slow'
(Hoyle, 1957). Stimulation of the quick fibre is followed by a large
electrical change and a rapid twitch, whereas stimulation of the slow
fibre is followed by a smaller electrical change and a tonic contraction.
In the cockroach Periplaneta americana L., as Pringle (1940) has
shown, the depressor reflex is evoked by stimulation of campaniform
sensilla on the trochanter; the levator response, by touch on the upper
side of the leg or on the tibial spines. Most of the campaniform sensilla
are located and oriented so as to be stimulated by forces produced in
the leg when the insect is standing normally. The situation may actually
be much more comphcated than this because of the possibihty of the

MECHANORECEPTION 63

muscles being compound ones (Bccht, 1959) and because of the large
number of mechanoreceptors (tactile hairs, hair plates, campaniform
sensilla, chordotonal sensilla) and their complicated innervation
(Nijenhuis and Dresden, 1952; Dresden and Nijenhuis, 1958). That
the situation is indeed complicated is further suggested by the detailed
cinematographic studies of Hughes (1952, 1957) on normal cock-
roaches and amputees. As various legs are amputated, the mechanical
aspects of walking change. There is a corresponding change in proprio-
ceptive feedback, with the result that gait and other aspects of walking
are altered.

In the stick insect (Carausius morosus) the posture in walking is
regulated by feedback control via hair plates which measure the angle
between the coxa and trochanter-femur (Wendler quoted by Mittel-
staedt, 1961). In the normal insect the body is held free from the
ground as a result of this feedback. The distance above the ground
remains the same even when the insect is carrying four times its weight.
If the sense organs are eliminated the insect touches the ground under
its own weight.

Flying

Flight, of all forms of locomotion, is the most demanding of infor-
mation. Whether one believes that the basic co-ordination of flight is
an inherent function of the central nervous system modulated by sen-
sory feedback (Wilson, 1961) or that flight is more exclusively a matter
of reflex control (Weis-Fogh, 1956; Pringle, 1957), it is certain that its
initiation, maintenance, adjustment, and termination can be effected
by means of elaborate sensory mechanisms. A brief survey of the prob-
lems involved will indicate the role of the mechanical senses. A com-
plete treatment is given by Pringle (1957).

The most specific and universal reflex for initiating flight is the
'tarsal reflex' originally described by Fraenkel (1932). When the feet of
most insects are removed from the substrate, flight commences. (There
is a similar reflex in giant water bugs, Lethocerus americanus and
Benacus griseus, whereby breaking of contact with the substrate ini-
tiates swimming movements [Dingle, 1961].) So essential to flying
insects is this reflex that in some species (Phormia, Calliphora, Vespa)
amputation of the legs interferes with the ability to cease flying. Con-
versely, contact with the substrate terminates flight. The electro-
physiological studies of Pringle (1938 b, 1940), have shown that at
least in Periplaneta the principal receptors involved are sensilla
campaniformia on the trochanter and femur. On the other hand.

64 THE PHYSIOLOGY OF INSECT SENSES

amputation of the tarsi of flies abolishes the reflex (Fraenkel, 1932;
Friedman, 1959).

In Periplaneta another flight-initiating reflex has been described by
Diakonoff'(1936). This is a change in the relative positions of the pro-
and mesothorax, the so-called 'fall-reflex'. The sense organs concerned
are small trichoid sensilla functioning in a manner similar to that of
hair plates. A similar mechanism can cause flight to commence in
giant water bugs (Dingle, 1961). Giant water bugs are also stimulated
to fly by action of wind on sensilla trichodea located on the head
between the eyes.

Although some insects {Drosophild) may continue to fly to ex-
haustion without further mechanical stimulation (Chadwick, 1939;
Wigglesworth, 1949), others (e.g., Muscina and Schistocerca) require
continuous stimulation. In the case oi Muscina stabulans a flow of air
against the antennae, causing the arista-bearing segment to move with
respect to the second joint, results in the legs being flexed in the flying
position. Air flow against the antennae also seems to be necessary for
sustained flight (Rollick, 1940).

In Schistocerca ten groups of sensilla trichodea bilaterally arranged
on the leading surfaces of the head are sensitive to wind. When they are
stimulated the forelegs are flexed in flight attitude and continuous
flight occurs (Weis-Fogh, 1949, 1956). They assist in stabihzation in a
horizontal plane (yaw), as is shown by the fact that the locust turns if a
jet of air is directed on the hairs from the side rather than from the
front (Weis-Fogh, 1949). Whether they are static or phasic receptors
or both has not been satisfactorily ascertained. It has been observed
that they do not vibrate in the wind; they are damped. The hair plates
on the legs oi Periplaneta, which these resemble, are sensitive both to
continuous deformation and the vibrations of a loudspeaker (Pringle,
1938 c). According to Weis-Fogh (1956), *wind on the wings' is also an
adequate stimulus for the maintenance of flight in Schistocerca.

Maintenance and alterations of flight velocity depend in con-
siderable measure upon information supplied by the Johnston's organ .
The detailed experiments with Calliphora (Burkhardt and Schneider,
1957), Aedes (Bassler, 1957, 1958), and Apis (Reran, 1957, 1959) have
revealed its hitherto unsuspected role. When insects are flying, air
currents impinging on the antennae cause the Johnston's organ to be
stimulated. As a result of the sensory information received, flight
velocity is reduced and the antennae are actively brought forward.
Forward alignment reduces the angle of attack of the air current, with
consequent reduction in the intensity of stimulation. Reran suggested

MECHANORECEPTION 65

that proprioceptive information about the position of the antennae
relative to the head supplements information from Johnston's organ
for steering flight velocity.

When the funiculi of the fly's antennae are rotated in their sockets
and fixed in the position which they would normally assume if wind
were blowing on them the flies fly much more slowly than controls
(Burkhardt and Schneider, 1957). In aphids flight is impaired when the
flagella of the antennae are removed; however, when artificial flagella
are supplied as replacements control is regained (Johnson, 1956).

Change in velocity of flight can be eff'ected in a number of ways. The
experiments of Hollick (1940) with Mucina showed that these flies
shift the path travelled by the wing tip cranially if they are subjected to
a current of air from in front and if they are in possession of their
antennae. In the absence of antennae, there is no shift. Wing-beat
amplitude is also altered as a result of information received through
the Johnston's organ. Bees reduce the wing-beat amplitude in the face
of air currents, but if they lack antennae the reduction is less marked
(Heran, 1959). In the case o{ Aedes there is a greater wing-beat ampli-
tude in air currents when antennae are lacking than in still air when
antennae are intact (Bassler, 1957, 1958).

Burkhardt and Schneider (1957) had pointed out that the antennae
of Calliphora also respond electrophysiologically to sound in the range
1 50-250 c/s, the same range as the wing-beat frequency. They proposed
that this response might possibly be employed to register the acceler-
ation due to each wing-beat, since the funiculi of the antennae vibrate
at this frequency as a result of the discontinuous air stream arising with
each wing thrust. Support for the hypothesis is derived from the obser-
vation that the resonance frequency of the bee's antennal flagellum
also agrees well with the wing-beat frequency, and the Johnston's
organ is most sensitive to vibrations of 200-350 c/s (Heran, 1959).
No tonic component similar to that detected in Calliphora can be seen
in electrical activity in the bee's antenna. Furthermore, if vibrations
too fast for the flagellum to follow are forced upon the antenna it is
bent outwards but no longer vibrates, and the bees do not respond to
the change in position (Heran, 1959).

Information from the Johnston's organ is also used to correct yaw.
In turning about the vertical axis the angle of attack of the antenna
on the outside of the turn becomes larger. Since intensified current on
an antenna reduces wing-beat on the same side, a passive rotation
produces an active torque in the opposite direction by which the
straight flight course is stabilized (Heran, 1959).

66 THE PHYSIOLOGY OF INSECT SENSES

Other monitoring of the movements of the wings occurs as a result
of stimulation of wing receptors. As many investigators (e.g., Vogel,
1911, 1912; Lehr, 1914; Erhardt, 1916; Eggers, 1928; Hertwick,
1931; Zacwilichowski, 1936) have shown, the wings of insects are
elaborately equipped with sense organs (Fig. 44). These are sensilla
trichodea, sensilla campaniformia, and chordotonal sensilla. They
are strategically situated to respond to wind on the surfaces of wings or
to forces acting on the veins. With the exception of one chordotonal
organ, all sensilla are located distal to the hinge; consequently, they
will be subjected to strains and distortions set up by aerodynamic and
inertial torques, but not elastic torques (Pringle, 1937). Several
investigators have shown by electrophysiological monitoring that
considerable information is fed back to the central nervous system by
these sense organs (Sotavalta, 1954; Wolbarsht and Dethier, 1958;
Wilson, 1961).

The sensilla trichodea are clearly tactile. Although Pringle (1957)
did not exclude the possibility that they react to air flow, he suggested
that this is improbable, because the small fibre diameter precludes
rapid enough impulse conduction to meet the demands of quick flight
reflexes. On the other hand, it is known that these hairs mPhormia fire
at great speed and adapt very rapidly (Wolbarsht and Dethier, 1958),
hence might respond to turbulence. The loss of control during the few
wing-beats necessary to allow time for nervous conduction might be
inconsequential considering the high rate of wing-beat in most insects.

The sensilla campaniformia are distributed singly and in groups.
The former are usually circular, hence not markedly directional in
their sensitivity. Grouped sensilla are usually oval. The direction of the
long axis is uniform within a group but differs from one group to
another. As Pringle (1938 b) has demonstrated, each group constitutes
a unitary organ selectively sensitive to strains whose compression axis
is oriented parallel to the direction of the long axis of the sensillum
dome.

The location of each group is such that the sensilla are maximally
sensitive to a particular direction of torque in the vein wherein they He.
Pringle (1937) has discussed at length the probable mode of action
of these sensilla in relation to the torques acting on the wing during
flight. Probable modes of action of the chordotonal organs have been
deduced with less certainty (Pringle, 1937).

Any flying machine must possess stability, that is, a tendency to
return to a characteristic attitude when displaced if it is to be con-
trollable in the air (Pringle, 1950). In this respect insects are able to

MECHANORECEPTION 67

control lift and to stabilize reflexly in all three planes of rotation, that
is, to correct for pitch, yaw and roll. In insects with four wings much of
the required information derives from mechanoreceptors on the wings.
As a result, changes in pronation, supination, and angle of attack of
the wings are made until stable flight is achieved. Details of the res-
ponse characteristics of the various sensory structures of the wing,
however, are poorly known.

More information is available about the halteres of Diptera. These
structures, homologues of the hind wings, are radically specialized as
balancing organs. They are lavishly equipped with all types of
mechanoreceptors whose homologies can be traced to those of the
wings of non-dipterous insects (Pflugstaedt, 1912; Zacwilichowski,
1936). Essentially, each is a heavy mass of tissue on the end of a thin
stalk. The folding of the base is very complicated, consisting of a main
hinge and a condyle of secondary articulation (Fig. 45) (Pringle, 1948).
In these regions are concentrated, in Calliphora, for example, about
418 mechanoreceptors. They include the following sensilla as des-
cribed by Pflugstaedt (1912) :

Campaniform Sensilla

Dorsal scapal plate
Ventral scapal plate
Basal plate
Dorsal Hicks papillae
Ventral Hicks papillae
Undifferentiated papilla

Chordotonal Sensilla

Large chordotonal organ
Small chordotonal organ

All but thirty-five are located in the three plates and the large chordo-
tonal organ. It is significant that the axons from these sensilla do not
fuse as is the case with the campaniform sensilla in the palps of the
cockroach (Pringle, 1938 b, 1948). The arrangement of sensilla is
basically the same in all families of Diptera (Brauns, 1939). It has been
suggested (Pringle, 1948) that strains produced by vertical oscilla-
tion of the haltere are detected by the dorsal and ventral scape plates,
the dorsal and ventral Hicks papillae, and the small chordotonal
organ. Strains produced by gyroscopic torques are presumed to be
detected by sensilla of the basal plate and the large chordotonal organ.

68 THE PHYSIOLOGY OF INSECT SENSES

The undifferentiated papilla may be sensitive to all strains in the
cuticle of the base.

The beat of the two halteres is synchronized. They are oscillating
masses which generate forces at the base of the stalk as the whole fly
rotates (Pringle, 1948). They are thus gyroscopes, but it is questionable
that they act as direct stabiHzing gyroscopes (Schneider, 1953). It is
more likely that their action is indirect, in that the activity of their
sense organs signals the nervous system to set up the necessary correc-
tions in the flight mechanism (Pringle, 1957). There is convincing
experimental evidence that they are instrumental in the control of yaw
(Pringle, 1948; Faust, 1952; Schneider, 1953), pitch (Faust, 1952), and
roll (Faust, 1952).

Finally, in the control of flight there are visual stimuli (e.g., Schaller,
1960). Movement in the visual field is employed directly by Calliphora
in incorrecting yaw (Schneider, 1956) and roll (Faust, 1952 ; Schneider,
1956). Vision is also employed indirectly (Antrum and Stocker, 1952;
Mittelstaedt, 1950) to correct roll, as the following example of dragon-
fly (Anax) behaviour illustrates.

The head is a broad, heavy object, and rotations of the body produce
movement in the neck region. The position of the body relative to the
head is signalled by two pairs of prothoracic hair plates. When the head
of a flying or stationary dragonfly is twisted there is a compensatory
twisting of the wings. In flight this corrects rolling; when the insect is
perched it assists in the maintenance of balance. In nature the insect
moves its head so that the light intensity is always greatest dorsally. By
means of proprioceptive feedback from the hair plates, the body is
then rotated until it is ahgned with the head (Mittelstaedt, 1950).

Swimming

Insects that swim beneath the surface of water are faced with orienta-
tion problems more nearly resembling those encountered in the air
than on land. They can move freely in three dimensions in a homo-
geneous medium and must possess stability, just as must flying insects.
In Notonecta, Naucoris, and Macrotorixa the normal swimming
position also represents the stable equiUbrium position. This position
is maintained in the imago oi Naucoris and Macrotorixa by the distri-
bution of air and the position of the legs; in Notonecta larvae, princi-
pally by the location of the centre of gravity of the body mass (Oever-
mann, 1936). When these insects are Winded, deprived of the air film
on the surface of the body, and artificially weighted so that the position

MECHANORECEPTION 69

of Stable equilibrium is altered they are still able to orient with respect
to gravity (Oevermann, 1936). No special localized static organ has
been found; however, in Notonecta, Plea, Naucoris, and Corixa the
Johnston's organ is employed in conjunction with a bubble of air to
mediate position sense (Rabe, 1953). Notonecta and Plea swim ventral
side uppermost. When they are properly oriented an air bubble
trapped between the antennae and the surface of the head causes by its
buoyancy the antennae to be deflected away from the head. If the
insect is turned upside down the antennae are then deflected towards
the head. In each instance the deflexion activates the Johnston's organ.
If the bubble is removed, Notonecta placed in darkness will swim dorsal
side up (that is, reversed) because the antennae by their own weight
lean away from the head. For Corixa, which swims dorsal side up, the
mechanism is the same but acts in reverse (Rabe, 1953).

Lack of a speciahzed static organ is characteristic of many aquatic
species, and it appears that much assistance is derived from extero-
ceptors. BHnded insects, while not losing their ability to retain their
normal attitude and position with respect to gravity entirely, do tend
to show locomotory aberrations (Oevermann, 1936; Tonner, 1938;
Hughes, 1958; Dingle, 1961). Compensatory movements to rotation
cease when eyes are blackened, and some species show a tendency to
swim in spirals when unilaterally bhnded. In larvae o^ Acilius sulcatus
and Dytiscus marginalis, which possess twelve simple eyes (stemmata),
Schone (1950) has shown that the structure and arrangement of these
eyes are especially suitable to mediate optical orientation in space by
sensing the distribution of brightness in the visual space. In rotations
around the longitudinal axis a displacement of the ratio of excitation
between the right and left groups of eyes occurs ; in rotations around
the transverse axis a change in the quotient front : back forms the
basis of regulation.

Exteroceptive information concerning water movement is received
via the antennae. It is assumed that Johnston's organ is involved. A
swimming Dytiscus holds its antennae at a slight angle to the median
line. Any deviation from the axis of swimming causes the antennae to
bend (Hughes, 1958). Some proprioceptive information is necessary
in any case. Giant water bugs possess hair plates on the trochanters at
the coxo-trochanteral joints which are important for the co-ordination
of swimming (Dingle, 1961). When the legs are flexed, as when grasping
a stem or when flying, the hair plates are stimulated by a covering fold
of cuticle. When the legs float freely in the water in an extended
position the hair plates are no longer stimulated by the fold, and

70 THE PHYSIOLOGY OF INSECT SENSES

swimming occurs. Destruction of the hair plates interferes with normal
swimming.

By analogy with marine invertebrates, especially Crustacea, it might
be expected that statocysts would be common in aquatic insects. Such
organs are ideally constructed for responding to gravity in a medium
where cues are few. Aside from a series of questionable cases studied
by Wolff (1922) and Studnitz (1932), however, no true statocysts have
been reported. In the larvae only of many Limnobidae (Diptera) there
is located on the terminal segment a pair of small sacs, open to the
outside, and equipped with muscles by which water can be pumped in

Fig. 50. The organ of pressure sense of Nepa. T, trachea ; N, nerve ; M,
membrane of overlapping expanded margins of scale sensilla; C,
closed spiracle ; S, sensory papilla. (Redrawn from Thorpe and Crisp,
1947.)

and out (Fig. 51). Invariably each pouch contains a few minute par-
ticles either formed by the larva itself or picked up from the outside. As
the muscles of the sac contract and relax, the particles are rattled
about the interior.

In the interior of the sac are usually two sensory hairs, one at the
blind end or deepest part of the sac, the other, in a lateral position.
Wolff (1922) was convinced that these organs were not static in
function. Studnitz (1932), on the other hand, after proving that uni-
lateral or bilateral destruction of the organs abolished the positive
geotaxis characteristic of the normal larvae, concluded that they were
indeed statocysts.

In water bugs (A^e/?^ and ^/j/ze/oc/ze/rw^) there are elaborate mechano-
receptive organs which appear to be designed to respond to pressure

MECHANORECEPTION 71

changes and are presumed to have a static function. The organs in
Nepa, situated adjacent to the spiracles of the third, fourth, and hfth
abdominal segments, consist of groups of peg-shaped sensilla alter-
nating with umbrella-shaped sensilla (Fig. 50). According to Hamilton
(1931), the whole is covered by a thin membrane which docs occlude
the adjacent spiracle, but Thorpe and Crisp (1947) considered the
membrane to consist of the overlapping umbrella-like portions of the
sensilla which do indeed cover the spiracle. Baunacke (1912) con-
sidered these organs to be hydrostats. After he ascertained that a
blinded Nepa placed on an underwater seesaw would reverse its
direction of crawling as the inclination was reversed, he eliminated the

Fig. 51. Statocyst in the larva of an aquatic dipteran (Limnobiidae). The
black objects are particles in the cavity. (Redrawn from Wolff, 1922.)

organs. Elimination abolished the bug's response to the change of
inclination. Oevermann (1936) and Thorpe and Crisp (1947) con-
firmed these observations. Oevermann, however, demonstrated that
there must be proprioceptive information from the legs, because a
weighted insect can still make position corrections.

Baunacke (1912) believed that the three pairs of organs act as a
system and respond to relative changes in pressure. In other words, the
lowest pair (on the fifth sternum) are under greater pressure than
the upper pair (third sternum) when the bug is in certain attitudes. For
this system to work the gas-filled cavities of all of the sense organs
must be in communication with one another via the tracheal system.
According to Hamilton, who described the organs as covered with a
membrane sealing them from the trachea, this is not so, but Thorpe

F

72 THE PHYSIOLOGY OF INSECT SENSES

and Crisp maintained that there is communication. They performed a
series of extirpation experiments that are in agreement with
Baunacke's hypothesis (Fig. 52). In the control (A) bugs on a seesaw
gave sixty-eight correct performances out of ninety- two. B did not

. Controls intact

B. Middle pair
extirpated

C. All on one side
extirpated

O
O
O

O

X

O

o

X

O

O X
O X
O X

D. Anterior and
posterior pair
extirpated

E. Posterior pair
extirpated

F. Middle of one
side only extirpated

X

O

X

O
O

X

A

O

o

X

O O
O X
O O

KP

Kl^

Kp

1 ^ :

Fig. 52. A. Diagram illustrating the various operations performed on the
three pairs of pressure organs in Nepa. (Redrawn from Thorpe and
Crisp, 1947.) B. Diagram to illustrate the principle of action, when
tilted in water, of a system of distensible membranes enclosing an air
space and connected by an air-filled tube, of three pressure receptors
of one side of Nepa. (Redrawn from Thorpe and Crisp, 1947.)

differ statistically from A, C and D were generally negative; E res-
ponded fairly well and F less well. It was concluded that the three pair
of organs in Nepa were indeed co-ordinated as a differential mano-
meter system.

A less-compHcated organ occurs in adults of the predatory water
bug Aphelocheirus (Fig. 53). There is a pair of oval sensory plates on

MECHANORECEPTION 73

the second abdominal sternum. Each consists of a shallow ovoid
depression in which the usual plastron hairs are replaced by large
hydrofuge hairs (± 60,000 per sq. mm.), among which are delicate
innervated sensilla trichodea (Fig. 53). It is assumed that increases of

53. A. Ventral view of the left
spiracular 'rosette' on the second
abdominal segment of Aphelo-
cheinis with the specific organ of
pressure sense and the col-
lapsible air sac. B. Section
through the edge of the organ
of pressure sense of Aphelo-
cheirus showing one sensory hair
and its bipolar neuron and the
plastron hair pile. (Redrawn
from Thorpe and Crisp, 1947.)

pressure due to water currents and turbulence, depth, or the animal's
position with respect to gravity are registered because the large slant-
ing hairs press down upon the sensilla (Thorpe and Crisp, 1947).
Unilateral damage causes the animal to swim in spirals, but this is only
indirect proof of their static function (Larsen, 1955).
The air space around the pressure organs is continuous with tracheal

74 THE PHYSIOLOGY OF INSECT SENSES

air spaces ; hence, if the organs are to give meaningful information
regarding external pressure changes the internal air pressure must
be kept constant. Thorpe and Crisp (1947) have suggested that large
tracheal air sacs near these organs damp internal pressure changes. A
different interpretation of the function of these air sacs has already
been discussed (Larsen, 1955).

Comparable, though simpler, so-called static organs have been
described in Ranatra, Lethocerus, and Belostoma, but little is known of
their physiology (MoUer, 1921).

On the surface of water insects do not encounter the same sensory
problems as underwater, but, because of the great speed of swimming
that is possible, obstacle avoidance is a problem. Some imaginative
experiments by Eggers (1936 b, 1927) with the whirligig beetles Gyrinus
marinus and G. natator suggest that the antennae play a prominent
role. Beetles on the surface of water in a small container (diam. 35 cm.)
are able to avoid collisions with one another and with the walls of the
container. If the walls of the container are coated with paraffin so that
the meniscus is convex instead of concave or if the surface of the water
is carefully cleaned of all dust particles the frequency of collisions
against the walls increases markedly. When swimming beneath the
surface of the water the beetles also hit the wall. Neither vision, nor
detection of waves bouncing from the walls, nor air pressure produced
by the beetle moving towards the walls appear to be involved in
obstacle avoidance. Eggers postulated that detection of the meniscus
and detection of the resistance of the surface dust layer as it is com-
pressed between the beetles and the wall informs the beetle of the
proximity of the obstacle. It was presumed that the antennae, and
possibly their Johnston's organs, are the sensing elements involved.
Ordinarily the Johnston's organ is not highly developed in Coleoptera
(as compared with Diptera), but in Gyrinus it is fairly complex, and the
antennae have pecuhar structural modifications which, according to
Eggers, permit the pedicel to glide on the water surface while the
flagellum sticks up into the air. The pedicel would be very sensitive to
irregularities (menisci) in the water surface and to the resistance of
floating particles such as dust; when moved, it would cause counter-
movements of the flagella, with a consequent stimulation of John-
ston's organ. Amputation of the antennae results in an increased
number of collisions.

MECHANORECEPTION 75

RESPONSES TO GRAVITY

As the foregoing discussions have impHed, insects lack receptors
designed specifically for the detection of gravitational force. Infor-
mation required for maintenance of primary orientation is derived
secondarily by the central nervous system from many receptors whose
primary function may have little to do with responses to gravity. The
precision with which some insects respond to gravity and the import-
ance of a critical evaluation suggests that there may be some receptors
whose information is critical. Buckmann (1954) has demonstrated, for
example, that the staphylinid beetle BUdius bicornis Grm., which
burrows in sea sand, responds to gravity with a precision of the order
of one angular degree. This precision is not inferior to that exhibited
by other animals which possess statocysts (Buckmann, 1955 a).
Centrifugation experiments have indicated that neither light nor
surface features are cues. Unfortunately the nature of the required
sensory information is not known. In any event, the antennae are not
required (BUckmann, 1955 b).

Another insect for which a precise response to gravity is critical is the
honeybee. In order to be able to communicate the direction of food
sources by transposing the angle of flight with respect to the sun to an
angle of dancing with respect to gravity in a dark hive, the bee must
be able to assess very accurately the direction of gravitational force.
Vowles (1954), working with ants, had shown that the antennae were
of importance in these species. Iron filings were cemented to various
parts of the bodies of ants climbing a vertical surface. The ants were
then suddenly subjected to a magnetic force. Only when the filings
were on the funiculi of both antennae did the ants suddenly change
orientation. This result plus the knowledge that different orientations
with respect to gravity normally impart different rotational forces on
the funiculus led to the hypothesis that the Johnston's organs are
involved in geotaxis.

With the honeybee, however, the antennae are not so critical.
Instead, hair plates in the neck region and at the articulation of the
thorax and abdomen play a prominent role (Lindauer and Nedel,
1959). When the bee is standing on a horizontal surface the hairs of the
pro-thorax are in even contact with the back of the head. When the bee
crawls up a vertical surface the head will incline towards the sternum
because of its low centre of gravity. In this position the more ventral
hairs of the hair plate will experience maximal shearing forces. The
less the substrate is inclined towards the vertical, the smaller the

76 THE PHYSIOLOGY OF INSECT SENSES

shearing forces on the whole organ. When the bee is crawHng down a
vertical surface the head will incline dorsally and the dorsal areas of
the hair plate will be maximally stimulated. When a bee on a vertical
surface turns to the right or left the hair plates on different sides of the
thorax will be unequally stimulated. In this mechanism there is
potentially a very fine capacity for analysing the position angle on the
vertical plane. The degree of bending of individual hairs and the
spatial distribution of bend in the hair plates as a whole could provide
the necessary information (Lindauer, 1961). Severing the nerves to
these organs, immobilizing the head, or changing the centre of gravity
of the head by weighting it results in impairment of responses to
gravity and ability to orient communication dances correctly with
respect to gravity. The hair plates at the articulation of the thorax
and abdomen function in a similar fashion. They appear to be subord-
inate to the cervical organs, however, because they alone are unable to
regulate responses to gravity when the cervical organs are eliminated
or when the head is fixed or weighted.

A comparative study of the occurrence and structure of hair plates
at the joints of representatives of five famihes of ants, the honeybee,
and Vespa saxonica has shown that these setal fields occur at the
antennal, cervical, petiolus, and gaster joints and on the joints
between the thorax and coxae and coxae and trochanters (Markl,
1962). By removing sense organs, disconnecting the various joints,
or cementing the joints in abnormal positions, Markl was able to
demonstrate that the cervical hair plates are the most important for
gravity reception. The others in decreasing order of importance are
the petiolus, the antennal, the coxal, and the gaster. In agreement
with Vowles (1954) it was found that the antennal organs serve as
gravity receptors but they represent only one, and by no means the
most important, receptor system for gravity. In supplementing the
results of Lindauer and Nedel (1959) it was shown that the receptors
at the coxal joints but not those of the antennae serve for gravity
reception in the honeybee. Since the hair plates are able to perceive
active movements of the joints, it is argued that multiple development
of hair plates is necessary if these organs are to serve as gravity
receptors. Only a message in the same direction from one or more of
the hair plates is related to gravity by the central nervous system,
whereas aberrant responses from one or more areas are interpreted
as joint movement independent of gravity (Markl, 1962).

CHAPTER IV

Sound Reception

The to-and-fro motion of a particle repeatedly displaced from equili-
brium is termed vibration. In a continuous medium a vibrating particle
causes periodic displacement of neighbouring particles by elastic
forces. The resulting disturbance takes the form of periodic com-
pressional v^aves. Sound reception is the process of detecting these
waves whether they occur in gases, liquids, or solids. Whether one
defines all vibration detection as 'hearing', restricts the term 'hearing*
to the detection of air-borne sounds by specialized receptors and
relegates other sound reception to a 'vibration sense' (von Budden-
brock, 1952), or limits 'hearing' to that condition where an animal
behaves as if it has located a sound source (sound being defined as any
mechanical disturbance whatever which is potentially referable to an
external locaHzed source) (Pumphrey, 1950) is not important to an
understanding of how the receptors work. The crucial question,
physiologically speaking, is to what extent insects can detect periodic
motion or vibration in their environment and which parameters of the
wave are detected. In this respect it is helpful to inquire first what kind
of sounds or vibrations insects are exposed to normally.

Many sounds are produced by insects themselves. Many are inci-
dental to ordinary movements, but nearly all orders of insects have
members that produce specialized sounds. Fringsand Frings (1958),
in their review of the subject, have listed the following specialized
methods of sound production: (1) tapping the substrate; (2) explosive
expulsion of air or other material through a small orifice ; (3) snapping
a prosternal spine from a cavity in the mesosternum ; (4) vibrating the
body without flight to produce a buzzing, whining, or piping;
(5) snapping the wings in flight; (6) snapping tymbals or tymbal-like
organs; (7) stridulation (rtibbing specialized surfaces of the elytra
together or stroking the elytra with the hind femora) ; (8) expulsion
of air over a specialized vibrating membrane. The sounds produced
by insects are principally connected with sexual activity, territoriality,
defence against predators, and maintenance of cohesion in flying
swarms (Haskell, 1957). In addition to responding to sounds of these

77

78 THE PHYSIOLOGY OF INSECT SENSES

origins, insects respond aggressively to sounds produced by prey (e.g.,
the antlion to its prey) and defensively to sounds produced by pred-
ators (e.g., moths to the sounds of bats). Insects can also detect a
certain amount of the white noise that fills the physical environment.
So far as is known, only two kinds of mechanoreceptive sensilla
are sensitive to sound, sensilla trichodea and chordotonal sensilla,
although there is nothing about the structure of campaniform sensilla
or stretch receptors that precludes the possibility of their also respond-
ing to sound. Indeed, stretch receptors accurately signal phasic
stimulation up to 5 c/s (Lowenstein and Finlayson, 1960). Hairs
sensitive to sound do not differ fundamentally from those concerned
with tactile and proprioceptive functions (hair plates on the legs of
Periplaneta respond to loudspeaker sounds [Pringle, 1938c] ) nor do
the chordotonal sound receptors differ appreciably in structure from
proprioceptive chordotonal sensilla. The more highly specialized
chordotonal receptors, however, are grouped together and associated
with elaborate accessory structures enhancing their sensitivity to
sound. These organs include: tympanic organs, Johnston's organ in
culicine mosquitoes, and subgenual organs.

TYMPANIC ORGANS
Morphology

Few external structural details escaped the eyes of nineteenth-century
taxonomists, so it is not surprising that the occurrence of tympana was
noticed in a number of insects. In some cases the structures were
examined only with a view of their taxonomic usefulness ; in others,
guesses were made regarding their function. Those who guessed
correctly believed the organ to be an instrument of sound production.
Miiller (1826), studying the paired tympana on the first abdominal
segments of Acrididae, decided that they were organs of hearing.
Tympanic organs were then discovered in the tibiae of the prothoracic
legs of Tettigoniidae and GryUidae (Siebold, 1844) and later in moths
and butterflies. Among the Lepidoptera tympanic organs occur in the
abdomen of adults in the super-families Geometroidea and Pyraloidea
and in the metathorax of Noctuidea (Swinton, 1877; Jordan, 1905;
Deegener, 1909; Eggers, 1911, 1919; von Kennel, 1912). Abdominal
tympanic organs occur in the Cicadidae (Vogel, 1923 a) and in Corixa
andafewother Hemiptera (Graber, 1882 a, 1822 b; Hagemann, 1910;
Wefelscheid, 1912; Schaller, 1951), although in some of these forms
the variation from an obvious tympanic organ makes those sensilla
associated with tracheae look suspiciously like the proprioceptive

SOUND RECEPTION 79

chordotonal organs usually associated with tracheae in other species
Some chordotonal organs in the wings of Satyridae are associated with
tympana (Vogel, 1912).

The detailed inner structure of tympanic organs was investigated
thoroughly by Schwabe (1906), Eggers (1911, 1919, 1924, 1928), Vogel
(1923 a), Hers (1938), Roeder and Treat (1957), and Treat (1959).
These studies have been extended by the electron-microscope studies
of Gray (1960).

All tympanic organs have certain features in common. These
include: a thin cuticular membrane, a closely appressed internal
tracheal sac, and a group of chordotonal sensilla. The sensilla may be

Fig. 54. Diagram to show the method of attachment of the neural elements
to the inner surface of the tympanum of the locust. A, auditory nerve;
Fo, fold; Si and S2, location of sense cells; E, elevated process; S,
styliform body; L, leg; B, base; D, drum; P, pyriform vesicle; Fi,
fusiform body. (Redrawn from Gray, 1960.)

attached directly in the middle of the tympanum (Acrididae), to its far
edge (Cicadidae), or to the trachea instead of the tympanum itself
(Tettigoniidae and GrylUdae). They vary in number from two in
certain moths to 1,500 or more in cicadas (Figs. 55, 57, 58, 59).

Of all these, the tympanic organs of Acrididae have undoubtedly
been studied in greater detail than any others. The tympanum is a thin
(2-3 (J.), imperfectly ovoid area of exocuticle lying in a crater-like
depression (sometimes likened to the auditory meatus of the human
ear) formed by an incomplete rim of thickened cuticle (Fig. 54). The
tympanum is further strengthened circumferentially as a consequence
of a thickened cuticular invagination. It is thus a rigidly supported
drumhead. It is hned internally, as is almost all cuticula, by a thin layer
of hypodermal epitheHum. Closely appressed to the internal side of the
tympanum is a tracheal sac fed by a branch from the main trachea

80

THE PHYSIOLOGY OF INSECT SENSES

Fig. 55. Tympanic organ of the prothoracic leg of Orthoptera. Ta, anterior
trachea; Tp, posterior trachea; Sb, subgenual organ; I, intermediate
organ; A, acoustic organ; SbN, subgenual nerve; TmN, tympanic
nerve; C, cap cells, S, scolopoid cells. (Redrawn from Schwabe, 1906.)

(analogous to the Eustachean tube), whose spiracle lies at the anterior
border of the tympanum. The tympanum is in fact lined by two layers
of epithelium, that of the hypodermis and that of the tracheal sac, and
is bordered by air on both surfaces.

The tympanal nerve, a branch of the tergal nerve of the third thor-
acic ganglion, enters the region in a fold of the tracheal epithelium,
branches, and attaches at several points of the drum (Fig. 54) : the
folded body (rinnenformiges Korperchen), the styliform body (stiel-

SOUND RECEPTION 81

formige Korperchen), the elevated process (zapfenformige Korper-
chen), and the pyriform vesicle (birnformiges Korperchen). These
bodies are merely complex thickened areas of the drum. A small
branch of the nerve goes to the folded body, where it innervates a few
hairs and other sensilla. The main portion of the nerve, before branch-
ing to each of the four points of attachment, swells. This swelling has,

Fig. 56. Schematic dorsal stereogram of noctuid tympanic organ and
related structures. B, Biigel; C, conjunctiva or accessory tympanic
membrane; CTC, countertympanic cavity with external orifice
indicated by arrow; CTM, countertympanic membrane; E, epaulette
or nodular sclerite ; EL, electrode site ; H, hood ; L, ligament ; M^, muscle
originating on antero-median portion of metascutum and inserting on
metapostnotum anterior to CTM; Ma, muscle originating dorsally
on metascutum and inserting on epimeral band; MP, mesophragma;
MS, metascutum; Pi, pocket I of tympanic frame; PTG, pterothoracic
ganglion; S, sensillum; Si, metathoracic spiracle; Sg, first abdominal
spiracle ; SP, scutal phragma ; TAS, tympanic air sac lined with tracheal
epithelium; TM, tympanic membrane; TN, tympanic nerve; TR,
tympanic recess. (Courtesy of Roeder and Treat.)

82 THE PHYSIOLOGY OF INSECT SENSES

unfortunately, been called a ganglion by Gray (1960). It is in fact the
aggregation of most of the cell bodies (of which there are 60-80)
of the neurons making up the sensilla. It is invested by two layers of
cells, one of which is the epithelium of the tracheal air sac surrounding
the nerve, the other being the accessory cells of the sensillum.

Prothoracic leg tympanic organs lie in the proximal region of the
tibiae, one member on the anterior side, the other, posterior. They lie

Fig. 57. Right tympanic air sac and associated structures of a noctuid as
seen from within the metathorax. A portion of the air sac is represented
as cut away to show the sensory elements. ATP, anterior tendon plate ;
B, Biigel; DLg, second dorsolongitudinal muscle of metathorax;
EP, epaulette ; L, ligament connecting chordotonal organ with scutal
phragma; PI, PII, pockets of tympanic frame; S, acoustic sensory
cells; TAS, tympanic air sac; TM, tympanic membrane; Tr, tracheal
twig from metathoracic spiracle. (Redrawn from Treat and Roeder
1959.)

thus back to back separated by two branches of a trachea, an anterior
and a posterior (Fig. 55). Whereas in the Gryllidae the tympanic mem-
branes lie on the exposed surface of the leg, they are sunken in the
Tettigoniidae and are exposed to the outside only via a slit. The
sensilla, of which there are about seventy in each leg, lie in a long row
on the dorsal side of the anterior trachea. The dendrites and scolopoid
bodies are oriented in such a way that they are not inserted on the
tympanic membrane but instead on a membrane, presumably derived

SOUND RECEPTION 83

from the basement membrane of the hypodermis and trachea, border-
ing on the blood cavity of the leg. The cells thus stretch between the
trachea and this membrane (Fig, 55). This organ is innervated by a
purely sensory nerve, the tympanic nerve, originating from the first
thoracic ganglion. This nerve also innervates part of the subgenual
organ and the intermediate organ. The rest of the subgenual organ is
innervated by a branch of the leg nerve.

The metathoracic tympanic organ of Lepidoptera is the simplest
organ of this type. It lies in the anterior wall of a deeply recessed cavity

Fig. 58. The tympanic organ of the cicada Tibicina haematodes as seen
when the abdomen is cut transversely. H, heart; G, gut; A, auditory
cavity in which a window has been cut; T, tympanum. (Redrawn from
Vogel, 1923.)

bounded posteriorly by parts of the first abdominal segment (Fig. 56).
As in the Acrididae, the membrane is lined with hypodermal epithel-
ium plus that of the associated tracheal sac. Two chordotonal sensilla
are attached to the inner surface of the tympanic membrane. The
axons, forming part of the tympanic nerve, extend anteriorly, receive
support en route from a thin ligament, turn at an angle to receive
further support from a cuticular projection, the Bugel, then continue
on through the tracheal sac ultimately to join the pterothoracic
ganglion (Fig. 57). Like the comparable structure in the locust, the
nerve is ensheathed in tracheal epithelium.

84 THE PHYSIOLOGY OF INSECT SENSES

The tympanic membrane faces only part of the tracheal air sac.
The rest of it is faced by another membrane whose upper surface is
exposed to the air by a narrow slit which is the orifice of a large
counter-tympanic cavity. This cavity and membrane are regarded as
accessory resonating structures.

The paired tympanic organs of cicadas are located ventrally in the
region of the first and second abdominal segments. They differ from
others described chiefly with reference to the point of attachment of
the chordotonal sensilla. These, approximately 1,500 in number, occur
in a single large bundle, their proximal ends attached to one corner of

Fig. 59. Frontal section through the base of the abdomen of Cicadetta
coriaria showing the left auditory capsule. S, sense cells ; E, enveloping
cells; N, nerve; B, blood; Sp, spiracle; O, outer air; I, inner air space;
T, tympanic membrane. (Redrawn from Vogel, 1923.)

the tympanic membrane and their distal ends to the cuticle of the body
wall (Figs. 58 and 59).

The tympanic organs of Orthoptera and Lepidoptera have two so-
called tympanic muscles associated with them. The function of these
muscles is unknown, although it has been suggested from time to time
that they might be concerned with altering stresses on the tympanic
organ. In moths it is certainly true that distortion of the skeletal
elements of the tympanic organ reversibly changes the excitability of
the sensilla (Roeder and Treat, 1957).

The Sensilla

The chordotonal sensilla, as viewed with the light microscope, are all
basically the same. So far only those of the tympanic organ of the

SOUND RECEPTION 85

locust Locusta migratoria migratorioides have been examined with the
electron-microscope. The relations between the various cells compos-
ing the unit are clearly revealed (Fig. 60). The axon and cell body of
the neuron are enveloped in a Schwann cell. Each Schwann cell may
envelope several neurons. Distal to the Schwann cell, at the base of the

Fig. 60. A diagrammatic section through the sensory region of the locust
tympanic organ to show the method of attachment of the sensory cells.
H, hypodermis; E, elevated process; C, cap cell; Ci, cilium-like pro-
cess; S, scolopoid body; Sc, scolopale cell; D, dendrite; Sw, Schwann
cell; A, axon; F, fibrous sheath cell. (Redrawn from Gray, 1960.)

dendrite, lies another enveloping cell, the fibious sheath cells (prob-
ably the Bindesubstanz of Schwabe). The remaining portion of the
dendrite is enclosed in the enveloping cell (Hiillzelle of Schwabe,
scolopale cells of Gray). Between this cell and the hypodermis lies the
cap cell (Kappenzelle of Schwabe, Deckzelle of Eggers, attachment

86 THE PHYSIOLOGY OF INSECT SENSES

cell of Gray). These cells, as Schwabe (1906) observed, are in contact
with each other by finger-hke processes and interdigitate with the
hypodermal cells by folds.

The scolopoid body exclusive of its cap is a hollow tube composed
of several concentric rods. It lies within the cytoplasm of the envelop-
ing cell. The apical body or scolopale cap is an extracellular body lying

L

Fig. 61 . Details of the scolopoid region of the tympanic organ of the locust.
Cs, scolopoid-cap ; Ci, cilium-like process; Sc, scolopale cell; D,
dendrite; A, axon. (Redrawn from Gray, 1960.)

in a depression of the cap cell and fitted into a rim of the enveloping
cell containing the fused ends of the scolopale.

As the dendrite proceeds from the cell body it presents a ring-shaped
dilation just before entering the scolopoid body, a less-pronounced
dilation followed by a marked reduction in diameter after it has entered
the scolopale, and, finally, a subterminal sweUing. Its tip is inserted
into a channel in the cap (Fig. 61).

The axial fibre which can be seen by light microscopy is shown by

SOUND RECEPTION 87

electron-microscopy to be a very complicated cilium-like structure
consisting of basal rootlets fusing into a root as they ascend within the
dendrite (Fig. 62). Marked periodic cross striations are present. The
distal portion of the cilium-like structure consists of nine fibrils

Fig. 62. Diagrammatic longitudinal section through the sensillum in the
tympanic organ of the locust. Cs, scolopale cap ; Ci, cilium-like process ;
Sc, scolopale cell; C, cap cell; F, fibrous sheath cell; Sw, Schwann
cell; Sr, scolopale rod; D, dendrite. (Redrawn from Gray, 1960.)

surrounding a central core. Its resemblance to the typical motile
cilium is so striking that Gray (1960) called it a cilium. Since there is no
evidence that it is functionally a cilium, it seems inadvisable to identify
it so completely with those structures. 'Cilium-like' is a more appro-
priate designation.

Mechanism of Action

It will be recalled that vibrational disturbance in a homogeneous
medium produced both a local increase in pressure and a displacement

88 THE PHYSIOLOGY OF INSECT SENSES

of contiguous particles away from the source of disturbance. It follows
that the waves resulting from this disturbance may be detected either
by a device sensitive to pressure change or by one sensitive to displace-
ment. As Pumphrey (1940) has stated in his discussion of the physics
of sound detection, the prerequisites of a pure pressure receptor are a
massive, opaque (to sound) chamber closed by a stiff diaphragm whose
displacement is vanishingly small. A recording system coupled to the
diaphragm would measure the excursions which are proportional to
the pressure amplitude. The orientation of the instrument with respect
to the source of sound is unimportant. The auditory mechanism of
the mammal is undoubtedly a receiver of this type, as are most
commercial microphones.

A displacement receiver requires a diaphragm or moving vane
whose mass and hinges are so slight as not to offer resistance to
motion. For maximum efficiency a receiver of this type should be
oriented in such a way that the incidence of sound is normal to the
plane of the moving element.

Judging from their structure, all insect sound-detecting organs are
displacement rather than pressure receptors. There is one ingenious
experiment reported which was designed to ascertain whether or not
the sound receptors were indeed displacement receptors (Antrum,
1936). Ants {Formica rufa and Myrmica spp.) can respond to loud
artificial sounds. Antrum directed sound vertically downwards upon
a reflecting surface, and so set up standing waves. In one experiment
ants were allowed to walk upon the reflecting surface ; in another, they
walked upon gauze suspended at a critical level above the surface.
Under these conditions of a standing wave the reflecting surface was
the region of maximum pressure change and minimum displacement,
whereas at the gauze, this being an antinode, the reverse was true.
Ants responded most vigorously when walking on the gauze ; that is,
they were most sensitive to displacement. Unfortunately, as Pumphrey
(1940) pointed out, the sounds employed were extremely intense
(10^1 times human threshold at 1,000 c/s), and there is no indica-
tion as to whether the ants were responding to air-borne sound
waves or to vibrations in the substrate. It is not known what organs
were involved. As Pumphrey further pointed out, it is not necessarily
true, as Antrum maintained, that velocity was the critical factor, since
displacement, velocity, and acceleration are all maximal at an
internode.

In any event, there is at this time absolutely no conception in terms
of structure of the manner in which chordotonal sensilla operate.

SOUND RECEPTION 89

As Gray (1960) pointed out for Locust a, displacement of the
membrane is presumably transmitted to the cap cell. The cap appears
to be rigidly seated on the scolopale, which in turn seems to be firmly
locked to the dendrite. The cilium-like tip of the dendrite lies freely in
the extra cellular cavity of the scolopale and cap. What moves in
respect to what is obscure. Now that the electron-microscope is
revealing more clearly the structural relationships of these sensilla,
earlier speculations upon their mode of action based upon the partial
knowledge provided by the light microscope have only historical
value.

Function - Orthoptera

The earliest attempt to analyse thoroughly the function of tympanic
organs was that of Regen (1908, 1912, 1913, 1914 a, 1914 b, 1924, 1926).
As experimental animals he employed the cnckti Liogryllus campestris
and the long-horn grasshopper Thamnotrizon apterus. He demonstra-
ted that females would orient to a telephone transmitting the sound of
a male in another room and that two males could sing in concert.
Between the Vorspiel (prelude) and the Nachspiel (coda) of the duet,
that is, the beginning and end, where the two were in synchrony, there
was almost always an intermediate period during which the chirps of
the two alternated regularly. Furthermore, naive (newly moulted
adults) males could be induced to sing in concert with many kinds of
artificial sounds ranging in frequency from i 400 c/s to + 28,000 c/s.
All of these experiments proved that the insects were sensitive to air-
borne sound. When both tympanic organs were amputated the res-
ponses were almost completely abolished (some residual sound
perception remained), thus proving that the tibial tympanic organs
were the principal sound receptors.

These conclusions were confirmed a few years later, when Wever
and Bray (1933) recorded nerve activity in the forelegs of crickets
{Gryllus assimilis) and tettigoniids {Amblycorypha oblongifolia and
Pterophylla camellifolia) by means of electrodes inserted therein. The
frequency range for crickets was 300-8,000 c/s; for the tettigoniids,
800-45,000 c/s. The electrical response was asynchronous at all fre-
quencies. These experiments alone were not conclusive, because, by
inserting their electrodes into the leg rather than directly upon the
tympanic nerve, Wever and Bray might conceivably have been record-
ing from tactile hairs and other chordotonal organs of the leg as well as
from the tympanic organ (cf. also Wever, 1935). An attempt to
circumvent these difficulties was made with Decticus by Antrum

90 THE PHYSIOLOGY OF INSECT SENSES

(1941), who recorded from legs in which the subgenual organs pre-
sumably had been destroyed. Unfortunately, the precision of the
operation was not confirmed histologically post mortem. In any event,
the sensitivity to air-borne sound waves extended from 1,000 c/s to the
upper limits of the stimulation apparatus (10,000 c/s). The higher the
frequency, the lower the threshold. Responses at frequencies below
1,000 c/s were obtained only by employing intensities so high that it is
doubtful that they truly fell within a physiological range.

The abdominal tympanic organs of two acridids {Locusta migratoria
migratorioides and Arphia sulphured) were sensitive over the range
300-10,000 c/s (Wever, 1935; Pumphrey and Rawdon-Smith, 1936 b)

30 —

o

V

<u
c

o
o

\)

U

c

t

TO

■^
C
03

CU

o

c
o

30

60

90

300

3,000

10.000

Frequency (c/sec)

Fig. 63. a, Threshold curve for tympanic organs of Arphia sulphurea
averaged from Wever's figures, b, Threshold curve for an isolated
tympanic organ of Locusta. c, Human threshold subjectively de-
termined (from Wegel, 1932). (Redrawn from Pumphrey, 1940.)

