620 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxix. 



:i part of the radial sector, has also atrophied. The atrophy of the 

 base of the radial sector results in the utiiou of the cells R and 1st Rj, 

 a condition not found in any Tenth redinoidea other than Oryssus and 

 ^]>hry)i<>pus. This character is probably common to the other two 

 genera of this family which are not known to the writer. That the 

 above interpretation is the correct one is confirmed by the examination 

 of the wings of the genera Rhogas^ AidneaK^ Gastervptlon , Pelopa'ui^ 

 (lig. 20, (), and Aj)lfi (tig. '20, jf). If the base of the vein starting off 

 from the stigma in the first four of these genera be examined, it is 

 found that it extends obliquely to the first transverse vein just as in 

 Megalodoiittx, Cej>h)M, and Janux. If this vein were the base of the 

 radial sector, it would proceed in a regular curve, as in the genera 

 Tenthredo and Cladius. The composite nature of the first transverse 

 vein is shown by an examination of this vein in certain specimens of 

 Apis, where it is not straight but angular, as in Oep/ius. In certain 

 genera of Apoidea, as Boinhxs, Psit/iy /•>/■-, and Osvu'a, a,nd of Larridae, 

 the base of the radial sector is preserved as a fine thread-like vein, 

 frequently entirely colorless, while in some other genera only the 

 transparent stubs remain. 



The superfamily Tenthredinoidea can be difi'erentiated from the 

 other superfamilies by the presence in the front wings of one or both 

 of the anal cells, or, if both be wanting, with the medio-cubital cross- 

 vein extending ])etween R + M and cubitus; the cells R and 1st Rj 

 separated by a vein as broad as any of the others, or if not separated, 

 with the first antd cell present. 



The superfamily Tenthredinoidea is divisible into nine families. 

 They are the Xyelidie, Lydid^e, Blasticotomidre, Tenthredinida?, Xiphy- 

 driida% Siricida^, ]Megalodontida\ Cephida', Oryssida*. They are all 

 represented in the American fauna except the Megalodontidaj and the 

 Blasticotomidte, and contain a very limited number of species except 

 the family Tenthi-edinidse, which embraces several hundred species and 

 a large numbei- of subfamilies. 



The close relation of these families is proven by characters showing 

 a difl'ereuce in degree of specialization of the same kind, but through 

 the loss of certain of the intermediate stages those characters indicat- 

 ing a difference in degree of specialization of the same kind are here 

 just as useful as characters indicating a difi'erence in kind of speciali- 

 zation for marking dichotomous divisions. Although each of these 

 families represents a period in the development of certain characters, 

 yet the series is not a lineal one; that is, the connecting links do not 

 lie between the various fanulies, but behind them. They have been 

 developed froiu a common progenitor which transmitted its charac- 

 ters to its offspring in an elementary condition, and these ofi'spring 

 have developed along several parallel lines. Fortunately for our studN , 



