METHODS OF INVESTIGATION 173 



against these could be plotted the corresponding data of 

 each day for temperatures, sunshine, humidity, wind, 

 evaporation, soil- water content, subsoil- water level, or 

 any other data which might be relevant to the subject 

 under in vestiga t ion . 



Presenting the state of the crop thus in the form of 

 an ideal average plant at once brings the agricultural 

 problems into directly botanical form, or, rather, pro- 

 pounds those problems in terms of plant physiology. 

 With such exact records it is possible to trace most minute 

 differences, of which one example may suffice: 



In former years the author had ascertained the period 

 of boll maturation for several varieties and strains by 

 marking open flowers, and watching for their date of 

 opening. Among these was a pure strain. No. 77, which 

 in 1910 had a maturation period of forty-eight days, with 

 a Probable Error of 3 per cent. In 1913 material was 

 taken for these lint-development investigations from this 

 strain, but no direct record of the maturation period was 

 repeated ; the question arose as to whether the change of 

 plot and of year had altered the maturation period, or 

 whether^ forty-eight days could still be taken as correct. 

 A full set of daily observations of flowering and boiling 

 upon the variety " Domains Afifi " were smoothed to 

 five-day means, and the two curves superposed; they 

 fitted with an interval of fifty-one days. The process 

 was repeated with the relatively imperfect data from the 

 observed group of No. 77, and indicated a slightly shorter 

 period; then the dates of irrigation of the group of No. 77 

 >vere taken, fifty-one days added to them, and the dates 



