198 Mr. H. Scott on 



only in passing, or who lias )irlden for some liours among 

 the precipitous mountains, dense forests, and countless 

 swift streams of such an island as Dominica, must admit at 

 least the possibility of fresh finds being made in them. 



Assuming the l^altostominee to be a natural group, then 

 this group not only contains all the Neotropical representa- 

 tives of the family [Paltostoma, KeUoggina, and Curupira), 

 but is also represented in South Africa by tlie only Blepharo- 

 cerid known from that region [Kelloyyina barnardi, Ed- 

 wards, op. cit.), and in New Zealand by the genus Neocuru- 

 pira, Lamb [op. cit.). 



Affinities of the Larva. — The Paliostoma-larva., though 

 agreeing with that of Curupira (as described by F, Miiller) 

 in its short 2-jointed antennae, in having the dorsal surface 

 spinose, and in having only one pair of lateral processes on 

 the sixth segment, yet differs from it in having the branchial 

 filaments disposed in tufts, not in series. It differs more 

 widely from the Cu7-upira-\ike larva from New Zealand 

 described and figured by Chilton {op. cit.), and considered 

 by Bezzi {op. cit. 1914) to be the larva of Neocurupira, for 

 that larva not only has the branchial filaments disposed in 

 series, but also has two pairs of lateral processes on the 

 sixth segment. 



On the other hand, the Pa/tostoniaAarxn has several 

 points of resemblance with that of the South-African 

 Kelloggina barnardi. There would be nothing surprising in 

 this, as both forms belong to the same group, Avere it not for 

 the difficulty that^ according to Bezzi's tables of larval 

 characters {op. cit. 1913, pp. 76-80), the larva of K. bar- 

 nardi falls next to that of Blepharocera fasciata, Westw. : 

 this being a member of a genus which differs widely from 

 Kelloggina and Paltostoma in venation, in possessing dichoptic 

 eyes, &c. Bezzi, remarking in a footnote {op. cit. 1913, p. 78) 

 that the Kelloggina-lavva falls in a group different from that 

 to which the perfect insect appertains, suggests that the 

 larva described as that of K. barnardi may really belong to an 

 unknown species of a different genus. But, after examining 

 the South-African material presented to the Cambridge 

 Museum, I find this not to be the case. From one of the 

 pupse I have dissected a ^ fly which is undoubtedly K. bar- 

 ?iardi, thereby settling the determination of the pu[ se. Also 

 the only larva which I possess has the pupal respiratory 

 horns already formed beneath its skin, and these agree in 

 form with those of the pupse : so there should be no reason- 

 uble doubt as to the identity of the larvse. Therefore, the 



