All Pseudomonas aeruginosa strains produce Sabath and Abrahams' (SA) enzyme, as inducible β-lactamase. Hydrolysis of cefotaxime by this enzyme, although slow in terms of V max was efficient when the parameter V max /Km (physiological efficiency (Pollock, 1965)) was considered. Hydrolysis of cefsulo–din was not detectable in assays used and enzyme binding (Ki) was poor, ensuring a very low physiological efficiency. Physiological efficiency represents a measure of enzyme function under low substrate (antibiotic) conditions, as apply in the periplasm; consequently SA enzyme might protect the cell against cefotaxime, but not cefsulodin. This depends on enzyme induction and retention within the periplasm. Agar checkerboard studies indicated the SA inducer cephaloridine antagonized the activity of cefotaxime against most (12/14) Ps. aeruginosa strains but cefsulo––din against only a minority (12/14). Cephaloridine/cefotaxime antagonism was lost in uninducible (SAI/) or constitutive (SAPcom) mutants where SA expression was independent of cephaloridine concentration. This indicated the antagonism was SA dependent. Cefotaxime plate MICs against parent SAI+/and uninducible SAI−/organisms were similar indicating cefotaxime did not induce SA enzyme in these tests. Cefotaxime was however much less active against the SAI+ organism than the SAI− when log phase broth cultures were exposed to antibiotic and incubation wasextended to 30 h. This correlated with observed SA induction in the SAI+ organism. Cefsulodin MICs against SAI+/and SAI– organisms were similar and no difference existed in activity, over long periods against broth cultures. Overall, results indicated that SA enzyme, if induced, constituted a defence against cefotaxime but not cefsulodin and this correlated with the physiological efficiency results. To what degree induction occurs in vivo during cefotaxime therapy of pseudomonal infections remains unknown.