The following sections discuss the application of different regeneration harvest methods to naturally regenerate oak forests.

• Even-aged systems: Systems that use regeneration harvest methods to create even-aged stands. By convention, the spread of ages in an even-aged stand does not differ by more than 20% of the intended rotation.

  • Clearcutting
  • Shelterwood cutting
  • Seed tree cutting

• Uneven-aged systems: Systems that use regeneration harvest methods to create uneven-aged stands. Commonly at least three distinct age classes are required for an uneven-aged stand.

  • Group selection cutting
  • Single tree selection cutting

• Two-aged systems: Systems provide for regeneration of shade intolerant species while carrying a sparse overstory of mature trees to offset the negative aesthetic impact of clearcutting.

For a primer on concepts of silviculture, see: Silvicultural systems.

Two-aged silvicultural systems involve a planned sequence of treatments designed to maintain and regenerate a stand with two age classes. Two-aged systems are emerging as a promising management alternative for oak forests.Two-age methods have the ability to deliver several important biologic and managerial advantages compared to clearcutting. Specifically they can provide: maintenance of sexual reproduction throughout rotation; maintenance of advance reproduction development throughout the rotation; reduced visual or aesthetic impact, especially compared to clearcutting; development of large diameter high value sawtimber or veneer trees; and, development of a wide range of multiple products. (Stringer 2000)

The maintenance of sexual reproduction and potential development of advanced regeneration throughout a rotation maybe critical in maintaining the ability to regenerate stands composed of oak. A typical clearcut may regenerate oak from stumps and high vigor advanced regeneration present at the time of the cut. However, clearcutting stops sexual reproduction for a portion of the rotation. While some oak trees grown in seed orchards can produce acorns between 10 and 20 years, it is probable that significant acorn production in naturally regenerating stands may take significantly longer. This coupled with the sporadic nature of good acorn crops limits the number of opportunities for developing advanced oak regeneration before the next harvest. Unlike a clearcut, the reserve trees in a two-age method ensure that sexual reproduction will occur throughout the rotation. In a sense, the reserve trees are acting as seed trees for the next rotation. This, in turn, provides a longer period of time for advanced regeneration development compared to even-aged methods. Therefore, two-aged systems offer more possibilities in helping with oak regeneration problems than clearcutting systems. (Stringer 2000)

A number of silvicultural research efforts are currently evaluating the two age system for use in eastern hardwood forests. Johnson (1998) evaluated residual tree quality and regeneration 2-5 yr after cutting in Monongahela National Forest, West Virginia. He found that the largest grade reductions were due primarily to epicormic branching and logging wounds, however tree grades were not significantly affected by the cutting. Also, regeneration of both shade-intolerant and tolerant species was prolific following cutting; in the mixed oak stands of the Ridge and Valley Province the number of tree seedlings and sprouts was 8,217stems/ac (Johnson 1998). The results of a 15 year study in southeastern Kentucky indicate that small sawtimber sized co-dominant reserve white oak trees are capable of successfully producing advanced regeneration. Results also indicated that dramatic increases in advanced regeneration density can be obtained when the combined understory, midstory, and overstory exhibit a densiometer reading >6 (Stringer 2000).

The seed tree method is an even-aged method that provides for seed production after most of the parent stand has been harvested. A typical application leaves 10 or fewer seed-producing trees per acre (Smith 1986). However, the method generally has not been recommended for regenerating oaks because the seed trees are likely to contribute too little too late to a stands regeneration potential. This method favors wind-blown, light-seeded species (Toliver and Jackson 1989) and not heavy-seeded species, such as oak, in which seeds are dispersed for relatively short distances from the parent tree by gravity, birds, rodents, or sometimes water.

Because most oak regeneration develops from advance reproduction present at the time of cutting or from the sprouting of stumps, there is little need or justification for seed trees (Johnson and Krinard 1976). In the few reported cases where the method has been applied to oak stands, the seed trees had little or no effect on regeneration (DeBell et al 1968, Johnson and Krinard 1983). Twenty-nine years after a seed-tree cut in bottomland sweetgum-red oak stands in Arkansas, Johnson and Krinard (1988) reported that the resulting stands were essentially similar to those produced through clearcutting. Most of the oak regeneration developed from advance reproduction and stump sprouts. Seed trees did not significantly contribute to the establishment of new seedlings. (Clatterbuck and Meadows, 1993)

Nevertheless, the seed-tree method may be useful in some ecosystems if combined with competition control, as discussed for clearcutting, and it may be useful if sustaining acorn production on harvested areas is important.If good seed producers are retained, substantial acorn production can be ensured. Since only a fraction of individual oaks are good acorn producer in a given population, leaving as little as 10 good acorn producers per acre could retain 40 percent or more of the acorn producing capacity of the original stand. Moreover, under the open-grown conditions created by the seed tree method, seed tree crowns can expand to their maximum, thus increasing acorn production per tree. However, identifying the seed producers requires long-term records on seed production, which few forest managers are likely to have. (Johnson 1993)

