1022 The American Naturalist. [November, 
the amphibian segmentation which so exactly imitat@ the teleostean types. 
. According to Rauber, the first equatorial furrow of the frog has 
bese fost: in the Teleost. Agassiz and Whitman would seem to believe 
that the æ riori improbability of such a loss taking place is so great 
that, in spite of the variations just described, it is preferable to regard 
the first three furrows as homologous in the two groups. I do not see 
the inherent improbability of the loss. On the contrary, the disap- 
pearance of segmentation in the ventral half of the egg, coupled with 
the early contraction of the protoplasm (belonging to this half) towards 
the upper pole, make it easy, I think, to understand how the loss was , 
brought about.” The author seems to have come to this conclusion 
largely on account of the close resemblance in the arrangement of the 
_ eight micromeres in the 16-celled frog’s egg with the eight-celled stage 
-in the bass; but inasmuch as there exist the greatest individual differ- _ 
ences in the arrangement of the micromeres in the frog, and also in the 
fish, we have every reason to believe such a general resemblance might 
have been independently acquired in each case, and that the eight-celled 
stage in the fish corresponds to the 16-celled stage in the frog ; and that 
the resemblance between the 16-celled stage (eight micromeres, eight 
macromeres) in the frog and the eight-celled in the fish is entirely super- 
ficial. Moreover, if, as the author attempts to show, the fish gastrula 
may by derived directly from the frog’s gastrula, we have every reason to 
believe that so fundamental a process as the cleavage stages must be 
similar, making a very strong ‘‘ a priori improbability ’’ that the third 
furrow of the frog has dropped out of the fish egg. 
As is well known, the eight-celled stage of the frog separates in gen- 
eral the micromere cells of the upper (anterior?) pole from the lower 
cells containing more yolk. Now if the view taken by Agassiz and 
Whitman be true, it would seem probable that in the eight-celled stage 
of the fish four (most probably the more central four) cells would be 
entirely cut off from the protoplasm covering the yolk, and that the 
other four cells would have their protoplasm continuous, in part at 
least, with the protoplasm covering the yolk. ‘To some extent the 
author’s figures bear out such an interpretation, although he does not 
seem to have examined the sections from such a point of view. ; 
The origin of the periblast in the bass is the same as is described in 
Ctenolabrus by Agassiz and Whitman. ‘These authors proved 

beyond a doubt that in Ctenolabrus the nuclei are derived from the __ 
marginal cells of the blastodisc, which from the earlie:t stages of nt 
mentation are connected with the tee or periblastic protoplasm.” 
(See above.) 

