420 THE AMERICAN NATURALIST [Vol. XLIX 



ceived as a particular chemical body at a given level in 

 the chromosome then any change in this body would be 

 expected to affect one, or more, or even, at times, all 

 characters of the complex that gives rise to the body 

 character or characters. The particular change might 

 involve no more than a rearrangement of the materials 

 of the locus or the addition of a chemical element (or com- 

 pound) or the loss of a chemical element (or compound) 

 — any one of these changes might lead to the loss of a 

 character in the soma. As to what happens in the locus 

 we can form no idea, and so far as the mechanism of 

 heredity is concerned it is a matter of no immediate im- 

 portance. If, however, any one finds a greater satisfac- 

 tion in the view that a loss of something from the locus 

 (an atom or a molecule) leads to a recessive character, 

 there is not the slightest objection to his doing so, unless 

 by loss he means the loss of the entire locus. He may do 

 this if he rejects the linear arrangement of different 

 material in the chromosomes, but if he accepts the latter 

 view the assumption of a literal absence involves him in 

 unnecessary difficulties. It is not as generally under- 

 stood as it should be that the facts which the presence 

 and absence theory was constructed to account for do 

 not require the assumption that the absence of a char- 

 acter means the absence of a factor in the germ-plasm. 

 It is entirely gratuitous to involve the theory of Mende- 

 lian heredity in such an interpretation which adds noth- 

 ing to the theory and by bringing in a new hypothesis may 

 involve the Mendelian theory in further difficulties. An 

 example may make this clear. It is known that when a 

 chocolate mouse is bred to gray and the F x grays that 

 result are bred together there appear in F 2 grays (9), 

 cinnamons (3), blacks (3) and chocolate (1). Gray was 

 written GBCh and chocolate gbC, which gave in F 2 GBCh 

 (9) GbCh (3) gBCh (3) and gbCh (1). The occurrence 

 of the black class of gBCh is accounted for through re- 

 combination. But the same end is accomplished if we 