Single tree selection is very restricted in its application in the southern Appalachians. Assessments following cutting intensity experiments show that single-tree selection has little effect on increasing understory light intensity. For example, to increase light levels by 30% in northern hardwoods, the treatment must remove about 60% of the stand basal area. Most Southern Appalachian forests are comprised of canopy species that are intolerant or intermediate in their tolerance of shade, such as oak, yellow-poplar, black cherry, sweetgum, and ash; single-tree selection does not regenerate these shade intolerant species. Therefore, the single tree selection method generally is considered inappropriate for regenerating oak forests (Hibbs and Bentley 1983, Sander and Clark 1971). Canopy openings the size of individual trees do not provide adequate light and space for the recruitment of oak while the development of a tolerant understory tends to offset any regeneration advantages resulting from reduced overstory density (Auclair and Cottam 1971, Della-Bianca and Beck 1985, Loucks and Schnur 1976, Schlesinger 1976, Trimble 1970). The measure of a successful application of single tree selection must be, in the long run, whether or not a sufficient number of new trees are established and recruited into successively larger size classes over time to replace trees that are harvested. Studies have indicated that in the single tree selection method, shade-intolerant species can and do become established as seedlings, but are unable to develop adequately through the sapling and pole stages (Della-Bianca and Beck 1985). For example, Crow and Metzger (1987) indicated that by dropping the residual basal area in single-tree selection cutting to 40 ft²/ac considerable regeneration of intermediate and intolerant species can be obtained, but this compromised their ability to grow large diameter trees. Carvell (1967) noted that a single-tree selection cutting did release understory oak seedlings in a West Virginia study, but seedlings with a defined apical growing point responded better than flat-topped seedlings. (Hicks 1998) Attempts to remove enough trees to encourage regeneration and development of the more valuable, light-demanding species also poses the risk of reducing stand growth and quality through understocking (Della-Bianca and Beck 1985, Trimble 1973). (Beck 1988) At Bent Creek Experimental Forest in North Carolina, studies of the single tree selection method have been conducted for almost 50 years. Results of these studies have shown that the single tree selection method promotes regeneration of dogwood, sourwood, red maple, sweet birch, but is unsuccessful in regenerating common overstory species such as oaks and yellow poplar (USDA Forest Service 1995). Clatterbuck (1993) also concluded that single-tree selection is inappropriate for white oak, because white oaks that grow in subordinate canopy positions become poor candidates for future management. In general, most studies of the single tree selection system conclude that it decreases stand value by causing lower volumes and promoting lesser valued tolerant species (Smith et al. 1983). The single tree selection method is not recommended for oak management in bottomland forests as well. There are few commercially valuable shade-tolerant bottomland species; most bottomland oaks range from intolerant to intermediate in shade tolerance (Putnam and others 1960, Toliver and Jackson 1989, Clatterbuck and Meadows, 1993). Another challenge of implementing the selection system in Appalachians hardwoods is creating and perpetuating a balanced uneven-age condition in existing stands that are even-aged. Although many central hardwood stands currently have a diameter distribution that approximates a reverse S shape, which differs from the theoretical bell-shaped distribution of pure even-age stands, this occurs as a result of the mixed-species composition and differential growth rates of different species as opposed to different age classes. Due to the developmental history of these stands, they currently contain many species that are intermediate or intolerant of shade and have few representatives of the shade tolerant species that would be needed for single-tree selection management. (Hicks 1998) Despite the many failures of the single tree selection system in the central hardwood region, this method has been used successfully in some cases. For example, the single tree method has been used successfully for 50 years on a large industrial forest in the Missouri Ozarks. However, in this case, the success of this method was largely due to the intrinsic diameter distribution of the forest, characterized by numerous small-diameter white oaks. Thus, the relatively large pool of smaller overstory oak; plus the accumulated reproduction characteristic of the Ozark forest are potentially available for sustained recruitment into the main canopy. The results of this study suggest that the single tree selection method deserves further study for application auto-accumulator ecosystems. (Johnson 1993)

Single-tree selection methods are appropriate for stands in which the desired species composition is to be composed of primarily shade-tolerant species (Smith 1980, Leak and Gottsacker 1985, Lamson 1983, Mills et al. 1987). These species include hemlock, white pine, sugar maple, or beech. Single tree selection has been successful in the beech-birch-maple forests of the northeastern United States, but the distribution of this type in the Southern Appalachians is extremely limited. Trials are currently underway in white pine, but research efforts to use single tree selection in mesic Southern Appalachian hardwoods, and in mesic to somewhat xeric oak stands, have been unsuccessful. Single-tree selection has been successful in loblolly pine stands in the South, but only with the application of herbicides to control hardwood competition. (SAMAB 1996e)

These conflicting conclusions regarding oak regeneration may derive from the historical background of the different stands; therefore, in each situation where silvicultural options are being considered in the central hardwood region the final decision will be an informed judgment based on ownership goals, economic factors, historical factors, site, and current stand conditions. (Hicks 1998)

Clearcutting has been the most widely recommended and applied regeneration method, albeit with mixed success, in the management of oak forests. It was most successfully applied in the drier ecosystems where oak reproduction naturally accumulated in the understory. Such forests occurred in the Missouri Ozarks and the oak-hickory forests of the Ohio Valley and Appalachians (Sander 1977, Ross et al. 1986). Clearcutting was less successful in regenerating oaks in the more mesic ecosystems of the Ohio Valley, Appalachians, and elsewhere (Beck and Hooper 1986, Gammon et al. 1960, Johnson 1976, Loftis 1983b). Failures were largely related to the inability of oak reproduction to accumulate in the heavily shaded understories of these forests (Johnson, 1993a).

In the 1960s, even-aged management became the modus operandi for much of the central and eastern hardwood region. It was enthusiastically accepted for several reasons. It met the ecological requirements of the commercially valuable, shade-intolerant species, including the oaks, white ash (Fraxinus americana L.), black cherry (Prunus serotina Ehrh.), and yellow-poplar (Liriodendron tulipifera L.). Even-aged management also was economically efficient. Logging, road building, and administrative costs were minimized because large timber volumes could be harvested from relatively small areas. Moreover, the system was simple to implement, in contrast to the relatively complicated timber marking rules used to create the diameter distributions needed in uneven-aged management (Roach 1963, Smith and Lamson 1982, Trimble et al. 1974). There also was growing uncertainty about the ability of the commercially valuable intolerant species, including the oaks, to sustain the age distributions required in uneven-aged management (Roach 1963, Schlesinger 1976). Other factors that contributed to the acceptance of clearcutting included its endorsement by wildlife managers, the development of even-aged stocking guides (e.g., Gingrich 1967), an established scientific basis for its application (Roach and Gingrich 1968), and its utility in regulating the distribution of age classes of stands. Clearcutting is also recommended where tree quality and stocking are poor (such as in high-graded stands) and there is little potential to upgrade the stand (Kellison and others 1988). The many advantages of clearcutting upland hard wood forests have been further discussed by McGee (1987). (Johnson, 1993a)

In the first two decades of clearcutting (1960s and 1970s), forest managers were not greatly concerned about a diminishing oak resource even though many were aware of the problem. Apathy was reinforced by the economic acceptability of the species that usually replaced oaks in mesic ecosystems. Although there was some concert over increases in the proportion of the less valuable shade-tolerant hardwoods such as sugar maple (Acer saccharum Marsh.), red maple (A. rubrum L.), and American beech (Fagus grandifolia Ehrh.), clearcutting usually provided an acceptable mix of species even though the oak component was greatly reduced (Beck and Hooper 1986, Johnson 1976, Loftis 1983b).

That apathy diminished during the last decade for two reasons. First, in the late 1970s, there was a large increase in the demand and value of oak logs (Hoover 1985). This spurred management specifically for oaks and a concern for the economic losses that accompanied a shrinking oak resource. Second, the rise of environmental activism drew public attention to the perceived negative consequences of clearcutting. These perceptions included decreased aesthetic value, biodiversity, old growth, and certain wildlife values along with increased forest fragmentation. (Johnson, 1993a)

Patric and Schell (1990) indicated that much of the controversy over clearcutting is based on the misconception that clearcutting is simply a harvest-only technique and is frequently confused with the practice of high-grading, or cutting only good and valuable trees and leaving the poor ones. The concern expressed that clearcutting leads to erosion is also generally unfounded, and most of the soil movement that occurs during forest harvesting is preventable since it results from poorly constructed roads (Kochenderfer 1970). (Hicks 1998)

To counter concerns about the possible adverse impacts of clearcutting oak forests, various modifications of the method were advocated. These included reducing the size of clearcuts; leaving snags, cull trees, and uncut islands of trees; deferring the removal of non-commercial residual trees; shaping cuts to fit more aesthetically into the landscape; and leaving uncut strips where clearcuts bordered roads, lakes, streams, and other sensitive areas (Evans and Conner 1979; Smith et al. 1989; USDA Forest Service 1973, 1980). Although all of these methods were compatible with attaining conventional oak regeneration objectives, they also were consistent with other management objectives such as improving wildlife habitat for some species and enhancing landscape aesthetics and biodiversity (Perry et al. 1990). Nevertheless, the use of clearcutting on public lands, especially the National Forests, has further declined in favor of systems that focus less on producing commodities and more on preserving biodiversity and other values (Gale and Cordray 1991, Hansen et al. 1991, Kessler 1991, Perry and Maghembe 1989, Salwasser 1990). (Johnson, 1993a)

The clearcutting regeneration method has been applied in the Southern Appalachians for several decades and research on this technique dates back to the 1920s. The accumulation of knowledge on this regeneration method has revealed several general patterns.

