Summary Liquid-elution counts in a non-ionic surfactant and imprint replicate counts with velvet pads were used to investigate the changes in total bacterial and colony counts on normal skin surfaces after bathing with water, soap, or detergent. The experiments revealed that the total bacterial count on several skin sites was reduced soon after washing, returning in 24 hr to the pre-wash numbers. In contrast, the colonial count from the same sites increased considerably after washing, again returning to the pre-wash count. An explanation is offered that both kinds of counts were sampled from extremely large aggregates of bacteria on the skin surface, and that these aggregates could be segregated into very much smaller viable units, each still capable of originating individual colonies by either counting method. It is suggested that the extent of separation depended on the mechanical processes of washing, the presence of surface-active agents in the washing water, and the use of a non-ionic surfactant in the sampling and subsequent counting procedure.