THE SCALES OF FRESHWATER FISHES. 369 



Sauvage on the fishes of Madagascar, containing numerous figures 

 of scales of Acanthopterygian fishes. In other writings, here 

 and there, are various figures of scales, usually with little dis- 

 cussion. The scales of Gadus (codfish and allies) have been 

 beautifully figured and fully discussed by Dr. H. W. M. Tims 



(I905)- 1 



Dr. Jordan, with great kindness, sent me a fine series of fresh- 

 water fishes from the collections of Stanford University, while 

 Dr. Evermann was equally good in supplying numerous species 

 from the Bureau of Fisheries. I thus obtained nearly all of the 

 principal forms of North American Cyprinidae or carp-like 

 fishes. The following summer I visited the British Museum, and 

 was indebted to Dr. Boulenger for the opportunity of investi- 

 gating the scales of nearly all the principal old-world cyprinids, 

 and a like series of African characinids. I also obtained through 

 him many African fishes of other families. Quite recently I 

 have received from Dr. Eigenmann, of the University of Indiana, 

 an extremely fine series of South American characinids. With 

 all these, and some others, it has been possible to test rather 

 thoroughly the value of scale-characters, and the result has been 

 to show that while they are not rarely deceptive, through con- 

 vergence, they are on the whole of great taxonomic importance. 

 As in most other taxonomic work, there are disturbing elements 

 due to individual variability and differences of age, but while these 

 are sure to lead to various minor errors, they will not much affect 

 the broader results. The key to the origin of the sculpture of a 

 teleostean scale is apparently to be found in that ancient type 

 the bowfin of North America, Amia calva. Fig. I shows part of 

 the base of a scale of this fish, which it w r ill be observed consists 

 of longitudinal strands or fibres, separable elements which fray 

 out basally. 2 In the apical field these are directed toward a 

 broad rough nuclear area. A close approximation to this is found 

 in a very old type of teleosteans, the lady-fish, Albula. In this, 

 however, appear also the beginnings of the radial lines, extending 

 from the nucleus of the scale to the margin. As we go higher 



1 Dr. B. L. Chaudhuri informs me that Dr. John McClelland published an account 

 of the scales of Indian Cyprinidae in the appendix to his work on these fishes, in 

 1839. This I have not seen. 



2 See Smithsonian Misc. Coll., Vol. 56, No. 3, p. 2, 1910. 



