32 PROCEEDINGS OF UNITED STlTES NATIONAL MUSEUM. 



0.12 to 0.145 ; that of mandible from 0.15 to 0.18. The diameter of the 

 eye enters 4^ times in the length of the head. Its width varies from 

 0.11 to 0.15 in very fat individuals. 



Shape of Body. — This is exceedingly variable, and the variation is 

 caused largely by the fatness of the individual. In very plump ones, 

 the expansion of the belly throws back the origin of the ventrals and 

 anal, and greatly changes the appearance of the fish. In the specimens 

 before me, the height of the body ranges from 0.31 to 0.38i. The table 

 of measurements subjoined shows the effect of increased height of body 

 ui)on the other measurements of proportion. 



Fins. — The range of variation in the position of the dorsal is indi- 

 cated in the diagnosis. There is no appreciable correlation between the 

 positions of the dorsal and anal in the same specimen. The insertion 

 of the anal is distant from the snout from 0.C8 to 0.75. The lengths of 

 the rays in dorsal, anal, ventral, and caudal vary much, as the table of 

 measurements indicates. In the caudal, the upper lobes vary from 0.16 

 to 0.25; the lower lobes from 0.18 to 0.27. The relation of the pectoral 

 and ventral fins is much affected by the length of the head, the inser- 

 tion of the former being thrown much further back in long-headed indi- 

 viduals. 



Scales. — The degree of serration varies much in individuals as well as 

 the squamation of the bases of the vertical fins and the number and 

 regularity of the body-scales. In young individuals, the scales are 

 arranged with much regularity; but, in adults, I have strong reason to 

 believe that scales are intercalated here and there, throwing the arrange- 

 ment into great disorder, and rendering an accurate enumeration imijos- 

 sible. 



Varieties. — The series before me embraces some two hundred specimens 

 of Brevoortia tyranniisoi various ages, seasons, and localities. Almost 

 every feature is subject to wide variations, and there is usually no decided 

 correlation between different characters except that a long head is accom- 

 panied usually by long jaws, and a pectoral set farther back and extend- 

 ing more nearly to the insertion of the ventral. There are, however, 

 certain groups of individuals which can be included within a diagnosis 

 which may serve to distinguish them from all the others of the same 

 species. To what extent it is desirable to define varieties which are 

 not separated geographically, I am not well satisfied. The exact mean- 

 ing of the terms " sub-species" and "variety" as employed by Cope, 

 Cones, Gill, Yarrow, and other recent writers has not been definitely 

 interpreted. It seems desirable, however, to designate in some way 

 the limits of variation from the normal specific type in different direc- 

 tions. With this purpose, and premising that by a variety I mean sim- 

 ply a divergent form, connected by intermediate forms with the typical 

 specific form, I have thought it desirable to name provisionally two 

 varieties, and to call attention to others which may possibly exist. This 

 is done with much hesitation, and only with a view to an attempt to 