SOUND RECEPTION 91

(Fig. 63). Again over the entire range the nerve response was asyn-
chronous. Since 10,000 c/s was the limit of the apparatus used, it would
appear that the tympanic organ is much more sensitive than the
human ear at frequencies above 10,000 c/s. In the neighbourhood of
3,000 c/s, according to one report of Pumphrcy and Rawdon-Smith
(1936 c), the tympanic organ ofLocusta exhibits maximum sensitivity.
At this point the frequency is only 20 db above the ear of man at its
maximal sensitivity. On the other hand, in an earlier paper (Pumphrey
and Rawdon-Smith, 1936 b) it was indicated that the threshold con-
tinues to fall with increasing stimulus frequency (see also Pumphrey,
1940). A continuous fall was observed by Antrum (1941).

In any case, the tympanic organ is insensitive in the frequency range
that is optimal for man and extends into the ultrasonic range. At its
optimum the organ responds to a stimulus of about 7 x 10"^^
ergs/sec, about the same as for man, and thus is operating at the
physical hmit. In contrast to the human ear, the tympanic organ of
Orthoptera does not fatigue readily. No diminution in the response to
a 2,000-c/s tone could be detected after continuous stimulation lasting
one-half a minute (Pumphrey and Rawdon-Smith, 1936 c).

Although the tympanic organs are not the only receptors possessed
by Orthoptera (there are in addition hair sensilla and chordotonal
organs not associated with tympana), it is clear that they are the
principal, if not sole, receptors involved in detecting noises produced
by stridulation of the same and other species. Since Orthoptera can
discriminate among the various kinds of calls and can distinguish
between some artificial and genuine calls (Regen, 1926), it is obvious
that the tympanic organ does have some specialized characteristics.

Songs of Acrididae are, on the whole, broad-band noises at
2-12,000 c/s with maxima at 4,000-8,000 c/s and intensities of
30-40 db, re 0-0002 dynes/sq. cm., at 10-30 cm. (Busnel, 1953 ; Haskell,
1955; Loher and Broughton, 1955). The Tettigoniidae also produce
wide-band noises at frequencies up to 100,000 c/s (maxima 8,000-
15,000 c/s) and intensities of 46-70 db (Horror, 1954; Busnel, 1953;
Busnel and Chavasse, 1950; Pasquinelly and Busnel, 1955; Pielemeier,
1946 a, 1946 b; Pierce, 1948). Gryllidae produce relatively pure tones
at 2,000-6,000 c/s and intensities of 40-60 db (Alexander, 1957;
Busnel, 1955 a, 1955 b; Haskell, 1955; Pasquinelly and Busnel, 1955;
Pierce, 1948).

The songs can be described physically by five parameters: fre-
quency, intensity, wave form, phase, and the temporal distribution
of sound units (Frings and Frings, 1958). Successive units may be

92 THE PHYSIOLOGY OF INSECT SENSES

combined to form more complex arrangements, and this in turn
further multipHed.

The tympanic organ is sensitive to the entire frequency range of
sounds produced by stridulation, but regardless of the frequency, the
pattern of discharge in the tympanic nerve is asynchronous. From this
it has been concluded that the organ is incapable of frequency discrim-
ination, and this may indeed be so. On the other hand, there have been
no recordings from single fibres in Orthoptera, and this should be done
to clinch the argument, even though Pumphrey (1940) has argued on
structural grounds that harmonic analysis, that is, different fibres
responding to different stimulus frequencies, is impossible. The fact
that no frequency discrimination has been observed in the moth
tympanic organ, where the presence of only two fibres permits a precise
analysis (Roeder and Treat, 1957), adds weight to the argument. It
also supports the idea that there is no frequency pattern in the nerve
mirroring the frequency pattern of the stimulus. On the other hand,
it is possible that the insect itself may be able to analyse frequency to
some degree by means of central interpretation of input from a
number of organs each having a different frequency range (Katsuki
and Suga, 1960; Suga and Katsuki, 1961).

Pumphrey and Rawdon-Smith (1939) proposed that the tympanic
organ, being sensitive to intensity, detected changes in intensity, that
is, amplitude of the sound wave, and that the different sounds of
insects differed most importantly by the periodic changes (modulation)
in intensity of the sound. In other words, any sound wave of a fre-
quency within the detectable frequency range was merely a carrier
wave upon which was imposed a pattern of amplitude modulation.
By contrast, the human ear is very sensitive to frequency changes in
the carrier wave but relatively insensitive to changes of the modulation
frequency. This hypothesis was supported by records showing that an
organ stimulated by an amplitude-modulated sound discharged
volleys of impulses at the modulation frequency of the stimulating
sound. These results were confirmed in a number of species by Haskell
(1956 b), and as he pointed out, it would be reasonable to assume that
the different pulse rates found in the different songs of various species
would be the parameter signalled by the receptors, and hence the key
to the recognition of songs. Busnel (1955) and his co-workers (Busnel,
Loher, and Pasquinelly, 1954), on the other hand, have suggested as a
result of experiments in which natural and artificial songs were played
to species of Chorthippus that the essential element in songs that aids
discrimination is 'transients' (sudden changes in intensity). To test the

SOUND RECEPTION 93

latter hypothesis Haskell (1956 b) played the identical stimulus again
and again to a tympanic organ and recorded the pattern of discharge
from the tympanic nerve. He found that the resultant volleys of spikes,
although synchronous with the pulse frequency of the stimulus, were
variable in themselves and that this variability bore no simple relation-
ship to any quality of the stimulus. On the basis of this evidence he

Fig. 64. Sensitivity of an isolated tympanic organ of Locusta plotted on
polar co-ordinates as a function of direction of incidence of the test
stimulus. Sensitivity (log reciprocal of threshold amplitude) is plotted
radially and the minimum sensitivity is arbitrarily taken to be zero.
The line 0-180 degrees lies in the sagittal plane of the animal, and
for angles of positive or negative sign the test stimulus is incident on
the external or internal aspect of the tympanic organ respectively.
(Redrawn from Pumphrey, 1940.)

concluded that the characteristic of stridulation that permits inter-
specific recognition is the pulse repetition of the songs. Many ques-
tions about sound recognition remain to be answered, however, and
Haskell's (1956 b) discussion should be consulted.

One feature of the early field experiments that turned up repeatedly
was the ability of the insects to locahze sound. Regen (1924), for
example, showed that unmated females could orient from a distance

94 THE PHYSIOLOGY OF INSECT SENSES

of approximately 10 metres to a chirping male. A female's path of
approach was nearly a straight line. Even with one tympanic organ
destroyed, orientation was possible, although the line of approach
now became more devious. These observations suggesting that the
tympanic organ has directional characteristics are supported by
experiments in which sensitivity was tested to a stimulus given
successively at different directions of incidence (Pumphrey, 1940). A
plot on polar co-ordinates of the sensitivity of an isolated tympanum
o^ Locust a as a function of the direction of incidence of the stimulus
illustrates the pattern of the directional characteristics (Fig. 64). This
finding is in agreement with the assumption that the tympanic organ
is a displacement rather than a pressure receiver (also Autrum, 1955 c)
Also, as these experiments revealed, the tympanic organ responds
equally well to 'push' and to 'pull'.

Function - Lepidoptera

Moths and butterflies also are known to be sensitive to sound (Stobbe,
1911 ; Turner, 1914; Turner and Schwarz, 1914; Eggers, 1925, 1926 a,
1928; Frings and Frings, 1956, 1957). Many of the sounds employed
for testing in early experiments were non-specific (e.g., squeaks of a
glass stopper twisted in a bottle), and the responses, too, were not
always specific (e.g., partial erection of the antennae, folding of the
wings, flexing of the antennae, running, or flying). It was difficult to
know what kind of sounds to employ as test stimuli for the simple
reason that no one knew what sounds moths responded to in nature.
One of the earliest speculations regarding the normal role of sound
perception in behaviour was the suggestion of White (1877) to the
effect that auditory receptors, in moths specifically, might assist in
detecting the approach of insectivorous bats. This idea gained plaus-
ibility from a number of field and laboratory observations (Schaller
andTimm, 1950; Webb, 1953; Treat, 1955), and was finally established
on a firm experimental basis by Roeder and Treat (1957, 1961 a,
1961 b).

The principal sound receiver is the tympanic organ. The sensitivity
of this organ in noctuids to air-borne sounds was demonstrated by
Eggers (1925, 1926 a, 1928), first confirmed by the behavioural studies
of Schaller and Timm (1950) and of Treat (1955), and ultimately by
electrophysiological techniques. Other sound receptors must exist
because, as is true with Orthoptera, there is still a residual response to
intense sound after destruction of both tympanic organs (Stobbe,
1911; Eggers, 1925; Treat, 1955; Frings and Frings, 1957). Further-

SOUND RECEPTION 95

more, species in which tympanic organs are not known to occur on
the body are somewhat sensitive to sound. For example, the satyrid
butterfly Cercyonis pegala responds to air-borne sounds of high
intensity (100-1 10 db). It should be remembered, however, that there
exist on the ventral surface of the basis of the wings of this insect
chordotonal sensilla associated with minute tympanic membranes
(Vogel, 1912), and some people have tendered the suggestion that
sound reception by these organs is possible (Vogel, 1912; Pringle,
1957).

The tympanic organ is sensitive to very high sound frequencies. Res-
ponses in the frequency range 3,000-20,000 c/s have been recorded
from the tympanic nerves of the moths Phalera bucephala (Notodon-
tidae) and Arctia caja (Arctiidae) (Haskell and Belton, 1956). The very
thorough investigations of Roeder and Treat (1957, 1959, 1961a,
1961b) with a number of noctuids have shown that the sensitivity
extends up to stimulus frequencies of 100,000 c/s. These investigations
have also revealed many of the characteristics of these organs. A rest-
ing discharge occurs at the ordinary noise level of the laboratory. It

A

t '^^^n^ l t SlWlVl^ BeH ^l*»

lAAAAAAAAAAAAAAAAAAAAAAAAAAA

Fig. 65. Tympanic nerve activity in the noctuid Prodenia eridania. A,
activity in the tympanic nerve at laboratory noise level. The large spike
belongs to the non-acoustic unit. B, response in one acoustic unit to
continuous sound of 70 kc/s. C, response in both acoustic units to
30 kc/s. Time marks, 100 c/s. (Courtesy of Roeder and Treat.)

C/5 <U k>

O ^ ^

60

C

O
o

u

\

w

':

1

I

':

:

'■_[

';

■

':':

':

,

'■:

'

j

,"

■':

;

;

:■

;

'\\

;

-O -

:':

:

\

SOUND RECEPTION 97

takes the form of a series of discharges of a large spike at a steady
frequency of 5-20 per second. Haskell and Belton (1 956) believed these
to be motor discharges, but in Rocder's and Treat's preparation they
were clearly afferent. In addition, two species of small spikes discharge
randomly, in all probability, as a result of low-level random noise in the
laboratory. These two spikes represent the acoustic units of the
tympanic organ; the large spike is a non-acoustic unit whose function
is not clearly understood (Fig. 65).

Both acoustic units respond to sound by a high-frequency discharge
which shows no change with stimulus frequency. The two differ only
in their threshold. For both, the greatest sensitivity lies between
15,000 and 60,000 c/s. In response to a click, with many transients, or
a short pulse (0-02-1 msec.) of pure sound at threshold, the more
sensitive fibre responds after a latency of 3-5 msec, with one spike
(Fig. 66). When the stimulus intensity is increased, the latency de-
creases and there is an after-discharge of two-three spikes at a fre-
quency of 700 per second. A further increase in stimulus intensity
brings in the second acoustic unit. The variety of overlap observed
indicates that there is no regular alternation of activity or coupling of
the units. Although one cannot state categorically that only two units
are firing, this appears to be the case. It fits well with the histological
facts.

Contrary to the situation found in orthopteran tympana, these
organs adapt rapidly. From an initial frequency of about 1,000 spikes
per second the discharge drops to 50 per cent of this value in the first
0- 1 second and to 25 per cent by the end of the second.

Various operations on the tympanic organ prove that the effects
observed electrophysiologically are truly indicative of activity in the
two chordotonal sensilla. Touching the sensilla or disengaging them
from the tympanic membrane terminates their activity. On the other
hand, tearing the membrane or opening the air sac behind it fails to
alter the response significantly. Occluding the external tympanic recess
narrows the band of frequency reception to its middle range and raises
the threshold.

These experiments are in agreement with the idea that this organ is a
displacement-sensitive receiver. They show that the tympanum must
be free to vibrate but that uniformly distributed elastic tension is not
necessary. The direction of the displacement, in so far as the sensillum
is concerned, is apparently not critical, since in many noctuids the axis
of the sensillum is not normal to the plane of the tympanic membrane ;
instead, it forms an acute angle with it.

98 THE PHYSIOLOGY OF INSECT SENSES

The function of the non-acoustic unit may be proprioceptive. If the
acoustic sensilla are rendered inoperative resting activity may be re-
corded from the non-acoustic unit alone, but only so long as the
attachment to that structure known as the Bugel is intact (Fig. 57). In
this region is located a single Type II neuron (Treat and Roeder, 1959).
The rate of discharge of this unit changes as various portions of the
tympanic frame are distorted, as they would be during flight or as a
result of activity of the two tympanic muscles (Fig. 56).

The fact that the moth tympanic organ exhibits after-discharge, a
form of physiological amplification for stimuli of short duration, and
rapid adaptation fits it very well for the reception of short, rapid bursts
of sound of the sort given off by hunting bats. Furthermore, its maxi-

FiG. 67. Polar plot of the distances at which a 1-msec. click of fixed intensity
elicits the same response from the right tympanic organ when placed
at various angles relative to the median axis of the moth (0-1 80 degrees).
The moth was headed towards degree and inclined upward at about
30 degrees to the plane of measurements. Distances in metres from the
moth are indicated on the 90-270-degree line. Open circles and solid
line, Acronycta; broken line, Gmphip/iora; solid circles and solid line,
Lucania. (Redrawn from Roeder and Treat, 1961.)

PLATE I. The cry of a Hying bat (Myotis) recorded by a Granith microphone
(upper trace) and the acoustic cells of the noctuid moth AK'ropcrina
diibitans (lower trace). Time mari<, 10 msec. Made in collaboration
with Dr Fred Webster. (Courtesy of Roeder and Treat.)

PLATE II. Triply innervated chemosensory hair of Phormia (pig-
ment layer omitted). HY, hypodermis; DF, distal fibres
(dendrites); PF, proximal fibres (axons); N, neuron cell
bodies; TR, trichogen; TO, tormogen; d, thin-walled
cavity; C2, thick-walled cavity; VA, vacuole; TA, tracheole;
SP, sensory papilla. (From Dethier, 1955.)

m

PLATE III. Electron-micrograph of the
tip of a tarsal contact chemo-
receptor of Stomoxys calcitrans
(X 20,000). (Courtesy of J. R.
Adams.)

PLATE IV. Photomicrographs of images of newsprint formed by the corneal
lens of ocelli of the caterpillar Isia Isabella, ex is seen through a unitary
type cornea; x, through a tripartite type. (From Dethier, 1942.)

SOUND RECEPTION 99

mum sensitivity is in the range ( 1 5,000-60,000 c/s) corresponding to the
predominant frequencies in the cries of flying bats of the family
VespertiHonidae (Griffin, 1950, 1953). Actual tests in which electrical
recordings were made from tympanic organs in the presence of bat
cries prove beyond all doubt that the organs could be used for detect-
ing these predators (Roeder and Treat, 1957) (PI. 1). Behavioural
studies (Roeder and Treat, 1961 a, 1961 b) have shown that free-flying
moths take evasive action when they are stimulated by bat cries. This
is random action, at least when the bat is near, in that it bears no par-
ticular relation to the flight path of the bat.

Like the tympanic organ of Orthoptera, the organs of moths are
directional receivers (Roeder and Treat, 1961 a). Localization is
theoretically possible with one, but is obviously more efficient with
two. At low sound intensities, that is, when the bat is not too near,
there is a differential response from the two organs, but as the intensity
increases, when the bat would be close, the differential nature of the
binaural response disappears (Roeder and Treat, 1961b) (Fig. 67).

Another role of the tympanic organ may be concerned with navi-
gation during flight. It has been suggested (Hinton, 1955) that some
moths may orient themselves by echo-location. Although no experi-
mental proof exists, there is direct evidence that moths can detect the
sounds of a flying moth and presumably those of their own flying
(Roeder and Treat, 1957). This evidence suggests that moths might be
able to detect the reflection of their own flight sounds from nearby
objects, and hence to echo-locate. For this to succeed, however, numer-
ous obstacles would have to be overcome, not the least of which would
be the ability to distinguish between direct and reflected flight sounds.

Function - Hemiptera

Little is known about the tympanic organs of this Order. Males of the
waterboatman, Corixa, stridulate. Females respond to the chirping,
and other males chirp in chorus as long as the tympanic organs are
intact (Schaller, 1951). These are the only insects for which a sensitivity
to water-borne sound has been demonstrated.

HAIRS AS SOUND RECEPTORS

Many insects that lack tympanic organs are sensitive to sound waves in
air. As early as 1779, Bonnet recorded that caterpillars respond to
sound by thrashing about of the anterior portion of the body. Since
that time many naturalists have recorded this phenomenon, a response
brought about by convulsive contractions of the longitudinal muscles

100 THE PHYSIOLOGY OF INSECT SENSES

in the anterior trunk. Some species respond by a complete cessation of
movement, a 'freezing'; some respond by 'freezing* followed by
thrashing. The response probably serves as a protective device. It re-
mained to Minnich (1925, 1936, 1937) to attempt a physiological
investigation of the response. He studied seven species of butterfly
larvae and eight species of moth larvae, bringing to a total of twenty-
eight the number of lepidopterous larvae known up to that time to be
sensitive to sound. It is likely that all caterpillars detect sound. The
range of frequency sensitivity extends from 32 to 1,000 c/s. By cutting
larvae into fragments and noting that the fragments responded,
Minnich showed that the receptors for sound were diffusely distributed
throughout the body.

Early experiments were confined to relatively hairy species, and in
these the responses to sound were abolished when the hairs were
loaded with flour dust or droplets of water. Upon removal of the dust
or water, responsiveness returned. A steady stream of air directed
against the hairs inhibited response. These experiments support
strongly the idea that the hairs are the sound receptors. On the other
hand, similar behavioural responses to sound were obtained with
'hairless' species. Although even these species possess minute setae,
Minnich (1936) entertained doubts that sound reception here was in-
deed mediated by hairs. The question is still open.

In the belief that the sensory hairs of the hirsute caterpillars were
resonant structures, that is, that different lengths and sizes would reso-
nate to different frequencies of stimulus, Minnich (1936) attempted to
fatigue the response to one frequency and test for a response to
another. He found, in fact, that fatiguing at low frequency inhibited
responses to all higher frequencies, but that the reverse was not true.
As Pumphrey (1940) pointed out, these results are inconsistent with a
hypothesis of reaction by resonance. He pointed out further that the
results obtained in the experiments on fatiguing are explicable on the
grounds that low-frequency stimuU will be much louder to the animal
than those of equal intensity but higher frequency and would therefore
probably have a greater fatiguing effect. This interpretation is based on
a study of action potentials recorded from cereal hairs of cricket and
cockroaches (Pumphrey and Rawdon-Smith, 1936 a, 1936 b, 1936 c).

The cerci of cockroaches and Gryllus and many other Orthoptera
are clothed with extremely delicate, long, lightly hinged hairs (Sihler,
1924). They are typical sensilla trichodea. In Periplaneta americana
there are several hundred on each cercus. Each hair is about 0-5 mm.
long and i 0-005 mm. in diameter. They move visibly to light puff's of

SOUND RECEPTION 101

air and to sounds of adequate intensity over a wide frequency range;
that is, the response is not resonance. As Pumphrey and Rawdon-
Smith(1936a, 1936b, 1936c) showed, these hairs are pure displacement
receptors sensitive up to about 3,000 c/s. Threshold measurements over

Frequency Q^/sec)

Fig. 68. Experimentally determined thresholds of the cercus of Gryllus
for pure tones at various frequencies (solid circles). The heavy line
represents a constant displacement amplitude of 560 A. The light line
represents the human threshold of hearing from Wegel's data.
(Redrawn from Pumphrey, 1940.)

the frequency range 50 c/s to + 1 ,000 c/s (Fig. 69) reveal that the res-
ponse is to a constant displacement amplitude of 560 A. (Pumphrey,
1940). If the hair acts as a rigid lever this finding suggests that
the threshold displacement of the dendrite at the base of the hair
cannot greatly exceed 0-5 A. Pumphrey (1940) suggested that the

102 THE PHYSIOLOGY OF INSECT SENSES

'spontaneous' activity recorded from nerves in silent surroundings
may represent random movement of the dendrites by Brownian
agitation of the molecules in their vicinity. Conditions are similar for
the cricket and the cockroach.

In contrast to the tympanic organ, the cereal hairs exhibit a number
of characteristics seen in recordings from the mammalian cochlear
nerve. One of these is synchronization. At low stimulus frequencies
the action potentials in the nerve are exactly synchronized with the
stimulus. At the low end of the frequency range the nerve may be
synchronized or show what Pumphrey and Rawdon-Smith (1936 a,
1936 b, 1936 c) term 'frequency doubling'. At high frequencies it may
show frequency halving or quartering. Thus, at a stimulus frequency
of 400 c/s the nerve may discharge at 400 c/s or at 800 c/s, and at a
stimulus frequency of 400 c/s, fire at 300 or 150 c/s. In the 400 c/s
range of stimulation the nerve that is firing in synchrony may begin
halving if the intensity of the stimulus is reduced by about 10 db.
Pumphrey and Rawdon-Smith (1936 c) suggested that these phe-
nomena may exemplify the condition known as 'alternation'. In the
mammalian Vlllth nerve the synchrony in the whole nerve at high
stimulus frequencies is believed to result from each fibre firing to
alternate sound waves and the different fibre groups being 180 degrees
out of phase. In the cereal nerve the alteration is diff'erent in that all
fibres fire to every other sound wave but all are in phase, that is, the
whole nerve alternates, hence, the response frequency is halved or
quartered. It was suggested, however, that true alternation of the
mammalian type also occurs, at least during the first fraction of a
second of response, when the nerve momentarily fires at 600-800 c/s.
At higher stimulus frequencies the nerve responds asynchronously.

Another resemblance to the mammahan Vlllth nerve is the occur-
rence of equilibration, that is, a decline with time in the amplitude of
massed spikes in the nerve. This phenomenon is interpreted as a
reduction in the number of impulses at each sound wave with time as
the result of the lengthening of the relative refractory period of each
fibre.

Electrophysiological studies of the cereal sensilla of a number of
Acrididae are in general agreement with the results just described;
however, no clear evidence of equilibration or frequency halving or
doubling was obtained (Haskell, 1956 a, 1956 b).

Hairs on other parts of the body, especially of Orthoptera, are
believed also to be receptors sensitive to air-borne sound waves, but
the evidence is not always convincing, because it is derived primarily

SOUND RECEPTION 103

from electrophysiological records of activity in segmental nerves
known to carry fibres from other types of sensilla, particularly chordo-
tonal sensilla. Pumphrey and Rawdon-Smith (1936 c), studying the
locust, believed that the responses originated in trichoid sensilla. Later
Pumphrey (1940) suggested that the segmental chordotonal organs
were perhaps involved (see also Hughes, 1952). Working with a num-
ber of acridid species, Haskell (1956 a) inclined to the view that res-
ponses did indeed emanate from hairs because smearing the abdomen
with vaseHne abolished the response. In any event, the receptors
involved fired asynchronously at all stimulus frequencies and required
high stimulus intensities (+ 7 dynes/sq. cm.). Even after smearing with
vaseline, however, a residual sensitivity to sound occurred, and
Haskell (1956 a) suggested that in this instance the chordotonal organs
might be involved. Other sensilla on the abdomen of grasshoppers,
presumed to be sensitive to vibrations of the substrate, will be dis-
cussed later.

Little is known about the functions of hair sensilla sensitive to
sound. It is unlikely that those in Orthoptera are concerned with
reception of noises produced by stridulation (Haskell, 1952 a). In the
cockroach stimulation of cereal hairs with a puff of air evokes the
'evasion response'. This kind of response is not well developed in
locusts. In Orthoptera in general mechanical stimulation of the cerci
seems to be of great importance during copulation, and perhaps this
is the principal role of cereal hairs.

Hairs most highly developed for the reception of air-borne sounds
differ in performance from tympanic organs in several respects : they
respond over much lower frequency ranges; they respond synchro-
nously with stimulus frequency over certain ranges and thus do
exhibit a limited frequency discrimination; they fatigue fairly rapidly;
they tend to equilibrate. Like tympanic organs, however, they are
displacement receivers.

THE JOHNSTON'S ORGAN OF CULICIDAE

In the males of Culicidae and Chironomidae the Johnston's organ
attains an extraordinary complexity manifested by such an enormous
multiplication of sensilla that the pedicel is bulbous and a special
internal cuticular framework has developed to serve as points of
attachment for the many sensilla. This exceptional organ is demons-
trably able to be stimulated by air-borne sound waves as Johnston
(1855) and Child (1894) originally surmised.
When Mayer (1874) demonstrated that the hair (or fibrillae)

H

104 THE PHYSIOLOGY OF INSECT SENSES

comprising the whorls on the antennae of male Culex vibrated in the
presence of a vibrating tuning fork he did not indeed prove, as many
authors have pointed out (e.g., Fulton, 1928), that the antennae are
auditory organs. Nor did his experiments impHcate the Johnston's
organ. Furthermore, since the different forks employed gave different
intensities, and hence stimulus energy could not be controlled, the fact
that the hairs vibrated most extensively at 512 c/s is not especially
significant. In an experiment with Culex pipiens pallens in which
intensity was controlled, Yagi and Taguti (1941) found that the hairs
which vibrated in the stimulus frequency range of 193-870 c/s vibrated
most widely at 217 c/s. As Mayer had shown, maximum vibration
occurs when the sound wave advances at right angles to the longi-
tudinal axis of the hairs.

The literature is replete with accounts of mosquitoes reacting to
sound. In a most extensive study Roth (1948) found that males of
Aedes aegypti respond to the sound of a flying female and to artificial
sounds of comparable frequencies by orienting to the source of sound
and displaying the mating response, which consists of seizing and
clasping. To sounds of other frequencies they: rub the antennae with
the forelegs, rub the hind legs together, rub the abdomen or wings with
the hind legs, jerk the body, suddenly fly, or suddenly become
immobile.

The frequencies which induce the mating response are different
for males of different ages. The effective range extends from about
100 to about 800 c/s, and the spectrum is narrower for non- virgin
males. For virgin and non-virgin alike the effective band widens with
age. This shift to higher frequencies has been correlated with the ex-
tension of the antennal fibrillae. When males first emerge, the hairs or
fibrillae He recumbent along the shaft of the flagellum. By 48 hours
they are fully extended. This correlation suggests either that the apical
fibrillae vibrate at higher stimulus frequencies than the lower or that
a larger total of fibrillae must be set in motion at higher stimulus
frequencies in order to mediate a response.

Males can be adapted to one frequency (as judged behaviourally)
and still retain responsiveness to other frequencies (Roth, 1948). Since
intensities were not controlled in these experiments, the significance is
in doubt. With minor deviations the same story holds for Anopheles
quadrimaculatus, Culex pipiens, and Psorophora confinnis.

The correlation between the degree of extension of antennal fibrillae
and the extension of responses to higher stimulus frequency suggests
that the fibrillae are concerned with sound reception ; however, they

SOUND RECEPTION 105

are not innervated in mosquitoes. Early workers postulated that the
vibrating fibrillae acted as mechanical amplifiers which set the flagellar
shaft in vibration. The movement of the shaft with respect to the
pedicel was presumed to be detected by Johnston's organ. In an
extensive series of experiments with Aedes aegypti Roth (1948) ob-
tained results that confirmed this hypothesis. The mating response of
males to sound is abolished by completely removing the flagellum,
joining the flagellum rigidly to the pedicel with shellac, or weighting
the tips of the antennae with large beads of shellac. In all cases res-
ponsiveness returns when the shellac is removed. When the fibrillae
are removed from the antennae greater intensities of sound are re-
quired in order to elicit responses.

Although no electrophysiological studies of the Johnston's organ
of mosquitoes have yet been reported, these experiments of Roth,
taken in conjunction with Burkhardt's electrophysiological analysis of
the Johnston's organ of Calliphora, establish beyond reasonable doubt
that this organ in Culicidae is indeed adapted to respond to air-borne
sound waves.

PERCEPTION OF VIBRATIONS IN SOLIDS

It had been suspected for many decades that insects were sensitive to
vibrations in solids because they gave clear behavioural responses
when the substrate upon which they were standing or walking was
subjected to this form of motion. No very meaningful experiments
were performed, however, prior to the introduction of electro-
physiological methods of analysis. By recording from leg and seg-
mental nerves it has now been possible to demonstrate that there are
rather specific responses by some insects to vibration. A few clues as
to the identity of the sensilla involved have been obtained.

The two types of sensilla suited to the task are trichoid sensilla and
chordotonal sensilla. It is to be expected that those parts of the body
immediately in contact with the substrate, that is, the legs and abdo-
men, would be especially equipped with the requisite organs ; however,
aside from a few experiments showing that in Locusta there are hairs on
the sternites which mediate responses to vibrations of the substrate
(Haskell, 1956 a), there is no extensive work on areas of the body other
than the legs.

The legs certainly possess elaborate sensory equipment. In addition
to trichoid sensilla they invariably possess numbers of chordotonal
sensilla (Eggers, 1928; Debaisieux, 1935, 1938). Generally the legs
contain chordotonal organs at four locations ; the femur, the proximal

106 THE PHYSIOLOGY OF INSECT SENSES

tibia (the subgenual organs), the distal tibia, the tarsus-pretarsus. The
degree to which the organs are developed varies from Order to Order
and species to species. The number of sensilla in a group varies from
one to two hundred or three hundred. Each group, with the exception
of the subgenual organ, is associated with an articulation of the leg. In
the subgenual region there may be one or two concentrations of sen-
silla or none at all. The true subgenual organ lies immediately distal to
the femur-tibia articulation. When a second group of sensilla occurs,
it lies distal to the true subgenual organ. True subgenual organs have
not been found in the thysanuran Machilis, the beetle Rliagonycha, or
in Diptera, Heteroptera, Homoptera, and Neuropterpodea. Even the
more distal organ is absent in Machilis, Rhagonycha, and Diptera
(Debaisieux, 1938).

Subgenual organs are unusually diverse in structure, shape, and
numbers of component sensilla. They are most highly developed in
Orthoptera, Hymenoptera, and Lepidoptera and take the form of
cones, or sails, or fans. They are unique in that they more or less
completely occlude the dorsal blood sinus of the leg (Figs. 47, 56). At
the same time they lie in close proximity to the main tracheal trunk.
They are innervated by a branch of the leg nerve or, in the case of
insects possessing a tibial tympanic organ, by a branch of the tympanic
nerve and a branch of the leg nerve combined. They are supported
proximally by their nerves and distally by accessory cells which form a
ligament attached most often to the cuticula of the leg. Standard
chordotonal sensilla comprise the sensory elements. In number they
range from ten to forty. Extensive descriptions of these organs have
been given by Schwabe (1906), Schon (1911), Eggers (1928), and
Debaisieux (1935, 1938).

It has not been easy to ascertain just which of the many sense organs
in the legs mediate responses to vibrations of the substrate. The most
extensive experiments have been those of Autrum (1942, 1943),
Antrum and Schneider (1948), and Schneider (1950). Data are based
upon electrophysiological recordings of total activity in leg nerves
while the preparation was 'standing' on a metal plate which was set into
sinusoidal oscillations of measurable amplitude (Autrum, 1941).
Tests were conducted with Orthoptera, Lepidoptera, Hymenoptera,
and Coleoptera. In many, the prothoracic legs are the least sensitive;
the mesothoracic, the most.

All of the insects tested fall roughly into two groups as regards
sensitivity. In the less-sensitive group are the beetles, Carabus, Silpha,
Pterostrichus, and the Hymenoptera, Vespa and Bombus, The more

SOUND RECEPTION

07

100,000

10.000 —

1000

100

I-

001 —

0001

50 100 200 400 800 1.000 2,000 4000 8,000

Cycles per Second

Fig. 69. Vibration thresholds in my. (ordinate) in relation to frequency
(abscissa) for insects representing three levels of sensitivity. (Redrawn
from Antrum and Schneider, 1948.)

sensitive forms include the beetles Geotrupes and Melolontha, the
butterfly Pyrameis, the cockroach Periplaneta, and the cricket
Liogryllus (Autrum and Schneider, 1948; Schneider, 1950). In the
insensitive group there are responses to stimulus frequencies ranging

108 THE PHYSIOLOGY OF INSECT SENSES

from 50 to about 1,000 c/s; in the sensitive group, up to 8,000 c/s. In
both groups the threshold drops with increasing stimulus frequency;
however, in the sensitive group it shows a minimum in the 1,000 c/s
band of the frequency spectrum (Fig. 69). The rise in threshold at
higher frequencies following the minimum may be an artefact
(Antrum and Schneider, 1948; Schneider, 1950). The smallest meas-
ured amplitude that causes any response (middle leg of Decticus at a
stimulus frequency of 2,000 c/s) is 0'36A, a distance equal to about
one-third the diameter of the hydrogen atom. From this it follows
that the sense organs involved are developed to the physical limit, but
the minimum energy required is greater than the energy content of the
unitary elemental process (i.e., collision of one molecule) of which the
stimulus consists (Antrum, 1943).

Threshold can be recorded either as amplitude or in terms of
acceleration. The acceleration b which a substrate vibrating sinusoid-
ally imparts to an object lying on it is given by the expression —a-cxi^
sin out (maximum aoj ^) in which a is the amplitude and w (= lirv) is the
angular frequency. A comparison of the thresholds as accelerations
of the two groups of insects, sensitive and insensitive, shows that a
different range of acceleration is necessary for each group. For the
sensitive group it ranges from 2- 10~^* to \0~^g{g = 981 cm. sec. 2); for
the insensitive group, from S-IO"^ to g). On the other hand, for a
given insect threshold acceleration is more nearly constant over the
range of stimulating frequencies than is threshold amplitude. For
this and other reasons Antrum and Schneider (1948) and Schneider
(1950) felt that acceleration was the crucial physical parameter to
which the sense organs respond and that regardless of the stimulus
frequency the organs must undergo equally large accelerations.

There is only indirect evidence as to which of the several organs in
the leg are actually responding to the vibrations. Of the species
studied by Antrum and Schneider (1948), the least sensitive of those in
the insensitive group lack true subgenual organs, while the sensitive
species possess them. When an operation designed to destroy sub-
genual organs was performed on those species possessing them the
threshold of response to vibration rose and the effective frequency
band became narrow so that these sensitive insects now resembled
insensitive ones (Antrum, 1941).

From these experiments it was concluded that the sensitive insects
possessed an organ especially adapted to detect vibrations in the
substrate and that this organ was the subgenual collection of chordo-
tonal sensilla. The insensitive insects, on the other hand, were pre-

SOUND RECEPTION 109

sumed to be reacting primarily to stimuli outside the normal biological
range and doing so through the agency of organs not finely adapted for
this function. Coating the legs with paraffin did not abolish the res-
ponse so it is unlikely that hairs or spines mediate the response. As
already pointed out, other chordotonal organs in the leg are associated
with articulations. The articulations are also equipped with hair plates.
Either of these groups could be concerned. Whichever types they are,
it is probably the tarsal ones that are important in perception, since
response to vibration is lost following amputation of the tarsi (Antrum
and Schneider, 1948).

In flies (CalUphora and Eristalis), which are among the least sensi-
tive insects, there is no subgenual organ of any sort. Responses are
probably mediated through a tibial chordotonal organ which con-
sists of three sensilla (Schneider, 1950). The upper frequency limit of
this organ is about 100/cs. At frequencies up to this, the discharge in
the nerve is synchronized with the stimulus. At higher frequencies it
is asynchronous.

Antrum and Schneider (1948) decided that sensitivity is connected
with the anatomical structure of the subgenual organs. Where these
are fan-shaped, thresholds are low (e.g., Orthoptera); where they are
conical and compactly constructed (e.g., Hymenoptera), thresholds
are higher. On the other hand, the beetles Geotrupes and Melolontha,
which lack true subgenual organs, are very sensitive to frequencies be-
tween 600 and 1,000 c/s and 600 and 1,500 c/s respectively and still are
responsive to 4,000 and 8,000 c/s respectively (Schneider, 1950).
Recordings obtained following successive amputations of lengths of
leg suggest that the receptors are the ten chordotonal sensilla in the
distal end of the tibia. In any case, sensitivity does not depend upon
the number of sensilla. Camponotus sp? with a subgenual organ
containing about twenty sensilla is as sensitive as Apis with about
forty.

The curious anatomical position of the subgenual organ with res-
pect to the blood sinus of the leg invites speculation about the func-
tional association of this organ with the blood. Work with a model has
shown that vibrations would establish vortices in the blood such that
a rather constant pressure would be applied to the outer surface of the
subgenual organ (Antrum, 1942). It has been postulated that the effect
is a rectifying one and that it serves to permit the nerve to transmit
frequencies higher than the 200-500 c/s that represent the top limit
permitted by the refractory period of nerves (Antrum, 1942; Antrum
and Schneider, 1948). While it is possible that vibrations act indirectly

110 THE PHYSIOLOGY OF INSECT SENSES

by means of disturbances in the blood, the argument that it is necessary
in order to permit the organ to respond to higher stimulus frequencies
is not valid. For example, alternation of firing in the various fibres, as
in the cereal hairs, would accomplish the same purpose.

5 10

100

300 400

4000 6000 8.000

600 800 1000 1.500 2,000 2.500

Cycles per Second

Fig. 70. Comparisons of the vibration thresholds of flies (O

men (O O), bullfinches (• •), bees (• • )

cockroaches (C €). (Redrawn from Schneider, 1950.)

•O),
and

Those who would consider sensitivity to vibrations in solids a
separate sense, compare the abilities of man, birds, and insects (the
only animals for which threshold values are known) in this respect and
find insects to be the most as well as the least sensitive. The cockroach
at its optimum range is more sensitive by a factor of 10,000 than the

SOUND RECEPTION 111

bird which in turn is more sensitive than man (Schneider, 1950).
Flies are among the least-sensitive animals (Fig. 70).

Pure vibrations at these high frequencies do not generally occur in
nature, but such high frequencies undoubtedly exist as transients in
the single pulse-hke vibrations of the substrate to which insects are
ordinarily exposed. Suddenness is the important factor, so perception
of these transients, even without frequency discrimination, may be
the contribution which the subgenual organs make to behaviour
(Autrum, 1959).

CHAPTER V

Chemoreception

The accomplishment of many sexual, reproductive, social, and feeding
activities of insects depends to a great extent upon the detection and
assessment of specific chemical aspects of the environment. With the
exception of feeding, where most of the chemical energy of the stimu-
lating materials is utilized for purposes remote from the activity of the
sense organ, the energy of compounds eliciting the other behaviour
patterns serves merely to trigger excitation in receptors.

These receptors are not chemosensitive simply because of fortuitous
anatomical location, or exposure to the environment, or the possess-
ion of permeable coverings (Dethier, 1962). On the contrary, they are
inherently highly specialized and specific. Because of their specificity it
v^ould probably be more meaningful to speak of salt, sugar, or sex-
attractant receptors than of olfactory and contact chemoreceptors ;
however, since our knowledge of their physiology is still fragmentary,
the two broad but ambiguous categories are still convenient (cf.
Dethier and Chadwick, 1948 a).

It was postulated long before being demonstrated that chemo-
receptors generated impulses as did other receptors. There is nothing
unique about this part of the train of events initiated by stimulation.
At the opposite end of the train, a beginning is being made towards an
understanding of the behavioural correlates (cf., e.g., Dethier, 1957).
The great unknown now is the basis of specificity and the nature of the
transducing mechanism (Dethier, 1956, 1962).

THE RECEPTORS

The identification of chemoreceptors (whose nature is never so obvious
as that of photoreceptors and sound receptors) has probably been
retarded more than that of any other sense modality by the tendency to
assign the function of chemoreception to all sensilla whose structure
conforms to some postulated norm. There is, therefore, a voluminous
literature on the structure of chemoreceptors real and supposed (for
detailed reviews consult Kraepelin, 1833; Rohler, 1906; Forel, 1908;
Mclndoo, 1914 a, 1914 b; von Frisch, 1921; Minnich, 1929 a; Mar-
shall, 1935; Dethier and Chadwick, 1948 a; Dethier, 1953).

112

CHEMORECEPTION 113

By observing the responses of insects (from which various append-
ages or parts thereof have been extirpated) to food, naturally occurring
attractants, and odours to which they have been conditioned, it has
been established that the principal sites of chemoreceptors are:
antennae, maxillary and labial palpi or their homologues, legs, and
ovipositors. Since all of these areas contain a large and heterogeneous
population of sensilla, the assignment of a chemoreceptive function to
specific ones has been an arduous and uncertain task (cf. Dostal, 1 958 ;
Schneider, 1961). Thus, despite extensive histological studies from
those of Hauser (1880), Schenk (1903), and Vogel (1923 b) to the
present time, the identity of chemoreceptors is known in fewer than a
score of cases.

Olfactory Receptors

An olfactory function has been assigned to the following receptor types
with a fair degree of certainty : in the honeybee, sensilla placodea (von
Frisch, 1921; Dostal, 1958), sensilla basiconica, and possibly sensilla
coeloconica (Dostal, 1958); in the dung beetles, Geotrupes sylvaticus,
G. vernalis, Necrophorus tomentosus, N. vespilloides, Silpha novebor-
acensis, and S. americana, sensilla basiconica (Warnke, 1931, 1934;
Dethier, 1947 a) ; in lepidopterous larvae, sensilla basiconica (Dethier,
1941; Morita and Yamashita, 1961); in the human louse, sensilla
basiconica (Wigglesworth, 1941); in housefly larvae, sensilla basi-
conica (Bolwig, 1946); in Drosophila, sensilla basiconica (Begg and
Hogben, 1946); in Phormia regina, sensilla basiconica (Dethier,
unpublished) ; in saturniid moths, probably sensilla trichodea, sensilla
basiconica, sensilla coeloconica (Schneider, 1961) ; in the grasshoppers
Melanoplus differentialis and Romalea microptera, sensilla basiconica
and sensilla coeloconica (Slifer, 1961).

The saUent features of sensilla placodea as seen with the light
microscope are illustrated in Fig. 11. Details of structure have been
clarified by the electron-microscopic studies of Richards (1951) and
Slifer (1961). The sensillum is essentially a thin circular or oval plate
connected to the surrounding cuticle by a delicate circumferential
membrane and equipped with twelve to eighteen bipolar neurons
(Vogel, 1921, 1923 b). Each dendrite possesses one or two minute
refringent bodies of unknown function, the Riechstabschen (Vogel,
1923 b). Furthermore, each dendrite ends in a cilium-like structure
(Slifer, 1961) similar to that described in tympanic organs. From each
cilium-like structure there extends a fibrous strand. The strands termi-
nate on the extreme edge of the plate (Slifer, 1961). The antenna of a

114 THE PHYSIOLOGY OF INSECT SENSES

honeybee possesses from 3,000 to 30,000 sensilla placodea (Schenk,
1903; Mclndoo, 1914b; Vogel, 1923 b; Melin, 1941 ; Kuwabara and
Takeda, 1956; Dostal, 1958). Although the behavioural experiments
of von Frisch (1921) and Dostal (1958) support the idea that these
sensilla are olfactory, it is still difficult to comprehend in what way
their structure is particularly suited to chemoreception (see also
Snodgrass, 1935; Slifer, 1961).

Sensilla basiconica are so variable in size, shape, and number of
neurons that a general description is meaningless. Extensive electron-

FiG. 71. Diagram of a thin- walled sensillum basiconicum from a grass-
hopper antenna. A, section through a perforation showing finger-like
processes of dendrite ; B, small portion of scolopoid sheath showing the
dendritic processes passing through; C, longitudinal section of peg
showing dendrite passing through scolopoid sheath at base ; D, external
surface showing permeable basal spot and numerous fine perforations
of peg. (Redrawn from Slifer et al., 1959.)

microscope studies have been made only of grasshoppers on the an-
tennae of which two types occur (Slifer, 1961). The length of a typical
thick-walled peg (sensillum basiconicum) is 20-50 [i (Fig. 14) (Slifer,
Prestage, and Beams, 1957). The tip possesses an opening about 2 (x in
diameter. A tubular cuticular sheath extends from this opening down
the lumen of the peg, constricts in the region of its base, then again
widens. It is the scolopoid sheath. The dendrites of the neurons, usually
five, are conducted within this sheath to the tip of the peg, where,
according to Slifer, Prestage, and Beams (1957), they are exposed to
the air and bathed in a fluid derived from the vacuole of the tormogen
and trichogen cells. At moulting the cuticular sheath is drawn out
through the orifice at the tip of the peg.

CHEMORECEPTION 115

The typical thin-walled sensillum basiconicum of the grasshopper
antenna is 16 [j. long and 3 fx in diameter at its base (Fig. 13). The scol-
opoid sheath, instead of extending to the tip as in the thick-walled
pegs, makes a right-angle turn at the base of the peg and comes to the
surface. At moulting the old sheath is pulled out through this area.
The dendrites of the forty to sixty neurons enter the scolopoid sheath
in the usual manner, but when the latter turns, they perforate its side
and continue into the lumen of the peg (Fig. 71). Within this fluid-filled
cavity each one branches once or twice and each branch terminates at
one of about 150 openings on the peg's surface (Slifer, Prestage, and
Beams, 1959). At its termination each branch is composed of about
twenty-four parallel finger-like processes (Fig. 71). The long, hotly
debated question as to whether the neurons of chemoreceptors are
exposed to air, a question for which the conventional answer had been
negative (cf. Dethier, 1954 b) would now seem to have been answered
in the affirmative.

Contact Chemoreceptors

Receptors sensitive to stimulation by chemical solutions applied
directly have been identified positively in Diptera, Lepidoptera,
Hymenoptera, and Coleoptera. They are trichoid sensilla located on
the legs and mouth-parts. Small sensilla basiconica on the labella of
flies have also been identified as gustatory receptors. There is no doubt
that there are also other types of gustatory receptors, but their identity
still requires confirmation. For details regarding the location of con-
tact chemoreceptors in various insects the review of Frings and Frings
(1949) should be consulted.

The contact chemoreceptive hairs, as exemplified by those on the
tarsi and labella of Diptera, are actually compound sensilla ; that is,
they are organs containing several kinds of receptors. In the black
blowfly, Phormia regina, the receptors occur as groups of three, four,
or five bipolar neurons associated with long (30-300 \l) hairs which
characteristically possess two lumina (PL II). The cell bodies of the
neurons lie in a subhypodermal position beneath the hair in association
with the tormogen and trichogen cells and one or more neurilemma
and tracheal cells. The entire group of cells is wrapped in a thin
membrane which is continuous with the basement membrane of the
surrounding hypodermis and with the sheath enveloping the nerve.
The nerve consists of the axons of the receptor cells, which are be-
lieved to extend, without synapsing, to the central nervous system. The
dendrites of the neurons extend into a scolopoid sheath and then up

116 THE PHYSIOLOGY OF INSECT SENSES

the thicker walled of the two lumina to the tip of the hair. In triply
innervated hairs as originally described in Phormia by Grabowski and
Dethier (1954) and Dethier (1955 a) the fibres from two neurons extend
to the tip of the hair while the third terminates on the base. The exis-
tence in Phormia of hairs equipped with four neurons was only recently
suggested by behavioural evidence (Dethier and Evans, 1961) and
confirmed electrophysiologically (Mellon and Evans, 1961) and by
electron-microscopy (Larsen, 1962). Hairs with three, four, and five
neurons have been found in Calliphora (Sturkow, 1960; Peters, 1961).
The only part of the hair that is sensitive to chemical stimulation is
the extreme tip, which in some hairs is prolonged into a terminal
papilla into which the dendrites extend (Dethier, 1955 a; Dethier and
Wolbarsht, 1956). In the stable fly Stomoxys calcitrans Adams (1961)
has found that the terminal processes of the dendrites extend through
a pore in the cuticle (PI. III).

OLFACTION
Acuity

Although the behaviour of insects in their natural environment
suggests that their olfactory sense may be extraordinarily sensitive
by human standards, attempts to measure acuity accurately have not
been outstandingly successful. With the exception of Schneider's work
with Bombyx mori and associated species, electrophysiological meas-
urements have not been made. All measurements of the performance
of olfactory receptors have been based upon behavioural thresholds.

80

- )^ 60

o

^ 40|-

UJ

20f-

UJ

o
cr

UJ Q
Q_ UJ

20-

THE STIMULATING EFFECT OF
■ ISO-VALERALDEHYDE ON MUSCA DOMESTICA

0^,,,--^

-

J

/'

—

/

o

/

/

1 1 1 1 1 1 1 1 1 J

J

J L

MM

1

1

IxlO'S IxlO'^ IxlO'

MOLAR CONCENTRATION

Fig. 72. Change in behaviour with change in concentration of the odour of
iso-valeraldehyde. (From Dethier, 1954.)

CHEMORECEPTION 117

As the concentration of a stimulating odour exceeds that necessary
to stimulate the most sensitive receptor element, a behavioural res-
ponse occurs in the form of antennal movements, movements of the
mouth-parts, occasionally salivation or grooming movements whereby
legs or antennae are drawn through the mouth-parts or rubbed against
one another. A further increase in concentration results in an oriented
movement which may carry the insect towards or away from the
source of odour. If the initial response is towards the odour, an addi-
tional increase in concentration may cause a reversal in orientation
(Fig. 72). An odour which is initially repellent usually remains so at all
higher concentrations. An attractive odour may become repellent at
higher concentrations (Dethier, 1947 b; Dethier, Hackley, and Wag-
ner- Jauregg, 1952).

