Individually marked red grouse (Lagopus lagopus scoticus) could be distinguished as territorial or non-territorial from November till the next spring. Territorial birds survived well and many reared young, but numerous nonterritorial birds died or disappeared, and none paired up or reared young. Differences in spring sex ratio between years had already been set in the previous autumn's territorial stock. Many old birds which had been territorial in earlier seasons lost territories after an annual re-distribution of territory each October-November. Thereafter very few of them were seen alive on the area over the winter, and many were found dead, whereas nearly all old birds that kept their territories were still there in spring. The number of birds in spring and the fate of individuals over winter were predicted reasonably accurately from observations of behaviour in the previous November-December. Thus, differences of social class in autumn were critical to a bird's survival to breed in the next summer and to its chances of contesting for territory and breeding in subsequent years.