the light of the most recently discovered principles of heredity. A very important conception in this connection is the continuity of the germ plasm, another is the variability of the potency of unit characters.
Admitting, then, that certain acquired characters have actually appeared in later generations, we should consider, first, whether or not the germ plasm has been changed by the stimulus which has produced the changes in the body. It has been shown that starvation in the larval stages of insects will produce dwarfs in later generations, but here it is assumed that the unfavorable conditions surrounding the germ plasm persist and that there is no real change in the composition of the germ plasm. Can Kammerer's results be explained in the same way? Of course a Lamarckian can not be asked to produce a form which will not revert. The only test that can be readily applied is that of Mendelian inheritance. It has been shown by the author that in one case at least the new factor behaved like a Mendelian factor. Tower also found this true in crossing a pale potato beetle, which he derived experimentally, with a beetle of the normal color. Such a test to discover a change in the composition of the germ plasm is certainly very significant.
Granted, then, that the germ plasm has been changed, we should next consider whether it has been changed directly or indirectly. The experiment of keeping tadpoles in water for an abnormally long time showed that in order to affect the next generation the stimulus must continue to act until the sex cells are mature. Tower also came to the same conclusion in his experiments on the potato beetle where heat was the stimulus. The changes, then, are probably due to the direct action of chemical and physical stimuli on the germ plasm contained in the ripe germ cells, exactly as MacDougal produces mutations, as he claims, by injecting chemicals into the ovary of a plant. But why should the stimuli not effect the germ plasm of the embryo as well, since, according to the theory of continuity, the same plasm is always present even in the youngest stages? It may possibly be claimed that, if any such effect is produced in the embryo, the change is repaired before reproduction takes place.
Granted, then, that the germ plasm in these cases is more or less directly affected by the environment, we should consider whether the change is more than a change of potency of a factor already present. According to Castle such potency may be increased by selection. Perhaps the new environment may increase in some way the potency of a factor which is present in a weak condition. For example, in the case of the spotted salamander, the potency of the factor represented by the yellow pigment may possibly be changed by the action of the yellow light, which actually increases the amount of the pigment in the body of the adult until perhaps the nature of the fluids of the body cavity