THE STRUCTURE OF ANIMAL COMMUNITIES An enormous amount of work has been done upon the food habits of animals, but scarcely half-a-dozen pieces of research upon the food-relations of a whole community. The complete food cycle for a community has only been worked out in one or two cases (very simple ones like those of arctic animals, or the community of animals living on one plant) ; but as a result of the work which has been done so far, several generalizations can be made. First, it is found that animals are usually arranged in "food-chains" leading from herbivorous or scavenging species up to terminal animals, through several stages formed by carnivores. For instance, in the sea, flagellated protozoa of various species are eaten by copepods, the latter by pteropod mollusca, the pteropods by herrings, and finally the herrings are preyed on by sea birds such as gannets, or by human beings. Again, in a pine wood, the young pine trees support a community of animals which are arranged in various f ood-chains, all radiating from herbivorous insects such as aphids and moth caterpillars; aphids are eaten by hover flies (syrphids), which are in turn eaten by spiders, and the spiders are preyed on by wasps, which use them as a food store for their larvae. In this community, there are also insectivorous birds, which are eaten by hawks. The part of this food cycle which refers to insects is shown in fig. i, and it illustrates very well a typical animal community, containing a number of different food chains, all radiating from plants through herbivorous animals and several carnivores. This form of organi zation or structure of animal communities is universal, and it will be evident that the study of animals, as limiting factors to the distribution or numbers of a species, is an exceedingly difficult one. No animal can be said to be dependent upon only one other species, since all are bound up together by food connections into one complex organization. The result of this state of affairs is that any upset or interference with the existence or numbers of species may have, and usually does have, very unexpected effects upon other species which one would not at first sight expect to be affected at all. For instance, it is said that the increase of the manufacture of bicycles affected the walrus-hunting industry in arctic regions, and that any shortage of walrus leather (which is used in polishing metals) is in turn followed by an increase in the amount of tuberculosis in certain French factories, which are then compelled to use felt instead of walrus leather (the felt pro ducing a noxious kind of dust). Or there is the case of two species of blue butterfly (Agriades coridon and A. thetis) which formed a local colony on one place upon the Downs, but were wiped out through a temporary over-increase in the numbers of rabbits, which ate up completely all the plants of the horse-shoe vetch (Hippocrepis comosa) upon which the butterfly cater pillars depended for food. Instances of this sort could be multi plied.
It is clear from what has been said, that in order to under stand the way in which any animal is affected in its numbers or distribution by the other animals living with it, it is necessary to study the whole animal community living in one habitat, and that it is useless to treat the animal as if it were completely isolated and acting as a separate unit. A general knowledge of the complete food cycle of an animal community is accordingly of great value to anyone who is investigating the biology of a particular species belonging to that community, since, armed with such knowledge, he is enabled to see at a glance a number of the points at which that animal is connected with other ones amongst which it lives, and this in turn throws light upon the manner in which its numbers are kept down and controlled. Supposing one is studying some insect which is becoming a dangerous pest among crops or in timber, a knowledge of its food relationships, gained from previous work upon the food cycle of the community to which it belongs, makes it possible at once to try various experi ments with counterpests (enemies or parasites) without it being necessary to wait for several years in order to work out these food relations—by which time the damage might have become colossal or possibly even uncontrollable. It seems clear, then, that the study of food relations of animals should form the next im portant step in any biological survey, after the first lists of species from different habitats have been made. It is at this point that the ways of animal and plant ecology part quite definitely : plants are living in communities where competition is usually for the same kind of food, whereas animals feed in in finitely more diverse ways than plants, and consequently have developed the complex communities which we find everywhere, and which require special methods for their study.
