This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain. How Forest Fragmentation Hurts Species and What To Do About It Daniel Simberloff1 HOW FRAGMENTATION HURTS SPECIES effect (Saunders et al. 1991). Of course light and moisture regimes near an edge also differ from those of the interior. In Wisconsin forests, increased light can pennit shade-intolemnt vegetation to invade 30 m inwards (Ranney et al. 1981). In a well-studied fragment of old-growth longleaf pine forest, every introduced plant was within 2 m of a road or an artificially maintained clearing (S. Hermann, pers. comm. 1991). Animals can also penetrate far from an edge. For example, tropical animals disperse seeds from secondary habitats into pristine forest tree falls 5 km away (Janzen 1983). Similar effects are known in other temperate and tropical forests (e.g., Janzen 1986; Wilcove et al. 1986). Most research on effects of animals penetmting an edge is on how they eat forest interior organisms. Fragmentation of habitat is the major global environmental change occurring today and the one most likely to devastate biodiversity and ecological pr<?cesses in the near future (Simberloff 1993a). Fragmentatioq always accompanies habitat destruction and the effects of fragmentation per se, as opposed simply to the loss of area, have been intensively studied only recently. An early suggestion that fragmentation could have important consequences concerned forests in Wisconsin (Curtis 1956), and most of the maps of habitat fragmentation that pepper conservation journals and texts are of forests. 1bere is no satisfactOlY geneml theoretical framework for fragmentation analogous to the species-area relationship for habitat destruction (Simberloff 1993a). However, a number of intensive studies of particular systems suggest certain potential effects should always be considered. I will focus particularly on a system that surrounds my home, the longleaf pine (Pinus pa/ustris) ecosystem of the Southeast. Increased Predation and Herbivory Nest predation may increase greatly in a fragmented landscape. Wilcove (1985, 1990) placed artificial nests with quail eggs in eastern U.S. forests of different size, ranging from small woodlots to the continuous forest of the Great Smoky Mountains National Palk. In the latter, only 2% of nests were preyed upon within a week, while nests in rural woodlots of 4-10 ha averaged 48% predation, and similar sized subutban woodlots reached 70%. This study and similar results (e.g., Andren and Angelstam 1988, Small and Hunter 1988, Yahner and Scott 1988) inspired the "intermediate predator hypothesis" (cf. Terborgh 1988), which states that medium sized predators-raccoons, squirrels, blue jays, crows, dogs, cats, etc., in eastern forest-are greatly increased in a patchwork quilt of housing, farmland, second growth, and forest fragments. 1bese predators, ~ tum, invade the forest fragments and prey on its denizens. In the continuous forests of the past, according to this hypothesis, large predators like wolves, mountain lions, and raptors were much more numerous and greatly suppressed populations of the intermediate predators. In longleaf pine forests, nest predation of the state-listed gopher tortoise (Gopherus polyphemus) is enhanced by habitat destruction and fragmentation, as elevated populations of skunks, raccoons, crows, and introduced fire ants (So/enopsis invicta) thrive in the agricultural and second-growth matrix that surrounds longleaf fragments but attack nests in the longleaf (references in Simberloff 1993a). Nest predation of two common Edge Effects As fragments become smaller, they increasingly comprise edge habitat. This is because areas within the fragment are affected physically and biotically by the presence of the edge. The "edge effect" (Moore 1962, Williamson 1975)-the presence of species near an edge that chamcterize neither of the adjacent habitats--<>ften results in increased diversity at an edge, so wildlife biologists have traditionally viewed edges as desirable (Harris 1988, Yahner 1988). However, species that colonize edges are often common elsewhere, while forest interior species that do not tolerate edges are often of special concem Exactly how far inside a forest the existence of an edge is manifested depends on the forest, but it can be surprisingly far. Changes in wind currents, for example, can sometimes be detected at a distance 100 times the height of the vegetation; thus a forest with 20 m tall trees might need to be 2 km wide before any part of it would not suffer a meteorological edge 1 Depattment of Biological Science, Florida State University, Tallahassee, Florida 32306. 