242 PROBLEMS OF RELATIVE GROWTH 



is on the time of onset of a process, and not on its rate (Ford 

 and Huxley, op. cit.). 



It should be possible by comparing related forms to discover 

 whether time-relations of this second type are involved in 

 addition to those of the first. We must further remember 

 that growth-coefficients may change during ontogeny ; they 

 may change from strong to less strong positive heterogony, as 

 in Uca large chela, or from isogony to positive heterogony, 

 as in Maia large chela, etc. These facts must also have their 

 genetic basis. 



With an analysis such as this, we may hope for a fuller 

 understanding of the processes involved in changes of propor- 

 tion. One alteration in a single rate-gene may delay the first 

 formation of an organ and also decrease the growth-coefficient 

 once it is formed. Further, although the processes of histo- 

 differentiation do not seem to follow the same laws of relative 

 growth as those of auxano-differentiation, the quantitative 

 intensity of the two kinds of growth-processes may well be 

 controlled by the same genes. 



In considering evolutionary changes in relative size, we 

 must accordingly try to distinguish the various agencies which 

 may be at work. It appears that these may be (a) mutations 

 affecting the primary gradient of the early embryo, on which 

 the time-relations of antero-posterior differentiation depend ; 

 (b) mutations affecting specific rate-genes ; (c) mutations 

 affecting specific ' time-genes ' — genes controlling time of onset 

 and not rate of processes. The processes controlled by the 

 rate-genes and time-genes will be processes concerned with 

 growth-gradients, whether of a major or minor nature : they 

 will therefore always affect a number of parts in a correlated 

 way. 



Finally, it is at least possible, as we have seen in an earlier 

 section, that the primary ' axial gradient ' of the developing 

 egg and early embryo, on which, we must suppose, depend 

 the facts subsumed under the law of antero-posterior differ- 

 entiation, itself continues to operate later as a growth-influenc- 

 ing gradient, as well as influencing the time-relations of 

 differentiation. This would mean that mutations primarily 

 selected because of their effect upon early development would 

 have an effect upon proportional size in later life — an interest- 

 ing example, if substantiated, of what Darwin called correlated 

 variation. 



