144 



INSECTS OF SAMOA. 



C anastomosing with SC to near end of cell, SC 2 stalked with R 1 in $, coincident 

 in M 1 connate or just separate. 



Genotype : Pseudomimetis semiviridis (Warren) = Gymnoscelis coquina ab. 

 semiviridis Warren, 1897. 



Differs from Hybridoneara Warren in having the £ antenna not dentate- 

 fasciculate, in the less long stalking of SC 1 , in the hind wing of the $ not being 

 distorted, but with SC 2 coincident, the wings of the $ not unusually broad, 

 M 1 of the hind wing not stalked. 



The few Larentiids of the CJdoroclystis group which have the areole sup- 

 pressed (or vestigial *), leaving all five subcostals of the fore wing to arise from 

 a common stalk, can be referred to four genera, distinguishable as follows : 



1. Hind tibia with 3 spurs, SC 1 of fore wing free 2 



Hind tibia with 2 spurs, SC 1 of fore wing anastomosing with C . 3 



2. Hind wing of J much distorted, cost-ally with large brush of hair . . Mariaba, 

 Hind wing of $ nearly amygdaloid, not distorted, above with a roundish 



sex-patch beyond cell Onagrodes. 



3. Hind wing of <J triangular, wings of ° very broad .... Hybridoneura. 

 Wings normally shaped Pseudomimetis. 



Onagrodes contains two known species, and probably also a third (found in 

 S.-W. Sumatra), of which the $ has yet to be discovered. Though closely related, 

 the two former species differ in the exact position of the sex-patches of the 

 that of the hind wing being placed between SC 2 and R 1 , near their origin, in 

 the Indian species and at the base of the radials in the Papuan. The former, 

 0. obscurata Warren (1896), is only known as occurring in the Khasi Hills ; the 

 latter, 0. recurva Warren (1907) = ? 0. viridis Warren (?) was described 

 from a specimen from New Guinea (Owen Stanley Range), but is also known 

 to occur in both forms in the Bismarck Archipelago. In the present genus, the 

 structure of the scent-patch shows a rather close resemblance to that of Melitulias 

 Meyrick. 



To Mariaba Walker (= Adeta Warren) belong also, if my suggested 

 synonymy is correct, only two known species, both with a wide distribu- 

 tion, though not as yet found in Polynesia : (1) M. convoluta Walker (1866, 

 Sarawak) = ? M. amph, Warren (1899, Lombok) ; additional known localities 

 for the collective species are Assam to Malay Peninsula, and Dutch New Guinea. 



* The latter is the case in several and at least one $ of Hybridoneura, including the type 

 specimen of the genotype H. abnomiis Warren, though overlooked by its describe*. 



