Under these circumstances a definite retrograde movement out of the wooded country in September on the part of many Caribou must indicate their preference for the Barrens at this season. In any event, the movement begins just after the insect menace has subsided to a negligible stage. Possibly another inducement for retreating from the wooded country in the early fall is the dearth of open areas in which the animals may spend their resting periods, in comparative safety from Wolves. It is only after the freeze-up that the surfaces of the lakes and rivers supply this desideratum. This condition lasts from November to June--precisely that part of the year in which the Caribou are present in the wooded country in the greatest numbers. But by November what are the conditions on the Barrens? The weather has become severe; snow has covered up a large part of the ground lichens; tree lichens are not to be had. And so at this season, with the coming of the first heavy snows, there is a final movement out of the Barrens into the shelter of the woods, leaving only a minority of the animals to face a bleak and bitter winter in the open country. The biggest herds of the year may then be seen passing southward. A few bucks are said to remain during most winters in the Windy River area. Far to the westward, toward Great Slave and Great Bear lakes, the retrograde movements are somewhat different and more complicated (Clarke, 1940: 96). “Mr. Carl Buchholz, of Churchill, describes the caribou migration at the 60th parallel, north of Churchill, as a southward migration in August, northward in September, and then south in the autumn” (Clarke, 1940: 97). (See also the next paragraph.) _Fall migration in the Churchill region_ The following notes for 1947 were kindly furnished me by Angus MacIver. He reported large numbers of Caribou moving southward across Caribou Creek (25 miles south of Churchill) about November 10, a day after the local freeze-up. He would then see thousands in a day. Prior to that time (perhaps in September?) there had been two “runs” to the northward and northwestward; these presumably represented the normal retrograde migration in the fall. The herds engaged in these two northerly movements must have previously passed southward farther inland. He reported also that the rutting season this year had commenced a little later than the usual October 15.