85 whether most species are maintained as metapopulations. CursOly reviews (Harrison 1991, Simberloff 1993b) cast doubt on the proposition Rather, it seems that many species are not metapopulations at all and many others are metapopulations of the sort envisioned by Boonnan and Levitt (1973) and Pulliam (1988). In this model, large central populations are continuous sources of colonists for smaller ephememl populations, which are "sinks" in the sense that they are maintained only by this recruitment and do not contribute to the persistence of the central populations. Until many more data are available on movement, any hypothesis of metapopulation-collapse induced by fragmentation is just that: an hypothesis. game species, bobwhite quail and tmkey, is apparently similarly elevated, with many of the same culprits (references in Simberloff 1993a). Not only nests are preyed upon Adult fox squirrels (Sciurus niger), gopher tortoises, and tuIkeys are all heavily preyed upon by species typical not of their favored old growth pine habitats, but of the varied, distutbed landscape that now prevails. Herbivory of forest plants can be increased in a fragmented landscape just as predation can, if numbers of herbivores and/or their access to forest habitat are increased. The national forests of northern Wisconsin were once ca. 80% old growth and contained small fragments of earlier successional stages genemted by fires and stonns. Nowadays, in the wake of intensive logging, the landscape- is the reverse: a patchwOlK quilt with 95% earlier successional stages dominated by aspen and only 5% old growth fragments of 5 - 200 ha. Alverson et al. (1988) have found that white-tailed deer populations have more than doubled in this landscape of excellent browse, and their browsing modifies even the Gld growth fragments. TIle deer select many old growth ground cover plants as well as seedlings of old growth trees like. eastern hemlock, white cedar, and Canada yew. Alverson et aI. (1988) believe it will be impossible to maintain old growth in small fragments unless deer populations are controlled. Seedlings of longleaf pine suffer a similar fate in some regions (references in Simberloff 1993a). Pocket gophers (Geomys pinetis) and especially wild hogs eat longleaf seedlings and both animals thrive in the mixed agricultural/early successional stage communities surrounding longleaf fmgments. Other Effects Other effects of fragmentation are not as geneml as edge effects, increased predation and herbivory, and (potentially) failure of metapopulation dynamics. Some, however, probably occur in many systems. Some species simply cannot maintain a population in a small fragment and, if fragments are sufficiently isolated, cannot maintain a population or metapopulation in the entire constellation of fragments. For example, whatever forces determine minimum viable population sizes (reviews by Shaffer 1981, Simberloff 1988), huge carnivores are likely to disappear from small fragments for thermodynamic reasons alone unless they are very good at getting from site to site. There simply is insufficient food and space to support a population of bears or bobcats in a 10 ha site. Introduced species are likely to be a far greater problem within forest fragments in a variegated landscape than they would be in intact large expanses. Not only are some introduced species highly adapted to the anthropogenous habitats that surround forest fragments but these habitats provide access to the forest proper (Simberloff 1994). Longleaf pine forests are almost devoid of introduced fIre ants (Solenopsis invicta) except along roads or edges (Tschinkel 1988). The same is true of introduced plants. Fmgmentation