THE FERTILITY OF SOILS IN THE TROPICS 43! 



The average yield of the experimental plots (KPN), 

 taking the average of eight experimental plots for a 

 period of three years, was 345 bunches, with 3,035 dozen 

 bananas per acre per annum, at a value of say 38, the 

 cost of the fertilizer being about 12 los. per annum. 



In the experiments 2(KPN) we obtained an average 

 yield of 457 bunches, with 4,330 dozen, of a value of 

 54 per acre, showing a considerably increased net profit 

 over the yield obtained from plots (KPN). 



The record yield was given by experiment 12 in 

 1912-13 with thirty dozen per stool (second year's crop), 

 the same plot yielded last year only 14! dozen. Experi- 

 ment 20, which yielded 2i| dozen per stool in 1912-13, 

 gave in the last harvest 28^ dozen per stool. Very re- 

 markable is the yield of the two unmanured plots (experi- 

 ments 15 and 16), giving n and 17 dozen respectively 

 in 1912-13, and last harvest, up to March, 1914, 5$ and 

 i6 dozen respectively, showing clearly the influence of 

 the long dry spell during the year 1913 on the limed 

 and unlimed, unmanured experiments, and proving the 

 great advantage of liming. 



The owner of the experimental field also conducted 

 on his own account an experiment with a commercial 

 mixed fertilizer, frequently used for manuring in this 

 district. Using this fertilizer at the rate of 6 Ib. per 

 stool, or 16 cwt. per acre, he applied, expressed in our 

 standard, j K 3^ P i N, but the results were exceedingly 

 disappointing, clearly indicating the want of potash in 

 the mixture. 



With regard to the application of phosphoric acid, it 

 appears that both Thomas's phosphate and super- 

 phosphate give good results. In this class of red soil 

 the soluble phosphoric acid changes soon into an in- 

 soluble form of sulphate. As superphosphates are 

 more easily obtained, they were exclusively used for the 

 second series of experiments. 



Nitrogen acts apparently best in the form of dried 

 blood, and as nitrate of lime. 



The application of salt in addition to other manures 

 does not appear to make any appreciable difference in 

 the yields, and it appears that the plant gets a sufficient 

 amount of salt from the ordinary class of soil. 



