GEOMETRIDAE. 



141 



filled ; postmedian wavy, especially anteriorly, slightly oblique outward and 

 generally thickened (especially between the radials) from costa to R 3 , here with 

 a pronounced, though generally rather rounded elbow, posteriorly more puncti- 

 form, about parallel with termen ; distal area always with a very noticeable, 

 though not sharply defined, pale longitudinal streak about the middle, more 

 or less completely interrupting the subterminal markings and sometimes almost 

 reaching the termen. Hind wing with termen strongly convex, almost regularly 

 curved from apex to near tornus, only behind M 2 straighter or very faintly 

 incurved ; SC 2 stalked, but generally shortly, not rarely (at least in the 2) 

 almost connate ; postmedian line strong, generally thickened, with proximal 

 black vein-dashes, a more or less V-shaped indentation between the radials, 

 generally produced proximad by a longer black dash which converts it into a 

 Y, a rather variable angle outward at R 3 ; pale subterminal mark weaker than 

 on fore wing, shorter (macular), not rarely more or less obsolete. Under side 

 weakly marked. 



Most of the Samoan series are darker, or at least more brownish, than 

 the pale greenish-grey G. concinna, from which one gets the best conception of 

 the Tonga form, so far as at present known ; but one Malololelei $ approaches 

 that form quite closely. One has the wings much clouded, especially in the 

 median area of the fore wing (unfortunately discoloured at the base). A $ has 

 the median area of the fore wing and the narrow band beyond almost clear 

 white, confluent with the distal pale streak, which is also whitened, while the 

 rest of the terminal area, together with the basal, is dark ; the hind wing dark, 

 with narrow white band outside the postmedian and small whitish mid-sub - 

 terminal spot. The only taken 25.ii.1924, is small, attenuated and pale, but 

 not in perfect condition. 



G. semipurpurea Rebel was founded on 2 sent by Friedrichs, hence pro- 

 bably from Apia, although only " Samoa " is given. Notwithstanding the 

 close agreement of Rebel's description with one of Buxton's Upolu forms, I 

 have felt constrained to add a query to the citation because its eminent and 

 experienced author has definitely differentiated his G. semipurpurea from 

 G. upolensis $ by the " non-dentate " condition of the antenna. In the present 

 species — as in nearly all the allied Gymnoscelis material that I can examine — 

 the segments in both sexes bear at their distal ends bands of more or less pro- 

 jecting scales, giving to the proximal part of the antenna a vague suggestion 

 of an Equisetum, but tending progressively to project at a steeper angle distally, 



