Corridor Use by Diverse Taxa

Ecology, 84(3), 2003, pp. 609–615 q 2003 by the Ecological Society of America CORRIDOR USE BY DIVERSE TAXA NICK M. HADDAD,1,7 DAVID R. BOWNE,2 ALAN CUNNINGHAM,3 BRENT J. DANIELSON,4 DOUGLAS J. LEVEY,5 SARAH SARGENT,6 AND TIM SPIRA3 1 Department of Zoology, Box 7617, North Carolina State University, Raleigh, North Carolina 27695-7617 USA 2Blandy Experimental Farm, University of Virginia, Boyce, Virginia 22620 USA 3Department of Biological Sciences, Clemson University, Clemson, South Carolina 29634-0326 USA 4Ecology and Evolutionary Biology IGP, Iowa State University, Ames, Iowa 50011-3221 USA 5Department of Zoology, University of Florida, Gainesville, Florida 32611-8525 USA 6Department of Biology, Allegheny College, Meadville, Pennsylvania 16335 USA Abstract. One of the most popular approaches for maintaining populations and con- serving biodiversity in fragmented landscapes is to retain or create corridors that connect otherwise isolated habitat patches. Working in large-scale, experimental landscapes in which open-habitat patches and corridors were created by harvesting pine forest, we showed that corridors direct movements of different types of species, including butterflies, small mam- mals, and bird-dispersed plants, causing higher movement between connected than between unconnected patches. Corridors directed the movement of all 10 species studied, with all corridor effect sizes .68%. However, this corridor effect was significant for five species, not significant for one species, and inconclusive for four species because of small sample sizes. Although we found no evidence that corridors increase emigration from a patch, our results show that movements of disparate taxa with broadly different life histories and functional roles are directed by corridors. Key words: biodiversity; bird; butterfly; conservation; corridors; dispersal; fragmentation; fru- givory; landscape experiment; movement; pollination; small mammals. Reports INTRODUCTION To date, most corridor studies have focused on sin- gle species, or on groups of closely related taxa (see Corridors are long, thin strips of habitat that connect recent reviews in Rosenberg et al. 1997, Beier and otherwise isolated habitat patches. They are thought to Noss 1998, Hess and Fischer 2001). For example, cor- reduce local extinction by ‘‘rescuing’’ isolated popu- ridors have been shown to serve as movement con- lations (Brown and Kodric-Brown 1977) and by pro- duits for species of birds (Haas 1995), mammals moting gene flow. Indeed, recent studies have dem- (Beier 1995), and butterflies (Sutcliffe and Thomas onstrated that corridors can increase animal movement 1996, Haddad 1999a). However, for other species of between patches (Haas 1995, Sutcliffe and Thomas birds (Date et al. 1991), butterflies (Haddad 1999b), 1996, Gonzalez et al. 1998, Haddad 1999a, Mech and and small mammals (Suckling 1984, Henderson et al. Hallett 2001), increase population sizes (Fahrig and 1985, Bowne et al. 1999), as well as a salamander Merriam 1985, Dunning et al. 1995, Haddad and Baum 1999), increase gene flow (Aars and Ims 1999, Hale et species (Rosenberg et al. 1998), corridors do not in- al. 2001, Mech and Hallett 2001), and maintain bio- crease movement. In addition, many major taxa have diversity (Gonzalez et al. 1998). Other studies, how- not been considered in corridor studies. For example, ever, have found no corridor effects (Arnold et al. 1991, there are no studies of corridor effects on the move- Date et al. 1991, Rosenberg et al. 1998, Bowne et al. ment of plants or of many insect groups. 1999, Haddad and Baum 1999, Collinge 2000, Dan- Using a large-scale experiment, we synthesized re- ielson and Hubbard 2000). In one recent review, Beier sults from studies of the effects of corridors on move- and Noss (1998) evaluated 32 corridor studies and ment by species from diverse taxa, including butter- found that, although many were inconclusive due to flies, bees, bird-dispersed plants, and small mammals. flaws in study design, 10 of 12 well-designed studies Although each taxon was studied independently, all demonstrated positive effects of corridors. Given that studies took place in the same experiment. In our anal- habitat conservation measures such as corridors are ysis, we incorporated new and previously published likely to affect many species in a landscape, an im- data to address two specific questions. (1) For which portant, unresolved question is: Which species in a taxa do corridors direct movement? (2) What are the landscape will benefit from corridors? magnitudes of the effects of corridors on movement across taxa? In the discussion, using previously pub- lished data from this experiment on a subset of the Manuscript received 7 December 2001; revised 14 June 2002; accepted 18 July 2002. Corresponding Editor: F. W. Davis. taxa considered in this paper, we examine a third ques- 7 E-mail: nick haddad@ncsu.edu tion: for which taxa do corridors increase emigration? 