KAXPAS UNIVI' 



Fig. 10. 

 Scales taken from single forewing of Megalopijge crixpa/a. 



Coupled with this variation in the form of the scales it is to be 

 noted that the more slender and hair-like forms are more or less 

 assurgent, and are not appressed to the wing membrane, while the 

 flattened forms with pedicel and blade lie closely against the wing- 

 surface. The direction of the long axis of the insertion cup varies 

 correspondingly; those cups from which flat scales arise lying more 

 nearly parallel with the wing membrane, while those from which 

 hair-like scales arise are directed more toward the interior of the 

 wing substance. 



Variations in the mode of arrangement of the scales on the wings 

 may be comprised within two extremes: a condition of arrangement 

 whereby the scales are scattered over the surface of the membrane, 

 their insertions at approximately equal distances apart, but with little 

 or no suggestion of any rows of scales, as shown by Microptcryx: 

 and a condition as that described and figured on p. 48 (fig 3), in 

 which the scales are uniformly arranged in rows, each of two tiers, 

 and running subparallel with adjacent rows. It is to be noted that 

 an arrangement of the scales in rows and tiers is only associated with 

 broad, flat scales; the long, slender, hair-like forms of scales are 

 never arranged in rows. But the converse of this, that all flat, broad, 

 scales are arranged in regular rows and tiers is not true. One may find 

 fiat scales covering a wing, but showing little arrangement into sub- 

 parallel rows. This may be well seen in Micropteryx. 



IV. 



The functions fulfilled by the scales, once understood, must give a 

 clue to the rationale of the tendency of the specialization of the 



