No.nai. BREEDING HABITS AND E(l(^ OF PIPEFISH— (IVD^IEB. 471 



The oil drops in the pipeiish eg-g- arc not numerous enough to make 

 it tloat, but from their location the}' maintain the germ in an upright 

 position. If the eggs are overturned, this buoyanc}^ causes them to 

 rotate quickly in the licpiid tilling the '" ])reathing- chamber-' of 

 Ransom. How long this rotation persists 1 can not say, but certainl}' 

 until after the closure of the blastopore. Rathke (1837) first noted 

 this in the eggs of Black Sea forms. He also described, as best I can 

 make it out, an albuminous material coagulable in water or in air, 

 which fills the "" zwischenraum " referred to above. Whatever may 

 be the liquid filling this space in ^S'. fforlda^, it does not coagulate in 

 water, air, or in any of the fixing fluids I have used. It might l)e well 

 to add here that this rotation of the Qgg is not a new phenomenon, 

 having been reported, notably by Ziegler (1882) and His (1899) for 

 the salmon family. 



My earliest preservations of eggs with forming germ disk were 

 made four to five hours after the eggs had been placed in the water, 

 hence I am not able to describe by sections its formation. In any 

 case, however, I could not hope to add anything to the classic paper 

 of Agassiz and Whitman on Ctoioldl/'Hs, or to the more recent 

 memoir of Behrens on the })rook ti'out. Since I preserved eggs at 

 intervals of from five to twenty-five hours, I have sections which 

 illustrate the progressive degeneration of the blastodisc. So far as 1 

 know this has never been shown, and hence it may be of interest to 

 give a few figures illustrating this phenomenon. 



Fig. 1, Plate V, represents the sharpl}- marked ott' blastodisc resting 

 on the yolk sphere. It shows the relative diameters of blastodisc, 

 ''^ disque liulleux^'' yolk sphere and ^<g<y membrane. Fig. 28, Plate VII, 

 is a central section of a germ disk five hours old. The concentration 

 of protoplasm is not yet perfect. As l)est I can make it out, all has 

 not yet emerged from the central .yolk. The dotted line marks ofi' a 

 region where protoplasm and 3'olk are so closely intermingled as to be 

 indistinguishable. Oellacher (1872, fig. 17) figures and describes a 

 similar germ disk for the trout. Fig. 29, Plate VII, shows a degener- 

 ating blastodisc ten hours and twenty minutes old. Such structures 

 are not unfrequent in unfertilized Qg^:^ found among others in the 

 four to sixteen celled stages in ages from eight to twelve hours. 

 They are also found in eggs which have been in water about ten 

 hours, and, I am inclined to think, are of fairly regular occurrence in 

 degenerating blastodiscs of unfertilized eggs. 



Strieker, in 1865, described what he called an entirely new mode of 

 cell formation in the blastoderm of the brook trout — that is a ))udding 

 ofi* of cells — which he thought originated in the auKpboid activities of 

 the protoplasm. His figures show blastoderms with from one to twenty- 

 three " buds," lumps, or vesicular swellings on the outer surface, and 

 his one section is very inconclusive. Unfortunately, I have no surface 

 views of pipefish eggs showing any of these structures. The following 



