240 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
furrow can not be said to be bomologons with the medullary groove, the causes which 
produce thetwo grooves are the same. Thisconsiderationshowsus,Ithink,thatitisper- i | 
missible to regard the furrow as representing the extreme upper part of themedullary J 
groove, the chief part of which is ideally present in the middle line of the keel. 
Minot has recently in a brief communication to the American Jlaturalist (Novem |' 
ber, 1889) urged that no good reason exists for believing the solid keel of Teleostei to be a |''i 
derived structure. He regards the keel as representing an ancestral nerve cord, which f I 
arose as a thickening of the ectoderm, and hence as directly akin to the nerve cord of f 
annelids. The formation of a canal within the keel is, for Minot, the primitive method. ^ j 
The open medullary groove is a secondary feature. This view does not, I think, find 
support in the ontogeny of the teleostean nerve cord, for all the details here point to | 
the derivation of the keel from an open groove. i 'l 
Like Gotte and subsequent investigators I have not found anything which could | j 
confirm Oalberla’s account of the development of the keel in Syngnailms (10). In the t ; 
Bass there is no special sheet of cells running down from the neural furrow into the I \ 
keel mass. | • 
The strange interpretation which Kupfler (26) has applied to the neural furrow in f | 
the Trout, regarding it as a primitive groove homologous with that of Amniota, has I 
been sufficiently criticised by Henneguy (18, p. 531). The account which Miss Johnson | ^ 
(23) has given of a fusion of layers in the Newt, along the median dorsal line, does | j 
not find support in the Teleost, and its presence in the Newt must, I think, still be I ? 
regarded as problematical. | j 
Further development of the heel. — In the series of transverse sections (Figs. 60-64, | 
Pis. xcv and xcyi) it is seen that the wide medullary i)late disappears last in the | ; 
posterior region. The ueurenteric streak need not detain us. It is composed of the 1 1 
fused neural cord, notochord, and hypoblast. The notochordal cells are sometimes f 
distinguishable from the rest (Fig. 66). The constriction of the keel from the surface i, 
ectoderm has begun in the neck region (Fig. 62). Fig. 63, through the brain behind I- 
the eyes, and Fig. 64, through the eyes themselves, call for no explanation. The 1; 
constriction of the entire neural cord from the ectoderm is finished by the time the iv' 
embryo is 45 hours old (Figs. 88 and 90, PI. xcviir, and Fig. 94, PL xoix, through i | 
the trunk, and Figs. 95, 96, and 97, PI. xcix, through the head). I,'; 
The formation of the central canal, cavities of the brain, and optic sacs, is accom- 
plished in the manner described by Henneguy (18) by the simple separation along the 
middle line of the constituent cells. Oellacher (33) and Hoffman (17) state that the I 
central cells disintegrate, and that thus the cavities are established. This is not so in : 
the Bass. 
In the stage when the optic sacs begin to form (Figs. 62, 63, and 64) the cells compos- I | 
ing the brain and spinal cord are elongated at right angles to the median plane of the l| 
embryo, but they are interlocked and disposed in an irregular fashion. The beginning I IT 
of this arrangement may be seen in Fig. 57, PL xcv. The cells thus arranged begin, f | 
as seen in transverse section, to dispose themselves in two rows (Fig. 72, PL xcvii) 1^ 
which, at first interlocked at their inner ends, gradually acquire a more even dividing .| 
surface (Figs. 75 and 78, PL xcvii). The process continues until the inner ends of the | 
cells form an approximate plane (Fig. 97, PL xcix). The separation of the opposing I 
surfaces takes place first, at least in the spinal cord, at the upper and lower edges of I 
the canal (Fig. 90, PL xcviii. Fig. 94, PL xcix). The characteristic shape of the I 
