The Crane-Flies of New York — Part I . 863 



most Limnophilini (Plates XXXVIII-XLI), or it may tend to retreat 

 proximad toward the base of the wing as in many Eriopterini (Plates 

 XXXIV-XXXVII), or it may be very far removed from the tip so that 

 it lies before the base of the sector (as in the tribe Pekiciini, Plates 

 XLI, XLII, and in the genus Ula, Plate XLI, 164). In some Anto- 

 chini it is apparently lost by atrophy. In the subfamily Tipulinae only 

 the more generahzed species retain Sci (Plate XLIII, 188 and 189), but 

 Sd is present and is bent strongly into Ri at its tip, thus forming a good 

 subfamily character. 



The radius (R, fig. 128, a) is the strongest vein of the wing, and, with 

 its sector, one of the most plastic. Ri runs straight to the wing margin, 

 but usually at about midlength of the wing it forks, sending off the radial 

 sector (Rs). This is primitively twice forked, being forked and the 

 branches forked again, dichotomously. These branches of the sector are 

 numbered from 2 to 5, the upper fork carrying with it R2 and R3 and the 

 lower fork carrying with it Ri and R^. The full complement of branches 

 of the radial sector is found only in the Tanyderidae (Plate XXX, 1). In 

 the Ptychopteridae (Plate XXX, 2-4) the upper fork, R2 + 3, is fused to 

 the margin; in the Tipulidae (Plates XXX-XLVIII) it is almost always 

 the lower of the dichotomous forkings, Ri+5, that is fused to the 

 margin. 



The various ways in which the full complement of veins has been lost, 

 by the fusing together of adjacent veins or else by the atrophy or dropping 

 out of one or more of the branches, may be here discussed. In the 

 Cylindrotominae (Plate XXX, 5-8) the appearance suggests the fusion 

 of the upper fork of the sector (R2 + 3) with Ri, forming a long, backward 

 fusion of R1 + 2 + S from the wing margin. As suggested by the author 

 in an earUer paper (Alexander, 1914 b: 604-605) and later proved by the 

 discovery of the Oriental genus Stibadocera Enderlein (Alexander, 1915 c: 

 178-179), the loss of these veins is by atrophy rather than by fusion, 

 and the vein that simulates /?i + 2 + 3 is, in reality, R3 alone and corresponds 

 exactly to this vein in other tribes of crane-flies. In the subgenus 

 Leiponeura of the genus Gonomyia (Plate XXXVI, 86-88), the vein 

 Ro + z is fused to the wing margin, or, possibly, R3 is atrophied after the 

 fusion has proceeded almost to the margin. In the more generalized 

 species of Gonomyia (Plate XXXVI, 89 and 90), the fork of R2 + 3 is 



