spawning could be expected. The buffalo fish acted just as they do in nature; they moved out into the shallow waters and spawned-doubtless the first buffalo fish to spawn in controlled ponds. Whether temperature or oxygen supply, or both, or something else is responsible for phenomena such as these, it is evident that the fish-culturist must look to the student of the physical conditions of enclosed waters for guidance in the construction and the control of ponds.

Just as the trees and the small plants and the grass are continually breaking up noxious gases in the air and replenishing the supply of oxygen, so in the water the submerged vegetation plays an important part in maintaining the oxygen supply for fish. In fact, they are probably the principal dependence for oxygen in ordinary fish ponds. In a very large lake where there are high waves and pronounced wind-driven currents, rolling movements, and upheavals, the vegetation plays a less part. The smaller the pond, however, the more essential are the submerged plants. Plants serve another useful purpose in taking care of the noxious carbon dioxide which is given off by animals in breathing and which is formed by the processes of decomposition.

In selecting plants for the pond for the purpose of oxygenation, it must be kept in mind that plants do not possess this function except in the presence of sunlight. The large lilypads which are so esthetically pleasing, but which, being at the surface, can contribute little to the oxygen supply of the water,

form a deep shade that must diminish the oxygen-producing capacity of other plants living in the water beneath. It is evident that submerged vegetation is wanted and preference may be given to those plants having an abundant growth of narrow leaves, or to those with foliage so finely divided as to be needle-formed or brush-like. Consideration must be given, too, to the species which remain green during the winter or which are the earliest to give rise to new growth in the spring, so that there may be the most effective production of oxygen at a time when it is so important to the breeding fishes, and when the surface absorption of oxygen is normally less adequate.

Here, then, is a problem which has scarcely been attacked. What species of plants are the most effective oxygenators, under different conditions and at different seasons? The experienced and observant fish-culturist has somewhat definite ideas, and his judgment in the matter is very valuable, but I think that very little has been done in the way of experimental determination of the questions just stated. We ought to know, as precisely as we can, the relative oxygenating values of the different species of aquatic plants-for wintering ponds, for spawning ponds, and for rearing ponds.

FOOD SUPPLY

The fish must have food and, under ordinary conditions of fish-culture, the food naturally produced in and about the pond is the principal dependence. Obviously, the productiveness of a pond in fish is directly limited by its productiveness in food;

hence, fish-culturists often say that the whole problem of fish culture is one of food supply. It may well be so, since this is not a single problem, but a complex of problems.

Biologically speaking, the food problem starts with the plants, as the source from which, or through which, all animal food must come. Plants form the basis of food supply-large plants or microscopic plants, green plants or dead plants, or the finely divided plant remains constituting the detritus. To what extent do plants, living or dead, enter directly into the food of fishes? I venture to say that we know yet very little of this. Only a few years ago, the forage value of plants was considered insignificant. Yet, very recently, an investigator associated with our Bureau, Dr. A. S. Pearse, of the University of Wisconsin, has prepared for publication a report of the food of 32 species of fish from lakes in Wisconsin, and, from one of his tables, it can be found that, with 23 species, plant remains or algae constituted an appreciable portion (one per cent. or more) of the stomach contents, ranging from 1 per cent. to 25 per cent. If we include silt and débris (probably plant material principally), 25 of the 32 species were plant feeders, and the ratio of such food to the total ranges from 1 to 40 per cent. Other uncompleted investigations of the bureau indicate that vegetable detritus constitutes a substantial, or perhaps the principal, element of diet for fresh-water mussels and for the young buffalo fishes. This is certainly true for many insect larvæ, and other small animals.

Undoubtedly, the direct food value of vegetation to pond fishes, especially to the young, is not inconsiderable; but even more significant is the part which this form of food plays in an indirect way. Generally speaking, as the fishes become older and larger (this is not true of all species, of course), they seek larger and more active prey, entomostraca are passed over for small insect larvæ, amphipods, small snails, etc., these in turn give place to larger insect larvæ, crawfishes and small fishes, and finally, larger fishes and frogs may become the special prey of the "big-game hunters" among the fish. But all the multitudinous members of this complex community of hunters and hunted derive their origin from plant matter. Now, one phase of this general problem of the relation of plants to food supply to which it is desired to direct attention is this: We have very little information as to the relative food values of the different species of plants. Undoubtedly, some species of plants are better forage plants than others. Dr. Emmaline

Moore, of Vassar College, and quite recently a special investigator for the Bureau, has already given us some valuable information about this, and I may be permitted to emphasize the point that, as her investigations show, plants of one species may be foraged upon, while those of a closely related species are left untouched. Presumably, too, some plants, when dead and disintegrated, give rise to a more palatable or nutritious detritus than others; of this we know little, if anything.

