[89] EMBEYOGKAPHY OF OSSEOUS FISHES. 543 



hypoblast, and it is to be inferred that blood cells are budded off directly 

 into them, the division of the free nuclei in the subjacent Plasmodium 

 probably multiplying and giving rise to these blood corpuscles. In this 

 •way it is conceivable that the yelk is gradually broken down, just as we 

 know that by a similar process the yelk of Alosa, which has no vitel- 

 line circulation, is absorbed. The lumen of the vit€ilhue vessels is also 

 depressed or somewhat flattened upon the yelk surface, and not round 

 as in other parts of the body, and in some cases {Tylosurus and Apeltes) 

 they have at first the form of exceedingly iiTegular channels, which are 

 evidently much more deeply excavated into the Plasmodium layer at 

 some points than at others. In Apeltes the first sign of any vessels is 

 the appearance of a large irregular sinus on one side of the body between 

 it and the yelk in which the blood corpuscles vibrate in unison with 

 the pulsations of the heart, there being as yet no complete open channel 

 or cyclical path for the passage of the blood back to and through the 

 heart. The vessels end blindly at first and are also progressively length- 

 ened, and possibly the rhythmical imjiulses given to the primitive 

 blood during pulsation helps to open up the channels still farther. 

 Such blind vascular terminations are found on the yelk of a number of 

 Bpecies at an early stage of development of the blood system and usu- 

 ally end acutely but finally i)usb towards and open into some pre-exist- 

 ing channel, when they at ouoc become wider. In such blind vascular 

 terminations the blood cells simply oscillate back and forth. In Tylo- 

 sums the early blood cells may form adherent masses in the great merid- 

 ional vessel of very uneven caliber, which is the first to be formed and 

 wherein these masses move fitfully or only oscillate with the pulsations 

 of the heart. They soon acquire a reddish tinge, but the fact that they 

 adhere together shows that possibly they are of the nature of confluent 

 white corpuscles or even masses detached from the Plasmodium layer 

 which here evidently forms the floor of the vessel, in the act of segment- 

 ing and becoming blood corpuscles. This primitive blood of Tylosurus 

 is also rich in serum and poor in blood-cells. In Apeltes the blood cella 

 are more numerous at a similar stage. 



In Gambusia the blood-vessels which traverse the yelk, like in 8alm0j 

 seem more or less deeply imbedded in the yelk-hypoblast layer, and I 

 find it difficult to determine the nature of their outer coverings; inter- 

 nally they seem to lie in immediate contact with the yelk, so that the 

 contained blood-cells in sections of hardened specimens are packed right 

 against and imi^ressed into the Plasmodium or yelk hypoblast. The 

 vascular network over the yelk of Gambusia is, however, much finer than 

 in Salmo, and relatively thicker as seen in sections, but the external 

 covering of the yelk-sack, unlike in the latter species, is not thick and 

 two-layered, but exceedingly thin and formed solely of epiblast on the 

 ventral and lateral portions. 



In Alosa no cellular elements are distinguished in the yelk-hypoblast; 

 it is a thick homogeneous coating over the yelk, with scattered free 



