24 



doubtful specimens must be made ; an examination of genitalia is, of 

 course, final but not always possible: 



( 1 ) in ncumocgctii the spots are often small, well separated, ir- 

 regularly rounded ; when forming a more or less coalescent band sf)ot 

 2 from anal angle is usually almost square and its inner margin is not 

 prominently wedge-shaped as in aryxna. 



(2) The fulvous hairing at base of both wings is much more ex- 

 tended in ncumocgcni, covering on the secondaries most of the area 

 between the subterminal spots and the base of wing. The presence 

 or size of yellow spots in the basal area beneath the fulvous hairs we 

 have found of no specific value. 



(3 The underside of secondaries in ncumocgcni is usuallj- dis- 

 tinctly paler in color, due to a greater sprinkling of white scales (com- 

 pare Figs. 2 and 7). The whitish subterminal band is very variable in 

 both species, in both distinctness and extent, and of no value for pur- 

 poses of separation. 



If it had not been for the fact that a certain difference in the geni- 

 talia went hand in hand with the above mentioned characteristics we 

 wolud have been disposed to regard the species as one single variable 

 one. This latter factor, however, cannot be overlooked, and as both 

 fonns occur in practically the same locality at the same time of year, 

 the difference can neither be regarded as racial nor seasonal ; with our 

 present lack of knowle<lge of the earlier stages nothing remains but to 

 separate the two forms as species for the present. 



There is considerable variation in size, as is commonly found in 

 species whose larvae are internal feeders. \'ery few, however, of the 

 (J $ attain the size given by Edwards in his description. The average 

 size is about 47 mm. 



The 9 9 have proved still more confusing than the ^ <J ; in the 

 first place our material has been much more limited and in the second 

 there is no means of being certain, even after separating two forms, 

 as to the correct association of male and female, both species occurring 

 in the same territory. Taking a 9 specimen which had been compared 

 with the 9 type of ncumocgcni as basis, we made a slide of its geni- 

 talia ; we then made slides of the genitalia of 6 other specimens and 

 imagine, rightly or wrongly as the case may be, that they may be separ- 

 ated into two series. We find, however, that the 9 genitalia are sub- 

 ject to considerably more variation than is exhibited in the <J sex, a 

 fact that is not so surprising when one considers that the vaginal plate 



