SUMMARY:A statistical experiment has shown that no significant proportion of the errors of determination of bacterial populations by the surface drop method could be attributed to the dropping technique employed. With both pure and mixed cultures the numbers of colonies/drop followed a Poisson distribution very closely.It is suggested that in order to determine bacterial populations at a consistent level of accuracy a constant number of colonies should be counted rather than a constant number of drops. Whenever it is justified, the confidence limits for the results of counts should be quoted.