Measurements of acuity thus depend in large measure upon which
behavioural threshold is chosen for testing. Numerous attempts to
measure threshold have been made (Barrows, 1907; von Frisch, 1919;
Wirth, 1928; Folsom, 1931; Warnke, 1931; Reed, 1938; Wietingand

Table 1

(From Schwarz, 1955)

Differences between the olfactory thresholds of man and honeybees

THRESHOLD CONCENl RATION I MOLECULES/CC AIR

Odour

Man

Author

Bee

Difference

Propionic acid

4-2 . IQii

v. Skramlik

4-3. 1011

10. 1010

Butyric acid

7-0. 10^

V. Skramlik

1-1 . 1011

10-3 . 1010

Valeric acid

60. 1010

v. Skramlik

iso-Valeric acid

4-5. 10i«

Schwarz

1-6. 1011

11-5. 1010

Caproic acid

20. 1011

V. Skramlik

2-2. 1011

2-0.1010

Ethyl caproate

1-3. 1011

Schwarz

3-8. 1011

2-5. 1011

Ethyl caprylate

3-7. 1010

Schwarz

5-4. 1010

1-7. 1010

Ethyl pelargonate

3-1 . 1010

Schwarz

3-7. 1010

6-0.10^

Ethyl caprate

4-2 . 10«

Schwarz

5-6. 10^

1-4. 10«

Ethyl undecylate

1-4. 1010

Schwarz

1-8. 1010

40. 10»

Methyl

anthranilate

2-6.1010

V. Skramlik

1-9. 10«

24-1 . 10»

Phenyl propyl

alcohol

6-5. 10»

Schwarz

2-2. 10»

4-3 . 10»

Nerol

5-7. 10^

Schwarz

3-2. 10«

2-5. 10»

lonon-a

31 .10«

Zwaardemaker

1-5. 1010

146-9 . 108

Eugenol

8-5.1011

Ohma

2-0. 1010

8-3 . 1011

Citral

4-0. 1011

V. Skramlik

6-0. 1010

3-4.1011

118 THE PHYSIOLOGY OF INSECT SENSES

Table 2

(From Schwarz, 1955)

Olfactory thresholds for man and various insects

THRESHOLD CONCENTRATION

Odour

Species

Author

Original value

Mol.jc.c

'

Man

3-2. 101" M/50ml.

1-84.108

V. Skramlik

Necrophorus

0-00009 g./lOOml.W.

4-17.101*

Abbott, 1937

Skatol ^

Geotrupes

0-003 -0-009 mg./l.

1-83.1013-
-4-61.101*

Warnke, 1931

-

Hydrous

0-0000625%

Ritter, 1936

Benz-
aldehyde I

Man

0-44 . 10-« g./l.

2-49.1012

Ohma

Pieris rapae

580. 10-'' g./l.

3-28.101*

Dethier, 1941

Malacosoma

435 . 10-' g./l.

2-46.101*

Dethier, 1941

r

Man

2-0 . 10^^ M/50 ml.

4-0 .1013

V. Skramlik

Benzol -

Habrobracon

0-5-3-0 mg./l.

3-87.101°-
-2-3 .10i«

Wirth, 1928

c

Man

1-6. 101' M/50 ml.

3-2 .101^

V. Skramlik

Musca

0-23 g./l.

3-01. 10i«

Wieting &

Ethyl
alcohol

domestica

Hoskins, 1939

Habrobracon

5 -20 mg./l.

4-7 .10i«-

-1-8 .101'

Wirth, 1928

Phormia

2-4. 10-*mol.

1-44.10'^°

Dethier &

-

regitia

Yost, 1952

Hoskins, 1939; Dethier, 1941; Crombie, 1944; Dethier and Yost,
1952; Hodgson, 1953; Schwarz, 1955; Fischer, 1957; Dostal, 1958;
and others [see Dethier, 1947 b] ). A few data are given for comparison
in Tables 1 and 2.

Attempts have been made to correlate the acuity of olfaction in the
various species with the numbers of sensilla, but the correlation is
highly conjectural in as much as the identity of the receptors is un-
certain in many species, and quantitative tests of acuity have not been
undertaken. It is certainly true that some correlation exists between
the number of receptors and behavioural response to odours. The
number of olfactory sensilla in an insect varies from species to species.
There may be as few as 9-10, as in the human louse, or there may be
as many as 30,000 in the drone honeybee (Schenk, 1903 ; Vogel, 1923 b;
Mclndoo, 1914a, 1914b). Lice and lepidopterous larvae, which live
on their hosts and have little apparent need for an extremely acute
olfactory sense, possess few sensilla. In the life of the honeybee, on the
other hand, many odours are encountered and are highly significant.

CHEMORECEPTION 119

Among the Diptera Liebermann (1926) has shown that the so-called
olfactory pits (Fig. 73) average 820 in dung-feeding species which de-
pend largely upon odour for locating their food, while in the flower-
visiting species, which depend more on vision, the average number is
494. Moreover, the pits are more numerous in males than in females.
This difference is presumed to be related to the fact that males employ
the olfactory sense in the search for females.

In an individual there is clearly a relation between threshold and the
number of sensilla stimulated. The phenomenon was first investigated

Fig. 73. Olfactory pit on the antenna of a dipterous insect. (From

Liebermann, 1926.)

with hygroreception. Pielou (1940) suggested that a threshold number
of receptors is required by Tenebrio molitor for a response. Detailed
confirmation was provided by the experiments of Roth and Willis
(1951 a), which showed that the percentage of response of a population
of Tribolium was closely correlated with the number of sensilla
basiconica remaining on each individual after surgical operation. The
change in threshold with unilateral antennectomy is illustrated by the
figures in Table 3 based on data of Roth and Willis (1951 b). For
olfaction the first quantitative data were those of Dethier (1952 b),
which showed that in Phormia regina there is a definite relation be-
tween the threshold and the number of receptors functioning. The
data suggest that for a response to occur a given number of molecules
I

120 THE PHYSIOLOGY OF INSECT SENSES

must hit a given number of receptors and that the probabihty of a
response can be increased by increasing either the concentration or the
number of receptors.

Table 3
(From Dethier, 1952 b)

Effects of unilateral and bilateral operations in the responses of beetles

given a choice between 0% and 100% R.H. at 27°C.

(Data from Roth and Willis, 1951b)

Species

Receptor areas remaining on
Antenna

% Response

Rhyzopertha dominica
(thin-walled sensilla
present on segments
8-10)

Latheticus oryzae (thin-
walled sensilla present
on segments 7-11)

Tribolium castaneum
(thin-walled sensilla
present on segments
9-11)

Tribolium confusum
(thin-walled sensilla
present on segments
7-11)

Segments
Segments
Segments
Segments
Segments
Segments
Segments

Segments
Segments
Segments
Segments

Segments
Segments
Segments
Segments
Segments

Segments

Segments

segments

antennae

Segments

Segments

Segments

Segments

Segments

Segments

Segments

-10 on both (normal)

-10 on one

-9 on other

-9 on both

-10 on one

-7 on other

-7 on both

-1 1 on both (normal)
-1 1 on one 1
-10 on other/
-10 on both

-1 1 on both (normal)
-9 on both
-9 on one
-8 on other
-8 on both

1 1 on both (normal)
-11 on one
on other

-7 on both
-7 on one
-6 on other
-6 on both
-11 on one
-6 on other
-6 on both

76±2-7
69±30
56±3-4
61±2-9
01±7-0

87±3-l
73±3-2
46d=2-7

80±10

75±2-9

78di3-6
31±5-4

86±l-7

76±30

-0-2±2-2
23±l-6

12±3-9

5±4-l

75±2-l
5±4-l

A more detailed investigation of this phenomenon has been made
by Dostal (1958), who studied the response of honeybees to the sub-
stance Nerol R (CjoHigO), a synthetic constituent of rose oil. She
showed that with successive amputation of one to six antennal seg-

CHEMORECEPTION 121

ments the acuity of olfaction decreases only slightly. After extirpa-
tion of seven segments a pronounced decrease occurs, and amputa-
tion of eight segments abolishes all response. A comparison of the
frequency distribution of sensilla on the antennae and the threshold
of response reveals a logarithmic relationship which Dostal relates
to the Weber-Fechner law. An examination of the data which Schanz
(1953) obtained by measuring thresholds in the potato beetle after
differential amputation of antennal segments reveals similar relation-
ships.

Not all olfactory receptor fields of an individual are equally sensi-
tive. As a rule response thresholds are lowest when the antennae are
intact. Many insects lose their ability to respond to attenuated odours
when the antennae are rendered inoperative, but can respond to
higher concentrations as long as the palpi are still operative (Warnke,
1931; Frings, 1941; Dethier, 1941, 1947 a, 1952 b).

The Stimulus

As with man the number of gaseous materials, natural and synthetic,
to which insects respond is very great. Many attempts have been made
to relate stimulating effectiveness to some molecular property, and
many data have been collected by workers interested in insect attract-
ants and repellents. When these data can be translated into terms of
threshold some interpretation is possible. One of the earliest of such
studies was that of Cook (1926). It was designed to ascertain the re-
lation between the optimum attractiveness to flies and the physical
properties of aliphatic alcohols from Q to C5 and esters from acetates
to valerates. Within a homologous series acceptance threshold de-
creases with increased boiling point. The relationship is logarithmic.
Instead of ascertaining the relative stimulating effect of members of
a homologous series by testing each member separately over a concen-
tration range as Cook had done, one may derive the same information
by testing a number of compounds simultaneously and employing the
number of insects attracted as an index for efficiency. Employing this
method, Speyer (1920) found that homologous alcohols and esters
become more attractive as the chain length is increased. Increasing the
chain length on the acid side of the ester molecule results in a more
pronounced and uniform increase in effectiveness than do similar in-
creases on the alcohol side. Comparable experiments with codling
moths yielded similar results (Eyer and Medler, 1940). Olfactometric
experiments with the blowfly Phormia regina have shown that the
rejection threshold to homologous alcohols decreases logarithmically

122 THE PHYSIOLOGY OF INSECT SENSES

as the chain length is increased until a cut-off point is reached at or near
Cii (Dethier and Yost, 1952). With aldehydes both rejection and
acceptance thresholds reveal similar relationships (Dethier, 1954 a).

Mechanism of Olfaction

The nature of the process whereby the odour molecule initiates de-
polarization in the olfactory receptor remains a mystery despite
numerous theories, of which those of Ehrensvard (1942), Davies
(1953a, 1953b), Mullins (1955 a, 1955b), and Davies and Taylor(1959)
are the more comprehensive. Analyses of relationships between thres-
hold of response and molecular characteristics have contributed much
of the groundwork for theorizing. On the basis of measurements made
with Phormia (Dethier and Yost, 1952; Dethier, 1954 a) and with man
(MulHns, 1955 a), it is clear that olfactory thresholds, stated as thermo-
dynamic activities, for members of homologous series of aliphatic
alcohols, aldehydes, and saturated hydrocarbons are nearly constant

<
o

-2

12 3 4

LOG foo (ZS^'C)

Fig. 74. Comparison in terms of thermodynamic activity of the stimulating
effectiveness of the first eight normal alcohols acting in aqueous
solution (open circles) on tarsal chemoreceptors and as gases (solid
circles) on olfactory receptors of blowflies. In each case the value
represents a threshold of rejection. The vertical lines represent 2-575
standard errors for aqueous thresholds and 2 for vapours. (From
Dethier and Yost, 1952.)

CHEMORECEPTION 123

over the middle range of chain length (Fig. 74). Olfactory thresholds
for fatty acids measured in the dog (Neuhaus, 1953 a, 1953 b) and the
honeybee (Schwarz, 1955) do not show such a precise relationship.
For these two animals the thresholds, expressed as molecules per c.c.
of air, are lowest in the region Co to C4. Dethier (1954 b) pointed out
that the relationship observed with alcohols suggests that olfaction in
these cases may involve the establishment of an equilibrium, and as
such represents a physical rather than a chemical process which in-
volves specific receptor substances (Dethier, 1956). Mullins (1954)
further analysed the date of Dethier and Yost (1952) by plotting the
product of the thermodynamic activity at threshold and the molal
volume of the compounds against chain length and assigning a cal-
culated membrane solubility parameter of 11 -5. The analysis suggested
that the vapour of the first three alcohols might not have been in
equilibrium with the organism. Additional data and analyses indicated
that olfaction does not involve an equilibrium process and that excita-
tion is probably a result of the ability of the odour molecules to pro-
duce local disorder in the oriented molecular structure of the cell
membrane (Mullins, 1955 a, 1955 b).

The theories of Ehrensvard (1942) and Davies (1953 a, 1953 b)
postulate that odour molecules must first adsorb into the plasma mem-
brane of the olfactory neuron. Ehrensvard suggested that the potential
changes due to the adsorption actually initiated impulses in the re-
ceptor. The Davies theory envisions the adsorbed molecules as dislo-
cating the membrane, thus permitting an exchange of sodium and
potassium ions across it, with the consequent initiation of impulses.
The effectiveness of compounds depends not only upon the concen-
tration of adsorbed molecules but also upon their size and shape
(Davies and Taylor, 1959). Olfactory threshold data for Phormia, as
well as for dogs and man, confirm that both shape and adsorption are
important factors (Fig. 75).

Electrophysiological studies thus far have not clarified our under-
standing of the receptor mechanism. Summed spontaneous activity in
antennal nerves has been recorded by Boistel and Coraboeuf (1953),
Boistel, Lecompte, and Coraboeuf (1956), Roys (1954), Smyth and
Roys (1954), Schneider (1955, 1957 a, 1957 b), Schneider and Hecker
(1956), and Morita and Yamashita (1961). Firing from individual
neurons has occasionally been detected in the antennae of Bombyx
mori and related species (Schneider, 1957 b), but it has not been pos-
sible to ascertain from which type of sensillum the activity originated
(Schneider, 1961). When the olfactory stimulus is extract of female sex

124 THE PHYSIOLOGY OF INSECT SENSES

attractant (Bombyx mori) hexadeca-diene-(10, 12)-ol-(l) (Butenandt,
Bechmann, Stamm, and Hecker, 1959), electrical activity increases in
the male antenna but not in that of the female. When cyclohexanon or
sorbinol are employed as stimuli activity increases in both sexes. The
spikes are superimposed on a complex slow potential which Schneider
(1957 a) has termed the electroantennogram (EAG). The amplitude

-70 —

-60

:2

o

^ -50

-40

-30

-20'^

Fig. 75. Plot of the logarithm of the olfactory rejection concentration for
blowflies against the adsorption constant at an oil/air interface
(logio Kojd). The broken line represents the slope expected if Ijp were
to increase regularly with chain length. Threshold data from Dethier
and Yost (1952). (Redrawn from Da vies and Taylor, 1959.)

depends upon the concentration of the odour, and the wave form, on
the kind of molecule. Many compounds in high concentration elicit
an EAG (ether, ethanol, propanol, butanol, xylene, etc.), but at
extreme dilution only the natural sex attractant elicits an EAG. Its
wave form is characteristically different (Schneider, 1961). There are
no significant differences in the normal EAGs of a male antenna
when the stimuli employed are the glands of related saturniid species,

CHEMORECEPTION 125

although the conspecific gland tends to elicit the largest response
(Schneider, 1962).

Recently Morita and Yamashita (1961) have been able to record
directly from a simple large sensiHum basiconicum on the antenna of
the larva of Bombyx mori. This had been shown to be an olfactory
organ innervated by approximately twenty neurons (Dethier, 1941).
When odours of substances contained in the essential oil of mulberry
leaves (e.g., Py-hexanol and /z-butylaldehyde) were applied to the sen-
sillum a slow negative potential accompanied by an increase in action
potential frequency occurred. Some compounds evoked a slow diph-
asic potential, from negative to positive, while anaesthetics evoked
slow positive potentials accompanied by a decrease in impulse fre-
quency. A complete interpretation of these results is rendered difficult
by the fact that the sensillum is multiply innervated; however, the
available evidence strongly suggests that the slow negative potential
evoked by naturally occurring food substances is a true receptor
potential.

Odour Qualities

It is clear from studies of attractants and repellents that for all insects
there are at least two odour modaUties, that is, 'acceptable' and
'unacceptable', but it is unlikely that these represent accurately the
olfactory world of insects. For many insects that respond in a highly
specific manner to one particular odour, as, for example, the male
silkworm to female sex-attractant, and many monophagous cater-
pillars to the odour of their one food-plant, it is possible that the ol-
olfactory receptors are tuned only to that odour. Electrophysiological
recordings from the antennae of male Bombyx mori thus far have not
revealed a sensitivity to any compound other than the natural and
synthetic sex attractants of Bombyx and the natural attractants of
closely related species (Schneider, 1962).

The honeybee, on the other hand, is obviously sensitive to many
odours, but does it confuse them as one or can it discriminate? By
training bees to associate a particular odour with food, von Frisch
(1919) was able to show that a number of odours were discriminated.
Furthermore, compounds with different structure, which to man
possessed similar odours, were confused as one by the bee, whereas
compounds with nearly identical structure but obviously different
odours were easily distinguished. Within the following pairs the two
members are easily distinguished by the bee : amyl acetate and methyl
heptenone, bromstyrol and phenylacetaldehyde, isobutyl benzoate

126 THE PHYSIOLOGY OF INSECT SENSES

and salicylic acid amyl ester, /?-cresol methyl ether and w-cresol
methyl ether.

CONTACT CHEMORECEPTION
Sensitivity

Many behavioural studies have been conducted with numerous species
of insects for the purpose of locating and mapping contact chemore-
ceptors, determining limits of qualities, and elucidating the nature of
receptor action. These studies, centred for the most part around feed-
ing behaviour, have usually been designed as measurements of accept-
ance and rejection thresholds. Extension of the proboscis, duration of
feeding, and measurement of crop loads have been a few of the criteria
employed (for details consult Dethier and Chadwick, 1948 a).

Table 4

(From Dethier and Chadwick, 1948 a)

Comparison of taste thresholds

Compound

Threshold Concentrations

Sucrose

Sodium chloride

Hydrochloric

acid
Quinine

Man, 0-02M; bee, 0-06-0-125M; butterfly (Pyrameis),
average ca. 0-OlM, in starvation as little as
8 X IQ-^M; Danaus,9'S x lO-^M;hoTSQfiy(Tabanus)
0005-0 IIM.

Man, 0-009M; bee, rejects ca. 0-24M in 0-5M sucrose;
various caterpillars reject at 0-2M, whereas others
accept over the full range up to and including 5 -OM.

Man, 000125M; bee, rejects OOOIM in 1 OM sucrose;
various caterpillars reject at 0-01-0-2M.

Man, 1 -5 x 10" "^M; bee , rejects at 8 x 10"* in 1 -OM
sucrose; various caterpillars reject at 0002-0-033M;
aquatic beetles were conditioned to respond to
1-25 X lO-^M.

Data on man are for specific thresholds, adapted from Moncrieff" (1944);
for the bee, from von Frisch (1934); for Pyrameis, from Minnich (1922);
for Danaus, from Anderson (1932); for caterpillars, from Eger (1937); for
aquatic beetles, from Bauer (1938). The figures for insects are thresholds
of response.

Threshold studies have revealed many of the characteristics of the
contact chemical sense. They have shown, for example, that the
thresholds are frequently lower than those of man for similar sub-
stances (Table 4). They have revealed among individuals differences
that follow a common pattern. For example, the scattering of

CHEMORECEPTION 127

thresholds in a population ofPhormia is not significantly different from
a normal distribution when plotted against the logarithm of concentra-
tion (Dcthier and Chadvvick, 1947). Where sufficient data have been
reported a similar relationship can be shown to exist with other insects
(e.g., Eger, 1937; Frings, 1946; von Frisch, 1935; Weis, 1930;
Hodgson, 1951).

Threshold determinations have revealed that sensitivity of response
varies with the receptor field stimulated. In the horsefly Tabanus the
mean labellar threshold for sucrose is 0-02 IM; the tarsal threshold,
0-060M (Frings and O'Neal, 1946). In CallipJwra erythrocephala the
labella are generally more sensitive than the tarsi (Haslinger, 1935). In
C. vomitoria, on the other hand, the tarsi are sixteen times more
sensitive than the labellum to sucrose (Minnich, 1931), but lactose
elicits a response only when applied to the labellum (Minnich, 1929 b).
Musca domestica, Cochliomyia americana, and Phormia regina exhibit
lower rejection thresholds following oral stimulation than tarsal
stimulation (Deonier, 1938, 1939;Dethier, 1955a). Similar results have
been obtained WiihPieris (Verlaine, 1927). The antennae of honeybees
are more sensitive to solutions than the legs (Frings, 1944; Marshall,
1935 ; Minnich, 1932) ; the proboscis, more sensitive than the antennae
(Kunze, 1933). In the beetle Hydrous the labial palpi are sensitive to
hydrochloric acid and insensitive to sodium chloride and sugar, while
the maxillary palpi are sensitive to sodium chloride and sugar (Ritter,
1936). In the related species Laccophilus the antennae are more
sensitive than either pair of palpi to acids, salts, and alcohols (Hodgson,
1951).

The aforementioned studies do not reveal whether the differences
in threshold reflect central phenomena, variations in the sensitivity of
the receptors, summation, or differences in sensitivity due merely to
the number of receptors stimulated.

Differences between acceptance thresholds following unilateral and
bilateral tarsal stimulation had suggested that some relation does
exist between sensitivity and the number of receptors stimulated
(Imamura, 1938). An analysis of thresholds following stimulation of
one tarus versus two tarsi in Phormia showed that the bilateral thres-
hold for sugar is lower than the unilateral threshold (Dethier, 1953 b).
This difference could be due, however, either to summation or to a
simple statistical bias. In the analogous visual case where comparisons
have been made between monocular and binocular vision, Pirenne
(1943) and Barany (1946 a, 1946 b) pointed out that the experimental
procedure by its very nature assures that the two eyes will see more

128 THE PHYSIOLOGY OF INSECT SENSES

clearly than one. 'Let us assume that the visual acuities (or other
thresholds) of both eyes fluctuate independently of one another . . .
and that the instantaneous thresholds for monocular vision have the
same distribution in both eyes, . . . then if the one eye alone has the
chance of a of seeing the symbol, both eyes together have the chance
2a— a^. As a is smaller than 1, this expression will always be greater

7.0

(f)
h-

d)
O

Q.
O

o

<

UJ

o
cr

UJ

6.0

5.0

• I LEG-RANDOM
O 2 LEGS-RANDOM

•3.0

2.0

1.0

0.0

LOG MOLAR CONC. SUCROSE

Fig. 76. Comparison of the distribution of acceptance thresholds for
sucrose, as a function of concentration, for flies stimulated unilaterally
or bilaterally. The broken line represents the theoretical distribution
of bilateral thresholds (two-legged flies) calculated from the expression
\—q^, where q equals the fraction of the population of one-legged flies
not responding. (From Dethier, 1953.)

than a - that is to say, two eyes will be able to see better than one solely
as a result of random combination' (Barany, 1946 b, p. 127). And as
Smith and Licklider (1949) went on to state, the same source of bias is
inherent in the procedure as appHed to the determination of thresholds
in other sense modalities.

The idea can be clarified still further by quoting from the study of
hearing by these authors (p. 279). *In order to estimate the magnitude
of the bias, it is necessary to define the null condition under which we

CHEMORECEPTION 129

should say that there is no binaural summation. We can imagine, for
this purpose, two monaural listeners, one with only a right ear, the
other with only a left ear. The two listeners have no means whatsoever
of communicating with each other, but both report to the same experi-
menter. To obtain measures of monaural and "binaural" sensitivity,
the experimenter tests the two listeners separately (successively), then
together, in the latter instance recording a positive response whenever
either hstener reports hearing the stimulus tone.' The case of unilateral
versus bilateral tarsal stimulation of Phormia is similar. The results
show that the bilateral threshold for sucrose is lower than the unilateral
threshold by a factor that is never greater than can be satisfactorily
accounted for on a simple probabihty basis, that is, the greater the
number of available receptors, the greater that chance that the number
required for a threshold response will be stimulated (Fig. 76) (Dethier,
1953 b).

Discrimination

Threshold measurements do not permit evaluation of intensity dis-
crimination over the complete effective concentration range of a com-
pound. Von Frisch (1935) had demonstrated that honeybees responded
differently to 0-125M and 0-156M sucrose and to 0-25M and 0-3125M

Table 5
(From Dethier and Rhoades, 1954)
Intensity discrimination of sucrose

Intensity

Nearest cone, which
can be distinguished

Mil

Mil

{geometric
mean)

0001

000007

0-99

2-4

0001

007

60

001

007

0-3

0-387

001

0015

0-5

01

0088

012

012

01

0112

012

10

0-70

0-3

0-346

10

1-40

04

2

10

0-5

0-418

20

2-7

0-35

130 THE PHYSIOLOGY OF INSECT SENSES

sucrose, a ratio of 1 : 1-24 (AS IS = 0-25). Frings (1946) and Frings
and O'Neal (1946) had found that Periplaneta americana and Tabanus
sulcifrons could also discriminate at this level. On the basis of tests
with two specimens Verlaine (1927) claimed a slightly greater sensi-
tivity for Pieris,

A technique permitting the measurement of A/// for the complete
effective concentration range of a compound was developed by Dethier
and Rhoades (1954) and employed with Phormia. The smallest ratio
found (0-12) indicates better discrimination than that obtained in
earlier studies. An examination of the results tabulated in Table 5
shows that discrimination is best over the middle ranges of concentra-
tion and is optimum at about the same concentration for which there
is a maximum preference. The discrimination factor for the chemical
senses of man is about 0-3 according to Moncrieff (1944) but earlier
Lemberger (1908) had recorded optimum values of 0-15 for sucrose
and 0-11 for sodium-saccharin and higher values at both extremes of
the concentration range (cf. also Dahlberg and Penczek, 1941;
Schutz and Pilgrim, 1957).

Rejection Thresholds - Electrolytes

The number of compounds that can be detected by insects is very great.
From a behavioural point of view the only manifestation of stimula-
tion by the majority of them is the rejection of acceptable solutions
(e.g., sugar or water) to which they have been added. Rejection thres-
holds, therefore, represent the concentration of substances necessary
to prevent responses to sugar or water. Because the test solutions are
always mixtures, the thresholds involve : the sensitivity of two different
receptors, possible interactions at the receptor level, and demonstrated
interaction at the central level. Despite these complications, it has been
possible to estabhsh rather precise relationships between the stimulat-
ing effectiveness of many compounds and their molecular properties.
Electrolytes were among the first compounds studied. Tests con-
ducted with the mouth-parts of the cockroach Periplaneta americana
showed that for several series of salts with a common union (acetates,
bromides, chlorides, iodides, nitrates, and sulphates) the stimulative
efficiency could be correlated directly with ionic mobilities (Frings,
1946). Similar results had been obtained earher with the oral receptors
of cecropia moth larvae (Frings, 1945). Extensions of these studies to
the tarsal receptors of the horsefly Tabanus sulcifrons (Frings and
O'Neal, 1946), the ovipositors of parasitic Hymenoptera (Dethier,
1947 c), oral receptors of the hQetle Laccophilus (Hodgson, 1951), the

CHEMORECEPTION 131

labellar receptors ofPhormia (Dethier, 1955 a), and the receptors of a
number of other species (Frings, 1948) showed that the same general
relationships held. According to Frings, the data are in substantial
agreement with most of the data reported in the literature for chemical
stimulation of other animals by the series of cations employed, namely,
in decreasing order of thresholds: Li+ > Na+ > Mg++ > Ca++ =
Sr++>K+ = Cs+ = Rb+>NH4+>>>H+. As pointed out by
Frings, the order of effectiveness is also that of partition coefficients,
which parallels that of ionic mobilities (Osterhout et ai, 1934). The
arrangement for anions is less clear. The work of Frings (1946) estab-
lished the following order of thresholds: PO4 > Ac~ > SO4 =
CI- ^ Br- > I- > NO3- > > > OH-.

The situation is actually much more complex than these generaliz-
ations imply. For example, the activity of divalent cations is anom-
alous in that they do not fit any of these series. Furthermore, the
behaviour of insects towards divalent ions is noticeably different from
their behaviour towards monovalent ions (cf. Hodgson, 1951). In
addition, there are species differences. It is possible that these may
reflect differences in the molecular structure of the receptors, as has
been proposed for mammals (Beidler, 1953, 1954, 1960; Beidler, Fish-
man, and Hardiman, 1955).

Rejection Thresholds - Non-electrolytes

The rejection thresholds of over two hundred aliphatic organic com-
pounds have been measured by employing flies that had been rendered
anosmic by removal of the antennae and labella (Dethier and Chad-
wick, 1947, 1948 b, 1950; Chadwick and Dethier, 1947, 1949). A
number of significant relationships were revealed by these data.

Within any homologous series exceptionally high correlations were
found between the stimulating power of the compounds and such
properties as boiling point, molecular area, oil-water partition coeffi-
cients, molecular moments, vapour pressures, and activity coefficients.
Molecular weight, number of carbon atoms, and osmotic pressure
were eliminated from consideration in part by the use of isomers,
which proved usually to have different thresholds, and the correlation
with vapour pressures was inverse. Because of the paucity of data
relating to the various chemical and physical properties enumerated
above, it became convenient to plot stimulative efficiency against the
chain length of the molecule. Graphs so constructed (Fig. 77) show
clearly that for all series of homologous aliphatic compounds studied
the members of the series are rejected at logarithmically decreasing

132 THE PHYSIOLOGY OF INSECT SENSES

concentrations as the carbon chain length is increased. It is apparent
from Fig. 77 that this relation is not a continuous one. For each series
the curve shows a sharp break in the region of a definite chain-length
characteristic for each series.

Taking a saturated straight-chain hydrocarbon as a starting-point,
one may, without altering appreciably the arrangement of the carbon

ijj —

-2.0

-3.0

20

LOG NUMBER OF C- ATOMS

Fig. 77. Rejection thresholds of glycols and alcohols by Phormia. (From
Dethier and Chadwick, 1948 b.)

linkages, substitute various kinds of polar groups for one or more of
the hydrogen atoms. Whereas all such substitutions raise the molecular
weight and the boiling point, the effect on solubility is variable. The
different polar groups may be arranged in increasing order of solu-
bility as follows: Br < CI < CHg < CHO < C =- O < OH. The CH3
radical represents the unsubstituted compound. This series also repre-
sents in reverse order the relative stimulating efficiencies of these polar
groups. When the number of substitutions of the groups to the right of

CHEMORECEPTION 133

the CH3 (in the above representation) is increased the stimulating
efficiency decreases. Thus, glycols are less effective than alcohols, and
diketones are as a rule less effective than ketones. Increase of the
number of substitutions of those polar groups to the left of the CH3
tends to increase stimulating effectiveness. Thus, dibromo compounds
are more effective than monobromo homologues.

The effects of the positions of the functional groups are illustrated by
experiments with the glycols. The following rules are found to hold :

(1) juxtaposition of two hydroxy! groups in a short molecule
(e.g., 1, 2-butanediol) makes for a high threshold; this effect is re-
duced as chains become longer;

(2) in a chain with no terminal OH groups, i.e., subterminal but
not adjacent (e.g., 1, 3-butanediol) the thresholds are low;

(3) if both OH groups are terminal thresholds are intermediate;

(4) branching tends to raise the threshold, other factors being
equal.

Taken all together, the foregoing results state essentially that the
length of the free alkyl group largely determines the stimulating
effectiveness and that its power is modified to varying degrees by the
nature of the attached polar groups. The length of the alkyl group is

4'

-i\\

M W. "

-3 -2

LOe SOLUBILITY GRAM MOLES PER UTRE

Fig. 78. Correlation between concentration required for stimulation and
the solubility of compounds at 25-27 degrees C. in water. Each of the
forty-six points represents a different aliphatic compound. (From
Dethier and Chadwick, 1950.)

134 THE PHYSIOLOGY OF INSECT SENSES

also of prime importance in determining the solubility characteristics
of compounds of this type ; hence the same structural characteristics
that decrease water solubility likewise decrease threshold.

There is then only one molecular property, for which data are
available, which brings all the data from the different series into a
single homogeneous system. This is water solubility (Fig. 78). The
order of stimulative efficiencies follows the inverse of the order of
water solubilities with fewer contradictions than appear in most of the
other comparisons attempted. Additional evidence that solubility is of
importance in this connexion has been presented by the work of
Dethier (1951a), which showed that the thresholds of alcohols are
altered as the alcohol is presented as an aqueous solution, a glycol
solution, or a mineral oil solution (Fig. 79).

When threshold values are expressed as thermodynamic activities
rather than as moles the differences between successive homologues of
a series are not so marked, but a plot of the logarithm of these values
against the logarithm of activity coefficients (Fig. 74) does not produce
a straight line of the sort that one is accustomed to expect from
parallel experiments on narcosis (for a complete discussion consult
Ferguson, 1951 ; Brink and Posternak, 1948; Dethier, 1954 b).

In spite of the generally good correspondence between low solu-
bility in water and high stimulating power, it seems likely, from the
data on oil-water partition coefficients, that plots of threshold values
against water solubihties would also yield a smaller slope for the
lower than for the higher range of compounds in each series if such an
analysis could be made. This type of relationship seems to have no
conterpart in any of the tabulated values for the physical properties,
and its consistent recurrence prompted Chadwick and Dethier (1949)
to consider the possibility that different forces may be of primary
importance in stimulation by the lower and higher members of each of
the types investigated. This amounts to postulating at least a two-
phase system for the limiting mechanism in contact chemoreception.
The hypothesis that small molecules gain access to the receptors in
part through an aqueous phase, whereas the larger aliphatic molecules
penetrate chiefly through (or accumulate in) a lipoid phase, would
appear to offer a basis for reconciling most of the contradictions en-
countered when it is attempted to fit the facts into a single-phase
system.

Movement of the smaller molecules through an aqueous medium
should occur at rates related inversely to the molecular weight, which
would help to account for their being more stimulating than is antici-

CHEMORECEPTION 135

pated from the relationship found for the higher members of the series,
although it is doubtful that the entire difference can be explained in
this way. It may be noted also (Fig. 80) that the inflections in the
curves relating thresholds to molecular size occur at increasing chain

bJ

o
cc

UJ
Q.

o
•o

>-

CO

Q

UJ

I-

Ul

cc

2

O

z
o
u

<

o

o

_J

00

-1.0

-20

\

\
\

O \

\

\

\

o \

\

o o

K'

\o

\

\

\ o

-3.0

•4.0

-- ALCOHOLS IN WATER
O ALCOHOLS IN GLYCOL

• ALCOHOLS IN MINERAL OIL

0.0

0.4

0.8

LOG NUMBER OF C ATOMS

12

Fig. 79. Rejection thresholds of aqueous, glycol, and oil solutions of
primary alcohols by Phormia. (From Dethier, 1951.)

lengths in passing from the less to the more water-soluble species. At
the same time the predominant importance of lipoid affinity is sugges-
ted by the logarithmically increasing stimulating power of both
lower and higher members of all series, as well as by the inverse
relationship between water solubility and stimulating effectiveness in
comparisons of the several series with each other.

K

136 THE PHYSIOLOGY OF INSECT SENSES

Further analyses of the date of Dethier and Chadwick (1948 b, 1950)
suggest that the important factor in stimulation by non-electrolytes
is adsorption (Davies and Taylor, 1959). When the logarithm of rejec-
tion thresholds (molar concentration) is plotted against the logarithm

1.0

0.0 -

I—

Ul

o
a:

o
to

>-

CD

O
UJ

O

UJ

o

o
o

q:
<

_j
o

o
o

-1.0

-2.0

-3.0

-4.0

O ALDEHYDES
A KETONES

NORMAL ALCOHOLS

• SECONDARY ALCOHOLS

0.0

0.4 0.8

LOG NUMBER OF ATOMS

12

Fig. 80. Rejection of aldehydes, ketones, and secondary alcohols by
Phormia. (From Chadwick and Dethier, 1949.)

of the adsorption constant for an oil/water interface the slope of the
resulting line is very near that expected for adsorption from solution
on to a pure lipid membrane.

Rejection Thresholds - Organic Electrolytes

The situation with regard to organic electrolytes is less clear than that
for the non-electrolytes. Tests with a number of fatty acids indicated
that while the H ion is the principal factor in stimulation with com-

CHEMORECEPTION 137

pounds of this type, a contribution is made also by the anion or un-
dissociated acid, which is effective in inverse proportion to the hydro-
phile nature of the molecule. The relationship is thus very similar to
that found with the same class of compounds for human taste (cf.,
e.g., Taylor er^/., 1930).

i.5r*'

Q

_J
O
X
CO
bJ

cr

I

z
o

t-
o

UJ

-3
UJ

on

■z.
<

Q
UJ

0.75

• Na

.SALTS OF FATTY ACI

A Bo

i Na SULFONATES

FORMAMIDE

9.55

ACETAMIDE

7.22

CHLOROACETAMIDE

L43

PROPIONAMIDE

3.21

BUTYRAMIDE

1.92

VALE

:ramide

0.37

0.50 -

0.25 -

4 5 6 7 8 9 10 II 12 13 14
NUMBER OF C ATOMS

Fig. 81. Relation between the stimulating efficiency of organic salts
and chain lengths of the anion. (From Dethier, 1956.)

Tests with homologous organic salts show that beginning with the
five-carbon compound there is a logarithmic decrease similar in all
respects to that observed with non-polar homologues (Fig. 81)
(Dethier, 1956).

Acceptance Thresholds

Among compounds acceptable to insects carbohydrates are the most
usual and important. Comparisons of the acceptability of various

138

THE PHYSIOLOGY OF INSECT SENSES

carbohydrates have been undertaken with the honeybee (Vogel, 1931 ;
von Frisch, 1934), Calliphora (Haslinger, 1935), water beetles (Bauer,
1938), the ants Lasius niger, Myrmica rubra, and M. rubida (Schmidt,
1938), wireworms (Thorpe et al., 1947), the butterfly Pyrameis
atalanta (Weis, 1930), the blowfly Phormia (Hassett, Dethier, and
Gans, 1950; Dethier, 1955). The number of sugars acceptable to these

Table 6

(From Dethier, 1953 a)

Acceptance thresholds of various insects for sugars

(Molar concentrations)

PHORMIA (Tarsi)
(Data from

CALLIPHORA (Data from
Haslinger. 1935^

APIS (Mouth) (Data

Sugar

Hassett,
Dethier, and

-~ J — — y

from von Frisch,
1930)

Gans, 1950)

Tarsi

Mouth

Sucrose

001

0-0006

0035

0625-0-125

Maltose

00148

0-00125

0-002

00125

Trehalose

0126

0-14

002

0-25

Lactose

*

*

0-1

*

Cellobiose

5 01

05

003

*

Melibiose

*

N:

*

*

Melezitose

063

07

0-01

0125-0-25

Raffinose

0-275

01

0-0071

*

Fructose

0076

0-0033

0-004

0-25

Fucose

0-087

0-02

0-01

1-0

Glucose

0-114

0-125

0-04

0-25

Sorbose

0-218

...

Xylose

0-400

0-2

0-14

*

Galactose

0-502

0-14

0-09

2-0

Arabinose

0-50

01

08

*

Mannose

7-59

0-2

0-1

*

Ribose

7-94

...

Lyxose

33-1

...

...

Rhamnose

*

*

10

*

* Non-stimulating at all concentrations.
. . . Not tested.

species varies, the blowflies being more catholic than either the honey-
bee or the ants. However, if the data of von Frisch, HasHnger, and
Hassett et al. are compared (Table 6), allowances made for statistical
variance, and the sugars divided into three categories (mono-, di-, and
tri-saccharides), certain general relationships appear. For bees the
order of effectiveness is : (dissaccharides) sucrose = maltose >

CHEMORECEPTION 139

trehalose ; (monosaccharides) fructose = glucose > fucose > galac-
tose. All other sugars fail to stimulate. For Calliphora legs the order is :
(disaccharides) sucrose = maltose > trehalose = cellobiose > lactose ;
(monosaccharides) fructose > fucose > glucose = xylose = galactose
= arabinose = mannose. For Calliphora mouth-parts the order is:
(disaccharides) sucrose = maltose > trehalose = cellobiose > lactose ;
(monosaccharides) fructose > fucose > glucose > arabinose = xylose
= galactose = mannose. For Phormia legs the order is : (disaccharides)
sucrose = maltose > trehalose > cellobiose > lactose ; (monosacchar-
ides) fructose > fucose = glucose > sorbose > xylose = galactose =
arabinose > mannose = ribose > lyxose.

For these three species of insects it is true in general that the most
stimulating disaccharides are the a-glucosides and the most stimulat-
ing monosaccharides are fructose, glucose, and fucose. From studies
with Phormia, in which 75 carbohydrates, 8 polyhydric alcohols, 53
amino acids, and 18 monobasic and dibasic acids have been applied to
individual labellar hairs, some of the limiting factors in stimulation
have been revealed (Table 7) (Dethier, 1955 a). For example, trioses,
tetroses, heptoses, and octoses are uneifective. This finding indicates
that there is an optimum chain length. The size of the molecule as a
whole appears to be critical only within certain limits. Thus, at the
smaller extreme some pentoses stimulate while at the other extreme
certain trisaccharides stimulate. However, polysaccharides are un-
effective. For additional details the work of Dethier (1955 a) should be
consulted.

Categories of Receptors

Ultimately, an understanding of the mechanism of action of the re-
ceptors themselves had to await the application of electrophysio-
logical techniques. Studies of this sort have been confined to a few
carefully selected species : the flies Phormia regina, Lucilia caeser, and
Calliphora vomitoria; the butterfly Vanessa indica; and the Colorado
Potato Beetle, Leptinotarsa decemlineata.

The most intensively investigated and well-known receptors are
those occurring on the tarsi and labellum of the black blowfly,
Phormia regina. They have been the subject of co-ordinated be-
havioural, histological, and physiological study and may be taken as a
model for discussion. The multiplicity of neurons associated with these
hairs complicated for a while studies of the mechanism of chemo-
reception. Because of the specificity of the neurons, however, it was
originally possible to arrive at a considerable understanding of

140

THE PHYSIOLOGY OF INSECT SENSES

.s:
Si

<

On

o

is «

O -^

c c

a g
o ^

o

a
o

c

a

X

4)

O
(-1

(A

>

« Si

GO (O ;2

73 in

CiO

e
3

e

Cfl

hi -^

^ to

c
o

o <=>
GO a O

I ^ ^

C3

'S 8

O Vh

X) ^

2 o

c ^

GO p

•6 2

•^ 'S
-d 5
c -^

o iH

O C/3

«/^ cx

ON 3

3 o x:

S:SQ

o

73 >

S *^

"SI »^

^ 00

= 1
g§

ffi 73

2 c

3 M

f5

0194*
0069*

§

1111+ +il+! 1+ ++ 1 !

m-Erythritol
Penta-erythritol
Mannitol
L-Arabitol
Inositol
Glycosides
a-D-methyl glucoside
[B-D-methyl glucoside
NOa-benzyl glucoside
/7-Aminophenyl glucoside
Monoacetyl NOa-^-glucoside

Phenyl p-glucoside

tetraacetate
a-/)-Methyl mannoside
/7-Aminophenyl

p-maltoside
Nitrophenyl maltoside
7- Acetyl /7-nitrophenol

cellobioside
Substituted sugars
//-Acetyl glucosamine

*
**»** 00**

1 1 1 oo"^fnT}-mONr>-| 1 1 Orn^toON
ooooocb<N oooor-

.V.

Co

! 1 1 + + + + + 1 i 1 1 i + + + + 1

s:
s:

a

Triose

DL-glyceraldehyde
Tetroses

D-erythrose

L-erythrose
Pentoses

£-Fucose

D-Arabinose

L-Arabinose

D-Xylose

L-Xylose

Z)-Ribose

D-Lyxoset

Ribulose

L-Rhamnose

2-Desoxyribose
Hexoses

L>-Fructose

D-Glucose

L-Sorbose

L>-Galactose

D-Mannosef

CHEMORECEPTION

141

6

oo
O

+ + -H +

I + I ^- I

o

Id

Ui

O

JG

O

o

l-l

T3

>»

xi

<L>

U)

(U

r>

c

o

S

3

p3

O

8

1

(/J

CM

D

o

oz

o

(U o

(U

o
o

'tjb

O

ert

3

^—4

-t->

O
f )

DO

C

1

O

rr!

o

ro

■*->

-^-j

■*->

(L)

p

lU

(N

o

1>

CJ

o

Ci

&

a

^

(1>

a

s

Lh

c«

-*-»

O

VO

>>

■*->

»— ^

§

<L)
C/5

«<

<L)

-«— >

D

1

o 8

s-^

00
O

S3

y-!'

3

o

^o

'BbtN

CO

Z

"fiJOQ

:^Q

O

cJ o G

o 2.2

3cie

■«-•

ci

CO

' O

■t-t

o

a

vo

o

»-H CO.

C/5

■ - o

<D 3 3

8 ^ "^

ci cJ

c4

Q

g

J5

C

a

o

o

O

ci

§1

o

J2

-maltose
-7-acetyl ma
-cellobiose
etyl glucuro

-«-<

3

r)

a

3

8

ci

glue
acid

3

O

■i-t

o

c«

3

o
o

o

acetyl
uronic

o

'S

u
o

-1

S

M CM <N CJ

L^

sZ O

o

0.5

o

3
C/3

H^

3

o

:zi:z;zQ

oLO

o

a

* *

oo ro *

<N

00

*

ON -^ ''t m * .

1 ^

r--

Tf *

O '— 1 O m ^H

o

o

vo O

O O O '.^ O

O

o

O (N

O O O O «A>

o o

+ + + + I I I + + + +

o

35 «

^ o

C/5

d (J

o

top 3

ci —•

o .3:

^9

s<

QQ

SQ

S^

o

4>

■rt

c«

O

CO

y

•*-*

a

O

*v

a^

o

^

o o

ci

o

d 3

ni

s

oses

a-Glucohe

Gluco-Z)-g

Gluco-L-g
ccharides

CO
O

O

CO

O
C

2

Maltose

Trehalose

llobioset

o

CO

o

vert sugar
[uimolar gl

mixture
Lccharides

&c^Q

o

■t-l
f )

C^ CO

3

3 ' ' W C3 1— <

3 W

C\J

CO

e5

ffi

O

Q

3 o
« J;? ^

CO ci 3

D CO
35 O

g O I- o c S^ '^

O2oci3 55>^o-^

.ti O r3 'So 60 S 'C C>

ii^ ^ 5^ ^-S >>-£

>^ 3

142 THE PHYSIOLOGY OF INSECT SENSES

chemoreceptor activity by behavioural techniques. It was shown, for
example, that a single labellar hair responded to touch, water, certain
sugars, and a wide variety of electrolytes and non-carbohydrate
organic compounds. In a hungry, thirsty fly bending a hair, applying
water, or stimulating with certain sugars eHcited a reflex proboscis
extension. The application of such compounds as salts, acids, alcohols,
aldehydes, ketones, glycols, ethers, etc., to the same hair prevented
extension or elicited retraction if the proboscis was already extended.

By adapting the hair successively to bending, to water, and to carbo-
hydrate, and by noting that removing the tip of the hair prevented
response to chemicals but not to bending, Dethier (1955 a) concluded
that the neuron whose process terminated at the base of the hair was a
mechanoreceptor. He concluded further that of the remaining neurons
one was concerned with acceptance (exclusive of mechanoreception)
while another was concerned with rejection. In other words, one
neuron was conceived of as being sensitive to water and certain
sugars while the other was considered to be sensitive to all of those
compounds that caused rejection. Later behavioural experiments
(Dethier, Evans, and Rhoades, 1956) indicated, however, that re-
jection was in reality not a simple modality because it could be initi-
ated either by stimulation of the rejection receptor or inhibition of the
acceptance receptor. Sugars such as mannose, rhamnose, and sorbose,
which of themselves do not cause rejection, are able- to interfere with
stimulation by certain other sugars that alone are acceptable. For
example, mannose inhibits fructose (but not glucose), rhamnose
inhibits glucose (but not fructose or sucrose), sorbose probably in-
hibits glucose and fructose. Furthermore, certain rejected salts,
notably HgClg and CuClg, render sugar receptors reversibly insensi-
tive.

Still more recently, behavioural investigations into the mechanisms
controlling thirst and water satiation suggested that acceptance was
not a simple modality either and that there must be a water receptor
distinct from the carbohydrate receptor (Evans, 1961 a; Dethier and
Evans, 1961). In female flies at certain stages of the reproductive cycle
acceptance is further broken down into carbohydrate acceptance and
protein acceptance (Dethier, 1961).

Conclusions drawn from behavioural studies were borne out to a
very satisfactory degree when success finally attended attempts to
record electrically from chemoreceptive hairs. After many people had
tried and failed to record action potentials by standard techniques,
Hodgson, Lettvin, and Roeder (1955) and Morita, Doira, Takeda,

CHEMORECEPTION 143

and Kuwabara (1957) independently developed a novel technique. It
consisted of placing a fluid-filled micropipette over the tip of a hair
and employing it both as a recording electrode and as a source of
chemical stimulation. More recently a greatly improved technique was
developed by Morita (1959). It involved puncturing the side wall of the
hair and recording with the pipette electrode from this point. The tip
of the hair was thus left free for stimulation by any kind of material
whether electrolyte or not. The indifferent electrode was inserted into
the crushed head.

Originally the potentials from only two neurons were recognized in
the records obtained from the fly (Hodgson et al. 1955; Hodgson and
Roeder, 1956). One neuron, designated the L (for large spike) fibre by

i«i»i% w i»'

A

>slV^v<^»»W^v^^//M>^wV^>^ft**^^ "1 ° ■ ^

«wvUMi»*«VW»*v«*.^')S«»f^

I -1

C ' ■ ■■*«>■ ■■»■■ Y " ■ ' v^ . ...UV% ^>.. ^ V;^A^»^-«y[*f/^V,»f^*^yv^^^ ^■ .■. ^ .^^^W'^,■^ . . ..«.^..., .w^twvw^. J ^y

\

Fig. 82. Electrical response of a labellar hair ofPhormia to stimulation by
salt, sucrose, and bending. A, intact hair stimulated with 0-01 M NaCl
and motion. B, same hair with tip removed. C, cut hair stimulated by
OOIM NaCl + O-IM i)-fructose. The hair is bent at arrow. D, same
hair responding to motion after adaptation to chemicals. Time, 0-2 sec.
Positive potential at recording electrode is down. (From Wolbarsht
and Dethier, 1958.)

