THE SOCIAL INSECTS

I. A queen Polistes wasp depositing an egg in a newly made cell (on right).
Other eggs can be seen at the top, larval cells in the middle and pupa cells,
closed over, beneath them. A sticky secretion fastens each egg to the side of

the cell.

i
O. W. RICHARDS

M.A., D.SC.

THE

SOCIAL INSECTS

MACDONALD : LONDON

First published in 1953 by
Macdonald & Co. (Publishers), Ltd.

16 Maddox Street, W.i
Made and printed in Great Britain by
Purnell and Sons, Ltd.
Paulton (Somerset) and London

ACKNOWLEDGMENT

It has become very difficult of late for any worker to know all
that has been recorded about any group of animals. Even such a
restricted subject as social behaviour in insects has by now an
immense literature. It is quite impossible for any one worker to
have made personal observation on all the species or to have seen
all the varieties of behaviour which have been recorded. Any of
my entomological colleagues who read this work will be aware
how far I have drawn on the writings of naturalists of many
countries. I have not thought it necessary to give references to the
literature in the text, but I have often mentioned the name of the
worker who first recorded some of the more striking facts.

My thanks are due to the University of Hawaii Press for per-
mission to quote the extract from Insects of Hawaii by E. C.
Zimmermann, Volume V, which appears on page 126.

I am indebted to my colleague Dr. N. WalofT for reading and
criticising the first draft of the book.

O.W.R.
London, February, 1953

Vll

CONTENTS

ACKNOWLEDGMENT vii

CHAPTER i THE INSECT WORLD . . .15

How insects work — Food and water — Growth — Reproduc-
tion— Behaviour — Finding the way — How social behaviour
develops — Fully social behaviour

CHAPTER 2 SOLITARY AND SOCIAL WASPS . 40

A hunting wasp — Varieties of wasp — Social wasps — Wasp
society — Males, queens and workers — Wasps in Britain

CHAPTER 3 VARIETIES OF WASP SOCIETY . 59

Polistes wasps — Polybia wasps — Wasps in general

CHAPTER 4 SOLITARY AND SOCIAL BEES . 76

Bees and flowers — Solitary origins — Development of worker
caste — Humble bees — Stingless bees — Theory of hereditary
castes

CHAPTER 5 HONEY BEES . . . .103

Wild honey bees — Drones and the nuptial flight — Bee econo-
mics— Language and dances of bees

CHAPTER 6 THE ANTS OR PISMIRES . .119

Nest — Ants and other animals — Food of ants — driver and
legionary ants — Gatherers of mixed foods — Honey-pot ants
— Harvesting ants — Leaf-cutting ants — Types of nest —
Founding new colonies — Castes of ants — Communication —
Trail-making and path-finding

IX

CONTENTS

CHAPTER 7 SOCIAL PARASITES . . .151

Ant guests — Social parasites — How cuckoo bees and wasps
arose — Parasitic ants — Parasitism and degeneration — Sum-
ming up parasitism

CHAPTER 8 THE TERMITES . . . .169

Nests — Nuptials — Food and agriculture — Termites as pests
— Termites and tropical plant-life — Castes — Theories of caste
formation — Growth of colonies

CHAPTER 9 INSECT SOCIETIES . . .191

Reproduction — Food and social life — Origin of social be-
haviour— Insect societies as families — Study for the future

FURTHER READING 211

INDEX 213

LIST OF PLATES

I.

2.

3-

4-

5-

6.

7-
8.

9-
io.
ii.

12.

i3-
14.
i5-
16.

17.
18.
19.
20.
21.

Queen Polistes wasp depositing egg
A mud-dauber wasp with spiders .
A sand wasp dragging caterpillar .
Hunting wasp, Ammophila sabulosa
Hunting wasp, Sphex maxillosus .

I Nests of German wasps

Nest of Norwegian wasp .
Nest of Norwegian wasp opened .
Fence rasped by wasps
Comb from wasp's nest .
Queen German wasp hibernating .
Polistes wasp feeding larva

f Wasp larva spinning cap .

Polistes worker in flight .

Polistes wasps on nest

A Polistes wasp and suspended nest

A hornet's nest

A hornet worker

Another nest of Polistes wasps

facing page 49

49
. 64

. 64

following page 64

. <54
. 64

22.

24.

25-

LIST OF PLATES

Nests of Polybia rejecta in mango tree following page 64

Close view of nests of P. rejecta .
Nest of P. occidentalis
Nest of wasp Parachartergus

26. Parachartergus wasps rebuilding nest

27. Anthophora acervorum on flower .

28. Solitary mason bee, male and female

29. Mason bee, larvae feeding

30. Mason bee, view of nest behind door lock

3 1 . Humble bee, male, worker and queen

32. Nest of humble bee

33. Humble bee about to take nectar

34. Young larvae of humble bee

35. Drone honey bee

36. Queen honey bee .

37. Worker honey bee

38. Larvae of drone honey bees

39. Larvae of worker honey bees
Worker pupae in closed cells
Worker bees on comb
Marked bees feeding
Bees marked for observation
Honey bee comb with "royal" cells

40

41

42

43
44

45

46. Honey bee swarm divided in two

facing page 65

. 65

80

80

facing page 81
81
81
81

»• 96

. 96

. 96

. 96

97

• 97

97

112

112
following page 112
112
112
112
facing page 113

Young worker honey bee emerging from cell

"3
113

xii

LIST OF PLATES

47-

A worker ant with aphides

facing page 128

48.

Nest of wood ant .

. 128

49-

Worker wood ant carrying twig .

. 129

50.

Nest of termite

129

5*-

Fertile female or queen termite

. 129

Acknowledgments for Illustrations

Certain of the drawings in this book are derived from other
published books. Acknowledgments are due to Macmillan & Co.,
Ltd., for Fig. 2 {Cambridge Natural History by David Sharpe),
Methuen & Co., Ltd., for Fig. 3 {The Principles of Insect Physiology
by V. B. Wiggles worth), the Author and the Clarendon Press for
Figs. 4, 5, 6 {The Study of Instinct by Dr. N. Tinbergen), the
Author and H.M.S.O. for Fig 7 {British Bark-Beetles by J. W.
Munro), Cornell University Press for Fig. 8 {Bees by von Frisch),
Columbia University Press for Fig. 9 {Ants by Wheeler), Rout-
ledge and Kegan Paul Ltd. for Fig. 10 {British Ants by Donis-
thorpe), Constable & Co., Ltd., for Fig. 12 {Social Life Among the
Insects by Wheeler).

Thanks are due to the following for permission to use half-tone
illustrations: the Director of the Pest Infestation Laboratory,
Slough, Bucks, for 19 and 20; Royal Entomological Society of
London for 21-26; the Clarendon Press, Oxford, for 42 and
43 (from The Honeybee by Dr. C. G. Butler) ; and the Common-
wealth Scientific and Industrial Research Organization, Melbourne,
for 50 {Termites [Isoptera] from the Australian Region by G. F.
Hill).

THE INSECT WORLD

Zoologists have sometimes divided the past history of the
world into periods named after the dominant type of animal.
The age of fishes was followed by an age of reptiles, and that
by an age of mammals. Finally, for the last few thousand
years, there has been an age of man. Throughout an immense
period, starting before the dominance of the reptiles, insects
have been abundant. Though not large and conspicuous
enough to give their name to an age, they are at this moment
the only group of animals which disputes our dominance.
They are infinitely more numerous than the mammals either
in species or in individuals ; many more specimens of insects
can be found in and on an acre of ground than make up the
human population of the British Isles. In some parts of the
world, biting or disease-carrying insects may temporarily
drive man out, and much more often they largely determine
the nature of the plant cover and of the sort of animals which
live in it.

There is one other respect in which insects seem to rival
man: their large and complicated societies are the only ones
which can be compared with ours. While they have excited
the wonder of mankind at least since the time of Solomon,
real understanding of how they work is only just being
reached. The ties which unite their societies and the laws
which they obey are so different from ours that our emotions
are not involved as they are apt to be when we study even the

15

THE SOCIAL INSECTS

most primitive of human societies. Their most marked
characteristic is that within the limits set by the abilities of
the species, everything appears to be done for the good of
the community and only the necessary minimum for that of
the individual. Such societies are well worth studying, not
because we are ever likely to wish to imitate them but because
of the light they throw on the general principles of organisa-
tion and because of the fascination of their extraordinary
behaviour. It is, moreover, a field in which many striking
discoveries have recently been made, suggesting how much
more still awaits the patient investigator.

HOW INSECTS WORK

All the many kinds of insects are built on a rather uniform
common plan, very different from the one seen in the
mammals. In the diagram of a queen ant on this page, the

Fig. i
Queen ant (Lasius), to show the parts of the body.

16

THE INSECT WORLD

important parts of the body are indicated. The skeleton of
insects is almost all laid down as an external armour. This
largely limits their size, for the internal organs of a really
large insect would be crushed under their own weight.
Moreover, insects breathe by means of a series of fine tubes
which ramify into every organ of the body. Oxygen passes
in mainly by diffusion and respiratory movements, corres-
ponding to our breathing, are less developed and only aerate
the larger tubes. This method of obtaining oxygen is very
efficient in small animals but becomes less so as size increases.
The blood of insects, also, is not in general enclosed in veins
and arteries and does not circulate rapidly round the body
like ours, and this again would be a drawback in a large
animal.

An insect has a pair of compound eyes made up of a large
number of small facets or transparent lenses. The whole eye
works on a different principle from ours and there is no way
of focusing the vision on objects at different distances.
Insects cannot see nearly so far as we can nor form such clear-
cut images, but their eyes may be very efficient for certain
specialised needs, such as catching quick-moving prey. So
much can be deduced from a careful study of the micros-
copical structure of the eye. The visual powers of an insect
like the honey bee can be tested by a process of training.
Small glass dishes, one of which contains sugar and water and
the others water only, are exposed on small squares of
coloured paper. Foraging worker bees can be trained to
associate the dish containing sugar with a particular colour.
If the dish placed on the colour used for training is replaced by
one containing water only, the trained bees continue to visit
it for some time, though they cease to feed from it. The
postions of the various squares on the table can be frequently
altered so that it is clear that the bees are not learning merely

b 17

THE SOCIAL INSECTS

the position of the dish. By including amongst the coloured
squares all shades of grey, including the whole scale from
dark to light, one can show that the bees are responding to
colour and not to mere brightness. A colour-blind animal,
such as a dog, cannot distinguish a grey square from another
colour of similar brightness.

By means of such experiments, it can be shown that the eye
of the bee is relatively insensitive to the red end of the
spectrum but much more sensitive than our eyes to the ultra-
violet end. A curious experiment which demonstrates the
same point in ants was carried out by Lord Avebury. He
found that if the lid of an artificial nest is made of ordinary
glass, the ants endeavour to move into shadow, but if
Crookes' glass (which cuts out most of the light at the violet
end) is used the ants will stay under it.

Some insects, such as grasshoppers and crickets which
produce a lot of sound themselves, have quite elaborate
organs of hearing, either in the forelegs or in the abdomen. A
female cricket no longer walks towards a singing male if the
organs in her forelegs are destroyed, and she can be attracted
to a telephone receiver conveying the song of a male from
another cage. In other insects, no such organs have as yet
been discovered. This is true for instance of the honey bee in
spite of the loud hum which it produces. It is possible that
some organ is present which has not yet been recognised.

The sense of smell in a honey bee can be investigated by
training experiments, similar to those used to study its vision.
A number of dishes can be placed in boxes which may be
entered through a small hole. The one dish which contains
sugar as well as water can have a scent such as lavender water
added to it. Foraging bees can then easily be trained to
associate sugar with a particular scent. The honey bee,
probably because of the importance of flowers in its economy,

18

THE INSECT WORLD

can distinguish a great variety of scents; it is found to make
some of the same errors as we do in confusing certain chemi-
cally distinct substances. To us, these substances give out a
similar odour. By removing the antennae, it can be shown
that the antennae are the chief seat of scent perception. Other
insects, such as male moths, are attracted to the scent of the
other sex, and will fly to any object with which a virgin
female has been in contact.

Fig. 2

Two types of ears in the
foreleg of grasshoppers, A
with the eardrum visible, B
with the drum concealed in
a slit

The sense of taste has also been studied in some detail in
the honey bee by means of training experiments. The sense is
much like ours, but its range is a good deal smaller. About
one-third of the substances which we call sweet attract bees
and some of the things which we call bitter repel them.

An insect's antennae are also important organs of touch
and have the faculty of distinguishing hot from cold. If
normal honey bee workers are put in a cage of which the floor
is hot at one end and cold at the other, most of them collect
near the warm end. If the antennae are removed, their
position in the cage no longer bears any relation to the
temperature of the floor.

19

THE SOCIAL INSECTS

In vertebrates, there is a rough correlation between the size
of the brain and the mental powers of the animal. Because
insects are so small, their brain is minute compared with that
of a mammal and on analogy one would not expect it to be
very efficient. One certainly would not expect it to be
capable of producing intelligent behaviour. It is probable
that owing to great differences in plan the analogy is a false
one. The behaviour of such insects as the honey bee shows
that the insect brain must be a remarkable structure in spite
of its small size.

Insects are " cold-blooded " animals. This does not mean
that they cannot raise their own temperatures, but only that
they have no mechanism for maintaining them constant,
particularly at levels above their surroundings. When flying
or otherwise very active they produce a considerable amount
of heat and may raise their temperatures several degrees, but
this excess is soon lost when the activity ceases. It is, of
course, physically much more difficult for a small body to
remain at a different temperature from its surroundings than
it is for a larger one. Scales or hairs may partly hinder the
loss of heat but they are not very efficient in small insects.
Actually, as will be seen later, most social insects are able to
maintain their nests at quite a high and often fairly constant
temperature. This is due to the contributions of a large
number of separate individuals as well as to the structure of
the nest which is usually such as to prevent the flow of heat.
Apart, however, from such special cases, insects tend to do
everything, including movement, feeding, and development,
at a speed which is related to the temperature of their sur-
roundings. Shapley, the American astronomer, measured the
speed at which ants travelled along their runs and also the
temperature of the soil. Here are some figures for the Argen-
tine ant, each figure being an average of twenty individuals.

20

THE INSECT WORLD

Temperature (°C). 25.6 29.4 30.3 33.0

Speed of walking (cms/sec.) 2.62 3.49 3.57 4.17

From the speed of walking, the soil temperature can be de-
duced with an accuracy of one degree. It is probably because
of this dependence on external temperature that insects are
so much more numerous in the tropics, and that it is there
that so much of their evolution has taken place. In many
groups, the temperate and still more the arctic species are
relatively recent invaders, often with special modifications
fitting them to the new conditions. This seems to be true, for
example, of our social wasps.

FOOD AND WATER

Insects are not only dependent on the external temperature
but also on some external source of water. Active life is
possible only if the body fluids are kept at a steady level.
Some features of the insects' external skeleton serve to retain
water as well as heat; such are the felted hairs which cover
some species. But this particular problem is more often
solved by modifications of physiology or behaviour. Most
species extract all water from their excreta before voiding
it, as will be seen later in the termites. Many species do not
actually drink water but obtain it from their food. Fraenkel
and Blewett showed in the flour moth that caterpillars reared
in very dry flour might produce a chrysalis containing more
free water than had been present in all the flour which had
been eaten. The caterpillars must therefore have combined
part of the flour with oxygen to form extra water. This
would explain their observation that caterpillars consume
more solid food on dry than on damper flour.

The nests of social insects have among their functions that
of conserving water, and in artificial nests ants run about as if

21

THE SOCIAL INSECTS

distracted if they are not kept moist. The termites, because
of their soft external skeleton, are the most sensitive of all the
social insects to water-loss, and not only are their nests
waterproof but termites usually travel everywhere in tunnels
which they burrow beneath or construct above the ground.

GROWTH

The location of the insect skeleton on the outside of its
principal organs has two important consequences. To get
movement, there have to be many joints, and the character-
istic insect limb is one result. It is jointed in much the same
way as the armoured suit of a deep-sea diver. Again, growth
is possible only if the armour is taken off, and this is done
in the process of moulting. The external skeleton becomes
loosened from the tissues beneath and a new and larger one
is laid down beneath it. It is at first soft and wrinkled but
very soon after the moult it expands and hardens. In the
majority of insects, there is a fixed number of moults during
the growing period and they take place at fairly regular
intervals. Once the adult form is reached, no more moulting
occurs and there is no more growth. Thus one can roughly
estimate the age of a honey bee worker by the amount of
wear which its wings and skeleton show.

The physiology of growth and of moulting in insects has
been very actively studied during the last twenty years. The
story is complicated and not yet fully understood, but it
seems that moulting and the changes of form to which it
gives rise are controlled by the brain and by certain ductless
glands, mostly associated closely with it. In the bug Rhodnius
Wiggles worth showed that the early stages do not moult if
decapitated during the first Hive days after the large meal of
blood which normally suffices for the whole period. Bugs
decapitated later moult normally. This suggested that a

22

THE INSECT WORLD

substance was being produced in the head; when enough had
been accumulated in the blood, the insect could moult even
if the head was removed. This interpretation was confirmed
by decapitating two bugs, one which had moulted less than
five days previously and one which had passed this critical
period. The two headless insects were connected by a fine
glass capillary tube and it was then found that they both
moulted at the same time. This control of form and structure,

Fig. 3
Two 4th instar nymphs of
Rhodnius decapitated and con-
nected by a capillary tube sealed
into the neck with paraffin wax
(after Wigglesworth).

partly nervous and partly glandular, may, it is beginning to
appear, have great significance in some social insects. In
termites especially, the body sometimes seems to be moulded
by the needs of the community, and the intervention of the
brain may make this possible.

REPRODUCTION

In most insect species, the mature female, after being
fertilized by one or more males, lays between fifty and one
thousand eggs; the actual number largely depends on the
manner of life of the species. These eggs hatch out into a

23

THE SOCIAL INSECTS

young animal which in the less specialized insects is termed
a nymph. This differs from the adult chiefly in its smaller
size and in its undeveloped wings and reproductive organs.
Amongst social insects, this type of metamorphosis occurs
among the termites, sometimes called white ants. In other
insects, there has been a gradual divergence between the early
and adult stages, so that one gets for instance the familiar
contrast between the caterpillar and the butterfly. What
hatches from the egg is a larva which lacks many of the
characteristic features of the adult. It often happens that
when the larva develops in a protected situation with a rich
source of food it becomes a legless maggot or grub. This is
found in the ants, bees and wasps, but the grub is not especi-
ally associated with social life, since it is found equally in
solitary species. The larval and adult stages are specialized
for different functions, the former for feeding and growth,
the adult mainly for reproduction. In the simpler insects, the
early and adult stages are still sufficiently alike for one to
change gradually into the other. In the more complex insects,
they have become so unlike that a resting stage, or pupa, is
interpolated between them, to allow of extensive reorgani-
sation of structure.

Although it was stated earlier that the number of moults
is usually fixed, there are many exceptions. Thus the larva
of the carpet beetle can stand weeks or months of starvation.
If it is not feeding, it gets smaller at each moult. By a process
of alternate starvation and feeding, the American entomolo-
gist Wodsedalek got one of these larvae to live for five
years, getting first smaller, then larger. This potentiality for
variation in the number of moults is of great consequence in
the social life of termites, though it is not in them known to
be controlled by nutrition.

THE INSECT WORLD

BEHAVIOUR

Probably insects appeal to most of us in the first place
through our sense of beauty and strangeness. Most of us
have seen a red admiral on the Buddleia in our garden or an
eyed hawk moth on a paling. If we probe more deeply, the
story of insect behaviour can be just as engrossing. It is
soon apparent that many of their actions have, in some
sense, a purpose, but if we are not careful every word we
use carries into our discussions implications which it has
acquired in human intercourse. The red admiral feeds on the
nectar of Buddleia, but the female goes to nettle to lay her
eggs and it is on nettle that the caterpillar feeds. The be-
haviour of the butterfly has a purpose in the survival of the
species, but it is in the highest degree improbable that the
female knows what she is doing; she has no purpose in the
human sense. Her behaviour is instinctive: that is, it is not
learnt but produced automatically at the appropriate time.
We have no way of knowing what an insect is thinking nor
whether it is conscious at all, and it is best to avoid, as far
as possible, language which implies that we do know. When
we poke a wasps' nest with a stick, the " angry " wasps buzz
out and attack us. We use the word angry to avoid a clumsy
circumlocution, but we should not suppose that the wasp
feels as we do when our house is broken into by a burglar.
We cannot tell what the wasp feels, we can only see how it
acts. If sometimes in describing social insects it is convenient
to use words which have human applications, this is only for
convenience and in order to be more concise.

The simplest type of insect behaviour consists in immediate
reactions to messages received from the outside world — the
rays of the sun, the water vapour emanating from damp soil,
the smell or appearance of food or of the other sex. It is an

THE SOCIAL INSECTS

immediate reaction of this type which leads a grasshopper
to sit on the sunny side of a mound and to place itself so
that the largest possible area of its body is exposed to the
sun. Similarly the fully grown maggots of blowflies crawl
away from the light, and in nature this leads them away from
their food to pupate in the soil. In an often-repeated experi-
ment, the maggots are placed on a board which is lighted
from two sides by beams of unequal intensity. The track of
the maggot then divides the angle between its original posi-
tion and the lights in proportion to their brightness.

Another simple type of behaviour is the reaction to internal
messages, such as those from the stomach or alimentary canal
which lead an insect to search for food, or from the ripening
ovaries which lead to a search for an egg-laying site. A
hungry insect may become restless and merely wander about,
more or less at random, until it finds food. External stimuli
such as light, shade, and wind, will partly determine its path.
If normal spider beetles are put in a cage of which one end is
kept moist, the other dry, most of them will collect in the
dry end; but if they have been kept very dry for some days
previously, they collect at the damp end. Their behaviour
is a resultant between an external gradient and their internal
condition. Behaviour of this sort, where the pattern is simple,
is not usually called instinctive. Where the behaviour is
more elaborate, as when the red admiral seeks out a nettle
and lays eggs in a characteristic way on a special part of the
plant, the pattern is said to form an instinct. Instinctive
behaviour has a complex, inherited pattern. The performance
is normally perfect on the first occasion, so that learning is
not involved. The sequence of acts requires a particular
internal condition of the animal, but it can be carried to com-
pletion only in a situation which also provides certain specific
external stimuli.

26

THE INSECT WORLD

The hunting behaviour of the solitary wasp, Philanthus,
described by Tinbergen, is a good example of a complex,
inherited pattern. A hunting female at first uses her eyes and
examines any object of the right size, hovering in the air
4-6 inches away. An attack is only made if the object has

PM/ LAN THUS

Fig. 4
Hunting behaviour of the solitary wasp, Philanthus. A hunting
female at first uses her eyes and turns towards and examines any
object of the right size, hovering in the air four to six inches away.
An attack is only made if the object has the smell of a honeybee
which is the normal prey (after Tinbergen).

the smell of a honeybee, which is the usual prey. The attack
is usually made upwind, along the gradient of the smell.

Besides following their instinctive behaviour patterns, all
insects are to some extent influenced by their past experience,
that is they have some power of learning. Learning is of
more than one kind and the simpler types are probably
possible for most species, though only a few examples have
been well investigated. One sort which may be universal

27

THE SOCIAL INSECTS

is called habituation. Most animals which avoid noise and
vibration can learn that some regularly recurring noise is
harmless to them, just as we acquire the power of sleeping
through the noise of a nearby railway. Thorpe showed that
the vinegar fly will not normally lay its eggs in an agar
medium flavoured with peppermint. But if eggs are put in
such a medium, the flies bred from them no longer avoid it
when they in turn are laying. In some insects, habituation
has also a positive side and they may learn to seek certain
stimuli, just as many people prefer their native cookery to
that of other lands. Several entomologists have shown that
bean weevils can be bred on beans of several kinds, such as
runner beans and cowpeas. The weevils bred from each kind
of bean have a marked tendency to return to it rather than to
one of the other kinds if they are offered simultaneously for
egg-laying. This sort of habituation may account for the
observation that some insect species are divided into strains
which are attached to different food-plants, and it may also
account for the observation that social insects will rarely
admit into their nests an individual of their own species from
another colony. The stranger evidently does not have the
right smell, though sometimes he may be able to acquire it.
A more complex type of learning which is also very widely
spread is known as trial and error learning. This is demon-
strated clearly when for example a cockroach is placed in a
maze containing food. The maze is a box divided by barriers
into a number of passages, some of which are blind alleys
whereas others eventually lead to the food-chamber. It is
arranged so that the correct turning is sometimes to the right,
sometimes to the left, and the time which it takes for the cock-
roach to find its way through the maze to the food can be
recorded. In these experiments, the insect, just as a man
might, takes at first many wrong turnings, but gradually the

28

THE INSECT WORLD

correct path is learnt. A good deal of insect behaviour could
be of this type; but unless conditions are controlled so that
it is known what stimuli are being received, it is impossible
to be sure that the insect is not making a series of instan-
taneous reactions. It might, for instance, reach a source of
food whose smell became stronger and stronger the more
nearly it was approached, without any need to learn the
way.

FINDING THE WAY

A more complex type of behaviour is described as insight
learning. What this means can be explained by an example,
though there is still some dispute whether any insect is cap-
able of it. (In my own view, it is no longer possible to deny
some power of insight learning to some of the Hymen-
optera, including not only social forms like ants, but also the
solitary, nest-building bees and wasps). All bees and wasps
make a careful study of the nest-site, walking or flying round
it in a number of widening spirals or circles. It is often
possible to prove that they are memorising small local land-
marks.

Thus, Tinbergen describes how a circle of pine-cones was
made round the burrow of the wasp, Philanthus, when the
wasp was inside. When she came out, she flew round
studying the site for six seconds. While the wasp was away
hunting the cones were moved to form a circle one foot
away from the nest. When the wasp returned after ninety
minutes with a captured honeybee, she looked for the nest
in the circle of pine-cones and in no case found the entrance
until the cones were again placed round it. Experiments with
ants, to be described later, show that both the position of the
sun in the sky, and also a scent-trail, may help to indicate a
path.

29

THE SOCIAL INSECTS

Fig. 5

NEST

Fig. 6

How the alteration of landmarks confuses a female Philanthus
when returning to her nest with prey (after Tinbergen).

30

THE INSECT WORLD

The evidence for insight learning is based on more
elaborate behaviour of this type. A solitary spider-hunting
wasp, Anoplius fuscus, has been much studied. When it has
stung and paralysed its prey, it hides it in a tuft of vegetation
and, after making a short locality-study on foot, searches for
a place to dig its nest. Once when the nest was several feet
away from the spider, the Swedish entomologist Adlerz put
a white towel, of a size 2 x 3 J feet, on the ground between the
wasp's nest and the spider. The wasp normally visits the
spider several times while the nest is being constructed. On
such a visit, when she came to the towel she was completely
nonplussed. She would only walk an inch or two on to
it and, eventually, after casting round for half an hour,
she found her way round the edge of the towel and ran
to the spider. Coming back to the nest, she took a bee-
line across the towel and later the spider was dragged to
the nest by the same path. This suggests that the wasp
appreciates the relative positions of the spider and the nest
and that once the intruding object has been studied she
can pass from one to the other without using detailed land
marks.

Still clearer evidence for insight learning has been found by
Thorpe in the solitary wasp, Ammophila pubescens, which,
as we shall see, is known to be highly endowed in other
respects. This wasp catches caterpillars and drags them back
to her nest. When metal screens were put in her path she
merely made a detour round them and then went straight
on in the right direction. Even if the wasp was caught and
carried some distance in a dark box, she was usually able,
when released, to set off straight for her nest. She gave the
impression of knowing all the landmarks on a considerable
area of rough ground so well that she could quickly find her
nest from any starting point.

31

THE SOCIAL INSECTS

The knowledge of the relative positions of numerous land-
marks is something different from being able to follow a
single beaten trail. Social wasps, also, learn to know a large
district surrounding their nest. An experiment by Gaul in
the U.S.A. provides a curious commentary on this fact. He
brought a nest from a mile away and put it in a hive. The
adult wasps all deserted their nest and brood and were found
building a new nest at a few feet from the original site. They
will only stay in their nest in the hive if they are shut up in
it for the first eight hours.

As will be described later the honey bee has the further
power of communicating to its nest-mates information about
the location of sources of food. Most social insects, also, have
an advantage over most solitary ones in having a less rigid
type of behaviour. In many solitary wasps the successive
phases of nest-construction follow one another in a regular
sequence which does not allow for unforeseen accidents.
In Fabre's study of the wasp, Pelopaeus (now called Sceli-
phron), he found that when the mud-cells had been stored
with spiders, the whole group of cells was covered over with
an irregular daub of mud. If he removed the cells, the wasp
still daubed over the place where they had been. This irra-
tional rigidity is characteristic of all instinctive behaviour,
but is especially marked in solitary species, though there are
exceptions. No such complete rigidity is possible in social
wasps. The nest is growing all the time, and many different
jobs must be done in an order which depends on circumstances.
The wasps may have to fetch building material, or water, or
food, or add cells to the nest, or feed the brood. Thus each
individual worker has to do a greater variety of things than
any solitary species and must be prepared to do them in a
different order each day, since specialisation for particular
tasks is little developed in the wasps.

32

Jf

1

2. A mud-dauber wasp (Sceliphron caementarium). The female captures
spiders and paralyses or kills them by stinging. She stores them in mud
cells, laying one egg on each cell-full, thus supplying food for the grubs.

3. A sand wasp (Ammophila sabulosa) dragging a caterpillar to its burrow.

- v-

,. TTHHaHMHH|

, * - * «

4- Hunting wasp, Ammophila
sabulosa.

5. Another hunting wasp,
Sphex maxillosus.

6 and 7. Young nests of the German wasp (Vespula germanica), before any
workers have hatched. Part of the envelopes of the nest on the right have
been removed to show the comb.

THE INSECT WORLD

HOW SOCIAL BEHAVIOUR DEVELOPS

We may now consider what we mean by the term " social
insect ". Many solitary insects are gregarious, that is they
share certain common needs or react in the same way to
certain external stimuli so that dense populations assemble
locally. Many unrelated species of caterpillar may be found
feeding on one plant, and large numbers of assorted insects
may be trapped at a street-lamp. A suitable earthbank may
harbour the nests of hundreds of solitary bees or wasps, but
they never help one another, and indeed frequently fight, like
two families trying to share one kitchen. Gregariousness has
never led to social life in insects, though in the locusts, where
dense populations of nymphs aggregate into marching bands
and where the adults may fly in enormous swarms, there is
at least a rudimentary organisation of the behaviour of the
group.

All animals must behave so that their young will find the
conditions necessary for survival and growth. In the swift
moth this instinctive maternal care amounts to no more than
scattering eggs broadcast over a field, where there is a good
chance that the young caterpillars will be able to burrow into
the soil and to find the roots on which they feed. In this
moth, as in most other insects, the female also requires the
co-operation of a male for a brief period, to fertilise her eggs.
The female of the large cabbage white butterfly lays her
eggs in a group on a cabbage, the food-plant of the cater-
pillar. These when young sit side by side on the leaf, and
it is thought that this gives them some mutual protection
owing to their unpleasant smell and conspicuous colour.
The same sort of behaviour is seen in the cinnabar caterpillar
on ragwort. Other caterpillars, like those of the small
ermine moth on the garden euonymus, spin a joint web in

c 33

THE SOCIAL INSECTS

which a number hatched from one egg-group hide together
by day and from which they come out to feed by night. Any
advantage acquired by such association is essentially a social
one, but their behaviour has never become very elaborate.
A caterpillar is a relatively simple insect compared with,
say, an ant, and is in any case only a transitory stage.

All the advanced insect societies appear to have arisen
from a progressive development of maternal care, the social
unit being the mother and her offspring rather than a col-
lection of brothers and sisters. The first essential step is that
the female should continue to have contact with her eggs
after she has laid them, and usually after they have hatched.
It is impossible even to list the immense variety of ways in
which insects have improved on the simple act of laying their
eggs on or near the larval food-supply. Some ichneumon
wasps, instead of laying their eggs straight into the body of a
caterpillar, carry them about hanging from the ovipositor
until they hatch. Only then do they put the young grubs on
to the sawfly larva which forms their food. Though this
particular type of behaviour seems to have been a blind alley,
anything which prolongs the contact between the mother
and her young is a start on a possible road to social life. A
number of plant-feeding or sometimes predatory bugs brood
over their eggs and younger nymphs. The nymphs tend
to form a compact group and the female stands over them
when anything approaches, rather like a hen brooding over
her chickens. This example is also comparable with the
cinnabar caterpillars already mentioned, since the bugs are
brightly coloured and have an evil smell.

A group of insects, known as the Embiids, which are
found in most of the warmer parts of the world, are often
cited as an example of rudimentary social behaviour. They
are also of interest because they seem to be allied to the

34

THE INSECT WORLD

termites, in which the social life of insects reaches one of its
peaks. The Embiids are able to spin silk which comes as a
viscous thread from hollow hairs in their forefeet. They
weave irregular chambers on or under the bark of trees and
there live together in groups of some size. The group includes
the developing young as well as a number of males and
females. The chief advantage is probably the protection
given by the communal web, since each individual finds its
own food.

Other examples of maternal care which has developed into
rudimentary social behaviour are found in the earwigs and
in various beetles. The common earwig lays her eggs in a
small hole which she excavates in the ground. She sits with
them and turns them over occasionally and licks them. When
the mother is removed, the eggs always seem to die, probably
because a fungus grows on them when they are not cared for.
The young stay with the mother only for a few days after
hatching. Hinton has shown that one of the common British
rove beetles has rather similar behaviour. The female
excavates a chamber in a pat of cow-dung and lays her eggs
in it. She will defend them against intruders, including the
larger larvae of her own species. When the eggs hatch, the
young larvae stay with the female until just before their first
moult.

Rather more elaborate social behaviour, in which the males
also share, is seen in several other kinds of beetle. The dung
beetles, some of whose habits were described by Fabre in his
book on the " Sacred Beetle ", are an example. In Geotrupes
typhoeus, which is found on some of the commons round
London, the male and female co-operate in digging a deep
burrow and in storing dung as food for the larvae. The
female, at any rate, stays with the young for most of their
development. In some of the other kinds of dung beetle

35

THE SOCIAL INSECTS

the male and female also co-operate in rolling pellets of dung
to the mouth of the burrow, and both parents may stay in the
burrow until the young beetles of the next generation are
ready to leave it. In the burying beetles a male and female
co-operate in burying small dead animals by digging away
the earth from underneath them. The pair of beetles then

Fig. 7
Left, an ambrosia beetle, Platypus cylindrus. Right, a bark beetle,
Xyloborus saxeseni. (Redrawn from Munro)

live in a chamber surrounding the mass of carrion. The female
lays her eggs in the earth, in the walls of a tunnel leading out
of the chamber. After about five days, the young larvae hatch
and burrow through the soil to reach the chamber. Here, up
to the first moult, and for a short period after each succeed-
ing moult, they are fed by the female on regurgitated food.
It has been found that though some of them may develop
without this attention a much smaller proportion do so
successfully.

36

THE INSECT WORLD

Two allied families of beetles, the bark beetles and the
ambrosia beetles, though very different from the kinds which
feed on dung or carrion, have developed essentially similar
habits. The females, in some species assisted by the males,
excavate a tunnel in solid wood. A special fungus grows in
the tunnel, its spores being carried there either in a hollow
in the head of the female or in her alimentary canal, an
interesting analogy with what happens in the fungus-culti-
vating ants and termites, described later. The female lays her
eggs in side chambers and the young larvae may be fed by
the female on pieces of the fungus. In some species, the male
may guard the main entrance, perhaps chiefly to keep out
other males. These are remarkable societies while they last,
but they are never prolonged beyond the one generation.

FULLY SOCIAL BEHAVIOUR

It appears that to obtain a larger and more elaborate social
organisation the female must live long enough to overlap
with several generations of her young. Some solitary insects
do live long enough for this to be possible, though in most
of them there is very little overlap between the lives of the
adults of successive generations. An even more important
step is for more than one female to co-operate in looking
after the young. This happens in none of the elementary
social groups ; it provides the best distinction between social
and sub-social insects. A true social insect may be defined
as one in which the female tends or helps to construct a
brood-chamber for an egg (or larva) laid by another female.
This condition is realised only in the ants, bees, and wasps,
belonging to the order Hymenoptera, and in the termites;
the latter are unlike the Hymenoptera in that the males play
as big a part in the colony as the females. Only a very few
insect species have been able to develop fully social habits.

37

THE SOCIAL INSECTS

Of the 20,200 species of insects in the British fauna, only
thirty-seven ants, about forty-four bees, and seven wasps
are social. The proportions in other parts of the world are
probably much the same.

