THE MUTATED GENE 245 



The difference is actually found in the velocity of hardening of 

 the chitin. As pigment is deposited in the chitin in colloidal 

 solution, a hardened scale cannot be pigmented any more, and 

 therefore the speed of hardening, different in different areas, 

 leads to the pattern of pigmentation. Recent work in the 

 author's laboratory has confirmed these basic facts, which there- 

 fore cannot be overlooked any longer. 



These two groups of facts together show that a primary pattern 

 has been formed early in development for the whole wing as a 

 unit. The pattern consists in the creation of zones of differential 

 velocity of differentiation, which later will lead in some way to 

 different pigmentation and sometimes shape and structure of the 

 scales. This unit pattern is influenced by mutant genes. (Sexual 

 difference might be included in this statement, because genetically 

 the difference is of the general type of a mutational difference.) 

 Can we now conceive of the gene-controlled process that leads 

 to the formation of such a complex pattern? In the case of the 

 intersexual wing, we have found the answer: here the pattern of 

 areas of different speed of differentiation was conditioned by the 

 determination stream spreading from the wing base across the 

 wing. There is no reason to suppose that the process is a 

 different one (in principle) in the other cases. What has to be 

 explained, obviously, is why the production and spreading of 

 such a substance leads to induction of definite growth velocities 

 only at definite points or, using an expression previously intro- 

 duced, why this induction becomes stratified. 



b. Pattern Localization. — In Goldschmidt's first publication 

 on the subject (19206), a rather generalized formulation was 

 applied to this problem. The special arrangement of the parts 

 of the pattern was referred to as the arrangement of the outlets of 

 pigment, which now has to be defined as the arrangement of 

 points from which the determination stream starts. It was 

 known from the genetic analysis of different patterns that a 

 number of genes may be concerned in controlling this arrange- 

 ment. Forthwith an attempt was made (Goldschmidt 1920d) to 

 change experimentally the processes involved by injuring the 

 young wing in different ways, setting defects, etc. But only the 

 very beginnings of success were attained. As is now known from 

 Henke's work (1933a) upon the same material and using partly 

 the same methods, the lack of complete success was due to 