Hodgson and Roeder (1956), clearly responded to salts, while the
other, designated as the S (for small spike) fibre, responded to sugars.
Subsequently, Wolbarsht and Dethier (1958) were able to detect the
spikes of a third neuron (designated M for mechanoreceptor) which
responded only to bending of the hair (Fig. 82). Mellon and Evans
(1961) have now detected spikes from a fourth neuron which responds
to water. The function of the fifth neuron, present in some hairs, is not
known.

Since it is now known that certain hairs contain four distinct re-
ceptors, conclusions drawn from earlier studies of electrophysiological
records must be re-evaluated, specifically the conclusions that: low
concentrations of salt sometimes stimulate both L and S fibres
(Hodgson and Roeder, 1956), that acids and alcohols stimulate the L
fibre (Hodgson and Roeder, 1956), that water stimulates L and 5 fibres

144 THE PHYSIOLOGY OF INSECT SENSES

(Wolbarsht, 1957), that fructose stimulates L and S (Hodgson, 1957;
Wolbarsht, 1957; Evans, 1958), that sucrose stimulates two fibres in
Calliphora vomitoria (Morita, 1959), that bending stimulates L and S
(Hodgson et a!. 1955; but see also Hodgson and Barton Browne,
1960), that the primary receptor cell is responsive to chemical, tactile,
and thermal stimuli within the normal physiological range and hence
is at variance with the usual concept of single specificities of receptor
cells (Hodgson and Roeder, 1956; Hodgson, 1958 a, 1958 b).

Another variable that must be considered is the possible differences
among hairs. Some hairs clearly respond more vigorously to bending
or to water than others. Some hairs in the female fly show discharge of
one fibre when protein such as crystalline haemoglobin or brain-heart
extract is applied, but in other hairs all fibres are silent when protein is
appHed (Dethier, 1961). In Vanessa also all hairs are not equally
sensitive to all compounds.

In the course of recording from chemoreceptive hairs when mixed
stimuli were appHed, all workers have observed an interaction between
activity in the L and S fibres (Hodgson, 1956a, 1957; Morita et al.
1957; Wolbarsht, 1958; Morita and Takeda, 1959; Morita, 1959;
Sturckow, 1959; Takeda, 1961). Hodgson (1957) found that the
presence of 5* impulses is characteristically accompanied by a decrease
in L impulses and conversely that the stimulation of the L fibre is
associated with a decrease of S spikes. Since the inverse relationship
between the frequencies of S and L impulses is not constant, it does
not appear that the facts can be explained simply as a result of partial
depolarization of one receptor unit by electrotonic spread from the
more active adjacent unit. Wolbarsht (1958) also noticed this situation
and decided that the activity of the L fibre is not related to the activity
of the S fibre but depends only on the character of the stimulating
solution. He believed that such properties of the salt as thermo-
dynamic activity and the diffusivity of the salt can account in part for
the change in activity reported by Hodgson. On the other hand,
Takeda (1961) has concluded that the depression of the frequency of
impulses from the sugar receptor of Vanessa in the presence of NaCl
represents a direct inhibitory effect.

At the present time the picture in the blowfly appears to be as
follows: one neuron is specifically sensitive to sugars; one neuron is
specifically sensitive to monovalent salts; the neuron whose process
terminates at the base of the hair is sensitive to mechanical stimula-
tion ; a fourth neuron, whose exact location has not yet been established
histologically, has as its adequate stimulus water.

CHEMORECEPTION 145

The situation in the butterfly Vanessa indica is not yet so clear.
Kuwabara (1951, 1952, 1953) has shown that water, sugar, and some-
times sodium chloride elicit proboscis extension. Both sodium
chloride and quinine inhibit. There is no behavioural response to
acids. Morita et al. (1957) concluded that there were more than three
kinds of impulses from this chemoreceptive hair; one neuron respond-
ed to sodium chloride, one to sugar; probably there were also fibres
responding to quinine and water. A later work (Morita and Takeda,
1959) concluded that there were four sensory neurons responding as
follows: (1) one responding to concentrations of sodium chloride
which were less than M/4; (2) one responding to sodium chloride in
the range IM to M/64; (3) one responding to sodium chloride con-
centrations greater than M/64; (4) one responding to sugar. Presum-
ably there is also a mechanoreceptor. A further analysis of the
situation in this insect has yielded the conclusion that the tarsal hairs
contain a mechanoreceptor, a sugar receptor, and, in some cases, a
sugar-NaCl receptor (Takeda, 1961). In 50 per cent of the hairs tested
there was no response to NaCl. No sensitivity to quinine or acetic
acid could be detected nor were any other kinds of chemoreceptors
found. The only histological evidence bearing on this question is the
description by Eltringham (1933) of the chemoreceptive hairs of
Vanessa atalanta. No statement is made in this work of the number of
neurons associated with each hair.

In Leptinotarsa there are apparently five receptor cells associated
with each hair, but the electrical picture is still unclear. Sodium chlor-
ide evokes impulses designated by Stlirckow (1959) as h and n\
potassium, impulses designated as h' and n\ Stiirckow suggested that
h and h' originate in anion receptors and n and n' in cation receptors.
Water gives isolated low potentials; a sugar receptor is inferred;
bending the hair provokes no clear spikes attributable to a mechano-
receptor, although there is an unruly pattern of activity; alkaloids
associated with plants that are unacceptable as food provoke salvos of
spikes from two cells.

Electrical Events Following Stimulation

It is clear in all cases thus far studied in detail that the hairs subserving
contact chemoreception are compound organs containing a number
of chemoreceptors and usually a mechanoreceptor. In the fly at
least each chemoreceptor is specifically different with regard to the
chemicals to which it is sensitive. In all cases each receptor performs
the dual function of reacting to the stimulus and generating nerve

146 THE PHYSIOLOGY OF INSECT SENSES

impulses. Each of these functions is currently the subject of intensive
study.

The suggestion was made by Dethier (1956) that chemicals de-
polarize the membrane of the distal process of the neuron at the tip of
the hair, that this depolarization travels down the nerve fibre to the
cell body at the base of the hair, and that the action potential is gen-
erated in this region. Studies of the effects of temperature on thresholds
and on frequency of impulse discharge lent some support to this idea.
When the temperature of the tip of the hair, and hence the receptor
surface, was altered in the range 2°-41° C, no change in behavioural
threshold occurred (Dethier, 1956), nor was there any change in the
frequency of impulses (Hodgson, 1956 b; Dethier and Arab, 1958).
When the temperature of the whole preparation, and hence the cell
bodies, was altered, however, marked changes in frequency occurred
(Hodgson and Roeder, 1956; Hodgson, 1956 b). According to
Hodgson (1956 b), the fibres in some hairs increased their frequency
with warming while others decreased with warming.

The most conclusive evidence bearing on the matter came, as might
be expected, from electrophysiological work. Wolbarsht (1958), seek-
ing evidence for a slow potential associated with stimulation, dis-
covered when he placed one electrode on the tip of a hair and the other
in the crushed head or labellum that there was a resting potential of the
order of 60 mV. Strangely enough, however, this remained unaltered
when the hair was stimulated. He concluded that the potential repre-
sented a difference across the basement membrane which separates
the general body cavity and the area containing the neurons, associ-
ated cells, and hypodermis. Later, working exclusively with the
mechanoreceptor component of the chemoreceptive hair (and with
simple mechanoreceptors), he recorded a graded slow potential which
occurred when the hair was stimulated by bending. It was always an
increase in negativity at the recording electrode. It varied directly with
the magnitude of the stimulus and showed no overshoot when return-
ing to the baseline. This is a receptor potential in that it occurs prior
to the initiation of any impulses, varies smoothly, and must attain a
critical level before any impulses are generated. It is also either the
generator potential itself or is associated with it. Evidence was pro-
duced that suggests that the receptor potential originates at a site that
is not invaded by the propagated impulses and that this site is distad
of the site of impulse generation. The geometry of the hair and the
mechanoreceptor indicates that this must be in the distal process of the
cell.

CHEMORECEPTION 147

Morita (1959), recording through the side wall of the chemorecep-
tor hair, detected a slow d.c. potential occurring when the tip of the
hair was stimulated (Fig. 83). When the stimulus was cither sugar or
sodium chloride the potential was negative at the recording point with
reference to the base of the hair; when the stimulus was quinine the
potential was positive. The closer the recording electrode was placed
to the tip of the hair, the greater the negativity when sugar or sodium
chloride was applied. On the basis of these recordings Morita has
concluded that this slow potential is indeed a generator potential

%U^> ""'^'

Fig. 83. A. Response of labellar
chemosensory hair of Calliphora
to 0-25M sucrose before (top)
and after (bottom) crushing. B.
Response of a hair to 0-25M
sucrose (top), 0-25M sucrose
plus 0-05M CaCla (middle), and
0-05M CaCla (bottom). Time,
-^Q- sec. (Redrawn from Morita
and Yamashita, 1959.)

arising at the point of stimulation and initiating impulses at the base
of the hair. He makes no distinction between receptor potential and
generator potential, and the presumption is that they are one and the
same in this hair.

The electrical sign of the impulses arising from stimulation proved
puzzling, and an attack on this problem threw some light on the
question of the site of impulse generation. All workers now agree that
the initial component of the spike is positive (Morita et al, 1957;
Wolbarsht, 1958 ; Hodgson, 1958 b). Wolbarsht (1958) interpreted this
condition as indicating that the distal processes of the chemoreceptive
cells act as somewhat poorly insulated extensions of the recording
pipette into the interior of the cell. The spikes of the mechano-

148 THE PHYSIOLOGY OF INSECT SENSES

receptors are conceived of as being recorded by electrotonic spread up
through the contents of the hair, which would be analogous to an
external electrode opposite a part of the cell membrane that is not
involved in the propagated impulse. This conception is supported by
the results of polarization experiments. Anodal polarization produces
stimulation; cathodal stimulation produces depression, as has been
seen with intracellular electrodes in the eye of Limulus.

Arab (1958) found that behavioural responses were evoked by
cathodal stimulation and would not occur with anodal stimulation.
Wolbarsht (1958) was able to reconcile this apparent discrepancy. He
showed that the behavioural responses observed by Arab were due to
the action of the current on the neurons at the bases of hairs adjacent
to the one on which the electrodes were placed. In this case the polariz-
ing current was extracellular; hence, cathodal stimulation was
excitatory and anodal stimulation, the reverse.

Working with the ^yLucilia caeser and the butterfly Vanessa indica,
Morita and his co-workers attempted to pinpoint the site of spike
generation (Tateda and Morita, 1959; Morita and Takeda, 1959;
Morita, 1959; Morita and Yamashita, 1959 a, 1959 b). First, they
showed that spikes were never recorded from the tip of a hair when it
was severed from its base. They then confirmed the findings of
Wolbarsht (1958) that anodal stimulation elicited impulses and
cathodal current blocked them. They found also, as Wolbarsht had,
that anodal stimulation of one hair blocked activity in the neighbour-
ing hair, but they did not do a similar experiment with cathodal
stimulation as Wolbarsht had.

In an attempt to explain how spikes can be recorded from the tip of
the hair, Morita (1959) cooled first the middle and then the base of the
hair, recorded in each case from the tip, and analysed the shape of the
spike. According to his records, cooHng the middle affected pre-
dominantly the falHng phase of the spike, while cooling at the base
aff'ected both the rising and falHng phases. On the basis of this evidence
he claimed that spikes originate at the base of the hair and fire in both
directions, that is, back along the distal process to the tip, as well as
centripetally along the axon to the central nervous system. If this
condition occurs, it would mean that spikes invade the area of the
generator potential contrary to the situation found in mechano-
receptors.

Morita and Yamashita (1959 b) recorded simultaneously from the
tip and the base and found that the spike at the tip is generally larger
than the spike recorded from a hole in the side wall. They observed that

CHEMORECEPTION 149

the side-wall spike peaked more quickly, and they took this to mean
that the differences in height were not due to a gradient of electrotonic
spread. They also concluded from this result that there is back firing,
and they argued that impulses are recorded extracellularly and not
intracellularly as Wolbarsht (1958) concluded.

Hanson and Wolbarsht (1962) re-examined the situation in
PJwrmia and found that the impulse is indeed conducted some dis-
tance down the dendrite, that distance being a large fraction, if not
the entire length, of the short ones, but only a small fraction of the
length of longer ones. Electrodes located at various places along the
dendrite all showed an initially positive phase of the impulse some-
times followed by a longer negative phase. The following interpreta-
tion was proposed : the action potential is generated at a considerable
distance from the recording electrode and is recorded as a positive
swing. As the action potential invades the dendrite, the active region
comes beneath the recording electrode and appears as a negative
deflection. Simultaneous records from several electrodes placed
along the dendrite show that the active region of the membrane pro-
ceeds from the cell body toward the trip. The record resulting from
an algebraic sum of the two phases shows the positive phase being
abruptly terminated by the longer lasting negative phase. Cocaine,
xylocaine, procaline, and chloral hydrate reversibly depress or abolish
this negative phase with a correspondingly longer lasting positive
phase. Continuous chemical stimulation or injury accentuates and/or
speeds up the onset of the negative phase.

The Sugar Receptor

Since different receptor cells in a contact chemoreceptive hair respond
specifically to different chemicals, the question of mechanism of
stimulation, that is, how the chemicals depolarize the dendrite mem-
brane, should be investigated separately in the various types. The types
are most clearly delineated in the blowfly. In this insect, as already
pointed out, there is one cell specifically sensitive to sugars and one to
monovalent salts. A modestly detailed story of the operation of the
sugar receptor has been constructed from behavioural studies. The
task now presented to electrophysiology is that of proof-reading the
story, filling in the details, and revising where necessary.

Carbohydrates are the most effective, if not the only, stimuli; how-
ever, this cell is highly specific and unequally sensitive to the various
compounds. Behavioural tests in which eighty-two different carbo-
hydrates were applied to the chemoreceptive hair housing the receptor

150 THE PHYSIOLOGY OF INSECT SENSES

revealed a spectrum of activity from complete unresponsiveness to
extreme sensitivity (Dethier, 1955 a). The most effective compounds
are certain pentoses, hexoses, and compound sugars possessing an
a-D-glucopyranoside link. In general, the a-form of a sugar is more
stimulating than the p-form. Among effective pentoses D-arabinose
is more stimulating than L-arabinose. It is clear from these tests that
the structural configuration of a sugar is its most important determi-
nant as an effective stimulus. On the other hand, no sense can be made
of the situation on the assumption that there is one key molecular
structure for stimulation.

The first intimation that there are multiple and different sites of
action on the sugar receptor cell arose from a series of studies in which
mixtures of sugars were found not to be effective at concentrations
predicted from the threshold values of the individual constituents.
There were instances of synergism and of competitive inhibition. The
inhibitory effect of a given sugar varied, however, depending upon
which other sugar it was mixed with at the time. For example,
mannose, a weakly stimulating sugar, inhibited fructose, but not
glucose; rhamnose inhibited glucose but not fructose. There was
some evidence that sorbose inhibited both fructose and glucose
(Dethier et aL, 1956).

Supporting evidence for the idea of multiple sites came from an
entirely different kind of study. Evans (1961 b) found that feeding
blowfly larvae on a medium containing a specific sugar depressed the
sensitivity of the adult to that sugar. Relative sensitivity to glucose
or fructose could be enhanced or depressed by rearing in the presence
of the appropriate sugar. Evans interpreted these results as indicating
that the sugar reduced either the number or the affinity of sites on the
receptor cell for that particular sugar.

The nature of the combining action of sugars, while not understood,
can be narrowed down somewhat by a process of elimination. An
outstanding feature of stimulation by sugars is that no change in
threshold can be demonstrated to take place as the temperature of the
stimulus is changed. Nor is any change in threshold observed when the
pH of the sugar solution is changed (Dethier, 1956). Another feature
of stimulation by sugars is that there is no inhibition by any of the
following metabolic inhibitors: phlorizin, fluoride, azide, idoacetate,
cyanide (Dethier, 1955 a). This finding indicates that the first step in
stimulation probably does not involve any steps in the glycolytic
cycle below those blocked by the compounds listed. The absence of a
marked temperature and pH change argues against an enzymatic

CHEMORECEPTION 151

reaction being involved. It was proposed tentatively (Dethier, 1956)
that stimulation by sugars involves combination of the sugar mole-
cules with a specific receptor site or substance by weak forces, such as
van der Waal, to form a complex which depolarizes the membrane,
after which (or simultaneously) sugar is removed passively by a shift
in concentration gradient.

At this point current knowledge of the sugar receptor ends. Few
electrophysiological investigations of stimulation by carbohydrates
have been reported thus far. Hodgson (1957) examined the effects of
twenty-three carbohydrates. At that time the existence of a water
fibre was not suspected, and the lack of this information complicated
the interpretation of records. It appears, nevertheless, that all accept-
able sugars stimulate one neuron, while unacceptable sugars either
fail to stimulate or actually inhibit.

In Vanessa indica the sugar receptor responds to i)-glucose, D-
fructose, and sucrose in increasing order of effectiveness. Lactose,
Z)-galactose, inositol, trehalose, and L-sorbitol are non-stimulating
(Takeda, 1961). Maltose and turanose elicit small deformed spikes
which have been interpreted by Takeda as indicative of a combined
stimulatory and inhibitory action. The sugar receptor is sensitive to
concentrations of sucrose as low as 2-'^ M. In addition to this receptor
there is one that responds in identical fashion to sucrose and to
NaCl.

The Salt Receptor

It was originally believed that rejection of a solution was triggered
solely by activity in one neuron, and since so many compounds elicited
rejection, it was assumed that the single neuron was grossly non-
specific. Studies on competitive inhibition and on the action of heavy
metal salts, however, hinted that rejection was not a single modality.
As a consequence, it was no longer possible to assume that every
compound that was rejected acted on the same receptor cell.

Electrophysiological studies, while only just beginning, have par-
tially answered the question of the adequacy of stimuH for rejection.
They indicate that many rejected compounds block activity in all fibres
of the chemoreceptive hairs, but that monovalent salts stimulate one
cell actively (Hodgson, 1956 a; Morita, 1959). It has been possible to
come to some understanding of the mechanism of action of this salt
fibre by acquiring enough data to make the same kind of theoretical
analysis which Beidler (1954) made of mammalian salt receptors
(Evans and Mellon, 1962 b).

L

152 THE PHYSIOLOGY OF INSECT SENSES

Beidler assumed that the stimulus reacts with some receptor sub-
stance and that the reaction obeys the mass action law. His second
assumption was that the magnitude of response is directly related to the
number of ions or molecules that have reacted with the receptors. He
derived an equation which related magnitude of response to the con-
centration of the applied chemical stimulus. This is clR = (clRm)-\-
iXjKRm), where c = concentration, R — magnitude of response,
Rm = magnitude of maximum response, and K = the equilibrium
constant. The vahdity of the equation can be tested by measuring c
and R. Beidler found the agreement to be excellent. Now, knowing the
equilibrium constant, he calculated the free energies of reaction from
the expression AF = RT\n K, where AF = change of free energy,
R = the gas constant, and T = the absolute temperature. The low
value which he found for AF(from —1-22 to —1-37 Cal./mole) for a
series of sodium salts was taken to indicate that physical rather than
chemical forces are involved in the interaction between the chemical
stimulus and the receptors. This conclusion raised the question of what
the receptor substance may be.

Beidler argued that the small temperature dependence and the
low values for AF suggest a reaction similar to those that occur with
ion binding by proteins and natural polyelectrolytes. Some properties
of the reacting groups of the receptor substance can be determined
from a study of the effect of changing pH. Since no effect could be
demonstrated over the range 3-0-1 1-0, one might conclude that the
molecules of the receptor substance are strong acidic radicals. The
relatively weak carboxyl radical of a protein, for example, cannot be
considered as the reacting group. The phosphate and sulphate radicals
of such natural polyelectrolytes as nucleic acids and certain poly-
saccharides are able to bind cations in a manner consistent with the
properties of the receptors described by Beidler.

The many complexities of the mammalian taste organs, as, for
example, the receptor cell being non-neural and synaptically con-
nected to a neuron, the multiple innervation of a receptor cell, and the
innervation of more than one cell by a single neuron, obstructed
further analyses of the nature of the receptor sites and the process of
stimulation. None of these handicaps are associated with the salt
receptor of the insect. A quantitative study of the repetitive response
of the labellar hair receptor to NaCl has shown that Beidler's theory
applies equally well to this receptor (Evans and Mellon, 1962 b). The
calculated relative free energy change of the reaction between salt and
receptor site in this case is the range to — 1 kcal./mole. This value

CHEMORECEPTION 153

suggests that weak physical forces are involved, and evidence has been
obtained that the salt-combining sites are anionic and strongly acidic.
As a consequence, the cation of a salt largely dominates stimulation.

The Water Receptor

The water receptor discovered by Evansand Mellon (1962 a) is the only
cell in the chemoreceptive hair of P/wrmia that is activated by water.
The most effective stimulus tested thus far is pure water. Aqueous
solutions of sucrose appear to depress the activity of the cell as a linear
function of the logarithm of osmotic pressure. At 5M sucrose the
frequency of impulses decreases to less than half of the maximum.
Other non-electrolytes (e.g., glycerol and mannose) also inhibit, but
no simple relation was found between parameters of the solution and
the degree of inhibition. Inorganic electrolytes also inhibit water res-
ponse, but in a more specific manner. Aqueous solutions of NaCl
begin to inhibit at 0-1 M. Inhibition appears to be complete at about
0-3M. Calcium chloride solutions inhibit at about 0-01 M. The
inhibition of water response is unrelated to spike activity in other
fibres.

Modalities

For the blowfly it is clear that water, sugars, protein, and unacceptable
compounds can be distinguished as different. To what extent these
four taste categories can be subdivided into other taste modalities is
not clear. Although there is evidence that different sugars act upon
different sites at the molecular level, they none the less stimulate the
same receptor. In the case of unacceptable compounds, that is, salts,
acids, alkaloids, alcohols, etc., it is probable that different sensory in-
put results from stimulation by the different compounds. If this
suggestion is confirmed it provides a peripheral mechanism for dis-
criminating among a number of compounds and indicates that the
modality 'unacceptable' is not a homogeneous one.

Aside from incomplete electrophysiological data obtained with
Phormia, evidence for taste modalities has been derived principally
from the following sources: (1) observations of the behaviour of indi-
viduals conditioned to respond to a given chemical ; (2) determinations
of the additive or non-additive capacity of diverse stimuli ; (3) measure-
ment of the amount of sucrose drunk by honeybees after contamina-
tion with subliminal concentrations of different chemicals ; (4) com-
parison of the threshold changes for different compounds during
starvation; (5) localization of specific kinds of receptors. Much of the

154 THE PHYSIOLOGY OF INSECT SENSES

information derived from these kinds of experiments is suggestive
rather than conclusive ; nevertheless, considering all of the available
evidence, it is highly probable that insects can do better than dis-
criminating simply between acceptable and unacceptable.

Bauer (1938) trained Dytiscus marginalis and Hydrous piceus to res-
pond positively to sucrose; these animals could not be trained simul-
taneously to avoid glucose, and also reacted positively in the majority
of cases to some fifteen or twenty other sugars and sugar derivatives.
As in the experiments of Schaller (1926) and Ritter (1936) with the
same or related species, the beetles learned readily to distinguish be-
tween pairs chosen from sucrose, sodium chloride, acids (hydrochloric
and acetic), and quinine, and could even be trained to accept quinine
and avoid sodium chloride. But specimens that had learned to avoid
hydrochloric acid also avoided acetic and could not be taught to
respond differently to the two; the same result was obtained when
quinine was matched against salicin or aloin. Bauer concluded that the
sweet substances (with the possible exception of mannose, which was
avoided by some individuals) constitute a single homogeneous
grouping; salts, acids, and bitter substances, normally avoided, are
distinguished from sweet and from each other, so that taste sub-
stances can be classified into the same four quahties for these beetles
as for man.

As mentioned above in the discussion of methods, von Frisch (1935)
found additive the stimulatory effects of all sugars acceptable to the
bee. Summation was noted also between sodium chloride and lithium
bromide, ammonium bromide, or hydrochloric acid, but the repellent
effect of quinine was lessened rather than enhanced by the addition of
acids (hydrochloric, acetic, sulphuric, citric, lactic). When sucrose
solutions, one containing sodium chloride and another quinine hydro-
chloride, were prepared and were accepted by equal proportions of
the bees they were drunk in different amounts. Almost as much of the
quinine was taken as of the control, but considerably less of the
solution containing salt. The results would indicate three taste
qualities: sweet, acid-salt, and bitter. But in other experiments von
Frisch found that starved bees show no better acceptance of sodium
chloride than those fully fed, although the threshold for rejection of
quinine rises eight times and that for hydrochloric acid by a factor of
five. He concluded therefore that salt represents a quality different
from either acid or bitter. Furthermore, although there is no summa-
tion of repellency between quinine and acids, other bitter substances,
such as aloin, arbutin, colocynthin, and salicin, are rendered more

CHEMORECEPTION 155

repugnant by the addition of hydrochloric acid, so that the category
bitter is not homogeneous for the bee. It is possible that investigations
of this kind would reveal similarly complex relationships in man.

Other evidence as to the separation of the taste qualities in insects
was provided by Ritter ( 1 936), who found that after amputation of the
maxillary palpi specimens of Hydrous piceus would still react to
0'007M hydrochloric acid but not to salt, sugar, or quinine. Removal
of the tips of the labial palpi then abolished the response to acid,
although animals lacking both sets of palpi and the antennae still
reacted positively to meat juice, presumably via receptors in the mouth.
These observations would seem to set the acids apart from the other
taste substances, but are at variance with the findings of Bauer (1938)
with the same and related species, and of Hodgson (1951) with
Laccophihis. Bauer found that amputation of the maxillary palpi left
the gustatory response to sucrose, etc., unaltered except for an increase
in threshold. Hodgson found by ablation experiments that all the head
appendages could mediate responses to acids, salts, and alcohols.
Thresholds were increased considerably when the antennae were re-
moved, but less so when either pair of palpi were extirpated.

CHAPTER VI

Response to Humidity

The behavioural responses of insects to water vapour mixed with the
permanent gases of the atmosphere indicate that these organisms
somehow perceive the water or differences resulting from changes in
the concentration of the vapour. A unique feature of water is its pres-
ence both in the organism and in the environment. The existence of
water on both sides of the integument poses problems that are not
encountered in the case of other stimuli. Conceivably water may act on
the organism independently of what is within (as do, for example,
odour molecules), or it may cause outward movement of internal water
so that the insect does not respond directly to a change in external
environment but rather to a change induced in its own internal
environment. For these reasons, and because all attempts to under-
stand the mechanism of hygroreception have been based upon experi-
ments which correlate the degree of activity or direction of movement
with the amount of water in the air, decisions must be reached as to
which relationship of water vapour to atmosphere gases is to be meas-
ured. If humidity receptors are conceived of as responding to the con-
centration of water molecules, then the absolute concentration of
water to which they are exposed should be measured. On the other
hand, if the humidity reactions are conceived of as operating through
the agency of evaporation or transpiration, then the drying power
of the air is the variable to be measured.

Conventionally there are a number of humidity definitions (Humph-
reys, 1 940). The absolute humidity is either themass of water vapour per
unit volume or the gas pressure exerted by the water vapour per unit
area. Relative humidity is the ratio of the actual mass of water vapour
in a small volume to the maximum mass that can exist in the same
volume at the same temperature. It is also defined as the ratio of the
actual to the maximum pressure of water vapour per unit area that can
exist in the presence of a flat surface of pure water at the same temper-
ature. Saturation deficit may be defined as: (1) the amount of water
vapour in addition to that already present, per unit volume, necessary
to produce saturation at the existing temperature and pressure;
(2) the difference between actual and saturation pressure; (3) the ratio

156

RESPONSE TO HUMIDITY 157

of the vapour pressure deficit to the saturation pressure at the existing
temperature.

Absolute humidity in mass per volume is difficult to measure. Con-
ventionally one measures relative humidity. In an attempt to measure
the drying power of the air many investigators measure saturation
deficit (Anderson, 1936). Leighly (1937) and Thornthwaite (1940)
have pointed out that evaporation is a physical process dependent
upon a vapour pressure gradient between the evaporating surface and
the air and is not directly related to relative humidity or to saturation
deficit in the air. Hence, saturation deficit is of no value as an indicator
of moisture losses by evaporation or transpiration (see also Ramsay,
1935; WeUington, 1949; Edney, 1957).

Responses to water vapour include: orientation from a distance to
high concentrations of water vapour, proboscis responses to water
vapour in the immediate vicinity, avoidance of regions of high and low
humidities, and aggregation in zones of preferred humidity.

Just as the degree of hunger or satiation influences the response of
insects to food, so also does the state of water balance (and starvation)
influence the behavioural response to humidity. Desiccated flour
beetles {Tribolium castaneum) reverse their preference for humid areas
after water has been provided. The intensity of the reaction to dryness
also increases with increased starvation (WiUis and Roth, 1950).
Similarly, the cockroach Blatta orientalis, which prefers lower
humidities when its water balance is optimum, becomes hygropositive
when desiccated (Gunn and Cosway, 1938). Comparable reversals
have been demonstrated for the beetle Ptinus tectus (Bentley, 1944),
Drosophila (Perttunen and Erkkila, 1952), Aedes, and Tenebrio molitor
(Dodds and Ewer, 1952). It is clear from these and other studies not
only that a water deficit in the body elicits a moist reaction but also that
the presence of a maximum amountof water in the body elicits a dry
reaction (Perttunen, 1951; Syrjamaki, 1962). Other factors that
influence response to humidity are age (Perttunen and Ahonen, 1956;
Perttunen and Salmi, 1956; Syrjamaki, 1962), diurnal rhythms
(Perttunen, 1953), stage of the reproductive cycle (Perttunen, 1955 a),
stage of development (Hafez, 1950, 1953), sex (Roth and WiUis,
1951 c; Perttunen and Ahonen, 1956; Syrjamaki, 1962).

Most of our speculation regarding the physiology of hygroreceptors
is based upon analyses of behavioural responses to humidity gradients.
The following species have been studied in some detail: Locusta
migratoria migratorioides (Kennedy, 1937), Blatta orientalis (Gunn
and Cosway, 1938), Culexfatigans (Thomson, 1938), Tenebrio molitor

158 THE PHYSIOLOGY OF INSECT SENSES

(Pielou, 1940; Pielou and Gunn, 1940), Pediculus humanus corporis
(Wiggles worth, 1941), Agriotes osbcurus and A. lineatus (Lees, 1943),
Ptinus tectus (Bentley, 1944), Choristoneura fumiferana (Wellington,
1949), Tribolium spp. and Sitophilus (Willis and Roth, 1950; Roth and
Willis, 1951a, 1951b, 1951 c), Aedes gracilis and Blethisa multipunctata
(Perttunen, 1951), Blatella germanica and Aedes aegypti (Roth and
Willis, 1952), Drosophila (Perttunen and Syrjamaki, 1958; Perttunen
and Erkkila, 1952; Syrjamaki, 1962), Neodiprion americanus bank-
sianae and A^. lecontei (Green, 1954), Oncopeltus fasciatus (Andersen
and Ball, 1959), Melanoplus binttatus (Riegert, 1958, 1959, 1960),
Schistocerca gregaria (Aziz, 1957 a, 1957 b). The strength of reaction
and the ability to discriminate differences in humidity varies among
the species. With Locusta migratoria migratorioides, which aggregates
in dry microclimates, more individuals respond as the gradient be-
comes steeper. With Culex fatigans, Tenebrio molitor, which avoid
high humidities, and Agriotes spp., which avoid low humidities, the in-
tensity of the reaction is related to the highest humidity. Ptinus tectus
reacts most intensely at low humidities. Culex discriminates 1 per cent
differences in the range near 100 per cent relative humidity, but reacts
only to 50 per cent differences in the range of relative humidity from
30 to 85. rr/Z>6>//wm c<3^/<3«ewm discriminates between humidities differ-
ing by 15 per cent over the entire range. Reactions to differences of
5 per cent are weak between R.H. and 5, absent between R.H. 40 and
45, 45 and 50, and 50 and 55, but strong between R.H. 95 and 100. The
louse also discriminates better at higher humidities. Wireworms detect
the difference between R.H. 100 and 95-5. Normal undesiccated
Drosophila melanogaster of both sexes exhibit an intensity of reaction
that is correlated with the degree of the higher alternative rather than
with the difference. When alternate humidities lie between 100 and
87 per cent R.H., the drier is preferred ; when they He between 77 and
20 per cent the wetter is preferred (Perttunen and Syrjamaki, 1958).

The reactions of some species to a given relative humidity have been
studied at different temperatures. The results have been interpreted as
indicating that the reactions of Culex and Tenebrio are more in accord
with relative humidity than with saturation deficiencies, while the
reverse is true of Agriotes. What this may mean is difficult to know.
Although the results are interpreted by some as indicating that the
receptors act as evaporimeters, the lack of correlation between satur-
ation deficit and evaporation weakens the argument (cf. Thorn thwaite,
1940). On the other hand, in studies conducted with spruce budworm
larvae (C. fumiferana) and sawfly larvae of the genus Neodiprion a

RESPONSE TO TTUMIDITY

159

ri

O

l-r

•*->

(U

E

t-i

o

a

OS

>

(L>

c

o

■t-t

o

c;

3

Uh

O

•«-•

TD

CO

(L>

(/i

O

00

O

a

C

ex

D

C

CO

C/l

OJ

Si

p

ci

o

1>

c^

>

r/O

03

Si

03

a

VO

<u o

as

0)

Vh cc

■^m^

^

Vh '—>

o r;

r«

+->

00

Id

8

s

B

^

-4->

O. C/D

pa
<

-.a

C/5

O

rece]
eptor

F=,

O

-(— >

il y

o

o.

^ ^

<L»
O

Si

2? "

>1

o3
O

•d

fli

§ s

3

o

^

vi

Uh

bO

o

>^

^

CO

a

••-»

o

C/3

o

c

a

o

O

</i

o

a

3

u>

O

■«-•

a
8

<u

1-1

^

<5J

En

OO

0^

C/5
O

u

c
c

3

o

OS

Cj

0\

a

S2

ON OS

o CiO

-i:^ (50

On m

'— ' ON

§ •£ •£
■ o o

to

ON

03

-7:: C^ ON

•o

ON

T3

-•-> -t->
o o

N

c
m

too rt
to >

- ON ^

3 ^ ^•~

O nT CO

S N =

On

>o

r~-

ON

»o

ON

— ^

cr

^BM

>-.

(A

3

Uh

r-

3

c

PQ <

On ._"
5^

s^j
^

^

to

•Si

+ - I 1 +

! 1 +

+ :S - + I + + + - - ^ - + 1 + - + - H-

■3 :=:

•Si

VJ

•S

k.

^

-0

•2
o

s:

s:

S* o3

"t^ >

to V-

^ 03

•2 «^

S K "^

o, ft^ ^ a.

> .^
fee

tn

C/3

a

CO

■§

§0

<u .5:
X) ^

(4-1 ^

O 3

^ 03

CO s:

O to

& u
■^ -S

JD <)

oj
>

03

•^ .^ .1:2 'ii -^

t^o fee ~

5i> «il O

tn 03 .CJ

^ ^ 6)
^ ^ ^

05

01

S -S -^

bo

S ^

C ,^ 2

s ^ ^

S C: "2

=0 ?|

§"-^

§:2

to
J3

160 THE PHYSIOLOGY OF INSECT SENSES

relation between locomotory behaviour and rate of evaporation has
clearly been demonstrated (Wellington, 1949; Green, 1954). No in-
formation is available with regard to the receptors involved. It is
possible that the response is mediated by rather general changes in the
internal milieu rather than by specific sense organs.

The identity of humidity receptors is even less firmly established
than that of olfactory receptors. In Tenebrio responses to humidity are
aboHshed when the antennae are extirpated or completely covered.
The most common sensilla are pits and pegs, the latter being confined
to the seven distal segments. The intensity of response is greatly de-
creased when the seven distal segments are removed, but complete
abolition requires removal of the additional segments which bear only
pits. Asymmetrical amputation has shown that a threshold number of
pits is required (Pielou, 1940). In Tribolium but not in Sitophiius,
similar ablation experiments have shown that there is a correlation
between the number of thin-walled sensilla and the intensity of res-
ponse (Roth and WilHs, 1951b). The sensilla are simple or multiply
branched pegs. On the antennae of the human louse there are humidity
receptors which are distinct from the olfactory receptors (Wiggles-
worth, 1941). They are tufted sensilla; each consists of a minute cone
bearing four delicate apical hairs. A number of neurons are associated
with each sensillum. Similar sensilla are the humidity receptors of the
larva of Musca domestica (Hafez, 1950, 1953) and of Drosophila
melanogaster (Benz, 1956). In grasshoppers the sensilla coeloconica
have been suggested as humidity receptors (Aziz, 1957 b; Riegert,
1960), but there is no supporting experimental evidence (Slifer, 1961).
By studying antennaless mutants of Drosophila Begg and Hogben
(1946) came to the conclusion that the antennal receptors mediated
positive responses to wet and that responses to dry were mediated by
receptors at other unknown locations on the body. Perttunen and
Syrjamaki (1958) came to exactly the opposite conclusions and pro-
posed an explanation of the discrepancy. Kuwabara and Takeda
(1956) conditioned honeybees to extend the proboscis when water
vapour was held near the antennae and then amputated various lengths
of antennae. They concluded that the sensilla ampullacea were most
probably the hygroceptors. Bursell (1957) found that the normal
orthokinetic response of the tsetse fly Glossina morsitans was abolished
when the branched hairs guarding the thoracic spiracles were
removed.

How the hydroreceptors operate is still a mystery. They could res-
pond directly to water molecules striking the surface; they could

RESPONSE TO HUMIDITY 161

permit evaporation of internal water, with a consequent change in
chemical composition or osmotic pressure in the milieu of the recept-
ors or by mechanical deformation ; they could respond to temperature
changes due to evaporation; they could act as hygrometers, i.e.,
structures with special hygroscopic properties (Pielou, 1940). In some
instances (e.g., spruce budworm) responses may not be mediated by
special receptors at all but rather by means of general changes in the
body fluids occasioned by evaporation through non-sensory areas.
For a recent discussion of these matters, the paper of Syrjamaki (1 962)
should be consulted.

CHAPTER VII

Photoreception

The range of radiations in the electromagnetic spectrum extends from
about 10^2 to about 10~^ m,a. Only a very small segment of this spec-
trum, from about 253 to 700 m[i, can be detected by organisms. The
prime requisite for receiving the radiant energy of this narrow band of
the spectrum is a pigment or pigments that will absorb in these wave-
lengths. For the organism to derive useful information from this
energy absorption, bioelectric potentials must be produced. In insects,
cells capable of accomplishing these ends are grouped together in three
organs: the compound eyes, the dorsal ocelli, the lateral ocelH. Of
these the compound eyes, by virtue of their predominant role in be-
haviour and their extraordinarily large complement of cells, are by far
the most important. These great sensitive spots on the surface of the
head, with no lids to shield them, are in a continuous state of activity
from the light impinging upon them during the lifetime of the insect.

THE COMPOUND EYE
Structure

The principal components of the compound eye are the monopolar
neurons that absorb radiant energy and generate action potentials and
the transparent areas of the cuticle and internal lenses overlying them.
Since cuticle overlies all of the body surface, transparency, or at least
translucency, is a sine qua non for the functioning of the photo-
receptors. Further modification of the cuticle, however, in respect to
shape has given it the added property of gathering light. Together with
clear cellular bodies, the crystalline cones, it constitutes a lens system,
the dioptric apparatus of the eye.

The compound eye is a collection of sensilla known as ommatidia
(Fig. 6). Each ommatidium is constructed of a corneal lens, a crystal-
line cone, a number of primary neurons, and enveloping cells. In
primitive insects (e.g., Machilis, one of the Thysanura) the two cells
which secreted the cornea lie immediately beneath it, that is, between
the cornea and the crystalline cone. In most insects these cells with-
draw to a position on either side of the crystaUine cone where, having
become pigmented, they are called corneal pigment cells.

162

PHOTORECEPTION 163

Beneath the corneal lens lies the crystalline cone. There are three
types of cones. In one, four cells undergo changes in which a fused
intracellular body consisting of protoplasmic ground substance and
glycogen is formed (Fyg, 1960). The nuclei are located at the outer or
basal border of the cone (eucone type). In another type the cone cells
secrete a transparent material while themselves remaining distinct
(pseudocone type). In still other eyes no vitreous body is formed, the
cone cells being the only occupants of the area (acone type).

The primitive number of retinula cells in an ommatidium is prob-
ably eight (Snodgrass, 1926), but one is usually rudimentary or
eccentrically placed. The arrangement of the remaining seven varies
greatly from one species to another. Sometimes one occupies a basal
position so that only six contribute to the formation of the rhabdom.
Sometimes the cells are clearly arranged in two layers, three proximal
and four distal. More commonly, all cells are in a single layer around
a central axis.

The sensory cells appear like bipolar neurons early in their develop-
ment (Fyg, 1960), but when fully developed are unipolar, dendrites
being absent. In the place of dendrites the surfaces of the cells facing
the central axis are modified into peculiar structures, the rhabdomeres
(Fig. 6). These are the receptive surfaces of the cells (Schultze, 1868;
Grenacher, 1879).

Together the rhabdomeres of all retinal cells form a rod, the
rhabdom. Some workers had considered the rhabdom to be cuticular
(Machatschke, 1936), but the specificity of the test upon which this
claim is based is open to question and other tests for chitin have given
negative results (Richards, 1951). Other students of the eye had con-
sidered the rhabdomeres to be modified neurofibrillae traversing the
interior of the cell. The structure of the rhabdom as revealed by
electron-microscopy is open to different interpretations. Information
is now available on Musca domestica. Apis mellifera, the phalaenid
moth Erebus odora, a species of skipper Epargyreus, two grasshoppers,
Dissosteira spp. and Schistocerca spp. (Fernandez-Moran, 1956,
1958), Drosophila melanogaster (Fernandez-Moran, 1956, 1958;
Wolken, Capenos, and Turano, 1957), Sarcophaga bullata, and
Anax Junius (Goldsmith and Philpott, 1957).

The ommatidia of Musca, Sarcophaga, Drosophila, and Apis are
similar. Each is of the apposition type. Of the eight retinal cells only
seven contribute to the rhabdom, the eighth being shorter. In cross-
section each cell is roughly wedge-shaped. The rhabdomere appears as
a rod or beading extending the length of the medial surface (Fig. 87).

164 THE PHYSIOLOGY OF INSECT SENSES

It is 60-70 [I long and about 1 -5 [x in diameter (Mused). A longitudinal
section of a rhabdomere reveals that it consists of closely packed
parallel tubules aligned at right angles to the long axis of the cell
(Fig. 84). Each is about 400 A. in diameter and 1,000-15,000 A. long.
The most reasonable interpretation of these tubules is that they are
microvilli, that is, finger-like evaginations of the retinal cell; conse-
quently, their boundaries are part of the cell membrane and their
contents are extensions of the cytoplasm (Miller, 1957; Goldsmith and
Philpott, 1957; Fernandez-Moran, 1958; Slifer, 1961). A fine parti-
culate component appears in the microvilli in fixed material. Except
for the small one, the rhabdomeres appear to occur as pairs, one

Fig. 84. A. Distal tip of retinal cell
from the compound eye of
Musca domestica as recon-
structed by Slifer (1961) from
electron-micrographs of Fer-
nandez-Moran (1958). B.
Cross-section of cell with rhab-
domer at right. C. Section
through that part of rhabdomer
containing microvilli. (Re-
drawn from Slifer, 1961.)

member of which lies opposite its twin on the far side of the ommati-
dium. The longitudinal axes of the microvilli in both members of a
pair are parallel but in a different plane from those of other pairs. The
core separating the rhabdomeres is a loose network of fine filaments
in fixed materials. In the eye of the grasshoppers Dissosteira and
Schistocerca the fine filaments in the matrix are closely packed and run
lengthwise in orderly courses. Distally this core disappears as the
rhabdomeres come together just below the crystalline cone. Here the
microvillar structure is replaced by an electron-dense structure. In the
eye of the dragonfly Anax Junius and in the superposition eye of the
moth Erebus odora there is no central matrix, since the rhabdomeres
meet axially. The rhabdom of the skipper consists of eight fused
rhabdomeres. Most workers believe that the visual pigment is located

PHOTORECEPTION 165

in the rhabdomeres, that is, in the lumina of the microviUi. In Erebus
there is an extensive underlying structure, the tapetum, composed of
tracheoles which are highly reflective (Fernandez-Moran, 1958).

The retinal cells may occupy one of the two positions with respect to
the crystalline cone (Fig. 6). In apposition eyes the retinal cells lie
immediately beneath the crystalline cone; in superposition eyes the
retinal cells are situated some distance proximad of the crystalline
cone. The intervening space is transparent. In all eyes each ommati-
dium is enclosed in pigment cells. The number of pigment cells is not
constant, but usually there is a distal set (iris pigment cells) surround-
ing the crystalline cone and another set (retinal pigment cells)
surrounding the retinal cells.

At the base of the ommatidia, and marking the proximal limit of the
retina, is a fenestrated basement membrane. Through the fenestrae
pass the proximal processes, the axons, of the retinal cells. There is
usually one axon for each cell. In Erebus, however, there are sometimes
nine axons from a single ommatidium. This occurrence indicates that
sometimes the axons begin to branch close to the cell body. Some
synapse almost immediately with neurons of the optic lobe; others ex-
tend a considerable distance into the lobe before synapsing. The
presence of two kinds of retinal fibres, long and short, has been known
for some time. Hanstrom (1927) had speculated upon the possible
physiological differences between the two and likened them to the
rods and cones of the vertebrate eye.

The optic lobes are in fact a part of the brain, but since so many
physiological studies of the eye treat them as part of the eye rather
than as the central nervous system, a brief description at this point will
be helpful. Most of our knowledge of these complicated structures
stems from the classical studies of Zawarzin (1914) on the larva of the
dragonfly Aeschna, Cajal (1909, 191,8) and Cajal and Sanchez (1915)
on the house fly and the blowfly. In addition, there are the accounts of
Giinther (1912) and Holste (1923) on Dytiscus, Bretschneider (1921)
on the moth Deilephila euphorbiae, and Power (1943) on Drosophila.
The general works of Hanstrom (1928) and Snodgrass (1935) should
also be consulted.

The optic lobe (Fig. 85) is divided into three areas: the lamina
ganglionaris, usually in immediate contact with the basement mem-
brane of the retina; the medulla externa; the medulla interna.

The lamina ganglionaris is the first or most distal synaptic layer of
the optic lobe. It consists of giant monopolar cells, small monopolar
cells, short retinal fibres, long retinal fibres, and centrifugal fibres.

166

THE PHYSIOLOGY OF INSECT SENSES

Fig. 85. Simplified diagram of the neural pathways within the optic lobe of
Calliphora. R, retina; I, lamina ganglionaris ; II, medulla externa;
III, medulla interna; OC, outer chiasma; IC, inner chiasma. (Redrawn
from Cajal and Sanchez, 1915.)

After the axons of the retinal cells traverse the basement membrane
about one from each ommatidium passes directly to the lamina,
while its five or six companions turn immediately at right angles,
forming local chiasmata with those of other ommatidia. The axons
then enter the lamina in groups or bundles, each bundle consisting of

PHOTORECEPTION 167

axons from several ommatidia. Each bundle in the lamina becomes
associated with the axon of a giant monopolar cell, with which most
of the retinal fibres synapse, and the axon of the centrifugal cell whose
cell body lies in the medulla externa. The units ('optic cartridges' of
Power) in the lamina thus consist of a long retinal fibre, a number of
short retinal fibres, a giant monopolar fibre, and a centrifugal fibre.
These fibres now enter the outer chiasma. Fibres from the anterior
ommatidia cross to the posterior region of the medulla externa;
fibres from the posterior ommatidia cross to the anterior area of the
medulla externa; fibres from the middle ommatidia pass directly
through the chiasma without crossing. All fibres from the lamina
ganglionaris end in the medulla externa. At the present state of our
knowledge of the eye there is little point in pursuing the complexities
of the optic lobe beyond this level.

The Function of the Retinal Cells

It is generally assumed that the photosensitive pigment is located in
the retina cells. As yet there is no direct evidence for this. Prior to 1957
all attempts to isolate a visual pigment failed (Goodwin and Srisukh,
1949; Wald and Burg, 1957; Wolken, 1957; Wolken, Mellon, and
Contis, 1957), but Goldsmith (1958 a) was finally able to isolate
retinene from the heads of honeybees. It had the specific absorption
peak 664 mii. Since no retinene was obtained from parts of the body
other than the head and since retinene previously has been found only
in eye tissues, Goldsmith concluded that it is the chromophore of a
visual pigment located in the compound eyes and/or ocelli. He found a
pigment that possesses a wavelength maximum at about 440 mu.. On
bleaching in light the maximum moves to 370 m[i, probably represent-
ing the formation of retinene (Goldsmith, 1958 b). Spectral sensitivity
measurements of the eyes of honeybees reveal several peaks at diff*er-
ent wavelengths, one at 440 mjx (compound eye of drone). This is the
only light-sensitive pigment containing retinene that has been ex-
tracted from insect eyes thus far. Bowness and Wolken (1959) isolated
from the house fly a yellow pigment with a spectral absorption maxi-
mum at 437 mil when unbleached and at 440-446 and 350-360 my.
when bleached, but neither vitamin A nor retinene was detected in this
pigment. The work on colour vision suggests that there are others with
maxima at different wave lengths.

M

168 THE PHYSIOLOGY OF INSECT SENSES

Electrical Events

An electrophysiological attack on the action of retinal cells has been
pushed with more vigour than the biochemical, but, for reasons that
will become apparent, the results so far have a high degree of
ambiguity.