The insect society, with its store of food for the young,
is bound to provide a tempting site for a parasite to lay its
eggs. Moreover, as explained in Chapter 7, other females of
the same or allied species will often attempt to usurp the
incipient colony. Thus a certain " jealousy " of oviposition
by other females is nearly always shown by the female
founder of the colony. In humble bees and in most kinds of
social wasps the queen attacks any other individual attempting
to lay eggs, and this behaviour is aroused much more by the
act of oviposition than by the mere presence of another fertile
female. It is also true that if many females all lived together
and laid eggs indiscriminately, the number of young would
soon be too great for the available food-supply and for the
number of nurses. In all social forms a distinction, not always
very sharp, has thus developed between egg-laying queens
and sterile workers. Even if the workers do, in some circum-
stances, lay eggs, these are not fertilised and normally produce
males only.

It appears that the special form of sex-determination which
is found in most Hymenoptera made the attainment of social
life more easy in that group. The female can store up inter-
nally for long periods the sperm transferred to her by the
male ; and she is able to expose some of her eggs to sperm
just before they are laid, whereas others can be laid un-
fertilised. In nearly all Hymenoptera the fertilised eggs
produce females and the unfertilised ones males. Thus for
the maintenance of the colony males are needed only at
long intervals, when a new brood of young queens requires
fertilisation. Among the stimuli which determine whether

38

THE INSECT WORLD

the queen lays fertilised or unfertilised eggs are probably-
some depending on the general condition of the colony which
can thus adjust the sex-ratio to current needs.

In the typical Hymenopterous colony no males are present
for most of the year : the workers are sterile females and no
corresponding caste of sterile males has been evolved. The
termites, in contrast, have both sexes equally represented in
the sterile castes. In this group the female seems to be unable
to store for long periods the sperm received at her first
marriage flight, and she therefore requires frequent fertilisa-
tion. Moreover, the method of sex-determination is different :
unfertilised eggs do not develop, while fertilised ones give
rise to about equal numbers of the two sexes. It may be that
the difficulty of evolving two sterile castes is one reason why
the bisexual type of society arose only once, in the termites,
whereas the female-dominated type evidently arose four or
five times in different groups of Hymenoptera.

39

SOLITARY AND SOCIAL WASPS

Among the social insects the wasps seem to have changed
least from their solitary ancestors. Their colonies, even when
large and elaborate, never seem to attain to the high organi-
sation of the beehive or the ants' nest. Moreover, there are
good grounds for thinking that both bees and ants were
evolved in the remote past from ancestral wasp-like creatures.
The potentialities of the wasp-stock were probably all present-
in the ancestor of the two other groups.

A HUNTING WASP

Even a solitary wasp may be capable of elaborate behaviour,
and there is no better example of this than the species Ammo-
phila pubescens, which has become famous in the last few
years and is found in southern England. Many observers
had built up an incomplete picture of its behaviour over the
last fifty years, but during the war a very long and detailed
study of its habits was made by Baerends in Holland. He
marked many females with coloured paints so that each one
could be recognised again. In this way a detailed picture of
the nesting-cycle could be constructed. Soon after fertili-
sation, the female digs a nest in the form of a burrow in sandy
soil. The gallery descends for about an inch and then bends
at right angles and ends in a small oval chamber. Baerends
removed such nests bodily from the soil and replaced them
with a split block of plaster of Paris in which an artificial

40

SOLITARY AND SOCIAL WASPS

nest had been cast. Such a block when dusted over with
sand was accepted by the wasp, and had the advantage over
the real nest that its contents could be examined without
doing much damage to the site.

After making the nest, the female catches a caterpillar,
stings it until it is paralysed, drags it into the nest, and lays
an egg on it. The nest is closed, and opened again only
after some days when the wasp-grub will have hatched and
may need a second caterpillar. Adlerz in Sweden had sug-
gested that each female might be able to look after more
than one nest at the same time, and this Baerends found to be
true. Sometimes as many as three nests were kept going,
each in a different stage and needing different treatment. In
one the egg might not have hatched, in the second the grub
might need more food, whereas the third might be ready for
final closure. The female at first sight gives the impression
of knowing what treatment each nest requires, but experi-
ment showed that the " knowledge " was of a very primitive
type. Normally, the wasp makes a morning inspection of the
nests and this determines how she treats each one for the rest
of the day. Thus if on the first inspection she finds a nest
needs food, food is put into it even if a caterpillar has already
been placed experimentally in the nest. She seems to be
unable to revise her first impression until another twenty-
four hours have passed. Nevertheless the successful accom-
plishment of several elaborate duty-sequences within one
day, together with her marked ability in finding her way
about, make this wasp one of the most gifted of the solitary
species.

THE VARIETIES OF WASP

In Britain the word " wasp " is usually applied to one of
the common social species which eat fruit and sometimes

41

THE SOCIAL INSECTS

interfere with picnics. There are many other insects, includ-
ing even moths, with a similar black and yellow livery,
which are often mistaken for wasps and may indeed
gain some protective advantage if such mistakes are made
by birds.

Entomologists, however, apply the word wasp in the broad
sense to several large groups of insects, many of which do
not show the black and yellow colours. The ichneumon
wasps, for instance, lay their eggs in or on other insects,
more rarely in spiders, and other animals. The act of laying
the egg seems to disturb the host very little, and it may
survive for a long time. But in the end, after the egg has
hatched, the ichneumon grub eats the entire contents of its
host, so that nothing but an empty skin is left. An insect
which does this is usually called a parasite; but it can just as
well be called a specialised predator which kills its prey at a
convenient moment.

Solitary hunting wasps paralyse or kill their prey (which
may be a caterpillar, spider or something else of that sort) and
store it in a nest. Here again the prey is eaten by the grub
which hatches from the egg which the wasp has laid in the
nest. This type of behaviour is associated with a change in
the structure of the wasp's ovipositor: it becomes a sting. In
the ichneumon, the egg can be passed down between the
stylet-like blades of an ovipositor which thus acts like a
hypodermic syringe for injecting the egg into the host. In
the hunting wasp an egg pore is found at the base of sting ;
the ovipositor is then only a syringe for injecting poison
into the prey, or into any marauding enemies. A sting is
found in the nest-making ants, bees and wasps, though in
some of them it is not actually syringe-like.

The solitary wasps themselves are a very large group
with a great diversity of appearance and habit. Some of

SOLITARY AND SOCIAL WASPS

them behave in almost the same way as the ichneumon.
The Scoliid wasps, which are common in most of the warmer
parts of the world, attack the underground grubs of cock-
chaferlike beetles. The wasp has powerful spiny legs by
which she can work her way underground to find her prey.
Some species sting the grub and lay an egg on it without
making any nest, leaving the grub in its own burrow. Others
do make a small underground cell to which they drag the
grub before laying on it. Most of the solitary wasps, how-
ever, make a nest, usually a gallery ending in one or more
cells either in the soil or in rotten wood. A few of them can
build beautiful mud cells, as described for instance by Fabre
in Pelopaeus (Sceliphron) ; less commonly they build cells
of vegetable matter such as leaves or resin.

The nest-making solitary wasps fall into three principal
groups which have a different pattern of behaviour. The least
advanced type are the spider-hunters which first catch and
paralyse a spider, then hide it and look for a nesting-site.
Usually, they make only one cell at each place, and when
they have stored the spider in it wander off and repeat the
process. This pattern is not very different from that of
Scolia, and only very few of the spider-hunters make mud-
nests or construct several cells in one place. This means that
the position of the single cell has only to be remembered for
a short time, at most for a day or so, after which it is visited
no more.

The sand wasps always make a nest before catching their
prey; this may be any sort of insect or spider. The nest
consists of several cells which are typically sub-divisions or
branches, often slightly enlarged, of a long tunnel. When
several prey are stored in the cell, the egg is normally laid
on the last one brought in and the cell is then closed. Thus
the mother has no contact with her young, but she must

43

THE SOCIAL INSECTS

become very familiar with the surroundings of the nest since
many, sometimes all, her cells are in the same spot.

A third group, the Eumenids, resemble the social wasps in
many details of structure, but build nests very like those of
the sand wasps. They have one habit which has been
retained in the social species : the egg is laid in an empty
cell and the prey, which are paralysed caterpillars or beetle
grubs, are brought in afterwards.

These methods of providing for the young are known as
mass-provisioning, since the egg is enclosed in a cell with
enough food for it to complete its development. The food
is supplied whole, in the form of paralysed or sometimes dead
insects. A few species both of sand wasps and of the Eumenid
wasps show an important advance in behaviour; they
practise progressive provisioning. These wasps lay an egg
either on the first prey brought in or in the empty cell and
then wait until the egg hatches before bringing more food.
In this way the mother has some contact with her young.
Ammophila pubescens, already described, provides one of
the best known examples of such behaviour. In some of the
other species which practise progressive provisioning, such
as the African wasp Synagris, the female may feed her
young with chewed fragments of prey instead of whole
specimens.

SOCIAL WASPS

It is from wasps with this sort of equipment, though not
from any living solitary species, that we imagine that the
social species arose. The ancestors of the social wasps were
thus sting-bearing hunters, capable of building nests of some
complexity, capable of learning their location with great
exactness, and beginning to have some contact with their
young. However, in the solitary species each nest is made

44

SOLITARY AND SOCIAL WASPS

by a single female and the individuals are either males or fully
fertile females. In the social species, at least some of the
females are relatively or quite infertile and these assist in the
care of the young of the fully fertile females, or queens. The
females with reduced fertility are almost always quite incap-
able of fertilisation by the males and form the worker caste.
A further feature in which the solitary wasps differ from the
social ones is that the female rarely lives much longer
than the male, and successive generations of adults hardly
overlap.

The most familiar social wasps in Britain and North
America are species of the genus Vespula. These will be
described first, although some of the less familiar tropical
species have simpler behaviour and seem to come nearer to
the solitary species. These common wasps have two sorts
of females, the egg-laying queens and the workers. The latter
are considerably smaller than the queens, have a slightly
different pattern tff yellow markings, and are not fertilised.
Males occur for a rather short period at the end of the sum-
mer, when they fertilise the young queens of the next genera-
tion. The workers cannot correctly be described as sterile,
since some of them may lay eggs, especially towards the end
of the season or even sooner if the queen dies ; but these
unfertilised eggs produce only males, so that the workers
alone could not carry on the species.

The young fertilised queens may stay for a little time in the
nest where they were born, but quite soon they begin search-
ing for winter quarters, since they are the only members of
the colony which hibernate. At the first frosts or earlier, the
males and workers die off and all that is left is the young
queens hiding in the ground, under dead leaves, or under
bark. A few of them may try to hide indoors, but it is usually
too dry for them there and the warmth prevents them from

45

THE SOCIAL INSECTS

resting undisturbed until the next spring. Under natural
conditions they pass into complete immobility with wings
and legs folded up, and they do not begin to move again
until March or April. Almost at once, on becoming active
they begin to look for a nesting-site. In the two commonest
English species, the common wasp and the German wasp,
this is usually a hole underground, often an old rat- or mouse-
hole. Here the queen builds her first nest. This, like those
of all the social wasps, is made of a sort of paper. The paper
is formed from wood fibres scraped from posts, palings,
stems of dry hogweed or rotten wood. The marks of wasps'
mandibles can be seen on almost any unpainted wood fence in
England, forming little scratches often tending to run into
lines. The wood fibres are chewed up with saliva to make a
paste which when spread out thin dries into a grey or
brownish paper. The common wasp uses mainly rotten
wood and its paper is yellower, softer and more brittle ; the
German wasp, so-called because the eighteenth-century ento-
mologists received specimens from that country, uses sound
wood, and its paper is greyer and tougher.

The beginning of the nest is a little pillar, hanging often
from a root which projects downwards from the roof of the
hole. At the end of the pillar (which is about half an inch
long) is built a cell, and round the sides of the first one others
are added. The beginning of the first cell is a hemisphere,
hanging upside down; the second cell starts as a sort of
crescent-shaped pocket on one side of the first. Later cells
are added in the angle between two earlier cells. When a few
cells have been started the queen begins building envelopes.
These are dome-shaped covers attached near the point from
which the pillar hangs or, in the later envelopes, to an earlier
one. They form cup-like domes over the cells and later com-
pletely enclose them except for a small more or less circular

46

SOLITARY AND SOCIAL WASPS

entrance at the bottom. The construction of several (two
to four) envelopes with air in between provides the best sort
of heat-insulation that can be imagined. Eggs are laid in
cells which have only been started, but when the grub hatches
and begins to grow the cells are lengthened to keep pace
with it. The first cell remains circular in cross-section, but
the later ones become more and more regularly hexagonal,
each of their walls, except those on the periphery of the
comb, being shared with another cell. It can be shown that
this arrangement makes possible the most economical use
of a given quantity of paper. It seems that as wasps do not
vary much in size, they can use their own bodies as a pair
of dividers for measuring a standard distance.

The queen wasp never makes more than about ten to
twenty cells, since when the grubs are growing actively it is
all that she can do to keep them fed unaided. It seems clear
that adult wasps can take only liquid foods, since the entrance
to their throat is'too small to admit more than the minutest
solid particles. Probably most of their food is nectar from
flowers, juices of fruit and sweet things of that sort. But to
feed their young they catch and cut up other insects, such as
flies or moths. These are pounced on when settled. They are
not stung but cut up and chewed into a paste. The wasps
will also cut small pieces from dead animals. Probably some
of the juices extracted from the paste form a necessary sup-
plement to their own diet. Whether the young receive much
sweet stuff as well as animal matter is uncertain. The grubs
live and grow upside down, since the opening of the cell is
at the bottom. When full-grown the grub spins a silken
cocoon which forms a white cap across the bottom of the
cell and a thin lining to the rest of it.

The life of the queen with her first cells differs in no
important respect from that of a solitary wasp practising

47

THE SOCIAL INSECTS

progressive provisioning, but when the first workers bite
their way out through the cocoon a real social unit has
arrived. With the help of the workers the comb begins to
grow rapidly, and a new one will soon be started. To contain
the combs, the envelopes will be reconstructed: the paper
is chewed away and used with new material to make bigger
ones. The queen gradually stops leaving the nest and soon
performs no duty beyond egg-laying. Later in her life, when
she has been laying eggs rapidly for a long time, her abdomen
takes on a discoloured greasy look and she loses all power of
flight, so that the old queen is easy to recognise.

The second comb starts as a pillar hanging from the centre
of the first comb, and the process of growth is much the same
as in the first ; but large combs are always supported from
the one above by several pillars, and their edges are attached
to the envelope by small lateral struts. The workers get from
one comb to the next by crawling round the edge between
the comb and the envelope. As the nest grows, it is nearly
always necessary for the wasps to enlarge the hole in which
it is hanging. They do this by bringing in water to soften
the earth, and in a populous nest workers can often be seen
flying out one behind the other carrying little pellets of mud.
A big root or stone which they cannot remove may cause
the nest to acquire an irregular shape.

As the comb starts from a cell near its centre the first eggs
are laid here too, and as a result concentric rings of develop-
mental stages are often seen in a comb. The centre may
consist of cocoons, surrounded by rings of grubs getting
progressively smaller towards the periphery, till finally there
are rings of cells either with eggs or not yet occupied. How-
ever, after a new worker has bitten its way out of the cocoon,
the cell is cleaned up and used again, so that in a somewhat
older comb the central cells will contain eggs once more,

8. Nest of the Norwegian wasp
(Vespula norwegica) in a black-
currant bush.

9. Nest of the same
species opened to show
structure.

io. Part of a fence from
which wasps have rasped
away wood - fibre for
nest-building ( " paper-
making ").

ii. Comb from the nest of a
wasp (Vespula) showing closed
pupa cells and other cells open.

.«& jr

12. Queen German wasp (Vespula germanica) found hibernating.

13. Polistes wasp feeding larva. Two other larvae are clearly seen and some
unhatched eggs. This wasp and those in 14, 15, 16 and 17 belong to an
American species of the genus Polistes.

#

14 and 15. When a wasp larva is fully grown, its lower lip exudes a sticky
secretion with which it spins a cap over its cell. On the left the larva has
begun spinning and on the right the cap is half done.

16. Polistes worker in flight by nest. The photographer reports that this
wasp had waged " successful warfare " against him.

SOLITARY AND SOCIAL WASPS

surrounded by a ring of cocoons. After a few cycles of this
sort the arrangement tends to get out of step, and any comb
in which all stages are found mixed together is known to be
an old one.

The development from egg to adult takes about thirty
days ; the exact time depends partly on the temperature and
partly on the age of the nest. A comb which has the eggs,
grubs and cocoons irregularly arranged must be at least ten
weeks old. A cell which has been used before can always
be recognised under the microscope by the remains of the
thin silken lining which formed part of the cocoon, and by
the dried larval excreta which accumulate at the bottom of
the cell.

A successful nest in the late summer may contain thou-
sands of cells and perhaps two or three thousand workers.
A nest of the common wasp, which I dug up in a field at
Slough on September n, 1935, contained a queen and
3,008 workers. ,It consisted of eight combs, with 11,299 cells
in all. The fifth and sixth combs were the largest, each with
about 2,188 cells. Nearly all the cells contained brood in
various stages, so that many more wasps, including young
queens and males, would have been produced later.

No nest with more than one or two hundred workers can
be safely examined without destroying the wasps. This can
be done most conveniently with the dangerous poison
potassium cyanide : a piece as big as a walnut placed at the
entrance will rapidly kill every adult wasp. Carbon bisul-
phide and some other poisons are almost equally effective.
The Slough nest was not unusually large, and much bigger
ones, with up to ten thousand workers, are found in some
years. In such nests a great deal of food has to be brought
in, and large numbers of flies and other harmful insects are
consequently destroyed. Unfortunately, wasps are not very

d 49

THE SOCIAL INSECTS

particular about what they catch, and they will pick the blue
butterflies off their evening resting place on grass-stems,
besides other insects, or butcher's meat if they can get it.
Wasps need sweet stuffs and water. It is doubtful if they
often damage sound fruit, but they will enlarge any hole
started by other insects or by birds. Soft, over-ripe fruit is
readily attacked and wasps are very fond of blackberries in
the hedges. They visit flowers not only to catch flies but also
to get nectar, being very fond of figwort and wild parsnip.
Wasps have a sting with microscopic barbs, but these do
not prevent them from withdrawing it after use. They use
the sting only in defence, especially of the nest, not for killing
their prey — an interesting social trait distinguishing them
from the solitary species. One soon finds that a wasp at or
near its nest is a much fiercer enemy than one flying about
in the open. Association with the nest seems to call out a
characteristic aggressive behaviour pattern. Moreover, the
wasps from larger nests are fiercer, but this may be because
such nests are usually hotter inside and this increases all
aspects of insect activity.

THE WASP SOCIETY

Several entomologists have measured the temperatures of
wasp-colonies. According to Himmer, a thriving colony of
the common wasp runs at 26-360 C which is 5-1 50 C above
the temperature of its surroundings ; the temperature varies
much less than that of the air outside. The protected situation
of most nests, and the good insulation provided by the
envelopes, make some degree of temperature control easy,
since any heat produced will be lost only slowly. The
numerous larvae continually digesting rich food and the
crowds of active workers are of course at all times providing
the necessary central heating.

50

SOLITARY AND SOCIAL WASPS

Since the length of the developmental period, and prob-
ably also the egg-laying rate of the queen, depend on tem-
perature, the control of temperature is very important. It
is well known that the developmental period is shorter in
large colonies than in small ones. If the nest gets too hot
some of the wasps stand at the entrance and fan with
their wings. This has been shown under experimental
conditions to be moderately effective in the hornet. When
a nest was heated with an electric heater the wasps began
actively fanning at yj° C. This kept the temperature
steady for a few minutes. When, in spite of fanning, it
began to get hotter, they flew off and abandoned their
brood.

We may now consider the behaviour of the colony in
two of its most distinctively social aspects, the relation of
the workers to the grubs and the division of labour amongst
the workers. The larger wasp grubs are able to call attention
to themselves by bending their heads backwards and scraping
the wall of the cell with their mandibles. When a number
of them do this together quite a distinct rustling sound may
be heard. It is possible that they do this more often when they
are hungry. But they have another and probably more
important way of attracting the workers. The labial gland,
which when they are full-grown produces silk for the cocoon,
produces earlier in life a sweetish secretion. This appears as
a little drop on the lower lip, and can be elicited by stroking
the grub's head with a straw. The workers lick up the
secretion whenever they feed the larvae. It has been sug-
gested that this mutual exchange of food is a very important
factor in maintaining social ties, but there is very little evi-
dence that the liquid obtained from the grubs is of much
consequence to the workers. It may be that the process is
really of more importance in nest-hygiene, the workers

5*

THE SOCIAL INSECTS

removing liquid that the grubs could not otherwise dispose
of conveniently.

It is clear that in a large flourishing wasps' nest there are
many different jobs to be done, and one would like to know
whether there is any division of labour amongst the workers.
In one sense there evidently is such division, since at any
one moment the workers are not all doing the same thing.
Some are standing guard over the nest, others are feeding
the young, others getting food or paper. The observations
of Weyrauch in Germany suggest that an individual may
tend to do a particular sort of work for some time, perhaps
a day or two, but that there is no rigid division of labour.
If some particular task, such as defending the nest, becomes
temporarily more important, all will take part in it, and in
any case during its life-time one worker will probably do a
little of everything. In the honey bee, in which the workers
are all very similar to one another in structure, there is a
division of labour largely based on the age of the individual,
each worker tending to perform different duties as it gets
older. In the wasps, the workers are also very similar to one
another, but the relation between age and duties has hardly
been studied. According to Gaul, the younger wasps work
in the nest and do not go out to forage, but from Wey-
rauch's records of the relatively rapid changes of occupation,
it is not likely that there is an elaborate cycle of duties.

Towards the end of the life of the colony the workers
suddenly start building larger cells. The line between these
and the smaller ones may cut sharply across the bottom comb,
but any combs added later below this will consist of large
cells only. The young queens come exclusively from eggs
which have been laid in the large cells. The males develop

5*

SOLITARY AND SOCIAL WASPS

in either sort of cell, but only in the lowest of the small ones.
How the queen comes to distribute her eggs in this way is not
known. The whole success of the colony is measured by
the number of large cells and therefore of young queens
which it can produce. In a temperate climate with a limited
breeding season, there must always be an optimum arrange-
ment. If worker production in small cells goes on too long,
there may not be time before the frosts to rear many queens.
On the other hand, the larger the nest and the greater the
number of workers, the more queens can be reared.

Weyrauch has published records of the number of queens
and males produced by colonies of the different species of
wasp in Western Germany. The figures show that the bigger
the nest and the more combs of worker cells, the more queens
are produced. On the average, something between twelve
and forty worker cells correspond to each queen cell, and
these figures give some measure of the effort required to
produce a large, crop of queens. The colonies produced
between thirty and two hundred queens each, depending on
the species. The males appear to be about twice as numerous
as the queens.

The sudden change in the type of cell constructed corre-
sponds to an important change in the whole behaviour of the
colony. After the change, it is supposed, either special queen-
producing eggs are laid, or else from then on the grubs are
given some special or more abundant food. But the question
of what makes queens develop instead of workers is the major
unsolved problem in the biology of wasps. There is really
no reliable information on which .to found a theory. There
is no evidence that a special or more abundant food is sup-
plied to the queen larvae. With progressive provisioning, a
larva will be given food as long as it accepts it, so that a
longer developmental period would mean that more food

53

THE SOCIAL INSECTS

was supplied and a larger wasp would be produced. But this
would only put the problem one step back, since we should
want to know why the grubs in the large cells took longer to
mature. Janet's records for the hornet show that the develop-
mental period is longer in the early part of the season, but
there is nothing to show that queens develop more slowly
than the workers which are being reared at the same time.

In the honey bee the queen, although larger than the
worker, develops more rapidly, probably because of the
richer food she is given as a grub. On the whole, the evidence
in wasps suggests that the important thing is likely to be
some change in the behaviour or physiology of the old
queen rather than a change in the food supplied to the
grubs. Common wasps and hornets are the only social
wasps in which the queen-cells are easily recognisable ; in
the others cell size is almost uniform.

When the young queens emerge from their cocoons they
mature in a few days and leave the nest rather rapidly. Pair-
ing takes place in the open, probably as often with males
from other colonies as with their brothers. Very soon after
pairing the young queens enter hibernation. Meanwhile, the
social organisation of the colony degenerates. At this stage
the workers may kill many of the remaining grubs, either
flying out with them and throwing them away, or using them
as food for the others, or even eating them themselves. No
convincing explanation of this behaviour can at present be
given, but from Deleurance's experiments on Polistes it may
turn out to be associated with a glandular degeneration of
the old workers. The failure of the nurses seems sometimes
to be associated with the exhaustion of the ovaries of the
old queen or with her death, but it may also occur in nests
where she survives. The final end of the colony is usually
determined by the onset of cold weather, a few sleepy

54

SOLITARY AND SOCIAL WASPS

stragglers being found in the nest up to some date in October.
According to Duncan, nests sometimes survive much longer
in the mild winters of California, and may even survive the
winter.

Besides the sort of cannibalism described above, there
appears also to be some eating of eggs, probably both by the
queen and by workers. Such behaviour is common in most
of the social insects, and a queen will sometimes eat her own
eggs soon after laying them. As will be explained in chapter
7, egg-eating may sometimes tend to protect the colony
against parasites which try to lay their own eggs in the cells ;
some egg-eating may be the price which has to be paid for a
type of instinctive behaviour which is at other times socially
valuable. It is also certain that the regulation of the number
of eggs is essential to the well-being of the colony, and egg-
eating might be regarded as a form of birth-control. It is
unlikely, however, that either of these explanations can cover
all the facts. In most insects some of the B vitamins are
essential for growth and normal health, and it is possible
that the eggs are a source of such substances. The needs of
wasps for vitamins have not yet been studied.

Later in the season the workers, or some of them, may lay
a number of unfertilised eggs, and unlike the queen they may
lay more than one in the same cell. If the queen dies, worker
laying may occur on a much larger scale and will lead to the
production of many male wasps.

It is a familiar fact that some years are " wasp years "
whereas in other years wasps are scarce. This seems to be
due much more to the number of nests which are successfully
established in the spring than to the number of queens pro-
duced in the autumn. In other words, the hazards of hiber-
nation are less than those encountered during colony-
foundation in the spring. For this reason, the campaigns

55

THE SOCIAL INSECTS

which local authorities in Britain have at times subsidised
(e.g. in Berkshire in 1901), paying school-children to collect
queen wasps, do not seem to be successful. In 1950, up to the
June 15, the Department of Agriculture for Cyprus destroyed
243,394 queens of the local species of hornet, paying rather
more than a farthing for each one brought in. Considerable
numbers were also destroyed by other interested parties.
Later in the same year the Department's officers destroyed
27,829 nests. This was one of the worst wasp years recorded!
Beirne, by comparing the occurrence of wasp years in
Britain with the meteorological record, has shown that an
important factor is the weather in April, May and June, when
the queens are establishing their first nests. Rain at that time
is very bad for wasps. Early warm weather followed by a
cold or wet spell is also harmful, since the queens all come
out of hibernation but are then unable to establish themselves.
Prolonged cold weather which merely prevents them from
leaving winter quarters is harmless. Nixon has maintained
that, apart from weather, the most important enemy of
wasps is other wasps. Suitable nesting-sites are usually not
sufficiently numerous and attempts at usurpation lead to
fights in which one or both queens may be killed. It is quite
common to find a dead queen lying in the tunnel leading
down to the nest. Other enemies are badgers which dig up
nests to eat the grubs : evidently their fur protects them from
stings. It is doubtful if this enemy is common enough to be a
serious check on numbers.

WASPS IN BRITAIN

There are five other species in Great Britain, all super-
ficially very like the two common wasps. One of them
is a parasitic wasp without workers and will be men-
tioned in Chapter 7. Another, the Norwegian wasp, builds

56

SOLITARY AND SOCIAL WASPS

nests suspended in bushes or trees instead of underground.
It is considerably commoner in the north and west than it is
in south-east England. The hornet, Vespa crabro is much
larger than the other species, and has lately been getting
much more common. It is chiefly found in southern, especi-
ally south-east England, and twenty years ago was rarely
seen, except in the New Forest and in some parks near
London, such as Richmond Park. It is now quite common
all round London, though being so conspicuous it does not
take many to produce an impression of numbers.

It seems that the hornet is not nearly so ferocious as its
appearance suggests. It is much the largest of our wasps, and
its colour makes it very conspicuous; this may indeed have
some value in preventing attacks by insectivorous birds.
But from the human point of view what matters is the willing-
ness of the wasp to use its sting. I have never been stung by
a hornet, and I have been told by a friend who had the task
of removing a nest from the roof of a thatched cottage that
he was able to do this without wearing any protective cloth-
ing and without being stung by the workers which were
flying around. One certainly cannot do this with the larger
nests of the common wasp, and this may be partly due to the
relative numbers of workers. As we have seen, the more
workers there are, the more aggressive their behaviour, and
a hornets' nest rarely has more than a few hundred in-
habitants.

Wasps very like the British ones are found in all countries
with temperate climates, except in South America and the
Australian region. The tropics are mostly occupied by social
wasps of other types. The hornet, though itself an inhabitant
of temperate climates, belongs to a tropical group found
chiefly in Asia and the East Indian islands. The hornet some-
times works very late on warm nights. Amongst its Malayan

57

THE SOCIAL INSECTS

allies are certain very peculiar wasps with a pale ghost-like
colour and enlarged eyes which are entirely nocturnal. Very
little is known as yet of the biology of any of these tropical
hornets. It may prove to be interesting, since there is already
a hint that in them the queen and worker are much less
different from one another than they are in the temperate
species.

58

VARIETIES OF WASP SOCIETY

The social wasps described in the previous chapter are
familiar because they are common in those parts of the world
in which most naturalists live. In other parts of the world,
especially in the tropics, several other groups are found.
Two of these groups, the Polistes wasps and the Polybiine,
wasps, are of special interest because of the light they throw
on the nature of social life and on its probable mode of origin
in the family.

THE POLISTES WASPS

The Polistes wasps are found almost everywhere, both in
temperate and in tropical climates. Though common in
North America and in Western Europe, they do not occur
in Britain except as an accidental introduction. They have
been much studied, since the whole history of the colony
can be followed with ease. This is because the nest is small,
is suspended above the ground, and is not surrounded by an
envelope. The European species build a nest rather like a
single, irregular wasp-comb, usually of less than one
hundred cells, but occasionally in the south with up to five
hundred.

In temperate climates the life-history is like that of the
common wasp. In external appearance the workers can
hardly be distinguished from the queen except by their
behaviour. It is, however, possible to mark each individual

59

THE SOCIAL INSECTS

on a comb with different coloured paints so that each one
can be recognised. It is then found that one individual, the
queen, stays in the nest and does all or almost all the egg-
laying. She receives a relatively large share of the food
brought in by the foragers. The workers do all the food and
paper gathering and most of the nursing. The workers are
rather smaller than the queen, on the average, but there is a
wide overlap in size. If the wasps are dissected, the queen is
found to have a much better-developed ovary and to have
sperm in the receptacle where they are stored. Neither of
these features, however, may be present in the young autumn
queens before hibernation, and these can be distinguished
from the workers only by their inactivity. The workers'
ovaries are smaller than those of the old queen though better
developed than in the workers of the common wasp. Workers
are never fertilised. No special queen cells are made when the
colony is mature.

Deleurance, who has kept nests in his laboratory in Paris,
has shown that if the queen is removed the workers may
begin laying male-producing eggs within twenty-four hours.
This proves that their failure to do so in the ordinary way
is not because they are unable to lay eggs, but because their
egg-laying is inhibited by the presence of the queen. This
inhibition may be partly due to active interference by the
queen. To lay an egg a wasp has to enter a cell backwards,
and this action might stimulate the queen to threaten or to
attack. In nature the queen is rarely observed to make such
attacks, and probably her mere presence inhibits egg-laying
by the workers. This interpretation is supported by Pardi's
observations in Italy. In many nests the queen dies rather
early in the history of the colony, and the workers then begin
laying eggs. Pardi found that in late summer some nests
produce queens and males in about equal numbers, whereas

60

VARIETIES OF WASP SOCIETY

others produce nearly all males. The second sort of colony
is one in which the queen has died early.

At the end of the season many of the grubs in the Polistes
colony die and may be thrown away by the workers. This
seems to be much the same as the autumnal destruction of
grubs by the common wasp, though in Polistes they are not
killed by the workers. Deleurance has obtained good evi-
dence that the production of "aborting brood" is due to some
lack in the food provided by the workers, probably in some
glandular secretion which they are no longer able to supply.
By rearing his wasps in captivity at a constant temperature,
he was able to have colonies at all stages of development
simultaneously. He could then show that brood from young
nests aborted if given to old wasps and, conversely, brood
from old nests could complete its development successfully
if the old workers were replaced by young ones. Since,
however, in both experiments the wasps were given plenty
of insect food, abortion cannot be due to simple scarcity.
It is probably due to the exhaustion of one of the glands
whose secretion is mixed with the food. This idea is supported
by an experiment in which the workers were kept cool (40 C)
at night, so that they did not age so rapidly. With this
treatment, brood could be reared normally in the laboratory
for seventy days instead of the usual maximum of thirty-five
days.

The species Polistes gallicus is particularly interesting
because its habits have been studied over a wide range of
climates, extending from north Germany to the oases of the
Sahara. In the north, colonies are usually founded in the
ordinary way by a single queen after hibernation. In south
Germany, more than one queen is often seen on a nest.
Heldmann noticed that several queens might begin to build
nests on one beam of a shed. When none had more than

61

THE SOCIAL INSECTS

three or four cells, there might be a desertion by one queen
who would join another, so that some nests were eventually
founded by two or three queens. At first all such queens
took a nearly equal share in the work and each of them would
lay eggs. But after a time one queen became dominant and
did more and more of the egg-laying, leaving the nest less
and less often. The others gradually ceased to lay eggs and
took on all the foraging duties.

In north Italy, Pardi found that nearly every colony of
this wasp was founded by three or four queens. He has
described a process of active " terrorisation " by which one
of them becomes the real queen. She will attack the others
with open jaws, making a loud buzz. After a time she
establishes her position so securely that force is not required,
and the ovaries of the other queens degenerate, either because
the wasps work harder and have less food or because they
are not allowed to lay. Amongst these worker-like queens,
which Pardi calls auxiliaries, there is a regular order of domin-
ance, in which the queen always comes first. The position of
a wasp in the scale is shown most clearly in the disposition
of food, but it appears also in the frequency with which it
shows aggressive behaviour and lays eggs, and in the amount
of time it spends on the nest. A wasp high in the scale is
more likely to be given food by one of its inferiors and less
likely to give it away. Similar orders of precedence are known
in many mammals and in birds, such as chickens, which
live in groups. According to Pardi, when the first brood of
workers emerges, the auxiliaries all disappear, so that they
serve only to get the colony started and as far as is known
never found a colony of their own.

In the Sahara, Weyrauch found that this same species of
Polistes founded colonies by swarming; that is, the founder
queen is accompanied by a number of workers. This is

62

VARIETIES OF WASP SOCIETY

possible only in climates where breeding is more or less
continuous. In the tropics also, many species of Polistes
seem to swarm, though colonies founded by single queens
are common too. It seems probable that in Polistes swarms
the queen is a young, recently fertilised female and not the
old mother as in the honey bee. Much more study of this
point is still required.

This variation in the behaviour of Polistes from north to
south may provide an important clue to the origin of social
behaviour. A colony in which only one female lays eggs
loses a large fraction of the potential birth-rate; the com-
pensating advantage is the greater safety and better care of
the young. Any auxiliary or worker is an egg-layer lost but
a nurse or hunter gained. According to what is the most
likely reason for failure of the wasp-colony, it may pay to
sacrifice more or fewer queens in favour of nurses. In the
north, where climate is perhaps the chief difficulty to be met,
direct multiplication may be more important. As climate
becomes less severe, it is possible that enemies of various
types become more numerous, and it becomes advantageous
to sacrifice up to two-thirds of all the queens in order that
nests shall be safely founded.

The process of swarming depends in the first place on a
climate sufficiently favourable to allow continuous breeding,
or on a worker caste capable of hibernating like the female.
On the southern edge of the South American tropics, where
the winter is short, whole swarms of Polistes apparently
do hibernate together. (The honey bee, of course, with-
stands the winter by storing honey for the period when
food cannot be obtained.) A few of the Polybia wasps
described below are able to do this too, but Polistes never
store more than small drops of nectar in some of the cells.
Swarming involves a great sacrifice of direct multiplication,

63

THE SOCIAL INSECTS

since every worker which accompanies the queen has been
bred from a cell which might otherwise have produced
another queen. One hundred cells, at the end of the season,
might produce ten queens, ten males and eighty workers.
The males would fertilise the females and die, and ten swarms
of one queen and eight workers could be produced. On the
other hand, in the absence of swarming, something like fifty
males and queens might have been produced, with the
potentiality of six times as many new colonies. The only
possible conclusion seems to be that swarming is advan-
tageous because the colony is at all times protected by a
number of workers.