can disrupt a fIre regime and thereby change an entire community. Longleaf pine forests are fIre disclimaxes maintained by frequent fires. Previously, lightning-induced fIres spread widely and every site was thus burned every few years whether it was struck by lightning or not. Now the situation is completely changed, because the forest fragments are widely separated by farms, commercial plantations, roads, housing, etc. (Simberloff 1993a). Managers must perform regular controlled burns. Disruption of fue regimes can also be induced by introduced plants, as has occurred with the introduction of Melaleuca quinquenervia (Ewel 1986). Failure of Metapopulation Dynamics Metapopulation dynamics as a hedge against extinction are all the rage nowadays. The fIrst model (Levins 1969) has been supplemented by numerous others (Hanski and Gilpin 1991), and the overall theoretical result is clear: populations that would not persist in one large population might do so in a metapopulation of populations, given sufficient mtes of intersite movement. Metapopulation theory has superseded island biogeogmphic theOIy as a way of thinking about nature among conseIVation biologists (Merriam 1991). Many authors (e.g., Carter and Prince 1988, Wilson 1992, Noss 1993) contend that most species are distributed as metapopulations, but there are few data. If species are, in fact, maintained by continual recolonization of tempomrily empty sites, it is easy to see how fragmentation could cause a metapopulation to collapse. As fragments get smaller and more isolated, the number of individuals moving from site to site decreases and may surpass a threshold below which the entire metapopulation collapses. But this is the rub: actual rates of movement between sites is mrely known, so it has proven almost impossible to assess 86 Uses" with countless signs. Thus there is no radical shift in direction indicated in the letter, "Ecosystem Management of the National Forests and Grasslands" sent by Chief ED. Robertson on June 4, 1992. Rather, he says that now "an ecological approach will be used to achieve the multiple-use management of the National Forests and Grasslands." There seem to be two main components to the new approach: more science and an ecosystem focus. Will a more ecosystem-focussed management and closer interaction with scientists lead to successful maintenance of biodiversity while allowing continued other uses of the habitat, such as recreation and harvest of wood, at levels acceptable to all users? Only time will tell. Just as with the rush to create corridors, however, there seems to be an element of faith in the New Perspectives. That is, one would expect a scientific approach to forest management to be founded on a falsifiable hypothesis and a commitment to discard the hypothesis if it is falsified. Neither the Chief's letter nor the more formal statement of the New Perspectives (Kessler et aI. 1992) really presents the approach in this way. Neither considers the possibility that adequate maintenance of biodiversity might be incompatible with other uses at desired levels. Worse, the terms of these manifestoes are sufficiently vague and general that it is difficult to imagine a possible future result in some specific ecosystem or landscape that would defInitively falsify the hypothesis. That is, is there a particular set of observations that could cause the Service, or its Chief, or its scientists, to proclaim that the New Perspectives cannot achieve their desired goal? The explicit method proposed by the scientists is adaptive management, in which "information from monitoring is used to continually evaluate and adjust management relative to predicted responses, management objectives, and predetermined thresholds of acceptable change" (Kessler et al. 1992, p. 225). It is unclear in this approach exactly when the entire framewotk for conceiving the problem might be rejected, if ever. The Service itself clearly views the New Perspectives as something vety different from what had gone before, and they are "new" in the sense of" recent." So it is important to remain optimistic and open-minded until some results are in However, the histoty