It has been known for nearly a century that retinae give an electrical
response to light. This response, the electroretinogram, is usually
detected by placing one electrode on the surface of the cornea and the
other at any point on the body. Considering the neural complexity of
the eye, this is a relatively crude measurement to make. It should be
apparent that the electroretinogram (ERG) is a mass effect dependent,
as Granit (1955) has pointed out, upon the favourable orientation of
certain retinal structures which conduct currents in one direction.
Furthermore, the second electrode, the so-called indifferent electrode,
is not necessarily indifferent (cf. Granit, 1955).

For these reasons and also because the profile of the potential
changes recorded from the eye differs in form and magnitude with the
location of the electrodes, the intensity of illumination, the duration of
illumination, the degree of Hght or dark adaptation, the time of day
(Jahn and Wulfif, 1941a, 1941b, 1943), the temperature (Ishikawa and
Hirao, 1960 c), the age of the preparation (Hassenstein, 1957; Naka
and Kuwabara, 1959), the amount of blood lost (Ruck and Jahn,
1954), the amount of pressure applied to the body (Autrum and Hoff-
mann, 1960), the species of insect, the individual (HartHne, 1928), and
the characteristics of the recording apparatus, it is understandable that
there is great difficulty in interpreting ERGs and reconciling the results
reported by one laboratory with those reported by another. These
difficulties will become apparent below. None the less, hidden in the
ERG is the code to some of the primary events that occur in photo-
receptors when they are stimulated by light, and ERGs have been
studied intensively in the hope that they will yield bits of this infor-
mation.

ERGs have been recorded from a large number of insects. The first
ever obtained (from Melanoplus differ entialis, M. femur-rubrum,
Chortophaga viridisfasciata, Vanessa atalanta, Bombus pennsyh aniens,
Musca domesticd) were described by Hartline (1928) as being essen-
tially similar to one another and consisting exclusively of negative
components. These included a rapidly rising wave (which he desig-
nated as A and later workers as b) whose decay was interrupted by a
slowly rising wave (his B and the c of others). The off-effect was merely

PHOTORECEPTION 169

a decay of the c wave. Up to 0-08 seconds the Bunsen-Roscoe law
held. The many differences in the profiles of the ERGs were interpreted
as being due to differences in the form and magnitude of the two
negative waves.

Melanoplus was again studied ten years later (Jahn and Crescitelli,
1938; Crescitelli and Jahn, 1939) as were the grasshoppers Trim-
erotropis citrina and T. maritima (Jahn and Wulff, 1942). The ERG
was essentially similar to that reported by Hartline; however, this
time it was reported as beginning with a weak positive wave (a) promi-
nent in some eyes, absent in others, and abolished by light adaptation.
In addition, a negative off-effect {d wave) occurred in the form of a
small hump on the c wave. The moth Samia cecripoa was reported to
have a similar ERG except that a d wave was never seen (Jahn and
Crescitelli, 1939). But for Galleria mellonella Taylor and Nickerson
(1943) reported a, b, c, and occasionally (i waves.

The ERG of the water beetle Hydrous triangularis showed in its
most complicated form (high intensity, night eye) a simple negative
profile consisting of b and c waves only, while Dytiscus fasciventris
in similar circumstances showed a, b, and c waves (Wulff and Jahn,
1943). On the other hand, Bernhard (1942) described the ERG of
Dytiscus as being a simple b and c wave.

In the grasshopper Dissosteira Carolina the profile obtained was
totally negative, consisting oib and c waves and only occasionally a d
wave (Taylor and Crescitelli, 1944). The most marked departure from
form encountered was an occasional ERG, in which, after the initial
negative response to illumination, there was a long, slow, positive
swing. A similar ERG had been reported by Hartline (1928) for
Chortophaga viridisfasciata. The ERG of the cockroach Periplaneta
americana was reported as a simple negative wave (Ruck, 1958 a;
Walther, 1958 a, 1958 b), as were also the ERGs ofTachycines, Dixi-
ppus, and the nymph of AescJma (Autrum, 1948 a, 1948 b, 1950).

For all of the species mentioned thus far everyone agrees that the
usual ERG is essentially a cornea-negative swing. Depending on con-
ditions and species, it may or may not have an initial a wave and a final
i/ wave. There may be negative and positive after-effects.

Disagreement arises when the ERGs for Diptera and Hymenoptera
from different laboratories are compared. For Musca domestica
Hartline (1928) stated that the ERG was similar to that of the other
species he studied, i.e., a negative potential consisting of components
b and c, but in a later record (Hartline, Wagner, and MacNichol,
1952) the ERG appeared with a small positive a wave at the on-point

170 THE PHYSIOLOGY OF INSECT SENSES

and a small short negative d wave at the off. For Eristalis and Sarco-
phaga Hassenstein (1957) recorded an ERG with a very pronounced
positive on-effect followed by a slower negative wave and terminated
at the end of stimulation with a sharp negative off-efifect. These on-
and off-effects have the sign and position of the a and d waves res-
pectively. For Lucilia Kuwabara and Naka (1957) and Naka and
Kuwabara (1959) showed a simple positive monophasic wave at low
stimulus intensities. For Phormia regina the records of Ruck (1958 b)
showed at high intensities a small positive a wave and complex multi-
phasic waves with prominent negative components. The ERGs for
Apis mellifera and the dragonfly Pachydiplax longipennis were similar.
In his original papers Antrum (1948 a, 1948 b) described the ERG of
Calliphora as a diphasic response consisting of a spike-like positive
on-effect and a similar but negative off-effect (in these reports the po-
tential sign was incorrectly reversed but corrected in later papers).
With short flashes of light the wave was smoothly diphasic; with a
stimulus of long duration the positive wave returned to a line of zero
potential which continued until the off-effect. At high stimulus in-
tensities Antrum reported a *negatives Zwischenpotential*, and in a
later paper (Antrum and Hoffmann, 1960) figured the 'normal' ERG
as having a pronounced negative wave between the on- and off-effect.
The positive on-effect is sometimes preceded by a small negative wave
(Antrum, 1950; Kirschfeld, 1959). Similar ERGs have been described
for Apis, Bombus, Vespa, and the adult of Aeschna (Antrum and
Stoecker, 1950; Antrum and Gallwitz, 1951).

Antrum decided that there are two fundamentally distinct kinds of
ERGs in insect eyes, the one, the Dixippus-typQ, purely negative and
monophasic, the other, the Calliphora-typQ, diphasic. In the first cate-
gory ('slow') he placed all of the slow-flying, night-flying, and aquatic
species (e.g., most Orthoptera, Blatteria, Lepidoptera, aquatic
Coleoptera, and dragonfly nymphs) ; in the second category ('fast'),
the rapidly flying species (Diptera and Hymenoptera).

The wave forms of 'slow' and 'fast' eyes do not appear so radically
different in the records of most workers as they do in the records of
Antrum, Hassenstein, and Kirschfeld. Whereas the on- and off-effects
are the prominent features of the 'fast' eye as described and the slow
negative potential is negligible, the reverse is generally true in records
from other laboratories. Antrum has maintained that carefully
handled, fresh preparations are required to give reliable ERGs, and
both Hassenstein (1957) and Naka and Kuwabara (1959) have shown
that the ERGs of flies become monophasic as the preparation ages

PHOTORECEPTION 171

(about three hours). In all of the earlier work the preparations were in-
tentionally allowed to stand for approximately three hours in the dark
to stabilize. On the other hand, it is clear from all records that the
ERGs of Diptera and Hymenoptera are more complicated than those
of Orthoptera. Autrum (1950) suggested that those of Lepidoptcra,
with their small a waves, may represent an intermediate form.

According to Autrum, there are fundamental physiological differ-
ences, other than wave form, between 'slow' and 'fast' eyes. The 'slow'
ERG increases in magnitude with increase in stimulus intensity, it
changes markedly with light and dark adaptation, it responds poorly
to flickering light. Above frequencies of forty-fifty flashes per second
it fails to follow the stimulus. For stimuli of durations up to about 0-08
seconds, the magnitude of response is related to energy (intensity x
time). The 'fast' ERG increases in magnitude with an increase of
stimulus intensity, it is unaffected by light and dark adaptation, its
absolute sensitivity is less than that of 'slow' eyes {Tachycines is
approximately 140 times more sensitive than Callfphora), it follows
flicker up to about 300 flashes per second, the rate depending upon
stimulus intensity and temperature. The on-effect for all stimuli longer
than 5 msec, is dependent only upon intensity; the off-effect in the
range 5-200 msec, is dependent upon the total energy (i.e., intensity x
time). Ruck (1958 a, 1958 b), on the other hand, maintained that 'slow'
and 'fast' eyes differ only in their flicker fusion frequency and that the
wave form of the ERG, absolute sensitivity, and rate of dark adapt-
ation are independent visual functions not necessarily related to each
other, to the flicker fusion frequency, or to the form of the ERG. In
contesting these conclusions, Autrum and Hoffmann (1960) demon-
strated that subjecting the eye of Calliphora to oxygen lack (0-1 per
cent O2 plus 99-9 per cent N2) caused the ERG to revert to a mono-
phasic negative type until reoxygenated. The flicker fusion frequency
of this negative ERG drops to 60 per second ; it is very sensitive to light
adaptation. It was also pointed out that damage to the eye and long
use of preparations leads to the conversion of the diphasic potential
into a slow monophasic potential. The observations of Hassenstein
(1957) and Naka and Kuwabara (1959) confirm this. It is noteworthy
that in practically all of the work reported prior to 1948 preparations
were held for a period of one to four hours in the dark in order to
stabilize.

The shape of even the simple 'slow' type ERG has suggested to all
workers that it is the summation of several components. Even the
ERG of Limulus, the simplest of them all, is now believed to be dual

172 THE PHYSIOLOGY OF INSECT SENSES

(Wulff, 1950). Hartline (1928) considered that there were two negative
waves in the ERGs of insects (his A and B). Bernhard (1942) also
believed that in Dytiscus the ERG is made up of two negative waves, a
slow (S) and a rapid (R). In the dark-adapted state and at low in-
tensities R predominates; in the light-adapted state and high
intensities, S. He believed that R represents receptor activity and S is
in some way related to adaptation. Jahn and Wulff (1942) beHeved that
the ERGs of the grasshoppers Trimerotropis citrina and T. maritima
represent the sum of a positive and a negative wave. Taylor and
Crescitelli (1944) also assumed two components of opposite sign in
the ERG of Dissosteira, but did not completely abandon the idea that
the negative component itself might in fact consist of two waves. In
Galleria the a wave of the ERG was considered by Taylor and Nicker-
son (1943) to result from the interaction of two components. For the
more complicated, i.e., the *fast' type ERG, Antrum (1950) considered
that the negative off-effect is the homologue of the R component of the
'slow' type ERG and that the R component is a constant feature of all
ERGs.

The compound nature of ERGs suggested that different components
have different origins. Roeder (1940) had found that extirpation of the
optic ganglion of Melanoplus did not abolish the slow wave of nega-
tivity characteristic of the ERG. For Dytiscus, neither removal of the
ganghon nor its cocainization altered the ERG other than to remove
the small spikes (nerve action potentials) that had been superimposed
on the slow wave (Bernhard, 1942). This smooth monophasic wave is
identical to that recorded when the electrodes are placed on the outer
and inner surfaces of the retina, and its size decreases as the leads are
removed proximally away from the retina. In some cases Bernhard
was able to record from a cornea from which the basement membrane
and all structures below it had been removed. In these cases the usual
large negative wave was obtained. This is clear evidence, at least for
Dytiscus, that the negative potential originates in the retinal cells and
not in the giant monopolar cells. Working with Trimerotropis, Jahn
and Wulff (1942) compared the ERGs of normal and deganglionated
eyes and concluded that the retina contributes a negative component
and the optic ganglion, a positive component.

The origin of some of the components in *fast* eyes has been eluci-
dated by two very ingenious experiments of Antrum and Gallwitz
(1951). Successive portions of the optic ganglia of Calliphora were
removed and the ERGs recorded. As more and more of the ganglionic
mass (medulla interna and medulla externa) was extirpated, the

PHOTORECEPTION 173

negativity of the ERG increased. When the medulla interna, medulla
externa, and all or part of the lamina ganglionaris were removed the
ERG became a negative monophasic curve resembling that of
Limulus (Fig. 86).

In the dragonfly Aeschna cyanea the optic ganglia of the young
nymph are remote from the retinal layer. As the nymph develops,
the distance between the ganglia and the retina decreases, until in the
adult eye the usual situation obtains (Fig. 87) (Viallancs, 1884).
The ERGs of nymph and adult are distinct - that of the nymph is

A B

Fig. 86. A. Simplified scheme of the optic lobe of Calliphora. AK, centri-
petal cells ; IK, centrifugal cells ; R, retinal cells ; a, ERG from isolated
retina; b, ERG from retina plus lamina ganglionaris; c, ERG from
intact lobe. B. Detail of centrifugal cells. (Redrawn from Autrum,
1959 after Cajal and Sanchez, 1915.)

the *slow' type, that of the adult, the 'fast' type. The ERG of the
deganghonated retina of the adult is the *slow' type.

As a result of these experiments, Autrum (1958) has concluded that
there is a negative component of the ERG of all insects assignable to
the retina; it is the receptor potential. He has pointed out that in
Calliphora under certain conditions (e.g., green or blue light) the posi-
tive on-wave is preceded by a negative deflexion. This negative wave,
assignable to the retinal cells, triggers a positive electrical response in
the centrifugal cells of the lamina ganglionaris. This positive wave
is considered to be of great importance in connexion with adaptation.
Dark adaptation has been measured for intact insects (e.g., the honey-
bee) by behavioural criteria, such as optomotor responses, and has

174 THE PHYSIOLOGY OF INSECT SENSES

been found to require up to thirty-five minutes for completion (Wolf
and Zerrahn-Wolf, 1935 a). In the intact eye of Calliphora, as measured
electrophysiologically, it has been found to require from minutes to
hours. In the isolated retina, however, adaptation is complete within
fractions of a second. It is proposed (Antrum, 1950, 1952, 1958) that
the positive potential from the centrifugal cells acts as a *bucking
potential' preventing sustained depolarization of the retinal cells. As
a corollary of this, resolution of faster rates of flicker becomes
possible. When the ganglia are removed the resulting monophasic
ERG is unable to follow flicker at the same high rates as the intact
eyes.

Support for the idea that the positive component arises post-
synaptically has been obtained by poisoning with nicotine, a specific

ABC

Fig. 87. Position of the retina (R) in relation to the optic lobe (I, II, III) in
the developing dragonfly. A, young larva; B, older larva; C, imago.
(Redrawn from Autrum, 1958 after Viallanes, 1884.)

synaptic poison. This treatment abohshes the positive components of
*fast' ERGs, leaving the negative portions unchanged, and has no
influence on *slow' ERGs, as in Dixippus (Autrum, 1958).

Ruck (1958 a, 1958 b) has not been able to confirm the relation be-
tween wave form of the ERG, flicker, and adaptation in the eyes of
Apis mellifera, the dragonfly Pachydiplax longipennis, and Phormia
regina. He suggested that the lowering of flicker fusion frequency in
Antrum's experiments is due to injury and that actually the flicker
fusion characteristics of an eye are determined by the retinal elements
themselves.

Other ideas on the subject of ERG components include Hassen-
stein's (1957) analysis of the ERG of Sarcophaga and Eristalis into a
diphasic component and a negative component and Naka's and
Kuwabara's (1959) conclusion that the ERG ofLucilia is composed of
a negative and a positive component, both of which originate in the
receptor layer. Burtt and Catton (1958), on the basis of an observed

PHOTORECEPTION 175

reversal of polarity of potentials (also noted by Naka and Kuwabara)
at a critical depth in the eye (retina plus optic lobes) corresponding to
the superficial part of the lamina ganglionaris, concluded that cells
in this layer, probably the giant monopolar cells, are the source of the
major part of the potential observed when the eye is illuminated. This
conclusion is in disagreement with the results of Bernhard (1942) from
Dytiscus, and has been criticized by Antrum (1958) as representing
pecularities of the recording technique. Ruck (1 957) has argued that an
interpretation of these data in a manner consistent with the distri-
bution of current in a volume conductor would in fact localize the site
of the origin of the ERG to the ommatidia.

Still another electrical phenomenon observed in the optic tract of
insects is rhythmic spontaneous activity first found in Dytiscus
(Adrian, 1937) and subsequently in all eyes in which it was sought
(Roeder, 1939, 1940; CrescitelH and Jahn, 1942; Jahn and Wulff, 1942;
Bernhard, 1942; Massera, 1952; Autrum, 1951, 1952; Burkhardt,
1954; Burtt and Catton, 1958). It is clearly of ganghonic origin, since
extirpation of the optic ganglion abolishes it (Roeder, 1940; Burk-
hardt, 1954). 'Fast' eyes produce faster rhythms than 'slow' eyes
(Autrum, 1951, 1952; Burtt and Catton, 1958). In Dytiscus there are
differences between the day and night eye, the former having a higher
flicker fusion frequency than the latter; furthermore, the dark-adapted
night eye is more than 1 ,000 times more sensitive than the dark-adapted
eye, and the light-adapted day eye more sensitive than the light-
adapted night eye (Jahn and Wulff, 1941). The differences are un-
related to pigment migration in the eye. In Dytiscus there are actually
two rhythms: a dark rhythm and a bright (fast) rhythm. The latter
occurs only under maximum illumination, and was visualized by
Adrian (1937) as representing firing of all neurons. By contrast,
rhythms in butterflies and grasshoppers are found under any inter-
mediate conditions (CrescitelH and Jahn, 1942). They occur at high
temperatures as an after-discharge to a single flash of light and also
as a result of repetitive stimulation. The dark rhythm in Dytiscus
was explained by Adrian (1937) as the resting potential of all neurons
in the system being favoured by injury. In Mantis religiosa and
Romalea microptera there is also a light and a dark rhythm (Roeder,
1939, 1940). The light rhythm is somewhat similar to that found in
Dytiscus and unHke that described by CrescitelH and Jahn (1942).
Burkhardt (1954) has localized the bright rhythm as originating in the
medulla externa. This rhythmic excitation is related to the stimulus in
its amplitude but not in its frequency. The frequency Hes between 100

176 THE PHYSIOLOGY OF INSECT SENSES

and 250 per second. The amplitude is related to the number of illumi-
ated ommatidia, the Hght intensity, and the state of adaptation of the
eye. As illumination continues the amplitude gradually decreases. If
illumination is interrupted before the amplitude has dropped to zero
the rhythm is abruptly terminated. The likelihood is great that these
rhythms are an important link in the causal chain of central processes
which occur when an eye is illuminated.

One conclusion that can be drawn with certainty from all of the
electrical data obtained from the compound eye is that the ERG, even
in its simplest form, is the algebraic sum of potentials originating at a
number of loci. Some of these loci are non-retinal; nevertheless, it is
now doubtful that even retinal potentials are as simple as once be-
lieved (cf. Wulff, 1950). The most elegant approach to the whole prob-
lem has been made in Limulus. As the classical work of Hartline (1928)
showed, this is a rather simple monophasic cornea-negative wave.
When a micro-electrode is placed intracellularly in an ommatidium in
the dark (HartHne, Wagner, and MacNichol, 1952; MacNichol,
Wagner, and Hartline, 1953) there is a resting potential, negative at the
recording electrode, of about 50 mV. When the ommatidium is stim-
ulated there is a change in the positive direction (with intracellular
recording the recording electrode goes positive when the cell is de-
polarized). Superimposed on this slow wave are spikes that are related
linearly to the magnitude of depolarization. These spikes are synchron-
ized with spikes recorded simultaneously from the axons of the om-
matidium. Curiously enough, the only cell in the ommatidium which
appears to be active is the eccentric cell; the others invariably are
electrically silent. There is little doubt that the slow negative wave
characterizing the ERG of the Limulus eye is the generator potential
(and probably also the receptor potential) giving rise to the action
potential in the nerves. There are now two reports of similar record-
ings having been made in insects. Burkhardt and Wendler (1960) have
recorded with intracellular electrodes the ERG from individual retinal
cells of Calliphora. The ERG is a simple negative monophasic wave
nearly identical in appearance to that o^ Limulus. Naka and Eguchi
(1962) have made similar recordings from the drone honeybee (see
also Ishikawa, 1962).

Many attempts have been made to deduce from ERGs some of the
characteristics of the photoreceptors themselves, but as the previous
discussion has indicated, the ERGs have been much too complicated
and too poorly understood to have contributed greatly to this end.
Such attempts have been made by Wulff (1943) and Wulff and Jahn

PHOTORECEPTION 177

(1943) and vigorously criticized by Granit (1947). More recently
Antrum (1953, 1958) has proposed a hypothesis which can be summar-
ized as follows: Light is absorbed by a photosensitive substance, as a
result of which the substance is altered and stimulates the retinal cells.
In 'slow* eyes it decomposes, and the photosensitive substance is re-
generated. Since these processes require time, adaptation is slow. In
*fast' eyes, however, decomposition and regeneration is blocked during
illumination, but when illumination ceases the process proceeds very
rapidly. In other words, adaptation is very rapid. The blocking is
accomplished by the positive potential from the centrifugal cells.
This is a very high potential (20 mV as compared with 1 mV in the
frog).

Wulff and his associates (Wulff, Fry, and Linde, 1955; Fry, Wulff,
and Brust, 1955) have made a different analysis. Following illumina-
tion of the eye there is always a latency before the first wave of the ERG
arises. It is assumed that this is the period required for photolysis to
initiate the retinal action potential. The coupling processes connecting
these two events is obscure. According to the hypothesis of Wulff e/ al.^
there are two processes involved in the sense cells : an electrical pro-
cess generating the retinal action potential and an auto-csLtalytic rate
process which controls the latent period. Light acting on a photo-
sensitive substance generates another substance (C) whose concentra-
tion manifests itself as a retinal e.m.f. C is presumed to be generated at
a rate proportional to the intensity of light. The retinal e.m.f. is pro-
portional to log cone. C. The latency for a flash of light is a hnear
function of log / over most of the intensity range tested and is relatively
insensitive to time. Wulff e^ ai. suggested that the failure of 'slow' eyes
to obey the Bunsen-Roscoe law after 30 msec, may be attributed to
decay of an electrically active substance rather than to a recovery
process as Autrum (1950) suggested.

COLOUR VISION

From the observations of Plateau (1 888) to the present time there have
been literally hundreds of recorded observations about insects show-
ing preferences for one colour or another. Consequently, there is not
the slightest doubt that insects can distinguish among the various
coloured objects in nature and that their eyes are sensitive to wave-
lengths from about 253 m[x (the near ultra-violet) to about 700 mii (the
infra-red). Lubbock (1886) had shown that ants were less apt to move
their pupae out of sunlight from which the ultra-violet had been

178 THE PHYSIOLOGY OF INSECT SENSES

filtered than out of normal sunlight. At the opposite end of the spec-
trum Buck (1937) had observed that the firefly Photinus pyralis could
respond to red flashes (approximately 560-690 mjji) emitted by other
individuals. What has not always been clear is whether or not the eye is
equally sensitive to all wavelengths and can distinguish one wave-
length from another when the energies of the various coloured
fights are equalized; in other words, whether insects possess colour
vision.

A large number of experiments designed to provide information on
these points have been undertaken. They fall into two categories:
behavioural and electrophysiological. In both instances some selected
response is measured at different wavelengths of fight whose spectral
purity and intensity had been controlled with various degrees of rigor.
In so far as response to different wavelengths is concerned, the results
of the majority of tests, regardless of their sophistication, agree in a
general way in showing that insects as a class are especiaUy sensitive to
ultra-violet and to blue-green fight (Fig. 88).

Behavioural experiments have been based on phototactic responses
(Peterson and Haeussler, 1928;Bertholf, 1931a, 1931b;Sander, 1932,
1933; Cameron, 1938; Weiss, Soraci, and McCoy, 1941, 1942, 1943;
Weiss, 1943a, 1943b, 1944, 1946; Weiss, McCoy, and Boyd, 1944;
Milne and Milne, 1945; Fingerman, 1952; Fingerman and Brown,
1953; Wolken, 1957; Wolken, Mellon, and Contis, 1957; Heintz,
1959; Goldsmith, 1960), optomotor responses (Schlegtendal, 1934;
Rokohl, 1942; Moller-Racke, 1952; Antrum and Stumpf, 1953;
Schone, 1953; Resch, 1954; Schneider, 1956), training (von Frisch,
1914; Koehler, 1924; Kiihn, 1927; Kuhn and Pohl, 1921; Bertholf,
1931a; Use, 1949; Kuwabara, 1957; Hertz, 1939; Daumer, 1956,
1958), and simple observation (Lubbock, 1886; Hamilton, 1922; Use,
1928, 1937, 1949; Buck, 1937). Phototactic responses have been em-
ployed exclusively for studying spectral sensitivity, while other be-
havioural techniques have been employed in the search for evidence
of colour vision.

The relative efficiency of different wavelengths in eliciting photo-
tactic responses has been measured for a number of species. Bertholf
presented to Drosophila and Apis two sources of light, one of which was
the test wavelength and the other, a white standard. In one series of
tests the intensity of the test colour was kept constant and that of the
standard varied until both sources were equally attractive. In another
series of tests each coloured light was matched with a white fight of
fixed intensity. After the ratio of attractiveness had been found a white

300

400

500

600

mfj

Fig. 88. Relative effectiveness of different vi'avelengths of light in evoking
positive phototaxis and similar responses of adult insects. In each case,
open circles indicate one set of experiments and the closed circles,
another. (Redrawn from Goldsmith, 1961.)

180 THE PHYSIOLOGY OF INSECT SENSES

light of variable intensity was substituted for the coloured light, and
the intensity was found which, when tested against the white stand-
ard, gave the same ratio of attractiveness as the test colour. Despite
these attempts at control, Bertholf did not have control over the
energy of his stimuU (see Weiss, 1943 a, 1943 b, and Goldsmith, 1961,
for a full discussion). The results of his experiments showed a peak
sensitivity in the ultra-violet (365 mjjt) and a slight indication of sensi-
tivity in the blue-green (487 my.) for Drosophila and in the yellow-green
(553 mji.) for Apis. Sander (1933) adjusted his lights so that the in-
tensities of the white standard and the test colours were equal. He then
used the number of insects {Apis) attracted to each as an index of rela-
tive effectiveness and found peaks in the yellow-green (570 mji.) and
in the blue (470 m^), but none in the ultra-violet. Cameron (1938)
employed all of the aforementioned methods in tests with Musca and
found a maximum in the ultra-violet. Weiss and his co-workers ex-
posed approximately 1 5,000 insects representing forty species to ten
wavelengths of light of equal intensities. The composite group be-
haviour pattern consisted of a peak at 492 m[ji and a maximum at
365 ra\i. Wolken determined for Drosophila the relative energy re-
quired to produce a constant phototactic response and found a peak
in the blue-green and a maximum commencing towards the ultra-
violet. Heintz measured, as a function of intensity, the number of bees
that crawled per unit time over a slit in a box illuminated by different
wavelengths and found a small peak in the green and the maximum
in the ultra-violet. He suggested that Sander's failure to detect a peak
in the ultra-violet stemmed from the very high light intensities em-
ployed. Since these intensities elicited maximum responses in the blue-
green, no higher response could possibly have been obtained in the
ultra-violet. Goldsmith (1960), perturbed by Sander's failure to find a
maximum for the bee in the ultra-violet, tested the phottoactic effective-
ness of green (546 mfx) and ultra-violet (365 m[j.) lights of equal in-
tensities and did find a peak at 365 mt^. In short, whether the curves
describing the relative effectiveness of different wavelengths in eliciting
phototactic responses are true action spectra (relative energy for a
constant effect) or not, all show that the most effective part of the
spectrum for Musca, Apis, and Drosophila is the near ultra-violet and,
to a lesser degree, the blue-green. This finding has been confirmed in
whole or in part for so many species by Weiss and his associates that it
is probably of general occurrence.

The danger, of course, in interpreting these behavioural tests with-
out caution as indicative of spectral sensitivity is that the response of an

PHOTORECEPTION 181

insect to specific wavelengths may vary with the physiological state.
Use (1928) has observed, for example, that cabbage butterflies {Pieris
brassicae) normally land preferentially on blue or yellow flowers or
paper models of the same colours, while gravid females ready to
oviposit shift their preference to green and blue-green (1937). On the
other hand, electrophysiological studies have confirmed the more
general behavioural findings (Crescitelli and Jahn, 1939; Jahn and
CrescitelU, 1939; Jahn, 1946; Jahn and Wulff, 1948; Donner and
Kriszat, 1950; Autrum and Stumpf, 1953; Goldsmith, 1958 a, 1958 b;
Goldsmith and Ruck, 1958; Walther, 1958 a, 1958 b; Walther and
Dodt, 1957, 1959). Spectral sensitivity curves were obtained in these
cases by relating the magnitude of some selected components of the
ERG to the wavelength of light or, more accurately, by plotting some
function of the quanta necessary to produce a given response against
wavelength. For the moth Samia cecropia (Crescitelli and Jahn, 1939)
and the grasshopper Melanoplus differentialis (Jahn and Crescitelli,
1939) the greatest sensitivity, exclusive of ultra-violet, which was not
tested, is in the blue-green (from about 460-530 mfi). The curve for
night and day eyes of Dytiscus fasciventris also shows a maximum
in the blue-green (about 520-575 m\L). It is based upon the magnitude
of the a wave. This was plotted against intensity. From the family of
curves (one for each colour) obtained, a constant response magnitude
was selected and plotted against wavelength (Jahn and Wulff, 1948).
Curves for Musca domestica, Lucilia caesar, Calliphora vomitoria,
Pollenia rudis, and Drosophila meJanogaster based upon measurements
of the on-efTect of the ERG at wavelengths of equivalent quantum
intensity show maxima in the green and in the near ultra-violet
(Donner and Kriszat, 1950).

A spectral effectiveness curve for Calliphora erythrocephala shows
a peak at 540 m(x (blue-green) and one at 630 m\i (red) (Autrum and
Stumpf, 1953) (Fig. 89). This curve is based on the amplitude of the on-
effect of the ERG, or the height of the flicker potential occurring as a
reaction to twenty-seven flashes, produced by stimulation by light of
different wavelengths but equal quantal value. No tests were made
with ultra-violet, but the rising curve at the shorter wavelength
suggested ultra-violet sensitivity. This expectation was realized in tests
subsequently conducted by Walther (1957) and Walther and Dodt
(1957, 1959). A spectral sensitivity curve (reciprocal of the quanta
necessary to generate a constant magnitude of on-effect plotted as a
percentage of the maximum of 507 m(a against the wavelength) for
Calliphora shows maxima in the red (about 630 mjx), the blue-green

182

THE PHYSIOLOGY OF INSECT SENSES

500

400

500

600

IVave Length mu

Fig. 89. The spectral sensitivity of the compound eye of Calliphora. Data
from Antrum and Stumpf (1952) (open circles), Walther and Dodt
(1957) (half-filled circles), and Walther and Dodt (1959) (solid circles).
(Drawn from Goldsmith, 1961.)

(540 mji.), and the ultra-violet (341-369 m\i). The curve for Periplaneta
is similar, but lacks the peak in the red.

A spectral sensitivity curve for the drone honeybee shows a maxi-
mum at 400 my. for the compound eye and maxima at 335-340 my. and
490 my for the median ocellus (Goldsmith, 1958 a, 1958 b; Goldsmith
and Ruck, 1958). The dorsal ocelli of the cockroaches Periplaneta

PHOTORECEPTION 183

americana and Blaberus craniifer show a peak at 500 my. (Goldsmith
and Ruck, 1958).

One of the more interesting aspects of all of these descriptions of
spectral sensitivity is the great stimulating effectiveness of the near
ultra-violet. Hess (1920) had questioned the sensitivity of the insect
retina to ultra-violet and suggested the possibility of stimulation
actually occurring as a result of fluorescence set up in various tissues of
the eye. Fluorescence may occur (Walther and Dodt, 1959), but
Merker (1929) and Lutz and Grisewood (1934) showed that crushed
eyes of Drosophila did not fluoresce in ultra-violet of the wavelength
to which flies respond. They showed, furthermore, that the cornea
of Apis and Sarcophaga can transmit in the 253-m[JL band of the
spectrum. Additional reasons for rejecting Hess's hypothesis are
mentioned by Walther and Dodt (1959) and discussed by Goldsmith
(1961).

Taken at face value, the spectral sensitivity data indicate what may
constitute brilliance for an insect, and a number of studies bear this
out wholly or in part (MoUer-Racke, 1952, with Dytiscus, but compare
Schone, 1953; Rokohl, 1942; Llidtke, 1953, and Resch, 1954, with
Notonecta). Demonstration of true colour vision has been technically
more difficult. There is now evidence for colour vision in thirty-three
genera of insects in six orders. The evidence has been derived princi-
pally from training experiments, optomotor responses, and electro-
physiological analyses.

Von Frisch (1914) trained honeybees to collect food from a dish
placed on a card of a particular colour. This card was then placed in
one square of a checkerboard of greys (white to black) of different
intensities. The position of the coloured card was changed constantly.
Under these conditions the bees could always pick out a blue or yellow
card from the checkerboard of greys. Thus, it was demonstrated that
bees conditioned to blue confused blue-violet and purple, while those
conditioned to yellow confused yellow, orange, and yellow-green.
None could distinguish red. Later tests of the papers used by von
Frisch showed that some of the blues and greens reflected ultra-
violet, whereas some of the yellows and greens reflected red and blue
(Lutz, 1924). Similar experiments employing spectral colours pro-
jected on a white background demonstrated that bees could dis-
tinguish four regions of the spectrum: yellow-green and orange
(510-650 m|j.), blue-green (480-500 ma), blue and violet (400-480 m\L\
and the near ultra-violet (300-400 m\3) (Kiihn and Pohl, 1921 ; Kiihn,
1927). These results were confirmed by the training experiments of

N

184 THE PHYSIOLOGY OF INSECT SENSES

Kuwabara (1957) and of Daumer (1956, 1958), who also demon-
strated that bees can even discriminate between wavelengths within
these bands, but with less precision.

Colour vision can also be investigated by colour-matching tests.
Daumer (1956, 1958) applied these with outstanding success to honey-
bees by training them to come to a particular colour. The training
colour was then paired with a colour produced by mixing wave-
lengths. Various mixtures were tested until one was found which was
as acceptable to the bees as the training colour. For example, bees can
distinguish between white light containing ultra-violet and that lack-
ing ultra-violet. White light with ultra-violet ('bee white') can be
matched by a mixture of 15 per cent ultra-violet (360 mjx) and 85 per
cent blue-green (490 mii). The complementariness of ultra-violet and
blue-green had already been suggested by Hertz (1939), and Use (1937)
had shown that purple adjacent to white sets up a green colour con-
trast in the white. Similarly, *bee purple' can be matched by mixtures
of yellow and ultra-violet. Goldsmith (1961) has summarized Daum-
er's results in a preliminary chromaticity diagram which is explained
in Fig. 90.

From the results of all of his matching experiments, Daumer con-
cluded that the eye of the honeybee possesses a receptor maximally
sensitive to ultra-violet, one for blue-violet, and one for green-yellow.
The existence of two of these has been confirmed by the electrophysio-
logical analyses of Goldsmith (1958 a, 1958 b, 1959, 1961). The wave
form of the ERG in the early studies with grasshoppers and Dytiscus
had shown no change as the wavelength of the stimulus was changed,
provided the intensities were matched (Crescitelli and Jahn, 1939;
Jahn, 1946; Jahn and Wulff, 1948). On the other hand, as Walther
(1958 a) found in the cockroach eye, there are some areas of the eye
where differences in the form of the ERG arising from differences in
the colour of the stimulus cannot be matched by adjusting intensity.
Goldsmith and Ruck (1958) also found that the ERG of the ocellus of
the bee had at low stimulus intensities a different contour for ultra-
violet than for green.

By adapting with lights of different wavelengths it is possible to
demonstrate the existence of different receptors (Hamilton, 1922;
Fingerman and Brown, 1953). The dark-adapted eye of the worker
honeybee shows a peak of sensitivity in the green (535 m^t) and a
smaller peak in the ultra-violet (Goldsmith, 1960). If the eye is
adapted with red or yellow light the peak at 535 miJ. diminishes and the
peak at 340-345 m[j. increases (Fig. 91). The complementary experi-

PHOTORECEPTION

185

375

Fig. 90. Tentative chromaticity diagram for the honeybee based on the
work of Daumer (1956). The behavioural experiments of Daumer led
to the prediction that any colour can be matched for bees by an appro-
priate mixture of three monochromatic reference stimuli, e.g., ultra-
violet (360 m[x), blue-violet (440 my.), and yellow (588 my.). Any light,
therefore, can be represented in Fig. 90 by a point; for example, 'W
is the white of the xenon emission spectrum. The proportions, in
fractions of the total energy - the chromaticity co-ordinates - of the
yellow and blue reference stimuli required for a colourimetric match
of the light represented by the point are given by the co-ordinates of
the point, 588 my. on the abscissa and 440 my. on the ordinate. Since
the sum of the three chromaticity co-ordinates is 1 , the fraction of the
energy at 360 my. can be calculated by difference. The locus of the
spectrum is given by the experimental points ; however, as is suggested
by the dashed line, it may actually lie somewhat outside the triangle, for
it is questionable whether this kind of experiment is sufficiently precise
to reveal the necessity for small negative values of one of the chroma-
ticity co-ordinates. The complementary to 360 my. is 490 my., for both
wavelengths lie on a straight line passing through the white point.
Similarly, the point 'p' is the 'bee-purple' complementary to 440 my.
and is composed of 79 per cent 588 my. and 21 per cent 360 ma.
(Redrawn from Goldsmith, 1961.)

186

THE PHYSIOLOGY OF INSECT SENSES

20

1-5

>
V)

C
-J

10

0-5

Dark Adapted

During adaptation
with red light

300

Fig. 91. Alteration of the shape of the spectral sensitivity of the worker
honeybee by selective light adaptation. During a constant level of
adaptation brought about by red light the ultra-violet receptor system
contributes more prominently to the spectral sensitivity function
(open circles) than it does in the dark-adapted eye (filled circles).
Ordinate: logarithm of the reciprocal of the relative number of quanta
required to produce a retinal action potential of constant size.
(Redrawn from Goldsmith, 1961.)

ment is less striking because the ultra-violet receptor system contri-
butes less to the ERG than the green system and because the sensitivi-
ties of the two systems behave differently to different coloured
adapting hghts. Nevertheless, these experiments confirm the existence
in the bee of two of Daumer's postulated receptor types. The existence

PHOTORECEPTION 187

of a receptor maximally sensitive at about 440 m[i has been demons-
trated in the drone (Goldsmith, 1958 b). Furthermore, Goldsmith
(1958 a) has found a retinene-containing pigment with a maximum
absorption in the blue-violet (440 m|i.).

Studies of the cockroach eye have also revealed the existence of
more than one type of receptor (Walther and Dodt, 1957, 1959;
Walther, 1958 a, 1958 b). In the dorsal area of the eye the sensitivity
maximum to ultra-violet is higher than that to blue-green; in the
ventral part of the eye it is lower. Whereas the ventral area shows a
single-peaked (507 m,u.) sensitivity curve that is unaltered by the state
of adaptation, the dorsal area shows indication of another peak in the
ultra-violet. Adaptation with monochromatic lights changes the rela-
tive spectral sensitivities, and colour specific differences occur in the
shape of the ERG.

For Drosophila Hamilton (1922) had produced evidence for the
existence of a system sensitive to blue-violet and one sensitive to blue-
green by specific wavelength adaptation. Fingerman and Brown
(1953), employing a similar technique, concluded that Drosophila also
possesses sensitivity in the red (650-675 mpt).

The blowfly Calliphora is unique among the insects studied in that
the spectral sensitivity curve of its eye, as measured electrophysio-
logically, exhibits three maxima: ultra-violet, blue-green, and red
(Autrum and Stumpf, 1953 ; Walther and Dodt, 1958 a, 1958 b). There
is a discrepancy of about 30 mjx between the positions of the green
maximum reported by the two papers.

In their search for evidence of colour vision, Autrum and Stumpf
combined electrophysiological and flicker fusion techniques. After
adjusting the intensities of two monochromatic lights they flickered
them alternately. Thus during the period of stimulation there was
constant intensity of illumination but flickering wavelengths. If the
ERG showed with heterochromatic flicker ripple that could not be re-
moved by adjustment of intensities, then one could conclude that the
eye discriminated between the two wavelengths. The findings were as
follows: In the region 690-620 m\L discrimination is slight, but this
region is sharply distinguished from all other regions ; yellow to blue-
green (620-650 m[x) is distinguished from all other regions and dis-
crimination in this band is sharp (a restricted region between 570 and
590 m.[L cannot be distinguished from colourless light); blue-green
(480 m.\i) is distinguished from all other colours; on both sides of
480 m\i there is a region (450-500 m\i and 500-560 m^t) which is dis-
tinguished from all colours between 560 and 690 mfx. Autrum and

188 THE PHYSIOLOGY OF INSECT SENSES

Stumpf concluded that there need be only two receptor systems to
account for the data, a blue-green receptor and a red receptor. The
maximum for the blue-green receptor lies between 500 and 550 my.. Its
sensitivity extends from the region of the near ultra-violet up to about
660 mil. The maximum for the red receptor lies in the vicinity of
625 my.. Its sensitivity extends from about 480 up to about 700 my..
Several flies were found which were colour-blind when tested with
flicker and whose spectral sensitivity curves were abnormal. The
spectral sensitivity curve of one possessed the usual peak in the red but
none in the blue-green. This fly possessed no colour discrimination in
the region 400-500 my. as judged by flicker. Another fly lacked the
peak in the red (630 m^y.) and was totally colour blind. This fly possessed
the usual red eye-pigments in the pigment cells of the ommatidium.
Later Antrum (1955 a) described the spectral sensitivity of a white-
eyed mutant which lacked these pigments. Since this mutant also
lacked a sensitivity peak in the red (630 my.). Antrum concluded that
the usual 630-m[x peak resulted from transmission of long wave-
lengths by the red shielding pigments.

Behavioural experiments conducted with a revolving drum and
light of low intensities did not reveal a marked sensitivity in the red
(Schneider, 1956). To reconcile the behavioural and electrophysio-
logical results Schneider proposed that since the enveloping pigments
of the ommatidia are partially transparent to red light, the ERG would
be higher at red than at other wavelengths because the red would be
stimulating a greater number of ommatidia. On the other hand, the
failure of the screening pigments to isolate the ommatidia when the
eye was illuminated with red light would lead to a loss in visual acuity
and since the behaviour in a revolving drum depends upon visual
acuity, it would fail to reveal a sensitivity to red. Antrum (1961) has
reported, however, that even in a white-eyed mutant, which completely
lacks screening pigment, the visual acuity equals that of wild-type flies.

Recently Burkhardt and Antrum (1960) have been able to record
electrical events directly from single visual cells. The spectral sensi-
tivity was determined for thirty-nine single retinal cells of the wild type
of Calliphora and from the white-apricot mutant which lacks screen-
ing pigments (Antrum and Burkhardt, 1961). The red screening pig-
ment does not absorb Hght in the region 616 m.y., and in weakly
illuminated elements the amount of scattered red light is much greater
than that amount of light coming through the lenses. Except in these
weakly illuminated elements and in the far red, the screening pigments
are not relevant to the shape of the sensitivity curves. All cells in the

PHOTORECEPTION 189

dorsal part of the eyes were found to have curves with a maximum at
489 ma and one at 345 mix. Three receptor types were found in the
ventral region of the eye: a green receptor (maxima at 491 m[x and
345 m[x); a blue receptor (maxima at 468 m[x and 345 mix); a yellow-
green receptor (maxima at 524 mix and 345 mix). Some elements show a
peak at 616 m^x. The numerical relationships among the green, blue,
and yellow-green receptors is 18:4:3. This is almost the ratio to be
expected if one assumed that of the seven retinula cells in an ommati-
dium five are green receptors, one blue, and the remaining one
yellow-green. Since all have a peak at 345 m[x, corresponding to a
stable position of the p-absorption in derivatives of vertebrate visual
pigments, it is possible that the most common visual pigment in
Calliphora is retinene with a maximum absorption at 490 mjx (Autrum
and Burkhardt, 1961).

Some workers have been tempted to think of the insect eye as
possessing analogues of rods and cones, especially since Hanstrom
(1927) demonstrated retinal cells with short and long axons and
suggested the analogy. But Wolf and Zerrahn-Wolf (1935 a), in dis-
cussing the relation between threshold intensities and time of dark
adaptation for the honeybee, called attention to the fact that the curve
describing this relation was, unlike that for the human eye, a smooth
one. Nor do curves relating visual acuity and flicker fusion frequency
to intensity exhibit any changes in slope that could be attributable to
two populations of receptors of different sensitivities analogous to
rods and cones (Hecht and Wolf, 1929; Wolf, 1933e;HechtandWald,
1934; Wolf and Zerrahn-Wolf, 1935 b). In Drosophila, however,
Fingerman and Brown (1952, 1953) reported a 'Purkinje shift', that is,
a shift of sensitivity of the eye towards shorter wavelengths as the light
intensity is decreased. They interpreted this as indicating a shift from a
photopic to a scotopic mechanism. According to their conclusions,
Drosophila loses its ability to discriminate among wavelengths as
intensity is lowered. No such change could be demonstrated in the
cockroach (Walther, 1958 a, 1958 b), nor in the honeybee (Goldsmith,
1960, 1961). The data of Fingerman and Brown (1952, 1953) have
been criticized by Goldsmith (1961). On the other hand, Weiss et al.
(1942) reported that the order of attractiveness of 365 my. and 470-
528 mjx at low intensities was reversed at high intensities. With
Calliphora Autrum (1955 a) found that the wavelength of maximum
sensitivity was 540 mpi at high intensities but shifted to shorter wave
lengths (480 m^x) as the intensity was dropped. Schneider (1955, 1956)
reinvestigated this problem by comparing spectral sensitivity curves

190 THE PHYSIOLOGY OF INSECT SENSES

obtained in optomotor-type apparatus at high and at low intensities
with the curve obtained electrophysiologically. The behaviourally de-
termined curve obtained at low intensities showed a maximum at
480 ray.. The high-intensity curve shifted towards 540 my.. This shift
could, as Schneider pointed out, be due to the transparency of envelop-
ing pigments to long wavelengths, but were this true, the white-eyed
mutant should exhibit a maximum at 480 my., whereas it actually shows
a maximum between 510 and 540 my. (Antrum, 1955 a). The shift must
be due to the participation of a receptor of low threshold, the 'receptor
for twilight vision', whose maximum sensitivity lies at 480 my. and a
less-sensitive receptor whose maximum lies above 540 my. (Schneider,
1956). Walther and Dodt (1957, 1959) have not reported any shift. In
the back-swimmer (Notonecta) a shift in which red and yellow become
darker to the animal with advancing dark adaptation (as measured
by optomotor reactions) has been noted by Resch (1954) and Liidtke
(1953) and referred to as a Turkinje phenomenon'.

POLARIZED LIGHT

In 1948 von Frisch (1948, 1949) opened a new area of investigation in
insect vision by demonstrating that honeybees could distinguish
different quadrants of the sky in which the plane of polarization of
sunlight differed. This conclusion was based on the fact that: (1) bees
in the hive could orient their communication dances as long as they
were able to see a patch of blue sky, and (2) the orientation of the dance
could be altered by interposing a piece of polaroid between the bee and
the sky. Bees act as though the sun is at right angles to the plane of
polarization. Since then directional orientation in the presence of
linearly polarized light has been demonstrated in many arthropods
(for reviews see von Frisch and Lindauer, 1956; Waterman, 1960).

Polarized Hght may differ in degree, in type (linear, elliptical, or
circular), or in orientation of major axis. Practically all polarized light
in nature is linear (Jander and Waterman, 1960), and only this type is
known to affect animals. Two basic explanations have been proposed
to explain its perception. The most direct and obvious explanation is
that the eye possesses a specific polarization analyser, that is, that
animals see the direction of vibration of polarized light as distinct from
such other characteristics as intensity and wavelength (von Frisch,
1949; von Frisch, Lindauer, and Daumer, 1960; Antrum and Stumpf,
1950; Vowles, 1950; Menzer and Stockhammer, 1951 ; Stockhammer,
1956, 1959; de Vries, 1956; Birukow and Busch, 1957; Jander, 1957;

PHOTORECEPTION 191

Liidtke, 1957; Jacobs-Jessen, 1959; Waterman, 1960; Jander and
Waterman, 1960; and a number of investigators studying Crustacea).
An alternative hypothesis suggests that differential responses to
various planes of polarization of light depend only upon detection of
different intensity patterns (Baylor and Smith, 1953; Stephens,
Fingerman, and Brown, 1953; Bainbridge and Waterman, 1957,
Baylor and Smith, 1958; Kalmus, 1958, 1959; de Vries and Kuiper,
1958; Baylor, 1959a, 1959 b; Smith and Baylor, 1960). This later
hypothesis is consonant to two basic facts: (1) differential reflection,
refraction, and scattering of polarized light by environmental features
or dioptric elements of the eye (Waterman, 1954) can produce
quadrants of maximal and minimal intensities whose positions are
determined by the direction of vibration ; arthropods respond to light-
intensity patterns. In further support of the hypothesis is the lack of
convincing proof of the existence of a polarization analyser in the eye.
No direct positive proof has been presented to contradict the opposing
hypothesis.

Support for the idea that the animals do see the plane of polariz-
ation as distinct from intensity patterns has recently been summarized
by Jander and Waterman (1960) in the following terms:

(1) In Daphnia and the beetle Bidessus, which could be made either
positively or negatively phototactic, reversal of this sign caused a
reversal of response to ordinary horizontal light intensity patterns, but
had no effect on the response to the plane of polarization of a vertically
directed beam. Therefore, polarized Hght orientation cannot be due to
horizontal patterns due to scattering and reflection of polarized light.

