292 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



blood corpuscles are clearly derived by budding off from the inferior 

 Lypoblastic walls of the vitelline blood channels. Clusters of adherent, 

 not fully formed, blood disks are circulating en masse through the ves- 

 sels. Some of them appeared to be amoeboid in character. But the 

 process of blood formation is in its most active phase in Figs. 15 and 

 16, where the vitelline vessels converge :o join the heart. Here it was 

 observed that the vitellus was breaking up into clear globular corjius- 

 cles from -r^o to -^^ of an inch in diameter; the largest corpuscles were 

 always observed to be most deeply imbedded in the yelk, or most re- 

 mote from the vascular channel. A progressive segmentation of these 

 corpuscles was also observed, from which it was concluded that they 

 were directly concerned in the formation of the nucleated oval blood 

 disks. The rapid formation of blood disks in this region had the effect 

 of piling them up into great adheren ; masses about the venous end of 

 the heart, which was also more distinctly marked as the red color of the 

 ovoidal corpuscles became developed as hfiemoglobin was formed. The 

 pulsation of the heart would for a long time sway these masses of cor- 

 puscles back and forth, until finally one after the other would be de- 

 tached from the mass and carried along in the current of blood. Kot 

 only were the corpuscles budded off in this way into the blood channel 

 itself, but they were also found to be held in suspension in great num- 

 bers in the great heart space p^ where every pulsation of the heart 

 ■would cause them to vibrate in the surrounding serous fluid. At the 

 upper part of the heart chamber great lumibers of blood disks were 

 found to be collected together below and in front of the origin of the 

 breast fin /. The hyi^oblastic origin of the blood in this species is 

 therefore undoubtedly a fact, as was learned from repeated observation; 

 ■whether the hypoblast was more than the intermediary parent of the 

 blood disks I am not in a position to state, but this was probably the 

 case, for as the hypoblastic structures w ere broken down into corpuscles 

 in the blood-forming region at the venous end of the heart, there ap- 

 peared to be a constant renewal of germinating cells from below which 

 "were clearly derived from the yelk. The actual phenomenon of cleavage 

 of the cells was not observed since the nuclei were relatively indistinct, 

 and their genesis at this point was assumed to be undoubted from the 

 constant!}" augmenting numbers which were developed independently 

 of any which might accumulate in consequence of eddies in the blood 

 current. The blood disks themselves were not measured, but as com- 

 pared with the size of the corpuscles from which they were derived they 

 ■were estimated to measure somewhat less than 3^0 of an inch in their 

 greatest diameter. 



What may lie beyond the stage represented in Fig. IG I am not able 

 to say, as we were unable to keep the eggs in a healthy state after this 

 jieriod. The species was found in abundance, in spawning condition, 

 at Cherrystone during July and August last, and I take this occasion 

 to express my appreciation of the assistance of Colonel McDonald and 



