This report explores the importance of characterizing short-term (peak) exposures for successful epidemiologic study of acute reversible effects. By selecting upper respiratory tract irritation, which is time-limited, reversible, and not critically affected by host factor sensitivities, the illustrated approach (directed at an acute respiratory irritant) provides insights on issues important in the general study of acute effects. The fundamental concern in a study of acute irritant effects is that such events occur in response to abrupt increases in exposure. This calls for a closer examination of how to distinguish bursts, or peaks, of exposure from continuous exposures measured simply as time weighted averages. Three parameters of peak exposures are discussed: duration, magnitude, and frequency. The importance of duration is illustrated by contrasting exposure patterns relevant for the study of acute vs. chronic health effects. The minimum magnitude of a peak relevant to the study of a specific physiologic response requires knowledge of the threshold below which no detectable response occurs. For any identified threshold level, only exposures exceeding that level are defined as relevant peaks for an exposure-response analysis. Finally, the issue of frequency, the minimum length of time between successive exposures necessary for the subject to return to baseline risk, is explored. A study of respiratory irritation associated with sodium borates is used to illustrate these issues.