Moderate caffeine consumers (n=64, mean caffeine intake 453 mg/day) were deprived of caffeine overnight and semirandomly allocated to four treatment groups, designated PPP, CPP, CCP and CCC, where P is placebo and C is caffeine (1.2 mg/kg, giving an amount of caffeine similar to that consumed in a serving of ground coffee). Caffeine or placebo (i.e. no caffeine) were administered double-blind in novel fruit juice drinks at 10:15, 11:30 and 13:00 h on the test day. Before (baseline), and 45min after each of these times the participants completed a mood questionnaire and begun psychomotor performance tests lasting 25min (1-min tapping task, and a long-duration simple reaction time task (SRT)). Caffeine significantly increased energetic mood and improved psychomotor performance relative to placebo. Caffeine had particularly marked effects on SRT performance, ameliorating the slowing of performance with time on task and removing the post-lunch dip in performance. However, the three caffeine treatments, CPP, CCP and CCC, were equally effective. That is, mood and performance were improved to the same extent by one, two and three spaced doses (totalling 86,172 and 258 mg) of caffeine. This result is consistent with previous findings indicating a flat dose-response relationship for the psychoactive effects of caffeine; and because of the adverse effects (e.g. fatigue) associated with overnight caffeine deprivation, it suggests that there is little net benefit to be gained from frequent caffeine use. At the very least, it appears that the psychostimulant effects of caffeine cannot on their own account for the typical pattern of consumption of caffeinecontaining drinks. Behav Pharmacol 1998; 9:611-618 © 1998 Lippincott Williams & Wilkins