Frequencies and patterns of browsing by large herbivores on conifer seedlings

Browsing by large herbivores on planted and naturally regenerated conifer seedlings (Picea abies and Pinus sylvestris) was recorded in 104 clear‐cuts in east‐central Sweden during 1990 and in 47 clear‐cuts in 1991. The number of seedlings browsed and the browsing patterns were analysed in relation to seedling type. Browsing frequencies were also compared between forest stands with different site productivities and subjected to different management practices. The variation in the number of seedlings browsed in 1990 was explained mainly by seedling category. Among both planted and naturally regenerated seedlings, pine was browsed more than spruce. Two‐year‐old containerized seedlings of spruce was browsed more than 4‐yr‐old bare‐rooted spruce. In 1991, browsing was more equal among species and seedling types. Number of seedlings with their leader browsed and the amount of biomass left after browsing differed significantly between seedling types. Seedlings that had been browsed in 1990 experienced significantly higher browsing frequencies in 1991 when compared with unbrowsed seedlings. The effects of stand characteristics were not found to be significant.     

Large-scale reduction of ungulate browsing by managed sport hunting

Browsing by free-living ungulates is a serious threat to forest regeneration. Although predictors of browsing and potential consequences are well researched, less information is available in the scientific literature about possible solutions to the problem. We investigated whether a large-scale game management system in a traditional sport hunting culture has the potential to effectively reduce browsing damage. Our analysis was based on browsing survey data from 2006 and 2009 covering more than 25,000 km² in Bavaria, Germany. After an initial assessment of browsing damage throughout Bavaria in 2006, game management plans were altered, and the effect of this intervention was assessed in 2009. Browsing damage clearly declined in 2009 in areas where the suggested deer harvests were increased in 2006 game management plans. This is the first report of a successful implementation of a large-scale game management system aiming at reducing browsing damage to facilitate unfenced forest regeneration.     

水分胁迫对栓皮栎幼苗生理特性及生长的影响

通过盆栽实验，研究了在不同的水分条件下栓皮栎幼苗生理特性的变化和苗木生长量的变化。研究表明，在水分胁迫初期，随土壤含水量的降低，栓皮栎叶片的净光合速率、蒸腾速率下降，而水分利用效率提高。在水分胁迫末期，由于长期的干旱胁迫，水分条件最差的一组苗木已干枯死亡，土壤含水量从38．123％下降到20．323％时，净光合速率和蒸腾速率均下降，水分利用效率也下降。随着土壤含水量的降低，苗高、地径和生物量均有下降的趋势，其中生物量下降的幅度最大。     

Effects of roe deer browsing and site preparation on performance of planted broadleaved and conifer seedlings when using temporary fences

Abstract

The effects of fencing and site preparation on performance of birch (Betula pendula Roth), oak (Quercus robur L.), pine (Pinus sylvestris L.) and spruce [Picea abies (L.) Karst.] seedlings were studied over four growing seasons in southern Sweden. The experiment was composed of four browsing treatments, from which roe deer (Capreolus capreolus L.) were excluded for 0, 12, 24 or 42 months, and four soil treatments. Natural browsing was combined with artificial browsing (clipping). It was demonstrated that browsing by roe deer could be isolated from browsing by moose (Alces alces L.) by selective fencing. Except for birch, browsing had little effect on seedling survival. Browsing reduced height growth by more than 100% for oak and pine, and more than 60% of pine seedling developed multiple stems. Except for oak, site preparation increased seedling survival. Inverting site preparation, in combination with or without fertilization, produced positive growth responses in seedlings, whereas patch scarification did not. There was no interaction between browsing and soil treatments on seedling performance. It was concluded that long-term protection against browsing by roe deer is needed for oak and pine, whereas short-term fences might be used for birch, and that spruce is relatively unaffected by browsing.     

Effects of timing,duration, and intensity of simulated browsing on Scots pine growth and stem quality

This study examines a large number of structural and growth-related responses of young Scots pine in relation to simulated browsing from large herbivores. We combined different modes of intensity, timing, and duration of clipping for nine consecutive years from 2000 to 2008. Data on tree responses were collected yearly during this period, as well as in 2011, and were analyzed using ordination techniques and generalized linear models. A key result demonstrated is the serious effects of clipping on the youngest pines. Mortality was high when clipping started the first year, indicating that the trees are vulnerable at that stage, but after that increasing intensity and duration of clipping were the main causes of mortality. Among the different treatment factors, intensity of clipping had the largest influence on altering the final growth responses, some of the quality responses, and mortality. Timing of clipping was the most important factor for explaining the occurrence of the majority of responses related to stem structure. Duration of clipping had the largest effect on the responses that included repeated-measures data, such as height development and number of stems. We discuss the long-term consequences of browsing for the trees and the implications for forest management.     