First, species composition following the clearcutting method is generally related to site quality. On high-quality mesic sites and in cove stands, species composition following the clearcutting method is generally dominated by yellow-poplar seedlings, sweet birch seedlings, stump sprouts of red maple, and root sprouts of black locust. It has been observed that stands become increasingly dominated by yellow poplar with development (Beck and Hooper 1986). At elevations of 3,500 to 4,500 feet in the southern Appalachians numerous black cherry seedlings often appear after clearcutting.(Hicks 1998)

On medium-quality sites dominated by upland oak species, species composition following the clearcutting method is generally dominated by the same species that were harvested, primarily white, scarlet, and chestnut oak with several pine species as associates. As previously noted, more sproutable stumps are usually present on such sites than on the better quality cove sites. Also, greater numbers of advance oak seedlings are frequently present due to the less competition from tolerant understory species. Therefore, if the clearcut stand is on southeast or northwest middle and upper slopes, we can expect to have a stand at about age 20 that can be molded into an essentially pure oak stand by thinning. On north and east aspects and lower slopes, the stand composition may be highly variable. (Sander and Graney, 1993)

Second, if advance growth dependent species, e.g. oaks, are desired on these high-quality sites, large advance reproduction of these species must be present at the time of harvest if they are to represented as dominant and codominant stems in the new stand. Herbicide application prior to, during, or immediately after harvest is often necessary to control sources of competing species. (USDA Forest Service 1995)

Third, an opening of at least 1-2 acres in size is required in order to create the openness needed to produce the characteristics of a clearcut (Sander 1992; Dale et al. 1994). Stands smaller than this have a large proportion of their area in a zone around the stand border where reproduction growth will be slow because of the influence of the surrounding trees. (Hicks 1998)

In spite of these general patterns, Loftis (1988b) acknowledges that the type and amount of regeneration following clearcutting can be quite variable, and there are instances where clearcutting has not achieved the desired objectives. For example, Elliott and Swank (1994) reported that a southern Appalachian hardwood stand clearcut in 1939 and again in 1962 showed a decrease in frequency of oaks, due in part to the regeneration strategies that are initiated when clearcutting a very young stand (notably sprouting).(Hicks 1998)

On many sites in the southern Appalachians, the clearcutting method increases unwanted competing or allelopathic vegetation, rather than the desired species (Boring et al. 1981; Leopold and Parker 1985). Horsley (1988) lists a number of woody and herbaceous species as undesirable competing vegetation, including ferns, grasses, brambles, rhododendron, mountain laurel, grapevine, striped maple, sourwood, dogwood, pin cherry, sassafras, and blackgum. It is important, before clearcutting, to assess the potential for such competitive interactions. The options for control vary, according to the type of competing vegetation being managed, but may range from selecting an alternative regeneration system to use of prescribed fire or herbicide treatment before or after cutting. Since most selective herbicides kill broadleaf species, it is not practical to use broadcast spraying after hardwood regeneration has already become established. In such cases, spot spraying, injection, or basal spraying may be required. Both herbicide injection and/or basal spray of individual stems shortly before harvest (Loftis 1985) and cut-stump treatment with herbicides at the tire of the harvest cut have met with success in the southern Appalachians (Zedaker 1987). Cutting competing vegetation without the use of herbicides is not effective due to the rapid regrowth of sprouts from the cut stumps. These treatments are expensive and labor intensive; thus, where such problems are expected, it may not be economically feasible to do them, particularly for the small private landowners typical of the southern Appalachians. Ferns and grasses have an advantage over tree seedlings in highly compacted soils so it is imperative that soil compaction be minimized during logging. (Hicks 1998)

Deer browsing also contributes to the failure of regeneration (Marquis and Grisez 1978). Smaller isolated clearcuts are particularly vulnerable since they serve as an attractant for deer. In addition to using larger clearcuts, leaving slash piles scattered through the clearcut helps in discouraging deer and promotes regeneration success. A last resort for obtaining regeneration in areas with very high deer population levels is fencing (Marquis and Grisez 1978) however, fencing alone may not be enough to ensure regeneration. (Hicks 1998)

In a few situations in the central hardwood region natural regeneration may need to be supplemented by planting or direct seeding following clearcutting (Pope 1993). Plantings can be successful but are expensive and will probably require weed control or protection from deer during the first few years, particularly on good sites (Stout 1986; Walters 1993). Although planting may be desirable under certain circumstances (Davidson 1988), cost and uncertainty of success limit its practice. (Hicks 1998)

1. Determine size of area to be designated as a stand. Stand size can vary but should be at least 2 acres. Stands smaller than this have a large proportion of their area in a zone around the stand border where reproduction growth will be slow because of the influence of the surrounding trees. A stand should preferably be restricted to a single condition or size class of timber and site quality. Size of the forest property will also influence stand size. In areas where deer populations are high, stand size may need to be relatively large in order to reduce the impact of browsing on reproduction growth.

2. Arrange and shape clearcuts so they mingle with uncut stands and blend into the landscape as much as possible.

3. Plan, construct, and maintain skid trails and logging roads to minimize erosion.

4. Harvest all merchantable trees.

5. Cut or kill remaining culls and small trees larger than about 2 in. d.b.h. Killing the unwanted trees instead of cutting them reduces sprouting and provides snags for nesting holes and perches for birds. Cutting some of them will provide habitat for other wildlife species such as ruffed grouse.

