Sleepy drivers should "take a break," but the efficacy of feasible additional countermeasures that can be used during the break is unknown. We examined a shorter than 15 min nap, 150 mg of caffeine in coffee, and a coffee placebo, each given randomly across test sessions to 10 sleepy subjects during a 30-min rest period between two 1-hr monotonous early afternoon drives in a car simulator. Caffeine and nap significantly reduced driving impairments, subjective sleepiness, and electroencephalographic (EEG) activity indicating drowsiness. Blink rate was unaffected. Sleep during naps varied, whereas caffeine produced more consistent effects. Subjects acknowledged sleepiness when the EEG indicated drowsiness, and driving impairments were preceded by self-knowledge of sleepiness. Taking just a break proved ineffective.