TRE CAUSE OF MIMETIC EESEMBLANCE. 601 



species, while the lowest figure shows a form which is found in 

 Ecuador. In certain parts of Ecuador the black borders and 

 bands of the Methona (recently distinguished as M. psamathe, 

 Godm. & Salv.) are much narrower than in the type form, and 

 it is seen that the Dlsmorphia from the same locality has 

 followed in the same direction. This interesting parallelism 

 was pointed out to me by Mr. Salvin, and is here illustrated 

 from lithographs of specimens in the Grodraan-Salvin Collectiou, 

 Eig. 2 shows that, ia Dlsmorphia, the transparency is attained, 

 unlike the Ituna, by a reduction in size of the scales ; while, 

 unlike the IfhomiincB, the normal shape and outline are pre- 

 served, almost unchanged. 



It is a little remarkable that in this large and dominant group 

 no member of the Seliconinoe has yet been shown to find a 

 place. 



The moths which join the group are shown in PI. 44. fig. 3. 

 I believe that all the convergent species of AntJiomyza and 

 Castnia are represented in the figure ; these are, at any rate, all 

 that are to be found in the collection of Mr. Herbert Druce 

 and the Hope Collection. There are, however, other species 

 of Hyelosia which I was unable to obtain at the time the photo- 

 graphs made use of in preparing these Plates were taken. 



Microscopic examination proves that the moths become trans- 

 parent in two ways, which differ entirely from each other, and 

 from any of the methods already described in the butterflies. 

 A representation of the scales of the transparent part of the 

 Castnias is shown in PI. 44. fig. 6. The scales are not reduced 

 in size, but they have lost their pigment and are transparent ; 

 they are furthermore set up on edge so that the light freely 

 passes between them. The scales further from the opaque border 

 of the transparent patches are much more upright than those at 

 the margin which have been selected for figuring. The arrange- 

 ment in Syelosia is shown in fig. 5, and it is seen to be closely 

 similar to that of Castnia, save that the scales are much reduced 

 in number. The lithograph of the scales of Anthomyza (fig. 4) 

 proves that the scales are normal in size and arrangement. 

 They lie flat on the wing-membrane with the usual overlap, 

 but are so transparent that the light freely passes through 

 them. Although transparent, they retain a more or less faint 

 yellow or greenish-yellow tinge, but this is also to be noticed iu 

 the transparent part of the wings of Methona, Dlsmorphia, &c. 



