When restoration work began at what is now Pleasant Valley Conservancy, the area was heavily infested with undesirable woody vegetation, both brush and trees. The original savanna oaks and hickories were still there, with their open-grown character visible, but surrounded and crowded with "bad" woody plants. The invaders were all fire-sensitive and had encroached during the 50 or so years since the last burns.

Early attempts to knock back the invasive woody plants by fire were unsuccessful, as there was insufficient fuel on the ground. In retrospect, fire should not have been used at that stage anyway, since all it would do was top-kill the woody plants, but not eradicate them. They would resprout from dormant buds and a bigger mess would be created. We were thus "lucky" that fire did not work.

The correct approach is to cut the invasives and treat the cut stems with herbicide to prevent resprouting and to kill the roots. Only when the bad actors are gone should fire be introduced.

Removing Woody Invasives

Clearing of both the oak savanna and the prairie remnants involved a lot of chain saw and brush cutter work. The early work was done by part-time students from the University of Wisconsin-Madison. When the work became more intense, a private contractor, Michler and Brown LLC, was hired. Volunteers occasionally helped with burning brush piles.

The decision was made to remove all trees that were not part of the original fire-dependent ecosystem. The only exception was that some birch trees were kept because of their attractiveness. Woody shrubs were removed at the same time the trees were cut. Tree species removed included elm, black walnut, cherry, and box elder. Woody shrubs included buckthorn, honeysuckle, and prickly ash. A number of black oaks were removed if they were crowding white or bur oaks. (Black oaks are fast-growing trees that tend to grow into the canopies of the whites or burs.) Some shagback hickories were also removed, although a hickory standing by itself was left.

An attempt was made to salvage the good wood, either for use as fire wood or as lumber, but none of the wood was sold. Heavy equipment was not used to remove logs. On the south slope, the wood was cut in short segments with a chain saw and rolled down to Pleasant Valley Road. Local residents quickly removed all this firewood. On the ridge top, the wood was removed by pickup truck, using the service road that runs along the length of the ridge. Although this involved some off-road driving, this was only done in the winter (at snow-free times), and care was taken to avoid making ruts. Frequent off-road driving over the same track was avoided. Growth of savanna plants the following season quickly hid whatever ruts were present. Firewood-size wood was brought down to Pleasant Valley Road where it was removed by local residents. Saw logs were taken by cooperators with front-end loaders.

The brush that was cut was stacked into large piles, together with the smaller limbs of the trees. The winter of 1999-2000 over 200 large brush piles were created on the south-facing slope. When snow came, these piles were burned (see photos below).

As much as possible, dead trees were left standing as habitat for woodpeckers and other wildlife. Such dead trees sometimes presented problems during controlled burns, but this risk was accepted in the interest of creating desirable wildlife habitat. (Two red-headed woodpecker nests identified in 2004 by Rich King of the U.S. Fish and Wildlife Service were both in dead birches.)

We had a number of aspen clones which had to be removed. Aspen cannot be simply cut like other tree species, but must be girdled. Details are given below.

Clearing Savanna Areas

The lower part of the south-slope, which was predominantly prairie, was cleared first, and the upper slopes and ridge top, which were savanna, were cleared later. This was primarily a budgetary decision. It was less expensive to clear the south slope because there were fewer undersirable trees and removal of the cut material from the hill was relatively easy. Although cutting trees on the ridge-top savanna was easier for the workers, because it was on more level ground, removal of cut material was more difficult and expensive.

The slower pace of savanna clearing was also necessitated by the need to have seeds to plant in the cleared areas. Since we were trying to use only local seeds, we had to phase in the clearing with the seed collecting possibilities. In general, around 10-15 acres of savanna were cleared in a single winter. The schedule of clearing is shown on the table below.

Although most of the cost of the prairie and savanna clearing was paid for by Savanna Oak Foundation, Inc., two helpful grants were obtained from federal agencies. The Wildlife Habitat Incentive Program of the U.S. Department of Agriculture provided a grant for the work in 1999 and 2000. The Private Lands program of the U.S. Fish and Wildlife Service provided partial support for the work carried out in January-March 2003. During the main period of tree removal, from 2003-2007, all restoration work has been funded by the Savanna Oak Foundation, Inc.

