THE SOCIAL LIFE OF ANIMALS
THE
SOCIAL
LIFE
OF ANIMALS
BY W. C. ALLEE
PROFESSOR OF ZOOLOGY THE UNIVERSITY OF CHICAGO
WW- NORTON & COMPANY • INC
Publishers, New York
Copyright, 1938, by W. W. Norton & Company, Inc. 70 Fifth Avenue, New York City
First Edition
Published by arrangement with The University of Chicago Press
printed in the united states of AMERICA
This hook is gratefully dedicated to the past and present members of our "Ecology Group"; without their enthusi- astic co-operation much of the underlying evidence could not have been collected during my lifetime, and without their critical attention the expression of these ideas would have been more faulty.
Contents
Foreword 13
I. Science versus Metaphysics 15
II. History and Natural History 20
III. Beginnings of Co-operation 50
IV. Aggregations of Higher Animals 90 V. Group Behavior 133
VI. Group Organization 175
VII. Some Human Implications 209
VIII. Social Transitions 244
Literature Cited 277
Index 289
52556
Illustrations
PLATES
FACING PAGE
I A. A hibernating aggregation of ladybird beetles 32
I B. A breeding aggregation of midges 32
II. A grassland-bison community 38
III. Aggregating behavior of brittle starfish 44
IV. Diagrams showing the effect of population size
on the rate of evolution 128
V. Castes of a termite from British Guiana 266
FIGURES
PAGE
1. Grasshopper nymphs on the march 36
2. The effect of numbers present on rate of bio-
logical processes 52
3. Group protection from ultra-violet radiation
for planarian worms 60
4. Another aspect of group protection for plana-
rians 62
5. The small marine flatworm Procerodes 64
6. Group protection from fresh water for Proce-
rodes 66
7. Bacteria frequently do not grow if inoculated
in small numbers 67
8. The common sea-urchin Arbacia 70
9. Arbacia eggs cleave more rapidly in dense
populations 72
10. Robertson found that two protozoans placed
together divided faster than if isolated 76
9
lO ILLUSTRATIONS
PAGE
11. Other protozoa reproduce more rapidly when a
certain number of bacteria are present 78
12. Some protozoans divide more rapidly in dense
bacterial suspensions if more than one is present 79
13. A and B. A recent suggestion concerning the
ancestral relations within the animal king- dom 86
14. Goldfish grow more rapidly if placed in slightly
contaminated water 95
15. An extract from the skin of goldfish frequently
has growth-promoting power 97
16. White mice grow faster in small groups than in
large ones 101
17. Flour beetles reproduce more rapidly if more
than one pair is present 105
18. The "spread" of time in which eggs are laid in
a colony of herring gulls affects the per- centage that survive ii2
19. In small populations, genes drift into fixation
or loss largely irrespective of selection 121
20. In medium populations complete fixation or
loss is less likely to occur 123
21. In large populations, gene frequency is held to
a certain equilibrium value by the opposing pressures of mutation and selection 124
22. As intensity of selection increases it becomes
more and more dominant in determining the end result 126
23. Manakin males establish rows of mating courts
in the Panamanian rain-forest 134
24. Many kinds of fishes eat more if several are
present 136
ILLUSTRATIONS 1 1
PAGE
25. An ant which works at an intermediate rate
may be speeded up if placed with an ant which works more rapidly, or vice versa 141
26. A simple maze used in training cockroaches 151
27. Isolated cockroaches make fewer errors during
training than if paired or if three are trained together 152
28. They also take less time 153
29. Parrakeets learn equally well if trained when
isolated, whether they are caged singly or in pairs 156
30. Parrakeets learn more rapidly if trained alone
than if two are placed together in the maze 157
31. Feeding a fish which has just come through the
opening from the larger side of the aquarium 160
32. Goldfish learn to swim a simple aquarium-maze
the more readily the more fish there are present 161
33. Isolated goldfish learn the problem set for them
less rapidly, and unlearn it more readily 162
34. The aquarium-maze used in training part of
the fish to come forward and part to go to the rear to be fed 164
35. Cyprinodon learn to move in a body more read-
ily than to split into two separate groups 165
36. Goldfish learn more readily if accompanied by
a trained leader 166
37. An aquarium-maze arranged to test the power
of observation of fish 168
38. Goldfish react more rapidly if allowed to watch
others perform ^ 169
39. Flocks of hens are organized into a definite so-
cial hierarchy 178
40. Cockerels also have a social organization 180
1 2 ILLUSTRATIONS
PAGE
41. In flocks of pigeons the organization is one of
peck-dominance rather than of peck-right 187
42. The Dionne quintuplets also show evidence of
a social organization among themselves 204
43. The percentage of births that were canceled by
deaths for the given years in Italy and Ger- many 220
44. The percentage which deaths were of births
steadily increased during the war years 223
45. Crepidula fornicata shows sex reversal 254
46. Mated males of Crepidula fornicata retain that
stage longer 256
47. Castes of the common honey-bee 260
48. Some ant castes 265
49. The brown locust of South Africa has a swarm
phase which is distinct from the solitary phase 273
Foreword
I WAS recently honored by an invitation to give the Norman Wait Harris lectures at Northwestern Uni- versity; the more so since as one of their side-door neighbors I live close enough for my personal foibles to be well known, thereby removing the chief source of any possible glamour. In this book which grew out of those lectures, as in the lecture series itself, I make no effort to pose as the remote purveyor of a mys- terious erudition; I could not in any case regard my- self as more than the exponent of the glorified com- mon sense which I more and more firmly believe all science should be.
Even more than most, this book is the outgrowth of years of co-operative effort. Some of the basic facts were collected with the aid of funds from the Rockefeller Foundation given to aid biological re- search at the University of Chicago. Other researches were supported directly by that university and more recently by a grant for the study of the effect of hormones on behavior from the National Research Council.
13
x\^
1 4 FOREWORD
In addition to the personal aid received from my scientific associates, many of whom will be named in the text, the kindly criticism of Professor Alfred E. Emerson has been particularly helpful in developing the work and in shaping the content and implica- tions of these lectures. My thanks are given also to Professor Sewall Wright for his criticism of Chapter IV, to Mr. Kenji Toda for preparing the illustra- tions and to Marjorie Hill Allee, whose command of the written word has been a constant resource.
W. C. Allee The University of Chicago.
Mm Science versus Metaphysics
THE RATE of obsolescence of material things is high. With consumers' goods we are well aware of this fact; and even capital goods usually become out of date in a long generation. Last summer an admirer of Will Rogers dedicated a lasting monument to the humorist. Although built for time and erected in our semi-arid West where decay is slow, the tower is ex- pected to last only a thousand years. Invested capital evaporates even with watchful care; there are few private collections of material wealth that remain in- tact a third of a thousand years.
Oddly enough, the most permanent contributions of our age appear to be the scientific discoveries we have made, the artistic beauties we have created, and the ideas we have evolved. To the extent that these advances are entombed in libraries and museums they share the impermanence of more material things. A nearer approach to immortality is per- mitted those bits of science and art that escape from the bindings of books and pass into the active life
IS
l6 THE SOCIAL LIFE OF ANIMALS
and traditions of people. The more widespread and firmly fixed these become in the minds of living men, the greater is their chance of longevity.
The most practical achievement of our extremely practical period is the habit of searching for new truths and for correct interpretations of those long known. The unique contribution of the present era is not that made by men of business and affairs, spec- tacular as it may be. Rather this age is and will be known as the time of the development and ap- plication of scientific methods. These contributions are being made by extremely impractical research workers who are supported by a tiny splinter from the great block of capital gains. Money spent effec- tively to this end, whether in the aid of research or other creative scholarship, or in teaching the results gained by research, makes the most lasting and im- portant of all modern investments. The most nearly permanent monument any man can erect is to have influenced directly or indirectly the growth of im- proved ideas and traditions among the men in the street, in the factory or on the farm.
It is in this spirit that I have undertaken to inter- pret one of the significant biological developments of recent years. It is my hope that from the work described in these pages, all social action may have a somewhat broader and more intelligent foundation.
SCIENCE VERSUS METAPHYSICS 17
We can gain the impression from some modern oversimplifications that science deals with empirical facts, that philosophy attends to principles and eternal truths, and that religion is concerned with values. In the following pages it will be necessary to shake aside such artificial limits and to present principles along with the evidence that supports them; to test these against experience and to attempt frequently to weigh the general biological values involved. This last process will be easier if we assay survival values only. Admittedly in dealing even with survival values we must be relatively rough and ready in our methods, and perhaps the conclu- sions will carry a strong odor of the laboratory in which they had their origin.
Basically the approach will be that of the experi- mental biologist rather than that of the theorist, which might be more polished, or of the philoso- pher, which would certainly be more abstract and would probably use a great many more words for the same number of ideas. Despite much practice to the contrary, any biological fact which concerns us can be accurately described and the conclusions from its study be clearly expressed in relatively sim- ple and direct language.
In research reports and scholarly discussions there is need for the conciseness and precision made pos-
l8 THE SOCIAL LIFE OF ANIMALS
sible by technical language. Science has no need, however, and is ill-served by any tendency to de- velop a cult of obscurity. Scientists must be free to attack the unknown as effectively as they can and in return for intellectual freedom they have an obligation, which rests heavily on those able to do so, to interpret research results in terms which can be understood by intelligent and interested people. There is current in at least one American uni- versity at present an attempt to organize all knowl- edge about metaphysics, and so secure a longed-for unity. In order to obtain a simplified system, the group of men occupied with this enterprise turn back to the days before the present scientific era to find a statement of eternal principles which will serve as a unifying nucleus for human experience and thought. Such efforts at establishing a Neo- Scholastic philosophy, while furnishing an excellent corrective for overconfident scientists, seem mis- chievously naive as a serious, present-day movement. We do need relief from our absorbed attention to conflicting scientific detail, but progress must needs come from newer syntheses which take account of the world and man as science sees them rather than by accepting almost as a whole some ancient system of historical significance. These systems are out of date primarily because they were evolved before
SCIENCE VERSUS METAPHYSICS IQ
one of the greatest advances in knowledge that man has yet been able to make, that of modern science.
Modern philosophical educational systems, if they are to survive, must have as their central core the well-tested evidence compiled by objective scientific methods. Such knowledge must have stood the test of being checked and re-checked by men constitu- tionally agnostic in their mental attitudes; who can say, "I don't know. What is the evidence?"; who are constantly seeking critical new evidence concerning the validity of their ideas.
An anecdote that is becoming classic among scien- tists will illustrate the point. Professor Wood, phys- icist of Johns Hopkins, was asked to respond to the toast "Physics and Metaphysics" at a dinner of some philosophical society. His response was somewhat as follows:
The physicist gets an idea which seems to him to be good. The more he mulls over it the better the idea appears. He goes to the library and reads on the subject and the more he reads the more truth he can see in his idea. Finally he devises an experi- mental test and goes to his laboratory to apply it. As a result of long and careful experimental check- ing he discards the idea as worthless. "Unfortu- nately," Professor Wood is said to have concluded, "the metaphysician has no laboratory."
History and Natural History
LIKE other human disciplines, science has its or- thodox and its heterodox views. The idea that un- conscious automatic co-operation exists has had a long history, and yet it is just now beginning to escape from the heterodox category.
My own interest in this subject dates not from a preconceived idea but from a clearly remembered bump against some stubborn experiments. Almost thirty years ago as a graduate student in zoology I was engaged in studying the behavior of some com- mon small fresh-water animals called isopods, tiny relatives of the crayfish. All fall and winter I had been collecting them from quiet mud-bottomed ponds, chopping the ice if necessary, and from be- neath stones and under leaves in clear small streams.
I kept them in the laboratory under conditions which resembled those in which they lived in na- ture. Then day after day I put lots of five or ten isopods into shallow water in a round pan that had a sanded wax bottom so prepared that the isopods
20
HISTORY AND NATURAL HISTORY 21
could crawl about readily. When a current was stirred in the water the isopods from the streams usually headed against it; but those from ponds were more likely to head down current, or to be indif- ferent in their reaction to the current. The behavior of the two types was sufficiently different so that at first I thought that stream and pond isopods repre- sented different species, but the specialist at the National Museum assured me that all belonged to the species appropriately called Asellus communis, the commonest isopod of our inland waters.
Rather cockily I reported after a time to my in- structor that I had gained control of the reaction of these animals to a water current. By the judicious use of oxygen in the water, I could send the indif- ferent pond isopods hauling themselves upstream, or I could reduce the stream isopods to going with the current. I had not reckoned with another factor that presently caught up with me.
After a winter in the laboratory it seemed wise as well as pleasant to take my pan out to a comfortable streamside one sunny April day, and there check the behavior of freshly collected isopods in water dipped from the brook in which they had been living. To my surprise, the stream isopods, whose fellows all winter had gone against the current, now went steadily downstream or cut across it at any angle to
22 THE SOCIAL LIFE OF ANIMALS
reach another near-by isopod. When I used five or ten individuals at a time, as I had done in the labo- ratory, they piled together in small close clusters that rolled over and over in the gentle current. Only by testing them singly could I get away from this group behavior and obtain a response to the current; and even this reaction was disconcertingly erratic.
It took another year of hard work to get this con- tradictory behavior even approximately untangled; (i) * to find under what conditions the attraction of the group is automatically more impelling than keep- ing footing in the stream; and that was only the beginning of the road that I have kept from that April day to this time, continuing to be increasingly absorbed in the problems of group behavior and other mass reactions, not only of isopods, but of all kinds of animals.
As the years have gone on, aided by student and other collaborators and by the work of independent investigators, I have tried to explore experimentally the implications of group actions of animals. And necessarily, too, I have had to turn to the world's literature to find what others have done and are doing along this line.
* Detailed citations to more complete statements will be found in the bibliography.
HISTORY AND NATURAL HISTORY 23
A Greek philosopher named Empedocles seems to have had the first recorded glimmerings of an idea of the universal and fundamental nature of co-operation which underlies group action, as well as a conception of the opposing principle of the struggle for existence. Empedocles lived in the fifth century B.C., and he was not only a thinker but so much a man of affairs that he was offered a king's crown, which he refused. (128)
He owes his present-day fame to two long poems in which he outlined the idea that there are natural elements: fire, earth, air, and water, which are acted upon by the combining power of love and the dis- rupting power of hate. Under the guidance of the building force of love the separate elements came together and formed the world. Separate parts of plants and various unassorted pieces of animals arose from the earth. These, Empedocles taught, were often combined and at first the results were mon- strous shapes, which in time became straightened around until, still guided by combining love, they clicked, to make the more perfect animals we now know. It has taken us almost two and a half mil- lennia to transmute this poetic conception into the less picturesque but more exact and workable ex- pression acceptable to modern science.
After the fertile Greek era there intervened in this
24 THE SOCIAL LIFE OF ANIMALS
field as elsewhere the long sterile period when Greek philosophy, if known, was dogmatically accepted, and shared with other authoritarian systems the re- sponsibility of explaining the world of reality as well as the universe of fancy.
It was not until my own experiments and think- ing and reading had begun to form in my mind a fairly definite pattern that, by the aid of Havelock Ellis's The Dance of Life (43) I stumbled upon the ideas of the third Earl of Shaftesbury, who lived be- fore and after 1700. He seems to have been the first intellectual in the modern period to recognize fairly clearly that nature presents a racial impulse that has regard for others, as well as a drive for individual self-preservation; that, in fact, there are racial drives that go beyond personal advantage, and can only be explained by their advantage to the group.
An unfriendly contemporary wrote pretty much these words: "Shaftesbury seems to require and ex- pect goodness in his species as we do a sweet taste in grapes and China oranges, of which, if any are sour, we boldly proclaim that they are not come to their accustomed perfection." Havelock Ellis, in reviewing this development, says that "therewith 'goodness* was seen practically for the first time in the modern period to be as 'natural' as the sweetness of ripe fruit." It is only fair to record that in the religious
HISTORY AND NATURAL HISTORY 25
world for at least fifty years previous there had been growing a similar conviction among certain heretics.
In 1930, after having written the text of a care- ful account of experimental evidence concerning the existence and non-existence of co-operation at sub- social levels, (3) I set down in the draft of a proposed preface that the existence of such a principle was now for the first time an established fact, for which the details to follow gave full proof. I still think that the proof is good. However, the preface as published does not contain any such claim, for at that point in the writing I went back and re-read Des societes animales by the French scientist Es- pinas, (44) which appeared in 1878 and which was the pioneering essay in this field so far as modern work is concerned. There I found Espinas affirming that no living being is solitary, but that from the lowest to the highest each is normally immersed in some sort of social life, a fact which he proclaimed sixty years ago, and added that he was ready to offer conclusive proof.
I turned through the pages past his detailed his- tory of the evolution of ideas about the origin and development of society, and read his massed evi- dence that communal life is not "a restricted acci- dental condition found only among such privileged
26 THE SOCIAL LIFE OF ANIMALS
species as bees, ants, beavers and men, but is in fact universal."
The evidence was largely based on observations of the existence of animal groupings in nature, which are found widely distributed in the different levels of the animal kingdom— facts such as I shall review later in this chapter. It was clear to me that the facts which Espinas had found so impressive had not convinced others and, while suggestive, did not seem compelling to me in the light of other indica- tions to the contrary. Perhaps, I cautioned myself, even the experimental evidence that I had accumu- lated in 1930 was not really crucial, and it would be well to avoid making too strong a claim in the matter. The same caution must continue even in the face of still stronger evidence known today.
The conclusions of Espinas coming in 1878 are the more important because the scientific world was then, as men in the street are today, under the spell of the idea that there is an intense and frequently very personal struggle for existence so important and far reaching as to leave no room for so-called softer philosophies. The idea of the existence of natural co-operation was apparently in the air despite the preoccupation with this phase of Darwinism. Kessler is said to have addressed a Russian congress of natu- ralists on this subject in 1880, and from this ad-
HISTORY AND NATURAL HISTORY 2?
dress sprang the remarkable if uncritical book by the Russian anarchist, Prince Kropotkin, on mutual aid. (74)
By combing the accumulated natural history rec- ords, Kropotkin was able to collect observation after observation which indicated that animals in nature do aid each other to live, as well as, on occasion, kill each other off. Kropotkin's work served the admi- rable purpose of keeping this idea alive and popu- larizing it. It has had also the less fortunate result of bringing Kropotkin's fundamental doctrine into disrepute among students who are critically sensi- tive to the value of evidence, and who find that Kropotkin's sources were not always reliable.
William Patten, an American biologist who taught for many years at Dartmouth College, made the next general statement of the fundamental nature of co- operation when in 1920 he gave it a central place in his analysis of the grand strategy of evolution, (go) It is of personal interest to me that at the scientific meetings in 1919 at which I presented my first ex- perimental results on this subject, Professor Patten gave a vice-presidential address in which he outlined, mainly from philosophical considerations, his con- clusions concerning the importance of biological co-operation. He was rightly impressed by the fact that cells originally were separate, as protozoans are
28 THE SOCIAL LIFE OF ANIMALS
today. Some, however, evolved the habit of remain- ing attached together after division. This made a beginning from which the many-celled higher ani- mals could develop. With each increase in the ability of cells to co-operate together there came power to increase the complexity of organization of the cell masses. The highly evolved bodies of men and of insects are thus an expression of increasing inter- cellular co-operation which finally reaches a point at which, for many purposes, the individual person becomes the unit rather than the co-operating cells of which he is composed.
About the same time the German, Deegener, (40) published an extensive treatise on the social life of animals, along the same lines as the book written by Espinas forty years before. Deegener 's distinctive contribution was a classification of the different social levels, from the simplest sorts of artificial col- lections of animals to parasitism and truly social life. His rating of these different aspects of sub-social and social life in one long outline has the great merit of showing that there are no hard and fast lines which can be drawn between social and sub- social organisms, but that social communities are the natural outgrowth of sub-social groupings. Un- fortunately, wdth Teutonic vigor and vocabulary, he designated the different categories in words as
HISTORY AND NATURAL HISTORY 29
unwieldy as they were exact. Bogged down by the weight of such terms as sympatrogynopaedium, syn- aporium and heterosymphagopaedium, Deegener's real contribution tends to be lost even to biological scholars.
