DARWINISM 159 



Como ; but at a height of 400-500 feet up the mountain side 

 above Cernobbio a form is developed during the height of the 

 flojiiering season with thirty-four rays, reverting at the end of it 

 to the twenty-one-rayed type. On the Mount Resegone above 

 Lecco, there is a still larger species, presumably with thirty-four 

 or more rays, which being called a species, is, I suppose, more 

 or less fixed, but I have not found it. At Taormina, where 

 the extensive observations on Calendula were made, I think, at 

 the end of them, if a batch of flowers gathered in one spot had 

 been given me, I could have inferred the habitat, at all events 

 approximately, by counting the rays." 



This writer does not allude to the biometrical observations 

 going on in Europe, but has studied them, independently, from 

 the adaptive point of view ; i.e., upon which I have urged the 

 importance above. 



The following, however, gives an illustration of an attempt 

 to combine locality with numerations. 



In a paper on the Variation and Correlation in Lesser 

 Celandine, from Divers Localities, ^ the study of the table gave 

 the following conclusions : — 



" I. Local races in plants cannot be defined or distinguished 

 by the existence of differences many times the value of their 

 probable errors between the means, variations or correlations. 



" 2. The influences of environment and season are for 

 plants of supreme miportance and very widely or indeed 

 entirely screen any differences due to local race." 



As an example : While three sepals with eight petals were 

 common in Dorsetshire ; five sepals with ten petals prevailed 

 at Gais (Switzerland). 



Again, selecting poppies from five different localities ; their 

 different standards of deviation became apparent from i"4S5 

 to I '898 upon which Prof. Karl Pearson observes : — 



" We have here illustration of how race, environment, selec- 

 tion, correspond to numerical differences in type and variability." 



He also suggests that one should "count in some hundreds 



^ Biometrika ii., p. 144. 