of conservation is littered with bright ideas of great intuitive appeal that turned out not to solve many or any conservation problems (Simberloff 1988), and one should take a lesson from this fact: remain skeptical and conceive of evety idea as an hypothesis. Partnerships play a key role in the New Perspectives--R.obertson's letter speaks of "partnerships with State and local governments, the private sector, conservation organizations... " . For maintenance of biodiversity, it is clear in some regions that partnerships are necessaty if only because the Service (in fact the entire federal government) does not control a large enough fraction of the land to ensure continued persistence of all species. Half of all federally listed species and subspecies are oot found on any federal lands; 64% of all occurrence records for these taxa in Natural Heritage Data Centers are oot from federal lands (Natural Heritage Data Center Netwotk 1993). Consider the South: 90% of southern timberland is privately owned (Norvell 1993). Corporate timber companies HOW TO COUNTER THE EFFECTS OF FRAGMENTATION Corridors The most highly publicized approach to mitigating problems engendered by fragmentation is to connect the fragments by corridors. 'These proposals range from rows of trees 10 m wide (Hussey et al. 1990) through mega<orridor proposals such as the Wildlands Project (Mann and Plummer 1993) or the proposed corridors 300 meters wide 'and thousands of kilometers long to provide for movement in the face of global warming (Hunter et al. 1988). What these proposals almost all have in common is a dearth of evidence that the target OIganiSms will actually use the corridors and scant consideration of the cost of the corridors relative to the cost of other possible conservation measures (Simberloff and Cox 1987, Simberloff et al. 1992). Because these problems have been tboroughly aired (e.g., Hobbs 1992), I will oot belabor them here. Suffice it to say that even defenders of the proposition that corridors will often be vety useful still qualify their defenses by admitting there are few data showing this, and that corridor proposals almost never include a discussion of possible alternative uses of funds. Landscape Management Another way to mitigate fragmentation, or one that might operate simultaneously with corridors, is to manage the entire landscape so that, as a whole, it supports a large fraction of the community. In other words, granted that small refuges are important but insufficient and that large eoough refuges may oot be attainable for ecooomic reasons, is there some way that the land outside refuges can be managed so that the refuges do oot appear, to the species of concern, as islands in an inhospitable sea? This is the premise behind the "new forestry" (e.g., Franklin 1989, Swanson and Franklin 1992): can timber be extracted from a major portion of the forest without major harm to resident species? 'The idea of" habitat variegation" (McIntyre and Barrett 1992) proposed for the oorthem tablelands of New South Wales is very similar. In both instances, the goal is to manage a landscape so that, even if it is far from pristine, and even if many resources are extracted, the threat to all species is vitiated. The Forest Service calls its version of this philosophy "New Perspectives" (Kessler et al. 1992). Is the new forestry truly new? It and related ideas seem to be versions of a multiple-use strategy at the landscape level. The Forest Service's planning regulations under the National Forest Management Act of 1976 (36 C.ER. pt. 219) require that the Service manage the land for multiple use (sec. 219.2(b)(l». Well before then, the Forest Service applied a multiple-use philosophy to forest management (Kessler et al. 1992), and the Service has for years proclaimed the national forests the "Land of Many 87 management of private lands has generally been even more inimical to the bird than management of the national forests. Not all private lands have been poorly managed from the standpoint of the woodpecker. The Red Hills Hunting Plantations of southern Georgia contain the sixth largest aggregation of birds (James 1994). This region is