A survey such as I am attempting here should not try to be exhaustive; I shall dismiss with a word the slight advance made by Alverdes (16) and the work of many others without that. There is, however, another phase of the literature whose reading has given me so much pleasure as well as useful infor- mation that I shall not pass it over: this deals with the social insects. Espinas, Kropotkin, Deegener and Alverdes of those mentioned, and a host of others, have written in detail and in general about these fascinating insects, but none more accurately or with greater insight and literary as well as scientific skill than the American entomologist, William Morton Wheeler. His book on Social Life Among the Insects, which appeared in 1923, is a noteworthy general summary. (120) In this he shows that among insects alone, and including such well-known forms as termites, bees, wasps and ants, and the less gen- erally known social beetles, the social habit has arisen some twenty-four distinct times in about one- fifth of the known major divisions of insects. It would seem that there is a general reservoir of pre-
30 THE SOCIAL LIFE OF ANIMALS
social traits from which, given the proper opportu- nity, society readily emerges. Wheeler, no less than Espinas, from whom he quotes, emphasizes that even so-called solitary species of animals are of necessity more or less co-operative members of associations of animals and that animals not only compete among themselves but they also co-operate with each other to secure mates and insure greater safety.
It did not, however, make for the full acceptance of these ideas that Wheeler drew his illustrative material primarily from, and based his conclusions mainly on, his knowledge of social life among in- sects. The existence of co-operation among nest mates in ants and bees does not prove that there are beginnings of co-operative processes among amoebae and other greatly generalized animals.
Man and the few species of highly social insects are a small part of the animal kingdom; in order to discover and distinguish the principles of general sociology it is necessary to look farther, to focus attention on the social and anti-social relationships of many animals usually regarded as lacking social life.
With and without this end in view there have been in the last twenty years simultaneous but inde- pendent outbreaks of experimentation on group effects among the lower animals. For a time just
HISTORY AND NATURAL HISTORY 31
preceding and following 1920 we, who in Aus- tralia, (107) in France (26) and in the United States (2) were engaged in these studies, continued unaware of each other's work. Relatively soon, how- ever, since biological world literature is today widely and promptly circulated, all such work, even that in Russia, (53) became generally known. It is these general experiments on population growth, on mass physiology and on animal aggregations, that are now the important aspect of the field of animal co- operation.
I have briefly traced here the history of the idea of innate co-operation. One reason for the slowness of accepting that idea is the obvious fact that co- operation is not always plain to the eye, and that competition in its most non-co-operative form, in which no social values are apparent, can readily be observed. With certain exceptions to be nientioned soon, it has seemed that, social species aside, crowd- ing, the simplest start toward social life which is easily apparent and a condition of nearly all society, was harmful alike to the individual and to the race. It has been known from experimental evidence since 1854 (62) that crowded animals may not grow at all, or, at any rate, gi-ow less rapidly than their uncrowded brothers and sisters. And under many conditions crowded animals not only do not grow.
?^a^ii
30
THE SOCIAL LIFE OF ANIMALS
social traits from which, given the proper opportu- nity, society readily emerges. Wheeler, no less than Espinas, from whom he quotes, emphasizes that even so-called solitary species of animals are of necessity more or less co-operative members of associations of animals and that animals not only compete among themselves but they also co-operate with each other to secure mates and insure greater safety.
It did not, however, make for the full acceptance of these ideas that Wheeler drew his illustrative material primarily from, and based his conclusions mainly on, his knowledge of social life among in- sects. The existence of co-operation among nest mates in ants and bees does not prove that there are beginnings of co-operative processes among amoebae and other greatly generalized animals.
Man and the few species of highly social insects are a small part of the animal kingdom; in order to discover and distinguish the principles of general sociology it is necessary to look farther, to focus attention on the social and anti-social relationships of many animals usually regarded as lacking social life.
With and without this end in view there have been in the last twenty years simultaneous but inde- pendent outbreaks of experimentation on group effects among the lower animals. For a time just
i
HISTORY AND NATURAL HISTORY
31
preceding and following 1920 we, who in Aus- tralia, (107) in France (26) and in the United States (2) were engaged in these studies, continued unaware of each other's work. Relatively soon, how- ever, since biological world literature is today widely and promptly circulated, all such work, even that in Russia, (53) became generally known. It is these general experiments on population growth, on mass physiology and on animal aggregations, that are now the important aspect of the field of animal co- operation.
I have briefly traced here the history of the idea of innate co-operation. One reason for the slowness of accepting that idea is the obvious fact that co- operation is not always plain to the eye, and that competition in its most non-co-operative form, in which no social values are apparent, can readily be observed. With certain exceptions to be nientioned soon, it has seemed that, social species aside, crowd- ing, the simplest start toward social life which is easily apparent and a condition of nearly all society, was harmful alike to the individual and to the race. It has been known from experimental evidence since 1854 (62) that crowded animals may not grow at all, or, at any rate, grow less rapidly than their uncrowded brothers and sisters. And under many conditions crowded animals not only do not grow,
32 THE SOCIAL LIFE OF ANIMALS
they die more readily, and frequently they repro- duce less rapidly than if living in uncrowded popu- lations.
All the older works in natural history taught fairly clearly that crowded groups, to have real sur- vival values, must be sufficiently well organized to contribute to group safety by warning of danger or by defense in case of attack. (3) If, in addition, these groups are organized on a basis of division of labor, such as occurs in the highly social colonies of ants or termites, with specialized reproductives, workers and soldiers, or according to the patterns found in human society, then the survival values of groups are readily seen.
Yet for some reason, under natural conditions and with very many sorts of animals, crowding in all degrees does occur and apparently always has oc- curred. Conceded that animals do not always act for their own best interests, still they must do so to a certain degree or be exterminated in the long run. The advantages of the long-established habit of a species may not be obviously apparent, but it is not safe to say offhand that advantages do not exist.
There are the dense crowds of certain animals, ladybird beetles (Plate la), for example, that with the approach of winter collect in restricted and fa- vorable places where they hibernate together. Ap-
PLATE I. a. Ladybird beetles cellect in dense ag- gregations in the autumn and hibernate. /;. During their breeding season, male midges gather in swarms and await the coming of .the females. (Photographs by Welty.)
HISTORY AND NATURAL HISTORY 33
parently, in the face of winter cold, there is some safety in numbers even among cold-blooded animals that collect in hordes without any organization.
A second plain exception to the general testimony that crowding of non-social species is harmful are the aggregations that form during the breeding sea- son. Like the hibernating groups, these are very widely distributed through the animal kingdom. Breeding aggregations of worms, crustaceans, fishes, frogs, snakes, birds and mammals or the midge in- sects shown in Plate lb, for example, have long at- tracted attention. Their numbers have been great enough and conspicuous enough to stimulate re- peated descriptions by naturalists.
A third exception is found during times of migra- tion, when animals frequently crowd together in great hordes and execute mass migratory movements, like those of many birds.
However, breeding, hibernation and migration aside, the older information indicated that up until the point that social life is developed, crowding is harmful.
But there are many other instances of crowding which do not fall under any of these classifications; and it will be worth while to consider here the ex- tent and the natural history of some of these dense animal aggregations. Here, as elsewhere, there will
34 THE SOCIAL LIFE OF ANIMALS
be no attempt to catalogue all known instances or to select merely the very best cases known. I shall try to use examples that are not too shopworn by repeated description.
Almost every observant person has seen the soft green "bloom" which covers many stagnant ponds. Under the microscope this "bloom" is often seen to be composed of myriads of the tiny plant-animal Euglena. These organisms are commonly one-tenth of a millimeter long, which means that in a char- acteristic layer of "bloom" there would be at least sixty to one hundred thousand animals per square inch; and acres of water are sometimes covered.
Lobster-krills are small crustaceans that occur com- monly in shoals about the Falkland Islands, Pata- gonia, New Zealand and other southern waters. (81) A larval stage of this animal, less than an inch long, occurs often on the surface of the water in such numbers that the sea is red for acres; and whales in those waters simply open their mouths and swim through slowly, feeding with no more effort than the process of straining them out. These shrimp- like animals may be piled up on the shore by tide and wind in stench-producing layers. Dampier wrote of them in 1700: "We saw great sholes of small lob- sters, which colored the sea red in spots for a mile in compass"; and they have been known to extend
HISTORY AND NATURAL HISTORY 35
along the Patagonian coast for as much as three hun- dred miles.
At Woods Hole, on Cape Cod, I have at certain seasons dipped up a bucket of sea water from the harbor and found more space occupied by clear, jelly-like ctenophores, each the size of a walnut, than was taken by water. Sometimes I have dipped up a fingerbowl of sea water and found it so filled with small pin-point-like copepods that again there seemed to be more of them than of the water itself. These tiny relatives of the lobster-krills are also the food of whales, and they, too, may discolor the ocean for miles.
Around bodies of fresh water, may-flies or midges may emerge in clouds. At Put-in-Bay, near the lights flooding the monument that commemorates Perry's victory, I have picked up living may-flies by the double handfuls from the millions that fly to- ward the lights; and near by our lake boat steamed through windrows of cast skins of the emerging may- fly nymphs. Nearer Chicago I have taken water isopods, the half-inch crustaceans mentioned earlier, by the bucketfuls from pools where they had col- lected in numbers only to be compared with those in twenty swarms of bees.
We have already spoken of the migratory hordes. Locusts in migration (116) swarm out of the sky in
36 THE SOCIAL LIFE OF ANIMALS
the Sahara borderlands, in southern Russia, in South Africa and on the Malay Peninsula in ter- rorizing numbers (Figure 1). They once did so on the Great Plains of the United States, leaving a lively memory of destruction that is still roused by the smaller migrations that may occur there any summer
^''' ^ ■.■:,:■:. . ■■ . -•■•■■■•• \
Fig. 1. A band of grasshopper nymphs on the march. (From Uvarov, by permission of the Imperial Bureau of Entomology.)
in spite of active control measures. I myself have seen the so-called Mormon cricket advancing from the relatively barren mountain pastures of Utah into the green fields in numbers which were not halted by the hawks, turkeys and snakes attendant on the swarm and feeding greedily; or the active assaults of men and children warned out to protect the cultivated lands. Migrating army worms and chinch bugs present equally impressive aggregations. The emergence of Mexican free-tailed bats from the Carlsbad cave of an August evening has been described as a black cloud pouring out in such den- sity as to be visible two miles away. (19) Such bats
HISTORY AND NATURAL HISTORY 37
are estimated to hibernate in these caves by the milHons; and they may be found through the day in sleeping masses a yard across, hanging from the roof like a swarm of bees.
Even larger mammals may collect into great, closely packed herds. The migrating caribou on the tundra are said to pour south in hordes that flow past a given point for hours or even for days. And of the antelope on the plains of Mongolia, (17) Roy Chapman Andrews says that he has seen thou- sands upon thousands of bucks, does, and fawns pour over the rim and spread out on the plain. Sometimes a thousand, more or less, would dash away from the fierd, only to stop abruptly and feed. The mass of antelope were in constant motion even when the animals were undisturbed. They scattered before his automobile only to re-form within a few hours. In that region only the grassland antelope gathers in such immense herds; the long-tailed desert species never does so, probably because there is not enough food to support them in their more arid dwelling place.
These are merely a few of the more dramatic instances of the collection of great masses of animals in a small space. They are more spectacular but probably less important than are the innumerable smaller aggregations of animals which are frequently
38 THE SOCIAL LIFE OF ANIMALS
encountered. The small dense crowds of whirligig beetles are a case in point. These occur in wide- spread abundance on the surface of our inland waters.
The more common condition of less intense crowd- ing does not mean that animals are usually solitary. Rather, the growing weight of evidence indicates that animals are rarely solitary; that they are almost necessarily members of loosely integrated racial and interracial communities, in part woven together by environmental factors, and in part by mutual attrac- tion between the individual members of the different communities, no one of which can be affected with- out changing all the rest, at least to some slight extent.
Let us take an example. Before the coming of the white man, and even a century ago or less, much of the Great Plains was occupied by what ecologists call a grassland-bison community. (4) Grasses could readily grow in the rich soil, even with the usual summer dry spells and the more severe cyclic drouths that occurred even then. By keeping the grasses fairly closely cropped the bison herds pre- vented the invasion of herbs and shrubs that might have withstood the severities of the climate but could not make headway against continual grazing (Plate II). In this function the bison were joined by
PLATE II. A giasslancl-bison community. (Photo- graph from the National Park Board of Canada.)
HISTORY AND NATURAL HISTORY 39
a myriad of grasshoppers, crickets, meadow mice and prairie dogs. All these were key-industry ani- mals. In one way or another they converted the grass into meat of different sorts, on which the plains Indians, buffalo wolves, haw^ks, owls, and prairie chickens fed. If the grass failed, then many of the key-industry herb-eaters and those that in turn fed on them must either starve, migrate into another community where they would be disturbing factors, or change their source of food and thereby disturb the balance in their own community.
It must be pointed out here that the plants of this community cannot be set off as separate from the animals. They divide the available space between them; they constantly interact upon each other and upon their physical environment; except for pur- poses of formal study or in limited fields, the biolo- gist must consider both as members of a given association.
In such a community the effects of the dominant bison were felt in times of stress by the humblest and least conspicuous grasshopper. In the spring of the year hundreds of square miles normally sup- ported populations of six to ten million insects and other invertebrate animals for every acre of land. As with warmer weather the predatory animals re- turned to the grasslands, these insects were eaten off
40 THE SOCIAL LIFE OF ANIMALS
until perhaps a tenth of their number could be found later in the season; with the autumn lushness they increased again, only to fall back to some half- million or so per acre during the winter cold.
Similar communities exist among aquatic forms. In fact one of the first demonstrations of such a community was made for the animals living in and on an oyster-bank. (82) A beautiful and penetrating description of the interrelations that may be found in a small lake was published not long after by the late Professor Forbes (48) of the Illinois Biological Survey, in which he pointed out that minnows com- peted with bladderwort plants for key-industry or- ganisms; and showed that when a black bass is hooked and taken from the water the triumphant fisherman is breaking, unsensed by him, myriads of meshes which have bound the fish to all of the dif- ferent forms of lake life.
The existence of these communities is now gen- erally recognized, and in order that they may exist it seems that there must be a far-reaching, even if vague and wholly unconscious, co-operation among all the living creatures of the community. It is to such relationships that Wheeler referred when he said, "Even the so-called solitary species are neces- sarily more or less co-operative members of groups or associations of animals of different species."
HISTORY AND NATURAL HISTORY 41
Within these communities aggregations of animals occur for a variety of reasons. Their nature can best be shown by a series of illustrations.
One variety of aggregations is that of colonial forms, in which many different so-called individuals remain through life permanently attached together. In the simplest cases all the individuals are alike. Each possesses a mouth and food-catching tentacles, and each feeds primarily for itself, although food caught by one individual may be shared with others near by. In more complex forms some individuals have the mouths suppressed, and receive all their food from those that do take food. They have be- come specialized as bearers of batteries of stinging cells; they strike actively when the colony is touched, and their stinging cells explode so effectively as to give protection to the colony. Other individuals in the same colony bear medusa-like heads which break away and swim off, producing eggs and sperm, dis- tributing them as they drift. Here is certainly a divi- sion of labor though these colonial animals are never rated as social.
Various modifications of such colonial animals are found particularly among the colonial protozoa, sponges and the coelenterates; they also occur higher in the animal kingdom, even among the lower chordates, the great phylum to which man himself
42 THE SOCIAL LIFE OF ANIMALS
belongs. It is interesting that animals whose struc- ture forces them to the sort of compulsory mutual aid that automatically follows such structural con- tinuity have never progressed far either in social achievement or in the evolutionary scale. When higher animals, such as the lower chordates, show this development they are usually regarded as de- generate members of their general stock. These colonial animals are seldom dominant elements in the major communities of which they are a part. One comes to the conclusion that the more nearly voluntary such co-operation is, the greater its ad- vantage in social life. It might on the other hand be pointed out that when an animal has achieved social organization and division of labor low in the evolutionary scale, the resulting colonies are so well adapted to their environment that there is not suffi- cient pressure to cause evolutionary changes.
A second type of aggregation occurs when animals are forced together willy-nilly by the action of wind or tidal currents or waves over which they have no control, and whose effects they cannot resist. Many of the masses which lend color to wide patches of the ocean surface are brought together by tempo- rary or permanent currents. Often animals so dis- tributed are thrown down more or less by chance on types of bottom on which they can develop, and
HISTORY AND NATURAL HISTORY 43
there, if favorable niches are somewhat rare, dense aggregations may result, like New England coral on a suitably hard bottom, or the animals found on a wharf piling.
These accidental animal groupings may persist only as long as the physical forces which brought them together continue to act. Usually, however, they last somewhat longer, as a result of a slightly positive social inertia which tends to keep animals concentrated in whatever place they happen to be found. If the groupings are to have much perma- nence this quality of social inertia, the tendency of animals to continue repeating the same action in the same place, must be reinforced by another quality: the social force of toleration for the pres- ence of others in a limited space. The densely packed communities of animals on a wharf piling can per- sist only if toleration for crowding is well developed.
Other dense collections may be brought about by forced movements of animals in response to some orienting influence in their environment. These oriented, compelled reactions are frequently called tropisms. They are shown by the moths or June beetles or may-flies that collect about lights. Such aggregations are a result of the inherited, internal organization of the animals; and the irresistible at- traction of the may-fly to the light is joined with
44 THE SOCIAL LIFE OF ANIMALS
active toleration for the close proximity of others.
Similarly close aggregations occur as a result of the less spectacular trial and error reactions, in which the animals wander here and there, more or less vaguely stimulated by internal physiological states or external conditions, and so come to collect in favorable locations. Collections of animals about limited sources of food give a good illustration. These, too, may show only the social qualities of inertia and toleration.
A decided advance is made when animals react positively to each other and so actively collect to- gether, not primarily because the location is favor- able or through environmental compulsion, but as the result of the beginnings of a social appetite. In early stages of such reactions, the movement together may come primarily because the collection of isopods or earthworms or starfishes are substitutes for miss- ing elements in the environment.
Take, for example, the snake or brittle starfishes of the New England coast. These are rare now along Cape Cod, but before the wasting disease swept away the eel grass they were abundant in favorable locali- ties, but were rarely found close together. I have spent hours peering down through a glass-bottomed bucket here and there and round about in one of these localities, and have not seen more than one
PLATE III. a. Brittle starfish aggregate readily when put into a bare vessel of sea watei . b shows con- ditions ten minutes after a was taken. (Photographs by Welty.)
HISTORY AND NATURAL HISTORY 45
at a time. And I have spent more hours wielding a sturdy garden rake in swathe after swathe through the short eel grass, very rarely pulling in more than one starfish at a haul.
Yet when a few brittle starfishes are placed in a clean bucket of sea water they clump together like magic (Plate III). In bare laboratory aquaria they remain so clumped for weeks; in fact the aggrega- tions become more compact as time goes on as the animals bring back their extending arms and tuck them into the mass. If, however, the aquaria are dressed up by the introduction of eel grass so that conditions approach those found in nature, the ag- gregations disperse and the starfishes climb actively about over the blades of the eel grass, feeding on organisms and debris found on their surfaces.
The idea that in clean laboratory dishes these star- fishes are substituting each other for the missing eel grass was obvious and easy to test. A kind of artifi- cial eel grass was made of glass rods twisted in various shapes so that they offered a supporting framework for climbing in much the same way as the true eel grass. So long as the rods remained the starfishes clambered about over the meshwork or hung motionless, usually isolated. If the rods were removed they again clustered together.
As I have said elsewhere, (3) it is a far cry from
46 THE SOCIAL LIFE OF ANIMALS
such aggregations to the groupings of foreigners in a strange city that result in Little Italy, or the Mexican settlement, or a German quarter; and yet basically some of the factors involved are similar. Perhaps there is a closer connection between such aggregations in the wide expanse of a clean aqua- rium and the schooling tendency found among many fishes of the open sea; perhaps the same phe- nomenon accounts for the flocking tendency of many birds, as well as mammals on the equally monotonous grassy seas of temperate plains.
