Iv. Jordan and N. 0. Rothschild 
o 
O 
identical are now recognised as being perfectly independent of each 
other. The nests of the sand martin yield no other flea than Gerato- 
phyllits styx, while henhouses are infested by Geratophyllus gallinae , 
which are easy to rear. Rabbit warrens contain quantities of Spilo- 
psyllus cuniculi, the insect being sometimes present in very large 
numbers on a rabbit, the individuals varying to a slight extent. From 
a careful comparative study of a very large series of specimens of such 
species, we arrive at a knowledge of the approximate limits of variation. 
The range of variation not being the same in the various organs, the 
most important point to ascertain is which organs are valuable and 
which are comparatively constant in such undoubtedly distinct species. 
The characters here observed to be specific are a guide and a basis of 
classification in the case of allied forms about the life history of which 
nothing is at present known. 
In the present treatise we are dealing with Pulicidae (. Loemopsylla 
cheopis being one of them) which can be recognised by the following 
characters: the complete absence of a comb (= pecten or ctenidium) 
from the genal edge of the head as well as from the pronotum, and 
the presence of a strongly developed eye. 
This group of fleas comprises six genera already recognised, namely, 
Pulex L., Lycopsylla Rothsch., Moeopsylla Rothsch., Parapsyllus 
Enderl., Rhopalopsyllus Baker, and Goniopsyllus Baker. We are, 
however, convinced that the division of this group into six genera only 
is an unnatural grouping. In the present article therefore a classification 
of these insects based upon a comparison of nearly all the known species 
has been attempted. This we think is especially desirable as con¬ 
siderable confusion appears to exist regarding the generic position of 
some of the species. The distinctions between the various species, too, 
do not appear to us to have been realized by all of those who have 
studied them. A concise exposition therefore of the generic and 
specific characters should, we think, be of use to those who have 
material they wish to name. It should, however, be clearly understood 
that the genera treated of in this paper are not more closely allied to 
each other than to some other Siphonaptera which have ctenidia. For 
instance, Ctenocephalus erinacei and Spilopsyllus cuniculi, which we 
exclude from this paper, are more closely allied to Pulex irritans than 
are the members of the American genus Rhopalopsyllus , which we shall 
deal with. 
We intended at first to restrict this paper to an exposition of the 
species closely allied to Loemopsylla cheopis (which plays so prominent 
1—2 
