Mills, 50, who is trim with sandy blond hair and whose own beard is turning white from the middle outward, began studying snowshoe hares in 1997. Just four years out of a Ph.D. program at UC-Santa Cruz -- where he studied the effects of forest fragmentation on voles, working with the prominent conservation biologist Michael Soulé -- and two years into an assistant professorship at the University of Montana, he began studying hares because they are a major food source for the Canada lynx, which was slated for listing under the Endangered Species Act. "Of course, I believe hares are more important than just as grocery items for lynx," Mills said, sitting in his creaky Missoula office in the university's 1920s-era Forestry building. "But that's how I got drawn in."

The project involved a relatively simple management question: Would certain types of logging disrupt the hare population, and ultimately threaten lynx survival? To answer it, Mills spent several years engaged in the slightly gruesome business of fitting hares with radio collars and essentially waiting for them to die. He'd record where and when they died, and, if possible, what ate them.

Out in the woods tracking hares, he began to notice certain patterns. During full moons when there was snow on the ground, the hares tended to move around much less than at other times -- and they were also less likely to live to see the dawn. "When they're glowing under a full moon on snow," said Mills, "they respond behaviorally. And there is a cost to making the wrong decision: They die." He found a similar pattern in open patches of forest, including clear-cuts: Hares of either color moved more freely, and survived longer, in closed-canopy parts of the forest than in the open. In other words, the hares' visibility has a big impact on their survival.

Mills also noticed a third pattern, one that seemed counterintuitive at the time: Far more hares died in the fall and spring than in winter or summer. "That was surprising to me when I first saw it," he recalled, "because you'd think, in winter it's 20 below zero, the predators are hungry, the snow is deep -- shouldn't hares die more in the wintertime?" But this was 10 years ago, before anyone was thinking about mismatches between animals' camouflage and their habitats.

Because the same amount of snow doesn't fall on the same date from year to year, and because it can take so long for hares to acquire their seasonal coats, inevitably there are times when the animals lack camouflage -- they're brown in a white world or white in a brown one. This was always true, even before humans intervened with global warming.
Recently, though, anecdotal reports of mismatched hares have been on the rise. Over the past decade, as word of his research has spread, Mills has received a growing number of autumn phone calls from hunters. ("They'll say, 'Hey, Scott, I was out deer hunting last week and I saw one of your white bunnies out hopping around,' " he said.) Many of the veteran hunters tell him the same thing: The snow is arriving later. "The old-timers will say, 'We used to always have snow by the first of November, I'd always be hunting deer on snow, and now it's really rare to have an opening day,' the third week of October, 'where there's snow on the ground.' "

If hares start turning white at the same time each year but the snow keeps arriving later, and if the hares are more likely to be eaten when they don't match their surroundings, then their future prospects could be bleak. Snowshoe hares, with their long legs, slender ears, and outsized feet, could become one more victim of climate change. In this scenario, as winter shrinks, the poor snowshoe hare will fall increasingly out of sync with the new seasonal cycles; eventually, like the polar bear, it'll become another cover model for collapsing ecosystems.

There's just one problem with this simple version of things: We have no scientific evidence that it's true.

Given that species that change their coat color -- weasels, ptarmigans, Arctic foxes, hamsters -- live across much of the world, our understanding of them remains exceptionally sketchy. We don't really know whether hares that are mismatched to their environment are more likely to die; it's entirely possible that more hares die in spring and fall for some other reason -- like changes in forest cover or shifting diets. We don't even know, in fact, whether there really are more mismatched hares now than there used to be.

It's even unclear whether the hares know if they're camouflaged. Coming upon one that's white when the ground is bare, or brown when the forest is snow-covered, is "like the most embarrassing thing that can happen in the field," said Mills. "It's like walking in on someone taking a shower. Because you look over and there six feet away from you is this totally white hare that's just sitting there. And it feels like that hare is thinking, 'Oh, yeah, he can't see me.' But then other times" -- when the hares bolt -- "it seems like they really can look down and say, 'Yikes, I'm mismatched!' So we're trying to quantify it."

One thing is certain, though: On average, there are fewer days of snowpack than there used to be. In fact, it was at a 2007 lecture on global warming's regional impacts that Mills had an epiphany about the future of his research. Steve Running, a Missoula colleague who received the Nobel Peace Prize for his work with the Intergovernmental Panel on Climate Change, put up a slide showing that the biggest sign of climate change across the world's temperate regions wasn't a change in temperature, but a decline in the number of days with snow-covered ground. "When I saw that slide of his," said Mills, "it all came together for me to realize, wow, OK, so these hares are changing no matter what, and they're confronting this reduction of snow on the ground. That's when it all crystallized."

Mills' excitement about this important new line of inquiry was quickly tempered by regret: Suddenly, his previous research represented a dozen years of missed opportunities. He hadn't collected a single piece of information on coat color. More than four years after Running's talk, he's still kicking himself. "I had all these radio-collared hares in the late '90s," he lamented, "but it just didn't occur to me to record their coat color when they were killed.

"And to put salt in the wounds," he continued, laughing, "one of the first times I submitted a grant on this project and I said that we've had 175 hares that were collared to look at sources of mortality, one of the reviewers wrote, 'It's too bad Mills didn't actually record whether or not the hares that got killed were mismatched.' Yeah, I know it's too bad!"