Ot) PROCEEDINGS OF THE NATIONAL MTSEt^I vol. los 



neglectus v>-as collected from the nest of another subspecies of the 

 Holarctic H. ruMica, namely H. r. erythroga^ira. Both subspecies 

 of barn swallow build sim.ilar nests in similar places, and it may 

 well be assmned now that H. rustica is the true host of C. a. 

 affinis. 



The differences in structure between these two forms of barn swallow 

 fleas are fairly pronounced, and some workers would perhaps consider 

 them to be full species. However, the differences between, e. g., 

 Ceratojphyllu^ styx siyx Rothschild and C. s. jordani Smit, which are 

 certainly subspecies and are parasites of the sand martin Riparia 

 riparia in Europe, are even more pronounced than those between the 

 two subspecies of C. affinis. Both barn swallow and sand martin 

 build very isolated nests and they live in a considerable degree of 

 isolation from other hosts. It is an acknowledged fact that isolation 

 favors evolution and this may explain why the differences between 

 subspecies of fleas which live in identical isolated ecological habitats 

 (e. g., the fleas of the swallow and the sand martin) are greater than 

 those between subspecies of fleas infesting hosts which occupy less 

 isolated ecological habitats (e. g., the two subspecies of the tree- 

 squirrel flea Tarsopsylla octodecimdentata, T. o. octodecimdentata 

 (Kolenati) in the Palaearctic region, and T. o. coloradensis (Baker) 

 in the Nearctic region). Thus the degree of differences between 

 C. a. affinis and C. a. neglectus might be interpreted as being the 

 result of the occurrence of these fleas in isolated habitats and therefore 

 strengthens the assumption that C. a. affinis, as well as C. a. neglectus, 

 is a parasite of Hirundo rustica. If the true host (and the ecological 

 factors concerning its nest) of one of the subspecies should prove to 

 be different, it would be necessary to regard affinis and neglectus as 

 full species, for, apart from considerations of geographical, morpho- 

 logical, and reproductive isolation, it seems that in fleas we must 

 usuaUy also stipulate that subspecies of a certain species must occupy 

 almost identical ecological niches, since it seems extremely probable 

 that if two fleas have diverged so much as to be able to occupy sub- 

 stantially different niches they will also have become too distant to 

 have retained the capacity to interbreed without impairment of 

 fertflity. Thus, though the structural differences between Xeno- 

 psylla ramesis (Rothschild) (a parasite of gerbils) and X. cunicularis 

 Smit (infesting the rabbit in Morocco) are very small, and the two 

 forms appear not to be sympatric, I regard them as full species be- 

 cause of the very different nesting habits of their hosts. 



It may be that another form, more or less intermediate between 

 C. a. affinis and C. a. neglectus, occurs in eastern Siberia. Since bam 

 swallows are widely distributed and travel over great distances, it 

 seems likely that the distribution of the subspecies of C. affinis will 