dominated by uneven-age management and selective cutting (James 1994), the antithesis of the methods primarily used on national forests and large timber plantations (Jackson 1994). The generally poor situation on private lands has led the U.S . Fish and Wildlife Service to formulate a strategy for private landholders that )they claim will aid the recovery effort (Costa 1993). They view small landowners as unlikely targets for this effort because they feel the costs would be too high for them to bear and the birds on their lands are doomed anyway, so they focus on large landowners. The strategy has three parts: 1) a procedures manual for private lands, 2) a rangewide habitat conservation plan, and 3) individual habitat conservation plans or memoranda of understanding. The last element, the memoranda of understanding, is viewed as "probably the best hope of maintaining the remaining, relatively large RCW populations on private lands" (Costa 1993, p. 13). Given the importance of these memoranda, it is not surprising that the first one, with the Georgia Pacific Corporation, by far the largest southern timberland holder (Norvell 1993~ pulp and paper companies not included), was front page news in the national press (e.g., Schneider 1993). Secretary of Interior Babbitt said that the most important effect of this agreement "could well be the precedent it sets in helping to establish a less politically incendiary approach to safeguarding endangered wildlife and their habitat" (Schneider 1993). It is thus crucial that this agreement be a sound one. The company agrees to restrict operations on some 20,000 ha in return for a government promise not to invoke the Endangered Species Act to restrict logging on the remaining 1.68 million ha of Georgia Pacific timberland in the South. The restriction consists of not clearing at least 4 ha of land around each colony site on those 20,000 ha, and reducing the stocking rate to 4.59 m2 basal area of pine/ha over 61 ha. On the face of it, this agreement seems quite remarkable, given that home ranges often exceed 80 ha and may exceed 400 ha in poor habitat (references in Jackson 1994). In prime homogeneous habitat in Florida, ca. 25 ha per social group apparently suffices for population persistence (James, pers. comm. 1993). This disparity is less mysterious, perhaps, in light of the heavily criticized 1985 recovery plan (Lennartz and Henry 1985), which calls for a 4 ha core area around each cavity tree. The stocking rate is more impressive than the area, as it is quite low and constitutes a 3-fold reduction of the original plan of the company (Wood and K1einhofs 1992). From the standpoint of the company, this agreement may be acceptable in that the sacrifice of income does not unduly affect profit margin (Wood and Kleinhofs 1992). own 17% of forested areas (Doster 1993). Evidently the private sector will have to be a partner if biodiversity is to be maintained. A Case-Study: The Red-Cockaded Woodpecker The red-cockaded woodpecker (Picoides borealis) is the most publicized problematic species in southern forests. Probably it would have achieved the global notoriety of the northern spotted owl but for two facts: 1) The longleaf pine forests that are its prime habitat, though beautiful, do not match in visual impact the majestic rainforests of the· Northwest. 2) Almost all of the southern old-growth was cut down long ago (Tebo 1985), before conservation of biodiversity was even an issue and before heavy logging operations moved to the Northwest. Of about 28 million ha of original longleaf pine fQrest in the Southeast, fewer than 600 ha remain (references in Simberloff 1993a). The woodpecker has been viewed as endangered since at least 1968 (U.S.D.!. 1968) and was listed as endangered in 1970 under the Endangered Species Conservation Act of 1969. It was one of the flISt species listed under the Endangered Species Act of 1973. Two recovery plans have been written; the flISt was never implemented, and the second (Lennartz and Henry 1985) was severely criticized by a committee appointed by the American Ornithologists' Union but has not been revised (Jackson 1994). The number of birds has declined