Owing to its small size, to the absence of an envelope,
and to the relatively exposed situation in which it is built,
the colony of Polistes cannot avoid being cooled at night.
In Europe, these wasps usually build the nest in a sheltered,
south-facing situation, so as to get the full heat of the sun,
and they have two effective methods of cooling it if it gets
too hot, as it sometimes does when the sun's rays fall directly
on it. In the first place they fan very vigorously with their
wings. If the temperature continues high, some of them
fetch drops of water which they put into the cells. The
fanning is continued and the nest is cooled by evapor-
ation. The nest may be maintained at 120 C below the
temperature of an unoccupied nest placed nearby for com-
parison.

According to Deleurance, the cooling of the nest at night
plays an important part in determining whether a grub will
turn into a worker or a young queen. He had earlier shown
that in artificial conditions queen-producing eggs might be
laid and the grubs might even have spun their cocoons before
any workers had hatched, so that the developmental stage of
the whole colony does not seem to decide whether or not

64

17- A group of three'Polistes wasps share a load of food brought in by a
forager. Other wasps are feeding larvae or adding to the comb.

1 8. A Polistes wasp and its suspended nest. The nest has no outer cover
like those of more advanced wasps.

19. A hornets' nest. Contrary
to proverbial lore, hornets
are not especially dangerous;
but they are larger than
wasps, and correspondingly
more alarming.

20. A hornet worker (Vespa
crabro).

/

21. Another example of Polistes wasps and their nest, taken by the author.

22. Nests of Polybia rejecta, hanging from a mango tree.

rm0y&'-

23. Close view of nest of
Polybia rejecta.

24. Nest of another species of
Polybia (P. occidentalis).

VARIETIES OF WASP SOCIETY

"queen-eggs " can be laid. Later he kept a number of nests
warm by day, but exposed either the grubs, or the adult wasps,
or both, to hot (25 ° C) or cold (50 C) conditions at night.
Thus there were four experimental conditions : grubs hot,
wasps hot ; grubs hot, wasps cold ; grubs cold, wasps hot ;
grubs cold, wasps cold. He found that when the wasps were
kept cold at night, the grubs turned into workers, but when
they were kept warm, into queens. The temperature at which
the grubs themselves were kept was immaterial. This suggests
that temperature in some way alters the behaviour or physio-
logy of the workers so that they give the grubs different food.
Nobody has yet detected any special queen-food in wasps,
but a change in the type of secretions produced by the
workers and mixed with the ordinary food might be very
difficult to demonstrate. It is not yet possible to relate these
experiments in any simple way to what happens under
completely natural conditions.

THE POLYBIA WASPS

Most wasp biology was studied first in Europe and much
more is known about the behaviour of the species of cooler
climates than of those in the tropics. This has led, almost by
accident, to the view, held at least implicitly, that the normal
thing is to have an annual colony, started by a single queen.
The truth is, however, that the majority of wasps, as of most
other kinds of insects, live in the tropics, and it is much more
likely that social life arose there. When it is realised that
social life, at any rate as soon as it makes any real forward
step, involves the sterilisation of a considerable proportion
of potential egg-layers, it seems almost certain that these
organisations first arose in warmer climates. Except in
conditions which make continuous breeding possible, many
of the advantages of a social organisation are lost. The kernel

e 65

THE SOCIAL INSECTS

of the argument is that the efficiency of this type of organi-
sation, measured by its rate of reproduction, increases
rapidly with time. In general, short-lived colonies are less
efficient than ones which last longer. Thus the winter rest
which is imposed by the temperate climates raises a very real
difficulty. It seems likely that this has been overcome in a
few species which have managed to evolve special modifica-
tions. The relatively large size of the queen in Vespula may
be one of these ; so also is their ability to lay up a large store
of internal fat which keeps them alive during hibernation.

The probability that social wasps first arose in the tropics
is one reason why it is interesting to examine the Polybia
wasps, the largest tropical group of social species. Most of
them are found in Central and South America, but a few
also in tropical Africa and Asia. One or two penetrate a little
way into the United States. In 1937, my wife and I made a
special study of them in British Guiana, and much of what is
written below is based on personal observations. The general
picture of their biology was already known, especially from
the work of the von Iherings and of Ducke, but there were a
great many gaps to fill in matters of detail.

With few exceptions these wasps found their colonies by
swarming. The swarms seem to consist of young fertilised
queens accompanied by a large group of workers. Since
males are never found in young nests, it is probable that they
die after fertilising the queens. Usually several egg-laying
queens are found in each nest ; the number varies a good
deal even in a single species and some colonies may have only
one queen. More often there are four to ten, and in some
species a hundred or more. It is hardly ever possible to tell
the workers from the queens at sight without dissecting them
to examine the ovaries. In some species, however, the
abdomen of the older queens becomes swollen after egg-

66

VARIETIES OF WASP SOCIETY

laying ; they can then easily be identified. In other species,
especially in the ones where queens are numerous, most of
the wasps have ovaries more or less developed and it may
be possible to decide which are queens only by seeing which
ones have been fertilised.

Most of these wasps build nests hanging from or attached
to trees, and they are often nests of very beautiful construc-
tion. Sometimes the outer envelope is covered with lichen or
bits of moss so as to make it almost invisible. There is also
wide variation in the plan of the nest. In the commonest
type each new comb is built on the outside of the lowest part
of the envelope and is enclosed by building out from the enve-
lope all round its edge. The exit hole, which is at the bottom
of the envelope, pierces each successive comb. All round their
periphery the combs are fused with the envelope, and the only
way the wasps can get from one to another is through the hole
which was the exit hole at an earlier stage in construction.

According to the species, there are probably two sorts of
colony, short-lived and long-lived. The former break up
into new swarms after a few months' growth, whereas the
others last for several years. It has been claimed that in one
Brazilian species colonies may persist for twenty-five years,
and records of three or four years are certainly authentic.
The difference between these two types of colony depends
on the physiology and in particular on the length of life of
the queens, but it may also be related to different ways of
attaining efficiency in reproduction.

We were very impressed by the dangerous life which these
wasps lead, especially by their constant exposure to attacks
from ants. Not only are they surrounded by a very much
greater variety and number of ants than one ever sees in
Europe, but the hordes of driver ants, which live exclusively
by hunting, are peculiar to the same parts of the world as the

67

THE SOCIAL INSECTS

Polybias. This danger can be met in several ways; perhaps
the best is to have friendly relations with the ants. A few
species have achieved this, in what way we do not yet under-
stand, but they build their nests almost exclusively on trees
harbouring enormous ants' nests. These ants do not molest
them and the wasp colonies are in many cases long-lived.
Several kinds of birds have learnt to build their nests on the
same trees and seem to have the same immunity. Probably
the small, short-lived type of colony is an illustration of
another way of meeting the same difficulty. A small nest
does not represent a large stock of capital and when the ants
make an attack the wasps can desert the nest without irre-
parable loss. The queens, because there are several of them
and none of them has been laying at a very high rate for
long, have lost none of their mobility. They are not nest-
bound like the old queen of the common wasp. With the
aid of the swarm of workers, a new nest can be reconstructed
in twenty-four hours. We witnessed an actual example of
such an escape near the Kaieteur Falls. A few days after we
had found a nest, we noticed one morning that it was
deserted except for a few workers which were biting away
the envelope. Tracking these wasps down, we found a new
nest about ten yards away (Plates 25, 26, facing p. 80). The
cells of the old nest were perfectly clean inside and we assumed
that it had been raided by driver ants during the night.

The advantage of the Polybia method of colony founda-
tion could be seen very clearly. The swarm consisted of
twenty-two queens and 118 workers and in two days they
were able to build a new nest of seven combs with 167 cells,
each one of which contained an egg. At this point, the nest
was raided again by driver ants, the tail of whose retreating
column we witnessed, and the cells were cleaned out once
more. This time we caught all the wasps for dissection.

68

VARIETIES OF WASP SOCIETY

In contrast to the chequered history of this colony is the
group of nests of Polybia rejecta which we found hanging
from a mango-tree. This tree was covered with enormous
ants' nests, probably the separate houses of a single gigantic
ant-city. The ants had built a nest of carton, a material
composed of earth and wood fibre which is formed into
chambers hanging from the branches of the tree. The ants
swarmed everywhere, so that it was extremely unpleasant
to climb on the tree, though the ants, a species of Dolicho-
derus, do not sting, but only bite and squirt poison. On the
same tree were at least eight large nests of the Polybia, some
of which were certainly one or two years old and probably
a good deal older. From the outer branches of the tree were
hanging several nests of the Oriole bird or Cassique which
frequently joins in this association. It was clear that the
Dolichoderus for some reason did not attack the wasps, but
they would certainly defend their tree if any horde of driver
ants tried to invade it. It is thought that the ants are the first
to colonise such trees, and that they provide most of the
mutual protection. The wasps come later, and though their
stings may make the tree unpleasant to some possible enemies,
such as monkeys, they seem to be almost helpless against ants.

The size of the Polybia colonies is as variable as their
appearance. One large group of species make small colonies
like those of Polistes, with only a few dozen cells and often
only half a dozen wasps. It used to be thought that these
colonies were founded by a single queen unaided. But we
found that though this was often true, sometimes a very
young nest would be occupied by two wasps. Since the cells
clearly had never contained a cocoon, neither of the wasps
could have been bred in it, and there must have been two
founders. We also found that some nests had more than one
egg-laying queen, so that the real difference between these

69

THE SOCIAL INSECTS

colonies and the more usual Polybia type was in their small
size rather than in the number of queens or in the absence of
swarming. The normal nest of a Polybia wasp in British
Guiana consists of 300-3,000 cells with 50-1,000 wasps.
One of the moderate-sized nests of Polybia rejecta consisted
of 7,758 cells and contained about 1,400 wasps. The most
populous nest we studied, belonging to another species,
had about 21,600 cells and 7,087 wasps of which 93 were egg-
laying queens. It seems to be a rule that Polybia colonies
are larger in the sub-tropical regions than they are in the
rain-forest near the equator. Some, for instance near Rio
de Janeiro, may be very large, with probably tens of thou-
sands of wasps. Some of the nests are over a yard long and
hang from trees like gigantic vegetable marrows.

Most species of Polybia seem to feed on other insects.
Sometimes one finds a nest which is temporarily packed
tight with winged ants or winged termites. These are caught
on their nuptial flight and are probably stored ; such flights
are not very frequent and the opportunity is taken to catch
a large number. Another group are active hunters for insects
and are found foraging widely all over the forest. These are
bad stingers and defend their nests with great ferocity. They
also bring home pieces of fresh meat or of carrion if they can
find it. Yet other species are flower-visitors and collect nectar
or fruit juices ; probably most Polybias do this to some
extent. A few have become regular collectors of nectar, like
the honey bee. The best known species with this habit has
been more or less domesticated in Mexico. The colonies
often hang in a citrus bush ; they are allowed to grow to a
certain size, and are then cut off or smoked out, and the honey
is drained off. If a stump of the old nest is left in the bush
the wasps will rebuild in the same spot. Apparently the same
kind of wasp in more tropical climates stores very much less

70

VARIETIES OF WASP SOCIETY

honey. The store of honey may well have one of the same
uses as it does in the honey bee, namely that of allowing the
wasps to get through a period (either dry or cold) when food
is scarce.

It was stated earlier that Polybia workers could as a rule
hardly be separated from the queens. A curious fact which
we discovered was that in several species careful measurement
showed that the workers and queens differed somewhat in
their proportions. The queens are a little larger and the base
of the abdomen is often relatively broader. These differences
could not arise after the wasps are adult, so that it seems that,
as in Polistes and the common wasp, the difference between
queens and workers is determined earlier in life. There may
well be some difference in feeding, though there is as yet
no evidence for it. We found that the number of workers
available to look after each grub increased with the age of the
colony. This is a simple consequence of the fact that a worker
lives longer than the time it takes for an egg to turn into an
adult, so that the number of workers relative to grubs gradu-
ally increases. It may be that the very slight differences
between queens and workers are produced by the increased
care which the grubs receive during the last part of the life
of the colony. This would be the simplest way of differentia-
ting queen and worker castes and would be automatic, with-
out any demand for special behaviour.

It seems clear that there must be some balance between
egg-laying and the construction of cells since either the queens
would lay below the rate of which they were capable or else
many cells would be built for which eggs could not be pro-
vided. We found that in Polybia nests on the average about
one-fifth of the cells at any time are empty. It was possible
to show that the rate of egg-production is probably the
limiting factor in the growth of the colony. The age of the

71

THE SOCIAL INSECTS

smaller colonies can be roughly determined by examining
the contents of the combs and by seeing how many cells are
being used for a second time. If the age of the colony is
divided into the number of cells it gives a measure of the
rate of construction — so many cells per day. This rate
depends partly on the number of workers in the nest and
partly on the number of queens and on the total of eggs
they have produced. But of course the number of workers
themselves, so far as they are not members of the original
founder swarm, also depends on the number of eggs which
has been laid. It is, however, possible by a statistical analysis
to show that the number of queens, and therefore the egg-
supply, controls both the growth of the colony and the
eventual number of workers. What is not yet clear is how
a large number of queens together limit their egg production.
It may be that production is controlled automatically, from
the necessity to share some essential food which is limited in
quantity.

WASPS IN GENERAL

It is generally held that the wasps have the least advanced
type of social organisation of the four principal social groups
(wasps, bees, ants, termites). The worker caste is least
sharply differentiated from the queen. The queen is relatively
short-lived except in some of the colonies where many
queens are present. Foraging still mainly takes the primitive
form of hunting for other insects. Such animal food can be
stored only for short periods. Thus only in the very few
kinds of wasp that store honey can the whole colony with-
stand a prolonged period of food scarcity.

The behaviour of wasps is rather uniform amongst the
different species. Almost without exception they construct
hexagonal cells of paper ; this tends to limit the size of the

72

VARIETIES OF WASP SOCIETY

colony, while the collection of wood-fibre involves a great
deal of hard work. Yet in spite of these limitations, the
behaviour of wasps is sufficiently remarkable. Although
there is no advanced type of division of labour, one does
find that groups of workers are simultaneously engaged in a
number of complicated tasks. This is seen especially in nest-
construction, when foraging, envelope-building, and the
making of cells and suspensory pillars may all go on at once.
It is perhaps easier to understand how the single founder
queen may do these things in the appropriate order than
how a large number of workers can share out the work at
one moment.

The behaviour involved in temperature regulation is
particularly noteworthy. When the nest gets too hot, it
appears to be only some of the workers which take the steps
necessary to cool it. In a more typical example of instinctive
behaviour, all individuals exposed to apparently identical
stimuli might be, expected to do the same thing. We do not
yet know whether the workers which go and fan the rest are
those which have themselves temporarily got hotter than the
others, or whether they take up the work for some other
reason. Fetching water to cool the nest, as in Polistes,
raises a further problem, since the wasps are not cooling
themselves but only the nest they sit on.

The workers are, it is true, very little different from the
queens, except in a few species like the common wasp. Even
here the difference is much less than in the honey bee or ant.
It is clear that we do not yet fully understand what turns a
grub into either a queen or a worker. It certainly looks as
if it is mainly due to automatic factors, such as the growth
of the colony, or of external ones, such as temperature. It
is perhaps wise here to make a qualification. There is always
a possibility that when two possible types of development,

73

THE SOCIAL INSECTS

such as those producing a queen or a worker, have been
evolved, the actual trigger which sets off the individual on
one or other path may change. It is conceivable that even
if temperature now controls the development of queen or
worker in some Polistes it was something else which had
this effect originally.

There are really two problems: first what is the evolu-
tionary advantage of having a worker caste? Second, what
mechanism leads to the production of workers at the present
day ? The advantages are fairly certain, namely a more
accurate control of egg-production and the release of energy
from that process for the other needs of the colony. The
mechanism of the development of the two types is much less
certain and need not be the same in every species.

Most species of social wasps live in the tropics, and species
like the common wasp or the European Polistes, though
extensively studied, are really exceptions. It is probable that
the primitive and simpler arrangement is to have a colony
with several queens, with workers very little differentiated,
with no queen-hibernation but with colonies breaking up
into swarms.

The life-cycle of the common wasp, with annual colonies
and hibernating queens, seems to be a modification for a
temperate climate. One very important feature is the
inhibition of the young queen's ovary in the autumn. The
quantity of food which the queen takes before the winter is
mostly laid down as fat to sustain her during hibernation.
In the spring, however, a similar process of feeding leads to
egg-development instead. This looks much more like a
special device for meeting the unfavourable climate than an
original feature of social life. The common wasp seems to
be much more firmly attached to temperate conditions than
is Polistes, and if they are both regarded as invaders from

74

VARIETIES OF WASP SOCIETY

the tropics, the invasion of the common wasp and its allies
must have taken place a much longer time ago.

Finally, the dominance of the queen in a wasp colony is
very marked. In her absence the social organisation tends to
degenerate. Many workers may start laying and the rule of
one egg to the cell may be broken. When the queen is
functioning normally she seems somehow to determine the
rate of cell-construction, and the length of her life seems to
decide how long the colony will last. There are no grounds
for thinking that she " gives orders " to the workers, except
to the limited extent that she may prevent illegitimate ovi-
position, but she seems to provide the central point round
which the whole complicated pattern of behaviour revolves.
Thus, the simple family character of the social group is more
fully preserved in wasps than it is in ants, honey bees or
termites.

75

SOLITARY AND SOCIAL BEES

Few insects are more familiar than the honey bee, one of the
very few insect species which has been domesticated and
which is consequently the subject of a large literature. Never-
theless the vast majority of bees are solitary species and
rarely noticed except by the specialist. There are two hundred
species of bees found in Great Britain, but less than one-
quarter of these are social. In most of them, such as the
Anthophora and the mason bee shown on Plates 27—30 (facing
p. 81), a single female constructs a nest in the soil, in rotten
wood, or in some crevice, quite in the manner of a solitary
wasp. Sometimes, as in the mason bee, many cells may
eventually be constructed side by side. There are good
grounds for thinking that in the remote past the bees evolved
from the same stem as the sand wasps, the beginning of the
divergence being a change in habit from hunting to collecting
vegetable food. Many wasps make considerable use of fruit-
juices or of the nectar of flowers and one curious group of
solitary wasps (Masaridae), found in warmer climates, lives
entirely on pollen and nectar in all stages. The bees made a
similar change and developed, gradually, a series of new
bodily structures which especially fit them for this mode of
life. Many of the hairs, which are usually numerous all over
the body, are branched. This helps the pollen grains to adhere
to them and enables a large load to be carried on each
journey. Often the hairs are grouped on special areas of the

76

SOLITARY AND SOCIAL BEES

legs which may be considerably broadened so as to make a
pollen-carrying device; in other bees most of the pollen is
carried by hairs on the underside of the abdomen. The tongue
is nearly always more or less lengthened and is sometimes
as long as the body. This enables the bee to reach to the
bottom of the nectary of flowers and to extract the sweet
liquid. This liquid, the nectar, is stored in the bee's crop
during the journey back. Here it is partially digested, so
that it becomes honey when stored. Solitary bees put a
mixture of pollen and honey in each cell before laying an
egg in it. Only social species such as humble bees and honey
bees store the two separately.

BEES AND FLOWERS

It is well-known that the visits of bees benefit the flower
by transferring pollen (male cells) from one flower to fer-
tilise the female cells of another flower of the same species.
The nectar acts essentially as a bait to attract the insects,
and the insects pollinate the flower while raiding it. For many
flowers, bees are the most important pollinating agents,
chiefly because of their methodical behaviour and of their
long tongues. As will be described later for clover, a crop
may fail to set seed if the right kind of bee is absent.

An insect looking after a nest has to work harder than one
which is merely feeding itself, like a butterfly or a fly. Thus
bees usually fly in a more or less regular way from one flower
to another, mostly keeping to one species on each journey
and returning to the nest only when they have a full load.
This can be shown sometimes by direct observation of
individual bees; or, again, it is often possible to identify the
pollen of different species of flower and to analyse the pollen
loads brought home by bees, as has been done by V. H.
Chambers. In the honey bee, individual workers have been

77

THE SOCIAL INSECTS

marked with coloured paints, so that their field routine can
be recorded.

From the point of view of the bee, there seem to be two
different ways of collecting food efficiently. One arrange-
ment, seen in some British solitary bees and in a very large
number of species in North America, is to specialise on a
single species of flower. This involves synchronising the
flight period with the flowering time of the plant and has
often led to the development of a special structure, of the
tongue for instance, adapted to deal with the particular
shape of the flower. As in a human population which depends
on a single industry, the arrangement may be extremely
effective as long as conditions do not change. In the long
run, however, it is very risky.

The alternative is to eschew specialisation and to obtain
food from whatever flowers are available at the moment.
Perhaps most solitary bees behave in an intermediate way :
they visit more than one kind of flower but by no means all,
and usually have a favourite one. Social bees, on the other
hand, are very rarely at all particular, and visit almost any
flower which is providing a good flow of nectar or has plenty
of pollen. This is seen in humble bees almost as much as in
the honey bee. A large colony can find enough food near
the nest-site only if it is willing to take what it can find.

SOLITARY ORIGINS OF SOCIAL SPECIES

It is thought that social life in bees developed from the
more usual solitary type of behaviour in much the same way
as it did in wasps. Females must have acquired the habit
of progressive-provisioning, so that they made contact
with their grubs, and as they became longer-lived their life-
span overlapped that of their daughters. Later still some of
these daughters became more concerned with running the

78

SOLITARY AND SOCIAL BEES

nest and less capable of reproduction, so that a worker caste
was produced. The stages of this history are much less well-
known in bees than in wasps, since fewer of the primitive
types of colony have survived. Social bees seem to have
started their evolution a longer time ago and to have pro-
gressed further, and the simpler types of colony have now
largely become extinct.

Nevertheless, one of the more primitive types, the genus
Halictus, does still exist and is represented by a vast number
of species all over the world. Thirty-six are found in Britain
and a much larger number in the United States. For reasons
which will appear in a moment the fact that they are social
was quite unsuspected until about 1923, and many of the
details of their organisation are still little understood. One
of the chief difficulties in their study has been the large number
of species, many of them very alike, especially in the female
sex. Many species will often nest together in one bank of
earth, and in any ordinary locality in southern England one
may expect to find at least sixteen species. Another trouble
is the variety of their behaviour, not only when different
species are compared but within a single species either in
different parts of Europe or in different years.

Some species are really solitary. Thus in Halictus xantho-
pus both males and females may be found flying together in
the home counties during the spring or early summer. Each
female after fertilisation enters winter quarters in a hole in
the ground. Early next spring she excavates a nest and raises
a brood of males and females. The males die soon after
mating, while the young females hibernate. Thus there is
one generation a year and it flies early. In the same species,
in Germany, both males and females may be found flying in
the autumn. Probably here there are two complete genera-
tions in the year, the females of the second one surviving

79

THE SOCIAL INSECTS

the winter to renew the cycle in the following spring. In
the more favourable climate of southern Europe the males
may not only appear in the autumn but also hibernate and
appear again in the spring. By these conditions the same males
and females may fly both in the autumn and in the spring.
Other species, such as the common British Halictus morio,
seem to have more than one generation of males and females
during the summer, but almost invariably only the females
of the last generation survive the winter. In this species, the
seasonal picture is as follows:

Spring Summer Autumn Winter

Female Male and female Male and female Female

DEVELOPMENT OF A WORKER CASTE

Life histories of the type hitherto described are solitary,
but in some other species the spring female may survive
long enough to be found in the nest at the time when the
summer females, her daughters, are emerging. Though the
mother and offspring do not in this case co-operate in nest-
construction, there is here the first hint of social behaviour,
since the presence of the mother in the nest must give her
young some protection throughout their development.
Another group of species, including in Britain Halictus
calceatus and several very similar ones, have a real social
organisation. Males and females are found in the late summer
and autumn, but only the fertilised females hibernate to be-
come the spring females of the following year. The spring
females, either alone or working in small groups (up to three
females), construct a nest and produce offspring which are
all female. This brood of females may or may not be extern-
ally different from the spring-mother, but in any case they
are unfertilised ; their ovaries develop only to a limited

80

25. Nest of the wasp Parachartergus when first found at Kaieteur, British Guiana.

26. The same colony rebuilding the envelope and combs on a new site after
the first nest had been raided by driver ants.

M t

.

^"*?. k '

27. A solitary bee (Anthophora
acervorum) on a flower.

29. Osmia rufa. Larvae feeding
on pollen masses in their mud
cells.

28. The solitary mason bee
(Osmia rufa), male above,
female below.

30. Osmia rufa. View of a
many-celled nest built behind a
door lock.

SOLITARY AND SOCIAL BEES

extent, and they are essentially workers. These workers now
do nearly all the work of the summer-nest, and the old spring
female becomes a queen who guards the nest and lays most
of the eggs. The result of this activity is the production of
another autumn brood of males and females. It seems from
the observations of Stockhert and Noll that the autumn
females all come from eggs laid by the queen, whereas the
males come from unfertilised eggs laid either by the queen
or by some of the workers. Thus throughout the whole
seasonal cycle females are always derived from fertilised eggs
which are all laid by the original spring female.

It seems, at any rate in the Halictus studied by Stockhert,
that where two or three spring females combine to make one
nest, only one of them survives to act as queen when the
workers appear. Thus the situation seems very like that
already described for the wasp Polistes gallicus in north Italy,
where colonies may be founded by several queens, but in
which one queen becomes the dominant egg-layer and is
the only one who remains in the nest after the workers have
hatched. Apparently in some species, such as Halictus
maculatus, rare in Britain, several spring females may con-
struct separate nests with a common entrance-gallery. When
such a nest is fully developed, with each separate part contain-
ing its own brood of workers, the single entrance is as busy
as the entrance of the much larger colonies of a humble
bee.

The colonies of Halictus show a type of social life which
is primitive because they are always small (probably six to
ten workers) while the workers are little different from the
queens and still lay some of the eggs. Probably the growth
of the colony is much restricted because they have retained
the normal subterranean habits of solitary species. Each
separate cell has to be excavated, and though a group of cells,

F 81

THE SOCIAL INSECTS

looking like a small comb, may be dug up by the entomolo-
gist, the comb has not been built up, cell by cell, as in a
wasps' nest. Rather the superfluous earth has been removed,
and a group of six to twelve cylindrical cells has been isolated
in a small underground chamber. The earth is strengthened
by a coating of saliva. In Halictus the cells are composed of
that earth which the bees have not carried out of the nest,
whereas in a social wasp the cells are built of new material
which has been brought into the underground chamber. The
nearest human analogy is seen when a dug-out is compared
with a house.

The great number of species of Halictus, many of which
are extremely similar to one another, suggest that the group
has evolved recently and is probably still evolving. It cannot
therefore be regarded as ancestral to the other social bees,
since these have a long history of social life. But Halictus
may be regarded as illustrating how social organisations in
bees may have originated. In them, the process has not yet
evolved very far.

Rudimentary social behaviour approaching that of Halictus
occurs in some other bees of groups which are mostly solitary.
Such are the South African species of Allodope which build
nests in hollow plant stems. There are no proper cells and
the brood lies in a rather disorderly mass with the different
stages mixed together. The colony is begun by a single
female, but later her daughters help in obtaining food for
the brood. So much was established by Dr. H. Brauns, for
long a successful observer of African insects. Clearly this
example needs much more detailed study before it is fully
understood.

82

SOLITARY AND SOCIAL BEES

HUMBLE BEES

The next social group, the humble bees, are much more
familiar and have been much more studied. Apart from the
cuckoo humble bees, social parasites which will be mentioned
in chapter 5, nineteen species have been seen in Great
Britain and thirteen of these are more or less common in the
southern part of the country. The group, however, is a very-
large one, with two hundred to three hundred species, found
all over Eurasia to Japan and Sumatra ; all over North and
South America and in Africa north of the Sahara. Thus, while
it is especially well represented in temperate regions, it is also
found in much of the tropics. The solitary bees most like
humble bees are now found in South America, and it may
have been here that the group originally developed.

Some people imagine that a humble bee does not sting like
the honey bee. This is unfortunately untrue, but it will very
rarely sting unless roughly handled or unless the nest is
disturbed. As in the wasps, the larger the nest the more
fiercely is it defended. In temperate climates the life-cycle is
very like that of the wasp. Males and females are produced in
the summer or autumn, but only the fertilised females survive
the winter. In the spring the queens feed on flowers of sallow
and other early flowering plants, and then begin searching for
a nesting-site. Some species nest under grass-tussocks on
the surface of the ground, others in holes underground. In
both types most nests are constructed in an abandoned mouse
or vole nest. The collection of pieces of grass and dead leaves
made by the mouse provides the essential protection and
insulation for the early stages. More grass or moss may
be piled up later.

Suitable sites of this type are often not very numerous,
and this has two very important consequences. First, there

83

THE SOCIAL INSECTS

may be severe competition for nests, and one or even more
dead queens may be found at the entrance to the nest, killed
by the rightful owner in defence of her property. Secondly,
there is a limit to the number of nests which can exist on an
area, and this may be very important in the production of
such crops as red clover-seed, since humble bees are the chief
agents in fertilising the flower. Darwin, in fact, in a well
known passage in The Origin of Species, records that red
clover sets very little seed if the flowers are covered and made
inaccessible to bees. He also quotes Newman as believing
that mice destroy many humble-bee nests and that near
villages and small towns humble bees tend to become more
numerous because cats destroy many of the mice. Although
this observation needs to be better substantiated, it suggests
that there may be a complex, double relationship between
mice and bees, the mice being not only enemies but also
providers of nesting-places.

The queen, after starting her nest, has to raise her first
brood of workers unaided, but once the workers have hatched
they do most of the work, especially all the foraging for
nectar and pollen. The queen gradually drops all work except
the laying of eggs, and eventually becomes incapable of flight.
In fact, she never leaves the nest again once the colony is
fully established. Sometimes, in bad years, very worn queens
are seen flying in midsummer, and these are trying to raise
a new brood after the first one has failed to develop. If the
nest goes well several groups of workers are raised, and
eventually males and young queens. The cycle is then com-
plete and the colony soon breaks up. This may happen
surprisingly early, even by late June in some species in a
good year. The young June females enter winter-quarters
at the height of summer. Apparently there is no time for them
to start and to bring safely to completion a new colony in

SOLITARY AND SOCIAL BEES

the same season. Bols showed that in Belgium certain banks
may be much favoured by the young queens, several dozen
digging holes near one another. These hibernation-holes are
of about the diameter of a cigar and three inches long. Not
all queens use such sites, and some may be found hiding in
moss or dead leaves in woods or hedgerows.

Apart from the dead vegetable matter which insulates
the nest and helps to protect it from the damp, the cells are
constructed of wax, a very characteristic product of social
bees. In the humble bee it appears as thin plates protruding
between the overlapping segments of the abdomen both above
and beneath. The wax plates are removed by a stiff brush on
the hind legs, and can be worked up by the jaws. All insects
produce at least minute quantities of a wax which serves to
make their outer skeleton waterproof. Several kinds, such
as scale insects or woolly aphids, produce much wax which
seems to serve as a partial protection from their enemies.
Only social beesy however, " handle " the wax which they
produce.

The humble bee does not construct regular combs of
hexagonal cells like the wasp or honey bee, and the nest
grows rather piecemeal. This makes its arrangement difficult
to understand without careful study. The first step of the
queen is to make a small cavity in the centre of the ball of
grass, and here she lies for considerable periods while wax is
being secreted. As in the honey bee, the production of wax
entails the eating of a good deal of nectar. The first wax is
used to make a small honey-pot which stands near the entrance
to the nest. This when filled provides a food supply during
bad weather. The next step is to make a small ball of pollen,
moistened with honey. On top of this a small waxen egg-
cell is built. In some species the egg-cell is built first and the
pollen collected afterwards. The first cell contains six to ten

85

THE SOCIAL INSECTS

eggs ; from these the grubs hatch in three or four days.
Each grub after a time produces a swelling on the sides of the
waxen cell, so that the cell is gradually enlarged. In some
species the queen opens up the cell at intervals and uses her
tongue to inject into it a mixture of honey and pollen. In
others the pollen is added in wax-covered lumps at the sides
of the original cell. In this way the queen has less contact
with her grubs and her behaviour is a modified form of mass-
provisioning, whereas the injection of liquid food is a
characteristic type of progressive provisioning. In older
nests a good deal of wax may be worked into the grass above
the cells, to form a waterproof roof.

The grubs are fully fed and begin to spin their cocoons
seven to eight days after they have hatched. The cocoons
are oval, made of yellowish silk, and in most species stand
in a close-packed group side by side. In fact the group of
cocoons looks not unlike an irregular comb, though made of
silk and not of wax. In the first brood the cocoons on each
side are placed higher, so that the central groove in which
the queen sits is preserved. Here she spends much of her
time keeping them warm. The first workers usually hatch
rather more than three weeks after the egg was laid, and the
queen or, later in the season, other workers often help them
to bite their way out of the cocoon. Those from the middle
of the groove develop more quickly than those from the sides.

The new-hatched worker is pale-coloured with somewhat
matted hair. Her first activity is to take some honey from the
honey-pot. However, in two or three days, she is ready to
leave the nest and to begin to collect honey and pollen.
Provided enough workers have been raised in the first brood
the queen will now stay in the nest until she dies, but some-
times, when the first lot of workers is too few, she will con-
tinue foraging a little longer.

86

SOLITARY AND SOCIAL BEES

Later batches of eggs are laid in little pockets attached
to the side of the cocoons. Mr. and Mrs. M. V. Brian,
working at Glasgow, have recently shown that there is a
tendency for the number of eggs in the group to be propor-
tional to the number of cocoons in the clump on which they
are laid. Taking the average of cocoon groups of different
sizes, there were about three eggs for each cocoon, with
six to sixteen eggs in a batch. This seems to be a form of
family planning which ensures that there shall be no more
brood to look after than there will be nurses to care for
it.

At this stage, the brood soon becomes too large to be
incubated by the queen, and the groups of larvae or cocoons
no longer form a groove but are convex on top. The develop-
ing brood produces enough heat from the digestion of its
rich food to keep the whole nest at a sufficiently high tem-
perature.

At the beginning of the eighteenth century, a Dutch
naturalist, Goedart, first propagated the curious myth of the
" trumpeter " humble bee. He observed that a bee often
rests on the top of the nest, fanning with its wings and making
a loud hum. This often occurred in the early morning when
the sun's rays first strike the nest, and Goedart thought that
the hum of the vibrating wings was a call to start the morn-
ing's work. During the last two hundred and fifty years
naturalists have argued whether the " trumpeter " really
existed, and if so what were his functions. The Austrian
entomologist, Hoffer, finally showed that trumpeting really
occurred by observing a colony in a nest-box on his window-
sill. Eventually, Plath in America showed that it happened
especially when the nest was exposed to the sun, and that
this could be checked by moving the nesting-box to different
positions. Fanning sometimes occurs also in the evenings

87

THE SOCIAL INSECTS

and may then serve to dry the nest or to remove odours.
Nests in natural situations will less often be exposed to the
sun than when they are kept above ground in nesting-boxes,
so that trumpeting seems to be observed much more fre-
quently in conditions of semi-domestication. It is behaviour
of the same kind as that observed in wasps when their nest
gets too hot.

At the height of its activity the colony is able to store
considerable quantities of food, especially in some species
and especially those nesting below ground in which
workers become very numerous. Such nests may have three
or four hundred workers, but one hundred and fifty or less
is more common. Nectar, later turning to honey, is stored
chiefly in empty cocoons which are not used again for brood
rearing as are the cells of a wasp-comb. Since the egg-pockets
are built near the top of the sides of the cocoons, the nest
grows upwards and the empty or honey-containing cocoons
are mainly in the lower part of the nest. Sometimes some
waxen honey-pots are also built round the edge of the nest,
but according to Plath these are temporary structures. The
food stored in them is for daily use and they are often broken
down and the wax used elsewhere. Some of the species also
store considerable quantities of pollen, either in empty
cocoons or in large waxen cylinders which may eventually
reach the length of more than two inches. The surface
nesting " carder bees " store pollen in little wax pockets at
the sides of the groups of grubs ; the quantity stored in this
way is quite small. Since the colony does not survive the
winter there is no point in making a large store of food as
the honey bee does. But, as explained later, the store of honey,
at any rate, has a very important function in preparing the
young queens for their winter rest.

