Increase in Number of Species 187 



as an isolated breeding population on the new host. Otherwise 

 genetic mixing would continue between populations on the two 

 hosts, no genetic divergence would develop, and no independent 

 phylogenetic line would evolve. In two insect groups suspected of 

 being parents of sympatric fission species, these objections are 

 circumvented. 



In the transfer of the rodent flea to the owl, adult fleas could have 

 been accidentally introduced on prey into the burrows by the 

 foraging owls, the fleas could have laid eggs there, and a population 

 of rodent fleas emerged in the owl nest. This sort of happening 

 undoubtedly occurs frequently without a change of host, the adult 

 fleas leaving the owl burrows in search of their normal rodent 

 host. However, on at least one occasion the fleas adopted the 

 owls as a host, became established on them, and evolved into a 

 distinctive species. Presumably the owl-liking fleas maintained a 

 series of populations in owl burrows effectively isolated genetically 

 from the parent populations of fleas on rodents. The isolation here 

 would be linked with the fact that the flea larvae develop and 

 pupate in the host nest; hence each nest is almost an isolated micro- 

 population as long as the nest is continuously occupied. 



Prey brought into the nest would introduce occasional indi- 

 viduals of the old rodent flea species into the new owl-nest popula- 

 tions. Presumably here the same forces and processes would 

 prevail as postulated for the colonization of geographically separated 

 areas (see p. 168). 



Complex multiple-host transfer is found in three North American 

 species flocks of the leafliopper genus Erythroneura, almost entirely 

 confined to the eastern deciduous forest region. In three lines of the 

 E. comes flock (Fig. 81), a species feeding on grape (Vitis) gave 

 rise to a sister species on redbud (Cercis). In the E. maculata flock 

 ( Fig. 82 ) , species arising from oak-feeding ancestors became estab- 

 lished on sycamore, maple, elm, and at least ten other host genera. 

 In the E. ohliqua flock, an even more complex series of host 

 transfers have occurred. Each detectable host transfer led to the 

 establishment of a new phylogenetic line, and some of these have 

 evolved subsequently into two, six, or as many as 40 species (Ross, 

 1958Z?). In these three species flocks, totalling about 500 species, 

 about 150 host transfers are known. 



The great bulk of the species are widely distributed, frequently 

 extending as far as their hosts. This circumstance does not fit a 

 pattern of unusual past fragmentation of ranges. Furthermore, the 

 earliest probable age of this cluster of flocks is the same as that of 