Schedule of Clearing Activities

Removal of Aspen by Girdling There are two species of aspen that are native to our area: trembling aspen (Populus tremuloides) and big-tooth aspen (Populus grandidenta). Pleasant Valley Conservancy had both species, although trembling aspen was the more common. We worked hard to get rid of these trees, and were rewarded for our work by the discovery that hidden among the aspens was real prairie sod. Aspen are native trees but remain undesirable inhabitants of prairies and oak savannas. Aspen are capable of spreading rapidly and crowding out other vegetation. Aspen are pioneer trees on open , burned, or cut-over land. Although important economically in the paper industry, they are a menace in prairies and oak savannas and eradicating them became an early goal of our restoration. Although aspens grow from seeds, the primary spread is asexual by underground runners. The typical aspen "grove" is a multi-stemmed clone in which all the roots are interconnected. If an injury to a root occurs, there will be a rapid response by the clone, and new shoots ("suckers") will be sent up all over the area. New shoots have been known to arise as far as 50 feet from the nearest aspen tree! The clone may expand simultaneously in several directions, as influenced by environmental conditions. In western United States, huge aspen clones have been found, the largest occupying over 100 acres. In our part of the country, aspen clones are smaller, but are often more than an acre in extent. Aspen clones are widespread in the Driftless area of southwestern Wisconsin. They are easy to spot in spring when the light green leaves first appear, and in the fall, when the leaves take on their distinctive fall color. Virtually every woodlot in our area has one or more aspen clones. This was not always the case. In earlier times, aspens were uncommon in our area. This was because aspens are fire-sensitive, and when prairie and savanna fires were widespread, they were held in check. The first air photo of Pleasant Valley Conservancy, taken in 1937, shows that the land was quite open. Fire protection and elimination of grazing resulted in a major change. Subsequent air photos showed our land gradually becoming wooded. Many of the trees were aspen. The air photos from the 1990s show virtually no open land at all, and our field studies showed that most of the trees were aspen. We have identified many trembling aspen clones, in different parts of the Conservancy. There was a small clone above unit 1 (the goat prairie), two or three very large clones in the White Oak Savanna (unit 12A), at least three clones that surrounded the upper agricultural field (see photo at top of page) that was later planted to prairie in 1998 (now called Toby's Prairie). There was also a large clone at the lower end of the White Oak Savanna, bordering the ag field that is now the Pocket Prairie. Also, there were two or three clones surrounding the ag field that is now the Ridge Prairie. Small clones were also present in the lower part of Unit 11. Finally, there was a clone of big-tooth aspen in Unit 19. This clone was also invaded with white sweet clover, a persistent weed that is devilishly difficult to eradicate. Girdling When looking at a large aspen clone, it may be tempting to go in with a chain saw and cut it down. Wrong! The roots remain alive and immediately send up a huge number of new shoots. An area that had perhaps 5 or 10 large aspen trees will soon have hundreds of aspen stems. This photo shows what happens when an aspen clone is simply cut . There are dozens (hundreds?) of resprouts occupying the space. (From a site in southeastern Wisconsin.) There is only one certain way of killing aspen and this is by girdling. Girdling means stripping a layer of bark and the underlying cambium and phloem in a band around the trunk. The phloem vessels translocate sugars and other nutrients to the roots, so if the phloem tubes are broken, the roots become starved of food. The xylem vessels, which translocate water to the leaves, are not affected by girdling. With girdling, the upper part of the tree still remains alive, since photosynthesis can continue. Eventually, however, the roots die, and the whole tree dies. The first year after girdling, the clone may appear almost normal, but by the second year the clone usually dies. The dead trunks can then be cut without stimulating resprouting. For girdling to be affective the whole clone must be treated. It is also important to make the girdle in such a way that the underlying xylem is not damaged. Damage to the xylem sends signals to the tree that something bad has happened, and the tree then sends up shoots. The photos below demonstrate the technique of girdling. Girdling is done in May or early June when the sap is running fast and the tree is growing. At that time, the bark can be easily cut and the girdled bark stripped off. Later in the summer, it is virtually impossible to do a girdle properly. We carried out a small girdling in 1995 and initiated major girdling in 1997. Our major girdling was done in 1998 and 1999, but some girdling was done in 2000, 2001, 2002, and 2003. The trees girdled in 1998 (see photos here) died in 2000 and were cut, stacked, and burned. Almost all the trees that were girdled are now dead and have been removed. The dead trees of the major aspen clone in unit 12B were finally cut and removed in the spring of 2002. Although we still have many aspens on the east end of Pleasant Valley Conservancy, aspens in the highest quality land are now gone. We did our final aspen girdling in May 2008 in the area we call the East Basin (Unit 22). This isolated basin was not cleared until the winter of 2007-2008. There were close to 100 aspen in this clone, many of them quite large.   An unexpected result An unexpected result of removing the aspens was the discovery that hidden within one of the clones was real prairie sod. Conservative species such as Panicum leibergii and Viola pedatifida had managed to survive within the clone. Once the aspen were gone, these species (and others) were able to thrive. Today we can identify the exact boundaries of this aspen clone from the "good" species that are now present. Although girdling is very effective, small aspen resprouts may still come up at the edge of girdled areas. These can be handled by cutting with a hand clippers and treating the cut stems with glyphosate, and we continue to do this every year or so.

