26 
Psyche 
[March 
the known New World forms include prominent dorsal placement 
of the first and second pairs of abdominal spiracles and modifi- 
cation of some setae to form pigmented scales; in this latter respect, 
they all resemble an arboreal African larva tentatively assigned 
by Withycombe (1925) to the neuroptyngine genus Tmesibasis 
McLachlan, which is described as possessing conspicuous swatches 
of white scale-like setae on its otherwise dark body. On the other 
hand, those aspects of New World neuroptyngine head and mouth- 
part morphology that are functionally associated with prey capture 
fit the generalized pattern seen in most ascalaphines. However, 
one of the Central American forms shows a slight modification 
of the ventral mandibular articulation that may be associated with 
its proven ability to open its jaws slightly beyond 180 degrees. This 
tendency toward a Ululodes type of specialization is much more 
completely expressed in Withycombe’s “ Tmesibasis ” just men- 
tioned. The plesiomorphic complex of head features, then, may 
not be any more characteristic of Neuroptynginae than it is of 
Ascalaphinae. 
In summary, it appears that we are dealing here with two genera, 
Ululodes and Ascaloptynx, that seem to have acquired numerous 
specializations independently of one another and of other Old 
World members of their respective subfamilies. In view of the 
distinctiveness of the New World owlfly fauna (Weele, 1908; Orfila, 
1949), it is probably safe to assume that the American radiations 
of both subfamilies are old ones and that many of the basic traits 
that ancestrally defined the subfamilies have been partially ob- 
scured or modified. Thus, it is extremely difficult, given the present 
state of our knowledge, to separate unequivocally the larvae of split- 
eyed owlflies from those of entire-eyed ones. Equally difficult is 
assessing which of the two subfamilies displays the more specialized 
complex of larval traits. We have seen that feeding-related adapta- 
tions of the head and mouthpart region are of the same basic, gen- 
eralized sort in the majority of known representatives of both sub- 
families, with presumably independent evolution of the specialized 
constellation of characters at least twice in the Ascalaphinae and 
once in the Neuroptynginae. Other evidence presented in this paper 
suggests that the placement and degree of flattening of the body 
scoli may be more reliable indicators of subfamily affinity than 
head features: known neuroptyngines have flattened scoli with the 
secondary series coplanar with the primary series, while ascala- 
