Floor dust has been collected for bacteriological examination by three methods: (1) with a miniature vacuum cleaner in which the ‘bag’ consists of a paper Soxhlet extraction thimable, (2)by sweeping, and (3)by blowing the dust on to the surface of a culture plate. The dust on to the surface of a culture plate. the dust from on to the first two methods was shaken with broth and an aliquot sample plated out on crystal-violet blood ager plates.

Surface inoculation of the plates was found to give somewhat higher counts than the use of pour plates, but otherwise minor variations in the routine were without great effect on the count.

Analysis of the results obtained in examining a number of specimerns from schoolrooms and day nurseries showed that, though the counts of streptococci in successive samples from one dust suspension agreed well, there was wide variation between the counts from duplicate samples of dust, however collected; the frequency distribution of the counts in rooms was of the log-normal type. It is clear that in routine work, all the dust collected from the whole floor of a room should be examined, in order to give the most reliable result.

Of 185 specimens in which four colonies of haemolyt streptococci were grouped, 67% had all four strains of group A. Ofniety-four samples in which four group A strains were typed, 55 % had all four strains of the same type; more than one type was found in 20 %.