A Revision of the genus Tarucus. 273 



selected as the types of the geiuis have both long har- 

 pagones (clasps), by no means typical of the bulk of the 

 genera, but in each case there are connecting links between 

 them. The two genera are evidently closely related. 

 They can, however, be separated by two characters; the 

 species of Tarucus have androconia (^. e. battledore scales), 

 none of the species of Castalius have them ; there are 

 only two or three that I have not had the opportunity of 

 examining, and I believe I am correct in thinking that 

 those I do not know are not likely to have these scales. 

 The second character is the genitalia, the general form of 

 the clasps is quite different, the aedoeagus differs also, 

 whilst the shape and position of the cingulum and tegumen 

 confirm the previous points; there is also no tectorius. 

 It will be seen that my conclusions have been brought 

 about in the first instance solely by my investigations of 

 the genitaha; these led to the necessity of re-grouping 

 most of the species and very many of the individual speci- 

 mens, and in so doing I have discovered other species, 

 evidently cpiite distinct, that had always hitherto passed 

 under other well-known names. Under these circum- 

 stances it was manifestly advisable to confirm these points, 

 and for this I turned to the androconial scales that are so 

 marked a character in the males of the majority of the 

 Lycaeninae; these amply confirm the correctness of each 

 of the specific identifications I had made, and they also 

 confirm the re-grouping of such species as ananda and 

 bowkeri, neither of which had hitherto been included in 

 the genus. The South Afiican species {bowkeri) is, I admit, 

 aberrant ; its pattern differs from the rest of its allies in 

 certain particulars, its clasps differ also, but the androconia 

 are so close to theophrastus that I have included it with 

 the others. In cases of this kind it appears to me to be 

 better to do this, indicating the small divergencies, rather 

 than to create a new genus for a single species that otherwise 

 groups itself very naturally with its close allies. 



Tarucus nara, Kollar. Plates XIV, figs. 1, la; 

 XV, fig. 1, and XIX, fig. 19. 



Lycaena nara, Kollar, Hiigel's " Kaschmir," iv, pt. 2, p. 

 421 (1848). 



The general consensus of opinion on the identification 

 of this species seems to me to be correct; it is almost as 



