SUMMARY:A series of co‐operative experiments was conducted to compare the bacterial colony counts of soil obtained by workers in different laboratories, using soil extract agar and other media for the determinations.In the earlier experiments it was not possible to obtain a reasonable degree of reproducibility of results between laboratories even when the plating technique was carefully prescribed. By modification and more rigid standardization of the technique closer agreement was obtained in subsequent experiments. It is suggested that when co‐operative investigations are contemplated the participating laboratories should check their technique by the examination of ‘control’ soils.No evidence could be obtained to support the suggestion that higher colony counts are obtained by the use of soil extract media containing extract prepared from the same soil as the sample tested. The source appears to be immaterial so long as the soil for extract preparation is not of extreme type and has been well manured and cul