406 Deconqwsition of Organic Matter in Soil 



Experimental. 



The mechanical analysis, specific gravity, and other data for the 

 soils are given in this Journal, 1915, 7, 44. 



The methods of samphng and of analysis are the same as those 

 described in the earlier paper. 



The determinations have in all cases been made at intervals of 

 about 10 days but more frequently during critical periods. The results 

 are then plotted, and the points joined to give the curves here shown. 

 This method has the drawback that it gives the appearance of a con- 

 tinuous record which, however, has not been attempted; on the other 

 hand it is so much more convenient than the alternative and more 

 correct method of drawing columns to indicate the separate points that 

 we consider its advantages on the whole to out- weigh the disadvantages. 



The data are given in Table I, and the curves in Figs. 1-5, but 

 the main fluctuations may be briefly indicated. 



Broadbalk Unmanured Falloiv (Fig. 1). 



This represents the simplest case; it is part of the Broadbalk un- 

 manured plot which instead of being under crop as usual was this year 

 (1915) divided into two parts, one cropped, and the other fallowed. No 

 manure has been added since 1839, and nothing but wheat has been 

 grown since 1843, so that the land is about as exhausted as possible. 



The winter level for bacterial numbers is about 6 to 9 millions per 

 gram; when the temperature rises in April this runs up to 21 millions. 

 But it speedily falls, and during May is about 10 millions and in June 

 below five — the lowest point of the year. Throughout July, August, 

 and September the numbers are approximate constant about 10, then 

 comes a rise in October, and a drop to the winter level in November 

 and December. 



The CO2 in the soil air begins in the same way ; the winter level is 

 0-1 to 0-15 %; then comes a rise in April to over 0-25 %, near which 

 it keeps till May 17th, when the land was ploughed and a fall set in. 

 During July there is a considerable rise to 0-45 % in August followed by 

 a fall to the minimum in October, and then a rise to 0-2 % in November 

 and December, when a new crop was beginning to grow. 



The nitrates were not satisfactorily determined during the winter 

 months so that the first record is in April at the rather high point of 

 over 50 lbs. per acre, then comes a fall, and then a rise till June when 

 a sharp drop sets in to 42 lbs. per acre, and again a rise to 67 lbs., about 