609  610 NICK M. HADDAD ET AL. Ecology, Vol. 84, No. 3 FIG. 1. Corridor experiment at the Savannah River National Environmental Research Park, South Carolina, USA. Black areas show patches (128 3 128 m, numbered 1–27) and 10 corridors (32 m wide, varying in length from 64 m to 384 m). Solid lines show roads, and stippled areas are ponds and streams. Reports METHODS habitats that we studied. These species do not require Our research was conducted within large-scale, ex- clearcut corridors for movement in unmanaged land- perimental patches and corridors at the Savannah River scapes, but are adapted to colonizing ephemeral open- National Environmental Research Park, near Aiken, ings within forests, some of which may be connected South Carolina, USA. Patches and corridors were early- by strips of secondary vegetation. successional vegetation and were created by harvesting Research on different taxa followed one of two ap- pine trees within large pine plantations. The experiment proaches. In the first, movements of two butterfly spe- consisted of 27 128 3 128 m open patches (1.64 ha), cies (Junonia coenia and Euptoieta claudia) and one some of which were connected by a 32 m wide, open small-mammal species (Sigmodon hispidus) were stud- corridor that varied in length from 64 to 384 m (Fig. ied between all pairs of adjacent patches (see Fig. 1). 1). These patch sizes were chosen because they are In the second, movements of four plant species (Ilex large enough to prevent shading of the entire patch by opaca, Myrica cerifera, Phytolacca americana, and surrounding trees, and because they match the scale of Rhus copallina), another small-mammal species ( Per- movement of many of the study organisms. For ex- omyscus polionotus), and one bee species (Xylocopa ample, the average lifetime movement distance of two virginica), as well as pollen of one plant species ( Pas- butterfly species, Junonia coenia and Euptoieta clau- siflora incarnata), were studied in ‘‘blocks’’ of three dia, are approximately the width of a patch (Haddad patches. In each block, one center patch was surround- 1999a), and home ranges of old-field mice, Peromyscus ed by two patches equidistant from the center patch. polionotus, are approximately one-third the area of a One of the two peripheral patches was connected to patch (Davenport 1964). The patches and corridors the center patch by a corridor, and the other was isolated were dispersed among five different stands of pine for- from the center patch (e.g., patches 14, 15, and 16, est (mainly Pinus taeda and P. palustris). where patch 15 is the center patch; Fig. 1). In this The landscapes of open patches and corridors sur- second approach, and except where noted, naturally rounded by forest were the inverse of landscapes with occurring individuals of a study species were elimi- forested patches and corridors that are typically con- nated from patches and corridors, and newly introduced sidered in conservation. The key aspect of this system, individuals were monitored. Except where noted, all similar to that of other landscapes where corridors are research was conducted in 1996. proposed, was that patches and corridors were suitable The results published here are a synthesis of studies for the species that we studied and contrasted with for many species, and the analysis of corridor effects unsuitable matrix habitat. All of our study species were on movement directions of butterflies was published common in the region and occurred naturally in the previously (Haddad 1999a). In addition, studies of P. March 2003 CORRIDOR USE BY DIVERSE TAXA 611 polionotus (Danielson and Hubbard 2000) and S. his- 3 4 grid of 16 perches in each patch. For P. americana pidus (Bowne et al. 1999) from this experiment have and R. copallina, naturally occurring individuals were been published, but the analyses of the effects of cor- removed from the peripheral patches, and seeds of these ridors on movement direction are new. Although un- species recovered in the peripheral patches were as- published data comprise the bulk of the paper, we in- sumed to have been dispersed from the center patch, clude the published results on corridor effects on move- which was the nearest seed source. The two other spe- ment directions of butterflies because they allow us to cies, I. opaca and M. cerifera, were not eliminated, and make clean cross-taxonomic comparisons. seed dispersal was tracked by marking seeds in the Butterflies were studied in daily surveys of all patch- center patch with flourescent microspheres, which are es between 3 April and 29 June (for detailed methods, defecated