608 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 117 



The greatest problems in the classification of the Percidae lie in the 

 largest genus, Etheostoma. Twelve subgenera were utilized by Bailey 

 (in Bailey and Gosline, 1955) and one more recently has been recog- 

 nized (Collette and Yerger, 1962). 



The subgenus Boleosoma is composed of at least two groups: the 

 nontuberculate nigrum group of five species and the tuberculate 

 stigmaeum group of three species, although in some other characters, 

 chlorosomum of the latter group tends to be somewhat intermediate 

 between the two. The nontuberculate monotypic subgenus loa 

 clearly seems to be derived from the nontuberculate nigrum group. 

 Further study may lead to synonymizing loa under Boleosoma, 

 and removing E. stiginaewn, E. jessiae, and E. chlorosomum from 

 Boleosoma and placing them in Vaillantia Jordan. Similarities in 

 spawning behavior between loa and the nigrum group in contrast to 

 the stigmaeum group reinforce this idea. 



The subgenus Etheostoma can be divided into two closely related 

 tuberculate groups and one nontuberculate group. The variatum 

 group of five species is the onl}'^ one in the Etheostomatini where both 

 sexes are tuberculate. Other authors have previously pointed out the 

 close relations which members of the variatum group have to one 

 another. The inscriptum group of five species differs from the 

 variatum group in that only the males develop tubercles. Four of the 

 remaining five species (rupestre, histrio, zonale, and blennius) presently 

 placed in the subgenus form a closely related nontuberculate group. 

 This group appears to be somewhat intermediate between the variatum 

 and inscrijHum groups on the one hand and the nontuberculate 

 subgenus Ulocentra on the other. It will be necessary to await the 

 description of some nine new species of Ulocentra presently in manu- 

 script by Reeve M. Bailey before a full appraisal can be made of the 

 relations between the zonale group and Ulocentra. The two type 

 specimens of E. sellare are females caught in the spring; they are not 

 tuberculate. Therefore, it probably should not be placed in the 

 variatum group, although a number of authors have indicated that its 

 relationships are with that group. E. sellare does not api)ear to be close 

 to either the inscriptum or zonale groups and I am not satisfied that it 

 belongs in the subgenus Etheostoma. 



The monotypic subgenera Allohistium and Austroperca are ap- 

 parently nontuberculate. Psychromaster is also apparently non- 

 tuberculate. The lack of tubercles in all species of Nothonotus and 

 Catonotus further indicates the compact nature of each of these subgenera. 



The largest subgenus, Oligocephalus, contains about 20 species 

 which belong in at least four species groups. The three forms of the 

 nianguae group are very closely related to each other as was noted by 

 Bailey (1948) and Kuehne and Bailey (1961). The radiosum group 

 is similar to the nianguae group in having tubercles on the body scales, 



