HEREDITY 489 



often shows telescoping; and it is truer of stages in organo- 

 genesis than of stages in the development of the embryo as 

 a whole. 



It has also to be remembered that one term in the com-' 

 parison, the phylogenj, is very imperfectly known, so that 

 assertions as to the exactness of the recapitulation must be 

 taken with reserve. Needless to say, one must beware of 

 the vicious circle of arguing from the development to the 

 presumed ancestor, and then from the ancestor to its recapit- 

 ulative rehabilitation in development. 



Another saving clause is that the individual development, 

 especially when there are larval stages, may have its recapitu- 

 latory features obscured by secondary adaptations to rela- 

 tively recent conditions of life. Thus one does not look 

 for recapitulation in the life-history of insects which have 

 sub-aquatic larvae, for these have been secondarily adapted 

 to a habitat which was not that of the ancestral stock. We 

 may also recall the idea that life-histories have been adap- 

 tively altered by lengthening out one chapter and telescoping 

 another. 



Another saving clause concerns specificity, the individu- 

 ality and uniqueness of every well-defined type. Increased 

 precision of embryological work has shown that from very 

 early stages in ontogeny an organism is itself and no other. 

 An expert can distinguish an embryo chick a few days old 

 from an embryo duck, before either of them shows any avian 

 characters. There is only a technical difficulty in the way 

 of distinguishing even the cells of an embryo mouse from 

 those of an embryo rabbit, or those of an onion from those 

 of a lily: the number of chromosomes is different. But a 

 recognition of specificity from first to last is not inconsistent 

 with admitting a significant correspondence between steps 