with seeds (Levey and Sargent 2000). Fecal see Haddad 1999a). Each butterfly was captured and samples were collected weekly from 26 August 1996 marked with a unique number. An average of 124 J. to 15 March 1997, and from 3 November 1997 to 10 coenia (range 38–532) and 11 E. claudia (range 1–38) April 1998, during the peak fruiting season of each were marked in each patch. Recaptures provided mea- species (fall 1996 for P. americana, winter 1997 and sures of movement between all pairs of adjacent patch- 1998 for M. cerifera and R. copallina, and winter 1998 es. for I. opaca). The numbers of seeds of each species Adult Sigmodon hispidus were captured at least 13 recovered from fecal samples in each patch were km from the experimental area, fitted with radio collars, and released into one of 10 different patches that were summed over the entire season. When a species was contained within one forest stand (Fig. 1, patches 1– studied for two years, analyses were conducted on the 10; detailed methods are provided in Bowne et al. average numbers of seeds dispersed at each site among [1999]). Although movement distances of released in- years. dividuals were expected to be greater than those of Pollination was measured for Passiflora incarnata, resident individuals, there was no evidence that move- a plant species pollinated by large-bodied bees and ment direction was biased toward the initial capture wasps. The experiment followed the same design as area (Bowne et al. 1999). In total, six individuals were that of the seed dispersal study. After naturally occur- released into each of two different patches, and their ring plants were removed from the patches and corri- Reports movements were monitored for 10 consecutive days. dors, 24 individual P. incarnata were planted next to These releases were repeated until all patches had been trellises in each patch. Pollination was estimated daily studied once (n 5 10 patches), and were repeated a from 1 July to 14 August by marking flowers in the second time in six of the same patches. Radio telemetry center patch with fluorescent powder and recording the permitted direct observations of animal movements be- presence of powder on flowers in peripheral patches. tween adjacent patches. In addition, pollinators, including large-bodied bees Peromyscus polionotus were eliminated from three and wasps, were marked during the same period with blocks of three patches, where the distance between unique tags. Because one species, Xylocopa virginica, center and peripheral patches was 128 m or 256 m. comprised 90% of marked individuals and was the only Between 16 and 27 adult female mice were ear-tagged species recaptured, we restrict our discussion to it. and introduced into nest boxes in the center patch of We used one-tailed t tests to test the hypothesis that each block (detailed methods are provided in Danielson plants and animals move more frequently between con- and Hubbard [2000]). In our system, introduced P. po- nected than between unconnected patches. For the sec- lionotus move more frequently between patches than ond approach just described, where connected and un- do naturally occurring individuals, but their movement connected patches both occurred in the same block, we orientation is not biased (i.e., by homing; J. Brinker- used paired t tests. Where appropriate, we arcsine- hoff, unpublished data). After acclimating to the boxes, transformed proportions or square-root transformed mice were allowed to move freely. Marked individuals counts (after adding 0.375) to normalize data (Zar were recaptured in 18 Sherman traps (Sherman Traps, Tallahassee, Florida, USA) arranged in a grid in each 1999). Because some of the studies involved small sam- of the nine patches. Live-trapping was conducted three ple sizes (as low as two or three replicates), we con- days per week from 1 May to 30 August. Recaptured sidered a corridor effect significant if P , 0.10. For mice were used to estimate movement between con- cases in which sample sizes were low (n # 3), we nected and unconnected patches. conducted power analyses to estimate the number of Seed dispersal of bird-dispersed plants (Ilex opaca, samples needed to detect a significant effect, given the Myrica cerifera, Phytolacca americana, and Rhus co- values that we recorded (Steidl et al. 1997). pallina) was studied in 2–4 blocks of three patches Although distances between patches were variable (depending on plant species), where the distance be- and replicated, many of the studies were not designed tween patches was 64, 128, 256, or 384 m. Dispersed to test for distance effects on corridor function. Hence, seeds were collected from avian fecal samples in seed we exclude examination of distance effects in this pa- traps that were placed under perches positioned in a 4 per.  