E. J. Russell and A. Appleyard 395 



numbers, and continuing after the numbers have begun to fall. It 

 might be argued that the similarity is only accidental and that the 

 curves for nitrate and for bacterial numbers have nothing to do with 

 one another; in other words that the bacterial numbers as recorded 

 by this method are not related to the decompositions in the soil. This 

 view is, we think, ruled out by the regularity mth which the result 

 has been obtained during the last three years (1914 — ^1916). The 

 method of counting is admittedly faulty, and it has not yet reached 

 the precise quantitative stage ; before it does this not only must it be 

 improved very considerably, but the part played in the decomposition 

 by moulds and fungi of the soil must be ascertained. In spite of its 

 defects, however, experience shows that it indicates the fluctuations in 

 numbers and gives results in fair agreement with, those obtained by other 

 methods^. On the whole, therefore, we prefer to explain the pheno- 

 mena on the assumption that bacteria are the active agents and that 

 the curves are connected. Admitting the truth of the relationship its 

 obvious meaning is that the formation of nitrate is dependent on some 

 previous change, which in turn is dependent on the bacterial numbers. 



There are two possibihties ; one might suppose that the first change 

 is ammonia production, which goes on simultaneously with the increase 

 in numbers, and that this is followed by nitrate production, which is 

 independent of these particular organisms. The lag between the curve 

 for bacterial numbers and for nitrate would then represent the time 

 required for the conversion of ammonia to nitrate. This view has in 

 its favour the fact that the ammonia-producing bacteria are countable 

 by the gelatine plate method, while the nitrifying organisms are not. 

 But we consider it to be ruled out because it is inconsistent with another 

 fact, already demonstrated here, that the amount of ammonia in the 

 soil is always a minimum, and consequently the rate of conversion of 

 ammonia to nitrate is greater than that of ammonia formation, so that 

 there cannot be a lag. The dividing up of the reaction must therefore 

 go further back and the formation of ammonia supposed to involve 

 two stages, the first being brought about by bacteria capable of grow- 

 ing on the plates, and therefore fluctuating according to the numbers 

 there recorded, while the second is subsequently, and somewhat more 

 slowly, brought about by other organisms or in another way. 



This hypothesis cannot be tested in the field, but only in the labora- 

 tory ; we hope to make the investigation at an early opportunity. 



^ See E. J. Russell and H. B. Hutchinson, this Journal, 1913, 5, 215 et seq. 



