Variation and Heredity 291 



fittest species." It is, therefore, not the study of Linnaean 

 species that has a physiological interest, but it is the study 

 of the elementary species of which the Linnaean species are 

 made up, that furnishes the all-important problem for experi- 

 mental study. 



De Vries gives a critical analysis of a number of cases in 

 which new races have been formed under domestication. He 

 shows very convincingly that, whenever the result has been 

 the outcome of the selection of fluctuating variations, the 

 product that is formed can only be kept to its highest point 

 of development by the most rigid and ever watchful care. 

 If selection ceases for only a few generations, the new form 

 sinks back at once to its original level. Many of our cul- 

 tivated plants have really arisen, not by selection of this 

 sort, but by mutations ; and there are a number of recorded 

 cases where the first and sudden appearance of a new form 

 has been observed. In such cases as these there is no need 

 for selection, for if left to themselves there is no return to the 

 original form. If, however, after a new mutation has ap- 

 peared in this way, we subject its fluctuating variations to 

 selection, we can keep the new form up to its most extreme 

 limit, but can do nothing more. 



Another means, frequently employed, by which new varie- 

 ties have been formed is by bringing together different ele- 

 mentary species under cultivation. For instance, there are a 

 large number of wild elementary species of apples, and De 

 Vries believes that our different races of apples owe their 

 origin in part to these different wild forms. Crossing, culti- 

 vation, and selection have done the rest. 



De Vries points out some of the inconsistencies of those 

 who have attempted to discriminate between varieties and 

 species. The only rule that can be adhered to is that a 

 variety differs from a species to which it belongs in only one 

 or in a few characters. Most so-called varieties in nature are 

 really elementary species, which differ from their nearest 



