E. J. Russell and A. Appleyard 397 



Turning to the connection between bacterial numbers and soil con- 

 ditions, the most striking fact is the failure of temperature to cause 

 a permanent rise in numbers. The usual result is that the numbers rise 

 in spring when the soil temperature first rises, but they then fall (Fig. 6). 

 There are subsequent variations, but neither the base line nor the 

 average are much above the winter Hne, and on the Broadbalk dunged 

 plots they fall below. This is completely different from plant growth, 

 which in the main follows soil temperature. The relationship is not 

 accidental, it has been obtained in the laboratory experiments here^. 

 Conn^ has obtained a similar result in the United States, the winter 

 level being higher than the summer; he explains it by assuming that 

 two sets of bacteria exist, — a summer and a winter set, — the latter 

 being more numerous than the former. 



It is possible that the flora undergoes seasonal variations, but even 

 if it does, this does not explain the whole of the facts, for Russell and 

 Hutchinson showed that the anomaly ceased as soon as the soils are 

 partially sterilised ; increases in the temperature then caused increases 

 in bacterial numbers. Their hypothesis of a complex soil population 

 seems more adequate. Lohnis and Smith^ have suggested that these 

 fluctuations may be due to life cycles, some steps in the cycle being 

 countable, but not others. As the cycles are not worked out it is 

 difi&cult to ascertain how far this new hypothesis meets the facts. 



The moisture curve shows only a general kind of relationship with 

 the biochemical activity, but it affords some explanation of the sluggish 

 period. During the period of active accumulation the moisture on the 

 dunged plot fluctuates between 18 and 21%; during the sluggish 

 period it is only 15 to 18 %. The connection is not altogether close 

 because for a period in May the soil moisture rises to 20 %, but there 

 is no corresponding rise in nitrate accumulation. But in the main 

 active nitrification only occurs when some 18 % of water is present. 



The low period following the spring rise does not fit in with any of 

 the recorded data, nor does the high period in autumn when rain first 

 comes after a time of drought ; the drop in the first case and the rise in 

 the second seems more than the conditions require. These phenomena, 

 however, are exactly parallel to those observed by Russell and Hut- 

 chinson in their studies of partial steriUsation, and no doubt are due to 

 the same cause. It is not possible to discuss these fluctuations in 



1 See E. J. Russell and H. B. Hutchinson, this Journal, 1913, 5, 157 et seq. 

 ' Conn, H. J., Genir. BaU. Par. 1910, Abt. n. 28, 422-434. 

 3 Lohnis and Smith, Journ. Agric. Research, 1916, 6, 675-702. 



