Three studies investigated the ability to perceive caffeine via taste or bodily effects, under different concentrations and in different brands of coffee. In Experiment 1, 40 subjects ingested capsules containing 0, 200, 400, or 600 mg of caffeine. After a 90-min delay, they made magnitude estimates and category ratings of the caffeine ingested. Results showed significant linear trends between actual and judged caffeine levels. In Experiment 2, 89 subjects tasted samples of coffee with concentrations corresponding to the doses of Experiment 1. The estimated levels of caffeine again showed a significant linear relationship to actual levels. In Experiment 3, 63 subjects tasted 24 samples of coffee and judged which of two brands each sample was and whether each was caffeinated or decaffeinated. Brand, but not caffeination, was identified with better than chance accuracy. It was concluded that the ability to perceive caffeine in coffee is lost somewhere between 200 mg/cup and 50 mg/cup.