(2) In all animals, responses to horizontal patterns of Hght intensity
are of two sorts only, that is, towards or away, while four basic
orientation directions have been observed in a vertical beam of linearly
polarized light. Therefore, two different sensory mechanisms must be
involved.

(3) Response to the direction of vibration of polarized light showed
smaller deviations from the maximally preferred horizontal intensity
patterns. Therefore, both reactions cannot be intensity responses.

(4) When the dark intensity patterns due to the scattering of
polarized Hght are made artificially brighter, the crustacean Mysidium,
Daphnia, and Bidessus still respond as before to the plane of polariz-
ation. Therefore response to polarized light is not simply a photo-
tactic response to brightness.

(5) The crustacean Mysidium can distinguish between horizontal

192 THE PHYSIOLOGY OF INSECT SENSES

light patterns and polarized light presented simultaneously by reacting
to each as if it were alone. Therefore, polarized light and intensity
patterns are distinct visual qualities.

(6) For Daphnia light contrast reaction deteriorates at low overall
levels of illumination, while reactions to the plane of polarization do
not.

(7) Honeybees orient at the same angle with respect to the plane of
polarization under a polarizer as to that of corresponding quadrants
of the blue sky, despite large differences in intensity in the two situa-
tions.

(8) 'Since bees know the regional distribution of polarized light in
the sky (von Frisch, 1948, 1949), they must know the direction from
which it is coming. Yet they cannot with ordinary image vision infer
the direction of the original source from a hght pattern estabHshed by
reflection and refraction.'

(9) 'Under natural conditions reflection and refraction patterns due
to polarized sky light must to a considerable extent cancel each other
out because the plane of polarization is different in various parts of the
sky. In addition such patterns as do arise will be confused by much
more marked intensity patterns due to direct sunlight, which com-
prises up to 80 per cent of the total sky light, and to clouds and surface
details of the earth. To claim that bees could learn sky polarization
patterns from these reflection-refraction cues observed on their field
trips seems highly unlikely in view of such conditions.'

(10) Tn this view, experiments with spiders (Papi, 1955; Corner,
1957), ants (Jander, 1957) and bees (Jacobs- Jessen, 1959), have shown
in some instances that when the main intensity pattern is dislocated by
transposing the sun 180° with a mirror, the animals nevertheless main-
tain their orientation direction relative to the blue sky.'

Additional evidence in support of this hypothesis is derived from
the work of von Frisch (1960) and von Frisch, Lindauer, and Daumer
(1960).

If insects do indeed possess a discrete polarized light vision it
follows that there must be a mechanism in the eye for analysing the
plane of polarization. This could be located either in the dioptric
apparatus or in the retinal elements. It could depend on single re-
fraction according to the Brewster-Fresnel law or upon double re-
fraction. There is no convincing evidence for the existence of an
analyser in the dioptric apparatus, and the proposition that analysis
depends on simple reflection-refraction phenomenon at the corneal

PHOTORECEPTION 193

surface (Stephens et al., 1953; Baylor and Smith, 1953) does not con-
form to a number of experimental facts. For one thing, the necessary
properties would be realized only in an optically isotropic medium,
and the corneal lenses of Drosophila are anisotropic (Stockhammer,
1959). Nor is a mechanism of double refraction present. When the
dioptric apparatus is observed through a rotating polaroid no change
in the pattern or intensity of transmitted light is observed (Autrum and
Stumpf, 1950; de Vries and Kuiper, 1958; Stockhammer, 1959).
Furthermore, analyses of the ERG have failed to support the hypo-
thesis (Autrum and Stumpf, 1950; Baylor and Kennedy, 1958; de
Vries and Kuiper, 1958).

The on-effect of the ERG is known to depend in its magnitude on
the intensity of light. When the action potential resulting from the
stimulation of a total of twelve-fifteen ommatidia by polarized light
was measured there was no change in magnitude of response as the
plane of polarization was rotated. It seems unlikely, therefore, that the
eye as a whole acts as an analyser. Similar results obtained by stimula-
tion of a single ommatidium also indicated that the single ommati-
dium does not act as an analyser. However, when the effects of polar-
ized and non-polarized light of equal intensities were tested the
polarized light always elicited the greater response (Autrum and
Stumpf, 1950). From this it was concluded that each of the retinal
cells (eight in the honeybee and seven in Calliphora) is maximally
sensitive to light of a definite vibration direction, namely, that per-
pendicular to the radial direction of the cell within the ommatidium.
With unpolarized light each cell receives less than maximal stimula-
tion; with polarized light some receive maximal and some minimal.
The net result in the ERG is a response of greater magnitude than to
ordinary light because maximal and minimal effects do not cancel out.
Thus, the unit of analyses seems to be the single retinal cell, and the
central nervous system is presumed to integrate the pattern of light and
dark from each ommatidium.

This hypothesis is supported by the electrophysiological work of
Liidtke (1957), Naka and Kuwabara (1959), and Burkhardt and
Wendler (1960), as well as by field experiments of von Frisch (1950 a,
1950 b).

Von Frisch (1950 a, 1950 b) provided support for the hypothesis by
comparing the actions of bees dancing under a polaroid material with
the changes observed in the light-intensity pattern of the sky when
viewed through an artificial ommatidium constructed of eight seg-
ments of polaroid material each oriented differently. When this model

194 THE PHYSIOLOGY OF INSECT SENSES

was held up to the blue sky a different pattern of intensity was seen
for every part of the sky, depending on the position of the sun. When
a sheet of polaroid was placed between the model and the sky in such
an orientation as not to alter the pattern and another sheet was placed
similarly over dancing bees the dances maintained proper orientation.
When the covering sheets were rotated until that over the model
altered the pattern the orientation of the dance changed. The artificial
pattern produced in the model could usually be found in another part
of the sky, and it was to this direction that the bees danced. When the
pattern produced by interposing the polaroid between the model and
the sky happened to have no normal counterpart in the sky the dance
of the bees under the same polaroid became disorganized. If the
pattern as seen through the model happened to occur simultaneously
in two areas of the sky the bees also oriented their dances in two
directions.

In so far as electrophysiological evidence is concerned not all
workers have been able to confirm the findings of Autrum and Stumpf.
Baylor and Kennedy (1958), employing especially refined control of
intensity and wavelength, were unable to detect any difference in the
ERG related to the presence or absence of polarization. Nor were de
Vries and Kuiper (1958) able to detect changes in the ERG when the
eye was subjected to flicker consisting alternately of vertically and
horizontally polarized light. The insects themselves failed to show
optomotor reactions in these circumstances. These results not only
challenge the proposition that analysers are located in the retinal
cells but also question even the abihty of the eye to possess any
analyser at all.

Contrary to these findings is the report of Burkhardt and Wendler
(1960). By measuring the ERG of individual retinal cells of Calliphora
with intracellular electrodes and controlHng for adventitious polariz-
ation of oblique incident light at the corneal and crystalline cone
boundaries, these workers were able to show that the amplitude of the
ERG is dependent upon light intensity and the direction of vibration
of Hnearly polarized light. A rotation of the plane of polarization from
the most-effective position to the least-effective position changes the
illumination potential of the cell quantitatively to the same degree as a
50 per cent reduction of light intensity.

If the retinal cells are indeed analysers it should be possible to find
within them the necessary structural modification. Menzer and Stock-
hammer (1951) proposed that birefringence in the rhabdoms coupled
with a mechanism for extinguishing one ray provided means of

PHOTORECEPTION 195

analysis. They suggested : (1) that the rhabdomeres consists of a doubly
refracting central column and a surrounding cylinder of different
refractive index, and (2) that the optical axis of each of the retinal cells
is different. They reported that there was birefringence in the rhab-
doms and that between crossed polarizers each rhabdomere ex-
tinguishes about four times during one complete rotation. Many ob-
jections to this hypothesis were raised by de Vries (1956), who also
pointed out that the report of birefringence was based upon examina-
tions of fixed material and that fresh material showed at best too httle
birefringence (less than 1 per cent) to meet the requirements of the
hypothesis (de Vries, 1956; de Vries, Spoor, and Jielof, 1953; de
Vries and Kuiper, 1958).

An alternative hypothesis to the effect that analysis is accomplished
by means of dichroic molecules was advanced by de Vries (1956).
Organic molecules tend to absorb light polarized in the direction of the
long axis of the molecules. If all of the photosensitive molecules of a
sense cell are arranged in parallel and different sense cells are oriented
differently the visual organ possesses the necessary mechanism for
analysing polarized light. Indeed, this mechanism works to a minor
extent in the human eye, which can detect the polarization of Hght by
Haidinger brushes. These are seen as a yellow shape (figure of eight)
and are due to orientation (about 50 per cent) of the molecules of the
yellow macular pigment. Nobody, however, has been able to demon-
strate dichroism in the insect eye, even though evidence from electron-
microscopy proves that the microvilli of each pair of rhabdomeres are
oriented in a different horizontal direction and the visual pigment is
presumed to be in the microvilH. De Vries (1956) concluded that if
dichroism is present it must be less than 10 per cent; nevertheless, he
did not consider that this negative finding disproved his hypothesis.
Later, however, he concluded that the oriented structure of the
rhabdomeres has nothing to do with the detection of polarized fight
and that insects do not *see' polarization (de Vries and Kuiper, 1958).
He and his co-worker were influenced in drawing these conclusions by
failure to detect analysers and by the experiments of Baylor and Smith
(1953, 1957). Stockhammer (1959), in discussing the dichroism
hypothesis, advanced a number of arguments from structure which
support the idea.

Baylor and Smith (1953, 1957) proposed that the responses of
animals to the plane of polarization of light is in fact a response to a
non-uniform distribution of intensities in the pattern of fight reflected
from a substratum (in the case of terrestrial animals) or scattered by

196 THE PHYSIOLOGY OF INSECT SENSES

various particles in water (Baylor, 1959 a, 1959 b; Smith and Baylor,
1960; Kalmus, 1958, 1959; Bainbridge and Waterman, 1957). Some
of the evidence upon which this view is based has already been given.
Additionally, it is claimed that the ability of insects to respond photo-
tacticly to polarized light depends upon whether the substrate is black
or white (e.g., Kalmus, 1958).

Evidence against the reflexion hypothesis has already been stated.
In so far as bees are concerned, recent experiments by von Frisch,
Lindauer, and Daumer (1960) emphasize the inadequacy of the hypo-
thesis. With bees, direct analysis of the sky's light during the dance on
a horizontal comb requires use of the dorsal areas of the eyes, while
recognition of reflexion patterns of the substrate requires the ventral
areas. Masking of the upper portions impairs orientation of the dance,
but masking of the lower portions does not. Furthermore, walking
bees orient as well on a white substratum as on a black, glossy one.

Whatever may be the mechanism, the fact remains that many in-
vertebrates behave as though they can detect differences in the plane of
polarization of light and utilize this abihty in their economy of living.

FORM PERCEPTION

The intrinsic characteristics of the stimulus for a photoreceptor are
intensity, wavelength, and plane of polarization, but the occurrence of
stimuli also varies in space and time. In the visual field there is at any
given time a definite spatial arrangement of the kinds and amounts of
radiant energy. In other words, there are patterns of light. The fineness
of form perception depends upon the accuracy with which photo-
receptors can detect this pattern, that is, upon their resolving power,
and ultimately upon the faithfulness with which the spatial and tem-
poral relations are represented in the central nervous system.

The capacity of the dioptric apparatus to form images that are
reasonably distinct by human standards is easily demonstrated by
taking photographs in the focal plane of excised lens systems
(Gottsche, 1 852 ; Exner, 1891; and many others). In those eyes in which
each ommatidium is optically isolated from its neighbours by envelop-
ing pigment cells and in which the rhabdom Hes immediately beneath
the crystalline cone, a small inverted image is produced at the base of
each cone. Eyes of this type are termed apposition eyes and are
characteristic of diurnal insects (Fig. 6). Miiller (1826) proposed that
these minute images were physiologically unimportant as images. He
visualized each ommatidium of the apposition eye as a device for

PHOTORECEPTION 197

gathering light from a narrow sector of the visual field and projecting
it as a point on the retinal field. Since each ommatidium sampled, as it
were, a mean intensity from a particular area, the point of light pro-
jected by all the ommatidia composed an erect mosaic image.

Nocturnal and crepuscular insects characteristically possess eyes in
which the rhabdoms are situated at a considerable distance from the
crystalline cone. The enveloping pigment migrates back and forth in
their respective cells as the eye is light- or dark-adapted (Fig. 92).

A B

Fig. 92. Image formation in the apposition eye, A, and the superposition
eye, B. a-f, paths of light rays; P, pigment; IP, position of pigment
when eye is in the light adapted condition; Rh, rhabdom. (Redrawn
from Wigglesworth, 1939 after Kuhn.)

When this eye is dark-adapted the iris pigment moves distally into a
position surrounding the cones. In this position each cone is isolated
from its fellows, but the rhabdoms are now no longer shielded. Under
these conditions light entering one ommatidium at an angle is not
confined to that ommatidium as in the apposition eye, but may travel
to neighbouring rhabdoms (Fig. 92). The image produced by a single
ommatidium is an erect image which becomes larger the farther away
from the crystalline cone it is projected. At the plane of the rhabdoms
it is so large that it overlaps many rhabdoms where it is superimposed
on the images similarly produced by the other lenses corresponding to

198 THE PHYSIOLOGY OF INSECT SENSES

those rhabdoms. Microscopic examination of the image at the level
of the rhabdom reveals a large, somewhat fuzzy, erect image. As the
microscope is brought to focus closer and closer to the crystalHne
cones, the rays of light forming the image are seen to retreat into each
cone. The image at the level of the rhabdoms is a compound of super-
imposed images; hence, this type of eye is called a superposition eye.
When the pigment in the superposition eye migrates proximally so that
the rhabdoms are shielded from one another the eye functions Hke an
apposition eye. It is noteworthy in this connexion that Antrum (1961)
in measuring the visual acuity of a white mutant of CaUiphora in which

Fig. 93. Concentric lamellae
of different refractive
index in the dioptric
apparatus of the om-
matidium. The refrac-
tive index is greatest
along the axis xy. (Re-
drawn from Exner,
1891.)

Proximal Surface

there was no screening pigment in the eye found no difference between
the mutant and the wild type. Since the superposition is not total, that
is, not all the images are congruent, it could be argued that the differ-
ence between the mosaic image and the superposition image is one of
degree.

In order to explain image formation in the two types of eyes Exner
(1891) proposed that the crystalline cone acted as a lens cyHnder. If it
were constructed of concentric lamellae of different refractive indices
(Fig. 93), the refractive index would be greatest along the longitudinal
axis and would decrease towards the periphery. The path of a ray of

PHOTORECEPTION 199

light in such a cylinder is diagrammed in Fig. 94. As a ray xm enters
the cylinder it is refracted by the surface of each lamella a 'b\ but
since the lamellae decrease in refractive index progressively from A^to
ab, the ray is bent less and less until it re-enters regions of higher re-
fractive index, where its path is now directed back towards xy. It finally
emerges at^*. If such a cylinder is as long as its focal length it will pro-

a

fc

^„^-' 1

..^^-^^^

^

h'

a

— ^

^N^- -

/

ym

"^

\^

c

d

Fig. 94. Course followed by light through a cylinder abed corresponding to
the dioptric apparatus of the ommatidium. A, path of a ray of light.
B, course of a spherical wave, a -b', layer of equal refringence ; x, point
of origin of wave; mm, mini, etc.", successive positions of wave front
indicating changes due to refraction accompanying passage through
cylinder; y, focal point. (Redrawn from Exner, 1891.)

duce an inverted image at its base. If a cyhnder is twice as long as its
focal length it will produce an inverted image at its midpoint (Fig. 95).
The rays will then diverge from the focal point and continue on straight
paths. Rays that enter the cylinder obliquely (e.g., from the right) will
emerge from the same side (i.e., the right).

If, in the apposition eye, the ommatidium is the optical unit, if it
perceives Hght primarily from points on or near its axis, if there is
minimal overlap of ommatidial visual fields and the image which has
o

200 THE PHYSIOLOGY OF INSECT SENSES

significance for the insect is a mosaic of point of light, then visual
acuity should depend upon the number and spacing of luminous points
received from the object in the visual field. There should be exact
correspondence between the minimum angle subtended by adjacent
ommatidia and the minimum angle subtended by two points that are
perceived as separate (the reciprocal of the visual angle is termed the
visual acuity).

c

11 ,

■ ''" ^--^-^ ,--- - -.

a

!'-'''

.--'•--''^^'^>3><'''^^----.-

~^-.^

__::-^'^"^^^- -

^~^-.^

^'--""' ,

^ -.^

b

■-/'

Fig. 95. Course of light rays through a lens cylinder abed whose length is
(A) equal to its focal length (B) twice the focal distance, m, p, n, q,
rays from objects in the visual field; q', n', p', m', course of rays after
passage through cylinder; xy, optic axis; zy, inverted image of object
at base of cylinder. Note that in B an inverted image is produced at
mid-point of cylinder. Rays continue through to produce an erect
image at the rhabdoms. (Redrawn from Exner, 1891.)

Acuity does indeed depend upon the angle between adjacent direc-
tional elements, but reducing this angle to increase acuity has limita-
tions. There is no improvement in acuity when the angle is so close that
visual overlapping occurs, and there is now evidence that such over-
lapping is much greater than would be predicted from Muller's theory
(de Vries, 1956; Burtt and Catton, 1954). Furthermore, as Barlow
(1952) has pointed out, visual acuity is related to the resolving power of
each ommatidial lens.

The resolving power of a lens is limited by its optical properties and
the wave nature of light. The diameter of the ommatidium is critical,
and diffraction must be taken into consideration. Because of dif-

PHOTORECEPTION 201

fraction, light coming through a small circular aperture produces an
Airy pattern, a central point of maximum intensity with two concen-
tric dark rings alternating with light rings. If two point sources are to
be resolved they must be far enough separated so that the bright image
centre of one falls no closer than within the first dark ring of the Airy
pattern of the other. Miiller (1826) had said that an apposition omma-
tidium was sensitive only to light from a point on or near its axis. It is
maximally sensitive on its axis and minimally sensitive away from its
axis, so the diameter determines how close two points can be if one is to
stimulate maximally and the other minimally. In the honeybee, if the
eye resolves points separated by twice the minimal ommatidial angle
the performance of the ommatidia has approached the theoretical
limit set by diffraction.

It is by no means certain, however, that the ommatidium is the
optical unit of the eye. Refined techniques which permit illumination
of single rhabdomers (in Diptera) have prompted the suggestion that
it is the retinal cell which is the functional unit of the eye (de Vries,
1956; de Vries and Kuiper, 1958; see also Burkhardt and Wendler,
1960; Antrum and Burkhardt, 1961). By cutting transversely through
an ommatidium in the region of the rhabdom and examining the cut
end of the rhabdom as light is transmitted through the crystalline cone
it can be seen that each rhabdom 'looks' at a different part of the visual
field. Each visual field has a diameter of about 4 degrees, and the fields
of neighbouringcells overlap so that the centre of one field lies approxi-
mately at the edge of the neighbouring field. The field of the whole
ommatidium is about 8 degrees, and it overlaps with that of its
neighbours. As with the entire ommatidium when it is analysed as a
unit, the field of the individual cell is controlled by the diameter of the
rhabdom and the diffraction of Hght. If a point source of light being
viewed at the cut end of a rhabdomer is displaced more than 2
degrees from the optical axis of the retinal cell, it shines again, but
more weakly, by reason of the diffracted hght outside of the first dark
ring of the Airy pattern. Once a beam has reached a rhabdomer it will,
unless it deviates too far from the optical axis, be reflected repeatedly
at the boundaries of the rhabdomer, since it has a higher refractive
index than surrounding tissues. In this manner the rhabdomer acts as
a wave-guide, and the cell captures light (see also Demoll, 1917).
Since the internal reflexion decreases as the angle between the incident
beam and the axis of the cell increases, there is a directional sensitivity
in these cells, the Stiles-Crawford effect of human vision (de Vries
and Kuiper, 1958).

202 THE PHYSIOLOGY OF INSECT SENSES

If one recalls that the ommatidium is a multi-innervated sensillum
and that the neurons of a sensillum may possess quite different
characteristics (e.g., the chemoreceptive hair of flies), the indepen-
dence of retinal cells is not unreasonable. On the other hand, the
significance of the pattern of responses from single retinal cells cannot
be simple because of the intricacy of synaptic connexions in the lamina
ganglionaris. For every retinal cell whose axon traverses the area there
are many that synapse together on giant monopolar neurons. Further-
more, there are many local chiasmata each involving several ommati-
dia. Some success towards understanding the extent to which an
ommatidium is a functional unit of the eye has been achieved by the
sophisticated functional analyses of movement perception by Hassen-
stein (1951, 1959) and Hassenstein and Reichardt (1956).

Behavioural measurements of visual acuity (usually derived from
optomotor responses) have not always corresponded to the minimal
ommatidial angle (Hecht and Wolf, 1929; Hecht and Wald, 1934;
Wolf, 1933 a, 1933 b). Compared with man, whose visual acuity is from
2 to 2-5, the maximal visual acuity of the bee is 0-017 and ofDrosophila,
0-0018. In the case of the bee, the value obtained corresponds exactly
to the smallest ommatidial angle (0-90-1-00 degrees) measured by
Baumgartner (1928). In other cases, however, the correspondence has
been less exact, the acuity being either greater or less.

Where the visual acuity has been less than expected in terms of
ommatidial angles a number of explanations have been offered. Von
Buddenbrock and Schulz (1933) proposed that ommatidia might be
acting not as units but as pairs or groups. This hypothesis would also
explain observed differences of visual acuity at high and low light
intensities. As with man, the relation between visual acuity and the
logarithm of intensity is described by a sigmoid curve (Fig. 96)
(Hecht and Wald, 1934) which has been interpreted as an integral
population curve denoting the number of retinal elements active at
each intensity. In both the bee and Drosophila visual acuity and in-
tensity discrimination, as measured by optomotor reactions, begin at
approximately the same intensity and change rapidly within two log
units (Wolf, 1933 a, 1933 b; Hecht and Wald, 1934).

It is also possible that visual-acuity measurements obtained from
behavioural studies depend upon what part of the eye is being stimu-
lated at the time. Viewed from the outside, the lens surface of a com-
pound eye is an approximate and imperfect hemisphere. The relative
area with respect to the head, and the number of ommatidia, vary from
species to species and often between the sexes. Each of the ommatidia

PHOTORECEPTION

203

200

Log I (mlLLUamberts)

Fig. 96. Relation between visual acuity and intensity of illumination
for the eye of the bee. (Redrawn from Hecht and Wolf, 1929.)

Log I (tn'dULamberts)
A

Log I (mUlUamberts)
B

Fig. 97. Relation between visual acuity and illumination in bees in which
A, the central area of the eye has been opaqued; B, the anterior region
of the eye has been opaqued. The broken line represents the curve for
normal individuals. (Redrawn from Hecht and Wolf, 1929.)

204 THE PHYSIOLOGY OF INSECT SENSES

faces in a different direction. As a result of these differences in align-
ment and the fact that the lenses may differ in size from one part of the
eye to another, the angle subtended by adjacent ommatidia is not
constant for any given eye (Baumgartner, 1928; del Portillo, 1936).
Measurements of visual acuity in insects in which various areas of the
eye have been blacked out have shown that the eye is not uniform
(Hecht and Wolf, 1929). When the centre of an eye, where the minimal
ommatidial angle is 1 degree as compared with 4 degrees at the
periphery, is blacked out the visual acuity decreases (Fig. 97). It also
decreases as the number of functional units is reduced (Fig. 97).

Another technique for measuring visual acuity was employed by
Burtt and Catton (1954) following their discovery in Locusta, Phormia,
and Calliphora that the movement of illuminated objects in the visual
field was accompanied by specific electrical discharges in the ventral
nerve cord. By employing these discharges as a criterion of movement
perception, Burtt and Catton were able to ascertain that the overlap of
ommatidial fields, at least in Locusta, is much greater than heretofore
realized. The angle of the visual field is 20 degrees; the average angle
subtended by adjacent ommatidia is about 2-4 degrees in the longi-
tudinal and 1 degree in the vertical meridian. Responses to movement
were obtained to angular displacements in the visual field as small as
0-16 degree of arc. The threshold of acuity was independent of light-
and dark-adaptation. Burtt and Catton suggested that the extensive
overlapping makes possible the perception of an angle smaller than
that subtended by adjacent ommatidia; however, the experiments of
de Vries (1956) suggesting independent action of individual retinal
cells offer an alternative explanation.

Whatever their visual acuity may be, there is no doubt that insects
can resolve patterns. A series of field experiments with honeybees has
shown them to be fairly efficient in this respect (Hertz, 1929 a, 1929 b,
1931, 1933a, 1934, 1935a, 1935b, 1935c, 1937a). By training bees to
associate specific black figures on a white background with food. Hertz
was able to demonstrate that contrast, contour, and the degree of sub-
division of the form are perceived. She concluded that the most greatly
divided pattern is preferred, but that the pattern is recognized as such.
The results of Zerrahn (1933) and Wolf and Zerrahn-Wolf (1937),
based on field and laboratory experiments in which bees were per-
mitted to choose patterns spontaneously, are in essential agreement ;
however, these workers derived different conclusions. They contended
that the pattern as such has no meaning. The number of choices of each
pattern was proportional to the length of its contours, that is, the

PHOTORECEPTION 205

length of edges of black against white. This finding suggested that
recognition and discrimination is based merely upon the degree of
transitory stimulation produced in the compound eye. To test this
hypothesis, Wolf and Zerrahn-Wolf mapped on translucent co-
ordinate paper the points of intersection of the axes of the ommatidia
with a place located at the same distance from the eye as were patterns
used in testing. They then placed the co-ordinate paper over a pattern
and counted the number of ommatidia whose visual fields would be
black and the number whose fields would be white. They also counted
the number of ommatidia in which there would be a black-white shift
when the pattern was moved over a unit distance in any direction.
Calculations could then be made of the relationships between the area
of a pattern and the degree of subdivision necessary to provide a con-
stant transition value. Predictions of discrimination based on these
calculations were realized in field tests. Other lines of evidence were in
agreement. For example, bees conditioned to flickering fields of equal
size but different flicker frequencies exhibited choices that were directly
proportional to the flicker frequency. These findings are in agreement
with observations on the natural behaviour of bees in the field (Wolf,
1933 c). The honeybee reacts much more effectively to moving flowers
than to stationary ones.

All of these experiments suggested that movement is of greater sig-
nificance in form perception than the ability to resolve stationary
patterns (cf. Exner, 1891). In its response to flicker the insect eye
exhibits many of the characteristics of the human eye. It obeys Talbot's
law (the brightness of a fused light produced by flashing it on and off" is
visually equivalent to the product of the actual source of illumination
and the proportion of the time it is on) and the Ferry-Porter law
(critical frequency, that is, the frequency at which a flashing light fuses,
is proportional to the logarithm of intensity) as does the human eye
(Wolf and Zerrahn-Wolf, 1935 b; Wolf, 1933 b). It is, however, more
eflflcient at detecting flicker than is the eye of man. Some of the early
values obtained for insects by optomotor methods were : 55 per second
for the bee (Wolf, 1933 b); i 60 per second for dragonfly larvae
(Salzle, 1932; Crozier, Wolf, and Zerrahn-Wolf, 1937). More highly
developed behavioural techniques indicated that maximum values are
actually much higher. The critical flicker frequency for Calliphora is
about 265 per second, depending upon the number of ommatidia
stimulated. The maximum of 265 per second drops to 60-165 per
second when only one to four ommatidia are stimulated. A similar
relationship holds for the honeybee (Autrum and Stoecker, 1950).

206 THE PHYSIOLOGY OF INSECT SENSES

Studies of the electrical response (ERG) of the eye to flickering light
agree beautifully with recent behavioural studies. As described
earlier in this chapter, there are oscillations in the ERG which are
synchronized with the stimulus flashes. In *fast* eyes (e.g., Calliphora,
Apis) the synchrony persists at even higher frequencies than those
judged by optomotor responses to be fusing. In 'slow' eyes (e.g.,
Dixippus, Tachycines, Periplanetd) the responses, both behaviourally
and electrophysiologically, are slower. These findings and the con-
clusions of others (Zerrahn, 1933 ; Wolf and Zerrahn-Wolf, 1937) that
the decisive feature of pattern recognition is the transitory stimulation
ofommatidia suggested to Antrum (1948 a, 1948 b, 1949, 1952, 1955 b)
the idea of temporal resolution.

The concept is based upon the assumption that the ommatidium is
the optical unit of the eye. In rapidly flying insects the angle subtended
by ommatidia in the horizontal direction is approximately twice that
in the vertical (del Portillo, 1936). During horizontal movement a
point being observed remains in the visual field of a single ommatidium
longer; consequently, the summation time for this stimulus is pro-
longed. Points of a pattern that succeed each other rapidly fuse less
readily. Accordingly, the capacity for discrimination is improved
through movement (Antrum, 1949). Studies of ERGs show that two
points of light succeeding each other within a critical time fuse if they
pass over the ommatidial grating vertically, but act as individual
stimuli if they pass horizontally. The electrophysiological responses of
*fast' eyes reveal features which tend to favour reception of repetitive
stimuli of short duration. The ERG shows that the on-effect for stimuli
of 1-To^ second depends only upon intensity. It is independent of
duration. The off"-effect depends on the product of intensity and
duration. Furthermore, in the 'fast' eyes, according to Antrum, a
positive off'-effect originating in the lamina ganglionaris prevents sus-
tained depolarization as occurs in 'slow' eyes, and thus 'prepares' the
eye for the next stimulus.

Insects also appear to be responsive to stroboscopic (apparent)
motion. In contrast to the negative findings of Gaffron (1934),
Antrum and Stocker (1952) demonstrated that this phenomenon
exists provided that certain critical conditions are met. For strobos-
copic motion to be perceived the time sequences must be short for those
insects ('fast') whose eyes possess high temporal resolution and long
for those ('slow') species whose eyes possess low temporal resolution.
A close relationship exists between the times that are important in
motion perception, whether real or apparent, and the times for which

PHOTORECEPTION 207

the receptors can still separate separate stimuli. Motion is perceived
as a vector without spatial connotations when two stimulated omma-
tidia are separated by an unstimulated one (Hassenstein, 1951).

OCELLI

Many insects possess dorsal ocelli in addition to compound eyes. The
distribution of ocelli within the Class is erratic and is of no assistance
in the problem of interpreting the role that these organs play.

The structure is basically similar in all ocelli. It consists of a trans-
parent cornea which may be perfectly flat on both surfaces, biconvex,
or plano-convex. Underlying the cornea there is frequently a trans-
parent layer of corneagen cells and beneath this a layer of retinal cells
very similar in structure to those of the compound eyes. In the more
complex ocelli the corneagen cells may be thickened to form a aux-
iliary refractory organ. The retinal cells are grouped together in units
of two, three, or four, and each cell of a unit contributes to a central
rhabdom. In the lateral ocellus of Sympetrum there are about 675
photoreceptor cells. Beneath them lies a reflecting layer, the tapetum.
As the sensory axons proceed proximally they soon form synaptic
connexions with the intra-ocellar terminals of the ocellar nerve fibres
(Cajal, 1918; Ruck, 1957). The ocellar nerve of the cockroach consists
of about twenty-five fibres, of which four are very large - 6-10 \l.
In the dragonfly {Sympetrum) there is one giant fibre (25-38 [x) and
three or four ranging in diameter from 4-13 [x. The remainder are
1 [I or less. The neurocytes lie in the brain (Cajal, 1918 ; Satija, 1958 a,
1958 b).

Although the corneal lens is capable of forming fairly sharp images,
everyone who has studied the optics of the system agrees that the image
has little significance, since it falls a considerable distance behind the
retina (Homann, 1924; Parry, 1947; Cornwall, 1955).

Electrical events have been studied in the ocelli of Locusta migra-
toria migratoroides (Parry, 1947; Hoyle, 1955; Burtt and Catton,
1958), Periplaneta americana (Ruck, 1957, 1958 a; Goldsmith and
Ruck, 1958), Blaberus craniifer (Ruck, 1957, 1961a), Melanoplus
bivittatus (Ruck, 1957), Apis mellifera (Goldsmith and Ruck, 1958;
Ruck, 1958 b). Pachydiplax longipennis and Phormia regina (Ruck,
1958 b), Libellula luctuosa (Ruck, 1961 a), Libellula vibrans (Ruck,
1961 b), Anax Junius, Aeschna sp., and Sympetrum ribicundulum
(Ruck, 1961 c). Because different workers employed different record-
ing situations it is difficult to compare their results and to reconcile

208 THE PHYSIOLOGY OF INSECT SENSES

differences in the wave form and sign of the ERG. The most compre-
hensive analysis is that based upon extensive studies of the ocellus of
the cockroach and dragonfly (Ruck, 1958 a, 1958 b, 1961 a, 1961 b). In
these insects the total electrical response to light of high intensity

oscilloscope

different

Fig. 98. Components of the ERG of the model ocellus. One photoreceptor
cell {left) with an expanded sensory ending, and an axon which makes
synaptic contact with one ©cellar nerve fibre. The two units are con-
tained in an electrolyte-filled compartment. A 'corneal' electrode
enters at left; a 'nerve' electrode lifts the nerve into air; an indifferent
electrode is placed far to the right. Four components of the ERG are
shown, each in one repetition of the model. Active sites for each
component are shaded. Current at active sites is indicated by arrows.
Each component appears at corneal and nerve electrodes. (Redrawn
from Ruck, 1961 a.)

consists of four components, of which two originate in the retinal cells
and two in the ocellar nerve (Fig. 101). The first event is a generator
potential arising in the distal end of the receptor. It evokes a depolariz-
ation in the axonal region of the receptor. This in turn evokes a
hyperpolarizing postsynaptic potential in the fibres of the ocellar

PHOTORECEPTION 209

nerve. This component is easily discernible in the ocelli of dragonflies
but rarely detected in cockroaches. Ruck (1961 b) has suggested that
the postsynaptic potential is not induced directly by the electrical
events in the receptor axon, but rather by the liberation of a trans-
mitter substance at the end of the receptor axon. In the absence of the
hyperpolarizing potential the fibres of the ocellar nerve discharge
impulses (component 4). In the dragonfly the ocellar nerve in the dark
discharges spontaneously, while the cockroach nerve usually dis-
charges only at off. It is inferred that there is spontaneous receptor cell
activity which modulates the rhythmic spontaneous discharges in the
ocellar nerve.

The ocellus of the dragonfly is a very sensitive receptor. Corneal
illumination as low as 10~^ ft.-candles produces an ERG (Ruck,
1958 b). Furthermore, it is to be expected that sensitivity is enhanced
by the high degree of convergence of receptor axons on ocellar nerve
fibres, multiple synapsing of single receptor axons with ocellar nerve
fibres, and the presence of a white tapetum (Ruck, 1961 a).

The ability of the ocellus to respond to flickering light varies from
species to species. The flicker fusion frequencies of Apis, Pachydiplax,
and Phormia are very high, being, respectively, 250-265, 200, and
+220 per second. For the cockroach maximum flicker fusion fre-
quencies range from 45 to 60 per second (Ruck, 1958 a). Thus, the
ocelli of those insects with *fast' eyes are *fast' and those with 'slow'
eyes, *slow'. As with the compound eyes. Ruck (1958 b) found that
there is no general relation between flicker fusion frequency and
sensitivity or rate of dark adaptation. Furthermore, since ocelli do not
possess a lamina ganglionaris and since in the dragonfly the generator
potential can respond to higher rates of flicker (220 per second) than
can the receptor axon responses, the postsynaptic potential and the
ocellar nerve impulses, the difference between 'fast' and 'slow' ocelli
is a fundamental characteristic of the photoreceptor cell itself. The
evidence of Antrum and Gallwitz (1951) that the abihty of 'fast' com-
pound eyes to follow high rates of flicker depends upon electrical
interaction between the receptor cells and neurons of the optic
ganglion cannot be extended to ocelH.

Dorsal ocelli are sensitive to the same spectrum of wavelengths as
are compound eyes. The ocelH of Periplaneta have a single sensitivity
peak at 500 mi^., while those of the honeybee show sensitivity maxima
at 490 mjj. and at 335-340 m[L (Goldsmith and Ruck, 1958). Because in
the honeybee the ERG produced by ultra-violet stimulation is quali-
tatively different from that produced by stimulation of light of 490 mjjt,

210 THE PHYSIOLOGY OF INSECT SENSES

Goldsmith and Ruck (1958) believe that there are two types of receptor
in this organ.

The exact nature of the contribution of dorsal ocelli to behaviour is
still elusive. Alone they are insufficient for phototaxis (Homann, 1924 ;
Bozler, 1926; Muller, 1931; Wellington, 1953; Cornwall, 1955; but
compare Gotze, 1927 and Wellington, 1953). They are not, however,
totally without effect in light-directed behaviour. Elimination of ocelli
may cause phototaxis to be temporarily reversed (Muller, 1931),
reduce the speed of response to Hght (Bozler, 1926; Muller, 1931), or
alter the direction of response to two light sources (Muller, 1931). In
Periplaneta the characteristic persistent daily rhythm depends upon
stimulation of the oceUi (Marker, 1956). As a consequence of these
actions the ocelli are generally considered to be * stimulatory' organs
(Bozler, 1926; Wolsky, 1930, 1931, 1933).

STEMMATA

With the possible exception of dermal receptors sensitive to Hght, the
lateral oceUi or stemmata are the sole visual organs possessed by many
larval insects. Structurally they vary from forms similar to the dorsal
ocelli of adults (e.g., larval ocelli of Tenthredinidae) to mere pigment
spots equipped with refractive bodies. In blowfly larvae the photo-
receptors are small groups of vacuolated cells pocketed among the
hypodermal cells of the cephalic region (Bolwig, 1946).

The most thoroughly studied stemmata are those of lepidopterous
larvae (Grenacher, 1897; Hesse, 1901; Dethier, 1942, 1943). Each
resembles an individual ommatidium. It possesses two lenses, one
corneal, the other analogous to the crystalline cone of compound eyes.

Table 9

(From Dethier, 1953 c)

Optical constants of the six ocelli of Isia Isabella

Ocellus

Minimum Angle

of Resolution

(Radians)

/ value

Dioptric
Value

Type of
Image

Focal
Distance

1

2
3
4
5
6

M X 10--^
1-2 X 10-2
5-3 X 10-3
7-1 X 10-3
7-8 X 10-3
1 X 10-2

0-93
0-89
0-50
0-67
0-73
100

14,285
14,084
14,285
14,285
13,888
14,084

Triple
Triple
Single
Single
Single
Triple

070
071
070
070
072
071

PHOTORECEPTION 211

The first is a thick converging concavoconvex or biconvex meniscus
lens.

If the secretion products of the three corneagen cells fuse incom-
pletely during the formation of this lens, as is typical of certain ocelli,
the cornea acts as a triple lens and forms three images (PL IV). The
crystaUine lens, like the cornea, is a thick, converging biconvex lens,
either unitary or tripartite to correspond with the cornea. It also forms
more or less distinct, real, inverted images. The receptive layer, like
that of the ommatidium, consists of distal and proximal retinula cells.
Some of the optical constants obtained from the lenses of the six ocelli
of the caterpillar Isia Isabella are given in Table 9. The peculiar optical
properties of the lens system of ommatidia as described by Exner
(1 89 1) are not characteristic of these lenses.

Whether the ocellus acts as an eye or merely as a photoreceptor
depends primarily on : (1) the ability of the dioptric apparatus to form
images ; (2) the location of the image plane with respect to the rhab-
dom; (3) the ability of the rhabdom to receive any images formed. By
determining the image space (that region of space in which all possible
positions of the image are situated) from the optical constants and by
actual measurement of the length of the vertical retinal elements,
Dethier has shown (Fig. 99) that regardless of the distance of an
object from the ocellus, the image falls somewhere along the length
of the rhabdom. In this manner a degree of accommodation is obtained
which would otherwise be impossible with a fixed-lens system. The
ocellus is analogous in this respect to a fixed-focus camera.

Vision in insects bearing simple eyes is the sum of the capacities
of all units operating jointly. Hundertmark's (1936, 1937) experiments
have shown that larvae which orient to black shapes on a white back-
ground are directed towards the black-and-white boundary. Dethier
has postulated that the basic principles of mosaic vision apply equally
well to lateral ocelH and compound eyes. Since each ocellus gathers
light from the area at which it is directed and concentrates this light
at some point along the vertical rhabdom, the six pairs of ocelli
together form twelve points of light. The ocelli are so arranged on the
head that little or no overlapping of the visual field exists, hence each
spot represents the intensity from a different area. Combined they
provide an exceedingly coarse mosaic of intensities. The paucity of
units is compensated for in part by the habit, characteristic of many
larvae, of moving the head from side to side while advancing. By this
klinotactic-like behaviour a larger visual field is examined, and the
recognition of changes in intensity, as at a black-white boundary, is

212

THE PHYSIOLOGY OF INSECT SENSES

*,-—

Fig. 99. A. Camera lucida drawing of an ocellus of the caterpillar of
Isia Isabella to illustrate the congruity of image space and retinal space
and the kind of accommodation offered by a vertical rhabdom.
Oi, O2, O3, representative objects; Ii, Ig, I3, positions of resultant
images; F, focal point; CO, cornea; CR, crystalline lens; DR, distal
retinula; PR, proximal retinula. The broken line in the cornea repre-
sents the ideal lens curvature. B. Graphic construction of images
formed. O, object; Iv, virtual image; Ir, real image; Fi, focal point of
corneal lens; Fg, focal point of crystalline lens; F, focal point of entire
lens system. (From Dethier, 1943.)

PHOTORECEPTION 213

facilitated by this motion. Even in those species (e.g., Cicindela) where
the ocellus consists of a larger number of retinal elements beneath
each lens, it is believed that all ocelli function together as a unit.

Little is known about the parameters of the stimuli for stemmata.
It has been demonstrated that larvae are sensitive to a wide band of
the spectrum and seem to be able to respond to the plane of polariz-
ation (Wellington, Sullivan, and Green, 1957). Electrophysiological
studies have only just been undertaken with these receptors. The
ERG recorded with the active electrode on the ocellar nerve and the
reference electrode in the abdomen is simple in appearance. It con-
sists of a prominent positive wave followed at high intensities of
stimulation by a slow negative wave. It is believed to consist of two
components (Ishikawa and Hirao, 1960 a, 1960 b). The stemmata
become completely dark-adapted in about one hour and completely
light-adapted in about ten minutes. They follow flicker up to
25-30 per second.

References

ADAMS, J. R. (1961) The location and histology of the contact chemo-
receptors of the stable fly, Stomoxys calcitrans L. Doctoral
Dissertation, Rutgers Univ., New Brunswick, N.J.

ADRIAN, E. D. (1937) Synchronized reactions in the optic ganglion of
Dytiscus. J. Physiol, 91, 66-89.

ALEXANDER, R.D. (1957) Sound production and associated behavior
in insects. Ohio J. Sci., 57, 101-13.

ALEXANDROWicz, J. s. (1951) Muscle receptor organs in the abdo-
men of Homarus vulgaris and Palinurus vulgaris. Quart. J. Micr.
Sci., 92, 163-99.

ALEXANDROWICZ, J. s. (1952 a) Receptor elements in the thoracic
muscles of Homarus vulgaris and Palinurus vulgaris. Quart. J.
M/cr.5'd., 93, 315-46.

ALEXANDROWICZ, J. s. (1952 b) Muscle receptor organs in the
Paguridae. /. Mar. Biol. Assoc. U.K., 31, 277-86.

ALEXANDROWICZ, J. s. (1954) Notes on the nervous system in
Stomatopoda. IV. Muscle receptor organs. Pubbl. Staz. Zool.
Napoli, 25, 94-in.

ALEXANDROWICZ, J. s. (1956) Receptor elements in the muscles of
Leander serratus. J. Mar. Biol. Assoc. U.K., 35, 129-44.

ALVAREZ-BUYLLA, R., and DE ARELLANO, J. R. (1953) Local res-
ponses in Pacinian corpuscles. Am. J. Physiol., 127, 237-50.

AMOURiQ, L. (1957) Optimum vital et role des scolopidies du scape et
du pedicelle antennaires dans la recherche du preferendum
hygrometrique de Gryllus bimaculatus L. Bull. Soc. Hist. Nat.
Afr.Nord., 48,137-47.

AMOURIQ, L. (1959) Histologic et role des organes chordotonaux de
I'antenne dans le comportement hygropathique de Gryllus.
Thesis, 163 pp. Alger.

ANDERSEN, L. w., and BALL, H. J. (1959) Antennal hygroreceptors
of the milkweed bug, Oncopeltus fasciatus (Dallas) (Hemiptera,
Lygaeidae). Ann. Ent. Soc. Amer., 52, 279-84.

ANDERSON, A. L. (1932) The sensitivity of the legs of common butter-
flies to sugars. /. Exp. Zool, 63, 253-9.

ANDERSON, D. B. (1936) Relative humidity of vapor pressure deficit.
Ecology, 17, 277-82.

ARAB, Y. M. (1958) Behavioural response to electrical stimulation of
chemoreceptors in the blowfly. /. Ins. Physiol., 2, 324-9.

214

REFERENCES 215

ARAB, Y. M. (1959) Some chemosensory mechanisms in the blowfly
Phonnia regina Meigen. Bull. College Sci, Univ. Baghdad, 4, 77-85.

AUGER, D., and FESSARD, A. (1928) Observations surl'excitabilit^de
I'organe tympanique. C.R. Soc. Biol. Paris, 99, 400-1.

AUTRUM, H. (1936) Ober Lautausserungen und Schallwahrnehmung
bei Arthropoden. I. Untersuchungen an Ameisen Eine allge-
meine Theorie der Schallwahrnehmung bei Arthropoden. Zeit.
vergl. Physiol., 23, 332-73.

AUTRUM, H. (1940) Ober Lautausserungen und Schallwahrnehmung
bei Arthropoden. II. Das Richtungshoren von Locusta und
Versuche einer Hortheorie fiir Tympanalorgane vom Locusti-
dentyp. Zeit. vergl. Physiol., 28, 326-52.

AUTRUM, H. (1941) Uber Gehor und Erschutterungssinn bei Locusti-
den. Zeit. vergl. Physiol., 28, 580-637.

AUTRUM, H. (1942) Schallempfang bei Tier und Mensch. Naturwiss.,
30, 69-85.

AUTRUM, H. (1943) Uber kleinste Reize bei Sinnesorganen. Biol. Zbl,
63, 209-36.

AUTRUM, H. (1948 a) iJber das zeitliche Auflosungsvermogen des
Insektenauges. Nachr. Akad. Wiss. Gottingen, Math.-Phys., 1948,
8-12.

AUTRUM, H. (1948 b) Zur Analyse des zeitlichen Auflosungsvermo-
gens des Insektenauges. Nachr. Akad. Wiss. Gottingen, Math.-
Phys., 1948, 13-18.

AUTRUM, H. (1948 c) Uber Energie-und Zeitgrenzen der Sinnesemp-
findungen. Naturwiss., 35, 361-9.

AUTRUM, H. (1949) Neue Versuche zum optischen Auflosungsver-
mogen fliegender Insekten. Experientia, 1, 271-7.

AUTRUM, H. (1950) Die Belichtungspotentiale und das Sehen der
Insekten (Untersuchungen an Calliphora und Dixippus). Zeit.
vergl. Physiol., 32, 176-7.

AUTRUM, H. (1951) Erregungsvorgange und Leistungen des Insekten-
auges. Verhandlungen Deutschen ZooL Gesellschaft Wilhelms-
haven, 1951, 133-43.

AUTRUM, H. (1952) Uber zeitliches Auflosungsvermogen und Primar-
vorgange im Insektenauge. Naturwiss., 39, 290-7.

AUTRUM, H. (1953) Elektrobiologie des Auges. Klinische Wochen-
schrift, 31, 241-5.

AUTRUM, H. (1954 a) Die Sehvorgange in den Augen der Insekten.
Zeit. Umschau Wiss. Tech., 54.

AUTRUM, H. (1954 b) Formensehen im menschlichen und tierischen
AugQ.Zeit. Umschau Wiss. Tech., 1954.

AUTRUM, H. (1955 a) Die spektrale Empfindlichkeit der Augen-
mutation white-apricot von Calliphora erythrocephala. Biol.
Zbl., 74, 515-24.

216 THE PHYSIOLOGY OF INSECT SENSES

AUTRUM, H. (1955 b) Die Zeit als physiologische Grundlage des

Formensehens. Studium Generale, 8, 526-30.
AUTRUM, H. (1955 c) Analyse physiologique de la reception des sons

chezles orthopteres. In Colloque sur U acoustique des Orthopteres,

ed. R. G. Busnel. Fascicule Hors Serie des Ann. Epiphytes,

pp. 338-55.
AUTRUM, H. (1958) Electrophysiological analysis of the visual systems

in insects. Exp. Cell Res. (SuppL), 5, 426-39.
AUTRUM, H. (1959) Non-photic receptors in lower forms. In Hand-
book of Physiology, ed. J. Field. Am. Physiol. Soc, Washington,

D.C.
AUTRUM, H. (1961) Die Sehscharfe pigmentfreier Fazettenaugen von

Calliphora erythrocephala. Biol. Zbl., 80, 1-4.
AUTRUM, H., and burkhardt, d. (1961) Spectral sensitivity of

single visual cells. Nature, 190, 639.
AUTRUM, h., and gallwitz, u. (1951) Zur Analyse der Belich-

tungspotentiale des Insektenauges. Zeit. vergl. Physiol., 33,

407-35.
AUTRUM, H., and Hoffmann, c. (1960) Diphasic and monophasic

responses in the compound eye of Calliphora. J. Ins. Physiol., 4,

122-7.
AUTRUM, H., and Hoffmann, e. (1957) Die Wirkung von Pikrotoxin

und Nikotin auf das Retinogramm von Insekten. Z. Naturf., 126,

752-7.
AUTRUM, H., and Schneider, w. (1948) Vergleichende Untersuch-

ungen iiber den Erschiitterungssinn der Insekten. Zeit. vergl.

