19 



forming two rings of hairs and two of scales on each joint {Incurvaria 

 prcelatella, female). 



4. From this many variations occur, (a) The hairs disappear in 

 the upper ring, leaving a basal ring of hairs and two rings of scales ; 

 and (a 1 ) the basal row of hairs becomes evanescent, leaving two rings 

 of scales only (Gracilariadje). (I?) Gradual disappearance of first 

 ring of scales {Tinea fitlvimitrella), leaving only a wide basal ring of 

 hairs and one ring of scales. This continues (Glyphipteryx), by the 

 loss of ventral scales, to there being only a dorsal terminal row of a few 

 scales to a segment. 



5. The main line, then, is a strong tendency for scales to dis- 

 appear ventrally and for hairs to disappear dorsally. In some (as 

 Coleophord) the dorsal disappearance of hairs has advanced greatly, 

 whilst scales still persist ventrally. In others the scales have disap- 

 peared ventrally, whilst one or two rows of hairs still exist dorsally. 

 This line is well illustrated in Tortrices. 



6. The final tendency seems to be the usual form in obtect Heter- 

 ocera, of hairs ventrally, two rows of scales dorsally. 



7. Pectination appears to result from an effort to increase the 

 antennal area, which does so dorsally as well as ventrally, so that the 

 double row of scales dorsally is increased, usually throwing the 

 scales out of alignment, whence the scales, instead of being in two 

 regular rows, are an irregularly disposed thatch. This persists even 

 after pectination has disappeared again (Cochlididse). 



8. There is a capacity in the final obtect form for hairs and scales 

 to encroach on each other's territory, but they never apparently occur 

 again mixed up together. The tendency is often seen for scales to 

 disappear terminally (Alypia, Psodos, Polyploca, etc.). 



9. Butterflies have followed a separate line of evolution. They 

 begin with a strong tendency for scales to be basal and hairs terminal, 

 and with scaling largely developed ventrally. The irregularity of 

 their scaling suggests that their ancestry included some pectinated 

 forms, notwithstanding that any tendency to pectination is now con- 

 spicuous by its absolute absence. 