Cost of Restoration

Restoration work of the kind described here is expensive. During the largest efforts described in the table above, the costs ran between $3000 and $5000 per acre. These costs were for an experienced crew who worked very efficiently, and includes cutting of trees and brush, stacking and burning, and all herbicide treatment of cut stems. It does not include the cost of removal of firewood except where the wood could be rolled down the hill to the town road.

The specified payment rates (dollars per acre) that the Federal Government uses in their granting programs covers only a small part of the operation. For instance, the per acre payment rate for removal of heavy brush is given at $190.00 per acre, which is only a small part of the cost. (Sites with heavy brush have more than 40% of the area stocked with woody vegetation 0-2 inces in diameter and up to 25 stems per acre larger than 2 inches in diameter at the ground line.)

Brush Control

Most people involved in restoration of native landscapes learn quickly about the evils of invasive woody plants. The principle culprits in our area are bush honeysuckle, common buckthorn, and prickly ash. The first two are exotics, introduced in this country from other parts of the world. Although prickly ash is a native, it is still a problem because it is so unpleasant to have in the woods. All three are fire-sensitive, and they are classic invaders of areas that have been protected from fire.

Another group of invasive plants that are a problem primarily in savannas are the various Rubus sp. (brambles; black and red raspberry, blackberry, dewberry). Although native, they are so invasive that their control is essential. Without control, large areas of restored savannas will become virtually bramble monocultures.

Honeysuckle has been especially bad at Pleasant Valley Conservancy in areas that had been disturbed, either by previous logging or by tree fall. Buckthorn was especially bad in the bur oak savannas. Prickly ash was specially bad in savanna areas that had once been grazed. In shady woods, brambles are only a minor component, but once increased sunlight becomes available due to savanna restoration, they flourish. All of these species are being removed as part of our savanna restoration work.

When we first started restoration, we focused on these three shrubs in the understory of the savanna. Because we had not yet begun serious cutting of larger trees, we hired an herbicide applicator to carry out basal bark treatment using trichlopyr (Garlon 4) mixed in diesel fuel. Garlon 4 is oil soluble and passes readily through the bark. It is translocated to the roots, where it acts.

The basal bark procedure used was to mark 100 foot wide swaths from the bottom of the hill at Pleasant Valley Road to the top of the ridge. The applicator then moved up the hill, back and forth, spraying the base of each stem with the Garlon mixture. The spray was spread in a zone about six inches high around the base of the stem. A blue dye was used to control the spraying pattern. Everything in the path that was bad: honeysuckle, buckthorn, prickly ash, was treated. Due to the volatility of the treatment, the work was done in the winter. This approach worked quite well, and surveys the following year showed that over 90% of the target shrubs were killed. This procedure has also been used on invasive shrubs in the oak woods.

