NOTES ON COSSID^. 97 



that of Z. pi/ri)ii, but of less pronounced convexity. The dorsum 

 of shaft is covered with numerous scales arranged in irregular 

 rows across each segment ; there are scales also on the dorsa of 

 pectinations and numerous ventral sense-hairs. Three segments 

 are transitional between the bipectinate and the terminal seg- 

 ments. These have a ventral appendage, with sense-cones, 

 sense-hairs, and bristles ; the short lateral extensions are of the 

 appendage, not of the shaft itself. 



The female antenna of Z. pijrini, C. cossus, and of C. expressa 

 are without pectination, and bear ventral extension with anterior 

 sense-cones, sense-hairs, and bristles. Comparing the male 

 unpectinate terminal segments of Z. pyriiii with the female 

 antennal segments, and remembering the transitional segments 

 of P. a r audi II is, one concludes that ventral extension preceded 

 bipectination among these insects. 



Thus, I should say, P. rohinice is a higher — more recent — 

 species than Z. pyrini, as in the former species both sexes are 

 bipectinate. The male pectination being slender, smooth, devoid 

 of scales, without scaling on the shaft, and very like Z. pyrini, 

 except that the segments are bipectinate to tip of antenna. The 

 female segments (hg. 15) have stout bipectination to the tip, and 

 one row of scales per segment on shaft. 



It is interesting to find that the male antennal segments of 

 C. cossus have simply a tongue-like ventral appendage, which, 

 viewed in section, is almost round (figs. 13, 14) ; two rows of 

 scaling per segment are represented by somewhat diminutive 

 scales on the dorsum of shaft. The female antenna has cones 

 on the anterior surface of the appendage. 



C. expressa male antenna (tig. 10) approaches that of P. 

 arandinis more nearly than either of the Cossids in regard to 

 scaling. The dorsum of shaft very closely covered with scales, 

 two rows per segment, and the dorsa of the pectinations are 

 likewise covered very closely with scales to the tips. The 

 female antenna (fig. 11) has a close resemblance to the Cossid 

 antenna, the dorsum of shaft having two rows of scales per 

 segment. 



From the foregoing it is evident C. expressa is not a Cossid ; 

 still, it has Cossid affinities, and if in other respects it is a 

 Tortricid, it must be rather low in that family. There is also 

 some reason to suggest that Zeuzera and Cossus are not such 

 near relatives as is sometimes supposed, but more material in 

 the newly-hatched larval stage must be examined. 