612 NICK M. HADDAD ET AL. Ecology, Vol. 84, No. 3 Reports FIG. 2. Plant and animal movement between connected and unconnected patches. Panels (A)–(D) and (J) show the mean (11 SE) proportion of individuals that were marked in one patch and moved to a connected or unconnected patch. Panels (E)–(H) show the mean (11 SE) number of bird-dispersed seeds that moved from center patches to connected or unconnected patches. Panel (I) shows the mean (11 SE ) proportion of flowers in connected or unconnected patches with fluorescent powder. Data for butterflies are adapted from (Haddad 1999a). Asterisks indicate significance levels: * P , 0.10; ** P , 0.05; *** P , 0.01. RESULTS For one rodent species, Sigmodon hispidus, the cor- ridor effect on movement direction was not significant Of the 10 species that we studied, five moved sig- (Fig. 2C; t test assuming unequal variance: n 5 22 nificantly more often between connected than between patch pairs, t 5 1.38, P 5 0.107). The apparent large unconnected patches (Fig. 2), and all species showed positive corridor effects (calculated as the percentage increase in movement between connected relative to unconnected patches) .68% (Fig. 3). As shown pre- viously (Haddad 1999a), both butterfly species moved more frequently between connected than unconnected patches (Fig. 2A,B; t test assuming equal variance: Ju- nonia coenia, n 5 48 patch pairs, t 5 2.34, P 5 0.012; Euptoieta claudia, n 5 47, t 5 2.14, P 5 0.019). As in Haddad (1999a), one outlier was dropped from the analysis of E. claudia because only two individuals were marked in that patch. One rodent, Peromyscus polionotus, moved more frequently between connected patches (Fig. 2D; paired t test: n 5 3 blocks, t 5 3.54, P 5 0.036). Finally, seeds of two bird-dispersed plants, Rhus copallina (Fig. 2E; paired t test: n 5 4 blocks, t FIG. 3. The corridor effect, measured as the percentage 5 4.71, P 5 0.009) and Myrica cerifera (Fig. 2F; paired increase in movement between connected relative to uncon- nected patches, presented in the rank order of the size of the t test: n 5 4 blocks, t 5 2.13, P 5 0.062) moved more effect. For Sigmodon hispidus, the black bar excludes one frequently between connected than unconnected patch- outlier, described in Results, whereas the white bar includes es. the outlier. March 2003 CORRIDOR USE BY DIVERSE TAXA 613 difference in movement into connected vs. unconnected that corridors influence the movement directions in patches (means 5 0.181 vs. 0.057, respectively) was these landscapes of many species with very dissimilar generated by a single outlier that, when removed, re- features. duces mean movement to connected patches from 0.181 Other previously published results from our exper- to 0.117. This statistical result, based on an analysis iment allow us to answer our third question for a subset with large sample size, led us to conclude that this of the taxa: for which taxa do corridors increase em- species does not move preferentially between con- igration? In those studies (Bowne et al. 1999, Haddad nected patches. 1999a, Danielson and Hubbard 2000), corridors did not Movements of four other species appeared to be di- influence the number of emigrants of any species for rected by corridors, but our results were inconclusive, which this response was measured, including butterflies probably due to low sample size. For one plant species (Junonia coenia and Euptoieta claudia) and small (Ilex opaca, Fig. 2H), the bee species (Xylocopa vir- mammals (Peromyscus polionotus and Sigmodon his- ginica, Fig. 2J), and the pollination study (Fig. 2I), pidus). Although there are species in other landscapes sample sizes were two blocks each and analyses were that are highly restricted in habitat use and that would not possible. In the pollination study, the plant, Pas- never leave a patch through unsuitable habitat (or siflora incarnata, established in only two of four blocks would be unsuccessful emigrants if they did), our re- where they were planted. For another plant species, sults suggest that emigration of our study species is Phytolacca americana, samples were gathered from not determined by landscape pattern. three blocks and analysis was possible, yielding t 5 Given the consistent patterns across most of the spe- 1.88 and P 5 0.100 (Fig. 2G). The importance of low cies that we studied, we are tempted to draw more sample size can be illustrated by a power analysis for general conclusions about the effects of corridors on P. americana. Despite the large corridor effect on movement. However, landscape connectivity is an at- movement (over five times as many seeds were recov- tribute both of the species and of the landscape in which ered in connected than in unconnected patches), power that species resides (Weins and Milne 1989, Beier and