Periodic drought may interact with spatially patterned soil moisture regimes, canopy disturbance and species seedling traits to influence tree regeneration processes in northern temperate forests. In northern Michigan, USA, we assessed the effects of regional upland soil moisture regimes (six sites classified from dry to mesic) and tree canopy shading (% light) on planted-seedling dry mass, probability of mortality and functional traits for eight tree species varying in shade tolerance and natural distribution over soil moisture gradients (i.e., drought tolerance): Quercus velutina, Quercus alba, Quercus rubra, Prunus serotina, Acer rubrum, Acer saccharum, Fraxinus americana, and Betula alleghaniensis. The assessment period included droughts during the seedlings' 2nd and 3rd growing seasons. High mortality (especially on drier-classified sites and in more open, high-light environments) and negative associations of soil water minima with mortality collectively suggest drought-driven mortality across sites. In contrast, canopy shade driven mortality occurred only on mesic sites at <10 % light. Overstory canopies facilitated seedling survival relative to more open environments, with survival maxima ranging from 3 to 20 % light among sites. Among species, probability of survival was 1) highest for Quercus species (similar to 0.70) and lowest for B. alleghaniensis (0.10), 2) positively and generally associated with drought tolerance, but not associated with shade tolerance, and 3) positively and strongly associated with species seed-size related differences in total dry mass, root dry mass and rooting depth. In turn, rooting depth (range 3-68 cm) was positively related to seedling water status, and water status to survival during drought. In summary, site soil moisture regimes, facilitation vs. competition via variation in canopy density, and size-related species traits drove patterns of young seedling mortality during an interval that included droughts. Results suggests current drought regimes likely have strong influence on canopy composition across the region's spatially patterned dry to mesic soil gradients, and may have larger impacts if drought intensifies with climate change. Our results can aid development of planting strategies for assisted migration and other emerging needs, including planting 1) species more drought adapted than those dominating current overstories, 2) large stock with well-developed root systems, and 3) in microsites that maximize canopy facilitation.

Single-tree selection (STS) silviculture has dominated management of northern hardwood forests (NHFs) in the Great Lakes region since introduction the 1960's. Periodic partial cuttings are assumed to promote and recruit well-stocked natural regeneration, resulting in balanced uneven-aged stand structure with sustainable harvest volumes. However, low sapling density of tree species desirable for management, including typically dominant and economically valuable sugar maple (Acer saccharum), suggests STS may not be effectively recruiting sugar maple in some regions. Few studies have analyzed sugar maple size-age structure to determine whether recruitment failure has been chronic since partial cutting began or is a more recent phenomenon. Here, we analyzed sugar maple age structure, regional patterns, and possible drivers for 51 selection-managed NHF stands located in northern Michigan, USA. Our data were 1499 aged basal discs collected from stumps of sugar maple trees >5 cm diameter at breast height (DBH) harvested winter 2017-2018. Our results offer little evidence to suggest that partial cuttings have promoted well-stocked ingrowth of sugar maple saplings. For most stands, age and diameter were non-linearly related, with a wide range of larger-diameter trees having similar ages. Among stands, sugar maple sawtimber (>24 cm DBH) averaged 106 years old and poletimber (11.4-24 cm DBH) 91 years, along with older than expected saplings (5-11.4 cm DBH) averaging 67 years. Most stands were dominated by a similar to 100-year-old upper canopy cohort, with a majority having <= 2 younger cohorts typically of low density and >50 years old. Areas with greater snow water equivalent and lower January minimum temperatures, where deer populations have been historically low, generally have younger sugar maple saplings. Low density and generally old age of smaller size classes led to much lower stocking of age classes <70 years old than expected under STS. Together, this suggests long-term STS management for sugar maple may be unsustainable over large portions of the study region if current trends in forest and deer management continue. Our age structure analysis supports the growing consensus that alternative management may be needed for NHF in parts of the Great Lakes region.

