556 CEPHALOCHORDA. 



While the above changes are taking place in the branchial region, 

 the number of gill-clefts is doubled, each cleft being cut up into two 

 halves by the down-growth of a conical process from its dorsal edge 

 (Fig. 293). So long as this process (the rudiment of the secondary 

 or tongue-bar) does not reach the lower edge of the cleft, the latter 

 is horse -shoe-shaped, recalling in appearance the gill-clefts of Bnlano- 

 gloszus. The process finally fuses with a prominence which rises to 

 meet it from the lower edge and each gill-cleft is thus divided into 

 two. In consequence of the clefts having developed in this way, we 

 are able to distinguish in the adult alternate series of primary and 

 secondary gill-bars, between which an essential anatomical distinction 

 exists, as has recently been pointed out by RAY LANKESTER (No. 1*2), 

 SPENGEL (N"o. 19), and BOVEKI (No. 2).* Here we can only mention 

 that the primary bars enclose a coelomic canal which is wanting in 

 the tongue-bars. The cross-bars or synapticula (Fig. 312, s), running 

 in an obliquely horizontal direction from one primary bar to the next, 

 only develop comparatively late. 



While the changes just described are gradually leading to the 

 perfection of the pharynx, a cavity develops round this region, the 

 peribranchial or atrial cavity. In accordance with KOWALEVSKY and 

 ROLPH, this was formerly thought to arise through two longitudinal 

 folds growing completely over the region of the branchial clefts like 

 the branchiostegite in the Crustacea and fusing in the ventral median 

 line so that, finally, of the original wide aperture between the 

 peribranchial folds only a small median postero-ventral opening is 

 left. This aperture, the atriopore, serves to put the atrial cavity 

 into communication with the exterior. Our view of the development 

 of the atrial cavity has, however, recently been modified through the 

 researches of RAY LANKESTER and WILLEY. The first rudiment of 

 this cavity is, indeed, found in the form of two folds (Fig. 294 .4), 

 known as the hdcral or metapleural folds (If and rf). Within these 

 folds a cavity develops (Fig. 311, of, p. 572), which, according to 

 KOWALEVSKY, represents an isolated part of the body-cavity. This 

 cavity, which is known as the metapleural canal (HATSCHEK'S 

 Oberfaltenhoh/.e), is not reckoned as belonging to the coelom by RAY 

 LANKESTER and WILLEY, but is regarded as a lymph-sinus (pseudo- 

 coele). According to HATSCHEK, on the contrary, it should be 

 considered to belong to the myocoele. 



*[See BENHAM'S more recent work, No. I. ED.] 



