64 U.S. NATIONAL MUSEUM BULLETEST 282, 



Sojt pectoral rays: (a) 126 (8-9) 8.62; (b) 68 (8-10) 8.97; (c) 102 

 (8-10) 8.74; (d) 94 (8-9) 8.89; (e) 390 (8-10) 8.78. 



Lower-half caudal rays: (a) 63 (21-26) 23.48; (b) 34 (22-26) 

 24.74; (c) 52 (22-27) 24.48; (d) 47 (21-26) 24.23; (e) 196 (21-27) 

 24.14. 



Upper-half caudal rays: (a) 63 (22-28) 24.81; (b) 34 (25-29) 

 26.50; (c) 52 (24-31) 27.52; (d) 47 (25-28) 26.91; (e) 196 (22-31) 

 26.33. 



Vertebrae seem to decrease in number northward. The upper 

 Mississippi and Missouri basin samples are low in number. The counts 

 from the White River system are highest: Iowa, Wisconsin, and Illinois 

 River, IlUnois 8 (36-38) 37.13; Missouri River drainage, Missouri 

 and Kansas 36 (36-39) 38.11; Arkansas River drainage, Oklahoma 

 and Arkansas 32 (38-40) 38.88; White River system 24 (39-41) 

 40.13; Tennessee River system 13 (39-40) 39.38; total 113 (36-41) 

 38.83. 



Internasal pores: A sample from Hickory Creek, Illinois has 9 of 

 70 circumorbital canals unconnected anteriorly (two internasal pores) ; 

 the other 61 are connected. Otherwise the species throughout its 

 range has a single internasal pore. This characteristic was utilized 

 as one of the identification checks for nearly all the lots examined. 

 Of the samples tabulated at random from areas other than Hickory 

 Creek, 466 of 471 canals have a single internasal pore. This variation 

 is not reflected in the number of preoperculomandibular pores (table 

 22). 



Variations in caudal ray and anal ray counts are given in tables 4 

 and 5. Note especially the greater number of anal rays in the White 

 River population, and the low number of caudal rays in the Arkansas 

 River system. 



Distribution. — The range (map 2) of Noturus exUis, a creek species, 

 may have been broken very recently into two or more populations, 

 one in the Ozarks and northward and one in the Tennessee and 

 Cumberland River basins. It was recorded by Goldsborough and Clark 

 (1908, p. 33) from the Guyandotte River, West Virginia [USNM 

 56615], on the basis of a single specimen, likely the result of a trans- 

 position of locality data. Dr. Frank J. Schwartz has failed to find 

 additional specimens in his survey of West Virginia fishes. Thus the 

 locality is now doubted, and not shown on the distribution map, 

 since other records (two from Indiana) from the Ohio drainage have 

 been re-identified as Noturus flavus, and presence in the Green River, 

 Kentucky (Charles, 1967, pp. 386-389) is doubtful and probably also 

 based on misidentifications of juvenile N. flavus. It does not seem 

 likely that A'', exilis has occurred in the Ohio basin within historical 

 times. 