6. There may be some other options for harvesting and getting rid of the unwanted and unmerchantable trees. These options are generally used for special purposes, such as to soften the aesthetic impact or provide special wildlife habitat. Their use depends on owner policy rather than silvicultural desirability. (Sander and Graney, 1993)

If the clearcut stand is on southeast or northwest middle and upper slopes, we can expect to have a stand at about age 20 that can be molded into an essentially pure oak stand by thinning. On north and east aspects and lower slopes, the stand composition may be highly variable. In the mixed and western mesophytic forest regions, yellow-poplar will likely be abundant. Other species such as white ash, black cherry, and red and sugar maples will also be present. However, if the oak advance regeneration is adequate, we can expect to have a predominantly oak stand 20 years after clearcutting. (Sander and Graney, 1993)

When clearcutting to regenerate oaks in bottomlands, the harvest should take place during the dormant season to maximize stump sprouting (Kellison and others 1988). Control of residuals either by herbicides or cutting is advocated, especially after a commercial clearcut (Golden and Loewenstein 1991), to reduce competition from undesirable midstory and understory species (Janzen and Hodges 1987). Clearcutting is also recommended where tree quality and stocking are poor (such as in high-graded stands) and there is little potential to upgrade the stand (Kellison and others 1988). (Clatterbuck and Meadows, 1993)

Probably the most frequently cited method for regenerating oaks is the shelterwood method (Beck 1991, Hannah 1987, Jacobs and Wray 1992, Korstian 1927, Sampson et al 1983, Sander 1979, Scholz 1952, Smith 1986). The method is potentially suited to regenerating northern red oak where the species most frequently occurs- in the troublesome mesic ecosystems where it is difficult to obtain the accumulation and development of oak advance reproduction. The essential feature of the shelterwood method is a reduction in stand density in one or more steps near the end of the rotation to encourage establishment and/or growth of reproduction (Smith 1986). Normally, the final overstory removal is made when established reproduction is deemed adequate for replacing the parent stand (Beck 1991, Smith 1986) (Isebrands and Dickson, 1994).

The shelterwood method appears to be especially well suited for regenerating species that are intermediate in shade tolerance and have slower initial growth. In the southern Appalachians, this fits the description of the oaks. The shelterwood method for oak regeneration is primarily used to establish new seedlings of the desired species and/or to encourage the development of large seedlings from existing smaller ones, so that they will be able to successfully compete with other vegetation when the overstory is removed (Hodges and Janzen 1987, Loftis 1990). In the central hardwood region, shelterwoods work for other species of intermediate tolerance such as ash, basswood, hickory and others whose advance regeneration must be relatively large in order to compete successfully. (Johnson, 1993a)

Where oak regeneration problems occur, the shelterwood method may be the most effective solution, especially in the troublesome mesic ecosystems (Isebrands and Dickson, 1994). In mesic ecosystems, successful application of the method requires coordinating overstory density control with competition control and, in some cases, acorn production. oak regeneration is easier to establish on the poorer upland sites (Carvell and Tryon 1961). (Johnson, 1993a)

Research in the southern Appalachians has demonstrated the potential of the shelterwood method to regenerate northern red oak on highly productive sites (site index 70 to 90 feet at 50 years) (Loftis 1983b, 1990a, 1990b). This research focused on finding a level of disturbance that enhanced growth of red oak seedlings, but also impeded the establishment and growth of yellow-poplar, the primary shade-intolerant competitor.The resulting prescription called for a single shelterwood cut that reduces stand basal area by only 30 to 40% of full stocking. In this initial cut, stand density is thinned from below by removing the subcanopy of tolerant non-oaks using herbicides and leaving the main canopy intact. The shelterwood is then removed about 10 years after the initial cut treatment. Studies in the southern Applachians have shown that if understory competition is not controlled after this initial cutting, sources of competition from red maple stump sprouts and dogwood will displace oak regeneration. These studies have also verified that over a broad range of residual basal areas (25-66 sq ft/ac), species composition of the new age class is a function of the regeneration sources present at the time of the initial cut. (USDA Forest Service 1995)

It is critical to understand that the shelterwood method practiced on non-accumulating oak ecosystems, such as mesic sites in the southern Appalachians, generally does not result in the establishment of new oak seedlings (Janzen and Hodges 1987, Loftis 1990).  Non-accumulating oak ecosystems are those where no significant new oak seedling establishment is occurring over time.  Loftis (1990b) points out the results of several authors (McGee 1975; Beck 1977; Auchmoody et al. 1993) who found that opening up oak stands did not necessarily result in increased acorn production. Oak regeneration takes many years to form, and often develops only in years with (1) a heavy seed crop; (2) low predation by seed-eating insects and animals; (3) seedbeds of only thin leaf litter; (4) favorable weather conditions; and (5) freedom from interference by understory vegetation or faster-growing competing plants. These conclusions seem to be verified by Schuler and Miller (1995) in central West Virginia, where they reported that 10 years after a shelterwood seed cutting, desired regeneration of northern red oak was still inadequate. Thus, Loftis proposed that the objective of the first cuts in an oak shelterwood was as a release cut to enhance already existing oak regeneration, rather than preparatory and seed cuts to create regeneration (Tryon and Carvell 1962). (Johnson, 1993a)

Contrasting experiences in the southern Appalachians and other areas in the central hardwood region emphasize the need for shelterwood prescriptions that are both species- and ecosystem-specific. The minimum duration of the shelterwood period, and the length of time between the seed cut and shelterwood removal, varies among sites. Xeromorphic oaks such as black oak, white oak, scarlet oak, and chestnut oak growing in xeric ecosystems may require two or more decades in the understory before they accumulate the root mass necessary for competitive shoot growth after overstory removal (Sander 1979a, Sander and Clark 1971). However, because seedlings of these species typically accumulate in the understories of xeric ecosystems in which they typically occur, use of the shelterwood method is often unnecessary there (Sander and Clark 1971). In contrast, studies on regenerating northern red oak in the Driftless Area of Wisconsin indicate that a shorter establishment period of 1 or 2 years may be sufficient. Results from one case history in southwestern Wisconsin even indicate that northern red oak can be successfully regenerated with little or no oak advance reproduction provided that the final harvest occurs during the dormant season following a good acorn crop (Johnson et al. 1989). In mesic and hydric ecosystems, a longer accumulation period may be required when applying the shelterwood method. To be effective, this may require control of both overstory and competing understory density (Johnson, 1993a).

The shelterwood system also can be used in combination with underplanting (Johnson et al. 1986, Johnson 1989). This method provides the opportunity to introduce genetically improved stock into stands. However, there is little theoretical evidence that application of the shelterwood system is dysgenic (Howe 1989).