The climax of the colony comes when the males and young

88

SOLITARY AND SOCIAL BEES

queens are produced. It is not known what exactly deter-
mines the date when this shall begin. It is not simply that
the vigorous, successful colony shall have passed a certain
size, for relatively small colonies may produce one or two
queens at about the same date as a larger colony of the same
species produces a much bigger number. It may be that there
is some change in the queen's rate of egg-laying, so that the
number of nurses tends to pile up compared with the number
of grubs to be looked after. Among the brood which hitherto
has produced nothing but workers a certain number of
males appear and a little later queens. After this no more
workers are produced. This sequence, in which males
tend to precede the queens but partly develop among
them, is exactly what is recorded in the common wasp.
The young queens work only in colonies in which the
workers are for some reason scarce. Normally, they stay in
the nest for a time to feed, and then, after pairing, enter
winter quarters. ,

At this season of the year, one may notice several bees
taking the same course in a garden or wood, though there
may be a few minutes between one flier and the next. At
intervals, they will fly down and hover for a moment in such
places as the hollow at the roots of a tree. These are male
humble bees on their characteristic circuit, usually a different
one for each species. Humble bees have a strong, often honey-
like smell in the male and a less obvious, often rather un-
pleasant smell in the queen. It has been suggested that the
males emit scent at the stopping places on their circuit and
that these are then attractive to. the young queens; the
suggestion has, however, not been studied experimentally.
But R. A. Cumber found that young queens could be
suspended by a piece of cotton in bushes in the circuit, and
that males would then soon come and pair with them. In

89

THE SOCIAL INSECTS

some species, however, the queens have been known to pair
with a brother, in or near the nest.

In most species queen and worker humble bees differ very-
little from one another except in size. In the larger, subter-
ranean species the gap in size between the two castes is
appreciable and there is almost no overlap. In the carder
bees, however, the two types almost run into one another,
probably because this group of humble bees has not evolved
so far as have the others.

A study of the differences between workers and queens
has been made by the author and by R. A. Cumber. A
picture of this type of investigation will be obtained if a nest
is described which was excavated at Silwood Park, Berks, on
July 4, 195 1, by Dr. N. Waloff and myself. It was a nest of
Bombus hortorum, about six inches underground, in an old
mouse-run in a hay field. The bees were divided into three
lots : first, those that were out foraging and returned to the
nest while it was being excavated, mostly carrying pollen
on their hind legs ; second, those that came out to defend the
nest while it was being dug out ; third, those that stayed on
the comb and were removed from it, one by one, in the
laboratory. The weights of these three lots of bees and their
status in the nest are shown in the table opposite. Although
there is some overlap, the average difference in weight of the
three lots of workers is very striking. The old queen
(weight 553) was recognisable by her rather battered appear-
ance. Only one of the other queens (weight 442) had been
fertilised, as was shown later by dissection. The colony was
a rather small one and it may be for this reason that one of the
young queens was carrying a load of pollen. Besides these
bees, the nest contained six eggs, twenty-four grubs, sixty-
six full cocoons, fifty-one vacated cocoons containing honey,
and one cell containing pollen. There were also twenty-six

90

SOLITARY AND SOCIAL BEES

empty cocoons not in use. Judging by their size, about thirty
of the cocoons might later have produced males or queens.
Though there were only forty-three bees in the nest there
were seventy-seven vacated cocoons. Most of the difference
can be accounted for by the deaths amongst the first broods
of workers. This was a moderately successful colony which
would probably have finished its effective working life two
or three weeks later. Most of the honey would probably
have been used in feeding the queens.

Weights of bees in nest of Bombus hortorum, excavated on
July 4, 195 1. Weights are given in milligrammes (1 mg.= about thirty-
five millionths of an ounce).

LOT I FORAGERS

LOT II

LOT III

REMOVED

RETURNING TO NEST

WORKERS

FROM COMB

DEFENDING

THE NEST

Weight of

Weight of

Nature

Workers'

Queens' Males'

bee

pollen load

of bee

Weights ■

weights

weights weights

Young

395

25

Queen

317

202

553

316

332

30

Worker

218

162

493

262

301

45

»»

217

149

465

215

278

62

5J

164

148

462

268

12

»

155

140

442

264

4i

}>

150

123

440

260

0

Male

144

122

432

220

30

Worker

139

102

398

155

40

»

136
115

96

87
86

381
368

Average

260 mg

176 mg.

129 mg.

weight of
workers

It seems from this record and from many other similar
ones that the variation in size of the workers has some

91

THE SOCIAL INSECTS

importance in the economy of the nest. The variation can
be referred back still earlier to what happens to the grubs.
Cumber, by weighing grubs or pupae extracted from different
positions in the groups, showed that those in the middle are
heavier and those from the periphery lighter. In fact, as
Sladen first suggested and Cumber later confirmed, the
variation in the size of the workers depends mainly on the
competition of the grubs in one pocket for their common
supply of food. Grubs which happen to be near the middle
of the pocket grow large, while those at each end are smaller.

The variation in size of worker so produced is the basis of
the later economy of the nest. It seems that once workers
are sufficiently numerous, a division of labour is established
between them based on the size-difference which itself
depends on the earlier competition for food. Weighing
samples of workers throughout the season shows that there
is no increase in the average weight of the whole set of
workers, but when the colony is at its largest, the large bees
do most of the foraging; the small ones mostly stay at home
to act as nurses. According to very recent work (1952) of
Mrs. Brian, small workers do not begin to forage till 15
days after leaving the cocoon, whereas large ones begin at
5 days. Small bees forage mainly for nectar while large bees
also collect pollen. Cumber found a further minor division
of labour between the foragers, for by measuring the length
of the tongue of humble bees caught on different flowers he
showed that the largest workers of the species, with the
longest tongues, went to flowers with long nectaries, whereas
the smaller, shorter-tongued foragers went to flowers in which
the nectar was more readily accessible.

In a humble bee such as Bombus agrorum there is a com-
plete overlap in size between the young queens and the largest
workers. There is, however, an essential difference in

92

SOLITARY AND SOCIAL BEES

physiology. Cumber found that young queens can be
confined in small cages and artificially fed for a week on
honey. When this is done they put on weight very rapidly,
laying down a large amount of white reserve tissues in their
abdomen. Their ovaries, on the other hand, show not the
slightest sign of development. Large workers when similarly
treated have no power to increase appreciably in weight, and
if they are dissected the reserve tissue is found to be dis-
coloured brownish as it is in the old queen.

At least some of the large workers usually show partial
development of their ovaries. If there is anything wrong
with the queen, many of them lay eggs, though these prob-
ably produce males only. We may conclude, therefore, that
there is a fundamental difference between the young queens
and the workers, caused in some way not yet understood,
such that queens can prepare themselves for hibernation
whereas workers cannot. There is no evidence that workers
ever hibernate in the climate of Western Europe or of North
America, and no evidence that they are capable of pairing.
These statements have, however, to be rather guarded owing
to the difficulty of separating the two castes by mere inspection.

As noted earlier, bees play an important part in the
pollination of flowers. This is because they are numerous,
are hard and methodical workers, and are relatively large
insects, with long tongues. Flowers in which the nectar
collects at the bottom of a deep nectary are nearly always
fertilised by bees, more rarely by large moths. In temperate
climates the bees with longest tongues are some of the humble
bees, and it is known that these insects are essential for the
setting of seed by such flowers as red clover or monkshood.
Both in the Alps and in Norway a special species of humble
bee is found which visits monkshood almost exclusively.
Without this bee, the flower would hardly ever set any seed.

93

THE SOCIAL INSECTS

However, the red clover is a much more important flower
to man and has been much more studied. Clovers make a
rich fodder-crop and their roots bear nodules in which
bacteria multiply and convert nitrogen into nitrates. This
means that besides producing good food for cattle, the clover
also enriches the soil it grows in. Some clovers, such as the
common white species, have relatively short nectaries and
can easily be fertilised by the honey bee. The red clover,
however, which for some purposes is particularly valuable,
has a larger flower and is rarely fertilised by anything other
than one of the long-tongued humble bees. This creates a
difficulty whenever large fields of clover are grown for seed,
since there are rarely enough of the right kind of bee to
fertilise all the flowers.

The problem has been studied in Britain at Aberystwyth,
in Scandinavia, in Russia, and elsewhere. It was proposed at
one time in Russia that wild humble bees should be
encouraged by the provision of large numbers of nesting-
boxes. For the shortage of bees is partly due to the absence
of sufficient nesting-sites, and this may in turn depend on a
shortage of mice. For other reasons, too many field mice
would be undesirable, so that it was thought that nesting-
boxes might be a substitute for abandoned mouse-nests.

Probably the only practical measure is to grow the clover
in small patches rather than in the large continuous areas
which are much more convenient to the farmer. The prob-
lem arose in its most acute form in New Zealand, where
there are no native humble bees. For many years the red
clover set no seed, but in about 1880 bees were introduced
from England. Unfortunately, the introduction was not
carefully planned and besides one useful, long-tongued bee,
a short-tongued species was brought over at the same time.
Some of these short-tongued humble bees have the habit of

94

SOLITARY AND SOCIAL BEES

biting a hole near the bottom of the nectary, so that they can
rob it without coming in contact with pollen and fertilising
the flower. Once the hole has been made long-tongued
species may use it, too, so that much harm may be done.
However, for many years New Zealand was able to obtain
all the clover seed that was needed, and it is only recently
that difficulties have again arisen. It is not clear whether for
some reason the long-tongued bee has become less numerous
or, as I think is more likely, that the area under clover has
been increased without any corresponding increase in the
possible nesting-sites.

The humble bees which live in tropical countries have
not been much studied, but it seems that some of them, as
in the tropical social wasps, live in colonies with several
queens and found new nests by means of a swarm of workers
and queens. This has been recorded by von Ihering in
Brazil, but further accounts of the habits of humble bees in
the equatorial region would be very interesting. Foundation
by swarms is probably developed because in that climate
hibernation is unnecessary ; hence any surplus workers can
accompany the young queens when they leave the colony.

Observations by Plath in the United States support the
idea that colonies might become perennial, the old queen
being replaced by a new one when she was worn out. In one
American species he observed this to take place in two
separate colonies at the end of August. In one of them one
large crop of new workers was produced in September and
a second crop of males and queens in October. This is
exceptional behaviour in a temperate climate, and might be
possible only in a favourable year.

In New Zealand Cumber found that one of the introduced
British species behaved on one occasion in a rather different
way. He dug up a large nest on November 4 (which is early

95

THE SOCIAL INSECTS

in the New Zealand spring) and found that it was a survival
from the previous season. The old queen was alive, but was
no longer capable of laying eggs, and this function had been
taken over by about twenty of the workers. The next also
contained more than forty young queens, of which about half
had been fertilised, and twenty-one males. In none of the
young queens was there any ovarial development, so that
they were really only hibernating in the nest. None of them
had taken over the functions of the old queen as in Plath's
observation. It is an explanation of facts of this sort in terms
of physiology which we most need to advance our know-
ledge of social behaviour. For two of the most important
differences between solitary and social species are that first,
in the latter, young females stay by their mother, and second,
that the development of the ovaries is subject to more
elaborate control.

THE STINGLESS BEE

A bee without a sting seems almost a contradiction in
terms, but in one large group of social bees, the Meliponidae,
the sting is reduced to a vestige and is useless for defence or
attack. The insects are, however, by no means helpless, since
they can bite and can smear their enemies with sticky sub-
stances, so that populous colonies are as much feared as
those of wasps. Stingless bees are found only in the tropics
and sub-tropics; they are most abundant in South and Central
America, but occur also in Africa, Asia, in some of the Pacific
islands and in Australia. The species are very numerous :
about two hundred are clearly recognised, while recent work
in Brazil suggests that there may be many more. The sting-
less bees fall into two groups. Melipona itself is confined
to South America and its species are rather large, often little
smaller than the honey bee. Trigona, which is found in the

96

■vW1

31. Top to bottom, male,
worker and queen of the carder
humble bee (Bombus agrorum).

33. Humble bee (Bombus
lucorum) about to take nectar
from lupin.

32. Nest of Bombus agrorum,
found in a disused mouse nest
on the ground.

34. Young larvae of a humble
bee (Bombus terrestris).

m^

e

35- Drone honey bee (Apis mellifera). Note the large eyes.

36. A fertile female, or queen honey bee, with large abdomen.

37. An infertile female, or worker
honey bee.

SOLITARY AND SOCIAL BEES

Old World as well as in America, has many more species,
and varies from a similar size down to some minute bees no
larger than house-flies.

Unlike the humble bees, the workers produce wax only
from the upper side of the abdomen. Some early workers
thought that the males also secreted it, but this has not been
observed by any modern naturalist. Besides their own wax,
they also collect plant waxes and juices and earth, and these
substances are often mixed with the wax for some purposes.
The most usual site for a nest is a hollow tree, but some
species nest in the open, in the forking of two branches, or
in crevices of stonework, or in holes in the ground. A few
build in deserted ant-mounds or in holes in large termite-
nests. Within such a large group some variety of habit is
to be expected, but even within one species more than one
type of site may be used. Most of the species construct a
spout-like structure at the entrance. This may project for
several inches from the hole leading into a hollow tree. It
is sometimes closed at night by a temporary wax membrane.
In a typical nest, the hollow branch is walled up at each end
by a barrier of wax, mixed with earth. The cavity so formed
is divided into two, one part for the brood, the other for a
store of honey and pollen. The part containing the brood is
surrounded by a number of irregular waxen envelopes which
doubtless serve, as in the common wasp, to conserve heat.
The brood is reared in waxen cells arranged in one of three
ways. The commonest is to have a series of horizontal
combs, one above the other. The lowest comb is built first,
the later ones being supported one above the other on little
pillars. The cells all open upwards, being constructed on
the upper side of each comb. Occasional cells are missing
so that there are gaps in the comb, allowing for the passage
of the bees and perhaps for ventilation. In large colonies,

g 97

THE SOCIAL INSECTS

several tiers of combs of this type may exist side by side.
In another type of nest the combs are arranged like a rather
irregular spiral staircase, so that though the combs rise
upwards, they are not close-packed one above the other.
Finally, in some species of Trigona no regular comb is
built at all, but the cells are arranged in a cluster or conical
heap. Nests of this type also lack the waxen envelopes sur-
rounding the brood. It has been suggested that bees which
made this imperfect type of nest are less highly evolved than
the others and have never acquired the habit of building
regular combs. This, like many other things about these
bees, needs much more study.

Apart from the brood, sometimes on one side of it or
sometimes on both, is the store of honey and pollen, but this
is not kept in cells similar to brood cells as it is in the honey
bee. The storage chamber contains globular or cylindrical
pots of honey and pollen. In a large colony there may be a
hundred or more of these, each one much larger than a
brood-cell. There may also be pots of resin, or plant-materials
may be kept in irregular lumps.

The Maya Indians domesticated these bees for their honey
long before the Spanish conquest of Mexico. Sections of
hollow logs were used as hives, as has also been the custom
in some parts of Europe for the honey bee. The stingless
bees were introduced into the log by various methods.
Sometimes natural wild swarms colonised them. Sometimes
the Indians put into the log part of the food store together
with some workers and the queen from an older nest. Most
commonly, the brood combs of an older nest were divided.
Apiaries of a hundred or more logs were kept, and such
property was an important part of an Indian's capital and
could be the largest item in his will. The Mexicans usually
removed honey from their hives twice a year, obtaining

98

SOLITARY AND SOCIAL BEES

about three and a half pints of liquid honey from each. This
honey does not seem to thicken and crystallise out even with
keeping. Some species or colonies may produce a great deal
more : an Australian nest, for instance, has yielded 50 lb.

The castes of the stingless bees are much more strongly
differentiated than those of the social species described so
far. The worker is the only type structurally adapted for
running the colony. For carrying pollen her legs are modified
in much the same way as those of humble bees. Her mandibles
often bear serrations or teeth which are thought to be of use
in manipulating wax. The queen, on the other hand, lacks
the pollen-collecting apparatus of the worker, has simpler
mandibles and a smaller head with somewhat smaller eyes.
She does not produce wax. She would be quite unable to
found a new colony unaided, and she must therefore be
accompanied by a swarm of workers. In Melipona the queen
is no larger than the workers, but in most species of Trigona
she is considerably larger; in both groups her abdomen soon
becomes very swollen by her enlarged ovaries. An older
queen is therefore easily distinguished and is too heavy for
flight. In fact she never leaves the nest once it is established.

Many of the details of swarming and of the foundation
of new colonies are still obscure. Pairing has not been
accurately described, but there is probably a nuptial flight
something like that of the honey bee. Probably the founding
swarms consist of a single queen and many workers. The
laying queen does however tolerate some young queens, and
a number of them may be found living in the colony. It is
doubtful, however, if there is more than one laying queen
at any one time; the others may be merely waiting to take
part in swarms.

The way in which the castes come to be differentiated is
still a matter of dispute. Many of the species of Trigona

99

THE SOCIAL INSECTS

build " royal " cells, usually on the edges of the ordinary
comb. These cells are considerably larger than the ordinary
brood-cells and produce queens which are larger than the
workers. In Melipona and in some kinds of Trigona no
special royal cells are recognisable and the queen is no larger
than the worker. It is also said that she has to feed up for
some time before her ovaries can become active.

The stingless bees are almost the only group of social
insect in which there is no contact between the adult and the
grubs. The cells are filled with a food, mainly honey and
pollen, and when the egg has been laid the cell is sealed with
wax. There is thus no opportunity for the type of differential
feeding which is well-known to occur in the honey bee.
Moreover, it is stated that several workers may contribute
to the storing of each cell, making it much more difficult to
suppose that any accurately balanced diet can be provided.
While enlarged royal cells would provide the queen grub
inside them with a greater bulk of food, this would not
apply to Melipona, since the queens and workers develop
in similar cells.

THE THEORY OF HEREDITARY CASTES

Because of these difficulties, Kerr has recently suggested
that the characters of the castes are inherited and are not the
result of special feeding. When he bred out the brood from
combs removed from nests he found that in three species of
Melipona the queens formed one-eighth of all the progeny,
while in a fourth species they formed one-quarter. These
fractions correspond to what are known in the Mendelian
scheme of genetics as a " trifactorial " and a " bifactorial "
backcross. The meaning of this description can be explained
most simply in the species which produces one queen for
every three workers, that is, one-quarter queens.

IOO

SOLITARY AND SOCIAL BEES

The hereditary factors which are supposed to control the
type of development of the individual are represented by
letters, the queen having the constitution Aa Bb. During the
development of her eggs in the ovary, there is a special cell
division (reduction division) in which each group of genetical
units (chromosomes) is halved. Thus the Aa Bb queen will
produce four types of unfertilised eggs with constitutions
AB, Ab, aB, and ab. The males are, as in all bees and wasps,
derived from unfertilised eggs, and so have only the half
complement of hereditary factors to start with. They omit
the reduction division and their sperm are, therefore, of the
same four types as the unfertilised eggs. When fertilisation
occurs, there will be an equal likelihood for any of the four
types of eggs and sperm to meet. Only four out of sixteen
possible combinations will give an individual constituted
like the queen. Thus sperm AB + egg ab, or sperm Ab +
egg aB, would both give a queen. We should therefore
expect, on the average, one-quarter of the offspring to be
queens and three-quarters workers. Males will be derived
from any eggs which are not fertilised.

The case where one-eighth of the females are queens would
arise if three factors were involved and the queen had a
constitution Aa Bb Cc. While such schemes can theoreti-
cally explain the facts, this is very far from saying that they
have been proved to be true. Kerr supposes that some
similar arrangement may also exist in Trigona and in the
honey bee, but that in them the queen grub has to be given
the right quantity or quality of food as well as inheriting
the right constitution. In Melipona,-if queens are produced at
a constant rate throughout the life of the colony they would
soon become too numerous, and Kerr finds that most of
them are killed by the workers.

If Kerr's ideas are substantiated it would seem that

IOI

THE SOCIAL INSECTS

Melipona has a rather wasteful and inefficient method of
providing the necessary proportion of new queens. In most
groups of social insects the argument whether castes are
dependent on hereditary constitution or are due to differ-
ential feeding has continued for the last fifty years. In most
groups, the weight of the evidence is in favour of control by
diet. It will be rather startling if the stingless bees turn out
to be an exception, or if our ideas on all the other groups
turn out to be wrong.

1 02

HONEY BEES

The last group of social bees is the most familiar, including
the honey bee and two or three similar species. These have
been domesticated by man from at least the time of the early
Egyptians, and an immense body of tradition and literature
has grown up around them.

It will be convenient to say something first about the
non-European relatives of the honey bee. In India and
elsewhere in the East there are two rather distinct species of
wild bee, Apis dorsata and Apis florea. Our familiar honey
bee is Apis mellifera and itself occurs in a variety of strains
or subspecies. The one which occurs wild and has also been
domesticated in India is Apis indica, while in Africa a variety,
Apis adamsoni, is also often found wild. Apis mellifera must
have been introduced into Europe very early in man's his-
tory; it has become markedly different from Apis indica,
though the whole group probably has an eastern origin. The
honey bee is to-day comsopolitan, but this is because it has
been carried everywhere by man during the last few hundred
years.

WILD HONEY BEES

Apis dorsata is considerably larger than the honey bee.
Its colonies build a single comb hanging from the branch of
a tree. This comb hangs vertically and has cells pointing
each way, their long axes being horizontal. All the cells,

103

THE SOCIAL INSECTS

including those in which the queens develop, are of the same
size. Some of them are used for storing honey, but the
colonies are relatively small and do not stay for long in any
one site. Apis florea is the smallest species, and it too builds
a single comb hanging from a branch. The cells are of four
sizes. Near the top are a number of honey cells and below
them a band of smaller cells for worker brood. Below this
are considerably larger cells in which drones are reared. All
these cells are hexagonal, but hanging from the bottom of
the comb are some irregularly cylindrical queen cells which
are the largest of all.

Apis mellifera in all its forms has a marked tendency to
nest in dark cavities, such as caves or hollow trees. Less
commonly its combs may be suspended from branches, and
this has been recorded occasionally in Europe when a swarm
has escaped into the woods. The honey bee builds several
combs hanging side by side, each somewhat resembling the
single comb of the other two species. The worker and honey
cells are of the same size, but the drone cells are a little larger.
They usually lie on the periphery or at the bottom of a comb,
while the honey cells are near the top. In the modern hive,
the frames which contain honey are above and somewhat
separated from the brood combs. In a very thriving
colony, the queen might wish to lay eggs in some of the
honey combs. But by making the only access to these upper
combs a hole which admits workers but is too small for the
queen, the honey and the brood can be completely separ-
ated. A small amount of honey stored with the brood will
be used as food by the bees.

It was early found that hives could be made in which some
of the walls were of glass. These windows are normally
kept covered with felt which can be removed at intervals
when the bees' behaviour is to be watched.

104

HONEY BEES

The honey bee occurs in three castes, much more different
from one another than those of wasps or humble bees. The
males or drones are considerably larger and stouter than the
workers and have greatly enlarged eyes which cover most
of the surface of the head. The tongue is not well enough
developed to be useful for gathering nectar, though the drone
can feed himself on honey from the cells. He never does any
work in the hive. He lacks the wax-producing glands and the
pollen-collecting apparatus. The queen, also, is in some
respects a "degenerate " or at least specialised type, for she
also is unable to produce wax or to gather pollen or nectar.
She is a little larger, especially longer, than the worker and
has immensely developed ovaries which allow her to lay
one to two thousand eggs a day and hundreds of thousands
of eggs in her life-time. The worker has wax-glands on the
ventral side of the abdomen, has a pollen-basket on her hind
legs, and has mandibles better developed for work in the
hive than those of the queen. Normally the worker's ovaries
are small and non-functional and the little sac in which
sperm are stored in the queen is rudimentary.

Occasionally, particularly in certain varieties, a worker's
ovaries develop and she lays eggs which normally produce
only drones. Very exceptionally, by a process which is not
understood, some unfertilised worker eggs may produce
new workers or even queens, especially in some African
strains. As a rule, however, as in other bees and wasps, only
the fertilised eggs of the queen can produce queens or workers.
The honey bee is thus committed to a completely social life,
since the queen cannot found a colony alone like a queen
humble bee, while the workers cannot normally reproduce
themselves.

105

THE SOCIAL INSECTS

DRONES AND THE NUPTIAL FLIGHT

The drones are produced right through the summer, the
season of the honey-flow, and leave the hive only on warm
days. Their single function is to fertilise the queens, of their
own or of some other colony. Fertilisation takes place only
in the air, a swarm of drones following the virgin queen as
she flies upwards. The male is always killed in the act of
fertilisation, since he can eject the sperm only by generating
great pressure in his abdomen with the aid of muscles and of
the fluid pressure of his blood. Under this pressure, the
hinder part of his genital system is forced out and is left
behind in the body of the queen. The pair then fall to the
ground and the queen pulls herself away and returns to the
hive, leaving the dead male behind. The remains of the male
genital system, from which all the sperm will have passed
into the queen, are removed from the end of her abdomen,
some hours later, either by the queen herself or by workers.

This natural pairing is necessarily uncontrolled, and the
male will belong to any variety which happens to be flying at
the time. To improve their stock beekeepers have recently
developed a process of artificial insemination by which care-
fully chosen queens and drones can be crossed. Sperm are
removed from the male with a fine syringe. Then while the
queen is held in a clip under a binocular miscroscope, her
genital orifice is held open with small hooks and the sperm
are injected into her. This makes selected high grade queens
available to all beekeepers. To use such a queen, the old
queen must be removed from the hive and the new queen
introduced. Such a stranger would be attacked if precautions
are not taken. Usually she is introduced in a small cage which
protects her until she is thought to have acquired the nest
smell. She is then allowed out into the hive.

106

HONEY BEES

At the end of the summer, when no more virgin queens
are likely to be produced for some months, the drones are
driven out of the nest by the workers and die of cold or
hunger. They are probably never or very rarely actually
killed or stung, but exclusion from the warm hive is just as
effective.

While the characters of the drone are determined by his
development from an unfertilised egg, the differences between
the queen and the worker are known to depend on the diet
which the grub receives. The queens develop in special,
large, irregular queen cells which usually hang from the
bottom of some of the brood-combs. These cells are built
when required and are taken to pieces when they have been
vacated ; they are not used again like the other cells. All
grubs receive the same food for the first two days — a sub-
stance known as "royal jelly " which is secreted by a gland
opening into the mouths of the workers. It is not produced
by queens or drones. Queen grubs receive this food through-
out their life, and since the queen cells are frequently visited
by the workers the supply of this food is always adequate.
After the third day, the food given to the grubs of develop-
ing workers and drones contains an increasing proportion of
honey. If a grub in a worker cell is transferred to a queen
cell before the third day it develops into a queen, and vice
versa.

About sixty years ago von Planta claimed that there were
relatively simple chemical differences between royal jelly
and the food given to worker grubs after the third day. More
recent analyses by more refined chemical methods show that
the nature of the food is much more variable than von
Planta supposed, and the difference may be more in the
quantity of the food than in its chemical composition; or at
least that the chemical differences must be subtle and not yet

107

THE SOCIAL INSECTS

properly understood. Nevertheless, the transference experi-
ments seem to prove that diet is the essential factor in pro-
ducing queens or workers.

The difference in diet also affects the time which is neces-
sary for the adult to develop, for, while all eggs hatch on the
third day, it takes the queen on the average thirteen, the
worker eighteen and the drone twenty-one more days to
complete development. These times seem to be related both
to the richness and quantity of the diet received, and to the
relative sizes of the three castes.

Queens are in general produced by the colony under two
conditions: when it is preparing either for swarming, or for
supersedure. Swarming is the normal method by which the
colony reproduces, and in preparation for it several and in
some varieties very large numbers of queen cells are con-
structed. Supersedure is the process by which a queen who
is old or whose oviposition is falling off is replaced by a
young, vigorous queen. In preparation for supersedure
only one queen cell may be constructed and rarely more than
three. A special case of supersedure occurs if the queen
unexpectedly dies, and the rearing of a new one becomes
necessary to meet the emergency. In this situation the queen
will be reared in a " post-constructed " queen cell, that is one
built from a worker cell already containing an egg or, more
often, a grub less than three days old. After supersedure,
the new and old queens may live and oviposit together for a
long period without fighting. The intense rivalry between
queens occurs only after swarming, or if two young queens
emerge at about the same time during supersedure.

The honey bee differs from all other social species whose
colonies reproduce by swarming, because it is as a rule the
old queen which accompanies the swarm. Exceptions
chiefly occur in secondary swarms, when the colony swarms

1 08

HONEY BEES

more than once during the season. The prime swarm accom-
panying the old queen is made up of a high proportion of the
older workers. Left behind in the hive is a correspondingly-
higher proportion of young workers, and several queens still
in their cells but approaching the time for emergence. Nor-
mally, the first one to hatch becomes the new queen, and she
destroys any others which emerge or which are in the adult
stage but have not left their cells. The remaining queens
cells which may contain grubs or pupae are then destroyed
too, usually by the workers. The new virgin queen will have
to fly out for her mating flight, usually when she is about
a week old, and she begins to lay eggs two or three days
later.

While some swarming is desirable in order to multiply
the number of colonies, the bee-keeper must regard it as a
wasteful process, particularly if it occurs more than once in
the season. Swarming must be preceded by the rearing of a
large brood to give a large population, and this uses up much
food and energy and means that less honey will be stored.
Moreover the workers in the swarm carry away large quan-
tities of honey in their crops ; this enables them to secrete
wax for the construction of new combs.

BEE ECONOMICS

The hive bee collects and stores five substances, nectar,
pollen, propolis, balm and water, as well as itself producing
wax. Nectar is obtained from a wide range of flowers by the
foragers. An individual bee tends to keep to one kind of
flower for long periods, probably ^as long as the secretion
of nectar makes it an attractive source. She also tends to
keep to one relatively small patch of the flower but she is
capable of comparing sources, and changing to a new one if
the nectar is more copious or more concentrated. Ribbands

109

THE SOCIAL INSECTS

has recently analysed much of this behaviour by marking
individual bees and watching what they do when collecting
pollen or nectar. Some flowers, such as buckwheat, secrete
nectar mainly at a particular time of the day ; the bees which
visit it may restrict their foraging to this period, apparently
spending the rest of their time inside the hive. By exposing
dishes of syrup at particular times of the day it has been
shown that the foragers have quite an accurate time-sense
and can be trained to come for it at a particular hour. When
the nectar is brought in it is handed over to a home bee which
spends some minutes in concentrating it by sucking it in and
out over the surface of the tongue before storing it in one
of the cells.

Pollen is just as important to the bees as nectar, since it is
their source of protein. Without it, the life of the workers
is very much shortened and the brood cannot be reared. A
considerable store of pollen is always kept through the
winter, and this makes it possible for brood-rearing to begin
early in the spring when fresh pollen is relatively scarce. A
small quantity of honey is added to the stored pollen ; this
mixture is " bee-bread ", and it has good keeping qualities.

Propolis is made from resin collected from various plants,
especially trees. It is used to stop up any small crevice in
the hive and to cover up objects which are too large to be
removed. Some varieties of bees collect much more of it
than others do. Balm is a particular variety of propolis,
more liquid and transparent than usual. There is some doubt
whether it is collected from particular sources of resin or
whether it is secreted by the bees themselves. It is used to
varnish the inside of the cells before eggs are laid in them.

Water is collected in quite large quantities, especially in
hotter, drier climates. It is apparently used both to cool the
hive by evaporation and to prevent the interior of the hive

no

HONEY BEES

becoming too dry to suit the grubs. It is sometimes also
used to dilute honey which has become too concentrated.
The hive is maintained at a very steady temperature, only
a little below that of our own bodies. The chief source of
heat is provided by the activities of the bees and their brood,
since the digestion of sugar automatically releases consider-
able energy. Workers regularly ventilate the hive by fan-
ning their wings at the entrance. This cools it to a consider-
able extent, even if water has not been brought in and
distributed, since the nectar loses much water as it is con-
centrated into honey. Ventilation serves to remove excess
water quite as much as to reduce the temperature. It is the
brood which is most sensitive to cooling and it always
occupies a roughly spherical space in the centre of the hive,
spreading through a number of the combs.

It is clear that all these manifold activities demand a
considerable versatility in behaviour, and it is known, largely
from the experiments of Rosch with marked bees in a glass-
walled observation hive, that each worker tends to undertake
a number of successive activities according to its age. During
the active summer months the workers live five or six weeks
and the succession of duties is as follows:

l IN DAYS

DUTY

0-3

Cleaning brood-cells and keeping brood warm

3-6

Feeding older grubs

6-14

Feeding younger grubs and the queen

14-18

Secreting wax, comb-building, cleansing hive

18-20

Guarding entrance of hive

20-40

Foraging

This schedule is not rigid. There is some overlap between
successive duties, and probably not all bees reach the same
physiological age in the same number of days. Still more

in

THE SOCIAL INSECTS

important, the schedule can be widely modified to suit the
needs of the colony. After swarming, for instance, there
may be an unusually low proportion of the older workers
and the foraging may then be done by much younger bees.
Such a result can easily be demonstrated experimentally by
making up a colony in which all the workers are of about
the same age.

The key function is that performed between the sixth and
fourteenth days, for this is the time when the workers are
producing the royal jelly. This is a secretion of the pharyn-
geal glands which is passed into the alimentary canal just
inside the mouth. This liquid, on ejection from the mouth
into the cells, thickens into a pale jelly-like paste. Glands of
the same type are found in all bees and wasps, but their
function in other species has not been demonstrated. In the
honey bee, the glands become active on about the sixth day
and shrink and cease to secrete on the fourteenth, so that this
age-group of workers is essential for brood-rearing. Appar-
ently, the development of the gland can be a little hurried up
in emergencies, and it is said that it may become functional
a second time if older workers receive a diet rich in proteins.

The existence of a schedule of duties which is capable of
modification for changing circumstances is characteristic
of the social organisation of bees, but we have no idea, as
yet, how this result is attained.

THE LANGUAGE AND DANCES OF BEES

The most remarkable advance in our knowledge of honey
bees is due to von Frisch's discovery of what he has termed
" the language " of bees. The first part of this work was
published in 1923, but von Frisch largely dropped the investi-
gation for twenty years. It was taken up again in Austria
during the Second World War and led to discoveries so

112

38. Full grown larvae of
drone honey bees about to
spin cocoons and to undergo
metamorphosis into adults.

39. Full grown larvae of
worker honey bees in capped
cells spinning cocoons.

k

40. Worker pupae in their closed cells.

&>

41. Worker bees on a comb whose upper cells are filled with honey. It is
in groups such as that on the left that the dances take place (see Fig. 8, p. 1 14).

» .^.^^

- 2 m

■■■£■ .k

« % s s*

4t> «*»

*£

.^

en O

<U tj

Oh W3

X 3
(U ■£*

£ 2 en"
£ °-< i>
b£ bb^

c o o
'S o -

CU c* CD

en cu 3

CU CU c

£ § £

<£

^ en
O <U

3 3 G

.2-3 o

cu u

cu ,o

r-H cn

§ ^

u cn

03 3

cu o

cn .i;

cu >-i

44- The honey bee. A comb with thimble-shaped "royal" cells in which
fertile females develop.

45. Young worker honey bee emerging
from the cell in which it was reared.

46. A honey bee swarm divided
in two.

HONEY BEES

amazing that some biologists have found it difficult to credit
them. At least the main outlines of the work have, however,
been confirmed in the last few years by other observers, for
instance at Rothamsted in England.

It has always been difficult to understand how such a
highly organised system as a colony of any social insect
could run without a general staff and with no means of com-
munication. While no system of government has yet been
discovered, very peculiar methods of conveying information
have been revealed.

The principal experimental method has been to put dishes
of sugar and water at points various distances from the hive.
When foraging workers begin feeding at the dishes and taking
the sweet liquid back to the hive, the bees are marked in
various ways on their backs with coloured paints, so that each
one can be recognised again. By watching the behaviour of
the marked bees, both at the dishes and after their return to
the hive, observers have discovered their methods of convey-
ing information.

It is known that worker bees have glands at the end of the
abdomen, between two of the overlapping plates of the upper
surface. When the bees are feeding on a rich source of food
these glands can be seen extruded, showing as small white
patches. Von Frisch found that if he covered the extruded
gland with a small patch of shellac, the dish did not in the
next hour or so attract nearly so many new bees as did a
similar dish at which the marked bees had not been treated
with shellac. As the structure of the gland resembles that of a
scent-producing organ, it is inferred that the dish becomes
impregnated with a scent which guides newcomers to it.

