474 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxix. 



arrang-ecl triinsversely to the plane of cleavage. Oellacher (1872, fig. 

 20) describes and figures a section through two cells of a four-celled 

 stage in the brook trout very like this. He saj^s an indistinct streak 

 made up of faint granulations runs vertically from the external groove 

 toward the base. Henneguy (1888, fig. 00) gives a figure of a two- 

 celled stage very like tig. 33, Plate VII, and says that the fine line divid- 

 ing the two cells is bordered on each side by clear protoplasm w hich is 

 traversed by very fine lines parallel to each other and perpendicular 

 to the median line, and that these fine lines lose themselves in the sur- 

 rounding protoplasm. His (1898, figs. 7, 8) illustrates and describes 

 similar structures in the S3'ncytium at the base of the trout germ in 

 early stages. In fig. 34: we have a high arched two-celled stage in 

 which the perfectly distinct cell wall is interrupted by a vacuole near 

 its center. This is plainly a derivative of fig. 32, as the preceding is 

 of fig. 28. 



Fig. 35 is a section through fig. 3, Plate V, in the plane «-/>, and 

 shows the split between the two cells dilated into a large vesicle at the 

 bottom. Very frequCntl}' the division between the two cells takes the 

 form of a deep cleft with nearly vertical walls, and at the bottom 

 the cleft ma}" or may not dilate to form a small vesicle. These struc- 

 tures are shown in fig. 36, and are oftentimes much larger than figured 

 here. In fig. 37 we see the split being formed by the breaking down 

 of the walls of a series of vesicles placed vertically over one another 

 in the center of the blastoderm. This formation of vesicles in the line 

 of cleavage was, so far as I know, first figured and described, for the 

 trout, by Oellacher in 1872. Balfour (1878, figs. 6, 6a, and 6b, Plate 

 I) illustrates and at some length describes vacuoles in the earl}- furrows 

 of the skate. He describes such a beaded structure, as shown in my 

 fig. 37, and thinks that these vacuoles are more common than supposed, 

 and that they play a considerable part in the segmentation. Brook 

 (1887) describes the like in the herring but gives no figures. Kowa- 

 lewski (1886, fig. 1, Plate XVII) finds vesicles at the bottom of the 

 furrows in the early stages of the goldfish. Agassiz and Whitman 

 (1889) figure, in surface views of blastoderms of Ctenolahrus^ rows of 

 small vacuoles extending along the whole length of the cleavage 

 planes in the two- and four-celled stages, but do not refer to them in 

 their text. Fusari (1890, figs. 4 and 5, Plate III) shows in both sur- 

 face views and sections blastoderms with vacuoles. Some of the sec- 

 tions show vacuoles with large dilatations at the bottom like those in 

 figs. 35 and 36, Plate VII. 



In the pipefish, the first furrow does not cut through to the 3'olk. 

 (See figs. 34, 35, 36, and 37.) In this respect it agrees with Crhticeps 

 (Fusari, 1890), the Herring (Brook, 1887), Cavd.sstiut^ (Kowalewski, 

 1886), the Bass (Wilso.n, 1891), the Salmon and Trout (His, 1898), but 

 is unlike Mcrlueiu-s (Kingsle}' and Conn, 1882), Gadus (Cunningham, 



