304 Mr. P. Schmklt on (lie 



teeth are still more enlarged and the series on eacli side 

 describes a strong- curve with its convex side toward tiie 

 median line." In JlipjuxiJossoides dtihiiis, after my descrip- 

 tion (Schmidt, lil()4:, }). 227), "the teeth are sharp, conical, 

 somewliat directed inward, in a single row; on the anterior 

 part ot" prcmaxillary 6 teeth of middle size and behind 

 them on each side a group of 2-3 large canines ; on the 

 posterior part are smaller teeth ; in the lower jaw are 4 

 canines on the symphysis.'" In Hijipoglossoides kalahurce, 

 after Snyder (1911, p. 54G), "teeth are small, slender, in a 

 single row on the sj-mphysis, where they are irreguhirlj 

 placed, enlarged, and somewhat canine-like/' 



Comparing all these descriptions, we see that the teeth o£ 

 the five species of Ilippof/lossoides are of the same general 

 structure. 'J'hey have all on the anterior part of the upper 

 jaws and on the symphysis of the lower jaw some enlarged, 

 curved, canine-like teeth, directed backward and set not so 

 close as other small conical teeth. 



The scales and their distribution are nearly identical in the 

 five species; nor do other morphological features give diffe- 

 rences sufficient for specific separation. 



This preliminary comparison of diagnoses has convinced 

 me that the five Pacific forms of Hippoglvssoides cannot be 

 regaided as well-defined sj^ecies. Two of them — Hippo- 

 glossoides hamilloni and H. katakurcc — I believe to be synonyms 

 of IJippor/lossoides rohustns and H. elassodoii. It is really 

 impocsible to find in what Ih'ppoglossoides hamihoni differs 

 from //. rohustns. The differeiices of four rays in the dor.'^al 

 and anal and of four pores in the lateral line is too insignifi- 

 cant, considering the high degree of variation of the fin-raj's 

 and pores in JJippoglostioldes {cf. Table II., p. 303). Jordan & 

 Evermann give as distinctive the length ot the pectoral — ^ the 

 length of head (=lo'8 ^/^ of total length) in Hippoghssoides 

 roiuhtus and ^ length of head ( = 16'3 °/q of total length) in 

 //. hamiltvni. But if we take into consideration that tlie 

 specimen of II. rohiisius was 318 mm. long and the specimen 

 of //. hann/toni ou\y 170 mm., and that, in general, yomiger 

 forms have compaiatively longer pectorals and caudals, we 

 shall understand that this difference cannot be regarded as 

 sufficient for separation. The wider interorbital space, the 

 smaller symphysial knob, the larger nasal tubes of Hippo- 

 ghssoides harniltoni are also features connected with youth, 

 and the roughness of the scales and the form of the anterior 

 jiart of the lateral line is highly variable in H ippoglossoides. 

 In the same manner it is impossible to distinguish IIippo- 

 glossoides katakurcr from £1. elassodon. The depth of the 



