64 PROCEEDINGS OF THE AMERICAN ACADEMY 



Fig. 2 can be best explained as an 8-cell stage, in which one of the 

 raisphicod nuclei escaped notice. The inclination of the axis of the 

 ovuwi and the method of treatment account for the illusive appear- 

 ances and the failure to discover any continuity between blastodisc 

 nnd periblast. 



Fig. 3 is probably a 32-cell stage (it cannot be earlier). It repre- 

 sents four amphiasters belonging to marginal cells as if they were in a 

 median plane. 



Figs. 4, 5, and 6 combine surface views with optical sections. They 

 become perfectly intelligible the moment we recognize the fact that 

 the nuclei seen beneath the blastodisc occupy an entirely unnatural 

 position. Their true position is superficial, near the margin of the 

 blastodisc. These nuclei are undoubtedly true periblastic nuclei in 

 Figs. 5 and 6. 



Figs. 7, 8, and 9 of the same plate represent stages in the division 

 of the Jirst cleavage-spindle. Fig. 7 of this plate corresponds very 

 nearly, if not exactly, to Fig. 3 of Plate III., the latter being another 

 of those perspective figures with the perspective left out. 



Fig. 10 is a puzzle: we know of no stage in the normal develop- 

 ment of the teleostean ovum that would present three cells in an opti- 

 cal section. 



According to Hoffmann's view, Fig. 7 ought to present two parallel 

 amphiasters. His explanation of the omission of the hypothetical infe- 

 rior amphiaster is, that, owing to its more central position, it could 

 not be distinctly seen. Our view would account for its omission on 

 entirely different grounds. 



IIotTaKvnn repeatedly calls attention to the fact that his " parablas- 

 tic" nuclei keep an exactly even pace with the "archiblastic" nuclei 

 in the process of division. This accords with our interpretation of 

 Figs. 1 and 2 ; for in these early stages all the nuclei divide syn- 

 chronously. 



The Origin of Periblastic Cells. — AVith the IG-cell stage before 

 described begins a most interesting chapter in the history of the 

 periblast. Reserving a detailed description of the cleavage for the 

 later and full account of our observations, we may here confine our 

 attention to the more important events of this history. The virtual 

 equator of the ovum, as we have said, lies in tlie marginal cells of the 

 blastodisc, and may be supposed to coincide very nearly with the 

 plane of division of these cells, as seen in Fig. 2. The peripheral 

 half of these cells contains the larger part of the material which, at 

 the outset, formed part of the periblast. Approximately speaking, 



