Attempts to measure the hydrophobicity of the cell surfaces of Gordonia amarae and Rhodococcus erythropolis, filamentous bacteria found in wastewater treatment plants, by several methods--microbial adhesion to hydrocarbons (MATH) or bacterial adhesion to hydrocarbons (BATH), contact angle, and micro-sphere adhesion to cells (MAC)--were unsuccessful. The results were erratic and inconsistent. This was in part because of the filamentous growth habit of G. amarae, but it was also a consequence of the fact that the 'hydrophobicity' of bacterial cells is not a clearly defined quantity. A technique is introduced in which bacteria are suspended in solutions of synthetic surfactants (non-ionic, cationic and anionic), and the suspensions aerated under defined conditions. The partitioning of bacterial cells between the foam and liquid phases was reproducible. The method was tested in model systems in which the bacteria were replaced by silica particles with defined surface modifications. Although this technique is not a direct measure of 'hydrophobicity', the partitioning of cells depends in part upon their surface hydrophobicity. In addition, qualitative information is gained about ionic interactions between the bacteria and the bubble surface. The results are pertinent to the problem of foaming in wastewater treatment plants.