Tree shelters and weed control: Effects on protection,survival and growth of cherrybark oak seedlings planted on a cutover site

In the southern USA oaks (Quercusspp.) are often favored by forest owners havingmultiple objectives for forest ownership as oaksprovide mast for wildlife, are consideredaesthetically pleasing, and are valuable for timberproducts. Regeneration and early seedling growth isa concern to those forest owners interested insustaining oaks as a component of their forests. Theeffects of tree shelters and herbaceous weed controlon second-year seedling survival, browse by deer andrabbits, and seedling growth of hand-plantedcherrybark oak (Quercus pagoda Raf.) wereexamined. The study was established on a cutovermixed pine-hardwood forest in Alabama. Fourtreatments were: weed control only, tree shelter only,tree shelter with weed control, and a controlconsisting of a seedling without a tree shelter orweed control. No significant difference in seedlingsurvival was found among the treatments after twoyears. Tree shelters were effective in preventingbrowsing. No seedlings in the tree-shelter-onlytreatment were browsed. There was no significant,difference, however, in the percentage of seedlingsbrowsed between the control treatment and the weedcontrol treatment. The use of tree shelters with weedcontrol was the most effective treatment for promoting2-year ground-line diameter, height, and stem volumegrowth.     

Influence of mulching and tree shelters on 4-year survival and growth of zeen oak (Quercus canariensis) seedlings

We assessed the effect of mulching and tree shelters on the establishment and early growth of zeen oak(Quercus canariensis Willd.) during the first 4 years after planting in Northwestern Tunisia. Five mulch types(Italian Stone Pine(Pinus pinea L.), Lentisk(Pistacia lentiscus L.),and a combination of Italian Stone Pine and Lentisk(organic mulches), gravel(inorganic mulch) and control), and three tree shelter types(non-vented and vented tree shelters, and control) were tested. An increase in the number of internodes occurred under the gravel mulch, while a reduction in survival was found for the lentisk mulch. Tree shelters had no effect on survival, but increased mean height growth and reduced mean diameter growth during the 4 years(excepting a non-significant effect for vented tree shelter at year four). Comparison of the annual shoots and growth units(GU) between sheltered and unsheltered plants according to year of formation revealed two growth phases. In first phase, shoots were totally or partially inside the shelters; mean length of annual shoots and GU were greater for sheltered plants. The second phase was characterized by shoots emerging from shelters; mean length of annual shoots and GU were similar for all plants, with or without tree shelters. Results suggest that the use of tree shelters, particularly vented shelters, could contribute to the improvement of the artificial regeneration of zeen oak.The use of mulching alone or in combination with tree shelters did not improve zeen oak performance in the field.     

Interactions among forestry regeneration treatments,plant vigour and browsing damage by deer.

In a large field experiment we studied the influence of regenerationmethods on the extent to which roe deer (Capreoluscapreolus L.) browse on spruce (Picea abies L.Karst) seedlings. Our objective was to evaluate if treatments that are good atstimulating regeneration may increase browsing damage by deer. Data werecollected between 1993 and 1995, and in the winters of 1993–94 and1994–95 the frequency of browsed seedlings was 13.2% and 13.4%,respectively. The frequency of browsing damage varied considerably betweensitesand clearcuts. Browsing damage was more frequent when the regeneration methodsincluded insecticide treatments. Among insecticide-free treatments, browsingwasheavier on scarified plots than on herbicide-treated or control plots (in whichseedlings were planted on untreated ground and given no post-plantingtreatment). Containerised seedlings were browsed more than bare-rooted ones.Neither the age of clearcut when planting, nor removal of slash, had any effecton browsing. It was demonstrated that regeneration methods that increased plantvigour (as assessed by leader growth and needle colour index) led to morebrowsing damage. Thus, our results support the plant vigour hypothesis, whichstates that many herbivores prefer to feed on vigorous plants because they aremore nutritious. However, the difference in browsing damage betweencontainerised and bare-rooted seedlings could not be explained solely by theplant vigour hypothesis.     

