1 Concentrations of phenytoin in mixed, parotid and submandibular saliva and serum were determined in normal subjects after an oral dose, using a specific double antibody radioimmunoassay which requires only 20 micronl fluid. 2 Semi-log concentration-time plots of phenytoin concentration in mixed saliva and serum gave good parallelism after the initial 14 h post-administration period. 3 The mean ratio of the mixed saliva: serum phenytoin concentration was 10.3% +/- 1.5 (s.d.) in seven normal subjects. 4 Phenytoin concentrations found in separate parotid and submandibular salivary fractions did not differ but were significantly greater (P less than 0.001) than those found in mixed saliva. 5 Phenytoin concentrations in all salivary fractions were independent of the volume of fluid produced and the degree of stimulation. 6 The rate of phenytoin secretion in the parotid and submandibular fluid was proportional to the salivary flow rate. 7 These data suggest that mixed saliva may be a suitable medium for the monitoring of phenytoin concentrations and may provide a non-invasive alternative to the direct determination of phenytoin in serum.