CLASSIFICATION. 17 
rockUon, Paraleyrodes. and Udamoselis. These genera include 
spec: :nga considerable variation in important characters and. 
with other material available, furnish a basis for the consideration 
of their relationships and probable lines of descent. Information 
on these question- is ssential for a correct understanding of the posi- 
tion of the family among the Homoptera and for the assignment 
within the family of the various species into natural groups. 
Until the recent paper by Dr. Enderlein 18 no attempt has been 
made to divide the family into subfamilies, and to date there has been 
no attempt to show the natural relations of the genera. Enderlein 
erected the subfamily Udamoselina?, in which, he included also 
Alex, . and the subfamily Aleyrodina?, in which were placed all 
other known forms. 
In the following classification we restrict Udamoselina? to the forms 
in which both media and cubitus are present. Our reasons for this. 
our ideas on the relation of the genera and the systematic position of 
the family, and a discussion of the subfamilies follow. It appears 
to us. from a careful study of the different fonns, that the Aley- 
rodida? are not intermediate in position between the Aphidida? and 
Coccida? but that they form an offshoot from the psyllid stem. This 
is indicated by the wing venation and by the structure of the mouth- 
parts, legs, and genitalia. 
The venation of the wings of this family has been compared with 
that of a psyllid wing tinder a separate heading (p. 9) and it is clear 
that the ancestor of the Aleyrodida? had a wing form and venation 
similar to the psyllid type. Under another heading (pp. 12-13) 
the genitalia of this family and those of the Psyllida? have been com- 
pared. The resemblance here is very striking. In the males only a 
slight modification of the structure found in the Psyllida? is necessary 
to produce the type found in Aleurodieus* while the Aphidida? have 
a widely different structure. la the females, too. the same condition 
is found to be true. The ovipositors in the two families are remark- 
ably alike, while in the Aphidida? this structure shows almost no 
resemblance. In this second character, then, the two families are 
d to be closely related. 
The mouthparts show the most remarkable resemblance. In fact, 
with a variation in form of some parts and a difference in size, the 
two are identical. Even - me number of similarly formed taste 
sensoria are present on the labium. 
In the legs, again. Larity is to be noticed. In the presence 
be pul villus and the segmentation of the tarsus, the Aleyrodida? 
are much closer to the Psyllida? than to the Aphidida?. 
The family itself shows two distinct lines of descent. In the one 
the media is retained, the cubitus is lost, and large compound wax 
pores are often developed in the pupa?. In the other the media is 
