18 



No. 3995, namely, 89.39 per cent, and the highest is found in No 

 3978, namely, 96.58 per cent. This sample therefore may be con 

 sidered as being more nearly pure sugar and water than any other of 

 the samples produced. 



SUMMARY OF FOUR-YEAR EXPERIMENT IN FERTILIZATION 0: 

 SUGAR CANE, 1902-1905. 



EXPERIMENTAL FIELDS. 



The experimental work on this subject having been concluded, n 

 is desirable to compare the data for the four years in order to get 

 view of the general effect .of the different methods of fertilizatioi 

 upon the yield of cane and upon its relative content of sucrose am 

 reducing sugar. The two fields on which the experimental cultui 

 were made are designated as A and B. The fields marked A and 

 however, w T ere not the same pieces of land, except for the years 19( 

 and 1905. The fields marked A and B for 1903 and 1904 were differ- 

 ent pieces of land. Therefore 110 direct comparison can be mad< 

 between fields A and B for the whole period, but only for the yei 

 1902 and 1905. 



Field A, for 1902 and 1905, is an old field continuously in cultivj 

 tion for twenty years, of a sandy soil, containing very little plant 

 food other than that supplied by the fertilizers. Field B is a land ( 

 similar composition, but only in cultivation six years, and therefoi 

 contains considerablv larger proportions of plant food than the old( 

 field. 



THE FERTILIZERS USED. 



Before beginning the comparisons it is well to submit a table show- 

 ing the average composition of the fertilizers used during the time 

 the experiment. This table contains the average composition of tl 

 fertilizers by direct analysis for the years 1903 and 1904. A direct 

 analysis of the fertilizers used in 1905 was not made, but as the fei 

 tilizers were of similar character they do not differ materially from t In 

 average of the three years, which is contained in the following table 



TABLE VII. Average analyst of fertilizer* nl in experimental work in 1902-1904. 





