280 ENTEROPNEUSTA FROM THE SOUTH PACIFIC, 



In the region of the collar funnels there is a pair of remarkable tubes lying 

 in the perihaemal cavities and opening like the collar canals into the first gill-pouch. 

 These are in fact perihaemal, i.e. truncal canals analogous to, and in all probability 

 homodynamous with the collar canals (PL XXXII. Fig. 52). 



As a rule, as mentioned above, the perihaemal spaces do not contain cavities 

 since they are quite filled up with muscular and connective tissue. In the present 

 species however, while the perihaemal tracts are solid in their anterior two-thirds, 

 posteriorly they develop a cavity which lies between the longitudinal and transverse 

 muscles of the perihaemal coelom (PL XXXII. Figs. 52 — 54). This in itself is an 

 interesting fact, but it becomes still more interesting when, on tracing the cavity back- 

 wards, the reason for its existence comes into view in the form of a genuine canalicular 

 extension of the first gill-pouch into the perihaemal coelom on each side. According to 

 Spengel, the collar canals themselves appear to arise as canalicular extensions of the 

 first gill-pouch, and, so far as I can gather, the observations of Bateson and Morgan 

 do not run counter to this view, in essentials. 



The truncal canals are smaller both in calibre and in extent tban the collar 

 canals and they are not provided with semilunar funnels, and I am not prepared to 

 assert positively that they open into the perihaemal cavities (see, however, PL XXXII. 

 Figs. 53 — 54). But in their quality of canals they are absolutely definite and so far 

 as is known are peculiar to the genus Spengelia. No truncal pores have hitherto 

 been described in Enteropneusta, and it is safe to add that none exist in previously 

 known species. 



That they existed at one epoch seem-; likely enough. We have already seen, in 

 the species described in this paper, how that vestiges of different structures crop up 

 in different species. 



One species may possess the vestige of a certain structure and another allied 

 species may be without it. It is not probable that truncal canals are essential to 

 one genus and non-existent in any other. 



For such reasons as the above I regard the truncal canals of Spengelia as being 

 functionally in a vestigial condition and comparable in this and in other respects with 

 the atrio-coelomic funnels described by Lankester in Amphioxus. 



The peripharyngeal spaces of S. alba commence anteriorly as in & porosa, but 

 they do not end blindly behind as in the latter, neither do they communicate with 

 each other across the ventral middle line ; each space passes separately into the 

 trunk coelom, as in Schizocardium (Spengel). 



The ventral septum only occurs in the posterior region of the collar, commencing 

 a short distance (nearly half a millimetre) in front of the posterior termination of 

 the cornua of the nuchal skeleton. 



TRUNK. 

 Branchial Region. 



The gill-pores of the first pair open coincidently with the posterior neuropore, 

 perforating the posterior commissure of the collar. 