Von Frisch further found that foraging bees, on their
return to the hive, perform a sort of dance, either on the
alighting board or on the comb. He found that if the source

h 113

THE SOCIAL INSECTS

from which the worker is returning is no more than about
ioo yards away, the bee does a " round dance ", running
round a circular track, first one way and then the other.
Some of these bees then leave the hive, usually before the
returned forager sets out again, and search for the source.
Some of the workers in the hive closely follow die dance and
may partly join in it. At these short distances the search
is mainly a matter of quartering the ground in all direc-
tions. It was found by experiment that if a worker returned
from a dish of syrup placed, say, at fifty yards north, then
after the dance the issuing foragers were no more likely to
find this dish than another one in a different place the same
distance away.

N^-W::>

Jt

Fig 8

Dances of the honey bee. Left, round dance ; right, tail-wagging dance,

If, however, the original or "training " dish contained
syrup scented with lavender water, then, if afterwards four
dishes were put out, one scented and three not, the foragers
all collected at the scented dish, even if it was exposed at a
different site from the training dish. Thus, the round dance
conveys the information that a rich source of food is to be
found somewhere near the hive. If the food has a scent, as
will be the case with most natural sources of nectar, this

114

HONEY BEES

provides an additional clue which helps in its discovery,
Finally, the bees themselves produce a scent which may give
additional help to the foragers.

If the source of food discovered by a forager is more than
about one hundred yards away, it performs a different dance
on its return. This is the " tail- wagging " dance, and con-
sists of running along a track shaped like a figure of eight
with the part where the two loops join straight. As the bee
runs along the straight piece she wags her abdomen. It was
found that the further away the source of food was, the more
slowly the dance was performed, but the more wags of the
tail were made on the straight part of the run. The figures
range from about eleven figures of eight in 1 5 seconds for
a source at 130 yards, to four completed in the same time at
2,000 yards. The number of tail-wags in the straight run
varied from two to three when the source was at 130 yards
to ten to eleven when it was at 830 yards. These are average
figures, and theyare subject to some variation. Moreover,
round about 100 yards transitions are sometimes seen be-
tween the two types of dance.

A relatively simple experiment demonstrates that the
dances do convey information about distances. It is neces-
sary to have a special observation hive with glass walls, so
that the dances can be seen. Then suppose a source of syrup
is exposed at 200 yards. After a time, scout bees find it and
they, and later other bees, begin taking syrup back to the
hive. The first dozen or so bees which come are caught and
if necessary killed. Next a new source is exposed much
nearer the hive and the number of new bees which come to
the two sources in the next few minutes is compared. It is
found that many more bees go to the distant source than
to the near one, though in general with random searching
near sources are discovered much more quickly. This is

115

THE SOCIAL INSECTS

taken to show that the dances which the marked bees have
been seen to perform have told other bees in the hive of the
distance (and as will be shown in a moment) also the direction
of the distant source. Thus until some bees have had time
to convey information to the hive about the nearer source,
the more distant one is much more visited. The number of
visits to a source depends at first on distance but afterwards
on whether the hive has been told about it. When told where
to go, workers find a source much more quickly than they
do by random searching.

The most extraordinary of von Frisch's discoveries was
of the method by which bees indicate the direction of a
source of food. The workers returning from a source may
perform the dance on the horizontal alighting board just
outside the hive. This is fully exposed to daylight and under
those circumstances the straight part of the track of the tail-
wagging dance points towards the source of food. With the
round dance, the direction is not indicated, as far as is known.
However, both dances are usually done in the dark, inside
the hive, on a comb which is vertical. In this case, the vertical
straight run of the tail-wagging dance indicates that the
source is on the imaginary straight line connecting the hive
with the sun. If on the straight run the bee moves upwards.
the source is towards the sun, if downwards away from it,
If the straight run is at an angle to the right or the left of the
vertical, the source is at the same angle to the direction of
the sun. This behaviour was observed in numbers of marked
bees collecting food from known artificial sources placed in
different positions relative to the hive and the sun. The
direction of the dance varies somewhat on each occasion the
straight run is made, and the average direction varies to some
extent in different workers, and at different times of the day.
In one experiment in which the direction of the dance was

116

HONEY BEES

recorded for twenty-four dances by fourteen workers over
a twenty-minute period, the angles recorded were as follows :
530 twice, 550 once, 6o° seventeen times, 65 ° once, 6j° once,
750 twice.

There is finally the problem of what the bees do when
the sun is obscured by cloud. When the sky is completely
overcast, no dance is performed, and in the dark the tail-
wagging dance on a horizontal surface becomes completely
disorientated. If, however, a patch of blue sky is anywhere
visible, the tail-wagging dance is performed as usual, the
direction being indicated with reference to the true position
of the hidden sun. It has been suggested by von Frisch
that this could be explained if bees were sensitive to the plane
of polarisation of light, for it is known that the light reflected
by blue sky is partially polarised, and that the extent of
polarisation depends on the distance of the patch from the
true position of the sun.

It should be explained that the light from the sun is emitted
in rays made up of waves vibrating in all planes, that is to
say unpolarised. But such light, when reflected from blue
sky, is more or less polarised, that is the light is orientated
so that the waves are all in one plane. Our own eyes are not
sensitive to the plane of polarisation of light, though motorists
know that a screen of " polaroid " which transmits polarised
light will cut down the glare of the light reflected from
puddles on the road or from snow. Von Frisch has very
recently found evidence that bees' eyes can respond to the
plane of polarisation of light, and that the direction of the
tail-wagging dance can be altered if a polaroid screen is
placed between the bees and the patch of blue sky which they
are using to detect the position of the sun. In fact, when the
polaroid is rotated through an angle the direction of the dance
is shifted through the same angle.

117

THE SOCIAL INSECTS

The subject is, however, still very little explored and there
are many details to clear up. Besides the honey bee, some
ants also react to the plane of polarised light.

The development of communication in the honey bee is
clearly an enormous advance in social organisation. Although
it was claimed earlier that social life starts when any female
looks after the young of one of her sisters, there is no com-
parison between the slight extension of normal nursing which
this involves and the process of conveying topographical
information. Moreover, there is no reason to think that we
have more than crossed the threshold of an immense new
field, since most of the more striking discoveries in it have
been made since 1940.

18

THE ANTS OR PISMIRES

In the ants we meet a group every member of which is social.
All of them, except a few degenerate types mentioned in the
next chapter, are found in three castes, often of very different
appearance. These are the queen, the male and the worker;
in some ants the worker may be of two types and when these
are sharply different and do not intergrade, the larger one
with a larger head is known as the soldier. As a rule the
male is winged and often like one of the lower solitary wasps
in appearance. He does not long survive pairing and can be
found only during a short part of each season. The queen
is usually considerably larger than the other castes and is
also winged. The wings are, however, removed after the
marriage flight. The winged females are thus found only for
a short season, though the queen herself may be very long-
lived, sometimes surviving for fifteen years. The workers
are usually different from the queen both in colour and
structure. They do not possess wings, and the absence of
wing-muscles has led to a reduction in the size of the thorax
and a simplification of its structure. Often the eyes are very
small compared with those of the queen and in many species
they are minute or absent.

Ants are much more flexible in their building habits than
bees or wasps, and they keep their brood piled in chambers,
not separated in cells. Most nests are built of soil or of small
vegetable fragments, and they are more easily adapted to

119

THE SOCIAL INSECTS

different habitats or to an increase in the size of the colony.
Some ants add saliva to vegetable matter, to construct a
stiff " carton ". A few use silk, which their grubs spin, to
join leaves together. Usually, however, no scarce materials
are required to make the nest. Perhaps one reason for the
success of the ants has been that their nests do not depend on
hardly won materials. They are at any rate more ubiquitous
than any other group of social insects, and the only habitats
on land in which they are not found are those of the Arctic.
They are also much the easiest of the social insects to keep
under observation in captivity. Many species can be main-
tained for years in small chambers of plaster of Paris with
a glass lid. It is necessary only to keep them damp and to feed
them on sugar and on occasional dead insects Much of our
knowledge of ants, especially of the early stages of the colony,
depends on observations made on captive colonies.

There are good grounds for thinking that ants have led
a social life for a much longer period than either bees or
wasps. Certainly the castes were fully developed at the time
the Baltic amber was fossilised, thirty to forty million years
ago, and even long before this. We know this, because we
find ants perfectly preserved in the amber. It is probable that
their social organisation will be found to be more elaborate
than that of the bees, but at the moment it is much less under-
stood. Although it is not difficult to keep ant colonies in
observation nests, no ant species has been as intensively
studied as the honey bee. An immense amount has been
recorded about the general natural history of ants, but most
of this information is purely descriptive. That is, though we
know in a general way what many species do under natural
conditions, we do not know how they do it. A real under-
standing of ant behaviour would require much more investi-
gation by means of experiment.

120

THE ANTS OR PISMIRES

ANTS AND OTHER ANIMALS

Let us consider the relation of ants to the aphides and
mealy-bugs which provide much of the sugary part of the
diet of many species. This is what Zimmermann says about
the pineapple mealy-bug, a scale insect which causes very
serious losses in the pineapple plantations of Hawaii.

Mealy-bugs become established on the pineapple planting
material during growth. These individuals persist on the future
plant sets while they are partially dried after trimming and can,
on occasion, even reproduce on this detached material. Hence,
a newly planted field may have a large initial mealy-bug popula-
tion. However, peculiar as it appears, most of these populations
disappear either because of the lack of adequate attending ant
populations or because of predator pressure or other causes.
The serious infestation of a new field comes from outside,
adjacent areas, especially old pineapple fields, waste or un-
cultivated lands. High infestations may build up from the out-
side within a period of six months. This movement is dependent
primarily upon the action of the attending ant Pheidole magace-
phala. If the ant populations are low, the subsequent mealy-bug
invasion of the fields will be slow and of low grade. If the ant
could be eliminated, the fields would be largely free of serious
mealy-bug infestation. Ants are essential for the proper develop
ment of mealy-bug colonies, for they tend them, shelter them,
protect them from parasites and predators and keep them clean
from detrius which, when massed with honeydew, has a
gumming-up and deleterious effect on the colonies. In short,
pineapple mealy-bugs have a difficult time maintaining them-
selves unless ant-attended.

There is another side to this picture. According to
Clausen, the date-growers of the Yemen have for centuries
carried certain ants' nests to their orchards to protect them
against other ants. In China, the citrus-growers collect nests
of the red ant, Oecophylla, and put them in their trees,

121

THE SOCIAL INSECTS

sometimes with bamboo runways from trunk to trunk. The
ant is stated to kill many caterpillars and to drive away
various beetles and bugs.

The fact that ants feed on the sweet, liquid excreta of
aphides and mealy-bugs has been known for at least two
hundred years. Linnaeus, indeed, said that aphides were
ant-cows. Ants often keep certain species in their nests.
Many years ago Sir John Lubbock showed that ants collect
the eggs of the bean aphis in the autumn and in the spring
plant them out again on the food-plant. But the relations
between the two sorts of insects have hardly been studied
experimentally. It is still doubtful in most cases how far the
ants protect their cows from enemies. Apart from a good
deal of practical information of the type quoted from
Zimmermann's work, we do not know very much more of
the relations between the two sorts of insects than Sir John
Lubbock did. The absence of a detailed experimental study
is felt in the same way in a great many other facets of ant
life.

THE FOOD OF ANTS

The methods by which ants provide themselves with food
are analogous to three types of human culture recognised by
anthropologists, namely, hunters, food-gatherers, and agri-
culturists. Many ants, especially those which in structure
most resemble the non-social wasps, are purely hunters.
They catch and kill insects or other small animals. As a rule,
these are brought home whole and then cut up, to be divided
amongst the members of the colony. Often all available prey
of the right size are captured; less commonly, an ant may
specialise on a particular sort of prey. The American Lepto-
genys, for instance, catches mainly wood-lice. Most of the
hunters have a well-developed sting.

122

THE ANTS OR PISMIRES

The food-gatherers collect either the seeds of various
plants, or nectar from flowers, or the liquid excreta of aphides
and scale insects. Some of them combine such collection with
hunting, but others, such as the harvesting ants, live entirely
on seeds, especially of grasses. In the harvesters the large-
headed soldiers act as seed-crushers for the colony. The large
head contains the very powerful muscles which work the
mandibles.

The most highly developed agriculturists are the leaf-
cutting ants. They subsist entirely on a special kind of fungus
grown on the fragments of leaves which they carry into
underground chambers. The queen of these ants takes a
little pellet of the fungus with her on her marriage-flight,
so that the food is passed on to each new nest.

Some of the ants which feed on the excreta of aphides can
also be classed as agriculturists. The common yellow ant,
Lasius flavus, which makes mounds in fields, obtains almost
all its food from, aphides living on the roots of plants in
underground chambers in or near the nest. These might well
be compared to herds of cattle, except that we do not yet
know how far the ants manage them in the way in which
man does his domesticated animals.

Some of both the food-gatherers and the agriculturists
have the sting greatly reduced, and in their attacks on their
enemies squirt the poison instead of injecting it with a stylet.
It is not clear what they have gained by the change, though
some of the most common and successful species retain the
sting only in a rudimentary form.

We may now consider in more detail the three main ways
by which ants feed themselves. A great many ants subsist
partly by hunting, as even civilised man does to a limited
extent — by fishing, for instance. The wood ant, which often
makes large mounds in pine-woods, brings in large quantities

"3

THE SOCIAL INSECTS

of small insects such as caterpillars. It may also find small
dead insects which are worth collecting. Ants are always a
serious nuisance to solitary wasps whose prey is stolen while
awaiting storage in the wasp's nest. But the wood ant also
collects large quantities of aphis excreta. Okland has cal-
culated that a large nest in Sweden may collect as much as
20 lb. of dry sugar in a season ; this means a very much
larger quantity of the liquid excreta. As a rule, hunting
workers act singly, so that the prey are no bigger than one
ant can overpower and carry by itself. In some ants, how-
ever, the workers co-operate to overpower much larger
insects, such as big beetles or even small birds.

Hingston has described how this is done by an Indian
ant, Oecophylla smaragdina. One of a number of hunting
workers happens to find a beetle and seizes one leg. It is too
powerful for the one worker. There is a struggle which soon
attracts others. Each one seizes some projecting part and
pulls, and soon the beetle is splayed out with all its members
pulled in different directions. The ants continue a steady
pull, if necessary for half an hour, until the prey is dead. The
workers then co-operate to bring it home; most of them pull
the legs on one side, but a few remain on the other side to
steady it. In this way it will be dragged home even if the nest
is high up in a tree.

There is one group of ants, the Ponerines, which are pre-
eminently hunters. Very few of them collect aphis excreta
or are attracted to sweet substances, though a few of them
feed, at least partly, on seeds. In structure the Ponerines are
the least specialised of the ants, in many ways most like what
we imagine the ancestors of the whole group to have been.
Both in their addiction to hunting living prey, in the small
size of their colonies, and in other features they seem to be
more primitive than other ants. They are now relatively

124

THE ANTS OR PISMIRES

rare and most live in retired situations, especially in the
tropical rain forests. Like the marsupials, which are primitive
mammals, they are specially numerous in Australia. Here
one of the dominant forms is Myrmecia, the bull-dog ant,
which is by no means retiring. The workers of some species
are nearly an inch long and are capable of jumping a foot at
a time. Froggatt describes how, if their home is approached,
they come jumping out one after another like a pack of dogs.
As they can bite and sting severely the nest must be treated with
great respect. Another Ponerine, Lobopelta, is a very active
hunter and tends to specialise on termites. Hingston has des-
cribed how an Indian species sallies out in dense armies to attack
termite nests. Some of the ants enter the nest and begin to kill
termites, throwing them out or bringing them to the entrance.
Other ants act as porters and carry the prey back to the ants'
nest. All who have observed these and other raiding parties of
ants are irresistibly reminded of a skilful military operation.

THE DRIVER AND LEGIONARY ANTS

This type of behaviour reaches its highest development
in the driver ants of Africa and the legionary ants of South
America. These together form a distinct section of the
Ponerines and have much in common. The workers of the
drivers are always blind, whereas the legionaries usually
have rudimentary eyes. Some kinds are almost entirely
subterranean, others merely nocturnal. One of the legionaries
builds a tunnel of grains of sand as it goes along and in this
way passes in shelter from log to log, looking for beetle-grubs,
etc. Others especially in the tropical rainforest, are fully
diurnal and easily observed, as I have myself proved in British
Guiana. A raiding army covers an area of several hundred
square feet, and as one approaches it one hears the character-
istic chirping of the ant thrushes. These are birds which have

"5

THE SOCIAL INSECTS

the habit of catching the numerous insects trying to escape
the ants; the birds themselves escape both by their size and
mobility. One can, in fact, stand still in the middle of a
raiding army and come to no harm if one moves on every
minute or two. But there is no such escape for any insect or
small animal which cannot fly away. Wasps' nests, for
instance, are deserted and the whole brood is captured.

The African driver ants are even more formidable and will
kill even pigs and fowls if these are enclosed. They have
been known to devour a large python which was so gorged
after a meal that it could not escape. These ants will enter
houses and clear them entirely of the vermin, such as cock-
roaches and spiders. Fortunately, human food other than
meat does not attract them, so that their visit is on the whole
beneficial, at least in retrospect.

Most driver and legionary ants construct only temporary
nests. Often they occupy some natural hole or crevice, such
as may be found at the roots of a large tree. Some of them
have the faculty of building a structure by using the bodies
of some of the workers themselves. By hanging onto one
another long chains can be formed, and the chains may be
arranged to arch over and enclose spaces. In this way a
temporary living nest is produced, inside which the queen
and brood can be protected and in and out of which the other
workers will be seen passing. The temporary nature of most
of these colonies appears to be the inescapable result of their
mode of life. Ants which occur in such large numbers and
which live exclusively by hunting soon denude any one area
of food and have to move on. There are many accounts
not only of the ordinary predatory raids but of the shifting of
the whole colony when all the brood is carried to a new site.

Some idea of the physiological background to the restless
activity of these ants has been obtained recently in America,

126

THE ANTS OR PISMIRES

by Schnierla. There are four types of raiders — subterranean,
arboreal (mainly), terrestrial swarm-raiders and terrestrial
column-raiders ; Schnierla's observations were made on the
last two types. They differ in the way their raids are
organised ; the first makes a swarm which fans out over a
considerable area, the second moves in much more compact,
narrow columns.

The ants have a regular alternation of two phases of
activity : one which lasts about three weeks during which
the nest or " bivouac " is stationary and only a few raids
are made, and one which lasts fourteen to seventeen days in
which the bivouac is moved almost every day and during
which there are several raids each day. This alternation of
relative inactivity with great activity depends on the ovi-
position-cycle of the queen. She goes through a short phase
in which her abdomen becomes greatly distended and she
then lays about twenty thousand eggs in a few days. When
these hatch and while the brood is growing, the colony is
very active and raids are frequent. When the grubs are
nearly full grown about two quarts of insect prey are brought
in every day. As these grubs begin to spin their cocoons,
raiding dies away rather suddenly and the bivouac becomes
stationary. Activity begins again just before the new lot of
ants emerge from the cocoons. The regularity of their
behaviour depends on the regular oviposition-cycle of the
queen and on the way in which almost all the brood keep in
step with one another during their development.

It is still not understood how the queen maintains a regular
cycle in egg-production. It is clear, however, that though a
stationary colony consisting of eighty thousand or more
ants could not feed itself for long by raiding, it is not the
shortage of food which directly controls their movements.
It is rather that the need of food for the growing brood

127

THE SOCIAL INSECTS

stimulates raiding and that when raiding is active the bivouac
is frequently shifted. On each occasion all the brood have
to be carried to the new site.

GATHERERS OF MIXED FOODS

Just as most ants do at least a little hunting, so most of
them will collect other foods of various types, especially the
sweet excreta of aphides, scale-insects, mealy-bugs and leaf-
hoppers. Some ants, like the common red stinging ants of
the genus Myrmica, are hunters as well as collecting some
nectar from flowers and some excreta from various insects.
At the other extreme, the common yellow mound-building
Lasius flavus subsists almost entirely on the excreta of root-
aphides, but also takes some dead caterpillars, etc. I once saw
on top of an ant mound a large dead caterpillar completely
surrounded by fifty or sixty ants dismembering it. From the
great excitement which this ant exhibits if given animal food
in an artificial nest it seems certain that this is an important
supplement to its sugary diet. The relative importance of
different articles of diet to the various kinds of ant is very
little known, but it is clear that the balance is different from
genus to genus.

None of the British ants builds elaborate shelters above
ground to shelter aphides or mealy-bugs, though they may
build up earth round the base of a plant which these insects
inhabit. Many exotic ants, however, build elaborate structures
around the suppliers of their sweet food and seem, at least
to some extent, to protect them from their enemies No
doubt the shelters are usually built round a natural colony
of insects which the ants happen to have discovered, but
there is a little evidence that they sometimes drive insects
into their sheds and prevent them from escaping from them.
Hingston describes an Indian ant, Polyrhachis, which builds

128

47- A worker ant (Formica fusca) with aphides. Gentle tapping of the
abdomen often makes the aphid produce a small drop of sweet excreta.

48. Nest of wood ant (Formica rufa). The mound is three feet high and
eight feet in diameter.

Vi

?«**£

49- Worker wood ants (Formica
rufa) combining to carry a pine
twig to nest.

50. Nest of termite (Amitermes
vitiosus) in Australian scrub.

51. Fertile female termite, or queen — nearly all swollen abdomen. Compare
with Fig. 12 on page 174.

THE ANTS OR PISMIRES

its shelters from varied pieces of debris which it spins together
with the silk obtained from its own grubs. These shelters
may be only an inch or two long or as long as a foot or more.
They surround part of the stem or a few twigs from which the
food-supplying insects suck the juice. This particular ant
often keeps leaf-hoppers of the family known as Membracids,
sheltering the adults and the early stages, both of which but
especially the young ones supply sweet excreta. Hingston
describes ants poking with their heads at a Membracid to
make it walk along a branch until it entered the shelter. He
also saw ants driving them back into the shelter when they
escaped through a hole torn in it by the wind. They also
seemed to prevent them from leaving by the entrances which
the ants themselves use. A great many ants behave in a more
or less similar way to Polyrhachis, but the exact meaning of
their behaviour will be understood only when many experi-
ments have been carried out.

HONEY-POT ANTS'

One well-known and extraordinary development of this
type of behaviour is shown in the honey-pot ants. They all
occur in the drier parts of the world such as the southern
U.S.A., Australia and South Africa. Here there is probably
a need to store a reserve of food to tide over periods of
scarcity, just as the honey-storing wasps, Brachygastra, do in
Mexico. The best-studied species is the one which occurs
in Colorado: this ant stores in the bodies of certain workers
the sugary liquid which it obtains from aphides, scale-insects
and from the surface of an oak-gall. These living storage-
tanks are known as " repletes " and have their abdomens so
greatly swollen that the hard plates which normally form a
continuous armour are widely separated by the distended
membrane connecting them. The repletes hang from the

i 129

THE SOCIAL INSECTS

roofs of special underground chambers in the nests and are
almost incapable of movement. If they happen to fall from
their perch they are unable to get up again. They are fed by
other workers who also bring the liquid which distends
them. The repletes start life as ordinary workers, but before
their skeleton has fully hardened they are given so much liquid
food that their crop distends their abdomen and this condition
becomes permanent, lasting for months or even years.
Though the repletes form only a small proportion of the
workers, there may be two hundred or three hundred in the
nest. The repletes give up part of their store when solicited
by other workers, and it is presumed that they can thus keep
the whole colony going during a period of prolonged
drought. It is a matter of common observation that many ants
when returning from a visit to plant-lice have the abdomen
considerably distended, but normally it contracts again as
soon as they have shared out the food. This facultative
distension has been much exaggerated in the honey-pot ants
and has become permanent in the repletes.

HARVESTING ANTS

One group of food-gathering ants has been familiar from
remote antiquity, namely the harvesting ants. It was to them
that Solomon directed the attention of the sluggard. They
are found in many of the drier parts of the world, especially
round the Mediterranean basin. Many ants make some use
of seeds as a source of food and the British red ant, Myrmica,
already mentioned as a hunter and a collector of aphis excreta,
has also been observed bringing cornflower seeds into its
nest. Many seeds have a small part which contains more fat
than the rest, and this is especially eaten by ants. Some
botanists have, indeed, maintained that the fat-reserve of the
seed has been evolved to make it attractive to ants, just as

130

THE ANTS OR PISMIRES

the nectar makes flowers attractive to bees. The ant thus acts
as an agent in seed-dispersal. Whether or not this is true, it
seems that ants do play some part in spreading plants such as
gorse which have heavy, edible seeds. In this example, the
ants eat only part of the seed; the seed consequently remains
capable of germination.

The true harvesting ants live mainly on seeds, especially
of grasses. Their behaviour has an extraordinary analogy
with that of the honey-pot ants, for they store the seeds in
underground chambers and use them as a source of food
during periods of drought. Tracks leading to the nest often
look like well-marked roads, since many species of these ants
forage in large companies. Usually the seeds are brought in
enclosed in the chaff, and this is removed inside the nest
and brought out again and thrown on a rubbish heap
which forms a sort of ring round the nest. In some species
the nest may be a large mound, sometimes as much as 20-30
feet across and penetrating 6 feet or more into the soil. The
entrance may be large and lie in a crater made up of coarse
soil pellets.

It is well established, and was indeed known to the ancient
Romans, that if the seeds get damp in their underground
chamber, the ants bring them out and spread them in the sun
to dry. Later they carry them in again. They may also bite
off the part of the seed called the radicle to prevent them from
germinating. In spite of these measures their stores do some-
times germinate; they are then thrown away on the rubbish
heap. This gave rise to a story that the ants planted seeds to
produce a crop near the nest for. the following year. Al-
though it is true that rejected seeds often germinate and grow
in the rubbish heap, and that the crop of seeds thus produced
may be harvested, there is no reason to suppose that this is
anything more than an accident. Many of the harvesting ants

131

THE SOCIAL INSECTS

have a special caste of large workers with enormous heads.
These are supposed to act as seed-crushers, making the food
available for the smaller ants who do all the collecting.

LEAF-CUTTING ANTS

The most elaborate agricultural operations are undertaken
by the leaf-cutting ants. These grow special fungi on frag-
ments of leaves or flowers which they carry into their nests
and store in underground chambers. Leaf-cutting ants are
found in America, from Texas to Patagonia, but are most
numerous in the tropics of South America. There are many
species, and there is a considerable range in structure and
behaviour. It will suffice here to describe the largest and most
specialised type, which belongs to the genus Atta.

In most parts of South America nothing is more familiar
than the processions of the leaf-cutter or parasol ants. Each
worker is carrying a fragment of leaf or flower, held upright
over its head, " like Sunday-school children carrying ban-
ners " as the Rev. McCook remarked in one of the early
accounts of their habits. In many districts they are a serious
menace to agriculture, since they strip all the leaves of
plantations, often seeming to prefer the non-native vegetation
grown by the farmer. The colonies are of very large size,
with well-beaten paths radiating in all directions from the
nest. The nest consists of large irregular mounds of earth
several feet high and 20 or 30 feet across. The entrances are
as large as rat-holes, and there often seem to be long under-
ground passage-ways leading to points more than 100 feet
from the nest. The underground part of the nest extends
downwards for 10 feet or more, and contains a large number
of fungus-chambers connected by wide galleries. The
chambers are quite large, perhaps a foot or more high and
1-4 feet wide. They may be connected to the surface by

132

THE ANTS OR PISMIRES

ventilating shafts which maintain suitable conditions of
moisture, etc.

The leaf-fragments brought in are chewed up and placed
in the chambers where a special fungus grows on them. This
is not a mere mould growing on decaying vegetable matter,
but a particular, very peculiar kind of fungus which seems
to be different for each sort of leaf-cutter. Indeed the fungus
is so peculiar that it is difficult to place in the botanical
system. Much of the nutriment of the fungus is provided
by the liquid faeces of the worker ants, which are added to
the leaves ; growth of the fungus penetrates all through the
mass of leaves and develops a number of little knobs, almost
like a cauliflower head, and on this all stages of the ants feed,
exclusively. Some of the youngest grubs may be fed on food
regurgitated by the workers, but the larger ones lie about
amongst the fungus on which they feed themselves. The
workers, as in so many ants, vary greatly in size. The
smallest ones stay in the nest, weed the fungus gardens of
moulds, and nurse the grubs ; the middle-sized ones, which
are the most numerous, cut and collect the leaves; the largest
ones with very big heads and mandibles guard the entrance
to the nest.

When the female flies off on her marriage-flight, a small
pellet of fungus and debris is carried in the pocket which is
found in all ants beneath the mouth. After mating, she sheds
her wings, excavates a small underground chamber, and
ejects this pellet onto the floor. The fungus soon begins to
grow and is nourished with her liquid excrement. A piece
of the fungus is taken in her jaws from the growing sheet and
held against her anus from which a drop of clear yellowish
fluid is ejected. It is then replaced in the young fungus
garden and this may be repeated every hour or so. Thus this
manuring of the fungus is not an accident due to living in a

i33

THE SOCIAL INSECTS

small enclosed chamber, but is an elaborate, instinctive
performance.

At this time, also, the queen begins to lay eggs. Some of
these are eaten and probably help to provide food for the
fungus. Others hatch and others again may provide food for
the youngest grubs. Soon after the first workers hatch, the
colony begins to collect vegetable fragments or to cut leaves.

TYPES OF NEST

Ants' nests are in many respects simpler than those of bees
or wasps, and this may give the colonies an advantageous
flexibility. Even so, there is a considerable variety in the
types of nests, some of which are remarkable structures. The
commonest type is built in the ground, either under a log or
stone or with a mound raised over it; more rarely there is
almost no external sign. The mounds are built by the ants,
bringing up the earth a few grains at a time, especially in
the early summer. Such mounds in England are often con-
fused with mole-hills, but they are made of much finer earth
and the ants can be found by digging into them. By tunnell-
ing in all directions, and leaving pillars to support the roof,
they are able to make as convenient a nest as a bee or wasp,
but in a much simpler way and with less work.

The larger mounds certainly take a long time to build,
and many of them must be twenty years old or more. Since
queen ants have been kept in captivity for more than fifteen
years, and since replacement of the queen is possible as in the
honey-bee, the colony can live for long periods without
difficulty. According to whether it is dry or wet, hot or cold,
the workers shift the brood about to different parts of the
mound, to give them the right conditions. Some ants are
said to have two nests, one for the summer and another, in a
more protected situation, for the winter.

i34

THE ANTS OR PISMIRES

Other nests are less simply constructed — for instance those
made of carton. This is a product of vegetable matter,
usually wood-particles, cemented together with saliva. It
ranges in texture from soft, leathery or friable material to
something which is fairly tough and may be very like wasp
carton. Some British ants build carton nests inside rotten
trees, but many tropical species build large external nests
which hang from the branches. There may be a number of
chambers, hanging from different branches and connected

Fig. 9

Brigade of Oecophylla smaragdina workers drawing edges of

leaves together while other workers bind them in place with the

silk spun by the larvae (after Doflein).

by runways, so that most of a large tree is occupied. Other
nests are more like a series of purses, each under a different leaf
of the tree, so that the separate chambers are very numerous.
The most remarkable of the tree nests are those of the
species of Oecophylla, already briefly mentioned. The way
the nest is built has been watched by a number of observers
and the same method has been seen in several other rather
different kinds of ants, in other parts of the world. Several
ants pull the edges of two leaves together ; towards the tip
of the leaves where the gap between them is wider, several
ants form a chain, each one holding onto the one behind with
its hind legs. The number of ants and the length of the chain

i35

THE SOCIAL INSECTS

is adjusted to the width of the gap at different points along
the leaf. When the edges of the leaf are drawn together,
another ant comes along bringing one of the larger grubs in
its mouth. With the silk which is emitted from the lower lip
of the grub, the leaves are sewn together by a regular criss-
cross of threads. Eventually enough silk is produced to close
the gap with a solid sheet.

This silk is the material which in Ponerine ants serves to
form the cocoon when the grubs change into the pupa. In
many of the higher ants no cocoon is formed, the pupa
lying naked in the nest. In forms like Oecophylla, the pupa
is also naked, but the ability to produce silk has been retained
and diverted to another use. The method of construction can
be seen if a hole is made in the nest and if the ants are
observed repairing it. This seems to provide one of the
best examples of the co-operation of individuals to perform
a complicated operation.

One further type of nest is interesting as an example of
the adaptability of ants. Many species build small nests in
natural cavities in plants, such as hollow stems or empty oak-
apple galls. A number of plants have natural preformed
cavities which are regularly used by ants. Such are many
species of Acacia, in which there are hollow stems or large
hollow thorns ; other plants have swollen, hollow bases to
their leaves. Certain kinds of ants are regularly and exclu-
sively associated with such plants, nesting in no other situa-
tion. It has been suggested that the plants have evolved in
this way in order to attract the ants ; the ants in turn are
supposed to protect them from the other insects which would
feed on their leaves. There has however been a good deal
of evidence that the leaves of such plants are no less eaten
than those of plants not inhabited by ants, and the question
needs much more study.

136

THE ANTS OR PISMIRES

HOW COLONIES ARE FOUNDED

The normal method of founding new ant-colonies is by
a single queen after her marriage flight. These marriage
flights often take place over wide areas simultaneously. One
day during July 195 1 the common black ant, Lasius niger,
was swarming over much of the country between Ascot and
Ealing, near London, a strip about 20 miles long. Such
observations are not at all uncommon, and millions of queens
and males are involved. It must lead to more crossing between
different nests than would occur if each one swarmed on a
different day. It is not known what conditions determine
the time of swarming, though it is often during warm,
thundery weather. Just before swarming the workers are
seen crowded round the exit-holes of the nest, and appear
to be restraining the young sexual forms from coming out
too soon.

Usually the queen is very much larger than the male or
worker. After mating, she falls to the ground, breaks off
her wings at a line of weakness near their bases, and hunts
for a crevice or digs a small cell under a stone. Here she
lives without food for nearly a year, being sustained by food
reserves derived from the breakdown of her wing-muscles.
Only in the Australian bulldog ants and in some of the other
Ponerines is she known to go out and forage for food during
this period. When she starts egg-laying some of the eggs are
eaten, so that however many she lays few survive. The small
number of grubs which she raises are fed on her saliva, or on
fragments of eggs or of other grubs. Eventually a brood of
unusually small workers is produced, and from then onwards
the queen can spend all her energies on egg-laying. This
method of colony-foundation is possible only when the queen
is much larger than the workers; otherwise she would not

i37

THE SOCIAL INSECTS

contain enough food-reserves to rear the first brood. Prob-
ably the Ponerine ants, in which the queen is still relatively
small and forages for food before the first workers emerge,
retain a more primitive condition.

One may suppose that the cloistered cell of the majority
of the queen ants has turned out in the long run to be less
risky. It does, however, put a considerable strain on the
colony, since it has to produce large numbers of big queens.
Accordingly several modifications of the usual method are
known. In some the size of the queen is reduced, and this
entails a change in the method of founding colonies. Some-
times the new queen is received back into her own or a nearby
colony. With many small queens the egg-laying rate can
be even higher than when there is only one large one. More-
over the colony is potentially immortal : its survival does
not depend on the life of a single queen.

The Brians have recently shown that the common British
ant, Myrmica rubra, occurs in two forms. In one the queens
are larger than the workers and colonies are founded by single
females. The other has relatively small queens which go
into the parent or some other nest after fertilisation. Ant
colonies of this second sort may multiply by " budding ",
either through the simple division of a very large and stragg-
ling colony or by a group of workers and queens seeking
a new site. The co-existence of several laying queens does
not seem to be an easy achievement for most species :
jealousy of other egg-layers is much more characteristic.
Moreover, budding is hazardous because the species cannot
spread so rapidly or so widely.

The commonest way round this difficulty seems to be the
type of behaviour known as "temporary social parasitism ".
The small queen finds her way into the nest of another
species, and succeeds in displacing the rightful queen. The

138

THE ANTS OR PISMIRES

alien workers bring up the early broods of the intruding
queen, and as they die off and are not replaced the colony
gradually becomes a society of the intruding species only.
This type of behaviour seems to lead in the end to more and
more pronounced parasitism of the sort which will be
described in the next chapter. A well-known example of a
temporary social parasite is the British wood ant, Formica
rufa, which may found colonies by budding but whose
queens also invade the nests of another ant, Formica fusca.

An entirely different method of solving this problem has
been evolved amongst the driver, legionary and a few other
kinds of ants. In them, the ant which lays all the eggs is
wingless, blind and has a thorax formed like a worker rather
than a queen. It is supposed that in these species the original
queen has entirely dropped out, and that her functions have
been taken over by a special type of egg-laying worker.
Since this acting queen has no wings she mates on the ground
and the species spreads only by the wanderings of the whole
colony. This arrangement is not found in many kinds of
ants, and probably has the same disadvantages as ordinary
budding, especially a loss of mobility for the species.

