350 Messrs. J. J. Joicey and W. J. Kaye's Notes on 



streak along the median vein, thus suggesting the well- 

 developed yellow streak in such races as nanna from 

 S. Brazil, and to a lesser extent in amandus from E. Bolivia. 

 Many specimens of the black hind-wing section of melpo- 

 mene from F. Guiana show a yellow basal blotch at the 

 submedian, but the extension along the median seems to 

 be very infrequently developed. 



There are one or two specimens showing white markings 

 partly in place of yellow, and these our friend Dr. Eltring- 

 ham will consider, and rightly so, as lending support to 

 his theory that the Colombian and Central American 

 white marked species oydno, galanthus and hermogenes are 

 really races of melpomene. 



There are four specimens of different fascies all showing 

 a white portion of what would be the true melpomene red 

 patch. Thus one is nearly an aglaopeia, with what is 

 usually the yellow spot in the cell partly white and partly 

 black scaled. 



Another of the cybele form has a large circular white 

 spot almost occupying the yellow spot between veins 2 

 and 3. Two others approximating to the faustina form 

 have white, in the one occupying the yellow spaces between 

 veins 5, 6, and 6, 7 on one side, and on the other occupy- 

 ing the costal yellow spot. In the second there is white 

 in the right fore-wing only, within the yellow discoidal 

 spot and within the yellow spots beyond the cell. 



The very long series of erato is equally variable with 

 the melpomene, and is chiefly different from the series 

 made in 1915 in that there are no fewer than 45 of the 

 tellus form, which was completely absent in the former 

 collection. There were 5 of the constricta form then 

 against 25 of the same form now, while the number of 

 andremona forms with a varying amount of white over- 

 laying pattern is much greater in the present series. One 

 of the erato (typical) has part of the yellow group of 

 spots in fore-wing white, for this form we propose the 

 name albida. The comparatively large number of the 

 fcliiis form emphasises what we said in our former paper, 

 thai the streaked forms of erato are essentially the 

 models for melpomene, as here there are 45 of this form 

 against 6 melpomene ab. penelope graduating into ab. 

 vicinus. In both collections the percentage of streaked 

 forms is very high, and the present collection is a remark- 

 able confirmation of the deduction arrived at from the 



