Such a supposition clarifies the behavior of some of the insects I have mentioned, but it has flaws. An example is the cabbage butterfly, which commonly restricts its feeding to plants of the cabbage family although its caterpillars sometimes appear on other dissimilar and unrelated plants. One is the nasturtium. The explanation is that nasturtium leaves have the same pungent odor and taste due to an essential oil similar to that in cabbage. A Dutch entomologist, E. Verschaffelt, who studied the behavior of the butterflies and caterpillars with reference to mustard oils, concluded that the presence of those chemicals was the factor that determined their choice. Other entomologists, particularly Vincent G. Dethier, have extended such studies to a large number of other insects and their food plants, and have found that some specific chemical (or more than one in combination) commonly forms the tie that binds them to an invariably constant diet.

It is evident now that odors, recognizable to the acute and discriminative chemical sense of the insects, ark the main factors involved in the selection of food plants by oligophagous insects. But knowledge of the multitude of chemical substances elaborated by plants is still too fragmentary to permit any broad generalization. Some species of less fastidious tastes will accept considerable variety, but may evince a dislike for some chemical to which they are not accustomed, even when it is combined with one that is highly attractive when not thus contaminated. That is not the whole story, however.

We must approach some other considerations with caution, as their meaning is not yet clear. They relate more directly to instinctive behavior.

Among the aphids, or plant-lice, there occurs quite commonly an alternation of generations, whereby the aphids migrate during the course of the seasonal cycle from one kind of plant to another and then return to the original host plant after the period of winter dormancy. These aphids frequently have a small series of acceptable plants, and one or both of the alternate food plants may not include more than a single botanical species. Early in the summer, several generations of wingless females rapidly succeed one another on the summer food plant, followed by a generation of winged females, which then migrate to another species of plant known as the alternate food plant. There they give rise to a generation of aphids of both sexes, which produce eggs that over winter. The next spring these give rise to a brood of winged females that migrate back to the summer food plant, after which the cycle repeats itself.

With some minor variations, that is the fundamental pattern two very different food plants are selected alternately. Thus the migrating aphids are conditioned to two diverse plants at different times, and we cannot attribute their attraction to a single specific chemical stimulus. Other factors obviously enter the picture, but they cannot be singled out further than to note that the appearance of the winged migrants and sexual phases is correlated with the season. This is a phenomenon similar to photoperiodism in plants, where vegetative growth and flowering are often closely related to seasonal variations in the duration of the daylight period. We cannot state definitely whether chemical changes in the plants may elicit a differential response by the aphids as the season progresses.

More surprising still is the relationship that exists among some gall-making insects. Many diverse insects induce the formation of abnormal growths or of highly modified specific structures. They are known as galls and are developed by the plant under some stimulus from the insects that lay their eggs in the tissues of the plant and undergo their growth feeding within the developing gall. The nature of the stimulus is not yet understood, but it is highly specific, as the galls produced by each species are always alike in form and structure.

The gall insects are quite uniformly oligophagous or monophagous. One group, the gall wasps, are small, wasp-like insects in some hundreds of species. Nearly all of them produce galls on oaks. They are highly specific for particular kinds of oaks each lays its eggs in some restricted part of the tree, whether leaves, twigs, buds, or roots. Furthermore, some species of gall wasps undergo an alternation of generations, whereby one generation induces galls on some aerial part of the plant and the next goes underground to induce a root gall, returning in the succeeding generation again to the previous location above ground.

In the gall wasps, the restriction of choice to particular food plants is similar to the one I outlined among the aphids, but the matter is further complicated by the fact that the response of the plant is an essential requirement for the maintenance of the relationship.

So we see that an insect's selection of food plants depends primarily on an acute and discriminative chemical sense, which enables it to recognize by smell and taste many essential oils and other less pungent substances in particular plants. As the presence of each such chemical is usually confined to some natural group of plants, they are the ones to which the insects are attracted. When the same chemical attractant appears sporadically in unrelated plants, they also may be chosen. This basic conception is supported by observations of the behavior of insects in nature and by the application of some experimental techniques. It explains the puzzling "botanical sense." It has already opened up a promising field for the study of attractants and repellents that should have great practical value.

The insect's selective appetite, so far as we can see at present, is purely an instinct to do thus and so, whether or no. Such instincts are innate and unalterable attributes of all insect behavior which excite our wonderment and captivate our curiosity because we can go no further than to catalog their manifestations.

Some matters relating to the correlation that exists between the larval appetite and the consistent choice of acceptable food plants on which to lay the eggs are amenable to analytical treatment. Even in insects that shift to another type of food when they reach the adult reproductive stage, we may assign some form of memory or nostalgia to account for the return of the gravid butterfly to a cabbage head after a round of sipping the nectar of sweetly scented flowers. That may seem a bit farfetched, but it is obvious that once it has been incorporated into the sphere of instinct, an identical response will inevitably be called forth. At this stage, any transgressions will be quickly eliminated through the most rigorous action of natural selection whereby any butterfly that failed to select a proper food plant is unable to pass on to posterity her careless or vacillating tendencies. That such aberrations of instinct do occur, although rarely is attested by actual observations of insects under natural conditions. Any such mutation of instinct may conceivably persist if compatible with the appetite of the larva and capable of weathering the competitive pressure imposed by the living environment.

In yet another way may memory have a part in the differential choice of food plants by insects that normally accept a variety of host plants. Where we cannot detect chemical attractants in common, it appears that memory of the larval food might lead the adult to prefer it to other acceptable plants. There are many cases where such strains, races, or clones appear to be established in nature. Experimental proof so far has been inconclusive, but it seems probable that when varied food plants are readily acceptable such strains do exist. We might even compare them to those racial or geographic components of our human population that consistently evince a preferential fondness for cabbage, garlic, red pepper, baked beans, macaroni, curry, or some other item of food.

CHARLES T. BRUES is professor of entomology emeritus in Harvard University, where he taught and engaged in research on various phases of entomology for 37 years. He has devoted much time to studies on the food habits of insects, a subject on which he has published extensively. A graduate of the University of Texas, he served as field agent for the Bureau of Entomology and later as curator at the Milwaukee Public Museum, before joining the biological staff of Harvard in 1909.