The steps in applying the shelterwood method are (Sander and Graney, 1993):

1. Determine the size of shelterwood cuts based on the overall management goals and the size of the property. In order to create even-age characteristics, they should be at least 2 acres in size. Shape and arrangement of even-age stands should blend with the character of the landscape.
2. Control the understory, including herbaceous species, shrubs, and undesirable advance reproduction, by cutting or preferably killing the non-oak species that will compete with the small oaks in the oak-hickory and oak-pine forest regions. Herbicide treatments have been used with success for controlling competition.
3. The initial shelterwood cut should be made at least 10 years before the end of the rotation (Johnson, 1992). On site indexes of 70, 80, and 90, recommended residual basal areas are 60, 65, and 70% of initial basal area, respectively. Reducing stand density below these limits increases the establishment of yellow-poplar seedlings to densities that are incompatible with regenerating red oak. Lower crown class trees should be removed first to create a uniform overstory which maintains crown closure to suppress intolerants, while still increasing light intensity at the ground level. Kill all stems of unwanted species. Trees of desired species and quality should be retained as a seed source and for future harvests. Leave the best dominant and codominant oaks as uniformly spaced as possible.
4. If possible, apply the understory and overstory treatments before seedfall in a good seed year. Red oak also can be successfully planted under shelterwood if large nursery stock is used and problematic understory vegetation is controlled. Plantings are most cost effective when they are designed to supplement a stands natural regeneration potential. (Isebrands and Dickson, 1994)
5. Monitor seedling establishment and growth and make additional light cuts to keep the overstory from restricting growth. Apply additional understory control if the understory redevelops to a point where it restricts the oak reproduction growth. This control may be desirable 5-10 years after the original treatment, particularly on high-quality sites.
6. When the regeneration potential of the oak reproduction is adequate to replace the stand, remove the remaining overstory trees in one cut. It is recommended that stand stocking should not be allowed to exceed about 75% to 80% stocking before the shelterwood removal. If this occurs then an additional cut to reduce stocking to 50-60% may be desirable. When the oak reproduction averages 4.5 feet and there are at least 435 trees per acre, the overstory trees should be removed (Mills et al. 1987, Sander and Graney, 1993).

Overall, this sequence of treatments may span at least twenty-five to forty years (Sander 1979; Sander et al. 1984).

Uses of site preparation techniques in the shelterwood and seed-tree methods can be similar to those used in the clearcutting method. Since natural regeneration is the focus of these two regeneration systems, site preparation is often used to improve seedbed and soil conditions, kill or set back interfering vegetation, and reduce hazards from fire and pests. However, in oak forests, existing advance regeneration may be utilized as the new cohort in the shelterwood method. Therefore, site preparation methods must be used that preserve advance regeneration. Foresters must also take special measures to safeguard the health and condition of seed and reserve trees, the seed stored in the litter layer, and the stumps and root systems that would contribute sprouts or root suckers to the new age class (Nyland 1996).

In the southern Appalachians, controlling competition, such as interfering herb, shrub, and tree layers, is crucial to the success of shelterwood methods, particularly in mesic ecosystems. Various methods have been tried to control competition in conjunction with the shelterwood method, including herbicides, scarification, and fire.

On mesic sites in the southern Appalachians, a precutting treatment (usually with herbicides) may be necessary in cases where an understory of undesirable tolerant species has become established under the canopy (Horsley 1981; Loftis 1985). Red maple also commonly develops in the understory of oak stands (Lorimer 1984), and, if these understory plants are not controlled, they will inhibit the new seedling stand that is established as a result of the shelterwood seed cut. At Bent Creek Experimental Forest, stream-line and cut surface herbicide application treatments were successfully used to kill competing vegetation (USDA Forest Service 1995).

Prescribed burning may also be used in combination with the shelterwood method for regenerating oaks (Nyland et al. 1982, Van Lear and Waldrop 1988). When burned by light surface fires, flat-topped oak seedlings resprout vigorously from their well-developed root systems and these seedling sprouts are often more competitive than the plants from which they originated. Frequent low-intensity back fires with low flame heights are most useful for building up oak advance reproduction under a shelterwood as such burns do not damage the overstory or site. In contrast, high intensity head fires are likely to cause excessive overstory mortality because the flames produced may reach into tree crowns. Such fires also may cause site damage. Single low-intensity burns are not likely to be effective (Johnson 1974. Nyland et al. 1982, Will-Wolf 1991). Doing site preparation work prior to seed cutting when no logging slash occupies the site often reduces costs and improves results. Logging slash will partially shade the new seedlings, and protect them from desiccation and heat damage. In addition, prescribed fires often become too hot and too difficult to control on sites covered with dry logging slash (Nyland 1996).

The shelterwood method has several disadvantages. First, if the objective is to regenerate oaks or some other species that are intermediate or tolerant of shading, these methods may not be effective, especially on above-average sites where shade intolerant species, such as yellow-poplar, are more likely to be competitive. This point is discussed in more detail below. Second, the reserve trees may be at greater risk of damage from lightning strike, windthrow, or ice damage. In soils that are prone to uprooting, the shelterwood technique is not advisable. Extreme care must be exercised to avoid logging wounds to the reserve trees since such wounds can result in decay that will greatly reduce their future value. The heavy cutting and opening of the stand can result in the formation of epicormic branches on reserve trees, which will reduce their value. Species that are prone to formation of epicormic branching, such as white oak, and individual trees with a tendency toward branching on the lower bole should be avoided for two-age silvicultural methods (Miller 1996). Removal of the reserve trees inevitably damages or destroys a portion of the new cohort, and in some cases, leaves poorly stocked new stand. Finally, there might be a temptation to perform a commercial high-grading and call it two-age silviculture. (Hicks 1998)

In the central hardwood region, the intensity of management required to apply the shelterwood method is also a disincentive, especially where small non-industrial landowners are concerned. Landowners may succumb to economic incentives to harvest shelterwood seed trees or reserve trees early (where two-age silviculture is proposed). The shelterwood method also requires a long time commitment on the part of the landowner. Therefore, although the shelterwood method appears to offer a compelling biological rationale in the central hardwood region, its application on a large scale will not likely occur unless some of these disincentives are nullified. (Hicks 1998)

Inconsistent results with the shelterwood method throughout the central hardwood is perhaps the foremost deterrent to applying this technique. Much of the inconsistencies with the shelterwood method is related to site quality. On fair-to-poor growing sites, oaks can be regenerated by even-aged silvicultural methods such as clearcutting or shelterwood (Roach and Gingrich 1968, Sander and Clark 1971). On better sites, regenerating oak becomes more difficult. Although there are examples of successful cases, these have occurred when overstory and understory treatments have been fine-tuned to meet site requirements. In the southern Appalachians, success has required finding a level of disturbance that enhances oak seedling growth, but impedes the establishment of yellow-poplar. Thus, applying the shelterwood technique has proved to be both an art and a science, and perhaps for this reason may not be an attractive choice for individual land owners. Smith (1993) noted that application of the shelterwood method is minimal in the Central Appalachians due in part to a lack of transfer of technical knowledge and a lack of confidence that the technique works or will work for the given situation.

A fair number of experimental trials of the shelterwood method conducted in regions outside the southern Appalachians have been reported as failures. On the Fernow Experimental Forest, a shelterwood study examined the effects of site quality (red oak site index 70 and 80), overstory residual stocking treatments (45, 60, 75, and 100 percent), understory herbicide treatments, and underplanting with red oak seedlings. Only on plots where advance oak regeneration existed before the initial shelterwood treatment, did red oak dominate the developing stands. However, on plots where advance regeneration did not exist, oak regeneration was absent or overtopped regardless of the treatments applied, even in plots with underplanted seedlings (C. Smith, 1993). A similar study in the Missouri Ozarks that compared overstory densities of 40, 50, and 60 percent stocking and three understory treatments was conducted on good and average sites with abundant oak advance regeneration. Although after 10 years the total number of oaks had increased in all treatments, stocking values were less than one-half that required for successful oak regeneration on the good sites and the oaks were still at a competitive disadvantage (Sander and Graney, 1993). Results from a study in southern Indiana showed that all shelterwood treatments failed to increase the number of oaks, due to the lack of acorn production the first 3 years after treatment and the rapid regrowth of the understory (Sander and Graney, 1993).

