262 
Psyche 
[September 
this genus is correctly referred to this subfamily. Our treatment of 
Naizema is based on an examination of the direction of the specializa- 
tions of this group. Thus the elongation and subfalcation of the 
wing; elongate, coiled mediuncus with setae at its tip; incipient de- 
velopment of hypocaudae; and the bulb-like enlargement of a portion 
of the spermatheca are all specializations of a berothine type and 
argue that Naizema has stronger phylogenetic relationships with the 
other genera of the Berothinae than with any other group. The gap 
between Naizema and other berothines is, however, considerable and 
tribal distinction may well be justified when more complete knowl- 
edge of the other genera of this subfamily is available. 
Relationship to the Mantispidae. On the basis of the similarities 
in the structure of their adults and on certain general parallels in 
their life histories, Tjeder (1959) pointed out the likelihood of a 
close phylogenetic relationship between the Berothidae and the 
platymantispine Mantispidae. Some of the items on Tjeder’s list of 
similarities (1959: 273), such as the pectinate branching pattern of 
Rs in the forewing, the sinuate course of the base of MA (“r-m 
vein”) in the hindwing, the presence of trichosors, the occurrence 
of similarly shaped mouthparts, and the similarly shaped distal tarso- 
mere of the anterior leg, actually prove nothing since they occur in 
many other, distantly related groups of Neuroptera as well. The 
additional similarity of the fusion of the 9th abdominal tevgite with 
the ectoprocts, noted by Tjeder in Plega , seems more likely to be 
the result' of parallel evolution, since these sclerites are unfu ed in 
Cyrenoberotha and in such generalized platymantispines as U 1 epanicus 
and Gerstaeckerella. We believe, however, that Tjeder s central 
thesis is certainly corrct and that it is further strengthened by several 
of the structural details of Cyrenoberotha and other generalized 
berothids. 
The form of the gonarcus and gonocoxites of unspecialized bero- 
thids such as Cyrenoberotha , Nosy bus J Naizema , and Trichoberotha 
are very similar to the pattern found in all of the Mantispidae known 
to us. In these groups the well-developed gonocoxities are widely sep- 
arated from the gonarcus distally, while these two sclerites form a 
strong articulation proximally. A strong similarity also exists between 
Cyrenoberotha and many mantispids in the form of the mediuncus 
which is separated from the gonarcus and has the form of a basal 
plate in the roof of the ductus ejaculatorius from which a free, pos- 
terior extension is developed. Comparisons have previously been d awn 
(Tjeder, 1959) between the extremely elongate posterior extension 