A somewhat different expression of a positive social reaction is shown when animals that are usually more or less isolated come together and pass the night grouped as though they were engaged in a slumber party. This type of behavior has been repeatedly described for different insects, even for the wasps that remain separate to such an extent that they are called solitary wasps. In some forms of solitary wasps both males and females are found in the sleeping group. With solitary bees, such as we have near Chicago, the overnight aggregations are composed of males only. A study which was made of the sleeping habits of a Florida butterfly species indicates that these Heliconii (69) come together night after night in the same location, in part at least as a result of place-memory. The assemblages
HISTORY AND NATURAL HISTORY 47
lack sexual significance. There is some protection in the fact that if one is disturbed the whole group may be warned. The presence of many butterflies would reinforce any species odor that might attract others of the same species, or repel possible predators.
The crowded roosts to which certain birds return not only for one season but sometimes for years are widely known. Here again we are concerned with a positive social appetite which grows stronger with the approach of darkness; the details as to why and how it operates are not known.
Animals which come together in intermittent groupings like these overnight aggregations are showing a social appetite which is none the less real because it is effective only at spaced intervals. In this it resembles other appetites such as those for food, water and sex relations. From such occa- sional or cyclic expressions of a social appetite it is a relatively short step to whole modes of life which are dominated by a drive for social relationships. As I have already said, in the insects alone this step has been taken some twenty-four distinct times and in widely separated divisions of that immense group.
Normally the development of highly social life comes by way of an extension of sexual and family relations over greater portions of the life span. Here again all degrees of increased length of asso-
48 THE SOCIAL LIFE OF ANIMALS
elation can be shown, from the sexual forms that meet but once and for a brief moment to the ter- mite kings and queens that live together for years. Also all stages exist in the evolution of the associa- tion of parents with offspring, from the insects like the female walking-stick, which deposits eggs as she moves about and pays no more attention to them, to the ants and bees whose worker offspring spend their entire lives in the parental colony or some colony budding off from it.
While the extension of family relations is very obviously one potent method by which social life is developed to a high level, there are other social groupings which also deserve consideration in con- nection with the problem as to the method of evo- lution of social life. Schools of fish arise, for exam- ple, under conditions in which there is no associa- tion with either parent after the eggs are laid. At times the eggs may be so scattered in the laying that the schools form from unrelated individuals. Here the schooling tendency seems to underlie rather than grow out of family life. The mixed flocks (22) of tropical birds which are composed of many spe- cies obviously did not grow directly from family gatherings, and the groups of stags of Scottish deer, probably the original stag parties, (38) appear to give evidence of a grouping tendency independent
HISTORY AND NATURAL HISTORY 49
of intersexual or family relations. This subject will be discussed in more detail in the final chapter.
The conclusion seems inescapable that the more closely-knit societies arose from some sort of simple aggregation, frequently, but not necessarily, solely of the sexual-familial pattern. Such an evolution could come about most readily with the existence of an underlying pervasive element of unconscious co-operation, or automatic tendency toward mutual aid among animals.
In the simpler aggregations evidence for the pres- ence of such co-operation comes from the demon- stration of survival values for the group. These are more impressive the more constant they are found to be. If they exist throughout the year they are much more important as social forerunners than if present only during the mating season or at times of hibernation.
Ill
■ Beginnings of Co-operation
WITH this chapter I begin the presentation of the evidence for the assertion that there is a general prin- ciple of automatic co-operation which is one of the fundamental biological principles. The simplest ex- pression of this is often found in the beneficial ef- fects of numbers of animals present in a population. Laboratory work of the last two decades still shows that overcrowding is harmful, but it has also uncov- ered a no less real, though somewhat slighter, set of ill effects of undercrowding.
To be sure, overcrowding always produces ill ef- fects, and these can always be demonstrated at some population density. On the other hand, the ill effects of undercrowding cannot always be shown, though frequently they can. In generalized curves the mat- ter may be summarized thus: Under certain condi- tions (g6) we find the curve running like the dia- gram in Figure 2 A, when height above base line gives the rate of the biological action being meas- ured, and distance to the right shows a steadily in-
50
BEGINNINGS OF CO-OPERATION 51
creasing population. Under these conditions only the ill effects of overcrowding are visible, and the optimum population is the lowest possible. This is the modern expression of what used to be called the struggle for existence. In the more poetic post-Dar- winian days this struggle was thought of as so in- tense and so personal that an improved fork in a bristle or a sharper claw or an oilier feather might turn the balance toward the favored animal. Now we find the struggle for existence mainly a matter of populations, measured in the long run only, and then by slight shifts in the ratio of births to deaths.
A second type of phenomena is represented by a curve with a hump near the middle (97) as shown in Figure 2B.
Again, height above the base line measures the speed of some essential biological process or proc- esses, such as longevity; distance to the right gives increasing population densities. The harmful effects of overcrowding, indicated by the long slope to the right, are still plainly evident, but there is also ap- parent a set of ill effects associated with undercrowd- ing which are shown by the downward slope to the left. Many have written pointedly about overcrowd- ing, and while there is still much to be learned in that field, it is in the recently demonstrated exist- ence of undercrowding, its mechanisms and its im-
52
THE SOCIAL LIFE OF ANIMALS
plications, that freshness lies. Without for one min- ute forgetting or minimizing the importance of the right-hand limb of the last curve, it is for the more romantic left-hand slope that I ask your attention.
Fig. 2. A. Under some conditions the rate of bio- logical action which is being measured is greatest with the smallest population, and decreases as the numbers increase. B. Under other conditions there is a distinct decrease in the rate of the measured biological reaction with undercrowding (to the left) as well as overcrowding (to the right).
Perhaps the simplest and most direct demonstra- tion of certain harmful effects of undercrowding comes from an experiment which I understand is carried on spontaneously among undergraduate men at certain universities and colleges of which X, or perhaps better, Y, is an example. A certain number of men gather together in a limited space under arti- ficial light and undertake to consume a more or less limited amount of stronger or weaker alcohol. If
BEGINNINGS OF CO-OPERATION 53
there are many men present in proportion to the amount of alcohol, relatively little or no harm will result from the experiment. If there are very few men and much alcohol there may be garage bills and other important repairs to be made.
In one way or another similar tests have been car- ried out in the laboratory with a variety of poisons, and many kinds of animals. Again I choose from the mass of available evidence the results of a simple and clean-cut experiment to illustrate the same point with non-human animals.
Everyone is acquainted with goldfish; they are hardy forms or else they would not be alive today in so many goldfish bowls. Colloidal silver in its commercial form of argyrol is also well known. Col- loidal silver, that is, the finely divided and dispersed suspension of metallic silver, is highly toxic to liv- ing things, including even the hardy goldfish.
In the experiment in our laboratory (8) we ex- posed sets of ten goldfish in one liter of colloidal silver, and at the same time placed sets of ten simi- lar goldfish, one each, in a whole liter of the same strength of the same suspension. This was repeated until we had killed seven lots of ten goldfish and their seventy accompanying but isolated fellows. Then when the results were thrown together we had the simple table on page 54.
t^- I
54 THE SOCIAL LIFE OF ANIMALS
TABLE I
Survival in minutes of goldfish in colloidal silver
NUMBER NUMBER DIFFERENCE STATISTICAL GROUPED ISOLATED PROBABILITY
7X lo 70x1 182 min. 507 min. 325 min. P < o.ooi
Any biological experiment has a large number of so-called variables, that is, of factors that it is diffi- cult or impossible to bring under such complete control that we can be certain that the experiment will be exactly repeatable next time. Hence it is customary to make experiments if possible as paired experiments, in which one set of conditions (those of the group in this instance) will differ from an- other lot (those of the isolated goldfish) only by the one difference, in this case of grouping and isolation. Such results with these fish can then be analyzed by statistical methods to find the probability of get- ting like results merely "by chance." These methods are now so simple that even I can make the calcula- tions. They are as accepted a technique as is the paired experiment.
With the goldfish there is less than one chance in a thousand of getting as great an average difference with the same number of trials. Technically we say that probability, or P, for short, is less than 0.0001.
BEGINNINGS OF CO-OPERATION 55
It means the same. Students of statistics have found that when P z= 0.05 or less, that is, when there are fewer than five chances in a hundred of such a thing happening as a result of random sampling or "chance," there is likely to be something significant in such results, the more so the smaller the fraction which P is said to equal.
We make such tests of our experimental results continually, to find how we are getting on, and I shall give probabilities repeatedly. In doing so it must be remembered that these test the data, not the theory— and that the data may vary significantly for unknown reasons, even when we think we are in full control of the situation; and that because there is only one chance in one hundred, or ten thousand, or a million that a thing may happen by "chance" does not mean that it will never happen through what we call an accident; merely that the chances of its happening so, our evidence being what it is, are on the order of one in one hundred, or ten thousand, or a million.
I will digress even further into the realm of coinci- dence. A Negro friend of mine spent a summer in Europe and while in Paris visited the art galleries of the Louvre. While there he saw a Negro woman busy looking at pictures and on coming closer dis- covered that she was his own aunt. Neither had any
56 THE SOCIAL LIFE OF ANIMALS
idea that the other was in Europe. With no pre- arrangement, what is the probability that an Ameri- can Negro from Chicago will meet his aunt in the Louvre? Yet it did happen this once without in any way shaking the probability principle.
Perhaps the digression is not so great as might ap- pear at first glance, for we need a slight common understanding of the practical working of statistical probability; all of modern science, the more as well as the less exact, is built on it.
To get back to our goldfish: those in the groups of ten lived decidedly longer than their fellows ex- posed singly to the same amount of the same poison; and significantly so. But why? Others had made that experiment w^ith smaller animals, and had decided that the group gave off a mutually protective secre- tion which would protect that particular species and none other. One reason that we were working with goldfish was because they are large enough so that we could use approved methods of chemical analysis in finding where the silver went. The balance sheet from such tests showed that we could account for all the silver present. With the suspensions which had held ten fish the silver was almost all precipi- tated, while in the beakers that had held but one fish almost all the silver was still suspended.
When exposed to the toxic colloidal silver the
BEGINNINGS OF CO-OPERATION 57
grouped fish shared between them a dose easily fatal for any one of them; the slime they secreted changed much of the silver into a less toxic form. In the ex- periment as set up the suspension was somewhat too strong for any to survive; with a weaker suspension some or all of the grouped animals would have lived; as it was, the group gained for its members a longer life. In nature, they could have had that many more minutes for rain to have diluted the water or some other disturbance to have cleared up the poison and given the fish a chance for complete recovery.
With other poisons, other mechanisms become effective in supplying group protection. Grouped Daphnia, (50) the active water fleas known to all amateur fish culturists, survive longer in over-alka- line solutions than daphnids isolated into the same volume. The reason here is simple. The grouped animals give off more carbon dioxide, and this neu- tralizes the alkali. Long before the isolated individual can accomplish this, it is dead; in the group those on the outside may succumb, though if the num- ber present is large enough even they may be able to live until the environment is brought under tem- porary control.
Frequently the protective mechanism is much more complex. With many aquatic animals, other things being equal, isolated animals consume more
58 THE SOCIAL LIFE OF ANIMALS
oxygen than if two or more share the same amount of liquid. By one device or another, grouping fre- quently decreases the rate of respiration. Several of these devices are known to us. Professor Child showed many years ago (31) that when animals are exposed to a strongly toxic material, those with the higher rate of respiration, though otherwise similar, die first. This has been applied to group biology by direct tests, and it has been shown that the group, by decreasing the rate of oxygen consumption of its members, makes them more resistant to the action of relatively strong concentrations of toxic materials.
Perhaps I have said enough to show that under a variety of conditions groups of animals may be able to live when isolated individuals would be killed or at least more severely injured by unaccus- tomed toxic, chemical elements, strange to their nor- mal environment.
Will the same relationship hold in the presence of changes in physical conditions? There is a con- siderable and growing lot of evidence that massed animals, even those that can be called cold-blooded, are harder to kill by temperature changes than are similar forms when isolated. (51, 126) This interests us because massing of such animals at the onset of hibernation was recognized as one of the early ex-
BEGINNINGS OF CO-OPERATION 59
ceptions to the rule, now outgrown, that crowding is always harmful.
The exploration of temperature relations is a time-honored field. I prefer to take up a newer though related area, that of the effects of ultra-violet radiation, in which I shall present some evidence so recently collected that it has never been reported extensively before. A year ago Miss Janet Wilder and I began exposing the common planarian worm of this region to ultra-violet radiation, to find whether there was any group protection from the well-described lethal effect of ultra-violet light on these worms. (12)
In lots of twenty, worms of similar size and the same history were placed together in a petri dish and exposed to the action of the ultra-violet light long enough so that they would disintegrate within the next twelve hours. Half of them, that is, ten worms, were then placed together in five cubic centimeters of water and each of the other ten was put into five cubic centimeters of similar water. Grouped and iso- lated worms were treated alike in every way, except that after irradiation together, half were grouped and half were isolated.
For one purpose or another we have repeated this simple experiment a great many times with a variety of waters, and with experimental conditions ade-
6o THE SOCIAL LIFE OF ANIMALS
quately controlled. Some of the things we have found out are:
If the worms are crowded under the ultra-violet lamp so that they shade each other, the shaded ones
RAOIATEO RADIATED RAOIATEO RADIATED
enOUPEO 6R0UPEO GROUPED GROUPED
reSTCO TESTED TESTED TESTED
6R0UPEO SINGLY GROUPED SINGLY
148*
247'
RADIATED RADIATED
GROUPED GROUPED
TESTED TESTED
GROUPED SINGLY
267'
IH 168' tI ItI ooooojjj HH^oooooJUJ I
137*
140 ^M P I
aooo4jU
WELL WATER
DISTILLED WATER
Fig. 3. Planarian worms which have been exposed to ultra-violet radiation disintegrate more rapidly if isolated than if grouped.
are definitely protected. When such crowding is eliminated and by constant watching and stirring, if needed, during exposure, the worms are kept ap- proximately equally spaced, even then the grouped worms survive longer than if isolated. Some of the relationships are shown in Figure 3.
Each block represents the survival time of several series of worms. The figures at the top of the block give the average length of survival in minutes. The blocks are constructed so that the worms surviving
BEGINNINGS OF CO-OPERATION 6l
longer, which in each case are the grouped worms, are given as lOO per cent, regardless of the time taken; while the isolated worms, which had been irradiated in the same dishes as their accompanying groups, survived on an average of 78 per cent and 77 per cent respectively in the two tests with well water, and only 61 per cent in the test in dis- tilled water. The numbers between the blocks show the number of worms averaged for each block; that is, the number of pairs of worms for which results are summarized. The statistical significance given in terms of 'T" is very high in each case.
The number present during exposure is impor- tant, as well as the number present during the time when it is being determined how long the animals will survive. Such data are summarized in Figure 4, which is built exactly on the same principle as that preceding. Worms radiated when crowded (left-hand block), and then tested when isolated, survived 517 minutes, while accompanying worms which had been radiated singly as well as tested when isolated, lived only 24 per cent as long. Those radiated in a group and tested singly (middle block) lived 55 per cent as long as those which had been radiated in a crowd and then were isolated to observe the effects of radi- ation. It will be remembered that these crowded worms actually shaded each other and so gave
62 THE SOCIAL LIFE OF ANIMALS
physical protection from the ill effects of ultra-violet light. Finally (on the extreme right) is diagramed the fact that worms radiated and tested singly lived only 62 per cent as long as those radiated in a group
RAOIATCO RAOIArCO RADIATES RADIATEO RADIATED RADIATED
CROWDED SINGLY CROWDED CROUPtO GROUPED SINGLY
TESTED TESTED TESTED TESTED TESTED TESTED
Singly singly Singly singly singly singly
517' 517' 107'
WELL WATER
Fig. 4. Planarian worms survive exposure to ultra- violet radiation better if much crowded while being radiated, or even partially crowded, even though all are isolated after a few minutes of irradiation.
of 20 per 20 cubic centimeters and also tested singly. Again the figures give the number of pairs tested and under "P" the statistical probability, which shows that all these must be taken seriously even though there is decreasing significance as the per- centage of difference of average survival time de- creases.
In the two cases just outlined mass protection has been demonstrated, first against the presence of toxic
BEGINNINGS OF CO-OPERATION 63
materials, and second against the ill effects of expo- sure to lethal ultra-violet rays. To complete the pic- ture I have now to describe the results of exposing animals to harmful conditions in which the difficulty is caused by the absence of elements normally pres- ent in their natural environment. The experiment has been made on aquatic animals in a number of ways, for example, by putting fresh-water animals into distilled water; but it is easier to demonstrate when marine animals are placed in fresh water.
Again I select one experimental case from several available. Near Woods Hole, on Cape Cod, a small flatworm Procerodes (Figure 5) lives in certain re- stricted areas in large numbers. They are most abun- dant along a stony stretch at about the low tidemark or a little beyond it. (5) There, if one finds the proper location, one may take from ten to fifty flatworms from the lower surface of a single stone. Usually they are more or less clumped together. They are not easy to see since each is only a few millimeters long and all are of a dull gray color. Once seen, they are hard to detach, for the posterior end has a muscular sucker, by means of which the animal can cling pretty securely even to smooth stones. When these worms are put into fresh water, pond water for example, they swell greatly and soon begin to dis-
integrate.
64
THE SOCIAL LIFE OF ANIMALS
If these flatworms are washed thoroughly to re- move sea water from their surfaces, and then placed in fresh water, a certain proportion of the grouped
Fig. 5. The small marine flatworm Procerodes.
animals survive decidedly longer than isolated worms. The first worms to die in the group do so almost as soon as the first isolated worms. As the dead worm disintegrates it changes the surrounding water; we say it conditions it; and as a result of this condition- ing the remaining worms of the group have a bet- ter chance of life.
BEGINNINGS OF CO-OPERATION 65
For more careful experimentation, a sort of worm soup was prepared by killing a number of well- washed worms and allowing them to remain in the water in which they had died and so condition it. Freshly collected Procerodes lived longer in such conditioned water than their fellows which were isolated into uncontaminated, clean pond water. The difference between the two waters was only that caused by the fact that in one the worms had died and disintegrated, while the other was clean. This difference in survival persisted even when, to make the test more revealing, the total amount of salt in the two waters was made identical by adding some dilute sea water to the clean pond water. Results from these experiments are shown in Figure 6. In this chart, distance above the base line gives the percentage of survival, and distance to the right shows time of exposure. It will be noted that the worms lived decidedly longer in the conditioned water than they did in dilute sea water of the same strength of salts.
The mechanism of this superficially mysterious group protection is now known. (86) The dead and disintegrating worms, or more slowly, the living worms, give off calcium into the surrounding water, and calcium has a protective action for marine ani- mals placed in fresh water or for fresh-water animals
66
THE SOCIAL LIFE OF ANIMALS
put into distilled water, a protective action which is out of all proportion to its effect in increasing the osmotic pressure of the water. We can demonstrate that this is in fact the mechanism of such group
1 Conditioned water
Fig. 6. Procerodes die more rapidly if transferred to pure fresh water than in dilute sea water, but live longer if placed in fresh water in which other Procerodes worms have died, even though the total amount of salt is the same as in the dilute sea water.
protection. For example, we can analyze the water which worms have conditioned, find the amount of calcium that has been added, and by adding that amount directly get the same results that we do from the conditioned water.
This explanation is not yet complete— no scientific explanation ever is— but we have demonstrated that what was for a time a very mysterious group pro-
BEGINNINGS OF CO-OPERATION 67
tection is in fact in this case an expression of calcium physiology. The further developments on the sub- ject await exact information concerning the details of the physiological effects of calcium.