analysis indicated that we would have needed a sample Noss 1998, Tischendorf and Fahrig 2000). Our study size of four to detect a statistically significant differ- species, which comprise a diverse set of natural-history ence. strategies and evolutionarily independent groups, are Reports not representative of all of the species that occur in our DISCUSSION landscape. In fact, we biased our selection of species Using the results of our study on 10 taxa, we can toward those that we thought were likely to respond to answer two questions raised in the Introduction about corridors based on their habitat preferences and, in corridor effects on movement. In answer to our first some cases, on their movement abilities relative to oth- question, our results demonstrate that corridors in our er species in their respective taxa. Our experiment was experimental landscapes consistently direct the move- designed to exploit the strong contrast between open- ment of diverse taxa. These results include the first ings (patches and corridors) and pine forest (the ma- demonstration that corridors affect interpatch move- trix). Thus, our choice of species within these land- ments of plants. Interestingly, corridor effects on move- scapes probably did impact our consistent finding of ment were detected even without controlling for other corridor effects on movement direction. known important factors that are not correlated with In addition, the life histories and movement behav- the presence of a corridor but that influence movement, iors of individual species influence their movement such as abundances of butterfly host and nectar plants through corridors. The birds that dispersed seeds in our (Haddad 1999a). Even though other sources of envi- study (primarily Yellow-rumped Warblers, Dendroica ronmental variation in our experiment may have influ- coronata, and Eastern Bluebirds, Sialia sialis) are large enced movement, the response to corridors overrode and highly vagile relative to the other species in our such effects. study, and frequently move across large distances re- In answer to our second question, the effects of cor- gardless of corridors. However, the dispersal patterns ridors on movement direction were uniformly large. of bird-dispersed seeds are not determined by bird dis- For all 10 species, at least 68% more individuals moved persal ability alone. Importantly, gut retention times to connected than to unconnected patches. The low are typically ,1 h in small (,30 g) frugivorous birds probability that 10 species would show the same pos- (Levey and del Rio 1999). Within this time frame, cor- itive responses to corridors by chance alone (calculated ridors appear to influence the local foraging behavior as 0.59, i.e., ,0.01) led us to conclude that the corridor of birds, thus determining the dispersal of seeds. Like- effect is generally significant. This effect was statis- wise, large bees are able to move many kilometers tically significant for five species, not significant for between nests and foraging sites (Cunningham 2000). one species, and inconclusive (because of low repli- However, they may use corridors during short-term for- cation) for four species. The species that we studied aging, which will impact the plants that they pollinate. spanned a large range of body sizes, locomotory Finally, although there was a remarkably consistent and modes, and trophic levels. Thus, our results suggest large corridor effect on movement direction across the  614 NICK M. HADDAD ET AL. Ecology, Vol. 84, No. 3 species that we studied, there is no obvious pattern in creases in movement to connected patches strongly the relationship between taxa and the size of the cor- suggest, but do not demonstrate, the value of corridors. ridor effect (Fig. 3). In the one group for which we have population data, With regard to species-specific responses to corri- butterfly species that used corridors for movement also dors, it is instructive to ask why one species, S. his- had higher population sizes in connected than in un- pidus, did not show a corridor response. We propose connected patches (Haddad and Baum 1999). However, two explanations. First, we expected species that pre- all species in our study make movements across the ferred second-growth habitat to disperse through cor- matrix (although not recorded in our capture–recapture ridors of the same habitat type. S. hispidus is much studies, Xylocopa virginica has been observed by one more abundant in open habitats than in forest (Golley of us (T. Spira) to move through the matrix). Such et al. 1965). In addition to habitat preference, however, movements between isolated patches might be suffi- movement behaviors, including aversion to movement cient to rescue populations and maintain genetic di- through unsuitable habitat, also