Abernethy, Kate, ERIC.W. SANDERSON and Gillian Woolmer. Wildlife Conservation Society Gabon Program, Lope Reserve, Gabon (KA), Wildlife Conservation Society, 2300 Southern Blvd., Bronx, NY 10460, USA (EWS and GW).
A GEOGRAPHIC ANALYSIS OF THE GAP CROSSING BEHAVIOR OF MANDRILLS IN A FOREST/SAVANNA MOSAIC IN GABON.
Surprisingly large groups (300 - 600 individuals) of mandrills (Mandrillus sphinx) have been shown to use a mosaic of disconnected forest corridors in a forest/savanna mosaic at the Lope Reserve, Gabon. These groups restrict themselves almost entirely to the forest corridors, rarely ranging more than 15 meters into the open savanna. As a result, some islands of forest appear to be effectively isolated because the gaps between forest corridors are too wide for the mandrills to cross. We examined this behavior using a geographic analysis of gap cross distances and hypothesized barriers to crossing, including roads, railroads and swampy areas. Based on over one year of field monitoring, we quantified the sizes of gaps the mandrills actually crossed. We compared these results to a measurement of all possible gaps using a forest / savanna vegetation map interpreted from a 6 m resolution radar image. We showed a distinct preference for gaps less than 110 m and showed that the swampy areas may pose a barrier to crossing, but roads may not. We suggest from these results that selective reforestation or changing the anthropogenic burning regime which maintains the savannas may result in increased habitat for these mandrills.

ABRAMS, PETER A. Dept Zoology, U of Toronto, 25 Harbord St., Toronto, ON M5S 3G5, Canada.
ABRUPT POPULATION CRASHES WITH STEADILY DETERIORATING CONDITIONS IN ECOLOGICAL MODELS.
When gradual changes in environmental conditions reduce individual fitness within a species, simple models suggest that population density may often not change significantly for many years, and then decline precipitously to extinction. The early lack of response to environmental degradation is due to compensatory changes in intra- or inter-specific interactions. Models of simple food webs show what conditions frequently produce abrupt rather than gradual declines in population size with deteriorating conditions. Weak density dependence, strongly saturating consumer functional responses, self-reproducing resource populations, over-exploitation of resources by consumers, and environmentally-caused reductions in consumption rates are the major factors identified with the pattern of abrupt declines in population size. Cycling consumer populations are especially likely to show initial increases in average population sizes, followed by abrupt crashes as the environment deteriorates. These theoretical results call for increased monitoring of parameters affecting individual fitness, particularly in predatory species that greatly reduce their prey.

ACCACIO, GUSTAVO. Departamento de Zoologia, IBUSP. Rua do Matão, Travessa 14, n 101 - 05508-900, Butantã-São Paulo, SP, Brazil.
FRUIT FEEDING BUTTERFLY COMMUNITIES IN FOREST REMNANTS AND SILVICULTURES AT UNA, BA - BRAZIL.
Butterflies are regarded as good predictors of diversity and disturbance in forest ecosystems. Here frugivorous butterflies were censused with bait traps at Una, in Brazilian Atlantic Forest. The landscape included interconnected rainforest tracts, silvicultures and pastures. Butterflies were sampled in large and small mature forest patches, forest edges and wooded matrices specified as young secondary forests, shadowed cocoa and rubber tree plantations. Butterfly sub-communities were compared using rarefaction curves, diversity indexes and species relative abundances. 3,704 butterflies of 89 species were trapped. Significant differences were found among sub-communities. The dominant species in mature forests, Pseudodebis valentina, is apparently sensitive to fragmentation, being very rare or absent in matrices and small forest patches. Other forest butterflies were able to use small patches and matrices at variable degrees. Forest edges owned greater species richness due to presence of more canopy-clearing individuals and invasion of non-forest, disturbance-adapted species. Such invasion also occurred in wooded matrices, but there were marked differences within species colonization capabilities. Among the matrices, secondary growth was favorable for most butterflies found in mature forest while shadowed cocoa strongly favored small understory species, usually restricted to light gaps. Conclusion focused the utility of highly disturbed patchy ecosystems for butterfly conservation.

ACOSTA, MARTÍN, Lourdes Mugica, Dennis Denis, and Ariam Jiménez. Museo de Historia Natural Felipe Poey, Facultad de Biología, Univ. de la Habana, Calle 25 entre J e I. Vedado. Ciudad Habana, Cuba.
RICE FIELDS AND THEIR IMPORTANCE FOR THE CONSERVATION OF AQUATIC BIRDS.
Rice covers an important portion of the planet, constituting the main diet of over half the world's population. Its cultivation cycle imitates the natural hydrology of wetlands providing conservation opportunities. Mobile aquatic birds use the patchily distributed pulses of rice production characteristic of these agricultural landscapes. We studied aquatic birds in rice fields within the 27,000 ha agroindustrial complex "Sur the Jibaro," Sancti-Spiritus. Rice fields harbored a wide diversity of species: 61 were aquatic, 50 granivores and 46 mostly insectivores and piscivores. Rice fields were an important foraging habitat for North American wintering waterfowl and for 16 species that breed in adjacent coastal swamps. Seven species use paddies for both breeding and feeding, while 54 use them mostly for feeding. The Yaguasa (Dendrocygna arborea), an endangered species endemic of the Caribbean, uses rice heavily by feeding nocturnally in harvested and flooded fields. Although pesticides have been diagnosed world-wide as a major wildlife problem in rice cultivation, in Cuba a 50% reduction in pesticide use through the implementation of biological pest control have reduced this hazard. Research results are used in an Environmental Education Program implemented with university biology students in schools, zoological gardens, and Forest Guard Corps throughout the country.

ADAMS, JENNIFER, Brian T. Kelly, and Lisette P. Waits. Department of Fish and Wildlife Resources, University of Idaho, PO Box 441136, Moscow, ID 83844, USA (JA, LPW), and USFWS, 708 North Highway 64, PO Box 1969, Manteo, NC 27954, USA (BTK).
MONITORING FOR GENETIC INTROGRESSION INTO AN ISLAND POPULATION: THE RED WOLF AS A CASE STUDY.
Reintroduced red wolves (Canis rufus) in Northeastern NC have been hybridizing with recently established coyotes (Canis latrans). Molecular genetic methods are being used to aid in the USFWS' goal to preserve the red wolf gene pool. The largest problem faced by field personnel is distinguishing morphologically between red wolves and hybrids. Genotypes were obtained at 15 microsatellite loci from the 14 red wolf founders to determine alleles present in the modern population. We combined this data with wild red wolf breeding data and microsatellite allele frequency data from SE coyotes to develop an assignment test based approach to identify unknown individuals. The second largest management problem is effectively screening vast areas of the recovery zone for red wolves, coyotes and hybrids. We addressed this problem using non-invasive genetic sampling of fecal samples and GIS technology. We created a species level diagnostic test using both restriction enzyme analyses on the cytochrome b region of mtDNA and sequencing of the control region of mtDNA. Over 400 scats were collected across 250,000 acres and 72% were successfully assigned to species. These techniques will help implement the adaptive management plan and play a key role in efforts to recover the red wolf.

ADAMS, JONATHAN, S., Carrie Brugger, Yi-lun Ding, Michael Grove, Naomi Nickerson, and Daniel Peerless. The Nature Conservancy, 4245 North Fairfax Drive, Arlington, VA 22203, USA.
CONSERVEONLINE: AN ONLINE RESOURCE FOR THE CONSERVATION COMMUNITY.
The Nature Conservancy, in partnership with the Society for Conservation Biology, the Association for Biodiversity Information, and Community of Science, has launched ConserveOnline, a website designed to create a knowledge base for conservation and to foster information sharing and collaboration among conservation practitioners and researchers. The goals of the site are to provide conservationists and land managers at conservation organizations and government agencies with the practical tools they need to advance biodiversity conservation, and to link academic researchers with field practitioners. The foundation of ConserveOnline is an indexed, searchable library of conservation tools, techniques, data, and experience. We welcome broad participation and invite anyone with relevant documents, data, slide presentations, spreadsheets, software, images, or maps to add their material to the library. ConserveOnline also features Discussion Groups and Listservs, information about Conservancy and SCB programs, and links to other sources of conservation-related information.

AGNESS, ALISON, Dan Goltz, Paul C. Banko, and Jeff Trainer. Pacific Island Ecosystem Research Center, U.S. Geological Survey/ University of Hawaii, P.O. Box 44 Hawaii National Park, HI 96718, USA (AA, DG, PCB), and Kalaupapa National Historic Park, Kalaupapa, Molokai, 96742, USA (JT).
MOVEMENT PATTERNS AND POPULATION TRENDS OF AXIS DEER (AXIS AXIS AXIS) ON THE KALAUPAPA PENINSULA, MOLOKAI.
The population of axis deer (Axis axis axis) on Kalaupapa peninsula increased for more than ten years following the removal of cattle in 1985. The overabundant deer population reduced the cover of Hawaiian vegetation, including naupaka, pohina and ilima, and converted grasslands to dense shrublands dominated by Lantana camara (lantana) and Schinus terebethifolius (christmas berry). The deer population has declined by 77% since 1998 due mainly to starvation resulting from low availability of palatable species and hunting pressure. Three of twenty (15%) deer survived eighteen months after being radio-collared in 1999. Several critical areas have been fenced to exclude browsing pressure. Documentation of axis deer movement patterns may help guide fence placement and eradication efforts. Daily movement patterns were determined by tracking radio-collared deer during day and night sessions. Deer moved to higher elevation, dense stands of Psidium guajava (guava) and christmas berry during the day and receded toward lower elevation, open grasslands at night. However, one deer migrated over 12 miles where it was shot outside the peninsula and another moved <1 km in dense christmas berry between day and night tracking sessions. This reclusive behavior will complicate eradication efforts on the Kalaupapa peninsula.

AIRAME, SATIE, Sean Hastings, and John Ugoretz. Channel Islands National Marine Sanctuary. Santa Barbara, CA 93103, USA (SA and SH). California Department of Fish and Game, 1933 Cliff Drive, Suite 9, Santa Barbara 93109, USA (JU).
THE ROLE OF ECOLOGY, ECONOMICS AND SOCIETY IN MARINE CONSERVATION: A CASE STUDY FROM THE CHANNEL ISLANDS.
Numerous marine species are becoming scarce due to a variety of factors, including increased levels of pollution, commercial and recreational fishing, management failure, and environmental fluctuations. Marine reserves have the potential to compensate for failed management and the impacts of fishing by increasing abundance, size, and reproductive output of fished species. In southern California, a group of state and federal agencies, commercial and recreational fishermen, environmentalists, and other community members worked together for two years to establish marine reserves for conservation and fisheries management in the Channel Islands National Marine Sanctuary. An advisory panel of marine scientists used a variety of innovative theoretical and empirical approaches to develop a recommendation for reserve size and location to meet the goals for conservation and fisheries management. A panel of economists and social scientists collected and evaluated data on commercial and recreational fishing effort and catch in the Sanctuary. The agencies, organizations, and individual stakeholders used the ecological recommendations and socioeconomic data to design a network of marine reserve areas for conservation and fisheries management. The Channel Islands marine reserves process provides a model for developing marine conservation policy using ecological and economic principles through cooperation and education of community.

AKÇAKAYA, H. RESIT, Jonathan L. Atwood, David Breininger, Charles T. Collins, and Brean Duncan. Applied Biomathematics, Setauket, NY 11733 (HRA), Antioch New England Graduate School, Keene, NH 03431, USA (JLA), Dynamac International, DYN-2, Kennedy Space Center, FL 32899, USA (DB, BD), California State University, Long Beach, CA, USA (CTC).
VIABILITY OF CALIFORNIA LEAST TERN METAPOPULATION AND EFFECTIVENESS OF MANAGEMENT ACTIONS.
The California least tern is federally listed as an endangered species. Its nesting habitat has been degraded, and many colony sites are vulnerable to predation and human disturbance. Most management efforts have concentrated on predation, an important source of reduced fecundity. We developed a metapopulation model of the California least tern. We defined each cluster of nearby colonies as a population. Within each population, the model included age-structure, year-to-year changes in survival and fecundity, regional "catastrophes" (strong ENSO events), and local catastrophes (reproductive failure due to predation). We simulated the effect of management by increasing fecundity at each population. The model predicted a continuing population increase and a low risk of a substantial decline in the next 50 years. However, this result was sensitive to the assumption about the vital rates. Under the assumption of low vital rates (pessimistic scenario), the model predicted a high risk of decline, although a low risk of extinction. The model results indicated that the effectiveness of management actions depend not only on their local efficiency (increase in fecundity due to protection at each population), but also on the number and spatial distribution of the populations selected for management.

ALIN, SIMONE, Andrew Cohen, and Manuel Palacios-Fest. Department of Geosciences, University of Arizona, Tucson, Arizona 85721, USA.
PALEOECOLOGY, BIODIVERSITY, AND HUMAN IMPACT: OSTRACODS AS INDICATORS IN LAKE TANGANYIKA, EAST AFRICA.
The ostracod fauna of Lake Tanganyika is exceptionally diverse, with approximately two hundred species and high levels of endemism. Ostracods valves are preserved in sediment cores in sufficient abundance to allow high-resolution paleoecological reconstruction of biodiversity turnover through the recent period of intensive land-use change in the lake's watersheds. Ostracods serve as conservative indicators of change in benthic communities, because they have a higher response threshold to sediment inundation than fish and molluscs. Using fossil ostracod assemblages, radiocarbon dating, and detrended correspondence analyses (DCA), we compared recent ostracod diversity trends in sediment cores offshore from two adjacent watersheds in northern Tanzania - one almost completely deforested, the other protected as part of Gombe Stream National Park. In the upper part of the cores from the deforested site, sedimentation rates tripled, and DCA plots and ostracod species abundance profiles revealed marked turnover in the dominant ostracod taxa through time. In contrast, no such changes were observed in the cores from the protected site. The paleoecological record of microinvertebrate diversity at these two sites suggests a recent, anthropogenic transition in the species composition of the benthic community offshore from the deforested watershed.

ALLEN, SARA G. National Park Service, Point Reyes National Seashore, Point Reyes, CA 94956, USA.
SELECTION CRITERIA FOR MARINE PROTECTED AREAS IN THE GULF OF THE FARALLONES, CA, USA.
Marine protected areas (MPA) comprise many forms, ranging from complete to limited restrictions on activities. Scientists and resource managers are closely examining the criteria for MPA designation. In the Gulf of the Farallones, California, several potential sites for marine protected area designation were identified based on simple criteria including ecological, sociological and regulatory components. Ecological criteria included elements such as species rarity and diversity, and source population significance. Sociological criteria took into account elements such as commercial and sport fishing effort. Regulatory criteria took into account jurisdiction, existing designations, and enforcement capabilities. Applying GIS models, we identified MPAs using simple boolean logic with the above criteria. Point Reyes Headland, for example, ranked high for all criteria. Ecologically, the site is significant, exemplifying one of the few locations in the world where major coastal upwelling occurs, and having great diversity and abundance of marine species. Fishing effort is limited because of remoteness and hazardous conditions. Finally, Point Reyes Headland comes under the jurisdiction of several agencies. These simple GIS models, using basic criteria, allow managers to promptly identify MPAs without extensive, long-term research to justify designation.

ALLENDORF, FRED W. Division of Biological Sciences, University of Montana, Missoula, MT 59812, USA.
THE POPULATION GENETICS OF INVASIVE SPECIES.
Genetic and evolutionary processes may be key features in determining whether invasive species establish and spread. Invasive species may evolve both during their initial establishment and during subsequent range expansion, especially in response to selection pressures generated by the novel environment. The rate of change in response to natural selection is proportional to the amount of genetic variation present. If genetic changes and thus evolution during and after colonization are characteristic of invasive species, it will be important to understand the role of genetic diversity during this process. For example, one common feature of invasions is a lag time between initial colonization and the onset of rapid population growth and range expansion. This lag time is often interpreted as an ecological phenomenon (the lag phase in an exponential growth curve). However, lag times are expected if evolutionary change is an important part of the colonization process. This could include the evolution of adaptations to the new habitat, the evolution of invasive life history characteristics, or recovering from the effects of inbreeding associated with the founding event. Population genetic considerations also may be important to effectively control and manage invasive species.

ALLENDORF, TERILYN, Khaing Khaing Swe, Kyaw Kyaw Moe, and Thida Oo. Global Environment Center, US Agency for International Development, Room 3.08, 1300 Pennsylvania Ave., NW, Washington, D.C. 20523, USA (TA), Chatthin Wildlife Sanctuary, Kanbalu Township, Sagaing Division, Myanmar (KKS, KKM, TO).
CREATING A CONSTRUCTIVE RELATIONSHIP BETWEEN LOCAL COMMUNITIES AND MANAGEMENT OF CHATTHIN WILDLIFE SANCTUARY, MYANMAR.
Effective management of a protected area depends on a constructive working relationship between management and residents living around the area. As one of the first steps in building this relationship, we conducted surveys in each of 33 villages surrounding Chatthin Wildlife Sanctuary (CWS), a 103 sq mi. (168 sq km) protected area in central Burma which has one of the few remaining remnants of indaing forest, and is the last stronghold of an endangered species of deer, the thamin (Cervus eldi thamin). The survey was designed to serve as an easy and effective tool to begin communication between management and local communities. The purpose of the survey was to provide baseline data on people's perceptions of CWS in order to inform management strategies. For example, along the southern boundary, crop damage by thamin is a problem for local residents; while along the eastern boundary, lack of access to CWS for the extraction of resources is a problem. We present an analysis of the results and discuss implications for management of CWS.

ALLISON, ALLEN. Hawaii Biological Survey, Bishop Museum, 1525 Bernice Street, Honolulu, HI 96817, USA.
GEOGRAPHIC PATTERNS OF RICHNESS AND ENDEMICITY IN ISLAND FAUNAS.
A worldwide survey of major oceanic islands and island groups was conducted to document patterns of faunal distribution and diversity and to put the Hawaiian Islands into a global perspective. In general, the faunas of oceanic islands differ from continental faunas in at least four significant ways. They are generally impoverished, disharmonic, and have a high percentage of endemic species. Island faunas are also thought to be unsaturated, although this is a matter of some debate. Beyond these four differences, there are a number of ecological and evolutionary trends that further distinguish island faunas from their mainland counterparts. Island species tend to have noticeably larger or smaller body sizes than do their closely related continental relatives. There is also a tendency for groups that can fly, such as birds and insects, to become flightless on islands. And as well there is a trend, most pronounced on isolated islands and archipelagos such as Hawaii, for initial colonizing species to undergo adaptive radiation. Although islands comprise only about 3% of the earth's surface and in general support smaller faunas than do similarly sized mainland areas, because of the high incidence of endemism on islands they harbor a large fraction of the earth's fauna.

ALLNUTT, THOMAS F., G.V.N. Powell, and A. Tye. World Wildlife Fund - U.S., 1250 24^th St. NW, Washington DC 20037, USA, and Charles Darwin Research Station, Puerto Ayora, Isla Santa Cruz, Galápagos, Ecuador.
A HABITAT GAP ANALYSIS OF THE GALÁPAGOS ISLANDS.
Although the Galápagos islands are probably most famous for Darwin's finches, land iguanas, and giant tortoises, the archipelago is also notable for numerous endemic trees, shrubs, herbaceous plants, and invertebrates, many of which are restricted to single islands. The sizable Galápagos National Park covers more than 95% of the islands, providing protection for many of these unique species. A habitat gap analysis, however, shows that the majority of the highland habitats do not receive the same level of protection as do the arid zones that make up the majority of the archipelago. These moist highland environments are among the island's richest habitats for archipelago and island endemics, particularly for plants. A comprehensive plan to protect the biodiversity of the Galápagos needs to consider adequate protection and restoration of these highland communities.

ALMASI, KAMA N., and Peter M. Eldridge. Pacific Coastal Ecology Branch, U.S. Environmental Protection Agency, 2111 S.E. Marine Science Drive, Newport, Oregon, 97365, USA.
A SPATIAL POPULATION SIMULATION MODEL OF A NON-NATIVE SEAGRASS.
Native seagrasses provide essential habitat for many estuarine species. Invasion by an exotic seagrass species with a different morphology may alter the habitat's community structure and thus is an important conservation concern. Zostera japonica (Zosteraceae) is a non-native eelgrass in Pacific Northwest estuaries. To help determine whether this seagrass is detrimental to native estuarine communities, we incorporate a non-linear patch expansion rate into a spatially explicit simulation model. Using field data, we parameterize the model with bathymetry data and size-specific patch expansion rates. We test the effects of competition for space between Z. japonica and native burrowing shrimp and the effects of "bad years", in which climatic and river flow conditions adversely affect Z. japonica survival. Results suggest that Z. japonica will spread, but patch centers are mobile and experience some dieback. The expansion ability of Z. japonica appears especially sensitive to shrimp population dynamics. We conclude that where abundant, burrowing shrimp will likely keep Z. japonica populations in check. In mudflats where burrowing shrimp or organisms with similar functions are not present, Z. japonica may be able to create monocultures across the habitat.

Almasi, Kama, and CHARLOTTE T. LEE. Coastal Ecology Branch, EPA, Newport, OR 97365 (KA), and Graduate Group in Ecology, University of California, Davis, CA 95616, USA (CTL).
MODELING THE SPREAD OF AN INVASIVE PLANT IN NATIVE FORESTS OF HAWAII.
Organisms that appear benign in their home environments may become noxious invaders when introduced to new habitats, particularly on islands. Tibouchina herbacea (Melastomataceae) is a South American plant that has grown from its introduction to the Hawaiian Islands in the 1970s to dense stands in open habitats. Sparser populations are also present in the wet forests of the islands of Hawaii and Maui. We use a spatially explicit, stochastic computer simulation to determine how quickly T. herbacea will be able to spread in a closed forest habitat. We parameterize the model using field data on seed set, seed dispersal, seed germination, seedling survival, and vegetative reproduction. We find that T. herbacea is likely to become a major pest in the closed forest of Hawaii within the next few decades. Because the slowest spread occurs under conditions of maximal canopy cover and maximal aggregation of any small gaps within the canopy, we predict that disturbance to canopy cover is likely to increase the threat of T. herbacea spread, particularly if the disturbance is distributed throughout the forest. We explore the implications of this result for targeting particular life-history stages of T. herbacea in developing management strategies to control its spread.

ALVA, YVETTE R., Susan Masta, and Eric Routman. San Francisco State University Biology Department,1600 Holloway Street, San Francisco, CA 94132, USA.
APPLICATION OF THE NESTED CLADISTIC ANALYSIS METHOD TO THE DISAPPEARING FROG SPECIES, RANA BLAIRI.
Frogs and other amphibians are disappearing from their historically occupied ranges at an alarming rate. Even widespread and abundant frogs, such as species in the Rana pipiens complex, are being affected in at least parts of their ranges. Information about population structure and historical biogeography of species of ranid frogs will be extremely useful for understanding how populations of Rana species will be affected by the spread of pathogens and other factors that may be causing extinction. We are surveying mitochondrial DNA variation in the species Rana blairi, which is undergoing rapid extinction in the southern part of its range (probably due to chytrid fungus infection). Templeton and colleagues have developed a new algorithm, the nested cladistic analysis, that uses molecular sequence data to differentiate several models of population structure from each other and from several historical phenomena. We will apply nested cladistic analysis to mtDNA allele phylogenies estimated from samples from forty populations of R. blairi. We will estimate levels of ongoing gene flow using appropriate models, as well as historically associated genetic subdivisions.

ALVAREZ, L. MARIA. Columbia University MC 5557, 2960 Broadway, New York, New York 10027-6902, USA.
ILLICIT CROPS AND BIRDS OF COLOMBIA.
The forests of Colombia, even in protected areas, are being fragmented for economic exploitation and development including the cultivation of illicit crops: coca, and poppy. Over the last five years illicit crops in Colombia have grown an average of 21% per year, and may account for half the total area deforested in 1998. I present the first geographic analysis of the distribution of illicit crops in relation to areas of conservation priority for threatened and endemic birds. The most important sites affected by illicit crops are the southern Andes, the northern West Andes and adjacent lowlands, the Santa Marta, Perijá, and San Lucas mountains. The largest forested areas threatened by illicit crops are in Amazonia and the Amazonian East Andes, sites of relatively low conservation priority. Given trends in the expansion of illicit crops, and the narrow endemicity of some bird species, the conversion of forests for illicit-crop cultivation may result in several extirpations. Hence conservationists should give high priority to the effective protection against illicit crops of existing and proposed forest reserves and parks. The conservation of threatened and endemic birds in Colombian forests may hinge on successfully curbing incentives for deforestation, including international trade in illicit drugs.

AMATO, GEORGE, and Michael Russello. Wildlife Conservation Society, Science Resource Center, Bronx, NY, Center for Environmental Research and Conservation, Columbia University, New York, NY, USA.
AN INTEGRATED APPROACH TO THE CONSERVATION OF CARRIBEAN PARROTS IN THE GENUS AMAZONA.
Amazon parrots in the Caribbean are threatened by both general factors that impact other psittacines as well as specific factors characteristic of this region. Specifically, severe habitat loss and fragmentation, and continued illegal harvest have impacted these high profile birds. These threats are compounded in a number of species by unusual life history traits coupled with the enormous monetary value placed on them by collectors. Consequently, these circumstances give added priority to ex situ management for this group compared with many other related taxa. Presented here is an integrated approach employing techniques in molecular ecology for in situ studies, conservation genetics for ex situ management, molecular forensics for limiting trade, and strategic linkages with governments and NGOs as a model for conservation of these endangered island endemics.

GIOVANNI, AMORI, and Spartaco Gippoliti. Evolutionary Genetics Centre, CNR, Via dei Sardi 70, 00185 Rome Italy
MAMMAL TAXONOMY AS A FURTHER THREAT TO ISLANDS BIODIVERSITY
As early as 10,000 years ago, humans have played an often unrecognized role in dispersing many mammal species in islands around the world. Early taxonomists described those isolated and sometimes morphologically distinctive mammals as species' and subspecies' which are often included in national and international Red Lists. If the origin of such taxa' is not taken into account in conservation inventories, the risks are: 1) to direct funding opportunities towards populations of artificial origin and of negligible evolutionary importance 2) to overlook the conservation value and status of true palaeoendemic' mammalian taxa 3) to inadvertently protect one of the historic causes of biodiversity loss on islands. In the Mediterranean islands almost all mammal species went extinct following human arrival and new species were introduced. The only palaeoendemic surviving mammals are two shrews of the genus Crocidura which represent the true mammal conservation priorities in the region, often obscured by the greater interest for more appealing and erroneously believed typical' Mediterranean species. As we find similar cases in other parts of the world (i.e., Procyon species considered endemic' of the Bahamas and Lesser Antilles) we suggest that mammalian nominal taxa of proven artificial origin should be listed separately in conservation lists.

AMSBERRY, KELLY, and Robert J. Meinke. Oregon State University, Corvallis, OR 97331, USA.
EFFECTS OF SIMULATED LOGGING ON A THREATENED FERN IN CENTRAL OREGON.
The diminutive, xeric-adapted fern Botrychium pumicola (Oregon Moonwort) is endemic to pumice substrates in central Oregon. Initially found only at high elevations in the Cascade Mountains, this unique pteridophyte has more recently been discovered at lower elevations within commercial stands of lodgepole pine, putting its conservation into potential conflict with timber harvest activities. Our study (now in its third year) investigates some of the effects of timber harvest on B. pumicola, with the goal of providing information to Forest Service land managers for the development of conservation plans for this fern. Individual plants at seven study sites on the Deschutes and Winema National Forests were subjected to one of six treatments selected to simulate potential effects of logging and roadbuilding (burial, compaction, scraping, clipping, shading, as well as untreated controls), and subsequently monitored for emergence, growth and reproduction throughout the following growing seasons. Data from the first two years after treatment indicate that burying dormant plants with soil is detrimental to emergence, while scraping and compaction have negative effects the first year, with the potential for recovery in subsequent years. Clipping and shading had positive effects on growth and emergence, at least in the first year after treatment.

ANDELMAN, SANDY J. National Center for Ecological Analysis and Synthesis, Santa Barbara, California 93101, USA.
SCIENTIFIC STANDARDS FOR ASSESSING SPECIES VIABILITY UNDER THE NATIONAL FOREST MANAGEMENT ACT.
The regulations implementing the National Forest Management Act (NFMA) direct the National Forests to provide habitat that will support viable populations of native and desired non-native vertebrate species well-distributed across National Forest lands. A proposed revision of these regulations would extend this requirement to other plant and animal species. The Forest Service has used a range of approaches to assess species viability, from individual expert opinion to detailed habitat and demographic modeling. Many Forest Plans have been challenged on either the adequacy of the management guidelines for species, or the adequacy of the process used to demonstrate that viability requirements have been met. At the request of the U.S.D.A. Forest Service, NCEAS convened a national panel of 14 scientists to review methods used to assess species viability on National Forests and make recommendations for standards and scientifically defensible but practical methods for evaluating species viability. I discuss the panel's findings and recommendation, as well as potential strategies for implementation.

Anderson, Ileene, MARGOT J. GRISWOLD, Dana Kamada, and ADRIAN WOLF. Earthworks Construction & Design, 2116 Arlington Ave. Ste 301, Los Angeles, CA 90018, USA (IA, MJG), Harmsworth Associates, 36 Bluebird Lane, Aliso Viejo, CA , USA (DK, AW).
COYOTE CANYON LANDFILL: NATIVE VEGETATION RESTORATION RESULTS IN HABITAT CREATION FOR A THREATENED SPECIES.
The Coyote Canyon Landfill was identified as an area for habitat restoration in the Central/Coastal Natural Community Conservation Plan (NCCP) for Orange County, California. Coastal sage scrub vegetation was seeded on 104 acres of the Coyote Canyon landfill in 1994 as part of the mitigation program for the San Joaquin Hills Toll Road. The mitigation project required initial coordination to develop specifications for the landfill closure plan to accommodate planting the coastal sage scrub community. Avian surveys of the restoration site demonstrated colonization by the federally listed California gnatcatcher (Polioptila californica californica) showing an annual two-fold increase over the site as the vegetation developed. Nest monitoring and banding studies documented the breeding pairs of gnatcatchers from one pair in 1996 to three in 1997, seven in 1998, and fifteen in 1999. Overall, gnatcatcher breeding success was relatively high on the landfill. Low nest failure rates, perhaps due to limited predator populations on the landfill, appear to be responsible for the high nesting success. The Coyote Canyon revegetation site shows promise as a future source population site for the California gnatcatcher, thus fulfilling a policy of the NCCP governing habitat restoration to establish a defined historic ecosystem.

Anderson, Stephen J., and JEREMY I. GOODING. National Park Service, Biological Resources Management Division, Haleakala National Park, P.O. Box 369, Makawao, HI 96768, USA.
PROACTIVE ALIEN PLANT MANAGEMENT PROTECTS HAWAII PARKS: PARTNERS BATTLE INVASIONS ACROSS PARK BOUNDARIES.
Hawaii parks are making major progress in stemming the tide of alien plant invasions threatening Hawaiian ecosystems by fielding an expert weed control team that identifies and eliminates weed threats on a multi-island basis. The National Park Service Pacific Islands Exotic Plant Management Team (PIEPMT) travels between six parks controlling incipient invasions to prevent spread into the national parks. The PIEPMT utilizes the local knowledge of existing crews and integrated weed management techniques, to monitor and battle alien plant invasions in the parks or preferentially before they impact the parks. PIEPMT efforts are a core component of a multi-agency and private industry effort to control the establishment and spread of invasive species in Hawaii in the form of Island Invasive Species Committees (ISCs). Currently there are three ISCs working to prevent and control invasions on Maui Nui, Hawaii and Oahu.

Antunes, Agostinho, Rui Faria, Steven Weiss, and Paulo Alexandrino. Centro de Estudos de Ciência Animal (CECA), ICETA-U.P., Campus Agrário de Vairão, 4485-661 Vairão (AA, RF, SW, PA), Departamento de Zoologia-Antropologia, Faculdade de Ciências, Universidade do Porto, Praça Gomes Teixeira, 4099-002 Porto, Portugal (AA, PA).
Complex evolutionary history in the brown trout: insights on the recognition of conservation units.
Sixteen populations of brown trout Salmo trutta L. across 8 Portuguese river basins were screened for variation at 5 diagnostic loci (mtDNA and allozymes). Results revealed low to non-existent levels of introgression from hatchery reared fish of allochthonous origin. Population structure based on mtDNA revealed a mosaic pattern driven by past fragmentation and restricted gene flow with little correspondence to major river drainages or recently proposed OCUs on the Iberian Peninsula. Such patterns of variation offer a challenge to conservation strategies which base themselves on defining units of conservation, particularly if such units intend to reflect an hierarchical evolutionary structure. We suggest that geographically mosaic patterns of evolutionary lineages, as well as adaptively significant traits are common characteristics of many freshwater fishes. Thus, large-scale units, even if diagnosed by mtDNA clades, are often too heterogeneous to consider a "unit" for conservation. Alternatively, a bottom-up perspective which prioritizes unique populations is both more practical and more efficient in recognizing and preserving evolutionary diversity.

ARANGO-CARO, SANDRA. Department of Biology, University of Missouri-St.Louis, 8001 Natural Bridge Road, St. Louis, MO 63121, USA.
EDGE IMPACT ON TREE COMPOSITION IN REMNANT ANDEAN FOREST FRAGMENTS (COLOMBIA): LANDSCAPE AND PATCH LEVEL.
The harsh contrast between forest and pasturelands, creates an inbalance in abiotic conditions at edges when compared to the relatively stable forest interior, affecting tree composition. At five sites, abiotic variables, vegetation structure and tree composition (3 size classes), were recorded from the edge towards the forest (50m) every five meters. Decreases in temperature and increases in humidity (air and soil), were found from the edge towards the forest interior. Understory foliage density was higher at the edges, while overstory foliage was higher in forest interior. Richness and densities of trees among the sites differed in their responses to edges. Direct gradient analyses at the landscape level (all sites) did not show a response of tree composition as a function of distance, but there was a significant relationship with the abiotic environment. At the patch level (by site), tree composition responded to the edges (threshold 20-35m), but a significant relationship was not found with the abiotic environment, although it increased with size class increase. This suggests that generalizations of edge-effects at the landscape level are difficult to make due to site variations. At the patch level, tree composition responded differently to abiotic and biotic edge-effects depending on size class.

ARMSTRONG, DOUG P. Wildlife Ecology Group, Institute of Natural Resources, Massey University, Private Bag 11222, Palmerston North, New Zealand.
TWO PARADIGMS FOR UNDERSTANDING BROAD-SCALE EXTINCTION PROCESSES.
In 1994, Graeme Caughley noted that two distinct paradigms had emerged for understanding (and preventing) extinctions of single populations: the small population paradigm and the declining population paradigm. He argued that conservation biology would proceed most effectively if this dichotomy were broken down. Over the last decade the main focus in conservation biology literature has not been on single populations, but on broad-scale extinction processes affecting multiple populations. Examination of this literature shows that two paradigms have also emerged at this broader spatial scale: the "metapopulation paradigm" and the "habitat paradigm". The metapopulation paradigm assumes that suitable patches may be occupied or unoccupied, and that occupancy is strongly affected by stochastic extinction and recolonization processes. The habitat paradigm assumes that distributions are determined solely by habitat quality. Predictive models can be rapidly developed using either paradigm. However, the predictions may be misleading because metapopulation variables (size and isolation) may be confounded with key habitat variables, which may not be easily measured. Management recommendations therefore depend on the paradigm adopted, the metapopulation paradigm emphasizing translocations and corridors, and the habitat paradigm emphasizing restoration. I argue that effectively synthesizing these paradigms requires research on local dynamics of multiple populations.

ARNDT, ALLAN D., Vitor Cantarelli, and Jack W. Sites, Jr. Brandon University, Brandon, MB, CANADA, R7A 6A9 (ADA), CENAQUA, Goiania, Goias, BRAZIL (VC), Brigham Young University, Provo, UT, USA (JWS).
CAN POPULATION GENETICS FILL AN ECOLOGICAL VACUUM?
The Giant Amazon River turtle, Podocnemis expansa, is found in rivers throughout northern South America. In large part due to its' colonial nesting behaviour, this endangered species has been heavily exploited in the past. In response to this situation, the Brazilian government created the National Centre for the Amazonian Chelonia, CENAQUA, to co-ordinate conservation efforts. One of the difficulties CENAQUA faces is that only limited information is available on many aspects of the natural history of this species, including migration ranges, nest-site fidelity, and levels of paternity. A collaborative study of the genetic diversity and structuring of populations across the greater Amazon basin was thus undertaken. This information will be used to improve the existing conservation management plan by 1) identifying unique genetic lineages and important new areas for protection, 2) testing the assumptions of colonial nesting behaviour, 3) providing genetic markers for monitoring breeding programs, and 4) evaluating the possible genetic basis of observed morphological/behavioural differences among populations. Preliminary evidence suggests a surprisingly low level of mitochondrial DNA diversity, although significant geographic structuring among haplotypes does exist. Future efforts will include analysis with nuclear markers and a comparison of these results with simulation studies.

ARRIOLA, LAURA, and Curt Kessler. Commonwealth of the Northern Mariana Islands, Division of Fish and Wildlife, PO Box 10007, Lower Base, Saipan, MP 96950 (LA), Zoology Unlimited, LLC P.O. Box 31132, Flagstaff, AZ, 86003, USA (CK).
SARIGAN: HEINOUS TO HOPEFUL: VEGETATION RECOVERY AFTER FERAL ANIMAL REMOVAL ON AN UNINHABITED ISLAND IN THE NORTHERN MARIANA ISLANDS.
Sarigan is a 500 ha uninhabited island in the Northern Mariana archipelago which has been populated by feral pigs and goats for approximately 50 years. Sarigan supports several native wildlife species including the endangered Micronesian megapode (Megapodius laperouse) whose habitat was severely degraded. In 1998 feral pigs and goats were eradicated from the island. Permanent vegetation plots were established and initially surveyed prior to eradication of feral animals. The objective was to measure floral recovery as feral animal habitat degradation was mitigated. Since eradication the number of plant species in permanent plots has doubled, the canopy and ground cover has increased and in 2000 there was a significant increase in the number of seedlings present in the plots. There has also been an increase in the invasive vine, Operculina ventricosa. Most forest edges and seedlings in open fields are covered with dense mats of this vine. It remains to be seen if this weedy vine will be a limiting factor to forest recovery.

ATHENS, J. STEPHEN, and Jerome V. Ward. International Archaeological Research Institute, 2081 Young St., Honolulu, HI 96826, USA (JSA); P.O. Box 14, Sweet Briar, VA 24595, USA (JVW).
LOWLAND FOREST LOSS AND PREHISTORY IN HAWAII: ARE HUMANS TO BLAME?
Pollen data from four wetland pollen cores from different lowland areas of Oahu are compiled to document lowland vegetation prior to the islands colonization by humans. Such wetland cores provide one of the few direct sources of information on the nature of the pristine (pre-human) late Holocene vegetation of Hawaii. The advent of prehistoric Polynesians is clearly evident in these cores with the appearance of charcoal particles, taxa indicative of landscape disturbance, introduced taxa, and the decline of forest taxa. The rapidity and extensiveness of forest loss with the advent of humans suggest that direct impacts as a result of human activities such as burning and gardening was minimal and do not account for the observed vegetation transformation. Several different lines of circumstantial evidence point to the introduced Polynesian rat, Rattus exulans, as a significant factor in the ecological collapse of the lowland forests. The implications of paleoenvironmental findings for conservation biology in Hawaii are apparent, including the provision of baseline information on the nature Hawaii's pre-human lowland vegetation communities, the significance of certain species in these communities that are now extremely rare, and the probable importance of excluding rats for propagating native taxa in areas undergoing habitat restoration.

ATKINSON, CARTER T., Dennis A. LaPointe, and Susan I. Jarvi. USGS Pacific Island Ecosystem Research Center, Hawaii National Park, HI 96718, USA (CTA, DAL), Department of Biology, University of Hawaii, Hilo, HI 96722, USA (SIJ).
THE ROLE OF INTRODUCED PATHOGENS AND VECTORS ON THE DEMISE OF AN ENDEMIC AVIFAUNA: LESSONS FROM THE HAWAIIAN ISLANDS.
The Hawaiian Islands are home to a critically endangered avifauna that is often cited as one of the most outstanding examples of adaptive radiation and speciation in the world. The introduction of mosquitoes (Culex quinquefasciatus) and both avian pox (Poxvirus avium) and malaria (Plasmodium relictum) to the islands has had an enormous impact on native forest bird communities and is believed to be one of the primary factors responsible for their collapse at lower elevations. We have documented the impacts of these diseases through studies of their distribution, pathogenicity, vector ecology, and epidemiology. Native Hawaiian honeycreepers are extremely susceptible to malaria with mortalities in some species as high as 90% following exposure to a single infective mosquito bite. Epidemic outbreaks are seasonal, dependent on vector populations, and may be driven by environmental factors, host immunity, and/or cycling of new parasite strains in the host population. Restoration of Hawaiian forest bird communities will depend on aggressive and creative management strategies that incorporate detailed knowledge about the dynamics of this disease system across large landscapes.

AUGERI, DAVID M., and David Bourn. Wildlife Research Group, Dept. of Anatomy, University of Cambridge, Cambridge CB2 3DY, England (DMA), Environmental Research Group Oxford Ltd, P.O. Box 346, Oxford OX1 3QE, England (DB).
INSULAR ISLAND EFFECTS ON THE ALDABRAN GIANT TORTOISE (GEOCHELONE GIGANTIA) POPULATION.
By 1900, tortoises were extirpated from all Indian Ocean islands, except Aldabra Atoll, where "a few" survived. By 1973-74, Aldabra's population recovered to 129,000 individuals. This 1997 revised estimate was based on the first Aldabran tortoise census in 1973-74, which used a 5% (292 ha) stratified random sample of Aldabra's land area. A 1997 replicate census using a 55.5% random sample (162 ha) of the 1973-74 plots revealed a decline (p= <0.001) to 100,000 individuals (Bourn et al. 1999). Evidence are: 1) declines (p= <0.001) in censused live tortoises from a) 1997 plots compared to 1973-74 plots and b) long-term monitoring transects; and 2) an increase in tortoise mortality (p= <0.001). Aldabra's largest island maintains the highest tortoise density yet smallest-sized individuals; however, its subpopulation declined by 33%-41% (p= <0.001). Low-density subpopulations increased by 60%-93% (p=0.0001) on Aldabra's middle-sized islands and maintain the largest individuals, yet represent <5% of Aldabra's population. Insular effects on density-dependent processes of a fragmented island population are attributed. These population changes present island colonization and evolutionary implications. Change directions and rates are associated with pressures on individual island's carrying capacities. Decreased forage productivity and accessibility and tortoise reproduction and growth constraints exacerbated these effects.

BABCOCK, ELIZABETH A., Murdoch K. McAllister, and Ellen K. Pikitch. Wildlife Conservation Society, Marine Conservation Program. 2300 Southern Blvd. Bronx, NY 10460, USA (EAB, EKP), Renewable Resources Assessment Group, Centre for Environmental Technology, T H. Huxley School of Environment, Earth Sciences & Engineering, Imperial College, Room 402 Royal School of Mines, Prince Consort Road, London SW7 2BP, England (MKM).
DECISON ANALYSIS IN FISHERIES STOCK ASSESSMENT: WESTERN ATLANTIC BLUEFIN TUNA
Western Atlantic Bluefin tuna (Thunnus thynnus thynnus) are subject to commercial and recreational fisheries, under the management of the International Commission for the Conservation of Atlantic Tuna (ICCAT). The population is depleted to less than 20% of historic levels, and the current harvest quota may not allow the population to rebuild. ICCAT's stock assessments of bluefin tuna have been controversial in the past, in part because of uncertainty about the relationship between spawning stock size and recruitment of juvenile fish into the fishery. Furthermore, the management actions needed to rebuild the population to its maximum sustainable yield level depend strongly on the assumptions that are made about the stock recruitment relationship (SRR). At the 2000 bluefin tuna assessment, we presented a Bayesian method to estimate the probabilities of various functional forms of an SRR given a time series of stock and recruitment data. Without using a Bayesian method, the assessment was restructured so that the uncertainty in the SRR was addressed, thus improving the credibility and usefulness of the assessment. We also discuss the role conservation biologists can play in improving the scientific basis of fisheries stock assessment and management.

