I will now proceed to the results obtained by investigation of the different 

 characters. It should here be mentioned, that all the measurements in connection 

 with the eePhave|been made by myself, the necessary calculations having been 

 undertaken by cand. mag. Aa. Strubberg and Frk. cand. mag. Kirstine Smith. 



II. The Characters investigated. 



1. Number of vertebrae. 



In Report I, 1913, the number of vertebrae has been exhaustively dealth with; 

 only, however, in the case of the European eel, the material available as regards 

 the other two species being then but scanty. The large sample of elvers from 

 W. Gloucester, Mass., U. S. A., referred to in the introduction, has therefore also 

 been employed for countings of vertebrae, with a view to obtaining a more com- 

 prehensive view of the variation in number of vertebrae for the American eel than 

 has hitherto been possible. A sample consisting of 502 specimens was taken, and 

 the number of vertebrae determined for each. The result will be seen in the 

 accompanying Table I, which further includes a summary of all countings of 

 vertebrae of eels from the American continent and the West Indies. It will be 

 been from the table, that the figure 111, which had not previously been found in 

 the case of the American eel, occours three times in the W. Gloucester sample, 

 which, however, lacks the figure 103 found once in the earlier material consisting 

 of 361 specimens. From all countings, of 863 specimens in all, we thus find 

 '{vide Table I, summary), that the number of vertebrae in the American eel varies 

 from 103 to 111, the figure of most frequent occurrence being 107, the next in 

 order being 108 and 106. 



The number of vertebrae is an excellent character by which to distinguish 

 between the two species of Anguilla which occur in the region of the Atlantic. An 

 idea of this may be obtained from Table II, showing the number of vertebrae in 

 the first 266 specimens of the large elver sample from W. Gloucester, Mass., U.S.A. 

 as compared with an equally large sample of elvers from Denmark examined for 

 the purpose. From this we see, that only one figure is common to both species, 

 viz.; 111. This occurs but three times in the total of 532 specimens contained in 

 both samples together. This is to say, that a mixed sample of Anguilla vulgaris 

 and A. rostrata, amounting to 532 individuals in all, would contain but 0.56 7olof 

 specimens which could not be distinguished with perfect certainty, which again 

 practically means that even single specimens of either species may always be 

 classified by the number of vertebrae. As will subsequently be seen, this is 

 generally impossible with other characters, even though the averages for these, as 

 determined by examination of numerous specimens may differ greatly for the 

 two species. 



