132 THE DEVELOPMENT OF RAW COTTON 



less oval, and instead of having one darkest spot in the 

 centre, it will show two spots. 



Again, if we take a pure strain which is suspected to 

 have been contaminated by crossing, and make the 

 diagram for all the plants, we shall probably be able to 

 detect the hybrid rogues, since it is probable that their 

 out-turn and length will be unlike that of the pure strain, 

 and consequently the dots representing them will be 

 likely to lie well outside the group. If we have four 

 such measurable characters we can plot six such diagrams, 

 and be practically certain of finding the rogue in one or 

 more of the six. 



With this description we may further examine the target 



diagrams (Fig. 16) which show the composition of the 



Im urit of P rmc ip a l commercial varieties of Egyptian 



Egyptian cotton, not in respect of any abstruse 



Varieties, botanical features, but in the directly com- 

 mercial characteristics of lint length and ginning out- 

 turn. Side by side with them are plotted (Targets 5, 6, 

 7, 10) the target diagrams for pure strains, to compare 

 the amount of scatter which need exist with that which 

 actually does exist. The objection may be raised that 

 the various plots might have received different treat- 

 ment, hence accounting for the different degrees of scatter, 

 but this is not the case; when the plants are grown for 

 such comparison, they are all mingled together on the 

 same piece of land, plant by plant, so that all share equally 

 in any variation of soil or of cultivation. 



One or two points in these diagrams, though of par- 

 ticular interest to Egyptian cotton-growers, have also some 