Conifer leader browsing by roe deer in English upland forests: Effects of deer density and understorey vegetation

At moderate to high densities ungulates can impact negatively on forest crops and these may be managed by lethal control. In production forestry an understanding of the relationship between ungulate density, habitat-related factors and the incidence of tree damage may promote more efficient ungulate damage management than by lethal control alone. In plantation forests in the north east of England, the incidence of conifer leader browsing by roe deer (Capreolus capreolus) was positively associated with an index of deer density (at the site scale) and planted crop tree density (at the feeding patch scale) and was negatively associated with the relative abundance of palatable understorey vegetation (at the site scale). An interaction between the deer density index and palatable understorey browse abundance was positively associated with leader browsing. Coefficients from the optimal model were used to construct a browse probability surface across a matrix of values of deer dung density and palatable understorey vegetation abundance. This illustrated a very different relationship between dung density and conifer leader browsing probability across the various levels of palatable understorey vegetation abundance. From this we hypothesise that in areas of limited palatable understorey vegetation abundance, higher levels of culling may be required to achieve reductions in the incidence of conifer leader browsing than would be necessary in areas with moderate to high palatable understorey vegetation abundance. Improvements to understorey vegetation in concert with roe deer population control may enhance the efficiency of deer damage management in upland areas with nutrient-poor soils.     

草本地下竞争对3种幼苗存活和生长的影响

为了研究不同土壤营养条件下草本地下竞争对华北落叶松、麦氏云杉、辽东栎3种1年生幼苗存活和生长的影响,以华北落叶松、麦氏云杉和辽东栎1年生幼苗为材料,以土壤营养条件(以施用肥料中的氮含量为基准:10 g·m^2和40 g·m^2)和草本地下竞争方式(完全地下竞争,部分地下竞争,无地下竞争)为处理组合,统计幼苗存活率并测定苗高。结果表明:1)完全和部分去除草本根系的地下竞争均可同样有效地提高华北落叶松幼苗的存活率,而不同土壤营养条件对其存活率和高生长均无影响,不同地下竞争方式不影响其高生长。2)对麦氏云杉幼苗而言,在40 g·m^2的土壤营养处理下其存活率显著降低;但高生长未受影响,而完全和部分去除草本根系的地下竞争均显著提高了其存活率,且高生长在完全去除草本地下竞争的处理下最高。3)辽东栎幼苗在所有处理组合下全部存活,高生长在40 g·m^2的土壤营养处理下以及完全和部分去除草本根系地下竞争的处理下显著较高。华北落叶松和麦氏云杉幼苗定居的关键并非是草本植物对土壤养分的争夺,而是草本植物地下部分对华北落叶松幼苗根系生长空间的争抢。在今后森林营造和恢复实践中,通过清除草本的地下部分,特别是宿根性草本,避免草本占据过多地下生长空间,协助华北落叶松和麦氏云杉成功渡过幼苗定居期,同时森林培育中的除草措施也在一定程度上有利于辽东栎幼苗的生长。     

Fertilization—Effects on deciduous tree growth and browsing by moose

Knowledge of plant–herbivore interactions is fundamental for understanding foraging patterns of herbivores and their effects on the ecosystem as well as the human use of natural resources. The nutrient resources available for plants may determine not only growth rate, but also the amounts of secondary compounds in the plant tissues. As a consequence, browsing pressure on plant communities by large herbivores can be affected by fertilization and nitrogen deposition. We conducted a field fertilization experiment to test the hypothesis that fertilization leads to increased sapling growth and increased browsing by a large herbivore, the moose (Alces alces). The study was conducted on fertilized and unfertilized plots (50 m × 50 m) in northern Sweden. The plots were located in Scots pine (Pinus sylvestris) stands of young succession stages with an intermixture of deciduous species. Fertilization with 600 kg calcium–ammonium–nitrate per ha (=200 kg N ha⁻¹), was conducted in spring on one of the plots in each of 25 paired plots. The other plot in each pair was kept as control. The effect of fertilization on plants and browsing pressure was evaluated on five deciduous tree species in June and August the year after fertilization. Both browsing from preceding winter (bites) and current summer (leaf strips and bites) was recorded. Shoot size and other shoot variables increased with fertilization. Fertilization increased winter browsing for most species, but few differences were statistically significant. The late summer recording (August) showed pronounced differences in browsing pressure (leaf strips and bites) between control and treatment. In June, only a trend of increased browsing (leaf strips) was found for all species on fertilized plots. Generally, the pressure of leaf stripping was low in June, probably due to the short time the annual shoots had been exposed to moose browsing. Number of summer bites (recorded in August) was significantly higher on fertilized plots for the Betula spp., and the trend was similar for the other tree species. Within species, there was no difference in strip length between the treatments. The result indicates that fertilization can be used as a habitat improvement for ungulates within wildlife management. But, fertilization is also likely to lead to an increased risk of browsing damage, which is undesirable if forests are being fertilized for increased forest production.     

