NO.U31. BREEDING HABITS AND EGG OF PIPEFLSII—GVDGER. 469 



In this connection the experiments of Huot (1902) are very interest- 

 ing. He took the eggs of Syiuindthnx dumerUii from the marsupium 

 of the male, and, being careful not to break the agg meml^ranes (these 

 eggs were presuma))ly fertilized), put them in clean aerated sea water. 

 This he did also with eggs just before deposition (ovarian eggs), but 

 in no case did development go on more than a few hours. Then he 

 put into the water larva? old enough to move freel3% but these too died 

 within forty-eight hours. I can contirm all his results. 1 have found 

 that eggs in segmentation will go on dividing for a short while, but 

 that within eighteen hours all die. The discoveries of Huot (1902) 

 and of Cohn (1901), that the pouch and its contents act as a physiolog- 

 ical placenta, ofl'er the explanation for the above phenomena. The 

 eggs and embryos, depending on this for oxygen and food, can not 



exist out of the pouch. 



IV. MATURATION. 



Unable to fertilize artiiicially the eggs of Sljyhosfoma foridfe, and 

 having found it impossible to get from the pouch eggs young enough 

 to show the formation of polar bodies, I am unfortunately not in posi- 

 tion to say anything of the pi-ocess of uiaturation. For the latest and 

 best work on this phenomenon the reader is referred to Behren's 



paper (1898). 



Y. FOLIATION OF THE GERM DIRK. 



In the pipefish, fertilization is not necessary to bring about the 

 formation of the germinal disk. Immersion in water supplies the 

 stinudus as it does in many other tishes. All workers on the Salmon- 

 oids. Ziegler (1SS2), His (1S99), and others, so report. Kowalewski 

 (1886) found it true for the goldfish, as did Agassiz and Whitman 

 (188.5) for Cfe/ioiahru.^, though they state that for ]^elagic eggs the germ 

 disk is generally not formed until after impregnation. Brook (1887) 

 confirms this for the herring, but I have found that the eggs of the 

 sargassum fish, Pterophrijne hlxtrlo. form the germ disk shortly after 

 extrusion. Hertwig says (Ilandbuch, p. 541): ''One can emphatically 

 say for almost all fish eggs that by their transfer into water such a 

 powerful force is l)rought into play that the concentration of the germ 

 disk results," but that ''if they are impregnated first, a more rapid 

 growth and larger size for the germ disk follows." 



All writers, notably Brook (1887) and Ryder (1887), describe this 

 formation as ))rought al)out by the streaming of the protoplasm to the 

 germinal pole. There are three modes in which this may take place: 



(1) By streams from the circumference only. This is the method 

 in most fishes, especially those with pelagic eggs. (See Brook, Ryder, 

 Kingsley and Conn, and many others.) 



(2) By streams from the circumference with the help of little 

 "processions" from the interior of the yolk (Ziegler, 1882, and Oel- 

 lacher, 1872, for the trout). 