It is probably of more direct human interest to
■;V*i.;::i:
%^
35
50
60
Fig. 7. Bacteria frequently do not grow if inoculated in small numbers; here different numbers of Bacillus coli were inoculated into a medium containing gentian violet.
know that under many conditions bacteria will not grow if only a few are inoculated into an animal, man for example; while with a larger inoculation they may grow abundantly. (33) Gentian violet is a poison for many bacteria and in regular medical use for that purpose. In one well-studied case (Figure 7) bacteria belonging to the species Bacillus coli failed to grow on agar containing gentian violet, if singly inoculated on it; only when thirty or more bacteria
68 THE SOCIAL LIFE OF ANIMALS
were inoculated did steady and regular growth oc- cur. With the goldfish spoken of earlier, the mass protection was largely or wholly inoperative when the group of ten was exposed to ten times the amount of toxic colloidal silver to which a single fish was exposed. With these bacteria, however, such quanti- tative limitations did not hold; thirty organisms were found to fix at least two hundred times the amount of poison normally neutralized by an iso- lated bacterium. This difference between the change which thirty bacteria can effect together as compared with what they can accomplish if isolated has been called an expression of the communal activity of bacteria. There is a fairly large and growing litera- ture on this subject which indicates that when only one or a few bacteria, even if strongly pathogenic, gain access to the human body, they are likely to be killed by various devices which aid in resisting infection. It is fortunate for their victims that bac- terial infections normally tend not to take unless the inoculum is somewhat sizable or unless a smaller dose is frequently repeated.
Mass protection is known to occur among sper- matozoa. Many animals, especially those that live in the ocean, shed their eggs and spermatozoa into the sea water, and fertilization takes place in that me- dium. Dilute suspensions of such spermatozoa lose
BEGINNINGS OF CO-OPERATION 69
their ability to fertilize eggs much sooner than if they are present in greater concentration. It is rou- tine laboratory practice in experimenting with such animals as the common sea-urchin, Arbacia, to keep sperm in a cool place, densely massed outside the body, for hours. Small drops can be withdrawn as needed for experimentation, greatly diluted and used almost immediately to fertilize eggs. When such dilute suspensions have long since lost their fer- tilizing power the sperm in the original dense mass are still potentially as active as ever.
So far we have been considering mass effects, the survival value of which, if any, was shown by in- creased length of life, often under adverse circum- stances. Under many different conditions and for a variety of organisms, the presence of numbers of forms relatively near each other confers protection on a part of those grouped together or even on all present.
It is possible to go a step farther and demonstrate a more actively positive effect of numbers of or- ganisms upon each other when they are collected to- gether. Again I select a fresh case for close scrutiny; that of crowding upon the rate of development in sea-urchin eggs.
Arbacia, mentioned above, is the common sea- urchin of coastal waters south of Cape Cod (Fig-
70 THE SOCIAL LIFE OF ANIMALS
ure 8). It has been much used in studies of various aspects of development, particularly by the biologists who gather each summer in the research laboratories at Woods Hole, Massachusetts. There are several reasons for its popularity. These urchins are abun-
FiG. 8. Arbacia, the common sea-urchin of southern New England, shown from the upper surface.
dant in near-by waters and are readily mopped up by the tubful. They can be kept in good condition for some days in the float cages, and eggs and sperm are readily procured as needed. Also the breeding season of Arbacia extends through July and August, which are favored months for research at the seaside. For years biologists at Woods Hole have studied the embryology and physiology of developing sea- urchin eggs. They have built up a painstaking, almost a ritualistic, technique for handling glassware,
BEGINNINGS OF CO-OPERATION 7 I
towels and instruments. The procedures require as rigid cleanliness as a surgical operation. Conse- quently it was not surprising when I first took up their study a few years ago, to have one of my frank- est friends among the long-time workers on the de- velopment of Arbacia, voice what was apparently a common feeling among them. He asked pointedly if I thought I could come into that well-worked field and without long training find something they had overlooked. Such frank skepticism was refreshingly stimulating and added to the normal zest of bio- logical prospecting.
The shed eggs of Arhacia are about the size of pin points and are just visible to the naked eye. The spermatozoa are tiny things; the individual sperm are invisible without a microscope although readily seen when massed in large numbers. When a few drops of dilute sperm suspension are added to well- washed eggs, one spermatozoan unites with one q^^.
After some fifty minutes at usual temperatures, the egg divides into two cells. We call this the first cleavage. Thirty or forty minutes later a second cleavage takes place and thereafter cleavages occur rapidly. Within a day, if all goes well, such an egg will have developed into a freely swimming larva. Other things being equal, (lo) the time after fer- tilization to first, second and third cleavage is speeded
72
THE SOCIAL LIFE OF ANIMALS
up for the crowded eggs. Typical results and some of the methods are shown in Figure 9. With appro-
4- mm.'
FifSt
Second % cleai/ed
f^frst Second % cieaued.
58.25 60.25
85.83 90.25
99 /GO
Fig. 9. Eggs of the sea-urchin, Arhacia, cleave more rapidly in dense populations than if only a few are present. Figures below the diagrams, unless otherwise indicated, give time in minutes.
priate experimental precautions, some eighteen hun- dred eggs were introduced into a tiny drop of sea water. Near by on the same slide forty similar eggs were placed in a similar drop and the two were connected by a narrow strait as shown in the figure.
BEGINNINGS OF CO-OPERATION 73
A few eggs from the larger mass spilled over into this strait. The whole slide was placed in a moist chamber to avoid drying, and examined from time to time. In a trifle over fifty-five minutes half the eggs in the densest drop had passed first cleavage. A half-minute later, 50 per cent of those in the strait were cleaved, and twenty seconds later half of the more isolated ones had divided. The time to 50 per cent second cleavage ranged between eighty-four minutes for the crowded eggs and over eighty-six and a half minutes for the isolated ones.
This was repeated with four thousand eggs or thereabouts in the denser population, almost six hundred of which spilled through and formed a flat apron over the bottom of the second drop, in which there were thirteen other eggs scattered singly about the relatively unoccupied space. Under these condi- tions the time to 50 per cent first cleavage was ap- proximately fifty-two, fifty-eight and sixty minutes respectively, and the difference at the middle of the second cleavage was even greater.
In association with Dr. Gertrude Evans, who is a good, skeptical research worker, this experiment was repeated in many different ways; and there remains in my mind no doubt but that under a variety of conditions the denser clusters of these Arhacia eggs
74 THE SOCIAL LIFE OF ANIMALS
cleave more rapidly than associated but isolated fellows.
Under the conditions tested, the stimulating effect of crowding could be detected when sixty-five or more eggs were present in the more crowded drop and twenty-four or fewer eggs made up the accom- panying sparse population.
Within twenty-four hours, under favorable condi- tions, one finds one's cultures full of free-swimming larvae with characteristic arms which are known as plutei. When all our available data collected the first day after fertilization are compared there is again no doubt but that the more crowded cultures usually develop more rapidly than accompanying but sparser populations. However, it must be recorded that throughout the whole series there were occa- sional isolated eggs that developed as rapidly as the best of the accompanying denser populations. Such eggs and embryos were exceptional in our experi- ence; the fact that they exist indicates clearly that under the conditions of our experiments crowding, while usually stimulating, was not absolutely neces- sary for rapid cleavage and early growth.
In this connection it is interesting to note that others have prepared an extract from sea-urchin eggs and larvae which is growth-promoting, (91) and one which is growth-inhibiting. As has also been found
BEGINNINGS OF CO-OPERATION 75
with goldfish, the growth-accelerating principle seems to be associated with the protein fraction of the ex- tract. When the whole extract is used, it is said to be growth-inhibiting and to produce the same re- sults as overcrowding. The point I have made is that with the sea-urchin eggs, under the conditions of our experiments, there is also an ill effect of un- dercrowding, and that there is an optimum popula- tion size for speedy development which is neither too crowded nor too scattered.
Much similar work has been done with the ef- fects of numbers on the rate of multiplication with various protozoans. Again I shall have to select re- sults from the mass of available evidence. The late T. Brailsford Robertson (107) of Australia an- nounced back in 1921 that when two protozoans of a certain species were placed together, the rate of division was considerably more than double that which resulted with only one present. It should be noted that during the time of these experiments and in all these protozoa which we are considering re- production was entirely asexual, by self-division of the original animal. I subjected the data in Robert- son's original paper to statistical analysis and found that there were only thirteen chances in a thousand of getting as great a difference by random sampling. Such results must be taken seriously (Figure 10).
76 THE SOCIAL LIFE OF ANIMALS
They were. And the period after 1921 was en- livened for some of us by denials from one first- class laboratory after another that there was anything significant in Robertson's data. Robertson himself
ISOLATED PAIRED
24 HOURS 20.5 92.4
RATIO 1 2.2
I 44
Wf
P = 0.0128
Fig. 10. Robertson found that when two protozoans were placed together each yielded over twice as many as when the same number of similar protozoans were iso- lated.
rechecked and confirmed his results, though his ex- planations of them tended to vary. For the moment we are not concerned with the explanations; but what are the facts? The first extensive corroboration from outside Robertson's own laboratory came from the work of Dr. Petersen at Chicago. When she cul- tured the common Paramecium in small volumes of liquid, she obtained the same results as had Rob- ertson's critics, but when she used relatively larger volumes of the same culture medium, a cubic cen-
BEGINNINGS OF CO-OPERATION 77
timeter more or less, she got an increase in division rate with the presence of a second individual, as Robertson had found it in the Australian form he had studied.
Still the critics were not convinced. Accordingly Dr. Johnson, now of Stanford University, repeated this whole study using a different protozoan, one of the Oxytricha. (68) When sister cells from pure-line cultures were used there was no difference at the end of the first day, whether the Oxytricha were in- troduced singly or in pairs into one or two drops of good medium. Later, the cultures started with one organism always were ahead. With larger volumes, two organisms showed a higher rate of reproduction per original animal at the end of the first day than if started with a single protozoan.
Again for larger volumes Robertson's results were confirmed, and those of his critics for smaller vol- umes. But Johnson had only started. He knew from the work of others that if a protozoan is washed through several baths of sterile water the associated bacteria are rinsed off. Then if the washed protozoan is put into a weak solution of the proper salts, into which has been introduced known numbers of the bacteria on which they normally feed, the problem can be studied with a controlled food supply, both as to kind and amount.
78 THE SOCIAL LIFE OF ANIMALS
This he proceeded to do. He found a common bacterium on which his sterile Oxytricha would grow
NUMBERS OBTAINED IN 24- HOURS FROM THE
ISOLATION OF OXYTRICHA INTO CONSTANT VOLUMES
WITH DIFFERENT CONCENTRATIONS Of BACTERIA
CONCENTRN 4X 2X X X/4 X/lO
I
I
3.5 9.0 11.4 5.4 3.0
Fig. 11. The ciliate protozoan Oxytricha reproduces more rapidly with a certain limited number of bacteria present than with either more or fewer. (From Johnson.)
and reproduce faster than in the ordinary medium. He made standard suspensions of these bacteria in sterile salt solution, at what we may call an X con- centration. The bacteria could reproduce little, if at all, in the salt medium, so that he knew how much
BEGINNINGS OF CO-OPERATION 79
and what kind of fodder he was feeding his washed protozoans.
The resuhs of varying the amount of food are
REPRODUCTION-RATE FOR 24 HOURS WHEN ONE OR TWO OXYTRICMA ARE SEEDED INTO TWO DROPS OF P. FLUORESCENS
CONCCMTRN 4 X
seeoiNG t 2
8.0 tO.2 tO.6 10.4
Fig. 12. In the denser suspensions of bacteria the protozoans divide more rapidly when cultures are inocu- lated with two protozoans than if started with a single individual. (From Johnson.)
shown in Figure ii. With X concentration, in twenty-four hours one animal produced about eleven progeny. With 2X concentration, isolated sister cells produced nine, and with a 4X concentration other isolated sister cells produced but three and a half. The rate of reproduction also decreased when less than X bacteria were present.
8o THE SOCIAL LIFE OF ANIMALS
Now he was ready for the grand Robertson test, except that by this time nearly all the factors were controlled. The results are shown in the following figure (Figure 12). With X concentration it made no difference whether he started his small cultures with one or with two sterile animals. With 2X con- centration, the cultures started with two individuals did as well as in X concentration, but those which were started with only one individual lagged defi- nitely, producing only 80 per cent as many animals in twenty-four hours. With 4X concentration even the culture started with two Oxytricha was slowed down, but not so much as that started with only one. He had shown that in the presence of an ex- cess number of bacteria, cultures seeded with more than one bacterium-eating protozoan thrive better than if but one is introduced. Not content with this Johnson took another species and tried it all over again with the same results.
From all this careful work we judge that the facts on this particular aspect of the effects of numbers present on the rate of asexual reproduction seem now to be straight; but what about their expla- nation? This, as it turns out, also interests us. Robertson advanced the following hypothesis to explain the results which he had observed. Dur- ing division each nucleus retains as much as pos-
BEGINNINGS OF CO-OPERATION 8l
sible of an essential, growth-producing substance with which it was provided, and adds to it dur- ing the course of growth between divisions. At each division, however, this substance is necessarily shared with the surrounding medium in a propor- tion that is determined by its relative solubility in the culture water, and by its affinity for chemical substances within the nucleus. The mutual speeding of division by neighboring cells is due to each cell's losing less of this necessary substance because of the presence of the other. The more of this growth-pro- moting substance there was in the cell, Robertson thought, the faster would be the division rate; so that any circumstance which would conserve the limited supply would tend to speed up processes leading to cell division.
Stripped to essentials this hypothesis says that as a result of the presence of a second organism both lose less of an unknown something which is essen- tial for division than would happen if but one were present. Returning to the problem after the criti- cisms of half a dozen years, Robertson affirmed that all the data and conclusions on the subject that had been issued from his laboratory remained valid save that they might apply to the ^ associated food or- ganisms and not to the protozoans themselves.
Johnson has paid considerable attention to this
82 THE SOCIAL LIFE OF ANIMALS
problem, and has concluded that the results which he has observed can be explained as due to bacterial crowding; that the larger number of protozoans in- troduced into dense cultures thrive best because they are able to reduce the bacteria to density optimal to the protozoa faster than their isolated sister cells can; and therefore they show a higher rate of re- production.
This does not seem to be the whole story; for from points as distant as Baltimore (79) and Jerusalem, (101) I have reports from trustworthy men that with still simpler protozoans they are getting results which suggest that some modification of Robertson's hy- pothesis may be correct after all. These organisms stimulate each other to more rapid growth merely by their presence in the same small space.
With fine courtesy, Professor Mast of Johns Hop- kins has placed a report of his experiments in my hands in advance of publication and has permitted me to summarize his results. He finds that popula- tions of a flagellate protozoan grow more rapidly in a sterile medium of relatively simple salts when larger numbers are introduced than if the cultures are started with only a few organisms.
I must not put too much stress on these reports, pending the appearance of yet more data, but I should expect to find here, as elsewhere, that com-
BEGINNINGS OF CO-OPERATION 83
plicated problems such as these that deal with the rate of population growth are controlled by more than one mechanism.
The suggestions from the simpler protozoans, taken together with other aspects of the mass physi- ology of protozoa which have been only partially reviewed here, and with the acceleration of devel- opment demonstrated for sea-urchin eggs, encourage me to renew a suggestion made some years ago, (3) which has, so far as I am aware, been overlooked to date.
Let us go back to consider the case of external fertilization among aquatic animals. When sperma- tozoa and eggs are shed into sea water by sea- urchins or other marine animals, their length of life is distinctly limited. If a sperm fails to contact an egg during the fertilizable period, death results probably from starvation for the spermatozoa, per- haps from suffocation for the egg. This means that the animals of the two sexes must be fairly close together if there is to be a union of the shed sexual products. The most vigorous sperm of the sea- urchin Arhacia can travel in still water about thirty centimeters, that is, about one foot and two inches. (55) Spermatozoa of these animals diluted a few thousands of times can survive from three to twelve hours; the majority succumb by seven hours.
84 THE SOCIAL LIFE OF ANIMALS
If a current catches it, such sperm can travel many times thirty centimeters, but even in sea water the sexes must be relatively aggregated if fertilization is to be successful. In fresh water, the life of shed gametes is much shorter. After ten minutes, eggs of the pike lose the power to be fertilized, (102) and the longevity of sperm of certain fresh-water fishes is said to be less than a minute, so that in fresh water the aggregation is even more essential. With animals that require internal impregnation the necessity for close co-operation between at least two individuals is obvious. Such considerations must be fundamental for the long-recognized breeding aggregations of animals, especially of those that shed eggs and sperm into surrounding water.
Mass relationships may be even more important sexually, and here I come to the new suggestion: perhaps they had a hand in shaping sex itself. Pre- sumably sexual evolution started, as it does today in plants, with a time when all gametes of any one spe- cies were similar. Under these conditions a first step toward the union of two reproductive elements could be supplied by the greater well-being fos- tered by the presence of more than one gamete within a limited area, as even the simpler proto- zoans are stimulated to asexual division today by the near-by presence of another of the same species. In
I
BEGINNINGS OF CO-OPERATION 85
the survival value existing for separate living cells before actual sexual union took place we can find a logical beginning for the action of selection, which would in turn, with present known values, result in the establishment of the sexual phenomena as they appear today. These fields have not been sufficiently explored to allow for more than this flash of imagina- tion, which future researches may verify or discard.
At this point it would be well to pause and look back over the road we have traveled thus far. The charts, (7) shown as Figures 13 A and B, show that most of our evidence has come from fairly well down among the simpler forms of life. I have called atten- tion to mass protection of one sort or another among bacteria, planarian worms, goldfish and the simpler crustaceans. Actually there are in scientific literature good cases of mass protection for almost all the ani- mals shown in these charts; and where exact informa- tion is lacking, as for example among the rotifers, this is a result only of lack of interest in conducting experiments on this point with these animals. I have little doubt that we could, overnight, demonstrate mass protection from colloidal silver for rotifers; but we have more interesting work to do.
I have also shown active acceleration of fundamen- tal biological processes as a result of numbers present for sea-urchin eggs and larvae, and for various pro-
Bi(rdM M&mmalls
Amphil^ani^ /
.JZOA oeba)
Ancesrral plants
ANCESTRAL CCELENTERATES
ANCESTRAL PROTOZOA "
Ancestral animal-plants — "^ Primitive protoplasm Fig. 13. A recent suggestion concerning the ancestral relations within the animal kingdom. The circles in A
B
CHORDATES
1^
ANCESTRAt COELENTERATES
ANCESTRAL PR0T02OA
Ancestral plants
Ancestral animal-plants —
Primitive Protoplasm
and B allow cross-identification. (From Allee in The World and Man.)
88 THE SOCIAL LIFE OF ANIMALS
tozoans. These have been given in some detail, which has not left time for similar demonstrations among regenerating cells of sponges; nor have I time to tell how hydra have been saved from depression periods by the use of self-conditioned water. I have men- tioned but not elaborated the fact that grouped ani- mals frequently have different rates of respiration as compared with their isolated fellows. This has been recorded widely in the animal kingdom, notably among planarians, certain lower crustaceans, some starfish, fishes and lizards, and for some, at least, asso- ciated survival values have been demonstrated. To this extent, then, I have given the crucial evidence I promised earlier that a sort of unconscious co- operation or automatic mutualism extends far down among the simpler plants and animals.
These charts should illustrate one other point. The insects stand at the apex of one long line of evolu- tion; mammals and birds are at the peak of another line of evolution; the two have been distinct for a very long time. This view of evolution indicates that the ancestral tree of animals is not like that of a pine tree with man at the very top and insects and all the other animals arranged as side shoots from one main stem. Rather, there are at least two main branches which start low, as in a well-pruned peach tree. Both rise to approximately equal heights, indicating cor-
BEGINNINGS OF CO-OPERATION 89
rectly that in their way the insects are as specialized as the birds or mammals. Since both insects and mammals have developed closely-knit social groups, this is further evidence that there is a widely dis- tributed potentiality of social life. We shall return to this subject later.
IV
Aggregations of Higher Animals
A GREAT deal of skepticism is necessary in science, if progress is to be even relatively steady and sound. Not only must the scientist be skeptical of advance reports of new results until he has seen the support- ing evidence, no matter how stimulating the thesis and how well it would explain material already gathered; but in fields which lie near his own re- searches it is necessary if possible to bring the prob- lem into his own laboratory and there examine the validity of the evidence itself. This repeating of ex- periments in order to check the first observer is some- times also a testing of scientific courtesy, but every real scientist must be prepared to submit to it with the best grace possible.
