pressions of the same character. One must be careful, how- 

 ever, to distinguish between correlation and mere associa- 

 tion of characters. Correlated characters move together 

 —for example, short-beaked pigeons have small feet, hairless 

 dogs have imperfect teeth, and blue-eyed tomcats are deaf. 

 It is not always easy to determine to what extent char- 

 acters really move together. It is not safe to trust to mere 

 impressions of a correlation; not until actual measurements 

 and calculations are made can one be sure of the movement 

 together of characters. 



Correlations often exist in plants between botanical 

 marks and the bleeding qualities. Nilsson's work at 

 Svalof, in Sweden, has shown the value of such correla- 

 tions. Take for example, the selection of Primus barley, 

 of different strains of peas, oats, and clovers, etc. This 

 principle of correlation of characters is being followed up by 

 many experimenters. 



Burbank^ uses intuitively this principle in his large 

 selections, for he makes the majority of them while the 

 plants are in the seedling stage. He can "predict one qual- 

 ity or one function from the study of others." 



This correlation of characters brings forward again the 

 idea of unit characters which forms the basis of Mendelism . 

 The idea of correlation forces on us the assumption that the 

 unit may express itself in many ways — in the leaf, seed, stem 

 or tissue. "The correlated external marks may be but the 

 expression of the same internal character." 



The recent work of Morgan with Drosophila brings out 

 clearly the fact that "every change in the germ-plasm 

 (variation) affects not one but a large number of characters ; 

 and conversely, every visible character is the result of the 

 concurrent action of a large number of factor-differences 

 or variations. 



(b) — Coefficient of Correlation 



The ratio of correlation between separate'characters is 

 called the Coefficient of Correlation. Whe'n the coeffi- 

 cient is 1 the two characters involved are perfectly correl- 

 ated, when they are indifferent, and when -1 they are 

 mutually exclusive. 



When inter-relations are expressed in the form of a co- 

 efficient or ratio many errors of judgment, and truths as 



(1). — Luther Burbank (1849- ) lives at Santa Rosa, Cal., where he 



conducts his experiments on the breeding of new plants. Some 

 of his productions are the Burbank potato, Burbank plum, plum- 

 cot, seedless prune, royal walnut spineless cactus, etc. 



51 