At present, only in a few cases have surveys of animal corn. munities been carried to this further stage. The preliminary list ing is in itself a long and formidable task, and furthermore, the systematics of many groups of animals have only been properly worked out within the last decade. Or, again, they have not been reduced to a simple and convenient system of reference which can be used by the ecologist, who is usually busy on other things. For this reason a number of the best ecological surveys up to date have been done by people who have paid attention mainly to one group of animals, in which they themselves are expert at identifying the different species (e.g., birds or beetles or molluscs) ; but it is obvious that such surveys cannot, by their nature, apply the food cycle method of study at all widely, since any one group of animals usually has rather similar food habits (e.g., butter Hies). This kind of survey, although of very real value in many other ways, is yet of limited application to fundamental research upon the numbers of animals. It seems likely, therefore, that future ecological surveys, which attempt to carry their work beyond the preliminary phase which has now been reached in many cases, will either have to employ teams of experts in differ ent groups (either in the field or by sending the material to be examined afterwards), or else will have to concentrate upon very simple communities, which contain few species and which have the minimum of complications in other respects. It is probable that the latter type of work will yield the most useful results, as far as the discovery of general principles is concerned.
Looking over the literature on the subject, one notes that food cycles have only been worked out with anything approaching completeness in a handful of cases : the land animal communities of arctic and high arctic regions (Elton) ; the animals of Calluna heath, pine woods, and associated habitats in the south of England (Richards), some of the land communities of temperate North America (Shelford) ; and the plankton community of the North Sea (Hardy). Otherwise, attention has been confined t0 single food-chains, or to the chains radiating from one species, followed out to a certain distance. (Of course, these, if followed farther, would give the food-cycle for the whole community, and this, as we have explained, is the whole point of doing such work.) There is, accordingly, a rather big blank space in our organized knowl edge of the food relations of wild animals, and even if we examine the few cases in which the f ood-cycle has been made out at all, it is found that the food habits of the different animals are only slightly known and indicated in a very rough way, that quanti tative data are usually lacking, and that all consideration of the parasites has had to be omitted. The value of this work, however, lies in certain ideas which have resulted from it, and which do re veal to some extent the way in which an animal community is constructed and how it works.
The first of these ideas is concerned with the size of animals and of their food. Every carnivorous animal is limited to a certain definite range of size in the matter of food, since on the one hand it cannot seize and overpower any animal above a cer tain size, while on the other hand it cannot find a sufficient num ber of animals below a certain size, to be worth while or provide it with enough food in the time available. Small spiders catch springtails, willow wrens eat caterpillars, the red-backed shrike attacks bumble bees, the merlin preys on the meadow pipit, and the eagle lives on ptarmigan—in all cases the size as well as the other qualities of the food comes into question. This rule applies also in many cases to herbivorous animals, but is less widely in operation, owing to the fact that plants are to a large extent de fenceless. However, the lower limit of size often plays a part in determining the food of seed or fruit-eating birds. In all cases the abundance of food is also an important consideration, since a small animal, if very numerous, may be obtained in sufficiently large numbers to attract animals which usually feed on things of a greater size. The significance of this idea about the size of food is that it explains to a large extent why food-chains exist at all in animal communities. Each stage in the chain has the effect of turning small food animals into larger ones, and since there are wider limits to the sizes of animals than to the size of the food of any one animal, we usually find some three or four stages forming the food chain. Sometimes there are more, often fewer, but there seem hardly ever to be more than six or seven stages in a chain leading from a vegetarian animal to the terminal carnivorous species. This leads us to ask why there should be a limit to the number of stages in such a chain. The answer is partly that animals themselves have limits of size; but there is another important reason, which holds good for most animal communities. This depends upon the relative numbers of the individuals forming the various stages in a chain. For instance, suppose we take the chain of freshwater animals : copepod-insect fish-bird. It will be found that the number of individuals in each species decreases very rapidly as we pass from the copepod to the bird. There might be, in a single lake, millions of copepods, many thousands of insects, several hundred trout, and half-a dozen grebe. Another way of putting it is to say that one grebe requires so many fish to supply its food requirements, each fish needs so many insects, and each insect so many copepods. At the same time there must be a certain number of each animal produced to perpetuate its own species in the face of old age and of checks other than being eaten. Thus each species is supporting, in ad dition to itself, a very great burden in the form of enemies. The copepod is supporting all the rest, and therefore has to be the most abundant of them all; but this it is, in fact, enabled to be on account of its smaller size, and consequent ability to grow up and breed more quickly. It is clear that as each stage in this chain is reached a smaller margin of living matter is left over to support any further species. The amount of transmuted copepod gradually diminishes, like a suit of old clothes which passes from one person to another, losing its substance by degrees until there is nothing left but a single patch. It can be seen that the size limits to the food of animals produce food-chains, and that these chains are able to exist because the smaller animals, which have the greater burden to bear, are by their smaller size enabled to support it the more effectively. We have not at present the neces sary data for solving in a quantitative manner these fascinating problems about the balance between size, rate of increase, and density of numbers; but presumably animal ecology will some day be in a position to present the phenomena of animal inter-relations in almost as accurate a form as those of chemistry. A beginning along these lines has been made by Chapman, who has worked upon populations of weevils kept at standard temperatures, and with controlled conditions of food and breeding rates.
