130 THE DEVELOPMENT OF EAW COTTON 



tion is poor, only those plants which happen to be 

 thoroughly well suited to the environment will produce 

 tolerable lint, and the presence of this lint in the sample 

 will increase the irregularity of the sample. 



Thus, if a commercial variety is well cultivated, it 

 tends to greater regularity; and if badly cultivated, to 

 greater irregularity. Since the brightness, cleanliness, 

 lustre, etc., of the lint are all indices of good cultivation, 

 they also go hand in hand with an approach to regularity. 

 On the other hand, no amount of good cultivation can 

 make a short-staple plant into a long-staple one, so far 

 TheLimita- as our P resen ^ knowledge of growth and 

 tions of Good especially of senescence can avail us. Con- 

 Cultivation, gequently, any approach to real uniformity 

 is impossible with impure varieties. It may possibly be 

 thought that undue insistence is here being laid on vari- 

 etal impurity; that the persons who introduce new 

 cottons are not likely to introduce them in an obviously 

 mixed condition; and that in speaking of such impurity 

 the author is applying some hyper-critical botanical test. 

 There is a very simple way of presenting data for varietal 

 composition in respect of two commercial characters 

 simultaneously, based on the statisticians' Correlation 

 Diagrams, which we may term Target Diagrams, since 

 the scatter of dots over the diagram is used in the same 

 way as the shot-pattern of a shot-gun. We have seen 

 that all these measurable characters, such as lint tength 

 and ginning out-turn, fluctuate to a definite degree round 

 a mean value, within a pure strain. If, therefore, we 

 grow a family of pure-strain plants, under uniform treat- 



