11 



passage from this form to Gracilaria, with one row of scales with a 

 slight circle of hairs at their base, and Lithocolletes, with only a 

 bristle or two peeping between them, is very direct. 



In Glyphipleryx {fuscoviridella, haworthaiia) this same type has 

 advanced by the disappearance of the ventral portion of the ring of 

 scales, so that the antenna is clothed all over with hairs, except that 

 at the dorsal apex of each segment are placed three or four short 

 squat scales, diminishing to two on the further segments of the 

 antenna. 



The Nepticulidas, with one row of scales to a segment completely 

 encircling the antenna, are very like Lithocolletes, and may have 

 a similar origin. They nre extremely alike ; Lithocolletes is, peihaps, 

 more slender. I have not been able to detect hairs on either. 

 Though they may have a similar origin, they certainly have not the 

 same origin, as Lithocolletes is clearly descended from the Tinea 

 stem, whilst Nepticula is in many respects more generalised than 

 Tinea. I should incline to place Nepticula on a special phylum of its 

 own in this matter, and as having acquired an antenna with one row 

 of scales because that was all it had room for, just as we imagine the 

 two rows of scales originated in a question of dimensions. The 

 scales of Nepticula are extremely large in proportion to the size of 

 the moth, and this circumstance would have its effect when the 

 scaling of the NepticuUi antenna (ancestral) was differentiated from 

 the hair clothing. 



The specialisation in the Psychidae and its lower allies, if these be 

 not included in the restricted family, gives us very many of the forms 

 that we find specialisation following in other groups. 



In the Lypusidae (Psychids with round cases and winged females) 

 we find in Lypusa maurella two rings of scales to the antenna with 

 hairs between them ; in the female the hairs are poorly developed and 

 sparse, and the further ring of scales is always complete. In the 

 male the scales are wanting ventrally, where the hairs are abundant 

 and long, and quite of a psychid type ; dorsally the further ring pre- 

 sents a patch of large strong scales; the nearer is nearly obsolete, 

 presenting only a scale or two of reduced size, and the dorsal por- 

 tions of the rings of hairs are fairly developed at both their proximal 

 and distal margins. 



In Psilothrix dardoinella and Melasina lugubris we find a divergent 

 specialisation that we find in the higher Psychids taking already two 

 separate directions. In these species the male antenna is pectinated. 

 In dardoinella the shaft is scaled dorsally, but long psychid sense- 

 hairs occur not only beneath the shaft and the pectinations, but occur 

 on all aspects of the pectinations, whilst in lugubris the pectinations 

 are scaled on their outer aspects. In lugubris female the scaling 

 covers the whole antenna except some area at the base in front in 

 each segment, where a few hairs ( ? tactile bristles) occur. The 

 scales might be described as being in two rows, but so irregularly as 

 almost to be without definite alignment. 