BACKLIN, ADAM, Robb Hirsch, Chris Brown, and Robert N. Fisher. USGS, Department of Biology, San Diego State University, San Diego, CA 92182-1015, USA.
CURRENT STATUS OF MOUNTAIN YELLOW-LEGGED FROGS (RANA MUSCOSA) IN SOUTHERN CALIFORNIA.
Populations of mountain yellow-legged frogs (Rana muscosa) continue to decline in well-protected areas in southern California. Monitoring of known populations and presence/absence surveys were conducted in the Angeles National Forest (ANF) and the San Bernardino National Forest (SBNF) in the summer and fall of 2000. Most techniques and locations were a continuation of work by others that extend back to 1993. Currently, there are less than three known remaining populations in the SBNF and less than six known populations in the ANF. During our surveys, reproductive success was detected from only two of these sites. Similar declines are being reported from the Sierra Nevada populations of this species, although the habitat requirements of the frogs there appear much different. No chytrid fungus or iridoviruses were detected during any survey. Introduced trout appear to be the primary threat and removal experiments at some locations should begin immediately in order to restore historic habitat. Human recreational activities, including hiking, fishing, and bathing appear to negatively impact mountain yellow-legged frog populations. Continued monitoring of known populations, in addition to further searches of historical localities, is imperative for developing a conservation strategy for this very endangered species.

BAEZ, SELENE, and Henrik Balslev. Department of Systematic Botany, Institute of Biological Sciences, University of Aarhus, Nordlandsvej 68, DK-8240 Risskov, Denmark.
PALM COMMUNITY RESPONSES TO EDGE INDUCED ENVIRONMENTAL VARIATION IN FOREST FRAGMENTS IN ECUADOR.
Edge effects on palm communities were evaluated in two 40-50 years old forest fragments. In each forest fragment three edges were examined through 9 10x20m plots placed at 0 (edge), 50 (middle), 100 (interior) m from the forest/no-forest border. All the individuals of 12 palm species were recorded and identified to species level. The environmental variables measured related with forest structure (number of trees 30 cm dbh, forest-phase), understory light, and undergrowth cover type were influenced by the edge closeness. Results show that the palm diversity and abundance are strongly influenced by the edge closeness. Edge plots have the lowest diversity levels of adult palms, while interior plots present the highest diversity. Palm tree density consistently diminished or disapeared from the edge plots. Understory palms may persist along forest edges if the forest structure is mantained. Low palm recrutiment along edges show that the core area' is not expanding. Edge effects for the palm community penetrates at least 100 m into the forest fragments.

BAILEY, SALLIE A., and Carol L. Boggs. Center for Conservation Biology, Department of Biological Sciences, Stanford University, CA 94305, USA.
EFFECT OF AN INVASIVE SPECIES ON THE POPULATION DYNAMICS OF A NATIVE BUTTERFLY (PIERIS NAPI).
Invasive species can disrupt co-evolved relationships in many ways. We are studying the effects of an invasive exotic weed on the population dynamics of a native butterfly. The invasive species, Thalaspi arvense spreads rapidly by invading disturbed areas and fulfills the same role as its native relatives in the life cycle of the butterfly Pieris napi, by acting as a larval host, but with one difference. Early instar larvae that feed on T. arvense die. We have collected field data on distribution of native and exotic host plants, habitat use and movement of the butterfly species, and the proportion of eggs laid on the exotic versus native hosts. Our previous work has shown that there is variation in oviposition preference of P.napi between native hosts and T. arvense which is heritable in a sex linked manner and depends on the length of exposure to plant species. We have used this information to construct a spatially explicit population model simulating the effects of the distribution of the invasive species on the population dynamics of P. napi. This model is of value to other systems where invasives are a threat or other factors are reducing a species' habitat quality.

BAKER, JASON D., Thea C. Johanos, and George A. Antonelis. National Marine Fisheries Service Honolulu Laboratory, 2570 Dole Street, Honolulu, HI 96822, USA.
Population Dynamics and Status of the Endangered Hawaiian Monk Seal.
Hawaiian monk seal counts have declined by 60% since the late 1950s and current abundance is estimated at 1300-1400 seals. Counts declined about 5% yr^-1 during 1985-1993, and subsequently remained stable through 1999. Population trends are influenced by the highly variable dynamics of the six main reproductive subpopulations in the Northwest Hawaiian Islands (NWHI). Overall pup production has increased in recent years, but juvenile mortality also increased in varying degrees at several subpopulations. The largest subpopulation is at French Frigate Shoals, where counts of non-pups have dropped by 60% since 1989, and the age distribution has become severely inverted due to high juvenile mortality. Laysan and Lisianski subpopulations have remained low but relatively stable. In contrast, the population at Pearl and Hermes Reef has increased 6-7% yr^-1 since the mid-1970s. Similarly, the population at Kure Atoll has consistently grown since the mid-1980s, after several management actions. Finally, the small population at Midway Atoll, which was nearly extinct, is showing signs of recovery due largely to immigration from Pearl and Hermes Reef and Kure Atoll. Future abundance trends for the entire species will likely depend upon whether predicted losses at French Frigate Shoals are countered by gains at other sites.

BALDWIN, ROBERT F., and Aram J.K. Calhoun. Deering Hall, University of Maine, Orono, Maine 04469, USA.
EFFECTIVE SCIENCE COMMUNICATION: A MAINE RESPONSE TO CONSERVATION BIOLOGY IN PRACTICE.
The goal of Conservation Biology in Practice is to "fill the gap between academic and practicing conservation biologists to put conservation science into practice and conservation practice into science." We evaluated the effectiveness of CBiP by asking 72 Maine environmental professionals (from state and federal government, non-profits, schools, and private industry) to review the first issue and respond to a survey. A subset (7) will be interviewed during a focus group discussion. Preliminary survey results (12) indicate approval for the new journal, and that it will be a useful tool. Several respondents not already SCB members indicated they would subscribe to CBiP. Although all survey respondents have marked "conserving biological diversity" as of primary importance to their work, even those with graduate degrees in science have rated conducting scientific research as less important to their work. Focus group discussion will focus on broader issues of how science is best presented to the public. will present full results, including survey responses and excerpts from focus group discussion.

BALLOU, JONATHAN D., and R. C. Lacy. Department of Conservation Biology, Smithsonian National Zoological Park, Washington, D.C. 20008, USA (JDB) and Department of Conservation Biology, Chicago Zoological Society, Brookfield, IL 60513, USA (RCL).
EVALUATING GENETIC MANAGEMENT STRATEGIES IN CAPTIVE BREEDING PROGRAMS USING VORTEX.
Genetic management strategies for captive populations have evolved over the last 20 years from simply avoiding inbreeding to the currently recommended strategy of minimizing kinship. This means that kinship strategy provides a theoretically robust and simple way to identify the genetically most valuable animals to breed to maintain gene diversity. Four factors are considered when selecting pairs for breeding: the pair's average mean kinship, the difference in the pair's mean kinship, the inbreeding coefficient of potential offspring, and the extent of uncertainty in the pair's ancestry. Optimal pairs are those with minimum values for all criteria. However, few pairs are optimal for all criteria and little is understood about the trade-offs in prioritizing one criterion over another. For example, how important is it to select pairs with similar mean kinship values and what are the cost in terms of long-term maintenance of genetic diversity if this criterion is relaxed? Furthermore, alternative breeding strategies have been recommended. The kinship value strategy weights mean kinship values by reproductive values to take into consideration an individual's expected future reproductive performance. A modified version of the VORTEX population modeling program is used to assess the relative performance of different strategies for managing breeding programs.

BALMFORD, ANDREW, Joslin Moore, Thomas Brooks, Neil Burgess, Louis A. Hansen, Paul Williams, and Carsten Rahbek. Conservation Biology Group, Department of Zoology, University of Cambridge, Downing Street, Cambridge, CB2 3EJ, UK (AB, JM, TB), Zoological Museum, University of Copenhagen, Universitetsparken 15, DK-2100, Copenhagen Ø, Denmark (JM, TB, LAH, CR), Center for Applied Biodiversity Science, Conservation International, 2501 M Street NW, Suite 200, Washington, D.C. 20037, USA (TB), Wildlife Conservation Society of Tanzania, Pamba House, P.O. Box 1668, Morogoro, Tanzania (NB), Biogeography and Conservation Laboratory, The Natural History Museum, Cromwell Road, London, SW7 5BD, UK (PW).
CONSERVATION CONFLICTS ACROSS AFRICA.
Recent studies have provided disturbing evidence that areas of outstanding conservation importance may coincide with dense human settlement or impact. We test the generality of these findings using 1^o -resolution data for sub-Saharan Africa. We find that human population density is positively correlated with species richness of birds, mammals, snakes and amphibians. This association holds for widespread, narrowly endemic and threatened species, and looks set to persist in the face of foreseeable population growth. Our results contradict earlier expectations of low conflict based on the idea that species richness decreases and human impact increases with primary productivity. We find that across Africa, both variables instead exhibit unimodal relationships with productivity. Finally, new prioritisation algorithms show that resulting conservation conflicts are not easily avoided at this scale, because many densely inhabited grid cells contain species found nowhere else.

BANKO, PAUL C., Steve Dougill, Daniel Goltz, Luanne Johnson, and Pete Oboyski. Pacific Island Ecosystems Research Center, U. S. Geological Survey - Biological Resources Division, Kilauea Field Station, P.O. Box 44, Hawaii National Park, HI 96718, USA.
HAWAIIAN BIRD RECOVERY: REMOVING INVASIVE THREATS, RESTORING HABITAT, AND REESTABLISHING POPULATIONS.
Recovering palila (Loxioides bailleui) and their dry forest habitat requires managing a broad assortment of invasive species. Palila are threatened by predators and food competitors, and their habitat is threatened by alien species that browse native vegetation and disturb the soil, increase fire fuel levels, and suppress forest regeneration. These and other factors have concentrated the majority of palila in only 30 km² of subalpine habitat on western Mauna Kea. This population may be responding slowly to habitat improvements resulting from the reduction of feral ungulates but is vulnerable to fire and other catastrophes; thus, viable populations must be reestablished elsewhere. Attempts to reintroduce palila to northern Mauna Kea by translocating wild birds from the western slope have not yet reestablished a resident population. However, some birds remain for over a year before returning home, encouraging us to continue developing this approach. Additionally, reforestation of downslope pastureland will increase the elevation gradient along which the availability of critical foods, including seeds and caterpillars, varies through the year. Recovery also must include protecting palila from feral cats and rats; preventing the spread of alien grasses, vines, and shrubs; controlling insect predators and parasites that impact food resources, and managing fire.

BARNES, MYRA E., and Joel Berger. Program in Ecology, Evolution, and Conservation Biology, University of Nevada, Reno, NV 89557 (MEB, JB), Carlsbad Caverns National Park, Carlsbad, NM 88220, USA (MEB).
IS CONSERVATION IN PROTECTED AREAS BASED ON AESTHETIC OR ECOLOGICAL CAPACITY? ELEPHANTS IN BOTSWANA.
Large populations of browsing elephants Loxodonta africana alter vegetation composition in riparian savanna woodlands in northern Botswana. During the dry season, elephant densities of 7-12/km² have been recorded along the Chobe and Linyanti Rivers. When natural processes are allowed to regulate populations, the increasing abundance of one species may reduce the relative abundance of other species. Before elephants reach an ecological carrying capacity in northern Botswana, vegetation changes may alter the aesthetic appearance of riparian woodlands and influence the abundance of other species, such as endemic Chobe bushbuck Tragelaphus scriptus ornatus. Chobe National Park and Moremi Game Reserve are often viewed as pristine wilderness areas. However, a review of historical events shows that in addition to changes in rainfall, surface water distribution, and browsing wildlife, human caused factors such as elephant hunting, livestock grazing, frequent fires, rinderpest, and logging influenced vegetation composition and appearance over the past 150 years. Management alternatives that simultaneously address natural regulation of ecosystem processes, preserving biodiversity, and aesthetic values may not be possible. Aesthetic values may influence conservation decisions; however, understanding factors responsible for ecological change over broad periods of time is critical for the long-term protection of conservation areas.

BARRACLOUGH, ROSEMARY K., and Michael N. Clout. School of Environmental and Marine Sciences, The University of Auckland, Private Bag 92019, Auckland, New Zealand.
THE EFFECTS OF TEMPORAL AND SPATIAL SCALES ON ECOLOGICAL RESTORATION MONITORING: A NORTHERN TE UREWERA, NEW ZEALAND, EXAMPLE.
The northern Te Urewera ecosystem restoration project is an ambitious management initiative restoring 50,000 hectares of forest in New Zealand's North Island. The project area includes the entire northern sector of one of the largest forest-fragments remaining in the North Island today, Te Urewera National Park. The restoration work is primarily oriented towards invasive mammalian pest control. The objective of this study was to determine the success of this initiative in achieving ecosystem recovery, through comparison of a core management region with non-treatment areas. However, defining an ultimate goal for such a restoration project is difficult, due to the absence of pristine mainland forests to provide a benchmark. Therefore outcome monitoring, for gauging success, focussed on characteristics of community ecology. It was found that spatial and temporal scales in monitoring had a dramatic influence on whether detectable differences can be identified. This is illustrated by results from: distance-sampling estimates of forest-bird densities; ground-transect collections showing differential predation and predator disturbance of fruit; canopy-surveys revealing differential fruiting and flowering; and seedling-plots. It is clear that spatial and temporal monitoring scales dictated by non-biological considerations can potentially lead to misleading or unclear information, and this has important implications for restoration management.

BARROWS, MORGAN E. University of California, Irvine, 248 Social Ecology I, Irvine, CA 92697-7070, USA.
NATURAL COMMUNITY CONSERVATION PLANNING ACT: UNIQUE APPROACH BUT IS IT WORKING?
This study carries forward the exploration of whether or not the Natural Community Conservation Planning Act of 1991 is successful at preserving biodiversity while allowing development to occur in the Coastal Sage Scrub habitat. This study examines the amount and quality of habitat preserved, a population count of the three target species (California gnatcatcher, cactus wren, and orange-throated whip-tail lizard), and a count of the number of Coastal Sage Scrub species that have been listed on the state and/or federal endangered species list since 1991. A series of comprehensive interviews will also be conducted. The subjects include land developers, public officials, scientists, and environmentalists who have actively participated in the development and/or implementation of the NCCP throughout the five Southern Californian counties: Orange; San Diego; Riverside; San Bernardino; and, Los Angeles. This study will conclude if the Act is successful, if it has deficiencies, and whether or not this Act should be implemented throughout the United States.

BASS, DAVID. University of Central Oklahoma, Edmond, OK 73034, USA.
COMPARISON OF MACROINVERTEBRATE COMMUNITIES IN STREAMS OF THE LESSER ANTILLES.
An ongoing survey of macroinvertebrates inhabiting the freshwater habitats islands composing the Lesser Antilles was initiated in September 1995. Islands sampled included Barbados, Tobago, Grenada, St. Lucia, Dominica, Montserrat, Nevis, St. Kitts, Saba, and Antigua. Qualitative collections were made by sweeping a dip net through the water column and by hand examination of rocks, plants, and debris submerged in both flowing and standing bodies of freshwater across the each island. In addition, water temperature was recorded at each site. Collections by previous investigators were also included. My collections have yielded over 250 species thus far, many of which are reported for the first time from these islands. Dominant taxa collected included several species of gastropods, decapod crustaceans, ephemeropterans, odonates, hemipterans, and coleopterans. Many of these taxa have fairly widespread distributions throughout the Lesser Antilles. Observations indicate that most stream species are associated with leaf packs and most pond species are associated with aquatic macrophytes. Generally the macroinvertebrate fauna of these islands is sparse, most likely due to the oceanic origins, small size, and human disturbance of freshwater environments of the islands.

BEEVER, E.A., and David A. Pyke. USGS-BRD Forest and Rangeland Ecosystem Science Center, 3200 SW Jefferson Way, Corvallis, OR 97331, USA.
A RESEARCH PROGRAM FOR RANGELANDS OF THE INTERIOR COLUMBIA BASIN AND SNAKE RIVER PLATEAU
The Interior Columbia Basin ecosystem (i.e., Columbia River, Klamath, Great, and Snake River Basins) encompasses portions of seven states and contains areas that possess some of the highest ecological integrity in the U.S. Here we report development of an integrative Plan to address priority research needs to support management of over 11 million ha of land administered by five Department of Interior agencies. In addition to reviewing assessments, contract reports, and EISs for the Basin, we have canvassed primary literature, agency and University researchers, and local resource managers to identify critical issues. Scientists and managers in 1999 agreed upon five research emphases: 1) restoration; 2) rangeland health; 3) aquatic-terrestrial interface; 4) development of monitoring and evaluation protocols; and 5) at-risk species and habitats. Issues include: exotic species; microbiotic soil crusts; fragmentation of shrub-steppe habitats; management in the face of anthropogenic disturbance, catastrophe, and uncertainty; restoring historic disturbance regimes (e.g., fire); riparian and wetland management; livestock grazing; multiple-scale monitoring; salmonids; and sagebrush-obligate species (e.g., sage grouse). The Plan supports a broad-scale approach to restore and maintain ecosystem health and ecological integrity through active, adaptive management of habitat and disturbance processes, framing local decisions within a consistent region-wide management direction.

BEIER, PAUL, Maryann van Drielen, and Bright O. Kankam. School of Forestry, Northern Arizona University, Flagstaff AZ 86011-5018, USA (PB), 4010 Lugano Way, Flagstaff AZ 86004, USA (MVD) and Forest Research Institute of Ghana, UST Box 63, Kumasi, Ghana.
AVifaunal collapse in West African forest fragments.
West African tropical forests are the most-fragmented and least-studied in Africa. To determine how forest birds respond to 5 factors (patch size, patch isolation, canopy density, abundance of large trees, proximity to forest edge), we surveyed 61 species of forest birds on 121 transects in 35 forest fragments in Ghana's semi-deciduous forests. Twenty-two species of forest birds were area-sensitive (15 were never found in small patches), and 9 were edge-sensitive, but patch isolation and forest structure had little influence on which species occurred in forest patches. Species richness per transect increased with patch size over the entire range of patch sizes observed (3 to 33,000 ha). The total diversity of forest birds on 18 transects in 13 small patches (25 species) was comparable to the diversity on a single transect in a large forest patch (maximum 22 species). Clearly, only large forests will conserve many of these forest birds. Edge-sensitivity was not due to altered forest structure near patch edges. Species were unlikely to occur in patches smaller than ~4 home range areas. Diversity of forest birds did not vary with isolation (distance to large forest), suggesting that island biogeographic mechanisms had little influence on birds.

BELFIORE, NATALIA M., and Bernie May. Department of Animal Science, Meyer Hall, 1 Shields Ave. University of California, Davis, California, 95616, USA.
EFFECTS OF CONTAMINANTS ON THE GENOMES OF RED SWAMP CRAYFISH IN THE SACRAMENTO VALLEY OF CALIFORNIA.
We test whether population genetic patterns can be used by conservation biologists to understand effects of environmental stress, namely contaminants, on populations. Three mechanisms are hypothesized by which genetic shifts may occur. 1. Selection on resistant alleles can cause allele fixation at one or multiple loci, or reduction in genetic polymorphism. 2. Drift from population reductions or reduced reproductive success can cause genetic bottlenecks and reduced gene diversity. 3. Direct mutagenesis or the cumulative effects of repeated genotoxic damage can cause increased mutational load indicated by rare alleles. Genetic variation in red swamp crayfish is assessed at twelve microsatellite loci. Replicate populations exposed to multiple contaminants are compared to replicate reference populations. Stochastic processes are factored out to assess genetic patterns attributable to chronic contaminant exposure. All populations show significant population structure (F_ST). Inbreeding coefficients (F_IS) are higher in all contaminated populations than in reference populations. Samples from the most heavily exposed population show less than half the allelic diversity than those from the cleanest population. Genetic patterns, therefore, conform to the predictions of selection or bottleneck processes, and correlate with degree of exposure. Multiple biomarker assessments are used as supportive evidence to associate genetic patterns with contaminant exposure.

BELLINGHAM, MARK. School of Planning and Global Studies, Massey University, Private Bag 11 222, Palmerston North, New Zealand.
The importance of edge habitat for alien and indigenous bird species in New Zealand's indigenous forest fragments.
Indigenous forest fragments are often all that remains of New Zealand's indigenous biodiversity in the extensive agricultural matrix of the lowlands. Biodiversity planning issues for these forest remnants include understanding changes in bird abundance with forest remnant size and habitat quality, and the minimum size of forest remnant necessary to sustain indigenous bird species. This study sampled forest vegetation and bird abundance in isolated indigenous forest patches in across 2000 square kilometres of the Rodney District, north of Auckland. It revealed a decline in alien bird densities away from forest edges and an increase in indigenous bird densities towards the interior of forest patches. The change in bird densities reflected changes in forest vegetation composition between the edge and interior of forest habitat islands. With increasing patch size the density of alien bird species decreased and the density of indigenous bird species increased. The edge effect explained some of these changes in bird density. Minimum patch sizes were established during the breeding season for four indigenous and three exotic bird species. This study may provide some answers for planners trying to identify how and where to protect and restore landscape elements to maintain functioning lowland forests.

BENDER, DAVID W. National Tropical Botanical Garden, Limahuli Garden, PO Box 808, Hanalei, HI, 96714, USA.
LOWLAND FOREST RESTORATION IN HAWAII: INVESTIGATING STRATEGIES FOR SEVERELY DEGRADED AREAS.
Restoring severely degraded lowland forests in Hawaii is recognized as one of the greatest challenges to Hawaiian conservation biologists. The National Tropical Botanical Garden (NTBG) manages Limahuli Valley, consisting of 1,000 acres of tropical wet forest that provides habitat for at least 12 endangered plant species. The research presented is a result of a collaborative effort between the NTBG and U.S. Forest Service. The experiment investigates different combinations of treatments of existing alien vegetation, and the response of selected species as both out-plants and natural recruits. Experimental plots either had the over-story completely removed, the under-story sprayed with glyphosate, or both in combination. Plots in which the over-story was removed and under-story was killed with glyphosate responded with the greatest amount of natural native recruitment after 2 years, and resulted in site conditions best suited to continued management. Out-planted species were more vigorous and survival was higher in plots where over-story vegetation was removed compared with plots where over-story was retained. Our results suggest that restoration of highly degraded Hawaiian forests may be more feasible than previously thought, particularly in areas where removal of alien canopy and reduction of alien under-story can stimulate an existing native seed bank.

BENSON, LORI K. Louisiana State University Museum of Natural Science and Department of Biological Sciences, 119 Foster Hall, Baton Rouge, LA 70803, USA.
DNA Sequence variation and morphoMETRIC Analysis of two types of Kuhlia sandvicensis in HAWAII.
Aholehole, Kuhlia sandvicensis, are believed to be endemic to the Hawaiian Islands and are an important and popular food fish. Local observers have noted the presence of two "types" of Kuhlia in Hawaii and have divided them into groups based on eye size. Currently, both morphotypes are identified as one species in the scientific literature. Because meristic counts in the two types of Kuhlia often overlap, a preliminary discriminant function analysis was conducted to determine if the specimens could be classified based on morphometric data. This analysis grouped the two types of Kuhlia with more than 95% correctness. Furthermore, a DNA sequence analysis (mitochondrial cytochrome b gene) revealed phylogenies with two monophyletic clades, which correspond to the aforementioned morphological types. Sequence divergences are high between the two morphotypes of Kuhlia and indicate species level differences. Due to their taxonomic classification as one species, management strategies currently in place are likely more beneficial for one species than the other. Thus, conservation and management strategies should be reconsidered, as these results indicate that the two Kuhlia morphotypes are genetically distinct populations. In addition, current research indicates differences in habitat preferences, and future research may elucidate variations in life history strategies as well.

BENSTEAD, JONATHAN P. Institute of Ecology, University of Georgia, Athens GA 30602, USA.
CONSERVATION STATUS OF MADAGASCAR'S RIVER ECOSYSTEMS: LESSONS FROM A LARGE ANCIENT ISLAND.
Madagascar is the fourth largest island in the world, comparable in size to Texas or France. It is also a continental fragment of Gondwana that has been isolated from other land masses for 90 million years. Large size and long isolation have resulted in great hydrographic complexity and a very high level of endemism and incidence of taxonomically basal taxa among the freshwater biota. Conservation of the river ecosystems of this large, ancient tropical island consequently faces fundamentally different challenges than those faced on small oceanic islands. Pressures on Madagascar's rivers operate overwhelmingly at large (catchment or supra-catchment) scales; deforestation and introduction of exotic species are the principal threats to these systems. These pressures have given rise to widespread ecological degradation (principally caused by sedimentation) and Madagascar now has the most endangered native fish assemblage in the world. Conservation of the island's river species and habitats faces two immediate challenges: the widespread and ongoing degradation already sustained; and the incomplete information available for making decisions and establishing priorities. Basic research, including biological surveys and taxonomic studies, combined with geographic information systems approach, is needed so that river systems can be more effectively considered in future conservation planning.

BERENDS, MICHAEL. P., and Mart R. Gross. Department of Zoology, University of Toronto, 25 Harbord St., Toronto, Ontario, Canada M5S 3G5.
TRADING NATIVES FOR ALIENS? FISH SPECIES BIODIVERSITY IN THE GREAT LAKES.
The Great Lakes are a model system for what is happening to fish species biodiversity throughout North America. The Great Lakes currently have 151 species, of which 17% (26) are established aliens. Fish stocking has been the single major source of aliens, accounting for 32%, and canals have been the next most important source (25%). There are multiple sources of aliens, but government sponsored stocking and aquatic transportation programs have made principal contributions. The Great Lakes were once managed for natives, but after alien Pacific salmon were stocked to control the canal-entered alien alewife, government management plans shifted to preserving alewife and Pacific salmon for the sports fishing industry. Although the direct cause-and-effect are poorly known, four native species that were endemic to the Great Lakes are now globally extinct and an additional 11 native species are extirpated. However, at the species level, across the entire Great Lakes system, fish biodiversity experienced an overall increase of 11 species (26 established aliens minus 15 natives lost from the system). The Great Lakes is thus a larger-scale example of the famous Clear Lake, California situation cited in many conservation biology textbooks, showing a decline in natives and an increase in total species.

BERRY, KRISTIN H., Bruce L. Homer, and Mary M. Christopher. U.S. Geological Survey, WERC, 6221 Box Springs Blvd., Riverside, CA 92507, USA (KHB), Dept. of Pathobiology, College of Veterinary Medicine, University of Florida, Box 110880, Gainesville, FL 32611, USA (BLH), Dept. of VetMed: Path, Micro, and Immunology, University of California, Davis, CA 95616, USA (MMC).
RECOVERY THREATENED: NEW DECLINES IN PREVIOUSLY STABLE POPULATIONS OF DESERT TORTOISES.
The Goffs population of desert tortoises in the eastern Mojave Desert, Mojave National Preserve, California, has long been considered the Gold Standard of stable populations for the species. Populations have been monitored since 1977. Between 1980 and 1994, population densities for all sizes of tortoises remained between 140 and 173/sq. km; densities of adults ranged from 66 to 88/sq. km. No statistically significant differences occurred between sample years. Between 1994 and 2000 populations declined 80% for all sizes and 92% for adults; the changes were statistically significant. Results of hematological and plasma biochemical tests, ELISA tests for Mycoplasma agassizii, and cultures collected between 1990 and 1995 indicate increasing severity of shell disease (cutaneous dyskeratosis), increasing frequency of mycoplasmosis, and liver disease. Necropsies of three tortoises salvaged in 2000 implicate metabolic disease, shell disease, and elemental toxicosis. Other potential contributing factors are herpesvirus infections and droughts. This population decline follows similar catastrophic declines in eastern California in the Chemehuevi Valley in the 1990's and on the Chuckwalla Bench (1982-1996). Consideration should be given to disease-related causes of mortality in recovery efforts. California populations may require a Federal listing upgrade from threatened to endangered status.

BERTHIER, PIERRE, and Gordon Luikart. Laboratoire de biologie des populations d'altitude, UMR 5553 CNRS / Université Joseph Fourier, BP 53 - F-38041 Grenoble cedex 9, France.
QUANTIFICATION OF THE EFFECT OF SUBSTRUCTURE ON EFFECTIVE POPULATION SIZE ESTIMATORS.
The effective population size (N_e) is a key parameter in conservation biology because it determines the rate of inbreeding and genetic change. Unfortunately, N_e is difficult to estimate demographically in natural populations. Consequently, genetic N_e estimators are becoming widely used, along with powerful markers (microsatellites) and statistical methods. However, these estimators are based on models that include unrealistic assumptions such as the absence of substructure in the studied population. We quantified the loss in accuracy and precision due to substructure on the most commonly used genetic estimator of N_e, which is based on the rate of change in allele frequencies over time. We used an empirical approach by simulating data. We show that hidden substructure (and immigrants) can lead to a substantial loss in accuracy and precision as soon as ~10-20 percent of the sample size is contributed by foreign individuals originating from another population differentiated by a F_ST of 0.1. Fortunately, such individuals should be detectable by clustering techniques or assignment tests. We quantified the effect of population differentiation (F_ST) and of the number of loci used on the estimator. This work provides guidelines and some assurance to users of genetic N_e estimators when limited substructure or immigration exists.

BEYERS, JAN L., Therese O'Rourke, Tom White, and Sandiann Engh. USDA Forest Service, Pacific Southwest Research Station, Riverside, CA, USA (JLB) and USDA Forest Service, Cleveland National Forest, San Diego, CA, USA (TO, TW, SE).
MANAGING ENDANGERED SPECIES AND HUMAN USES IN HABITAT "ISLANDS": THE SOUTHERN CALIFORNIA NATIONAL FORESTS.
The four National Forests in southern California - the Los Padres, Angeles, San Bernardino, and Cleveland - comprise islands of relatively natural habitat in a sea of expanding human development. Demands on the Forests for recreation, transportation and utility corridors, water development, and other uses are increasing. The Forests are also home to 64 species listed or proposed under the Endangered Species Act. The four Forests recently completed joint consultation with the US Fish and Wildlife Service over the effects of their current management plans on these species. To analyze the impacts of myriad uses on many species, a GIS database was set up to overlay Forest activities on known and potential species habitat. Spatial habitat models were developed with input from species experts. Another database was created to list and track interim conservation measures proposed by interdisciplinary specialist teams to protect habitat. Forest management teams approved a subset of the measures for the consultation package. The database tools - GIS layers and proposed conservation measures - will continued to be refined and used as the Forests begin a management plan revision process involving public collaboration. These tools may have utility for other organizations and are explained in this presentation.

BIERWAGEN, BRITTA G. Bren School of Environmental Science & Management, University of California, Santa Barbara, Santa Barbara, CA 93106, USA.
Impacts of land use changes on southern California butterflies.
The landscape of southern California has changed dramatically over the last century. Many native habitats have been converted to other uses, thus changing habitat patterns in terms of fragmentation and isolation. The land use matrix surrounding habitats may contribute to changes in the flora and fauna within fragments. The short generation time of butterflies makes them well suited for an investigation of the effects of land use change and fragmentation. Researchers in England, the Netherlands, and Poland have observed morphological responses by butterflies to fragmentation. I measured museum specimens taken over the past 50 years at collection localities in southern California. I then looked at land use changes surrounding these habitat using a time series of aerial photographs. Differences seen have important implications for conservation planning. Individual habitat areas need to be evaluated not only in the context of their current isolation, but also the history of their surrounding land uses.

BISER, JENNIFER A., Laura A. Vogel, Sabine S. Loew. Department of Biological Sciences, Illinois State University, Normal, Illinois 61761, USA.
THE EFFECTS OF ENVIRONMENTAL POLLUTANTS ON IMMUNOCOMPETENCE IN PEROMYSCUS LEUCOPUS.
To what degree habitat deterioration from pollution contributes to the long-term survival of natural populations is currently unknown. Laboratory models cannot reproduce environmental conditions experienced by natural populations, and may therefore not be sensitive enough to assess the effects of particular contaminants in the wild. The study of contaminants in natural situations has lagged far behind laboratory investigations of this subject, especially in terms of sublethal data, as well as the quantitative effects of pollutants on wild mammals at the population level. Previous work has documented the detrimental effects of exposure to high levels of heavy metals and polychlorinated biphenyls (PCBs) on laboratory mammalian immune fitness. To assess such effects in natural populations, wild white-footed mice, Peromyscus leucopus, were trapped from known polluted and pristine locations. Immunocompetence was evaluated via antibody and cytokine production as well as immunopathology. Initial results show no significant difference between clean versus contaminated sites. Uptake analysis, however, revealed low levels of tissue contaminants in mice. This may have failed to cause a threshold reduction in immunocompetence, as detected by the methods employed in this study.

BJORK, ROBIN. Wildlife Conservation Society and Department of Fisheries and Wildlife, Oregon State University, Corvallis, Oregon 97331, USA.
PATH OF THE PARROT: REGIONAL HABITAT LINKAGES AND CONSERVATION PLANNING IN TROPICAL LOWLAND FORESTS.
Little is known about regional resource gradients in lowland tropical forests or the animals that may rely on them. My case study demonstrates reliance of a large avian frugivore on a suite of lowland forest habitats arrayed in a specific spatiotemporal pattern, creating ecological interdependencies among habitats. For three years, I have documented that the adult population of Mealy Parrots breeding in northeastern Guatemala engages in predictable regional migrations within mature moist forests. Although the overall area covered by the parrots is considerable (10,000 km² ), of greater significance is the consistency of their movements and specific locations they utilize. Even over distances of a few hundred kilometers there is variation in fruiting phenology, forest composition, and rainfall and these probably play a significant role in shaping the site-area requirements of the birds. Large protected areas, such as the 576 km² Tikal National Park and even the 21,000 km² Maya Biosphere Reserve, do not adequately protect the annual requirements of Mealy Parrots breeding there. My results illustrate the asymmetry of this landscape and underscore the importance of regional gradients and habitat connectivity in maintaining animal populations, and thus in conservation planning, in lowland tropical forests.

BLUNDELL, GAIL M., Nathan P. Nibbelink, Merav Ben-David, Pamela Groves, R. Terry Bowyer, and Eli Geffen. Alaska Cooperative Fish and Wildlife Research Unit, University of Alaska Fairbanks, Fairbanks, AK 99775, USA (GMB), Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82071, USA (NN, MB); Institute of Arctic Biology, University of Alaska Fairbanks, Fairbanks, AK 99775, USA (GMB, PG, RTB), and Institute for Nature Conservation Research, Tel Aviv University, Ramat Aviv, 69978, USA (EG).
ARE REINTRODUCTIONS OF RIVER OTTERS NECESSARY? PATTERNS OF GENE FLOW IN POPULATIONS WITHOUT BARRIERS.
Historically, river otters (Lontra canadensis) were present throughout most of North America but urbanization, pollution, and overharvest greatly reduced their distribution by the early 1900s. Recently, expensive projects have successfully reintroduced otters in many areas. We studied genetic diversity among populations of river otters inhabiting marine environments in wilderness areas of Prince William Sound, Alaska, to determine characteristics of sex-biased dispersal and gene flow in an environment without barriers to dispersal. With those data we constructed a model to evaluate the potential for natural repopulation under optimal conditions, versus the need for reintroduction in areas with geographic barriers to dispersal (i.e., inland habitats and terrestrial distance between watersheds). Fifty-five otters in three populations were radiotracked to determine patterns of dispersal, and genetic material (microsatellite DNA) was obtained from 110 individuals in seven populations to determine patterns of gene flow. Both telemetry and genetic data indicated that female otters generally showed site fidelity and males were more likely to disperse. Despite lack of geographic barriers in a marine system, dispersal distances were relatively short, indicating that the likelihood of natural repopulation from adjoining watersheds in inland habitats is limited when watersheds are distant.

BOARMAN, WILLIAM I. U.S. Geological Survey, Western Ecological Research Center, Riverside, CA 92553, USA.
THE ECOLOGY OF A SUBSIDIZED PREDATOR: COMMON RAVEN.
Subsidized predators are species whose populations survive and often thrive on resources (e.g., food, water, safety, etc.) provided by humans. Common ravens (Corvus corax) are an excellent example of a subsidized predator. Their populations have grown precipitously in recent years as a result of the proliferation of human activities in the desert. In a multi-year survey, significantly more ravens were found at landfills and sewage ponds than at other human-dominated and natural areas. Radio-tagged ravens primarily moved between anthropogenic resource sites. Nestling and fledgling survivorship is higher in nests located near human-provided resources. Ravens prey on juvenile desert tortoises (Gopherus agassizii), a Federally-listed threatened species, over much of the tortoise's range. However, not all ravens prey on tortoises, and few appear to prey on them in large numbers. It is not known if raven predation is high enough to prevent tortoise recovery rangewide, but it is likely high enough in some areas to alter success of recovery efforts. Removal of selected ravens known to prey on tortoises will likely aid short-term recovery in some areas, but such efforts should be coupled with aggressive reductions in anthropogenic resources (e.g., garbage) made available to ravens.

BOERSMA, P.D. Department of Zoology, University of Washington, Box 351800, Seattle, WA 98195-1800, USA.
MARINE RESERVES AND MARINE ZONING AS TOOLS FOR THE CONSERVATION OF MARINE MAMMALS AND SEABIRDS.
Ecological theory is an underlining construct in the design of terrestrial and marine reserves. The general theory for reserve design is widely accepted (i.e. bigger is better, importance of fragmentation, invasives, dispearsal, and management). Although, most models of marine reserves were developed for mid-trophic level fish, marine conservation should also focus on top-predators such as marine mammals and seabirds. In the terrestrial realm, "wildlands" advocates have argued the only way to maintain top carnivores such as elephants, grizzly bears, or wolves is by designating large tracts of land as protected areas. This strategy however, does not fit the "commons" of the marine realm. The designation of small marine reserves, protected areas, and sanctuaries are unlikely to benefit long-term conservation of wide-ranging marine mammals or seabirds. For example Albatrosses breeding in the Southern Ocean Sanctuary are killed off the Brazilian coast. Effective conservation strategies require detailed life history knowledge. We need to better understand how different species use the environment and how uses change based on season, breeding location, and climatic variation. Because whales and seabirds use large and variable areas, large scale and flexible marine zoning may offer the most conservation benefits. This in turn necessitates international cooperation.

Bohorquez, Amy S., BARRY A. NICKEL, Emma K. Grigg, Deborah E. Green, Robin M. Bouse, Bruce E. Jaffe, Sarah G. Allen, and Hal Markowitz. San Francisco State University, Department of Biology, 1600 Holloway Avenue, HH222, San Francisco, CA 94132, USA (ASB, BAN, EKG, DEG, HM), National Park Service, Point Reyes National Seashore, USA (SGA), U.S. Geological Survey, 345 Middlefield Rd, MS 472, Menlo Park, CA 94025, USA (RB, BJ).
THE HIGH PRICE OF GOLD: POSSIBLE EFFECTS OF HYDRAULIC MINING ON HARBOR SEALS IN SAN FRANCISCO.
The Gold Rush of the 1853 brought prosperity to California; however, the costs of the methods used to extract that gold are still being assessed today. The use of highly pressurized water washed over one billion tons of sediment from the Sierra Nevada foothills, a portion of which remains in the San Francisco Bay (SFB). The SFB also has the highest proportion of harbor seals with a red discoloration of their pelage. The red color results from iron oxide adherence to the keratin surface on the shaft of the coat hairs. This iron may adhere while seals are foraging in sediment contaminated by re-exposed hydraulic mining debris. In order to test this hypothesis, a model will be developed from a collaborative approach integrating behavioral, geological and chemical methods by, (1) synthesizing behavioral information obtained from VHF radio and satellite tagged harbor seals into a Geographic Information System (GIS); (2) classifying habitat from a digital terrain model based on bathymetric and hydraulic mining sediment data; and (3) chemically analyzing the red pelage for the hydraulic mining signature and other metals that may be linked to the debris. We present this plan to demonstrate the benefits of collaborative analysis for conservation issues.

Bohorquez, Amy S., Deborah E. Green, Emma K. Grigg, Hal Markowitz, and Sarah G. Allen. San Francisco State University, Department of Biology, 1600 Holloway Avenue, HH222, San Francisco, Ca. 94132 (ASB, EKG, DEG, HM), National Park Service, Point Reyes National Shoreline, USA (SGA).
CURRENT STATUS OF RED-PELAGED HARBOR SEALS WITHIN THE SAN FRANCISCO BAY.
A previous study of the San Francisco Bay (SFB) red-pelaged harbor seals suggests a significant difference in haul out use by these seals at three sites within the bay. Allen et al. (1993) reported that red pelage resulted from iron oxide adherence to the keratin surface on the shaft of the hairs, likely related to foraging behavior. This coloration tends to make the red fur more brittle, leading to the loss vibrissae. Using several methods of comparison, a significant difference in the number of seals with red pelage was found between the central bay and the north and south bays. Continued analysis of this population from June 1999 to December 2000 showed significantly more of red-pelaged seals used the north and south bay haul out sites (p<0.05). We found red-pelaged mothers gave birth earlier in both pupping seasons (p<0.05). These results may indicate that red-pelaged mothers are older or foraging in nutrient rich areas, presenting a possible balance between foraging efficiency and the cost of the red-pelage. We recommend using this demographic data and the red-pelage as an indicator of optimal foraging areas and therefore identifying these areas for protection.

BOOTH, D., W.W. Crozier, and P.A.P. Prodöhl. Department of Biology and Biochemistry Queen's University of Belfast, The Medical Biology Centre, 97 Lisburn Road, Belfast, BT9 7BL (DB, PAPP), The River Bush Salmon Station, D.A.R.D.N.I. (WWC).
MICROSATELLITE DNA PROFILING: MONITORING ATLANTIC SALMON (SALMO SALAR L.) FISHERIES IN NORTHERN IRELAND.
The Foyle Atlantic salmon Fisheries comprised of over 4000 km2 catchments area and more than 3000km of main river systems, is one of the largest in the UK with catches of over 50,000 fish/year (estimated at 70-90% of all adult fish in the system). Recently a large EU funded project has been commissioned to supplement wild stocks within the system. Suitable management of mixed stock requires characterisation of the many individual components and their relative contribution to the fisheries. Little is known, however, on the genetic composition within the Foyle catchments. To address this situation, an extensive sample set of over 3000 fish, comprising of some 30 populations, was exhaustively collected from the major catchments areas of this region including temporal and spatial structure. Samples were screened for 7 highly polymorphic microsatellite loci. The combined result of the sampling strategy and extensive analysis of the microsatellite data, provided us with an unique data set investigating the effectiveness of DNA profiling in the management of this fishery, and also yielded an essential genetic baseline to monitor potential subsequent genetic changes that may occur due to anthropogenic effects (e.g. habitat alteration, stock supplementation).