influence responses to versity (Simberloff et al. 1992, Beier and Noss 1998). corridors. Such behavioral responses were measured in More work is needed to understand how differences in another group of published studies from our experi- movement impact plant and animal populations. ment. The movement paths of individuals were mapped We have shown that corridors are effective at di- by radio-tracking Sigmodon hispidus (Bowne et al. recting the dispersal of diverse taxa. These taxa are 1999) and by visually tracking three butterfly species important in a broad range of community functions, (Haddad 1999b). S. hispidus preferentially emigrated including herbivory, pollination, seed predation, and from patches through corridors (Bowne et al. 1999), seed dispersal. Given the breadth of species that re- although, as previously demonstrated, this behavior did spond to corridors, our results suggest that corridors not affect their ultimate destination. For butterflies, two may have broader, community-wide impacts. This pos- of three species (Phoebis sennae and Eurema nicippe) sibility is typically overlooked by studies of single spe- were found to preferentially leave patches through cor- cies. Taken together, our results suggest that corridors ridors and to use corridors for movement (Haddad have the potential to be valuable tools for landscape- 1999b), whereas a third (Papilio troilus) showed no scale conservation of diverse taxa and the biological preference for, or use of, corridors. A second expla- processes that they direct. Reports nation for the lack of corridor use in S. hispidus relates ACKNOWLEDGMENTS to the observation that it is the largest species consid- ered in this study, with an average home range size of We thank Robert Cheney and the U.S. Forest Service Re- search, Fire and Forest Products staffs at the Savannah River 0.247 ha (Cameron and Spencer 1985). The scale of National Environmental Research Park for their role in cre- our landscape manipulations may have been too small ating and maintaining the experimental sites. We thank Frank for S. hispidus to have perceived corridors as a sig- Davis, Jeremy Lichstein, John Orrock, and Dan Rosenberg nificant component of habitat heterogeneity. In short, for comments on the manuscript, and Ken Pollock for statis- corridor effects depend on corridor size relative to the tical advice. Funding was provided primarily by the Depart- ment of Energy–Savannah River Operations Office through scale at which a species perceives the landscape. the U.S. Forest Service, Savannah River, under Interagency After considering the caveats associated with the Agreement DE-AI09-76SR00056. types of species and landscape used in our study, we feel that our results support more general conclusions LITERATURE CITED about the effects of corridors. In total, including results Aars, J., and R. A. Ims. 1999. The effect of habitat corridors from this and previous studies in our experimental on rates of transfer and interbreeding between vole demes. landscapes (Haddad 1999b, Haddad and Baum 1999), Ecology 80:1648–1655. Arnold, G. W., J. R. Weeldenberg, and D. E. Steven. 1991. there were only two species of 13 (Sigmodon hispidus Distrubution and abundance of two species of kangaroo in and Papilio troilus) for which corridors did not direct remnants of native vegetation in the central wheatbelt of movement. All of the species in this study showed a Western Australia and the role of native vegetation along positive effect of corridors on movement (Fig. 3), and road verges and fencelines as linkages. Pages 273–280 in D. A. Saunders and R. J. Hobbs, editors. Nature conser- none showed a negative response to corridors. Another vation 2: the role of corridors. Surrey Beatty, Chipping result that has been consistently supported in our work, Norton, New South Wales, Australia. for a smaller set of species, is that although corridors Beier, P. 1995. Dispersal of juvenile cougars in fragmented do direct emigrants, they do not increase the numbers habitat. Journal of Wildlife Management 59:228–237. of emigrants. In sum, there appear to be consistent Beier, P., and R. F. Noss. 1998. Do habitat corridors really provide connectivity? Conservation Biology 12:1241– responses across a range of habitat-restricted taxa that 1252. may support the use of corridors in landscape planning. Bowne, D. R., J. D. Peles, and G. W. Barrett. 1999. Effects Preferential movement through corridors is neces- of landscape spatial structure on movement patterns of the sary, but not sufficient, for corridors to rescue other- hispid cotton rat (Sigmodon hispidus). Landscape Ecology 14:53–65. wise isolated patches from extinction and to increase Brown, J. H., and A. Kodric-Brown. 