Declining temperate forest biodiversity is a global challenge to future sustainability and resilience. Canopy tree diversity has declined in northern hardwood forests (NHFs) of the Great Lakes region since European colonization, with current tree regeneration patterns suggesting this trend may continue. However, regional (10(6) ha) patterns and drivers are underexplored. We used data from 141 managed NHF stands, Michigan, USA, to examine diversity indices across tree size classes (seedling, sapling, and canopy); stand-level drivers of regeneration (seedling and sapling) diversity and individual species density; and patterns among species in relative abundances across size classes. Diversity was similar across size classes (mean asymptotic species richness of 6-7 species and <3 effective common species); however, some species contributing to sapling diversity are unlikely to contribute to future canopy diversity (Fraxinus americana and Fagus grandifolia, insect/pathogen limited; Ostrya virginiana, small maximum size). For the 11 most common species, conspecific canopy density (+, seedlings, significant for 10 species) and stand basal area (-, saplings, 3 species) most consistently associated with density; deer use, site quality, and substrate were less predictive. Patterns of relative density by size class among 18 species present on >10% of sites were consistent with species life history strategy and/or establishment/recruitment barriers. For seven species (including Tsuga canadensis, Betula alleghaniensis), relative density was lowest in the seedling layer, suggesting seedling establishment substrate and/or early shade mortality limitations. For two species (Acer saccharum, Quercus rubra), reduced sapling relative density compared with seedling or canopy strata suggests a sapling recruitment bottleneck from deer browsing/shade, whereas six species (e.g., F. grandifolia, O. virginiana, Pinus strobus) had highest relative density for saplings, suggesting sapling to canopy recruitment limitations. Lastly, seedlings were relatively most dense for Acer rubrum and Prunus spp., indicating disproportionally high seedling establishment, if not sapling or canopy recruitment, and for F. americana, consistent with recent pest-related canopy mortality. Collectively our analyses suggest sustained low diversity and shifting composition are driven by limited local seed availability, limited seedling establishment substrate, deep shading, and abundant deer. Management aimed at overcoming these limitations may be necessary to promote future NHF resilience.

Red pine plantations may promote more desirable hardwood regeneration than hardwood stands because of differences in: 1) the composition and quantity of tree seeds available, and 2) abiotic and biotic properties of growth and establishment environments. However, information on how distance from hardwood edges, thinning of overstory pine, and site quality shapes hardwood regeneration properties in red pine plantations is scarce. Such information would aid in understanding and managing red pine plantation effects on hardwood regener-ation processes and guide spatial and temporal rotations of such effects across landscapes. To elucidate re-lationships, we monitored hardwood seedlings and canopy openness across transitions from hardwood edges to red pine plantation interiors at 40 sites spanning a site quality gradient in northwestern-lower Michigan. Hardwood seedling density: 1) declined substantially from edges to interiors at all but low-quality sites before stabilizing 30-m into plantations, 2) was reduced, or not affected, by thinning despite positive effects on seedling light environments, and 3) was greater at intermediate than high-and low-quality sites. From edges to interiors, species composition shifted towards Acer rubrum, Quercus species, and Prunus serotina and away from Acer sac-charum, Fagus grandifolia, Fraxinus americana, and Ostrya virginiana. Species increasing in relative abundance and/or density from edges to interiors were generally more drought-tolerant, shade-intolerant, large-seeded, and animal-dispersed than decreasing species. Shifts towards less shade-tolerant species from edges to interiors were consistent with increases in canopy openness and were most pronounced where changes in light were greatest (i. e., thinned plantations on high-quality sites). Changes in hardwood seedling density and composition were likely driven by differences across species in dispersal traits, seed characteristics, and relative adaptation to interior conditions (likely hotter, drier, and nutrient poor). To promote commercially valuable hardwoods, increase mast for wildlife, or improve climatic resiliency, the most suitable targets for conversion to hardwoods may be plantations with significant interiors (i.e., area > 30-m from hardwood edges) on higher quality sites adjacent to desirable seed sources (e.g., Quercus rubra, Prunus serotina, Acer saccharum). More desirable hardwood regener-ation may result from reevaluating thinning timings and intensities originally designed to maximize red pine production. Our results also indicate that intentional management of surrounding hardwood forest composition (i.e., seed sources), within plantation retention of hardwood islands (e.g., mature oak stems), and greater control of Acer rubrum competition may further increase the desirability of hardwood regeneration in red pine plantations.

