WITH NOTES ON THE WEST INDIAN SPECIES. 283 



may be still further defined as occurring at the level of the anterior dilated end of 

 the ventro-lateral glandless epidermal tract shown in Figures 9 B and 9 C on PI. XXVII. 



Similar pores, in varying numbers, occur in the same region in Bal. meresch- 

 kowskii (Schimkewitsch), Bal. koiualevskii (Spengel), Sch. brasiliense (Spengel) and 

 Gl. hacksi (Spengel). In addition to these paired pores belonging to the posterior 

 end of the genital region Spengel has described a number of similar structures 

 following close behind the branchial region, i.e. at the anterior end of the genital 

 region. These anterior pores may be either paired or unpaired and have been found 

 bv Spengel in Sell, brasiliense, Gl. hacksi and Gl. talaboti. 



Although generally limited in distribution, in the last-named species they occur 

 in 9 groups .liMributed at unequal intervals over the anterior four-fifths of the long 

 genital region. 



In So. alba the intestinal canals are not provided with a special ring-shaped 

 thickening of basement-membrane such as Spengel has described in Bal. koiualevskii, 

 nor with a sphincter muscle such as occurs in Sch. brasiliense. I have not found 

 any pores other than those here described at the posterior end of the genital region. 



Having now become personally acquainted with these remarkable structures, I 

 agree with Schimkewitsch 1 , who was the first to record their existence, in regarding 

 them as vestigial gill-clefts (see below p. 298). 



The hepatic region follows immediately behind the intestinal canals. It is charac- 

 terised by the presence of internal hepatic saccules having essentially the same 

 topographical relations to the wall of the gut as the hepatic saccules of the Ptycho- 

 deridae and Schizocardium, but they are quite internal and are not associated with 

 permanent external sacculation of the body-wall. This is all the more striking because 

 the internal saccules are of large size and perfectly definite; a fact which serves to 

 distinguish this species, and perhaps the genus, from other Enteropneusta. 



Although no dermal elevation (or at most a slight arching) accompanies the 

 hepatic saccules we nevertheless find remarkable intersaccular involutions of the 

 epidermis, the walls of which sometimes present complicated corrugations (PL XXXII. 



Fig. 58). 



The occurrence of these intersaccular epidermal involutions would seem to indicate 

 that the external hepatic saccules of the Ptychoderidae are not merely due to the 

 mechanical effect of the hepatic diverticula causing elevations of the skin but have a 

 more fundamental physiological causation. Because it seems probable that the invo- 

 lutions in question are related to the intersaccular, i.e. interannular intervals in the 

 Ptychoderidae. 



If, as I believe, the Ptychoderidae are relatively primitive, we ought to find 

 vestiges of their hepatic saccules in less primitive forms, and from this point of view 

 we might regard the internal saccules and intersaccular involutions of Spengelia as 



such vestiges. 



The ventro-lateral epidermal tract shown in Plate XXVII. Fig. 9 B, 9 c has been referred 

 to above. It is characterised in general by the absence of gland cells and by low 

 cubical or flattened ectoderm similar to that which forms the floor of the dermal 



1 W. Schimkewitsch. -'Uber Balanoglossus raereschkowskii Wagner," Zool. Am. xi. 1888, p. 280. 