Honeysuckles

There are several species of bush honeysuckles that cause problems in our area, including Lonicera morrowii, L. maackii, L. tatarica, and the hybrid Lonicera X bella. However, there is no reason to attempt to distinguish them since they are all nonnative and they are all bad. The native honeysuckle in our area (Lonicera reticulata and L. dioica) can easily be distinguished from the bad ones because the natives are all woody vines rather than bushes.

Bush honeysuckles are upright shrubs ranging from a few feet to 15 feet tall. They form many branches from the base, and the spreading branches shade other plants. In a honeysuckle "thicket", almost nothing will be found under the canopy. (After the honeysuckle is removed, the soil is often bare.) Honeysuckles form fragrant tubular flowers, followed later by red fruits. Birds are attracted to the fruits and spread the seeds. Bush honeysuckles have a wide environmental tolerance, but they prefer partial to full sunlight and are most commonly found in abandoned fields, forest edges, roadsides, and other open upland habitats. They are extremely invasive and can easily take over and dominate a habitat.

Bush honeysuckle is one of the plants that will invade a habitat if it is protected from fire. Once honeysuckles have conquered a habitat, there is no possibility of fire because there is no fuel. In order to reintroduce fire, it is essential first to eliminate the honeysuckles and then reseed with native plants, preferably seed mixtures containing grasses that will carry a fire.

Both mechanical and chemical methods are used on honeysuckle, and often both together. The most assured method is to cut all the stems of a plant and treat each cut stump with a 20% solution of glyphosate. The concentration given here is percent of the active ingredient.

Honeysuckle can be cut with either a brush cutter or a hand lopper. For an occasional plant, a hand lopper is fine, but for any extensive honeysuckle thicket, a brush cutter is essential. With a brush cutter, a sharp saw blade is preferable. For very large bushes, a chain saw must be used.

With a group of volunteers, an ideal way to work is with one person operating the brush cutter and several persons following with herbicide. (A skilled brushcutter can cut enough to keep three or four volunteers busy treating!)

However, a single person can also make significant inroads into a honeysuckle thicket. Here is a procedure guaranteed to work: Cut each stem with a hand lopper or handsaw, counting the stems as you cut. Cut the stems as close to the ground as possible, but still leave a small amount of stem showing above the soil layer. Pull all cut stems away from the base. Now treat each cut stump carefully with the glyphosate mixture. It is strongly recommended that a red or blue dye be added to the herbicide mixture, so that treated stumps can be distinguished from untreated ones. (Dyes suitable for herbicide use can be obtained from an agricultural chemical supply house.) As you treat, count each stump again, and do not stop treating until you have treated every stump you have cut. If a spray bottle is used, do not spray the whole base, since this wastes herbicide and spreads it around. Instead, place the tip of the spray bottle onto each cut stump, press gently to bring up several drops of solution, and spread them around the cut stump with the tip of the bottle. The whole cut stump should be colored with the dye/herbicide mixture. With practice, this procedure works quite well and the honeysuckle plants should not resprout.

Honeysuckle can be cut and treated at any time of the year, although the winter is often preferable because of the lack of foliage. Glyphosate does work in winter! (Here is a research paper on this topic. Ecological Restoration Vol 22, June 2004, pp. 145-146.)

Small honeysuckle plants can also be killed by spraying the leaves with a 1-2% solution of glyphosate (active ingredient). It is important that all leaves be sprayed. Honeysuckle is so sensitive to glyphosate that the plant should be killed within a few weeks. More importantly, the root system is also killed, and within a year the dead shrub can be readily pulled up.

Honeysuckle is very persistent, and will resprout readily if not treated with herbicide. Please note: There is no point in cutting honeysuckles if the cut stumps are not going to be treated with herbicide.

We do not recommend hand pulling, as some authorities do, because it disturbs the soil and opens it up for establishment of weeds.

