Current techniques for remotely monitoring wildlife lack the capability to survey a wide area or to transmit data in real time. We addressed these technology gaps by developing and testing a new video and telemetry system for remotely sampling wildlife abundance, distribution, and behavior across large open areas. The system consisted of 2 pan–tilt–zoom video cameras equipped with 20–200× lens, and an automated telemetry scanner and data logger. All components were charged by wind and solar power and located on a hilltop overlooking an open valley (23 km2) in Yellowstone National Park, USA. A satellite up-link to the internet transmitted data in real time to the University of Minnesota-St. Paul and relayed commands from undergraduate students who controlled the cameras and systematically scanned the area at 2-hour intervals 6 times/day (0800–2000 hr) 7 days/week for about 20 consecutive weeks (Dec–Jun). During each scan, students recorded presence, activity, and location of bison (Bison bison), coyote (Canis latrans), grizzly bear (Ursus arctos), red fox (Vulpes vulpes), and wolf (Canis lupus). The telemetry system continuously scanned frequencies of 100 tagged wolves, including 4 members of the one resident pack. We determined wolf presence in real time by viewing the incoming data stream, or we assessed presence later after off-loading the data logger. Matched pairs of simultaneous observations taken by remote and on-site observers during a 13-day double-sampling period were highly correlated (r = 0.71–0.94), but remote observations were biased toward larger, more visible mammals (e.g., bison) and tended to underestimate their abundance. Nevertheless, the system was deployed longer than was practical for on-site observers and was, therefore, useful for detecting long-term, fine-scale trends such as daily changes in bison numbers as a function of snow depth.