BORCHERS, JEFFREY G., Gay A. Bradshaw, Carol Spinos, Dean Apostol, and Bart R. Johnson. Forest Community Research, Taylorsville, CA, PNW Research Station, Corvallis, OR, USA (GAB), and Rogue River National Forest, Medford OR, USA (CS), 23850 SE Borges Rd., Gresham, OR 97080, USA (DA), and Landscape Architecture Department, University of Oregon, Eugene, OR 97403, USA (BRJ).
DECISION MAPPING FOR LANDSCAPE RESTORATION UNDER ADAPTIVE MANAGEMENT.
In conservation planning, collaboration is essential to creating landscape designs that are both ecologically and socially sustainable. On public lands, the process requires the balanced participation of communities, scientists, and managers, for each possesses unique and necessary stores of knowledge, experience, wisdom, and values critical to making sustainable decisions. In the Applegate Adaptive Management Area (Rogue River National Forest, Oregon, USA), a landscape design for the Little Applegate River watershed was completed. This community-based collaboration with natural resource management agencies resulted in a set of landscape issues and goals, as well as a watershed map of desired vegetation 200 years in the future. These mapped goals are now the basis for developing a 10-year conservation plan under NEPA using decision mapping, a process that integrates the values and knowledge of all stakeholders to improve the quality of decision making. "Value trees" and utility models were constructed to provide support for public involvement and analyses in the preparation of an environmental impact statement that delineates alternative restoration strategies for the watershed. Because they reveal areas of stakeholder differences, they were critical in resolving conflicts and assessing specific trade-offs among alternative strategies for landscape restoration.

BOUZAT, JUAN L. Department of Biological Sciences, Bowling Green State University, Bowling Green, OH 43403-0212, USA.
GENETIC DIVERSITY AND CONSERVATION.
A commonly accepted "conservation dogma" states that levels of genetic diversity provide a good indicator of extinction probabilities in natural populations. As a result, a considerable number of conservation genetic studies focuses on the detection of low levels of genetic diversity at the molecular level to suggest detrimental effects on fitness and, therefore, increases in extinction probabilities. This prevalent dogma assumes a simple relationship between genotypes and phenotypes and consistent detrimental effects on fitness through inbreeding depression. In addition, it disregards the potential effects of fine scale population structure, and mutation as an extinction factor. These issues have direct practical implications for conservation. For example, the relation between molecular and adaptive genetic variation has direct relevance on the preservation of phenotypic diversity and the characterization of conservation units. Inbreeding does not necessarily imply detrimental effects on population fitness, and fine scale population structure may be an important mechanism for maintaining genetic diversity in small populations. Finally, mutation may play a relevant role in extinction through the process of mutational meltdown. Experimental and field research in each of these areas is essential if we want to better understand the practical implications of genetic diversity in conservation.

BOYD, IAN L., and John Croxall. British Antarctic Survey, NERC, Madingley Road, Cambridge CB3 OET, UK; Sea Mammal Research Unit, NERC, Gatty Marine Laboratory, University of St Andrews, Fife KY16 8LB, UK.
PRINCIPLES UNDERLYING THE USE OF MARINE PREDATORS TO DEFINE PROTECTED AREAS IN THE SOUTHERN OCEAN.
Land-based marine predators, including seals and seabirds, are particularly abundant in the Southern Ocean. A large proportion of these species use sub-antarctic islands for breeding and raising offspring. Based upon satellite tracking data collected from individuals of several species, it has been possible to define and examine variation in their foraging habitats. The spatial scale of such habitats varies extensively between species, ranging from 10s to 100s of kilometres from the breeding site for several penguin and seal species, to 1000s of kilometres for albatross species. Examination of these foraging ranges and, in particular, the detection of specific regions of importance to these predators may highlight areas of importance to both predators and their prey. In addition, the breeding performance of these predators can provide a measure of the status of the ecosystem, and in particular, changes in the relative biomass of krill. The integration of spatial information regarding foraging with temporal variation in overall breeding performance could be used as a tool for the identification of areas of recurring benefit to these species, and thus may allow us to define zones for protection in the pelagic environment.

BRADSHAW, GAY A., and Joseph Mitchell. PNW Research Station and Rogue River National Forest, 6941 Upper Applegate Road, Jacksonville, OR 97530, USA, Cherokee Nation and USDA Forest Service, P.O Box 948, Tahlequah, OK 74465, USA.
RE-VISITING THE CONTRACT: METHODS AND THEORY FOR INTEGRATING AMERICAN INDIAN AND EURO-AMERICAN APPROACHES TO CONSERVATION.
American Indian and Euro-American approaches to conservation are often perceived as incompatible because of significant contrasts in philosophy, practice, communication styles, and social organization. In collaboration with tribal and non-tribal scientists, we developed methods, analysis, and theory for integrating Indian and non-Indian approaches to conservation planning. We assert that meaningful engagement of American Indians and the adoption of Indian approaches to conservation must explicitly consider law, concepts, ethics, and ecology as an integral whole in the planning process. To this end, and to understand how Euro-American and Indian perspectives can complement each other, we employ cognitive maps. A standard method in cognitive and decision sciences, cognitive maps are visual tools that we employed to improve inter-cultural communication and understanding of the socio-ecological relationships of the two communities. Specifically, cognitive maps elucidated the delicate relationships between "values" and "facts" in both cultures. This format provided an equitable medium to examine both scientific viewpoints and helped unravel the tangle of cultural assumptions that block collaboration. Cognitive mapping revealed the multiple and shifting relationships of ecological phenomena and beliefs. We illustrate this process using specific examples of fire re-introduction as part of current federal and tribal restoration plans in the United States.

Braggins, John E., John J. Engel, MATTHEW VON KONRAT, Peter J. de Lange, and Matthew Renner. Plant Science, School of Biological Sciences, The University of Auckland, Private Bag 92019, Auckland, New Zealand (JEB, MR), The Field Museum, Chicago, Illinois 60605-2496, USA; and Science & Research Unit, Department of Conservation, Private Bag 68908, Auckland, New Zealand (PJL).
RARE AND THREATENED LIVERWORTS: AN OVERLOOKED AREA OF CONSERVATION IN THE NEW ZEALAND BOTANICAL REGION.
New Zealand is an archipelago that is both geographically and climatically diverse. Recently, it has been stated that the decline in biological diversity is New Zealand's most pervasive environmental issue. The impact of potential pressures such as micro-habitat loss and fragmentation on the liverwort flora have not been previously investigated. This is alarming because it is estimated that New Zealand has c. 10% of the world's liverwort species, and more importantly, has 48 of the 73 liverwort families. Because of these factors together with the high level of endemism, the liverwort flora of New Zealand is considered to be of worldwide significance. First, this study identified potential threats to the New Zealand liverwort flora, and second, developed a preliminary checklist of threatened liverworts to be considered as candidates for the World Red List. Selection of species was based on field studies spanning three decades, extensive herbarium records, and bibliographic references. Case studies illustrating the different categories of threat occurring in New Zealand include examples from a monotypic family and a monotypic genus. We conclude that research relating to the conservation biology of the New Zealand hepatic flora requires great urgency and recommend areas of priority.

BRANDHAGEN, MICHAEL D., Mary Cogliano, Larry L. Rockwood, Lee M. Talbot, Debra J. Ellis, and Patrick M. Gillevet. Department of Biology and Environmental Science and Public Policy, George Mason University, 4400 University Drive, MSN 3E1, Fairfax, VA 22030, USA (MDB, MC, LLR, LMT), Department of Biology, University of Massachusetts Dartmouth, 285 Old Westport Road, North Dartmouth, MA 02747, USA (DJE), School of Computational Sciences, George Mason University, Prince William Campus, 10900 University Boulevard, MSN 4E3, Manassas, Virginia 20110, USA (PMG).
NONINVASIVE POPULATION MONITORING OF NORTH AMERICAN RIVER OTTERS (LONTRA CANADENSIS) USING FECAL DNA.
Conservation and monitoring of North American river otters (Lontra canadensis) is challenging as they are a semi-aquatic, wide-ranging species, and typically occur at low population densities. Mark-recapture census methods are inadequate due to the difficulty of capturing individuals, and to the stress and potential injuries suffered by the otters. To date, noninvasive methods developed for monitoring river otter populations have been considered impractical for general use. We report here a pilot study of a noninvasive molecular method for monitoring L. canadensis populations. Fecal (scat) samples were collected during a survey of latrine sites from September through November 1999, at the Chopawamsic creek watershed on the US Marine Corps Base at Quantico, Virginia. Protocol was developed for extraction of otter DNA from scat using a soil DNA extraction kit. Extracted DNA was successfully amplified with mitochondrial 12S rRNA, microsatellite, and Y-chromosome primer sets in order to assign species, sex, and individual identity to otters based on their scat samples. From these data, sex ratio and minimum population size were determined. This study demonstrates the feasibility of population monitoring of L. canadensis using DNA from fecal samples, and should allow more precise population estimates in the future.

BRANNON, REE, and Michael L. Zientek. USGS/BRD Gap Analysis Program, 530 S. Asbury, Moscow, ID 83843, U. S. Geological Survey, 904 W. Riverside Ave., Room 202, Spokane, WA 99201, USA.
Biodiversity and Nonfuel Mineral Development - Competition or Coexistence?
As human populations continue to increase, we see a corresponding increase in the demand for consumptive uses of our natural resources. This in turn hastens loss of biodiversity by fragmenting habitat and disrupting ecological processes. Our traditional approach to development has been generally haphazard and driven by accessibility and opportunity rather than careful comprehensive planning. The Gap Analysis Program and the USGS Geologic Division investigate the feasibility of identifying which domestic mineral resource tracts could be explored with the least affect on biodiversity. Recognizing that exploration and development can have disruptive effects to biodiversity, either from the direct consequences of mining or from opportunistic growth by other sectors, it would be advantageous to integrate biological analyses early in the process of prioritizing mineral exploration. Geologists model "permissive" and "non-permissive" areas for potential mineral exploration as biologists model predicted distribution of species and vegetation alliances in the GAP Program. We use data from Montana to analyze mineral sites and GAP data. We spatially integrate the results to generate a ranking of sites indicating mineral potential and value for biological resources (richness and representation). We describe the feasibility of creating practical scenarios for future development of resources.

BRASHARES, JUSTIN S., Peter Arcese, and Moses K. Sam. Centre for Applied Conservation Biology, University of British Columbia, Vancouver, BC V6T 1Z4 (JB, PA), Ghana Wildlife Division, P.O. Box M.239, Ministry Post Office, Accra, Ghana (MKS).
Reserve Size, Human Demography, and Wildlife Extinction in West Africa.
Species-area models have become the primary tool used to predict baseline extinction rates for species in isolated habitats, and have influenced conservation and land use planning worldwide. In particular, species-area models are used to predict faunal relaxation as a result of natural demographic processes after habitat patches are reduced in size. However, where human influences such as hunting or the introduction of exotic species are common, extinction rates higher than those predicted by species-area models are expected and conservation plans based solely on species-area models may fail. Here, we show that extinction rates for 41 species of large mammals in six nature reserves in West Africa were 14 to 307 times higher than those predicted by reserve size under the species-area model. Ninety-eight percent of the observed variation in extinction rates among reserves was accounted for statistically by human population and reserve size. Extinction occurred at higher rates than predicted by species-area models for carnivores, primates and ungulates, and at the highest rates overall near reserve borders. Our results indicate that, where the harvest of wildlife is common, conservation plans should focus on increasing the size of reserves and reducing the rate of hunting.

Brasher, Anne M. U.S. Geological Survey, 677 Ala Moana Blvd. #415, Honolulu, HI 96813, USA.
EFFECTS OF HABITAT ALTERATION ON COMMUNITY STRUCTURE IN HAWAIIAN STREAMS.
Streams throughout Hawaii have been altered by water diversion, flow reduction, channel modification, introduced species, and water-quality degradation. As population increases and land-use changes, watersheds are becoming far different from those that once sustained native stream communities. During a 3-year study on Molokai, stream diversions were shown to correlate with reduced habitat availability and higher species overlap for fish and macro-crustaceans, potentially resulting in increased competition and predation. Dewatering can also inhibit downstream dispersal of larvae and upstream migration of post-larvae, critical to the life cycle of native amphidromous species. On Oahu, the U.S. Geological Survey's National Water-Quality Assessment Program has shown urbanization, associated with channel modification and reduced canopy cover, to result in higher water temperatures and increased temperature fluctuations, habitat characteristics favoring non-native fish. Benthic invertebrate community composition also was influenced by habitat alteration; the ratio of Diptera to Trichoptera varied positively with siltation and negatively with velocity. Total invertebrate abundance was greatest at forested sites, while taxa richness (predominantly non-native species) was higher at degraded sites. As Hawaii's population increases, the potential for watershed modification also increases. Understanding the relation between habitat alteration and aquatic community structure is critical for the development of sound management strategies.

BRIDGES, CHRISTINE M., and Raymond D. Semlitsch. USGS Columbia Environmental Research Center, 4200 New Haven Rd., Columbia, MO 65201, USA (CMB), University of Missouri, Division of Biological Sciences, 110 Tucker Hall, Columbia, MO 65211, USA (RDS).
VARIATION IN PESTICIDE TOLERANCE OF RANID TADPOLES AND PATTERNS OF AMPHIBIAN DECLINE.
There is significant variation among and within amphibian species with respect to reports of population decline; declining species are often found in environments that are physiograpically similar to environments where the same species is thriving. Because variability exists among organisms in their sensitivity to environmental stressors, it is important to determine the degree of this variation when undertaking conservation efforts. We conducted both lethal (time-to-death) and sublethal (activity change) assays to determine the degree of variation in sensitivity of tadpoles to a pesticide, carbaryl, at three hierarchical levels: among ranid species, among several populations of a single ranid species (Rana sphenocephala), and within populations of R. sphenocephala. We observed significant variation in time-to-death among the nine ranid species, and among the 10 R. sphenocephala populations we tested. Four out of eight R. sphenocephala populations exhibited significantly different times-to-death among families. The magnitude of the activity change when exposed to sublethal carbaryl levels was significantly different among species and within R. sphenocephala populations. Chemical contamination, at lethal or sublethal levels, can alter natural regulatory processes (e.g., juvenile recruitment) in amphibian populations and should be considered as a contributing cause of declines in amphibian populations.

BRITTEN, HUGH B., and Joseph W. Glasford. Department of Biology, University of South Dakota, Vermillion, SD 57069, USA.
DAKOTA SKIPPER POPULATION GENETIC STRUCTURE IN A FRAGMENTED NATIVE PRAIRIE LANDSCAPE.
The Dakota skipper (Hesperiidae: Hesperia dakotae) is a lepidopteran native to central North American tallgrass and mixed-grass prairie. Agricultural conversion has resulted in the loss of approximately 98% of its native habitat in Iowa, Minnesota, North and South Dakota, and Manitoba, Canada. The species has been extirpated from Iowa and only remnant fragments of tall and mixed-grass prairie remain in government and private preserves. As a result, it has been considered for protection under the US Endangered Species Act and is listed as Threatened by the State of Minnesota and the Province of Manitoba. We used protein electrophoresis to assay genetic variability at nine sample locations throughout its current range. Twenty-one isozyme loci were assayed, 12 of which were polymorphic. Significant isolation by distance was detected among the seven Minnesota and South Dakota sites suggesting historic dispersal among these populations in continuous habitat. The two northwestern-most sites in Manitoba were similar in allele frequencies and showed significant genetic divergence from the other sample locations. These results suggest that preservation of the remaining genetic diversity of Dakota skipper populations will require preservation of habitat fragments from throughout its geographic range.

BROOKS, MARGARET A. National Park Service, 1415 N. Sixth Ave., Tucson, AZ 85705, USA.
VOLUNTARY CONSERVATION OF SIGNIFICANT NATURAL RESOURCES ON PRIVATE LANDS: MAKING IT WORK.
The National Natural Landmarks (NNL) Program recently saw the end of a 10-year moratorium that had limited program activity and new designations. This long moratorium, enacted when property rights groups questioned the effects of NNL designation on private property, caused program coordinators to move into a more service oriented role to support the voluntary conservation efforts of landowners already participating in the program. Assistance to NNL owners and managers has ranged from grants to support research, monitoring, and education to "hands on" projects that have forged local partnerships and site improvements. All 587 designated NNLs are nationally significant for their biological or geological resources. They are widely dispersed around the United States and its territories. The National Park Service, through the NNL Program, provides an important link between these sites by serving as a conduit for information and by providing outreach through technical specialists located at parks. The results have been increased conservation and knowledge about these important areas, more efficient use of public funds through partnering, and improving conditions at some of the sites. Now that the program is fully functioning again, the NNL staff is poised to continue this level of support to existing and future participants.

BROOKS, THOMAS. Center for Applied Biodiversity Science, Conservation International, 1919 M Street, NW, Suite 600, Washington, DC 20036, USA .
Habitat loss and extinctions in the hotspots.
Nearly half of the world's plant species and a third of terrestrial vertebrates are endemic to 25 hotspots' of biodiversity. None of these hotspots retains more than 30% of its pristine habitat. Due to this habitat loss, we expect many hotspot endemics to have become extinct or to be threatened with extinction. We use the IUCN Red Lists to test this expectation. Overall, between a half and two-thirds of all threatened plants and 57% of all threatened terrestrial vertebrates are hotspot endemics. For birds and mammals predictions of extinction based on habitat loss generally match numbers of species independently judged extinct or threatened. For reptiles, amphibians, and plants, many less hotspot endemics are considered threatened or extinct than expected. However, this mismatch is small in temperate hotspots, suggesting that many endemic species in the poorly-known tropical hotspots have yet to be listed as threatened. Finally, we show that the Eastern Arc, Philippines and Pacific islands can least afford to lose any further habitat, and that if current deforestation rates continue the Caribbean, Tropical Andes, Philippines, Mesoamerica and Sundaland will lose the most species in the near future. Without urgent conservation intervention, we face mass extinctions in the hotspots.

BROTONS, LLUÍS, Mikko Mönkkönen, Ari Rajasärkkä, and Ari Nikula. Dept. of Biology, University of Oulu, PL 3000, 90014, Oulu, Finland (LB, MM),,Metsähallitus, Ostrobothnia Park area, Oulu, Finland (AR), Finnish Forest Research Institute, Rovaniemi, Finland (AN).
Is ISOLATION DECREASING OLD-GROWTH FOREST BIRd OCCURRENCE IN NORTHERN FINNISH RESERVES?
Old forests birds have been sharply decreasing in numbers during the last decades in Fennoscandia mainly due to the fragmentation associated with commercial forest harvesting. This decrease has been particularly evident in northern Finland where forestry has been very intensive, whereas adjacent areas in Russian Karelia have remained relatively untouched. Using a hierarchical approach, we ask how landscape characteristics of Northern Finnish forest reserves and their spatial location affect the occurrence of resident old forest birds. Proportion of old forest habitat had a clear positive effect on species number only in reserves further away from Russia. In areas close to Russia, with higher overall proportions of old-growth forest, proportion of habitat was negatively associated to species number. Reserve isolation was only marginally associated with species richness in areas close to Russia. Our results suggest that the effects of landscape structure on old-forest bird occurrence were not homogeneous across the reserve network, and that this variation might be related to thresholds in species responses to habitat availability and to the spatial location of the reserves within the network and in relation to source areas in Russia.

BROWN, CARRIE J., Bernd Blossey and Steven J. Joule. Department of Natural Resources, Fernow Hall, Cornell University, Ithaca, NY 14853, USA (CJB, BB), Wildlife Bureau, 21 South Putt Corners Road, New Paltz, NY 12561, USA (formerly with Environmental Division, Public Works, 85 First St West, Fort Drum, NY 13602, USA (SJJ).
DECREASED FITNESS AND SURVIVAL OF NATIVE AMPHIBIAN LARVAE IN A MARSH INVADED BY AN EXOTIC PLANT SPECIES.
The worldwide decline of amphibian populations and the increase of exotic species are issues of concern in natural resource management and conservation. The presence of exotic vertebrates is known to contribute to amphibian declines, but little information is available on how exotic plant invasions affect amphibians. This study assesses the impact of purple loosestrife (Lythrum salicaria) on Bufo americanus tadpoles in northern New York. I monitored survival, developmental rate, and diet of B. americanus tadpoles in field enclosures in a marsh invaded by purple loosestrife and in a marsh dominated by native Typha spp. Tadpoles developing in the purple loosestrife marsh had significantly lower survival and slower developmental rates compared to those in the cattail marsh. Preliminary gut analyses suggest that the tadpoles in the cattail marsh consumed algae of higher nutritional value than tadpoles in the purple loosestrife marsh. Decreased fitness and survival of tadpoles in purple loosestrife may ultimately cause reductions in adult breeding populations. These results suggest that invasion of exotic plants may contribute to amphibian declines.

BROWN, CHRIS W., Robert N. Fisher, and Ted J. Case. USGS, Department of Biology, San Diego State University, 5500 Campanile Dr., San Diego, CA 92182-1015, USA (CWB and RNF). Department of Biology, University of California San Diego, 9500 Gilman Dr., La Jolla, CA 92093, USA (TJC).
IMPACTS OF FRAGMENTATION OF A COASTAL DUNE/MARSH COMPLEX ON TERRESTRIAL VERTEBRATES.
Coastal wetlands in California have been reduced by 95% and the remaining habitat is often highly disturbed and fragmented. The last remaining intact dune/marsh complex in California south of Point Dume is the Tijuana River National Estuarine Research Reserve. This reserve encompasses approximately 2,500 acres: 1,660 acres are upland consisting of marsh upland, coastal sage scrub and coastal dunes. Coastal dunes form the western border of the reserve, urban and military development and agriculture comprise the remaining boundaries with relict marine terrace mesas along the south. This island of habitat is at the mouth of the 1,735 square mile Tijuana watershed shared between California and Mexico. Reptiles, amphibians and small mammals have been inventoried and monitored using passive pit-fall trapping and active searching since March 1997. 200 days of sampling at 15 points throughout the reserve has yielded over 4,250 captures across 3 species of amphibians, 17 species of reptiles, and 13 species of small mammals. The results indicate a decline and possible extirpation of several sensitive and protected species while other sensitive species remain common. Long term management of this unique ecosystem must balance the needs of both upland and wetland species in this fragmented landscape.

BROWN, GLENN. Environmental Science Program, Capilano College, North Vancouver, British Columbia V6K 1G50, Canada.
CONSERVATION BIOLOGY REQUIRES LEARNING EXPERTISE RATHER THAN KNOWLEDGE ALONE.
I present ideas derived from contemporary cognitive psychology about how and what people learn, how those ideas modify educational goals for science and what they imply about adjustments in teaching methods. Study of the performance of experts in many fields led psychologists to a refined concept of 'expertise,' meaning the knowledgable and flexible ability to address and resolve novel problems in a technical domain. This view of expertise seems a reasonable educational goal for conservation biology. Acquiring expertise turns out to require not knowledge alone, although it is essential, but intellectual traits and practices collectively called 'a self-regulated metacognitive perspective.' (Metacognitive refers to knowing management of one's thoughts, technical knowledge and intellectual abilities.). These attributes can, in turn, be learned through guided participation in a variety of goal-directed, problem-solving projects, which should include the means for students to plan and to reflect upon their own performance. My school's program has implemented a perspective on 'professional practice' to help students understand and set goals, teaches problem-solving directly and indirectly, and uses learning projects monitored with student portfolios to support self-regulated performance.

BROWN, JONATHAN M., and Christina Caruso. Biology Department, Grinnell College, Grinnell, IA 50112, USA.
PRAIRIE RESTORATION: AN INTRODUCTION TO BIOLOGICAL INQUIRY.
As a part of a major curricular reform, biology faculty members at Grinnell College have created a suite of new introductory courses, each of which explores a particular biological problem. Prairie Restoration introduces students to basic concepts in conservation biology, while emphasizing the ways that biologists ask questions, test hypotheses through observation and experimentation, and communicate their results. The course is taught in workshop format at our biological field station and local remnant prairies. It is structured in investigative blocks. Following two initial blocks that introduce students to methods of comparative and experimental analysis of prairie communities, teams of students carry out long-term investigations of their own design. Students publish the results of these projects in an electronic scientific journal (www.grinnell.edu/cera/tillers.html). We also include discussions of the history of the prairie region, and the ethical and economic challenges to the conservation and restoration of grasslands. We are assessing whether this approach to teaching introductory biology succeeds with students whose learning styles conflict with traditional lecture formats. In our second year, 75% of the students reported that the course changed their perspective on science. A majority indicated higher confidence in their ability to learn about new scientific topics.

BROWN JR., KEITH S., André Victor L. Freitas, and Márcio Uehara-Prado. Museu de Historia Natural, Instituto de Biologia, UNICAMP, C.P. 6109, Campinas, SP 13.084-970, Brazil.
ATLANTIC FOREST "ISLANDS" CLASSIFIED THROUGH BUTTERFLIES: AREA, HABITAT, AND POPULATION INSTABILITY.
The minimum size of a reserve for the preservation of options (species, communities, interactions) varies with the organisms and resources studied, habitat heterogeneity, disturbance, and isolation. The Brazilian Atlantic Forests (an ancient system of complex topography and high endemicity) have a long history of natural and anthropic disturbance: over 90% are strongly fragmented, leaving thousands of forest "islands" with 10-100 ha. Green, humid, and pesticide-free vegetation, even though anthropic, gives potential connectivity between such fragments throughout much of the region. Butterfly communities have been inventoried and followed in many sectors for over 50 years, seeking to establish baselines for conservation monitoring, planning, and management. Area classes range from "intact communities" (usually > 10,000 ha, heterogeneous habitat) down through different degrees of instability (in 100-1000 ha fragments, over a third of the species present may be transient, showing strong population imbalances) to subminimal systems requiring intensive management (< 10 ha or very uniform, greatly deplenished). Resident/migrant forest-interior butterflies including bait-attracted Nymphalidae, Ithomiinae, Riodininae, and Papilionidae are especially useful in classifying the local system. Many rare or threatened taxa on a 102-species list are recorded more rapidly and regularly in heterogeneous larger areas, and are useful for evaluation of system "integrity."

BROWN, MARK, and Elise Granek. National Park Service, NCRC-RTCA, Room 3612, 1849 C Street, Washington, D.C. 20240, USA (MB), Smithsonian Institution, 1100 Jefferson Drive, Suite 3123, Washington, D.C. 20560-0705, USA (EG).
ECOSYSTEM CONSERVATION IN MOHELI, COMOROS ISLANDS: A CO-MANAGEMENT APPROACH.
The Comoros Islands, a biodiversity hotspot in the West Indian Ocean, contain a high level of endemism. Political instability and limited funding, creating high potential for continued biodiversity loss, hamper existing efforts for biodiversity conservation. The history of international conservation and development projects in Comoros has been largely unsuccessful with few long-term results. The dire situation in Comoros warrants a new conservation approach: co-management, conservation from the bottom-up. Co-management, empowering local communities, offering alternative enforcement and a means of sustainable funding, is a viable solution. We conducted a year- long interdisciplinary evaluation of the co-management approach implemented by an inter-organizational project in Comoros. The Moheli Marine Park, a model for this new approach, offers tangible examples and insight applicable to similar situations. Though co-management is not immune to inter-village conflicts, weakness of government law enforcement, and political instability and unrest, with 80% of the responsibility in the hands of communities, the Moheli Marine Park has made notable strides. Methodical interdisciplinary examination of the strengths and shortcomings of the co-management approach reveal its versatility for application elsewhere and offer a template for effective scientific research and monitoring, policy-making, and management of protected areas in island nations.

BROWN, TRACEY K., Tandora D. Grant, and Allison C. Alberts. Center for Reproduction of Endangered Species, Zoological Society of San Diego, P.O. Box 120551, San Diego, CA 92112, USA (TKB, TDG, ACA).
Biological impacts of human disturbance on two horned lizard species (Phrynosoma) in southern California.
The dramatic urbanization of the southwest over the past century presents a potent threat to horned lizards (PHRYNOSOMATIDAE: Phrynosoma), as several aspects of their biology makes them particularly susceptible to the effects of human development and habitat disturbance. Since 1995, we have studied the foraging ecology and habitat use of San Diego coast horned lizards (P. coronatum blainvillei) occupying undisturbed and disturbed (by fire or grazing) coastal sage scrub habitats. We found Coast horned lizards were significantly smaller and lighter in the grazed habitat, while lizards from the undisturbed habitat had significantly richer and more diverse diets, smaller home ranges and moved shorter distances. In another study, we compared the eco-physiology of desert horned lizards (P. platyrhinos) occupying undisturbed Mojave desert habitat with lizards from an area used by the US Army for military training maneuvers. In contrast to the Coast horned lizard results, desert horned lizards were significantly heavier in the disturbed areas and had higher rates of food intake, but did not differ in home range size from lizards in the undisturbed habitat. Thus the ecological consequences of habitat disturbance for horned lizards likely depends on habitat type and species, and may occasionally be beneficial.

Bruegmann, Marie M., and James D. Jacobi. U.S. Fish and Wildlife Service, 300 Ala Moana Boulevard, Room 3122, Honolulu, HI 96813, USA (MMB), U.S. Geological Survey, Pacific Island Ecosystems Research Center, Kilauea Field Station, P.O. Box 44, Hawaii Volcano National Park, HI 96718, USA (JDJ).
AN INTEGRATED PROGRAM FOR PLANT CONSERVATION IN HAWAII.
The Hawaiian archipelago is one of the most isolated island chains in the world. The large diversity of habitats found here has resulted in the evolution of an extremely unique biota, exemplified by 90% endemism of the Hawaiian flora. However, following human contact, starting approximately 1,500 years ago, many of the Hawaiian ecosystems have been altered and more than 50% of the native flora has either gone extinct, is listed as endangered or threatened, or appears to be declining. Conservation of Hawaiian plant diversity requires the implementation of an integrated strategy of species recovery and ecosystem management that is well coordinated among the various agencies, organizations, and individuals. The U.S. Fish and Wildlife Service's Hawaii and Pacific Plants Recovery Coordinating Committee is facilitating the development of a conservation strategy to help with the definition and implementation of conservation programs to stabilize and restore this unique flora. Major components of the strategy are: 1) initiate emergency actions to keep species from going extinct, 2) implement species and community maintenance and/or restoration efforts to recover the species and the habitats the native plants depend upon, and 3) conduct other activities, such as public education and outreach programs, in support of plant conservation.

Bruner, Aaron G., Raymond E. Gullison, Richard E. Rice, and Gustavo A.B. da Fonseca. Center for Applied Biodiversity Science at Conservation International, 1919 M. St., NW Washington, DC 20036, USA (AGB, RER, GABF), Centre for Biodiversity Research, University of British Colombia, 13810 Long Lake Road, Ladysmith BC, V9G 1G5, Canada (REG)
THE EFFECTIVENESS OF PARKS IN PROTECTING TROPICAL BIODIVERSITY
Over the past two decades, parks have been criticized as unable to protect the biological resources within their borders against growing human pressures, a claim which holds critical implications for conservation strategies. To test the hypothesis that parks are in fact an effective means to protect tropical biodiversity, we gathered and analyzed information on threats, impacts, local conditions and management activities from 93 protected areas in 22 countries. Analyses showed that these parks were largely successful at stopping land clearing, and to a lesser degree effective at mitigating logging, hunting, fire and grazing. Our most basic finding was therefore that tropical parks have been surprisingly effective in addressing major threats, often in a context of significant underfunding. Despite this success, a clear need exists to improve park effectiveness in some regards, perhaps especially against illegal hunting. The most effective parks were characterized by relatively strong basic management activities, such as enforcement, boundary demarcation, and direct compensation to communities. These findings suggest that creating new parks and increasing support for existing parks will both make a significant contribution to long-term biodiversity conservation in the tropics.

Buddenhagen, Christopher, and CLAYSON HOWELL. Department of Conservation, PO Box 10-420, Wellington, New Zealand.
WEEDS DATABASE - WHAT, WHERE, HOW, WHEN.
Dealing with weeds means dealing with all the planet's plants. By their nature, islands are vulnerable to weeds. Efficient weed control demands accessible information: What is the weed? Where is it? How to kill it? This lesson has been learned in New Zealand, a set of islands, with half its flora being exotic. The Department of Conservation (DOC) has developed a database to contain information to address these questions. Much of the important information existed but as scattered gems inaccessable to managers. After ONLY 6 months of activity it ALREADY contains more than 30, 000 observations, 1700 species, 1000 images and 2600 control techniques. Information is retrieved by staff through the DOC intranet using web browsers. This is despite the geographical isolation of many DOC offices. The database is fully interactive. This means users from all over New Zealand can enter new information in any of the fields at any time. Staff are required to enter data, gained through surveillance, monitoring and weed control activities. Data are entered continually, and immediately available to all users, thus supporting DOC's innovative approach to weed management. Ultimately, the database will be made available through the world wide web.

BURBIDGE, MARYANN L., Hugh A. Robertson, Rogan Colbourne, and Allan J. Baker. Centre for Biodiversity and Conservation Biology, Royal Ontario Museum, 100 Queens Park, Toronto, ON M5S 2C6, Canada, and Department of Zoology, University of Toronto, 25 Harbord St., Toronto, ON M5S 3G5, Canada (MLB, AJB), Department of Conservation, P.O. Box 5086, Wellington, New Zealand (HAR, RC).
GENETIC AND FIELD STUDIES SUPPORT SPECIES STATUS FOR THE HIGHLY ENDANGERED OKARITO BROWN KIWI POPULATION IN NEW ZEALAND.
Several decades ago Brian Reid from the Department of Conservation recognized that Brown Kiwis from Okarito on the South Island of New Zealand had different vocalizations from other populations of Brown Kiwi. Subsequently, genetic studies revealed that Okarito birds were genetically closer to the North Island Brown Kiwi than to other South Island populations, and they were therefore classified within the Northern species (Apteryx mantelli). To further determine if the Okarito population should be considered as a distinct species from that of the North Island, we sequenced the control region and the Cytochrome b, ATPase 6 and ATPase 8 genes of mtDNA. Field studies were also conducted on the ecology, behaviour, morphology and parasites of Okarito and North Island Brown Kiwis. The sequence data demonstrate that the remnant Okarito population is genetically depauperate, and constitutes a distinctive phylogenetic lineage 9% divergent from other Brown Kiwi lineages. Field data provide additional support for the evolutionary divergence of the Okarito Brown Kiwi, and along with the genetic analyses warrant its recognition as a new species. With a total population of around 300 birds this species is highly endangered and must be accorded priority conservation status.

BURGESS, SARAH L. Dept. of Genetics and Molecular Biology, Univ. of Hawaii-Mänoa, HI, USA.
PHYLOGENY AND INFERRED COLONIZATION PATTERN OF HAWAII'S ENDEMIC FLYCATCHER, THE ELEPAIO.
The five currently recognized subspecies of elepaio constitute a monotypic genus (Chasiempis) which is endemic to the Hawaiian archipelago. Elepaio from Kaua'i (C. s. sclateri), the geologically oldest of the major Hawaiian islands, are often presumed to be the ancestors of the populations on the youngest island of Hawaii (C. s. sandwichensis, C. s. ridgwayi, and C. s. bryani). However, whether the populations on the youngest island of Hawaii were established by geographically intermediate O'ahu lineages, or whether they were the result of a colonization event that bypassed O'ahu is questionable since the three islands forming the Maui-Nui complex located between O'ahu and Hawaii have neither fossil records nor contemporary populations of elepaio. Two mitochondrial regions totaling approximately 800 bp were sequenced and analyzed using three different methods of cladistic analysis. Phylogenetic estimations based on neighbor-joining, parsimony, and maximum-likelihood cladistics yield evidence of an interesting colonization pattern for this species. Since habitat loss does not sufficiently account for the rapid decline of the endangered O'ahu elepaio, understanding the founding of these populations and examining the genetic implications of how the O'ahu subspecies differs from its stable counterparts on Kaua'i and Hawaii is an essential piece of the elepaio conservation puzzle.

BURGETT, JEFF M., Glenn D. Klingler, David B. Ledig, and John L. Klavitter. U.S. Fish and Wildlife Service, P.O. Box 244, Honaunau, HI 96726, USA.
THE RECOVERY PROGRAM FOR THE ALALA (CORVUS HAWAIIENSIS): PROGRESS AND CHALLENGES.
The alala (Hawaiian crow) is the only extant Hawaiian corvid. Abundant prior to 1900 on the island of Hawaii, it is now critically endangered, with a population of 29 individuals. Suspected causes of decline include mammal predation, shooting, avian disease, and habitat modification and fragmentation from logging and agriculture. Recovery actions began in earnest in 1993 and were focused on the remaining wild flock in South Kona, on the western slope of Mauna Loa. Release of captive-reared juveniles sourced from wild nests was aimed at expanding the population in the wild. Retention of some juveniles in two captive breeding facilities enhanced the size and genetic diversity of the initially small captive flock. The 27 juveniles released over 5 years foraged well, but social integration with wild adults did not occur. Tracking of telemetered juveniles revealed unsuspected mortality factors: infection with novel pathogens and predation by the endangered io (Buteo solitarius). Released alal suffered high mortality but did not breed despite reaching adult age. This first experimental phase was concluded in 1999. Through this recovery program, a viable captive flock has been created that will allow for future experimental releases in protected areas where habitat management has reduced mortality factors.

BURGMAN, MARK A., and Julian Fox. School of Botany, University of Melbourne, Parkville, Victoria, 3010, Australia.
PROBLEMS WITH MINIMUM CONVEX POLYGONS: A CONTRIBUTION FROM THE NCEAS EXTINCTION RISK PROJECT.
In many instances, presence data are the only kind of data available for the estimation of the geographic ranges occupied by species. Minimum convex polygons (convex hulls) are a standard tool for estimating species' ranges. They are used for making area statements and for assessing trends in occupied habitat, and the results contribute to the assessment of the conservation status of species. We show by simulation that estimates are very likely to be biased. The biases may be substantial and increase with sample size, the underlying shape of the species habitat, the magnitude of errors in locations, and uneveness in the spatial and temporal distribution of sampling effort. The errors affect both area statements and estimates of trends. Some of these errors may be reduced through the application of Delauney triangulation, randomisation tests and heuristic rules, but they cannot be eliminated entirely. Other methods for estimating the extent of species' ranges should be preferred.

BURNEY, DAVID A. Dept. of Biological Sciences, Fordham University, Bronx, NY 10458, USA.
CAN PALEOECOLOGY GUIDE RESTORATION? FINDING A LOST WORLD ON KAUA`I.
Developments in paleoecological research and ecological restoration on Kaua`i offer an opportunity to explore the linkages between the two fields. Coring and excavations in the Maha`ulepu cave system highlight a remarkable fossil site preserving a well-dated stratigraphy of vertebrate bones, invertebrate shells, diatoms, pollen, plant macrofossils, ancient DNA, and human artifacts. Integrated site analysis documents the decline and extirpation, after human arrival, of endemic birds and snails and their replacement by a depauperate assemblage of cosmopolitan exotics, and shows that many plant species with highly restricted ranges today were formerly widespread. In combination with results from 11 other dated sites, this work provides evidence for island-wide transformation by human activities and biological invasions, resulting in perhaps three waves of extinction. These probably correlate with initial human arrival, prehistoric overpopulation, and European settlement. Benefits of this research to restoration include: 1) proposals for reintroduction of extirpated taxa that are extant elsewhere; 2) identification of challenges to restoration in the form of missing adaptive zones and trophic interactions; and, 3) recognition of integrated sites as resources for scientific study and public education. Potential for closer collaboration between paleoecologists and restoration biologists is ultimately limited only by the availability of suitable sites.

BURRIDGE, CHRIS P., and John R. Gold. Center for Biosystematics and Biodiversity, Texas A&M University, 2258 TAMU, College Station, TX 77843-2258, USA.
CONSERVATION GENETICS OF CAPE FEAR SHINER (NOTROPIS MEKISTOCHOLAS).
Cape Fear shiner, Notropis mekistocholas, is a small cyprinid fish endemic to the Cape Fear drainage of North Carolina, USA. This species was listed as endangered in 1987, and only five small populations have been documented. Current and potential threats to Cape Fear shiner include the existence of dams and hydroelectric plants, land use changes, and habitat deterioration. All known populations are presently separated by at least one impediment to inter-population movement. A genetic study of this species has been implemented to assess the presence of population structure, inbreeding, and recent changes in effective population size (N_e). This information will be useful for future conservation efforts, which may include supplementation. A suite of 16 dinucelotide repeat microsatellite markers (nuclear-encoded) have been developed for Cape Fear shiner, and samples from three populations provided by the USFWS have been analysed. Assessment of variation in the mitochondrial genome by sequencing and single strand conformation polymorphism (SSCP) analysis has also been proposed.

BUSTAMANTE, JAVIER, Javier Seoane, and Ricardo Díaz-Delgado. Department of Applied Biology, Estación Biológica de Doñana, CSIC, Avda. María Luisa s/n, 41013 Sevilla, Spain.
ARE LAND-COVER MAPS INCORRECT OR IS VEGETATION A POOR PREDICTOR OF BIRD DISTRIBUTION?
Bird species are selective on the vegetation types in which they are found but predictive models of bird distribution based on variables derived from land-use/land cover maps tend to have limited success. In this paper we test the hypothesis that the poor predictive power of some land-use/land-cover maps is partially due to the use of land-cover categories of little relevance to birds. In two areas of 4900 km² of Western Andalusia, Spain, we compared the predictive ability of models derived from the 1995 land-use/land-cover map of Andalusia from the Sinamba (Consejería de Medio Ambiente, Junta de Andalucía) and from a vegetation map built from satellite images for this study. We sampled the presence/absence of bird species at 1144 points using point-count surveys and measured vegetation structure at sample points. Statistical models for each species were built with Generalized Additive Models using as predictors land-cover variables derived from the two land-cover maps. We compared for each bird species the percentage of correct classification of the best model derived from each map. The results show that model predictions can be improved with land-use/land-cover maps built considering explicitely bird vegetation requirements.

BYRD, G. VERNON, Steve Ebbert, and Arthur Sowls. Alaska Maritime National Wildlife Refuge, 2355 Kachemak Bay Dr., Homer, AK 99603, USA.
INTRODUCED PREDATOR REMOVAL: A TOOL FOR RESTORING THE ENDANGERED ALEUTIAN CANADA GOOSE.
The Aleutian Canada goose (Branta canadensis leucopareia) was nearly extinct by the 1930s largely due to predation by introduced arctic (Alopex lagopus) and red foxes (Vulpes vulpes) on their remote nesting islands in Alaska and Siberia. The nesting islands historically had no native mammalian predators. When foxes were introduced for fur production, they quickly extirpated geese and other species of native birds. Most of the historic nesting islands are now part of Alaska Maritime National Wildlife Refuge, and a project has been underway since 1949 to eradicate introduced foxes to restore native biological diversity. By 2000, foxes had been removed from 38 islands totaling more than 400,000 (ha), and Aleutian Canada geese had increased from less than 1,000 birds to more than 35,000. This increase, along with protective measures at wintering areas in California, justified removing the goose from the endangered species list. Other native birds, particularly seabirds and endemic taxa of terrestrial birds, also have increased 2 to 10 fold on islands where recovery has been studied.

CABEZA, MAR, and Atte Moilanen. Department of Ecology and Systematics, Division of Population Biology, P. O. Box 17 (Arkadiankatu 7), FIN-00014 University of Helsinki, Finland (MC, AM).
DESIGN OF RESERVE NETWORKS AND THE PERSISTENCE OF BIODIVERSITY.
Site-selection algorithms are used in reserve design to identify networks of sites that maximize biodiversity given some constraints (e.g. cost, total area, etc.). These algorithms are mostly based on a snapshot of species occurrence and they ignore the question of how well species persist in the selected sites. I will present a theoretical approach that considers spatiotemporal dynamics in order to assess the effectiveness of site-selection algorithms in terms of species persistence. Once a set of sites is selected by a common algorithm, we apply a metapopulation model to simulate species occurrence through time. We look at the species extinctions at the selected sites in two situations: first, when all sites are included in the simulation, and second, when the non-selected-sites are excluded from the simulation (i.e. assumed to be destroyed if they do not become a reserve). The distinction is done to emphasize the importance of the non-selected sites for the spatial dynamics of the system. The results show that species persistence may be strongly dependent on patches not included in the selection. In summary, our results support the call for the integration of spatial population modelling in reserve network design.

