Düsing maintains that several of these factors may play a part in determining the sex of the embryo, and if this be true, the problem becomes a very complex one. He also suggests that there are self-regulative influences of such a kind that, when one sex becomes less numerous, the conditions imposed in consequence on the other sex are such as to bring the number back to the normal condition; but this idea is far from being established. The fact that in some species there are generally more individuals of one sex than of the other shows that this balance is not equally adjusted in such forms. Of far greater value than these speculations as to the origin of sex are the experiments that appear to show that nutrition is an important factor in determining sex. Some of the earlier experiments in this direction are those of Born and of Yung. By feeding one set of tadpoles with beef, Yung found the percentage of females that developed to be greatly increased, and a similar increase was observed when the tadpoles were fed on the flesh of fish. An even greater effect was produced by using the flesh of frogs, the percentage rising to 92 females in every hundred. These results have been given a different interpretation by Pflüger and by others, and, as will be pointed out later, there is a possible source of error that may invalidate them. Somewhat similar results have been obtained by Nussbaum for one of the rotifers. He found that if the rotifer is abundantly fed in early life, it produces _female eggs_, that is, larger eggs that become females; while if sparingly fed, it produces only small eggs, from which males develop. It has been claimed also in mammals, and even in man, that sex is to some extent determined by the nourishment of the individual. Some experiments made by Mrs. Treat with caterpillars seemed to show that if the caterpillars were well nourished more female moths were produced, and if starved before pupation more males emerged. But Riley has pointed out that since the larger female caterpillars require more food they will starve sooner than the males, and, in consequence, it may appear that proportionately more male butterflies are born when the caterpillars are subjected to a starvation diet. This point of view is important in putting us on our guard against hastily supposing that food may directly determine sex. Unless the entire number of individuals present at the beginning of the experiment is taken into account, the results may be misleading, because the conditions may be more fatal to one sex than to the other.