Establishment and growth of oak (Quercus alba, Quercus prinus) seedlings in burned and fire-excluded upland forests on the Cumberland Plateau

Recurrent problems with regeneration of oaks (Quercus spp.) have been documented across a wide range of ecosystems. In oak-dominated forests of the central and Appalachian hardwood regions of the United States, a lack of competitive oak regeneration has been tied, in part, to fire suppression in these landscapes, and managers throughout the region are using prescribed fire to address this concern. To examine fire effects on oak regeneration, researchers have generally relied on inventories or population studies of existing seedlings. These studies are valuable but do not permit examination of the role of fire in enhancing the establishment and growth of new oak seedlings stemming from oak mast events. In this study, white (Quercus alba) and chestnut oak (Quercus prinus) acorn mast crops serendipitously occurred in year three (fall 2005) of a landscape-scale prescribed fire experiment. We examined establishment, survival, height and diameter of new seedlings on sites on the Cumberland Plateau in eastern Kentucky. Treatments were fire exclusion, a single prescribed fire (1x-burn; 2003), and repeated prescribed fire (3x-burn; 2003, 2004, and after acorn drop in 2006), all conducted in late spring. Initial densities of newly established chestnut and white oak seedlings were statistically similar across treatments (P = 0.42), despite fires on the 3x-burn site having occurred after acorns were on the ground. Oak seedling density was significantly predicted by oak basal area on all sites (R² = 0.12–0.46), except for chestnut oak on fire-excluded sites (R² = 0.04). Litter depth was less on 3x-burn sites compared to 1x-burn and fire-excluded sites, whereas canopy openness was greater on both burn treatments compared to fire-excluded sites. Seedling mortality was generally higher on fire-excluded sites compared to burn sites, especially for white oak. Oak seedling mortality in the first two growing seasons was significantly predicted by initial litter depth and open sky, with greater litter depth and lower percent open sky leading to higher mortality. In the third growing season none of the measured variables predicted chestnut oak seedling survival; for white oak, percent open sky remained a significant predictor of mortality. Initially, seedlings on the fire-excluded sites had similar height but smaller diameter; after three growing seasons there were few differences in seedling height or diameter among treatments. Our findings suggest a potential role for prescribed fire in establishing forest floor and light conditions that may enhance the success of new oak germinants, although different responses among species may suggest the need to target management for individual oak species.     

Establishment and growth of coniferous seedlings in an altered forest floor after long-term exclusion of deer

We evaluated differences between the forest floors and the establishment and growth of coniferous seedlings in fenced (13 years) and unfenced plots on Mt Ohdaigahara where the sika deer (Cervus nippon) population density is high. Large coniferous seedlings (height > 0.05 m) were less abundant in the unfenced plot, as a result of deer browsing. Small coniferous seedlings (height < 0.05 m), however, were more abundant in the unfenced plot, where most seedlings of Abies homolepis were found on bare ground and those of Picea jezoensis var. hondoensis were found on buttresses and fallen logs. The large area of bare ground in the unfenced plot was caused by deer browsing. Deer therefore have an indirect effect on the emergence and growth of small coniferous seedlings by modifying the forest floor.     

Effects of treeshelters and polyethylene mulch sheets on survival and growth of cork oak (Quercus suber L.) seedlings planted in northwestern Tunisia