CABIN, ROBERT J., Susan Cordell, Stephen G. Weller, and Lisa J. Hadway. USDA Forest Service, 23 E. Kawili St., Hilo, HI, USA 96720 (RJC, SC), Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697, USA (SGW) and Hawaii State Department of Land and Natural Resources, P.O. Box 4849, Hilo, HI 96720, USA (LJH).
DRY FOREST RESTORATION IN HAWAII.
On the island of Hawaii, remnant native dry forests are dominated by fountain grass (Pennisetum setaceum), a highly invasive exotic bunchgrass. To investigate cost-effective methods for re-establishing native plant populations in these areas, we established 8 experimental blocks in a highly degraded treeless area dominated by fountain grass. Each block consisted of 4 plots randomly assigned to 1 of 4 fountain grass control treatments: control, weed whack/herbicide, weed whack/black plastic mulch, and bulldoze. We also erected 50% shade cloth structures over 4 blocks to simulate light conditions found under intact native forests. We divided each plot into 4 quadrats: 1 outplant quadrat, 1 direct-seed quadrat, and 2 control quadrats. At the end of the experiment 2 years later, we found that the percent cover of native plants was greatest in the bulldozed plots, intermediate in the weed whack/herbicide and weed whack/plastic treatments, and least in the control plots. The percent cover of native species was also greater in the shaded vs. full sun plots, and greatest in the outplanted quadrats, intermediate in the seeded quadrats, and least in the control quadrats. These results suggest that relatively cheap and efficient techniques can be employed to restore Hawaii's native dry forests.

CALHOUN, ARAM J.K., and Phillip deMaynadier. University of Maine, 5722 Deering Hall, Orono, ME 04469, USA (AJKC), Maine Department of Inland Fisheries and Wildlife, 650 State St., Bangor, ME 04401, USA (Pd)
Conserving pool-breeding amphibians: Best management practices (BMPs) for timber harvesting and development near vernal pools.
Vernal pools are poorly regulated because of their small size (typically less than ½ acre) and seasonal hydroperiod. The cumulative loss of small, isolated wetlands threatens populations of obligate pool-breeding amphibians and invertebrates as well as facultative wildlife. One of the most effective tools for conserving small, isolated wetlands is citizen conservation focused on responsible landuse practices. In the northeastern United States, Best Management Practices (BMPs) for timber harvesting near vernal pools have been extremely general and lack prescriptive guidelines. We have developed comprehensive voluntary BMPs for timber harvesting that provide specific management recommendations. These BMPs were developed with input from major stakeholders and reflect the best available science on breeding habitat and landscape needs of pool-breeding amphibians in our region. Three management zones are recommended: (1) the vernal pool depression, (2) the vernal pool protection zone (100 ft around the pool), and (3) the amphibian life zone (100 to 500 ft). For each zone, a description, management objective, rationale, and specific recommendations are provided. These BMPs are applicable to harvesting throughout the northeastern United States. BMPs for residential development, developed by scientists throughout the northeastern United States, reflect the needs of pool-breeding amphibians in fragmented, developing landscapes.

CAMPBELL, BRIAN D., and Jay R. Malcolm. Faculty of Forestry, University of Toronto, 33 Willcocks Street, Earth Sciences Centre, Toronto, ON M5S 3B3, Canada.
HOW IMPORTANT IS OLD-GROWTH?: RESPONSES OF BREEDING BOREAL BIRD SPECIES TO FOREST AGE AND LOGGING DISTURBANCE.
Although successional stage preferences are relatively well documented in Ontario's boreal mixedwood forest breeding bird community, little is known about the sensitivity of this group to changes in stand age as a result of disturbance by fire or logging. We surveyed boreal forest bird communities using point counts in 18 post-harvest and 16 post-fire, closed-canopy mixedwood sites in northeastern Ontario that varied in stand age from 29-98 years. Differences in species habitat occupancy with respect to age and disturbance origin independently of age were compared using logistic regression models. Several canopy-foraging Parulid warblers and one cavity nesting resident, Red-breasted nuthatch (Sitta canadensis), showed threshold relationships with stand age, indicating potential sensitivity to changes in stand age relative to other species. Few species showed significant effects of disturbance origin using stand age as covariate in the model. Our results highlight the importance of older boreal mixedwood stands in conserving habitats for bird species that are showing negative population trends elsewhere in North America.

CAMPBELL, DAVID G. Dept. of Biology, Grinnell College, Grinnell IA 50112, USA.
ARE THE FORESTS OF THE EASTERN PETÉN ANTHROPOGENIC?
In order to examine the relationship between patterns of human settlement during the Classic Maya Civilization (1,100 YBP) and the contemporary tropical forests of the Petén of eastern Belize, we conducted quantitative inventories of all woody stems > 1.5 cm. DBH on three forested sites: El Pilar, Terra Nova and Ix Chel. The sites have been abandoned since Classic times, and represent a gradient of Classic settlement densities (200, 50 and 2 structures km-2). All three sites had low species richness (94, 89 and 103 species, respectively, at asymptote on a species-area curve), and low beta diversity (Sorenson's index of similarity of between 0.58 - 0.71). Although the species richness of the three sites was independent of the density of structures, the total basal areas of the forests, as well as the percentage of vines (in terms of both individuals and species), were inverse functions of settlement density. An enumeration of the species of economic value to the Maya suggests that the oligarchies of these forests are the result of high grading over the past 4,000 years by the Maya. We therefore conclude that the forests of the eastern Petén are anthropogenic.

CAMPBELL, EARL W. USDA - National Wildlife Research Center, Hawaii Field Station, P.O. Box 10880, Hilo, HI 96721, USA.
FEDERAL REGULATory ISSUES AFFECTING ExotiC VERTEBRATE CONTROL IN INSULAR AREAS IN THE USA.
The intention of this presentation is to familiarize the audience with federal regulatory and registration issues that impact proposed exotic vertebrate species control efforts for conservation purposes in insular areas the United States. Several techniques that have been successfully used for vertebrate control (e.g., toxicant use patterns) for endangered species restoration in insular areas outside of the United States are not approved for use (or have limited approval) in this country. Federal, state, and non-governmental wildlife management agencies are investigating methods to gain regulatory approval for several conservation-oriented rodent control techniques using toxicants in insular areas in several sites in the United States. Multi-agency efforts in Hawaii to develop chemical control techniques for introduced rats and Eleutherodactylus frogs will be used to illustrate federal regulatory issues. The use of specialized techniques to control vertebrate exotics in insular areas in the United States is realistic but will need to be conducted or developed in a manner that meets current federal regulations. Wildlife management agencies in the United States must realistically consider research costs to gain federal regulatory approval of novel control techniques.

CAMPBELL, EARL W., Gordon H. Rodda, Fred Kraus, and Thomas H. Fritts. USDA - National Wildlife Research Center, Hawaii Field Station, P.O. Box 10880, Hilo, HI 96721, USA (EWC), USGS - Biological Resources Division, 4512 McMurry Avenue, Fort Collins, CO 80525, USA (GHR, THF), Hawaii Department of Land and Natural Resources, 1151 Punchbowl St., Room 325, Honolulu, HI 96813, USA (FK).
MANAGEMENT AND PREVENTION OF ALIEN REPTILES AND AMPHIBIANS: BrOWN TREESNAKES AND CARIBBEAN TREE FROGS.
The introduction of alien terrestrial reptiles and amphibians to new localities worldwide has gained increasing public and scientific attention in recent years. Much of this attention has been focused on the establishment of the brown treesnake (Boiga irregularis) on the island of Guam and this introduced predator's subsequent ecological, economic, and societal impacts to the island. Concerns over the introduction of this pest snake to Hawaii coupled with the recent introduction of a relatively conspicuous species of Caribbean tree frog (Eleutherodactylus coqui) continues to make the impacts of alien reptiles and amphibians an issue in this state. This is a realistic concern when one considers that the Hawaiian Island chain originally had no herps and currently has 30 species of established alien herps. The rate of alien reptile and amphibian species introductions to the state continues to increase at a steady rate while this pattern is stable for all other vertebrate taxa. We will review current issues associated with management and prevention of brown treesnakes and Caribbean tree frogs and discuss herp management issues in Hawaii as they relate to the observed synergistic ecological relationship between the highly fecund smaller introduced herps and the brown treesnake on Guam.

CAMPBELL, KARL J., and Felipe Cruz. Charles Darwin Research Station and the Galápagos National Park Service, Casilla 17-01-3891, Quito, Ecuador.
FERAL GOATS IN THE GALÁPAGOS ISLANDS: PRESENT SITUATION, ERADICATION EFFORTS, AND THE FUTURE.
Feral goats (Capra hircus) are present on all five inhabited and one uninhabited island within the Galápagos Archipelago. Eradication has been successfully conducted on six uninhabited islands ranging in size from 12 to 12,996 ha. The effects of eradicating goats have been dramatic, with bare earth being converted to native forest in some cases. Pinta Island (5940 ha) is the most recent successful eradication, being declared goat free in 2000 after the removal of over 41,000 animals. Goat eradication is currently being conducted on uninhabited Santiago Island (58,465 ha) which has a large well-established population. Plans for eradication on Isabela Island (458,812 ha) are awaiting funds for implementation, which have currently been approved. Isabela's goat population is growing rapidly and continues to expand its range. The campaign on Isabela is expected to start in 2002, and last for 5 years, while Santiago is expected to be eradicated within 5 years. Eradication or low-level control are planned for the remaining four inhabited islands. Recent advances in the Judas goat technique, involving sterilization and increased estrus behavior, are expected to facilitate eradication success.

CAMPBELL, STEVEN P., Shelly L. Thomas, Susan E. Hayhurst, and Regina Purtell. Department of Wildlife Ecology, 5755 Nutting Hall (SPC, RP), Department of Biological Sciences, 202 Deering Hall (SLT), School of Marine Sciences, 217 Murray Hall (SEH), University of Maine, Orono, ME 04469, USA.
SCIENTISTS AS EDUCATORS: N.S.F. GRADUATE TEACHING FELLOWS IN K-12 EDUCATION AT THE UNIVERSITY OF MAINE.
In an increasingly human-dominated landscape, public understanding of environmental issues is especially important. Consequently, conservation biologists will have to play a more active role in explaining why biological concerns must be considered in decisions on how to use land and natural resources. Here we provide information on a National Science Foundation sponsored program at the University of Maine that helps to meet this need. The program underwrites fellowships for 12 students (ten graduates and two undergraduates) to demonstrate science in local K-12 classrooms. A major goal is to enhance science education of K-12 students by providing expertise, equipment, activities, and role models that would not otherwise be available to teachers and students. In the first year of the program, 28 teachers from 12 schools and four districts have participated. Although the program enhances learning in many areas of science, several fellows have presented units relevant to conservation biology. These include diversity of life, effects of forestry practices on biodiversity, science and public policy, food webs, alternative energy sources, global warming, acid rain, and oil spills. We provide a synopsis of the different objectives of each unit and the corresponding activities that were used to illustrate them.

CARRIZOSA, SANTIAGO. Genetic Resources Conservation Program, University of California, One Shields Avenue, Davis, CA 95616-8602, USA.
ARE BIODIVERSITY-RICH COUNTRIES AND BIOPROSPECTING INITIATIVES COMPLYING WITH THE CONVENTION ON BIOLOGICAL DIVERSITY?
In 1992, the Convention on Biological Diversity (CBD) recognized the sovereign rights of countries to control the use of their resources. The CBD was also called the biotrade convention and encouraged countries to develop laws to regulate access to genetic resources. Thus, scientists that are collecting organisms with useful properties had to modify their bioprospecting approaches. Today, these groups must take into account not only the principles stated by the CBD, but also the obligations of national laws that regulate access to genetic resources. Based on two conceptual frameworks, I analyze whether biodiversity-rich countries and bioprospecting groups are complying with the CBD. The results indicate that access legislation has been difficult to implement in some countries. It has become a barrier not only for bioprospecting groups that have commercial interests, but also for scientists who are driven by basic research purposes. Current bioprospecting groups have also failed to propose benefit-sharing packages that ensure the equitable distribution of benefits derived from biodiversity. Technologies transferred by these groups have not met the expectations of local scientists and proposed conservation initiatives need to be improved. Therefore, it can be argue that some countries and bioprospecting groups are not fully complying with the CBD.

Carroll, Carlos, REED F. NOSS, and Nathan H. Schumaker. Klamath Center for Conservation Research, Orleans, CA 95556, USA (CC), Conservation Science, Inc., Corvallis, OR 97330, USA (RFN), US Environmental Protection Agency, Corvallis, OR 97333, USA (NHS).
THE GREATER YELLOWSTONE ECOSYSTEM AS A HABITAT ISLAND FOR MAMMALIAN CARNIVORES: DESIGNING A MULTI-SPECIES RESERVE NETWORK USING INDIVIDUAL-BASED SPATIAL MODELS.
We developed spatial habitat models for the Greater Yellowstone Ecosystem (GYE) for several large and medium-sized mammalian carnivores, including grizzly bear, gray wolf, and lynx. The models, which were constructed using resource selection functions, were used to parameterize a spatially-explicit individual-based model, PATCH, which predicted population viability under current and potential future landscape conditions. Even in species with similar habitat associations, contrasting demography and social structure resulted in divergence in critical regional-scale habitat requirements. For the wolf, incorporating pack structure into the simulation model increased resilience to human-associated landscape change. When added to the species' high vagility, this resulted in effective demographic linkages between the GYE and adjacent regional populations. Similarly for the lynx, relatively low levels of population cycling were found to greatly increase extinction risk when the region was isolated from boreal lynx populations, emphasizing the dependence of the GYE population on broader-scale dynamics. In contrast, regional grizzly bear population viability in the model was most influenced by incremental habitat degradation of potential sink habitats within the GYE. We integrated these contrasting results into a regional multi-species reserve design for the GYE using both the reserve selection program SITES and the PATCH model.

CASSEY, PHILLIP. Australian School of Environmental Studies, Griffith University, Nathan - Brisbane, Queensland 4111, Australia.
RELATIVE SUCCESS OF INTRODUCED LAND BIRDS ON CONTINENTS AND ISLANDS.
It is commonly reported that introductions of species among different regions have a disparate rate of success. For instance, many researchers state that introductions to oceanic islands are more successful than those to continental mainland regions. This pattern has attracted considerable interest, and several ecological mechanisms have been proposed as its cause. However, the validity of this pattern has rarely been questioned or empirically tested. If some fraction of the disparity simply results from more species being transported to one region than to another, then the observed difference is an artifact rather than an ecological phenomena. Here I test this idea using global land birds, and I show that there is no difference in the success of introduced species between island and mainland introductions. However, introductions within biogeographic regions have been more successful than those between regions and both island size and native geographic range are associated with increasing introduction success.

CASTAÑEDA G., GABRIELA. Departamento de Ciencias Biológicas, Pontificia Universidad Católica del Ecuador, 12 de Octubre y Roca, Apt. 17-01-2184, Quito, Ecuador.
THE IMPORTANCE OF EPIPHYTES IN BIRD COMMUNITIES: THE CASE OF ARACEAE AND BROMELIACEAE IN A CLOUD FOREST.
Epiphytes provide a variety of resources exploited by birds: fruits, flowers, seeds, water, invertebrates (e.g., in bromeliads "ponds"), nesting materials and nest sites. Here I document the use by birds of various resources provided by four species of bromeliads and three species of aroids in an Ecuadorian cloud forest. I recorded eleven species of birds visiting bromeliads during 580 hours of observation: eight species of hummingbirds visited bromeliads to take nectar and water, and one barbet, one furnariid and one woodcreeper took invertebrates from the clumps of dead leaves or among live leaves. For the aroids (230 hours under observation), I noted that one species of guan and three tanagers feeding on its fruits. While some bird species showed a tight association with the observed epiphytes, the bird-epiphyte interactions did not appear to be highly species-specific. The variety of resources provided by epiphytes to a number of bird species suggest that they may be important contributors to the maintenance of cloud forest bird communities. Furthermore, the bird species recorded may be important seed dispersers and pollination agents for the observed epiphytes.

CASWELL, HAL, and Takenori Takada. Biology Department MS-34, Woods Hole Oceanographic Institution, Woods Hole MA 02543, USA (HC), Department of International Cultural Relations, Hokkaido Tokai University, Minami-ku, Sapporo 005, Japan (TT).
THE RELIABILITY OF ELASTICITIES OF POPULATION GROWTH RATE.
Demographic models are used in conservation biology to assess population status, to diagnose the causes of population problems, to prescribe management tactics, and to make prognoses of population viability. The sensitivities and elasticities of population growth rate play an important role in these calculations. Sensitivities and elasticities describe infinitesimal perturbations of linear models, but conservation applications involve large perturbations of populations that are subject to density-dependent nonlinearities. We report here a series of analyses to examine the reliability of the usual elasticity calculations. We constructed matrix population models including density-dependent adult survival, juvenile survival, fertility, or development rate, and compared the linear and nonlinear elasticity calculations. The linear calculations are reliable until the nonlinearities become so strong that the population exhibits severely chaotic behavior. We also examined the effects of making large perturbations of multiple parameters in several published projection matrices. We show that the elasticities do an excellent job of predicting the results of such perturbations, non-infinitesimal though they may be. These conclusions provide additional justification for the use of demographic analysis in conservation biology, especially in situations where only limited population data are available.

CEGELSKI, CHRISTINE, Lisette P. Waits, and Neil Anderson. Department of Fish and Wildlife Resources, University of Idaho, Moscow, Idaho 83843, USA.
MITOCHONDRIAL AND MICROSATELLITE ANALYSES OF WOLVERINE POPULATION STRUCTURE IN MONTANA.
The wolverine (Gulo gulo) is the largest of the terrestrial mustelids and least known. Limited scientific knowledge combined with its elusive nature has made the wolverine a management and conservation enigma. The wolverine has declined across most of its range in North America, particularly in the lower 48 states. The largest and most viable population of wolverines in the lower 48 states is in Montana. We are currently assessing genetic variability, gene flow, and population structure of the Montana wolverines using mtDNA sequencing and nuclear DNA microsatellite analysis. A 377 base-pair segment of the mtDNA control region was amplified and sequenced in 67 individuals. Four variable sites defined four haplotypes (h=,4419) with pairwise nucleotide differences of p = .0033. Pairwise Fst estimates indicated that there is some restricted female gene flow, however an AMOVA analysis failed to detect significant structuring across the state. In comparison, wolverines in the Northwest Territories of Canada had higher levels of haplotype diversity (H= .636 to .758) and equivalent nucleotide diversity ( p = .0018 to .0055). We are currently completing nuclear microsatellite analysis using 10 loci to evaluate genetic diversity and gene flow.

CERASALE, DAVID J., Luanne Johnson, Rachel A. Rounds, T. Colleen Murray, and Paul C. Banko. USGS-BRD, Pacific Island Ecosystems Research Center, P.O. Box 44, Hawaii National Park, HI 96718, USA (DJC, LJ, TCM, PCB), Department of Environmental Science, University of Virginia, 2450 McCormick Road, Charlottesville, VA 22903, USA (RAR).
NEST SITE SELECTION OF THE ENDANGERED PALILA (LOXIOIDES BAILLEUI) IN A RECOVERING HABITAT.
Understanding nest site selection of endangered birds is important in assessing habitat requirements and recommending recovery strategies. The palila (Loxioides bailleui), a Hawaiian honeycreeper, is concentrated in <30 km² on Mauna Kea's western slope following decades of habitat degradation by ungulates. While previous studies characterized nest sites in mamane (Sophora chrysophylla) prior to (1971-1974), and shortly after (1988) large-scale ungulate removal, this study (1999-2000) focused on nest site selection in the same habitat after more than a decade of low browsing pressure. We found palila nesting higher (4.3m ± 0.2 vs. 3.9m ± 0.1) and in taller trees (6.0m ± 0.1 vs. 5.2m ± 0.1) than in past decades. Nest trees were significantly taller than trees available in the surrounding habitat. Contrary to previous studies, we found no correlation between nest site characteristics and nest success, suggesting that factors influencing nest fate in recovering habitat may vary over time. Palila also nested in previously undocumented tree and shrub species, and may increasingly use these species as well as large, healthy mamane trees for nesting as the forest recovers.

CHAVAN SHIVAJI GANPAT, and Anjana Pant. Wildlife Institute of India, P.B.No.18, Chandrabani, Dehradun-248001, Uttaranchal, India.
INTEGRATING gis AND REMOTE SENSING FOR Ecological analysis of the Corbett Landscape, INDIA.
Vegetation mapping has been carried out using IRS-1B data. Patch characterization parameters like fragmentation, patch density, juxtaposition, interspersion etc have been studied and presented in this paper. Patch sizes are significantly different in each vegetation type. Sal, Sal-mixed, mixed-sal, miscellaneous and Mixed-bamboo/Bamboo-mixed show higher patch size while pure bamboo shows the minimum average patch size. The dominant forest types viz. Sal, Sal-mixed, mixed-sal, miscellaneous and Mixed-bamboo/Bamboo-mixed show a very high index of shape demonstrating a very high irregularity of the edges of these patches. Interspersion image shows a fairly uniform distributed variation throughout the study area. However central part of the patch shows the homogeneity status, which is mainly due to the presence of pure sal and sal-mixed forests. 80% of the total area is intact. Only 0.35% of the total area towards the southern part shows high interspersion and can be attributed to the presence of varying forest types which are located on a highly varying physiography. The highest juxtaposition index was given by sal forest which accounts for nearly 57% of the total area followed by the non-forest class representing about 26% of the area. It was found that 92.5% of the area has low patch density.

CHAVES, JAIME A. Departamento de Ciencias Biológicas, Pontificia Universidad Católica del Ecuador, Apto. 17-01-2184, Quito, Ecuador.
IDENTIFYING HABITAT CHARACTERISTICS AND TREE PREFERENCES OF FRUGIVOROUS BIRDS AS A MODEL FOR BIRD AND HABITAT CONSERVATION.
Identifying habitat associations of bird communities is a crucial first step in conserving avian diversity and -- in the case of important seed dispersers - in conserving the plant communities on which they depend. A one year study of 13 frugivorous birds in the cloud forests of Ecuador revealed a high degree of both habitat selection and preferences for certain trees. Correspondence Analysis (CA) was performed to determine the habitat variables and tree associations from point counts. Canopy height, tree densities and tree architectures were the variables selected by the model. Small species (tanagers) were found to prefer foraging in lower tree densities, medium size species (tucanets and mountain tanagers) were found in habitats with low canopy height (< 10m) and large frugivorous birds (toucans) were correlated with high tree densities. Bird species selected fruiting trees preferentially with the strongest associations between toucans and Lauraceous trees and guans and Rubiaceous trees. Ficus and Cecropia fruits were used by many species and likely represent a keystone resource for avian frugivors. Results emphasize the importance of identifying the fundamental associations between frugivores and their foods in conservation planning.

CHEN, JIQUAN, Kimberly D. Brosofske, Eugenie S. Euskirchen, Radley Z. Watkins, Treneice J. Marshall, Sari C. Saunders, Thomas R. Crow, and Eric J. Gustafson. Michigan Technological University, Houghton, MI 49931, USA (JQ, ESE, RZW, TJM, SCS), University of Rhode Island, Kingston, RI 02881, USA (KDB) and USDA Forest Service NCRS, Rhinelander, WI 54501, USA (TRC, EJG).
CONTRIBUTION OF LANDSCAPE ELEMENTS TO PLANT DISTRIBUTION IN A MANAGED LANDSCAPE.
The patch-corridor-matrix framework has been widely used in describing landscape structure. Two critical questions are: (1) Should we incorporate additional landscape elements to more conclusively describe the structure?, and (2) How much does each element contribute to the cumulative richness and abundances of plant species? To answer these questions, we conducted inventories of plant species from 1994-2000 within a 28x33 km managed landscape in northern Wisconsin. Of the 333 species detected, 98 (29.4%) were found in all 20 patch types, including the riparian and edge zones. No species were unique to any patch type, but 32 species were restricted to the same eight patch types. The majority of exotic species were found in all 20 patch types, with the exception of bog interior. Our previous studies indicated that past management activities had fragmented the study site to a landscape dominated by areas influenced by edges and roads (>50%). However, we found no species is unique to these patch types. We conclude that a miss-matched relationship exists between species and habitats in this landscape; and the traditional species management approach based on habitat type needs to be revisited.

CHOW, NOLA, Jim Douglas, and Kathy MacKinnon. Environmentally and Socially Sustainable Development Network, The World Bank Group, 1818 H Street, Washington, D.C. 20433, USA.
NATURAL RESOURCE CONSERVATION IN PAPUA NEW GUINEA AND SAMOA WITH THE WORLD BANK.
Papua New Guinea still has 33 million ha of closed natural forest, however logging and forest degradation have been increasing problems. Also, the forest lands are essentially all owned and controlled by local indigenous groups. The World Bank will be assisting the government program to conserve forest resources with a multi-faceted conservation and development project that will have a landowner decision-making unit, conservation trust fund, sustainable forest management, and environmental assessment and monitoring components. With dual goals of improving rural livelihoods and sustainable management and conservation of forest resources, this project involves all levels of society, and goes beyond the existing policy and institutional framework. A cross-sectoral aspect requires that changes to current forest policy occur as a pre-condition for a structural adjustment loan. The project will utilize both participatory, and non-governmental mechanisms in addition to the more traditional government agency institution building. Similarly, other recent World Bank projects have been focused on resource conservation and livelihood improvements through traditional and non-traditional project designs.

CHRISTIANSEN, SARAH, Laura Valutis, Jonathan Loh, Pete Coppolillo, Adam Henson, Jen Divis, and Mohammed Bakarr. WWF-International, Avenue du Mont-Blanc, 1996 Gland, Switzerland (JL), The Nature Conservancy, 2404 Bank Drive, Suite 314, Boise, ID 83705, USA (LV), World Wildlife Fund, 1250 24^th Street NW, Washington, DC 20037, USA (SC), Wildlife Conservation Society 2300 Southern Boulevard, Bronx, NY 10460, USA (PC), African Wildlife Foundation 1400 16^th St. N.W. Suite 120, Washington, DC 20036, USA (AH), Enterprise Works Worldwide, 1828 L Street, NW Suite 1000, Washington, DC 20036, USA (JD), Conservation International, 2501 M Street NW Suite 200, Washington, DC 20037, USA (MB).
Large-scale conservation planning - Who's doing what where?
Several international conservation organizations have individually identified and prioritized globally important areas for biodiversity conservation efforts on a large geographic and temporal scale (e.g. World Wildlife Fund, Conservation International, The Nature Conservancy, Birdlife International, etc.). While these analyses reflect a diversity of criteria and scale, they also show significant overlaps in areas that are - or should be - global priorities for biodiversity conservation. To help understand how we are meeting the ambitious but necessary goals reflected in these global priorities, information from several international conservation organizations is synthesized to understand the larger and landscape scales and areas each organization is currently active in. This poster presentation helps to identify existing or potential gaps in conservation programs in relation to global priorities as well as potential opportunities for greater collaboration. A general overview of each organizations' mission and strategies for large-scale planning is provided. This overlay map and data will be used to inform and promote biodiversity conservation in areas where it is lacking and catalyze greater collaboration amongst existing efforts to avoid duplication and redundancy. The future of global conservation depends on this synergy.

CIMO, LAURA FAITEL, and Tracy Dobson. Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI 48824, USA.
GREAT LAKES FISHERY POLICY: ASSESSING LAKE TROUT CONSERVATION AND SOCIAL CONFLICT RESOLUTION.
Despite its historic role as a top predator, native lake trout stocks (Salvelinus namaycush) in the Laurentian Great Lakes have diminished, presenting significant challenges for fisheries conservation. Since 1985, lake trout conservation and management in the Great Lakes has been governed by a court-imposed fisheries management policy--the 1985 Consent Order--due to growing violence between treaty-right Native commercial fishers and non-Native, state-licensed sport fishers over competition for the Great Lakes fishery. This policy utilized a variety of unique provisions to promote conservation and rehabilitation of lake trout while reducing social conflict. To assess the effectiveness of the 1985 Consent Order at achieving its goals, catch data of lake trout were examined and individual, in-depth interviews were conducted with members of the following stakeholder groups: tribal and state biologists; tribal commercial fishers; state-licensed sport fishers; and tribal, state and sport fishing representatives. Under this policy, lake trout abundance increased and violence decreased. However, unsuccessful lake trout fry recruitment and survival remain obstacles for conservation. Proposed reasons for failure include lack of scientific decision-making, TAC limit enforcement, and regulatory flexibility. Importantly, fishing opportunities have diminished for some Native small-boat fishers, which may lead to further conflict in the future.

CLARK, TIMOTHY B. Department of Wildlife & Fisheries Sciences, Texas A&M University, College Station, TX 77845-2258, USA.
MOLECULAR PHYLOGYENY AND THE CONSERVATION OF THE MANTA RAY, MANTA BIROSTRIS.
Manta rays have recently been listed as a protected species in Baja and the Philippines due to conflict between local fisheries and recreational divers who want to protect this charismatic megafauna for eco-tourism. Despite their protected status, manta rays continue to be taken in remote regions. The manta ray is a circumtropical, vagile marine species capable of unlimited dispersal, however photo-identification studies reveal that many individuals are highly site specific. Limited dispersal of individuals would make manta rays highly susceptible to fishing pressure. Molecular sequence data was used to determine the amount of gene flow between locations in the Pacific Ocean and the Gulf of Mexico. While significant population structure was revealed between the Eastern and Western Pacific samples, no structure was found within either region. The low nucleotide sequence divergence within the Western Pacific suggests a recent founding of the Western Pacific from an unknown source. While protection for this species is warranted based on photo-identification studies, molecular analysis was not able to detect limits to gene flow within the Eastern or Western Pacific.

CLEARY, DANIEL. Institute for Biodiversity and Ecosystem Dynamics, P.O. Box 94766, 1090 GT Amsterdam, The Netherlands.
Community response to large scale disturbance in Bornean butterflies.
During severe El Niño Southern Oscillation events large-scale fires may severely effect tropical lowland rainforests. Although the effects of fires are poorly known a large proportion of remaining forests have already been affected. Both logging and fires increase the probability of future fires. When a landscape is affected by fire or logging we expect a breakdown of the original habitat. Other research has shown that original habitat breaks down from a single continuous patch to several patches when 60% of the original habitat is still present. In landscapes with a low proportion of original habitat (< 20%) additional loss will result in an exponential increase in distance between patches of original habitat. Patch isolation and size are not linearly related to the proportion of original habitat, but experience rapid changes at critical proportions of original remaining habitat. Butterflies make excellent bio-indicators for forest disturbances. Data is presented on community response to single and multiple disturbance events between ecotypes within an ecoregion and between ecoregions with different disturbance histories. There is a positive response (species richness) due to a single disturbance event within an ecoregion, but a dramatic negative response as an ecoregion is subjected to additional large scale disturbance events.

COLLEVATTI, ROSANE G., Dario Grattapaglia, and John D. Hay. Pós-Graduação em Biotecnologia Genômica, Universidade Católica de Brasília, SGAN 916, Brasília, DF. 70790-160, Brasil (CRG, DG), Departamento de Ecologia, Universidade de Brasília, Brasília, D.F. 70910-900, Brasil (JDH).
PHYLOGEOGRAPHY AND CONSERVATION GENETICS OF THE ENDANGERED BRAZILIAN TREE SPECIES CARYOCAR BRASILIENSE Camb. (CARYOCARACEAE).
Caryocar brasiliense is an endangered Cerrado tree species, with high ecological and economic importance. Cerrado fragmentation, the commerce of fruits, and the high frequency of fires due to agricultural practices have been affecting population dynamics of the species. We report the population genetic structure and phylogeography of C. brasiliense, based on variability at nuclear and chloroplast microsatellite loci, and discuss the potential consequences for conservation. Chloroplast based network showed four distinct clusters, but major breaks among populations were not related to current geographical barriers. Low haplotype diversity within populations were found for cpDNA polymorphisms, in sharp contrast to the high diversity within and modest differentiation between populations found for nuclear markers. Our results support the hypothesis of restriction of ancient populations to moist refugias during extended droughts coinciding with glaciations in the Northern hemisphere and a subsequent spread to favorable areas throughout Central Brazil. Seed dispersal has been restricted to local populations, and gene flow occurred mainly by pollen movement, due to the extinction of dispersers and current isolation of populations. The comparison of the distribution of genetic variability in the nuclear versus cpDNA microsatellites allows an in depth understanding of the genetic structure and provide important information for conservation.

COOPER, LONI, and Jan A. Randall. Department of Biology, San Francisco State University, 1600 Holloway Avenue San Francisco, CA 94132, USA (LC, JAR).
SPATIAL DYNAMICS OF THE GIANT KANGAROO RAT (DIPODOMYS INGENS): A PROPOSAL FOR MANAGEMENT AND TRAPPING.
The Carrizo Plain Natural Area (CPNA), California is not only one of the last refuges for many federally listed endangered species but also a representation of what California's San Joaquin Valley once was. The Giant kangaroo rat (Dipodomys ingens), a keystone species, is one of several federally listed species on the CPNA. Information on the spatial organization of D. ingens is limited. We radio tracked 20 animals in summer of 1999 and 17 animals in winter 2000 to assess the spatial dynamics of this solitary rodent and provide recommendations for trapping and management strategies. In the summer, D. ingens remained predominately at their home burrows. During the winter breeding season, males visited female's burrows and home range overlap increased. Female average home range sizes during both seasons were .03 ha. Males averaged .02 and .10 ha for summer and winter, respectively. Nearest neighbor distances were 2-5m. To gain accurate census estimates we recommend trapping at active burrows or, when trapping in lines, set traps every 5 meters. Since D. ingens occurred at healthy densities during both field seasons, (48 animals/ha) we recommend these results as the model for space requirements of this species.

COPPOLILLO, PETER B., Eric W. Sanderson, Amy Vedder, Robert B. Wallace, R. Lilian E. Painter, Humberto Gomez, Jeffery P. Jorgensen, Amanda B. Jorgensen, Bryan Curran, Fiona Maisels, Paul Elkan, and Sarah Elkan. Wildlife Conservation Society, International Conservation, 2300 Southern Blvd., Bronx NY 10464-1090, USA, (PBC, EWS, AV), Wildlife Conservation Society, Casilla 3-35181, San Miguel, La Paz, Bolivia (RBW, RLEP, HG), Wildlife Conservation Society, Mariscal Foch 635 Y Reina Victoria, Oficina 2B, Casilla 17-21-168, Quito, Ecuador (JPJ, ABJ), Wildlife Conservation Society, B.P. 14537 Brazzaville, Congo (BC, FM, PE, SE).
USING THE LANDSCAPE SPECIES APPROACH TO FOCUS CONSERVATION ACTION: DATA FROM CONGO, ECUADOR AND BOLIVIA.
The landscape species approach provides a wildlife-based framework for site-based conservation in heterogeneous landscapes. Conservation actions are prioritized by identifying spatially and temporally explicit intersections (and conflicts) between human activities and the requirements of landscape species. Suites of landscape species were selected for three sites: Madidi National Park in northwestern Bolivia, Yasuní National Park in eastern Ecuador, and Nouabalé-Ndoki National Park in northern Congo Republic. At each site, the spatial distributions of requirements for each landscape species were mapped at appropriate temporal scales and overlaid on distributions of human activities for the same periods. Based on these analyses, focal landscapes' were identified for conservation action at each site. In all three cases the landscapes identified for conservation action differed from those predicted by a traditional core-buffer-corridor' model. The subset of threats identified as critical' was also substantially smaller than the suite of all threats identified for each of the three landscapes. While the ultimate tests of the long-term utility of the landscape species approach will be ecological and performance monitoring data, the approach appears to be an effective way to focus site-based conservation efforts in large and heterogeneous landscapes.

CORDELL, SUSAN, Robert J. Cabin, and Lisa J. Hadway. USDA Forest Service, Institute of Pacific Islands Forestry, 23 E. Kawili St., Hilo, HI 96720, USA (SC, RJC), Hawaii State Department of Land and Natural Resources, P.O. Box 4849, Hilo, HI 96720, USA (LJH).
RESOURCE PARTITIONING AMONG NATIVE HAWAIIAN DRY FOREST TREES.
Hawaii's dry forests are among the most endangered of all ecosystems in the archipelago. Unfortunately, these once extensive and diverse communities have been severely fragmented and degraded by deforestation, development, fire, nonnative ungulate grazing, and invasions by alien plant species such as fountain grass (Pennisetum setaceum). Understanding factors such as determinants of growth, water relations, and seasonal patterns of phenology can help us interpret current forest composition and evaluate the threat of invasive species. We studied the mechanisms for partitioning limited water sources of seven co-existing dominant trees from the Kaupulehu dry forest preserve on the island of Hawaii. Species such as Diospyros sandwichensis and Santalum paniculatum exhibited more negative stem water potential and lower rates of net photosynthesis during periods of low soil moisture availability while making opportunistic uses of water when available. In contrast, other species such as Colubrina oppositifolia maintained more positive values of stem water potential with relatively high rates of photosynthesis over a large range in soil moisture availability. This may be attributed to the ability to tap deep sources of soil water during periods of drought. These results point to spatial and temporal partitioning of resources among Hawaiian dry forest canopy tree species.

CORY, COLEEN, and Samuel Gon III. The Nature Conservancy of Hawaii, 923 Nuuanu Avenue, Honolulu, HI 96817, USA.
RECRUITING SCIENTISTS FOR PRIORITY CONSERVATION RESEARCH.
The Nature Conservancy strives to obtain and utilize the best available scientific information to influence conservation planning and land management. To that end, the Hawaii office (TNCH) works cooperatively with scientists in government, academia, and the private sector. However, land managers have expressed concern that some research is inadequately targeted to address immediate conservation management needs. Therefore, starting in 1997, TNCH managers and biologists compiled a list of priority research needs that were designed to better link researchers and managers. The Secretariat for Conservation Biology promulgated this list, which was made available to the graduate students, in the Ecology, Evolution, and Conservation Biology (EECB) program at the University of Hawaii, many of whom intend to pursue careers in conservation. This list has been updated annually to reflect changing management concerns, and the format has evolved to describe the expected management application of each potential study. In Fall 2000 the list was posted to the EECB web site for easy access by researchers outside Hawaii. A dozen research projects have been undertaken that reflect topics on the research needs list. This research has provided valuable data and guidance to improve management of Hawaiian ecosystems. Some instructive examples are briefly reviewed.

COURTNEY, MARK W., and Elizabeth E. Lyons. National Science Foundation, 4201 Wilson Blvd, Arlington, VA 22230, USA (MWC, EEL).
FUNDING OPPORTUNITIES AT THE NATIONAL SCIENCE FOUNDATION.
The National Science Foundation is charged with facilitating the health of the scientific enterprise in the US. It does this by providing funding for research and education in science and engineering. Clear understanding of this role and the opportunities provided by the NSF serves both the scientific and public communities. Information will be provided on the nature of NSF support, with emphasis on funding opportunities in the biological sciences. A handout with descriptions and URLs of relevant NSF competitions will be provided.

COVICH, ALAN P. Fishery and Wildlife Biology, Colorado State University, Ft. Collins, CO 80523, USA, and National Center for Ecological Analysis and Synthesis, Santa Barbara, CA 93101, USA.
HOW DO DIFFERENT INSULAR TROPICAL STREAM ECOSYSTEMS FUNCTION?
Tropical insular rivers are relatively short in length and have small, linear drainage patterns rather than large reticulate networks found on continents. Isolated insular freshwaters generally have fewer native species than mainland rivers. For example, the diversity of aquatic insects is often low because of their limited dispersal ability. Insular headwaters are frequently dominated by native species of gobies, neritid snails, as well as atyid and palaemonid shrimp. These species are widespread because their larvae drift in oceanic currents to many different estuaries and migrate up rivers. Low-diversity insular streams are also highly vulnerable to human-introduced species. Consequently, food webs are often complex assemblages of relatively few native species mixed with variable combinations of non-native species. These ecosystems provide opportunities for studies of major ecological questions. How do food webs function with relatively reduced redundancy? How resilient are these communities to highly variable flow conditions? Studies of tropical streams on Caribbean and Pacific islands are providing new insights regarding energy flow through complex species assemblages. Geographic and temporal comparisons illustrate how food webs reorganize and respond to extreme hydrologic events such as floods and droughts. Understanding these dynamic responses is essential for managing native species and their highly variable habitats.

COWIE, ROBERT H. Bishop Museum, 1525 Bernice Street, Honolulu, HI 96817, USA.
DECLINE AND HOMOGENIZATION OF PACIFIC FAUNAS: THE LAND SNAILS OF AMERICAN SAMOA.
Native Pacific island land snail faunas are highly diverse and exhibit high levels of endemism but are also severely threatened. Many alien snails are being introduced, leading to faunal homogenization across the Pacific. Field work in American Samoa aimed to evaluate this homogenization. We detected 19 of the 42 previously recorded native species, 11 of 12 aliens, and 3 of 6 cryptogenic (unknown origin) species. We recorded 8 species from American Samoa for the first time (3 described and presumed native, 4 alien, 1 cryptogenic); 2 undescribed endemics; and one unidentified species. We discovered 1 species previously considered extinct. Comparisons with surveys in the 1920s and 1930s, 1975, and 1992 showed that: most native species are declining; a few native species may be stable or increasing, with 1 species abundant; only 7 of the alien/cryptogenic species are declining, while 6 are increasing; for others no trend was assessed; some aliens are extremely abundant. Threats include: alien predators, predominantly rats and predatory snails; possible competitors, including alien snails; invasive plants and habitat modification; development, including recreational, agricultural, and domestic/urban development. The diverse and highly endemic native fauna is being replaced by a small number of pan-tropical aliens.

COX, JOHN J., Jeffrey L. Larkin, Michael W. Wichrowski, and David S. Maehr. Departments of Forestry and Animal Sciences, University of Kentucky, Lexington, KY 40546-0073, USA.
COLONIZING PATTERNS OF A RESTORED POPULATION IN EASTERN KENTUCKY.
Restoration of large mammals is often controversial and policy may require animals to be confined to "restoration zones." Predicting post-release movements of translocated animals may enhance restoration success. Restoration of elk (Cervus elaphus nesonii) to the Cumberland Plateau region of Kentucky began in the winter of 1997, and population augmentation has continued for the past 4 years. We used multiple linear regression to determine what landscape and demographic variables influence post-release movements of translocated elk. Most animals moved between 2-20 km from their release sites. Movements were limited by roads and natural barriers. Adults moved significantly farther than yearlings. Eighty-one percent of the animals that traveled >30 km were adult females. Early population growth is enhanced by translocating pregnant females, however, 5% of the population leave the release area during colonization behavior associated with parturition. Such losses from the population and colonization-related Allee effects may be ameliorated if the founder population is large or multiple release sites are used.

CRAIG, DOUGLAS A. Department of Biological Sciences, University of Alberta, Edmonton, T6G 2E3 Canada
Evolution of running water habitats on hotspot island archipelagos.
As a hotspot island forms and proceeds through the shield-building phase and senesces into an atoll (4-6 Mya), running water habitats change. Observations from Society, Marquesas and Austral Islands, and Rarotonga, Cook Islands, indicate that original lotic habitats were small heavily shaded streams. More specialized flows developed at early mid age (1.0 Mya) then, in old age, again to smaller shaded streams but with smooth rock chutes. Importantly after the shield building phase were regular eruptions associated with erosion of the magma chamber. Nonconformities between lava layers resulted in springs on cascade faces and elsewhere. Further, eventual collapse of the caldera provided sufficient catchment area for formation of large rivers. Rainfall is fundamental to lotic habitats and while related to island altitude and area, is influenced by climatic region. Lotic habitat evolution will be overlaid on a reconstructed phylogeny of a subgenus of central-Pacific black flies (Simuliidae). Larvae of these flies require permanent running water and some species need highly specialized habitats. The specialized habitats are normally found on younger islands, hence age of aquatic habitats influences biogeography of these simuliids - younger highly derived species cannot colonize older islands because the required habitat is not available.

