INVESTIGATIONS ON THE PRODUCTION OF PLANT FOOD. 189 



(2) Measuring the rate at which ammonia or nitrate is formed in 

 the soil. 



(3) Estimating the changes in numbers of bacteria in the soil. 

 The first two can be determined as accurately as is desired, but 



the accurate estimation of bacterial numbers is not yet possible and 

 the values are comparative only ; nevertheless they are of considerable 

 value for our work. 



First of all we may take it as a general rule that the soil organisms, 

 being living creatures, are dependent on suitable temperature and water 

 supply, that they must have food, and also sources of energy to enable 

 them not only to live but also to carry on those reactions which involve 

 the accumulation of energy, or, in other words, resemble the rolling 

 of a ball up a hill. These are general requirements that can safely 

 be predicted of any living organism. In addition there is the special 

 requirement that has been discovered by experiment : the need for 

 calcium carbonate, without which many soil organisms will not act 

 efficiently. 



The application of general rules to soil problems is a very treacherous 

 business; it is commonly the unexpected that happens, and experi- 

 mental confirmation is therefore required at every stage. In order 

 to get at the general nature of the effect of temperature and moisture 

 content on the decomposition process it is necessary to do experiments 

 in the laboratory, where all the conditions can be carefully controlled. 

 Experiment shows that the effect of rising temperature on the bacterial 

 numbers is quite different from what one expects : instead of rising, the 

 numbers remain fairly constant up to about 8o° F., and then they 

 begin to fall. 



April 



May 



June 



July 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



Jan. 



Feb/ 



, Mar. 



April 







































A- 











l/\ 







s 



\ 



\ 









/' 



% 

 % 









. . // 





1 \ 



* V 

 \ \ 







\ 







<s- 



/ 



✓ 



r 



V 









// 



y 1 



<r 5 





* 1 

 \ 1 



M 

 1 1 



4-B 





> 











3— — 











11 































Fig. 61. — Bacterial Numbers in Soil at Different Seasons of the Year. 



[Plots ib and 4B. Cropped with millet : unmanured. The curve showing 

 moisture content is very similar to that for 4B (Conn.)] 



In like manner, increases in water content of the soil do not lead 

 to regular increases in bacterial numbers ; there is a rise at first, 

 but it is not sustained. So in natural conditions the numbers 

 of bacteria do not show the expected fluctuations in the rise 

 in temperature or moisture content. The discrepancy has been 



VOL. XLI. O 



