CLASSIFICATION OF THE CEPHALOCIIORDA. 271 



the same time recorded B. lanceolatum from the German 

 Ocean ; and in the next year 1 Tie introduced the system 

 of enumeration of the myotomes as a means of dis- 

 tinguishing the species of the group, and on this new 

 character he diagnosed the 4 species already known, as 

 follows, using three numbers to indicate (1) the myotomes 

 in front of the atriopore, (2) those between atriopore 

 and anus, and (3) those behind the anus: — 



B. lanceolatum, with the myo. form., 36, 14, 11. 



B. elongatum, with the myo. form., 49, 18, 12. 



B. belcheri (formula not then known). 



B. caribbceum (a new T form he describes from Rio 

 Janeiro), 37, 14, 9. 



The difference between the myotome formula? of B. 

 lanceolatum and B. caribbo?um appears very slight indeed 

 from Sundevall's figures, and it leads us to think that 

 he either regarded the formula as perfectly constant or 

 that he had few specimens to work with, as the examin- 

 ation of a large number of specimens of each species would 

 have convinced him of the great variation that exists 

 within the limits of a single species. These forms were 

 not at once regarded as specifically distinct, for GHinther 

 in 1870 2 classed them all as B. lanceolatum. 



In 1876 Peters 3 formed a new genus for specimens from 

 Torres Strait, Australia, which he called Epigonichthys 

 cultellus, basing his genus on (1) the supposed median 

 position of the vent, and (2) the shape of the fins. The 

 fins, however, are such delicate structures, and, moreover, 

 vary so within the limits of a single species that their 

 value in determining genera is very slight indeed. Only 

 in the best preserved material would they retain their 



ilbid. Vol. x. 1853. 



2 Cat. Fish. Brit. Mus. 1870. Vol. vii. 



3 Monat. K. Preuss. Akad. 1876. 



