THE MUTATED GENE 89 



This discussion touches the decisive points (if we leave out of 

 account here the reflections upon the nature of the genes involved). 

 (1) Facts of this type are expected if different processes are 

 involved. (2) If one process is involved, any special feature of a 

 developmental process that is in some way connected with the 

 main process will result in such phenomena as are discussed here. 

 Such developmental variation may depend on the control of a 

 threshold, but it may also be conceived in many different ways. 

 Most cases of this type have been reviewed in Stern's Monograph. 

 If we consider them from the standpoint of Wright's discussion, 

 there will be no difficulty in reducing the discrepancies of seriation 

 to such interferences as different thresholds, all-or-none effects 

 for certain processes, or changes of general situations of develop- 

 ment in an orderly way, such changes interfering only with the 

 differentiation of one organ (see the discussion in Goldschmidt, 

 1927c). Probably most frequent are those cases in which the 

 time relations of development (order of growth, etc.) are different 

 in different organs and the alleles affect some process at a definite 

 time. Let us assume A, B, C, D to be successive points in 

 development at which the alleles m l , m 2 , m 3 , m 4 act. One action 

 determines the quantity of pigment in a linear series proportional 

 to the time of onset (A, B, C, D) of the reaction controlled by the 

 alleles. Another consequence of these reactions would be an 

 influence upon the growth of another organ, say speeding up 

 growth. If, now, the normal growth of this part should occur 

 at the times A, B, C, D in the following seriation: grow thin 

 breadth/growth in length, growth in length, growth in breadth, 

 then the alleles m 1 , m 2 , m 3 , m 4 would produce the following seria- 

 tion of the length-breadth index in favor of length: 1- H , i.e., 



a different seriation from the pigment series. This model, which 

 might be varied in many ways (see Goldschmidt, 1932a, 1933c), 

 allows for all cases of not parallel pleiotropic action of multiple 

 alleles, without changing the general interpretation that such 

 genie series control different rates of one primary reaction. 



Whether there are cases of multiple allelomorphs that act in 

 such an irregular way that no simple explanation is forthcoming 

 is another question. Many such cases are known, and they are 

 reviewed in Stern's monograph. Some of those may be brought 

 in line when the actual facts of development are known. Others 

 may turn out to belong to a different category of facts; and, 