These questions of the relative values of plants, viewed either as oxygenators or as food makers, are not of theoretical or scientific interest only. This can be made clearer from an analogy. A stock farmer may have no interest whatever in plants as plants. Nevertheless, he sows alfalfa under certain conditions and burr-clover under others; he knows when and where he wants to plant red clover, and he knows that he never wants to plant sweet clover. All of these legumes are fairly closely related, yet the grower of stock has learned to discriminate between them, to use each to best advantage, or to let them alone, as his purpose may require. The grower of fishes, on the other hand, lets grow what will, practically speaking— and who can now advise him intelligently?

Our problem does not stop with the plants-it only begins there, biologically speaking. Small crustacea, insect larvæ, and molluscs feed upon plants or plant remains, and then upon each other. The problem becomes complex and peculiarly ecological, but its solution may be approached very directly. Here is an insect larva which feeds upon certain things and is preyed upon by certain other forms: it attacks and destroys small fishes and is itself devoured by larger fishes; it feeds upon materials which the fish that we wish to foster can not directly consume, thus adding material to the fish's food supply, while it competes with the fish for other forms and so diminishes the food supply; it destroys certain enemies of fishes, but who knows if it harbors some injurious parasite of fishes? The significance of this larva, and it is not altogether an imaginary one, is evidently not to be appraised as the result of casual observation. A great deal of data must be accumulated, the points of contact searched out in various directions, the evidence carefully analyzed, checked by experiment if possible, and weighed with sound judgment before a just conclusion is reached. Common sense will make the final ruling, but it will be common sense seated upon a secure bench of scientific observation and experiment. It would be an excellent thing for

fish culture if one after another of the typical inhabitants of a pond could be taken up for systematic study along such lines as have been suggested.

Since this paper must be kept within reasonable limits and as the ecological rather than the biochemic aspects of fish-cultural problems are primarily in mind, the important subject of the artificial feeding of fishes in ponds must be passed over at this time. More nearly ecological is the question of the fertilization of ponds-the adding to the water of organic or inorganic substances, so as to promote an abundant growth of desirable aquatic organisms, without impairment of the conditions of existence for fish. In this connection, I will merely hint at two very important subjects; that is, the character and composition of the bottom soils and the chemical composition of the water itself. We know that plants and animals have definite chemical requirements, and the requisite substances must come, directly or indirectly, from the soil or from the water. We strongly suspect, at least, that certain chemicals have subtle but significant physiological effects, favorable or unfavorable, upon the growth of aquatic plants and animalseffects that can be discovered not so readily by inference from analysis, as by experimental determination.

So far, we have kept strictly within the confines of the pond itself, but the ecological problems of fish culture extend well beyond the reach even of the highest waves that wash the margins. The sloping banks, the green sward, the meadows beyond, do not these contribute to the food supply of the pond? No one can be doubtful of this after walking around a pond, and noting the small frogs that leap from the banks to be snapped in by a hungry bass, or observing the grasshoppers and crickets resting on the lotus leaves or in the stems of Persicaria or of cattails, or watching the dragonflies and mosquitoes and dozens of other insects that pass from bush or grass to pond and back again (if luck is with them). Read the reports of stomach examinations by Forbes and others, and note the extent to which non-aquatic insects and other animals enter into the food of fishes. Mr. H. W. Clark, of the Bureau of Fisheries, tells of trout feeding upon masses of woolly plant lice as fast as they fell from overhanging alders. Professor C. B. Wilson, while working at the Fairport Laboratory, finds a certain dragonfly that, like others, through its larvæ supplies food to fish, but that almost invariably completed its metamorphosis on a hillside slightly removed from the pond, although in order