OF ARTS AND SCIENCES. 63 



phases of the nucleus, do not appear to support the opinion that the 

 chromatine is merely passive food-material, as maintained by Brass.^^ 

 The entire behavior of the chromatic elements during the process of 

 division seems to be opposed to associating them with such passive 

 material as food-yolk. The partial disappearance of chromatine in 

 the nuclei of starved animals cannot be regarded as conclusive evi- 

 dence that it is surplus nutritive material. 



The early araphiastral divisions are very nearly synchronous ; but 

 it is rather rare to find the nuclei keeping equal pace sufficiently long 

 to enable one to trace the exact genetic relationship of 64 blasto- 

 meres. "VYe have obtained only a few preparations showing precisely 

 64 cells, and none containing exactly 128 cells. The periblastic nuclei, 

 at the time when they first appear around the margin of the blasto- 

 disc, all divide nearly simultaneously, nearly all the amphiasters 

 taking a radial direction. 



The Cleavage-cavity. — Ryder has traced the history of the cleavage- 

 cavity with much greater care and thoroughness than any other ob- 

 server. He has not, however, given us the early history of this cavity. 

 We have traced it from its beginning in the 4-cell stage up to the 

 time when it becomes a spacious cavity, roofed by the expanding 

 blastodisc and floored by the periblastic cell layer. During the cleav- 

 age stages it remains a very shallow cavity. Its outline, traced in 

 Fig. 1, is quite distinct in most of our preparations and sections. 



The Germ Layers. 



The Periblast. — Among previous investigators, Rauber ^^ has 

 made the nearest approach to the discovery of the true origin of 

 periblastic cells. His observations were made on hardened ova of 

 Gobius, in which no periblast (" plasmodium ") was found till towards 

 the end of cleavage. At this time a differentiated ring of marginal 

 cells (" Randschicht ") was found, which were regarded, somewhat 

 doubtfully, as the primary periblastic cells. The following remark 

 will show how the case stood in his mind : — 



" Die Centralzellen sind ihrerseits immer umsjiumt von einer flachen 

 Randschicht, als einem Rest der ursprlinglichen Randschicht, von der 

 sich neue Zellen abgeschniirt und den vier ersten Centralzellen bei 

 gesellt haben. Anf diese Weise kommt es bei Gobius., so viel ich aus 



65 Brass. Zool. Anz., VI., No. 15G, p. 681, December, 1883; and Zeitschr. 

 f. wiss. Mikroskopie, I., No. 1, pp. 30-51, 1881. 

 « Rauber, 1. c, pp. 2ti8-200. 



