SUGGESTIONS FOR TRACING RELATIONSHIPS OF 
INSECTS. 
By NATHAN BANKS. 
In studying any group, especially when one is trying to 
make a synoptic table, we become interested in the relationships 
or affinities and try to arrange the species or genera according to 
our ideas of their phylogeny. Yet, I fear in many cases we 
proceed without any clear idea of a basis for decision. It is 
evident that in different groups different methods may be nec- 
essary, but there are a few considerations which I think may 
apply to many cases. 
Some authors try to put first those forms that possess prim- 
itive characters, or the greatest number of such characters. 
Others take certain synthetic forms which seem to show rela- 
tionships in several directions as a starting point for the group. 
Everyone has observed that in any large group, as an order, 
there is contradictory evidence as to what is the most primitive 
family or genus. In Coleoptera for example, certain genera 
have more free ventral segments than usual, other genera have 
ocelli, or traces of a median suture on the head, yet some of 
these will not have the five-jointed tarsi. 
A case familiar to me is the Hydropsychid caddice-flies. 
Their ancestors were near the Rhyacophilidz and had 3, 4, 4, 
spurs, ocelli present, and the female with two little appendages 
at tip of the body. We find in the Hydropsychids that some 
have ocelli, but do not have the 3, 4, 4 spurs, while others have 
the 3, 4, 4 spurs, but not the ocelli, and various genera have the 
primitive abdominal appendages. 
In other words primitive characters are inherent in the 
descendants and may be developed in various parts of the 
descendant series, or, more properly, retained by varying lines 
of descendant series, so that taking any family of existent forms 
several arrangements are possible according to what primitive 
character is chosen as the criterion. 
Genera differ from other genera by at least two sets of char- 
acters. One is the positive characters, the presence or absence 
of a structure, the other is in accrescent characters, or developing 
136 