Physiol., 31, 77-88.
AUTRUM, H., and stocker, m. (1952) tJber optische Verschmel-

zungsfrequenzen und stroboskopisches Sehen bei Insekten.

Biol. Zbl.,11, 129-52.
AUTRUM, H., and stoecker, m. (1950) Die Verschmelzungsfre-

quenzen des Bienenauges. Zeit Naturf., 5b, 38-43.
AUTRUM, H., and stumpf, h. (1950) Das Bienenauge als Analysator

fiir polarisiertes Licht. Zeit. Naturf, 5b, 1 16-22.
AUTRUM, H., and stumpf, h. (1953) Electrophysiologische Unter-

suchungen iiber das Farbensehen von Calliphora. Zeit. vergl.

Physiol., 35, 71-104.
AZIZ, s. A. (1957 a). The reactions of the Desert Locust, Schistocerca

gregaria (Forsk.), (Orthoptera, Acrididae) to physical factors,

with special reference to relative humidity. Bull. Entomol. Res.,

48, 515-32.
AZIZ, s. A. (1957 b) Probable hygroreceptors in the Desert Locust,

Schistocerca gregaria (Forsk.) (Orthoptera, Acrididae). Indian J.

Entomol, 19, 164-70.

REFERENCES 217

BAINBRIDGE, R., aiid WATERMAN, T. H. (1957) Polarized light and

the orientation of two marine Crustacea. J. Exp. Biol., 34,

342-64.
BARANY, E. (1946 a) A theory of binocular visual acuity and an

analysis of the variability of visual acuity. Acta. Opthal Kbh.,

24, 63-92.
BARANY, E. (1946 b) Some statistical observations on the methods in

threshold determinations in general with particular regard to de-
termination of visual acuity and subliminal addition. Acta

Opthal Kbh., 24, 113-27.
BARLOW, H. B. (1952) The size of the ommatidia in apposition eyes.

/. Exp. Biol, 29, 667-74.
BARROWS, w. M. (1907) The reactions of the pomace fly, Drosophila

ampelophila Loew, to odorous substances. /. Exp. ZooL, 4,

515-37.
BASSLER, u. (1957) Zur Funktion des Johnston-Organes bei der

Orientierung der Stechmiicken. Naturwiss., 44, 336-7.
BASSLER, u. (1958) Versuche zur Orientierung der Stechmiicken: Die

Schwarmbildung und die Bedentung des Johnstonschen Organs.

Zeit. vergl. Physiol., 41, 300-30.
BAUER, L. (1938) Geschmacksphysiologische Untersuchungen an

Wasserkafern. Zeit. vergl. Physiol., 26, 107-20.
BAUER s, c. (1953) Der Fixierbereich des Insectenauges. Zeit. vergl.

Physiol., 34, 589-605.
BAUMGARTNER, H. (1928) Der Formcnsinn und die Sehscharfe der

Bienen. Zeit. vergl. Physiol., 1, 56-143.
BAUNACKE, w. (1912) Statische Sinnesorgane bei den Nepiden.

Zool. Jahrb. Abt. Anat. Ontog., 34, 179-346.
BAYLOR, E. R. (1959a) Polarized hght responses mLimulus. Biophysical

Soc, Program and Abstracts, 1959 Meetings.
BAYLOR, E. R. (1959b) The responses of snails to polarized light. /.

Exp. Biol, 36, 369-76.
BAYLOR, E. R., and KENNEDY, D. (1958) Evidence against a polariz-
ing analyzer in the bee eye. Anat. Rec, 132, 411.
BAYLOR, E. R., and SMITH, F. E. (1953) The orientation of Cladocera

to polarized light. Amer. Nat., 87, 98-101.
BAYLOR, E. R., and SMITH, F.E. (1957) Diurnal migration of plankton

crustaceans. In Recent Advances in Invertebrate Physiology . Univ.

Oregon Press, pp. 21-35.
BAYLOR, E. R., and SMITH, F. E. (1958) Extra-ocular polarization

analysis in the honey bee. Anat. Rec, 132, 41 1-12.
BECHT, G. (1959) Studies on insect muscles. Bijdragen Tot de Dier-

kunde, 29, 5-40.
BEGG, M., and HOGBEN, L. (1946) Chemoreception of Drosophila

melanogaster. Proc. Roy. Soc. Lond. B, 133, 1-19.

218 THE PHYSIOLOGY OF INSECT SENSES

BEiDLER, L. M. (1953) Properties of chemoreceptors of tongue of rat.

/. NeurophysioL, 16, 595-607.
BEIDLER, L. M. (1954) A theory of taste stimulation. J. Gen. Physiol.,

38, 133-9.
BEIDLER, L. M. (1960) Physiology of taste. The Physiologist, 3,

5-12.
BEIDLER, L. M. (1961) The chemical senses. Ann. Rev. Psychol., 12,

363-88.

BEIDLER, L. M., FISHMAN, I. Y., and HARDIMAN, C. W. (1955)

Species differences in taste responses. Am. J. Physiol., 181,
235-9.

BENTLEY, E. w. (1944) The biology and behaviour of Ptinus tectus
Boie (Coleoptera, Ptinidae), a pest of stored products. V.
Humidity reactions. /. Exp. Biol., 20, 152-8.

BENZ, G. (1956) Der Trockenheitssinn bei Larven von Drosophila
melanogaster. Experientia, 12, 297-8.

BERLESE, A. (1909) GH Insetti. Societa Editrice Libraria, Milano.

BERNHARD, c. G. (1942) Isolation of retinal and optic ganglion res-
ponse in the eye ofDytiscus. J. Neurophysiol., 5, 32-48.

BERNHARD, C. G., GRANIT, R., and SKOGLAND, C. R. (1942) The

breakdown of accommodation-nerve as a model sense-organ.

/. Neurophysiol., 5, 55-68.
BERTHOLF, L. M. (1931 a) Reactions of the honeybee to light. J. Agr.

Res., 42, 379-419.
BERTHOLF, L. M. (1931 b) The distribution of stimulative efficiency in

the ultraviolet spectrum for the honeybee. J. Agr. Res., 43,

703-13.
BERTHOLF, L. M. (1932) The extent of the spectrum for Drosophila

and the distribution of stimulative efficiency in it. Zeit. vergl.

Physiol., 98, 32-64.
BIRUKOW, G. (1953) Menotaxis im polarisierten Licht bei Geotrupes

silvaticus Panz. Naturwiss., 40, 611-12.
BIRUKOW, G., and busch, e. (1957) Lichtkompassorientierung

beim Wasserlaufer Velia currens F. (Heteroptera) am Tage und

zur Nachtzeit. II. Orientierungsrhythmik in verschiedenen

Lichtbedingungen. Zeit. TierpyschoL, 14, 184-203.
boistel, j. (1953) Etude functionelle des terminaisons sensorielles

des antennes d' Hymenopteres. C.R. Soc. Biol. Paris, 147,

1683-8.
boistel, J., and coraboeuf, e. (1953) L'activit^ electrique dans

I'antenne isolee de Lepidoptere au cours de I'etude de I'olfaction.

C.R. Soc. Biol. Paris, 147, 1172-5.
boistel, J., lecompte, J., and coraboeuf, e. (1956) Quelques

aspects de I'etude electrophysiologique des recepteurs sensoriels

des antennes d'hymenopteres. Insectes Sociaux, 3, 25-31.

REFERENCES 219

BOLWiG, N. (1946) Senses and sense organs of the anterior end of the

house fly larvae. Videns. Medd. Dansk Natur. Forening., 109,

82-217.
BORROR, D. J. (1954) Audio-spectographic analysis of the song

of the cone-headed grasshopper, Neoconocephalus ensiger

(Harris) (Orthoptera: Tettigoniidae). Ohio. J. Sci\, 54, 297-

303.
BOWNESS, J. M., and wolken, j. j. (1959) A light-sensitive yellow

pigment from the house-fly. /. Gen. Physiol., 42, 779-92.
bozler, e. (1926) Experimentelle Untersuchungen liber die Funktion

der Stirnaugen der Insekten. Zeit. vergl. Physiol., 3, 145-82.
BRAUNS, A. (1939) Morphologische und physiologische Unter-
suchungen zum Halterenproblem. Zool. Jahrb., 59, 245-390.
bretschneider, f. (1921) Uber das Gehirn des Wolfsmilchsch-

warmers (Deilephila Euphorbiae). Jena. Zeit. Naturwiss., 57,

423-62.
BRINK, F., and POSTERNAK, J. M. (1948) Thermodynamic analysis of

the relative effectiveness of narcotics. /. Cell. Comp. Physiol., 32,

211-33.
BUCK, J. B. (1937) Studies on the firefly. II. The signal system

and color vision in Photinus pyralis. Physiol. Zool., 10, 412-

19.
BUCKMANN, D. (1954) Die Leistungen der Schwereorientierung bei

dem im Meeressande grabenden Kafer Bledius bicornis Grm.

(Staphylinidae). Zeit. vergl. Physiol, 36, 488-507.
BUCKMANN, D. (1955 a) Zur Leistung des Schweresinnesbeilnsekten.

Naturwiss., 42, 78-79.
BUCKMANN, D. (1955 b) Zur Frage der Funktion der Insektenfiihler

als Schweresinnesorgan. Naturwiss., 42, 79.
VON BUDDENBROCK, w. (1935) Fine neue Methode zur Erfor-

schung des Formensehens der Insekten. Naturwiss., 23, 98-100.
VON BUDDENBROCK, w. (1952) Vergleichcndc Physiologic. Sinnes-

physiologie (Vol. I). Birkhauser, Basil.
VON BUDDENBROCK, w., and SCHULZ, E. (1933) Beitrage zur

Kenntnis der Lichtkompassbewegung und der Adaptation des

Insektenauges. Zool. Jahrb. Zool. Physiol., 52, 513-36.
BULLOCK, T. H. (1953) Comparative aspects of some biological trans-
ducers. Fed. Proc, 12 (3), 666-72.
BULLOCK, T. H. (1957) Neuronal integrative mechanisms. In Recent

Advances in Invertebrate Physiology. Univ. Oregon Publications,

pp. 1-20.
BULLOCK, T. H.(195 8) Parameters of integrative action of the nervous

system at the neuronal level. Exp. Cell Res., 5, 323-37.
BULLOCK, T. H. (1959) Initiation of nerve impulses in receptor and

central neurons. Rev. Modern Physics, 31 (2), 504-14.

220 THE PHYSIOLOGY OF INSECT SENSES

BURKHARDT, D. (1954) Rhythmischc Erregungen in ben optischen
Zentren von Calliphora erythrocephala. Zeit. vergl. Physiol., 36,
595-630.

BURKHARDT, D. (1960) Action potentials in the antennae of the blow-
fly {Calliphora erythrocephala) during mechanical stimulation.
/. Ins. Physiol., 4, 138-45.

BURKHARDT, D., and AUTRUM, H. (1960) Die Belichtungspotentialc
einzelner Sehzellen von Calliphora erythrocephala Meig. Zeit.
Naturf., 15b, 612-16.

BURKHARDT, D., and SCHNEIDER, G. (1957) Die Antennen von
Calliphora als Anzeiger der Fluggeschwindigkeit. Z. Naturf.,
12b, 139-43.

BURKHARDT, D., and WENDLER, L. (1960) Ein direkter Beiveis fiir
die Fahigkeit einzelner Sehzellen des Insektenauges, die Schwing-
ungsrichtung polarisierten Lichtes zu analysieren. Zeit. vergl.
Physiol, 43, 687-92.

BURSELL, E. (1957) The effect of humidity on the activity of tsetse
flies. /. Exp. Biol, 34, 42-51.

BURSELL, E., and EWER, o. w. (1950) On the reactions to humidity
of Peripatopsis moseleyi (Wood-Mason). /. Exp. Biol., 26,
335-53.

BURTT, E. T., and CATTON, w. T. (1952) Nerve impulses originating
from the compound eye of the locust. Nature, 170, 285.

BURTT, E. T., and CATTON, w. T. (1954) Visual perception of move-
ment in the locust. /. Physiol., 125, 566-80.

BURTT, E. T., and CATTON, w. T. (1958) Visual acuity of the com-
pound eye in three species of insects. XV Internat. Congr. ZooL,
London.

BUSNEL, M. c. (1953) Contributions a Tetude des emissions acous-
tiques des Orthopteres. Ann. Inst. Nat. Recherche Agron., Ser. C,
Ann. Epiphyties, 4, 333-421.

BUSNEL, M. c, and BUSNEL, R. G. (1956) Sur une phonocinese de
certains Acridiens a des signaux acoustiques synthetiques. C.R.
Acad. Sci. Paris, 242, 292-5.

BUSNEL, R. G. (ed.) (1955a) Colloque sur V acoustique des Orthopteres.
Fascicule Hors Serie des Ann. Epiphyties, 448 pp.

BUSNEL, R. G. (1955 b) Misc en evidence d'un caractere physique
reactogene essentiel de signaux acoustiques synthetiques
d^clenchant les phonotropismes dans le regne animal. C.R. Acad.
Sci.,Paris,240,Un-9.

BUSNEL, R. G., and CHAVASSE, P. (1950) Recherches sur les emis-
sions sonores et ultrasonores d'Orthopteres nuisibles a I'agri-
culture. Etude des frequences. Nuovo Cimento, Suppl. 7 (Series
IX), 470-86.

REFERENCES 221

BUSNEL, R. c, and LOHER, w. (1953) Recherches sur le comporte-
ment de divers Acridoidea males seumis a des stimuli acous-
tiques artificiels. C.R. Acad. Sci., Paris, 237, 1557-9.

BUSNEL, R. G., and LOHER, w. (1954) Recherches sur le comporte-
ment de divers males d'acridiens a des signaux acoustiques
artificiels. Ann. Sci. Nat. ZooL, Ser. 11, 16, 271-81.

BUSNEL, R. G., LOHER, w., and PASQUiNELLY, F. (1954) Recher-
ches sur les signaux acoustiques synth^tiques reactogenes pour
divers Acrididaec^. C.R. Soc. Biol. Paris, 148, 1987-91.

BUTENANDT, A., BECHMANN, R., STAMM, D., and HECKER, E.

(1959) Ober den Sexuallockstoff des Seidenspinners Bombyx

mori. Reindarstellung und Konstitution. Zeit. Naturf., 14b,

283-4.
CAJAL, s. R. (1909) Nota sobre la estructura de la retina de la Mosca

{Musca vomitoria L.). Trabajos deb Lab. de investig. biologicas

Univ. Madrid, 7, 217-57.
CAJAL, s. R. (1918) Observaciones sobre la estructura de los ocelos y

vias nerviosas ocelares de algunos insectos. Trabajos del Lab. de

investig. biologicas Univ. Madrid, 16, 109-39.
CAJAL, s. R., and sanchez, d. (1915)Contribucionalconocimiento

de los centros nerviosos de los insectos. Trabajas del Lab. de

investig. biologicas Univ. Madrid, 13, 1-167.
CAMERON, J. w. M. (1938) The reactions of the housefly, Musca

domestica Linn., to light of different wave-lengths. Canad. J. Res.,

D, 16, 307-42.
CHADWiCK, L. E. (1939) Some factors which affect the rate of

movement of the wings in Drosophila. Physiol. ZooL, 12,

151-60.
CHADWICK, L. E. (1947) The respiratory quotient of Drosophila in

flight. Biol. Bull., 93, 229-39.
CHADWICK, L. E., and DETHiER, V. G. (1947) The relationship be-
tween chemical structure and the response of blowflies to tarsal

stimulation by aliphatic acids. /. Gen. Physiol., 30, 255-62.
CHADWICK, L. E., and DETHIER, V. G. (1949) Stimulation of tarsal

receptors of the blowfly by aliphatic aldehydes and ketones. J.

Gen. Physiol., 32, 445-52.
CHILD, c. M. (1894) Ein bisher wenig beachtetes antennales Sinnes-

organ der Insekten, mit besonderer Berucksichtigung der

Culiciden und Chironomiden. Zeit. wiss. ZooL, 58, 475-528.
COOK, w. c. (1926) The effectiveness of certain paraffin derivatives in

attracting flies. /. Agr. Res., 2il, 347-58.
CORN WELL, p. p. (1955) The functions of the ocelli of Calliphora

(Diptera) and Locusta (Orthoptera). /. Exp. Biol., 32, 217-37.
CRAGG, J. B., and COLE, p. (1956) Laboratory studies on the chemo-

sensory reactions of blowflies. Ann. Appl. Biol., 44, 478-91.

222 THE PHYSIOLOGY OF INSECT SENSES

CRESCiTELLi, F., and J A HN,T.L. (1939) The electrical response of the
dark-adapted grasshopper eye to various intensities of illumina-
tion and to different quahties of Hght. /. Cell. Comp. Physiol., 19,
105-12.

CRESCITELLI, F., and JAHN,T.L. (1942) Oscillatory electrical activity
from the insect compound eye. /. Cell. Comp. Physiol., 19, 47-
66.

CROMBIE, A. c. (1944) On the measurement and modification of the
olfactory responses of blowflies. /. Exp. Biol., 20, 159-66.

CROZiER, w. J., WOLF, E., and ZERRAHN-WOLF, G. (1937) Critical
illumination and critical frequency for response to flickered light,
in dragonfly larvae. J. Gen. Physiol., 20, 363-92.

DAHLBERG, A. c, and PENCZEK, E. s. (1941) The relative sweetness
of sugars as affected by concentration. A^. Y. State Ag. Exp. Sta.
Tech. Bull, 258, 3-12.

DANNEEL, R., and ZEUTZSCHEL, B. (1957) Uber den Feinbau der
Retinula bei Drosophila melanogaster. Zeit. Naturforsch., 126,
580-3.

D AUMER, K. (1956) Reizmetrische Untersuchungen des Farbensehens
der Bienen. Zeit. vergl. Physiol., 38, 413-78.

DAUMER, K. (1958) Blumenfarben, wie sie die Bienen sehen. Zeit.
vergl. Physiol., 41, 49-110.

DAViES, J. T. (1953 a) Olfactory stimulation; some ideas and possible
model systems. Inter. Perfumer, 3, 17-22.

DAVIES, J. T. (1953b) L'odeur et la morphologic des molecules. In-
dustrie de la Parfumerie, 8, 74.

DAVIES, J. T., and taylor, f. h. (1959) The role of adsorption and
molecular morphology in olfaction : the calculation of olfactory
thresholds. Biol. Bull, 111, 222-38.

DAVIS, H. (1961) Some principles of sensory receptor action. Physiol.
Rev., 41, 391-416.

DAY, M. f. (1950) The histology of a very large insect, Macropanesthia
rhinocerus Sauss. (Blattidae). Australian J. Sci. Res, Ser. B,
Biol. Sci., 3, 61-75.

DEBAisiEUX, p. (1935) Organes scolopidiaux des pattes d'insectes.

I. Lepidopteres et Trichopteres. La Cellule, 44, 273-314.
DEBAISIEUX, p. (1938) Organes scolopidiaux des pattes d'insectes.

II. La Cellule,41,ll-\02.

DEBAUCHE, H. (1935) Recherchcs sur les organes sensoriels anten-

naires de Hydropsyche longipennis Curt (Trichoptera-Hydro-

psychidae). La Cellule, 44, 46-84.
DEBAUCHE, H. (1936) Etude cytologique et compar^e de Torgane de

Johnston des insectes. La Cellule, 45, 77-148.
DEEGENER, P. (1909) Obcr ciu neues Sinnesorgan am Abdomen der

Noctuiden. Zool. Jahrb. Anat., 27, 631-50.

REFERENCES 223

DEMOLL, R. (1917) Die Sinnesorgane der Arthropoden ihr Bau und

ihre Fimktioju Frieder. Vieweg & Sohn, Braunschweig, 243 pp.
DEONIER, c. c. (1938) Effects of some common poisons in sucrose

solutions on the chemoreceptors of the housefly, Musca domes-

tica L. J. Econ. Ent., 31, 742-45.
DEONIER, c. c. (1939) Responses of the blowflies Cochliomyia

americana C. & P. and Phormia regina Meigen, to stimulation of

the tarsal receptors. Ann. Ent. Soc. Amer., 32, 526-32.
DETHIER, V. G. (1941) The function of the antennal receptors in lepi-

dopterous larvae. Biol. Bull., 80, 403-14.
DETHIER, V. G. (1942) The dioptric apparatus of lateral ocelli. I.

The corneal lens. /. Cell. Comp. Physiol., 19, 301-13.
DETHIER, V. G. (1943) The dipotric apparatus of lateral ocelli. II.

Visual capacities of the ocellus. /. Cell. Comp. Physiol., 22,

115-26.
D E T H I E R, V. G. ( 1 947 a) The role of the antennae in the orientation of

carrion beetles to odor. J.N. Y. Ent. Soc, 55, 285-93.
DETHIER, V. G. (1947b) Chemical Insect Attractants and Repellents.

Blakiston, Philadelphia.
DETHIER, V. G. (1947 c). The response of hymenopterous parasites to

chemical stimulation of the ovipositor. J. Exp. Zool., 105,

199-208.
DETHIER, V. G. (1951) The limiting mechanism in tarsal chemo-

reception. /. Gen. Physiol., 35, 55-65.
DETHIER, V. G. (1952a) Tastc sensitivity to homologous alcohols in

oil. Fed. Proc.,ll,?>A.
DETHIER, V. G. (1952 b) The relation between olfactory response

and receptor population in the blowfly. Biol. Bull, 102, 111-

17.
DETHIER, V. G. (1953 a) Chemorcception. In Insect Physiology, ed.

K. D. Roeder. John Wiley, New York, pp. 544-76.
DETHIER, V. G. (1953 b) Summation and inhibition following contra-
lateral stimulation of the tarsal chemoreceptors of the blowfly.

Biol. Bull, 105, 257-68.
DETHIER, V. G. (1953 c) Vision. In Insect Physiology, ed. K. D. Roeder,

John Wiley, New York, pp. 488-522.
DETHIER, V. G. (1954a) Olfactory responses of blowflies to aliphatic

aldehydes. /. Gen. Physiol, 37, 743-51.
DETHIER, V. G. (1954b) The physiology of olfaction in insects. Ann.

N. Y. Acad. Sci., 58,139-51.
DETHIER, V. G. (1955 a) The physiology and histology of the con-
tact chemoreceptors of the blowfly. Quart. Rev. Biol., 30,

348-71.
DETHIER, V. G. (1955b) Tarsal chcmorcccptors of the houscfly. Proc.

Roy. Ent. Soc. Lond., (A)3, 87-90.

224 THE PHYSIOLOGY OF INSECT SENSES

DETHiER, V. G. (1956) Chemoreccptor Mechanisms. In Molecular

Structure and Functional Activity of Nerve Cells. A.I.B.S.,

Washington, D.C., pp. 1-30.
DETHIER, V. G. (1957) Chemoreception and the behavior of insects.

In Survey of Biological Progress, Vol. Ill, ed. B. Glass. Academic

Press, New York, pp. 149-83.
DETHIER, V. G. (1961) Behavioral aspects of protein ingestion by the

blowfly Phormia regina Meigen. Biol. Bull, 111, 456-70.
DETHIER, V. G. (1962) Chemoreceptor mechanisms in insects. Symp.

Soc. Exp. Biol. 16, 180-96.
DETHIER, V. G., and ARAB, Y. (1958) Effect of temperature on

the contact chemoreceptors of the blowfly. /. Ins. Physiol., 2,

153-61.
DETHIER, V. G., and CHADWiCK, L. E. (1947) Rejection thresholds

of the blowfly for a series of aliphatic alcohols. /. Gen. Physiol.,

30, 247-53.
DETHIER, V. G., and CHADWICK, L. E. (1948a) Chemoreception in

insects. Physiol. Rev., 28, 220-54.
DETHIER, V. G., and CHADWICK, L. E. (1948 b) The Stimulating effect

of glycols and their polymers on the tarsal receptors of blowflies.

J. Gen. Physiol, 32, 139-51.
DETHIER, V. G., and CHADWICK, L. E. (1950) An analysis of the re-
lationship between solubility and stimulating effect in tarsal

chemoreception. J. Gen. Physiol, 33, 589-99.
DETHIER, V. G., and EVANS, D. R. (1961) The physiological control of

water ingestion in the blowfly. Biol. Bull, 111, 108-16.

DETHIER, v. G., EVANS, D. R., and RHOADES, M. V. (1956) SomC

factors controlling the ingestion of carbohydrates by the blowfly.
Biol Bull, 111, 204-22.

DETHIER, V. G., HACKLEY, B. E., and WAGNER-JAUREGG,T. (1952)

Attraction of flies by iso-valeraldehyde. Science, 115, 141-2.
DETHIER, V. G., and RHOADES, M. V. (1954) Sugar preference-
aversion functions for the blowfly. /. Exp. Zool, 126, 177-

203.
DETHIER, v. G., and WOLBARSHT, M. L. (1956) The electronmicros-

copy of chemosensory hairs. Experientia, 12, 335.
DETHIER, v. G., and YOST, M. T. (1952) Olfactory stimulation of

blowflies by homologous alcohols. /. Gen. Physiol, 35, 823-39.
DiAKONOFF, A. (1936) Contributions to the knowledge of the fly

reflexes and the static sense in Periplaneta americana L. Arch.

Neerl Physiol, 21, 104-29.
DINGLE, H. (1961) Flight and swimming reflexes in giant water bugs.

Biol Bull, 111, 117-28.
DODDS, s. E., and ewer, d. w. (1952) Effect of desiccation on the

humidity response of Tenebrio. Nature, 170, 758.

REFERENCES 225

DONNER, K. o., and KRISZAT, G. (1950) Die electrophysiologisch

bestimmte Sensitivitatsverteilung des Fliegenauges im sichtbaren

Spektrum. Arkiv. Zool., 42A, 1-7.
DOSTAL, B. (1958) Riechfahigkeit und Zahe der Riechsinncselemente

bei der Honigbiene. Zeit. vergl. Physiol, 41, 179-203.
DRESDEN, D., and NiJENHUis, E. D. (1958) Fibre analysis of the

nerves of the second thoracic leg in Periplaneta americana. Proc.

Kon. Nederl. Akad. Wetensch. Amsterdam, C, 61, 213-23.
EDNEY, E. B. (1957) The Water Relations of Terrestrial Arthropods.

Cambridge Univ. Press, 109 pp.
EGER, H. (1937) Ober den Geschmackssinn von Schmetterlingsraupen.

Biol. Zbl, 57, 293-308.
EGGERS, F. (191 1) Ober des thoracale Tympanal-Organ der Noctuid-

en. Sitzber. Naturf. Ges. Univ. Jurjew (Dorpat.), 20, 138-45.
EGGERS, F. (1919) Das thoracale bitympanale Organ einer Gruppe

der Lepidoptera Heterocera. Zool. Jahrb. Anat., 41, 273-376.
EGGERS, F. (1923) Ergebnisse von Untersuchungen am Johnstonschen

Organ der Insekten und ihre Bedeutung fiir die allgemeine

Beurteilung der stiftfUhrenden Sinnesorgane. Zool. Anz., 57,

224-40.
EGGERS, F. (1924) Zur Kenntnis der antennalen stiftfUhrenden

Sinnesorgane der Insekten. Zeit. Morph. Okol. Tiere, 2, 259-349.
EGGERS, F. (1925) Versuche iiber das Gehor der Noctuiden. Zeit.

vergl. Physiol., 2, 297-314.
EGGERS, F. (1926a) Beobachtungen iiber die Verkniipfung von Gehor

und Fluchtinstinkt bei Eulenschmetterlingen. Schriften Naturw.

Vereins Schleswig-Holstein, 17, 325-33.
EGGERS, F. (1926 b) Die mutmassliche Funktion des Johnstonschen

Sinnesorgans bei Gyrinus. Zool. Anz., 68, 184-92.
EGGERS, F. (1927) Nahere Mitteilungen iiber Johnstonsche Sinnes-
organe und iiber das Ausweichmermogen der Taumelkofer. Zool.

Anz., 71, 136-56.
EGGERS, F. (1928) Die stiftfiihrenden Sinnesorgane. Zool. Bausteine,

2, 1-353.
EGUCHi, E., NAKA, K.-L, and KUWABARA, M. (1962) The develop-
ment of the rhabdom and the appearance of the electrical

response in the insect eye. J. Gen. Physiol., 46, 143-57.
EHRENSVARD, G. (1942) t)ber die Primarvorgange bei Chemozep-

torenbeeinflussung. Acta Physiol. Scand., 3 (Suppl. 9), 1-151.
ERHARDT, E. (1916) Zurkenntnisdcrlnncrvierung und Sinnesorgane

der Fliigel von Insekten. Zool. Jahrb. Anat. Ontog., 39, 293-

334.
ERNST, E. (1957) Der Einfluss der Luftfeuchtigkeit auf Lebensdauer

und Verhalten verschiedener Termitenarten. Acta Tropica

(Basel), 14, 97-156.

226 THE PHYSIOLOGY OF INSECT SENSES

EVANS, D. R. (1958) The differential sensitivity of the two neurons in

the contact chemoreceptors of the blowfly. Anat. Rec, 132,

433-4.
EVANS, D. R. (1961 a) Control of the responsiveness of the blowfly to

water. Nature, 190, 1132-3.
EVANS, D. R. (1961 b) Depression of taste sensitivity to specific sugars

by their presence during development. Science, 133, 327-8.
EVANS, D. R., and mellon, dcf. (1962a) Electrophysiological stud-
ies of a water receptor associated with the taste sensifla of the

blowfly. /. Gen, Physiol., 45, 487-500.
EVANS, D. R., and mellon, dcf. (1962 b) Stimulation of a primary

taste receptor by salts. /. Gen. Physiol., 45, 651-61.
exner, s. (1891) Die Physiologic der Facetierten Augen von Krebsen

und Insecten. Franz Deuticke, Leipzig and Wien, 206 pp.
EYER, J. R., and MEDLER, J. T. (1940) Attractiveness to codling moth

of substances related to those elaborated by heterofermentative

bacteria in baits. /. Econ. Ent., 33, 933-40.
EYZAGUiRRE, c, and KUFFLER, s. w. (1955) Proccss of excitation

in the dendrites and in the soma of single isolated sensory nerve

cells of the lobster and crayfish. /. Gen. Physiol, 39, 87-119.
FAUST, R. (1952) Untersuchungen zum Halterenproblem. Zool.

Jahrb. Allg. Zool. Physiol., 63, 325-66.
FERGUSON, J. (1951) Relations between thermodynamic indices of

narcotic potency and the molecular structure of narcotics. In

Mecanisme de la Narcose. Coll. Internat. Centre Nat. Rech. Sci.,

26, 25-39.
FERNANDEZ-MORAN, H. (1956) Fine Structure of the insect retinula

as revealed by electron microscopy. Nature, 111, 742-3.
FERNANDEZ-MORAN, H. (1958) Fine Structure of the light receptors

in the compound eye of insects. Exp. Cell. Res., Suppl. 5, 586-644.
FINGERMAN, M. (1952) The role of the eye-pigments of Drosophila

melanogaster in photic orientation. /. Exp. Zool., 120, 131-64.
FINGERMAN, M., and BROWN, F. A. (1952) A Turkinje shift' in insect

vision. Science, 116, 171-2.
FINGERMAN, M., and BROWN, F. A. (1953) Color discrimination

and physiological duplicity of Drosophila vision. Physiol. Zool.,

26, 59-67.
FINLAYSON, L. H., and LOWENSTEiN, o. (1955) A proprioceptor in

the body musculature of Lepidoptera. Nature, 116, 1031.
FINLAYSON, L. H., and LOWENSTEIN, o. (1958) The structure and

function of abdominal stretch receptors in insects. Proc. Roy.

Soc, B, 148, 433-49.
FISCHER, w. (1957) Untersuchungen liber die Riechscharfe der

Honigbiene. Zeit. vergl. Physiol., 39, 634-59.
FOLSOM, J. w. (1931) A chemotropometer. J. Econ. Ent., 24, 827-33.

REFERENCES 227

FOLSOM, J. w., and wardle, r. a. (1943) Entomology with Special

Reference to Its Ecological Aspects, 4th Ed. Blakiston, Phila.,

p. 30.
forel, a. (1908) The Senses of Insects. Methuen, London.
FRAENKEL, G. (1932) Untcrsuchungen iiber die Koordination von

Reflexen und automatischnervosen Rhythmen bei Insekten. I.

die Flugreflexe der Insekten und ihre Koordination. Zeit. vergl.

Physiol, 16, 371-93.
FRIEDMAN, s. (1959) Sustained flight in Phormia (by a new method)

and its effect on blood pH. J. Ins. Physiol, 3, 118-19.
FRIEDRICH, H. (1929) Vergleichende Untersuchungen uber die

tibialen Scolopalorgane einiger Orthopteren. Zeit. wiss. ZooL,

134, 84-148.
FRINGS, H. (1941) The loci of olfactory end-organs in the blowfly,

Cynomyia cadaverina Desvoidy. /. Exp. ZooL, 88, 65-93.
FRINGS, H. (1944) The loci of olfactory end-organs in the honey-bee,

Apis mellifera Linn. /. Exp. ZooL, 97, 123-34.
FRINGS, H. (1945) Gustatory rejection thresholds for the larvae

of the cecropia moth, Samia cecropia (Linn.). Biol. Bull., 88,

37-43.
FRINGS, H. (1946) Gustatory thresholds for sucrose and electrolytes

for the cockroach, Periplaneta americana (Linn.). J. Exp. ZooL,

102, 23-50.
FRINGS, H. (1948) A contribution to the comparative physiology of

contact chemoreception. J. Comp. Physiol. Psychol., 41, 25-34.
FRINGS, H., and FRINGS, M. (1949) The loci of contact chemorecep-

tors in insects. Am. Midi. Nat., 41, 602-58.
FRINGS, H., and FRINGS, M. (1956) Reactions to sounds by the wood

nymph butterfly, Cercyonis pegala. Ann. Ent. Soc. Amer., 49,

611-17.
FRINGS, H., and FRINGS, M. (1957) Duplex nature of reception of

simple sounds in the scape moth, Ctenucha virginica. Science, 126,

24.
FRINGS, H., and FRINGS, M. (1958) Uses of sounds by insects. Ann.

Rev. Ent., 3, 87-106.
FRINGS, H., and O'NEAL, B. R. (1946) The loci and thresholds of con-
tact chemoreceptors in females of the horsefly, Tabanus sulci-

frons Macq. /. Exp. ZooL, 103, 61-80.
VON FRiscH, K. (1914) Der Farbensinn und Formensinn der Biene.

ZooL Jahrb. Physiol., 37, 1-238.
VON FRISCH, K. (1919) Obcr den Geruchsinn der Biene und seine

blutenbiologische Bedeutung. ZooL Jahrb. ZooL Physiol., 37,

1-238.
VON FRISCH, K. (1921) Uber den Sitz des Geruchsinnes bei Insekten.

ZooL Jahrb. ZooL Physiol., 38, 449-516.

228 THE PHYSIOLOGY OF INSECT SENSES

VON FRiscH,K. (1930) Versucheuber den GeschmackssinderBienen.

Naturwiss.y 18, 169-74.
VON FRiscH, K. (1935) Uber den Geschmackssinn der Biene. Zeit.

vergl. Physiol., 21, 1-156.
VON FRISCH, K. (1948) Geloste und ungeloste Ratsel der Bienen-

sprache. Naturwiss., 35, 38-43.
VON FRISCH, K. (1949) Die Polarisation des Himmelslichtes als orien-

tierender Faktor bei den Tanzen der Bienen. Experientia, 5,

142-8.
VON FRISCH, K. (1950 a) Die Sonne als Kompass im Leben der

Bienen. Experientia, 6, 210-21.
VONFRISCH, K. (1950b) Bees. Their Vision, Chemical Senses, and

Language. Cornell Univ. Press.
VON FRISCH, K. (1960) Wahmehmung der Schwingungsrichtung

polarisierten Lichtes bei Bienen. Bayerische Akad. Wiss., Math.

Naturwiss. KL, Sitz. 3 Juni 1960, 1-3.
VON FRISCH, K., and LiNDAUER, M. (1956) The 'language' and

orientation of the honey bee. Ann. Rev. Ent., 1, 45-58.
VON FRISCH, K., LINDAUER, M., and DAUMER, K. (1960) Uber die

Wahrnehmung polarsierten Lichtes durch das Bienenauge.

Experientia, 16, 289-301.
FRY, w. J., wuLFF, V. J., and BRUST,M. (1955) Retinal action poten-
tial-effect of temperature on magnitude and latency in the

grasshopper. J. Cell. Comp. Physiol., 45, 265-72.
FULTON, B. B. (1928) Sound perception by insects. Sci. Monthly, 11,

552-6.
FYG, w. (1960) Ober die Kristallkegel in den Komplexaugen der

Honigbiene. Mitteilungen Schweiz. Ent. Ges., 33, 185-94.
GAFFRON, M. (1934) Untcrsuchungcn iiber das Bewegungssehen bei

Libellenlarve, Fliegen und Fischen. Zeit. vergl. Physiol., 20,

299-337.
GETTRUP, E. (1962) Thoracic proprioceptors in the flight system of

Locusts. Nature, 193, 498-9.
GOLDSMITH, T. H. (1958 a) The visual system of the honeybee. Proc.

Nat. Acad. Sci., 44, 123-6.
GOLDSMITH, T. H. (1958 b) On the visual system of the bee (Apis

mellifera). Ann. N. Y. Acad. Sci., 74, 223-9.
GOLDSMITH, T. H. (1960) The nature of the retinal action potential,

and the spectral sensitivities of the ultra violet and green receptor

systems of the compound eye of the worker honeybee. /. Gen.

Physiol., 43, 775-99.
GOLDSMITH, T. H. (1961) The color vision of insects. In Light and

Life, eds. W. D. McElroy and B. Glass. The Johns Hopkins Press,

Baltimore, pp. 771-94.

REFERENCES 229

GOLDSMITH, T. H., and PHILPOTT, D. c. (1957) The microstructure

of the compound eye of insects. /. Biophys. Biochem. CytoL, 3,

429-40.
GOLDSMITH, T. H., and RUCK, P. R. (1958) The spectral sensitivities

of the dorsal ocelli of cockroaches and honeybees. J. Gen.

Physiol., 4h 1171-85.
GOODWIN, T. w., and srisukh, s. (1949) The carotenoids of the

integument of two locust species {Locusta migratoria migratori-

oides R. & F. and Schistocerca gregaria Forsk.). Biochem. 7.,

45, 263-7.
CORNER, P. (1957) Die optische und kinasthetische Orientierung der

Trichterspinne Agelena labyrinthica (CI.). Zeit vergl. Physiol., 41,

111-53.
gottsche, c. m. (1852) Beitrag zur Anatomie und Physiologic des

Auges der Krebse und Fliegen. Archiv. fiir Anatomie undPhysio-

/og/^, 1852,483-92.
GOTZE, G. (1927) Untersuchungen an Hymenopteren liber das

Vorkommen und die Bedeutung der Stirnaugen. Zool. Jahrb.

Zool. Physiol., 44, 211-68.
GRABER, V. (1882 a) Die chordotonalen Sinnesorgane und das

Gehor der Insecten. Archiv. f. mikroskopische Anatomie, 20,

506-640.
GRABER, V. (1882b) Die chordotonalen Sinnesorgane und das Gehor

der Insecten. Archiv. f. mikroskopische Anatomie, 21, 65-145.
GRABOWSKi, c. T., and dethier, v. g. (1954) The structure of the

tarsal chemoreceptors of the blowfly, Phormia regina Meigen.

/. Morph., 94, 1-20.
graham, c. h., and hartline, h. k. (1935) The response of single

visual sense cells to lights of different wavelengths. J. Gen.

Physiol, 13, 105-12.
GRANiT, r. (1947) Sensory Mechanisms of the Retina. Oxford Univ.

Press.
GRANIT, R. (1955) Receptors and Sensory Perception. Yale Univ.

Press, New Haven, 369 pp.
GRAY, E. G. (1960) The fine structure of the insect ear. Phil. Trans. Roy,

Soc. Lond., B, 243, 75-94.
GRAY, E. G., and pumphrey, r. J. (1958) Ultra-structure of the

insect ear. Nature, 181, 618.
GRAY, J. A. B., and SATO, M. (1953) Properties of the receptor poten-
tial in Pacinian corpuscles. /. Physiol., 122, 610-36.
GREEN, G. w. (1954) Humidity reactions and water balance of larvae

of Neodiprion americanus banksianae Rob. and A'', lecontei

(Fitch) (Hymenoptera : Diprionidae). Canad. Ent., 86, 261-74.
grenacher, h. (1879) Untersuchungen iiber die Sehorgan der

Arthropoden. G6ttingen,pp. 1-188.

230 THE PHYSIOLOGY OF INSECT SENSES

GRIFFIN, D. R. (1950) Measurements of the ultrasonic cries of bats.

J. Acoust. Soc. Amer., 22, 247-55.
GRIFFIN, D. R. (1953) Bat sounds under natural conditions, with

evidence for echolocation of insect prey. J. Exp. ZooL, 123,

435-66.
GUNN, D. L., and coswAY, c. A. (1938) The temperature and humi-
dity relations of the cockroach. V. Humidity preference. /. Exp.

Biol, 15, 555-63.
GUNN, D. L., and PiELOU,D. p. (1940) The humidity behaviour of the

mealworm beetle, Tenebrio molitor L. III. The mechanism of the

reaction. J. Exp. Biol., 17, 307-16.
GiJNTHER, K. (1912) Die Sehorgane der Larve und Imago von Dytis-

cus marginalis. Zeit. wiss. Zool., 100, 60-1 15.
HAFEZ, M. (1950) On the behaviour and sensory physiology of the

house-fly larva, Musca domestica L. I. Feeding stage. Parasitol.,

40, 215-36.
HAFEZ, M. (1953) On the behaviour and sensory physiology of the

house-fly larva, Musca domestica L. II. Prepupating stage. J.

Exp. ZooL, 124, 199-225.
HAFFER, o. (1921) Bau und Funktion der Sternwarzen von Saturnia

pyri Schiff". und die Haarentwicklung der Saturnidenraupen. Ein

Beitrag zu dem Thema: Das Arthropodenhaar. Arch. Natg.

Berlin, A, SI (2), 110-66.
HA GEM ANN, J. (1910) Beitragc zur Kenntnis von Corixa. Zool.

Jahrb. Anat. Ontog., 30, 373-426.
HAMILTON, M. A. (1931) The morphology of the water-scorpion, Nepa

cinerea Linn. (Rhymchota, Heteroptera). Proc. Zool. Soc. Lond.,

1067-136.
HAMILTON, w. F. (1922) A direct method of testing color vision in

lower animals. Proc. Nat. Acad. Sci., 8, 350-3.
HANSON, F. E., and WOLBARSHT, M. L. (1962) Dendritic action

potentials in insect chemoreceptors. Amer. Zool, 2, 528.
HANSTROM, B. (1927) Ubcr die Frage, ob funktionell verschiedene,

Zapfen-und Stabchenartige Sehzellen im Komplexauge der

Arthropoden vorkommen. Zeit. vergl. Physiol., 6, 566-97.
HANSTROM, B. (1928) Vergleichende Anatomic des Nervensystems der

mrbellosen Tiere unter Beriichsichtigung seiner Funktion. J.

Springer, Berlin.
HARKER, J. E. (1956) Factors controlling the diurnal rhythm

of activity of Periplaneta americana L. /. Exp. Biol., 33, 224-

234.
HARTLiNE, H. K. (1928) A quantitative and descriptive study of the

electric response to illumination of the arthropod eye. Amer. J.

Physiol., 83, 466-83.

REFERENCES 231

HARTLINE, H. K. (1930) The dark adaptation of the eye of Limulus, as

manifested by its electrical response to illumination. J. Gen.

Physiol, 13, 379-89.
HARTLINE, H. K. (1934) Intensity and duration in the excitation of

single photoreceptor units. /. Cell. Comp. Physiol, 5, 229-47.
HARTLINE, H. K. (1935) The discharge of nerve impulses from the

single visual sense cell. Cold Spring Harbor Symp. Quant. Biol, 3,

245-9.
HARTLINE, H. K., and GRAHAM, c. H. (1934) The spectral sensitivity

of single visual sense cells. Amer. J. Physiol, 109, 49-50.
HARTLINE, H. K., and McDONALD, P. R. (1947) Light and dark

adaptation of single photoreceptor elements in the eye of

Limulus. /. Cell. Comp. Physiol, 30, 225-53.

HARTLINE, H. K., WAGNER, H. G., and MaCNICHOL, E. F. (1952)

The peripheral origin of nervous activity in the visual system.

Cold Spring Harbor Symp. Quant. Biol, 17, 125-41.
HASKELL, p. T. (1955) Intensite sonore des stridulations de quelques

Orthopteres Britannique. In Collogue sur Vacoustique des

Orthopteres, ed R. G. Busnel. Fascicule Hors Serie des Ann. Epi-

phyties, pp. 154-67.
HASKELL, p. T. (1956a) Hearing in certain Orthoptera. I. Physiology

of sound receptors. /. Exp. Biol, 33, 756-66.
HASKELL, p. T. (1956 b) Hearing in certain Orthoptera. II. The nature

of the response of certain receptors to natural and imitation

stridulation. J. Exp. Biol, 33, 161-16.
HASKELL, p. T. (1957) The influence of flight noise on behaviour in

the desert locust Schistocerca gregaria (Forsk). /. Ins. Physiol,

1, 52-75.
HASKELL, P. T., and BELTON, P. (1956) Electrical responses of

certain lepidopterous tympanal organs. Nature, 111, 139-40.
HAS LINGER, F. (1935) Dber den Geschmackssinn von Calliphora

erythrocephala Meigen und die Verwertung von Zuckern und

Zuckeralkoholen durch diese Fliege. Zeit. vergl Physiol, 22,

614-40.
HASSENSTEIN, B. (1951) Ommatidicnrastcr und afiferente Bewegungs-

integration. Zeit. vergl. Physiol, 33, 301-26.
HASSENSTEIN, B. (1957) Uber Belichtungspotentiale in den Augen

der Fliegen Sarcophaga und Er is talis. J. Ins. Physiol, 1, 124-

30.
HASSENSTEIN, B. (1958) Ober die Wahrnehmung der Bewegung von

Figueren und unregelmassigen Helligkeitsmustern. Zeit. vergl

Physiol, 40, 556-92.
HASSENSTEIN, B. (1959) Optokinetische Wirksamkeit bewegter

periodischer Muster (Nach Messungen am Russelkafer Chloro-

phanus viridis). Zeit. Naturf., 14B, 659-74.

Q

232 THE PHYSIOLOGY OF INSECT SENSES

HASSENSTEiN, B., and REiCHARDT, w. (1956) Analyse der Zeit-,
Reihenfolgen- und Systemtheoretische bei der Bewegungsper-
zeption des Vorzeichenauswertung Riisselkafers Chlorophanus.
Zeit. Naturf., IIB, 513-24.

HASSETT, c. c, DETHiER, V. G., and GANS, J. (1950) A comparison
of nutritive values and taste thresholds of carbohydrates for the
blowfly. Biol. Bull., 99, 446-53.

HAUSER, G. (1880) Physiologische und histologische Untersuchungen
Tiber das Geruchsorgan der Insekten. Zeit. wiss. ZooL, 34,
367-403.

HECHT, s., and wald, g. (1934) The visual acuity and intensity
descrimination of Drosophila. /. Gen. Physiol., 17, 517-47.

HECHT, s., and wolf, e. (1929) The visual acuity of the honey-bee.
/. Gen. Physiol, 12, 727-60.

HEINTZ, E. (1959) La question de la sensibilite des abeilles a I'ultra-
violet. Insect es Sociaux, 6, 223-9.

HENKE, K., and RONSCH, G. (1951) Uber Bildungsgleichheiten in der
Entwicklung epidermaler Organe und die Entstehung des
Nervensystems im Flugel der Insekten. Naturwiss., 38, 335-6.

HERAN, H. (1957) Die Bienenantenne als Messorgan der Flugeigenge-
schwindigkeit. Naturwiss., 44, 475.

HERAN, H. (1959) Wahrnehmung und Regelung der Flugeigenge-
schwindigkeit bei Apis mellifica L. Zeit. vergl. Physiol., 42, 103-63.

HERS, J. (1938) Organe tympanal du Schistocerca gregaria Forsk.
Proc. 5th Internat. Locust Conference, Brussels, 434-45.

HERTWiCK, M. (1931) Anatomie und Variabilitat des Nervensystems
und der Sinnesorgane von Drosophila melanogaster (Meigen).
Z. wiss Zool., 139, 559-663.

HERTZ, M. (1929 a) Die Organisation des optischen Feldes bei der
Biene. I. Zeit. vergl. Physiol, 8, 693-748.

HERTZ, M. (1929 b) Die organisation des optischen Feldes bei der
Biene. II. Zeit. vergl Physiol, 11, 107-45.

HERTZ, M. (1931) Die organisation des optischen Feldes bei der
Biene. III. Zeit. vergl. Physiol, 14, 629-74.

HERTZ, M. (1933) Uber figurale Intensitaten und Qualitaten in der
optischen Wahrnehmung der Biene. Biol. Zbl, 53, 10-40.

HERTZ, M. (1934) Zu Physiologic des Formen-und Bewegungssehens.
I. Optomotorische Versuche an Fhegen. Zeit. vergl. Physiol,
20, 430-49.

HERTZ, M. (1935 a) Zur Physiologic des Formen-und Bewegungsse-
hens. II. Auflosungsvermogen des Bienenauges und optomotor-
ische Reaktion. Zeit. vergl. Physiol, 21, 579-603.

HERTZ, M. (1935 b) Zur Physiologic des Formen-und Bewegungsse-
hens. III. Figurale Unterscheidung und reziproke Dressuren bei
der Biene. Zeit. vergl. Physiol, 21, 604-15.