Well-stocked natural tree regeneration is critical to sustainable management of northern hardwood forests (NHF) by selection silviculture, but explicit goals and stocking criteria for sapling recruit size classes are lacking. For stems 0–2 in. (0–5.1 cm), we define 171 and 691 stems ac−1 (422.5 and 1,707.5 ha−1) as lower and upper full stocking thresholds for basic NHF management goals (wood and biomass productivity) and overabundant/pest–pathogen-challenged species. We define component stocking targets as 15

Group harvesting is often used to regenerate mid-tolerant tree species. However, in managed northern hardwood forests of the Great Lakes region, regeneration failures are common among mid-tolerant species regardless of gap size. Coppicing advance regeneration has the potential to alter recruitment patterns within harvest gaps, as some mid-tolerant species have large dormant season carbohydrate reserves in the root system. However, the comparative sprouting abilities of northern hardwood saplings are not well quantified. Furthermore, variation in light availability and chronic browsing from white-tailed deer may modify comparative sprouting abilities. In this study, we tracked the effects of dormant season coppicing and deer exclusion on the initial sprouting response and five-year survival, height growth, and final height of seven-year-old volunteer white ash saplings and planted yellow birch, paper birch, northern red oak, sugar maple, red maple, and American beech saplings across a gradient of harvest gap sizes (10-35 m diameter, <0.5-1.25 tree height) in a mature even-aged northern hardwood stand in northern Lower Michigan, USA. In addition, saplings were given a growing position ranking, which integrated sapling final height and survival status relative to competitors to evaluate the effects of coppicing and deer exclusion on sapling development. Overall, coppicing produced a strong initial sprouting response with nearly all red maple and northern red oak stems initially sprouting. In the absence of deer, appreciable differences in sprout survival and height growth occurred among mid-tolerant and intolerant species and contributed to changes in growing position. Northern red oak sprouts significantly gained growing position over all species through a combination of high survival and modest height growth across all gap sizes. In contrast, coppiced yellow birch and paper birch sprout survival was significantly lower than uncoppiced stems, which contributed to declines in growing position for both species across all gap sizes. Coppicing in the absence of deer had less effect on the growing position of mid-tolerant red maple and white ash and shade tolerant sugar maple and American beech. Without deer exclusion, browsing limited the effectiveness of coppicing by reducing height growth and final height of all mid-tolerant species. Collectively, our results indicate that coppicing does not benefit mid-tolerant species equally and is unlikely to benefit browsing-preferred species in areas with high deer density. Nevertheless, in areas of lower deer density or in situations where sprouts can be protected, coppicing may improve the competitiveness of northern red oak saplings within harvest gaps.

For temperate forests, Michigan, USA, we asked: (1) does evidence exist for a height-dependent sapling recruitment bottleneck caused by shrub layer competition and white-tailed deer (Odocoileus virginianus) browsing and, if so, (2) how do these factors interact with light availability in shaping species-specific sapling recruitment niches? The combination of deer browsing effects on tree population height structure and the effects of deer and shrubs on tree seedling height and mortality indicate a shrub-deer bottleneck for sapling recruitment. The maximum height of deer browsing and shrub canopies as well as height dependent seedling mortality rates revealed a critical height threshold of approximately 2 m, above which trees escape from shrub and deer effects (i.e. sapling recruitment). Species’ sapling recruitment niches differed across a harvest gap size (i.e. light availability) gradient, over which tree seedling height growth potential and shrub competition increased with gap size, but deer browsing probability decreased. Specific to forests with high deer populations, niche differences were related to species growth and survival properties, with species generally falling into three groups: High Light, Broad, and Nowhere. High Light species incur high mortality from shade, shrub competition, and deer browsing. However, recruitment can occur in recently formed large gaps, where seedlings growing within developing shrub canopies avoid competition via rapid height growth and are partially protected from browsing deer. Broad species have lower mortality in shade and grow slower than High Light species and, compared to other groups, exhibit height growth and mortality rates that are less-sensitive to deer browsing pressure. These properties allows sapling recruitment of Broad species over a wide range of gap sizes, including larger gaps, where, via advance regeneration, they can avoid competition with developing shrub layers following canopy disturbance. Nowhere species are similar to Broad species, except their height growth and mortality are sensitive to deer browsing pressure. Heavy browsing pressure in smaller gaps and an inability to maintain taller advance regeneration to contend with shrub competition when large gaps form leave Nowhere species without a sapling recruitment niche. In conclusion, we provide evidence that shrub competition, ungulate browsing, harvest gap size and their interactions constrain height-specific sapling recruitment among tree species in the temperate forests of Michigan. Similar mechanisms may operate in other forests with high ungulate populations. Given these drivers and current management regimes, increasing tree species diversity and resilience to future disturbances may hinge on changing and integrating forest and ungulate management paradigms.

