Many semi-arid coniferous forests in western North America have reached historically unprecedented densities over the past 150 years and are dominated by shade-tolerant trees. Silvicultural treatments generally open the canopy but may not restore shade-intolerant species. We determined crossover-point irradiance (CPI) (light at which the height growth rank of pairs of species changes) for seedlings in Sierra Nevada mixed-conifer forest and used these to interpret light environments produced by fuels-reduction thinning and group selection with reserved large trees. Nine of 21 species pairs had well-defined CPIs. The CPI of the most common shade-tolerant and intolerant species (white fir ( Abies concolor (Gordon & Glendl.) Lindl. ex Hildebr.) and ponderosa pine ( Pinus ponderosa Douglas ex P. Lawson & C. Lawson)) was 22.5 mol·m^–2·day^–1or 41% of full sun. Median understory irradiance increased from 9.2 mol·m^–2·day^–1(17% full sun) in pretreatment forest to 13 mol·m^–2·day^–1(24% full sun) in lightly and 15.5 mol·m^–2·day^–1(28% full sun) in moderately thinned stands and 37 mol·m^–2·day^–1(67% full sun) in group-selection openings. We estimate that 5%–20% of ground area in lightly to moderately thinned stands would have enough light to favor shade-intolerant over shade-tolerant growth compared with 89% of ground area in group-selection openings. The CPI provides a tool to assess regeneration implications of treatment modification such as increasing heterogeneity of thinning to enhance regeneration or reserving large trees in group-selection openings to maintain wildlife habitat.

Northern hardwood forests can exhibit considerable temporal stability in their species composition, which litterfall may help to maintain by modifying the soil environment to create an ecological inheritance. We evaluated the evidence for this niche‐construction perspective by carrying out a spatially explicit analysis of redistribution of calcium (Ca), magnesium (Mg), and potassium (K) via leaf litterfall of the six dominant tree species in a temperate forest in northwest Connecticut, USA. We combined stem maps with litterfall measurements to fit models of leaf litterfall for individual species, and we chemically analyzed leaves, surface soil extractions, and unweathered soil particles.Analysis of elemental oxides in unweathered soil particles suggested that species did not differ in ability to gain access to rock‐bound nutrients. Nevertheless, species had a significant effect on quantity of Ca cycled in litter—the two calcicoles (species with affinity for calcium‐rich soils) occurred most commonly on the richest soils, and the Ca concentration in their senescent leaves increased sharply across the soil fertility gradient. No species effects on Mg or K redistribution were seen. Litter dispersal distances cut across calcium affinities: e.g., of the four calcifuges (plants with affinity for acidic soils) two had distant litter dispersal (modal dispersal distance ≥ 4 m) and two had close litter dispersal (modal dispersal distance < 0.3 m). Dispersal distance was associated with leaf size: the three species with the smallest leaves had the most concentrated litterfall shadows. One of these, eastern hemlock, has a recognized ability to exclude other species in part by creating an acidic low‐nutrient soil environment. Overall, the ability of trees to generate soil conditions that were aligned with their own resource use was expressed most clearly in the calcium content of their leaves. We concluded that tree‐environment interactions at the site displayed several of the criteria for niche construction.