The conditions in the egg of the bony fishes are very different from those in the preceding forms. The protoplasm, from which the embryo is produced, accumulates at one pole to make the blastodisc. After the cleavage of this blastodisc, the blastoderm that has resulted grows over the yolk sphere at the same time that the embryo is forming along one meridian. I carried out some experiments, in 1895, on the eggs of _Fundulus heteroclitus_. If one of the first two blastomeres of the egg of fundulus is destroyed, the remaining one produces a whole embryo. If three of the first four blastomeres are removed, the remaining one may produce a whole embryo of small size. The problem of development is, in the case of the fish, different from the other cases described, inasmuch as the whole yolk sphere is left attached to the remaining blastomere and is covered over by cells derived from this blastomere. The smaller embryo that is formed lies on a yolk of full size.[122] Wilson’s work on amphioxus has been already described in connection with the experiments on the sea-urchin’s eggs. Later I (’96) also obtained whole larvæ from one-half and one-fourth blastomeres, and I also found that the one-eighth blastomeres do not develop beyond the blastula stage. The number of cells of which the one-half larva is composed is half that of the normal larva, and the one-fourth larva is made up of one-fourth of the total number of cells. [Illustration: FIG. 66.--Ctenophore-egg and embryo. _A._ Normal sixteen-cell stage. _B._ Half-sixteen-cell stage. _C._ Later half-segmentation stage. _D_. Later half-embryo. _E._ Corresponding whole embryo. _F._ Half-embryo seen from side. _G._ Same seen from apical end. In _F_ and _G_, four rows of paddles present, three endodermal sacs and ectodermal stomach.]