MUSEUM OF COMPARATIVE ZOOLOGY, 89 



sible rate of cell division and migration, especially in consideration of the 

 scarcity of parablast cells in the yolk. The position of the cells also fails 

 to give any evidence of such migration. The close proximity of the lat- 

 eral plates to the gut, and the impossibility of immigration from behind 

 at this early period of development, are sufficient arguments in my mind 

 to refute the theory of parablastic origin. It is not strange, however, 

 that one should defend the view of a parablastic origin when studying 

 only the later stages, and especially such conditions as are shown in 

 Ziegler's plates ('87, Taf XXXVIII. Figs. 31 Bj, 32 Bi, 32 B3, "hz.") 

 on the salmon, where the cells of the intermediate mass are found for 

 a considerable distance closely packed between the parablast and lateral 

 plates. 



It was also by reason of this close relation between the parablast and 

 intermediate cell-mass that Oellacher ('73), after stating that the results 

 of his work showed the endocardium to be mesodermic, added that he 

 thought some parablastic cells must be included also. 



In the cod, however, this close relation of parablast to intermediate 

 cell-mass is clearly not present, and although in places the parablastic 

 cells are in a position where migration into the intermediate cell-mass 

 would be possible, still I could not find in any of the sections examined 

 the least evidence of such, and I am confident that these cells are not as 

 a mass derived from the parablast ; moreover, in all my study of the 

 works of others I have failed to find satisfactory proof that any of the 

 cells of the mass come from the parablast. Many observers have con- 

 jectured such an origin, but none have actually seen cases in which the 

 cells gave evidence of migration from the parablast, and their statements 

 therefore still remain to be proved. 



2, Endoderm. — Another possible origin of the intermediate cell-mass 

 is the endoderm. 



In stages of development from five to six and a half days old, the fore- 

 gut in the cod is considerably flattened, and is sending out lateral evagi- 

 nations to form gill slits ; at this time, where the lateral plates are 

 approaching the median plane, the intermediate cell-mass certainly has 

 a position that suggests its endodermic origin. Take, for example, Fig- 

 ures 6 to 8 (Plate II.), where the intermediate cell-mass is bounded by 

 the parablast, the endoderm, and the well defined latei'al plates, and 

 where the cell layer between the lateral plates and the gut is only one 

 or two cells deep. The parablast cells at this stage show no evidence of 

 migration, the lateral plates are too well defined for the loose cells in 

 question to have come from them, and consequently the endoderm, since 



