Paleontologists and placer miners have dug up enough fossil bones to prove that large mammals once lived in Beringia. This ice-free refuge stretched from eastern Siberia across Alaska and the northern Yukon during the last Ice Age, when ice sheets covered much of the continent. But researchers still debate how grazing animals as big as mammoths found enough to eat in this cold northern land.

The so-called "Beringian Production Paradox" can be boiled down to a case of bones versus pollen: the unearthed bones prove that there was a diverse array of animals in Beringia, but pollen studies indicate that the vegetation was not productive enough to support many grazing animals.

Pollen samples taken from the northern Yukon indicate that Beringia had only sparse tundra vegetation, and tundra is not very productive. Based on this evidence, researchers have argued that Beringia could only have supported small numbers of grazing mammals. But other scientists contend that Beringia was covered by arctic-steppe, productive grassland unlike any found in the north today.

"The question is, in the middle of a glaciation, how can all of these animals live in what is apparently an inhospitable environment?" says Chris Burn. A geography professor at Carleton University, Burn has done extensive work on soils and permafrost in the Yukon. Based on a study he helped conduct in the Kluane area, he thinks that loess is a key to the production paradox.

Loess is wind-deposited dust, made up of fine particles of sand or coarse silt. Deposits of loess are usually found in areas that once bordered continental glaciers. It's estimated that about 30 percent of North America has at least some loess covering it. Loess renews the nutrients in the ground it falls on, enhancing plant growth.

During the spring and early summer, meltwater from the Kaskawulsh Glacier carries large quantities of soil particles into the Slims River. When the water level later goes down, exposed sediments along the riverbanks dry out. Winds blowing down the valley pick up the finer particles and carry them away.

Burn and the other researchers wanted to test whether loess deposits could have formed productive "hot spots" with the forage needed by big grazing mammals. The study was conducted along the southern shores of Kluane Lake, where loess from the Slims River delta is steadily being deposited.

A number of study sites were picked along a gradient extending away from the Slims River delta. The sites were mainly grasslands that had developed on clearings in the forest, and many had been used for grazing horses. Earlier studies had shown that loess deposits all over the Kluane Lake region make the area significantly more productive than surrounding regions.

Burn, Nicky Laxton, a graduate student, and Scott Smith, of Agriculture Canada, examined the soils and the vegetation at 30 sites along this gradient. They found that loess deposits decreased with increasing distance from the delta. They also found that soils with more silt in them supported thicker growths of vegetation.

Based on this study and the work of others, Burn says he is convinced that Beringia was a mosaic of steppe communities. He pictures a landscape with a fair amount of tundra, but pockets of grasslands in some valleys. Sedges and grasses would have grown there, as well as some shrubs, but he does not think that many trees could have survived during the glacial period.

Burn thinks that along the edges of the glaciers, a combination of water and loess could have produced "a vibrant environment with lots of plant productivity. Where glacial meltwater flowed in large valleys to the north and west of the ice sheets, there would have been areas where animal populations could have been abundant," he says.

Burn thinks that the grazing mammals in Beringia probably moved fairly regularly, feeding in one place and then moving on to the next site. Barrenground caribou survive in the North today because they migrate for huge distances, lowering their dependence on the vegetation in one area.

Burn says that another key to the Beringian environment was its dryness, particularly in the last glacial interval. In a dry climate without much snowfall, the sun's energy can be used to warm the soil rather than melt snow, and provide a suitable environment for plant growth. The loessal soils found in Kluane today are also warm and dry.

Burn says that the saiga antelope is one clue that not much snow fell on Beringia. Measuring about one metre at the shoulder, Burn said that this animal's height gives a lot of information. "If there was deep snow it would not have been able to travel and would not have been able to find food buried by the snow. So that tells us that there was very little snow and grasses would have been peeking up through the snow for most of the year," he says.

Burn says that when you look at the evidence, it is hard to accept that all of Beringia was a hostile barren environment. He points out that the pollen samples used to support the theory that Beringia was covered by tundra were taken from the Richardson and Barn Mountains in northern Yukon.

"There is still not much growing there. It is a pretty grim place today, although it can be beautiful too. But then, 400 kilometres south at Dawson, the environment is totally different. The glaciers did not reach the Klondike, and there was lots of loess there that might have supported a more varied fauna," he says.

Professor Chris Burn can be reached through Carleton University.