The ability of subjects to attend to vibrotactile patterns presented to two fingers was explored in several experiments. The patterns were generated by two 6 column x 24 row tactile arrays. In one set of measurements, the arrays were placed in contact with the subjects' left index and middle fingers. Both discrimination and identification tasks suggested that there was an attentional deficit in processing patterns presented simultaneously to these two fingers. A pattern presented to either the middle or index finger was identified both more accurately and more rapidly than was the same pattern divided in half and presented to two fingers. Patterns were also presented to two fingers on opposite hands, and performance measures were taken on patternidentification tasks, on discrimination tasks, and on tasks that required subjects to combine pattern information from two fingers. All three measures showed performance to improve when the patterns were presented to two fingers on opposite hands relative to when they were presented to two fingers on the same hand. The results are interpreted to suggest, first, that subjects can process patterns simultaneously presented to two fingers on the same hand, but with some deficit due to attentional mechanisms, and second, that information from patterns presented to two fingers is processed differently depending on whether the two fingers are on the same or on different hands. There is much less of an attentional deficit in processing patterns presented simultaneously to two fingers on opposite hands.There has been relatively little work on the role of attention in the perception of tactile stimuli. In considering a task such as detection of a signal, the question of whether or not a subject can attend to more than one site has received mixed answers. Franzen, Markowitz, and Swets (1970) concluded that subjects could not attend to more than one finger at a time, whereas Craig (1968) concluded that subjects could combine weak signals from two fingers, implying that the two fingers could be attended to simultaneously. Subsequently, Shiffrin, Craig, and Cohen (1973) presented data to support an unlimited attention model for detection of simple stimuli by the skin. These studies did not, however, deal with the question of whether and how subjects could attend to patterned stimuli presented to more than one site at a time. Here the question is more than simply whether or not a stimulus was presented at a particular site; it is whether or not information from more than one site can be combined.It has been demonstrated that subjects can use information presented to several sites on the skin at the same time, implying an ability to attend to clearly suprathreshold stimuli presented simultaneously to more than a single location. In one study, subjects were presented with pairs of patterns consisting of up to 10 vibrators, contacting 10 different sites on the body, acti-