NATURAL SELECTION V. ADAPTATION. 



77 



recent utterances of botanists and zoologists it is evident that the mathe- 

 matical and statistical study of variation is destined to occupy a prominent 

 part in the development of biological research."* Dr. Davenport is 

 reported to have said : — " The science of variation is therefore one of 

 those that we hope to see established in this century. I feel convinced 

 that statistical studies are first of all necessary to lay the foundations of 

 the science." 



I cannot help thinking that this belief is erroneous, being probably 

 based on Darwin's mistake in supposing individual differences to be a 

 source of the origin of species. The author referred to "treats of the 

 numerical variation found in the ray-florets of Helianthus annum, 11 1 He 

 examined 1103 heads and found he could represent the numbers in 

 maxima of frequency. The most prominent, about thirty times as high as 

 others, was 21 ; but numbers oscillated about 13 and 34. 



Those who are familiar with phyllotaxis will at once recognise these as 

 representing the " angular divergences," T % / x , and Jf ; indeed they are 

 just what might have been expected on a priori grounds. The author's 

 object was to search for some bearing on evolution, whereas there is none 

 at all, these high fractions being merely the result of crowding. The 

 same fractions represent the arrangement of the scales of Fir cones &c. 



Other writers have given statistics of the number of parts of flowers, as 

 Mr. Cockerell did with the carpels of the Marsh Marigold,J but it leads to 

 nothing ; any number, more or less about an average, may be expected in 

 •examining hundreds of individuals as the result of varying amounts of 

 nutrition. Such do not lead to varietal forms, much less to species. 

 The explanation is that plants do not, or very rarely, grow with any 

 mathematical exactness ; so that mathematical statistics can hardly be 

 expected to lead to practical origins of new forms. Consequently 

 systematists invariably ignore these individual " differences " as not being 

 strictly "varietal" in amount. If they be recognised at all, it is by 

 expressing such variable numbers ; thus, e.g. 4-6, which would apply to 

 the number of the petals in individual flowers on one and the same 

 •corymb of Elder. It is only a mere question of nutrition. 



Another example is worth giving. We know that the Wild Carrot, 

 Parsnip, and Radish do not vary ; but under cultivation they have given 

 rise to several "races." Now, when the experiment was made of raising 

 these plants in a prepared garden soil from seed taken from wild in- 

 dividuals, several " forms " appeared to each. Thus M. Carriere in 

 experimenting with Baphanus Raphanistrum has figured some dozen 

 different shapes of roots having also different colours. Similarly, 

 Professor Buckman found the seed of the wild Parsnip gave him several 

 types. These might be called " indefinite variations." In a sense they 

 were: i.e. they were merely exaggerated "individual differences." 



Now follows an important observation. Nature was only doing in a 

 garden what she does elsewhere, only in a far less pronounced form ; 

 that is, variability is stimulated by the changed conditions of life. The 



* M Numerical Variation of the Ray Florets of Composite?" Bot. Gas. vol. xxxiii. 

 j). 462, June 1002, by E. Mead Wilcox, Alabama Polytechnic Institute, Auburn, 

 t Other illustrations will be found in Biometrika. 

 X Nature, March 21, 1895, p. 487. 



