374 T - D - A - COCKERELL. 



these scales there is much variation, which is at times confusing, 

 although it usually does not affect the fundamental character 

 of the pattern. Thus I described the scale of Chondrostoma soetta, 

 from Italy, as having few apical radii, distinguishing it thus from 

 certain Spanish species. Noting later that this did not agree 

 well with Fatio's figure, I asked Mr. Regan to look into the matter 

 and he found that on the same fish immediately adjacent scales 

 had many and few radii respectively. Consequently my diag- 

 nostic character proves of no value, but it will be noted that the 

 variation affects the development rather than the character of 

 the markings. Greater differences are noted when the scales 

 are taken for different parts of the body, and for purposes of 

 comparison I always take them from near the middle of the side, 

 close to the lateral line. Fig. 28 and 29 show an extreme case, 

 that of the curious little fish Ericymba buccata, from Indiana. 

 Fig. 28 is a normal scale, and Fig. 29 is from the subdorsal region 

 about 5 mm. in front of the dorsal fin. These two scales are 

 from the same fish. 



Occasionally very marked abnormalities occur. Thus on a 

 specimen of Ostebrama fecB I found a greatly enlarged scale on 

 the middle of the side, below the lateral line. It was about 8.5 

 mm. long, six times as long as the exposed parts of the normal 

 scales, but not differing essentially in sculpture The age of 

 the fish is also often of importance. Fig. 30 shows a scale of 

 Gila robusta from Arizona. It was taken from a young fish, 

 and for some time I did not know that the adult scales of Gila 

 have a characteristic basal lobe, as indeed was figured long ago 

 by Girard in the report of the Mexican Boundary Survey. 



There is an interesting case outstanding, in which the value 

 of scale characters for the separation of very closely allied fishes 

 has yet to be determined. From Stanford University I received 

 a number of specimens of Myloleucus symmetricus , collected in 

 different localities. These were sent principally because Dr. J. O. 

 Snyder had already noticed the variability of this fish in other 

 than scale characters, and it was suspected that it might include 

 some separable forms. A study of the scales revealed four ap- 

 parently distinct types, (i) Scale broad, small. Navarro 

 River, Gualala River (California) and north to Drew Creek in 