CRAIG, JOHN L., and Anne M Stewart. School of Environmental & Marine Sciences & Department of Marketing, University of Auckland, Private Bag 92019, Auckland, New Zealand.
FUNDING BIODIVERSITY CONSERVATION: MOVING BEYOND SUPPLY TO ATTRACTING DEMAND.
Achieving adequate funding for biodiversity conservation is important for success. Biodiversity loss is a worldwide problem that is exacerbated by a lack of community understanding and finances. Often run primarily as a supply-driven government monopoly, conservation rarely ranks as a high government economic priority or rarely attracts private investment. Because conservation can deliver both public and private benefits, the future of biodiversity conservation must include determination of attitudes and willingness to invest in conservation. This paper reports on such a study in New Zealand. This survey of visitors to three different conservation sites showed that the public desire more conservation while expecting taxes to achieve much of this. In addition, those that enjoy conservation experiences favour conservation the most and are prepared to pay directly. Currently most tourists are forced to behave as takers of biodiversity, yet it is likely that these people would change to investors if mechanisms were available and the income was used to sustain biodiversity as well as enhance visitor experience. A lack of informed debate and supply oriented organisational structures appear to deny sustainable biodiversity gains and public satisfaction.

CRAINE, IAN T., and Mart R. Gross. Department of Zoology, University of Toronto, 25 Harbord St., Toronto, Ontario M5S 3G5, Canada.
COLONIZATION OF LAKE ONTARIO BY INTRODUCED PACIFIC SALMON.
Non-native Pacific salmon released from hatcheries have been a major component of the Lake Ontario recreational fishery for 30 years. The hatchery-produced Chinook and Coho were not anticipated to colonize because they lacked the stream-specific environmental cues that would drive them to return to their natal streams to reproduce. Recently, however, thousands of mature adults have begun entering the stream watershed to spawn. In addition, juveniles have been found in streams. Thus, it is possible that alien Pacific salmon are now colonizing Lake Ontario. The history of stocking alien species into Lake Ontario is well documented and we have good data on the environmental changes occurring in the watershed over that period including changes in the biotic community induced by exploitation and invasions by other species. Using GIS and other statistical methods, we relate the recent pattern of colonization by Pacific salmon to a variety of environmental factors, including changes in the levels of disturbance, pollution, community dynamics and watershed rehabilitation. We predict that natural reproduction by Pacific salmon will increase to the point where it will have a significant impact on fisheries management and affect restoration efforts to restore native biodiversity in the Lake Ontario ecosystem.

CRANDALL, KEITH A. Department of Zoology, Brigham Young University, Provo, UT 84602-5255, USA.
EVOLUTIONARILY SIGNIFICANT UNITS: GENETIC AND ECOLOGICAL COMPONENTS FOR CONSIDERATION.
The ESU concept has evolved significantly since its first inception to recent definitions by Moritz. These changes are associated with which biological properties to emphasize and/or de-emphasize in the designation of distinct populations for conservation status. Here I explore some of the similarities and differences among various concepts, the differential use of the ESU concept in practice, and make recommendations for the incorporation of ecological and genetic data in a multifaceted framework for describing various levels of population differentiation.

CREED, JOHN, Jiquan Chen, Graeme Cumming, Stevens Heckscher, Paul Massicot, Loyal Mehrhoff, Kathy Ramsey, and Georgia Valaoras. Michigan Tech University, Houghton, MI 4993, College of Charleston, Charleston, SC 29424, USA (JC), University of Wisconsin, Madison, WI 53706, USA (GC), Natural Lands Trust, Inc., Media, PA 19063, USA (SH), Maryland Department of Natural Resources, Annapolis, MD 21401, USA (PM), National Park Service, Ft. Collins, CO 80525, USA (LM), Umatilla National Forest, Pendleton, OR 97801 (KR), and University of La Verne, Athens 14510, Greece (GV).
THE THREE GORGES DAM IN CHINA: CONSERVATION CHALLENGES AND REFLECTIONS.
A delegation of 12 spent two weeks in China to explore and learn about the history, progress, and future of the Three Gorges Dam (TGD) that is currently under construction across the Yangtze River. The 185m-high TGD is the largest hydroelectric project in the world and will create a 660 km long reservoir. Sedimentation and flow alterations above and below the dam will be a major problem for engineers and local flora and fauna alike. The dam will also have less tangible consequences on the scenic and ecological value of the TGD. Major human activities such as coal mining, deforestation, agriculture, and factory production along the riverside have already produced much damage to the river ecosystem. Yet, ecological baseline data (e.g., species lists) of both the terrestrial and aquatic systems are limited. To promote conservation of the giant river ecosystem and many associated rare and endangered species (e.g., river dolphin, Chinese sturgeon, etc.), we call for increasing communications among all parties involved in the project, development of an adaptive management framework for the river and its watershed, and rapid development of a database and monitoring program before pre-construction information about the Gorges and the River vanish.

CRNOKRAK, PETER, and Derek A. Roff. Department of Botany, University of Toronto, 25 Willcocks Street, Toronto, Ontario M5S 3B2, Canada (PC), Department of Biology, McGill University, 1205 Docteur Penfield Ave., Montreal, Quebec H3A 1B1, Canada (DAR).
INBREEDING DEPRESSION IN THE WILD.
Despite its practical application in conservation biology, the cost of inbreeding in natural populations remains to a large degree unknown. In this review we have gathered estimates of inbreeding depression from the literature for wild species monitored in the field. Our data set includes seven bird species, nine mammal species, four species of poikilotherms and 15 plant species. Mean inbreeding depression ±SE for homeotherms was 0.509±0.081; for poikilotherms, 0.201±0.039; and for plants, 0.331±0.038. Levels of inbreeding depression this high will be biologically important under natural conditions. We compared our mammalian data (traits related to juvenile mortality) to estimates for zoo species published by Ralls et al. (1988) to determine if, as hypothesized, natural estimates of inbreeding depression are higher than captive estimates. The cost of inbreeding in mortality traits for wild mammals corrected for F=0.25 (as is the case for Ralls et al.), was substantially and significantly higher than that for captive species (2.155 compared to 0.314). Therefore, the cost of inbreeding is 7 times larger in the wild compared to captive conditions. 53% of the estimates gathered were significantly different from zero, indicating that inbred wild species frequently exhibit moderate to high levels of inbreeding depression in fitness traits.

CROOKS, KEVIN R., Andrew V. Suarez, Douglas T. Bolger, and Michael Soulé. Department of Wildlife Ecology, University of Wisconsin, Madison, WI 53706-1598, USA (KRC), Department of Entomology and Center for Population Biology, University of California, Davis, CA 95616, USA (AVS), Environmental Studies Program, Dartmouth College, Hanover, NH 03755, USA (DTB), The Wildlands Project, P. O. Box 2010, Hotchkiss, CO 81419, USA (MS).
EXTINCTION AND COLONIZATION OF BIRDS ON HABITAT ISLANDS IN URBAN SOUTHERN CALIFORNIA.
We used point count and transect surveys to estimate the distribution and abundance of eight scrub-breeding bird species in 34 habitat fragments and the urban matrix in southern California. We then calculated local extinction and colonization rates by comparing our data with surveys conducted in 1987. We classified factors that influence extinction and colonization rates into two types: (1) extrinsic factors, characteristics of the habitat fragments (area, age, and isolation), and (2) intrinsic factors, characteristics of the species that inhabit fragments (body size and population density). Over the past decade, at least one species went locally extinct in over 50% of the fragments, and local extinctions were almost twice as common as colonizations. Fragment size and to a lesser extent fragment age were the most important extrinsic factors determining extinction and colonizations. Density indices of scrub birds were the most important intrinsic factors determining extinction rates, predicting the number of sites occupied, probability of local extinction, relative area requirements, and time to local extinction.

CROUSE, DEBORAH, Loyal Mehrhoff, Mary Parkin, Diane Elam, and Linus Chen. US Fish & Wildlife Service, 4401 Fairfax Dr., Room 420, Arlington, VA 22203, USA (DC, MP, DE, LC), National Park Service, 1201 Oak Ridge Dr., Suite 200, Fort Collins, CO 80525, USA (LM).
ENDANGERED SPECIES RECOVERY AND THE SCB STUDY: THE U.S. FISH AND WILDLIFE SERVICE PERSPECTIVE.
Several current and former U.S. Fish and Wildlife Service (USFWS) biologists provide their perspective on endangered species recovery planning and the comprehensive recovery planning study recently conducted by the Society for Conservation Biology (SCB). After considering the review findings and recommendations, and real-life constraints and limitations the USFWS regularly faces, the authors identified six key areas that they believe have the greatest opportunity to improve the quality and effectiveness of recovery plans: (1) Continue to expand ties to academic and professional communities; (2) Increase efforts to expand the diversity of recovery plan contributors; (3) Expand training opportunities for USFWS employees; (4) Revise guidance on developing recovery plans; (5) Improve access to information on emerging theories, techniques, and technologies; and (6) Address key policy questions raised by the SCB recovery plan review papers. Furthermore, the authors identified a need to increase the internal consistency of recovery plans; improve species-specific details in multi-species recovery plans; standardize criteria for determining when recovery plans need revision; and improve the treatment of monitoring in recovery tasks. The authors also identify several policy issues the USFWS may want to address.

CROW, GERALD L., and Mark B. Heckman. Waikiki Aquarium, University of Hawaii, Honolulu, HI 96815, USA.
OCEAN MANAGEMENT AT MIDWAY ATOLL: RELATIONSHIPS BETWEEN DIVING, FISHING, AND GALAPAGOS SHARKS.
As part of its management plan for Midway Atoll National Wildlife Refuge, the U.S. Fish and Wildlife Service combines the protection of endangered avian and marine species with ecotourism ocean recreation. At Midway snorkeling, diving and tag-and-release fishing are all in close proximity to each other, and shark-human interaction occurs during all of these activities. Midway is a remote location--2,100 km from Hawaii. Its protected coastal environment has a large population of juvenile Galapagos (Carcharhinus galapagensis) and gray reef sharks (C. amblyrhynchos). In order to understand potential interactions we determined the sex, size and movement patterns of Midway's shark species, which are regularly attracted to diving and fishing locations. Both male and female sharks in a 1:1 sex ratio were captured, tagged, and measured before being released. They ranged in size from 87 to 118 cm total length. Sharks tagged at fishing sites were later observed at dive sites, and, based on tagged shark resightings, sharks moved freely between existing diving and fishing sites. A thorough understanding of shark movement patterns, growth rates, and reaction to dive boats and divers is needed for effective refuge management and ocean-recreation safety at Midway.

Cruz, Alexander, HEATHER SWANSON, Jameson Chace, and John Prather. Department of EPO Biology, Campus Box 334, University of Colorado, Boulder, CO 80309, USA.
EFFECTS OF THINNING AND PRESCRIBED BURNING ON PONDEROSA PINE FOREST BIRDS IN THE COLORADO FRONT RANGE.
The foothill ponderosa pine forests along the Colorado Front Range are an important, unique and understudied habitat. Ponderosa pine forests are important to a variety of wildlife species, including Neotropical migrant and resident songbirds. Specific stands of ponderosa pine on City of Boulder Open Space are slated for thinning and/or burning in accordance with the City of Boulder's management objectives. As part of this goal, in the spring and summer of 2000, we set up 6 study sites and gathered data on abundance, diversity and nesting success of Neotropical migrant birds. Our data describe the pre-treatment avian community present in each plot and the appropriate management for enhancing this community. Our data show that two of the plots support a depauperate avian community and a low rate of nesting. As a result, they offer a great opportunity for use of more extreme management measures. Two additional plots show a diverse and abundant breeding bird community so lower intensity management will be appropriate for these stands. Finally, the two control sites adequately reflect a healthy, diverse bird community and therefore will serve as appropriate controls for the four treatment stands in future comparisons of bird community responses to management.

CRUZ, JUSTINE B., and Grant M. Beauprez. Division of Fish and Wildlife, Commonwealth of the Northern Mariana Islands, P.O. Box 10007, Lower Base, Saipan, MP 96950.
FEATHERS AND SCALES: FOREST BIRDS AND LIZARDS ON THE NORTHERN ISLANDS OF THE MARIANAS ARCHIPELAGO.
The herpetofauna and forest avifauna of the Northern Islands of the Marianas Archipelago have not previously been systematically documented. Only 8 forest bird species are known from the Northern Islands, compared with 16 from the southern islands of the chain. In June-August 2000 we surveyed six islands to the north of Saipan. We estimated forest bird populations using Variable Circular Plot methodology and documented the presence of reptiles using adhesive rodent traps, loop-snare traps, and visual surveys. We recorded two island firsts for lizard species: the rock gecko, (Nactus pelagicus) on Alamagan, and the tide-pool skink (Emoia atrocostata) on Alamagan and Guguan. The resulting bird population estimates are the first for forest birds on these islands. We found that populations of Micronesian Honeyeaters (Myzomela rubratra) and Micronesian Starlings (Aplonis opaca) were wide-spread and abundant. Micronesian Megapodes (Megapodius laperouse) were also present on all islands, but abundance was related to presence or absence of feral animals. Other birds exhibited disjunct distributions, including the endangered Nightingale Reed-warbler (Acrocephalus luscinia). The Northern Islands have the potential, through translocation and establishment of new populations, to guard the archipelago's biodiversity should an infestation of Brown Tree Snake (Boiga irregularis) occur.

CRUZ-TORRES, MARILUZ. Department of Anthropology, University of California, Riverside, CA 92521, USA.
ENVIRONMENTAL DEGRADATION, COASTAL ECOSYSTEMS AND HUMAN COMMUNITIES IN NORTHWESTERN MEXICO.
The growth of the commercial, export-oriented agriculture, fishing and aquaculture industries are currently contributing to the degradation of coastal ecosystems in northwestern Mexico. Agricultural runoff and shrimp pond wastes usually end up in the many lagoons and estuaries found in the region, affecting water quality, mangrove forests and aquatic organisms. Habitat fragmentation is now a common process found in these coastal ecosystems. This research also examines the implications of natural resource degradation for the human population who inhabits the region. The results of the study show that local households have developed a set of responses to cope with the various problems posed by environmental degradation. The study concludes that management policies integrating the knowledge and socio-cultural characteristics of the local population are indispensable for the sustainable development of agriculture, aquaculture and fishing industries, and for the conservation of coastal ecosystems in northwestern Mexico.

CUNNINGHAM, ANDREW A., David Espinosa, and Peter Daszak. Institute of Zoology, Regent's Park, London NW1 4RY, UK (AAC & DE), Consortium for Conservation Medicine, Lamont Doherty Earth Observatory, PO Box 1000, 61 Route 9W, New York, USA (PD).
PARASITES, DISEASE AND CONSERVATION OF ISLAND ENDEMICS: LESSONS FROM THE PARTULA SNAILS OF POLYNESIA.
Tree snails of the genus Partula were extirpated from much of their South Pacific island range following the introduction of a predatory snail, Euglandina rosea, as a biocontrol for feral alien Achatina snails. Many Partula spp. are critically endangered and over 10 occur solely in captivity, where populations undergo periodic crashes suggestive of infectious disease. Here we report an emerging disease of partulid snails caused by a new species of microsporidian parasite. This disease caused a population crash to extinction of Partula turgida - the first proven case of extinction by infection. The origins of the parasite are currently unknown, but possibilities include allopatric hosts encountered in captivity, co-introduction to Polynesia with alien hosts, or endemicity in one or more species of partulid snail. We demonstrate an association between microsporidian infection and increased mortality in P. turgida, but not in three other infected Partula spp., although parasite-induced disease was present in individual snails. The reasons that infected populations of these species, but not of P. turgida, appear unaffected remain unknown, but one possibility is that they co-evolved with the parasite. Further work is urgently required to determine the prevalence and impact of microsporidian infection in other Partula spp.

CURRY, R. L. Dept. of Biology, Villanova University, Villanova, PA 19085, USA.
CONSERVATION STATUS AND THREATS TO ISLAND-ENDEMIC MOCKINGBIRDS AND ALLIES.
The family Mimidae includes several island-endemic forms that are among the world's most endangered avian taxa. I reviewed available information to assess whether these species face common threats. The Floreana Mockingbird (Nesomimus trifasciatus) disappeared from Isla Floreana (Galápagos) in the late 1800s due to predation and habitat degradation caused by introduced mammals; it persists on two mammal-free islets with a total population of < 200 individuals. Barring mammalian introductions, continuing threats include El Niño-associated climatic variability, and demographic and genetic stochasticity. Introduced mammalian predators and grazers likewise account for the decline of the Socorro Mockingbird (Mimodes graysoni) to < 350 individuals. Cyclones, potential for hybridization, and lack of formal habitat protection represent additional threats. White-breasted Thrashers (Ramphocinclus brachyurus) total < 200 birds on St. Lucia and Martinique. On both islands, anthropogenic habitat loss may compound effects of introduced predators; hurricanes are an additional threat given severe range contraction. Concern exists regarding three additional endemics (Cozumel Thrasher, Toxostoma guttatum; San Andres Mockingbird, Mimus magnirostris; and Tres Marias Blue Mockingbird, Melanoptila caerulescens longirostris) because of hurricanes, introduced mammals, and habitat loss; however, basic ecological data for these taxa are lacking. Other island mimids are comparatively safe where common threats are absent.

CZECH, BRIAN. U.S. Fish and Wildlife Service, National Wildlife Refuge System, 4401 N. Fairfax Drive - MS670, Arlington, VA 22204, USA.
IMPLICATIONS OF THE NATIONAL WILDLIFE REFUGE SYSTEM IMPROVEMENT ACT FOR CONSERVATION LAND ACQUISITION.
The National Wildlife Refuge System is the only federal system of lands devoted to wildlife conservation in the United States. The importance of the Refuge System to biodiversity conservation rose immensely as a result of the National Wildlife Refuge System Improvement Act of 1997. Among other things, the Refuge Improvement Act: 1) provided a unifying mission for the Refuge System; 2) mandated the maintenance of biological integrity, diversity and environmental health on the Refuge System; 3) calls for the Refuge System to contribute to the conservation of the ecosystems of the United States. Policies have been developed that elaborate the mission of the Refuge System and the maintenance of biological integrity, diversity, and environmental health. For purposes of contributing to the conservation of the ecosystems of the United States, an ecosystem conservation component was added to the U.S. Fish and Wildlife Service's Land Acquisition Prioritization System (LAPS), which is used to score potential acquisitions for developing budget proposals pursuant to the Service's trust responsibilities. For the conservation estate at large, a transdisciplinary approach that integrates natural and social sciences may help to assess in holistic terms the relative conservation merits of lands and, therefore, to optimize conservation land acquisition strategies.

DAEHLER, CURTIS C. Department of Botany, University of Hawaii at Mänoa, Honolulu, HI 96822, USA.
POPULATION ECOLOGY OF AN INVADING GRASS VERSUS A NATIVE COMPETITOR: IMPLICATIONS FOR CONSERVATION AND MANAGEMENT.
In Hawaii, African fountain grass (Pennisetum setaceum) is a highly successful invader that has displaced native pili grass (Heteropogon contortus) from many arid lowlands. Both species are fire adapted, perennial C4 bunchgrasses that rely on apomictically produced seeds for population growth. To gain insights into why the invader has been more successful than the native, we compared the reproductive ecology and growth responses of theses species under various conditions. We found that seeds of the invader germinated faster than those of the native. Likewise, seedlings of the invader grew faster, particularly in nitrogen-rich soils. Seed production by the invader averaged 6-fold greater than the native under high-to-moderate water and nutrient conditions. Nevertheless, some conditions favored the native: under drought and low nutrient conditions, the invader failed to produce seeds. The native's seedlings also had higher drought tolerance than those of the invader. In composite, these findings help explain the native's current restriction to some of the driest, nutrient-poor habitats in Hawaii while also suggesting that these habitats are suitable for pili grassland restoration because of natural resistance to fountain grass. The spread of alien nitrogen-fixing legumes into these habitats may exacerbate attempts to maintain or restore native pili grasslands.

DASZAK, PETER, and Andrew A. Cunningham. Consortium for Conservation Medicine, Lamont Doherty Earth Observatory, PO Box 1000, 61 Route 9W, New York, USA (PD), Institute of Zoology, Regent's Park, London NW1 4RY, UK (AAC).
"CO-EXTINCTION" IN THE PARTULA SNAIL CONSERVATION PROGRAM?
Tree snails of the genus Partula have undergone a series of dramatic extinctions over the last two decades, following the introduction of predatory snails throughout their Polynesian range. Captive breeding is critical for the survival of this group of island endemics, which contains over 10 species that are extinct in the wild but still exist as small captive populations. We repeatedly found high numbers of a diverse assemblage of flagellated protozoa in the gut of captive and wildcaught snails. Fluorescent microscopy demonstrates that these protozoa ingest particulate matter and are likely to be non-parasitic commensals. Newborn mortality is higher in generations with low flagellate prevalence. We conclude that these protozoa may assist digestion of the snails cellulose-rich diet. Prevalence and diversity of flagellates is significantly higher in wild caught snails and decreases significantly with each generation in captivity. These data suggest that captive management should be modified counter-intuitively, so as to increase flagellate transmission rates and prevent their permanent loss from captive stocks. Furthermore, they highlight the need for consideration of parasite conservation and "co-extinction" rates in ex situ conservation programs.

DAUGHERTY, CHARLES H., Susan N. Keall, and Nicola J. Nelson. School of Biological Sciences, Victoria University of Wellington, New Zealand.
RECOVERY OF TUATARA ON LITTLE BARRIER ISLAND, NEW ZEALAND: SCIENCE MEETS POLITICS.
Little Barrier Island is a 3083 ha volcanic island in the Hauraki Gulf, New Zealand. Its biological richness - rare birds, bats, mega-invertebrates, and kauri forests - makes it one of the most significant wildlife reserves in New Zealand, but the island is seriously degraded by the presence of introduced Pacific rats (Rattus exulans), which threaten many of the rarest species. A remnant population of tuatara (Sphenodon), an ancient reptilian lineage endemic to New Zealand, was re-discovered on Little Barrier Island in 1991. We report on the progress of recovery programs for this population, which have successfully met the following scientific challenges: (1) protection of the 8 surviving tuatara from rat predation; (2) assisted reproduction of these survivors, resulting in approximately 60 hatchlings now in a head-start facility; (3) confirmation of the pattern of temperature-dependent sex determination and its consequences for population recovery; and (4) determination of the taxonomic status of the population. The primary remaining obstacle to long-term recovery of tuatara and other rare species on Little Barrier Island is the persistent failure of management authorities and indigenous M ori people affiliated with the island to establish a working relationship that will allow eradication of Pacific rats.

Day, Jenny, A., and MARCUS J. WISHART. Freshwater Research Unit, Department of Zoology, University of Cape Town, Rhodes Gift 7701, South Africa (JAD), CRC Freshwater Ecology, Australian School of Environmental Studies, Griffith University, Nathan - Brisbane, Queensland 4111, Australia (MJW).
AN ENDEMIC ISLE ON AFRICA'S SOUTHERN TIP: THE AQUATIC FAUNA OF THE CAPE FLORISTIC REGION.
An area of outstanding conservation importance the Cape Floristic Region is recognized as a region of extreme floristic species richness and endemism. Considering the CFR as an island of varying size, many of its aquatic ecosystems act as islands, isolated from each other for millions of years. We hypothesize that these effects contribute substantially to the narrow distribution ranges and extreme endemism within the aquatic fauna. Analyses show that while species richness of aquatic invertebrates may not be comparable to that of the vegetation, levels of endemism are greater. Morphological revisions, increasing data synthesis and molecular techniques have lead to greater recognition of endemism within this important biological repository. MtDNA data for some lotic species reveal some of the highest levels of intra-specific divergence recorded, pointing to possible species flocks. Spared climatic oscillations experienced across the rest of the continent and the effects of glaciation, this region has remained relatively stable over the past 200 MY allowing the formation of myriad water bodies, variable in physical and chemical characteristics. Strong orographic gradients and edaphic diversity, oligotrophic and often seasonal or ephemeral water bodies all impose strong selective pressures. These factors have provided environments enabling an array of endemic species.

DE ANGELO, MARIA R., and Jan A. Randall. Department of Biology, San Francisco State University, 1600 Holloway Ave., San Francisco, CA 94132, USA.
DIPODOMYS INGENS: KING OF THE PLAIN? PREDATION PRESSURE AND FOOTDRUMMING ON THE CARRIZO PLAIN.
The endangered kangaroo rat, Dipodomys ingens, exists on 2% of historic habitat on the Carrizo Plain Natural Area (CNPA) in California. A keystone species, D. ingens is prey to owls, snakes, weasels, coyotes and the endangered kit-fox. In predator encounters, D. ingens footdrums, a well known and highly efficient way to combat snake predation. Footdrumming is seen anecdotally in encounters with kit-fox (Vulpes macrotis), its primary mammalian predator, but has not been reported for other mammalian predators. Coyotes (Canis latrans) and kit-foxes compete for food items, including D. ingens, on the Carrizo Plain. As human encroachment increases, driving more predators onto CNPA and increasing competition, how may D. ingens respond to burgeoning numbers and types of predators? Our study tests the anti-predator response of D. ingens to mammalian predation. We introduced a taxidermy kit-fox and a stuffed tiger, representative of non-specific four-legged predator, to sixteen individuals on the Elkhorn Plain within CNPA. In predator avoidance tests, D. ingens did not avoid either predator. Predator recognition tests indicate D. ingens may footdrum in response to four-legged predator and not specifically to kit-foxes. This ability to adapt anti-predator behavior may maintain D. ingens' ecological viability under continually changing predation pressures.

Dever, Jennifer A., LUCY A. DUECK, Nancy (Mandy) Schable, and Travis C. Glenn. Lander University, Greenwood, SC, USA (JD), Savannah River Ecology Laboratory - University of Georgia, Aiken, SC, USA (LD, NS, TG).
GENE FLOW BETWEEN SUBSPECIES' OF CAPTIVE RED PANDA -- BRIDGING THE GAP, OR BURNING THE BRIDGE?
The red panda, Ailurus fulgens, is endangered in the wild because its declining population is severely fragmented (IUCN Red List). Zoos have developed a long-range captive management plan (SSP) as insurance against species loss, but require genetic information to prevent inbreeding or outbreeding depression. Animals collected from both extremes of red panda distribution across the Himalayas (19-20 each) have been managed separately as purported subspecies. In nature, western animals (A.f. fulgens) may be near extinction due to human encroachment. Heightened efforts to preserve captive A.f. fulgens also raise concerns about loss of genetic diversity from founder effects in long-term breeding. Managers are therefore considering crossing western with eastern animals (A.f. styani), but must understand potential consequences. We sequenced a 330-bp segment of the cytochrome b gene of red panda mitochondrial DNA from a small sample of zoo animals to determine genetic divergence between the subspecies. We observed five haplotypes - one shared by both subspecies, one exclusively found in fulgens, and the other three found only in styani. These results suggest that subspecific classification is still inconclusive, and that additional sampling of individuals and DNA segments is required before managers can be advised on interbreeding red panda subspecies.

DINERSTEIN, ERIC, Anup Joshi, Mingma Sherpa, Anil Manandar, Eric Wikramanayake, and David Smith. World Wildlife Fund - United States, 1250 24^th St. NW, Washington DC 20037, USA, and Dept. of Fisheries and Wildlife, University of Minnesota, St. Paul, MN, USA.
THE TERAI ARC: A LANDSCAPE-SCALE PLAN TO CONSERVE WILD TIGERS IN LOWLAND NEPAL AND INDIA.
The subtropical jungles to the south of the Himalayan foothills in Nepal and India, known as the Terai Zone, support perhaps the densest populations of tigers on Earth. Tiger populations are protected in a network of 11 national parks and wildlife reserves. The biology of tigers and their prey in several of these reserves have been studied in detail. Only one of the 11 reserves is larger than 1000 km² and none currently supports more than 60 breeding adults. The small size of the reserves poses a serious obstacle for long-term tiger conservation: despite their large size, ability to swim, and large home ranges, tigers are relatively poor dispersers. Empirical data from India indicates that tigers are reluctant to cross gaps greater than 5 km of human-dominated landscapes. Existing forest corridors have the potential to reconnect all 11 reserves, although there are a few major dispersal bottlenecks requiring restoration. We created a landscape-scale conservation plan, the Terai Arc, designed to manage tigers as a single metapopulation spanning a range of almost 800 km. We identify the steps now underway to make the Terai Arc the most ambitious wildlife recovery program in Asia.

DOBSON, ANDY. EEB, Eno Hall, Princeton University, NJ 08544-1003, USA.
MARINE RESERVES: ARE THERE ANY LESSONS TO BE LEARNED FROM TERRESTRIAL RESERVE SYSTEMS?
The last fifty years of the 20th century produced legislation that set a significant area of land for the preservation of terrestrial biological diversity. During the last ten years significant progress has been to conserve areas for marine biodiversity. As optimists we have to believe that we can continue to increase the amount of land set aside for as nature reserves. How much of what we have learned from studying terrestrial systems will apply to marine (and freshwater) systems? Here, I review some of the techniques for nature reserve design that have focused on specific species, communities or human uses. I will then discuss the degree to which the major innovations in terrestrial reserve design theory (e.g., SLOSS, corridors, edge effects) apply to the design of marine reserves - particularly those designed to conserve migratory marine mammals.

DOMROESE, MEG. Center for Biodiversity and Conservation, American Museum of Natural History, Central Park West at 79^th Street, New York, NY 10024, USA.
PROMOTING BIODIVERSITY CONSERVATION IN BOLIVIA THROUGH COMMUNICATION AND COLLABORATION.
To address the challenge of promoting broad participation in biodiversity conservation, the Museum's Center for Biodiversity and Conservation (CBC) has initiated nonformal education and training efforts in association with its international field projects. This poster presents a process of bringing together community and scientific expertise to enable effective local and regional planning and sustainable resource management. Begun in 1998, Conservation of Biodiversity through Integrated Management is a collaboration in Bolivia that includes the CBC, the Museo Nacional de Historia Natural, and the Museo de Historia Natural Noel Kempff Mercado. Along with interdisciplinary research and professional development, the program catalyzes interactions among stakeholders in selected protected areas. Training workshops for community representatives and park staff have been instrumental in building relationships among partners, legitimizing interpretation as a component of conservation projects, and increasing participants' capacity to effectively communicate and interpret environmental information. Local partners--museum educators in particular--are taking on a progressively greater role in facilitating training and follow-up with workshop participants. Small grants offer an opportunity for participants to apply what they have learned in community-based conservation and development projects. Project partners will continue to build on these efforts to support effective management and interpretive strategies in Bolivia's protected areas.

DONLAN, C. JOSH, Bernie R. Tershy, Don Croll, Brad Keitt, Bill Wood, and Jose Angel Sanchez . Island Conservation and Ecology Group, University California, Santa Cruz, CA 95064, USA (BRT, CJD, DC, BK, BW), Department of Biology, University of California, Santa Cruz, CA 95064, USA (DC), Institute of Marine Sciences, University of California, Santa Cruz, CA 95064, USA (BRT), Grupo de Ecología y Conservación de Islas , AP 162 ,Punta Banda, BC 22791, México (JAS).
ISLAND ENDEMISM AND EXTINCTIONS IN NORTHWESTERN MEXICO.
To guide conservation action, we developed a database of the biodiversity and threats to the islands of Northwestern Mexico. There are >250 islands in the region with a total area of ~5,277 km2. These islands are diverse in vertebrates and have high levels of endemism at the species and subspecies levels: ~141 land mammals (75% endemic), ~179 reptiles and amphibians (43% endemic), and at least 48 endemic birds. There have been 23 possible extinctions of endemic vertebrate species and subspecies. Of these, 21 were likely caused by introduced mammals (primarily feral cats), one was caused by over-hunting, and one by natural environmental stochasticity. Possible extinctions were not evenly distributed across taxa. Approximately, 12% of endemic mammals, ~20% of endemic birds, and 0% of endemic reptiles and amphibians may be extinct. Prior to 1995, introduced mammals were present on >49 islands in the region. The Island Conservation & Ecology Group, National Autonomous University of Mexico, Center for Biological Investigations, and the National Protected Areas Department have completed 28 mammal removals from 19 of these islands.

DOUGILL, STEVE J., Laurie S. Gold, and Paul C. Banko. USGS - BRD, Pacific Islands Ecosystem Research Center, Kilauea Field Station, P.O. BOX 44, HAVO, HI 96718, USA.
MONITORING THE DISTRIBUTION AND IMPACTS OF ALIEN WEEDS TO FACILITATE RECOVERY OF A DRY, SUBALPINE hAWAIIAN fOREST.
Alien plants have greatly impacted native ecosystems in Hawaii, especially in the lowlands where most plant communities are dominated by introduced species. High elevation areas (above 2000m) are believed to be impacted less. However, many aliens originating from temperate and Mediterranean regions have become widely established at high elevations, and others with very local distributions have the potential to spread more widely. On Mauna Kea alien plants have drastically changed the natural fire regime, reduced regeneration of native species, and have potentially provided additional resources for alien insect pests. In this paper, we describe the distribution of a number of alien plants that are currently a threat, or pose a threat, to the subalpine forests of Mauna Kea and discuss their implications to this unique environment and the palila (Loxioides bailleui), an endangered songbird restricted to this habitat. We recommend continuing to monitor alien plants on Mauna Kea because, 1) it is often difficult to predict what species will eventually spread and become a problem, and 2) the plant community is very dynamic as the forest recovers from > 150 years of ungulate damage making it difficult to predict the outcomes of interactions between native and alien plants.

DRAKE, JOHN M., David M. Lodge, Greg Dwyer, Kevin L.S. Drury, and Norman Yan. University of Notre Dame, IN 46556, USA (JMD, DML), University of Chicago, IL, USA (GD, KLSD), York University, Ontario M3J 1P3, Canada (NY).
PREDICTING THE SUCCESS OF INVADING SPECIES: APPLYING STOCHASTIC MODELS OF POPULATION GROWTH AND THE ROLE OF ALLEE EFFECTS.
Because they are often very small, inoculating propagules of invading species are subject to the caprice of two sources of stochasticity: demographic and environmental. These stochastic factors supply a condition under which some inoculating populations will subsist while others fail. The development of stochastic models in ecology in general, and population viability analyses (PVA) in particular, presents a formalized probabilistic approach for studying invasions of exotic plants and animals. Here we present results from a two-pronged exploratory investigation for the Great Lakes invasive zooplankter Bythotrephes longimanus. First, we use time series data to estimate population growth parameters and predict invasiveness. Second, using life history data available in the literature, we simulate possible population trajectories to estimate the probability of invasion and corroborate the analytic result. Results suggest that Allee effects and quasi-Allee effects have significant consequences for the growth of introduced populations, possibly accounting for often observed "lag-times" during the initial stages of population establishment. Where Allee effects are not present, as few as 10 introduced individuals results in a high (>.5) probability of invasion. This result is achieved independently through both simulation and analytical techniques.

DUNCAN, JEFFREY R., and Julie L. Lockwood. National Park Service/Rivers, Trails, and Conservation Assistance, 424 Georgia Ave. Ste. 2B, Chattanooga, TN 37403, USA (JRD), Department of Environmental Studies, Natural Sciences II, University of California, Santa Cruz, CA 95064, USA (JLL).
EXTINCTION IN A FIELD OF BULLETS: INTRINSIC VS. EXTRINSIC CAUSES IN THE DECLINE OF THE WORLD'S FRESHWATER FISHES.
Because human actions alter aquatic ecosystems similarly worldwide, the extinction risk among many freshwater fishes that share particular life-history traits may also be similar. Determining whether taxonomic selectivity exists among the world's freshwater fish families is then a key step in predicting future species declines and triaging conservation efforts. We use binomial statistics to look for taxonomic patterns among the world's freshwater fish families. Families are identified as being at risk of extinction if at least one species within a family is classified as either extinct or at risk of extinction by the IUCN Redlist of Threatened Animals. Eighteen freshwater families have more threatened species than expected if extinction risk was evenly distributed across all families. Next, we use a series of chi-squared analyses to determine if known family-level characteristics produce this taxonomic pattern. We find that families inhabiting well-studied regions of the world contain more threatened species. However, we find no indication of a unifying set of extinction-promoting biological or ecological traits that contribute to extinction risk among freshwater families. A possible explanation for this discrepancy is that aquatic alterations worldwide are so severe that extinction is driven by extrinsic rather than intrinsic factors.

DUNHAM, JASON B., Bruce E. Rieman, and Deborah L. Myers. U. S. Forest Service, Rocky Mountain Research Station, Boise, ID, 83702, USA.
ISLANDS OF COLD WATER IN THE DESERT: DEFINING HABITAT FRAGMENTATION, ITS EFFECTS, AND WHAT TO DO ABOUT IT.
To understand the effects of habitat fragmentation, it is essential to identify the causes of fragmentation, and to define habitat structure and its effects. In many systems, such as oceanic islands or lakes, the structure of habitat may be obvious. In others, fragmentation of habitat may be less apparent. We developed models to predict the effects of both local and geographic environmental gradients on the distribution of threatened cutthroat trout in the Great Basin, USA. These models were linked to a geographic information system to create watershed "patches" of suitable habitat. This model was used to predict occurrence of cutthroat trout at two scales: among patches, and among patch networks within a large (70,000 km² ) basin. This information forms the template for a spatially explicit landscape classification to identify specific threats (habitat loss and nonnative species) and alternative conservation management strategies (habitat restoration, control of nonnative species, reintroductions). Our approach to addressing the problem of habitat fragmentation should be generally relevant for imperiled species with widespread distributions and restricted habitat requirements.

DUNHAM, SUSIE, Thom O'Dell, and Randy Molina. Dept. of Forest Science, Oregon State University (SD), USDA Forest Service, Pacific Northwest Research Station (TO and RM), 3200 Jefferson Way, Corvallis, OR, 97331, USA.
WITHIN POPULATION GENETIC STRUCTURE IN GOLDEN CHANTERELLES ASSESSED WITH MICROSATELLITE LOCI.
The Northwest Forest Plan has drawn attention to conservation needs for many species of fungi including the commercially harvested golden chanterelle (Cantharellus formosus). Cryptic growth habits make fungal individuals difficult to observe and life history information critical for management efforts is lacking. In broadly distributed species like C. formosus, knowledge of within-population genetic structure is a prerequisite for planning larger scale studies upon which management decisions are based and contributes to our understanding of life history parameters difficult to measure without genetic data. Fruit bodies from 167 C. formosus individuals were mapped and collected from a 50 year old, 50 ha stand of naturally regenerated douglas fir (Pseudotsuga menziesii) and western hemlock (Tsuga heterophylla). Sixty-five unique genotypes resulted from scoring these collections at four microsatellite loci. The spatial distribution of the 10 most frequent alleles was studied using Moran's I statistic. Significant but weak spatial autocorrelation was detected in the four smallest distance classes estimating a 250-400 m patch size for C. formosus. This result indicates that either limited spore dispersal possibly coupled with inbreeding, long lived individuals fragmented over large areas, or some interaction between these processes works to maintain fine scale genetic structure in this species.

DUNK, JEFFREY R., and William J. Zielinski. USDA Forest Service, Redwood Sciences Laboratory, 1700 Bayview Drive, Arcata, CA 95521, USA (JRD and WJZ). Department of Natural Resources Planning and Interpretation, Humboldt State University, Arcata, CA 95521, USA (JRD).
GEOGRAPHIC RANGE AND HABITAT OF RARE MOLLUSKS IN NORTHERN CALIFORNIA.
Under the Northwest Forest Plan, national forests are mandated to survey for a variety of "Survey and Manage" species prior to on-the-ground activities. Almost no quantitative data exists for the majority of the 400+ "Survey and Manage" species. We sampled at 308 randomly-selected points within a 5.4 km grid network (n = 1055 points) over 23,000 km² in northern California. We detected 93 species or subspecies of mollusks and we estimated geographic ranges (extent of occurrence) and habitat associations of 9 terrestrial "Survey and Manage" mollusks. We also evaluated whether these species occurred disproportionately in reserves that were developed for northern spotted owls. In addition to discovering several new mollusk species, we found that the 9 selected "Survey and Manage" mollusks were as likely to be detected on reserved lands as lands where timber harvesting is allowed. We developed areas of occupancy models for each species using spatial data in conjunction with vegetation data from several scales. Geographic ranges that were produced by connecting the outermost locations for each species were smaller than those generated by including locations that were not sampled but which had high predicted probabilities of occurrence. A relatively small region of the study area accounted for most of the diversity of mollusks.

DUNLEVY, PETER A., and Earl W. Campbell. USDA - National Wildlife Research Center, Hawaii Field Station, P.O. Box 10880, Hilo, HI 96721, USA (PD, EWC).
Assessment of avian non-target hazards from the broadcast of rodenticide bait in Hawaiian forests.
The presentation will summarize the results of a field study designed to assess the relative risk that the broadcast application of rodenticide bait poses to Hawaiian forest bird communities. This study was conducted in four different forest habitats in Hawaii using two different formulations of placebo bait. Non-target risk assessment has two components: bait acceptance and toxicity (if bait is consumed). This project was designed to assess the former issue, bait acceptance. Risk was assessed by comparing placebo bait uptake to the relative abundance of specific avian species within a particular avian community. Vertebrate uptake of placebo bait was monitored in each site using forty infrared monitors and cameras. With data from three of eight replicates analyzed (approx. 7,600 slides of vertebrate bait consumption) only eighty-five slides from one site documented bait consumption by a single avian species [the introduced red-billed leiothrix (Leiothrix lutea)]. Rats were documented consuming bait 7,500 times. This data suggests that there is a relatively low direct risk to Hawaiian forest birds from the broadcast of pelletized rodenticides and will be used to support multi-agency efforts to obtain regulatory approval of the aerial broadcast application of rodenticide bait for conservation purposes in the state of Hawaii.

DUVALL II, FERN P. Department of Land and Natural Resources, Division of Forestry and Wildlife, 54 South High Street, Room 101, Wailuku, HI 96793, USA.
FERAL CAT (FELIS CATTUS) PREDATION ON LOW ELEVATION NATIVE SEABIRD COLONIES ON MAUI ISLAND.
Feral house cats, introduced to many areas of the world, have been implicated in many cases of predation on native mammalian and avian wildlife, often leading to extinctions of unique island species. Native seabirds, such as the common Wedge-tailed Shearwater (Puffinus pacificus) and more rare Bulwer's Petrel (Bulweria bulweri), which nest in coastal colonies at low elevations on Maui were negatively impacted by predation by feral cats. Cat predation on the seabirds was investigated at several locations on the main island of Maui to determine overall predation scope and predation efficiency, to characterize temporal patterns to predation, and to provide an assessment key for identifying cats as predators. Peaks in predation at colonies of sufficient size were found to be diphasic, and tied to the bird's biology. Small colonies were vulnerable to total failure and larger colonies to losses of returning adults and late-stage chicks and adults. Comparison of cat free Molokini islet illustrated cat predation has a sustained negative impact on established Maui native seabird colonies, expansion of colonies, and colonization of new areas by native seabirds.

