This was the condition of the subject when it was seriously taken up at Rothamsted. Before turning to agriculture, the writer had been studying the mechanism of certain slow chemical oxidations, and one of his first experiments in agriculture was to examine the phenomena of oxidation in soil. The results accorded with the biological explanation of Schloesing: when the soil was completely sterilised oxidation almost ceased. But the striking discovery was made, as the result of an accident to an autoclave, that partial sterilisation increased the rate of oxidation, and therefore presumably the bacterial activity. This remarkable phenomenon had, however, already been observed, and it had been shown that both bacterial numbers and soil fertility were increased thereby. A full investigation was started in 1907 by Dr. Hutchinson and the writer.[9] From the outset the phenomena were recognised as dynamic and not static, and the rates of change were always determined: thus the bacterial numbers, the nitrate and ammonia present were estimated after the several periods. Close study of the curves showed that the chemical and bacterial changes were sufficiently alike to justify the view that bacteria were in the main the causes of the production of ammonia and of nitrate; although non-biological chemical action was not excluded, there was no evidence that it played any great part. Thus the importance of micro-organisms in the soil was demonstrated. The factor causing the increased bacterial numbers after partial sterilisation was studied by finding out what agents would, and what would not, allow the numbers to increase, e.g. it was found that the bacterial increases became possible when soil had been heated at 56° C., but not at 40° C. Again, it was shown that the high numbers in partially sterilised soils rose for a time even higher if a little fresh untreated soil were incorporated into the partially sterilised soil, but afterwards they fell considerably. Putting all the results together, it appeared that some biological cause was at work depressing the numbers of bacteria in normal soils, but not--or not so much--in the partially sterilised soils. Studied in detail, the data suggested protozoa as the agent keeping down bacterial numbers, and they were found in the untreated, but not in the treated, soils. The hypothesis was therefore put forward that bacteria are not the only members of the soil population, but that protozoa are also present keeping them in check, and therefore adversely affecting the production of plant food.