After the honeysuckles have been taken care of, the area should be reseeded with native species. This is especially important because when the honeysuckles were removed a "hole" has been created, into which weeds will readily move. In fact, if the honeysuckles were almost solid, it might be preferable not to remove them all at once, but to gradually cut them back, seeding with native species as you go. It may take several years to eliminate the honeysuckles in this way, but this may be preferable to creating a habitat full of weeds!

Once large honeysuckles have been eliminated from a natural habitat, the work is not finished. There will be a seed bank, so that small honeysuckles will appear next year. Fire will top-kill these new plants but will not eliminate them.

It is essential to return to the area periodically to remove new growth.

Buckthorn

Common buckthorn (Rhamnus cathartica) has been a serious invader of the oak savanna areas at Pleasant Valley Conservancy. It is a small tree or shrub and can reach a height of 20-25 feet. Buckthorn has a characteristic orange inner bark which provides a good way of identifying it. The outer bark is dark gray or brown, and when cut the inner bark is brown, red, or orange.

Buckthorn has separate male and female plants; the latter are often easy to recognize because they produce copious amounts of deep purple berries. It is especially important to remove the berry-producing plants, because they will otherwise be a constant source of infection for the area.

Buckthorn is of Eurasian origin and was introduced into North America as an ornamental. It was planted for hedge rows in Wisconsin as early as 1849. Despite its insidious nature, it is still legally sold as an ornamental. It has become naturalized and has spread over most of the southern and eastern parts of the Wisconsin. Buckthorn is an especially troublesome invader of natural oak savanna and oak woodland areas of southwestern Wisconsin.

Many bird species relish buckthorn berries. However, the berries contain a chemical which acts as a laxative (hence the species name cathartica). The defecation by the birds insures the spread of the seeds throughout the habitat. Since female trees may produce abundant fruit, within a few years there can be thousands of buckthorn seedlings in the area of a mature tree. Buckthorn seeds are able to remain alive in the soil, and new seedlings will continue to appear for years after the plants have been removed from an area.

Early identification, before seed production has started, is vital. Small buckthorn seedlings can be readily removed by hand pulling. However, this mechanical removal disturbs the soil and encourages reinfestation or colonization of other weeds so that loose soil should be tamped down to make a firm surface.

Herbicides for buckthorn control There are several herbicides that are very effective in control of buckthorn. One of the most effective is triclopyr (Garlon; Dow Agrochemical). An effective way to control buckthorn is by the use of basal bark treatment with Garlon in oil. This is best done in winter, when native vegetation has all died back and will not be affected. A concentration of 12-15% triclopyr (active ingredient) in diesel fuel or kerosene is recommended by the manufacturer. Use the herbicide in a backpack sprayer (see photo above) with a nozzle that produces a solid cone or flat fan spray. Spray the lower part of the trunk in such a manner that it becomes thoroughly wet, including the root collar, but not to the point of runoff. Each stem of the plant must be treated. Properly done, this basal bark treatment is extremely effective and the plant will not leaf out the following growing season. Once dead, the plant can be cut and removed, or allowed to stand to rot.

We have used this procedure extensively at Pleasant Valley. It has the advantage that it is quick and does not require any cutting. However, if we are removing trees at the same time, the cut stump procedure described below is preferable.

Cut stump treatment Another very effective way of eradicating buckthorn is to cut the plant just above the ground level (hand lopper, brush cutter, or chain saw) and treat the cut stump with 20% glyphosate. Glyphosate is the preferred herbicide because it is inactivated by soil particles so that there is no residue in the soil. It should be emphasized that cutting buckthorn without herbicide treatment should never be done, because the cut plants will resprout heavily from the roots, leading to a situation that might be worse than if the plant had not been cut at all. It is useful to include a blue or red dye in the herbicide mixture so that the cut stump treatment can be monitored. A backpack sprayer or hand spray bottle can be used. Be sure that the stump is thoroughly wetted with herbicide. This procedure is economical of herbicide and confines the chemical to the stump itself, but is more labor-intensive than basal bark treatment. However, it has the advantage that the buckthorn plants themselves are being removed from the habitat.

Although the cut stump procedure can be used at any time of the year, the winter is preferable because nontarget plants are not affected. This procedure is effective with plants of any size, even large ones. Remove all the cut material to a pile for subsequent burning.