Single-tree selection silviculture management of northern hardwood forests relies on natural tree regeneration for long-term sustainability, yet current trends in tree regeneration and recruitment elicit concern. Low densities of economically valuable sugar maple (Acer saccharum Marsh.) in understories are often common, likely driven by many factors, including deer browsing, management-dictated stand structure, and site fertility/moisture regimes. However, landscape sugar maple regeneration patterns and relationships with underlying factors are largely unknown. We quantified associations of spatially varying factors with sugar maple regeneration using detailed vegetation and white-tailed deer winter fecal pellet surveys from 141 northern hardwood stands in Michigan, managed for decades with single-tree selection silviculture. We developed models of plot-level sugar maple regeneration counts for three key size classes as a function of plot- and stand-level predictors, including deer use, forest structure, and site quality. Among our 141 stands, sugar maple seedlings (50 cm tall) were consistently abundant, averaging 69,000 stems ha(-1) and present in 76% of plots per stand, on average (25 plots per stand, each 2 m(2)). In contrast, small (50-137 cm tall) and large (5 cm DBH) sugar maple sapling densities were much lower, averaging 2,300 and 1,100 stems ha(-1), respectively, and occurring, on average, in 31% and 32% of plots per stand (25 plots per stand, each 12.6 m(2)). Under a wide range of potential sugar maple stocking criteria, most stands are understocked in saplings; e.g., only 29% of stands had 2,500 small sugar maple sapling stems ha(-1) and 35% of stands had > 1,000 large sapling stems ha(-1). Based on our models, sapling densities negatively associated with deer use and were more abundant on medium than high quality sites. Across all size classes, negative associations with subcanopy trees and/or shrub densities suggest light limitation, whereas positive associations with sugar maple canopy trees > 25 cm DBH suggest persistent seed limitations. Overall, our study supports need for alternative forest and/or deer management strategies over much of the range of northern hardwood forests in Michigan to promote higher densities of large sugar maple regeneration for canopy recruitment. Medium quality sites with abundant large sugar maple canopy trees and low deer browsing pressure (for example, the deep snow region in the north western Upper Peninsula) are the exception; high sugar maple sapling densities suggest forests in this region are thriving under single-tree selection management.

Maintaining richness and diversity of bird species on forest landscapes has long been a major theme in sustainable forest management. However, effects of management decisions on species diversity are uncertain because of difficulties in accounting for all species simultaneously. Utilizing a diversity contribution approach, we applied an integrated coarse filter-mesofilter conservation strategy to inform avian biodiversity conservation in the western Upper Peninsula of Michigan, USA. Diversity contribution is a measure of relative contribution of a given area to regional species diversity based on species richness and distinctiveness, with region defined according to the study design. At the coarse filter level, we compared and ranked forest types by bird conservation value based on proportional representation of high-contribution sampling plots (sampling plots in the top 25% of diversity contribution) relative to proportional representation in the dataset. At the mesofilter level, we identified how plot-level compositional and structural attributes affected diversity contribution within forest types. We found that forest types less common in the dataset, such as cedar and upland conifer, had the highest bird conservation value, while the most common forest type (northern hardwoods) provided the greatest number of high-contribution plots. Stand-level attributes that increased structural complexity, such as higher densities of large trees, saplings, and snags, or lower basal area, tended to increase the likelihood of being a high-contribution plot within a given forest type. However, the importance and effect of different attributes varied among forest types. Our novel application of the conservation filter strategy provides a way to identify high priority forest types for avian conservation, and create stand-level prescriptions for increasing avian conservation value. Given data on forest inventory and bird community composition, managers can use this application as a planning tool in the face of uncertainty about how management decisions affect larger-scale avian biodiversity.

Well-designed, robust partnerships are important for addressing forest resource problems that intersect diverse ownerships, management, and research expertise. Here, we describe a case study of a research–management partnership. First, we describe the rationale and design of a large-scale, long-term experiment focused on increasing the sustainability and resilience of managed northern hardwoods. With the experiment as context, we next discuss details of the partnership contributing to its success, as well as some challenges. The experiment is a Michigan-wide evaluation of silvicultural alternatives to the current paradigm of selection silviculture. Our goal is to identify combinations of silvicultural treatments that increase the diversity and density of sapling recruits over forested landscapes where deer-browsing pressure and other limiting factors are at play. Pretreatment data from our 140-site study bolstered justification for the research, with sapling and pole classes generally lacking adequate stocking and diversity of desirable species over large geographic extents. We anticipate that the wildlife and forestry research–management model used to develop this project will be useful for other rapidly developing challenges that cannot wait decades to be addressed without potentially catastrophic losses to forest function. Key beneficial elements of our partnership include: (1) sensitivity to information needs at all management levels (i.e., policy to practice), (2) capitalizing on differences in strengths and reward systems of managers and scientists, (3) structured, frequent scientist–manager meetings to maintain focus on research and management application outcomes, (4) long-term relations and support that build trust, and (5) explicit wildlife and forestry collaboration on problems requiring joint expertise. Challenges to our partnership generally revolved around communication, workload, and implementation issues, and were partially alleviated by planning.