Application of the shelterwood method for regenerating bottomland oaks has also been discouraging with many more failures than successes. In southern bottomland hardwoods, most stands range from closed-canopy stands with little advance reproduction to cutover, mismanaged, or unmanaged areas where species composition and tree quality are poor. In theory, the shelterwood method should nurture oak reproduction if oak seed sources are present. However, in practice, a heavy shelterwood cutting that creates gaps in the overstory usually favors the development of faster-growing intolerant species rather than oak, while in other instances shelterwood cutting actually encourages the growth of undesirable tolerant species already established in the midstory and understory. Both scenarios exclude optimum oak regeneration. Single-tree selection also is not recommended for regenerating bottomland oaks because this method favors the growth and establishment of shade-tolerant species. Group selection can be used to regenerate bottomland oaks if the size of the opening is adequate to provide sufficient light to the forest floor to encourage the establishment and development of oak reproduction. (Clatterbuck and Meadows, 1993)

Hodges (1989) reported that, in practice, heavy shelterwood cuts favor the development of fast-growing, intolerant species rather than the oaks, while lighter cuts may encourage more tolerant and less desirable species. Factors contributing to the variable success of the shelterwood method to regenerate bottomland oaks as compared to oak regeneration in the Southern Appalachians include (1) higher site productivities; (2) a preponderance of tolerant species such as sugarberry, boxelder, beech, elms, maples, hickories, hornbeam, and hophornbeam; and (3) the unique hydrologic regime of bottomlands. (Clatterbuck and Meadows, 1993)

The group selection method is an uneven-aged system that also is potentially suitable for regenerating northern red oak. The method requires creating small openings in the main canopy to provide the conditions necessary for the establishment of new reproduction and/or the growth of advance reproduction.

Interest in uneven-age regeneration methods has increased greatly in recent years due to political and sociological pressures against the adverse visual impacts clearcutting. In the southern Appalachians, group selection is viewed as a more viable alternative to clearcutting than single-tree selection methods because it provides more adequate conditions for regenerating and maintaining a variety of commercial hardwood species. Similar to even-aged systems, group selection creates favorable conditions for regeneration such as increased light, nutrients, and moisture. However, trees surrounding group patches modify environmental conditions making the changes less dramatic than with clearcutting methods.

Compared with single-tree selection, group selection also offers the following advantages (Smith 1986):

• There is less damage to residual growing stock during periodic cuts.
• Trees growing in dense groups in openings have better form.
• Wildlife habitat is more diverse within a stand.
• Patches of low-quality trees are removed and replaced by new reproduction in only one periodic cut.
• Sale preparation and logging are more economical.
• Openings made at each cut can be monitored as recognizable age classes.
• Proper application of the group selection method is unlikely to produce dysgenic effects.

Group or patch selection, with openings of between 0.13 and 0.4 ha (0.33 and 1 acre) in size can be used to secure reproduction of most of the desired species, such as white ash, green ash, black cherry, cucumber tree, red maple, and, to some extent, red oak and yellow-poplar. Research on the effect of opening size on regeneration has demonstrated that oaks can be reproduced in small openings as well as in large ones (Minckler and Woerheide 1965, Sander and Clark 1971, Smith 1981). However, as with clearcutting and shelterwood methods, oak advance reproduction of adequate size and in adequate numbers must be present where openings are created in order to achieve successful oak stand replacement. (Sander and Graney, 1993)

Experience with applying group selection in the central hardwood region has been varied; for the most part, group or patch cutting methods have not consistently established oak species but have been more successful with other commercial species (Beck 1970; Sander and Clark 1971; Sander et al. 1984; Parker and Merritt 1995). Based on their experience with selection management of southern Appalachian hardwoods, Della-Bianca and Beck (1985) recommended using larger openings than were created by single-tree removal, coupled with herbicide control of unwanted competition to obtain regeneration.Others recommend using the group shelterwood method, where openings are centered over clusters of well-developed ad vance oak regeneration at least 3-4 ft (0.9-1.2 m) tall, and having well-established root systems (Watt et al. 1979; Smith 1981; Parker and Merritt 1995; Sander and Graney 1993). (Nyland 1996, Hicks 1998)

A long-term study of patch cutting on the Fernow Experimental Forest on near Parsons, West Virginia ( site index 70 for northern red oak) showed that group selection does regenerate many commercial species, but is less successful at managing oak. In this study, circular openings 150 feet in diameter (0.4 acre) were cleared every 10 years to harvest merchantable trees and establish new regeneration over a specified area of each stand. All openings surveyed after 9 and 19 years had adequate reproduction for future development into sawtimber. Younger openings surveyed 9 years after cutting had 800 to 1,200 codominant stems per acre, with 60 to 70 percent classified as potential crop trees of good quality. Older openings surveyed 19 years after cutting had 300 to 450 codominant stems per acre, with 70 to 85 percent classified as potential crop trees of good quality. Species composition was dominated by black birch, sugar maple, black cherry, yellow-poplar, and American basswood; white ash, and northern red oak were present in lesser numbers. (Miller et al. 1995)

Anderson-Tully Company of Memphis, TN, successfully uses variations of group selection and patch cutting in its management of bottomland oak forests. Anderson-Tully practices classical uneven-aged silviculture (Smith 1986) by maintaining at least three age classes in its stands and by using volume regulation of the cut. The companys biggest challenge has been the regeneration of oaks under the uneven-aged system (Stephano 1992). Larger openings are used to provide optimum conditions for oak regeneration. This company also practices intensive control of midstory vegetation to further enhance establishment and development of oak reproduction. (Clatterbuck and Meadows, 1993)

When applying the technique to northern hardwood stands, Leak and Filip (1977) found that most of the regeneration was from tolerant species, in spite of the larger openings created using the technique. Weigel and Parker (1995) found that group selection did regenerate intolerant hardwoods (yellow-poplar, black cherry, ash) when applied in southern Indiana. In their study, aspect and time-since-cutting had the greatest effect on stem density of the various species of regeneration.

