WITH NOTES ON THE WEST INDIAN SPECIES. 273 



extension, commencing a short distance behind the region of bifurcation of the nuchal 

 skeleton ; it is a much folded membrane containing blood-spaces. 



The collar-canals have the usual semilunar funnel opening into the posterior dorsal 

 portion of the collar cavity. The ciliated columnar epithelium of the canals has a 

 striated inner portion free from nuclei, and a basal two-thirds, with densely crowded 

 nuclei which stain nearly black with ordinary haematoxylin. The dorsal wall of the 

 canals is lightly plicated, but there is no definite tongue-like fold such as occurs in 

 many other species (PI. XXXI. Fig. 42). 



Perihaemal Cavities and Canals. 



The perihaemal • cavities, as is known, are prolongations from the truncal coelom 

 into the collar. In Spengelia, as in all Enteropneusta except the Ptychoderidae, they 

 contain transverse muscles below the longitudinal muscles. Their most noteworthy 

 feature in S. porosa, however, is due to the presence of a pair of canals, analogous 

 to the collar canals, which open like the latter, into the first pair of gill-pouches from 

 which they are derived (PI. XXXI. Fig. 44). These structures also occur in S. alba, 

 n. sp. In the present species they lie imbedded in the spongy connective-tissue which 

 is abundant in the posterior region of the perihaemal cavities. They are long canals, 

 and perfectly definite, and I knew of their existence long before realising their probable 

 significance. Their epithelium is of a spongy nature, and contains mucous cells. The 

 lumen is to a large extent occluded in my preparations, so that it is not possible to 

 assert positively that these canals open into the perihaemal cavities. The latter are 

 cavities only in name, being filled up by muscular and connective tissue, so that there 

 could hardly be an effective opening into the cavities. Nevertheless in the following 

 species which I have to describe, the conditions are more favourable for observation, 

 and there is more reason to suspect the existence of an internal pore in that case. 



In the present species, at least in the adult, my impression is that these structures 

 are not of great functional importance, they are in fact vestiges of a former condition of 

 which we know nothing definitely. I regard them as truncal pores homodynamous with 

 the collar pores and the proboscis pores, and the true homologues of the atrio-coelomic 

 funnels (brown funnels) described by Lankester in Amphioxus. These latter structures 

 have fallen into desuetude phyletically, since the evolution, and historically since Boveri's 

 discovery, of the nephric tubules'. 



Splanchnic layer of Nerve-fibres. 



There is a well-defined layer of " Punktsubstanz " at the base of the throat 

 epithelium. It is thicker in front than behind. It also occurs at the base of the 

 epithelium of the oesophageal portion of the pharynx. 



Spengel has also described such "Punktsubstanz" in places where one might not 

 have expected to find it. I have seen it at the base of the branchial epithelium of 

 the septal bars in Pt. flava. 



1 For further remarks, see below p. 310 et seq. 



38—2 



