VIVIPAROUS FISHES OF THE PACIFIC COAST. 
427 
tenth segmentation, at least it is not so evident, and it disappears some time after 
the diplastic gastrula has been formed. These outer cells seem to become distended 
with extraneous matter. 
The changes are so rapid from one segmentation to another that a much larger 
series of eggs than I possess is necessary to illustrate all the phases. At this point 
it will be worth while to compare the results ot the successive stages of the segmenta- 
tion of Cymatogaster with those of some large-yolked teleost. Since Agassiz & Whitman 
have so ably discussed and figured the segmentation of Ctenolabrus , that species may 
be taken. Agassiz & Whitman, 1883, p. 45, state that the ninth generation of 
amphiasters is reached in two hours and sixteen minutes, and that the time between 
the eighth and ninth generations is fifteen minutes. From these data I estimate 
that their figure 3 contains amphiasters of the ninth generation, or more than 
250 cells and less than 500 ; their figure 4, amphiasters of the thirteenth generation, 
or more than 4,000 cells. As stated above, the segmentation could not be directly 
observed in Cymatogaster , and in order to determine the serial number of my seg- 
mentation I had recourse to counting the nuclei, a method more laborious and 
inexact with each generation. Since, however, the number of the segmentation, as 
already stated, can not be missed unless the number of nuclei present is actually 
doubled or diminished by half, except at the beginning or end of a segmentation, 
the result thus obtained is sufficiently exact for purposes of comparison. 
If we compare Agassiz & Whitman’s fig. 3 (24a), containing amphiasters of the 
ninth segmentation, with fig. 24, at the beginning of the ninth segmentation, and 
their fig. 4 (30a) during the thirteenth segmentation, with figs. 41 and 42 during the 
twelfth segmentation, the contrast will be seen to be very great, and the work accom- 
plished by Cymatogaster , between the ninth and the thirteenth segmentation, com- 
paratively enormous. The actual time consumed in reaching these stages (three 
hours and sixteen minutes in Ctenolabrus ) is in all probability different, but that 
does not concern us since the time of successive generations of amphiasters differs 
greatly in oviparous fishes in which the generations are more alike. Cymato- 
gaster accomplishes in twelve generations of nuclei what is not accomplished by 
Ctenolabrus till much later. To anticipate somewhat, the stage of Cymatogaster just 
before the closing of the blastopore does not correspond to the same stage in Ctenola- 
brus, but rather to the beginning of the epibolic growth of the blastoderm. Just 
before the marginal infolding of the blastoderm of oviparous fishes it spreads over 
the yolk. This spreading probably corresponds to the sinking of the yolk into the 
blastoderm in Cymatogaster. By the time Cymatogaster reaches a stage homologous to 
that reached by Ctenolabrus at the closing of the blastopore, Cymatogaster is hatched. 
The hastening in the inclosing of the yolk on account of its small size in Cymato- 
gaster and the rapidity of the spreading of the blastoderm at the time of the infolding 
of its edge in most oviparous fishes result in the diplastic gastrula in Cymatogaster and 
the familiar condition in other fishes. The lack of a two-layered gastrula in fishes 
seems to be due to the rapid spreading of the epiblastic portion of the blastoderm. A 
large vitellus permits this spreading to take place more rapidly than the margin grows 
inward, so that the inner anterior and posterior margins of the embryonic ring do not 
ordinarily meet. An interesting egg bearing on this subject is that of Stolephorus, 
in which the blastoderm is placed at the end of an ovate egg. The lateral spreading 
of the blastoderm is here almost impossible, and at the beginning of gastrulation the 