Beck speculates that group selection might work for cove hardwoods. Most cove species, even the intolerants, seem capable of regenerating and developing in openings of about 1/5-acre or larger. In fact, openings of 0.1- to 0.2-acre in the southern Appalachians were dominated by yellow-poplar, sweet birch, and red maple after 25 years, the same species expected in larger clearcut openings (unpublished data, USDA Forest Service). However, Beck admits that there is no indication that oaks and other difficult-to-regenerate species would fare any better on cove sites under this method than in any other site without special provisions to develop sizeable advance reproduction. (Beck 1989)

Although group selection may work well under some circumstances, it also has several limitations. Limitations of the group selection include the following biological and operational factors as well as those discussed for the single tree selection method.

Biological limitations

1. Initiating group selection management is complicated by the fact that trees in uneven-age stands do not occur naturally in well-distributed groups that occupy comparable areas of the stand.
2. The relatively small openings created during group selection cuts is less effective at establishing and developing oak reproduction than even-aged systems. In oak-pine types, the pine component is usually reduced.
3. The edge-to-center ratio of openings is very high, especially for smaller openings typical of group selection; thus, the adjacent trees have a disproportionate effect on the group of regeneration. For example, 10 one-third-acre circular openings (3.3 acres) create 3.2 times more forest edge than a single 3.3-acre circular clearcut. Oaks are prone to develop epicormic branches when exposed to the high light intensities occurring along forest edges which reduces their bole quality and value (Trimble and Seegrist 1973).
4. Because of the shape and size of "natural" groups, the surrounding trees may quickly close the opening as crowns spread.
5. Group selection may not be successful in areas with high deer populations because of excessive browsing in the small openings. (Sander and Graney, 1993)

Operational limitations

1. The group selection method requires a "hands-on" intensive silvicultural approach. Regulation of stand structure and development is complicated, time-consuming, and expensive.
2. The size and arrangement of groups within a forest make them difficult to map, locate, and mark. Adequate methodology for inventorying the spatial dispersion and sizes of family groups across a stand is lacking.
3. Difficulty in seeing from one mature group to the next makes it difficult to establish proper ground control. In fact, past trials suggest that after about two cutting cycles the arrangement of family groups often becomes confusing, and foresters cannot visualize the dispersal and juxtaposition of them during marking (Roach 1974b; Leak and Filip 1977; Smith 1995; Murphy et al. 1993).
4. It is difficult to fit new openings among older ones with increasing time.
5. The scattering of groups through the forest makes it difficult to construct and maintain an access system through the forest. In fact, in order to effectively remove the dispersed clusters of adjacent mature trees, foresters must open a skidding corridor to each separate group as it matures. Otherwise, a contractor will have difficulty in moving the heavy skidding machines to and from the groups of fallen trees, or in preventing damage to many immature residual trees that grow between an existing skid trail and the group center. (Nyland 1996)

In summary, group selection, like single-tree selection, can provide many benefits, such as aesthetics and an even flow of products, while making it possible to regenerate shade intolerant species. Unfortunately, for non-industrial private owners in the central hardwood region, group selection shares many of the same limitations of single-tree selection, such as being difficult to apply and requiring long-term commitment. Also, because group selection focuses more on moving the stand toward its idealized uneven-age state than on current markets and economic considerations, small landowners may find it difficult to adopt in lieu of a lucrative timber sale when high market conditions exist. (Hicks 1998)

With group selection, merchantable volume is removed periodically every 10 to 20 years depending on management goals and stand conditions. This periodic cut includes volume from groups of trees removed to create openings plus that from individual trees removed to improve residual stand quality (improvement cut).

Miller and others (1995) describe a practical method for initiating group selection in second-growth central Appalachian hardwoods. The methods they present have been developed for use in central Appalachian hardwoods, though the techniques can be modified for use in other forest types based on local growth estimates and stocking requirements. Their guidelines cover inventory requirements, regulation of the cut, placement and size of group openings, and silvicultural treatments of residual trees between openings.

To prepare for cuttings in the group selection method, an inventory is needed to develop a stand table by 2-inch?diameter class. This information is used to regulate the volume and basal area removed and to develop guidelines for marking individual group openings. It is important to estimate the improvement cut during the initial inventory. The improvement cut removes high-risk and undesirable trees to increase residual stand quality. The improvement cut includes individual trees that are removed between group openings throughout the stand. By reviewing the initial stand table and the estimated improvement cut, the marking crew can become familiar with the unique characteristics of an individual stand. During marking, clumps of mature trees within the stand can be designated as potential locations for group openings. If oak regeneration is a management objective, then it is important to evaluate oak advance reproduction at this stage. Evaluate the potential of this advance reproduction to fill each opening that will be created by cutting. (Miller et al. 1995)

The second step in the group selection method is determining the total volume to be removed at each entry. For sustained yield, the cut volume should be equal to periodic stand growth. Periodic stand growth depends on stand acreage, average annual stand volume growth, and the length of the cutting cycle. Therefore:

Most forest managers have reliable local estimates of average stand volume growth. For second-growth central Appalachian hardwood sawtimber stands, the following estimates of volume growth by northern red oak site-index class can be used (Trimble 1968): SI 80: 400 BF/acre/yr; SI 70: 300 BF/acre/yr; SI 60: 200 BF/acre/yr.(Miller et al. 1995)

In managed stands, the minimum cutting cycle is 10 to 15 years depending on productivity. Short cutting cycles result in a more variable stand, with age classes closer together and lower cut volumes. Longer cutting cycles entail greater cut volumes per acre, but result in a less variable stand. Length of cutting cycle also depends on volume needed for an operable cut in local markets. In most central Appalachian hardwood stands where access roads are in place, a minimum total volume of 20 Mbf (international 1/4-inch rule) are considered operable (at least 2.0 to 2.5 Mbf/acre). (Miller et al. 1995)

The maximum volume cut includes both volume removed in group openings plus volume removed between openings to improve the residual stand. Before actual marking begins, the estimated improvement cut is subtracted from the maximum cut to determine how much additional volume can be removed in openings without overcutting the stand. (Miller et al. 1995)

In addition to computing the desired cut volume, adequate residual stocking must be maintained to assure sustained yield of desirable merchantable products in the future. For applying group selection in Appalachian hardwoods, residual stand basal-area goals for sawtimber-size trees (11.0 inches d.b,h. and larger) by northern red oak site class are: SI80: 70 to 86 ft²/ac; SI70: 55 to 70 ft²/ac; SI 80: 40 to 55 ft²/ac. These basal area guides assume that the residual stand will contain 15 to 20 ft²/ac in poletimber trees 5.0 to 10.9 inches d.b.h. and that logging damage may reduce residual basal area by as much as 10 percent (Trimble and others 1974). Similar to the guidelines described for volume, basal area in the estimated improvement cut is subtracted from the maximum basal area cut to determine how much additional basal area can be removed to create group selection openings without overcutting the stand. (Miller et al. 1995)