In the northern Appalachian region of North America, mortality of mature American beech (Fagus grandifolia Ehrh.) via the introduced beech bark disease (BBD) can result in dense thickets of beech saplings that inhibit the regeneration of other species. It is unknown if similar structures characterize more recently infested managed forests in the Great Lakes region. If these dense beech sapling layers do exist, management would be aided by knowing which site/regional factors they are associated with and by identifying particular sapling structures that may threaten the sustainability of these forests under current management paradigms. To examine these patterns, we used a natural experiment with sample plots in 69 unevenly aged, selection silviculture-managed, maple (Acer spp.)-dominated northern hardwood stands. Our stands were dispersed across northern Michigan, USA and had undergone BBD-motivated partial harvests favoring beech removal (mean = 5.5 years before measurement). In each stand, we quantified tree regeneration structure in relation to winter deer use (fecal pellet count density), site quality (habitat type), geographic region (Eastern Upper Peninsula and Northern Lower Peninsula), and multiple measures of overstory stand density. We also examined the density effects of taller regeneration strata on subordinate strata. Across sites, the small sapling recruit class (i.e., >137 cm tall and <5 cm diameter at 137 cm tall) was dominated by beech and was often dense (44% of subplots > 2000 stems ha−1 and 16% of subplots > 5000 ha−1) but never exceeded the > 10,000 stems ha−1 reported in the northern Appalachian region. Beech sapling density was higher in the Northern Lower Peninsula, on lower quality sites, at lower postharvest overstory densities, and on sites with higher densities of preharvest overstory beech. In contrast to the beech-dominated small sapling recruit class, seedlings (i.e., <25 cm tall) were generally more species diverse than sapling strata and were dominated by maple species. Although generally dense, seedling density was negatively related to small sapling recruit density, suggesting that saplings may suppress the seedling stratum. The general pattern for the small sapling recruit layer of browsing-insensitive beech (and ironwood, Ostrya virginiana Mill. K. Koch) dominance and low representation of browsing-sensitive species (e.g., Acer spp.) circumstantially supports the notion that regeneration structure is heavily influenced by deer. However, current deer use was generally low in our stands, and relationships with tree regeneration structure were weak. Instead, regeneration structure is likely shaped by a combination of factors operating at long time scales (i.e., legacies of deer browsing pressure, selection silviculture (given beech and ironwood are shade tolerant), overstory composition, and site quality) and by those effects that are more proximal, such as postharvest overstory density. Minimum stocking criteria for species considered desirable for management (e.g., sugar maple and Acer saccharum Marshall) suggest many stands are inadequately stocked in the sapling recruit classes. Although future regeneration dynamics are unclear, current patterns suggest that many stands with high beech/ironwood small sapling recruit densities may require management intervention to overcome insufficient recruitment of species targeted for management.

