228 PROCEEDINGS OF THE AMERICAN ACADEMY. 



rational explanation of the outcome. It thus meets the two-fold require- 

 ments of a scientific theory, a statement of phenomena and an explana- 

 tion of them ; the " law of ancestral heredity " attempts only the first of 

 these two things, and even here fails lamentably. It will thus be seen 

 that the claims of Mendel's law are much greater than those of Galton's 

 law. If it fails, its failure is as much more signal. 



The same test may be applied to Mendel's law as to Galton's. Can 

 we, on the basis of Mendel's law, make predictions concerning the vari- 

 ous generations of von Guaita's mice with greater accuracy than has 

 been found possible under Galton's law ? Before we can frame an an- 

 swer to this question, we must know precisely what the Mendelian pre- 

 dictions are. 



Mendelian predictions are based, not on the somatic character of the 

 parents^ but on the character of the germ-cells formed by the parents. 

 The simplest way of determining the character of the germ-cells formed 

 by an animal or plant is by experimental breeding tests. In cases where 

 this is not practicable, one can often predict with equal confidence from 

 a knowledge concerning the grandparents, not as to their somatic charac- 

 ter, but as to the character of their germ-cells as evidenced by the nature 

 of the offspring produced by them. Stated in the terminology of present- 

 day biology, the principles which underlie the Mendelian predictions are 

 these : — 



1. Every gamete (egg or spermatozoon) bears the determinants of a 

 complete set of somatic characters of the species. Accordingly when 

 two gametes (an egg and a spermatozoon) have met in fertilization, there 

 are present in the fertilized egg the representatives of two sets of somatic 

 characters, which may or may not be the same. If they are the same for 

 a given character, as, for example, coat-color in mammals, the individual 

 which develops from the egg must inevitably have that same character. 

 Thus when gametes formed by one white mouse meet in fertilization 

 gametes formed by another white mouse, the offspring are invariably 

 white. Similarly when a wild gray mouse is bred to another wild gray 

 mouse the offspring are invariably gray. And when a pure-bred spotted 

 black-white mouse is bred to a mouse like itself, the ofl^spring are all 

 spotted black-white. 



2. But when the two gametes uniting bear each what represents a 

 different somatic character, only one of these characters may be mani- 

 fested by the individual (or zygote) formed. Thus, when wild gray 

 mice are mated with white mice, only gray offspring are produced. The 

 gray character is, in Mendel's terminology, dominant, the white character 



