INTRODUCTION 



Since Neumoegen and Dyar issued their preliminary revision of 

 the Bombycidae north of Mexico (Jour. N. Y. Ent. Soc. II, p. i6o, 

 1894) no attempt at a classification of the Cossidae of North America 

 has been made. Our attention was first called to the matter when we 

 endeavored to place in their proper genus several new species which 

 had come into our possession and found that the key to the genera 

 given in the above mentioned work was not very practical. A further 

 examination of more material proved conclusively that the point of 

 difference on which the separation of the two main groups was based 

 was an exceedingly variable quantity and of no value in separating 

 subfamilies. We refer to the presence or absence of a cross-vein 

 between subcostal and radius (veins 7 and 8) of secondaries; this has 

 been used by Dyar to separate the two subfamilies Zcuzerinae and 

 Cossinae from the Hypoftiuac. As a matter of fact in a long series 

 of bertholdi Grt. examined, about half the specimens showed no trace 

 of the cross-vein, in the other half the vein was very distinctly pres- 

 ent, and in one specimen it was present on one wing and not on the 

 other. The same is true of several other species, e. g. nanus Strkr., 

 marga B. & McD., etc. In the subfamily Zeuacrinae the presence or 

 absence of this vein appears to be a constant feature and can, we con- 

 sider, be safely used for generic purposes : in Cossns itself (type C. 

 cosstis L.) the vein is not present and even in Acossus Dyar we have 

 met with a few cases where it was wanting. 



The synopsis of Cossid genera given by Dyar in Schaus' paper on 

 new S. American moths (Proc. of Nat. Mus. XXIX, p. 339, 1906) 

 has in our opinion several grave disadvantages in that secondary sex- 

 ual characters are largely used for separation purposes, and the error 

 of the previous paper to a great extent repeated, features which Dr. 

 Dyar himself has hoped that we should eliminate. An attempt has 

 therefore been made in the present paper, firstly, to base the Cossid 

 genera on some constant feature of structure and venation and, se- 

 condly, to group the various N. American species under their proper 

 genera. It has been found necessary to create several new genera, 

 but before doing so we have examined, as far as it was possible, all 

 allied European and S. American types of genera and trust that we 



