ANTENNAE OF LEPIDOPTERA. 39 



AA. Hairs of second type present and well devoloi)ed. 

 B. Antenna filiform. 

 C. Hairs of third type absent. 



D. Some hairs of second type on ventro-distal edge of sepments. 



GEOMBTRINA. 



DD. Hairs of second tyjie not present as in D AGABISTIDJ3. 



CC. Hairs of third tyjjc present. 



D. Hairs of third type without definite arrangement. 



E. Hairs nearly straight, segments of usual shape ARCTHD.^. 



EE. Hairs strongly curved, segments swollen, almost sul)glol)ose. 



LITHOSHD^. 

 DD. Hairs of third type inserted iu rows, usually on the compressed ven- 

 tral expansion in a semicircle opening ventrad. 

 E. Ventral expansion narrow and slightly compressed. GEOMETRINA. 

 EE. Ventral expansion large and usually compressed. . . .NOCTUID.^. 

 BB. Antenna pectinate. 

 0. Pectinations relatively short. 



D. Scai)e large, twice the diameter of pedicel NOCTUID^Sj. 



DD. Scai)e small, little larger in diameter than pedicel- • • .ZYG.ffiNID^. 

 CC. Pectinations long and well develoi)ed. 



D. Cones small, or apparently wanting PERICOPID.^. 



DD. Cones easily seen. 



E. Cones on shaft, or, if on pectinations, the latter have a single, strong 

 hair of second type directly at the ajicx. 

 F. Pectinations of cephalic and caudal sides subequal. 



GEOMETRINA. 



FF. Pectinations of caudal side longer ARCTIID.^. 



EE. Cones on pectinations, the latter with two strong hairs of second 

 type at the ai>ex. 



F. Cones very short ZYGiENlD^. 



FF. Cones large LYM ANTRIID^. 



The extent of my study and the limits of this thesis will not per- 

 mit me to discuss each family of the Frenulum-conservers, but cer- 

 tain features are especially worthy of note. The evidences presented 

 by antennal structures which might point to some relationships dif- 

 ferent from those shown in the classification of Professor Comstock 

 are as follows : The antennte of the Cymatophoridje and the Thy- 

 rididie seem to be closely related, and neither of them are widely 

 separated from the higher forms of the Pyralidina. The antennae, 

 therefore, would seem to indicate a lower position for these families 

 than that now assigned them. The antennie of the Dioptidte are 

 very close to those of the Notodontidie. The pectinations are more 

 higlily developed in the former, but the cones have not even begun 

 to migrate along the pectinations. Instead of that, they are situated 

 upoTi slight ventral elevations of the segnients between and a little 

 cephalad of the bases of the pectinations. In the Notodontidse, the 



TRANS. AM. ENT. SOC. XXIII. FEBKUAEV, 189ti. 



