A "barrel" is an interconnected network of layer IV neurons that is an important component of a functional cortical column in the whisker area of the rodent primary somatosensory cortex. The present study was undertaken in order to resolve apparently conflicting findings from single-unit studies of barrel neurons conducted in rats maintained under different anesthetic conditions. Multiunit responses to controlled deflections of mystacial vibrissae were recorded from the whisker/barrel cortex of awake, undrugged rats, and responses at the same recording site were reexamined after the animal was anesthetized with urethane. In contrast to the awake condition, stimulus-evoked responses under urethane were characterized by a large late component. Such effects were more pronounced for deflections of noncolumnar or "adjacent" whiskers than for the columnar whisker. Latencies to peak responses were virtually identical for the columnar whisker in awake and urethane states (11.9 vs 11.8 ms) but were considerably longer for adjacent whisker deflections in urethane-anesthetized animals (15.5 vs 29.0 ms). The magnitudes of adjacent whisker responses, relative to the response evoked by the columnar whisker, varied with the laminar location of the recording site in awake but not in urethane-anesthetized animals; in awake rats, receptive fields were clearly smallest in the layer IV barrels. Results in the awake condition confirm those of previous studies conducted in unanesthetized or lightly sedated animals, and data obtained with urethane are comparable to others' results in urethane-anesthetized rats. The former have important implications for how barrel cortex processes information in behaving animals.