The ability of trees dependent on arbuscular mycorrhizal (AM) fungi to establish in ectomycorrhizal forests is unknown. On northern Vancouver Island, Canada, there are sharp boundaries between mixed red cedar (Thuja plicata)-hemlock (Tsuga heterophylla) (CH) stands, and stands of hemlock and amabilis fir (Abies amabilis) (HA). We tested differences in AM colonization of red cedar between ectomycorrhiza-dominated (HA) stands and stands containing red cedar (CH), across a range of light levels. We used a soil bioassay approach to determine whether there was sufficient AM fungal inoculum in the HA tree stands to colonize red cedar seedlings. Seeds of hemlock and red cedar were sown in forest floor samples collected from the two types of forests, and shade treatments ranging from < 1 to 53% of full sunlight were imposed. After 6 months, seedling survival and root and shoot biomass were quantified, and red cedar seedlings were sampled for AM fungal colonization. Hemlock survival and growth did not differ between soil types, suggesting there was no substrate-associated limitation to its establishment in either forest type. Red cedar colonization by AM fungi was significantly correlated with light levels in CH soils but arbuscular mycorrhizas were absent in roots of red cedar seedlings grown in HA soil. Red cedar survival and relative growth rate were significantly greater in the CH than in HA soil; higher growth was due primarily to greater shoot growth in CH soils at high light levels. The low soil inoculum potential for red cedar in ectomycorrhiza-dominated stands may account for the virtual exclusion of red cedar seedlings from these forests.