Tiles t 22 mm Γ 22 mmt made of glass, polyvinylchloride, polystyrene, stainless steel, and Formica were colonised with Staphylococcus epidermidis NCTC 11047 or Pseudomonas aeruginosa PaWH, by submersion, for 2 min, in inoculated Tryptone Soya broth (37':C). After colonisation, the tiles were triple rinsed in saline before being transferred to the surface of pre-dried Tryptone Soya agar plates. After 1 min in contact with the agar surface the tiles were transferred to a second plate and the first plate was spread. The transfer process was repeated through 15 plates. Tetrazolium plates incubated with the tiles in situ indicated that very few organisms remained on the tiles after successive blotting. In all instances, the number of recovered colonies decreased exponentially with plate succession number according to the relationship CFU = A x 10 ~~', where CFU is the number of colonies transferred, N is the plate succession number, A is the intercept and k is an exponent describing the ease of removal of the cells. The gradients (kt, derived from eight replicate tiles, differed significantly between organisms (t" >0.95) and depended on the nature of the test surface. The hydrophobicity of the conditioned surfaces was determined as water conlact angle. AUC calculations, based on the relationship CFU = A Γ 10 ~ and integer wilues of N, showed R aeru<~inosa to colonise the test surfaces to a greater extent (cells cm 2) than S. epidermidis. The topography of the suithces was assessed using laser profilometry, expressed as R~ and lo values. The extent of colonisation was independent of substratum-hydrophobicity and surface roughness for both organisms. Removal exponents (k), however, indicated that whilst P. aeruginosa was attached most firmly to the substrata, only the strength of attachment of S. epidermidis was dependent upon substratum-hydrophobicity.