COTTON-GROWING 157 



built up day by day over a period of two or three months, 

 ordinary casual (or even skilled) observation breaks 

 down. 



The first step in developing cotton-growing for a new 

 country should be the procuring of records showing how 

 the yield was built up under optimum conditions. This 

 involves also determination of the optimum conditions. 



The optimum conditions of the site and year in which 



the trials were made are very simply defined as those 



which produced the largest crop, which also 



Experiinen- _ other things being equal will be the best 



I ill 1 1 lit IS. 



crop. The conditions which produce the 

 best crop can only be ascertained by trial and error, but 

 there are right and wrong ways of so doing ; much of this 

 work as done to-day consists of very little trial and very 

 much error. Certain conditions of the environment can 

 be controlled in the trials, while others cannot. Weather 

 is uncontrollable as regards temperature, and only as 

 regards water when there is no rain; in irrigated land 

 the water-supply should be made one of the subjects of 

 experiment. Although the weather cannot be changed, 

 its incidence on the plant may be changed, by making 

 the time of sowing another subject of experiment. The 

 soil may be modified by manurial treatment, though 

 experiment in such direction should be restricted initially 

 to those methods which are likely to be practicable for 

 the general crop. Lastly there is the arrangement of the 

 plants on the area cultivated, which is of more impor- 

 tance than has been realized in the past; there can, for 

 example, be little doubt that the American crop would 



