388 Action of the Genetic Material 



result. All intermediates between these two extremes are imaginable, 

 and some were described by Csik. In Waddington's experiments these 

 types occur, but also some more complicated effects which cannot be 

 described simply as "compromise." Thus the effects of loci acting in the 

 same way are, if combined, more than additive. This means that a 

 development which is already abnormal is easier to alter. I think that 

 this is the same phenomenon found in multiple factor inheritance 

 (e.g., in my work on tetraltera, 1953Z?), in which the effect sometimes 

 is not additive but multiplicative. In terms of genie action this may 

 have different reasons which might lie entirely in the type of reaction 

 involved. For exaggerations of the combination effect, Waddington 

 proposes this explanation. Suppose that the sequence of develop- 

 mental processes involves two steps P and Q; further, that in P a set 

 of reactants k, 1, m, n are involved; and in Q, p q r s. One at least 

 of p q r s must have been determined by reaction P, because we 

 started with the idea that P and Q are involved in the same set of 

 reactions. This common step might be p. Then we might distinguish 

 loci which affect reactants belonging to the same reaction (q and r) 

 and those which affect reactants belonging to a different reaction 

 (e.g., k and r). If the reactions P and Q did not involve any tendencies 

 toward the attainment of equilibrium states, the two types of loci 

 would not differ significantly: one affecting k would produce a change 

 in p and, in the presence of the other mutant affecting r, the result 

 would be the same as that of a single mutant affecting q and r. But 

 developmental processes tend to some sort of equilibrium. Thus, in a 

 homozygote for a mutant locus affecting k, the interaction between 

 the other components of the reaction P proceeds in such a way that 

 the result is little altered. Such a mutant might therefore produce a 

 less disturbing effect on the process Q than would be caused by a 

 similar locus acting directly on that process. Therefore, more exag- 

 geration is expected in combinations of mutants acting on the same 

 process at the same time than in a combination of successively acting 

 mutants. If the same things are considered as processes proceeding in 

 time like monomolecular reactions, the differential equations obtained 

 suggest that mutant loci produce changes in the rate constants, while 

 differentiation depends on the composition of the system. If both 

 mutants affect the system, each will have its full effect in altering this 

 composition. Quoting from Waddington: "But if one mutation affects 

 the system and produces a change in the quantity of as [a substance 

 produced by an irreversible reaction which is transported out of the 

 system at a known rate], which passes on to become involved in a 



