2 Darbishire, Mendelian Principles of Heredity. 



anything can be imagined to be). Above all, cases of 

 continuous variation can be described by curves of error, 

 a fact which seems to me to have a very deep meaning. 

 Discontinuous variation is the quality which animals and 

 plants possess of giving rise, from time to time, to offspring 

 bearing characters specifically different from those of 

 their parents : in fact by some these offspring are 

 termed ' sports.' This ' sport ' is a new species : that is 

 to say (to use Bateson's words) " Variation, in fact, is 

 evolution " ('99 p. 6) ; but while this can be said of dis- 

 continuous, continuous variation is merely the material 

 upon which natural selection operates (if we may thus 

 personify that process). From this it follows that 

 according to the latter view it is impossible to say v/here 

 one species begins and the other ends ; but that according 

 to the former view there is no such difficulty. 



Continuous variation may be looked upon as normal, 

 while discontinuous may be regarded as abnormal ; but 

 this aspect of the matter is perhaps only justifiable as a 

 help to understanding the difference. For if species have 

 arisen by discontinuous variation it is hardly fair to call it 

 abnormal. ( Vide Ewart '99 p. Ixxxiv.) 



The word " Variation " may represent an abstract 

 or a concrete thing : in the continuous sense it is 

 usually a name given to the whole phenomenon of 

 variation : in the discontinuous, the so-called sport is often 

 spoken of as ^a variation ' (whereas a man with a cephalic 

 index slightly less than his neighbour's would hardly be). 

 This twofold use of the term conveniently recalls the two 

 conceptions of variation. 



(2) Gallon's theory of Heredity. 



I refer to Galton's theory of heredity because I want to 

 show how the Mendelian theory differs from the statistical 



