12 D. S. FLETCHER 



15 (14) Arm of sacculus bluntly tapered and sinuous (Text-fig. 7) ; wing-span of moth 



usually 30-35 mm. ; endemic on Comoro Is. . . transversaria (p. 18) 



Arm of sacculus finely tapered ; wing-span of moth usually 40-45 mm. ; not 



found on Comoro Is. . . . . . . . .16 



16 (15) Aedeagus six times as long as broad, one and one-half times as long as dorsal 



margin of valve (Text-fig. 14) .... flavivenata (p. 22) 



Aedeagus almost eight times as long as broad, twice as long asdorsal margin of 



valve (Text-fig. 11) . . . . . . betularia (p. 20) 



17 (8) Arm of sacculus of almost even width, slightly spiral and extending parallel 



with and almost to apex of folded ventral margin of valve ; vesica without 

 cornuti (Text-figs. 41, 42) ...... carcassoni (p. 36) 



- Sacculus not so formed . . . . . . . .18 



18 (17) Arm of sacculus slender and spatulate, dorsal margin partly fused with dorsal 



margin of valve, ventral margin serrate and parallel with ventral margin of 

 valve ; long, very slender digitate process at base of sacculus (Text-figs. 47, 



48) 19 



Sacculus not so formed ........ 20 



19 (18) Serrate-edged process on sacculus one and one-half times as long as greatest 



width of valve (Text-fig. 47) ; endemic on Sao Thome I. 



subcincta subcincta (p. 43) 

 Serrate-edged process on sacculus twice as long as greatest width of valve 



(Text-fig. 48) ; continental Africa. . . subcincta longifibulata (p. 43) 



20 (18) Arm of sacculus extending parallel with apical half of ventral margin of valve, 



rasp-like, being densely and coarsely spined for full length of dorsal surface ; 

 a short, blunt digitate process at base of sacculus (Text-fig. 44) . . 21 



Sacculus not so formed ........ 22 



21 (20) Vesica with two cornuti (Text-fig. 45) .... herbuloti (p. 38) 



- Vesica without cornuti ...... acerata (p. 41) 



22 (20) Sacculus with one or more well sclerotized, tapered processes, each as long as, 



or longer than uncus, the apices curving through 90 ° . . . .23 



- Sacculus not so formed ........ 24 



23 (22) Sacculus with two tapered processes, each smooth to tip ; vesica with one 



cornutus equal in length to width of aedeagus (Text-figs. 72, 73) bicornis (p. 57) 



- Sacculus with one tapered process, the apex serrate-edged ; vesica with two 



stout, tapered cornuti fused at base, each two-thirds as long as aedeagus 

 (Text-figs. 66, 67) ...... quadrimaculata (p. 53) 



24 (22) Cucullus almost circular and projecting beyond ventral margin of valve ; 



sacculus with a curved, tapered process arising from near mid-valve, with or 

 without additional processes ....... 25 



Cucullus and sacculus not so formed ...... 26 



25 (24) Sacculus with a short truncate process at one-third ventral margin of valve ; 



vesica with one long and one short cornutus (Text-figs. 25, 26) ; endemic 



on Principe I. ....... tamsi (p. 28) 



- Sacculus smooth along ventral margin and without additional process ; vesica 



with one tapered cornutus (Text-figs. 28, 29) ; endemic on Sao Thome I. 



viettei (p. 30) 



26 (24) Scobinate medial plate of gnathus broader than greatest width of cucullus and 



with two scobinate arms projecting dorsally, one at each side ; process 

 from sacculus extends along mid-valve and is tipped with stout spines, 

 varying in number from 1-13 ; vesica with two cornuti, one scobinate, one 

 tapered and smooth (Text-figs. 51, 52) .... lacrymata (p. 44) 

 Genitalia not so formed . . . . . . . .27 



