328 Mr. A. R. Wallace on 



in Sumatra and also in Celebes ; the former locality is so impro- 

 bable that I unhesitatingly reject it in the absence of direct evi- 

 dence, and as to the latter perhaps a specimen of T. celebensis has 

 been mistaken for it. 



34. Terias puella, Boisduval. 



Terias puella, Bd. Voy. Astrol. p. GO, pi. 2, fig. 8, S ', Sp. Gen. 



p. G74 (Candida, var.). 



Hah. — Batchian, Gilolo, Waigiou (Coll. Wall.). 



The female resembles the male, but has the border rather wider, 

 and is of a pale sulphur-yellow above. In the male the black 

 border is more regularly curved towards the costa than in T. Can- 

 dida, $ . Vollenhoven (Mon. Pier. p. 69) says his specimen of a 

 female from Ternate is white. This would agree with my T.virgo, 

 Is it not an error of locality, as he does not say that the specimen 

 was obtained from any recent collector or other trustworthy 

 source, and many Ternate vessels visit the Aru Islands, the habitat 

 of T. virso ? 



o 



So. Terias virgo, n. sp. 



Male. — Above, rich yellow, the black border of the upper wings 

 as iti T. Candida, of the lower wings as in T. puella. 



Female. — Colour above, pure white as in T. Candida, J ; border 

 as in T. puella, $ . 



Hub.— Aru Islands (Coll. Wall.). 



The curious combination of characters as indicated above obliges 

 me to separate this as a distinct species. 



36. Terias impura, Vollenhoven. 

 Terias impura, Voll. Mon. Pier. p. 70, pi. 7, fig. 5. 

 Hah. — Timor. 



The form of this small species is more like that of some of the 

 South American species of Terias. 



PlERIS. 



This genus will now contain all the species of the old genus 

 Pieris which have two subcostal ncrvules given off before the end 

 of the cell, and the males of which have neither the costa serrated 

 nor the anal valves tufted. The third subcostal nervule is gene- 

 rally well developed, but it varies greatly, becoming very short in 



