One of the major difficulties in determining the cleanliness of blankets for use in hospital wards is the estimation of bacterial contamination remaining on the blanket after washing. Present procedures are either cumbersome or not readily reproducible. Procedures for measuring bacterial contamination on fabrics may be based on dispersion of the bacteria as an aqueous or as an airborne cloud. As aerial sampling has the advantages that it may be rapid and simple and can be used to sample blankets which have been washed with a bactericide the possibilities of this principle were investigated. A simple method has been developed (Blowers & Wallace, 1955) in which the blanket is scraped manually by the edge of an agar plate so that the bacterial dust is thrown on to the exposed agar surface, but this procedure is difficult to control. There are wide differences in surface properties of blankets, and consequently in the amount of fibrous material shaved off by the Petri dishes. Puck, Robertson, Wise, Loosli & Lemon (1946) worked on the principle of hitting an area of taut blanket and sampling the aerial cloud so formed, but as their apparatus was not convenient for routine sampling in wards an improved version has been developed.