CocKERELL: NorTH AMERICAN SPECIES OF HymENoxys 467 
superficial appearance. I have even found a species of Phileozera 
labeled Richardsoni! Those who could make such identifications 
ought not to feel very indignant at the way I have “lumped” the 
northern races under H. Richardsoni, and the southern annuals 
under H. chrysanthemotdes. ‘When one compares the typical 
chrysanthemoides with typical multiflora, the idea of specific unity 
seems ridiculous to the highest degree, although in the Avthe- 
mideae, at least, nearly as much difference may sometimes be pro- 
duced by the direct action of the environment. I confess that I 
cannot myself contemplate these two plants together without a 
strong feeling that they constitute a reductio ad absurdum of my 
classification. At the same time, when the whole series of plants 
is arrayed on a table so as to be directly compared, one does not 
know where to draw specific lines, and no good structural char- 
acters are apparent upon microscopical examination. I have tried 
consistently to follow the rule that species are not marked by the 
amount of divergence they exhibit, but by the fact of their discon- 
tinuity in nature. Species to me are like islands; sub-species are 
like peninsulas. A peninsula may be vastly more important and 
more distinct than the little island off the shores ; but nevertheless 
it remains a peninsula, and the island, however small, is an island. 
Thus in the Phileozera series, multiflora and anthemoides are great 
peninsulae, and Davidsonii is a little island close to the latter, much 
nearer to it in characters than it is to the chrysanthemoides main- 
land. 
I do not find any distinct evidence of the origin of species by 
mutation in Hymenoxys. On the contrary, the characters of the 
species are highly adaptive. Differences of environment appear to 
result in morphological differentiation, but when the environments 
are intimately connected, as they are in the north and south, not 
in specific segregation. When we find the plants growing on 
more or less isolated mountain ranges, as in New Mexico, there 
appears a strong tendency to form species, closely allied but defi- 
nitely separated. Thus isolation brings about the separation of 
“‘species,”’ without (in all probability) having anything to do with 
their differentiation. It is exactly similar to the case of mountain 
peaks which, submerged in the ocean, become islands, but the 
ocean had nothing to do with forming the peaks. Of course it 