It is demanded also that from time to time one should be skeptical of views long held, and of the evidence on which they were built up, particularly of the inclusiveness of the conclusions that have been drawn. Without my own fair share of this skepticism I should never have been drawn into what I knew
90
AGGREGATIONS OF HIGHER ANIMALS Ql
from the beginning would be a long and laborious series of experiments concerning the effects of num- bers present upon growth.
As long ago as the eighteen-fifties Jabez Hogg, (62) an Englishman, found by experimenting that crowding decreased the rate of growth of snails and produced stunted adults. From that day to this there has been almost no break in the reported evidence that overcrowding reduces growth; the number of reports that crowding in any degree increases growth are relatively few.
This phenomenon has, however, been observed by enough workers using animals widely distributed through the animal kingdom to show that the retard- ing effect of undercrowding on growth is real. Before considering the implications of this statement let me review briefly some of the evidence. (3) Here as else- where I shall make no attempt to catalogue all the available evidence; the list would be impressively long but tedious.
It is relatively easy to show that mixed populations of many animals grow faster than if the same number of some one species are cultured together. The com- mon experience of aquarium enthusiasts that the presence of the snails in aquaria increases the rate of growth and well-being of their fishes is a case in point. Their rule-of-thumb experience has been fully
9^ THE SOCIAL LIFE OF ANIMALS
verified by careful laboratory experiments. A more crucial test involves individuals of the same species: all snails, let us say, or all goldfish. Is there some optimum size of the population at which individuals grow most rapidly?
For years I have been studying different aspects of this problem with the aid of a succession of com- petent, critical research assistants and associates. The names of these young scientists are interesting and, I think, important. They include Drs. Bowen, Welty, Shaw, Oesting and Evans, and Messrs. Livengood, Hoskins, and Finkel, all of whom have independ- ently obtained the basic results I am about to de- scribe. (13, 14, 76)
We have used goldfish for our experimental ani- mals, because these are inexpensive, easy to obtain, hardy under laboratory conditions, and able to stand daily handling.
In order to have a consistently constant water we make up a synthetic pond water by dissolving in good distilled water salts of high chemical purity. Into such water goldfish about three inches long are placed in sufficient number so that they will give a conditioning coefficient of about twenty-five. Let me explain: this coefficient is obtained by multiplying the number of fish by their average length in milli- meters and dividing by the number of liters of water
AGGREGATIONS OF HIGHER ANIMALS 93
in the containing vessel. Living in this water the fish condition it by giving off organic matter and carbon dioxide. They are left in the water for twenty-one hours or so, while a similar amount of the same water stands near by under exactly similar conditions ex- cept for the absence of fish.
At the end of this time the clean control water is siphoned into a number of clean jars, and a small measured goldfish is placed in each. At the same time the conditioned water is siphoned, either with or without removing particles (that is, of excrement, etc.) that may be present, into similarly clean jars. A set of small measured goldfish, like those used in the control jars, are transferred into the conditioned water. These small "assay" fish have been feeding for about two hours before being transferred; the larger conditioning fish are allowed to feed for a somewhat longer time before being washed carefully to remove food residues and replaced in another lot of water to condition that.
Meantime the jars, 120 of them, are all washed carefully; and after this is done the experimenter has nothing more to do until the next day, except to put the laboratory in order, keep the temperamental steam distilling apparatus running, test the water chemically in several ways, keep his records in order,
94 THE SOCIAL LIFE OF ANIMALS
and otherwise see that nothing untoward happens to make him or anyone else question the results.
After some twenty, twenty-five or thirty days of such care, in which Sundays are included, again each fish is photographed to scale, as they were also photo- graphed at the beginning of the experiment; the photographs are measured and the relative growth determined for the fish that have daily been placed into perfectly clean synthetic pond water, as com- pared with those which daily have been put into conditioned water, that is, into the water in which other goldfish have lived for a day.
During the course of an analysis of this problem we have performed this simple basic experiment many times. The first forty-two such tests involving 886 fish gave on the average about two units more growth for the fish in the conditioned, slightly con- taminated water, than for those in clean water (Fig- ure 14). These results have a statistical probability (P) of about one chance in a hundred million of being duplicated by random sampling. Hence we have demonstrated that under the conditions of our experiments the goldfish grow better in water in which other similar goldfish have lived than they do when they are daily transferred to perfectly clean water.
The problem that has been occupying us for some
AGGREGATIONS OF HIGHER ANIMALS
95
time is why this is so. What are the factors involved
that make this slightly contaminated water a better
medium for young goldfish than a clean medium?
We have said that the conditioning fish are fed
EfFECT OF SELF CONTAMINATED WATER ON GROWTH OF GOLDFISH
«0. 290 274
I
GROWTH 1.8 -0 2
no. 180 142
I
I
GROWTH 1.65 1.00
MO. 210 120
CONCErfTRATEO
I
I
GROWTH 2.28 t.ia
ma t6l 114
II
GROWTH 1.92
Ha 220 217
tl
GROWTH 2.59 2.20
Fig. 14. Goldfish grow more rapidly if placed in vari- ous kinds of slightly contaminated (conditioned) water. The numbers above the columns show the number of fish tested. The longer column represents the growth in conditioned water.
for two or more hours daily and are then washed off and placed in a fresh batch of water. Although the fish are never fed in the water they are condi- tioning, within a few hours after their transfer into it from the feeding aquarium the water becomes more or less cloudy with regurgitated food particles. These bits of food are large enough so that the growth-assay
96 THE SOCIAL LIFE OF ANIMALS
fishes can strain them out of the water. When such particles are removed by filtering, the growth-promot- ing power of the conditioned water is greatly les- sened, but it is not completely lost. In our experi- ments we found that suspended food particles ac- counted for 80 per cent or more of the increased growth in conditioned water over that given in clean control water.
These experiments give certain suggestions con- cerning some other conditioning factors that may be acting. For example, we know that the skin glands of fish secrete slime (Figure 15). When we have made a chemical extract of this material we have frequently recovered a growth-promoting substance, apparently a protein, which was effective in stimulating growth when diluted 1 to 400,000, or even 1 to 800,000 times. At these dilutions it is not probable that this factor is affecting growth by furnishing food material.
There are, of course, other possibilities, many of which we have checked. The increase in growth is not due, for example, to a change in the total salt content of the water, for this does not change in our experiments; nor to differences in acidity or oxygen, nor, so far as careful quantitative analyses have re- vealed, to changes in chemical elements present. We may be dealing with some sort of mass protection, such as was discussed in the last chapter, in which
J
AGGREGATIONS OF HIGHER ANIMALS 97
the conditioning fishes remove some harmful sub- stance, but of this we have no real evidence.
Whatever the explanation, we are certain of the facts, and we know that we have demonstrated a de-
EFFECT OF PROTEIN EXTRACT FROM SKIN OF GOLDFISHES ON GROWTH OF GOLDFISH
NO. 56 59
EXTRACT
VS. ■■ P= 0.0106
CONTROL
.
GROWTH 1,95 0.54
NO. 61 122
EXTRACT
VS.
SALT CONTROL
GROWTH 3.22 0.61
P= 0.0006
NO. 26 28
EXTRACT
VS.
CONDITIONED
WATER
II
GROWTH 1.92 1.55
P= 0.26
Fig. 15. An extract from the skin of goldfish fre- quently has growth-promoting power. The arrangement of the figure is on the same plan as was used in Fig. 14.
98 THE SOCIAL LIFE OF ANIMALS
vice such that if in nature one or a few fish in a group find plenty of food, apparently without will- ing to do so they regurgitate some food particles which are taken by others, a sort of automatic shar- ing. Again, in water that changes rapidly, such stag- nant-water fishes as goldfish, if present in numbers, are able to condition their environment, perhaps by the secretion of mucus, so that it becomes a more favorable place in which to live and grow.
Perhaps I have lingered too long over this one case; I am so close to the facts and to the tactics used in collecting them that they may seem rnore interest- ing to me than they will ten years hence. We have run the same experiment with positive results with a few other species of fishes; and we have also found by experimentation that certain fish will regenerate tails that have been cut off if several are present in the same water more rapidly than if each is isolated. (112) The same is true for the young tadpoles of sala- manders, with which we have had experience. The explanation of the more rapid regeneration of such cut tails is probably relatively simple. The several animals together more readily bring the surrounding fresh water to approximately the salt content of the cut and regenerating tissues than can be done by a single animal placed in the same amount of water.
AGGREGATIONS OF HIGHER ANIMALS QQ
This may not be the whole of the story but it is prob- ably a significant part of it.
In both of these cases the additional growth of aquatic animals, which occurs as a result of the pres- ence of other animals of the same species, is produced in response to some sort of chemical which has been given off into the surrounding water. This may be nothing more than the unswallowing of surplus food by the conditioning fish. With animals whose tails have been freshly cut off the addition of salts to the water by the group may balance the osmotic tension at the cut surfaces and so favor re-growth. The excit- ing result of these studies lies in the suggestion that some less obvious growth-promoting substances may also be secreted into the surrounding water.
Animal aggregations frequently produce physical as well as chemical changes, and while we are con- sidering the effect of numbers of animals present on the rate of growth of individuals it is interesting to examine one case in which growth-promotion appears to have been produced largely by changes in tempera- ture. Such an effect has been reported more than once; it is most simply illustrated in a warm-blooded animal, this time the white mouse. The experiment was first performed in Poland, but the causal factors were then only partially recognized. It has been re-
lOO THE SOCIAL LIFE OF ANIMALS
peated in our laboratory where significant steps have been taken towards its further analysis.
Vetulani, the original experimenter, (117) used closely inbred mice for his experimental animals. He measured the growth of males and females separately from the sixth and on through the twenty-second weeks of their lives. After rearrangement he followed them for ten weeks longer as a sort of control. Fresh food was supplied in abundance each day, and proper experimental conditions seem to have been main- tained.
Growth during the first sixteen weeks of the ex- periment is shown in the accompanying graphs (Fig- ure 16). All started off at approximately the same rate. After the fifth week of the experiment, however, it is clear that the isolated mice were growing most slowly, and they continued to do so as long as the experiment ran. The most rapid rate of growth was observed in those mice which were placed two to four per cage; those five to six per cage grew next best, and only slightly below these came those living nine to twelve per cage.
Under the conditions of this experiment the iso- lated young mice were most handicapped, those most crowded were next, while those that were somewhat but not too crowded grew most rapidly. When the mice were rearranged for a continuing period of ten
OK — I 1 1 1 1 1 1 1 1 t I I I I . I
6 7 8 9 10 11121314 15 1G17 1819 20 2I22 A;,'c in weeks
Fig. i6. White mice grow faster in small groups than in large ones; they grow slowest . when isolated (solid line). (From Vetulani.)
102 THE SOCIAL LIFE OF ANIMALS
weeks the same relations held, showing that it was the state of aggregation rather than individual dif- ferences between mouse and mouse which was impor- tant in producing the differences in growth rates.
Mr. Retzlaff, (105) the student who brought this work into our laboratory, tried first to repeat Vetu- lani's experiments in a room held at relatively high temperatures (29-30° C). Under these conditions he found that insofar as significant differences existed they showed that most rapid growth occurred with the isolated mice. When, however, he lowered the room temperature to about 16° C. he obtained the same general effect as reported by Vetulani. It would seem then that in this case the opportunity to keep warm in a chilling temperature is one of the main factors in promoting growth of the crowded, but not too crowded, animals. This conclusion is strength- ened by recent analyses of the temperature relations of mice, made by French physiologists, (30) which show that a mammal as small as a mouse has great difficulty in maintaining a constant temperature and rarely does so for extended periods of time. A change of external temperature from 30° to 18° C. will cause a lowering of 0.4° in the body temperature of a resting mouse.
With such temperature lability it is easy to see that a few mice huddled together as is their habit could
AGGREGATIONS OF HIGHER ANIMALS IO3
help each other maintain their internal temperatures, conserving energy for growth, while if isolated they must use much of their energy in keeping warm.
Vetulani observed another factor at work. Some of his mice had lesions of the skin which they treated by licking. When these were in the head region they could only be treated by another individual. Some of his isolated mice had such lesions when at the end of the first experimental period they were re-grouped for further observation; these wounds were soon cured by their new nest mates.
When one turns from studying the rate of growth of individuals to that of populations of these higher sexual animals, many of the same principles can be observed working as were outlined in the last chap- ter for the growth of asexual populations of proto- zoans in which overcrowding retards population growth, while optimal crowding, at least in many instances, favors it.
With experimental populations of mice, for exam- ple, three long, laborious experiments made in Scot- land (36) and in Chicago (106) have indicated that, under the conditions tried, the least crowded mice reproduce most rapidly. The same holds true for the well-studied fruit-fly, Drosophila. (96)
Neither with these flies nor with the mice is there any indication to date of a more rapid rate of repro-
104 THE SOCIAL LIFE OF ANIMALS
duction per female when more than the minimal pair is present. I have a strong suspicion, however, that one would get a more rapid rate of increase per number of animals involved if, instead of keeping the sexes equal in numbers, there were a ratio, let us say, of two females to one male.
We do know that with Drosophila the greatest numbers are produced when the feeding surface is relatively great but not too great; (60) this result may be explained by the assumption that with too great space, or in other words, with too few flies present, wild yeasts or molds grow more rapidly than the Drosophila can keep under control.
Another well-studied laboratory animal, the flour beetle, Triholium, under certain experimental con- ditions gives most rapid population growth at an intermediate population size rather than with too few or too many present. A study of data collected by Chapman showed that in a flour beetle's little world, a microcosm of thirty-two grams of flour, these beetles, during the early stages of population growth, reproduce most rapidly per female with two pairs present (Figure 17). Reproduction is more rapid when four pairs or even sixteen pairs are present, than if there is only one pair. (3)
This work of Dr. Chapman's was done for another purpose. We took it for an indication of possibilities.
AGGREGATIONS OF HIGHER ANIMALS IO5
and Dr. Thomas Park looked into the matter inde- pendently. (88) He found the situation very much as it had originally appeared to be. A Scotsman named Maclagen had a curiosity along the same line
|
8 |
. |
|
|
S^'^ |
/' |
^N< |
|
"§ |
/' |
"^^^^^^ |
|
^6 |
Ji |
\*^^>^^ |
|
\ ^^^^^^^^^ |
||
|
/' |
\^ ^""X^ |
|
|
554 |
/ |
\ ^^frdays |
|
^ 5 |
P |
v.. |
|
i? |
/' |
*"'••. ^ |
|
2 |
// |
|
|
""-..SSc/ays |
||
|
1 |
1 « I 1 1 |
2 4 8 16 32 64
/nitiaC population per ^2 Sms, of f/oar
Fig. 17. Flour beetles reproduce more rapidly if more than one pair is present.
and independently re-checked the whole matter with the same results. (77) Three separate workers in three different laboratories have now obtained essen- tially similar results with these same beetles, and the chances that all are mistaken are rather remote.
One of them, Dr. Thomas Park, has proceeded to analyze the factors involved. (89) He finds that the results come from the interaction of two opposing tendencies. In the first place, adult beetles roam at
106 THE SOCIAL LIFE OF ANIMALS
random through their floury universe. They eat the flour, but they may also eat their own eggs as they encounter these on their travels. This habit of egg- eating tends to reduce the rate of population growth, the more so the denser the population.
The second factor is the experimentally proven fact that up to a certain point copulation and suc- cessive re-copulation stimulate the female Tribolium beetles to lay more eggs, and eggs with a higher per- centage of fertility. Thus the more dense the beetle population the more rapid its rate of increase. The interaction of these two opposing tendencies results in an intermediate optimal population in which more offspring are produced per adult animal than in either more or less dense populations.
It may be felt that I have been keeping too closely to the more or less artificial conditions found in the laboratory. It is true that in an attempt to bring the various aspects of the population problem under ex- perimental control we have avoided those field obser- vations which can only be recorded as more or less interesting anecdotes. We have now come to a point in our inquiry, however, at which it is necessary to move directly into the field.
Given the evidence at hand, that optimal numbers present in a given situation have certain positive survival values and some definitely stimulating effects
AGGREGATIONS OF HIGHER ANIMALS 107
on the growth of individuals and the increase of populations, we strike the problem of the optimal size of a population in nature. This is an exceedingly difficult question on which to obtain data. Suppose, therefore, that we simplify it by asking what minimal numbers are necessary if a species is to maintain itself in nature?
This inquiry is a direct attempt to find under nat- ural conditions the application of the statement by Professor Pearl that "this whole matter of influence of density of population in all senses, upon biological phenomena, deserves a great deal more attention than it has had. The indications all are that it is the most important and significant element in the bio- logical, as distinguished from the physical, environ- ment of organisms."
Over and over again in the last half-dozen years I have asked field naturalists, students of birds, wild- life managers, anyone and everyone who might have had experience in that direction, how few members of a given species could maintain themselves in a given situation. Always until this last summer I have found that, stripped of extra verbiage behind which they might hide their ignorance, the real answer was that they did not know.
And then I had two pieces of luck; I found a man and a scientific paper. My friend. Professor Phillips
lo8 THE SOCIAL LIFE OF ANIMALS
of South Africa, came to spend some weeks with us. He told us that the Knysna Forest, a protected wood- land in South Africa, has an area of 225 square miles, fifteen miles on a side, and that this forest is the home of a herd of eleven elephants, which can also range outside the forest limits. On the other hand, the Addo Forest, of twenty-five to thirty square miles, supports a herd of twenty-four elephants. (98) Dr. Phillips thinks that the smaller herd is not maintain- ing itself, and that the larger one, under apparently less favorable conditions as regards available area of range, is at approximately the lower limit for keep- ing up its own numbers. He estimates that an ele- phant herd of about twenty-five individuals could maintain itself in an unrestricted range providing civilized man were absent.
He gave us a second example, of a herd of some three hundred springbok on a protected reserve of six thousand acres in the Transvaal, which was un- able to maintain its numbers and became reduced to eighty or ninety, on its way toward total ex- tinction.
It is well known that in the life of equatorial Africa the tsetse fly plays an important part. It carries the trypanosomes which cause the deadly disease, "sleeping sickness," of man and his domestic animals, and which affect native game as well. The British
AGGREGATIONS OF HIGHER ANIMALS lOQ
colonial governments have been active in attempts to control the density of these fly populations. In general they are restricted to damp, low-lying forest. In districts where this is confined to the borders of water-courses, and hence where the fly belt has nat- urally a definite limit and is restricted in size, an ingenious fly trap has been used successfully. The trap takes advantage of the natural reactions of the tsetse fly. These are strongly positive to a slightly moving dark object a few feet above ground. With appropriate screening they can be caught as they fly toward such an object; they will fly up and fall back until they literally wear themselves out. It was at first thought that such a trap would be helpful chiefly in reducing the excess fly population; then, to the delight of the control officials, they found that when in these restricted fly belts the tsetse flies had been trapped down to a certain minimum population there was no need to catch the very last flies; below the minimum level those remaining disappeared spontaneously from the area. Nor did they return unless brought back in considerable numbers accom- panying movements of game, or as a result of the slow extension of range from other infested areas. The work of the control officials in such regions thus was very much easier than had been anticipated. Two pertinent cases concerning the minimum
no THE SOCIAL LIFE OF ANIMALS
number below which a species cannot go with safety have come in part under my own observation. In 1913, my first summer at the Marine Biological Lab- oratory at Woods Hole, Massachusetts, the veteran scientists of the laboratory, at least those who still were willing to exhibit naturalistic enthusiasms, were greatly pleased at the visit of a flock of laughing gulls to the Eel Pond near the laboratory. The main breeding ground of these gulls is on Muskeget Island off Nantucket. In 1850 the laughing gulls were abun- dant there; but they were exposed to the depreda- tions of egg takers and later, about 1876, to the attacks of men interested in obtaining their striking wings and other feathers to satisfy the millinery de- mand for feathers of native birds, which was then at its height. (49) Under this slaughter the colony was nearly wiped out; at its low point about 1880 there were not more than twelve pairs of laughing gulls left on Muskeget Island, and only a few of these bred. A warden was employed in a somewhat extra- legal capacity by certain ornithologists who regretted seeing the species die out, and he was assisted by the captain of the local life-saving crew in protecting the gulls from raids. Later changes in laws regarding protection of birds and the use of plumage in mil- linery gave more secure protection for the growing colony. For the first ten years the birds increased
AGGREGATIONS OF HIGHER ANIMALS 111
slowly, but thereafter more rapidly, until there are now thousands breeding on the island, and their range has spread to the mainland. In Woods Hole, at the present time, these birds whose return in 1913 excited so much comment are as common as the terns. In this case, a few breeding pairs, nesting in a relatively safe place, were able to regenerate the local population in less than fifty years; all that was needed was protection from the predations of man.
