( xcvi ) 



structure, there is an intermediate region where the markiugs 

 become crowded and indistinct, the a2>pen ranee presented 

 being that of a dense accumulation of granules. This is no 

 doubt chiefly due to the presence of pigment, and is especially 

 well seen in certain species of the genera Ganoris, Pina- 

 coptenjx, Delias and Neplieronia. 



The lamina varies much in shape in different groups. From 

 its distal margin, which is usually pointed, but may be 

 rounded or nearly straight, the fimbriae take their oi'igin. 

 These are usually from twenty to thirty in number, and 

 appear to be in direct connection with the chitinous divisions 

 of the lamina indicated by the longitudinal striation above 

 mentioned. It is not easy to say from actual observation 

 whether their distal extremities are open, though their aspect 

 under a high power suggests this ; and it would seem to be 

 necessitated by the view here taken of the function of the 

 plume-scale, that the fimbriae should consist essentially of 

 tubular prolongations of the intra-laminar cavity, provided 

 with terminal orifices. That the lamina itself is permeable 

 to fluids can be easily demonstrated by applying moisture in 

 the form of breath to a glass plate on which some of these 

 structures are displayed. Under a moderate power of the 

 microscope it can be seen that the plume-scale readily fills up 

 by capillary attraction, and it is frequently obvious that the 

 invasion of fluid does not proceed uniformly along the lamina, 

 but follows the longitudinal lines marked out with more or 

 less distinctness by the ribbed or scalariform internal chitinous 

 structure of the scale. As the fluid is apt to run more raj)idly 

 along some of these lines than along others, the general line 

 of advance tends to be uneven. 



This appearance, it may be noted, is not confined to the 

 plume-scale, but is observable also in wing-scales of the 

 ordinary character. 



In view of the structural features that have now been 

 described, the interpretation suggests itself that in the most 

 usual form of plume-scale the greater part of the cavity of the 

 lamina is divided more or less completely into longitudinal 

 channels, with or without lateral communications. These 

 channels arise by divergence from that portion of the base of 



