Résumé (en anglais):

In this paper, we report initial findings from an ongoing isotope study examining the paleodiets, paleoecology and chronology of late Quaternary mammals in eastern Beringia and the climate in which they lived. Stable isotope levels (del 13 C and del 15N ) were measured in purified collagen extracted from fossil bones of 12 species of late Quaternary mammals, mostly from central and northern Alaska. The difference in del 13 C and del 15 N are proving useful in revealing difference in the ecology, physiology, and diets of each species and for reconstructing specific predator-prey relationships. Pleistocene-age caribou and muskox display del 13 C and del 15 N values indicative of a typical tundra diet consisting of mixed herbs and shrubs, much like modern caribou and muskox. Bison display isotope values suggesting a homogenous diet consisting of mostly C3 grasses. Elevated levels of del 15 N in mammoths indicate that late Pleistocene ecosystems in Beringia were arid-based, and that changes in mammoth del 15 N values over time can be used as an indicator of changing levels of environmental aridity. Limited data on horses show that they also had a broad range of del 15 N values, although absolute del 15 N values were not were not as high as in mammoths. Elevated and varying del 15 N values and mammoths and horses are most likely indicative of varying levels of water-stress induced by environmental aridity, which id predicted to be more evident in monogastric grazers like mammoths and horses. In terms of predator-prey relationships, preliminary data suggest the following: 1) Scimitar cats (Homotherium spp.) may have been mammoth-eating specialists, 2) lions preyed mostly on bison and perhaps muskox and caribou, but very little on horses or mammoths, 3) wolves most likely preyed on caribou and muskox, and 4) Beringia's two purported scavengers?short-faced bears and wolverines?fed eclectically on a variety of herbivores. We also report on 50 new AMS radiocarbon dates on purified bone collagen from bison bones collected in Alaska's arctic slope. The dates indicate a continuous bison presence in the region from at least 40,000 to 10,000 B.P. There is a possible hiatus between about 13,000 to 16,000 B.P., but we believe this gap is best explained by taphonomic processes. Results also returned a bison date of 6400 B.P., the only known Holocene bison for Alaska's arctic slope.