The nesting colonies of gulls have attracted atten- tion from many; a report by Darling has recently appeared concerning certain relations between num- bers of herring gulls in a colony and breeding be- havior, and survival of young gulls on Priest Island off the northwest coast of Scotland. (39) There are indications that the members of larger colonies stim- ulate each other to begin mating activities earlier than when the colonies are smaller and, what is apparently more important, there tends to be a shorter spread in the time from the laying of the first egg until the last one is laid. This means that the breeding activities are more intense while they last.
The period between hatching and the growth of the first adult plumage is a crucial time in the life of young gulls. While they are in the downy stage they are preyed upon by outside predators; also at
112
THE SOCIAL LIFE OF ANIMALS
this time the gull chicks that wander from their home nests may be pecked to death by other members of the colony. The toll of the chicks is comparatively less, the shorter the time from the hatching of the
Survivor)
Time
Fig. i8. The "spread" of time in which eggs are laid in a colony of herring gulls affects the percentage that survive. The smaller the colony the longer the spread, and the fewer survivors. [From Darling (39) by permis- sion of The Macmillan Co.]
first fuzzy young gull until the last one changes to a young fledgling with adult feathers. These relations are graphically shown in Figure 18.
Darling thinks that the greater success of the larger colonies does not lie in any vague factor of mutual protection, but in the nearer approach to simultane- ous breeding throughout the colony. This is a phase of social facilitation which will be discussed more fully in a later chapter.
These observations need to be extended and con-
AGGREGATIONS OF HIGHER ANIMALS 1 I 3
firmed. They suggest one mechanism, that of mutual stimulation to mating, which may have operated to produce social nesting among birds, and which seems capable of giving added survival value to the larger colonies, once the habit of collecting into breeding flocks is established. We have here a suggestion that these social colonies of birds have evolved far enough so that there has come to be a threshold of numbers below which successful mating does not take place. The numbers that constitute this threshold probably vary under a variety of conditions.
In one case, when only two pairs were present, nests were built but no eggs were laid, while in a more favorable season, with three pairs, eggs were laid and one chick out of eight that hatched lived through the downy stage.
I saw the laughing gulls myself at Woods Hole last summer; and I also found a paper by Gross giv- ing the case of another almost extinct population which could not be revived. The heath hen, prob- ably a representative of an eastern race of the prairie chicken, was formerly very abundant in Massachu- setts, and may have been distributed from Maine to Delaware, or perhaps even further south. It was grad- ually isolated by the killing of birds in the intermedi- ate region and was driven back, until about 1850 it was found only on Martha's Vineyard and the near-by
114 THE SOCIAL LIFE OF ANIMALS
islands, and among the pine barrens of New Jersey. (56) By 1880, except for attempted and unsuccessful introductions elsewhere, it was probably restricted to Martha's Vineyard. In 1890-92 it was estimated that one hundred to two hundred birds remained on that island. Then several things happened at about the same time: prairie chickens were introduced and probably interbred with the vanishing heath hen, protection of the birds was stiffened, and collectors' prices went up! It is an interesting commentary that most of the museum specimens, of which 208 are known at present, were collected between 1891 and 1900, when the probable extinction of the heath hen was noised abroad. This is one of the modern handi- caps of small numbers; let a species or race become known to be rare, and museum collectors feel it their special duty to get a good supply laid in, just in case it does become extinct.
By 1907, when the Heath Hen Association was formed and employed a competent warden, the count had been reduced to seventy-seven. Massachusetts became aroused and purchased six hundred acres of heath hen range and leased a thousand acres more. The reservation was near a state forest which added another thousand acres of protected range. The birds responded to increased care and by 1916 it was esti- mated that there were two thousand in existence.
AGGREGATIONS OF HIGHER ANIMALS 11 5
Then came a fire, a gale, and a hard winter, with an unprecedented flight of goshawks, and in April, 1917, there were fewer than fifty breeding pairs. The next year, when there was an estimated total popu- lation of 150, the heath hen range was invaded by several expert photographers who took motion pic- tures of mating behavior. In the face of this disturb- ance at a critical time, still a good year allowed the birds to increase and again to spread over Martha's Vineyard. In 1920, 314 were counted; but thereafter a decline in numbers set in which was never stopped. The figures for those five successive years are: 117, 100, 28, 54, 25. At this point extra wardens were put on the job, who killed more cats, crows, rats, hawks, and owls, the enemies of the heath hen. The next year's count was 35; in 1927, there were 20; but in 1928, in a census that lasted four days, only a single male was found. No other bird was seen thereafter, though a reward of a hundred dollars was offered for the discovery of another. This single male was banded and released and was last seen alive on Febru- ary 9, 1932. With his death the heath hen became extinct. (18)
When this much is known of the decline in num- bers of a given species there should be some knowl- edge of the factors involved in its extinction. There is. In the earlier years, as I have indicated with re-
Il6 THE SOCIAL LIFE OF ANIMALS
gard to museum collecting, there was undoubtedly a considerable amount of poaching; but as population of heath hens declined, local sentiment turned in favor of protection and poaching decreased, both because of a more intelligent public reaction to the birds, and because of closer patrol by wardens. Dr. Gross, whose account I have been following, thinks that there was evidence of an inadaptability of the species, an excessive inbreeding, and, at the end, an excessive number of males. In such small populations the sex ratios frequently become highly abnormal. Disease and parasites took their toll. Predators, par- ticularly cats and rats, were active. The females hid their nests well and were faithful in remaining on them, so that they were killed off by the fires which at times whipped over the breeding grounds.
Over sixty thousand dollars was spent in trying to save the heath hen, but without success. In contrast to the laughing gull, which nested in a relatively safe place and which came back from a population as low as the heath hen's until the very last, this unfortunate species was not able to adjust itself and continue existence, even with as intelligent human help as could be mustered in its favor.
The general conclusion seems to be that different species have different minimum populations below
AGGREGATIONS OF HIGHER ANIMALS 1 1 7
which the species cannot go with safety, and that in some instances this is considerably above the theo- retical minimum of one pair.
By way of the laboratory, the coastal regions of Massachusetts, and South African grassland and for- est, we are arriving at a general biological principle regarding the importance of numbers present on the growth, survival and, as we shall see, upon the evolu- tion of species of animals.
Lacking definitive information on this last phase of the subject, we shall turn to mathematical explo- rations of its possibilities, as made primarily by Pro- fessor Sewall Wright. (127, 41) Although the ideas to be presented are essentially simple in principle, they are sufficiently novel and unfamiliar to challenge the closest attention.
I shall not indulge here in the details of the mathe- matical analyses, for the very good reason that I do not understand them. If I were not convinced, how- ever, that Professor Wright does understand them I should not present this outline. It is only fair to say that, in my opinion, in dealing with these ideal popu- lations Professor Wright cannot bring into sharp focus at one time all the factors that may be acting in nature. This is what he Has been courageous enough to attempt; the more nearly he succeeds, the more likely is the calculation to be too complex for
1 1 8 THE SOCIAL LIFE OF ANIMALS
presentation in detail except to highly specialized readers.
The environment is in a state of constant flux and its progressive changes, whether slow or fast, make the well-adapted types of the past generations into misfits under present conditions. The result may be rectified either by the extinction of the species, if it is not sufficiently plastic, or through reorganization of the hereditary types. In such a reorganization the simple Lamarckian reactions apparently do not op- erate; that is to say, when confronted with new, critical conditions, species cannot go to work and produce needed changes to order. The reactions are much more complicated than that.
To present the modern interpretation of this re- organization I need three technical terms which I shall define before using. Genes are bits of proto- plasm too small to be seen through the microscope, which are located in all cells and which are thought to be the bearers of heredity. They behave as indi- visible units, that is to say, a gene if present in an organism is either transmitted as a whole or not at all. Gene frequency is the term applied to the fre- quency with which a given gene is found in a popu- lation, relative to the total possible frequency (two in every individual). By mutation is meant a large or small hereditary change which appears suddenly,
AGGREGATIONS OF HIGHER ANIMALS IIQ
usually in the sense in which I shall use it, as a result of a change in one or more genes. With these three terms in mind we are ready to try to understand how the hereditary types may become reorganized.
Such a reorganization implies a change in gene frequencies. By this I mean now that there will be a decrease in the abundance of the genes which were responsible for the past adaptations that are now obsolete, and an increase in the frequency of those genes which allow an adaptation to the new condi- tions. Gene frequencies remain constant in a large population unless changed by mutation, selection or immigration. This is because of the unitary charac- ter, without blending, and the symmetry of the Mendelian mechanism of heredity.
These life-saving genes may have been present in the species for a million years as a result of long past mutations, without having been of any value to the species in all that time. Now under changed conditions they may save it from extinction. It is important to note that organisms do not usually meet changed conditions by waiting for a new muta- tion; frequently all members of a species would be dead long before the right change would occur. This means that since a species cannot produce adaptive changes when and where needed, in order to persist
120 THE SOCIAL LIFE OF ANIMALS
successfully it must possess at all times a store of concealed potential variability.
I may interject parenthetically that at times this appears to call for the presence of a considerable number of individuals as a necessary condition to provide the needed variations. A part of this reserve of variability may be of no use under any circum- stances; some characters may be useful, some may never meet with the circumstances under which they would have survival value; while others, though of no use or even harmful when they appear, may later enable the species to live under newly changed con- ditions.
Hereditary changes tend to be eliminated as soon as they run counter to decided environmental selec- tion. In large populations the results of mutations tend to stabilize about some average gene frequency, which represents the interaction between the rate of mutation and the degree of selection. Frequently mutation pressure pushes in one direction and selec- tion in another and the resulting gene frequency in the population represents a point or zone of equi- librium between these forces. In small populations which are not too small, selection between genes becomes relatively ineffective, and the gene fre- quencies drift at random over a wide range about a certain mean position. In very small breeding popu-
AGGREGATIONS OF HIGHER ANIMALS
121
lations, even though these may be small isolated colonies of a large widespread species, gene fre- quencies drift into fixation of one alternative or an- other more rapidly than they are changed by selec-
CHANCE I
0.5
LOO
Fig. 19. In small populations, genes drift into fixa- tion or loss largely irrespective of selection; the fre- quency of fixation or loss depends in the long run on the relative frequency of mutation and reverse mutation. (After Wright.)
tion or by mutation. Mutation, however, prevents permanent fixation. The condition at any given moment is largely a matter of chance.
Perhaps a diagram will help at this point. In Fig- ure 19 the horizontal axis shows the different gene frequencies in a population, and the vertical axis gives the chances of the population under considera- tion possessing any given gene frequency. At the left,
122 THE SOCIAL LIFE OF ANIMALS
the gene frequency is zero; that is, the gene in ques- tion is absent from the population for the time being. The height of the curve shows that there is a good chance of this happening. At the extreme right the gene has become fixed and all animals in the popu- lation have it; they are a pure culture so far as this gene is concerned. Again there is a high degree of probability that this may happen when numbers are few. But the intermediate condition, when the gene is present in some but not all of the animals, shows little chance of occurrence.
In such small populations, as has been said before, the gene frequency is determined mainly by chance; any given hereditary unit tends to disappear com- pletely or become fixed and occur in all members of the small inbreeding colony. Such a condition may have been reached in the inbred population of the heath hen on Martha's Vineyard.
With populations that are intermediate in size there is a greater variety of possibilities. Some genes are lost, others reach chance fixations, and others fluctuate widely in frequency from time to time. These conditions are shown in Figure 20.
If a given species is isolated into breeding colonies in such a way that but little emigration occurs be- tween them, a condition known to exist in nature, in the course of time, as Professor Wright shows, the
AGGREGATIONS OF HIGHER ANIMALS 123
species will become divided into local races. This will happen although at the time of separation the populations were all homogeneous and the environ- ment of all remains essentially similar.
If the environment does remain steady the larger
SELECTION
4
HUTATIOrt MUTATIOM
CHANCE
0 0.5 100
Fig. 20. In medium populations, gene frequencies drift at random about an intermediate point but not so much so that complete fixation or loss is likely to occur. (After Wright.)
colonies will tend to keep the same hereditary consti- tution as that which the whole species formerly had. (Figure 21.) Small breeding colonies will, how- ever, become pure cultures for different characters, and it is impossible to predict the course of the hereditary drift in any of these populations. As illus- trated in Figure 20, the fixation will be a matter of chance, and local races will result without any neces- sary reference to adaptation.
The snails in the different mountain valleys of Hawaii afford the classical illustration of this point.
124
THE SOCIAL LIFE OF ANIMALS
Each individual mountain valley has its separate species of snails. They are distinguished by size, by color markings, and by other characters which may be wholly non-adaptive.
Colonies which are intermediate in size will pre-
UTATION
CHANCE
A
I
0 0.5 1.00
Fig. 21. In large populations, gene frequency is held to a certain equilibrium value as a result of the oppos- ing pressures of mutation and selection. (After Wright.)
serve a part of the variability that will be lost in the smaller colonies. Even so, there will be some inde- pendent drifting apart of the various gene frequen- cies, so that these, too, will give rise to new local races. Professor Wright's calculations show that with mutation rates of the order of i:io,ooo or 1:100,000, such intermediate populations, optimal for evolution, will consist of some thousands or tens of thousands of individuals.
With small breeding populations, then, genes tend to become fixed or lost. Even rather severe selection
AGGREGATIONS OF HIGHER ANIMALS 125
is without effect. Individual genes drift from one state of fixation to another regardless of selection. In large populations, gene frequencies tend to come to equilibrium between mutation and selection, and if selection is severe, there tends to be a fixation of the gene or genes that carry adaptive modifications, and evolution comes to a standstill.
With a population intermediate in size, when there are enough animals present to prevent fixation of the genes on the one hand, but on the other, not enough animals to prevent a random drifting about the mean values determined by selection and muta- tion, then evolution may occur relatively rapidly. The results obtained will depend upon the balance between mutation rate, selection rate, and the size of the effective breeding population.
In one more case the effect of differences in sever- ity of selection was worked out by Professor Wright (Figure 22). With a moderate mutation rate, if the selection is relatively weak, mutation pressure may determine the result and the given character will then drift to fixation or, as shown in the diagram, to extinction. As selection pressures increase, selection tends to take charge of the end products, and, if slight, there is a wide variation about a mean; if more intense, the amount of variation becomes less and less.
126
THE SOCIAL LIFE OF ANIMALS
When a species is broken up into different breed- ing colonies, as it is with the snails in the Hawaiian
fN5=80
Fig. 22. As intensity of selection increases it becomes more and more dominant in determining the end result, and the degree of variation is lessened; 4Ns gives selec- tion pressure. (From Wright.)
valleys, (57) it can be similarly shown that the effects produced depend on the rate of emigration between colonies, as well as selection pressure, mutation pres- sure, and population size, other factors being con- stant. Cross-breeding introduces genes into a popula-
AGGREGATIONS OF HIGHER ANIMALS 127
tion in a way that is essentially identical with muta- tion in its mathematical consequences; however, similar results may be obtained in a much shorter time by cross-breeding. And in fact all the different results which have just been illustrated can be dupli- cated by varying the numbers of the emigrants.
This is not the place to explore all the implica- tions and possibilities of these interesting analyses. The highly significant conclusion has been reached that if a species occurs not as a single breeding unit but broken into effective breeding colonies which are almost isolated from each other, the members of different colonies, given sufficient vigor, may evolve into dissimilar local races. If one of these becomes well adapted to its environment it may increase in numbers and send out numerous emigrants. If these emigrants find and interbreed with members of other less advanced colonies they will grade these up until they resemble the most adapted colony. This part of the process resembles a stock breeder's grading up of a mediocre herd of cattle by repeated infusions of new and improved ''blood" into his herd. The sig- nificant thing here is that the random differentiation of local populations furnishes material for the action of selection on types as wholes, rather than on the mere average adaptive effects of individual genes.
The end results will vary even when the original
128 THE SOCIAL LIFE OF ANIMALS
population was homogeneous, and when mutation rates are similar throughout, even though selection is in the same direction in all parts of the different colonies. The primary factor under these conditions will be that of effective breeding population size, and there will be greater chance for varied evolution among the populations that are intermediate in size, as contrasted with those which are small or large, and still greater chance for evolution when a large species is broken into small breeding colonies which are not completely isolated from each other.
This argument, even as I have simplified it, is not too easily followed the first time one goes over it. Perhaps my use of an old teaching trick, that of repe- tition of the same ideas with different words and different illustrations, may be forgiven. In doing so I am still leaning heavily on Professor Wright. The series of diagrams shown in Plate IV are built on one fundamental background. In perspective we see two elevations, one higher than the other, and two depressions which are the low points in a valley between the two peaks. Every position is intended to represent a different combination of gene fre- quencies. The peaks represent gene combinations which are highly adaptive; the depressions represent those that lack adaptive value. The degree of adap- tiveness is shown by the height occupied by the given
I#
|
€ |
o^ |
.X- O
clI
PLATE IV. A population originally possessed a set of gene combinations of some slight adaptive value (dotted line). With increased mutation rate it can ex- pand to less adapted levels (A); with increased selec- tion it contracts (B); if the environment changes the gene frequency must shift (C); with small numbers and close inbreeding the course of evolution is erratic and extinction usually follows (D); with larger num- bers, evolution takes place more readily (E); most read- ily, when a large population is broken into local colonies with inter-emigiation (F). (Modified from Wright.)
AGGREGATIONS OF HIGHER ANIMALS 129
population. The variability of the population is shown by the size of the area that is occupied. Every individual in a species may have a different gene combination from every other, and yet the species may occupy a small region relative to all the possi- bilities.
We may call the lower peak Mount Minor Adap- tation and the higher one Mount Major Adaptation. In Figure A we find a population which is fairly well-adapted, but not so much so as if it occupied the higher peak. Its original position and its variability are shown by the dotted circle. As a result of increased rate of mutation or of reduced selection, or both, the variability of the population has increased and it now spreads down to lower positions on this Mount Minor Adaptation. It contains more aberrant indi- viduals and even freaks than when subject to less frequent mutation or to more severe selection, and a freak may appear that is more adaptive; but this important end has been achieved at the expense of the variability which might have made a major ad- vance possible.
Figure C introduces a different situation. As a result of environmental change Mount Minor Adap- tation has disappeared and the adapted population has been able to move to a new location at about the same level formerly occupied; now it is on the slope
130 THE SOCIAL LIFE OF ANIMALS
of Mount Major Adaptation, and if selection con- tinues may be expected to move up that adaptive peak. A continually changing environment is un- doubtedly an important factor in evolution.
The effects of population size are illustrated in the next three diagrams. The general background is the same as in Figures A and B. In Figure D is shown the effect of a decided reduction in population size, and consequently in variability, in the species that formerly occupied Mount Minor Adaptation. It is in fact so small that selection has become ineffective and the different hereditary qualities shift to chance fixations. As non-adaptive characters become fixed at random the species moves down from its peak over an erratic, unpredictable path. With reduction of population size below a certain minimum, control by selection between genes disappears to such an extent that the end can only be extinction.
With the species population intermediate in size, with the same mutation and selection rates as before, gene frequencies move about at random but without reaching the degree of fixation found in the preced- ing case. Since it will be easier to escape from low adaptive peaks, the population will tend finally to occupy the more adapted levels. The rate of progress is, however, extremely slow.