1977. Turnover rates in gene flow. In our study, we lack data on population insular biogeography: effect of immigration on extinction. viability and genetic diversity, and our dramatic in- Ecology 58:445–449. March 2003 CORRIDOR USE BY DIVERSE TAXA 615 Cameron, G. N., and S. R. Spencer. 1985. Assessment of Hale, M. L., P. W. W. Lurz, M. D. F. Shirley, S. Rushton, R. space-use patterns in the hispid cotton rat (Sigmodon his- M. Fuller, and K. Wolff. 2001. Impact of landscape man- pidus). Oecologia 68:133–139. agement on the genetic structure of red squirrel popula- Collinge, S. K. 2000. Effects of grassland fragmentation on tions. Science 293:2246–2248. insect species loss, colonization, and movement patterns. Henderson, M. T., G. Merriam, and J. Wegner. 1985. Patchy Ecology 81:2211–2226. environments and species survival: chipmunks in an ag- Cunningham, A. 2000. Landscape distance and connectivity ricultural mosaic. Biological Conservation 31:95–105. effects on pollination of Passiflora incarnata L. Thesis. Hess, G. R., and R. A. Fischer. 2001. Communicating clearly Clemson University, Clemson, South Carolina, USA. about conservation corridors. Landscape and Urban Plan- Danielson, B. J., and M. W. Hubbard. 2000. The influence ning 55:195–208. of corridors on the movement behavior of individual Per- Levey, D. J., and C. Martı´nez del Rio. 1999. Test, rejection, omyscus polionotus in experimental landscapes. Landscape and reformulation of a chemical reactor-based model of gut Ecology 15:323–331. function in a fruit-eating bird. Physiological and Biochem- Date, E. M., H. A. Ford, and H. F. Recher. 1991. Frugivorous ical Zoology 72:369–383. pigeons, stepping stones, and needs in northern New South Levey, D. J., and S. Sargent. 2000. A simple method for Wales. Pages 241–245 in D. A. Saunders and R. J. Hobbs, tracking vertebrate-dispersed seeds. Ecology 81:267–274. editors. Nature conservation 2: the role of corridors. Surrey Mech, S. G., and J. G. Hallett. 2001. Evaluating the effec- Beatty, Chipping Norton, New South Wales, Australia. tiveness of corridors: a genetic approach. Conservation Bi- Davenport, L. B., Jr. 1964. Structure of two Peromyscus po- ology 15:467–474. lionotus populations in old-field ecosystems at the AEC Rosenberg, D. K., B. R. Noon, J. W. Megahan, and E. C. Savannah River Plant. Journal of Mammalogy 45:95–113. Meslow. 1998. Compensatory behavior of Ensatina esch- Dunning, J. B., Jr., J. R. Borgella, K. Clements, and G. K. scholtzii in biological corridors: a field experiment. Ca- Meffe. 1995. Patch isolation, corridor effects, and colo- nadian Journal of Zoology 76:117–133. nization by a resident sparrow in a managed pine woodland. Rosenberg, D. K., B. R. Noon, and E. C. Meslow. 1997. Conservation Biology 9:542–550. Biological corridors: form, function, efficacy. BioScience Fahrig, L., and G. Merriam. 1985. Habitat patch connectivity 47:677–687. and population survival. Ecology 66:1762–1768. Simberloff, D., J. A. Farr, J. Cox, and D. W. Mehlman. 1992. Golley, F. B., J. B. Gentry, L. D. Caldwell, and L. B. Dav- Movement corridors: conservation bargains or poor in- enport. 1965. Number and variety of small mammals on vestments? Conservation Biology 6:493–504. the AEC Savannah River Plant. Journal of Mammalogy 46: Steidl, R. J., J. P. Hayes, and E. Schauber. 1997. Statistical 1–18. power analysis in wildlife research. Journal of Wildlife Gonzalez, A., J. H. Lawton, F. S. Gilbert, T. M. Blackburn, Management 61:270–279. and I. Evans-Freke. 1998. Metapopulation dynamics, abun- Suckling, G. C. 1984. Population ecology of the sugar glider, Reports dance, and distribution in a microecosystem. Science 281: Petaurus breviceps, in a system of fragmented habitats. 2045–2047. Australian Wildlife Research 11:49–75. Haas, C. A. 1995. Dispersal and use of corridors by birds in Sutcliffe, O. L., and C. D. Thomas. 1996. Open corridors wooded patches on an agricultural landscape. Conservation appear to facilitate dispersal by ringlet butterflies (Aphan- Biology 9:845–854. topus hyperantus) between woodland clearings. Conser- Haddad, N. M. 1999a. Corridor and distance effects on in- vation Biology 10:1359–1365. terpatch movements: a landscape experiment with butter- Tischendorf, L., and L. Fahrig. 2000. On the usage and mea- flies. Ecological Applications 9:612–622. surement of landscape connectivity. Oikos 90:7–19. Haddad, N. M. 1999b. Corridor use predicted from behaviors Weins, J. A., and B. T. Milne. 1989. Scaling of ‘landscapes’ at habitat boundaries. American Naturalist 153:215–227. in landscape ecology, or, landscape ecology from a beetle’s Haddad, N. M., and K. A. Baum. 1999. An experimental test perspective. Landscape Ecology 3:87–96. of corridor effects on butterfly densities. Ecological Ap- Zar, J. H. 1999. Biostatistical analysis. Fourth edition. Pren- plications 9:623–633. tice Hall, Upper Saddle River, New Jersey, USA.