Success of plantation establishment depends on the ability of seedlings to withstand animal browsing and competition from vegetation. This study in northwestern Tunisia examined the influence of browsing protection and vegetation control on seedling mortality, diameter, and height growth components of an evergreen sclerophyllous oak species, cork oak (Quercus suber L.) for the first 2 years after planting (2005 and 2006). Browsing protection was controlled by varying both treeshelter type (non-vented, vented, and control) and height (1.2-m and 1.8-m tall). Competing vegetation was controlled through increasing sizes of square-shaped black polyethylene mulch sheets: 0 m, 0.7 m, 1 m and 1.3 m large. Seedling mortality rate was relatively low during the first year (9%) and was not affected by any of the studied factors. During the following year, mortality was lower for seedlings in both types (non-vented and vented) of 1.2-m tall treeshelters than for unsheltered ones. Mortality was only slightly higher for 1.8-m tall shelters than for 1.2-m shelters. Compared to unsheltered seedlings, basal diameter was reduced during both seasons for seedlings inside non-vented treeshelters (1.2-m and 1.8-m tall), whereas for vented shelters it was reduced during the first year and then increased during the second year. Height growth was positively affected by tree shelters during both years. Seedlings inside non-vented 1.8-m tall shelters were the tallest having an average height of up to 140 cm at the end of the second year, while the unsheltered seedlings had only an average height of up to 42 cm. The greater height of the sheltered seedlings was attributed to the increase of: (1) the number of shoot flushes yearly established on the main stem (up to three), and (2) the length of all the growth units produced during each flush. Seedlings inside vented shelters had balanced growth with a low proportion of seedlings unable to support themselves without a stake (7%). However, non-vented shelters had unbalanced height and diameter growths resulting in both a higher height-to-diameter ratio and in a higher proportion of seedlings having stability problems (47%). Polyethylene mulch sheets did not affect early survival, seedling basal diameter, or number of shoot flushes established yearly, but did slightly improve seedling height at the end of the second season, mostly due to the significant improvement in shoot elongation of the first flush. Results suggest that both 1.8 m treeshelters and plastic mulches may enhance the growth of cork oak seedlings planted on harsh sites in northwestern Tunisia.     

Effects of 13-year fencing on browsing by sika deer on seedlings on Mt. Ohdaigahara, central Japan

We examined the effects of fencing on deer browsing on seedlings 13 years after the building the experimental fence on Mt. Ohdaigahara, central Japan, where the sika deer (Cervus nippon) population is high. There was no difference in the species number of seedlings between 1991 and 2004. The density of seedlings in the fenced plot in 2004 was three times higher than in 1991. Although the unfenced plots had no seedlings higher than 20 cm, there were seedlings up to 120 cm in the fenced plot. These results suggest that 13-year fencing promoted regeneration of seedlings in this area.     

The effect of seaweed concentrate on the growth of Pinus pinea seedlings

Seedlings of Pinus pinea L. growing in plastic containers were treated with seaweed concentrate (SWC). Different concentrations of SWC were applied, 0 to 3 times, to the roots or shoots of the seedlings. Shoot application increased plant weight mainly by increasing shoot growth. This was manifested as increased shoot length and weight and a decrease in the root/shoot ratio. Root drenches did not change the total plant weight but it accelerated root growth and increased lateral root dry weight. Root growth capacity (RGC) tests for both shoot and root applications indicated an increase in root length and some increases in root number when applied as a root drench. This study indicates that root application of SWC improved seedling quality and increased the ability of seedlings to survive transplanting into pots.Abbreviations GC-MS Gas Chromatography-Mass Spectrometry - RGC root growth capacity - SWC seaweed concentrate     

Effects of site conditions and methods of cultivation on growth of sawtooth oak plantations

The effects of site conditions and cultivation on the growth of sawtooth oak (Quercus acutissima Carr.) plantations were evaluated at the Hongyashan forest farm, in Chuzhou City, Anhui Province, China. The results indicate that the position on the slope, the amount of gravel and the thickness of the soil were important factors in the growth of the sawtooth oak. Lower slope positions with small amounts of gravel and a thick soil were better for the growth of this species than middle slope positions with more gravel and a thin soil. Given the site conditions of the hilly and mountainous areas in Chuzhou City, the mixed Chinese fir (Cunninghamia lanceolata Hook.) and sawtooth oak forests did not improve forest productivity compared with pure sawtooth oak forests. Both urea and compound fertilizers promoted the growth of sawtooth oak, as did site preparation and intercropping. Two years after planting, the height growth of ordinary seedlings with a starting height of 0.6 m was higher than that of supper seedlings with a starting height of 1.0 m. Compared with planting, the early growth of the coppices was faster, but the later growth of the coppices was slower. __________ Translated from Journal of Fujian College of Forestry, 2008, 28(2): 130–135 [译自: 福建林学院学报]     

Effect of animal browse protection and fertilizer application on the establishment of planted Nuttall oak seedlings

