The responses to the Jenkins Activity Survey of groups of 40 short sleepers, 40 long sleepers, and 40 variable sleepers were compared. As predicted, it was found that the variable sleepers showed significantly less Type A behavior than either of the other two groups. These data further elaborate the possible role of sleep behavior in the management of stress.Recently, the San Jose group (Hicks, Pellegrini, Martin, Garbesi, Elliott , & Hawkins, 1979) reported finding a significant inverse relationship between normal habitual sleep duration and level of Type A behavior as defined by score on the Jenkins Activity Survey (JAS) . In discussing the implications of this relationship, we speculated that the relatively higher stress susceptibility of Type A individuals might, in part, be a consequence of the relatively shorter sleep durations that tend to be characteristic of these individuals. That is, the Type A person tends to adopt the habit of sleeping less, which in turn may reduce the amount of REM sleep but leave intact the deeper stages (3 and 4) of NREM sleep.In this regard, Hartmann and Brewer (1976) demonstrated that the sleep duration of a group of variable sleepers tended to fluctuate as a function of the stressfulness of their life situations. Specifically, their variable sleepers reported needing more sleep during periods of increased " mental or emotional stress." In commenting on this result , Hartmann and Brewer suggested that it was probably the requirement for REM sleep that changes as a function of the need to resolve worry or stress and , hence , the need to change the duration of sleep.While the results of Hartmann and Brewer's (1976) study are not incongruent with the aforementioned inverse relationship between normal habitual sleep duration and score on the JAS, their data suggest a possible qualification of that relationship . To elaborate, Hicks et al. (1979) systematically excluded self-reported variable sleepers from consideration as subjects. Thus, their results may be somewhat misleading in the sense that stable sleep duration was emphasized as the sleep factor that seemed to be implicated in the formation of Type A behavior. Clearly, Hartmann and Brewer's (1976) comments suggest that, because of their ability to adjust sleep appropriately to the stressfulness of life, variable sleepers should show less Type A behavior than nonvariable sleepers, regardless of mean sleep duration.Therefore, the purpose of this research was to deter-We are grateful to Signe Gary for her help in preparing this report. Requests for reprints should be addressed to