Controls on seed production in Eriophorum vagina tum were studied at two Low Arctic sites, Latnjajaure in Sweden and Toolik Lake in Alaska. At Latnjajaure, seed production was monitored for 3 years (1993-1995) in control plants and in plants that were experimentally warmed using ITEX open-topped chambers (OTCs). At Toolik Lake, experimental treatments included a factorial greenhouse x NP fertilizer experiment and artificial shading; data collection occurred in 1995, after eight seasons of treatment. Temperature differences between sites, years, and treatments all had significant effects on components of Eriophorum seed production. Pre-flo ration and post-floration times were decreased in warmer treatments and years. Ovule number per inflorescence was greater in warm years and at the warmer site, Toolik Lake. Although seed set was reduced in the OTCs at Latnjajaure, probably due to low pollination rates, in the larger greenhouses at Toolik Lake seed set was increased. The most consistent responses to temperature were in seed weight, where treatment effects were greatest at lower temperatures, such that, for example, at Latnjajaure the increases in seed weight in the OTCs were greatest in the coldest year. Fertilizer treatment had no significant effect on seed number or weight at Toolik Lake, and in the fertilizer + greenhouse treatment there were no inflorescences because Eriophorum had been nearly eliminated from the treated plots. In the shade treatment at Toolik Lake, inflorescences were fertilized but only a single mature seed was collected. Overall, the results indicate that seed production by E. vagina tum is strongly responsive to environmental variation and that only a small increase in mean summer temperatures would result in a much larger and more stable seed production.