480 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxix. 



tig. 20, Plate VI. Tliis is probably a descendant of a blastoderm like 

 fig. 14, on the same ]>late. No description of it is needed, beyond 

 calling attention to the fact that the live upper cells are cut out by an 

 equatorial furrow. This is seen by referring to fig. 21, which is a 

 ventral view of the same blastoderm. Here only five of the vertical 

 planes seen from above cut all the way through. The ones marked o 

 in tig. 20 have not reached the base, l^'he small segmentation cavity 

 {s. c.) recalls that of lig. 15. Let us compare with this the next, 

 fig. 22, which is a view from below of a similar high-piled sixteen- 

 celled stage. Here there are nine basal cells resting on the yolk, 

 six in the second tier, and a central one forming the keystone of 

 the arch, the whole inclosing a spacious segmentation cavity. Barring 

 the fact that the segmentation cavity (.y. c.) extends under the marginal 

 cells, tig. 48, Plate VIII, may be given as a section through tig. 22 in 

 any vertical plane passing through the keystone cell. The central cell 

 has not yet completely cut itself otf from its neighbor to the right, and 

 the cell to the left has a resting nucleus curiously elongated. 



There have now been tigured and described in surface views and 

 sections, such sixteen-celled structures as may be considered typical 

 for the pipetish. Of these, two are sutiiciently like the usual teleost 

 form as to be called normal, but a great majority, fully 90 per 

 cent of those studied, are like tigs. 18, 19, and 20, Plate VI. In this 

 connection Hertwig's statement (Handbuch, pp. 64.5-64()), with refer- 

 ence to the fourth segmentation and formation of the sixteen-celled 

 stage, is of interest. He says: "The end result is everywhere the 

 same, a ' checkerboard-like ' arrangement of sixteen blastomeres, four 

 in the center, and a circle of twelve marginal cells.'' How untrue this 

 is for the pipetish, a glance at the tigures given and at the table shown 

 on page 478, will demonstrate. 



EQl'ATORIAL PLANE OF SEGMENTATION. 



All investigators are agreed as to the homology between the first 

 and second furrows in teleost and am[)hibian eggs, but whether or 

 not the third furrows coi"i"espond is a ver}' del)ated question. 



Hotfmaiui (1881) tigures and descri))es in pelagic fish eggs the tirst 

 segmentation as ecjuatorial. dividing germ from periblast; but later 

 (1888), he acknowledges his error and declares that in Sdlnia the third 

 furrow is equatorial. Ziegler says that the third furrow in the sal- 

 mon and trout is equatorial and divides eight upper from eight lower 

 cells, the latter not l)eing as vet marked otf from a periblast. Rauber 

 (1883) made a careful study of the subject based on the well-known 

 fact that the fourth amphibian furrow in a great many cases is not 

 truly meridional but avoids the polo and forms many structures like 

 figs. 7 and 9, Plate V. He concludi's that the tirst <-quator'ial furrow 

 of the frog has been lost in the Teleost, and homologizes the third 



