CEsophajeal Teeth of the Stromateidie. 253 



supported mainly by the epibrancliial and pharyngo-brancliials 

 of the fourth gill-arch. Tlio toothed upper j)haryiigeal3 are 

 not rounded patches, but are in the form of two ridges whicli 

 project backwards and cud in pointed free extremities |)ro- 

 jectiiig into the opening of each sac. There are a few reduced 

 anterior gill-rakers on the fifth branchial arch, and tlio 

 opening of the oesophageal pouch is immediately behind the 

 gill-arch, so that there is a more apparent transition between 

 the gill-rakers and the oesophageal teeth, wliich might there- 

 fore be readily interpreted as the posterior gill-rakers of the 

 fifth branchial arch. These teeth are apparently of the same 

 nature as in the last case. 'J'here is a marked difference, 

 however, in their basal expansion, for they are firmly fixed 

 in the muscular wall of the pouch, not by a circular scale-like 

 structure, but by a number of root-like processes, which are 

 more or less curved at their pointed extremities (text-fig. 2). 



Psenes [Atimostoma, Smith ; Cabiceps, Giiuther) capensis. 



The oesophageal sacs of this species do not, as in P. nata- 

 lensis, lie parallel with the oesophagus, but more or less across 

 it, following the general contour of the branchial arches. 

 They thus project below the level of the cesophagus. They 

 are closer to the branchial arches, and are mainly supported 

 by the fourth. The most noteworthy difference is tiie 

 greater development of the toothed upper pharyngeals. 

 These, in the first arcli, are poorly developed, and have 

 no teeth, those of the second have a small patch of teeth, 

 those of the third are well developed, and those of the 

 fourth extend backwards as large toothed lobes into the 

 cesophagus, past the openings of the two oesophageal pouches. 

 The pouches are provided with rounded toothed lobes or 

 papilla similar to the pharyngo-brancliials of the fourth 

 arch. The long horny processes resembling gill-rakeis are 

 entirely absent in this species, and the inner lining of tlio 

 oesophageal sacs seems to be entirely derived from a backward 

 extension of the tooth-bearing epithelium of the superior 

 pharyngeals. This epithelium can readily be distinguished 

 from the oesophageal epithelium, which in this and some other 

 species of the Stromateidai is characterized by longitudinal 

 foldings. These extend forward on the floor of the oesophagus 

 between the sacs, and pass over the lip of the sac but not 

 into it. 



Nomeus gronovii. 

 In this fish the oesophageal sacs follow still more closely 