THE CASTES OF ANTS

As a rule, in any one species the male and female are rather
constant in form and colour, while the workers are much
more variable. In the common, yellow, mound-making ant,
the workers vary in size only to a moderate extent. In the
wood ant the variation in size is much greater, and there is
also a good range in the amount of black markings.

In many non-British ants the variation or polymorphism
is much more marked. In driver ants or leaf-cutting ants the
biggest workers are many times the size of the smallest ones.
Moreover the shape of the mandibles, or of spines on the

139

THE SOCIAL INSECTS

thorax, may vary in a marked fashion. Finally, in some
groups, such as the large genus Pheidole, the workers are
found in two types with no intermediates, the small workers
with normal heads and the larger soldiers with dispropor-
tionately large heads. While some variation in size might be
expected in any species, especially if there is any likelihood
of competition for food amongst the brood, the variation
within the leaf-cutters, for instance, seems far more than
would be expected from such a cause. Besides, the grubs
which give rise to the queens are brought up in very similar
way, yet the queens vary little in size. Where the workers
are dimorphic, i.e. of two discontinuous types, there must
be some special explanation. However, it might be argued
that the females of nearly all ants are dimorphic, developing
either into queens or workers. The existence of the two sorts
of worker in Pheidole might then be regarded as no more
than a special case of the more general problem of female
polymorphism. If we could explain why it is that some grubs
produce queens and others workers, the variation within the
worker caste might also be understood.

The question of what determines the development of a
queen or a worker is one about which a debate has gone on
for years; it is still not all settled. The answer is not neces-
sarily the same for every kind of ant, though it is usually
assumed that it is likely to be so. There are two main sorts
of theory. It is suggested first, that the queen lays two or
more sorts of eggs (genetic theories) ; second, that originally
similar grubs are given different quantities or types of foods
(trophic theories). In the midst of disagreement, there are
at least two certain facts: that the differences between males
and females (or workers) are genetic, and are due to the laying
of two sorts of eggs ; and that by starving ant-grubs at least
some variation in the size of queens or workers, running

140

THE ANTS OR PISMIRES

parallel with the simplest kind of natural polymorphism, can
be produced experimentally.

In the past, the chief support for the trophic theories has
been derived first from what is supposed to occur in bees and
wasps ; in these caste-production has been supposed to be
entirely due to differential feeding. Secondly, in most ants
the polymorphism of the workers is mainly a matter of size.
Species with a sharply discontinuous soldier caste are few.
Thus the most usual type of variation within the worker
caste is very much what would be expected from an unequal
division of food. It is known from observation nests that
worker ants may give exceptional attention to grubs which
happen to be larger than the others, so that a slight natural
variation in size might be accentuated by the behaviour of
the nurses.

Recently, some experimental evidence has been obtained
which supports the trophic theory of the origin of castes.
Goetsch and Gregg have studied the relation of food supply
to the production of soldiers in Pheidole. This is an ant in
which the large, swollen-headed soldiers are sharply differen-
tiated from the small workers; no intermediates between them
are found. Goetsch showed that a group of grubs given
honey or some other liquid food turned into workers, where-
as a group provided with pieces of dead insects produced
one or more soldiers ; the number depended on the quantity
of solid food. Liquid food is taken into the worker crop and
then some is regurgitated for each grub individually, so that
each one gets a more or less similar share. Solid food, on the
other hand, is merely put near the grubs, so that one or more
of them, by reaching it and feeding themselves, get a much
bigger share than the others. This solid food must be received
during the first five days of life, otherwise the grub irre-
vocably becomes a worker. The feeding on the fourth and

141

THE SOCIAL INSECTS

fifth days seems to be most important, and there seems to be
a short period when the grub can be most readily diverted
from a worker into a soldier. The first group of eggs laid
by a young queen is, however, incapable of producing
soldiers even if the eggs are removed and given to an older
colony for rearing. Soldiers can be produced only from eggs
laid in the second or later batches of the queen.

In experiments by Wesson, on an American species of
Leptothorax, an effect of food on the production of queens
was shown. Grubs from one colony were made up into two
groups of forty-four and put with workers in artificial
nests. One lot were given abundant food, the other just
enough food to allow growth and to prevent cannibalism.

The results were as follows :

WELL FED POORLY FED

Initial number of grubs 44 44

Died during the experiment 6 9

Formed pupae 38 35

Produced males 3 2

„ queens 32 10

„ workers 3 23

It seemed, however, that food supply was not the whole
explanation, since grubs present during the summer did not
respond to such treatment. Only grubs which had over-
wintered were responsive in such experiments.

Brian has very recently found some confirmatory evidence
in the British ant, Myrmica rubra. He finds that the grubs
of this ant all pass the winter in the third moult stage,
though they vary greatly in size ; the biggest ones are ten
times as heavy as the smallest. In the spring, when the
colony wakes up to life, the biggest ones develop into queens,
the rest into workers. If the ratio between nurses and grubs
is altered experimentally the proportion of queens and

142

THE ANTS OR PISMIRES

workers produced can be similarly altered. With one nurse
for each grub, provided enough food is given to the colony,
all may become queens, whereas with one nurse for ten
grubs, all will become workers. A similar result can be
obtained by partial starvation of the half-grown grubs and
in this way, in certain conditions, intermediates between
queens and workers were obtained.

These experiments have not yet been described in full,
and it is not known why the grubs vary so much in size at

Fig. io

Male-worker mosaic of the ant,

Formica sanguinea, found at

Bewdley, Wore, 20.vii.1909

(Bondroit). Right half male.

the beginning of the winter. There is always the fundamental
difficulty that worker ants may recognise the caste of newly
hatched grubs although we cannot do so. This might lead
them to treat each type of grub in the appropriate way. On
the other hand, in the leaf-cutter ants, as Wheeler pointed
out, fungus appears to be almost the only food of all grubs
which have access to an almost unlimited supply, and yet
the workers are highly polymorphic.

What has been regarded as the best evidence for the

M3

THE SOCIAL INSECTS

genetic origin of castes comes from the study of various
abnormal individuals. Very occasionally specimens are found
which may Be called sex-mosaics. These have part of the
body female, part male, with the areas often sharply marked
off from one another. Where the male and female are different
in size, shape and colour, such mosaics are asymmetrical and
look very peculiar. Analogous mosaics which are male-
worker, male-soldier, queen-worker, soldier-worker and
soldier-queen, have also been found.

The origin of one kind of male-female mosaic, known as
a gynandromorph, is fairly well understood. During develop-
ment, some particular cell divides abnormally, so that one
daughter cell receives a full set of the hereditary material
(chromosomes), whereas the other gets only a half set:
the latter, in the Hymenoptera, results in a male. All those
parts of the body which later develop from the subdivision
of the abnormal cell become male, and form an island or
islands of maleness in a predominantly female individual.
It was suggested by Wheeler that the other mosaics, such as
the queen-worker, might arise in an analogous way. This
would imply that the caste of each individual is genetically
determined in the egg.

There is, however, another type of sex mosaic, the inter-
sex, which may be described as a mosaic in time rather than
in space. In the gypsy moth, for instance, every cell of the
body produces substances which influence development in
the direction of either maleness or femaleness. If the physio-
logy of development is disturbed, there may at first be an
excess of the female-promoting substance but later of the
male-promoting one. The parts of the body which in develop-
ment reach their final form first will then be female, while
those which develop later will be male. The mosaics so pro-
duced tend to be of a type in which male characters chiefly

144

THE ANTS OR PISMIRES

appear in the head region and female characters in the more
posterior parts. Whiting has suggested that in ants the
mosaics are really intersexes or intercastes, and that the
second type of anomaly at least might be due to a change
in nursing behaviour during development. This is in fact
supported by Brian's observations. It cannot, therefore, be
claimed that the existence of caste-mosaics proves that the
castes have a genetic origin. To carry the argument further
it would be necessary to study the order in which the parts
of the ants' external skeleton are laid down, and to test
whether after such treatment as starvation the later developed
parts show any change in their caste character.

No clear picture can as yet be given of the type of here-
ditary mechanism which could be operative in the more
polymorphic type of ant. Some scheme, like the rather
theoretical one already mentioned for stingless bees, might
be imagined. There are two special features about the social
life of ants which might make some such theory more tenable.
The possibility of the differential feeding of grubs whose
caste the worker ants had recognised at the very beginning
of development, has already been mentioned. It is further
possible that some of the egg eating which is common in ants,
especially in young colonies, may also be differential. Thus,
only worker or queen eggs might be allowed to survive at
certain seasons, though both types of egg might be produced
at all times. This would explain why the supposed here-
ditary mechanism, which would be expected to produce all
castes in fixed proportions, does not in practice do so, the
castes being produced rather according to the needs of the
colony.

There is another process which operates, not only in the
ant colony but in any social animal. The unit whose
efficiency determines whether the species shall survive or

k 145

THE SOCIAL INSECTS

become extinct is the colony rather than the individual. An
individual which is useful to the colony may survive though
it would be quickly eliminated in a solitary species. This
happens to a considerable extent in man. In civilised societies,
many members are supported who contribute only very
indirectly to the provision of the food and shelter necessary
for life. Others who contribute nothing are enabled to
survive because our social behaviour benefits the whole
species and not merely the bread-winners. In an ant-colony
there is an analogous situation: the worker which is sterile,
or capable only of producing male offspring, is a good
example of a type which could not survive apart from the
colony. Some of the more fantastic types of soldier ants
seem to be an even more extreme example of the same thing.
They might be described as freaks for whom, during the
evolutionary process, a use has been found, just as the circus
has found a use for dwarfs.

COMMUNICATION

The social life of ants involves extensive co-operation
between individuals for the good of the whole colony. The
spinning together of leaves by Oecophylla is a striking
example, but a great deal of ant behaviour is of the same
type. To say that each individual is reacting instinctively
to the needs of the colony is a description rather than an
explanation of what is observed. Some ant behaviour might
be better understood if ants were known to have means of
communication of the type described in the honey bee.
Although, so far, nothing of this sort has been discovered,
this does not prove very much. Facts as strange as those
revealed by von Frisch for bees may yet come to light.
Meanwhile, the little that is known about communication and
the related topic of path finding may be briefly summarised.

146

THE ANTS OR PISMIRES

Some ants living on trees tap the leaves with their heads
as a danger signal. When a number of workers do this
simultaneously a clearly audible sound is produced ; but this
can hardly be more than a generalised signal, like the blowing
of a policeman's whistle. Other ants are able to produce
a slight sound by rubbing together the roughened surfaces
of two overlapping parts of the abdomen. This appears to
be used as a danger signal, but it seems unlikely that it
conveys any very precise information : some information
is also conveyed in the exchanges of food which are such a
regular feature of the ant-colony. A single worker bringing
in a crop full of Aphis-excretions shares it out with many of
the nurses inside the nest. A worker may also solicit another
worker for food, standing face to face and tapping or strok-
ing the other with her antennae. Clearly the donation of food
at least conveys the information that such food is available
at some point outside the nest. Conversely, the eagerness
with which food is demanded, depending on the state of the
grubs as well as on that of the adults, conveys to the foraging
workers information on the needs of the whole colony. But
it has not yet been possible to detect the communication of
precise information, either on the location of food or on
the needs of the colony. In recent years, however, a good
deal has been learnt of how ants find their way, and this bears
on the method by which a worker stimulated to start foraging
might successfully find a source of food already discovered
by other workers. After all, even the marking of a trail
involves a simple form of language.

TRAIL-MAKING AND PATH-FINDING

As might be expected, there are considerable differences
between different kinds of ants : many of them, though
rather versatile, tend to use one type of information more

i47

THE SOCIAL INSECTS

than another. Thus the shining jet black ant, Lasius fuli-
ginosus, mainly follows a scent-trail, like a blood-hound.
In this species well-beaten trails are formed, often visible to
the naked eye since most movable rubbish has been cleared
away. Down these trails rather dense, close columns of the
foragers may often be seen marching. On the other hand,
the common British black ant, Lasius niger, mostly uses its
eyes to see the way and has much more indefinite trails,
along which it usually goes singly. Most species, however,
can when necessary use senses other than the favourite one.
The black ant, for instance, certainly makes use of scent as
well as of sight.

One method by which ants can find their way is known as
the " sun compass reaction ", and was first studied in North
Africa by Santschi about forty years ago. When using the
sun as a clue to the direction, ants will continue to walk on a
path parallel to the one leading to the nest even if they are
picked up and put down some way off the track. They also
distinguish between the way to and the way from the nest.
This is probably accomplished by arranging that the sunlight
falls on one eye so that the ants walk at a constant angle to it.
In the short time necessary for making one journey the sun
does not move enough to make the method inaccurate. But
when an ant is confined for two or three hours in a dark box,
placed over the track, the sun will move through an appreci-
able angle. As a result, when the ant is liberated it takes a
course which makes about the same angle to what would
have been the correct course at the time it was first confined.

Carthy, in London, has recently studied the way-finding
of the jet black ant and the black ant in a small artificial arena.
The glass floor of the arena can be turned round independ-
ently of the circular sides. Some grubs are placed in the
centre and a worker is allowed to enter through a hole in the

148

THE ANTS OR PISMIRES

sides, find the grubs, and start carrying them back to the nest.
While the ant is at the centre, it can be confined in a small
thimble and the arena rotated. Some species, such as the
black ant, are not much disturbed by this rotation, but
usually set off in the right direction. They probably mainly
use their eyes to find the way. The jet black ant, on the
other hand, is usually nonplussed. It tends to follow the old
scent-trail, and this, after the rotation, no longer leads to the
door in the wall. Carthy was able to make the scent tracks
visible by dusting them with very fine Lycopodium powder.
The track appears to be formed by liquid excreta which is
ejected when the abdomen is dragged over the floor. In
another ant, Myrmica, Macgregor showed that the track
consisted of a discontinuous series of pear-shaped spots.

Some of these ants are also sensitive to the plane of polari-
sation of light, like the honey bee. The light by which the
ant finds its way across the arena can be polarised by passing
it through a sheet of " polaroid ". If then, before the return
journey, the sheet of polaroid is rotated through an angle,
the ant is much less certain of the correct way back. There
seems little doubt that with further experiments much more
will be discovered about how ants find their way to and from
the nest. In so far as all the ants in one nest use the same
tracks, this already has some social significance, but any
discovery of definite system of intercommunication would
have much more.

The special feature of the social life of ants, compared
with that of bees or wasps, is the much wider range in
behaviour and in type of social organisation. This is partly
because there are many more kinds of ants than there are of
the other social Hymenoptera. This in turn is to some
extent a reflexion of the long period over which they have

149

THE SOCIAL INSECTS

gradually perfected their organisation. Some members of the
earlier types of social species, such as the Ponerines, still
exist, and the most highly organised types have existed long
enough to have evolved in widely different directions. But
it does seem also that many species of ant are less rigidly
specialised in their behaviour than any wasp or bee. Many
of them are able to use a variety of foods, and their apparently
simple nests are really more adaptable to a varying environ-
ment. In this way the ants have become one of the dominant
groups of insects in all the warmer parts of the world.

150

SOCIAL PARASITES

The long history of insect evolution has largely been a
struggle for food and shelter. It has often evidently proved
easier to make use of the energy and labour of another
species than to be completely independent. Any habitat in
which many animals live will be attractive to a large set of
hangers-on which depend in various ways on the activities
of other species. The more diversified the fauna, the more
opportunities for still more species to behave as predators,
parasites or guests of various kinds. The nests of social
insects, providing excellent shelter and rich stores of food,
have proved particularly attractive to other species. More-
over, the nursing instincts which become so highly developed
in the most successful social species make them especially
liable to deception. The nests of ants and termites, and to a
less extent those of bees and wasps, provide a home for a
large variety of strange insects. Many of these are merely
scavengers and survive by being inconspicuous, by having
impenetrable armour or by some type of chemical defence,
such as a repellent glandular secretion.

ANT GUESTS

The most successful and remarkable of the guests have
developed special secretions to which the social insects are
powerfully attracted. In some instances the attraction may
be regarded as dangerous socially as drug addiction in man.

Mi

THE SOCIAL INSECTS

A number of beetles, for instance, have tufts of yellow hairs
which produce some aromatic substance to which ants are
strongly attracted. The ants lick the hairs and take as much
care of the beetles as they do of their own brood. Yet often
the beetles feed on the ant-grubs and are dangerous parasites
of the colony. The caterpillars of some of the blue butterflies
have a gland which produces a sweet secretion which ants
eagerly lick up. Some of the caterpillars are carried into the
nest by the ants and carefully tended. And yet the food of
the caterpillars is either the aphid ant-cows, or else the brood
of the ants themselves. Other insects, like the flies observed
by Farquharson in Africa, while not parasites of the colony,
have evolved a way of soliciting a meal from the foraging
workers so that they can obtain a share of the food intended
for the grubs.

SOCIAL PARASITES

All these insects are parasites of social insects, but are not
themselves social. But there is another group of parasites
whose behaviour is more like that of the cuckoo who lays
her eggs in the nests of other birds. These are the real social
parasites : species of wasps, bees and ants, mostly without
a worker caste of their own, which lay their eggs in the nest
of some other species of their own group. This behaviour
is evidently not so much a response to the direct struggle for
food and shelter as to the difficulties and dangers of founding
a new colony. Amongst the social wasps certain species both
of Vespula and of Polistes have become social parasites of this
type. The cuckoo Vespula, of which there are three in
western Europe and one, V. austriaca, in England, can be
recognised in the queen by slight differences in the shape of
various parts of the head. The differences are not very pro-
nounced, but have been supposed to be advantageous in

152

SOCIAL PARASITES

fighting. There is little sign of loss of the tools used in nest-
ing, such as occurs in cuckoo bees, though the shape of the
mandibles is thought to be less suited to the manufacture of
paper. The cuckoo wasps have no workers.

It is curious that V. austriaca is a short-headed wasp and
lives with a short-headed host, V. rufa, whereas the two other
cuckoo Vespula are long-headed wasps and live with long-
headed hosts. Yet all three cuckoo species show similar
changes in head-shape which must have been acquired inde-
pendently as a result of their mode of life.

Our knowledge of how the cuckoo wasps establish them-
selves in the nest of the host is still very fragmentary. The
most extensive observations are those of Weyrauch in
western Germany. The cuckoo queens come out of hiber-
nation a good deal later than their hosts, and attempt to enter
the nest at about the time that the first comb of queen cells
is being constructed. At some point, perhaps when the
cuckoo begins to lay, a big fight ensues with the workers,
many of which are found dead in or near nests containing a
cuckoo queen. The host queen may survive for a time, but
apparently is killed or dies before long. Eventually, the combs
produce some workers and a few males of the host, but no
young host queens. Most of the later brood consist of young
queens and males of the cuckoo wasp.

Amongst the wasps of the genus Polistes also, there are
three cuckoo species in western Europe. Their existence was
hardly suspected until it was first established by Weyrauch
in 1937. More detailed studies were published in Switzerland
by de Beaumont, in 1945. All the species have a characteristic
groove on the mandibles, due to a thickening of the upper
and lower margins. This, together with certain other features
of the head, are thought to be useful in fighting, but the
cuckoo Polistes are less different from the industrious species

M3

THE SOCIAL INSECTS

than are the cuckoo Vespula. It is not even certain whether
one of them may not produce a few workers, for in this group
it is difficult to separate queens from workers unless the
wasps are dissected.

The way in which the cuckoo queen establishes herself is
not known, but it seems that the host queen is somehow
eliminated, perhaps after a fight. The brood which is later
produced consists mainly of males and females of the cuckoo
species, with a few workers and males but no queens of the
host. Even the later host workers hatch only from cells
occupied before the cuckoo queen takes over. It seems
significant that in both these kinds of cuckoo wasp it is the
young queens of the host brood that are particularly elimin-
ated. This is not to the advantage of the parasite, since it
means that there will be fewer nests to invade in the following
year. It suggests rather that queens are, so to speak, " expen-
sive " for the colony to produce. It would then be difficult
either from the point of view of labour or of the supply of
some scarce food to produce queens of two kinds.

The cuckoo humble bees of the queens Psithyrus are more
different from their hosts than are the cuckoo wasps. The
name of these bees means "whisperer", because the female in
flying makes a much less pronounced hum than does a true
Bombus ; but the most striking difference in this sex is the
loss of the pollen-collecting apparatus. There is no pollen-
basket on the hind leg and instead of a shining, rather con-
cave surface there is a convex, densely hairy one. The sting
is more powerful, and the whole skeleton is stronger and
more thoroughly articulated against attack. The worker
caste is absent.

As in the ordinary humble bees, males and females are
produced in the autumn, and after fertilisation the females
hibernate in a shallow hole in a bank. When they come out

i54

SOCIAL PARASITES

rather later in the spring, instead of beginning a nest of their
own they search for a nest of a humble bee of the correct
species, since each kind of parasite is attached to one or a few
host species. The entry into a nest is a dangerous process,
and is often unsuccessful if the host colony has reached a
moderate size. The smaller the colony, the more easily it is
invaded. The resistance is due entirely to the workers, and
if there are enough of them the Psithyrus queen may be
killed. One or more Psithyrus may be found dead in the
tunnel leading to a humble-bee nest. If an entry is success-
fully accomplished, probably after a few of the workers have
been killed, the cuckoo queen becomes established in the
nest and leaves it no more. For some time she frequently
mauls workers when they meet inside the nest, usually with-
out killing them. Eventually an uneasy truce is established.
There has been disagreement about the relations between
the cuckoo humble bees and the host queens. Sladen, who
published in 19 12 a well-known work on British humble bees,
maintained the host queen was always killed. Hoffer in
Austria and Plath in the United States, however, have
numerous records of the two queens living side by side and
taking little notice of one another. Cumber also found a
nest in which both the Bombus and its parasite had been
laying eggs. Some of these discrepancies may arise from the
different behaviour of different species, but it seems almost
certain that Sladen exaggerated the antagonism between the
queens. Nevertheless, once the colony has been invaded the
Bombus never produces any young queens, very rarely any
males, and probably no more workers. This would be
quite understandable if the Bombus queen were killed, but
it is not clear how it happens when she is not. Possibly the
cuckoo queen eats the eggs; or it may be that, as described
below in some of the ants, the behaviour of the workers

i55

THE SOCIAL INSECTS

becomes perverted; the colony therefore produces a brood
of Psithyrus males and queens and then dies out rather earlier
than usual, since the host workers are not being replaced.

HOW CUCKOO BEES AND WASPS AROSE

It is possible to form some idea how this type of behaviour
may have arisen in bees and wasps, and this applies also to
the solitary kinds of these insects, amongst which there are
quite a number of cuckoo species. In both wasps (Vespula)
and in humble bees, suitable nesting sites are not very
numerous, and this seems often to set a limit to the abundance
of a species in any locality. As has been stated earlier, this
leads to a certain amount of fighting even between the
ordinary industrious species. Queens which have tried to
force their way into an already occupied site are quite
commonly found dead at the entrance. Sladen showed that
in two very similar British humble bees, B. terrestris and
B. lucorum, the former sometimes succeeds in invading young
nests of the latter and taking them over. Thus, after a short
period in which there are workers of both species present,
the B. lucorum workers die out because they are not being
replenished, and the colony becomes a pure one of B.
terrestris. This behaviour, though exceptional for these
species, is very similar to the " temporary social parasitism "
of certain ants, mentioned in the previous chapter. Such
successful invasions by normally industrious species seem
to be very rare in wasps, but Nixon in 1935 recorded one
example: a nest of Vespula vulgaris taken over by a queen
of V. germanica.

In 1927 I showed that cuckoo bees, and in particular the
species of Psithyrus, tend to have the colours and shorter
hairs which are often seen in southern races of industrious
species. Bischoff and Weyrauch have noted that cuckoo

156

SOCIAL PARASITES

wasps also have a " southern " appearance. It has been
suggested that if a southern species works its way into more
northern territory already occupied by some allied species,
the situation would be one in which cuckoo habits might well
arise. The more southern species would, as is known from
the observation of industrious species, tend to leave hiber-
nation later in the spring. The conditions for founding a
colony might well be more unfavourable than in southern
regions. At the same time, the northern species, having come
out earlier, would already have occupied many of the suit-
able nesting places.

This is in fact what seems to happen with Bombus terrestris
and B. lucorum. In Europe as a whole, terrestris has a more
southern distribution than lucorum and where, as in England,
they occur side by side, terrestris tends to come out of
hibernation later. This probably explains the occasional
invasion of nests of lucorum by terrestris, and it is possible
that over a long 'period (perhaps to be measured in millions
of years) such a conjunction of species might lead the
southern intruder to become a cuckoo species.

This hypothesis may account for the evolution of many of
the cuckoo wasps and bees, both social and solitary, in
Europe and North America. It can hardly account for the
origin of the numerous solitary cuckoo species in the tropics.
So far the only known social cuckoo species of bee or wasp
outside the temperate regions are two Psithyrus found in
eastern Asia. Possibly in tropical climates cuckoo habits
might tend to develop quite apart from climate whenever one
species invaded the territory occupied by one of similar
habits. However, much more observation of the tropical
species is required before such speculation can have much
value.

i57

THE SOCIAL INSECTS

PARASITIC ANTS

The story of the development of cuckoo habits in ants is
more complicated. It has happened in a much bigger range
of species, and may have arisen in more than one way.
Moreover,' all sorts of stages in the development of such
behaviour are known, ending up in workerless, degenerate
parasites.

One manifestation of such behaviour is the temporary
social parasitism of such species as the wood ant, Formica
rufa, mentioned in the previous chapter. A rather different
example is seen in the North African ant, Bothriomyrmex
decapitans. The queen of this species, after her marriage
flight, enters the nest of another ant, Tapinoma nigerrimum.
She does this by alighting near the nest and allowing herself
to be dragged in by the workers. She is subjected to some
rough treatment, but once inside the nest, she takes refuge
on either the head or the back of the much larger Tapinoma
queen. In either of these positions she seems to be safe from
the workers. After a time, she bites off the head of the
Tapinoma queen, but since by that time she has (presumably)
acquired the characteristic odour of the nest, she is readily
adopted as the new queen and her brood is reared. Gradually,
all the Tapinoma workers die off, and a pure colony of the
Bothriomyrmex remains.

Similar and, from a human point of view, even less pleasant
behaviour has been observed by Bruch in an Argentinian ant,
Labauchena. Several of the very small queens of this species
invade a nest of the fire ant, Solenopsis, and climb onto the
back of the much bigger queen. There they co-operate in
gnawing off her head, an act which may take them many
days.

Such methods of colony foundation are appropriate to

158

SOCIAL PARASITES

species with small queens which could not found colonies
unaided. They can scarcely be considered degenerate, though
they may have led to the slave-making habit of such species
as the blood-red ant, Formica sanguinea. When the young
queen of this ant founds a new colony, she invades a small
nest, usually of Formica fusca, and kills the queen and any
of the workers which resist or attack her. Before long, all
the adults of the host ant will be dead and the sanguinea
queen tends the remaining brood. When new young workers
of fusca emerge they look after the brood which hatched from
the eggs laid by the sanguinea queen. In some regions, and
in some varieties of the blood-red ant, there is no further
development of parasitism, and as the fusca workers die out,
the colony becomes pure sanguinea. This behaviour is no
more than a form of temporary social parasitism in which the
relations between the host and the invader are even less
cordial than usual.

More commonly, particularly in western Europe, including
Britain, the later history of the colony is different, since the
population of slave workers is maintained by a regular series
of raids on other nests. These are carried out by sanguinea
workers which leave the nest in a large straggling group
which may cover several yards. Sometimes several smaller
groups follow one another at short intervals. When this
happens the first to arrive surround the nest of the prospective
victims. When the whole group of raiders has assembled
they enter the nest and begin seizing grubs and cocoons.
Workers which resist are killed, but the others are not
molested. Eventually, each sanguinea worker returns, carry-
ing a grub or cocoon.

This remarkable behaviour attracted the attention of
Darwin, and in The Origin of Species he describes a number of
his own observations on this ant. Darwin was puzzled how

i59

THE SOCIAL INSECTS

the habit of raiding could be transmitted from generation to
generation if, as he supposed, it was manifested only by the
sterile workers. Much later, however, it was discovered that
the behaviour of the queen in founding new colonies is not
so very different from that of the workers. Moreover, as
Darwin remarks, the queens whose worker offspring be-
haved in a way which made the colony as a whole most
successful would be the ones which leave most offspring.
That is, the process of natural selection involves evolutionary
change in the whole colony, rather than in single individuals.
The sterility of the workers from this point of view is
immaterial.

Darwin also suggested that the raids are partly for obtain-
ing food, like the raids of driver ants, and that the slaves
hatch from grubs or cocoons which fail to get eaten. There
is little doubt that this theory is partly true. Sanguinea does
eat a varying proportion of the grubs and cocoons which it
brings home, and does sometimes raid nests of species which
it never uses as slaves. Indeed, some other species, such as the
wood ant, whose queens found colonies as temporary social
parasites, do also raid other ants' nests for food, at least
occasionally.

It may be that the slave-making habits of sanguinea have
developed from two directions; partly from the way in which
the queen founds new colonies and partly from food raids
undertaken by the workers. One remarkable feature of the
raids is that they may be made on a nest as distant as ioo
yards from the sanguinea nest, and that on the way to the
chosen nest the ants may pass and ignore other nests of the
slave-providing species. This suggests that scouts have
previously reconnoitred the way. The ants are not, however,
led by the scouts, since the position at the head of the raiders
is occupied by new individuals at frequent intervals.

160

SOCIAL PARASITES

Although most of the nursing in the sanguinea nest is
done by the slaves, at least in those colonies which have them,
the species is not at all degenerate. The workers are fully-
capable of running the whole colony if necessary ; they are
extremely active and, judging by their behaviour on raids,
very well endowed. Nevertheless, it is probable that the
behaviour of some of the more degenerate parasites described
in the next few paragraphs developed out of slave-raiding.
Once the ability to run the colony independently has been
lost it is evidently very difficult to regain it.

Heads of some wprkers of ants and of their slavemakers or para-
sites. A, the slave-maker, Formica sanguinea; B, its slave, Formica
fusca; C, the amazon ant, Polyergus rufescens; D, Strongylog-
nathus, parasite; E, Tetramorium, host. The mandibles of types C
and D are suitable only for offence; such ants have to be fed by
their slaves or hosts.

This is seen in the more specialised slave-keeping ants
known as the amazons, the various species of Polyergus.
Some of these are found in western Europe (France, Switzer-
land), others all the way across central Asia to Japan and
again in North America.

The mandibles of the amazons are sickle-shaped and
pointed, without the broad, finely-toothed edge of those of
an ordinary ant. These jaws are unsuited for anything but
fighting, and the amazons are incapable of making a nest or
of looking after their own young. The tongue-like organ of
the mouthparts is also short, and they are almost incapable

l 161

THE SOCIAL INSECTS

of feeding themselves. At the most they can take a little
liquid food if they happen to find a conveniently placed drop.
The amazons are thus clearly degenerate, and can live for
only a short time in the absence of their slaves. There are
few things more remarkable in an amazon colony than the
sight of large groups of workers, much bigger and more
formidable than their slaves, spending the whole day doing
nothing but occasionally cleaning themselves or soliciting the
slaves for food.

The small slaves, which are usually workers of Formica
fusca, construct and enlarge the nest, forage for food, feed
the amazons and tend all the brood. This is in marked
contrast to Formica sanguinea, whose slaves never leave the
nest but only tend the brood.

During the height of the summer the amazons make a
number of slave-raids. When engaged in this way their whole
behaviour alters. There is a period of intense activity when
the amazons gather round the mouth of the nest, and then
they all start off in a rather compact army. Normally they
go straight to some nest of the slave species, and it seems that,
as in Formica sanguinea, the way must be surveyed in pre-
paration, though the army is not led by any particular
individual. When they arrive at the slave nest they pour in
through the entrance and begin to come out with grubs or
cocoons. Adult ants are not attacked unless they resist,
when they are killed instantly by being pierced with the
sickle-like mandibles. The return with prey is more leisurely
and not made in a compact group. Much of the prey serves
later as food for the colony, but some of the cocoons are kept
to produce new slaves.

When the amazon queen founds a new colony, after her
marriage flight, she enters a colony of Formica fusca or one
of the similar species. She kills the queen but does not molest

162

SOCIAL PARASITES

the workers ; these will later feed her and tend her brood.
The workers which she has taken over are afterwards
replaced by new slaves obtained by the raids of amazon
workers, so that the colony never becomes pure for one
species as in a temporary social parasite.

PARASITISM AND DEGENERATION

There are a great many more kinds of parasitic ants, and
they all have sickle-shaped mandibles and are in various
degrees incapable of maintaining the colony unaided. A few
seem to be slave-keepers rather after the pattern of the
amazons, with pugnacious workers. But many are more
degenerate and occupy a relatively subordinate position in
the host colony both as regards numbers and behaviour.
They have, in fact, become much more like true parasites,
and their chief effect in this case is to prevent normal repro-
duction by the host species. All the more degenerate
parasitic ants have lost the worker caste. Many of them
are found in only a few localities and are rarely seen. It
appears that such behaviour is not ultimately a reliable way
of perpetuating the species. Moreover, once a species has
reached the stage realised in the amazons, there is no
return.

In Great Britain there are three species of ant which are
permanent social parasites and which, though not closely
related to one another, are examples of three stages in degen-
eration. Formicoxenus nitidulus lives exclusively in the nests
of the very much larger wood ant, Formica rufa, and some
other very similar species. It excayates small chambers in the
recesses of the nest and looks after its own brood. Owing to
its small size the galleries it uses are inaccessible to the workers
of its much larger host, but it is able to wander freely through-
out the Formica nest. Its natural food is unknown, but it will

163

THE SOCIAL INSECTS

eat honey or ant grubs in observation nests. It does not seem
to attack the Formica brood, and the two sorts of ants take
very little notice of one another. It is really dependent on the
host only for shelter and protection, and is thus hardly a para-
site. Structurally, it shows few signs of degeneration, since
workers are numerous and have normal broad mandibles.
The male, however, is wingless and very worker-like in
appearance. There can therefore be no marriage flight, and
mating takes place on top of the Formica nest.

The virgin queens are winged, and presumably the majority
of them fly off to found new colonies, after mating with their
wingless brothers. New colonies are founded in small cells
excavated on the edge of a Formica nest. The queen and
her young brood must move into the interior at a later stage.
In full established colonies, if the Formica moves its nest to
a new site the Formicoxenus follow them, carrying their own
brood.

The various species of Strongylognathus are essentially
slave-makers, more dependent than the blood-red ant but
somewhat less so than the amazons. They have the charac-
teristic sickle-shaped mandibles of all the true parasites. One
species, S. huberi, which is found in Switzerland, is known
to make raids to obtain the brood of the ant Tetramorium
caespitum. The raids are apparently made at night when the
Tetramorium are less active and less resistant to aggression.
The case is unusual in that some Tetramorium slaves may
take part in the raids, but few observations on the subject
have so far been made. While the Tetramorium slaves are
necessary for rearing the brood, the Strongylognathus may
do some of the work of nest-construction. Another species,
S. testaceus, is quite common in parts of western Europe and
is more degenerate. The worker caste is never numerous
and seems to be on the way to disappearance. It does not

164

SOCIAL PARASITES

make slave-raids, but the young queen invades a Tetra-
morium nest and lives there without killing the host queen.
Both queens continue to lay eggs, but those of the host queen
produce thereafter nothing but workers, whereas the brood
of the parasite consists of all three castes but especially of
males and queens. The species is thus harmful to its host —
since it prevents normal reproduction — and degenerate,
because it is unable to rear its own brood. This is done for
it by the Tetramorium workers. In 1935 a new species of
Strongylognathus, S. diveri, was discovered by Diver, in
Dorset. It is very similar to S. testaceus, and probably has
similar habits. It certainly lives in the nests of Tetramorium,
and it has few and rather feeble workers.

The third British parasitic ant is Anergates atratulus. It
too lives in the nests of Tetramorium. In England it has been
found only in the New Forest, and there only on very few
occasions. It is also rare on the continent. It is an example of a
group of rare species, each belonging to a different genus,
and probably showing the last stage in a long parasitic history.
Such forms are rare, it is thought, because they have passed
the point where parasitism is even a temporary advantage
and are now approaching extinction.

Anergates has no workers. The male is wingless, of a pale
colour, and with a rather soft skeleton, more like a pupa than
an adult insect. The virgin queens are relatively small ;
after mating with their brothers in the colony which pro-
duced them, they fly off to try to invade new colonies. If
successful, they kill the host queen and begin laying eggs at
a great rate, their abdomen swelling to an enormous size.
The host workers care for the brood, but the whole colony
must die out rather quickly since the workers are no longer
being replaced. Although the species can live only in places
where Tetramorium colonies are numerous, it tends to reduce

165

THE SOCIAL INSECTS

their number by its parasitic action. The life of the species
is therefore rather precarious.

