226 
GENETIC STUDIES ON THE SILKWORM 
blank space of the "eye-spots" was filled up (Text-fig. 8). All these traits 
have been proved to persist in the offspring. 
The facts above given are, perhaps, enough to show that one or both of 
the complementary factors CP,ft) of the normal marking may be subject to 
minor changes which are inherited to the offspring. I am consequently 
forced to conclude that the so-called unit characters are not unit in the strict 
sense, but are of compound nature somewhat comparable to chemical 
compounds. There are one or more genes which are indispensable for the 
development of the essential features of a given character; they correspond, I 
venture to assume, to the element or elements forming the nucleus of a chemical 
compound. This genetic nucleus combines with a number of determinants or 
groups of them which correspond to the chemical radicles, and forms a com- 
pound. The cluster of genes thus combined may be inherited as such and 
appears as a "unit character", but it may also undergo slight changes caused 
by the dropping off of one or more than one genes or radicles or their replace- 
ment by another, or rearrangement of them in consequence of some physi- 
ological disturbance, say, in chromosomes. Any variations must be genetic 
in nature, as far as they are transmissible, yet they do not alter, the essential 
feature of the character, unless the variations affect the factor-nucleus itself. 
The forms produced by minor genetic variations may be compared to chemical 
isomers and derivatives. 0 
3) Multiple Factors Concerning the Moulting Character. 
Silkworms usually pass through four moults before spinning cocoons, i. e. 
they are 4-moulters. There are, however, some races which spin, as a rule, 
already after the third moult and are, therefore, 3-moulters. These 3-moulters 
are not distinguishable from the 4-moulters until the second moult which is 
gone through by the latter earlier. In the 3-moulters, the third and fourth 
stages last, moreover, a few days longer than in the 4-moulters which have, 
however, to moult once more, and commence to spin ("mount") 2 to 5 days 
1 . Whether the present cases are comparable with those enumerated by Castle (1912) as examples 
of the inconstancy of unit characters, is still doubtful. 
