SCIENCE 



that the modified condition is recessive to G 

 proves nothing whatsoever in regard to the 

 nature of the transformation that has occurred 

 in the germ plasm. Assuming that G is not 

 actually lost, but modified into another kind 

 of substance g, the recessiveness of g may be 

 due to the fact that its activity is manifested 

 in a different way, relative slowness of its 

 metabolism, or to various other conceivable 

 causes. 



There is, I believe, no good reason for con- 

 sidering the recessiveness of a character as 

 due to the relative simplicity of its germinal 

 basis. Many variations of a minus character 

 are recessive, but there are numerous excep- 

 tions to this rule, as is illustrated by the domi- 

 nance of the hornless condition in cattle, the 

 short tail of Manx cats, and the lack of beards 

 in certain kinds of wheat. Suppose we have 

 two allelomorphic units (assuming for the 

 present that there are such things as germinal 

 units) A and A', one of which tends to pro- 

 duce a relatively simple development of a part 

 and the other a relatively complex development 

 of a particular part of the body. The one A 

 calls forth, say, simple horns, the other 

 branched horns. A and A' presumably differ 

 chemically, and the development of the part 

 in question depends not upon A or A' alone, 

 but upon how these agencies affect other parts 

 of the body during development. Will the 

 simpler substance or organic unit call forth 

 the simpler structure in the adult body? In- 

 asmuch as the development of any organ in- 

 volves activities in which a larger number of 

 elements are concerned it seems not at all im- 

 probable that the simpler substance or unit 

 might conspire to produce the more complex 

 organ. Now suppose that the forked horn 

 proved to be dominant over the simple horn. 

 What conclusions would we be entitled to draw 

 from this fact concerning the germinal basis 

 of these characters? Obviously none. 



Whether we interpret a variation as a gain 

 or a loss is in most cases a purely arbitrary 

 matter. In sugar corn there is a loss of starch 

 but there is a gain of sugar. Does sugar corn 

 therefore represent a plus or a minus raria- 

 tion? Consider the familiar cases of rose 

 comb and pea comb in poultry. Both of these 



variations are dominant over the primitive 

 condition of single comb. Yet both breeds 

 carry the basis for the production of single 

 comb in their germ plasm. It is commonly 

 assumed that both conditions represent single 

 comb plus something. We may suppose that 

 in a certain chromosome a change has taken 

 place which results in the development of rose 

 comb. This change, for all that we know, 

 may be due to the loss or impairment of a 

 portion of germ plasm, or it may be due to a 

 change not properly describable as either a gain 

 or loss. We may regard rose and pea comb as 

 more or less pathological deviations based on 

 germinal defect, as true progressive varia- 

 tions, or simply as normal variations neither 

 progressive nor retrogressive. So far as com- 

 plexity of structure is concerned it may be a 

 matter of dispute whether rose comb, pea 

 comb, or single comb represents the higher 

 grade of development. 



But, it may be asked, are not color varieties 

 commonly due to loss, and is not this obviously 

 the case with albinism? In many varieties 

 there has certainly been a loss of pigment, but 

 has there been a dropping out of factors? It 

 by no means follows. The factors represented 

 by small letters in our color formulas are by 

 no means missing entities. They are changed 

 so that they occasion a diminished production 

 of certain pigments, but in other respects they 

 may be as potent as before. The albino does 

 not produce pigment, but there may be other 

 substances in the place of pigment that would 

 distinguish the albino as a positive variation 

 when judged by other standards. The animals 

 whose gametic formulas contain a number of 

 small letters are not necessarily more imper- 

 fect or perhaps I should say incomplete than 

 their congeners which carry a large number 

 of dominant characters. 



Of course there may be varieties due to 

 losses of germinal material. Considering the 

 complex mechanism of mitosis, and the op- 

 portunities afforded for the loss of chromatin 

 during this process, such variations are not 

 improbable a priori. But there is not the 

 slightest warrant in the fact of recessiveness 

 per se for the doctrine that all recessive varia- 

 tions are produced by this method. The origin 



