In a residential research ward coffee drinking was studied in 9 volunteer human subjects with histories of heavy coffee drinking. A series of five experiments was undertaken to characterize adlibitum coffee consumption and to investigate the effects of manipulating coffee concentration, caffeine dose per cup, and caffeine preloads prior to coffee drinking. Manipulations were double-blind and scheduled in randomized sequences across days. When cups of coffee were freely available, coffee drinking tended to be rather regularly spaced during the day with intercup intervals becoming progressively longer throughout the day; experimental manipulations showed that this lengthening of intercup intervals was not due to accumulating caffeine levels. Number of cups of coffee consumed was an inverted U-shaped function of both coffee concentration and caffeine dose per cup; however, coffee-concentration and dose-per-cup manipulations did not produce similar effects on other measures of coffee drinking (intercup interval, time to drink a cup, within-day distribution of cups). Caffeine preload produced dose-related decreases in number of cups consumed. As a whole, these experiments provide some limited evidence for both the suppressive and the reinforcing effects of caffeine on coffee consumption. Examination of total daily coffee and caffeine intake across experiments, however, provides no evidence for precise regulation (i.e., titration) of coffee or caffeine intake.