96 THE DEVELOPMENT OF KAW COTTON 



the probable uncertainty of the mean of all the observa- 

 tions according to a definite law, it follows that this rise 

 in strength, which culminates on August 9, is not due to 

 deficient precision in the methods used, but was a real 

 rise which the plants actually experienced. 



We can now go back to the other alternative, assume 

 that the general antagonism between length and break- 

 ing strain in this series was mere accident, and see 

 whether our knowledge of the structural development 

 will help. We decided in the previous chapter that the 

 most rapid increase in lint length took place about the 

 fifteenth day, while in the thickness of the lint hair wall 

 it took place at about the thirty-sixth to the thirty-ninth 

 day. 



We will first consider the fortunes of some flowers 

 opening after July 29, remembering that this date was 



near the end of a long period of water 

 Length de- 

 graded by shortage, when the plants were scarcely re- 

 Water taining any of their flowers, and were prob- 

 Snortage. a biy mO re or less poisoned or senescent. 

 Flowers opening on August 29 were in their fourth day 

 of development when the plot was irrigated, which we 

 have seen in the last chapter would imply that their lint 

 was about 10 millimetres long; they went on to maturity, 

 and produced lint of 28-2 millimetres length. Flowers 

 which opened after this day, up to those which opened on 

 August 11, were of course younger when the water was 

 given ; those which opened on August 1 1 were then young 

 buds which had only just begun to form their pollen- 

 grains, for example. The younger these buds were when 