For establishment to be successful, planted oak must emerge above vegetative competition and browse level. A 3 year study was established in 2000 in a cutover bottomland hardwood forest in west Alabama to examine the influences of seedling browse protection and fertilizer use on growth of Nuttall oak (Quercus nuttallii Palm.) seedlings. The treatments included: control, 1.2 m tall wire tree shelter, and 1.2 m tall plastic tree shelter. Fertilizer (20–10–5) was applied to one-half of all seedlings at planting. Competing vegetation was controlled around all seedlings with mulch mats and herbicide. Fertilization did not affect height growth or successful establishment after 3 years. Four percent of protected seedlings (plastic and wire tree shelters combined) were browsed compared to 95% of control seedlings. After 3 years, plastic tree shelters were the most effective treatment for promoting height growth and successful seedling establishment of Nuttal oak seedlings (193 cm, 92.71% compared to 120 cm, 68.75% and 52 cm, 8.33% for wire tree shelter, and control seedlings, respectively). Animal browse protection is deemed essential for successful seedling establishment in southern bottomland forests where deer density is high (density estimated at 27 km⁻² in vicinity of study area).     

Impact of deer on temperate forest vegetation and woody debris as protection of forest regeneration against browsing

A 3-year field experiment with paired exclosure (fenced areas, excluding deer) and control plots (unfenced areas, free access to deer), with two treatments with and without woody debris, was carried out at two sites in a temperate forest in eastern France. The aim of the experiment was to assess the effect of browsing by roe deer (Capreolus capreolus) and red deer (Cervus elaphus) on the diversity and richness of plant species and to test the effectiveness of using woody debris to protect seedlings and saplings from deer browsing. In presence of deer, both plant species richness and diversity were reduced the first year of the study, but this negative impact of deer then disappeared after 3 years. Deer browsing mostly affected species composition of plant communities. We observed a decrease in the abundance of preferred species such as Carpinus betulus, Rubus fructicosus, Rubus idaeus, Anemone nemorosa and Epilobium angustifolium, and palatable species such as Acer spp., Carex spp., Festuca spp. and Mycelis muralis, whereas unpalatable species such as Lamium spp., or species particularly resilient to browsing such as grasses (Brachypodium spp. and Luzula spp.) increased in abundance. The use of woody debris as protection against browsing by deer did not limit damage to seedlings and saplings of the main commercially valuable species, Abies alba and Quercus spp. Instead of limiting deer impact, use of woody debris seemed to increase the negative effect of deer browsing on regeneration in control plots relatively to those without protection.     

Influences of woody encroachment and restoration thinning on overstory savanna oak tree growth rates

Midwestern savannas historically covered >10 M ha in central North America, but are now rare due to agricultural conversion and anthropogenic modifications to disturbance regimes - particularly fire suppression. Throughout this range, Midwestern savannas are characterized by scattered overstory trees; however, with fire suppression, these systems are invaded by non-savanna trees. Restoration of encroached savannas involves removal of invading trees, yet little is known about the impacts of encroachment or encroachment removal on the relict savanna overstory trees, which define these systems. Here, we use tree ring analysis to investigate savanna tree growth rates in encroached, non-encroached, and experimentally restored Midwestern oak savannas in central Iowa. We found that woody encroachment led to pronounced declines in growth rate (ring width) of relict overstory white oak (Quercus alba), relative to Q. alba trees in competition-free, open-grown conditions, or in an encroachment-free remnant woodland. To further understand effects of encroachment removal on relict Q. alba savanna trees, we conducted a large-scale restoration experiment, where encroaching trees were mechanically removed from four encroached savannas, with an additional four savannas retained as encroached controls. Restoration led to elevated tree growth rates, with these changes generally persistent through 7 years post-restoration (2003-2009). Over the course of this post-restoration study period, ring width, basal area increment, and relative basal area increased by 49%, 59%, and 55%, respectively, in trees from restored sites, relative to trees from encroached, control sites. These results suggest that woody encroachment has strong influence on overstory savanna trees, through increased competitive dynamics; however, woody encroachment removal may help to restore relict savanna tree growth rates, even after prolonged periods of encroachment (>40 years). To restore the oak savannas at our sites, and perhaps elsewhere, we advocate a three step process: (1) mechanical woody encroachment removal, (2) maintenance of the encroachment-free state through prescribed fire, and (3) promotion of a diverse understory layer, characteristic of oak savanna in our region. While promoting oak regeneration will be important for the long-term maintenance of these sites as oak savanna, relict savanna trees appear responsive to restoration and should maintain overstory conditions through the near-term.