SIMULIIDAE OF AFRICA 59 



(with the exception of rivuli) there are more than four filaments (6, 10, or 12) in the 

 pupal gill ; this group clearly corresponds to Rubzov's annulum-group, and there- 

 fore has no equivalent in the Ethiopian and Malagasy fauna. One of its species, 

 S.(E.) cougar eenarum, is apparently very closely allied to S.(E.) dogieli (Ussova) from 

 Karelia, which Rubzov (1959-1964) in his Palaearctic monograph first placed in 

 Eusimulium but later (p. 591, op. cit.) transferred to the genus Greniera : this re- 

 assignment, whether justified or not, emphasizes the fact that many of the more 

 primitive Holarctic forms placed in the subgenus Eusimulium are so similar in many 

 of their characteristics to certain Prosimuliine genera that it is even doubtful 

 whether they should be placed in the genus Simulium at all. In the Ethiopian and 

 Malagasy Regions such difficult intermediate forms do not occur, and the Pro- 

 simuliine and Simuliine faunas of these areas are readily differentiated. 



The batoense-gvoup of Eusimulium, delimited by Rubzov for a small number of 

 species from Japan, contains some forms in which the katepisternum is fully haired 

 (such as batoense Edwards itself) and others in which it is bare. The species showing 

 the former character appear to have all the diagnostic features of the Oriental sub- 

 genus Gomphostilbia, and have been assigned to this subgenus in an earlier paper 

 (Crosskey, 1967a : 38). The remaining Japanese species in the batoense-group in 

 Rubzov's sense have a gill with four or six filaments, multiple parameral hooks, and 

 the katepisternum bare : these species (S.(E.) yamayaense Ogata & Sasa, S.(E.) mie 

 Ogata & Sasa, and S.(E.) sasai Rubzov) are assignable to Eusimulium, not to 

 Gomphostilbia, and appear to show some affinity with S.(E.) euryadminiculum 

 Davies or S.(E.) furculatum (Shewell) from the Nearctic Region. They also have 

 a striking resemblance to the loutetense-group from the Ethiopian Region (defined 

 below). 



The African fauna contains no species that fit with the rather weakly defined 

 montium-group and alpinum-group of Rubzov. 



In the present work four species-groups are recognized for the fauna of the African 

 area and can be distinguished by the following key. 



Key to the Species-Groups of EUSIMULIUM in Africa and its Islands 



1 Ventral plate of ^ with small subtriangular body and large outwardly directed basal 



arms (Text-fig. 102) ; style very small in relation to coxite and with character- 

 istic shape (Text-fig. 142). Postnotum with scale patch on either side. Cocoon 

 simple ......... aureum- group (p. 65) 



- Ventral plate of q* with large transverse lamellate body and small forwardly directed 



basal arms (Text-figs. 100, 101, 103); style large, not of this shape. Postnotum 

 bare (except in occasional specimens of latipes-group) . Cocoon simple or with 

 triangular or long horn-like anteromedian process ...... 2 



2 a* hypopygium with one very long strong parameral hook on each side (Text-fig. 88). 



Postgenal cleft of larva almost absent or small rounded or subquadrate, subequal in 

 length to or much shorter than postgenal bridge (Text-figs. 241-244). Cocoon 

 without neck ............ 3 



- o* hypopygium with several (usually 3-6) parameral hooks on each side. Postgenal 



cleft of larva large and mitre-shaped (Text-fig. 245), very much longer than post- 

 genal bridge. Cocoon with or without neck . . loutetense-group (p. 63) 



