Two pollen diagrams from lakes north of treeline in northwestern Quebec indicate that Picea never extended north of its present-day limit during the past 6000 yr BP. Alnus crispa was slightly more abundant around 5000 BP, but there are few major changes in the vegetation of the region during the Holocene. A third site in the tundra along Hudson Bay has a slightly longer sequence (7000 yr BP) which indicates more open conditions in the early and recent part of the record. Picea may have been more abundant locally around 3000 BP. Few major changes in these diagrams can be unequivocally attributed to local changes in plant abundance; changes in tree and shrub pollen abundance parallel those seen south of treeline.