122 BIOLOGICAL LECTURES. 



ungsorgane) they assume the same function in the ontogeny. 

 Through the stimulus or by the aid of these organs, now be- 

 come rudimentary, the permanent parts of the embryo appear 

 and are guided in their development ; when these have 

 attained a certain degree of independence, the intermediary 

 organ, having played its part, may be placed upon the retired 

 list."i 



This principle is unquestionably of fundamental importance ; 

 and it remains true whether or not we accept the doctrine of 

 " substitution," with which it was connected by Kleinenberg. 

 It is thus that I would interpret, for example, the gastrula 

 stage of development. The diblastic embryo is a necessary 

 stage in the formation of a complex multicellular body of which 

 the most fundamental characteristic is the differentiation of an 

 internal part devoted in the main to the functions of nutrition 

 from an external part which serves as a protection and as the 

 medium of communication with the environment. In a broad 

 sense, therefore, the diblastic embryo does represent an 

 ancestral phase such as still exists in the lower metazoa, biU 

 only by virtue of the persistence of the original fmctional con- 

 trast between the iniier and outer parts. For if the diblastic 

 embryo be simply an inheritance from the ancestral type 

 (" Gastraea," etc.), why has not its ancestral mode of origin 

 likewise persisted "by inheritance," and why should it arise 

 by processes (invagination, delamination, etc.) so widely diverse .? 



We are still too ignorant of the nature and distribution of the 

 forces at work in development, and so of the causal connection 

 between the successive stages of ontogeny, to determine how 

 far this principle can be applied. But it seems clear that when 

 once a particular train of events has been established in the 

 ontogeny, it must form, as it were, a path of least resistance 

 along which the idioplasmic transformation will continue to pro- 

 ceed until definite causes operate to divert it into a new path. 

 Only upon such a view can we form any conception of the 

 physiological meaning of recapitulation in the case of function- 

 less ancestral stages. And it is equally clear that we cannot 

 successfully analyze the morphological aspect of development 



1 Kleinenberg: Lopadorhynchus, p. 223. 