Finally, in Figure F, we see the case of a large
AGGREGATIONS OF HIGHER ANIMALS 13I
species which has become broken up into many small local races, perhaps as a result of restricted environ- mental niches. Each of these local races breeds largely within its own colony, but there is an occasional emi- gration from one to another. Each tends, if it is small in number, to give rise to different variations which shift about in a non-adaptive manner. The total number of relatively stable variations will be much greater since the total number of individuals is so much larger than in E. Under these conditions the chances are good that some of the local colonies will escape from the influence of Mount Minor Adapta- tion and manage to cross the valley to Mount Major Adaptation. Here the race will expand in numbers and will send out more and more emigrants which will interbreed with the stocks in the less adapted colonies and tend to grade them all up toward a higher adaptive level.
The conclusion is as Professor Wright says: "A subdivision of a large species into numerous small, partially isolated races gives the most effective setting for the operation of the trial and error mechanism in the field of evolution that results from gene com- binations."
In the rate of evolution, therefore, population size is as important as we have seen it to be in the growth
1S2 THE SOCIAL LIFE OF ANIMALS
of individuals or in the gTO\\ih of popnlation num- bers: and the optimal population size does not coin- tide \sith either the largest or smallest possible but lies at some iiuermediate point.
V.
Group Behavior
IN THE second chapter I told of how I stumbled on the fact that in the breeding season the normal be- havior of isopods is affected by numbers present. Such effects have long been known for many types of behavior, and it would not be profitable here to catalogue and analyze all the cases that are on record. Rather, as before, I shall select certain well-authenti- cated examples of breeding reactions and of other types of behavior. Those which are chosen are espe- cially noteworthy because of the behavior pattern which is involved, or because freshly observed, or both.
And here is a shift in emphasis. I have been stress- ing the existence of a widespread, fundamental auto- matic co-operation which has survival value, and have given evidence that it is a common trait in the animal kingdom. In this chapter I shall discuss group be- havior which may or may not have immediate sur- vival value. In each instance, and throughout the discussion as a whole, I shall be engaged in trying
133
154 THE SOCIAL LIFE OF ANIMALS
to find to what extent behavior is influenced by the presence of others, and shall not consistently attempt to assay possible values which may or may not be involved.
With many more or less social animals the group up to a certain size facilitates various types of be- havior. This is frequently called social facilitation.
Shore Line
Fig. 23. Manakin males establish rows of mating courts in the Panamanian rain-forest. (From Chapman.)
One phase of social facilitation is illustrated by some observations of the mature student of birds, Frank E. Chapman, (28) near the tropical laboratory on Barro Colorado Island in the rain-forest of Panama. Mr. Chapman found that males of Gould's manakin establish lines of courting places (Figure 23). The manakin is a small warbler-like bird, delicately colored and relatively inconspicuous. Each of the courting places is occupied by a single male; the line thus formed extends for many yards through the undergrowth of the rain-forest. From time to time each day during the long nesting season, the males resort to their individual cleared spots on the forest
GROUP BEHAVIOR 1 35
floor and make their presence known by a series of snaps, whirrs and calls which may be heard as far as three hundred yards. The females, who are more quiet and retiring, apparently are attracted by the line of males; they come individually from the sur- rounding thickets and each mates with one of the males. The evidence suggests that they are attracted from a greater distance by the spaced aggregation of males than they would be by isolated courting places. The more or less organized line of males in breeding condition apparently facilitates the mating of these jungle birds.
This is a highly specialized example of the wide- spread phenomenon of territoriality which can be recognized even among breeding fishes, (103) and which has been much studied of recent years in birds. (65) Typically the male birds arrive first in the spring and take up fairly well-defined territories in the same general area, which they defend from in- truding males. Then the females come in and flit from territory to territory before settling down to raise a brood with one particular male. There is always the strong suggestion that the presence of a number of singing males, even if spaced about in different territories, attracts and hastens the accept- ance of some one of them by an unmated female.
Group stimulation of the amount of food taken
136
THE SOCIAL LIFE OF ANIMALS
has been reported for various animals, including rats, (59) chickens (23) and fishes. (118) I shall illus- trate by some of the experiments conducted in our laboratory by Dr. J. C. Welty. These have been
150
125
a te.
s:ioo
^ 75
<
<
o
u.
o
le 25 la <0
E
C
• GROUPS OF FOUR O ISOLATED
DAILY FEEDmS 5
Fig. 24. Many kinds of fishes eat more if several are present than if they are isolated. (From Welty.)
amply verified by other research workers. In connec- tion with experiments on the effect of numbers on the rate of learning in fishes, which will be discussed later. Dr. Welty undertook to find whether grouped fish ate more or less than if they were isolated. The results of a typical experiment are illustrated in Figure 24.
Goldfish were photographed to scale, and those of
GROUP BEHAVIOR I37
similar size were selected for experimentation. Two groups of four each were placed in separate crystal- lizing dishes and eight others were isolated each into a wholly similar dish. The different dishes were sep- arated by black paper so that vision from one to the other was impossible. A known number of the small crustacean, Daphnia, were introduced daily into each dish. These living Daphnia had been screened so as to select the large animals only. As shown by the fig- ure, fish in all groups of four ate decidedly more on the first three days of the experiment. At this time the two lots were shifted. Those that had been grouped were now isolated, and vice versa. There was an immediate shift in the numbers of Daphnia taken, with the newly isolated animals now eating less than the accompanying groups. This indicates that we are dealing with an effect of numbers present rather than with chance differences in individual appetites. This difference kept up steadily until the last three days of observation, when an interesting complication arose. By this time the grouped fish were receiving a total of over six hundred Daphnia daily, including those which were eaten and the extras added to insure an economy of plenty. Each isolated fish was receiving only one-fourth as many. Now six hundred and more large Daphnia, each about an eighth of an inch long, make quite a swarm
9,8 THE SOCIAL LIFE OF ANIMALS
in a none-too-large crystallizing dish. The consump- tion of food per animal by the grouped fish fell off, and as was shown by appropriate tests, this was due to the action of a so-called confusion effect. When fewer Daphnia were present, a fish might be observed to swim after an isolated crustacean and eat it, whereas a dozen Daphnia or so in the immediate field of vision seemed to offer conflicting stimuli that blocked the feeding response. Working on this suggestion, one group of four was given the usual quota of some six hundred Daphnia all at once; another group was given only one hundred at a time, and when these were approximately all eaten then another hundred would fjc introduced, and so on until the end of the regular feeding period. This prevented the Daphnia from being too dense at the beginning of the hour's feeding time. The isolated fish were fed as usual. Under these conditions the grouped goldfish which were fed one hundred Daphnia at a time ate defi- nitely more than those given the whole confusing mass at once.
Here we come upon two, not one, mass effects. In the first place we see that the fish in groups of four were stimulated to eat more food than if isolated, and this depended on their state of aggregation. But, incidental to this demonstration, we hnd that in the presence of too many animated food particles a con-
GROUP BEHAVIOR 1 39
fusion effect arises which decreases the feeding effi- ciency of the fish.
It has been suspected for years that such a confu- sion effect exists and has survival value for small animals flocking together in the presence of a preda- tor, such as small birds in the region of a hawk. These observations of Welty's make the best demon- stration that I know of the existence of such an effect, in this case the Daphnia in the presence of the fish. I am less interested in this confusion effect at present than in the demonstration of social facilita- tion in feeding, a phenomenon which has been shown to exist for a number of fishes, including zebra fish, paradise fish, goldfish and guppies of the more usual aquarium varieties, and the lake shiner, No- tropis atherinoideSy as well.
None of these fishes is very social, that is, none of them group into close schools. For evidence of similar social stimulation among social animals it is interesting to examine the effect of numbers present on the digging behavior of the highly social ants. The account of this work was published in 1937 by Professor Chen of Peiping, China. (29)
These ants, a species of Campanotus, dig their nests in the ground. It was found that all the worker ants of this species are capable of digging a nest when in isolation, but that the rate of work varies
140 THE SOCIAL LIFE OF ANIMALS
with different individuals. If marked ants, whose reaction time has been tested in isolation, are placed together in pairs or in groups, they will start work sooner and will work with greater uniformity than if alone.
With oriental patience. Professor Chen and his assistants collected and counted the number of the tiny pellets of earth which were dug by different individual ants when isolated, and when members of groups of two or three ants. They found that the number of pellets removed is greater when the ants work in association with others than when each works alone. This accelerating effect is greater for slow than for rapid workers; when ants with inter- mediate working tendencies were tested (Figure 25) they were found to be speeded up when in com- pany with a rapid co-worker and relatively retarded when placed with a slowly working ant. Interestingly enough, there was no difference between the stimu- lating effect of one additional ant and of many ants on the rate of work of a given individual. The social facilitation seemed maximum for these digging tests when only a second individual was present.
Ants which regularly work rapidly were found to be physiologically different from those that work more slowly. The faster workers were more suscepti- ble to starvation, to drying, and to exposure to ether
0 5 ra e 20 25 30 35 40 45 50 55 60 5 10 15 20 25 30 35 40 45 50 55 60 TIME IS MINUTES
Fig. 25. An ant which works at an intermediate rate (Ml) may be speeded up if placed with an ant which works more rapidly (Rl) and slowed down if put with a slower worker (SI). (From Chen.)
142 THE SOCIAL LIFE OF ANIMALS
or to chloroform. Tests that have been made by others indicate that animals that are more active physiologically usually succumb sooner under such adverse conditions, just as these rapidly-working ants were found to do. These are exceedingly interesting results because here we see that ants with apparently innate differences in speed of fundamental processes are affected in their speed of digging by the presence or the absence of a nest mate. The ant of intermediate speed, presumably with an intermediate underlying reaction system, is most interesting of all, because it can be either speeded up or retarded according as it is placed with an active or a more passive individual.
In this connection it has been known for over a decade scientifically what was common sense before that time, namely, that human animals, whether adults or children, can accomplish more mental and physical work, at least of certain kinds, and will work with greater uniformity when in association with others doing similar tasks, than if obliged to work in isolation. (15, 84)
Such considerations lead directly to problems con- cerning the effect of numbers present on the rate of learning in man. Here we find a set of questions that have great and immediate human significance. The world over, the training of the young animals of their own species is one of the major preoccupations
GROUP BEHAVIOR 143
of mankind. This is particularly true in the United States, where we are engaged in mass education on an unprecedented scale. This teaching of the young to the extent to which we are attempting it is an expensive business in time, in effort and in money. We need to know, therefore, the number of these interesting young animals that can be trained to- gether with best results. In other words, what is the optimal class size for the various levels of training from pre-school days through the preparation for the doctor's degree and further?
In part, the proper answer to this question calls for a statement of educational objectives. The devel- opment of strong individuality, for example, is not necessarily accomplished by the same teaching meth- ods and class size which favor the growth of conform- ity to group patterns; and the rapid development of mastery of so-called skills may call for difiEerent num- ber relations than those needed for the mastery of logical thought.
Even without positive information we can guess that the tutorial method with individuals or very small groups will best serve some ends while others will be achieved most readily in larger groups. The question, or a simplified part of it, thus becomes: What class size favors optimal rate of learning of the usual class material presented at different ages?
144 THE SOCIAL LIFE OF ANIMALS
As might be anticipated, the difficulties of human experimentation being what they are, it is hard to collect accurate information on this point. Much depends on the comparative accuracy of the sam- pling, and also on more subjective factors, such as the attitude of the teacher and of the students toward large and small classes. There is also a factor which I have not seen mentioned in the literature on the subject, the effect on the student of realizing or sus- pecting that he is an object of experimental interest, an educational guinea pig. This stimulus is more likely to be potent, in my opinion, when the student is a member of a class which is unusual in size.
In the more careful studies, results of which have been published, the class numbers have ranged from "small" through ''medium" to "large." The "small" experimental classes apparently have about twenty to twenty-five members; this represents a more usual experience to the student, and he is more likely to be conscious of class size when he is a member of a large class of seventy-five or more than when he is in a small class or a medium-sized one of thirty-five to forty. The sizes that are counted "large" or "small" vary greatly, sometimes in the same experimental treatment, so that frequently the comparisons are between larger and smaller classes, both medium in size, rather than between real extremes in numbers.
GROUP BEHAVIOR I45
Frequently, too, the teaching practice varies in the two classes. Thus in one experiment the smaller classes in high-school geometry contained about twenty-five, while the large ones had about one hundred members. In the large classes a student helper was present for every ten class members. These helpers were superior students in geometry of the preceding year. As nearly as I can discover, there were no student helpers in the small classes. Under the conditions it is perhaps not unexpected that a better showing was made by those in the large classes. With them, there were present not only more in- structors per student but these were people of nearly their own age, who could be approached without hesitation not only in class but out of class and even out of school hours. Every mature teacher knows that even with the best intention and the most demo- cratic attitude, age differences widen the gap between the teacher and the taught, whatever other compen- sations there may be.
The most comprehensive experiments I have seen reported in this field are those of the sub-committee on class size of the committee on educational re- search at the University of Minnesota. (66) These were carried on at the college level and involved 109 classes under twenty-one instructors in eleven de- partments of four colleges in the University of Min-
146 THE SOCIAL LIFE OF ANIMALS
nesota. Forty-two hundred and five students were observed in large classes, and 1,854 in small ones; of these 1,288 were paired as to intelligence, sex and scholarship before the experiment began. One of each pair was assigned to a large and one to a small class in the same subject taught by the same instruc- tor. In this way the obvious variables were controlled as well as is humanly possible, unless we could have a large number of identical twins with which to experiment.
In 78 per cent of the experiments a more or less decided advantage accrued to the paired students in the large classes, and at every scholarship level tested, the paired students in the large sections did better work than their pairs in the smaller ones; the excel- lent students appeared to profit somewhat more from being in large classes than their less outstanding fellows.
Of the available data, a re-examination of the sum- maries indicates that there is on the average a dif- ference in the means in the final grade of 4.1 points, favoring the students in the larger classes. There is a statistical probability of matching this by random sampling of four chances in ten million (P = 0.0000004), and this despite the fact that the majority of the class comparisons did not give signifi- cant differences when considered alone.
GROUP BEHAVIOR I47
The numbers in the smaller classes usually ranged from twenty-one to thirty, but in some classes dropped as low as twelve; in the larger classes there were usually thirty -five to seventy-nine students; in the largest, one hundred and sixty-nine. Under the conditions which prevailed in these classes in psy- chology, educational psychology and physics, the stu- dents in the larger class sections made slightly but significantly higher final grades than those in smaller sections of the same subject taught by the same instructor.
So much for objective experiments. It happens that subjective estimates, made both by teachers and by students at Minnesota, favor the smaller rather than the larger classes. It was even true that the students were better satisfied with the marks re- ceived in smaller classes than they were with the slightly higher grades given them in the larger sec- tions.
The general attitude seemed somewhat like that toward a friend of mine who teaches general mathe- matics at Purdue University. He is an experienced and excellent teacher. His program for one semester required that he should meet a normal-sized class of thirty to thirty-five at eight o'clock, and that at nine o'clock he should meet a class of double the size in a larger room, to repeat the same subject matter.
148 THE SOCIAL LIFE OF ANIMALS
At the close of the semester the two sections were asked to rank the instructor on many different points. Uniformly the students in the larger section rated him lower than those in the smaller section, in such matters as teaching skill, pleasantness of voice, neat- ness of appearance and personal attractiveness!
I have had a fairly extensive teaching experience, which has included work in grade- and high-school teaching, as well as over twenty-five years of teaching at the college and university level, during which time I have taught classes of almost all sizes, from those of over six hundred at the University of California to the graduate classes of three or four that come my way; and I must confess to a personal prejudice against these very large classes. Even when using the same lecture notes, I do not give the same lecture to five hundred students that I give to forty or fifty. On the other hand, even with graduate classes and advanced seminars I am prejudiced in favor of hav- ing enough students, which means at least eight to ten, to give a certain esprit de corps to the group. Such personal opinions have their value, particularly when they click with experimental results such as those outlined by Hudelson from the experiments at Minnesota. It is unfortunate that those experi- ments did not test either the upper or the lower limits of class size which are conducive to good class-
GROUP BEHAVIOR I49
room performance on the part of the students; and I know of none that does test these points adequately.
Some of the difficulties which are inherent in ex- perimentation on the effects of class size on the rate of learning in man can be obviated by the use of non-human animals. This procedure does not solve all the requirements for elegant objective experimen- tation, and has the additional real difficulty of elim- inating all possibility of adding subjective impres- sions to objective findings, a point which makes one of the strongest arguments for experimentation on man when feasible.
In some respects the most completely controlled experiments on the effect of numbers present on the rate of learning are those that Miss Gates and I performed some years ago, using common cock- roaches as experimental animals. (52) Earlier work by two independent investigators had shown that cockroaches can be trained to run a simple maze, and can show improvement from day to day. In our experiments we found that the cockroaches could be trained to run the maze we used by fifteen to twenty- five successive trials on a given day, and showed defi- nite improvement both in time taken to run the maze and in number of errors. However, unlike the experience of our predecessors, these University of
150 THE SOCIAL LIFE OF ANIMALS
Chicago cockroaches could not carry over the effects of training from one day to the next.
The reason for this difference between our cock- roaches and those around St. Louis and in Germany is not known. It may be that at the University of Chicago, despite our reputation for scholarship, the local cockroaches have a low IQ, or it may be that since we used animals from the bacteriological lab- oratory, because of their unusual size and physical vigor, we were unconsciously selecting the dumber sort. Or perhaps, contrary to our plan, we set them a problem which is intrinsically more difficult for the cockroach mentality. In any event, it is important to remember that our cockroaches forgot overnight any- thing they may have learned the day before. As it turns out, this was fortunate for the experiments we were carrying on, because we could match up indi- vidual cockroaches with the same speed of learning in pairs or groups of three for later tests without fear of a carryover from their previous experience.
The maze used is shown in Figure 26. It consisted of a metal platform from which three runways ex- tended, each about two inches wide and a foot or so long. The two side runways ended blindly, but the center one led to a black bottle, which allowed the cockroaches to escape from the light. This apparently
GROUP BEHAVIOR I5I
was a reward for cockroaches which, when possible, give a negative reaction to light.
The three-pronged set of runways was mounted about half an inch above a pan of water, which the majority of the cockroaches tended to avoid, and so kept on the runways. The tests were all made in a
Fig. 26. A simple maze used in training cockroaches.
dark room and light was furnished by a single elec- tric bulb mounted just above the point where the central runway left the main platform. In other words, the cockroaches, which are negative to light, had to learn to run through the area of strongest illumination in order to reach the dark bottle which served as a reward. After two minutes' rest in the dark bottle the cockroaches were literally poured out onto the platform of the maze without being touched by the experimenter, and observation of them began again.
The problem as set was about at the limit of cock-
152
THE SOCIAL LIFE OF ANIMALS
roach ability. Approximately one-third of the insects tested never learned to stay on the maze; whenever they were placed on it they proceeded immediately to run off into the underlying water. Of the two-thirds
n
16 15 14 13 12 U
9-
8-
isoIa."tecL * paired J roup of 3
\
I
15 Trials
— r- 20
25
Fig. 27. Isolated cockroaches make fewer errors on the maze than the same animals paired, and still fewer than if three are being trained together.
GROUP BEHAVIOR
153
that did learn to remain on the maze, a half, or an- other third of all those tested, did not show improve- ment in speed of reaching the bottle, after repeated
|
12- |
isola-tecL •— * |
|
|
U- |
paired ^roup of 3 •—• |
|
|
10- 9- 8- 0 7- t (>■ c f 5- |
\ \ |
|
|
A- |
>s^^^ |
|
|
3 |
^ . |
|
|
2- |
||
|
1- |
10
15 Trials
20
25
Fig. 28. They also take less time.
trials. Thus only one-third of the cockroaches we tested showed improvement with experience, and, as I said before, they forgot overnight all that they learned during the day.
