358 GEORGE JOHN ROMANES 1393 



selection is not wanted at all [i.e. if all variations 

 are definite in Nature]. 



Moreover, it is contended that climatic variations 

 are of no great, even of any useful, importance. 

 This may be so, for all I know, with animals ; but it 

 is precisely the reverse with plants. I took my illus- 

 trations from desert plants, and showed that their 

 remarkable characteristics, which give the fades to 

 desert plants, are on the one hand the direct results 

 of the excessive drought, heat, light, &c. On the 

 other, they are just those features which enable the 

 plants to live under their extremely inhospitable 

 environment. These characters are the minute 

 leaves, hardening of woody tissues, thick cuticle, 

 dense clothing of hair, wax, storage of water tissues, 

 &c. ; so that the whole economy of the plant, in- 

 cluding its specific characters, is all climatically 

 acquired. Although some may vary when the plants 

 are grown in ordinary gardens, such is no more than 

 one would expect on a priori grounds to be the case. 



I would limit natural selection, as far as plants 

 are concerned, to three things : 



1. Mortality among seedlings with the survival of 

 the strongest. 



I do not say ' fittest,' because it is ordinarily 

 understood to mean that the survivors have some 

 morphological features, by which they are benefited, 

 which lead on finally to specific characters. 



I do not find this to be the case. Take an 

 instance of great contrast. Sow 100 seeds of the 

 water (submerged) Ranunculus fluitans in a garden. 

 They all grow up as aerial plants, i.e. they vary as 



