What's the best way to find out what crows living in the South Pacific really do when people aren't watching? Equip them with mini-cameras and have them make their own home-, er, nest-movies, of course.

University of Oxford zoologists are hoping that this hands-off approach to studying New Caledonian crows—aka Corvus moneduloides—will lead to a wealth of information about these infamous aviators, known to be one of the few nonhuman species to use tools to accomplish daily tasks.

These birds are particularly difficult to observe in their natural habitat, because they are sensitive to human disturbance and live in forested, mountainous areas of New Caledonia—a Pacific island east of Australia, roughly the size of New Jersey—where visibility is limited. But researchers report in the online edition of Science that they captured on video the birds using sticks, grass and stems to forage for food.

"The use of tools is one of the defining criteria for humans," says Christian Rutz, a postdoc candidate at Oxford's Behavioral Ecology Research Group, who conducted the study. "When it was revealed that here was a bird species that does something that's supposedly exclusive to humans, and to some extent chimpanzees, that was pretty important."

Rutz and fellow Oxford researchers Lucas Bluff, Alex Weir and Alex Kacelnik taped tiny camera units (which included a 2.4-GHz video transmitter, voltage regulator, timer chip and VHF radio-tag, along with batteries) to the two inner tail feathers of 18 crows to learn more about how they use tools. Each unit weighed about 14 grams (0.5 ounce), which was key given that the crows themselves each weigh only about 320 grams (11 ounces). "It was low-tech, but we wanted to have a fairly weak attachment so the crows could remove the camera if they wanted to," Rutz says. The camera lens protruded through the feathers and shot images from between each crow's legs, a position that researchers say did not interfere with the birds' movements and allowed them to shed the equipment via the normal molting process.

Sure enough, the video cameras captured images of the crows using 10- to 15-centimeter (four- to six-inch) sticks to probe holes where they would find grubs. Rutz and his team used the footage to determine that the birds forage around eight small food items per hour. This indicates, Rutz says, that the birds may have begun using implements to overcome challenging foraging conditions. The researchers also discovered that the birds' tool-use and choice of tool materials—including sticks and dry grasslike stems—were more diverse than previously believed, and that birds will save, rather than discard, tools that they find particularly useful.

"One of our main goals with this was to observe natural, undisturbed behavior," Rutz says. "If you want to understand why these crows developed this unusual behavior, namely tool use, you have to understand what role these tools play in their daily lives."

Rutz and his team crafted the camera units themselves, because they could not find any technology that could be affixed to the birds without inhibiting their movement. "We pieced together the smallest equipment we could find," Rutz says. The VHF radio-tag let the researchers track the movement and behavior of individual crows, whereas the video footage was broadcast as a microwave signal.

The timer chip was crucial because it prevented the camera from switching on too soon and running out of battery power before capturing meaningful behavior. "When you first release a bird, the first thing it will do is sort out its feathers," Rutz says. "We didn't want to film this because we only had about 70 minutes worth of battery life." In fact, the average length of the footage recorded by each of the cameras was 38 minutes.

The researchers rounded up 18 crows in about a week using a baited five- by three-meter (roughly 16- by nine-foot) net placed on the ground. They ended up getting footage from 12 of those birds.

The researchers relied on the VHF radio-tags to find the cameras after the birds shed them but wound up recovering fewer than half because each tag's signal lasted only about three weeks. Rutz says he and his colleagues did not have to retrieve the cameras to recover their data (although each unit was worth several thousand dollars), because the video feeds were broadcast to a receiver that fed the footage into a camcorder.

The crows filmed themselves using tools to search for grubs in the rotted wood of fallen trees and to fish for ants in the ground. Researchers were surprised to discover that they used dry grass stems in addition to sticks and leaf stems. "We now know we were getting only a fraction of the whole picture," Rutz says. "This cautions us to broaden our view."

Rutz and his team are considering strapping cameras onto dozens of other crows in an effort to create a comprehensive video catalogue that will help them understand, among other things, whether females or males use tools more. "Does each crow have its own tool kit or are there crows that serve as tool specialists for the group?" Rutz asks. "These are important questions."

Rutz plans to rejoin colleague Lucas Bluff in New Caledonia within the next week to continue their studies. They have already begun to work on improving the camera units and are hoping that their "crowcam" technology can be used to study other types of birds and animals in the wild. "I'd like to see video tracking become routine for field ornithologists, or biologists in general," he says.