The purpose of this study was to examine attitudes toward the duration of survival in different health states. Health professionals were asked to consider hypothetical situations in which a specified period of time would be spent in a defined health state. Preferences for these different situations were assessed using two methods--a Preference Questionnaire and a Certainty Equivalence method. The results obtained indicated that attitudes toward duration of survival depended strongly on the amount of time to be spent in a hypothetical health state, and on the quality of the state. As the duration of survival increased, or the state under consideration became more dysfunctional, a higher proportion of raters expressed a preference for immediate death over further survival in the dysfunctional state. These results contradict the view that time preference curves always increase monotonically, independent of the quality of survival. Instead, subjects appeared to identify a variable, the "maximal endurable time" in a given state. When this time was exceeded, attitudes toward additional increments of survival changed dramatically.