SUMMING UP PARASITISM

The story of the parasitic wasps, bees and ants which
avoid some of the difficulties of colony foundation by their
dependence on other species is suggestive in several ways.
There is of course no need to draw the moral that — even in
organic evolution — crime does not pay. Many parasites,
both amongst the insects and in other animal groups, are
highly successful and show no signs of dying out. Some
dependence on other organisms is universal amongst animals,
for all depend ultimately on the green plants. A successful
parasite, however, is one which does least damage to its
host, which allows if not the individual at least the species
to flourish. Apart from Formicoxenus, all the types dealt
with in this chapter prevent their host from producing sexual
forms. This by itself might not be a fatal objection to their
mode of life, from the evolutionary point of view. Yet once
the parasitic habit has begun, there seems to be a general
tendency for the worker caste to be lost.

This loss can be looked at in two ways. First, in any
species, varieties lacking structures or types of behaviour
which contribute towards successful reproduction and sur-
vival will tend to be eliminated. In most social insects, the
presence of many workers is essential to successful repro-
duction in the colony ; any tendency to reduce the number
of workers would be suppressed. But in a parasitic species
whose brood may be reared by the host workers, the workers
of the parasite are no longer essential to its survival. A
situation therefore exists in which the suppression of the
worker caste will not be an immediate disadvantage.

Secondly, while we are uncertain what determines whether

166

SOCIAL PARASITES

a particular egg produces a queen or a worker, it seems that
the balance between the two possible types of development
may be quite easily tipped one way or the other. It may well
be that the original balance of the colony is upset by the
presence of the parasite in such a way as to suppress queen-
production in the host. It has been suggested, for instance,
that in Strongylognathus it requires less effort to rear the
small sexual forms of the parasite than the much larger ones
of the host, Tetramorium. This might lead both to the
suppression of the sexual forms of the host and to the con-
version of the majority of the parasite brood into sexual forms
rather than workers. This argument, of course, raises again
the question whether castes are determined by feeding or
genetically.

Probably in the more degenerate parasites such as Aner-
gates the power to produce a worker caste has been com-
pletely lost as the result of a long process of selection. Once
the workers have lost almost all their functions it is more
advantageous to have all the females fully fertile. It is possible
that Strongylognathus provides an example of what has been
called social regulation, a process (as we shall see) much
better known in the termites. This is the method by which
different types of individuals and of developmental stages are
produced in the proportions best suited to the needs of the
colony. It arises from the natural interplay of the normal
behaviour of the species with its environment. In Anergates,
social regulation has been replaced by an innate physiological
process.

Social parasites are not very numerous, perhaps because
species which have evolved along these lines are rarely
successful for long. Amongst the eighty-eight social species
in Great Britain, eleven are parasitic, if one counts the slave-
making ant. According to our present knowledge, the

167

THE SOCIAL INSECTS

proportion is lower in tropical regions, and it is possible
that the difficulties of colony foundation in a temperate
climate have favoured the development of such parasites.

It is a curious point that slavery and parasitism of this type
do not seem to occur in the termites. A few cases of com-
pound nests, rather like those of Formicoxenus and Formica,
are the nearest approach to it. It is not clear why termites
have not shown this type of evolution. It may be partly
because in them the queen is always attended by a male and
mating has often to be repeated. The invasion of a strange
nest by the two sexes together might be more difficult.
Another possibility is that it would be harder to eliminate the
reproduction of the host species, since when the original
queen is killed alternative queens can readily be produced.
A similar argument might be applied to the honey bee and
its allies, since these are not subject to attacks by cuckoo
species. Such suggestions, however, are at the best very
tentative, and there is still much in the evolution of the
parasites which we do not understand.

1 68

THE TERMITES

The termites, sometimes called " white ants ", are the least
familiar of the social insects to Europeans. Although one
or two species are found in France and Italy and fifty-five
in the United States, the group is essentially tropical, and
even in warmer lands the insects are usually inconspicuous
apart from some of their large nests. There are some remark-
able analogies with other social insects, and since the majority
of individuals are wingless, it is with the ants that they would
naturally be compared. Yet in reality the differences between
the termites and all other social insects are profound and
more important than the resemblances. It is almost true to
say that what they have in common is little more than the
minimum which allows us to apply the one word " social "
to both groups.

The peculiarities of the termites are traceable to two basic
facts about their societies. First, males and females are
represented at all stages and in all castes, and this is associated
with a different method of sex-determination. Secondly,
development is gradual, over a series of moults, with no help-
less grub-stage. At least after a few moults, each individual
is a potential worker and its bodily form and functions can
partly be determined by the number of moults it has experi-
enced. Thus the whole caste system is much more flexible
and complicated. In the ants, apart from the exceptional

169

THE SOCIAL INSECTS

forms which use the grub to spin silk, the whole work of
the colony is done by adults, and by definition the form
and behaviour of these is in the main fixed for life. Though
there may be variation in fertility and changes in the size
of the abdomen, as in the queen legionary ant or the worker
honey-pot ant, the form of most individuals is constant and
none can alter parts of the body other than the abdomen.
In the termites, on the other hand, the individual is usually
capable of changing both its appearance and its functions
over a considerable period after it has begun to play an
important part in the life of the colony. This introduces the
possibility of " social regulation ", the moulding of the
individual to the needs of the community at any particular
moment.

NESTS

Termite nests often form a conspicuous feature of the
landscape in tropical regions. They vary much in shape and
structure, even within one group. They are built of earth,
of rotten wood, or of the insects' own excrement, or of a
mixture of these. Often a sort of cement is made of saliva
and clay, and this may set extremely hard. One of the diffi-
culties of preparing the site at Kongwa in East Africa for
planting ground nuts was the problem of removing the
numerous large termite nests. Often a nest required a large
charge of dynamite.

Termites, apart from a few exceptional species, avoid all
contact with light and with the free air. Their nests are
therefore always closed. When the winged forms wish to
leave, or in species which send foraging columns into the
open, temporary exit-holes are made, but they are soon
closed again. A few species are known to carry their excre-
ment outside the nest and these also make temporary exits

170

THE TERMITES

for the purpose. More commonly, if the excrement is not
used in building, it is piled in certain chambers in the peri-
phery of the nest.

The most conspicuous type of nest is a tall, steeple-like
structure which may be 10-20 feet high and up to 50 feet
in circumference at the base. The outer walls of these nests
are as hard as concrete. Inside, the nests are usually divided
up into a number of chambers of varying size and shape,
connected by passages. The more central of these chambers
are inhabited by the royal couple — king and queen — by the
young brood and by various stores of food. Such nests may
also extend several feet underground. In other species there
is a central chamber suspended inside the much larger outer
part of the nest, almost like a wasps' nest hanging inside a
hollow tree.

Some of the large Australian nests have the steeple-like
structure very flattened, so that it is almost tongue-shaped.
These are known as " compass termites " because the two
broader faces always point east and west. Although very
useful to the bushrangers, the value of this arrangement to
the termites is not known.

Another type of conspicuous nest is built On trees. These
may be of irregular shape and placed flat against the trunk
or in a fork of the branches. Others are spherical and are
built all round thin branches. These may be conspicuous
objects in tropical rain forest and have been known as
" niggers' heads ". The nests on tree trunks are sometimes
surmounted by a number of V-shaped ridges built out from
the trunk which seem to protect the nest from water running
down the trunk. Another construction which has the same
effect is a series of overlapping flaps, shaped almost like a
hand with fingers and projecting all round one side of the
tree-trunk. Nests with such protections against rain are built

171

THE SOCIAL INSECTS

by various species in the damp forests of Africa and South
America.

The nests so far described are the conspicuous ones, but
many kinds, probably the majority, are hidden. Some,
especially perhaps those in drier regions, are entirely subter-
ranean. These may be merely a series of irregular intercon-
nected underground chambers or they may have a much more
definite structure, an enclosed, walled-in underground space
within which the actual chambers are constructed. Some of
the desert-living species make vertical tunnels many feet
long to reach the deep-lying water table so that they can
bring up water to keep the nest suitably damp.

The two groups of termites which have been most in-
tensively studied (because they happen to be represented in
Europe and North America) make very simple nests in
decayed wood. Irregular branching tunnels are excavated,
sometimes widening into chambers ; but there is no special
provision for the royal couple or the brood : these merely
occupy the more central part of the colony.

THE NUPTIALS

Just as in ants, new colonies of termites normally arise
from flights of winged sexual forms. Commonly, these
forms are produced only during a short season when they
may appear in millions. In parts of Africa, when camping in
the bush, it may be difficult to eat one's supper by lamp-
light because of the immense numbers of winged termites
which are attracted to the light and fall into the food. Other
species, however, fly by day.

Some time before the marriage flight the workers construct
special passages to the outside, each wide enough to allow
several winged adults to leave simultaneously. One species
is known to construct special waiting-rooms for the winged

172

THE TERMITES

forms to stand in, near the exit-holes. At the moment of
departure some of the workers and soldiers, which for some
time have shown signs of excitement, pass out of the exit-
holes and form a ring round them, outside the nest. The
winged forms then leave and fly off over a period of two or
two-and-a-half hours ; thereafter the holes are stopped up
again.

Few termites are capable of powerful flight. Usually they
travel less than ioo yards, only occasionally as much as a
mile. Mating does not usually take place in the air, though
in Pseudacanthotermes the male seizes the female in flight,
losing his wings at the same moment. The female flies on
with the male and loses her wings only on landing. In most
species, however, both sexes get rid of their wings on landing,
pressing them against some plant stems so that they snap off
at the preformed line of weakness.

The wingless adults are still very excited and dash about
in all directions on the ground. If a male and female meet,
they stand face to face, tapping one another with their mouth-
parts and antennae. If the male is accepted, the female
turns round and walks off, followed by the male, and they
undertake what is known as " the nuptial promenade ". This
is essentially a search by the female, often lasting several
hours, for a suitable site for a new nest. This will very often
be found at the side of a fallen log, or against a post or tree,
preferably where conditions are a little damp.

The male and female together dig a tunnel, later enlarged
into a small chamber. The entrance is gradually blocked by
debris from inside, the complete work taking a day or two.
When they have finished they become inactive, and at this
moment, for some obscure reason, they each remove the last
few segments of their antennae. Mating occurs for the first
time either at once or some time up to a fortnight later,

i73

THE SOCIAL INSECTS

according to the species. It can be well imagined that flights
of winged termites cause great excitement in all the beasts
of the field and birds of the air, and that enormous numbers
of the termites are eaten before some are able to dig their
nuptial chamber. In South America nests of social wasps are
often found to be stocked with winged termites, and some
South American Indians are fond of eating them.

Fig. 12

Scene in the royal chamber of the African Termes bellicosus

showing the king, queen and attendant soldiers and workers

(after Escherich).

THE COLONY AND ITS FOOD AND AGRICULTURE

In the less specialised termites the young king and queen
eat rotten wood during the early part of colony-foundation ;
from it they elaborate food which they supply to the first
stages of their brood. In them, too, the brood quite soon are
themselves able to feed on wood and later to take over the
supply of food for the royal couple, who eventually no longer
feed themselves at all. In the most specialised termites the
nuptial chamber is often made in the soil where there is no
external source of food. The first brood is brought up
entirely by the royal couple, who rely on what was in their
intestines at the time of flight, on the enormous fat-reserves

174

THE TERMITES

which have been built up inside their bodies, and on material
derived from the degenerating wing-muscles. In these
termites, too, the brood do not become self-supporting until
they are much bigger. In some species it is perhaps only the
fully grown workers which actually obtain new food, and
all the other inhabitants have to be fed by them.

Termites are pre-eminently eaters of wood; to a less
extent they eat other vegetable matter. The less specialised
types are almost exclusively wood-feeders, and these also
nest in timber. More highly organised forms make use of
other supplies, such as dead vegetable matter in general. A
few collect grass and other green leaves which they cut like
leaf-cutter ants. Others collect grass seeds which they store
in underground chambers after the manner of harvesting
ants.

Many, especially those which make the large conspicuous
nests, store varied vegetable material in underground chambers.
These collections, especially those of thoroughly chewed
wood, become impregnated with fungus just as happens in
leaf-cutting ants, and some of the fungi are peculiar to this
situation. The termites feed on the fungus, but it does not
seem that they are so closely adapted to the diet as the ants
are. Almost certainly it is one of the ways in which they can
supplement their diet with proteins or with special growth-
promoting substances. Termites also eat all the moult skins
of the developing brood, having a digestive juice which can
dissolve insect skeletons. Any wounded or damaged indi-
vidual is quickly eaten by the others. A few species are
known to collect other termites anAto store them in chambers
just as they do vegetable materials. Like the queens of other
social insects, the termite royal couple may eat an appreciable
number of their own eggs, even when there seems to be no
shortage of food.

i75

THE SOCIAL INSECTS

The termites which are more especially wood feeders have
a remarkable digestive specialisation. Their intestine con-
tains vast numbers of protozoa belonging to a special group
which is found only in this situation. These microscopic
unicellular animals are able to digest the cellulose which
forms the major part of all woody tissues. They convert the
cellulose into sugars or into other substances which the
termites can assimilate. Although wood might provide a
complete diet, it would be an extremely wasteful one if
none of the cellulose could be absorbed. The relation be-
tween termites and their intestinal fauna is a typical example
of a symbiosis : the protozoa receive chewed wood in a
suitable environment, while the termites receive back digested
cellulose and also assimilate a fair proportion of the protozoa
themselves.

The young brood quickly become infected with the
protozoa by eating the faeces of older members of the colony,
and the royal couple carry some infected material with them
when they go on their marriage flight. The process has an
interesting analogy with the digestion of such animals as
the cow, in which the rumen, a special compartment of the
intestines, contains vast numbers of bacteria which help to
digest grass or hay.

About twenty-five years ago, Cleveland showed that the
protozoa in the termite intestine could be eliminated by
exposing the insects to a raised pressure of oxygen. After
this has been done the termites can no longer assimilate
cellulose, and, if given no food but purified paper, die rather
rapidly. Conversely, normal untreated termites can subsist
on such a diet for a long time. Those termite species which
seem to have a more varied natural diet lack the intestinal
fauna ; exactly what fraction of their diet they assimilate and
how they do it is unknown.

176

THE TERMITES

Much of the food which termites eat does not come
directly from outside the nest but has already been partly
assimilated or even secreted by other members of the colony.
As already noted, in some species it is only the workers
which feed on unprepared foods, such as wood-fragments.
Treated food is of three kinds. Thoroughly chewed and
probably partially digested material may be received from
the mouth. This source also provides saliva, which is the
richest and most thoroughly prepared food. Finally, termites
can pass two types of faeces ; one is relatively dry and is
thoroughly digested, waste material. It is thrown away or
built into the nest. The other consists of much more liquid
material which is really food which has passed through the
intestine without having all nutritive substances extracted
from it. It contains large numbers of the intestinal protozoa
in those termites which are infected. Generally, in a well-
established colony, the youngest brood are given only
prepared food, and soldiers are incapable of feeding them-
selves. The royal couple, particularly the queen, are fed
continuously on saliva. At the other end of the queen"
an almost continuous stream of eggs and also of certain
liquid excretions is continuously removed by bands of
workers.

Termites are continually exchanging food with one another
and also licking one another's bodies, probably to ensure
cleanliness. It is supposed that the mutual interest which
such behaviour engenders is an important link in the social
organisation.

TERMITES AS PESTS

The wood-feeding habits of termites are the cause of very
serious losses, chiefly in tropical countries. In Queensland
Mastotermes makes it impossible to construct sheep-pens

m 177

THE SOCIAL INSECTS

with wooden posts. In many tropical countries the wooden
parts of houses have to be protected, especially where they
come in contact with the earth. Even the use of concrete
pillars to support the ground floor is not infallible, since the
termites will build covered runways up the concrete. The
damage is particularly insidious since it is entirely internal :
the first sign of trouble is apt to be the collapse of a beam.
During the Second World War crates of military stores which
had to be piled on the ground in such countries as New
Guinea were often attacked. A few weeks later, when the
crates were moved, they fell to pieces.

During his travels at the end of the eighteenth century,
Humboldt noted how rare it was to find old books in South
America, and this was attributed to the attacks of termites.
In such countries, to-day, all books are treated with a solution
of a poison such as corrosive sublimate, to protect them
against termites and other insects. The protection of houses
is more difficult, but special modes of construction, employ-
ing concrete, have been devised, and there are expensive ways
of impregnating wood with preservatives.

A curious but less important source of trouble occurs when
runways are constructed for African aerodromes. It may be
very expensive to level out the hard, projecting termites'
nests and the strips may require repeated treatment because
the nests are continually being reconstructed.

THE IMPORTANCE OF TERMITES TO TROPICAL PLANT-LIFE

Termites play an important part in the economy of tropical
nature in other ways which impinge on human interests,
though less directly. Trapnell has shown that where they
are really numerous, as in parts of Rhodesia, they greatly
influence the type of vegetation. Just as in England the
vegetation of chalk downs differs from that of sand or clay,

178

THE TERMITES

so in Africa termite country may have a characteristic
flora. This will indirectly affect all the other animals, and
will also determine what crops can be most successfully
grown.

The natives are well aware that the soil formed from broken
termites' nests is unusually fertile. To some extent these
insects do the work which in temperate countries is done
by earthworms. Worms drag dead leaves underground
and eat them, and incorporate dead vegetable matter into
the humus in a form which is suitable for plant growth.
Termites do much the same thing, and may be very im-
portant in dry climates in which ordinary decay would be
rather slow.

CASTES

So far the different individuals of which the colony is
made up have been only incidentally mentioned. The subject
is much more complicated than in other social insects, but
the nature of the castes and the factors determining their
development lie at the root of their whole social organisation.
There is a larger number of castes than in other social insects
and also a bigger difference between the simpler and most
advanced types.

The eggs are laid singly except in the primitive Australian
genus Mastotermes ; in the latter they are laid in groups of
sixteen to twenty-four, stuck together in two rows. The
appearance of this egg-group is rather like the similar groups
laid by cockroaches. The eggs develop slowly, and even
at quite high temperatures (20-25 ° Q may take three to
twelve weeks to hatch, according to species and circumstances.
In the less specialised termites the rate of egg-laying is not
very high, perhaps two hundred to three hundred a year,
but in the higher forms extraordinary figures have been

179

THE SOCIAL INSECTS

recorded. Emerson kept queens of some of the South
American species and actually observed that 1,600-7,500
eggs were laid in twenty-four hours. One of the species
which builds the large steeple-like nests in Africa has been
seen to lay 36,000 eggs in twenty-four hours ; this corre-
sponds to thirteen million a year.

The method of sex-determination is different from that
of ants, bees and wasps. If the developing sexual cells are
stained and examined under a microscope, it is found that
those of the female contain an even number of heavily stained,
rod-like bodies or chromosomes. The corresponding cells in
the male have one less chromosome. Thus as far as sex is
concerned, the hereditary constitution of the queen termite
may be represented by the symbols AAXX and of the male
by AAX. The unfertilised eggs all have the constitution
AX, but the sperm may be A or AX. Thus when the eggs
are fertilised, females (AAXX) and males (AAX) are pro-
duced in equal numbers.

Termites are one of the groups of insects in which develop-
ment up to the adult stage is gradual. This means that there
is no grub, nor an immobile pupa enclosed in a cocoon.
The individual which hatches from the egg is like the adult
except that it is small, has no traces of wings, and is sexually
immature.

It is convenient to use the French terminology for refer-
ring to the immature stages, though it is not universally
accepted. The form which hatches from the egg is the larva,
and it continues to be one in spite of moulting, as long as it
has no traces of wings. The nymph resembles the larva but
has visible wing-pads. To produce a fully winged adult
from the nymph will require several moults, at each of which
the wing-pads will get a little longer.

The soldier is typically without wing-pads, is relatively

180

THE TERMITES

large, is at least partly pigmented, has the eyes small or
absent, and has a modified head. The most usual type of
soldier has large mandibles, rather like the typical ant soldier,
but there is another type in some termites known as the
nasute. This type of soldier has reduced jaws, but has the
top of the head produced into a long snout, at the tip of
which is an orifice. This is the opening of a gland which
produces a copious, sticky, often white, secretion. They can
use this to gum up their enemies, principally ants. While
there is some doubt whether, in certain species, the soldiers
are much use in defending the colony, this does seem to be
their function in most. It is very striking in the species
which make daylight forays above ground that the immense
columns of workers are flanked on each side by rows of
soldiers ; this strongly supports the idea that their function
is protective.

Soldiers are incapable of feeding themselves and do none
of the work of the colony. They might be compared to
the workers of the amazon ants except that they do not
make slave-raids. In some of the more specialised termites
there are two sorts of soldiers, quite distinct in size, but
mandibulate and nasute soldiers never occur in the same
species.

In the less specialised termites there is no definite worker
caste and the main work of the colony is performed by the
nymphs and the older larvae. There may, however, be a
group known as false workers whose form is semi-stabilised
though under certain conditions they may recommence
growing and moulting. If they do so, they become either
larger false workers or else the substitute sexual forms
described in the next paragraph. In the specialised termites, a
true worker caste is found and may, as in the soldiers, be of
two distinct sizes. Members of this caste resemble large

181

THE SOCIAL INSECTS

larvae, having no vestiges of wings, and no or only minute
eyes, but they have a fixed form and are no longer capable
of growth and moulting.

The substitute sexual forms are individuals capable of
reproduction, though not winged and not leaving the nest
for a marriage flight. They appear whenever one or both of
the royal couple die or is removed from the colony. In
bodily form they resemble larvae or nymphs, depending
on their mode of production, but the wing pads are often
smaller than in the average nymphs. Once a termite has
become a soldier, it can no longer become a substitute sexual
form, and the same holds for the true workers of the higher
termites.

Grasse and his colleagues in France have obtained the
conclusive experimental result in some of the lower termites
that an artificial group of, say, twenty or thirty identical
third stage larvae can differentiate in isolation into a colony
in which all castes are represented. This self-differentiation
of a group is the fundamental fact in caste-production.
The higher termites, with their more elaborate caste system,
have been much less studied, and little can be said about
them with certainty. There is, however, no reason at the
moment to suppose that their development is essentially
different.

Another element of great importance in the flexible caste
system is the influence upon it of the age and size of the
colony. In general, the younger and smaller the colony, the
fewer moults are necessary to produce the adults of any caste.
In Zootermopsis, for instance, a soldier will be produced
after five, six, seven or eight moults according as the colony
consists of about twenty-five, fifty, one hundred or five
hundred individuals, corresponding to ages of one, two, two
to three, and four to five years, respectively. In such termites

182

THE TERMITES

the winged sexual adults are not produced until the colony-
is several years old. This last point is another example of the
natural balance which is maintained between the members
of the various castes. Artificial removal of the royal couple
results in the production of substitute sexual forms.

An interesting experiment has been performed by Castle
and others with the soldiers. In the normal course, in a young
colony only one larva will be transformed into a soldier.
But if the single soldier is removed, another larva or nymph
will quite soon be transformed. Moreover, the appearance of
the first soldier can be suppressed by introducing one from
another nest into the young colony. These results suggest
that the presence of sexual forms or of soldiers inhibits the
conversion of further nymphs into these castes, at least until
the nymphs become very numerous, when the balance may
be restored by the production of more individuals of the
required caste.

In another curious experiment Grasse and Noirot made up
isolated pairs of fourth stage nymphs of Calotermes. They
found that if the pair was one of each sex both became
sexually mature after the next moult. But if the two were of
the same sex, only one became sexually mature — another
example of balanced production of the required castes. In
normal colonies of Calotermes, if the royal couple is removed,
substitute sexual forms are produced, but only one pair
becomes the royal couple and the others are destroyed by
the workers. In some other termites, however, a number
of substitute sexual forms may reproduce simultaneously,
perhaps because their rate of eggrlaying is lower than that
of the original queen.

A final example of social regulation is provided by an
experiment by Miller on the American Prorhinotermes. He
put a group of 545 similar nymphs into an artificial nest

183

THE SOCIAL INSECTS

and after several moults the following individuals were
produced :

Large workers Soldiers Substitute Stayed as nymphs

sexual forms or died

103 37 82 323

This result illustrates another characteristic of termite
development : that moulting can lead either to progressive
or to retrogressive development. Thus the large workers
and soldiers have actually lost the wing-pads which were
present in the nymphs, whereas the heads and mandibles of
the soldiers and the reproductive organs of the sexual forms
show a positive change. The possibility, within limits, of
developing either forwards or backwards gives the caste
system great flexibility.

THEORIES OF CASTE FORMATION

It has been mentioned that both the original royal couple
and the soldier seem to inhibit the production of forms like
themselves, either temporarily or permanently. This obser-
vation led certain American students to propose the theory
of caste control by " social hormones ". It was suggested,
for instance, that the secretions of the royal couple, which
are so eagerly licked up by the workers attending them,
would be shared all through the colony by the mutual inter-
change of food, and might, like the effect of some drug,
inhibit the development of sexual maturity in other members
of the group. A similar theory might explain the inhibitory
effect of the soldier, though in that caste the secretions to be
handed round are much less copious.

Such theories meet two initial difficulties. First, in many
termites the variety of castes would require a surprising
number of special substances ; secondly, winged sexual forms

184

THE TERMITES

are produced in colonies containing a royal couple. It is
only the substitute sexual forms which are inhibited. The
difference is that the substitutes would reproduce in the
colony already containing the royal couple, whereas the
winged forms would leave it to found new colonies else-
where. The inhibition, therefore, is of a peculiar type.

Eventually Castle and Light tried to test the theory by
experiments with Zootermopsis. Royal couples were killed
and extracts of them were made in various solvents, such as
alcohol or ether, and the extracted material, after removal of
the solvent, was added to the food of groups of isolated
nymphs. It is broadly true to say that in all the most satis-
factory experiments the nymphs nevertheless transformed
into sexual adults without much delay. Such experiments are
not quite conclusive, since it must always be uncertain
whether it is possible to extract an unidentified organic sub-
stance unchanged. Yet since there is no objective evidence
for the existence of the inhibitory substances, and since the
mere absence of inhibition is a very incomplete explanation
of the differentiation into several castes of an isolated group
of similar nymphs, the theory has now little to commend it.

The social hormone theory is but one of several " trophic "
explanations of caste production in termites. Some authors
have supposed that foods of the ordinary sort, together with
varying amounts of salivary secretions, might control the
process. At present it is very difficult to accept any such
theory. The food varies considerably with the age and the
condition of the colony. In young colonies, the sexual forms
are largely reared on vegetable matter, and the same is true
of substitute sexual forms at any time. There is no evidence
for a copious supply of glandular food, such as is given to
the queen-producing brood of the honey bee. Moreover, the
self-differentiating group of nymphs scarcely seems to have

185

THE SOCIAL INSECTS

the possibility of providing itself with the necessary variety
of food.

The theory put forward by Grasse is that caste-determina-
tion is essentially a " group effect ", acting through the
nervous system. This implies that the physical development
and behaviour of a termite is determined in part by the
sensory impressions which it receives from the other members
of the colony. This might be automatically reinforced by
differences in feeding which become established as the
behaviour of different individuals diverged. Group effects
of a kind are, of course, common in human beings. We learn
the language and often the vocabulary and trade of our
immediate neighbours. But this is the result of imitation, of
learning and of the legacy of traditions, none of which is
operative amongst termites. A closer but still false analogy
is with the revolt, often subconscious, which children some-
times make against the ideas or way of life of their parents.

Group effects are, however, well known in many insects,
though none is quite comparable with what is assumed to
happen in termites. One example has already been described,
namely the inhibition of oviposition in humble-bee and
Polistes workers as long as the queen is present. In Polistes,
the removal of the queen releases egg-laying within a day or
two. Group effects have also been noted in various beetles
and flies when bred in laboratory cultures. Grain weevils,
for instance, do not lay at the highest rate of which they are
capable unless each female has at least fifty grains of wheat,
so that twenty females on six hundred grains lay less well
than the same number on one thousand grains. Moreover,
the weevils bred from a moderately crowded culture are
larger and are capable of laying more eggs than are those
reared from cultures which were either less or more crowded.
This last effect seems to be due to physical changes in the

1 86

THE TERMITES

culture such as the accumulation of the water given out
during respiration. The effect of crowding on the rate of
egg-laying is, however, likely to be due to nervous stimu-
lation.

A closer analogy to what is presumed to happen in termites
is found in the migratory locust. In this insect, both in
nature and in laboratory cultures, the individual reared in a
crowd is profoundly different from one reared in isolation.
It takes on a different colour, the shape of the thorax becomes
perceptibly different, the wings become longer and the
behaviour is entirely altered. The nymphs tend to march
about together in bands instead of hopping about as indi-
viduals, and the adults tend to undertake migrations in large
swarms. It is certain that these differences are mainly due
to crowding, and probable that they arise mainly from the
sensory impressions received through the eyes. Living
together in a crowd and seeing their fellows moving, makes
each individual more active, and this in turn in some way
alters the colour and shape during development. No doubt
this is a much simpler process than those which would have
to be invoked in the case of termites, since in the latter several
different developmental paths are open and each path is
chosen in approximately the right proportions. The theory
of the group effect in termites cannot at the moment be
regarded as more than an interesting hypothesis requiring
further work.

GROWTH AND DIVISION OF COLONIES

The multiplication of termite colonies by the marriage
flight of the winged adults, though the normal method, is
not the only one. In some species, e.g. Calotermes, if the
nest becomes very diffuse and spreads a long distance under-
ground, the inhibition exercised by the royal couple may be

187

THE SOCIAL INSECTS

no longer effective throughout, and substitute sexual forms
may develop. Such outlying sections of the colony may
eventually bud off, like the small bulbs found round a larger
one. Much less commonly, in a few kinds of termites, more
active splitting up of the colony has been observed, with all
its members walking out of the nest, carrying the eggs and
young larvae, and breaking up into groups, one with the
original royal couple, the others with substitute sexual forms.

The ability to produce these substitutes means that a
termite colony is potentially immortal, since its repro-
ductive functions can always be renewed. It seems, however,
that in many species substitute sexual forms do not play the
important role which they do in laboratory experiments.
Except in certain species, they seem to be relatively un-
common, probably coming into play when one of the royal
couple is lost in some rare accident. There is some evidence
that as the colony gets older, substitute forms are even less
effective, so that in reality most termite nests probably last
only a moderate number of years.

Calotermes colonies usually appear to survive for twelve
to fifteen years. Other species in which substitute sexual
forms are produced more regularly and in larger numbers
may well last longer. The very large colonies of the higher
termites must certainly be long-lived, if only because the
huge nests must be constructed over a long period. Hill has
recorded that the top of a large Australian steeple-like nest
was knocked off in 1872 to allow a telegraph wire to pass over
it and that the colony was still flourishing in 1935, when it
must have been much more than sixty-three years old, per-
haps nearly one hundred years. Apparently, the original
royal couple are indeed able to live for a period as long as
this, a much longer period than has been recorded in any
other insect.

188

THE TERMITES

These old colonies may contain an enormous number of
individuals. Thus in a South American species, with a nest
measuring about 2 by 1 yards, Emerson found three million
individuals. These large numbers can be arrived at by
counting samples. Many of the large African nests certainly
contain many millions of inhabitants. On the whole the
nests of the less specialised termites are much smaller, having
only a few thousand termites and flourishing for a much
shorter period.

One interesting exception is the Australian Mastotermes,
which as far as structure goes is the most primitive of all
termites, but whose colonies may nevertheless contain several
million individuals. It was mentioned earlier that the eggs
of this termite are glued together like those of a cockroach.
The insect too is cockroach-like in a number of details of
bodily structure, and this strongly supports the view that
it was from an insect of the cockroach-type that the termites
were evolved in the remote past. Some of the species of
cockroach which live out of doors show a certain resemblance
to termites in the care which they devote to their young and
in their wood-feeding habits.

Mastotermes itself seems to be the one surviving relic of a
type of termite which was once found all over the world.
Fossils closely resembling it have been found in the rocks in
a number of places, including the Isle of Wight, which at
the time the insects were alive must have had a much warmer
climate than it has to-day.

This account of termite life has been built up from the
observations and experiments of a Jarge number of workers
in all the continents, mostly during the last sixty years. The
whole body of information makes an impressive picture of
their strange social organisation. Yet one may well feel that
our knowledge of them is still quite rudimentary, that there

189

THE SOCIAL INSECTS

is far more to discover than has so far been revealed. We
know least about the termites which make the largest and
most perfectly organised nests. We know nothing about
their powers of communicating with one another, which
may well be as effective as in the honey bee. The great
attention which has recently and is now being paid to them
may lead to startling discoveries.

190

9

INSECT SOCIETIES

Insects, including even the most successful social species,
are so unlike man that their resemblances to one another
seem much more important than the differences. We feel, I
think rightly, that even the solitary cat is more like us than
the most socially-minded ant or the most docile termite.

Although the problems which have to be solved by any
animal which becomes social have always something in
common, the possible solutions depend mainly on the
structure and behaviour which the species has previously
acquired in the course of its evolution. The principal diffi-
culties which have to be mastered are, first, how to control
reproduction ; second, how to obtain enough food from a
limited territory ; third, how to substitute co-operative and
docile for solitary, aggressive behaviour ; and fourth, how
to adjust behaviour to the varying current needs of the
community. The solutions of these four problems by the
various types of social insect may be compared with what
we find in man and to a less extent in other vertebrates. The
problems are not, of course, independent of one another,
though it is convenient to consider them separately. For
instance, if the food supply were unlimited, control of
reproduction would be less important.

Explanation of biological observations can often be made
at several levels. Consider the fact that ants often eat a
proportion of the eggs which they lay. The immediate cause

191

THE SOCIAL INSECTS

of this behaviour in the individual ant may be merely hunger
or more specifically, need for extra proteins. But the circum-
stances of the whole colony for several preceding weeks may
have subjected that ant to protein starvation. Over a much
longer period it may have been advantageous for the species
to develop a habit of egg-eating when the rate of multipli-
cation was out of balance with the food supply. In the dis-
cussion which follows, many of the arguments are chiefly
concerned with the more remote levels. An individual insect
may behave in a way which is advantageous to it or to its
colony in the long run, without any suggestion that it knows
what it is doing. Some behaviour, indeed, such as egg-eating,
may seem harmful in the short run.

REPRODUCTION

While all species, including man, are capable of rapid
multiplication if enough food is available, there are good
grounds for thinking that the rate characteristic of each species
has been subject to evolutionary change. A species which
produces more young will not benefit unless more of them
are also brought to maturity. Many marine animals, such as
fish or molluscs, produce immense numbers of eggs of which
very few survive. In many of these animals the eggs and
sperm are merely liberated into the sea, and fertilisation is a
haphazard process. A less wasteful method of pairing is a
prerequisite of a smaller egg-production. Insects commonly
produce a few hundred eggs each, and of these a higher but
still small proportion reach maturity. It is advantageous to
reduce the number of eggs only if at the same time they have
more chance of survival, probably through the development
of maternal care.

Most insects do, in fact, lay their eggs in some more or
less protected situation on or near the food which their young

192

INSECT SOCIETIES

will eat. The solitary bees and wasps, which lay still fewer
eggs than most, look after them even more. Although no
solitary ants or termites are known, we can be certain that
the ants had solitary, wasp-like ancestors, to whose habits
the colony-founding queen of most species reverts. The past
history of the termites is less well understood, and it is
possible that co-operation between males and females, rather
as in the bark beetles, may have existed before they were
really social.

The reduction in the egg-number of nest-making insects is
a redistribution of physiological energy. If the female is to
live longer and to work hard at making a nest, it is impossible
for her to produce large numbers of eggs at the same time.
It may be in a sense an accident of the past history of a
species which determines whether it is capable of balancing a
reduction in eggs by taking better care of them.

Nevertheless, some maternal care has appeared again and
again in the course x)f evolution. Nest building occurs in vari-
ous fish, such as the stickleback. Here the eggs are guarded
by the male and the number laid by the female is very small.
The land vertebrates, especially birds and mammals, nearly
always produce fewer young than any solitary bee or wasp,
but show a similar or higher degree of care for them. Most
species restrict their breeding to a definite season of the year.
Even in tropical countries there are usually seasonal differ-
ences in rainfall or in the quality of the light, and the breeding
of most animals is adjusted to the seasons. Thus one very
common restriction on reproduction is to confine it to a
relatively short breeding season. This not only slows down
multiplication but adjusts it to the most favourable time of
the year.

