Structure of Scent Organs in Male Danaine Butterflies. 167 



structure. Both transverse and longitudinal sections 

 show that there are, entangled in the hairs, large numbers 

 of minute spherules — apparently a coagulated secretion. 

 They are certainly not chitinous particles of dust material, 

 since they dissolve fairly readily and not slowly in eau-de- 

 javelle. From the appearance of these granules in a longi- 

 tudinal section, one is tempted to believe that they are 

 extruded from the hairs themselves, since they adhere 

 thereto very closely in many places. The absence, how- 

 ever, of any glandular structure in the cells of the brush-bag 

 lends greater support to the theory that they are in fact a 

 product of the secretion produced in the wing-pocket. 

 The brush-hairs are round in section for part of their 

 length, but become much flattened towards their distal 

 extremities. PI. XIX, fig. 26, shows a section of one of 

 these hairs ; fig. 27 a longitudinal surface view. The 

 surface of the hair is seen to be traversed by longitudinal 

 curved ridges, each of which bears two rows of small 

 projections arranged alternately on each side of the crest. 

 As in T. petiverana there is no pocket in the unexpanded 

 wing, but the portion which afterwards becomes invagin- 

 ated consists of a diverticulum. PI. XVI, fig. 1, is a photo- 

 graph of a section of a wing before emergence. On the 

 right may be seen the outgrowth from the wing-surface 

 covered on its outer side by a layer of small thick scales. 

 The double row of delicate structures in the centre of the 

 section represents the ordinary scales of the upper surface 

 of the wing. PI. XX, fig. 28, is a drawing of a small portion 

 of that part of the unexpanded wing which will form the 

 scent-pocket in the imago. The cells which produce 

 scales are highly vacuolated, and alternate with cells 

 attached to cup-hke projections somewhat similar to those 

 already described in other species. The cup-cells at this 

 stage seem mainly to produce those long processes described 

 by Mayer,* which ultimately form the fibres holding to- 

 gether the upper and lower surfaces of the wing during 

 expansion. Nevertheless some of them ultimately develop 

 into cells which appear to differ but little from those 

 attached to the scale-sockets. The cause of the folding 

 in of the diverticulum during expansion may be easily 

 discerned by comparison with PI. XIX, fig. 25, which shows 

 in section a part of the ordinary unexpanded wing-mem- 

 brane. Here the chitin is thrown into deep folds which 

 * Bull. Harvard Mus., vol. xxix, p. 209 et seq., 1896. 



