256 EMBRYOLOGY OF TEREBRATULINA. 



the division between the two segments. The alimentary tract is seen here, its anterior por- 

 tion blending with the cell contents of the cephalic ring, while posteriorly it hangs in the 

 cavity supposed to be the incipient peduncular segment. Whether this is the proper inter- 

 pretation I cannot say, and the difficulty arises from the fact that in some embryos farther 

 advanced this structure does not appear. Thus in Fig. 83, Plate ix., an embryo is shown in 

 which no sign of the peduncle appears, and yet it seems in advance of the one just 

 described, in showing the widening of the thoracic ring, one side of which shows a bulging 

 which represents the future ventral fold. An irregular cavity is thus left by the bulging 

 of the external wall, and in this cavity at a later stage I have seen the circulation of 

 minute granules. Within this cavity is another cavity bordered by thick walls which hang 

 from the suture dividing the two segments. The alimentary tract is seen suspended between 

 the two segments, and showing a faint connection with this cavity. The cephalic segment 

 shows no change in its granular contents save the presence of the alimentary tract. 

 The parts surrounding this tract are continually contracting in a spasmodic way, though no 

 trace of a muscular band could be detected. The interpretation of these cavities has per- 

 plexed me exceedingly. The most reasonable supposition is that the outside cavity in the 

 thoracic ring marked pi. represents the cavity from which the pallial sinuses are to arise ; in 

 this cavity I observed the circulation of minute granules, as before stated. The cavity 

 within surrounded by thick walls is the peduncular cavity, and a faint interspace running 

 from the upper part of this cavity to the base of the digestive tract appears to be the open- 

 ing up of the peduncular cavity into the cephalic segment, a cavity in which the stomach is 

 afterwards to be suspended, and this may be called the perivisceral cavity. In Lingula and 

 Discina the peduncular and perivisceral cavities are in open communication. In Discina the 

 large azygos opening is very apparent. Future study will undoubtedly modify the inter- 

 pretation of these cavities, and possibly show it to be altogether wrong. 



In the hundreds of embryos examined, a few only were of sufficient transparency to re- 

 veal any structure within, and from this stage onward, the difficulty in this respect increases, 

 as the embryo becomes more opaque, and with the after formation of the dorsal and ventral 

 plates, combined with their minute size, render the study of their internal structure almost 

 hopeless. Plate ix., Fig. 86, represents the peduncular portion just ready to protrude. 



Third Stage. In this stage the peduncular segment has made its appearance as a 

 rounded knob, varying in size in different embryos. Soon after the appearance of the 

 peduncle, granules are seen gathering about its end, apparently adhering to the mucus 

 secreted by it. Soon afterward adhesion takes place, the embryo remaining fixed when the 

 water is disturbed, though even before the embryo becomes attached, the bulging of the 

 thoracic ring becomes more conspicuous, leaving a clear cavity within. The walls of this 

 bulging portion present a number of highly refractive granules scattered evenly over its 

 surface. This portion first loses its cilia (Plate ix., Fig. 87) just before attachment ; the 

 embryo becomes sluggish, or at least, in most cases, ceases to swim actively. In two cases 

 a cluster of delicately barbed setae, to the number of 'thirty-five, was seen to project 

 directly backward from the posterior end (Plate ix., Fig. 79). That these were very tran- 

 sient is evident from the fact that they were only observed in two cases out of three hun- 

 dred and fifty-two different embryos examined. The head continues to move freely, and 

 shows a slight fold at its base. As the head contracts, the thoracic ring widens ; this is shown 

 in Plate i., Figs. 42, 43, where the same is sketched with the segments extended and 



