The Recapitulation Theory. 86 



formed in some instances and only a nucleus in others. In every 

 case this ventral canal cell or nucleus is only a reminiscence which 

 continues to reappear as a terminal stage although in its origin it 

 dates back at least to the Liverworts. But it appears to be finally 

 eliminated in the Taxodinea;. The cases just cited, however, do not 

 involve recapitulation so far as the individual ontogeny is concerned. 

 They are only seen to be closely related to such phenomena from a 

 comparative phylogenetic point of view. 



Such a reduction series Coulter (1915) no doubt justly con- 

 siders to represent the expression of an orthogenetic tendency, 

 though whether it is the result of climatic differentiation is not so 

 clear. The series derives its interest from the fact that the higher 

 Gymnosperms, in which the archegonium is quite eliminated, must 

 have had ancestors in which that reduction gradually took place. 

 From our present point of view, the explanation cannot be found 

 in a series of successive nuclear variations or mutations, for these 

 would make themselves felt in other parts of the organism, affecting 

 many characters in a correlated fashion. Though the argument 

 is by no means conclusive as regards this matter, yet it seems 

 most reasonable to consider these as organismal characters and to 

 explain the shortening of the gametophyte generation in the same way 

 that recapitulatory characters find their explanation. The direction 

 in which that explanation is to be sought will be discussed below. 



Another support for this view may be found in the fact that 

 the chromosome number is remarkably uniform throughout the 

 Gymnosperms, being, with few exceptions, 24. If the evolution of 

 the Gymnosperms had taken place largely through mutations, i.e., 

 through changes arising in the germinal chromatin, one would expect 

 it to have produced some effect on the chromatin morphology in the 

 various species. The Angiosperms, by contrast, in which there is 

 evidence that much mutation has taken place and is now going on, 

 are characterized by remarkable variety in number, size and shape 

 of their chromosomes even within single families or genera. The 

 same is true of insects. 



Other cases of ancestral reminiscence in gametophytes, as 

 though the organism repeated certain stages from force of habit or 

 from some source of energy impelling its development forward, are 

 to be found in Angiosperms. In the nuclear divisions within the 

 embryo sac an evanescent cell-plate sometimes appears on the 

 spindle. But this reminiscence of wall-formation soon disappears 

 along with the spindles. The same is true of the prothailial cells 



