236 BULLETIN OF THE 



these cells form a complete endothelial lining, which follows closely the 

 delicate basement membrane. They doubtless represent the loose cells 

 alluded to in Stage IV. as occurring within the fundament of the glomus. 

 Their origin 1 have already discussed in connection with that stage, 

 where 1 showed that they were probably not derived from the outer 

 layer of the glomus. Although the structure of the central mass of 

 cells is, it must be admitted, somewhat obscure, 1 have found no evi- 

 dences of the complication which Hoflfmann ('8G, p. 591) has recently 

 maintained for it. 



On the contrary, a comparison of many individual cells in this mass 

 with the loose cells in the cavity of the aorta has made me confident 

 that most of the cells contained in the glomus are embryonic blood cor- 

 puscles. It is possible, however, that others are derived from infolded 

 portions of the outer layer of the glomus. They appear to have no 

 representatives in early stages of the organ. In my opinion, then, the 

 glomus is essentially a blood sinus, the wall of which projects into the 

 body cavity, carrying before it tlie peritoneal layer. 



The junction of the two aortic roots takes place very nearly opposite 

 the first nephrostome (compare Plate IV. Fig. 28, rx. ao.). The aortic 

 trunk thus formed (Fig. 10, ao.). since it occupies the space between the 

 chorda and mesentery, passes close to the attachment of each glomus. 



The precise relations of the aorta to the glomus are rather diffic\dt to 

 observe, since the former is peculiarly liable to injury in sectioning. 

 The interior of the chordn at this stage consists of a frail vesicular tissue, 

 whereas its outer sheath is tough, and resistant to cutting instruments. 

 In sectioning, therefore, it collapses, and occasions serious distortion of 

 the adjacent parts. 



In the younger individuals of this stage, the cavity of the aorta did 

 not seem to be sharply marked off from the root of the glomus ; in sev- 

 eral instances, indeed, I was able to observe a continuity between the 

 endothelium of the aorta and that lining the glomus. In older indi- 

 viduals I have repeatedly noticed distinct branches from the aorta pass- 

 ing into the glomus (Plate I. Fig. 10, va. sng.). These observations 

 were made, however, only on the most favorable sections, and I have 

 been unable to ascertain the number or distribution of such branches. 

 In both of the two most obvious cases, however, the vessel entered the 

 hinder end of the glomus. Occasionally, the veissel to the glomus seems 

 to be only a lateral branch given off from a vessel which can be traced 

 betw-een entoderm and splanchnopleure for some distance ventral to the 

 glomus (Fig. 10, va. sng., the lower of the two dotted lines). 



