2 The Non-Combed Eyed Siphonaptera 
In a paper by Rothschild (1906, p. 483 *) it was stated that the 
species of flea usually found on rats in tropical and subtropical countries 
is Palex cheopis Rothsch. When investigating the fleas which occur on 
any given host it appears to us to be essential to ascertain if there is 
more than one species associated with the spread of disease in the host, 
and to demonstrate precisely, in order to avoid confusion, the characters 
by which the various allied species can be distinguished. 
It is obvious that false conclusions will be drawn by investigators 
who maintain that numerous similar (though in fact specifically 
distinct) forms are identical. Such observers have usually examined 
their specimens with eyes untrained for the special line of investigation 
upon which they have entered, with the result that they deny the 
existence of differences which are obvious to the specialist. 
We shall, therefore, before describing the various species of Loemo- 
psylla with which we are acquainted, point out the reasons which 
guide us in the discrimination of the differences which we consider to 
be specific and those which we consider to be non-specific. 
The criterion of a “species” is a biological one. The fauna of 
a district consists of a certain number of units, each producing its own 
kind only; each being independent of the other in that respect; their 
own organisation only keeping these units apart. Such a unit we call 
the “ species.” Each species varies more or less and all individuals 
composing a species are different from one another. Again a species 
may be split up into well marked varieties, sometimes with and some¬ 
times without the existence of intergradations. All individual varieties 
are, nevertheless, specifically identical, standing either in the relation 
of brothers and sisters or parents and offspring. It is for the specialist 
to determine in each group which forms are mere varieties of one unit 
and which forms are species. The ultima ratio is breeding. In a group 
where breeding from the egg has been carried out, at least in some 
instances, the systematist knows which differences are specific in these 
cases and which are not; and it is perfectly legitimate—in fact the only 
possible course that can be adopted—to determine by analogy the 
taxonomic value of differences observed in other allied forms. In the 
case of Siphonaptera we are at all events certain as to the specific 
distinctness of a number of common species. The flea of the domestic 
fowl, that of the domestic pigeon, the one from the house martin, and 
the flea from the sand martin, though formerly considered to be 
1 For literature relating to the transmission of plague by fleas see Journal of Hygiene, 
vol. vi. p. 433 (1906). 