Carex pensylvanica (Pennsylvania sedge) dominance of managed temperate mesic forest understories is associated with high deer browsing pressure, frequent partial timber harvests, invasive earthworms, and low seedling-to-sapling recruitment rates of canopy-dominant Acer sacchaiwn (sugar maple). There is concern that Carex dominance may be difficult to reverse because Carex may decrease soil resource availability and tolerate these conditions better than Acer seedlings. However, evaluation of this notion and of management aimed at reducing possible Carex impacts on Acer seedlings is lacking. Here, we examine the effects of (1) removing Carex on soil water and the growth and survival of naturally-established Acer seedlings over a forest understory light gradient, (2) drought on growth, survival, water relations, and soil mineral N pools of potted monocultures and mixtures of Carex and Acer seedlings, and (3) summer vs. autumn glyphosate herbicide treatments on Carex and Acer seedling dynamics in group-selection harvest openings. Carex removal increased soil water and growth of naturally established Acer seedlings but only in harvest gaps > 8% light. Carex exhibited greater tolerance of drought than Acer by closing stomates at lower pre-dawn water potentials (- 4.2 MPa vs. - 3.0 MPa), and incurring lower post-drought mortality (similar to 0% vs. 50%). In well-watered Carex-Acer mixtures, Carex reduced soil NH4+ but not Acer growth or survival relative to Ater monocultures. Glyphosate treatments reduced Carex by > 90%, but possible benefits to young Acer seedlings may have been offset by greater deer herbivory on Acer in glyphosate treatments. Our results indicate that Carex reinforces its dominance of mesic hardwood forest understories via water competition and drought tolerance mechanisms that slow the establishment and recruitment of canopy dominant Acer. These effects could be exacerbated by increased intensity and frequency of droughts in the future, such that Carex dominance of understories could accelerate the xerophication of these systems in the face of climate change. Herbicides can be used to reduce Carex but may need to be combined with deer population reductions to promote adequate Acer seedling to sapling recruitment for sustainable management by selection silviculture. The sustainability of other temperate mesic forests globally characterized by high ungulate populations and associated high graminoid density may be similarly challenged.

Declines in the diversity of herbaceous and woody plant species in the understory of eastern North American hardwood forests are increasingly common. Forest managers are tasked with maintaining and/or promoting species diversity and resilience; however, the success of these efforts depends on a robust understanding of past and future system dynamics and identification and application of appropriate silvicultural interventions. We review how historical timber harvesting and land use, increases in deer population sizes, invasive species, and contemporary forest management practices interact to erode ecological memory and increase resilience debt of hardwood forests of eastern North America. The erosion of ecological memory and growing resilience debt in these forests pose significant challenges for forest managers because they alter the response of forests to management from the understory to overstory. Differences in how much ecological memory these systems retain, as well as the mosaic of interacting factors influencing contemporary dynamics, preclude a one size fits all management approach. That being said, our review has identified a host of common factors and pathways that can be manipulated. The approach we propose requires a more thoughtful understanding of the forest understory as the foundation upon which resilient systems are built. Which silvicultural levers will have the greatest utility clearly depend on the ecological context of the forest, and a willingness to experiment and adapt. We offer a conceptual model and recommendations for managers confronting novel plant communities and uncertain system dynamics.

Many managed northern hardwood forests are characterized by low-diversity tree regeneration. Small harvest gaps, competition from shrub-herb vegetation, and browsing by white-tailed deer (Odocoileus virginianus) contribute to this pattern, but we know little about how these factors interact. With a stand-scale experiment, we examined the effects of gap size (0-3234 m(2)), vegetation (weeded: unweeded), and deer (fenced: unfenced) on seedling growth and survival for 18 tree species. With increasing gap size and light, shrub-herb vegetation density increased, while deer browsing on seedlings in unweeded plots decreased. Fenced: weeded seedlings of all species increased in height up to 35-45% light, with optimal growth in large-group selection and patch cut harvest gaps. Height growth rank order among tree species changed between gap sizes, but growth varied little in small, low-light gaps. Instead, a low-light survival (i.e., shade tolerance) vs. high-light growth tradeoff we observed is likely more important for species sorting of gap sizes. Shrub-herb vegetation decreased seedling survival and growth, especially in larger harvest gaps, shifting gap size optima to smaller gaps, but had little effect on growth/survival rank order among species. In contrast, deer had strong impacts on growth rank order, especially in larger gaps where species differences in growth potential were trumped by differences in deer browsing pressure responses. However, contrary to their consistently negative main effects, vegetation and deer had two positive interacting effects: dense shrub-herb vegetation in large gaps protected seedlings of faster-growing species from browsing and deer browsing of shrub-herb vegetation modestly increased light and growth of short, suppressed, browsing-avoided species. In summary, harvest gap size-mediated light availability, shrub-herb vegetation, and deer herbivory had strong interacting effects on tree seedling interspecific performance ranks and intraspecific optimal gap sizes. For management, a broad range of harvest gap sizes and rapid establishment of tree regeneration (naturally or planted) to minimize shrub-herb competition should increase tree diversity in forests with few deer. However, with deer browsing pressure, a more limited set of lesser-browsed species are likely to recruit successfully regardless of gap size, except in large patch cut gaps, where recruitment of faster-growing, shade-intolerant species is possible.