E. J. Russell and A. Applp]yard 411 



on the Harpenclen Field curves has gone further so that some of the 

 rises in the dunged plots become falls here, and vice versa. 



The bacterial numbers in winter time he between 15 and 20 millions : 

 in April they rise rapidly to 30 millions, the maximum for the year: 

 then they drop to a minimum of 12 milhons in June : they rise again in 

 July, following the high rainfall : then once more they drop, and again 

 rise at the end of September, also accompanying high rainfall, finally 

 they drop again. These fluctuations are similar in general character 

 to those on Broadbalk, and show the same initial connection with 

 temperature and subsequent relation with rainfall. 



The CO2 to begin with follows the curves for bacterial numbers, but 

 in May it shoots upwards to 1 %, and finally in August it becomes 2-5 % 

 of the soil air. Then there is a fall to 0-3 % in October, and the fluctu- 

 ations again follow the curves for bacterial numbers. Here, as in the 

 other cases, there is a great outpouring of CO2 in the soil at the time 

 of ripening of the crop and simultaneously a depression of bacterial 

 numbers although the rainfall curve would have indicated a rise. 



The nitrate begins in April at 50 lbs. per acre in the top 18 inches : 

 towards the end of the month it rises suddenly to 67 lbs. just as on the 

 Broadbalk unmanured plot : then comes a fall, and in June a great rise 

 to 94 lbs. followed by a rapid drop to the minimum, 35 lbs., after which 

 a rise sets in, then a fall, then after ploughing another rise, finally after 

 October a fall to 40 lbs-, per acre in the top 18 inches. The curve differs 

 from the Broadbalk dunged plot mainly in that it rises so much higher 

 early in July, which may be attributed partly if not wholly to the 

 dressing of sulphate of ammonia. 



The rate of nitrification of sulpJiate of ammo7iia in the soil. 



We had hoped to obtain definite information on this point from 

 these observations, but we were disappointed. The efEect of the 

 sulphate of ammonia is not at all clear. It was apphed on April 5th 

 and the amount used (1 cwt. per acre) contained 22 lbs. of nitrogen. 

 Twenty-four days after the apphcation the nitric nitrogen in the soil 

 rapidly gained 9 lbs. which might be attributed to the added ammonia 

 except for the circumstance that the Broadbalk unmanured plot showed 

 a similar rise. The only period when Harpenden field falls out of line 

 with the rest is as already stated in early July. It is difficult to beheve 

 that the ammonia had so long remained unnitrified if it had been within 

 the range of action of the bacteria. The explanation may be that it 



