NO. 1431. BREEDING HABITS AND EGG OF PIPEFISH— GUDGER. 473 



ONE-CELLED STAGE. 



Thi.s is shown in iig. 1, Plate V, from al)ove, and in iig-. ;>2, Plate 

 VII, in section. It is high arched and falls steeply into the outer peri- 

 l)last, from which it is clearl}^ marked off by the circumferential 

 furrow of the authors. This furrow is sometimes so pronounced in 

 the germ disk of the Salmon famil}" that the disk literall}^ overhangs its 

 l)ase. See His (1898, tig. 1) for the trout and (tig. 2) for the salmon. 

 Kuprt'er (18<i8), however, sa3\s that in a European Syn(jnatJiHs{ii\)Qc\Q<, 

 not given) the germ disk is not sharply marked off from the periblast, 

 and that this condition holds till the end of the four-celled stage. 

 Most workers on the Salmonoids, Behrens (1898), and, nota))ly, His 

 (1899), represent the unsegmented blastodisc as somewhat sunken in a 

 saucer-shaped depression. In the pipefish, however, the blastodisc, 

 fig, 1, Plate V, underlaid with oil glol)ules, rests on a slightly flattened 

 area at the upper pole. Below it is not sharply marked oft' from the 

 yolk, but across its l)ase extends a band, about as wide as the periblast 

 to the right, composed of mixed yolk and protoplasm. The section 

 shows several vacuoles to the right, which in the living eg^ were 

 probabh^ filled with oil. Brook (1887) describes in the herring a 

 ])lastodisc with yolky base; His (1899), the like in the salmon. 



This blastodisc was found in a batch of eggs in the eight to sixteen- 

 celled stage (eight to twelve hours). Plis (189H) says the germ disk 

 in the Salmon is formed in from one to four days. Hertwig (1903) 

 says that the formation of the germinal disk in the herring takes place 

 in two hours, and in the trout from seven to eight hours. Evidentl}^ 

 the time varies with the kind of fish, the temperature, and the purity 

 of the water. In the pipefish I have found it to take place in from 

 four to six hours. It is noteworthy that in none of the ])lastodiscs 

 which were sectioned have I ever found a nucleus. Brook (1887) could 

 find no nuclei in the herring until after the appearance of the third 

 furrow. 



TWO-CELLED STAGE. 



As in Teleosts generall}^, the blastodisc elongates slightly before 

 the appearance of the first furrow, and, as a result, one axis is some- 

 what longer than the other. This is shown in fig. 2, Plate V, the nor- 

 mal two-celled stage, in which the blastomeres are equal. In fig. 3, 

 however, we have an irregular segmentation, with one cell much larger 

 than the other and with a vacuole in the line of division. Of this tjpe 

 quite a number were found. 



Fig. 33, Plate VII, shows a fiat two-celled blastoderm, not definitely 

 marked ofl' on the right from the outer periblast, in which the nuclei 

 have divided, the external furrow has formed, but the cell wall has 

 not yet come into existence. In the line of division, the protoplasmic 

 reticulum has formed a verv delicate network of dendritic fibrils 



