SELECTION 151 



slightly different species e.g. rose, bramble 

 and hawkweed owe their protean character 

 to their variations being of no service or 

 disservice, and consequently not being acted 

 on by natural selection.) In determining 

 whether groups of similar forms should be 

 ranked as species or as varieties, the opinion 

 of naturalists of sound judgment and wide 

 experience is the only guide, yet this lacks 

 unanimity : for example, of the polymorphic 

 genera (i.e. rich in species with a small range 

 of differences) in the British flora alone, Ben- 

 tham reckons 112 species, but Babington 251, 

 Wallace has shown that no certain criterion 

 can be given by which to define his own con- 

 venient working categories of Malayan butter- 

 flies and moths, viz. variable forms, local 

 forms, sub-species, and representative species. 

 As De Candolle concluded from his mono- 

 graph on oaks (in which he shows at least 

 two-thirds of his 300 species to be provisional), 

 " so long as a genus is imperfectly known and 

 its species founded upon a few specimens " 

 they seem clearly limited; but, "just as we 

 come to know them better, intermediate 

 forms flow in and doubts as to specific limits 

 augment." The terms variety and species 

 are thus arbitrarily applied to indefinable 

 groups of more or less closely similar in- 