more than 20% during the last decade, and much of the decline has been in populations designated as recovery populations in the 1985 recovery plan, including populations in national forests (James 1994). The bird is important not only in its own right, but because the cavities it laboriously constructs in large, diseased trees are used by many other species (Engstrom 1993). Human activity has affected the woodpecker primarily in two ways. First, through loss of active and potential cavity trees, and second, through fragmentation and loss of foraging habitat through conversion of forest to other habitats or change in forest type because of short-rotation, even-aged management or limitation of fire (Jackson 1994). The bird does not regularly disperse more than about 8 km (Walters et al. 1988), and small, isolated sites that lose their woodpecker colonies yet appear to constitute suitable habitat often remain without birds for a long time. The decline of numbers is undoubtedly partly due to this loss of isolated populations, but it is an open question whether the failure of this aspect of metapopulation dynamics threatens the larger aggregations~ over half of all sites are in six areas (James 1994). Given the large fraction of southern forests in private ownership, it seems that recovery of this species could be greatly aided by partnerships with private landholders. Of approximately 4,000 known active sites, half are on national forests, a fourth on Department of Defense lands, and only an eighth on private lands (Costa 1993). This disparity between fraction of privately held lands and ftaction of woodpeckers reflects the fact that 88 Carter, RN., and S.D. Prince. 1988. Distribution limits from a demographic viewpoint. pp. 165-184 in A.J. Davy, MJ. Hutchings, and A.R. Watkinson (eds.), Plant Population Ecology. Oxford: Blackwell. Copeyon, CX. 1990. A teclmique for constructing cavities for tre red-cockaded woodpecker. WildI. Soc. Bull. 18:303-311. Costa, R. 1993. U.S. Fish and Wildlife SeIVice white paper on tre issue of red-cockaded woodpeckers on private lands. Unpublished. Curtis, IT. 1956. 1k modification of mid-latitude grasslands and forests by man. pp. 721-736 in W.L. 11x>mas (ed.), Man's Role in Changing the Face of the Earth. Chicago: Univ. of Chicago From the standpoint of the U.S.F.W.S., it is hard to imagine why this agreement is acceptable; certainly it does not reflect an abundance of scientific evidence. The rangewide habitat conselVation plan of the U.S.F.W.S. seems peculiar as well. It will apparently consist of memoranda of understanding with large landholders, such as the one with GeoiWa Pacific, plus a global agreement with smaII landholders by which woodpeckers on their lands would be moved to federal Iands or to larger private Iands (Costa 1993). It is again clear that such cooperation benefits the landholders: it relieves them of the onus of managing for an endangered species. It is again not clear that the U.S.F.W.S. will benefit. The U.S.F.W.S. sees these birds as a potential aid in designated, larger recovety populations. 1\vo recent developments (reviewed by Jackson 1994) spur such reasoning: movement of young females from natal sites to clans lacking a female (DeFazio et al. 1987), and the construction of artificial cavities (Copeyon 1990, AIlen 1991). Although both techniques may be of great use in recovety, they are sufficiently new~ that one cannot be certain how well they will wotk in the long term, and they are expensive. Although the Endangered Species Act does not address the fate of species that interact with a listed species, an ecosystem approach to forest management surely would, and there has been no substantial study of how artificial cavities are used by the many species who depend on red-cockaded woodpecker holes. In fact, if the new thrust of forest management is to be ecosystemic, a goal proclaimed not only by the Forest SelVice but by Secretaty Babbitt (U.S.D.I. 1993), species-specific remedies such as movement of individuals and provision of artificial habitat will likely play a reduced role. Rather, a real ecosystem approach necessitates addressing entire native ecosystems. The longleaf pine forests are ripe for such an approach, with numerous species of special concern, some understanding of the problems of many of them, and a good scientific basis for management (Hermann 1993). One hears much talk about "getting ahead of the extinction cUlVe" nowadays; we will never get ahead of this CUlVe if we attempt to save one species at a time. Press. DeFazio, IT., Jr., M.A. Hunnicutt, M.R. Lennartz, G.L. Chapman, andIA. Jackson 1987. Red-cockaded woodpeckertrnnslocation experiments in South Carolina. 