1967] 
MacLeod & Adams — Berothidae 
259 
in Tables I and II. In this analysis, twenty- three morphological 
characters, which are judged to be of phylogenetic importance, have 
been coded into “primitive” and “derived” states by a study of the 
form of these characters in other groups of present-day Neuroptera. 
We have regarded as primitive states those which, because of their 
occurrence in generalized members of related families, are believed 
to reflect the state of these characters in the common ancestors of 
these families and, presumably, in the direct ancestors of the Ber- 
othidae as well. Derived states are presumed to have departed from 
the primitive states in the manner specified in Table I. In deducing 
the states of these characters, particular attention has been paid to 
their condition in the Mantispidae, Dilaridae, and Neurorthidae 5 , 
since study of both the larval and adult stages of these groups has 
indicated a close phylogenetic relationship of these taxa with the 
Berothidae (MacLeod, 1964; MacLeod and Adams, unpublished). 
Study of these data emphasizes the evolutionary mosaicism of each 
of the subfamilies and shows that although there is a preponderance 
of primitive states in the Cyrenoberothinae, Rhachiberothinae, and 
Nosybinae, important specializations are present in each of these 
groups which makes it unlikely that either is ancestral to the other 
two or to the Berothinae. Thus in the Cyrenoberothinae, the elongate 
head, non-tuberculate vertex, and divided 8th abdominal sternum of 
the female are specializations which eliminate this taxon as a candidate 
for the ancestor of the other subfamilies. Likewise the specialized 
prothorax and prothoracic legs and the reduction of 8th abdominal 
sternite in the female rule the Rhachiberothinae out as direct ancestors 
of the Nosybinae and Berothinae, while the additional specializations 
of the fusion of 9th abdominal tergite and ectoprocts and the loss of 
the recurrent humeral vein in the forewing eliminates the Rhachi- 
berothinae from the ancestry of the Cyrenoberothinae as well. Finally 
the specializations of the fusion of the stems of the radius and media 
in the forewings and the greatly elongate 9th gonocoxites of the 
females argue that the Nosybinae could not serve as ancestors of the 
Rhachiberothinae or Berothinae, while the additional features of 
the fusion of 9th abdominal tergite and ectoprocts and the loss of the 
cereal callus removes them as ancestors of the Cyrenoberothinae. 
This analysis also emphasizes the rather generalized nature of 
Naizema within the Berothinae and raises the question as to whether 
°A study of the adults and presumptive larvae of this group has convinced 
the senior author that this group should be treated as a distinct family. The 
evidence justifying this conclusion will be presented in a separate publication. 
