638 riWCEEDINGf-i OF THE NA TloyAL MUSEUM. voi,. xxix. 



appear to be of an}^ phyllogenetic value. Although the Hylotominse, 

 Schizocerina^ and Perreyinee are so closely related in their wing char- 

 acters that there is a gr-eat dearth of prominent differences for dividing 

 them, yet the Perreyinte are readily separated from the other sub- 

 families on antennal characters. 



The anal cell Jooxers. — The genera included under this heading are 

 generally divided into three su])families, the Lobocerinie, Pterygo- 

 phorinffi, and Perginte, Such a grouping combines forms that are not 

 closely related and they have therefore been divided into the following 

 subfamilies: The Incaliinte, including the genera hicalla and Paralypla., 

 which are limited to South America; the Lobocerinse, including the 

 genera Lohoceras (tig. 82), Aulacomeru^, Syzgonia^ and Corynoplvilus^ 

 which are also limited to South America; the Acordulecerinfe, includ- 

 ing the single genus Acordulecera (fig. 83), found in North and South 

 America; the Pterygophorina? including the genera Pterygojjhorus 

 (tig. 81), Pterygophorinus^ Liqjhyrotoiiia^ and Phlloiuastix, which are 

 limited to Australasia; and the Perginje, including the genera Cerealces 

 and Perga (fig. 84), and its subdivisions, which are limited to Aus- 

 tralasia. It has been impossible to examine specimens of all these 

 groups, and the writer has had to depend in many cases on figures, 

 so that the characters given in the table on a later page may not be of 

 any more vahie than to indicate the regions which are undergoing 

 modifications. These subfamilies are set off from all the other Ten- 

 thredinida^ ])y the reduction of both anal cells of the front wings. The 

 wings are long and slender, and the anal area of the hind wings is 

 generally greatly reduced. In the front wings the costal area is broad, 

 and the free par-t of Sci is preserved in the Lobocerinee, but in the 

 other subfamilies the costal area is hardl}^ more than a line, and the 

 free part of Scj is wanting. The radial cross-vein is wanting. The 

 cell R^+„ is append iculate in the Incaliinw, Lobocerina\ Pterygophor- 

 ini¥, and Pergina?, ending at or before the middle of the cell Rg, the 

 vein Rj being continued to the apex of the wing. 



In the Acordulecerinje this cell is not appendiculate, and ends 

 distinctly beyond the middle of the cell R3. The medio-cubital cross- 

 vein joins Sc+R+M at or very near the origin of M. In the Ptery- 

 gophorina^ the free part of R^, is wanting, in Acordulecera and certain 

 species of Perginie the radio-medial cross-vein is also wanting. The 

 free part of M^+Cuj joins the cell M^ just before the middle in the 

 Loboceriiuv and Pterygophorina', and insterstitial with the medio- 

 cubital cross-vein in the Perginre. In the hind wings the cell Rj+g is 

 appendiculate in the PerginiB and Lobocerina^, while in the Acordule- 

 cerina' and Pterygophorinjv it extends as a long, pointed cell to the 

 apex of the wing. The free part of R^ is alwa3's present, and the 

 transverse part of M.^ always wanting. As in the Hjdotominse and its 

 allies, the base of cubitus apparently forms a continuous vein to the 