1987 Proc. Annu. Conf. Southeast Assoc. Fish and WildI. Agencies 41:311-317. Doster, GL. 1993. Timber company biologists opposed to feral swine. Southeastern Cooperative Wildlife Disease Study Briefs 9(1):3-4. Engstrom, R.T. 1993. Characteristic birds and mammals of longleaf pine forests. In press in S. Hermann (ed.), Proceedings of a Conference on LongleafPine Forests. TaIIahassee, Florida: Tall Timbers, Inc. Ewe~ JJ. 1986. Invasibility: Lessons from south Florida pp. 214-230 in H.A Moorey and IA Drake (eds.), Ecology of BiolOgical Invasions ofNorth America and Hawaii. New York: Springer-Verlag. Franklin, IF. 1989. Toward a rew forestty. Amer. Forests. Nov.lDee. 1989: 1-8. Hanski, 1., and M.E. Gilpin 1991. Metapopulation dynamics: brief histoty and conceptual domain BioI. I Linn Soc. 42:3-16. Harris, L.D. 1988. Edge effects and conselVation of biotic diversity. ConselV. BioI. 2:330-332. Harrison, S. 1991. LocaI extiIdion in a metapopulation context: an empirical evaluation BioI. J. Linn Soc. '42:73-88. Hennann, S. (ed.) 1993. Proceedings ofa Conforence on Longleaf Pine Forests. In press. Tallahassee, Florida: Tall Timbers, Inc. Hobbs, R.J. 1992. The role of corridors in conselVation: Solution or bandwagon? Trends in Recent EcoI. EvoI. 7:389-392. Hunter, M.L., G.L. Jacobson, Jr., and T. Webb m. 1988. Paleoecology and the coarse-futer approach to maintaining biological diversity. ConselV. BioI. 2:375-385. Hussey, B.M.J., R.J. Hobbs, and D.A. Saunders. 1990. Guidelines for Bush Corridors. Chipping Norton, N.S.W.: Organizing Committee of conference "Nature ConselVation: the Role of Corridors." Jackson, J.A. 1994. Red-cockaded woodpecker. In press in A. Poole, P. Stettenheim, and F. Gill (eds.), Birds of North America. Philadelphia: A.O.U. J~ F.e. 1994. The status oftre nrl-<:OCkackrl woodpecker in 19W ani tre prospect for recovery. In p-ess in R Costa, D. Kullnvy, an:l R H~ (aIs.), Proceedings of the Red-Codmded Woodpecker Symposiwn m. Charleston, S.Car.: U.SDA FonS: Service. REFERENCES AIlen, D.H. 1991. An insert teclmique for constructing artificial red-cockaded woodpecker cavities. U.S.D.A. Forest Service, Southeast For. Exp. Sta., Gen Techn Rep. SE-73. Alverson, W.S., Waller, D.M, am S.L. Solheim. 1988. Forests too deer: Edge effects in northern Wisconsin ConselV. BioI. 2:348-358. Andren, H., and P. AngeIstam 1988. Elevated predation rates as an edge effect in habitat islands: experimental evidence. Ecology 69:544-547. Boonnan, S.A., and P.R Levitt. 1973. Group selection on the boundaIy of a stable population 1koret. Pop. BioI. 4:85-128. 89 Janzen, DR. 1983. No Jmk is an is1aId: haease in intetfereoce from outside as patk size decreases. Oikos 41:402-410. Janzen, DR. 1986. The external threat. pp. 286-303 in ME. Soule (eel), Conservation Biology: The Science ofScarcity and Diversity. SwdetiaOO, Mass.: Sinauer. Kessler, WB., H. Salwasser, C.W. Cartwright, Jr., am lA Caplan 1992. New perspectives for sustainable natural resources management. Ecol. Appl. 2:221-225. LennaI1:4 MR., am V.G. Hemy. 1985. EOOangered species recovety plan Red-cockaded woodpecker Picoides borealis. Atlanta: U.S. Fish am Wildlife Setvice. Levins, R 1969. So~ deIOOgrnphic am geretic consequeoces of enviromnental rererogeneity for biological control. Bull. Entomol. Soc. Amer. 15:237-240. Mann, C.C., am ML. PIUlIlIrer. 1993. The high cost of biodiversity. Scieoce 260: 1868-1871. McIntyre, S., am G.W. Banett. 1992. Habitat variegation, an ahemative to fragrrentation ConseIV. Biol. 6:146-147. Meniam, G. 1991. Conidors conrectivity: animal populations in reterogeneous enviromrents. pp. 133-142 in D.A Sawders and Rl Hobbs (eds.), The Role of Corridors. Chipping Norton, N.S.W., Australia: Surrey Beatty. Moore, N.W. 1962. ~ 1raths of Dorset am kir conservation 1 EcoI. 50:369-391. Naturnl Heritage Data Center Netwotk. 