EARNHARDT, JOANNE M., Jon D. Ballou, Steven D. Thompson, Robert C. Lacy, and Kevin Willis. Department of Conservation and Science, Lincoln Park Zoo, 2001 N. Clark St., Chicago, Il 60614, USA (JME, SDT), Smithsonian National Zoological Park, 3001 Connecticut Avenue, NW, Washington, DC, USA (JDB), Brookfield Zoo, 3300 Golf Road, Brookfield, IL 60513, USA (RCL), Minnesota Zoological Garden, 13000 Zoo Boulevard, Apple Valley, MN 55124, USA (KW).
MateR_X: A PRESCRIPTION FOR GENETIC MANAGEMENT OF CAPTIVE POPULATIONS.
Genetic management of captive populations can increase long-term population viability and fulfills goals to maintain genetic diversity of founding populations. Rigorous genetic analyses of populations guide decisions selecting specific animals to breed and pair with each other. We present a software model, MateR_X designed to standardize and simplify selection of genetically appropriate pairs. Four genetic factors are evaluated in the pairing process: change in gene diversity, inbreeding coefficient for prospective offspring, parity of parental mean kinships and degree of unknown pedigree information. Currently, specially trained advisors interpret and synthesize these parameters to make their genetic decisions. MateR_x quantifies each of the four factors based on a combination of genetic theory and current management practices and in a second step, integrates them into a single Mate Suitability Index (MSI). The MSI identifies pairs as beneficial or detrimental to the population's genetic health. Okapi and cinereous vulture, populations with different genetic structures, are compared and contrasted using MSI. MateR_x standardizes genetic management because parameter scores are applied consistently across species and repeatable over time.

EASON, CHARLES T., Eric B. Spurr, and Elaine C. Murphy. CENTOX (Centre for Environmental Toxicology), Landcare Research, P.O. Box 69, Lincoln 8152, New Zealand (ECT, EBS), and Department of Conservation, Private Bag 4715, Christchurch, New Zealand (ECM).
BIOACCUMULATION RISKS ASSOCIATED WITH BRODIFACOUM VERSUS OTHER ANTICOAGULANTS: OPTIONS FOR RODENT CONTROL.
Aerial application of cereal baits containing brodifacoum has been used successfully for eradication of rats on islands around New Zealand. Brodifacoum has also found increasing favour for rodent and possum control on the mainland of New Zealand. Concurrent with the increased field use of brodifacoum, concerns have increased with regard to primary and secondary poisoning and contamination of wildlife. Brodifacoum has been detected in significant numbers of pigs, deer, cats, stoats, and bird life from these areas. In view of the potential impact on pig hunters and dogs consuming wild pig meat and offal and, to a lesser extent, deer, the wide-scale field use of brodifacoum baits on mainland New Zealand is under review. The comparative toxicokinetics of alternative less persistent toxicants such as coumatetralyl, diphacinone and warfarin is being re-evaluated in rats and compared with brodifacoum, and their tendency to bioaccumulate is being reassessed as part of a selection process prior to field use. Use patterns of anticoagulants on small to medium sized islands (for eradication) and on mainland areas varies and the pros and cons of different anticoagulants for different uses will be presented.

ECKERT, KATHERINE A., Nina E. Hahn, Michael D. Stuart, and Hal Markowitz. Dept. of Biology, San Francisco State University, HH 520, 1600 Holloway Ave. San Francisco, CA 94132-1722, USA (KAE and HM), Office of Laboratory Animal Care, University of California at Berkeley, NW Animal Facility Rm. 203, Berkeley, CA 94720-7150, USA (NEH), Dept. of Biology, University of North Carolina, Asheville, NC 28804, USA (MDS).
THE UTILITY OF PARASITE DATA FOR THE CONSERVATION OF BLACK HOWLER MONKEYS IN BELIZE.
The implications of specific parasite infection may be important in discerning ecological, behavioral and evolutionary trends in the host-parasite relationships for a species. For non-human primate populations, there is an added impetus to understand their parasitic relationships due to the fact that they are susceptible to many human pathogens, and may be particularly sensitive to human encroachment on their habitat. Since little published information currently exists on the presence and taxa of parasites of black howler monkeys (Alouatta pigra), a coprological survey was performed of A. pigra living in the 385 ha. Lamanai Archaeological Reserve, in Belize. Fecal samples (n=269) were analyzed from monkeys living throughout the Reserve, including some areas adjoining tourist destinations and human settlements. The formalin-ethyl-acetate sedimentation procedure was used for each sample to create six wet mounts that were scanned systematically. Over forty percent (40%) of the individuals in the sample population were found to have parasitic infections, the majority of which were of a trematode believed to be in the family Dicrocoelidae. Future comparisons can be made with these baseline data to help thoroughly assess impacts of anthropogenic disturbance, natural disasters or other stresses to this population.

EGGERT, LORI S., and David S. Woodruff. Ecology, Behavior and Evolution, University of California, San Diego, La Jolla, CA 92093-0116, USA.
CENSUSING ELUSIVE ANIMALS: THE FOREST ELEPHANTS OF KAKUM NATIONAL PARK, GHANA.
One quarter to one third of the remaining African elephants belong to the forest subspecies (Loxodonta africana cyclotis). These animals are difficult to observe in the dense rainforests, and previous studies have relied on indirect methods to estimate population sizes. Using multilocus genotyping of non-invasively collected samples, we performed a genetic census of the forest elephant population at Kakum National Park, Ghana. We estimated population size, sex ratio and genetic variability from our data, then combined this information with field observations to divide the population into age groups. Our population size estimate of 225 (confidence interval 173 to 308) was very close to that obtained using dung counts, the most commonly used indirect method of censusing forest elephant populations. For forest elephant populations to persist as their habitat is fragmented by expanding human populations, management will be increasingly important. Our genetic census provides data that will allow managers to plan for the conservation of what may prove to be the largest population of forest elephants in west Africa.

ELDERD, BRET, and M. Philip Nott. Department of Environmental Studies, University of California, Santa Cruz, CA 95064, USA (BE), and Institute for Bird Populations, Point Reyes Station, CA 94956, USA (MPH).
EXPLORING THE SENSITIVITY OF AN ENDANGERED SPECIES TO CHANGES IN DEMOGRAPHY AND HABITAT USING AN INDIVIDUAL-BASED MODEL FOR THE CAPE SABLE SEASIDE SPARROW (AMMODRAMUS MARITIMUS MIRABILIS).
The Cape Sable seaside sparrow (Ammodramus maritimus mirabilis), a US endangered species, occurs in the isolated short-hydroperiod marl prairies of the Everglades of South Florida. Several sub-populations are threatened by changes in hydrology and subsequent habitat change. In order to explore the effects of a changing landscape on sparrow population dynamics, we constructed a spatially-explicit individual-based Monte Carlo model. The model (SIMSPAR) includes the topography, vegetation and hydrology of sparrow habitat as well as field-derived demographic parameters. We explored the population's sensitivity to behavioral, demographic and landscape level changes. Sparrow populations are highly sensitive to changes in maximum number of clutches, mortality rates, and female dispersal distance. Degradation of sparrow habitat resulted in a greater than proportional decrease in breeding population size and increased the population's coefficient of variation. If habitat was degraded by allowing shrub invasion of higher elevation breeding areas, the population levels declined to a greater extent than if habitat was degraded by other means. Although this population is sensitive to changes in demographic and behavioral parameters, these results highlight the sparrow's sensitivity to habitat quality and the importance of using spatially explicit models to explore a species response to landscape level as well as demographic changes.

ELPHICK, CHRIS S., J. Michael Reed, and J. Marcelo Bonta. Department of Ecology and Evolutionary Biology, University of Connecticut, 75 North Eagleville Road U-43, Storrs, CT 06269, USA (CSE), Department of Biology, Tufts University, Medford, MA 02155, USA (JMR), and 150 Oswego Summit, Lake Oswego, OR 97035, USA (JMB).
Correlates of population recovery goals in endangered birds.
Endangered species recovery plans commonly set population size goals that are used to define the success of recovery efforts. We examined variation in these goals for bird species listed under the United States Endangered Species Act to determine whether there were simple predictors of recovery population size. The median population sizes that must be met for a species to be removed from the list or down-listed to the threatened category are 4,000 and 1,500 respectively, but there was much variation. Most variation in population recovery goals (³ 75%) was explained by the population's size when the recovery plan was written. Species listed when their population's size was relatively large have higher population recovery goals, whereas those listed when populations were small have lower population goals. Population sizes set for recovery also increased over time and were higher for species listed throughout the United States. In combination, these three variables explained 86% of the variance in de-listing population goals and 94% of the variance in down-listing goals. Body mass, annual fecundity, maximum life-span, whether the population was listed as threatened or endangered, and whether a population viability analysis was conducted were not significantly associated with population recovery goals.

ERASMUS, BAREND F.N., Albert S. van Jaarsveld, Steven L. Chown, Mrigesh Kshatriya, and Konrad J. Wessels. Conservation Planning Unit, Department of Zoology & Entomology, University of Pretoria, Pretoria 0002, South Africa (EBFN), Centre for Environmental Studies, University of Pretoria, Pretoria 0002, South Africa (AsvJ), Department of Zoology & Entomology, University of Pretoria, Pretoria 0002, South Africa (AsvJ, SLC), Conservation Planning Unit, Department of Zoology & Entomology, University of Pretoria, Pretoria 0002, South Africa (MK), Institute for Soil, Climate & Water, Agricultural Research Council, Private Bag X79, Pretoria 0001, South Africa (KJW).
A BROAD SCALE ASSESSMENT OF SPECIES' DISTRIBUTION RESPONSES TO CLIMATE CHANGE.
We explore the spatial response of the distribution of selected species to climate change using a climate envelope model. The model can interpolate the distribution of poorly-sampled taxa as well as predict responses to a changing climate. Species from the arid western parts of South Africa are expected to experience range contraction and an eastward range shift to the possible detriment of resident eastern species. Shifted ranges may conflict with current land use practices that are unsuitable to sustain viable populations. Given an already fragmented habitat matrix within which these range changes are expected to happen, the ability of species to change their ranges as a successful response to climate change is called into question. Such a scenario will lead to climate sensitive species being limited to shrinking islands of suitable habitat. Given the expected extent of the spatial response, conservation planners can no longer afford to ignore the effects of climate change when engaged in long-term planning.

ERVIN, E. L., Robert N. Fisher, and K. R. Crooks. United States Geological Survey, San Diego Field Station, San Diego State University, San Diego, CA 92182-1015, USA (ELE, RNF; ELE), Department of Wildlife Ecology, University of Wisconsin, Madison, WI 53706-1598, USA (KC).
ROADS AND TOADS: AMPHIBIAN MORTALITY IN RELATION TO RAINFALL, ROADWAY SURFACE MOISTURE, AND TRAFFIC VOLUME.
Roads constitute a direct and often permanent loss of wildlife habitat, can serve as physical or psychological barriers to animal movements, and are often the source of exceedingly high levels of animal mortality. Our goal was to better understand the effects of roads on amphibian populations in a planned landscape corridor in southern California. Road cruising was employed to examine the usage of roadways and related mortality levels of amphibians. Two hundred and fifty four evening road cruising surveys were conducted between February and April, 1999 and 2000. During 93 road nights with 'wet roads', we recorded 465 dead animals and 505 live animals on roadways, yielding an overall mortality rate of 48%. In contrast, during 161 road nights that were classified as having 'dry roads', we recorded only 25 dead animals and 105 live animals, yielding an overall mortality rate of 19%. In addition to rainfall, road mortality was also influenced by the type of road surveyed; as expected, highest mortality rates occurred on highly traveled roads, with small two-lane roads exhibiting the lowest mortality rates. Our result suggest that roadways in the region are negatively impacting herpetofauna populations, particularly when roads are wet in areas of high traffic volume.

ERWIN, R. MICHAEL, David Allen, and David Jenkins. USGS Patuxent Wildlife Research Center, Dept. of Envir. Sciences, Univ. of Virginia, Charlottesville VA 22903, USA (RME), NC Wildlife Resources Comm., 355 Paul Dr., Trenton NC 28585, USA (DA), and NJ Div. of Fish and Wildlife, P.O. Box 400,Trenton NJ 08625, USA (DJ).
CREATED VS. NATURAL COASTAL ISLANDS: DEGRADING HABITATS, ATLANTIC WATERBIRDS, AND MANAGEMENT NEEDS.
In the mid Atlantic region of the U.S., breeding colonial waterbirds such as gulls, terns, Brown Pelicans (Pelecanus occidentalis) and Black Skimmers (Rynchops niger) nest on barrier islands, created dredged material islands, and to a lesser extent natural marsh/ shell pile islands. We compare breeding distribution and population changes of selected species in New Jersey, Virginia, and North Carolina from the 1970s to the 1990s In New Jersey and North Carolina, species of concern such as Common Tern (Sterna hirundo), Black Skimmers and Gull-billed Terns (S. nilotica) that depended heavily upon dredged material islands in the 1970s have experienced declines in populations in the 1990s, with a simultaneous decline in quality (and use) of dredged material islands. Rather than maintaining high quality sandy dredge sites, managers are diverting to "beach nourishment" projects instead. In Virginia, few dredge material islands are found, however, here populations of terns and skimmers on barrier islands are declining during the past 25 years and marsh and artificial sites are becoming more important; predators (fox, raccoon, coyote) have increased on barrier islands. A strong plea is made for coastal managers to better maintain or create additional small dredge, shell, or artificial islands as larger barrier islands become increasingly degraded by human activities or mammalian predators.

EVANS OGDEN, LESLEY J. Centre for Wildlife Ecology, Simon Fraser University, Vancouver, British Columbia V5A 1S6, Canada.
FARMING FOR SHOREBIRDS: USING STABLE ISOTOPES TO QUANTIFY THE VALUE OF AGRICULTURAL LAND FOR WINTERING DUNLIN.
The Fraser River Delta represents the only major over-wintering habitat in Canada for shorebirds. However, the relative dietary importance of farmland to shorebird populations was not previously established, and has been determined in this study by means of stable isotope analysis. Comparing blood and tissue samples from Dunlin (Calidris alpina) with those of terrestrial-exclusive American Robins, (Turdus migratorius) and marine-exclusive White-winged Scoters, (Melanitta perspicillata), analysis of Carbon-13 and Nitrogen-15 indicates considerable variation between individuals in the relative contribution of terrestrial prey to overall diet. Individual values range from highly marine signatures to highly terrestrial ones, but a 3-year population average (d¹³ C =-17.55 %o ± 2.6) suggests that use of fields is widespread throughout the population. High variance in isotopic values suggests that field feeding is a widely adopted strategy amongst wintering Dunlin rather than a habit adopted by a small subset of the population. Isotopic values were significantly different between years (c² = 30.19, df = 2, p<0.0001). Stable isotope values also indicate a mid-winter peak in field feeding.
Land use changes in the Fraser Delta are reducing the availability of farmland to wildlife. This research highlights previously undervalued agricultural habitat as a conservation concern.

EWEL, KATHERINE C., Kimberly Bonine, Stacy M. Rowe, Moses Palik, and Erick E. Waguk. USDA Forest Service, 1151 Punchbowl St., Rm. 323, Honolulu HI 96813, USA (KCE), Institute for International Studies, Stanford University, Stanford CA 94305-6055, USA (KB), Ngatpang State Government, Republic of Palau 96940 (SMR), and Kosrae Island Resource Management Program, PO Box DRC, Kosrae, FSM 96944 (MP, EEW).
MANGROVE CRAB (SCYLLA SPP.) HABITAT IN THE CAROLINE ISLANDS.
The mangrove crab Scylla spp., also called mud crab, is commonly harvested for consumption throughout the tropical Indo-Pacific. Decreased trapping success and increased mangrove tree harvesting rates in some places have raised questions about specific habitat requirements and the need for management. We established transects through mangrove forests in both the Republic of Palau and the State of Kosrae, Federated States of Micronesia, in order to determine basic characteristics of crab burrow location. Crab burrows were found in riverine and interior mangrove forests, with no affinity for particular species of trees. Burrow density was much higher in Palau (1.63/100m² ), where sampling was concentrated primarily in riverine mangrove forest, than in Kosrae (0.20/100m² ), where sampling was dispersed in and around fringe, interior, and riverine forests. Burrow density was correlated with presence of logs and, particularly in interior stands, with presence of forest gaps. In Kosrae, it was also correlated with the width of the stand, i.e., distance from land to shore. Mangrove crabs appear to be habitat generalists. Their abundance is not likely to be diminished by gaps resulting from natural tree mortality and firewood harvesting. Instead, burrow habitat appears to be enhanced by logs on the forest floor.

FAGAN, W.F., W.L. Minckley, Peter Unmack, and Colleen McFadden. Department of Biology, Arizona State University, Tempe, AZ 85287-1501, USA.
LINKING SPATIAL RARITY AND EXTINCTION DYNAMICS: ACROSS-SCALE STUDIES OF DESERT FISHES.
Extinction is one of the most fundamental processes in ecology. It shapes evolution, community structure, and patterns of biodiversity. Yet studies of extinction are scarce. We have taken advantage of an extraordinary database concerning occurrence records for native fishes in the southwestern US and northwestern Mexico to investigate relationships between extirpation and patterns of rarity and life-history traits. This unique database includes over 25,000 locality records spanning 160 years of field research for all 50 known taxa of fishes from the Sonoran Desert region. Key to our project is the development of new methods for quantifying patterns of rarity that are independent of the spatial scale of analyses. We discuss how the patterns of spatial rarity these fish exhibit relate to their extinction dynamics (e.g., frequency and pattern of extirpation events) and current levels of endangerment. We also suggest ways in which patterns of rarity and extinction can be predicted by life-history attributes.

FARABAUGH, SUSAN M., Noelle Tornatore, and Nancy C. Harvey. Zoological Society of San Diego, Center for the Reproduction of Endangered Species, P.O. Box 120551, San Diego, CA 92112-0551, USA.
COMPARING BEHAVIOR OF HAND-REARED OF PARENT-REARED SHRIKES, IMPLICATIONS FOR RELEASE.
A captive breeding population of the endangered San Clemente loggerhead shrike (Lanius ludovicianus mearnsi) was established in 1991. Chicks are produced both by pulling clutches for artificial incubation and hand-rearing, and by allowing the shrikes to raise their own chicks. Behavioral observations of foraging and flight skills of juvenile release candidates and parental care behavior of the captive flock have been conducted for both parent- and hand-reared birds. Timed samples of juvenile release candidates showed that parent-reared birds were more wary than hand-reared; they spent significantly less time in "dangerous" areas (i.e., on low perches or on the ground). Observations of captive breeding pairs revealed that parent-reared adults were more wary as parents. Once the chicks hatched, parent-reared adults had significantly higher rates of alarm calling than did the hand-reared adults. Once young shrikes fledge, parent shrikes actively sing in close proximity to the chicks, and young birds learn their parents' songs. Parent-reared adults sang significantly more during pair introductions. Thus parent-reared birds learn song repertoires in a social setting, they subsequently used song at higher rates than did hand-reared shrikes. Thus parent-reared shrikes may be better release candidates both because of higher anti-predator vigilance and improved song learning.

FAUST, LISA J., Steven D. Thompson, Joanne M. Earnhardt, Michelle Sherman, and Ellen Brown. Department of Conservation and Science, Lincoln Park Zoo, 2001 N. Clark St., Chicago, IL 60614, USA.
USING STAGE-BASED DEMOGRAPHIC MODELS TO GUIDE CAPTIVE MANAGEMENT: WESTERN LOWLAND GORILLAS.
Captive population management integrates genetics and demography to guide populations towards sustainability and stability. Age-based demographic models currently predict population trajectories and future space needs. However, stage-based or system models may be more appropriate because of their flexibility in representing complex dynamics, when vital rates are determined less by age than by non-demographic factors such as social or developmental state, husbandry or health issues, or the logistics of multi-institutional management. Western lowland gorillas (Gorilla gorilla gorilla) in North American zoos are currently managed via age-based matrix models. Yet stage models are appropriate because gorillas exhibit clear stages (infants, silverbacks, reproductive females, post-reproductives), are polygynous with bachelor groups of non-breeding males, and have other social, logistic, and husbandry concerns that provide complex structure to management plans. Of particular interest are the demographic and genetic effects of breeding and bachelor group structure. The gorilla population has exhibited significant bias in birth sex ratio, creating concern about population viability and management implications. We present a stage model that confirms impending management, but not demographic, problems under a variety of sex ratio biases. This model will assist decision-making about future birth rates and space needs for breeding and non-breeding groups.

FAY, KERRI, Katie Cassel, David Alexander, Ellen Coulombe, Laura Arnold, and Kay Koike. Koke'e Resource Conservation Program, Koke'e Museum, P.O. Box 100, Kekaha, HI 96752, USA.
CONSERVING KAUAI'S RESOURCES.
The Koke'e Resource Conservation Program (KRCP) is a volunteer based alien species control program sponsored by the non-profit Hui o Laka/Koke'e Museum in collaboration with the Hawaii Department of Land & Natural Resources (DLNR), State Parks Division. KRCP involves the public in protecting native ecosystem resources. Objectives are: 1) Conduct invasive plant removal in selected areas on Kauai, particularly those that still have relatively intact native forest. 2) Eliminate priority incipient weeds before they become widespread. KRCP is involved in eradication of Miconia calvescens, which was recently re-discovered in Wailua State Park. 3) KRCP has also taken action in the vicinity of endangered plants: weeding, rat baiting and fencing, which has resulted in survival of seedlings of the endangered Kokia kauaiensis in Kalalau Valley and Pteralyxia kauaiensis, and renewed growth on Solanum sandwicense that had been over-run by Rubus. Over the past three years staff and volunteers weeded 870 acres with 21,581 volunteer hours, have found new locations of several endangered species and discovered many more intact, botanically rich areas of forest that need protection now. Funded by the Hawaii Community Foundation, the U.S. Fish & Wildlife Service, the U.S. Forest Service, and private donations.

Feldheim, Kevin A., Samuel H. Gruber, and MARY V. ASHLEY. Department of Biological Sciences (M/C 066), University of Illinois at Chicago, 845 W. Taylor St., Chicago, IL 60607, USA (KAF, MVA), Division of Marine Biology and Fisheries, Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, FL 33149, USA (SHG).
STOCK STRUCTURE AND THE ROLE OF THE NURSERY IN LEMON SHARK RECRUITMENT.
Little information is available on breeding patterns, demographics or population genetic structure of most elasmobranchs, but such information is clearly needed to develop appropriate management plans. Here we characterize the local recruitment dynamics of the lemon shark (Negaprion brevirostris) and population genetic structure at greater spatial scales. DNA microsatellites developed for lemon shark were used to characterize relatedness among cohorts of juveniles at three nursery lagoons in the western Atlantic. Significant differences in reproductive dynamics were found among nurseries as indicated by levels of relatedness among juveniles, annual growth rates, and recapture rates. Lemon sharks samples from the northern and southern parts of the range of the species in the western Atlantic were used to characterize gene flow patterns. Estimates of genetic structure yielded small but significant values. No sharp discontinuities were found between Caribbean and Brazilian waters, indicating that gene flow occurs throughout the western Atlantic with no evidence for distinct stocks.

Felling, Candace, Sherri Hiraoka, Teresa Restom, Jennifer Rodwell, ELIZABETH STAMPE, Georgina Lillich, Shari Lyons, Sean Mark, Scott Murakami, Frank Parrish, Shana Brickman, Kim Moffie, Gustav Bodner, Jennifer Garrison, David Hopper, Anita May, and Rebecca Rundell. Botany Department (CF, SH, TR, JR, ES), Geography Department (GL, SL, SM, SM, FP), Hawaii Institute of Marine Biology (SB), Law School (KM) and Zoology Department (GB, JG, DH, AM, RR), University of Hawaii at Mänoa, Honolulu, HI 96822, USA.
A SYSTEMS-BASED EVALUATION OF FOREST BIRD CONSERVATION EFFORTS IN HAWAII.
Although tremendous amounts of funds, research, time and energy have gone into the preservation of Hawaii's forest birds, the birds are still declining. What aspects of the system of forest bird conservation in Hawaii are not working? As the culmination of a graduate seminar, we examined the roles of management institutions, the public, and scientists, and the interaction among these groups. Our resources included interviews, literature and Internet research, case studies, and an analysis of newspaper coverage. We found that considerable scientific understanding of the problem already exists; however, several factors still hinder progress. These include a lack of public awareness and of consistent, long-term funding, a lack of institutional flexibility, and a lack of trust among stakeholders. Successful public outreach could incorporate a sense of pride of place and community involvement in conservation efforts. Increased public awareness would in turn bolster funding. Institutional communication and evaluation are improving, but more follow-up is needed. More communication is also necessary across groups, among stakeholders with different goals. We provide specific recommendations to make the system of bird conservation in Hawaii more effective. Our analysis is not limited to Hawaiian bird conservation; any conservation effort can be approached from this systems perspective.

FISCHER, MARKUS, Michael Hock, and Melanie Paschke. Institut fuer Umweltwissenschaften, University of Zurich, Winterthurerstr. 190, CH-8057 Switzerland.
Low genetic variation reduces plant fitness in populations of a narrow endemic plant.
The sporophytically self-incompatible Cochlearia bavarica is known from <30 localities in Bavaria, SE Germany, where it occurs in isolated populations with up to 5000 flowering plants. Larger populations have a higher isozyme variability than smaller populations. After free pollination, compatibility and the cumulative fitness during 14 months in a common garden were larger for plants from larger field populations. Compared with free pollination, hand pollination (with self pollen or pollen of one donor) resulted in reduced compatibility and offspring fitness. This suggests, that several pollen donors contribute to free pollination. Positive effects of free pollination compared with hand pollination on compatibility and cumulative offspring fitness were smaller for plants representing smaller populations, which suggests increased relatedness of mates in small populations. Compared with hand-outcrossing, hand-selfing resulted in reduced compatibility and reduced offspring fitness. Our study demonstrates twofold fitness reductions for smaller populations of C. bavarica (i.e. Allee effects) associated with their reduced genetic variability: reduced cross-compatibility and reduced offspring fitness.

FISHER, ROBERT N., Boris Kuperman, Victoria Matey, Edward L. Ervin, and Manna Warburton. USGS, Department of Biology, San Diego State University, San Diego, CA 92182-1015 (RF, EE, MW), Department of Biology, San Diego State University, San Diego, CA 92182-4614 (BK, VM).
THE SILENT INVASION: PARASITES OF EXOTIC AQUATIC SPECIES IN SOUTHERN CALIFORNIA.
Exotic species are known to be the cause of declines in many species, with islands and freshwater aquatic systems showing the greatest impacts. Southern California is now the home of scores of exotic aquatic species ranging from copepods to beavers, and much of the native diversity of these systems is either extinct or declining. Surprisingly little is known of the parasitology of the aquatic organisms in southern California in general, and what exotic parasites may be associated with the exotic species. We analyzed the parasite communities of two exotic frogs and several exotic and native fish across different aquatic habitats. We found many exotic parasites associated with the exotic species, and that some had moved to native hosts. In particular, the African clawed frog was maintaining about 30% of its endemic parasite fauna in California, and fish parasites common to the aquarium trade were now occurring in natural habitats. Parasite transmission appeared tied to similar habits between native and exotic species thus putting some species at greater risk of invasion. Management and conservation implications of this study are that the removal of the exotic species is only one step and needs to be followed by intensive parasite analyses.

FLASPOHLER, DAVID, Stanley Temple, and Robert Rosenfield. School of Forestry, Michigan Technological University (DF), Dept. of Wildlife Ecology, University of Wisconsin-Madison (ST), Dept. of Biology, University of Wisconsin-Stevens Point, USA (RR).
SPECIES-SPECIFIC EDGE EFFECTS ON NEST SUCCESS AND BREEDING BIRD DENSITY IN A FORESTED LANDSCAPE.
Using natural nests of eight bird species, we provide one of the first multi-species tests for edge effects on reproductive success in a forested landscape. Our primary objective was to assess whether distance to the edge of recent clearcuts was related to nesting success in intact northern hardwood forests. Estimated nest success was generally lower for the two ground-nesting species than for the six canopy nesting species. Brood parasitism was < 3 % for species which typically accept eggs of the Brown-headed Cowbird (Molothrus ater), and nest predation was the most common cause of nest failure. Probability of nest failure was influenced by distance to forest edge for the ground-nesting Hermit Thrush (Catharus guttatus) and Ovenbird (Seiurus aurocapillus), but not for six canopy-nesting species. Our data suggest that the effect of proximity to edge on nest success for ground-nesting species may penetrate 300 m into intact forest while the effect of proximity to edge on nest density may penetrate farther. Because areas of contiguous forest (e.g., publicly owned forest) in the Upper Great Lakes remain relatively intact, they may serve as source habitat for regional songbird metapopulations.

FLEISHMAN, ERICA, and Robert B. Blair. Center for Conservation Biology, Department of Biological Sciences, Stanford University, Stanford, CA 94305-5020, USA (EF), Department of Zoology, Miami University, Oxford, OH 45056, USA (RBB).
EMPIRICAL VALIDATION OF A NEW METHOD FOR UMBRELLA SPECIES SELECTION.
Empirical validation that putative umbrella species protect many co-occurring species is rare. Using data from two taxa and three ecoregions, we tested the effectiveness of a recently-developed index for selection of umbrella species. We also tested whether species identified with the index were more effective umbrellas than species selected at random, whether sample size and intensity affect selection of umbrellas, and whether the index could identify cross-taxonomic umbrellas. The most realistic scenario, conservation of subsets of locations with relatively high numbers of umbrella species, generally would protect 0.75 of each assemblage. Randomly selected sets of species often required that more locations be designated for protection than did sets selected using the umbrella index. The umbrella index tended to identify fewer locations that offered an equivalent level of species protection. Sampling intensity affected which species were identified as umbrellas, but not the proportion of species that would be protected. Umbrella species were no more effective than randomly selected species for cross-taxonomic applications; nonetheless, neither group was significantly less effective than same-taxon umbrellas. It may indeed be feasible to identify effective umbrella species, but our studies do not support unconditional utility of the umbrella index or umbrella species concept.

FLESSA, KARL, David Dettman, Bernd Schoene, Carlie Rodriguez, David Goodwin, Miguel Tellez-Duarte, Guillermo Avila-Serrano, Michal Kowalewski, and Glenn Goodfriend. Department of Geosciences, University of Arizona, Tucson, AZ 85721, USA (KF, DD, BS, CR, DG), Facultad de Ciencias Marinas, Universidad Autonoma de Baja California, Ensenada, Mexico (MT, GA), Department of Geological Sciences, Virginia Tech, Blacksburg, VA 24061, USA (MK), Department of Geology, George Washington University, Washington, DC 20052, USA.
NEW USES FOR THE DEAD: RECONSTRUCTING BASELINE CONDITIONS ON THE COLORADO DELTA.
Paleoecological and geochemical techniques can be used to reconstruct the species composition, abundance, and environmental tolerances of marine shelly invertebrates prior to human alteration of the environment. The problem on the Colorado Delta, like so many other places, is that people didn't start making scientific observations until people had already modified the habitat. We estimated benthic shelly productivity before the upstream water diversions that began in the 1930s. We used field counts, satellite images, radiocarbon-dated shells and analyses of shell growth to estimate pre-diversion population densities of ~50 clams/m² . Surveys of the living shelly fauna indicate densities of only 3 clams/m² . The oxygen isotope composition of prehistoric shells of the bivalve mollusk Mulinia coloradoensis show that this once-dominant species thrived when salinity was lower than at present. The reduction in the number of shellfish has probably meant a diminished food supply for migratory waterfowl. Upstream dams and irrigation projects have profoundly changed the diversity and biological productivity of the Colorado River Delta in Mexico. Prehistoric shells can be used to reconstruct past diversity, composition, abundance, ecological interactions, growth rates, survivorship, salinity regimes, and environmental preferences. Dead shells provide a baseline to assess environmental impact in coastal areas.

FLETCHER, JENNIFER O., David N. Bengston, and Kristen C. Nelson. Conservation Biology Graduate Program, University of Minnesota, St. Paul, MN 55108, USA (JOF), USDA Forest Service, North Central Research Station, 1992 Folwell Ave., St. Paul, MN 55108, USA (DNB), and Department of Forest Resources, University of Minnesota, 1530 Cleveland Ave. No., St. Paul, MN 55108, USA (KCN).
POLICIES FOR MANAGING LANDSCAPE CHANGE: LESSONS FOR HABITAT CONSERVATION.
Conservation of wildlife habitat is assisted through public policies directed at land use choices. A wide range of public policies have been applied at varying spatial scales to achieve ecological, social and economic goals in managing landscapes. An extensive literature describes policies for managing landscape change, but few evaluations of policy effectiveness and impacts have been conducted. A framework for analyzing landscape change policy and program evaluations is presented; such policies range from urban growth management, to preserving natural habitat. The full range of policy tools for managing landscape change are classified, including regulatory, incentive-based, and voluntary. Lessons for policy makers, planners, and scientists include observations that: 1) spatial scale and political level should be carefully considered; 2) collaborative and incentive-based approaches may be associated with greater policy effectiveness; and 3) a dearth of empirical evaluations have been carried out. Researchers have only recently begun the task of evaluating the successes and failures of landscape change policy approaches and programs, and little is known about their effectiveness or impacts (ecological, social, economic) in spite of the longevity of some of these policies (e.g. the first state-wide growth management efforts began in Hawaii in 1961).

FLOYD-HANNA, M. LISA, Thomas L. Fleischner, and David Hanna. Environmental Studies Program, Prescott College, 220 Grove Ave., Prescott, AZ 86301, USA.
EFFECTS OF HISTORIC LIVESTOCK GRAZING AT CHACO CULTURE NATIONAL HISTORIC PARK, NEW MEXICO.
Livestock grazing is the most ubiquitous land management practice in western North America, yet it has rarely been studied in a controlled manner due to lack of large areas free of grazing. We compared the ecological effects of three grazing treatments--long-term protection, short-term protection, and currently grazed--at Chaco Culture National Historic Park (CCNHP) in northern New Mexico. CCNHP has a long history of human habitation, and is now one of the largest grazing exclosures in the American West. The effects of livestock grazing on shrub, grass, and soil crust cover and plant diversity were studied at six sites with different potential natural vegetation. Species richness was higher under long-term protection than current grazing at all six sites. There was significantly greater cover of potentially nitrogen-fixing black soil crusts at all six sites. Trends in shrub and grass response varied with the site's potential; shrub cover increased with long-term protection at 4 upland sites, and grass cover increased with protection at 4 sites. The complexities of shrub "invasion" and the role of biological potential in driving post-grazing succession will be discussed. From a conservation biology perspective, these results affirm significant ecological impacts by livestock grazing in the semiarid West.

Fonseca, Gustavo A. B., MOHAMED I. BAKARR, and Kristen Walker. Center for Applied Biodiversity Science at Conservation International, 1919 M Street, N.W., Suite 600, Washington, DC 20036, USA.
BRINGING SCIENCE AND INNOVATION TO SAVING EARTH'S BIOLOGICALLY RICHEST AND MOST ENDANGERED ECOSYSTEMS.
Conservation International's (CI) mission is to conserve our planet's biodiversity and demonstrate that human societies can live harmoniously with nature. In 1998, with a major grant from the Co-Founder of Intel Corporation, Gordon Moore, and his wife Betty, CI launched the Center of Applied Biodiversity Science (CABS) to focus on generating and disseminating science-based information to support the broader goals of the conservation community. Anchored on multi-institutional partnerships, the mission of CABS is to strengthen our ability to accurately identify and quickly respond to the emerging threats to Earth's biological diversity. CABS draws together its own researchers with other scientists in universities, research centers, multilateral government and non-governmental organizations around the world to tackle the urgent, global-scale questions of conservation science. Established partnerships span over 30 research fellows in 40 institutions distributed in 12 countries. Through this effort, CABS mobilizes leading experts in science and technology to collect and interpret data about biodiversity. Armed with this information, CABS and its partners develop strategic plans for conservation, forge key partnerships to promote conservation goals, and invigorate public awareness of and involvement in the challenges we face in saving the planet's living resources.

FONT, WILLIAM F. Department of Biological Sciences, Southeastern Louisiana University, Hammond, LA 70402, USA.
Native Hawaiian Stream Fishes: Differential Impacts of Native and Exotic Parasites.
Threats from parasites to the conservation of native Hawaiian stream fishes have been assessed. Two sources have been identified for species of native helminths presently parasitizing these native gobioid fishes: marine fishes and piscivorous birds. During their marine planktonic phase, amphidromous stream fish larvae occasionally acquire juvenile stages of marine fish helminths when feeding on copepods. Stream fishes are considered dead-end hosts for these rare parasites. Naturally colonizing birds such as herons have been responsible for introduction of helminths that use stream fishes as intermediate hosts. Because of their scarcity, native helminths are not likely to present disease threats to Hawaiian stream fishes. Exotic helminths have invaded Hawaiian streams as parasites of alien fishes, principally livebearers (Poecilidae), introduced for mosquito control and as aquarium escapees. These broadly specific parasites, including a roundworm Camallanus cotti, a tapeworm Bothriocephalus acheilognathi and a leech Myzobdella lugubris, have transferred from exotic to native fish hosts and now are the most widespread of all stream fish parasites. Because of their high prevalence and intensity, potential for pathogenicity, and rapid rates of transmission, these three exotic parasites represent the greatest disease threat to the conservation of native Hawaiian stream fishes.

FOOTE, DAVID, Cedric C. Muir, J. Kiyoko McMurry, William Haines, and Karin Schlappa. USGS Pacific Island Ecosystems Research Center, P. O. Box 44, Hawaii National Park, HI 96718 USA (DF, JKM, WH, KS), Natural Sciences, University of Hawaii at Hilo, HI 96720 (CM).
SUSTAINABILITY OF HAWAIIAN DROSOPHILA IN MANAGED LANDSCAPES OF HAWAII VOLCANOES NATIONAL PARK.
Approximately one third of the world's species of Drosophila occur only in Hawaii, including giant "picture-wing" flies with bizarre legs and mouthparts. These flies are illustrated in many biology texts as a premiere example of historical biogeography and the radiation of species. Over the past three decades, Hawaii Volcanoes National Park has lost approximately 10% of its picture-wing fauna. Local species richness at bait stations has declined much more dramatically. Olaa Forest, for instance, has lost 4 of 14 species. Mark-recapture data supports the contention that populations of individual species have declined dramatically. Morphological and genetic data suggest that hybridization is also occuring between species. The causes of picture-wing decline appear to include the loss of host plants, introduction of alien predators (such as the Western yellowjacket wasp in 1978), and reduction in overall habitat. All of these limiting factors are being remedied under current park resource management programs. In January 2001, the U. S. Fish and Wildlife Service proposed listing twelve species of Hawaiian picture-wing Drosophila under the Endangered Species Act. Long-term sustainability of picture-wing populations may depend on a combination of population protection and active habitat restoration such as that currently underway in the park.

FORSMAN, JUKKA, Mikko Mönkkönen, and Janne Seppänen. Department of Biology, University of Oulu, P.O.Box 3000, FIN-90014 Oulu, Finland.
THE EFFECT OF OTHER SPECIES ON COLONIZATION RATE AND FITNESS OF MIGRANT BIRD IN A FRAGMENTED LANDSCAPE.
We have earlier shown that some migrant birds use resident birds as a cue of profitable breeding patch. This process, heterospecific attraction, brings about variability on bird diversity among patches with respect to density of residents in the landscape. We further studied the effects of heterospecific attraction on colonization and fitness of Pied flycatchers (Ficedula hypoleuca). Study was conducted on nine isolated forest plots embedded on agricultural landscape. Experiment had two treatments: removal and addition treatment of residents (Parus spp.) with equal amount of flycatcher nest boxes per plot. Three plots received removal and six plots addition treatment. We predicted that addition plots are colonized earlier and they have higher fitness than birds on removal plots. Male and female flycatchers tended to arrive earlier on resident addition plots. The average hatching day was 1.7 days earlier and the average number of fledglings was 0.6 chicks higher on addition than on removal treatment. The results suggest that the presence and density of species itself is an important characteristics of the landscape potentially affecting colonization and fitness of later arriving species. Fragmentation coupled with heterospecific attraction may hinder later arriving species obtaining potentially good breeding patches, if fragmentation affects negatively the residents.

FORSYTH, David M., and Richard P. Duncan. Landcare Research, PO Box 69, Lincoln 8152, Canterbury, New Zealand (DMF), Soil, Plant and Ecological Sciences Division, PO Box 84, Lincoln University, Canterbury, New Zealand (RPD).
THE EFFECTS OF ISLAND AREA AND LATITUDE ON SPECIES EXTINCTION RATES.
The Theory of Island Biogeography models the number of species present on an island as a dynamic equilibrium between immigration and extinction rates. An important assumption of the theory is that smaller islands support fewer species relative to larger islands because smaller islands suffer higher extinction rates. While many data are consistent with this hypothesis, there have been few direct tests. We tested if extinction rates decrease with increasing island area by collating information on the fates of 151 deliberate introductions of six mammal species (pigs, goats, sheep, rabbits, cats and possums) to 84 of New Zealand's offshore islands. The introductions occurred decades to centuries ago, and the islands varied in size from 0.4 - 175,000 ha. Extinction rates for all six species did decrease with increasing island area. However, extinction rates also decreased with decreasing latitude, and this effect was far stronger than that of island area. Our results therefore confirm that smaller islands do experience higher extinction rates, and provide a mechanism for understanding how latitude can shape range-size.

FORSYTH, STACEY A. Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ 85721, USA.
POLLINATION BIOLOGY OF TWO MAUI ENDEMICS: POTENTIAL IMPACTS OF ALIEN INSECTS ON PLANT REPRODUCTION.
Alien species can potentially disrupt interspecific interactions in ecological communities. I investigated the pollination biology of two Maui endemics in the Hawaiian Silversword Alliance, Argyroxiphium sandwicense ssp. macrocephalum and Dubautia menziesii, and examined the impact of alien ants and bees on plant reproduction. I measured self-incompatibility, the relative effectiveness of different insect visitors as pollinators, and the degree of pollen limitation in each plant species. I assessed Argentine ant (Linepithema humile) impacts on plant reproduction by comparing seed set in ant-infested and non-infested areas. A. sandwicense and D. menziesii are both strongly self-incompatible, and are dependent on insect-mediated pollen transfer. In both species, pollination is effected by diurnal visitors, primarily native yellow-faced bees (Hylaeus sp.). Honeybees are common visitors, but are not effective pollinators. A. sandwicense was pollen-limited in non-infested areas in both 1998 and 1999, while the more widespread D. menziesii was not pollen-limited in either year. In 1997, A. sandwicense was pollen-limited in ant-infested areas, but not in non-ant-infested areas. Reduced seed set in ant-infested areas may be due to reduced numbers of Hylaeus in these areas. Results suggest that alien insects may decrease seed set in native plants through competition with, or predation on, native pollinators.

FOSTER, JEFFREY T., Erik J. Tweed, Bethany L. Woodworth, Richard J. Camp, and Corey D. Adler. University of Illinois, 606 E. Healey St., Champaign, IL 61820, USA (JTF), USGS-BRD Kilauea Field Station, P.O. Box 44, Hawaii National Park, HI 96718, USA (EJT, BLW, RJC, CDA).
AVIAN POPULATION TRENDS IN THE ALAKAI SWAMP, KAUAI: NATIVE DECLINES AND EXOTIC EXPANSIONS?
The Hawaiian avifauna has declined severely since human colonization, with losses exceeding 80 native bird species. Kauai's Alakai Swamp, the only remaining native forest bird habitat on the island, has not been immune to the problems plaguing Hawaii's birds, where five native bird species have gone extinct within the last 30 years. We conducted avian surveys of the entire Alakai Swamp in March and April 2000 and compared them to a similar survey from 1973 to determine long-term population changes of native and introduced species. Our work revealed unchanged or increasing bird populations for five of the seven most common native species. At least four introduced species increased in number, one was extirpated, and one new species invaded. Additionally, when compared to surveys conducted in the interior of the swamp in 1981, 1989, and 1994, only two native species in the 2000 survey showed significant population declines. Thus, most native bird populations in the interior and periphery of the Alakai Swamp remain healthy despite numerous limiting factors. Nonetheless, the avifauna of Kauai remains extremely vulnerable and active management of introduced plants, animals, and disease is essential in preventing further habitat degradation and native bird loss.

