Spencer explains these transitional forms, and I agree with him, by supposing that the stoppage in food supply has taken place at different times after development has begun. ('It must happen that the stoppage of feeding will be indefinite.') Thus, the existence of transitional forms presents no difficulty on the theory of the agency of food. But how can the doctrine of determinants be applied to it? 'If he is consistent' (says Spencer, _Contemporary Review_, lxiv., p. 901), 'he must say that each of these intermediate forms of workers must have its special set of "determinants," causing its special set of modifications of organs; for he cannot assume that while perfect females and the extreme types of workers have their different sets of determinants, the intermediate types of workers have not. Hence we are introduced to the strange conclusion that, besides the markedly distinguished sets of determinants, there must be, to produce these intermediate forms, many other sets slightly distinguished from one another--a score or more kinds of germplasm, in addition to the four chief kinds. Next comes an introduction to the still stranger conclusion, that these numerous kinds of germplasm producing these numerous intermediate forms are not simply needless, but injurious--produce forms not well fitted for either of the functions discharged by the extreme forms, the implication being that natural selection has originated these disadvantageous forms. If, to escape from this necessity for suicide, Professor Weismann accepts the inference that the differences among these numerous intermediate forms are caused by arrested feeding of the larvæ at different stages, then he is bound to admit that the differences between the extreme forms, and between these and perfect females, are similarly caused. But if he does this, what becomes of his hypothesis that the several castes are constitutionally distinct, and result from the operation of natural selection?' My course of thought leaves me with little to add to this criticism by Spencer. In this case, as in many others that I have pointed out, Weismann makes his usual mistake. He incorporates in the rudiment what really are stimuli coming from external conditions during the process of development; he makes a grave confusion between the rudiment and the conditions of its development.