G 



'I'lKtsc r:inii('i-s \\li(» art' more ju-cusIoiikmT to tin* woi-J 

 aiiimoiiia limn to ilu^ tfU'iu iiitrog(Mi, can change? (lie 

 ligiiirs for nilr(»g('n into their aniiiionia equivalents by 

 multiplying by 1 {\. 



Tlie phosphate and cotton seed >Nere purchastMi at 

 niarkci ])rices. Most of the kainit >vas donated by the 

 CJernian Kali AA'orks. 



In determining the increase over the unfertilized 

 plots, the yield of the fertilized plots, Nos. 4, 5, G and 7, 

 is compared with both unfertilized plots, Ij'ing on either 

 side, giving to each unfertilized X)lot a weight inversely 

 proportional to its distance from the plot under com- 

 parij«-on. This method of comparison tends to compen 

 sate for variations in the fertility of the several plots. 



It should be remembered that seasons, as well as soils* 

 determine the effects of fertilizers, so that to be abso- 

 lutely reliable a fertilizer experiment should be repeated 

 for several years on tlie same kind of soil. Abnormal 

 weather conditions in 1S99 and 1900 resulted in an un- 

 usually large proportion of inconclusive experiments. 



THE WEATIIEK IN 1899 and 1900. 



The following data are taken from the records of the 

 .vlabama Section of the AVeatlier Bureau for 1S99 and 

 1900 and give average results of a number of stations: 



1899. 1900. 



Kainfall for April, inches 2.80 9.00 



Kainfall for Mav, inches 2.03 2.64 



l^ainfall for June, inches 2 54 11.80 



Iiainfall for July, inches G.7() 4.93 



iiainfall for August, inches 3. OS 2.89 



Kainfall for September, inches GO 4.00 



Kainfall for October, inches 2,18 5. 04 



Kainfall for November, inches 3.04 3.88 