We have dealt with food chains, the food cycle, and with the idea which may be conveniently termed the "pyramid of num bers" in animal communities. Another useful conception is that of niches. We can divide up animals into different classes accord ing to their food habits, e.g., herbivores, carnivores, etc. But it is possible to go farther than this, and divide the herbivores into those of different sizes, or belonging to different plants, e.g., aphids, grasshoppers, mice, rabbits, grouse, sheep, cows, etc., or, on the other hand, aphids of grass, of trees, and so on. Then we can classify carnivorous animals according to the size and nature of their prey. When we begin to classify the food-habits of animals in some such manner it becomes gradually clear that there are in each animal community a great many different occupations, jobs, or niches (just as there are in a human community), and that each niche is filled by a particular species. If we now turn to another animal community, occupying a different habitat, we find that although many of the same niches occur there, the actual species filling them are different in most cases. For instance, there is one mouse (Microtus) in grassland, which is preyed upon by the kestrel ; in woods there is another kind of mouse (Apodemus) which is preyed on by owls and weasels; while in houses there is a third kind of mouse (Mus) which is preyed on by the domestic cat and by human beings. In this example, the same niche (small mouse) is filled by three different species, whose habits and numbers, and size are to a large extent similar. The term "niche" is meant to be an elastic one, and it is hardly worth while attempting to define it exactly; but in spite of this, the idea is a useful one, since it emphasizes the fundamental similarity of all animal communities—a fact which is of some importance, since it is probable that any general principles which may be found out about the workings of animal inter-relations in any one com munity will be found to have a wide application elsewhere. It would therefore seem likely that intensive work, carried out corn pletely on very simple communities such as those of arctic regions or deserts, would afford the strongest chance of discovering the fundamental laws governing the inter-relations of animals and therefore the regulation of their numbers. The fifth idea of im portance in the study of animal communities is concerned with the significance and place of parasites in the general scheme. It has been customary in the past to treat parasites as if they were really different from carnivorous animals. In their morphology and in the fact that they get free transport on their host, they un doubtedly differ in many ways profoundly from the ordinary free living animals. But the resemblances are really much more than these differences ; parasites feed in essentially the same manner as carnivores, and a complete graded series can be traced from typical parasites which are completely dependent upon their host for food, lodging and transport, leading to species which lead an entirely free-living, carnivorous existence. Fleas are a good case of the transitional type of animal, while mosquitos and lampreys and jackals are other examples. From an ecological point of view, the chief significant difference between these two classes of pred atory animal is that a carnivore usually destroys its prey at once, whereas a parasite usually confines itself to abstracting little bits of its host, which are often scarcely missed by their owner. The principle followed is the same as that employed by a manufacturer who sells to an enormous number of people a very small article (like matches or pins) and by making a very small profit on each, amasses a huge fortune without harming his clients to any great extent. Another way of putting it is that a carnivore employs the method of a burglar, the parasite those of a blackmailer. In mak ing out a complete scheme of the food relations in a community it is clearly necessary to include parasites as well as carnivores; but, in practice, it is often possible to manage without this, owing to the small absolute bulk of the parasites, and the fact that they are often eaten at the same time as their host, and so in a sense can be considered to be part of it. However, ecological work on numbers, especially medical work, has shown the pro found inter-dependence of the two classes of animal, carnivore and parasite, especially where they carry any smaller parasite (bac terial or protozoan) which is capable of generating epidemics in the host species.