[83] EMBRYOGRAPHT OF OSSEOUS FISHES. 537 



a plate of cells; secondly, its cavity appears as a round opening in the 

 plate; thirdly, it grows in length so as to form a nearly straight tube; 

 fourthly, it bends upon itself so as to carry its venous end upwards and 

 backwards; fifthly, it is differentiated into three distinct regions, as 

 shown in Figs. 41, 43, and 46. In freeing itself and becoming tubular, 

 a serous space is formed around it ; the pericardiac space, which is con- 

 tinuous, at least by way of the open venous end of the heart with the 

 segmentation cavity, from which also the fluid filling the heart and peri- 

 cardiac cavity appears to be derived. Not uufrequently colorless blood 

 cells or white cori^uscles may be seen in both the segmentation and peri- 

 cardiac cavities during this stage and later, which are moved or swayed 

 in their bath of serum by the pulsatile action of the heart. These phe- 

 nomena may be observed in numerous species, and seem to be a normal 

 accompaniment of the development of the Teleosteau heart. The presence 

 of now and then a colorless blood cell in the pericardiac ca\ity during 

 the early stages of Gadus, and even of colored ones in the pericardiac 

 spaces of Salmo and Tylosums, shows that this cavity is not wholly shut 

 off" from the yelk hypoblast {'■^couchc Juematogene^^), and therefore not 

 altogether discontinuous with the segmentation cavity. In Elacate the 

 segmentation cavity extends under the head, back towards whore the 

 heart originates, and so close to it that a relation of continuity seems 

 altogether probable at an early stage. The thin veil like fringe shown 

 at the posterior end of the heart of the young cod in Fig. 41 1 have found 

 in longitudinal vertical sections of the recently-hatched embryos oi Alosa 

 to be actually continuous with the exceedingly thin pericardiac mem- 

 brane below and behind the heart, the venous end of the organ actually 

 opening through it into the segmentation cavity. The latter is regarded 

 by me as synonymous with the body cavity; so that the body cavity it- 

 self is derived from the segmentation cavity. The body cavity is divided 

 from the heart or pericardiac space by the posterior pericardiac mem- 

 brane, which is developed concurrently \/ith the heart itself. The peri- 

 cardiac membrane incloses the heart space ventrally and posteriorly 

 and is of splanchnopleural origin. In Alosa sections through it show 

 it to be a membrane of filmy tenuity, with nuclei imbedded in it here and 

 there. It is much thinner in that species than even the outer epiblastic 

 covering of the yelk-sack. In Gadus it is a transverse fold of notable 

 thickness, during the early condition at least, or off the middle line of 

 the body, as shown in Figs. 30 and 40. 



The differentiation of the heart tube into regions is a gradual process; 

 the first portion to be marked off is the dilated anterior end or venous 

 sinus, then the ventricle and bulbous aortfe. A glance at Figs. 29, 46, 

 42, and 41 will show the steps of the process; how the heart tube has 

 been bent upon itself in shifting the venous end round to its final ob- 

 lique position in the middle line. I have not been able to make out the 

 double tube or one within the other, as represented by several investi- 

 gators. Doubtless there is an outer pericardiac layer, but neither in 



