G. A. CowiE 125 



that of the controls. This indicated that the ammonia was originating 

 from a slight decomposition of the compound itself as the direct effect 

 of the reagents, and it was accordingly decided to test the accuracy of 

 the method in presence of the nitrogenous dressings used. Equivalent 

 amounts of nitrogen were, therefore, added to the soil in the same 

 proportions and determination of the nitrates made forthwith before 

 any nitrification of the added materials could take place. The results 

 showed the method was subject to the following errors with the different 

 dressings: 



I. Cyanamide 

 II. Cyanamide 3 parts ] 



Dicyanodiamide 1 part ) 

 in. Cyanamide 1 part ) 



Dicyanodiamide 3 parts I 

 IV. Dicyanodiamide 

 V. Ammonium sulphate ... 

 The above table should therefore be considered in association with 

 the nitrification results under Table IV and plotted in Fig. 1. It must 

 be noted, however, that with the rapid conversion of cyanamide to 

 ammonia, even in the presence of dicyanodiamide, the above errors 

 would be substantially reduced even at the first period with the first 

 and second dressings and would probably only continue to hold good 

 in respect to the proportion of dicyanodiamide present. For further 

 corroboration, however, the nitrates were also determined at the final 

 period by the phenol-sulphonic acid method, and the figures, also in- 

 cluded in the table, confirm substantially the results obtained by the 

 reduction method. 



In the subsequent tests the phenol-sulphonic acid method was 

 adhered to throughout. 



Percentage Nitrification of the added Nitrogen. 



To arrive at the percentages nitrified of the added nitrogen a difficulty 

 was introduced through the rise in the nitrate content of the control 

 during the progress of the experiment and the consequent inability to 

 determine what figure should be deducted from the amounts of nitrate 

 in the treated soils. As a comparison of the results will show, the 

 control starting with a relatively low nitrate content has shown in most 

 cases a steady rise in nitrates, amounting to lO-l.'j parts N per million 

 dry soil, according to the length of time the series was allowed to run. 

 In such a case it would be unjustifiable to assume that the nitrification 



