Graduation date: 2006I measured responses of free-ranging Rocky Mountain elk (Cervus elaphus) to recreational disturbance at Starkey Experimental Forest and Range, Oregon from April to October, 2003 and 2004. Resting, feeding, and travel activities of 13 cow elk were recorded at 5-minute intervals using Actiwatch™ motion sensors. Elk were subjected to four types of recreational disturbance: all-terrain vehicles (ATV), mountain biking, hiking, and horseback riding. Individual disturbance activities were recorded for five consecutive days following a nine day control period of no human activity. Elk alternated their activity budgets between feeding and resting bouts during the controls, with little time spent traveling. Travel time increased during the disturbances and was highest in the mornings. Traveling was significantly different among disturbances and was greatest for ATV, followed by mountain biking, hiking, and horseback riding. Feeding time decreased during the ATV disturbance and resting decreased during mountain biking and hiking in 2003. Little or no reduction in feeding or resting was evident during hiking in 2004 or for horseback riding during both years. Elk returned to behavior patterns similar to those of the controls once each disturbance ended. There was less travel time during disturbances in 2004 compared to 2003, suggesting that elk became habituated to these recreational activities. However, travel time during 2004 remained above that measured during the control periods. For each of the four treatments I collected visual observations on the distance (m) from an observer that elk took flight (flight distance) and the type of vegetation elk occupied for each flight distance. Radio-collared elk locations were used to estimate mean distance (m) from observer GPS locations when elk displayed movements greater than the control periods (i.e., a flight response). Visual detection rate of elk was depended upon the treatment; the greatest numbers of elk observations were for horseback riding (128), followed by hiking (67), ATV (47), and mountain biking (35). Using direct visual observations, I found no significant difference between the four treatments in elk flight distance. This was in contrast to the activity sensors where a difference between treatments in the time elk spent traveling was detected. Direct observations also produced significantly shorter mean flight distances compared to those of the GPS/radio telemetry estimates. It is likely that direct observations of elk in this study underestimated the effects of recreational disturbance on their behavior patterns. The more detailed activity sensor and GPS/radio telemetry results provide managers with information that can be used for balancing objectives for off-road recreation with those for elk.