As shown in the summarizing graphs (Figures 27 and 28), isolated cockroaches made fewer errors per trial throughout the whole training period. They
154 THE SOCIAL LIFE OF ANIMALS
also took less time to run the maze than when the same animals were members of pairs or of groups of three. Turning the comparison around, paired cock- roaches took longer time per trial and made more errors than when isolated, and groups of three took still longer and made more errors than those in pairs.
A study of the rate of improvement shows that during the early part of the training, as is indicated by the slant of the graphs, so far as time spent is concerned, paired cockroaches improved more rap- idly than they did if isolated or in groups of three, and those placed three together on the maze im- proved somewhat more rapidly than they did when isolated. Thus, while the presence of one or two extra cockroaches slowed down the speed of reaction on the maze and increased the number of errors made at all times, yet the rate of improvement in speed of re- action was higher when more than one was present. There was, however, no significant difference in rate of improvement as measured by number of errors.
Excluding this one aspect of rate of improvement in time spent on the maze, in all other phases of the experiment isolated cockroaches turned in a bet- ter learning performance than they showed when more were present. Evidently under the conditions of our experiments the tutorial system usually works best with cockroaches.
GROUP BEHAVIOR 155
Essentially the same sort of experiment was tried with isolated and paired Australian parrakeets, which are commonly called love birds. (11) Rather naively, perhaps, I thought that since these birds so readily pair off, perhaps two might learn to run a simple maze more rapidly than a single individual would. This turned out to be entirely a mistaken idea. I shall spare you the details concerning this maze; it was adequate in size, so that two birds could pass through practically abreast. Almost all the ninety- odd birds that were tested learned easily to run the maze and normally reduced their time per trial from about two minutes to a few seconds, after six or seven days of training. Errors also were reduced, and several of the birds were trained so that they ran the maze day after day with no errors at all.
The selected summarizing graphs (Figures 29 and 30) will outline the results obtained. It made no dif- ference whether the birds were caged in pairs or separately; if placed alone in the maze the perform- ance was similar. If, however, two birds were put together in the maze, the speed was reduced and errors increased as compared with the scores made by isolated parrakeets. It made no difference whether two males, two females or a male and a female were trained in the maze together; there was always in- terference. The tendency was for the more rapid
156 THE SOCIAL LIFE OF ANIMALS
bird to slow down rather than for the slower bird to speed up. The paired birds tended to take the same time and to make the same errors. Given suf- ficient training, they might make perfect scores so
Trials
Fig. 29. Parrakeets learn equally well if trained when isolated, whether they are caged singly or in pairs. A, time per trial; B, errors per trial.
GROUP BEHAVIOR
157
far as errors were concerned, but even after long training the performance of pairs was always more
Fig. 30. Parrakeets learn more rapidly if trained alone than if two are placed together in the maze. A, time per trial; B, errors per trial, (The upper curve is unsmoothed; the lower three have been smoothed mathe- matically.)
erratic than that of isolated birds. When birds that had been trained to a consistent level of excellence were exchanged so that those formerly isolated were paired and those formerly paired were isolated, their behavior in the maze took on the characteristics
158 THE SOCIAL LIFE OF ANIMALS
usually shown by paired and by isolated birds, prov- ing that the type of reaction given was a result of the numbers present rather than of the working of other factors. With these love birds then, contrary to the original assumption, all indications were that being paired in the maze slowed down the rate of learning and increased the erratic character of their behavior.
Our experience with the general problem did not end here. I teach at the University of Chicago a favorite course called Animal Behavior. In this class the beginning research students attempt some small problem and frequently make good progress toward its superficial solution. One of these student projects has been the training of the common mud-minnow to react to traffic lights. The fish were trained to jump out of water and obtain a bit of earthworm when red was flashed. Under the green light they were conditioned to retire to one of the bottom cor- ners. If they did jump under green light they were fed filter paper soaked in turpentine. Within two months a lot of fishes, isolated one in each small aquarium, could be trained so that they would have been given an A for the project if they had been properly enrolled students.
When, however, several fishes were placed together in the same aquarium and an attempt was made to
GROUP BEHAVIOR 159
train all at the same time the rate of learning was retarded. Paired fish reacted as well as if they had been isolated, but the reactions of groups of four were slowed down, and those of ten even more so. Two fish would rarely jump at once, and when some one individual was getting set to jump for the food under the red light, another would frequently come along and give him a jab in the belly which would stop all tendency to jump for the time.
One more instance remains to be reported. Dr. Welty, who has been mentioned before, undertook to train goldfish to move forward from the rear screened-off portion of an aquarium through a door into a small forward chamber where each was fed just after it came through the opening. (118) An aquarium-maze, similar to those used, is shown in Fig- ure 31. The signal to the fish that it was time to react came from increasing the intensity of light in the aquarium and opening the door between the two compartments. Under Dr. Welty's careful coaching the fish improved rapidly in their speed of reaction and usually had reached a good level of performance by the sixth day of training.
In his experiments almost a thousand fishes were trained at one time or another. The results of a sample experiment are recorded in Figure 32. In this test there were eight goldfish, each isolated in indi-
i6o
THE SOCIAL LIFE OF ANIMALS
vidual aquaria, four sets of paired goldfish, two lots of four placed together, and one group of eight in one aquarium. As shown by the graph, there was a
Fig. 31. Feeding a fish which has just come through the opening from the larger side of the aquarium. (From Welty.)
marked group effect on the rate of learning. The speed of first performance of the untrained fishes was most rapid with eight present and slowest with iso- lated goldfish. In the early days of rapid learning the same order held. This experiment was repeated sev- eral times with identical results. Under these condi- tions there seems to be little doubt but that the
QE37
100-
O ISOLATED
Q PAIRS
• GROUPS OF FOUR
€ GROUPS OF EIGHT
TRIAL 5 10 14
Fig. 32. Goldfish learn to swim a simple aquarium- maze the more readily the more fish there are present. (From Welty.)
i62
THE SOCIAL LIFE OF ANIMALS
groups of goldfish learned to move forward and se- cure food more rapidly than the same number of isolated fish.
The conditions of the experiments allow certain
TRIAL
Fig. 33. Isolated goldfish learn the problem set for them less rapidly, and unlearn it more readily. (From Welty.)
types of analyses to be made. One of these is to test the tenacity with which the newly acquired habit will be retained. A set of fish was trained as usual (Figure 33). After ten days, when the grouped fish had been letter-perfect for four days, although the isolated goldfish were still taking some three min- utes per trial, the experiment was changed; when-
GROUP BEHAVIOR 163
ever the fish came forward through the gate they were offered pieces of worm soaked in acetic acid. The isolated fish, perhaps because they had not learned to perform so well, perhaps because they were isolated or for some other reason, ceased to react rapidly, and on the twenty-ninth day they were averaging fifteen minutes per trial. The grouped fish were much more steady in behavior, and per- sisted in coming forward with relatively little change until the twenty-seventh day; and even then the old conditioning held for most of the fish most of the time. Many individuals persisted in coming forward through the gate for a long time after they ceased biting or even swimming toward the acid-treated worm.
When a group of fish are reacting together, if a given individual moves forward through the gate to the feeding space, others may follow because of a group cohesion. It is obvious that if a fish is isolated and moves forward, the faster reaction cannot affect the behavior of other isolated fish.
With this in mind. Dr. Welty undertook a series of experiments in which there were two partitions in the aquarium, with one door opening forward and another door opening through the other parti- tion toward the rear of the aquarium (Figure 34).
164 THE SOCIAL LIFE OF ANIMALS
The fish were placed in the central space and those in half the tanks were trained to come forward as usual. In the other half, two selected fish were con- ditioned to come forward and two were similarly
A ^
V
Fig. 34. The aquarium-maze used in training part of the fish to come forward and part to go to the rear to be fed. (From Welty.)
trained to move to the rear compartment to be fed. The experiment was tried several times with gold- fish, the minnow, Fundulus, common at Woods Hole, and another marine minnow, Cyprinodon. For one reason or another, only one series in which the fish were comparable was successfully completed. The re- sults are shown in Figure 35. Generally speaking, the cohering groups of Cyprinodon learned more rapidly and reacted more steadily than the separat- ing groups. This, then, is one factor that is working,
GROUP BEHAVIOR 165
at least at times, in causing grouped fish to learn more rapidly in a simple aquarium-maze than iso- lated fishes under similar treatment.
As the goldfish move forward in the usual divided
ojo.i'
e SEW\RATinG GROUPS O COHERIMQ GROUPS
TR»AL 5 10 J5 19
Fig. 35. Cyprinodon learn to move in a body more readily than to split into two separate groups. (From Welty.)
aquarium there comes a time when one or more fish may be in front of the screen, and the others in the rear of this advance guard. It was obviously a part of the investigation to find the effect these more rapidly reacting fish had upon their fellows merely as a result of being in the forward chamber. Con- ceivably they may have served as a lure. Another pos-
i66
THE SOCIAL LIFE OF ANIMALS
sibility is that a rapidly learning individual becomes a leader in the reaction of the whole group.
Both of these possibilities were tested experi-
Qli5*
|
70 |
• 9 |
|
^ OCOMTROL |
|
|
60 |
\ ©PLUS LURE |
|
\ • PLUS LEADER |
|
|
50 |
• #1 |
|
40 |
■ u |
|
30 |
■ n |
|
20 |
■ \l |
|
to f^io D C |
- \_ |
OCOMTROL ©PLUS LURE • PLUS LEADER
TRIAL
4 5
A
Fig. 36. Goldfish learn more readily if accompanied by a trained leader than if there is a fish in the proper position to act as lure. (From Welty.)
mentally by Dr. Welty, with results which are sum- marized in Figure 36. Three sets of aquaria were established. In the control aquaria all the goldfish, of which there were four in each tank, were fish which had had no previous experience in these ex- periments. These were trained as usual. In another set, an untrained fish was kept in each forward com-
GROUP BEHAVIOR 167
partment as a lure and four untrained fish were placed in the rear compartment. These fish were trained as usual; the so-called lure-fish was fed after the first of the untrained lot came through the gate- way. In the final set of aquaria a trained fish was in- troduced along with the four untrained fish. When the light was admitted and the gate was raised this trained fish moved forward, came through the gate- way, and was fed immediately. The others followed. As the graphs show, after the first day there was lit- tle difference in the reactions given by the control fish and by those which had a lure-fish in front of the screen. The fish with a trained leader generally gave more rapid reactions than either of the others. There is always a temptation to make comparisons between the learning behavior of these laboratory animals and that of men. Direct comparisons should usually be avoided. However, in human terms, the goldfish reacted more rapidly in the presence of a trained leader which went through the whole be- havior process with them, than they did to the pres- ence of one of their kind as a lure-fish in the forward compartment, a sort of signpost to proper behavior. Evidently leaders working with these goldfish can in- fluence them more than fish which by their posi- tion merely show them where they can come. It seems fair to say that with these fish demonstration
i68
THE SOCIAL LIFE OF ANIMALS
teaching is the most effective method yet discovered.
Still another attempt was made to study group
cohesion in these goldfish. For this purpose aquaria
Fig. 37. An aquarium-maze arranged to test the power of observation of fish placed in the side compart- ment. (From Welty.)
were arranged like those in Figure 37. At the side of the usual aquarium-maze a narrow runway was placed into which untrained goldfish were intro- duced. In half of the tanks the glass partition was clear and allowed the fish to see the reaction of those
GROUP BEHAVIOR
169
in the larger aquarium-maze. In the other half, the partition was of opaque glass, cutting off the view.
O CLEAR GLASS 0 OPAQUE GLASS
t 2 TRIAL
Fig. 38. Goldfish react more rapidly if allowed to watch
others perform. (From Welty.)
Trained fish were placed in the aquarium-maze
and were run through their performance from ten
to twenty times in different experiments. The same
170 THE SOCIAL LIFE OF ANIMALS
treatment was given the fish in the aquaria with opaque partitions and those with clear glass. The trained fish were then removed and those from the small side chamber were gently transferred to the larger side. An hour later they were given an ordi- nary test such as had been given to the trained fish. As is clearly shown by the graphs in Figure 38, the fish which had been able to watch the others react behaved decidedly more like trained fish than those which had not been able to see their fellows perform.
As a final check, the whole test was repeated, ex- cept that no fish were placed in the larger side of the aquarium. Fifteen times each aquarium was lighted up, the door opened, and the experimenter stood ready to feed any imaginary fish that might come through. Then when those in the side passages were transferred, there was no essential difference in the behavior of the fish from the two types of aquaria, and the experimenter was free from any sug- gestion that he might have been signaling the fish.
The results of these experiments suggest that there is such a thing as imitation among goldfish. Whether there is or not depends, as Dr. Welty rightly says, largely upon the definition given to the word imita- tion. These fish probably do imitate each other on a relatively simple instinctive level. The untrained fish that watched the reaction of their trained fel-
GROUP BEHAVIOR 171
lows through the clear glass became conditioned in two ways which were not open to the fish behind an opaque glass. In the first place they saw the fish move forward on the reception of a given stimulus, pass through the gate, receive food, and give no evi- dence of an avoiding or "fright" reaction. This prob- ably gave what might be called a certain reassurance. Secondly, they showed group cohesion, and moved forward with the reacting fishes; at times they were even seen to move forward in advance of the fishes on the maze side of the aquarium.
When transferred to the aquarium-maze and given the releasing stimulus of an increase in light, accom- panied by the opening of the gate, both types of previous experience probably played a role in pro- ducing a faster reaction. Fish behind the opaque glass could have neither of these helpful experi- ences. When their narrow aquarium was flooded with light they ordinarily moved back to the far end and remained there. There was nothing to train them to overcome this normally negative reaction. So reviewed, it must be said that this behavior has some points of resemblance to what is called imita- tion in other animals.
There is also an element of imitation in the greater food consumption of grouped fishes. One fish sees another pursue, attack and consume a bit of
lyS THE SOCIAL LIFE OF ANIMALS
food and its own feeding mechanism is set off as a result of this visual experience, even though its own hunger might not have been sufficient to stimulate feeding behavior. It is difficult to say to what ex- tent such behavior is an expression of competition as contrasted with unconscious co-operation. The two types of motivation overlap here and elsewhere.
The evidence which we have been considering furthers our understanding of the fundamental na- ture of group activities among many animals, some of which are not usually regarded as being truly so- cial. The whole emphasis of this chapter has been laid upon facilitation as the result of greater num- bers being present. This kind of social facilitation has been described for such diverse processes as breed- ing behavior, eating, working and learning.
Added numbers do not always facilitate these ac- tivities, as was shown by the analyses of the effect of numbers upon the rate of learning. With some animals, for example men and goldfish, under cer- tain situations, learning is more rapid with several present; but with others, such as parrakeets and mud- minnows, under the conditions tested, increased num- bers lead to a lower rate of learning. It seems that no all-inclusive positive statement can as yet be made in this field. One can, however, make the affirmation that in the general realm here considered the pres-
GROUP BEHAVIOR 173
ence of additional numbers by no means always re- tards, and is frequently stimulating. As before with regard to other processes, we find that in certain cases there are ill effects of undercrowding as well as ill effects of overcrowding. Without careful ex- perimental exploration, we cannot predict which effect will emerge from a given situation.
One other result comes from these studies which will help us to clarify evidence still to be presented, as well as to review that already given. We have come upon another measure of the existence of so- cial behavior. Reactions may be regarded as social in nature to the extent that they differ from those that would be given if the animals w^ere alone. Such differences are frequently quantitative, as they have been in the cases we have discussed, although quali- tative differences occur as a result of a change in the numbers present.
From this point of view social behavior may have or may lack positive survival value. All that is nec- essary is that the behavior be different from that which would be given if the animal were solitary. In this sense all the animals whose behavior we have been discussing are social to a considerable degree; the more so, the greater the difference between their behavior when grouped and when isolated.
When the behavior of such animals as cockroaches,
174 THE SOCIAL LIFE OF ANIMALS
fishes, birds and rats shows evidence of distinct modi- fication as a result of more than one being present, we have another suggestion that there exists a broad substratum of partially social behavior. There are many indications that this extends through the whole animal kingdom. From such a substratum, given suitable conditions, societies emerge now and again as they have among ants and men. At these higher social levels, as is to be expected, the type of behavior shown under many conditions is related even more closely to the number of animals present than with less social cockroaches and fish.
VI.
Group Organization
WE ALL know that human society is more or less closely organized. Sometimes, as in military circles, some business organizations, and certain universities, there is a line organization which extends in a defi- nite order, step by step from the highest official to the lowest rank. Frequently, however, the organiza- tion is more complex, intricate and temporary.
We have known for some time, too, that in herds of the larger mammals, where one can distinguish different individuals, the group may be organized to some extent with a dominant leader and frequently with sub-leaders that stand out above the common run of the herd. (16)
Despite this knowledge we have found with sur- prise that other animal groups, a flock of birds for example, in which the different birds are indistin- guishable to the human eye, also are organized into a social hierarchy, frequently with a well-recognized social order which runs through the entire flock. The situation that has been revealed in these flocks
175
176 THE SOCIAL LIFE OF ANIMALS
of birds is amusing, interesting and important enough to warrant more attention than it is receiv- ing at present.
Studies of the sort I am going to describe were initiated by a Norwegian named Schjelderup- Ebbe. (108) They were made possible by the use of colored leg bands and other markings by which the different individuals could be recognized by a human observer. Apparently the birds themselves knew the individual members of the flock without such artificial aids.
Not because it is the most important work on the subject, but because I can best vouch for it in de- tail and in general, I shall present certain analyses of group organizations that have been made in our own laboratory.
The organization of flocks of chickens is fairly firmly fixed. This is particularly the case with hens. The social order is indicated by the giving and re- ceiving of pecks, or by reaction to threats of peck- ing; and hence the social hierarchy among birds is frequently referred to as the peck-order.
When two chickens meet for the first time there is either a fight or one gives way without fighting. If one of the two is immature while the other is fully developed, the older bird usually dominates. Thereafter when these two meet the one which has
GROUP ORGANIZATION 177
acquired the peck-right, that is, the right to peck another without being pecked in return, exercises it except in the event of a successful revolt which, with chickens, rarely occurs.
The intensity of pair contact-reactions varies greatly. A superior may peck a subordinate severely, or lightly, or it may only threaten to do so. It usually turns its head, points its bill toward the subordinate and takes a few steps in that direction. It may then give a low deep characteristic sound which fre- quently accompanies an actual peck, and stretch its neck up and out without the resulting peck which it seems just ready to administer.
The peck, when actually delivered, may be light, heavy, or slashing. These vigorous pecks may be painful even to man, as anyone can testify who has tried to take a setting hen off her nest; and particu- larly painful if repeated in the same spot. The peck- ing bird may draw blood from the comb or may pull feathers from the neck of the pecked fowl. The peck is frequently aimed at the comb or the top of the head; often it is not received with full force, for the pecked bird dodges. Less often the peck is di- rected toward back or shoulders.
The severity of a peck which lands as aimed is illustrated by a recent observation in one of our small flocks. One bird received a vicious peck di-
178
THE SOCIAL LIFE OF ANIMALS
rectly on the top of its head; it walked backward two or three feet, staggered and fell, arose and again walked backward in a blind course that took it into the bird that had given the original peck. By that
|
RW pecks |
all 12 |
A, BG, BB, M, Y, YY, BG^, |
GR, |
R, |
GY, RY, RR. |
|
RR pecks |
11 |
A, BG, BB, M, Y, YY, BG^, |
GR, |
R, |
GY, RY. |
|
RY pecks |
10 |
A, BG, BB, M, Y, YY, BG2, |
GR, |
R, |
GY. |
|
GY pecks |
9 |
A, BG, BB, M, Y, YY, BG^, |
GR, |
R, |
|
|
R pecks |
8 |
A, BG, BB, M, Y, YY, BG^, |
GR, |
||
|
GR pecks |
7 |
A, BG, BB, M, Y, YY, BG2. |
|||
|
BG2 pecks |
6 |
A, BG, BB, M, Y, YY. |
|||
|
YY pecks |
4 |
A, BG, BB, M. |
|||
|
M pecks |
4 |
A, BG, BB, Y. |
|||
|
|