Conducting scientific research in the Arctic is expensive, hard and dangerous. Studying the creatures who live there sometimes calls for a novel approach, as Tim Flannery reveals in a piece for The New York Review of Books about the work of the conservationist Joel Berger:

Berger decided to try to determine whether musk oxen fear bears, reasoning that if they did, then bears must be significant predators. So he dressed in a bear costume and approached herds of musk oxen, recording their response. Just to be sure that it was the bear costume they were responding to, he also approached the same herds dressed in a caribou outfit.

Berger discovered that the approach must be made from at least a mile away and, like that of a bear bent on attack, it must not be direct. With a wind-chill factor of–15° C and a skin of ice over the snow, on his first attempt Berger took an hour and a half to get within forty-five yards of the herd. Then a bull charged—from twenty-five yards away. Instinct kicked in, and he tossed the head of his bear costume skyward, causing the confused bull to halt. Berger then struggled through the deep snow toward his colleagues, who were approaching on their snowmobiles.

The astonishing thing is that Berger did not give up but repeated the exercise again, and again and again, over deep snow, sharp rocks, and permafrost, enduring hours of agonizing cold. At most, he got to record two encounters per day, but often only one. Over the years, he built a data set of more than one hundred encounters and got charged “seriously” by bulls four times. Always, in the back of his mind, a question lurks: What if, while dressed in his costume, he meets a real bear?