We must now consider whether there is any direct way of impairing good heredity. It is currently believed that there are certain substances, popularly known as "racial-poisons," which are capable of affecting the germ-plasm adversely and permanently in spite of its isolation and protection. For example, the literature of alcoholism, and much of the literature of eugenics, abounds with statements to the effect that alcohol _originates_ degeneracy in the human race. The proof or disproof of this proposition must depend in the last analysis on direct observation and carefully controlled experiments. As the latter cannot be made feasibly on man, a number of students have taken up the problem by using small animals which are easily handled in laboratories. Many of these experiments are so imperfect in method that, when carefully examined, they are found to possess little or no value as evidence on the point here discussed. Hodge, Mairet and Combemale, for example, have published data which convinced them that the germ-plasm of dogs was injured by the administration of alcohol. The test was the quality of offspring directly produced by the intoxicated animals under experiment. But the number of dogs used was too small to be conclusive, and there was no "control": hence these experiments carry little weight. Ovize, Fere and Stockard have shown that the effect of alcohol on hen's eggs is to produce malformed embryos. This, however, is a case of influencing the development of the individual, rather than the germ-plasm. Evidence is abundant that individual development can be harmed by alcohol, but the experiments with eggs are not to the point of our present purpose. Carlo Todde and others have carried out similar experiments on cocks. The conclusions have in general been in favor of injury to the germ-plasm, but the experiments were inadequate in extent. Laitinen experimented on rabbits and guinea pigs, but he used small doses and secured only negative results. Several series of experiments with rats indicate that if the dosage is large enough, the offspring can be affected. Nice, using very small numbers of white mice, subjected them not only to alcohol, but to caffein, nicotin, and tobacco smoke. The fecundity of all these sets of mice was higher than that of the untreated ones used as control; all of them gained in weight; of 707 young, none was deformed, none stillborn, and there was only one abortion. The young of the alcoholized mice surpassed all others in growth. The dosage Nice employed was too small, however, to give his experiment great weight.