REFERENCES 233

HERTZ, M. (1935 c) Die Untersuchungcn iiber den Formensinn der

Honigbienc. Naturniss., 23, 618-24.
HERTZ, M. (1937 a) Beitrag zum Farbensinn und Formensinn der

Biene. Zeit. vergl. Physiol., 24, 413-21.
HERTZ, M. (1937 b) ZurTechnik und Methode der Bienenversuchcmit

Farbpapieren und Glasfiltern. Zeit. vergl. Physiol., 25, 239-50.
HERTZ, M. (1939) New experiments on colour vision in bees. J. Exp.

Biol., 16, 1-18.
HESS, c. (1920) Neues zur Frage nach einem Farbensinne bei Bienen.

Naturwiss., 8, 927-9.
HESS, w. N. (1917) The chordotonal organs and pleural discs of

cerambycid larvae. Ann. Ent. Soc. Amer., 10, 63-78.
HESSE, R. (1901) Untersuchungcn liber die Organe der Lichtemp-

findung bei niedere Tieren. VII. Von den Arthropoden-Augen.

Zeit. wiss. ZooL, 70, 347-473.
HICKS, J. B. (1857) On a new organ in insects. J. Proc, Linn. Soc.

(ZooL), 1,136-40.
HINTON H. E. (1955) Sound producing organs in the Lepidoptera.

Proc. Roy. Ent. Soc. Lond., C, 20, 5-6.
HODGSON, E. s. (1951) Rejection thresholds of an aquatic beetle,

Laccophilus maculosus Germ., to salts and alcohols. Physiol.

ZooL, 24, 131-140.
HODGSON, E. s. (1953) Chemoreception in aqueous and gas phases.

^/o/.^w//., 105, 115-27.
HODGSON, E. s. (1956 a) Physiology of the labellar sugar receptors of

flies. Anat. Rec, 125, 555.
HODGSON, E. s. (1956 b) Temperature sensitivity of primary chemo-

receptors of insects. Anat. Rec, 125, 560-1.
HODGSON, E. s. (1957) Electrophysiological studies of arthropod

chemoreception. II. Responses of labellar chemoreceptors of the

blowfly to stimulation by carbohydrates. /. Ins. Physiol., \^

240-7.
HODGSON, E. s. (1958 a) Chemoreception in arthropods. Ann. Rev,

Ent., 3, 19-36.
HODGSON, E. s. (1958 b) Electrophysiological studies of arthropod

chemoreception. III. Chemoreceptors of terrestrial and fresh

water arthropods. Biol. Bull, 115, 114-25.
HODGSON, E. s., and BARTON BROWNE, L. (1960) Elcctrophysio-

logy of blowfly taste receptors. Anat. Rec, 137, 365.

HODGSON, E. S., LETTVIN, J. Y., and ROEDER, K. D. (1955) Physio-

logy of a primary chemoreceptor unit. Science, 122, 417-18.
HODGSON, E. s., and ROEDER, K. D. (1956) Electrophysiological
studies of arthropod chemoreception. I. General properties of
the labellar chemoreceptors of Diptera. /. Cell. Comp. Physiol. ^
48, 51-76.

234 THE PHYSIOLOGY OF INSECT SENSES

HOLLiCK, F. s. J. (1940) The flight of the dipterous fly Mus-

cina stabulans Fallen. Phil. Trans. Roy. Soc. Lond., B, 230,

357-90.
HOLSTE, G. (1923) Das Gehirn von Dytiscus marginalis L. Zeit. wiss.

ZooL, 120, 251-80.
HO MANN, H. (1924) Zum Problem der Ocellenfunktion bei den

Insekten. Zeit. vergl. Physiol., 1, 541-78.
HOYLE, G. (1955) Functioning of the insect ocellar nerve. /. Exp.

Biol., 32, 397-401.
HOYLE, G. (1957) Nervous control of insect muscles. In Recent

Advances in Invertebrate Physiology. Univ. of Oregon Pub., 1957,

pp. 73-78. ^
HSiJ F. (1938) Etude cytologique et comparee sur les sensilla des

insectes. La Cellule, 47, 1-60.
HUGHES, G. M. (1952) Abdominal mechano-receptors in Dytiscus and

Locusta. Nature, 170, 531-32.
HUGHES, G. M. (1957) The co-ordination of insect movements. II.

The effect of limb amputation and the cutting of commissures in

the cockroach {Blatta orientalis). J. Exp. Biol., 34, 306-33.
HUGHES, G. M. (1958) The co-ordination of insect movements. III.

Swimming in Dytiscus. Hydrophilus, and a dragonfly nymph.

/. Exp. Biol, 35, 567-83.
HUMPHREYS, w. J. (1940) Physics of the Air, 3rd ed. McGraw-Hill,

New York, 676 pp.
HUNDERTMARK, A. (1936) Helligkeits-und Farbenunterscheidungs-

vermogen der Eiraupen der Nonne (Lymantria monacha). Zeit.

vergl. Physiol, 24, 42-57.
HUNDERTMARK, A. (1937) Das Formunterscheidungsvermogen der

Eiraupen der Nonne (Lymantria monacha). Zeit. vergl. Physiol,

24, 563-82.
ILSE, D. (1928) tJber den Farbensinn der Tagfalter. Zeit. vergl.

Physiol, 8, 658-92.
ILSE, D. (1937) New observations on responses to colours in egg-laying

butterflies. Nature, 140, 544-5.
ILSE, D. (1949) Colour discrimination in the dronefly, Eristalis tenax.

Nature, 163, 255-6.
IMAMURA, s. (1938) Studies on the chemical susceptibility of the

Kyosofly, Sturmia sericariae Cornalia. Bull. Imp. Seri. Exp. Sta.

(Tokyo), 9, 219-69.
ISHIKAWA, s. (1962) Visual response patterns of single ganghon cells

in the optic lobes of the silkworm moth, Bombyx mori L. /. Ins.

Physiol 8, 485-92.
ISHIKAWA, s., and hirao, t. (1960 a) Electrophysiological studies

on vision of the silkworm, Bombyx mori. (I) Electroretinogram of

stemmata. /. Sericultural Sci. Japan, 29, 8-14. (In Japanese.)

REFERENCES 235

SHIKAWA, s., and hirao, t. (1960 b) Electrophysiological studies
on vision of silkworm, Bombyx mori. (II) Analyses of ERG in
stemmata with special reference to dark- and light-adaptation.
J. Sericultural Sci. Japan, 29, 21 1-17. (In Japanese.)

ISHIKAWA, s., and hirao, t. (1960 c) Electrophysiological studies
on vision of silkworm, Bombyx mori. (Ill) The effects of temper-
ature on the ERG of the compound eye. /. Sericultural Sci. Japan,
24, 334-6. (In Japanese.)

JACOBS-JESSEN, u. (1959) Zur Orientierung der Hummeln und
einiger anderer Hymenopteren. Zeit. vergl. Physiol., 41, 597-641.

JAHN, T. L. (1946) The electroretinogram as a measure of wave-length
sensitivity to light. J. N. Y. Ent. Soc, 54, 1-8.

JAHN, T. L., and CRESCiTELLi, F. (1938) The electrical response of
the grasshopper eye under conditions of light and dark adapta-
tion. J. Cell. Camp. Physiol, 12, 39-55.

JAHN, T. L., and CRESCITELLI, F. (1939) The electrical response of
the Cecropia moth eye. J. Cell. Comp. Physiol, 13, 113-19.

JAHN, T. L., and CRESCITELLI, F. (1940) Diumal changes in the elec-
trical responses of the compound eye. Biol Bull, 78, 42-52.

JAHN, T. L., and WULFF, V. J. (1941 a) Influence of a visual diurnal
rhythm on flicker response contours of Dytiscus. Proc. Soc. Exp.
Biol and Med., 48, 660-5.

JAHN, T. L., and WULFF, V. J. (1941b) Retinal pigment distribution
in relation to a diurnal rhythm in the compound eye of Dytiscus.
Proc. Soc. Exp. Biol Med., 48, 656-60.

JAHN, T. L., and WULFF, V. J. (1942) Allocation of electrical responses
from the compound eye of grasshoppers. /. Gen. Physiol, 26,
75-88.

JAHN, T. L., and WULFF, V. J. (1943) Electrical aspects of a diurnal
rhythm in the eye of Dytiscus fasciventris. Physiol Zool, 16,
101-9.

JAHN, T. L., and WULFF, V.J. (1948) The spectral sensitivity of Dytis-
cus fasciventris. J. N. Y. Ent. Soc, 56, 109-16.

JANDER, R. (1957) Die optische Richtungsorientierung der Roten
Waldameise (Formica rufa L.). Zeit. vergl. Physiol, 40, 162-238.

JANDER, R., and WATERMAN, T. H. (1960) Sensory discrimination
between polarized light and light intensity patterns by arthropods.
J. Cell. Comp. Physiol, 56, 137-60.

JOHNSON, B. (1956) Function of the antennae of aphids during flight.
Austral. J. Sci., 18, 199-200.

JOHNSTON, c. (1855) Auditory apparatus of the Culex mosquito.
Quart. J. Micr. Sci., Old Series, 3, 97-102.

JORDAN, K. (1905) Note on a pecuHar secondary sexual character
found among Geometridae at the sensory organ situated at the
base of the abdomen. Nov. Zool, 12, 506-8.

236 THE PHYSIOLOGY OF INSECT SENSES

KALMUS, H. (1958) Responses of insects to polarized light in the

presence of dark reflecting surfaces. Nature, 182, 1526-7.
KALMUS, H. (1959) Orientation of animals to polarized light. Nature,

184, 228-30.
KATSUKi, Y., and suga, n. (1960) Neural mechanism of hearing in

insects. /. Exp. BioL, 37, 279-90.
K ATZ, B. (1950) Depolarization of sensory terminals and the initiation

of impulses in the muscle spindle. /. Physiol., 11, 261-82.
KENNEDY, J. s. (1937) The humidity reactions of the African migratory

locust, Locusta migratoria migratorioides R. & F., gregarious

phase. /. Exp. Biol, 14, 187-97.
VON KENNEL, J. (1912) Uber Tympanalorgane im Abdomen der

Spanner und Ziinsler. Zool. Anz., 39, 163-70.
VON KENNEL, J., and EGGERS, F. (1933) Die abdominalen Tym-
panalorgane der Lepidopteren. Zool. Jahrb. Anat. Ontog., 57,

1-104.
KEPPLER, E. (1958) Uber des Richtungshoren von Stechmiicken.

Zeit. Naturf., 13b, 285-6.
KIM, c.-w. (1961) Development of the chordotonal organ, olfactory

organ and their nerves in the labial palp in Pieris rapae L. Bull.

Dept. Biol. Korea Univ., 3 (4), 23-30.
KiRSCHFELD, K. (1959) Quantitative Beziehungen zwischen Lich-

treiz und Reaktion beim diphasischen Elektroretinogramm.

Zeit. Naturf., 14b, 212-13.
KOEHLER, o. (1924) Sinnesphysiologische Untersuchungen an

Libellenlarven. Verhandl. deut. Zool. Ges. Berlin, 29, 83-91.
KRAEPELiN, K. (1833) Uber die Geruchsorgane der Gliedertiere.

Osterprogramm. Realschule Johanneum, Hamburg.
KRUMii^s, R. (1952) Die Borstenewicklung bei der Wachmotte

Galleria mellonella L. Biol. Zbl, 183-210.
kDhn, a. (1927) Uber den Farbensinn der Bienen. Zeit. vergl.

Physiol., 5, 762-800.
KUHN, A., and POHL, R. ( 1 92 l)Dressurfahigkeit der Bienen aufSpek-

trallinien. Naturwiss., 9, 738-40.
KUNZE, G. (1933) Einige Versuche iiber den Antennengeschmacks-

sinn der Honigbiene. Zool. Jahrb. Zool. Physiol, 52, 465-512.
KUWABARA, M. (1951) Eff'ects of inorganic ions on the tarsal chemo-

receptor of the butterfly, Vanessa indica. Zool. Mag., 60, 9. (In

Japanese.)
KUWABARA, M. (1952a) Effects of inorganic salts on the tarsal chemo-

receptor of the butterfly, Vanessa indica II. Zool. Mag., 61,

121.
KUWABARA, M. (1952 b) Uber die Funktion der Antenne der Honig-
biene in bezug auf die Raumorientierung. Mem. Fac. Set.

Kuyshu Univ., Ser. E, 1, 13-64.

REFERENCES 237

KUWABARA, M. (1953) Effects of successive stimulation on the tarsal

chemoreceptor of Vanessa indica. Zool. Mag., 62, 154. (In

Japanese.)
KUWABARA, M. (1957) Bildung des bedingten Reflexes von Pavlove

Typus bei der Honigbiene, Apis mellifica. J. Fac. Sci. Hokkaido
Univ., Ser. YI, 13, 458-64.
KUWABARA, M. and NAKA, K. (1957) ERG from the compound eye

of the honeybee. Zool. Mag., Tokoyo, 66, 1 12. (In Japanese.)
KUWABARA, M., and TAKEDA, K. (1956) On the hygroceptor of the

honeybee. Apis mellifica. Physiology and Ecology, 1, 1-6. (In

Japanese.)
LARSEN, J. R. (1962) The fine structure of the labellar chemosensory

hairs of the blowfly, Phormia regina Meigen. /. Ins. Physiol, 8,

683-91.
LARSEN, o. (1955) Spezifische Mechanorezeptoren bei Aphelocheirus

aestivalis Fabr. nebst. Bemerkungen liber die Respiration dieser

Wanze. Lunds Univ. Arsskrift N. F. Avd. 2, 51 (1 1), 1-58.
LEES, A. D. (1942) Homology of the campaniform organs on the wing

of Drosophila melanogaster. Nature, 150, 375.
LEES, A. D. (1943) On the behaviour of wireworms of the genus

Agriotes^sch. (Coleoptera, Elateridae). I. Reactions to humidity.

J.Exp.Biol.,2^,A3-5?>.
LEES, A. D. (1948) The sensory physiology of the sheep tick, Ixodes

ricinus L. /. Exp. Biol., 25, 145-207.
L E H R, R. ( 1 9 1 4) Die Sinnesorgane der Beiden Fliigelpaare von Dytiscus

marginalis. Z. wiss. Zool, 110, 87-150.
LEIGHLY, J. (1937) A note on evaporation. Ecology, 18, 180-98.
LEMBERGER, F. (1908) Psychophysische Untersuchungen liber den

Geschmack von Zucker und Saccharin (Saccharose und Krystal-

lose). Arch.f. ges. Physiol., 123, 293-311.
LEYDiG, F. (1851) Anatomisches und Histologisches liber die Larvae

von Corethra plumicornis. Zeschr. wiss. Zool., 3, 435-51.
LiEBERMANN, A. (1926) Correlation zwischen den antennalen

Geruchorganen und der Biologie der Musciden. Zeit. Morph.

Okol. Tiere, 4, 1-97.
LiNDAUER, M. (1961) Communication Among Social Bees. Harvard

Univ. Press, Cambridge, Mass.
LINDAUER, M., and NEDEL, J. o. (1959) Ein Schweresinnesorgan der

Honigbiene. Zeit. vergl. Physiol., 42, 334-64.
LISSM ANN, H. w. (1950) Proprioceptors. In Physiological Mechanisms

in Animal Behaviour. Academic Press, New York, pp. 34-

59.
LOHER, w., and broughton, w. b. (1955) Etudes sur le comporte-

ment acoustique de Chorthippus bicolor (Charp) avec quelques

notes comparatives sur les especes voisines (Acrididae). In

238 THE PHYSIOLOGY OF INSECT SENSES

Collogue sur Vacoustique des Orthopteres, ed. R. B. Busnel.

Fasciules Hors Serie les Ann. Epiphyties, pp. 248-77.
LOTMAR, R. (1933) Neue Untersuchungen liber den Farbensinn der

Bienen, mit besonderer Beriicksichtigung des Ultraviolettes.

Zeit. vergl. Physiol., 19, 673-723.
LOWENSTEiN, o., and FiNLAYSON, L. H. (1960) The response of the

abdominal stretch receptor of an insect to phasic stimulation.

Comp. Biochem. Physiol., 1, 56-61.
LOWENSTEIN, w. R. (1956) Excitation and changes in adaptation by

stretch of mechanoreceptors. J.Physiol., 133, 588-602.
LOWNE, B. T. (1890) The Blow-fly. London.
LUBBOCK, J. (1886) ^«/^, Bees, and Wasps. Pp. 152-68.
LtJDTKE, H. (1953) Retinomotorik und Adaptationsvorgange im

Auge des Riickenschwimmers (Notonecta glauca L.). Zeit.

vergl. Physiol., 35, 129-52.
LUDTKE, H. (1957) Beziehungen des Feinbaues im Ruckensch-

wimmerauge zu seiner Fahigkeit, polarisiertes Licht zu analys-

ieren. Zeit. vergl. Physiol., 40, 329-44.
LUTZ, F. E. (1924) Apparently non-selective characters and combina-
tions of characters, including a study of ultraviolet in relation

to the flower-visiting habits of insects. Ann. N. Y. Acad. Sci., 29,

181-283.
LUTZ, F. E., and grisewood, e. n. (1934) Reactions of Drosophila

to 2537 A radiation. Am. Mus. Novit., 706, 1-14.
MACHATSCHKE, J. w. (1936) Der cuticulare Aufban des Rhab-

doms im Arthropodenauge. Vest. Ceskoslov. Zool. Spol., 4, 90-

109.

MaCNICHOL, E. F., WAGNER, H. G., and HARTLINE, H. K. (1953)

Electrical activity recorded within single ommatidia of eye of

Limulus. XIX Internat. Physiol. Congress, Montreal, p. 582.
MARKL, H. (1962) Borstenfelder an den Gelenken als Schweresin-

nesorgane bei Ameisen und anderen Hymenopteren. Zeit. vergl.

Physiol, 45, 475-569.
MARSHALL, J. (1935) The location of olfactory receptors in insects: a

review of experimental evidence. Trans. Roy. Ent. Soc. Lond., 88,

49-72.
MASSERA, M. G. (1952) Osservazioni preliminari sur potenziali

elettrici delle libellule per stimoli luminosi. Experientia, 8,

271-3.
MAYER, A. M. (1874) Experiments on the supposed auditory apparatus

of the mosquito. Am. Nat., 8, 577-92.
MciNDOO, N. E. (1914 a) The olfactory sense of insects. Smithson.

Misc. Pub., 63, 1-63.
MciNDOO, N. E. (1914 b) The olfactory sense of the honeybee. /. Exp.

Zool, 16, 265-346.

REFERENCES 239

MciNDOO, N. E. (1916) The sense organs on the mouthparts of the

honey bee. Smithson. Misc. Coll., 65 (14), 1-55.
MELIN, D. (1941) The function of the pore-plates in Hymenoptera.

Zool. Bidr. Uppsala, 20, 303-344.
MELLON, D., and EVANS, D. R. (1961) Electrophysiological evidence

that water stimulates a fourth sensory cell in the blowfly taste

receptor. Amer. Zoologist, 1, 372.
MENZER, G., and STOCKHAMMER, K. (1951) Zur Polansationsoptik

der Fazettenaugen von Insekten. Naturwiss., 38, 190-1.
MERKER, E. (1929) Die Fluoreszenz im Insektenauge, die Fluoreszenz

des Chitins der Insekten und seine Durchlassigkeit fiir ultraviol-

ettes Licht. Zool. Jahrb. Zool. Physiol, 46, 483-574.
MILLER, w. H. (1957) Morphology of the ommatidia of the compound

eye of Limulus. /. Biophys. Biochem. CytoL, 3, 421-8.
MILNE, L. J., and MILNE, M. J. (1945) Selection of colored lights by

night-flying insects. Ent. Americana, 24, 21-86.
MiNNiCH, D. E. (1922) The chemical sensitivity of the tarsi of the red

admiral butterfly, Pyrameis atalanta Linn. J. Exp. Zool., 35,

57-81.
MINNICH, D. E. (1925) The reactions of the larvae of Vanessa antiopa

Linn, to sounds. /. Exp. Zool., 42, 443-69.
MINNICH, D. E. (1929 a) The chemical senses of insects. Quart. Rev.

Biol, 4, 100-12.
MINNICH, D. E. (1929 b) The chemical sensitivity of the legs of the

blowfly, Calliphora vomitoria Linn., to various sugars. Zeit.

vergl. Physiol., 11, 1-55.
MINNICH, D. E. (1931) The sensitivity of the oral lobes of the proboscis

of the blowfly, CaUiphora vomitoria Linn., to various sugars.

J. Exp. Zool., 60, 121-39.
MINNICH, D. E. (1932) The contact chemoreceptors of the honey bee.

Apis mellifera Linn. /. Exp. Zool., 61, 375-93.
MINNICH, D. E. (1936) The responses of caterpillars to sounds. /. Exp.

Zool, 72, 439-53.
MINNICH, D. E. (1937) The reactions of fragments of the larvae of

Aglais antiopa Linn, to sounds. Bull. Mount Desert Biol. Lab.

1937, 19-20.
MiTTELSTAEDT, H. (1950) Physiologie des Gleichgewichtssinnes bei

Fliegenden Libellen. Zeit. vergl. Physiol., 32, 422-63.
MITTELSTAEDT, H. (1952) Uber den Beutefangmechanismus der

Mantiden. Verh. Dtsch. Zool. Ges., 1952, 102-6.
MITTELSTAEDT, H. (1954) Regelung und Steuerung bei der Orientier-

ung der Lebewesen. Regelungstechnik, 2. Jahrg. 1954 (10),

226-32.
MITTELSTAEDT, H. (1957) Prey capture in mantids. In Recent Ad-
vances in Invertebrate Physiology. Univ. Oregon Pub., pp. 51-71.

Q2

240 THE PHYSIOLOGY OF INSECT SENSES

MiTTELSTAEDT, H. (1961) Control systems of orientation in insects.

Ann.Rev.Ent.,1, 177-98.
MOLLER, H. (1921) Ober Lethocerus uhleri Mont. Zool. Jahrb. Abt.

Out. Tiere, 42, 43-90.
MOLLER-RACKE, I. (1952) Farbensinn und Farbenblindheit bei

Insekten. Zool. Jb., Jena Allg. ZooL, 63, 237-74.
MONCRIEFF, R. w. (1944) The Chemical Senses. Leonard Hill,

London.
MORiTA, H. (1959) Initiation of spike potentials in contact chemo-

sensory hairs of insects. IIL D. C. stimulation and generator

potential of labellar chemoreceptor of Calliphora. J. Cell. Comp.

Physiol, 54, 189-204.

MORITA, H., DOIRA, S., TAKEDA, K., and KUWABARA, M. (1957)

Electrical response of contact chemoreceptor on tarsus of the

butterfly, Vanessa indica. Mem. Fac. Sci. Kyushu Univ., Ser. E

(Biol.), 2, 119-39.
MORITA, H., and TAKEDA, K. (1957) The electrical resistance of the

tarsal chemosensory hair of the butterfly, Vanessa indica. J. Fac.

Sci. Hokkaido Univ., Ser. VI, Zool, 13, 465-69.
MORITA, H., and TAKEDA, K. (1959) Initiation of spike potentials in

contact chemosensory hairs of insects. II. The eff"ect of electric

current on tarsal chemosensory hairs of Vanessa. /. Cell. Comp.

Physiol., 54, 177-87.
MORITA, H., and YAMASHITA, s. (1959 a) Generator potential of

insect chemoreceptor. Science, 130, 922.
MORITA, H., and YAMASHITA, s.( 1959 b) The back-firing of impulses

in a labellar chemosensory hair of the fly. Mem. Fac. Sci. Kyushu

Univ., Ser. E (Biol.), 3, 81-87.
MORITA, H., and YAMASHITA, s. (1961) Rcccptor potentials recorded

from sensilla basiconica on the antenna of the silkworm larvae,

Bombyx mori. J. Exp. Biol., 38, 851-61.
mOller, e. (1931) Experimentelle Untersuchungen an Bienen und

Ameisen iiber die Funktionweise der Stirnocellen. Zeit. vergl.

Physiol, 14, 348-84.
MULLER, J. (1826) Zur vergleichenden Physiologie des Gesichtssinnes

des Menschen und der Tiere. Leipzig.
MULLiNS, L. J. (1954) Some physical mechanisms in narcosis. Chem.

Rev., 54, 289-323.
MULLINS, L. J. (1955 a) Olfactory thresholds for some homologous

series of compounds. Fed. Proc, 14, 105-6.
MULLINS, L. J. (1955 b) Olfaction. Ann. N. Y. Acad. Sci., 62, lAl-16.
NAKA, K., and EGUCHI, E. (1962) Spike potentials recorded from the

insect photoreceptor. /. Gen. Physiol, 45, 663-80.
NAKA, K., and KUWABARA, M. (1959) Electrical response from the

compound eye of Lucilia. J. Ins. Physiol, 3, 41-49.

REFERENCES 241

NEUHAUS, w. (1953 a) Uber die Riechscarfe des Hundes fur
Fettsauren.Z^/r. vergl. Physiol. ,35, 527-52.

NEUHAUS, w. (1953 b) Die Riechschwellen des Hundes fiirJonon und
Athyl-Mercaptan und ihr Verhaltnis zu anderen Riechschwellen
bei Hund und Mensch. Zeit. Naturf., 9b, 560-7.

NEWTON, H. c. F. (1931) On the so-called 'olfactory pores' in the
honeybee. Quart. J. Micr. Sci., 74, 647-68.

NiJENHUis, E. D., and DRESDEN, D. (1952) A micro- and morpho-
logical study on the sensory supply of the mesothoracic leg of the
american cockroach, Periplaneta americana. Proc. Kon. Nederl.
Akad. Wetensch. Amsterdam, C, 55, 300-10.

OEVERMANN, H. (1936) Das statische Verhalten einiger Wasserwanz-
enarten. Zeit. wiss. ZooL, 147, 595-628.

ORLOV, J. (1924) Die Innervation des Darmes der Insekten (Larven
von Lamellicorniem). Z. wiss. ZooL, 122, 425-502.

OSBORNE, M. p. (1963) An electron microscope study of an abdominal
stretch receptor of the cockroach. /. Insect, Physiol, 9 (in press).

OSBORNE, M. p., and FiNLAYSON, L. H. (1962) The structure and
topography of stretch receptors in representatives of seven
orders of insects. Quart. J. Micr. Sci., 103, 227^2.

OSTERHOUT, W. J. V., KAMERLING, S. E., and STANLEY, W. M. (1934)

Kinetics of penetration. VII. Molecular versus ionic transport.

/. Gen. Physical., 17, 469-80.
OWEN, w. B. (1962) The contact chemoreceptor organs of the

mosquito and their function in feeding behaviour. /. Insect

Physiol., 8 (in press).
PAPi, F. (1955) Richerche sull 'orientamento astronomic© di Arctosa

perita (Latr.) (Araneae-Lycosidae). Pubbl. Staz. ZooL Napoli,

27, 80-107.
PARRY, D. A. (1947) The function of the insect ocellus. /. Exp. BioL,

24,211-19.
PASQUINELLY, F., and BUSNEL, M. c. (1955) Etudes preliminaires

sur les mecanismes de la production des sons par les Orthopteres.

In Collogue sur Vacoustique des Orthopteres, ed. R. B. Busnel.

Fascicules Hors Serie Ann. Epiphyties, pp. 145-53.
PERTTUNEN, V. (1951) The humidity preferences of various Carabid

species (Col., Carabidae) of wet and dry habitats. Ann, Ent

Finnica, 17, 72-84.
PERTTUNEN, V. (1953) Reactions of diplopods to the relative hum-
idity of the air. Investigations on Orthomorpha gracilis, lulus

terrestris, and Schizophyllum sabulosum. Ann. ZooL Soc.

'Vanamo\ 16, 1-69.
PERTTUNEN, V. (1955 a) The reversal of the humidity reaction at the

onset of the egg-laying period in the diplopod Schizophyllum

sabulosum. Arch. Soc. 'Vanamo\ 9, 231-34.

242 THE PHYSIOLOGY OF INSECT SENSES

PERTTUNEN, V. (1955 b) The effect of antennectomy on the humidity

reactions of normal and desiccated specimens of Schizophyllum

sabulosum L. (Diplopoda, lulidae). Ann. Ent. Fenn., 21, 157-62.
PERTTUNEN, V., and AHONEN, u. (1956) The effect of age on the

humidity reaction of Drosophila melanogaster (Dipt. Drosophi-

lidae). Ann. Ent. Fennica, 22, 63-71.
PERTTUNEN, V., and ERKKiLA, H. (1952) Humidity reactions in

Drosophila melanogaster. Nature, 169, 78.
PERTTUNEN, V., and SALMI, H. (1956) The responses of Drosophila

melanogaster (Dipt. DrosophiUdae) to the relative humidity of

air. Ann. Ent. Fennica, 22, 36-45.
PERTTUNEN, V., and SYRJAMAKi, J. (1958) The effect of antennect-
omy on the humidity reactions of Drosophila melanogaster

(Dipt., DrosophiUdae). Ann. Ent. Fennica, 24, 78-83.
PETERS, w. (1961a) Morphologische Untersuchungen an chemischen

Sinnesorganen der Schmeissfliege Calliphora erythrocephala

Mg. (Diptera). XI Internat. Congress Entomology Vienna 1960,

1., 407-9.
PETERS, w. (1961b) Die Zahl der Sinneszellen von Marginalborsten

und das Vorkommen multipolarer Nervenzellen in den Labellen

von Calliphora erythrocephala Mg. (Diptera). Naturwiss., 48

(10), 412-13.
PETERS, w. (1962) Die propiorezeptiven Organe am Prosternum und

an den Labellen von Calliphora erythrocephala Mg. (Diptera).

Zeit. Morph. Okol. Tiere, 51, 211-26.
PETERSON, A., and HAEUSSLER, G. J. (1928) Responsc of the oriental

peach moth and codling moth to coloured Ughts. Ann. Ent. Soc.

Amer.,21,353-79.
PFLUGSTAEDT, H. (1912) Die Haltcren der Dipteren. Zeit. wiss.

Zool., 100, 1-59.
PIELEMEIER, w. H. (1946 a) Supcrsonic insects. /. Acous. Soc. Amer.,

17, 337-8.
PIELEMEIER, w. H. (1946 b) Seeing summer sounds. Sci. Monthly, 62,

450-2.
PIE LOU, D. p. (1940) The humidity behaviour of the mealworm beetle,

Tenebrio molitor L. II. The humidity receptors. /. Exp. Biol.,

17, 295-306.
PIELOU, D. P., and gunn, d. l. (1940) The humidity behaviour of

the mealworm beetle, Tenebrio molitor L. I. The reaction to

differences of humidity. /. Exp. Biol., 17, 286-94.
PIERCE, G. w. (1948) The Songs of Insects. Harvard Univ. Press.
PIRENNE, M. H. (1943) Binocular and monocular threshold of vision.

Nature, 152, 698-9.
PLATEAU, F. (1888) Recherches experimentales sur la vision chez

arthropodes. Bull. Acad. Roy. Belg., Ser. 3, 15, 28-46.

REFERENCES 243

DEL PORTILLO, J. (1936) Beziehungen zwischen den Offnungswink-

eln der Ommatidien, Kriimung und Gestalt der Insektenaugen

undihrerFunktionellenAufgabe.Z^//. verg/.P/2>;5/o/.,23, 100-45.
POWER, M. E. (1943) The brain of Drosophila melanogaster. J.

Morph., 72, 517-60.
PRINGLE, J. w. s. (1938 a) Proprioception in insects. I. A new type of

mechanical receptor from the palps of the cockroach. J. Exp.

Biol., 15, 101-13.
PRINGLE, J. w. s. (1938 b) Proprioception in insects. II. The action

of the campaniform sensilla on the legs. J. Exp. Biol., 15,

114-31.
PRINGLE, J. w. s. (1938 c) Proprioception in insects. III. The

function of the hair sensilla at the joints. /. Exp. Biol., 15, 467-73.
PRINGLE, J. \v. s. (1940) The reflex mechanism of the insect leg.

J. Exp. Biol., 17, 8-17.
PRINGLE, J. w. s. (1948) The gyroscopic mechanism of the halteres of

Diptera. Phil. Trans. Roy. Soc. London, 233, 347-84.
PRINGLE, J. w. s. (1950) The flight of insects. School Sci. Rev., 31,

364-9.
PRINGLE, J. w. s. (1957) Insect Flight. Cambridge Monographs in

Experimental Biology, Cambridge Univ. Press, 132 pp.
PRINGLE, J. w. s., and wilson, v. j. (1952) The response of a sense

organ to a harmonic stimulus. /. Exp. Biol., 29, 220-34.
PUMPHREY, R. J. (1936) Slow adaptation of a tactile receptor in the

leg of the common cockroach. J. Physiol., 87, 6P-7P.
PUMPHREY, R. J. (1940) Hearing in insects. Biol. Rev., 15, 107-32.
PUMPHREY, R. J. (1950) Hearing. In Physiological Mechanisms in

Animal Behaviour. Academic Press, New York, pp. 1-18.
PUMPHREY, R. J., and RAWDON-SMiTH, A. F. (1936a) Synchronized

action potentials in the cereal nerve of the cockroach (Periplaneta

americana) in response to auditory stimuli. /. Physiol, 87, 4P-5P.
PUMPHREY, R. J., and RAWDON-SMITH, A. F. (1936b) Sensitivity of

. insects to sound. Nature, 137, 990.
PUMPHREY, R. J., and RAWDON-SMiTH, A. F. (1936 c) Hearing in

insects : The nature of the response of certain receptors to auditory

stimuli. Proc. Roy. Soc. Lond., B, 121, 18-27.
PUMPHREY, R. J., and RAWDON-SMITH, A. F. (1939) Frequency

discrimination in insects: A new theory. Nature, 143, 806-7.
RARE, w. (1953) Beitrage zum Orientierungsproblem der Wasser-

wanzen. Zeit. vergl. Physiol., 35, 300-25.
RADL, E. (1905) tJber das Gehor der Insekten. Biol. Zbl, 25, 1-5.
RAMSAY, J. A. (1935) Methods of measuring the evaporation of water

from animals. /. Exp. Biol, 12, 355-72.
VOM RATH, o. (1888) tJber die Hautsinnesorgane der Insekten. Z.

wiss.Zool.,46,A\3)-5A.

244 THE PHYSIOLOGY OF INSECT SENSES

VOM RATH, o. (1895) Uber die Nervenendigungen der Hautsinnes-

organe der Arthropoden nach Behandlung mit der Methylen-

blau-u. Chromsilbermethode. Berichte Naturforsch. Gesell-

schaft Freiburg. I. B., 9, 137-64.
VOM RATH, o. (1896) Zur Kenntnis der Hautsinnesorgane und des

sensiblen Nervensystems der Arthropoden. Z. wiss. ZooL, 61,

499-539.
RAU, p. (1940) Auditory perception in insects, with special reference

to the cockroach. Quart. Rev. Biol., 15, 121-55.
REDiKORZEW, w. (1900) Untersuchungen iiber den Bau der Ocellen

der Insekten. Zeit. wiss. Zool., 68, 581-624.
REED, M. R. (1938) The olfactory responses of Drosophila melano-

gaster Meigen to the products of fermenting banana. Physiol.

Zoo/., 11, 317-25.
REGEN, J. (1908) Das tympanale Sinnesorgan von Thamnotrizon

apterus Fab. ^ als Gehorapparat experimentelle nachgewiesen.

S. B. Math. Naturwiss. Akad. Wiss. Wien., Ill, 487-90.
REGEN, J.(1912) Experimentelle Untersuchungen iiber das Gehor von

Liogryllus campestris L. Zool. Anz., 40, 305-16.
REGEN, J. (1913) iJber die Anlockung des Weibchens von Gryllus

campestris L. durch telephonisch libertragene Stridulationslaute

des Mannchens. PflUg. Arch. ges. Physiol. (Arkiv. ges. Physiol.),

155, 193-200.
REGEN, J. (1914 a) Untersuchungen iiber die Stridulation und das

Gehor von Thamnotrizon apterus Fab. (^. Sitzungsberichte

Math. -Naturwiss. Akad. Wiss. Wien., 123 (1), 853-92.
REGEN, J. (1914b) Haben die Antennen fiir die alternierende Stridula-
tion von Thamnotrizon apterus Fab. S eine Bedeutung? PflUg.

Archiv. ges. Physiol., 155, 245-50.
REGEN, J. (1924) Uber die Orientierungdes Weibschens vonL/ogr>^//M5

campestris L. nach dem Stridulationsschall des Mannchens.

Sitzber. Akad. Wiss. Wien., Abt. 1, 132, 81-88.
REGEN, J. (1926) Uber die Beeinfllissung der Stridulation von

Thamnotrizon apterus Fab. ^ durch kiinstlich erzeute Tone und

verschiedenartige Gerarsche. Sitzber. Akad. Wiss. Wien., Abt. 1,

135, 329-68.
RESCH, B. (1954) Untersuchungen uber das Farbensehen von Noton-

ecta glauca L. Zeit. vergl. Physiol., 36, 27-40.
RICHARD, G. (1951) L'innervation et les organes sensoriels des pieces

buccales du termite a cou jaune (Calotermes flavicollis Fab).

Ann. Des Sc. Nat. Zool., 13, 397-412.
RICHARD, G. (1952) L'innervation sensorielle pendant des mues chez

les insectes. Bull. Soc. Zool. France, 11, 99-106.
RICHARD, G. (1956) Ontogenese de I'Organe de Johnston chez

divers insectes. Tenth Internat. Congr. Ent., 1, 523-4.

REFERENCES 245

RICHARD, G. (1957) L'ontogenese des organcs chordotonaux

antennaires de Calotermes flavicollis (Fab.). Insectes Sociaux,

4,107-11.
RICHARDS, A. G. (1951) The Integument of Arthropods. Univ. Minn.

Press, Minneapolis, Minn.
RIEGERT, p. w. (1958) Humidity reactions of Melanoplus bivittatus

(Say) and Camnula pelhicida (Scudd.) (Orthoptera, Acrididae):

Reactions of starved and of moulting grasshoppers. Canad.

EntomoL, 90, 680-4.
RIEGERT, P. w. (1959) Humidity reactions of Melanoplus bivittatus

(Say) and Camnula pellucida (Scudd.) (Orthoptera, Acrididae):

Reactions of normal grasshoppers. Canad. Entomo!., 91, 35-40.
RIEGERT, p. w. (1960) The humidity reactions of Melanoplus bivi-
ttatus (Say) (Orthoptera, Acrididae): Antennal sensilla and

hygroception. Canad. EntomoL, 92, 561-70.
RiTTER, E. (1936) Untersuchungen liber chemischen Sinn beim

schwarzen Kolbenwasserkafer Hydrous piceus. Zeit. vergl.

Physiol., 23, 543-70.
ROEDER, K. D. (1939) Synchronized activity in the optic and proto-

cerebral ganglion of the grasshopper, Melanoplus femur-rubrum.

J. Cell. Comp. Physiol, 14, 299-307.
ROEDER, K. D. (1940) The origin of visual rhythms in the grasshopper,

Melanoplus femur-rubrum. /. Cell. Comp. Physiol., 16, 399-401.
ROEDER, K. D. (1948) Organization of the ascending giant fiber system

in the cockroach {Periplaneta americand). J. Exp. ZooL, 108,

342-62.
ROEDER, K. D. (1959) A physiological approach to the relation

between prey and predator. Smithson. Misc. Coll., 137, 287-306.
ROEDER, K. D., and TREAT, A. E. (1957) Ultrasonic reception by the

tympanic organ of noctuid moths. /. Exp. Zool., 134, 127-57.
ROEDER, K. D., and TREAT, A. E. (1961 a) The reception of bat cries

by the tympanic organ of noctuid moths. In Sensory Communica-
tion, ed. W. Rosenblith, M. I. T. Technology Press.
ROEDER, K. D., and TREAT, A. E. (1961b) The detection and evasion

of bats by moths. Amer. Sci., 49, 135-48.
ROGOSiNA, M. (1 928) tJber das periphere Nervensystem der Aeschna-

Larve. Z. Zellforsch., 6, 732-58.
ROHLER, E. (1906) Beitrage zur Kenntnis der Sinnesorgane der

Insekten. Zool. Jahrb. Anat. Physiol., 22, 225-88.
ROKOHL, R. (1942) Uber die regionale Verschiedenheit der Farben-

tiichtigkeit in Zusammengesetzten Auge von Notonecta glauca

Zeit. vergl. Physiol, 29, 538-76.
ROTH, L. M. (1948) A study of mosquito behavior. An experimental

laboratory study of sexual behavior of Aedes aegypti (Linnaeus).

Amer. Midi Nat., 40, 265-352.

246 THE PHYSIOLOGY OF INSECT SENSES

ROTH, L. M., and WILLIS, E. R. (1951a) Hygroreceptors in adults of
Tribolium (Coleoptera, Tenebrionidae). J. Exp. ZooL, 116,
527-70.

ROTH, L. M., and WILLIS, E. R. {1951b) Hygroreceptors in Cole-
optera. /. Exp. ZooL, 111, 451-88.

ROTH, L. M., and WILLIS, E. R. ( 1 95 Ic) The effects ofdesiccation and
starvation on the humidity behavior and water balance of
Tribolium confusum and Tribolium castaneum. J. Exp. ZooL,
118, 337-62.

ROTH, L. M., and WILLIS, E. R. (1952) Possible hygroreceptors in
Aedes aegypti (L.) and Blatella germanica L. J. Morph., 91, 1-14.

RO YS, c. (1954) Olfactory nerve potentials a direct measure of chemo-
reception in insects. Ann. N. Y. Acad. Sci., 58, 250-5.

RUCK, p. (1957) The electrical responses of dorsal ocelli in cock-
roaches and grasshoppers. J. Ins. Physiol., 1, 109-23.

RUCK, p. (1958 a) Dark adaptation of the ocellus in Periplaneta
americana: A study of the electrical response to illumination.
/. Ins. Physiol, 2, 189-98.

RUCK, P. (1958 b) A comparison of the electrical responses of com-
pound eyes and dorsal ocelli in four insect species. J. Ins. Physiol,
2, 261-74.

RUCK, P. (1961a) Electrophysiology of the insect dorsal ocellus I..
Origin of the components of the electroretinogram. /. Gen.
Physiol, 44, 605-27.

RUCK, p. (1961b) Electrophysiology of the insect dorsal ocellus. II.
Mechanisms of generation and inhibition of impulses in the
©cellar nerve of dragonflies. /. Gen. Physiol, 44, 629-39.

RUCK, p. (1961c) Electrophysiology of the insect dorsal ocellus. III.
Responses to flickering Hght of the dragonfly ocellus. /. Gen.
Physiol, 44, 641-57.

RUCK, P., and jahn, t. l. (1954) Electrical studies on the compound
eye of Ligia occidentalis dana (Crustacea: Isopoda). /. Gen.
Physiol, 37, 825-49.

RUDERMAN, M. A., and ROSENFELD, A. H. (1960) An explana-
tory statement of elementary particle physics. Am. Sci., 48 (2),
209-17.

SALZLE, K. (1932) Untersuchungen an Libellenlarven iiber das
Sehen bewegter Objekte. Zeit. vergl Physiol, 18, 345-68.

SANDER, w. (1933) Phototaktische Reaktionen der Bienen auf
Lichter verschiedener Wellenlange. Zeit. vergl Physiol, 20,
267-86.

S ATI J A, R. c. (1958 a) A histological and experimental study of
nervous pathways in the brain and thoracic nerve cord of
Locusta migratoria migratorioides (R. & P.). Res. Bull Punjab
Univ., 138, 13-32.

REFERENCES 247

SATIJA, R. c. (1958 b) A histological study of the brain and thoracic

cord of Calliphora erythrocephala with special reference to

the descending nervous pathways. Res. Bull. Panjab Univ., 142,

81-98.
SCHALLER, A. (1926) Sinncsphysiologischc und psychologische

Untersuchungen an Wasserkafern und Fischen. Zeit. vergl.

Physiol., 4, 370-464.
SCHALLER, F. (1951) Lautcrzeugung und Horvermogen von Corixa

(Callicorixa) striata L. Zeit. vergl. Physiol., 33, 476-86.
SCHALLER, F. (1960) Die optomotorische Komponente bei der

Flugsteuerung der Insekten. Zool. Beit rage (Neue Folge) 5,

483-96.
SCHALLER, F., and TiMM, c. (1950) Das Horvermogen der Nacht-

schmetterlinge. Zeit. vergl. Physiol., 32, 468-81.
SCHANZ, M. (1953) Der Geruchssinn des Kartoffelkafers. Zeit.

vergl. Physiol., 35, 353-79.
SCHENK, o. (1903) Die antennalen Hautsinnesorgane einiger Lepi-

dopteren und Hymenopteren mit besonderer Beriicksichtigung

der sexuellen Unterschiede. Zool. Jahrb. Anat. Ontog., 17,

573-618.
SCHLEGTENDAL,A.(1934) Beitrag zum Farbensinn der Arthropoden .

Zeit. vergl. Physiol, 20, 545-81.
SCHLiEPER, c. (1927) Farbensinn der Tiere und optomotorische

Reaktionen. Zeit. vergl. Physiol., 6, 453-72.
SCHMIDT, A. (1938) Geschmacksphysiologische Untersuchungen an

Ameisen. Zeit. vergl. Physiol., 25, 351-78.
SCHNEIDER, D. (1955) Mikro-Elcktrodcn registrieren die elektri-

schen Impulse einzelner Sinnesnervenzellen der Schmetter-

lingsantenne. Industrie-Elektronik (Elektro-Spezial, Hamburg),

3, 3-7.
SCHNEIDER, D. (1957 a) Electrophysiological investigation on the

antennal receptors of the Silk Moth during chemical and mech-
anical stimulation. Experientia, 13, 89-91.
SCHNEIDER, D. (1957 b) Elcktrophysiologischc Untersuchungen von

Chemo und Mechanorezeptoren der Antenne des Seidenspinners

Bombyx mori L. Zeit. vergl. Physiol., 40, 8-41.
SCHNEIDER, D. (1961) The olfactory sense of insects. Dragoco Report,

Monthly Information Service, Gerberding & Co., Holzminden,

West Germany, 6, 135-51.
SCHNEIDER, D. (1962) Electrophysiological investigation on the

olfactory specificity of sexual attracting substances in different

species of moths. /. Ins. Physiol., 8, 15-30.
SCHNEIDER, D., and HECKER, E. (1956) Elcktrophysiologic der

Antenne des Seidenspinners Bombyx mori bei Reizung mit

248 THE PHYSIOLOGY OF INSECT SENSES

angereicherten Extrakten des Sexuallockstoffes. Zeit. Naturf.,

lib, 121-4.
SCHNEIDER, G. (1953) Die Halteren der Schmeissfliege {Calliphora)

als Sinnesorgane und als mechanische Flugstabilisatoren. Z.

vergl Physiol, 35, 416-58.
SCHNEIDER, G. (1955) Die spektrale Empfindlichkeit des Rezeptors

fiir das *Dammerungssehen' bei Calliphora. Verhandl. Deut.

Zool. Gesellschaft, 1954, 346-51.
SCHNEIDER, G. (1956) Zur spektralen Empfindlichkeit des Kom-

plexauges von Calliphora. Zeit. vergl. Physiol., 39, 1-20.
SCHNEIDER, w. (1950) Ober den Erschtitterungssinn von Kafern und

Fliegen. Zeit. vergl. Physiol., 32, 287-302.
SCHON, A. (1911) Bau und Entwicklung des tibialen Chordotonal-

organe bei der Honigbiene und bei Ameisen. Zool. Jahrb. Anat.,

31, 439-72.
SCHONE, H. (1950) Die Augen als Gleichgewichtsorgane bei Wasser-

kaferlarven. Naturwiss., 10, 235-6.
SCHONE, H. (1953) Farbhelligkeit und Farbunterscheidung bei den

Wasserkafern Dytiscus marginalis, Acilius sulcatus und ihren

larven. Zeit. vergl. Physiol., 35, 27-35.
SCHULTZE, M. (1868) Untersuchungen liber die Zusammenengesetzten

Augen der Krebse und Insekten. Bonn.
SCHUTZ, H. G., and pilgrim, f. j. (1957) Differential sensitivity in

gustation. /. Exp. Psychol., 54, 41-48.
SCHWABE, J. (1906) Beitrage zur Morphologie und Histologic der

tympanalen Sinncsapparate der Orthoptera. Zoologica, Stutt-
gart, 20, 1-154.
SCHWARZ, R. (1955) liber die Riechscharfc der Honigbiene. Zeit.

vergl. Physiol, 37, 180-210.
SIEBOLD, T. (1844) tjber das Stimm-und Gehororgan der Orthoptera.

Wiegmanns Arch. Naturg., Jahrg., 10, 52-81.
siHLER, H. (1924) Die Sinnesorgane an den Cerci der Insekten. Zool

Jahrb. Anat., 45, 519-80.
SLiFER, E. H. (1935) Morphology and development of the femoral

chordotonal organs of Melanoplus differentialis (Orthoptera,

Acrididae). /. Morph., 58, 615-37.
SLiFER, E. H. (1961) The fine structure of insect sense organs. Inter-
national Review of Cytology, 2, 125-59.
SLIFER, E. H., and FiNLAYSON, L. H. (1956) Musclc rcccptor organs

in grasshoppers and locusts (Orthoptera, Acrididae). Quart. J.

Micr. Sci., 97, 617-20.
SLIFER, E. H., PRESTAGE, J. J., and BEAMS, H. w. (1957) The fine

structure of the long basiconic sensory pegs of the grasshopper

(Orthoptera, Acrididae) with special reference to those on the

antenna. /. Morph., 101, 359-96.

REFERENCES 249

SLIFER, E. H., PRESTAGE, J. J., and BEAMS, H. w. (1959) The chemo-
receptors and other sense organs on the antennal flagcllum of the
grasshopper (Orthoptera; Acrididae). J. Morph., 105, 145-91.

SMITH, F. E., and BAYLOR, E. R. (1960) Bees, Daphnia and polarized
light, ^fo/