The nature and occurrence of carry-over effects, i.e. the response to a stimulus is influenced by previous samples, were examined for selected bitter compounds. A time-intensity procedure was used to rate the bitterness of six compounds (caffeine, denatonium benzoate, limonin, naringin, quinine and sucrose octa-acetate). For each subject concentrations of these compounds were determined that were approximately equal in intensity to 1.18 x 10(-5) M limonin. To test carry-over effects of each compound the 36 paired sequences (pairs) were evaluated. Within a session three pairs were tested, between which two-stage rinses were used to remove any effects of the previous pairs. Within a pair only water rinses were used between stimuli. For all compounds carry-over or sensitization effects were observed in which values for maximum intensity, rate of onset and total area under the time-intensity curve were higher for a compound when tested in the second position than in the first. In addition, the degree of sensitization and susceptibility to sensitization were compound-specific. Caffeine increased the bitterness by the largest amount for all other compounds, while it was least affected. Regardless of the compound in the first position, the bitterness of quinine and denatonium were most enhanced.