Once the volume of cut has been determined, the placement of group cuttings needs to be selected. A good rule of thumb for locating group openings is to cut parts of the stand where potential growth or returns are low compared to other parts of the stand. Although large, mature trees continue to grow, they have lower rates of return than smaller trees (Miller and Smith 1993). In second-growth Appalachian hardwood stands, clumps of trees more than 18 to 20 inches d.b.h. are candidates for removal due to declining rates of return. Also, parts of the stand occupied by undesirable species or poor-quality trees have lower revenue potential.In a variation of group selection, the <group shelterwood method> places openings around where advance oak regeneration has accumulated. (Miller et al. 1995)

The initial inventory provide information of where to locate openings. Some d.b.h. classes may have surplus trees, more than needed according to residual number-of-tree goals for uneven-aged stands (Smith and Lamson 1982). Removing trees in surplus d.b.h. classes, particularly those growing in clumps where it is convenient to make an opening, improves the residual stand structure for sustained yield. In future cuts, it is helpful to locate new openings against older openings to eliminate narrow bands of branchy, low-quality trees. (Miller et al. 1995)

The size of a group selection opening varies with management and silvicultural objectives and the biological requirements of desired reproduction. The size of openings can be designed to achieve the ecological conditions suitable for the regeneration of intolerant species. Or, aesthetic goals for a particular stand can limit opening size or require openings of different sizes and shapes. (Miller et al. 1995)

Forest managers must define clearly the stand goals for species composition and aesthetics, and then decide how many openings must exceed the minimum size to achieve those goals. If the improvement cut between openings requires a relatively high-volume removal, it is better to reduce the number of openings in the initial cut. Once residual stand quality is improved, future cuts can focus on regenerating the desired number of intolerants by creating group openings. (Miller et al. 1995)

Opening sizes for the group selection method in Appalachian hardwoods range from one-tenth acre to several acres (Law and Lorimer 1989, Marquis 1989, Roach 1963). Openings must be 0.4 acre (150 feet in diameter) to one acre to regenerate intolerant species (Sander 1971, Carvell and Tryon 1961, Schlesinger 1976, Smith 1980). A more general rule of thumb is to make openings with diameters at least twice the height of mature, co-dominant trees. However, some researchers consider this the upper limit on group selection openings; larger openings should be referred to as patch cuts or clearcuts (Smith 1986, Marquis 1989).

Murphy et al. (1993) suggest using the silvical characteristics of the species to be regenerated in deciding opening size rather than an arbitrary size based on height of the residual stand. Using a quantitative approach, the size needed to achieve a given amount of sunlight in terms of the height of surrounding trees can be computed (Smith 1981, Fischer 1979, Fischer 1980, Minckler 1961). Shape of clearing can also influence the amount of sunlight an opening receives. Circular clearings receive more sunlight than linear clearing. Openings as wide as one time the height of adjacent trees on south slopes will have light levels similar to those twice as wide on north aspects. (Mills et al.1987,Nyland 1996)

Patch selection methods

For implementing the patch selection method in the southern Appalachians, each patch or opening should cover 0.5-1.0 ac (0.2-0.4 ha), with a minimum width of one to two times the height of surrounding residual trees (Hawley 1921; Troup 1928; Watt et al. 1979; H.C. Smith 1981; D.M. Smith 1986; Law and Lorimer 1989; Matthews 1991; Murphy et al. 1993; Parker and Merritt 1995; Sander and Graney 1993). Within smaller canopy gaps, residual trees along the borders shade too much of the opening, and light levels remain too low for good growth of oaks (Sander and Clark1971; Sander et al. 1983; Smith 1981; Murphy et al. 1993; Parker and Merritt 1995). The center of circular openings of at least 0.5 ac (0.2 ha) will receive the equivalent of about one-third full insolation.Experiments with circular patches maintained a component of shade-intolerant species (Minckler and Woerheide 1965; Smith 1980, 1981), except where a dense understory of shade-tolerant trees and shrubs interfered with their development (Sander et al. 1983; Loftis 1990; Sander and Graney 1993). Subsequent cuttings should enlarge the circles along their east-west axis. Pre- or post-cutting site preparation must also reduce understory and mid-story competition, or regeneration of oaks and some less shade-intolerant species will fail (Hannah 1987; Loftis 1990; Murphy et al. 1993).

Field crews should follow several basic steps in marking group selection openings. Based on the initial inventory and experience from working in the stand, the crew first locates a potential opening using stand characteristics and regeneration goals to determine the size and shape of the opening. Once a satisfactory opening is defined, individual trees within the opening can be marked for harvest. The marking crew should keep a running tally of cut volume and basal area. Local volume tables can be used to estimate volume per tree.(Miller et al. 1995)

The improvement cut has priority over creating openings, particularly for initial cuts in unmanaged stands. Part of the improvement cut can be marked during the initial inventory. The remainder can be marked as openings are located and marked. Once total volume or basal area marked for harvest (improvement cut plus openings) reaches the limits computed for the stand, the mark is complete and no additional openings are needed. Adhering to guidelines for residual stocking will help to ensure that the stand will continue to produce regular timber yields. Removing all surplus trees can result in overcutting, particularly when the improvement cut is heavy. (Miller et al. 1995)

During the cutting or removal process, it is important to clean the openings, cutting all trees 1.0 inch d.b.h. and larger. However, for desirable advanced regeneration, guidelines could be revised to leave these trees or develop them for the future. If the oak advance reproduction is adequate, harvest the merchantable trees in the group and cut or kill remaining culls and small trees as described under clearcutting. The reproduction response will be similar to the responses after clearcutting except that reproduction growth will be retarded in a large part of the opening area because of the influence of the surrounding stand (Sander and Graney, 1993).

If the oak advance reproduction is not adequate to fill the opening, cutting the trees to create the opening will not result in oak reproduction and the opening will be filled by whatever species are present in the understory. In this case, follow the procedure for the shelterwood method where the openings will be located (Sander and Graney, 1993).

Plan to develop advanced regeneration in parts of the stand where group openings can be made in future cutting cycles. Removing sub-canopy overstory trees and the understory competition may be all that is needed to increase the amount of light reaching the forest floor and enhance oak reproduction establishment and growth.

Group selection openings are expanded during subsequent harvests to maintain the benefits of the openings. The size of the openings and the rate of crown expansion of the surrounding tree will determine the interval between harvests if subsequent maintenance or expansion of the openings is required or desired.

During periodic group selection cuts, a variety of cultural practices such as vine control and crop-tree release, can be scheduled for other parts of the stand.Grapevines can pose a problem on better growing sites by creating arbors that choke out desirable trees. Where vines already pose a problem, cut existing vines and keep the canopy closed; shading prevents sprout development and the vines will die in a few years (Smith 1984). In established openings containing saplings with vines in their crowns, cut vines near ground when the canopy has closed. Vines usually can be cut where saplings are 10 years old. Pre-commercial crop-tree release is an option for saplings growing in openings created previously. Crop trees should be co-dominant, well-formed stems with potential to become high-value sawtimber. Release a maximum of 50 to 75 crop trees per acre if they are available (Perkey and others 1994). (Miller et al. 1995)