THE SHORE FISHES OF PERU 33 



According to the specimens studied, M. dorsalis and M. mento 

 differ constantly only in the structure of the teeth and in color. In 

 mento the individual tooth has an almost smooth rounded margin, 

 whereas in dorsalis it has a prominent sharp cusp. Although the 

 teeth vary somewhat in shape according to the position in the mouth, 

 the differences pointed out are readily recognizable in any of them. 

 The teeth of these species, as well as those of maculatus, are well 

 described and figured by Bigelow and Schroeder (1940, p. 428, pi. 17). 

 The dermal denticles as figured by these authors seem to be distinc- 

 tive. However, a further study of these structures from specimens 

 at hand reveals so much variation that I am now unable to point 

 out an absolutely constant difference. In general, the denticle of 

 mento has a prominent median keel, and usually one and sometimes 

 two shorter and lower keels on each side of the median one. In 

 dorsalis a median keel sometimes is present, with a shorter and l<Mver 

 keel on each side. More usually, however, two keels of about equal 

 size occupy the middle of the denticle, just off the median line, and 

 a very short low keel occurs on each side of the pair. 



The color in dorsalis is plain gray above at all ages, as revealed by 

 embryos and adults up to 510 mm. in length. 



It has been thought and stated by some observers that M. mento 

 (or abhotti) and M. dorsalis differ in the proportionate length and width 

 of the snout, and in the position of some of the fins. However, so 

 much variation exists among specimens that intergradation obviously 

 occurs. This is especially true of Peruvian fishes. In specimens 

 from Panama, the only other place from which examples are available, 

 the snout actually is rather longer and narrower than in the Peruvian 

 specimens. The material now available is inadequate, however, to 

 show the significance. 



In several embryos of each species at hand, and in specimens up 

 to 390 mm. in length, the spiracle in mento is fully as long as the pupil, 

 and only about half that long in dorsalis. In the larger specimens it 

 is difficult, however, to demonstrate a difference. 



The diagnostic characters of dorsalis and related species are dis- 

 cussed in part by Meek and Hildebrand (1923, pp. 32-35) and by 

 Bigelow and Schroeder (1940, pp. 417-438). According to Beebe and 

 Tee-Van (see reference above) a length of about 90 cm. is attained. 



AI. dorsalis is represented in the present collection by two specimens, 

 370 and 390 mm. long, taken on a line trawl in Sechura Bay. Ten 

 embryos, each only about 60 mm. long, from Lobos de Tierra, col- 

 lected by R. E. Coker, from two females, each about 80 cm. long, are 

 also at hand. The proportions given in a preceding paragraph are 

 based on the two larger specimens. 



Range.— GwM of California to or beyond Callao, Peru. 



