260 Transactions. 



wing to be very different from N. hudsoni Lamb in the form of the anal 

 angle of the wing and the radial fork, while the form of the anal angle of 

 the wing of N. hudsoni is similar to that of Apistomyia elegans (fig. A, 2), 

 though the radial fork shows a different stage in reduction in these two 

 forms. Bezzi has also pointed out (1913, p. 68) the close resemblance 

 of Apistomyia elegans to Apistomyia collini from Australia. Further, the 

 subcostal vein is evanescent in P. turrifer (figs. 48 and 53) ; it is small in 

 N. hudsoni (figs. 44 and 50) ; and is still well developed in the first new 

 species about to be described (figs. 47 and 52). All these characters 

 in the venation, and the larval characters, require, in my opinion, that 

 N. hudsoni and P. turrifer be placed in the Apistomyinae, while the new 

 fly, herein described under the name Curupira chiltoni, should be placed in 

 the Paltostominae. Moreover, the larva A of Bezzi is a Paltostomid, and is, 

 I believe, that of Curupira chiltoni, while larvae B and C, being Apistomyids, 

 belong respectively to N. hudsoni (or a similar fly) and P. turrifer. Support 

 is given to these relationships by the distribution of the larvae and imagines — 

 e.g., in a stream at Purau Peritheates and larva C are found together with 

 Curupira chiltoni and larva A, but I have never taken either larva B or 

 N. hudsoni. I hope, however, to confirm or disprove the suggested relation- 

 ships, at least of larvae A and C, during the coming season. 



Preferring, in a letter to the author, to the position as stated so far, 

 Professor Bezzi points out that there may be other larvae more nearly 

 approaching his description of B and C — i.e., " dorsum unarmed and bare " — 

 and that other Paltostomid flies may be found The discovery of the fourth 

 fly from Ohakune has in at least one respect confirmed Professor Bezzi's 

 opinion. This fly closely resembles N. hudsoni Lamb, but the vein R 

 has lost its fork, and the wing is smaller. The assumed larva (from 

 Ohakune) of this fly, which I propose to name Apistomyia harrisi. 

 closely resembles the Otira larvae, which I take to be those of N. hudsoni. 

 I have also from Queenstown a Blepharocerid larva (fig. 29 ) which 

 has distinct characters, and is probably a closely related but undis- 

 covered fly. 



Returning again to Bezzi's figure of Apistomyia elegans (fig. A, 2), it 

 should be noted that Bezzi's enumeration differs from that used in this paper. 

 His M 1+2 is R 3 ; R 2 b & s disappeared ; the basal portion of his M 1+2 , up 

 to where it touches R 3 , is really part of R4+5 ; M 1+2 is not represented, 

 except that basal part marked M and r-m (fig. 46). The lost veins are 

 dotted. This appears to be confirmed by the primitive wing-venation 

 of Edwardsina, chilensis Alexander; and, furthermore, fig. 49 shows the 

 bases of insertion of the macrotrichia of R2+3 on the wing-membrane 

 of A. harrisi. If N. hudsoni and P. turrifer be Apistomyids, we have, 

 with A. harrisi n. sp., three distinct stages in wing- reduction in this 

 ancient subfamily. 



Curupira chiltoni n. sp. (Figs. 55 to 75.) 



Head : Vertex small, finely pubescent, occupied by the large ocellar 

 turret, around which is a number of stiff bristles. One ocellus anterior, 

 the other two placed laterally. Space between the eyes projecting out to 

 form the raised keel described by Lamb. This space is narrower in the <?• 

 and in both sexes appears broader as it approaches the base of the labrum 

 Eyes hairy, dichoptic, bisected in both sexes, upper eye-facets larger. 