FOSTER, SOPHIE E. and Daniel A. Soluk. University of Illinois, Department of Natural Resources and Environmental Sciences, Urbana, IL 61801, USA (SEF, DAS), Illinois Natural History Survey, Center for Aquatic Ecology, 607 E. Peabody Drive, Champaign, IL 61820, USA (SEF, DAS).
HABITAT USE OF MALE AND FEMALE HINE'S EMERALD DRAGONFLY, SOMATOCHLORA HINEANA, A FEDERALLY LISTED SPECIES.
Understanding habitat requirements of endangered species is key to ensure their protection. Knowledge of the distribution and habitat use of female Hine's Emerald Dragonfly is limited. At two sites in Door County, Wisconsin, we monitored male and female Hine's Emerald Dragonfly (HED) in breeding and non-breeding habitats. Two to four netters, with at least one netter in each habitat type caught, sexed, recorded behavior, and individually marked HED with bee tags. Sampling was conducted between 10AM and 3PM on sunny days with low winds. Netting sessions consisted of 3-50 minute periods repeated 3-4 times at each site. Females were more abundant in the non-breeding habitats representing 70% of the HED caught there. Female use of non-breeding habitats was highly variable and may be influenced by temporal availability of food resources. Males were more abundant in the breeding habitats, however the sex ratio there was less consistent and ranged from 50% to 80% male. This study links wetlands to their adjacent habitats and demonstrates that preserving both is important for the management and recovery of the Hine's Emerald Dragonfly. These links may be equally as important for other wetland species that depend on terrestrial habitats for part of their lifecycle.

FOUFOPOULOS, JOHANNES, and Martin Wikelski. Department of Ecology and Evolutionary Biology, Guyot Hall, Princeton University, Princeton, NJ 08544-1003, USA.
INVASIVE PATHOGEN THREATS TO BIRDS ON ISLANDS: LESSONS LEARNED FROM HAWAII AS APPLIED TO THE GALAPAGOS.
Introduced diseases like avian malaria and pox have been implicated in the extinction of birds in Hawaii and other places. The Galapagos islands harbor a unique set of endemic birds and have suffered few, if any extinctions. Nonetheless, some avian populations are now in decline and several exotic pathogens appear to have already been introduced to the islands. This talk outlines the conclusions from an October 2000 workshop that evaluated the potential threat of diseases and parasites to the avifauna of the Galapagos. A panel of 27 scientists and managers identified the most likely avenues of invasion, as well as the pathogens posing the highest risk to local bird populations. Some of the recommendations made to prevent the introduction and spread of exotic diseases include the establishment of a realistic monitoring system, enhancements to the current quarantine program, the evaluation of different intervention options and the establishment of contingency plans. These guidelines can be adapted to birds or other vertebrates on other oceanic islands.

FOX, GORDON A., and Bruce Kendall. Dept. of Biology (SCA 110), University of South Florida, 4202 E. Fowler Ave., Tampa, FL 33620, USA (GAF), and Donald Bren School of Environmental Science and Management, University of California, Santa Barbara, Santa Barbara, CA 93106, USA (BK).
HOW DEMOGRAPHIC VARIATION AMONG INDIVIDUALS AFFECTS EXTINCTION RISK OF POPULATIONS.
Individuals in populations have differing demographic properties. We have recently shown that treating them as identical (as in PVA models) causes potentially serious misestimation of the importance of demographic stochasticity. Here we present results on how demographic variation among individuals affects extinction risk. We can calculate these effects by using the sensitivity analysis of a demographic model to estimate the variance of the population's growth rate. Populations experiencing high variance in their growth rate are at increased risk of extinction. In comparison with a population of identical individuals, variation among individuals in survival always reduces the effect of demographic stochasticity, and thus decreases the extinction risk. Variation in fecundity has more complicated effects: it can either increase or decrease the extinction risk, depending on whether variance in fecundity has a convex or a concave relationship to the mean fecundity.

FRANKHAM, RICHARD. Department of Biological Sciences, Macquarie University, Sydney, NSW 2109, Australia.
POPULATION SIZE, GENETIC DIVERSITY AND CONSERVATION.
Genetic diversity is of conservation concern as it is required for evolutionary change. Further, loss of genetic diversity is related to inbreeding and this reduces reproductive fitness and so directly increases extinction risk. Endangered species have, by definition, small, or declining populations. Theory predicts that loss of neutral genetic diversity is related to effective population size. Experimental data indicates that heterozygosity is related to population size for both molecular and quantitative measures of genetic diversity, as predicted. The main reason for deviations from the predicted relationship between heterozygosity and population size in small populations is associative overdominance. Quantitative genetic variation determines evolutionary potential, but overwhelmingly information on genetic diversity comes from molecular measures. Based upon a meta-analysis, molecular measures of genetic diversity are only weakly related to quantitative genetic variation, and there is no relationship for fitness characters. The predicted correlationship between genetic diversity and population fitness has been confirmed in a meta-analysis. Evolutionary change in the medium to long-term depends upon initial additive genetic variation, upon effective population size and upon reproductive excess. Consequently, most endangered species have severely compromised ability to evolve as compared to non-endangered species.

FREED, LEONARD A. Department of Zoology, University of Hawaii at Mänoa, Honolulu, HI 96822, USA.
LIFE HISTORY CHARACTERS UNDERLIE ENDANGERMENT IN HAWAIIAN HONEYCREEPERS.
While causes of endangerment are well known in general for individual species, particularly plants, less attention has been directed on determining why some animal species in a community are endangered and others are not. Hawaiian Honeycreepers (Drepanidinae) provide a system for studying comparative endangerment. Based on census data from the Hawaii Forest Bird Survey, there is a strong correlation between distribution and abundance of six species of necarivorous and insectivorous Hawaiian Honeycreepers on the Island of Hawaii. Three of these are listed as endangered and, not surprisingly, these taxa have more restricted range and lower density than the unlisted taxa. A comparison of life history characters, such as habitat selection, clutch size, number of broods, and duration of the fledgling period, reveals substantial differences between the endangered and unlisted species. In all of these characters, endangered species are more conservative than unlisted species. One implication is that endangered honeycreeper populations are more susceptible to habitat disturbance and slower to recover. A second implication is that the lower reproductive rate decreases the rate at which new alleles appear that may be relevant for adapting to novel selective pressures.

FREILE, JUAN F. Departamento de Biología, Pontificia Universidad Católica del Ecuador, and Center for Tropical Research, San Francisco State University, USA.
CONSERVATION STATUS OF ANTPITTAS IN THE GENERA GRALLARIA AND GRALLARICULA (FORMICARIIDAE) IN ECUADOR: AN ASSESSMENT OF CONSERVATION STATUS RANKING MODELS.
Grallaria and Grallaricula antpittas are among the least known birds in the Neotropics, making conservation status assessment inaccurate. Due to their restriction to dense forest interior most of them are highly vulnerable to habitat alteration and fragmentation. I evaluated the conservation status of eighteen species recorded in Ecuador based on data on their geographic distribution. Using the IUCN evaluation model, two species previously considered as Endangered were ranked as Vulnerable, and two species not previously included in any threat category were ranked as Endangered. Only one species was ranked in the same category as before. These marked differences reveal the importance of basic distributional data on poorly-known species for improving the accuracy of their status ranking. Another evaluation model was also used in the species conservation status analysis. This model takes into account the habitat restrictions for each species, the current threat status of their habitats, and the protection level of their ranges within the country. Five species were classified as Near-threatened, one as Vulnerable and one as Endangered, with only one species having the same status as proposed by the IUCN system. This alternative model requires further evaluations to assess its limitations and applicability.

FRETZ, J. SCOTT. State of Hawaii, Division of Forestry and Wildlife, 1151 Punchbowl Street, Room 325, Honolulu, HI 96813, USA.
RELATIONSHIP OF FOOD AVAILABILITY TO DISTRIBUTION AND ABUNDANCE OF THE HAWAII AKEPA.
Extinction rates among Hawaiian forests birds have been high since humans colonized the islands. Habitat loss, introduced predators, and diseases have been implicated in these losses and are thought to limit existing populations. However, no studies have explored the importance of food to avian populations in wet and mesic forests. I developed methods to estimate food availability for the endangered, insectivorous Hawaii Akepa (Loxops coccineus coccineus) at site where akepa persist in relatively large numbers. I explored the relationship of food availability to reproductive success, and then used this baseline data to explore the interaction of habitat degradation and food at a site where akepa have declined drastically. I found that food is associated with reproductive success at the site where akepa are abundant. At the site where akepa are rare I found no evidence of direct disturbance to the arthropod community that provide food for akepa. However, integration of habitat structure data with the estimates of food availability per unit of habitat area revealed that food was far lower at that site. These results suggest a role for food in limiting akepa populations, and suggest the need to consider the importance of food at sites where akepa are rare.

FRIDAY, JAMES B., Paul G. Scowcroft, Adrian Ares, and Sean M. Gleason. University of Hawaii, Dept. of Natural Resources and Environmental Management, 875 Komohana St., Hilo, HI 96720, USA (JBF), USDA Forest Service Institute of Pacific Islands Forestry, 1151 Punchbowl St., Rm. 323, Honolulu, HI 96813, USA (PGS), University of Hawaii Dept. of Natural Resources and Environmental Management, 1910 East West Rd., Honolulu, HI 96822, USA (AA and SMG).
INVASIVENESS OF TROPICAL ASH IN A NATIVE HAWAIIAN FOREST.
Tropical ash (Fraxinus uhdei) was introduced in Hawaii in the 1930s as a timber species and has naturalized in native forests. We re-surveyed 32 plots in the Laupahoehoe forest reserve which were originally established after selective logging for koa (Acacia koa) in 1973 and included both native forest and ash stands. Basal area of ash increased from an average of 0.8 m2/ha in 1973 to 9.9 m2/ha in 2000. However, ash did not spread more than 500 m from where it was found in 1973. Basal area of native trees also increased slightly, from 18.9 m2/ha to 19.8 m2/ha, with koa showing an increase and Metrosideros polymorpha and other native species decreasing. Abundance of tree ferns (Cibotium spp.) decreased from an average of 47 stems/ha to 9 stems/ha. Basal area of native trees and number of understory species were both inversely related to basal area of ash. Ash seed rain was heavy, with up to 16,000 seeds/m2 collected in six months. If the ash plantation is opened up for harvest, managers should plan ways of regenerating the more valuable native koa rather than the ash and preventing the spread of ash into adjacent native forest.

FUJIWARA, MASAMI, and Hal Caswell. Woods Hole Oceanographic Institution, Biology Department MS#34, Woods Hole, MA 02543, USA.
INCORPORATING TEMPORARY EMIGRATION IN MARK-RECAPTURE STATISTICS.
Mark-recapture statistics are used to estimate survival probabilities of threatened species from capture histories of marked individuals. Such data often include temporary emigration of individuals from the study area, which causes biased estimates of survival. Emigrated individuals have a true capture probability of zero, but mark-recapture methods assume all individuals, inside or outside of the study site, have the same non-zero capture probability. This results in under- and over-estimation of the capture probability for individuals inside and outside of the study site, respectively. Because capture probability is inversely related to survival probability in the likelihood function, estimated survival probability is biased. Here, we present new statistical methods to estimate survival probability from mark-recapture data including temporary emigration. We show how to solve this problem by using stage-structured models that include one or more stages that represent individuals that have temporarily emigrated, and computing the likelihood function from this stage structure. We apply the method to simulated data representing different life histories, for which underlying parameter values are known, and demonstrate unbiasedness of estimators. Our results apply to seabird, sea turtle, and marine mammal data where individuals are sampled only on their breeding grounds.

FULLER, JULIE A., and L. Scott Mills. Wildlife Biology Program, Univ. of Montana, Missoula, MT 59812, USA.
QUANTITATIVE MOTION ANALYSIS AND SPATIAL DYNAMICS OF SNOWSHOE HARES IN DIFFERENT HABITAT TYPES.
Snowshoe hares are the primary prey of the federally Threatened Canada lynx. Forest structure affects hare population dynamics, and the U.S. Forest Service has ceased pre-commercial thinning of National Forest in the Northern U.S. until this relationship is understood. As part of a larger study investigating snowshoe hare population dynamics, we employed a novel, theoretical approach - Quantitative Motion Analysis (QMA) - for assaying fitness without having to capture the animal (i.e. a non-invasive sampling approach). During two winters across four forest thinning treatments in Northwestern Montana, we located radio-collared hares and backtracked their paths over snow. We measured turning angles, distance to shelter, and behavior for 70m of path, and compared the tortuosity (lack of directional bias) of paths in each treatment. Using this information as a baseline, we elicited realistic predator-avoidance behavior by flushing hares with a domestic dog. Tortuosity, considered indicative of habitat fitness, was highest in closed mature forest (.59), but dropped to .28 in open young forests. Hares in predation trials exhibited similar tortuosity to open young forest (.23), implying that pre-commercial thinning decreases habitat value for hares. We believe QMA may be a useful supplement to traditional invasive approaches that evaluate wildlife responses to human perturbations.

Furman, Bonnie J., Roger Greenwell, Mitisha Hudson, and William S. Dvorak. Department of Biology, Western Kentucky University, 1 Big Red Way, Bowling Green, KY 42101, USA (BJF, RG, MH) and Department of Forestry, North Carolina State University, Raleigh NC 27695, USA (WSD).
Evidence, Direction, and Effects of Interspecific Hybridization in Fragmented Populations of Pinus Species.
Disturbance due to agriculture and the harvest of trees for fuelwood and for timber has resulted in a sharp reduction in the population numbers of many of the tropical pine species. It is proposed that fragmentation accelerates the process of interspecific hybridization and that such hybridization may increase genetic diversity and diminish genetic integrity of individual tree species. Population-level assessment of RAPD marker diversity yielded several species diagnostic markers potentially useful in the assessment of interspecific hybridization for a number of pine species. Populations of Pinus tecunumanii and P. caribaea were collected from the Mountain Pine Ridge, Belize, an area that has undergone extensive fragmentation and where these two species are sympatric. Many individual trees in these populations appear to be morphologically intergraded. RAPD marker data on individual trees suggest that interspecific hybridization is occurring, as individuals contain a number of marker loci considered diagnostic for the alternate species. Chloroplast microsatellite data were obtained to investigate directionality of hybridization, as chloroplasts are paternally inherited in pine. These populations provide an invaluable resource for studying the effects of fragmentation on both genetic structure and diversity, the results of which would be immensely valuable for planning and implementing global conservation efforts.

GALBRAITH, MEL. School of Landscape and Plant Science, UNITEC Institute of Technology, Private Bag 92025, Auckland, New Zealand.
PUBLIC REWARDS AS PATHWAYS TO SUCCESS IN ECOLOGICAL RESTORATION: THE TIRITIRI MATANGI ISLAND (NEW ZEALAND) EXPERIENCE.
Slow rates of natural revegetation and limited funding are common factors that impact on the success of ecological restoration projects. Public involvement on Tiritiri Matangi Island, New Zealand, has shortened the restoration time frame and has allowed access to funding unavailable to government authorities. Public involvement on Tiritiri Matangi was initiated in 1984 with a replanting program using volunteers from conservation groups. This involvement has since broadened to encompass a wider cross-section of the community contributing to an increasing range of management and advocacy activities, firmly establishing public partnership in the project. On-going rewards', such as open access to a scientific reserve and participation in the release of translocated species, have sustained public interest in the project and facilitated advocacy for conservation. The establishment of a non-profit non-governmental organisation, the Supporters of Tiritiri Matangi, has enhanced and reinforced the opportunities for rewards, with the organisation itself contributing significantly to the project's success. The involvement of public was innovative (and controversial) at the time that the project was initiated, but is now an accepted, and indeed expected, management strategy in New Zealand conservation. The establishment of an autonomous management system with representation of key stakeholders is currently being explored.

GALVES AGUILERA, XIOMARA. Empresa Nacional para la Protección de la Flora y Fauna, MINAG, Ave. 42 No.514 esq. 7ma. Playa, Ciudad de La Habana, Cuba.
CURRENT SITUATION OF CUBAN GLOBALLY THREATENED BIRDS: CASE STUDIES OF CONSERVATION PROGRAMS.
In 1974 Buide et al. listed 21 species of Cuban birds as threatened. Twenty years later, Perera et al., and Gonzalez and Llanes listed 46 and 17 species, respectively, as having conservation problems. Bird Life (1998) considers 24 Cuban species to be globally threatened. We evaluated the status of the 20 bird species mentioned by most authors, of these 13 are forest residents and 9 are endemics. All forest and some savannah species are threatened by habitat degradation and fragmentation. Unjustified killing of "predators" (Cuban Kite and Gundlach's Hawk), poaching of nests and adults (Cuban Parrot and Cuban Parakeet), and illegal hunting (Plain Pigeon, West Indian Whistling-Duck and Cuban Sandhill Crane) are additional causes of endangerment. 100% of threatened species are represented in over 73 protected areas established for their conservation, many with infrastructure and specialized field staff conducting management and research. The Parrot, Parakeet and Crane conservation programs are the most advanced and new programs for other species are being initiated. Conservation of endangered species have been linked to educational programs targeted to city and rural communities. The charismatic, endangered species have served in Cuba as "umbrellas" under which ecosystems are conserved by generating support from the local population.

GARCELON, DAVID K., Kate R. Faulkner, Jan K. Larson, and Peter T.Schuyler. Institute for Wildlife Studies, Arcata, CA 95518, USA (DKG), Channel Islands National Park, Ventura, CA 93001, USA (KRF), Navy Region Southwest, San Diego, CA 92135, USA (JKL), Santa Catalina Island Conservancy, Avalon, CA 90704, USA (PTS).
FERAL ANIMAL REMOVAL ON THE CALIFORNIA CHANNEL ISLANDS.
The California Channel Islands were the site of many introductions of domestic livestock that eventually went feral. The four federal agencies or private organizations that administer the islands independently determined that feral animals were causing sufficient long-term environmental damage that their removal was mandated. Between 1972 and 1991 the US Navy removed 28,000 feral goats and 2,200 feral pigs from San Clemente Island using a combination of live-capture/transport and lethal measures, leaving the island free of exotic herbivores. On Santa Rosa Island, the US Park Service successfully eradicated feral pigs over a 3-year period. On Santa Cruz Island, The Nature Conservancy lethally removed 37,000 feral sheep over 5.5 years on 90% of the island. The remaining 9,700 sheep were recently removed by the US Park Service. On Santa Catalina Island, animals removal is on-going, with over 8,000 feral goats and 11,000 feral pigs removed primarily by shooting. Most of these programs have spanned several years, have been challenged by animal rights groups, and have cost hundreds of thousands to millions of dollars to complete. Eradication of introduced animals can be successful if sufficient resources and commitment are dedicated to the project.

GASSMANN-DUVALL, RENATE, and Anne S. Carter. Maui Invasive Species Committee, Cooperative Park Studies Unit, Department of Botany, University of Hawaii at Mänoa, 3190 Maile Way, Honolulu, HI 96822, USA.
FREE RANGING PARROTS ON MAUI, HAWAII : DISTRIBUTION, THREATS AND CONTROL POSSIBILITIES.
Ten parrot species have been documented from Maui since 1977. The most detectable species to date occupying the windward lowland rainforest on East Maui are the mitred conures (Aratinga m. mitrata). Originating from a pair of pet birds released in 1986/87 in Huelo the population is now estimated to be 150 to 200 birds. Our observations showed that these birds roost in seacliff areas from Waipio-nui to Huelo point, forage in higher valleys during early morning and late afternoon hours. They fly mostly in pairs or triplets in noncohesive flocks of up to 50 birds. We saw them feeding on Chinese banyan (Ficus platypoda), common guava (Psidium guajava) and rose apple (Syzigium jambos). The rapid mitred conure population growth, expanding range, potential to transmit diseases to Hawaiian birds and the proven dispersal of nonnative tree species, poses serious threats for native Hawaiian flora and fauna. Furthermore, the ability to damage commercial fruit and seed farms is an economic concern. This poster provides baseline data about the dynamics, range and feeding pattern of mitred conures on Maui. Maui's parrot distribution is mapped and possible management techniques are presented.

GELBARD, JONATHAN L., and SUSAN HARRISON. Department of Environmental Science and Policy, University of California at Davis, Davis, CA 95616, USA.
EFFECT OF SPATIAL ISOLATION (FROM ROADS) ON NATIVE AND EXOTIC PLANT DIVERSITY.
We examined the effect of isolation from roads on native and exotic plant diversity in a California foothill grassland landscape. We measured native and exotic plant diversity and cover in sites stratified by isolation (10 m, 100 m, and > 1000 m from roads), soils (serpentine and non-serpentine), and slope (cool, warm, and neutral). In non-serpentine grasslands, native cover was greatest in sites > 1000 m from roads (22%) and least in sites 10 m from roads (8%), and habitats > 1000 m from roads contained a significantly greater percent of native species (44%) than those 10 m (29%) or 100 m (35%) from roads. In serpentine grasslands, the percent of native species and native grass and forb diversity were greatest in isolated sites. Two exotic species that have recently been observed to be spreading (Aegilops triuncialis and Centaurea solstitialis) were least prevalent, and one native bunchgrass (Nassella pulchra) was most prevalent, in isolated sites. Native species were generally most prevalent on infertile sites on both soil types, and were most prevalent on cool, ungrazed slopes on nonserpentine soils. Roadless areas are significant refuges for native species, but careful management is important to protect these habitats from continued invasion.

GERBER, LEAH R. National Center for Ecological Analysis and Synthesis, University of California, Santa Barbara, CA, USA.
DEMOGRAPHIC SENSITIVITY ANALYSIS AS A TOOL FOR MARINE RESERVE DESIGN.
I will explore the idea that a sensitivity analysis conducted on stage-specific demographic rates can be used to help guide the design of marine reserves, whereby reserves target life history stages for which protection will accomplish the most in terms of fostering population growth. In the marine realm, an interesting facet of pleas for reserves is that most reserves are quite small in area, and represent only a small portion of the total range of species. Recognizing that we may not be able to protect all portions of the range for wide-ranging marine organisms, I use demographic analysis to help focus management efforts on critical life stages for a representative range of species life history characteristics. We synthesize data for marine mammals, fish and marine invertebrates and cluster species by elasticity in order to identify critical life history stages to be targeted in marine reserve. Our results indicate that reserves for marine mammals may have fundamentally different goals than those for fish and invertebrates (i.e., protection of juveniles in fixed spatial locations vs. protection of adult rockfish). Standardized demographic analysis may be useful in identifying disparate conservation goals for marine reserve design for species with distinct life histories.

GERWEIN, JOEL B., and Richard V. Kesseli. Biology Department, University of Massachusetts at Boston, 100 Morrissey Blvd., Boston, MA 02125, USA.
EFFECTS OF HISTORIC DEFORESTATION ON GENETIC DIVERSITY OF OLD-GROWTH AND SECONDARY QUERCUS RUBRA.
Most forests of New England were cleared in the last 200 years. This study assesses the genetic consequences of this deforestation by comparing genetic diversity of old-growth (OG) and secondary Q. rubra populations in four areas of Massachusetts. These areas underwent a reduction of forest cover from >80% to as low as 29% in the mid-1800s, followed by recovery to the current level of >70%. Genetic diversity was assessed using six nuclear simple sequence repeats (SSRs) and three chloroplast SSRs. Q. rubra stands in all four areas show high within-stand diversity and low differentiation between stands. Chloroplast SSRs show greater differentiation than nuclear SSRs, consistent with short-distance seed dispersal. Preliminary data show no correlation between forest history and allelic richness of chloroplast SSRs, but nuclear SSR data show that OG stands contain 32 alleles lacking in the secondary stands, while secondary stands contain only 17 alleles lacking in OG. This may be due in part to lower secondary sample sizes. These data suggest that historic deforestation in Massachusetts may have caused a loss of allelic richness. Thus, OG stands may be valuable genetic resources that should be preserved to enhance long-term evolutionary potential of the species.

GHATE, UTKARSH, Shonil Bhagwat, Yogesh Gokhale, and Raghunandan Velankar. Research and Action in Natural Wealth Administration, C-26/1, Ketan Heights, Kothrud, Pune, 411 029, India (GU, SB, YG, RV), Oxford Forestry Institute, University of Oxford, OX1 3RB, UK (SB), Bombay Natural History Society, S. B. Road, Mumbai, 400 023, India (GU,YG).
SACRED FORESTS AS MODEL CONSERVATION ISLANDS.
Sacred groves are relict forest pockets with harvests prohibited or regulated by the local people under religious pretext. Ranging from few trees to hectare in size, such groves represent several small conservation islands scattered allover the country. Tree enumeration along 0.1 ha strip transects reveals that these forests are no less diverse but shelter more rare or keystone or endemic species than the surrounding landscape, including large forest patches. Compared to larger, harvested forest patches in the neighbourhood, sacred groves harbour greater proportion of primitive and animal dispersed but lower proportion of animal pollinated trees. Higher assemblage diversity amongst groves as evident from higher beta diversity between transects distinguishes their conservation potential relative to larger forest patches. Their vital ecosystem and social services confer them long-term sustainability and competitive advantage in the context of SLOSS debate. Groves reconcile social and ecological concerns and offer important lessons for the state driven protected areas (PA) paradigm that employs few large wildlife reserves, threatened by industrial exploitation besides villager's resistance given curbs on their livelihood activities. In contrast, despite its recent erosion driven by socio-economic changes, the sacred groves tradition still shows signs of restoration and resurgence.

GHAZANFAR, SHAHINA A., Gunnar Keppel, and Shafia Khan. University of the South Pacific, P.O. Box 1168, Suva, Fiji. Present address: 6A Burnt Close, Grantchester, Cambridge CB3 9NJ, UK (SAG)
THE ASTROLABE ISLANDS OF FIJI: STATUS OF THE TERRESTRIAL VEGETATION.
Eleven islands lie in the Great Astrolabe Lagoon of Kadavu, Fiji. The largest three islands are inhabited, have subsistence and commercial plantations while five other islands are used as pastures for goats. A survey of the vegetation of all islands showed that species richness was higher in the inhabited islands but that they contained more introduced species (39% to 47% of total species). Species richness on goat-grazed islands was low (24% to 12 %) and contained a higher proportion of grasses. Over-grazing resulted in soil erosion, exposed tree roots and bare rocks. An uninhabited island with least disturbance had the most native species (95%) with several species not found on other islands. In terms of conservation, the vegetation of goat-grazed islands is vulnerable and loss of species through natural phenomenon may result in the local extinction of rare or endemic species. Lack of soil cover would make the establishment of species impossible. On small oceanic islands where the natural establishment of species is dependent on several abiotic factors including suitable coastline, currents and substrate, and where the natural loss of species can occur through cyclones, man-made disturbance must be kept at a minimum to avoid habitat and species loss.

GIBBONS, STEPHEN T. National Park Service, 909 First Avenue, Seattle, WA 98104, USA.
THE NATIONAL NATURAL LANDMARKS PROGRAM: A CONSERVATION PROGRAM ISOLATE BUT IN NEED OF CONSERVING.
Established by the Secretary of the Interior back in 1962, the National Natural Landmarks (NNL) Program is an honorary designation that recognizes the conservation efforts of state and federal agencies, along with private landowners. Unlike its much more recognizable National Historic Landmarks (NHL) sister-program, the NNL program has managed to sustain itself amidst a self-imposed, but necessary, moratorium and repeated attacks from program detractors. This feeling of isolation frustrated program officials for over a decade until the lifting of the servicewide moratorium in 1999. In retrospect, this scrutiny has proven to be a godsend both for the 587 existing NNL's, as well as, strengthening the resolve of the National Park Service to build upon an already successful conservation program. From well-known Diamond Head NNL, Hawaii to little-known Dinosaur Trackway NNL, Connecticut conserving biological and geological resources in partnership with all people is the name of the game, or better put, the National Natural Landmarks Program.

GIBBS, JAMES P., and W. Gregory Shriver. College of Environmental Science and Forestry, 350 Illick Hall, State University of New York, Syracuse, NY 13210, USA.
CAN ROAD MORTALITY LIMIT TURTLE POPULATIONS?
All tortoises and about one third of aquatic and semi-aquatic turtles in the United States currently require conservation action. Turtle life histories are characterized by high adult survival rates and delayed sexual maturity, which together severely constrain the ability of turtle populations to cope with additive mortality sources. We examined the potential for road-associated mortality to limit regional turtle populations by integrating simulations of turtle movements with maps of actual road networks and the traffic volumes upon them. Combinations of road density and traffic volume in the Northeastern, Southeastern, and Great Lakes regions were predicted to generate annual additive road-associated mortality in land turtles well in excess of 10%, a threshold level of additive mortality most turtle species likely cannot absorb and still maintain positive population growth rates. States in the Pacific, Mountain-Prairie, and Southwest regions were at or below the threshold. We conclude that the demographic traits of land turtles in combination with their mobility jeopardize their persistence within the road networks characteristic of many regions of the United States. This may place land turtles in the company of grizzly bears and gray wolves as fauna for which road networks may be a key limiting factor to population recovery efforts.

GILBERT, GREGORY S., and Ingrid M. Parker. Environmental Studies Department, University of California Santa Cruz, CA 95064 , USA (GSG), Department of Ecology and Evolutionary Biology, University of California Santa Cruz, CA 95064 (IMP).
ISLANDS AND CONNECTIVITY IN AN EPIDEMIC OF AN INVASIVE PALM DISEASE IN KUNA YALA, PANAMA.
Porroca is a newly emerging, invasive, lethal disease of coconuts in Panama. Formerly restricted to areas around Cartagena, Colombia and the Panama-Colombia border, the disease recently entered an epidemic phase, spreading along the Caribbean coast and inland to 40 km west of the Panama Canal. In 1998, 1999, and 2000 we mapped and determined the health status of several hundred thousand coconut palms on more than 300 km of coastline and 300 islands in the Comarca of Kuna Yala, the indigenous reserve of the Kuna Nation. Disease infections were patchy, mostly restricted to mainland sites, and most severe in long stretches of contiguous coconut palms. During the study period there was a 13-fold increase in the number of infected islands, but with most new infection sites restricted to a small number of palms. Epidemiological patterns indicated rare long-distance dispersal of the pathogen is important in establishment of the disease in previously uninfected coastal areas and islands, and that continued disease development is highly dependent on local conditions. Coconut is the primary economic resource of the Kuna, and loss of coconuts is likely to have a major impact on management intensity of their forest resources.

GINSBERG, JOSHUA, and Rosie Woodroffe. Wildlife Conservation Society, 2300 Southern Bvd., Bronx, NY 10460, USA (JG), Department Of Wildlife And Fisheries Biology, University Of California - Davis, Davis, CA 95616, USA (RW).
SOURCES, SINKS AND MANAGEMENT OF LANDSCAPE SPECIES: THE AFRICAN WILD DOG.
Landscape species use large, ecologically diverse areas. Their requirements in time and space render them especially susceptible to human use and alteration of natural landscapes. Using data from a five-year study in Zimbabwe, and from the literature, we examine how movement across a matrix of human land-use types affects the persistence of a population of African wild dogs (Lycaon pictus). Our analyses suggest that restoration of habitat in a human-dominated matrix (population sink) can have the perverse effect of drawing wild dogs out of a protected area (population source) thus reducing the short-term probability of persistence of the population overall. Such complex dynamics suggest that if restoration efforts are to be successful for wide-ranging species they must address direct and indirect causes of human-induced mortality across a landscape.

GIOVAMBATTISTA, GUILLERMO M., M. Verónica Ripoli, J. Pedro Lirón, Pilar Peral-García, and Juan L. Bouzat. Universidad Nacional de La Plata, La Plata B1900 AVW, Argentina (GGM, MVR, JPL, PP), and Bowling Green State University, Bowling Green, OH 43403, USA (JLB).
MHC DIVERSITY IN CONTINENTAL ISLANDS OF SOUTH AMERICAN CREOLE CATTLE.
Genetic diversity at the major histocompatibility complex (Mhc) has been shown to be important for the organisms' adaptive immune response. The predominance of polymorphisms at peptide binding sites of Mhc molecules, even allele frequency distributions, and excess of non-synonymous substitutions suggest that balancing selection may be the major evolutionary force driving Mhc dynamics. Genetic analysis of the MHC-DRB3 locus from 11 populations of South American Creole cattle suggests that founder events and genetic drift may have played important roles in the maintenance of genetic diversity in this system. Both average expected heterozygosity (0.845) and allelic diversity (13.4) varied considerably among populations. We detected 33 DRB3 PCR-RFLP-defined alleles, with individual populations having from 9 to 22 alleles. All populations showed highly significant levels of genetic differentiation (F_ST=0.071; P<0.001). Contrary to what has been observed in other species, all populations studied revealed no excess of heterozygotes (F_IS= 0.004; P=0.061). In addition, neutrality tests showed no evidence for either balancing or directional selection (Slatkin's exact test; 0.095 < P < 0.856). The characteristic population structure of the South American Creole cattle, with semi-isolated herds and low levels of artificial selection, may have increased the effects of stochastic processes such as founder events and genetic drift.

GISLER, STEVEN, and Robert J. Meinke. Oregon Department of Agriculture Native Plant Conservation Program, and Oregon State University, Department of Botany and Plant Pathology, 2082 Cordley Hall, Corvallis, OR 97331-2902, USA.
BALANCING THE VALUE OF SEEDS AND NATIVE SEED PREDATORS IN THE RECOVERY OF THE THREATENED PLANT SPECIES, SIDALCEA NELSONIANA (MALVACEAE).
Pre-dispersal predation of seeds by insects is widespread among flowering plants, and among host-specific predators, is generally considered a stable predator-prey system. However, for rare plants already threatened by anthropogenic pressures, seed predation (even at equilibrium levels) may tip the scales farther towards extinction. In this study we examined the efficacy of insecticide application in reducing seed predation by host-specific weevils in the threatened species, Sidalcea nelsoniana (Malvaceae). Because these weevils are themselves rare native species, and also host an undescribed parasitic wasp, we continued seed predation measurements the year following treatment to estimate long-term impacts to predators. Results showed that insecticide application dramatically reduced seed predation in the study population to less than 1 percent, compared to a mean 82.5 percent over the previous 2 years, and 66.8 percent in a nearby untreated population. Resulting enhanced seed yields contributed significantly to off-site seed banking and re-introduction projects. In the year following insecticide application, seed predation increased to 28.7 percent, indicating a partial rebound in the predator population. This study suggests that periodic insecticide applications may temporarily bolster seed production in rare, predation-limited plants, without permanently crippling populations of associated native predators.

GLEASON, SEAN M, Adrian Ares, James B. Friday, and Paul G. Scowcroft. University of Hawaii Dept. of Natural Resources and Environmental Management, 1910 East West Rd., Honolulu, HI 96822, USA (SMG, AA), University of Hawaii Dept. of Natural Resources and Environmental Management, 875 Komohana St., Hilo, HI 96720, USA (JBF), USDA Forest Service Institute of Pacific Islands Forestry, 1151 Punchbowl St., Rm. 323, Honolulu, HI 96813, USA (PGS).
TROPICAL ASH INVASION INTO NATIVE HAWAIIAN WOODLANDS: AN ANALYSIS OF PHOTOSYNTHETIC CHARACTERISTICS AND CARBOHYDRATE STORAGE STRATEGIES.
Fraxinus uhdei (tropical ash), a species introduced to Hawaii from Mexico, is known to invade disturbed native forests and have higher growth rates when nitrogen is not a limiting factor (Ares and Fownes 2001). We examined physiological, structural, and morphological characteristics of koa and tropical ash to explore possible mechanisms of invasion. Seedlings of both species were grown in a greenhouse under three shade treatments: 100% sun, 60% sun and 30% sun. Structural and non-structural carbohydrates in roots and stems were measured to examine the carbohydrate reserve strategies of both species. Light compensation points, maximum photosynthesis, and dark respiration differed significantly among light treatments but did not differ between the species. A defoliation experiment indicated that ash is much better suited to survive defoliation, especially under low light conditions. Tropical ash had much higher non-structural carbohydrate concentrations in roots and stems than did koa. A significant carbohydrate species by light interaction suggests that koa may not produce enough carbon reserves to survive prolonged periods of stress particularly under low light conditions.

Godfrey, Laurie R., William L. Jungers, and MITCHELL T. IRWIN. Dept. of Anthropology, University of Massachusetts, Amherst, MA 01003, USA (LRG), Department of Anatomical Sciences, State University of New York at Stony Brook, Stony Brook, NY 11794, USA (WLJ), Dept. of Anthropology, State University of New York at Stony Brook, Stony Brook, NY 11794, USA (MTI).
A PALEOBIOLOGICAL PERSPECTIVE ON EXTINCT AND ENDANGERED PRIMATES IN MADAGASCAR.
The work of paleontologists can serve conservation biology in a number of ways. Paleontological data can be used to test models of extinction, including, for example, those derived from species-area curves. They can be used to explore temporal changes in community "ecospace" (or the niche characteristics of species belonging to communities), and to elucidate the differential vulnerability to extinction of distinct guilds or ecological components of communities. They can be brought to bear on the history of still-extant species. Documenting recent cases of geographic range contraction can help us to understand aspects of the vulnerability of living species that may not be evident from their current population densities and distributions alone. The Late Pleistocene and Holocene records of the primate fauna of Madagascar reveal dramatic changes in the niche characteristics of primate communities and in the geographic ranges of still-extant species. Folivores and seed predators suffered the greatest losses. Some of the still-extant taxa whose geographic ranges were much greater in the recent past than today are Hapalemur simus and Indri indri. Almost all of the extinct lemurs were (and all of the extant lemurs are) forest-dependent and their extinction/endangerment pattern conforms reasonably well to expectations derived from estimates of forest-habitat loss.

GOLDSTEIN, MICHAEL I., R. Neal Wilkins, and Thomas E. Lacher, Jr. Mendocino Redwood Company, P.O. Box 489, Fort Bragg, CA 95437, USA (MIG), Department of Wildlife and Fisheries Sciences, Texas A&M University - 2258 TAMU, College Station, TX 77843-2258, USA (RNW, TEL).
RIPARIAN BUFFERS AND HERPETOFAUNA IN BOTTOMLAND HARDWOOD FORESTS.
We analysed the influence of timber harvest treatment on the spatial and temporal variability of amphibians and reptiles in an East Texas bottomland hardwood forest. The dataset represented a time-series of 5 years post-treatment. A total of 18,645 amphibians and reptiles were captured across 144 pitfall arrays (16 arrays in each of 9 plots; 3 plots in clearcut, select cut, and untreated controls; each plot bisected by one of 3 streams). Pitfall captures represented 46 species (16 amphibians and 20 reptiles). Broad-scale analyses revealed that reptile species richness increased in response to clearcut treatments, while amphibian richness did not seem to respond. The time-series data indicated that site use by species (and species groups) fluctuated, and that these fluctuations were independent of treatment effects. Small-scale (within patch) dynamics revealed that both species richness and relative abundance of common species were confined to specific array locations in clearcut treatments. The spatio-temporally "stationary" refugia in clearcuts were principally contained within the riparian management zone. Richness and abundance hotspots in control and select cut treatments, however, were mobile in space and time.

GOLTZ, DAN, Christopher Murray, Alison Agness, and Paul C. Banko. Pacific Islands Ecosystem Research Center, U.S. Geological Survey-Biological Resources Division, Kilauea Field Station, PO Box 44, Hawaii National Park, HI 96718, USA.
FERAL CAT HOME RANGE, HABITAT UTILIZATION AND MOVEMENTS ON MAUNA KEA, HAWAII.
Feral cats (Felis catus) are predators of palila (Loxioides bailleui) and other native forest birds on Mauna Kea. Since 1998, 8-11% of monitored palila nests have been depredated annually by cats. Predation inhibits efforts to restore palila, an endangered Hawaiian honeycreeper. Little is known of the movements, ranges, and habits of feral cats in high elevation dry forests. We captured and attached radio collars to 5 male and 3 female cats and tracked them for 18 months. Male cats occupied much larger ranges than female cats, but did not defend the entire range against other cats. Home range size (10 - 95 km² ) and overlap (37 - 96%) was highly variable among males. Female cats occupied spatially discrete ranges. Three of the male cats remained on the western slope while two roamed extensively, up to 25 km between sites. This suggests that local control efforts will be complicated by immigration, and that landscape level control will be very difficult.

GOMPPER, MATTHEW E., Center for Environmental Research and Conservation, Columbia University, 1200 Amsterdam Ave, New York, NY 10027, USA.
PREDICTING EFFECTS OF TOP-CARNIVORE COLONIZATION ON MESOCARNIVORES USING ELASTICITY ANALYSES.
Mesopredator release is a phenomenon whereby the absence of a top-predator removes competitive constraints on smaller carnivores, allowing increased population densities of the latter. This in turn may result in dramatic effects on the faunal and floral community. In several regions of North and Central America, terrestrial top-predators are now recolonizing habitat (wolves, grizzly bears, mountain lions) or dramatically expanding their historical range (coyotes). What affects will this have on the populations of mid-sized predators that have been implicated in mesopredator release studies (raccoons, foxes, oppossum, skunks, coatis)? Based on information on the life cycle of mid-sized carnivore populations, I use population projection matrices to assess the growth rates and then use elasticity analyses to identify the relative importance of vital rates of various stages. I then examine the behavior of top-carnivores to show the likelihood of theses species having a significant impact on the growth rates of mesocarnivore populations. While adult survival is typically the "most important" vital rate, this is not always the case, suggesting that the entry of a top-predator into a region may have varied indirect effects on the community depending on the structure of the mesopredator guild.

GON, SAMUEL, III, Coleen Cory, Tina Lau, Anders Lyons, Edwin Misaki, Alenka Remec, Pauline Sato, Mark White. The Nature Conservancy of Hawaii, 923 Nuuanu Avenue, Honolulu, HI 96817, USA (SG, CC, AR), The Nature Conservancy of Hawaii, P.O. Box 220, Kualapuu, Molokai, HI 96757, USA (TL, EM), The Nature Conservancy of Hawaii, P.O. Box 1716, Makawao, Maui, HI 96768, USA (AL, MW), The Nature Conservancy of Hawaii, P.O. Box 971665, Waipahu, Oahu, HI 96797, USA (PS).
WATERSHED PARTNERSHIPS: A MODEL FOR LANDSCAPE SCALE CONSERVATION IN HAWAII.
Effective conservation across large landscapes often requires involvement and cooperation of adjacent landowners whose lands comprise the target landscape. Public-private conservation partnerships, based on mutual interests to preserve watersheds, have become an increasingly popular way to facilitate this kind of cooperation in Hawaii. The watershed partnerships may involve federal, state, county, and private landowners and managers. In November 1991, major landowners and managers of lands on the windward slope of East Maui formed the East Maui Watershed Partnership. Since then, several other cooperative watershed partnerships have been formed in the Hawaiian Islands, including: West Maui Mountains Watershed Partnership (1998), Koolau Mountains Watershed Partnership (1999), and East Molokai Watershed Partnership (1999). Formation of another forest-watershed partnership is underway on the island of Lanai. Although the partners have different mandates, priorities, and constituents, all share a common commitment to long-term protection of watershed resources. It is fortuitous that even when a stakeholder's commitment and efforts are based on protection of water, mitigating threats to watershed landscapes is largely congruent with biodiversity conservation goals. This successful model of watershed partnerships has led to other cooperative management efforts in the state based on mutually agreed upon conservation goals.

