( -iVo ) 



of cloaiitliKs doanthus, and so they do in the form of the veutri-dorsal fold. The sub- 

 dol•^^al tooth is in every specimen shorter than that off. l.il (normal length of tooth 

 of doanthus cloonthiis), bnt there occur individuals wliich have the tooth nearly 

 as long as it is in f. I.j2. A s])ecimen of doanfhiis d>/menus with the tooth of a 

 length which is normal for d>/nicnus is represented by f. 1.53. We have examined 

 some individuals which externally are not different from Sikkim specimens, and 

 observed the highly interesting fact that in these individnals the tooth is as short 

 as in f. I.jS, wliile in some other examjjles whicli differ in the extent of the black 

 colour on the wings considerably from doanthus doanthus the tooth approaches 

 in length that of f. 152. This is a remarkable illustration of what we have said 

 in the introduction, namely that specimens can be similar or identical in one set 

 of characters, while in another, independent, set they are dissimilar ; the Chinese 

 individuals of doanthus which in pattern are like Indian doanthus doanthus are 

 nevertheless individuals oi doanthus dymenus, distinguished from doanthus doanthus 

 by a character of the genital armature not found, to our knowledge, in any 

 individuals from India. "We shall have to refer again to this fiict later on. 



The dorsal ridge of the Chinese doanthus is similar to that of the Indian 

 specimens ; we have not found any ditference that can be pronounced constant ; in 

 many individuals the ridge is a very little longer, and the teeth are often more 

 numerous and smaller. 



c. P. doanthus sumatranus from the mountainous districts of N.E. Sumatra: 

 f. 1.50, 154, 155. 



The sinus of the valve is much smaller than in tlie two jireceding subspecies ; 

 the dorsal edge of the dorsal lobe is more rounded, and the ventral lobe is con- 

 siderably blunter. The dentition of the ventral edge of the valve is, e.si)ecially near 

 the bhmt angle (/), extended upon the outside of the valve to such a degree that 

 five or six small, but strongly chitinised, teeth stand irregularly one above the other. 



The ventri-dorsal fold (e) is straighter and more raised than in doantliux 

 rlounthus and doanthus cbjmcnus, its edge is less rounded off and, especially near 

 the subdorsal tooth, slightly notched or faintly denticulate. 



The subdorsal tooth is as short as in doanthus dymenus, l)ut at the bas(> 

 broader, in conse(iuence of the fold, of whicii it is a process, l)eing liiglier ; mostly 

 it bears two or three faint teeth at the ventral edge. 



The dorsal ridge is obviously shorter than in the rest of the species, but is of 

 the same height; the free edge is less dentate, often simply sinuate. F. 154 and 

 155 illustrate the degree of variation in the dorsal ridge and the subdorsal tooth 

 noticed by us. 



The princijjal facts of variation as illustrated liy f. 149 to 155 are as follows:.^ 



(1) The three subspecies of P. doanthus agree in the valve and its armature, 

 but exhibit some differences in the detail of structure. 



(2) The Indian and tiie Chinese forms differ constantly in rlio lengtli of the 

 subdorsal tooth, though the extremes come very close. 



(3) The Sumatran form is aberrant in the form of \\u'. valve, the ventri-dorsid 

 ridge, the subdorsal tooth, and t he dorsal ridge. 



(4) In the length of the subdorsal tooth the Sumatran and Chinese forms agree 

 with one another, while thry ilisagree with thi' Indian subsiiecirs wiiich inhabits 

 interjacent countries. 



