322 Mr. C T. Regan on the Osteology and 



the same conclusion in the case of P(jecilia and Anubleps 

 (Hist. Nat. Poiss. xviii. I8i6, pp. 126, 258), and Philippi 

 has recently shown that Glaridichthys and Cnesterodon are 

 phvsoclists (Sitzungsb. Gesellsch. uatuif. Freund. 1906, 

 p. 232). 



Starks lias published some notes on the osteology of 

 Ambli/opsis, Cy/jrinodon, Fundulus, and Pcecilia (Biol. Bull, 

 vii. 1904). 



Suborder 1. Amblyopsoidea. 



Mouth rather wide, not protractile ; teeth in jaws and on 

 palatines ; gill-membranes attached to istlimus ; pelvic fins 

 small or absent; vent jugular. Parietals large; opisthotic 

 normal. Palatine distinct from pterygoid ; metapterygoid 

 present. No post-cleirhrura. Pracaudal vertebrae with 

 parapophyses from the fourth ; firot two ribs sessile. 



Family Amblyopsidse. 



This family includes Cliologasttr and the blind cave-fishes 

 Amhlyopsis and TypkUchthys [cf. Jord. & Everm. Bull. U.S. 

 Nat. Mus. xlvii. 1896, p. 702). I am able to confirm Stark's 

 account of the skeleton of Amhlyopsis (Biol. Bull. vii. 1904, 

 p. 261), except for the statement that all the praicaudal 

 vertebrae have parapophyses. Eigenmann lias studied the 

 degeneration of the eyes in the blind cave-fishes (' Cave 

 Vertebrates of America ' : Washington, 1909). 



Suborder 2. PcECILIOIDEA. 



Mouth small, unless the jaws are produced ; prgsmaxillaries 

 usually protractile ; maxillary little movable, adnate to pr£e- 

 orbita! ; teeth in jaws and sometimes on vomer, never on 

 palatines ; gill-membranes free from isthmus ; pelvic fins 

 usually well-developed, 6-rayed, sometimes absent. Parietals, 

 when present, small or moderate ; opisthotic, if present, 

 adherent to outer surface of exoccipital. Palatine and ptery- 

 goid ankylosed ; no metapterygoid, A single laminar post- 

 cleithrum on each side (except in Anahleps). Prgecaudal 

 vertebrae with parapophyses from the first or second; all the 

 ribs on parapoptiyses. 



The fishes of this suborder have usually been regarded as 

 a single family, Cyprinodontidse or Poeciliidfe ; in that case 

 several subfamilies should be reco^^fuized. 