Many birds, such as our common finches, aggregate into
flocks in the autumn and winter, and separate into pairs only

N 193

THE SOCIAL INSECTS

when they breed in the spring. This illustrates a difficulty
in adjusting reproduction to social life more fundamental
than the mere avoidance of over-eating the food supply.
Sexual reproduction tends to have a disruptive effect on social
organisation. In solitary bees and wasps one can often see
the nest-building females being pestered by ardent males
which still attempt to pair though the busy females no longer
need their attentions. Man is almost the only social animal
in which nearly all the adults of the right age are capable of
breeding at any season of the year. Probably all the other
social vertebrates, such as deer or beavers, have a relatively
short breeding season, outside which the relations of the
sexes are platonic. The majority of birds and mammals also
have some sort of territorial system. The male robin, for
example, at the beginning of the nesting season, occupies a
territory out of which other males are driven. His song
advertises his presence to unpaired females and also warns off
other members of the species. In this way he claims an area
of about the right size to supply one pair and their young
with food. In most seasons, some of the males fail to obtain
a territory and do not breed. Social insects, though they may,
as in ants, only forage in a limited area, defend the territory
only in the immediate neighbourhood of the nest.

The problem of too rapid multiplication in the social
insects has been met in a different way. One of the first steps
in the evolution of their social life was to establish a sterile
caste which either did not breed at all or did so only to a
limited extent and under certain conditions. They established
a distinction between egg-laying to produce workers and
egg-laying to produce sexual forms ; only the second process
needs severe regulation. In the less advanced societies, such
as those of the common wasps, the production of sexual forms
leads to the break up of the colony. This does not seem to be

i94

INSECT SOCIETIES

merely a modification of behaviour to meet the difficulties
of our winter climate, since colonies of some tropical species
disperse after a comparable developmental cycle.

The termites, owing to the nature of their sex-determining
mechanism, at all times produce males and females in equal
numbers. It is probable that in them the sterilisation of most
individuals of each sex must have developed very early in
their history.

The Hymenoptera, the group of which ants, bees and
wasps form a part, normally have a type of sex-determination
which makes it possible for purely female broods to be
produced for long periods. Once a female has mated the
actual fertilisation of each egg is under nervous control by
the female, and normally only unfertilised eggs will produce
males. While this might be expected to lead to a colony
rather different from that of the termites, the problems to be
solved are much the same.

In solitary bees,and wasps, unfertilised females do not seem
to make nests, though there is still a large field here for
experimental study. On the other hand, fertilised egg-laying
females are commonly quarrelsome and tend to interfere
with each other's egg-laying. Thus the sterilisation of most
of the females without depriving them of the urge to work
must in most species have occurred quite early. Moreover,
since unfertilised eggs will develop into males, merely to
deprive the female offspring of the queen mother of their
opportunity to mate does not prevent reproduction.

In the colonies of Polistes wasps it seems that the pre-
vention of egg-laying is still a mainly psychological in-
hibition. In the more advanced social species, the sterilisation
of the workers has gone much further, though egg-laying
can still occur, at least occasionally or in some individuals.

Sterilisation might be described as the least wasteful way

i95

THE SOCIAL INSECTS

of preventing reproduction. Social insects do, however, also
practise infanticide, the eating of eggs and larvae. The honey
bee is peculiar in also destroying adult males in the prime of
life as soon as they become superfluous.

While the less advanced insect societies break up as soon
as the next generation of sexual forms has been produced,
many species are able to emit huge swarms of males and
females without disrupting the colony. Ultimately, this seems
to depend on evolving a sufficiently long-lived queen or
royal couple around whom the workers are organised. The
break-up of the colony is probably more associated with a
decline in the oviposition rate than with the actual appearance
of the reproductives. In the termites and the honey bee the
production of substitute sexual forms gives the colony further
stability. Some ants may achieve the same end by receiving
back some of the young fertilised queens.

It is the peculiar triumph of man to have established social
communities more elaborate than those of ants and termites
without sterilising the majority of the adults. Other social
mammals have a relatively simple community and only a
short breeding season. In some primitive human societies
infanticide, especially of females, has been practised. In all
our societies a great variety of taboos and customs tends to
discourage promiscuity and partially to reduce the birth-
rate. Until very recently a high infant mortality and periodic
famine were the final but wasteful agents by which the
population was kept in equilibrium with the food supply.
If the progress of medicine and agriculture save us from
these, we shall presumably have to rely on family limitation.

It is no accident that the queens (or in termites, the royal
couple) of social insects live much longer than the females of
most solitary species. The full development of maternal care
demands a wide overlap between successive generations. For

196

INSECT SOCIETIES

species which are not social, this overlap need only last until
the next generation is self-supporting. In man, social life
has developed through the learning of crafts and the handing
on of traditions. Thus the great length of childhood in
man compared with any other animal has both made a long
period available for education and has required, in compen-
sation, a relatively long life for the adult.

Insects, in contrast, seem to depend very little on learning.
Their elaborate nests, the hunting territories, and some of
their food supplies may, like human traditions, be handed on
to the next generation, but would be quite insufficient in
themselves to preserve a stable society. The fundamental
reason for the greatly extended life of the queen is that she
stabilises the whole system by means of her oviposition.
Apart from periods of cold weather, no insect society is in a
stable condition unless the queen is laying regularly. The
whole system of inhibiting unwanted reproduction depends
on her, so that the colony usually survives very little longer
than the queen. The colonies of termites and of the honey
bee which no longer depend on the survival of the original
queen are potentially immortal.

FOOD AND SOCIAL LIFE

The importance of food in the early stages of social
evolution may have been much less in insects than it was in
man and some of the other social mammals. In early man and
in such animals as wolves, co-operation in hunting must
have been one of the original social motives.

In the social insects co-operation for this purpose is only
found in the ants, and in them chiefly in the more specialised
types. Wasps never hunt in packs, and bees forage alone,
though in the honey bee workers may tell one another where
to go. Termites are only rarely hunters and most of them eat

197

THE SOCIAL INSECTS

the same sort of vegetable food as probably satisfied their
solitary ancestors. This was at first made possible by their
special relations with the protozoa in their intestines. Similar
intestinal inhabitants are known in some solitary, wood-
feeding cockroaches. If this specialisation is lost, a more
varied diet at once becomes necessary.

The disadvantages of living by hunting other animals are
that the supply is too variable and not really large enough to
allow many hunters to live in one place. The driver and
legionary ants have carried this type of existence to its extreme
development, at the expense of being permanently nomadic.
They appear to be a successful sideline from the main stream
of ant evolution. Bees and wasps have attained to a different
solution. They have never developed any private source of
food, but compete with a large number of other animals.
It is true that bees and flowering plants have had a profound
influence on one another's evolution, but nectaries which
are open to bees but closed to other insects are possible
only to a limited extent.

The wasps are still mainly hunters, and only a very few
of them, such as a few of the South American Polybiines,
compete seriously for nectar. Bees and wasps have, however,
developed great powers of flight and an ability to work hard.
While honey-bee workers will get their nectar and pollen
from near at hand if they can, it is not at all unusual for them
to fly 2 miles or even considerably farther.

It is well known that in man large communities were
possible only after agriculture had been developed. This
not only provides much more food per acre than hunting
does, but also supplies it in a form in which it can be stored
for use in winter. The honey bee and the honey-gathering
wasps, although not agriculturists, both lay up similar
reserves. The termites seem on the whole to manage by

198

INSECT SOCIETIES

feeding on abundant widely available substances for which
competition is not very keen, since special modifications are
needed if they are to be utilised economically. Only the most
advanced families cultivate fungi, and the value of this crop
as food is uncertain. Most termites, however, can and do
lay up food-reserves.

Apart from man, it is the ants which have developed the
most varied and also probably the best solutions of this
problem. While the more primitive types still live almost
entirely by hunting other insects, the others have exploited
a gradually widening range of foods. This has been associated
with a great increase in the size of the colony. The greater
availability of food may, however, be only a part of the
reason for this.

The exact relations between ants and their green-fly cattle
are in doubt, but it is certain that they have much less control
of their beasts than we have of ours. Cattle by themselves,
without an agricultural industry which provides feeding-
sturTs, do not much increase the yield per acre.

This is a point which it would be interesting to study
in such species as the British yellow mound-making ant,
Lasius flavus, which is almost completely dependent on root-
aphides. The cattle are suppliers of sugar and proteins, so
that the ants have access to the riches of the plant kingdom
without needing the special digestive arrangements of the
termites.

Many ants also feed on plant seeds, and the harvesting
ants do so extensively ; but here again no great yield per
acre is possible without a proper agricultural system. This
has been developed only in the leaf-cutter ants with their
fungus gardens, which they providently manure with their
own excreta. It is probably not a coincidence that some of
these ants make the largest nests with the most numerous

199

THE SOCIAL INSECTS

inhabitants of any member of their family. It is not yet
possible to say whether their solution of the problem of
food-supply has enabled them to make other cultural
advances. Our knowledge of their behaviour is still in a
rudimentary state.

The general conclusion is that social insects have increased
the variety of their diet, that is they use a greater proportion
of what is available near at hand. Apparently very few have
been able to increase the quantity of any one kind to be found
within reach of the nest. It is possible that some of the ants
which keep cattle have done so, and it is almost certain that
the leaf-cutter ants have solved this problem.

THE ORIGIN OF SOCIAL BEHAVIOUR

The third point to be considered is the acquirement of
social behaviour by species which were originally solitary.
This also involves the idea of division of labour, and ulti-
mately the question of communication.

The simplest example of the division of labour is seen
where two sexes share the duties of reproduction. There is a
number of animals which show us that such a division need
not be made. In many molluscs, such as our common land-
snails, every individual performs the function of both sexes.
Nevertheless, in the great majority of species the sexes are
separate. This applies to almost all insects, and to all verte-
brates.

It is characteristic of insect societies that the division of
labour amongst their members is founded chiefly on physical
differences which are often well marked. Apart from
differences in behaviour which depend upon age, all the
individuals which have distinct roles in the society differ
from one another in structure. The differences between the
queen and the worker, between the soldier, the worker major

200

INSECT SOCIETIES

and the worker minor, are often very marked and quite^as
big as the differences between species of solitary insects.

If workers have a uniform structure, then at different times
they contribute to all the activities of the colony. The castes
of social wasps are feebly differentiated, and in them per-
manent division of labour, except between the queen and her
workers in the mature colony, scarcely exists. In the termites,
by way of contrast, we find the worker and the soldier often
with a strikingly different appearance and with quite distinct
functions. Even more curious is the example of the amazon
ants, in which the amazons are in effect the soldiers and the
worker caste is represented by slaves of another species. It
seems that the essential nature of an insect society must be
looked for in the physiology of the individuals.

Before pursuing this point further, we may notice how
different has been the evolution of social vertebrates. Structural
and physiological differences within the species are relatively
trivial apart from the differences between the sexes. Perhaps
because of this, only man has been able to build societies
of a complexity even approaching that of the ant or termite
colony. Without physiological differentiation, it is very
difficult to organise a society until there are elaborate means
of communication and a brain capable of understanding
what is conveyed.

There are a few exceptional animals like the beaver about
which one would like to know much more, but in general
the non-human social mammals combine only in a rather
rudimentary way, either for hunting or for defence. Any
division of labour depends either on age or on sex ; for
example, bulls defending the cows and young amongst wild
oxen. It might be argued that in man there is a division of
labour based on innate differences in abilities, but there is
very little physiological differentiation. The superficial

20 1

THE SOCIAL INSECTS

human variation in stature and in colour of hair and skin
are not related to social duties. With the requisite training,
most men can do most of the essential work of the com-
munity.

To return to insect societies, clearly the fundamental
problem of their origin is, " How did physiological differ-
ences arise within the species ? " Solitary bees and wasps
construct nests, and only a slight lengthening of the life of
the mother would be enough to make it overlap those of her
offspring. They are also capable of producing exclusively
female broods, though none of the solitary species is known
to do this except occasionally by chance. Such Hymenoptera,
therefore, could at once set up a primitive society of the
Halictus or Polistes type if they could produce a brood of
young females who would work without laying eggs.

There is little doubt that an important factor has been the
modifying influence which social life has on the process of
natural selection. In a solitary species, individuals with
reduced fertility will not often survive in competition with
others which are more fertile. But in social species any
change which benefits the group as a whole is likely to be
preserved.

In insects generally a reduction in size of the female is
nearly always associated with the laying of a smaller number
of eggs. It is significant that in most species the worker is
smaller than the queen. It is possible that originally, as is
still true to some extent, queens and workers differed chiefly
because of chance variations in the amount of food which
they received as larvae. Later, a system could have been
evolved in which differences in the food received had a greater
effect on the adult, whereas in solitary species the tendency
would be to keep the species as insensitive as possible to
variations in food supply. When physiological studies have

202

INSECT SOCIETIES

been made of the effects of nutrition on the size, structure and
fertility of Hymenoptera it may be possible to argue the case
in greater detail.

The beginning of termite societies was probably quite
different. There are several kinds of insects, such as the
Embiids, in which all the developmental stages are known to
live together, sharing a common silken shelter but not help-
ing one another in any other way. What seems to have hap-
pened in the termites was that the developmental stages began
to play a part in the life of the community as important as
that of the adults. This was possible in them because of their
gradual metamorphosis and the general similarity of the
early stages to the adult. The caste system of the termites
would originally have been founded exclusively on age, the
half-grown larvae or nymphs doing most of the work while
the queen laid the eggs. This system would increase the
chances of variations in nutrition during the developmental
period. As in the Jlymenoptera, an increased response to the
effects of nutrition might have been evolved later. Such a
scheme would give at least a hypothetical explanation of the
origin of the more primitive termite societies. We still
know so little of the complex caste system of the higher
termites that any discussion of them is premature.

INSECT SOCIETIES AS FAMILIES

An insect society is a family, a group of offspring sur-
rounding the mother or royal couple. Their mutual toler-
ance is founded on growing up in the same nest. If ant
colonies are split into two in artificial nests, the two halves
may fight if reunited after an interval of some months. Judg-
ing by ants and honey bees, mutual tolerance depends on
sharing a common nest-smell, in natural conditions acquired
during development. Ants of the same species but from

203

THE SOCIAL INSECTS

different colonies fight bitterly, and colonies of the honey bee
can be combined in one hive only by devices which enable
them to acquire a common smell. It is very unusual, perhaps
even unknown, for separate colonies to combine in natural
conditions.

Thus while the insects have species and families or colonies,
they do not have tribes or nations. The failure to develop
co-operation on the larger scale which occurs in man may
be partly due to their much higher fertility. In man, several
families had to combine to form a group big enough to master
a mammoth, let alone build a cathedral. But insect families
are so much larger that there would never have been the
same advantage in combining with strangers.

When two insects meet they do not normally fight unless
one forms the natural food of the other. What might be
called nasty aggressive behaviour, worthy to be compared
with our own, is always in essence the fight for an ovipositi on-
site. Some flies which lay their eggs in fruit will knock other
females off in order to get their place on the fruit, but
aggression is usually found only in females with a nest to
defend, and they will drive off any other insect which hap-
pens to come too close.

Members of one colony would, therefore, live together
amicably if they all smelt the same and if there were only one
ovipositing female. In itself, this would not lead to much in
the way of co-operation, and probably at first the nest would
be an aggregation of the single nests of each individual.
Something like this seems to be still true of Halictus. But
if one started with an insect which already had behaviour
as elaborate as that of the solitary Ammophila, it is reasonable
to think that the great advantages of co-operation within the
family group would provide plenty of opportunities for
selection. A great deal of the social co-operation of insects is

204

INSECT SOCIETIES

no more than the accomplishment of the same task by indi-
viduals which are all in the same physiological state. It is
only in the most advanced species, in ants like Oecophylla,
in the honey bee, and in the specialised termites that co-
operative behaviour can clearly be seen.

It may seem far-fetched to suggest that all ethical systems
are ultimately founded on the necessities of social life. Some-
thing of course depends on the definition of the word ethical ;
if it is restricted to behaviour determined by a conscious plan
of action, all animals other than man would seem to be
excluded. It is perhaps more reasonable to regard the
refined forms of altruism which we consider the highest type
of ethical behaviour as developments from the maternal care
of many solitary animals, and especially from the family
loyalty of social species. If a bird defends her nest against
an enemy, her behaviour is often said to be instinctive or
innate and therefore no more creditable than eating a worm.
In the social insects, workers will defend their nest to the
death, and will care for young which may be only remotely
related to them. In human beings, on some occasions and
by some individuals, the whole species is treated as a family
to be defended or cherished. Probably very little of our
altruism is instinctive or even unconscious. Nevertheless,
the instinctive altruism of the primitive family may have
made the pattern from which the traditions of conscious
altruism were later evolved.

Our assessment of ethical values is associated with a
conscience, a sense of right and wrong. In its lowest form
this is a desire to win the approval or to avoid the dis-
approval of the rest of the community. Domesticated and
therefore semi-social animals like the dog can be trained until
they have a rudimentary conscience, founded on an experi-
ence of rewards and punishments.

o 205

THE SOCIAL INSECTS

The process of natural evolution can be looked on, by
analogy, as a slow but very prolonged process of training in
which the rewards are food and successful reproduction and
the punishments hunger and extinction. In a social mammal
or insect, other members of the community are an essential
part of the environment without which the individual is
almost as helpless as a fish out of water. This reaction to
other individuals is innate and is the only element of social
behaviour which is common to all such species.

It cannot be supposed that social insects have any sense of
right or wrong. So far as they are obeying the laws of their
own physiology, they are very unlikely to have moral
scruples. But from the evolutionary point of view motives
are unimportant ; ethical conduct is what benefits the com-
munity, even at the expense of the individual life. The code
developed during evolution, because it is such a slow
growth, is apt to be too rigid to succeed on all occasions.
It leads to those " errors " of instinctive behaviour which
Fabre and some other naturalists have treated as contra-
dictions of the Darwinian theory. This, however, is a mis-
apprehension ; natural selection is seen just as clearly in the
failure as in the success of a behaviour pattern, provided that
in the long run the results determine survival.

The adjustment of individual behaviour to social needs
is closely related to the topics already discussed. An insect
living in a society has to do a greater variety of tasks, at least
until an advanced division of labour has been evolved, than
does a solitary species. Its action cannot follow any very
rigid time-table. Perhaps the various methods used in cooling
the nest when it gets too hot are the most striking example
of sudden adjustment of behaviour to the needs of the colony.
If a solitary species gets too hot, it may try to escape, but it
has never been known to do anything to make its nest cooler.

206

INSECT SOCIETIES

Although reactions to novel situations have not been much
studied, they are clearly better developed in the more
specialised social species.

Most wasps make no use of old nests, even in the tropics
where one can often see deserted ones in good condition.
It is a curious fact that whereas various solitary wasps will
build mud cells in the old combs of Polistes nests, the
Polistes themselves never seem to use an old nest, though
they may occasionally suspend a new one from it. The honey
bee, which will use a machine-made wax foundation as the
basis for a new comb, shows a higher form of behaviour.
Ants and termites, probably because they practice a less rigid
type of architecture, seem to be much more adaptable. This is
shown in the ease with which they can be domesticated in
artificial nests. It is, however, in the rigidity of their behaviour
that insect societies are most different and inferior to ours.
The adjustments which they can make to an unfavourable
or varying environment are relatively trivial. This rigidity
is one of the basic differences between the insect and the
mammal.

Even such mammals as rats and voles explore their habitat
apart from an immediate hunt for food. One of the reasons
why they get trapped is that they tend after a time to examine
any new object placed in their territory. Exploratory be-
haviour is much more developed in monkeys, and is carried
still further in man. This sort of activity has allowed man
to develop complex societies without the physiological
differentiation which was essential in the social insects. We
have to pass education acts to prepare our citizens for social
life, whereas insects are, to a very large extent, hatched
already educated.

There is another point of some importance. Man has been
social for, probably, less than a million years, whereas the

207

THE SOCIAL INSECTS

higher social insects have lived in complex societies for at
least thirty times as long. Few biologists would care to predict
where we shall be when thirty million more years have
passed. H. G. Wells, in his story The Time Machine^ fore-
casts one possible result of such a period of evolution,
assuming that we have a somewhat ant-like future. Such a
course seems much less likely now than it may have done
fifty years ago.

The efficiency of some of the different types of ant societies
can be judged by their success when in competition with one
another. The Argentine ant, a native of the southern part of
South America which has been accidentally introduced into
many parts of the world, is a good example, as Haskins has
shown. It is a small, fragile species with a soft, easily
damaged skeleton. The workers have no sting, but they can
produce a sticky, scented, anal secretion which may be of
some defensive value. It has now established itself in many of
the warmer parts of the world and in some, such as the island
of Madeira, it has completely eliminated the local ant fauna.
Though it cannot tolerate cool climates like ours, even in
the British Isles it has several times survived for some years
in the shelter of buildings.

It seems to have two main advantages. First, the workers
forage in large groups which are capable of an unusual degree
of co-operation. Secondly, it has a specialised type of repro-
duction. The queens are small, hardly larger than the
workers, but there are many of them in each colony : they
have solved the difficulty of getting several egg-laying females
to tolerate one another. This means that it is very difficult
to destroy a colony, and that the queens are cheap to produce.
This type of organisation is very successful, but not very
mobile. The queens cannot found new colonies unaided and
the colonies reproduce by " budding ", the splitting off of

208

INSECT SOCIETIES

groups of queens and workers. This is the chief reason why
the species did not spread until human transport was avail-
able to carry it about. But once this was provided its other
superiorities took effect. When it is not given such transport,
the other ants whose queens may fly long distances to found
new colonies have a compensating advantage, at any rate
outside small islands.

THE STUDY OF SOCIAL INSECTS IN THE FUTURE

The progress of biology in the last hundred years has been
associated with the gradual application of the experimental
method to more and more varied fields. The experimental
study of insect societies is quite recent. A beginning was made
in the nineteenth century, notably in the study of Italian
termites in the 1890s, by Grassi and Sundias. But most of
the tools necessary for such a study have become available
only recently. This is especially true of the basic facts of
insect physiology ,and of the methods and ideas used in
analysing the simpler behaviour of solitary species.

We can be sure that the experimental study of the organi-
sation of social insects will lead to many more strange and
even startling discoveries. For a long time to come, clear and
imaginative thinking will be quite as useful as elaborate
apparatus. A beehive or an ant-colony is easier to keep than
a dog, and although it will hardly be so satisfying to the
affections, it will provide a host of unsolved problems on
which anyone may exercise his disinterested curiosity. It is
the application of this distinctively human faculty, not only
to insect societies but also to our own, which is most needed
in the world to-day.

209

FURTHER READING

Butler, C. G. The Honey Bee, Oxford University Press,
Oxford.

Duncan, C. D. 1939. A Contribution to the Biology of
North American Vespine Wasps. Stanford University
Publications, Biological Sciences, 8 : 272 pp.

Imms, A. D. i 93 i. Social Behaviour in Insects. Methuen,
London.

Imms, A. D. 1947. Insect Natural History. Collins, London.
(This is a good general introduction to the biology of
insects).

Frisch, K. von. 1950. Bees. Their Vision, Chemical Senses,
and Language. Cornell University Press, Ithaca, N.Y.

Schwarz, H. F. 1948. Stingless Bees (Meliponidae) of the
Western Hemisphere. Bull. American Mus. Nat. Hist.,
90 : xvii. + 546 pp.

Sladen, F. W. L. 191 2. The Humble-bee, Its Life-history
and How to Domesticate It. Macmillan, London.

Snyder, T. E. 1935. Our Enemy the Termite. Comstock
Publishing Co., Ithaca, N.Y. (The most recent full
account of termites is in French and by P. P. Grasse,
in Traite de Zoologie, vol. 9. Masson et Cie., Paris,
1949).

Wheeler, W. M. 19 13. Ants. Columbia University Press,

N.Y.

Wheeler, W. M. 1923. Social Life among the Insects.
Constable, London.

211

INDEX

INDEX

Acacia, ants' nests in, 136

Adlerz, 31, 41

Aerodromes, termite nests on, 178

Aggressive behaviour in insects,
204

Amazon ant: see Polyergus

Ambrosia beetles, 37

Ammophila pubescens: way-finding,
31J nesting, 40

Anergates, 165-6

Anoplius fuscus, 31

Ant thrushes, 125

Antennae, function of, 19

Anthophora, 76

Ants: and aphides, 121-2, 129; castes
of, 119, 139-46; enemies of termites,
125; food of, 122; founding of
colonies, 137; geological history,
120; guests, 1 5 1-2; nests, 134;
sting, 125

Aphides, 121

Apis (honey bee), species of, 103

Argentine ant, 208

Atta, 132

Auxiliaries in Polistes, 62

B

Baerends, 40

Balm, no

Bark beetles, 37

Beaumont, de, 153

Bees: and flowers, 77-8; development

of social behaviour, 78-9; solitary,

76
Beirne, 56
BischofT, 156

Black ant, common: see Lasius niger
Blewett, 21
Blood-red ant: see Formica sanguinea

Blue butterflies: caterpillars in
ants' nests, 152; eaten by wasps,
5o

Bols, 85

Bombus: see Humble bee

Bombus hortorum, nest census, 91

Bombus lucorum, 156-7

Bombus terrestris, 156-7

Bothriomyrmex decapitans, 158

Brachygastra, 129

Brain of insects, 20

Brauns, 82

Brian, 87, 92, 138, 142

Budding by ant colonies, 138

Bulldog ant: see Myrmecia

Burying beetle, 36

Cabbage white butterfly, 33

Calotermes, 183, 187, 188

Cannibalism in termites, 54-5, 175

Carpet beetle, 24

Carthy, 148-9

Carton made by ants, 135

Cassique, 69

Castes: bees (Halictus), 64-5; bees
(honey), 105; bees (humble), 90,
92; bees (stingless), 99-102; ter-
mites, 179-84; wasps (Polistes), 64-
5; wasps (Polybia), 71; wasps (Ves-
pula), 45, 53

Castes, determination of: ants, 140-6;
honey bee, 107; stingless bees,
100-1; termites, 184-7

Castle, 183, 185

Cellulose, digestion of, 176

Chambers, 77

Cinnabar moth, 33

Clausen, 121

Cleveland, 176

Clover: see Red clover

215

INDEX

Cockroaches, similarities to termites,
189

Colony and individual in evolution,
145-6

Comb of wasps, 46

Common wasp, 46

Communication: in ants, 146; in
honey bees, 11 2- 17

Compass termites, 171

Co-operation in insects, 204-5

Cuckoo bees and wasps, 152-7; evo-
lution of, 156

Cumber, 89, 90, 92, 95, 155

Cyprus, hornets in, 56

Dances of bees, 1 1 1

Darwin, 84, 119

Degeneration in parasitic ants, 163

Deleurance, 54, 60, 61, 64

Digestion in termites, 176

Direction, indication of, by honey

bee, 116
Diver, 165
Division of labour: in honey bee, in;

in humble bee, 92; in wasps, 52
Dogs, social insects more interesting

than, 209
Dolichoderus, 69
Dominance in Polistes, 62
Driver ants, 67-8, 125
Ducke, 66
Duncan, 55
Dung beetles, 35

Earwig, 35

Egg eating: in ants, 145; in termites,

175; in wasps, 55
Egg production: in grain weevil

(group effect), 186; honey bee, 105;

humble bee, 87; legionary ant, 127;

Polybia, 71; termites, 179-80
Embiids, 34, 203
Envelopes of wasps, 46
Ethics, origin in social behaviour, 205
Eumenid wasps, 44

Fabre, 35, 206
Farquharson, 152

Fire ant: see Solenopsis
Food and social life, 197-200
Formica fusca: as host of F. rufa, 139;

as slave, 159, 162
Formica rufa (wood ant): colony
foundation, 139; food, 124; host of
Formicoxenus, 163; parasitism by,
158; variation of workers in size,
139
Formica sanguinea (blood-red ant),

159

Formicoxenus nitidulus, 163
Fraenkel, 21
Frisch, von, m-17
Froggatt, 125

Fungus: cultivated by leaf-cutting ants,
133; cultivated by termites, 175

Gaul, 32, 52

Genetic theories of castes, 100, 140

Geotrupes typhoeus, 35

German wasp, 46

Goedart, 87

Goetsch, 141

Grain weevils, egg-laying, 186

Grass seeds: collected by ants, 131; by

termites, 175
Grasse, 182, 183
Grassi, 209
Gregg, 141
Group effects, 186-7
Growth in insects, 22
Gynandromorphs, 144

H

Habituation, 28

Halictus, 79-82

Harvesting ants, 130-2

Haskins, 208

Hearing in insects, 18

Hereditary castes in stingless bees,

100
Hibernation: in humble bees, 84-5;

in wasps, 40
Himmer, 50
Hingston, 124, 125, 128
Hives of stingless bees, 98
Hoffer, 87, 115
Honey: see Nectar
Honey bee (Apis mellifera): absence of

cuckoo species, 168; castes, 105;

2l6

INDEX

Honey bee — continued

determination of castes, 107; dis-
tance flown, no; division of labour,
in; drones, 106-7; language and
dances, 1 12-17; nuptial flight, 106;
relation to flowers, 109—10; relatives
of, 103; substances used by, no;
swarming and supersedure, 108

Honey-pot ants, 129

Hornet: in Cyprus, 56; effect of heating
nest, 51; in England, 57

Humble bees, (Bombus), 83; cuckoo,
see Psithyrus; and flowers, 93-5;
hibernation, 84-5, 92; nest site,
83-4; nest size, 88; pairing, 87; sting
83; tropical, 95

Humboldt, 178

Hunting wasps, 40

Ichneumon wasps, 34, 42
Ihering, von, 66
Insight learning, 3 1
Instinctive behaviour, 25
Intersex, 144

J

Jet black ant: see Lasius fuliginosus

Labauchena, 158

Language: see Dances and Com-
munication

Larva: definition, 24; in termites, 180

Lasius flavus (common yellow ant),
120, 199; variation in worker size,
139

Lasius fuliginosus (jet-black ant), 148

Lasius niger (common black ant),
148

Leaf-cutting ants, 132-44, 199

Learning, 27; unimportance of, in
insects, 197

Legionary ants, 125; nature of queen,
139

Leptogenys, 122

Leptothorax, 142

Light, 185

Lobopelta, 125

Locust: see Migratory locusts

Longevity of queen: essential to social

life, 196; in termites, 188
Lubbock, 122

M
Macgregor, 149
Mason bee, 76
Mass provisioning, 44
Mastotermes, 179, 189
Maternal care, 34, 193
McCook, 132
Mealybugs, 121
Melipona, 96
Membracids, 129
Metamorphosis, 24
Migratory locusts, effects of crowding1-,

187
Miller, 183
Monkshood, 93
Moulting: in insects, 22; in termites,

182
Myrmecia (bull-dog ant), 125
Myrmica (red ant), 128, 130, 137;

scent trails, 149
Myrmica rubra, 138, 142

N

Nasute termite, 181

Nectar: collection by honey bee, 109;
collection by wasps, 63, 70; ex-
ploitation of, 198

Newman, 84

Nixon, 56, 156

Noirot, 183

Noll, 81

Norwegian wasp, 56

Nuptial flight: of honey bee, 99; of
humble bees, 89; of stingless bees,
99; of termites, 173

Nuptial promenade, 173

Nymph: definition, 24; of termites, 180

Oecophylla, 121, 124, 135-6
Okland, 124
Osmia: see Mason bee
Oviposition, provoking jealousy in
queen, 38

217

INDEX

Parasite, definition, 42

Parasitic: ants, 158-68; humble bees,
see Psithyrus; termites, absence of,
168; wasps, 152-4

Pardi, 60, 62

Path-finding: see Way-finding

Pelopaeus: see Sceliphron

Pheidole: megacephala, 121; determina-
tion of castes, 14 1-2; soldiers, 140

Philanthus: hunting, 27; nest-finding,
29

Planta, von, 107

Plath, 87, 88,95, 155

Polarised light: sensitivity of ants to,
149; of the honey bee to, 117

Polistes, 59; abortion of brood, 61;
behaviour in different climates,
61-2; economics of swarming, 63-4;
hibernation in S. America, 63;
parasitic species, 152-4

Pollen: stored by honey bee, no;
stored by humble bees, 88

Polybia rejecta, 69

Polybia wasps, 65; egg production, 71;
food, 70; nests, 67; relations with
ants, 67-9; size of colonies, 69-70

Polyergus (Amazon ant), 161-2

Polymorphism in ants, 140

Polyrhachis, 128

Ponerine ants, 124

Progressive provisioning, 44

Propolis, no

Prorhinotermes, 183

Protozoa in gut of termites, 176

Pseudacanthotermes, 173

Psithyrus (cuckoo humble bee), 154-7

Pupa, definition of, 24

Queen cells: in honey bees, 104; in

wasps, 52
Queens: see Castes, Halictus, Honey

bee, Humble bee, Termites, Wasps

R

Red Admiral, 25
Red ant: see Myrmica
Red clover, 93-5
Repletes, 129

Reproduction, 23; control of, 192-7;

disruptive effect of, 194
Respiration, 17
Ribbands, 109
Rigidity of behaviour in insects, 32,

207
Rosch, in
Rove beetle, 35
Royal jelly, 107, 112

Sand wasps, 43

Sceliphron, failure of instinct, 32; nest,

43

Scent: produced by honey bees, 108;
trails in ants, 148-9

Schnierla, 127

Scoliid wasps, 43

Sex determination, 38, 101, 180

Sex mosaics, 144

Shapley, 20

Sladen, 92, 155

Slave-making ants, 159

Small ermine moth, 33

Smell, sense of, 18

Social behaviour: definition, 37; de-
pendence of on structure and phy-
siology in insects, 201; difficulties of,
65, 191; as family life, 203-9; origin
of, 200-3; types of, 33

Social hormones, 184

Social insects, number of kinds of,

38
Soldiers: in ants, 119; in harvesting

ants, 132; in Pheidole, 140; in ter-
mites, 1 80

Solenopsis, 158

Solitary wasps, 42

Speed of walking in ants, 20

Spider-hunting wasps, 43

Sting: of ants, 122-3; °f humble bee,
83; of wasps, 50

Stingless bees, 96

Stockhert, 81

Strongylognathus, 164-5

Structure of insects, 16

Substitute sexual forms, 182

Sundias, 209

Supersedure, 108

Swarming: in honey bee, 108; in wasps,
63,66

Swift moth, 33

218

INDEX

Tapinoma nigerrimum, 158

Taste, sense of, 19

Temperature: of humble bees' nest,
87; of insects' bodies, 20; perception
of, 19; of Polistes' nest, 64; of wasps'
nests, 50

Temporary social parasitism, 138

Termites: ants as enemies, 125; castes,
179-84; caste determination, 184-7;
digestion, 176; effects on vegetation,
178-9; egg production, 179-80;
food, 174-7; growth of colonies,
187-9; longevity of colonies and
queens, 188; nests, 170; number of
individuals in colony, 189; origin of
social behaviour, 203; as pests, 177-
8; sex determination, 180; wasps as
enemies, 174

Territorial system, 194

Tetramorium, 165

Thorpe, 31

Tinbergen, 27, 29

Tracks: of ants, 148; of harvesting ants,
131

Trapnell, 178

Trial and error learning, 28

Trigona, 96 /

Trophic theories of caste differentia-
tion, 140, 185

Trumpeter bee, 87

Wasp, German (Vespula germanica),
nest usurped by common wasp,

156

Wasps: British species, 45; colony

foundation, 46; deserting nest when
moved, 32; destruction of colonies,
49; division of labour, 52; as enemies
of termites, 174; food, 47, 50; grubs,
secretion of fluid by, 51; males, 52;
nocturnal, 58; queens, 45, old
queens, 48, numbers produced, 53;
queen cells, 52; temperature of nests,
50; treatment of brood, 51, des-
truction of, 54; types of social
organisation, 72-4; varieties of, 41;
wasp years, 55; worker caste, 45; egg
laying by, 55

Water: collected by honey bee, no;
dependence of insects on, 21;
Polistes using, 64

Wax, 85

Way-finding, 29; in Ammophila, 31; in
Anoplius, 31; in ants, 147-9; in
Philanthus, 29

Wells, H. G., 208

Wesson, 142

Weyrauch, 52, 53, 62, 153, 156

White ants: see Termites

Whiting, 145

Wiggles worth, 22

Wodsedalek, 24

Wood ant: see Formica rufa

Wood damaged by termites, 177-8

Workers laying eggs: in honey bee,
105; in Polistes, 60

Vespula, 45, 56; V. austriaca, 152-3; V.

germanica, see Wasp, German; V.

vulgaris, see Wasp, Common
Vision in insects, 17

Yellow ant: see Lasius flavus

W
Waloff, 90

Wasp, Common (Vespula vulgaris),
size of colony, 49

Zimmermann, 121
Zootermopsis, 182, 185

219