Brambles

The term "bramble" refers to a whole group of plants of the genus Rubus, which include blackberry, red and black raspberry, and dewberry. Brambles are a minor component of the degraded savanna, but once the habitat is opened up and light reaches the forest floor, brambles can grow rampantly. Although our brambles are native, we still consider them undesirable because they tend to take over the savanna. (Brambles are less of a problem in prairie restoration.)

Brambles are biennial plants but have a perennial root system. The roots continue to grow for the life of the plant, but new above-ground shoots (generally called "canes") develop each year. The first-year shoots grow vegetatively but do not flower. In the second year these shoots flower, set seed (berries), and then senesce and die.

The photo below shows the arrangement of canes, shoots, and roots at the base of the plant. There are several dormant shoots. Iif the above-ground shoots are killed by fire or cutting, one or more of the dormant buds will begin to grow and form a new canes.

Some brambles (black raspberry, northern dewberry) exhibit a phenomenon called "tip-rooting." Canes whose tips reach the soil can form new roots, enabling the brambles to colonize new bare areas. Bramble patches with tip-rooted canes are especially difficult to walk through.

Since flowers only develop on the second year shoots, annual fire will keep brambles from flowering and setting seed, but will not eradicate them. In the next growing season, each killed cane will develop a rosette of leaves from an underground bud. By mid-summer, a new vigorously growing cane has been established.

Control of brambles in savanna restoration Although they are very fire-sensitive, they are not eradicated by burns, since the roots remain alive and resprout. Eradication of brambles in savanna restoration can only be done with the use of herbicide. Even then, removal requires careful monitoring and consistent control.

Bramble canes can be cut and the cut stems treated with glyphosate (20% active ingredient), as described above for honeysuckle. This procedure works well for blackberries, and reasonably well for black raspberries, but is generally ineffective for red raspberries. This latter species is a prolific clone former, and has an extensive underground root system. Herbicide treatment of cut stems does not work.

Brambles are easily cut with hand clippers, and a single person can cut and treat an area, with clippers in one hand and herbicide bottle in the other. It is essential that every cane be cut and treated. Some of the canes may be dead. These do not need to be treated, although it is still desirable to cut them. Living cut stems are easily recognized because they will be green and moist, even in the winter. Eradication by the cut-and-treat method is very time-consuming, but is the surest way. Once the brambles have been removed, annual burns should be carried out for a number of years, since there will always be a seed bank, and in the absence of fire the area would eventually return to an unrestored state.

In areas with large bramble patches, especially those where there are no "good" plants, foliar spraying can be done. Roundup (glyphosate; 1-1.5% foliar spray) is labeled for blackberry and should work on other members of the genus Rubus. According to the manufacture, best results are obtained when plants have reached full leaf maturity in late summer or fall. Garlon (triclopyr; 0.5-1% foliar spray) and Krenite (fosamine; 5-10% foliar spray) are also labeled by their manufacturers for blackberry and would presumably work for the others. However, all of these herbicides have the potential for harming nontarget plants and should only be used in areas where there are no desirable plants.

One additional procedure that may aid in bramble control is to cut the plants in mid-summer, at flowering time. At this time of year, most of the nutrients are in the stems, and if these are severed the roots will be starved. We have found this procedure to be helpful, although it does not completely eliminate the brambles. One advantage of cutting at flowering time is that berries will not yet have been made, thus eliminating the chance of adding to the seed bank.

Brambles are very sensitive to fire and are readily top-killed by a good burn, but the dormant buds (see photo above) soon begin to grow and form a rosette of new leaves. These rosettes can be treated with herbicide, such as glyphosate or triclopyr, at the low concentrations suitable for foliar spraying.

Rubus always has an extensive seed bank, so that effective bramble control requires frequent, preferably annual, burning as well as reseeding with herbaceous savanna species, since competition from other plants is an important factor in keeping brambles from becoming reestablished. Because fire will not eliminate the seed bank, it is also essential to return to previously restored areas periodically and repeat the cut-and-treat method.