( 483 ) 



characters of the wiug and valve, we have no reason whatever to say that monticolus 

 is derived from sarpedon sarpeclon, or from H'lrpcdon dodinyenKix, and not from 

 mrpedon milon; and hence a conclusion, based upon the occnrreuce of a special form 

 of F. mrpedon on Bonthain Peak, as to the proltable geological history of Celebes — 

 connection with the Snnda Islands or with the Northern Moluccas — would lack the 

 necessary facts. We shall have to refer again to monticolus in the last chapter of 

 this paper. 



The valve is nearly shaped as in .-Mrpedon milon, i.e. the ventral lobe is mnch 

 longer and narrower than in .mrpedon .virpedon. The dorsal ridge (f 146, from 

 apical side) bears a transverse dentate ridge, similar in position to that of dodingensis 

 (f. 140), but differently shaped; the dentate ridge is much less extended than in 

 ordinary examples of milon (f. 144). F. 14() gives a dorsal view of the ridge. 



0. P. mrpedon milon from Celebes, the Sulla Islands, and the island of 

 Talaut ; f. 14:j, 144, and 145. 



This long-winged and narrow-banded subspecies has the blue cnlour of the 

 wing-band in common with the Moluccaii races ; the nearest ally in respect to the 

 pattern of the wing is dodingensis, which has, like milon and monticolus, an 

 additional rod spot before the cell on the underside of the hindwing, and has the 

 median band of both wings also obvinusly narrower than it is in the subspecies 

 from Amboina and Ceram. 



The sinus of the valve of milon (f. 143) is deep and broad ; the ventral lobe 

 of the valve is very slender and long, resembling somewhat that oi jugans from 

 Sumba (f. 121) and teredon from Ceylon and S. India (f 13:3). The apical 

 ])rocessus of the dorsal ridge is nearly as long as in the two Moluccan races. The 

 transverse ridge has developed to a broad saw-like organ, which when seen from 

 above sometimes almost extends to the sinus of the valve. In the ventral view of 

 the dorsal ridge (f. 14.j, Celebes specimen) the denticulate portion of the dorsal 

 ridge itself (jl) is jjlainly visible ; this portion of the ridge is not developed in the 

 Moluccan races, but in monticolus from Bonthain Peak (f 147). The transverse 

 ridge (/<) is in this specimen almost perpendicular. A view of the same organ from 

 the apical side of the valve is given in f 1 44 ; the ridge is larger than in any 

 other subspecies, but there is some iudividiwl variation iu the size of the ridge. 



AVe have only one S from the island of Maugiola (Sulla Islands), which, like 

 the female, differs from the Celebes individuals in having a narrower median band 

 to the wings. The valve and its armature is in this tj very slightly different from 

 that of our Celebes examples; but, of course, we cannot tell from one individual 

 whether the difference is duo to individual or to local variation. 



The more prominent facts of variation illustrated by f 9(i to 148 are as follows:— 



(1) The genital armature varies according to locality; the ('hinese subspecies 

 agrees in the apparatus, however, with tlie Indo-Malayan one, in spite of very 

 prominent external differences. 



(2) The distinguishing characters in the genital armature of the subspecies are 

 found in the valve as well as in the harpe. 



(3) The individual variation within the Indo-Malayan subspecies, of which a 

 large material has been examined, is so great that the difference between every two 

 nearest allied subspecies is small compared with the difference exhibited by the 

 extreme individuals of tlic Indo-Malayan subspecies. 



