EMBRYOLOGICAL CRITERION OF HOMOLOGY. 113 



doubt true. This explanation contains, however, a fatal ad- 

 mission ; for if secondary modification may go so far as com- 

 pletely to destroy the typical relations between the germ-layers 

 and the parts of the adult, then those relations are not of an 

 essential or necessary character, and we cannot assume that 

 the germ-layers have any fixed morphological value, even in 

 the gastrula. 



Let us finally consider the study of cell-lineage or cytogeny. 

 The contradictions here reach a climax. In some cases, it is 

 true, there is a really marvelous agreement in the cytogeny of 

 related forms (annelids, gasteropods), so that adult homologies 

 are accurately foreshadowed by cell-homologies, even in the 

 earliest cleavage-stages. But as we extend the comparison 

 extraordinary contradictions arise. Lilly has recently shown 

 that the lamellibranch Unio agrees very precisely with the 

 gasteropod Crepidida (Conklih) up to a certain point, but then 

 shows a sudden and at present inexplicable departure in the 

 origin of the larval mesenchyme. The cephalopod suddenly 

 presents us with a totally different type of cleavage in which 

 no homologies whatever can be drawn between the individual 

 blastomeres and those of other mollusks or of annelids. In 

 another direction we find (in the Polyclade) a cleavage very 

 closely resembling the annelid type in form, yet the individual 

 blastomeres have from the very start an entirely different mor- 

 phological value. 



II. 



The puzzling facts reviewed in the foregoing brief survey 

 leave no escape from the conclusion that embryological develop- 

 ment does not in itself afford at present any absolute criterion 

 whatever for the determination of homology. Homology is not 

 established through precise equivalence of origin nor is it ex- 

 cluded by total divergence ; and this statement holds true for 

 all the stages of development, though on the whole the later 

 stages seem to show a closer agreement than the earlier. But 

 it does not by any means follow that the embryological method 

 has therefore failed and must be abandoned as a means of in- 

 vestigating homologies. The most skeptical critic of the re- 