1993. Perspectives on Species Imperilment (revised printing). Arlington, Va: ~ Nature Conservaocy. NOIVell, S. 1993. So1J1:lan comfort for a timber giant. N.Y. Times, March 28, p. 6F. Noss, RF. 1993. Wildlife conidors. In press in D. Smith and P. Hellmund (cds.), Ecology of Greenways. Minreapolis: Univ. of Minresota Press. Pulliam, H.R 1988. SoUICeS, sinks, am population regulation ~r. Natur. 132:652-661. Ranney, lW., M.C. Bruner, and lB. Levenson 1981. The importance of edge in the structure and dynamics of forest islands. pp. 67-95 in R.L. Burgess and D.M. Sharpe (eds.), Forest Island Dynamics in Man-Dominated Landscapes. New YOlk: Springer-Verlag. Sawders, D.A, Rl Hobbs, am C.R Margules. 1991. Biological consequeoces of ecosystem frngmentation: A review. ConseIV. BioI. 5: 18-32. Schreider, K 1993. AccoId is rea;hed to aid forest biId. N.Y. Times, April 16, pp. Al,A9. Shaffer, M.L. 1981. Minimum population sizes for species conseIVation BioScieoce 31: 131-134. Simberloff, D. 1988. ~ contnbution of population am commwlity biology to conservation science. Annu. Rev. Ecol. Syst. 19:473-511. Simberloff, D. 1993a. Species-area and frngmentation effects on old growth forests-prospects for longleaf pine communities. In press in S. Hennann (ed.), Proceedings of a Conference on Longleaf Pine Forests. Tallahassee, Florida: Tall Timbers, Inc. Simberloff, D. 1993b. ~ ecology of extioct:ion In press in Acta Pahmntologica Polonica Simberloff, D. 1994.1nfroduced species. In press in M Yelles (ed.), Encyclopedia of Environmental Biology. San Diego: Academic Press. Simberloff, D., and 1 Cox. 1987. Consequeoces am costs of conservation conidors. Conserv. BioI. 1:63-71. SinDerloff, D., lA Farr, I Cox, am D.W. Mehlman 1992. Move~nt corridors: Conservation baIEIDns or poor imrestments? Conserv. BioI. 6:493-504. Small, M.F., am ML. Hunter. 1988. Forest fragrrentation am avian rest predation in forested lardscapes. Oecologia 76:62-64. Swanson, FJ., ani IF. Franklin 1992. New forestIy prin;;iples from ecosystem analysis of Pocmc Northwest forests. Ecol. AppI. 2:262-274. Tebo, ME. 1985. The Southeastern Piney Woods: Describers, Destroyers, Survivors. Master' s thesis, Dept English, Florida State Univ., Tallahassee, Florida. TeIb01gh, I 1988. ~ big ~ that run k world - A sequel to E.G. Wilson Conserv. BioI. 2:402-403. TschinkeL W.R 1988. Distnbution of k fire ants Solenopsis invicta am s. geminata (H~mptera: Fonnicidae) in rorthm Florida in relation to habitat am disttnbaoce. Annals EntoIOOl. Soc. AIrer. 81:76-81. United States Department of the Interior. 1968. Rare am eOOangered fish am wildlife of k United States. WashingtOn: U.S. Bur. Sport Fish Wildl. Resoun:e PubI. 34. United States Department of the Interior. 1993. Gnatcatcrer to be listed as " threatened~" Interior's Babbitt promotes regional consetVation efforts. News release, March 25. Walters, JR, P.D. Doerr, am JR. Carter, ill. 1988. The cooperative breeding system of the red-cockaded woodpecker. Eth>logy 78:275-305. Wilcove, D. 1985. Nest predation in forest trocts am the declire of migratOlY songbiIds. Ecology 66:1211-1214. Wilcove, D. 1990. Empty skies. Nature Conservarx;y Magazire 40(1):4-13. Wilcove, D., CR. McLellan, am AP. Dobson 1986. HIDitat fragrrentation in the temperate zooo. pp. 237-256 in ME. Soule (ed.), Conservation Biology: The Science ofScarcity and Diversity. StnJerlam, Mass.: Sinauer. Williamson, M 1975. ~ design of wildlife reseIVes. Nature 256:519. Wilson, E.O. 1992. The Diversity ofLifo· Cambridge, Mass.: Harvard Univ. Press. Wood, G.W., and 1 Kleinhofs. 1992. Integration of timber management and red-<X>Ckaded woodpecker conseIVation goals on Georgia-Pacific CoIpOration tinDerlaOOs in k somrem United States. Unpublished white paper, Oct. 9. Yahrer, RH. 1988. Changes in wildlife comnnmities near edges. Conserv. BioI. 2:333-339. Yahner, R.H., and D.P. Scott. 1988. Effects of forest fragmentation on depredation of artificial nests. J. Wildl. Mgt. 52:158-161. arx\ 90