examine) have two pairs (PI. Ill, figs. lo, 12, 14), which in Cossula 

 are very prominent; in the Hypoptinae the genus Hypopta itself 

 (PI. Ill, fig. i) has two well developed pairs, the other genera but one 

 (PI. III. figs. 2 and 4). From the fact that in the two genera Cossula 

 and Hypopta we find two pairs of prominent spurs associated with a 

 well developed frenulum, whilst on the other hand along with the reduc- 

 tion of the frenulum there has been a corresponding reduction in the 

 size of the spurs {Cossinae) or a complete loss of one pair {Hypopti- 

 nae), the conclusion might perhaps be drawn that the most primitive 

 form was a two-spurred one. In this case the Zeu::crmae would show 

 reduction of the spurs combined with retention of the frenulum. This is, 

 however a mere hypothesis on our part and a careful study of far more 

 forms than we at present possess would be necessary to prove the truth 

 or error of the statement. 



The Tarsi are always spined on their ventral surface, show, how- 

 ever, considerable variation in the strength of the spining. In Zeuzera 

 (PI. Ill, fig. 6) they are strongly spined for their full length; in Ham- 

 ilcara (PL III, fig. 8) the spines simply form a group at the distal end 

 of each tarsus with occasionally a stray spine or two towards the cen- 

 tral portion ; a somewhat similar condition exists among all the genera 

 of the Hypoptinae (PI. Ill, fig. 2), with perhaps even more reduction 

 in the number of spines; the Cossiiiae (PI. Ill, figs. 12, 14) on the 

 other hand, as far as we could examine them, are strongly spined forms. 



The Palpi of each subfamily, with one exception, show character- 

 istic points of structural difference. In the Zeuzerinae they are very 

 minute (Pl. Ill, figs. 18, 19), appearing merely as slight tufts of hair 

 below the eyes; in the Hypoptinae (Pl. Ill, figs. 15, 16, 17) they are 

 better developed, porrect, usually projecting slightly beyond the front, 

 and with the 3rd joint conical and easily distinguishable from the pre- 

 ceding. In the Cossinae they are upturned, reaching to about the 

 middle of the front and more or less closely appressed against same; 

 they may be covered with long hair as in C. bertholdi (Pl. Ill, fig. 21) 

 or merely scaled as in P. robiniae (Pl. Ill, fig. 25), but the 3rd joint is 

 not conical nor well separated from the second. The one exception 

 is Cossula (Pl. Ill, fig. 20), in which the palpi, although somewhat up- 

 turned, have the third joint as in the Hypoptinae. The genus stands 

 in this respect midway between the two subfamilies. 



The Antennae are (in the male sex) mostly bipectinate, the 

 length of the pectinations varying in the different species. In the 