GONZALES, EMILY K., Rob De Loe, and Thomas D. Nudds. 642 East 2nd Street, North Vancouver, BC, V7L 1E3 (EKG); Department of Geography, University of Guelph, Guelph, Ontario, N1G 2W1 (RDL); Department of Zoology, University of Guelph, Guelph, Ontario, N1G 2W1 (TDN).
USING GIS TO PREDICT THE SPREAD OF AN INTRODUCED SPECIES.
Dispersal has traditionally been modeled and analyzed in ecology by assuming continuous, random spread over homogenous landscapes. Squirrel dispersal on Vancouver Island in Canada has been stratified and non-random over a heterogeneous landscape. I estimated squirrel habitat preference and rates of spread, and then classified the landscape into friction values that reflected presumed ease of movement. Using weighted surface analysis in GIS, I was able to identify the most likely patterns of future squirrel dispersal. This technique could be applied to other systems.

GOOD, THOMAS P., Amy Gilbert, and Mary Ruckelshaus. Conservation Biology, National Marine Fisheries Service, Northwest Fisheries Science Center, 2725 Montlake Boulevard East, Seattle, WA 98112, USA.
MATCHING THE MATRIX AND METRICS IN THE MAELSTROM.
Understanding the link between habitat characteristics and abundance, density, and productivity of threatened or endangered salmonids is a critical step in recovery planning. To evaluate whether proposed land use may impact fish populations, the National Marine Fisheries Service assesses baseline habitat conditions using a matrix of environmental pathways and indicators. However, the empirical link between this proxy for salmonid population health and actual fish population metrics (density, abundance, productivity, life stage) is not well described. We examined the relationship between fish metrics and categorized habitat conditions from the matrix using data from areas having assessments of fish metrics and baseline habitat conditions. The strength of the relationship of fish abundance/productivity to the categorized habitat characteristics varied with the number and identity of habitat variables exceeding requirements for long-term salmon survival as well as with the spatial scale of data collection. Habitat quality or quantity may thus poorly predict fish quality or quantity due to habitat inaccessibility, incomplete or incorrect habitat or fish information, or other factors driving fish responses. Using more direct measures of fish response to habitat characteristics will better predict how salmonid populations should respond to changes in habitat conditions and should enhance recovery planning efforts in general.

GOOSEM, MIRIAM W. Cooperative Research Centre for Tropical Rainforest Ecology and Management, School of Tropical Environment Studies and Geography, James Cook University, P.O. Box 6811, Cairns, Qld 4870, Australia.
EFFECTS OF ROADS AND POWERLINE CORRIDORS ON SMALL MAMMALS IN AUSTRALIAN WET TROPICAL RAINFOREST.
Roads and powerline corridors restrict small mammal movements in many habitats but little was known of this internal fragmentation' in structurally diverse rainforest habitats. Mark and recapture studies in the World Heritage Wet Tropics rainforest of north Queensland, Australia have demonstrated that several rainforest species are inhibited from crossing such linear barriers'. In a series of trapping experiments it was shown that degree of movement inhibition was influenced by several factors. These included the type of habitat in the clearing (grassland, woody weeds, road surface), the presence or absence of canopy above the clearing and the presence of road underpasses with rainforest habitat close to the entrances. Variations in low-level traffic intensity did not increase movement inhibition on narrow rainforest roads. Edge effects in small mammal community composition adjacent to roads and powerline clearings and intrusions by small mammals alien to rainforest habitats were also demonstrated. Several species were observed to use sub-road tunnels as a crossing route. Mitigation of linear clearing impacts on rainforest small mammals may be at least partially achieved by maintenance of canopy closure above clearings, by strengthening of rainforest connections in gullies and by provision of sub-road underpasses.

GRAHAM CATHERINE, H., and Craig Moritz. Museum of Vertebrate Zoology, 3101 Valley Life Science Building, University of California - Berkeley, Berkeley, CA 94720-3160, USA.
SPATIAL MODELING TO INVESTIGATE PATTERNS AND PROCESS IN SPECIES DISTRIBUTION FOR CONSERVATION PLANNING.
The distribution of biodiversity and the processes that influence this distribution are poorly understood. In the face of incomplete knowledge, we explore the use of predictive modeling as a surrogate of biodiversity and a way to test hypotheses about processes. We combined current environmental data (i.e. vegetation, temperature, raw radar data) and point localities to create predictive species distribution models for birds in Cameroon, West Africa. We used GIS models to create these distributions and evaluated models with independent point locality data. We used these predictive models as a surrogate of biodiversity and evaluated areas of high species diversity and endemism. To examine processes that may have generated this distribution of biodiversity, we used predictive models in combination with palaeoclimate information. For example, in Cameroon, sister species occur on either side of the "Adamaoua Cliff," a hypothesized geographical barrier. Predictive species models showed that suitable habitat for a sister species exists on either side of the hypothesized geographic barrier, which indicates that the allopatric species pattern may be a result of the historical barrier and not a change in habitat. Using these methods we generated broader patterns of complementarity, which are important for conservation planning.

GRANEK, ELISE and Brent J. Sewall. Smithsonian Institution, 1100 Jefferson Drive, Suite 3123, Washington, DC 20560-0705, USA (EG), Conservation Biology, University of Minnesota, 180 McNeal Hall, 1985 Buford Ave. St. Paul, MN 55108, USA (BS).
FRUIT BAT CONSERVATION IN THE COMOROS ISLANDS: PAST EFFORTS, CURRENT CHALLENGES, AND FUTURE STEPS.
Six bat species, including three Megachiropterans, inhabit the Comoros Islands of the West Indian Ocean. Here we examine the conservation of Megachiropterans in the Comoros. Pteropus seychellensis comorensis, and the endemics Pteropus livingstonii, and Rousettus obliviosus all utilize Comoros' rapidly disappearing forests for feeding and roosting habitat. Despite challenges including limited resources, poor infrastructure, and infrequent government enforcement of environmental regulations, various organizations working in the Comoros have undertaken conservation projects targeting bats. Action Comores focuses on P. livingstonii conservation through ecological research and roost monitoring. Several organizations developed a Species Action Plan for P. livingstonii. The World Conservation Union (IUCN) and the Comorian Environment Ministry are executing a five-year biodiversity conservation project that includes outreach, training, development of a Conservation Action Plan for P. livingstonii, and research on R. obliviosus. To gain local support, any long-term conservation strategy must promote environmental education and conservation-related income generation for rural Comorians. Effective bat conservation must include feeding habitat and roost site protection, long-term monitoring, further research on feeding ecology, population dynamics and reproductive biology of the Megachiropterans, and re-evaluation of the conservation status of R. obliviosus. Conservation of the Comorian rainforests may depend on the protection of these keystone bat species.

GRAY, ELIZABETH M. The Nature Conservancy, 201 Mission Street, 4th Floor, San Francisco, CA 94105, USA.
THE UTILITY OF CONSERVATION EASEMENTS IN PROTECTING LANDSCAPE SCALE SITES.
For many land conservation organizations, conservation easements have become an important tool for protecting large, landscape-scale sites. Conservation easements, legal documents which limit how a property may be used and are binding in perpetuity, provide a flexible and affordable way to protect biological diversity in states such as California, where costs of fee simple acquisitions are high and threats of rapid urbanization are imminent. Since its inception in 1998, The Mount Hamilton Project of The Nature Conservancy has used conservation easements to protect over 75,000 acres of wilderness and open space in the southern region of San Francisco Bay. Use of these easements has allowed The Mount Hamilton Project to design landscape scale protection strategies using relatively few sites (n=6). Long-term protection, however, is dependent on appropriate site-specific management and monitoring plans. Because this approach involves a more hands-off approach compared with traditional ownership and on-site stewardship, long-term success depends on collaboration with federal, state and local agencies, volunteer programs and university scientists. A statewide review suggests that conservation easements are a viable tool for landscape scale conservation; however, a comprehensive idea of how well this tool works long-term requires more systematic monitoring data collected for at least another decade.

GREENE, CORREIGH M. Section in Evolution and Ecology, University of California, Davis, CA 95616, USA.
HABITAT SELECTION STABLIZES POPULATIONS SUBJECT TO AN ALLEE EFFECT.
Theoretical studies indicate that a single population under an Allee effect will decline to extinction if reduced below a particular
threshold, but the existence of multiple local populations connected by random dispersal increases stability of the global population. An additional stabilizing process ignored in this literature is the existence of habitat selection by dispersers. Using a simulation model of population change, I found that when habitat patches exhibiting Allee effects are connected by dispersing individuals, habitat selection by these dispersers increases the likelihood that patches persist at high densities, relative to results expected by random dispersal. Populations exhibiting habitat selection also stabilize more quickly than randomly dispersing populations. These effects are particularly important when Allee effects are large and more than two patches exist. Integrating habitat selection into population dynamics may help address why some studies have failed to find destabilized populations, despite well-known Allee effects in many species.

GREGOR, AARON J., Lisa B. Passerello, John W. Slotterback, Justin P. Weber, Peter T. Oboyski, and Paul C. Banko. USGS-BRD, PO Box 44, Hawaii Volcanoes National Park, HI 96718, USA.
INVASIONS OF ALIEN INSECT PREDATORS IN MID- AND UPPER-ELEVATION DRY FOREST HABITATS ON HAWAII ISLAND, HAWAII.
Alien generalist predators, such as ants and yellowjackets, prey on native Hawaiian arthropods, which can jeopardize food web interactions and ecosystem integrity. We surveyed mid-elevation (1550-2000m) and upper-elevation (2200-2800m) dry forest habitats on Mauna Kea and Mauna Loa volcanoes, Hawaii Island, Hawaii, from 1999-2001 to document the distribution and seasonal fluctuations of western yellowjacket (Vespula pensylvanica) and alien ant species. Yellowjacket activity varied seasonally in both mid- and upper-elevation sites. Although workers died-off in February and March in the upper-elevations, they were active year-round at mid-elevations, suggesting that nests persist throughout the winter. Six ant species were detected (Cardiocondyla venustula, Linepithema humile, Monomorium pharaonis, Pheidole megacephala, Tapinoma melanocephalum, and Technomyrmex albipes). L. humile and C. venestula were collected across the widest elevation range and were detected higher than any of the other species, which were all found below 2050m. L. humile was distributed in large defined areas, whereas the other species were detected in scattered, isolated pockets. Distributional patterns are likely influenced by vegetation and substrate variation, species interactions, mode of queen dispersal, and origin and time of release. Understanding pest species' distributions, seasonal patterns, and natural histories is the first step in designing a control program.

GRESSWELL, ROBERT E., Douglas S. Bateman, and Martha H. Cavit. USGS-FRESC, Corvallis, OR 97331, USA (REG, MHC), Department of Forest Sciences, Oregon State University, Corvallis, OR 97331, USA (DSB).
PATTERNS OF FISH DISTRIBUTION AT THE WATERSHED SCALE.
There is a growing consensus that knowledge of the form, function, and historical context of landscapes is essential to research and management of aquatic ecosystems. In an attempt to gain new insights into fish habitat relationships within, and among watersheds, we are censussing potamodromous populations of coastal cutthroat trout Oncorhynchus clarki clarki from 45 third-order watersheds in western Oregon that are isolated above barriers to anadromous fish. A probability-based sample yielded watersheds with a range of land-management activities and landscape-scale variation in geology, geomorphology, and climate. Within-basin variation in physiography, climate, and land management is being assessed at various levels in a spatial hierarchy (e.g., stream, segment, reach, and channel unit). Preliminary results suggest that coastal cutthroat trout are not distributed randomly among channel units, and although there is a strong preference for pools, none of the metrics used to describe pools can be used to predict distribution of cutthroat trout in a watershed. On the contrary, distribution appears to be clustered around dispersed focal points, often associated with tributary junctions. This distribution differs substantially from patterns observed in less extensive surveys, and these results have important implications for future management and restoration efforts.

Griffen, Lauren R., and J. MICHAEL REED. Department of Biology, Tufts University, Medford, Massachusetts 02155, USA.
ACID DEPOSITION, SOIL MINERAL CONTENT, AND BREEDING BIRD DECLINES.
Adequate calcium is required for eggshell formation and chick growth in birds. Studies show that birds living in habitats with low levels of soil calcium can have reduced reproductive success. Magnesium is also important for eggshell development, and like calcium is selectively ingested by reproducing female birds. We hypothesized that population trends of birds could be predicted by soil mineral content and acid deposition, which can deplete soil minerals. We evaluated bird population declines of seven forest-specialist species in the northeastern U.S.A. relative to soil calcium and magnesium content, soil pH, and acid deposition for two 20-year time spans since 1968, and one 10-year span. The amount of variability in population trends explained by soil chemistry ranged from 1.3% to 65%. Of the seven species, all exhibited at least one consistent pattern between population trend and soil chemistry, but few consistently fit all a priori predictions. The two exceptions were Swainson's Thrush from 1968-1988 and Northern Waterthrush from 1988-1998, where models explained 65% and 34% of the observed variation in population change, respectively, and relationships were in the predicted patterns. For the other species, there was no consistent evidence for soil chemistry being correlated with long-term, forest bird population changes.

GRIFFIN, ANDREA S., Christopher S. Evans and Daniel T. Blumstein. Animal Behaviour Laboratory, Macquarie University, NSW 2109 Australia (ASG, CSE, DTB), Department of Psychology, Macquarie University (ASG, CSE), Cooperative Research Center for the Conservation and Management of Marsupials (ASG, DTB), Department of Organismic Biology, Ecology and Evolution, University of California Los Angeles, USA (DTB).
LEARNING SPECIFICITY IN ANTIPREDATOR TRAINING.
Captive-bred individuals that are reintroduced to the wild appear to be vulnerable to predation. There is evidence that antipredator behavior can be enhanced by pairing predator models with aversive stimuli. However, if acquired fear is not specifically evoked by the predator for which training is undertaken, it may extinguish quickly after release. We enhanced responses of tammar wallabies (Macropus eugenii) to a realistic model fox by pairing it with an aversive stimulus (a human who simulated a capture procedure). To determine whether the animals' acquired responses were specific to the fox, we also quantified their behavior to an array of visual models, both before and after training. Following training, animals responded to the fox by briefly increasing locomotion and maintaining sustained vigilance. Although neither had been paired with the aversive stimulus, a cat also evoked sustained vigilance, whereas a goat had no effect. These results suggest that trained animals learned that a predator-like model predicted the onset of a capture procedure. Learned response to the cat likely reflects generalization based upon convergent predator morphology. Antipredator training can thus produce a relatively specific fear response to a model predator and may improve survival after release.

GRILL, ANDREA, Roberto Crnjar, Paolo Casula and Steph Menken. Institute for Biodiversity and Ecosystem Dynamics, P.O. Box 94766, NL-1090 GT Amsterdam, The Netherlands and Dipartimento di Biologia Sperimentale, Università di Cagliari, Strada Statale 554, km 4,500, I-09042 Cagliari, Italy.
APPLYING THE IUCN RED LIST CRITERIA TO ISLAND SPECIES: SARDINIAN BUTTERFLIES (EUROPE, ITALY).
International nature conservation documents often reflect the celebrity of a species rather than its actual degree of threat and preference is given to more conspicuous and better studied taxa when establishing nature protection laws. We assessed the threat status of Sardinian butterfly species using the IUCN Red List Criteria. A threat factor analysis identifies potential risks towards butterflies in Sardinia arising from increasing human activities. This assessment again underlines the incompleteness of legislative conservation documents at the European scale. The Sardinian endemics Pseudophilotes barbagiae and Lysandra coridon gennargenti were identified as VULNERABLE in our analysis but are not mentioned in any European nature conservation documents. There is also evidence from many other invertebrates that priorities assigned in European nature legislation do not correspond to real threats. The inclusion of these two butterfly species in Appendix II of the Bern Convention and Annex II of the European Habitat Directive is therefore strongly recommended as well as a general update of European nature legislation.

Groombridge, Jim J., JAMES C. BRUCH, J. Gregory Massey, Trent R. Malcolm, and Chris N. Brosius. Maui Forest Bird Recovery Project, 2465 Olinda Road, Makawao, HI 96768, USA, Department of Land & Natural Resources, Division of & Forestry & Wildlife, 2600 Piiholo Road, Makawao, HI 96768, USA (GM).
TRANSLOCATION TECHNIQUES AND STRESS LEVELS IN MAUI CREEPERS; PREPARATION FOR TRANSLOCATING THE PO'OULI.
The Po'o-uli (Melamprosops phaeosoma) is possibly the world's rarest bird. Endemic to Maui, this honeycreeper has a population of only three wild individuals, located on Haleakala. The three individual territories do not overlap, being separated by a distance of 1-2 miles. Translocation of one Po'ouli across rugged terrain into the territory of another to form a wild breeding pair has presented a challenge to conservationists. A successful translocation will require that the amount of stress during transport be kept to a minimum, in order to optimize the chance of effective behavioral interaction upon release. We address this concern by using the non-endangered Maui Creeper as a surrogate species to explore the feasibility of hiking a Po'ouli between translocation points. We use haematologic parameters and hormone levels to test the degree of stress placed upon Maui Creepers translocated using two different transit container designs, and we incorporate radio telemetry and re-sight data to monitor post-release behavior. Clinical pathology results from Maui Creepers suggest that container design can substantially reduce stress levels during transit. Our cumulative results illustrate the feasibility of a hiked Po'ouli translocation, and indicate that monitoring stress levels in addition to post-release behavior can be a valuable interpretive tool.

GROOMBRIDGE, JIM J. Maui Forest Bird Recovery Project, 2465 Olinda Road, Makawao, HI 96768, USA. Formerly Institute of Zoology, Zoological Society of London, Regents Park, London NW1 4RY, United Kingdom.
NEW PERSPECTIVES ON THE RESTORATION OF THE MAURITIUS KESTREL.
The Mauritius kestrel (Falco punctatus) was the world's rarest bird in 1974, when the population consisted of four wild birds, with only a single breeding pair. Habitat loss and insecticidal organochlorine use were major contributing factors to its demise. An intensive captive-breeding and reintroduction program from 1984-94 restored the free-living population to 101 monitored breeding pairs by 1998, and the current wild population continues to be monitored. The well-documented historical decline of the population, and the completeness of demographic records for the kestrels' restoration have recently offered two new perspectives on this species recovery. Firstly, an analysis of 25 years of banding records has produced demographic estimates of the severity of the population bottleneck since 1974, and provided an assessment of the effectiveness of future population monitoring on Mauritius. The findings indicate that the reintroduction program was highly instrumental in accelerating the species recovery, and provide new insight into future habitat carrying-capacity. Secondly, a molecular survey of genetic diversity across the bottleneck has enabled an evaluation of the population recovery in the light of genetic impoverishment. Together, these recent studies offer a valuable re-interpretation of previous ecological and demographic information on the Mauritius kestrel.

GROSS, MART. Department of Zoology, University of Toronto, 25 Harbord St., Toronto, Ontario M5S 3G5, Canada.
ALTERNATIVE LIFE HISTORIES AND CONSERVATION BIOLOGY.
Many species have alternative life histories in which individuals of the same sex develop into different phenotypes. For example, sunfish have precocious cuckolder and adult parental males while salmon have precocious jack and adult hooknose males. Here I report on the evolutionary fitnesses of the alternative male life histories and show that, at equilibrium, the precocious males have higher average fitnesses than adult males. In addition, there is a genetic linkage between precocious maturity and fitness, thus precocious males are the reservoirs of the fittest genes in the population. This knowledge is critical for the management of wild populations and for the selection of individuals in conservation breeding. A case in point is that hatchery production of salmon typically discards jack males and uses only the hooknose phenotype as breeders. This bleeds the best genes for growth (e.g., enzymes, immunocompetence, visual and escape responses) from the population and leads to deterioration. Captive breeding programs should consider alternative phenotypes for breeders to retain the genetic health of populations, and conservation biologists should include alternative phenotypes as objectives for preservation. [Collaborator: Joe Repka, Department of Mathematics, University of Toronto.]

GURUNG, BHIM, and J. L. David Smith. Conservation Biology Graduate Program, 200 Hodson Hall, University of Minnesota, 1980 Folwell Ave., St. Paul, MN 55108, USA.
A VILLAGE RANGER BASED TIGER MONITORING NETWORK THROUGHOUT THE LOWLANDS ON NEPAL.
This reseach assesses the extent to which tigers use land outside of protected areas as breeding or dispersal habitat. Our premise for working outside reserves is based on the following observations: (1) existing reserves are not large enough to maintain viable tiger populations, (2) extensive forest lands exist outside reserves, (3) these forest lands may serve as critical tiger habitat, and (4) local people are increasingly interested in forest restoration. Formerly, experienced biologists surveyed the area inadequately. To overcome this problem we established a network of 30 village rangers to map the location of livestock kills. The results show that tigers still disperse through even degraded habitat.

GUSTAFSON, ERIC J., and Larry A. Leefers. USDA Forest Service, North Central Research Station, Rhinelander, WI 54501, USA (EJG) and Department of Forestry, Michigan State University, East Lansing, MI 48824, USA (LAL).
USING STRATEGIC MODELS TO COMPARE THE LANDSCAPE PATTERN EFFECTS OF TIMBER MANAGEMENT ALTERNATIVES.
On public and industrial timberlands, optimization models such as Spectrum often guide timber management decisions. Spectrum produces harvest schedules that are weakly spatial, yet many habitat consequences of harvest have a strong spatial component (e.g., forest interior habitat). We developed an automated procedure to simulate Spectrum harvest schedules developed for a portion of the Chequamegon-Nicolet National Forest (WI) using a spatially explicit harvest simulator (HARVEST). We used HARVEST to simulate the harvest schedules produced for the Forest Plan revision alternatives required under federal law. HARVEST produced maps of forest interior and forest edge habitat expected under each alternative, and calculated the patch structure of the resulting landscapes using seral stage to define patches. We found that the alternatives varied in their effects on landscape structure, but not as much as expected. We projected the habitat impacts of each alternative related to forest interior species, edge-dependent species, and seral stage-dependent species. Our methods provide important information needed to choose the preferred alternative for the final Forest plan.

GUSTAFSON, RICHARD G., Richard D. Methot, Bruce B. McCain, Cyreis C. Schmitt, and W. Stewart Grant. NOAA Northwest Fisheries Science Center, 2725 Montlake Blvd. E., Seattle, WA 98112, USA (RGG, RDM), NOAA Northwest Fisheries Science Center, 2030 South Marine Science Drive, Newport, OR 97365, USA (BBM, CCS), and Biodiversity and Genetic Resources Research Program, ICLARM-World Fish Center, P.O. Box 500, 10670 Penang, Malaysia (WSG).
IDENTIFYING DPSs OF PUGET SOUND MARINE FISH: PACIFIC HAKE, WALLEYE POLLOCK AND PACIFIC COD.
In response to a petition to list 18 species of marine fish in Puget Sound, Washington under the U.S. Endangered Species Act, the National Marine Fisheries Service (NMFS) initiated separate status reviews and formed separate Biological Review Teams (BRTs) for three taxonomic groupings: rockfish, Pacific herring, and the gadiforms (Pacific hake, walleye pollock and Pacific cod). The gadiform BRT examined environmental, geologic, historic, biogeographic, life history and genetic information in the process of identifying distinct population segments (DPSs) that satisfy ESA and joint NMFS/U.S. Fish and Wildlife Service interagency policy definitions of "discreteness" and "significance". The BRT found data on spawning aggregations, tagging, biogeography, ecological and habitat factors, seasonal migration patterns, parasite incidence, and genetic population structure to be most informative for this process. Demographic data (growth rate, age at maturity, fecundity, etc.) and morphometrics and meristics were less informative for DPS delineation. The gadiform BRT analyzed these data (within the guidelines of the joint-agency policy) and identified the following DPSs, all of which are substantially larger than Puget Sound: the Georgia Basin Pacific hake DPS, the Lower-boreal Eastern Pacific walleye pollock DPS, and a DPS for Pacific cod that extends from Puget Sound northward to at least Dixon Entrance.

HADDAD, NICK, David R. Bowne, Alan Cunningham, Brent Danielson, Doug Levey, Sarah Sargent, and Tim Spira. Department of Zoology, Box 7617, North Carolina State University, Raleigh, NC, 27695, USA (NH), University of Virginia, Department of Environmental Sciences, Charlottesville, Va 22904-4123, USA (DRB), Department of Biological Sciences, Clemson University, Clemson, SC 29634-0326, USA (AC), Department Animal Ecology, Iowa State University, Ames, IA 50011-3221, USA (BD), Department of Zoology, University of Florida, Gainesville, FL 32611-8525, USA (DL), Department of Biology, Allegheny College, Meadville, PA 16335, USA (SS), and Department of Biological Sciences, Clemson University, Clemson, SC 29634-0326, USA (TS).
CORRIDOR USE BY DIVERSE TAXA: A COMPARATIVE SYNTHESIS FROM A LARGE-SCALE EXPERIMENT.
A number of recent studies have demonstrated that corridors influence movement rates of individual species. However, it is unclear which types of species will benefit from corridors, and which types will not. We investigated the effects of corridors on movement rates of a number of taxa, including species of butterflies, bird-dispersed plants, small mammals, and large bodied bees. Investigations of each taxa were conducted independently, but within the same experimental sites. In the experiment, large, open patches (each 1.64 ha) were surrounded by plantation pine forest. Patches varied in whether or not they were connected by an open corridor, and in their distance from each other (64 - 384 m). Habitat-restricted butterflies, bees, and plants all used corridors by moving more frequently between connected patches than between unconnected patches, or by leaving patches preferentially through corridors. A habitat generalist butterfly and two small mammals did not use corridors. Even for species that were moderately restricted in habitat use, corridor effectiveness depended on interpatch distance relative to animal movement rates. Corridors appeared to have mainly positive and some neutral effects - but no negative effects - on plants and animals in our experiment.

HAEFNER, JAMES W. Department of Biology and Ecology Center, Utah State University, Logan, UT 84322-5305, USA.
COEVOLVING METACOMMUNITY EXTINCTION PROBABILITIES: A QUANTITATIVE GENETIC MODEL.
Most predictions of extinction probabilities ignore evolution and community interactions, yet recent studies document rapid evolution in some plant and animal species. Quantitative genetic recursion (QGR) models combine population dynamics with the evolution of quantitative traits (body size). I describe a QGR model of 2 prey and 1 predator populations in which parameters are size-dependent so that populations are dynamic and body sizes (parameters) coevolve in all populations. In numerical simulations without spatial structure, the presence of predation induces complex dynamics of population numbers and evolving body sizes depending on genetic heritabilities. Evolutionary dynamics vary from stable fixed points at low heritability, to regular oscillations and chaos-like dynamics at intermediate heritabilities, and back to fixed points at extremely high heritabilities. I have added to this QGR model size-dependent dispersal among metacommunities and examined a range of parameter values in which, when evolution does not occur, one or more populations either go extinct or reach extremely low numbers. When evolution is added, the range of parameters at which extinction occurs is reduced, effectively lowering the probability of extinction. These model results suggest that nature reserve designs should incorporate community interactions and the potential for evolution of ecologically relevant traits.

HANNON, SUSAN J. Dept of Biological Sciences, University of Alberta, Edmonton, AB T6G 2E9
SOURCE/SINK DYNAMICS IN REPRODUCTIVE SUCCESS OF AMERICAN REDSTARTS IN HABITAT ISLANDS.
Populations in fragmented landscapes may operate as metapopulations or source/sink systems. Conservation of these populations requires a knowledge of patch level reproductive success, survival and recolonization rates in order to define sources and sinks and research on the patch and landscape features that characterize source and sink patches. For 5 years I have studied American redstart reproductive success in 20 forest patches surrounded by agriculture in northern Alberta. Most patches flipped between being sources or sinks annually in an unpredictable way, but some patches were consistently sources. Contrary to predictions of island biogeography theory and many empirical studies on patch size and isolation, I found that consistent sources were smaller and more isolated than other patches and tended to have lower number of redstarts and other potential cowbird hosts in them. I suggest that cowbirds and nest predators may not enter these patches as much because of lowered profitability in finding hosts/prey.

HARCOURT, ALEXANDER H., and Mark W. Schwartz. Dept. Anthropology, University of California, Davis, CA 95616, USA (HAH), Dept. Environmental Science & Policy, University of California, Davis, CA 95616, USA (MWS).
A BIOLOGY OF HOLOCENE PRIMATE EXTINCTION AND SURVIVAL ON THE SUNDA SHELF ISLANDS.
Deeper understanding of the biology of extinction risk should help us refine conservation effort. To add to that biology, we employ the well-established method of comparing the biology of taxa that disappeared from small islands (taxa at risk) with the biology of taxa that survived on the islands (persistant taxa). The taxon investigated is the mammalian order, Primates. The islands are those created at the beginning of the Holocene as sea levels rose 120 m over the Sunda Shelf (S.E. Asia). Nine primate genera inhabit these islands, from the 40 kg orangutan, Pongo, to the 120 g tarsier, Tarsius. We use four statistical comparisons, along with phylogenetic control to ensure independence of data. As expected, large body mass, low population density, and large annual home range (high resource requirements) appear as traits of risk. Unexpectedly, neither high group mass (high resource requirements), nor low dietary variety (specialisation), nor low interbirth interval (low reproductive rate) appear as risky traits. Unexpectedly also, we identify a new trait of risk for terrestrial vertebrates, low maximum latitude (lack of adaptability?). Because different sorts of taxa are differentially prone to extinction, the community structure of small islands is substantially different from that of large islands.

HARRIS, LIZ, Chase Huntley, Bill Mangle, NAUREEN RANA, Julia Wondolleck, and Steven Yaffee. The University of Michigan, School of Natural Resources and Environment, 430 East University, Ann Arbor, MI 48109, USA.
TRANSBOUNDARY RESOURCE MANAGEMENT AND THE FLATHEAD RIVER BASIN.
Historically, the successful management of natural resources across North America's borders has been complicated by a lack of coordinated conservation policies. Stakeholders in the Flathead River basin, stretching from southeastern British Columbia to northwestern Montana, are currently struggling to resolve this dilemma. Established transboundary management efforts, like the Gulf of Maine Council and the International Sonoran Desert Alliance, are tackling issues as diverse as international fishing rights and localized flood control and providing important lessons from which others can learn. Case studies of these two programs plus six others reveal a variety of successful policy frameworks and on-the-ground strategies that can inform those wishing to implement cross-border environmental management in the Flathead region and beyond. While these successful efforts share numerous characteristics, such as high degrees of stakeholder diversity and attention to ecosystem processes, they vary across their levels of institutionalization and formality. Our findings indicate that process-oriented aspects such as open channels of communication are often more critical to the effective management of shared environments than outcome-oriented legal mandates for joint protection. In general, our conclusions suggest that any thoughtful approach to transboundary resource management be rooted in certain guiding principles while simultaneously maintaining sensitivity to area-specific issues.

HARRISON, DANIEL J., David C. Payer, Jeffrey A. Hepinstall, Angela K. Fuller, and Donald D. Katnik. Department of Wildlife Ecology, University of Maine, 5755 Nutting Hall, Orono, ME 04469, USA.
LANDSCAPE THRESHOLDS AND NONLINEAR RESPONSES TO FRAGMENTATION BY AMERICAN MARTEN.
Marten (Martes americana) are regarded as a forest specialist with high susceptibility to fragmentation. Animals responding solely to direct habitat loss should exhibit linear declines as extent of unsuitable habitat increases; however, fragmentation can lead to catastrophic declines in animal populations that may be nonlinear or additive to habitat loss. Our goals were to evaluate whether marten respond to forest fragmentation and exhibit nonlinear, threshold-level declines in response to forest harvesting. Marten occupancy declined precipitously in response to an increasing percentage of unsuitable habitat (forest < 6m) in the landscape. Occupancy declined nonlinearly and approached zero near percolation (i.e., matrix fracture). We observed steep, nonlinear, threshold-level declines in occupancy by marten in response to fragmentation. Further, effects of partial harvesting in landscapes with a previous history of clearcutting were additive. Marten were virtually absent when clearcutting resulted in percolation. The decay in occupancy rate was steeper than expected based on 98 randomly simulated home ranges in areas that were not occupied by marten. Marten exhibited strong responses to forest fragmentation that must be considered across time and space. Inferred linear responses of wildlife to direct habitat loss may not be adequate to predict landscape-level responses to fragmentation.

HART, PATRICK J. Dept. of Biology, University of Hawaii at Hilo, Hilo, HI 96720, USA.
AN IDEAL FREE DISTRIBUTION BASED ON NEST SITE AVAILABILITY IN AN ENDANGERED HAWAIIAN FOREST BIRD.
The Ideal Free Distribution (IFD) can account for geographic variation in density of an organism without corresponding variation in fitness, and has the potential to be a powerful tool for biological conservation. I assessed the "habitat matching" prediction of the IFD for the maintenance of different densities of endangered Hawaii Akepa, Loxops coccineus coccineus, within and between two nearly adjacent, 100 ha study sites in a Hawaiian rain forest. For at least the past 20 years, this insectivorous, secondary cavity nesting bird has persisted at approximately 3 times greater density at one study site than the other. Diverse aspects of forest structure, cavity availability, and home range size between the two areas were compared to evaluate the role of food and nest-site limitation to the IFD of Akepa. Analysis of over 7000 trees between sites revealed there were 2.94 times more potential nest-cavities available to birds at the high density site. There was little relationship between Akepa density and parameters known to be associated with food availability, such as tree density, basal area, canopy cover, and home-range size. Nest-cavities provided a precise habitat match with bird density and may be a strong determinant of population size in certain areas.

HARVEY, ERIK, Jonathan M. Hoekstra, Raymond J. O'Connor, and William F. Fagan. Department of Biology, Arizona State University, Tempe, AZ 85287-1501, USA (EH), Department of Zoology, University of Washington, 24 Kincaid Hall, Box 351800, Seattle, WA 98195, USA (JMH), Department of Wildlife Ecology, University of Maine, 238 Nutting Hall, Orono, ME 04469-5755, USA (RJO).
THE ROLE OF RECOVERY PLAN REVISIONS IN MANAGING ENDANGERED SPECIES.
We utilized the database developed by a project funded by the Society for Conservation Biology, National Center for Ecological Analysis and Synthesis, and U. S. Fish and Wildlife Service (USFWS) to compare three types of recovery plans: (1) those never revised, (2) original versions of subsequently revised plans, and (3) revised versions. Vertebrate species with designated critical habitat were nearly four times more likely to have their recovery plans revised than other species, but recovery priorities assigned by the USFWS did not appear to influence likelihood of plan revision. Paired comparisons between revised recovery plans and original versions suggested that understanding of general biology and status had improved, and that recognition of threats had increased since the original plans were drafted. However, these improvements did not appear to produce recovery criteria or monitoring actions that were more clearly justified by biological information. We recommend that recovery plan authors strive to maximize benefits from improved information in defining management actions that are more biologically justified. We also urge that USFWS establish a consistent priority system for recovery plan revisions that affords consideration to the full diversity of listed species.

HARWOOD, JOHN, and Roseline Beudels. Gatty Marine Laboratory, University of St Andrews, Fife KY16 8LS, Scotland (JH), Institut Royal des Sciences Naturelles de Belgique, Rue Vautier 29, B-1040 Bruxelles, Belgium (RB).
SPECIAL AREAS FOR CONSERVATION OF MARINE MAMMALS IN EUROPE: PURPOSE, IDENTIFICATION AND DESIGN.
Member states of the European Union are required by law to ensure the favourable conservation status of 632 species of animal and plant (including 5 predatory marine mammals - harbor porpoise Phocoena phocoena, bottlenose dolphin Tursiops truncatus, Mediterranean monk seal Monachus monachus, habor seal Phoca vitulina, and grey seal Halichoerus grypus). The principal mechanism for achieving this involves the establishment of a network of Special Areas for Conservation (SACs) that protect critical habitat for each species. We will describe what we, and the European Commission, understand by "favourable conservation status", what constitutes critical habitat for the 5 marine mammal species, how this can be identified, the processes that are involved in establishing SACs, and how SACs can be monitored to ensure that they achieve their desired purpose.

HATFIELD, JEFF S., Peter B. Sparks, Clytie Mead-Sparks, and Jack Jeffrey. USGS Patuxent Wildlife Research Center, 11510 American Holly Drive, Laurel, MD 20708-4017, USA (JSH), P.O. Box 3313, Honokaa, HI 96727, USA (PBS, CMS), and USFWS Hakalau Forest National Wildlife Refuge, 32 Kinoole Street, Suite 101, Hilo, HI 96720, USA (JJ).
GROWTH RATES OF TREES AND SHRUBS IN TWO MONTANE FORESTS ON THE ISLAND OF HAWAII.
Understanding the dynamics of native Hawaiian forest is necessary for the preservation and restoration of plant and animal communities. Growth rates based upon diameter at breast height (DBH) of 2 canopy species (Acacia koa, Metrosideros polymorpha) and 11 subcanopy species (Cheirodendron, Coprosma, Hedyotis, Ilex, Melicope, Myoporum, Myrsine, Styphelia, Vaccinium) were measured at 6 month intervals, 1996-99, in the relatively undisturbed Kulani Forest on Mauna Loa. For comparison, growth of the canopy species was measured in a restoration area at Hakalau Forest National Wildlife Refuge on Mauna Kea, beginning in 1997. Also measured were height growth of saplings of the canopy species and trunk height of the tree fern, Cibotium glaucum. Repeated measures analysis of covariance was used to compare species, 6-month periods, and study sites, with growth depending on DBH or height by assuming a quadratic relationship. For most periods, growth of A. koa was greater than M. polymorpha, and growth at Hakalau was greater than Kulani. Growth of most species was lower during the 1997-99 El Niño and La Niña events, implying significant ramifications if these events increase in frequency with global warming. Data will be collected in 2001 and future years to determine if these patterns are consistent.

HEAN SUN, and J. L. David Smith. Forest Protection Office, Department of Forests and Wildlife, Phnom Penh, Cambodia.
A HUNTER-BASED ASSESSMENT OF THE DISTRIBUTION AND RELATIVE ABUNDANCE OF TIGERS AND THEIR PREY THROUGHOUT CAMBODIA STATUS.
Following decades of political instability, Cambodia is currently undergoing rapid development and changes in land use. Although the government made important progress protecting biodiversity on the 1990s, more accurate information is urgently needed to develop a comprehensive, landscape scale tiger conservation plan. This study surveyed 153 local hunters to determine the relative abundance of tigers and 35 other species of mammals and birds. Hunters were shown photos to insure correct identification of species. Each hunter's homerange was delineated on 1:250,000 maps. Logistic regression was used to analyze information on tiger distribution and relative abundance in relation to the relative abundance of the major prey species, vegetation type, roads and villages. Based on our analysis tigers were predicted to occur in 10 populations ranging in size 1,678-20,348 km2. A total of 68,836 km2 was considered potential breeding tiger habitat. To verify the model we have chosen the 3 best tiger areas and are conducting extensive field surveys using 30 hunters, which have been hired as permanent wildlife rangers.

HEDDLE MANDY L., and Rosemary G. Gillespie. University of California, Berkeley, Division of Insect Biology, Department of Environmental Science, Policy and Management, 201 Wellman Hall, UC Berkeley, Berkeley, CA 94720, USA.
Losing Key Components of HAWAII's Forest Ecosystems: the status of a diverse group of hawaiian moths.
Arthropods constitute over 70% of Hawaii's native terrestrial biota, however little effort has been directed toward their conservation. Lepidoptera represent at least one fifth of Hawaii's terrestrial arthropods and are a key component of Hawaii's forest ecosystems in their role as herbivores, pollinators and prey for vertebrates and parasitic hymenoptera. Recently, the U.S. Fish and Wildlife Service sponsored an evaluation of the conservation status of Scotorythra (Lepidoptera: Geometridae), a diverse group of endemic macrolepidoptera suspected to be experiencing population declines. Between 1996 and 2000 current populations of these moths were surveyed throughout the Hawaiian Islands, the genus was revised taxonomically, and ecological data were gathered for several species. The results of these surveys were compared with historical collections held at Bishop Museum and at The Natural History Museum in London. The results suggest that 5 species are extinct, while 3 species are extremely rare. Ecological observations show several species of Scotorythra to be highly specialized, feeding on a single host plant. The rarity of these species in conjunction with their ecological specialization and increasing habitat degradation renders them susceptible to extinctions. Currently, no legal protection is afforded Scotorythra. However, the presented data demonstrate that their conservation status merits re-evaluation.

HEKKALA, EVON R., George Amato, and Rob DeSalle. Center for Environmental Research and Conservation, Columbia University, NY, USA (ERH), Wildlife Conservation Society, Science Resource Center, Bronx, NY, USA (GA), Molecular Systematics Laboratory, American Museum of Natural History, 79^th Street at Central Park West, New York, NY, USA (RD).
GENETIC RELATIONSHIPS AMONG MALAGASY AND MAINLAND AFRICAN NILE CROCODILES (CROCODYLUS NILOTICUS).
We evaluated wild populations of Nile crocodile (Crocodylus niloticus) from mainland Africa and Madagascar for unique molecular sequence characters in the 12s and 16s mtDNA gene regions. Historically the Malagasy crocodile was described as a distinct species, Crocodylus madagascariensis, found only on the island of Madagascar and the Comoros (Grandidier 1872). Subsequently, this species was synonomized with mainland African C. niloticus. Malagasy crocodile populations are currently managed under CITES (Appendix II) restrictions pertaining to the global Nile crocodile population. This research examines whether the current populations of C. niloticus persisting in Madagascar are genetically distinct from mainland Africa and from one another. The population structure and validity of species designations for Nile crocodiles are addressed. The methods developed by this project will provide the basis for a much finer approach to Nile crocodile management and conservation.

HENDERSON, SCOTT, Steve Evans, Dave Faucette, Lance Koerte, Lena Schnell, Debbie Scott, Laila Tamimi, and Shayne Veriato. Environmental Office, U. S. Army Garrison, Pohakuloa Training Area, P.O. Box 4607, Hilo, HI 96720-0607, USA.
ECOSYSTEMS MANAGEMENT OF THE POHAKULOA PLAIN, ISLAND OF HAWAII.
The U.S, Army continues an ecosystem-based approach to effectively manage natural resources on the 109,000-acre Pohakuloa Training Area (PTA). PTA is located in a tropical, sub-alpine, dryland ecosystem, one of the rarest of its kind on the planet. It contains three designated bird and plant preserves. Twelve federally listed plant species occur on PTA. During routine botanical surveys over the past year, new rare plant individuals and populations have been discovered at PTA. Many of those plants have been protected from ungulate browsing with temporary fences. Rare plant locations are recorded using GPS and are stored in a GIS database to provide comprehensive information regarding federally listed plants at PTA. Rare plant species are monitored on a regular basis to assess population health. Methods to revegetate denuded areas using native plants are being developed. Issues such as water and fertilizer requirements, planting techniques, species study and weed control are being tested. PTA continues to monitor and control incipient weed populations such as Russian Thistle (Salsola kali). During the past year, introduced predators were controlled over 70 acres of native forest bird habitat. Rodents were also controlled around rare plant species where seed depredation is a threat.

HENNEMAN, M. LAWRENCE, and Jane Memmott. University of Bristol, School of Biological Sciences, Woodland Road, Bristol BS8 1UG, UK.
INFILTRATION OF A NATIVE FOOD WEB BY BIOLOGICAL CONTROL AGENTS IN THE ALAKA`I SWAMP, KAUAI.
There exists a heated debate between conservationists and biological control workers about the extent of so-called "non-target" effects of biocontrol agents, or the rate of attack of unintended host species. To examine the impact on native Hawaiian moths by alien parasitic wasps, introduced both accidentally and as biological control agents, over two years we collected and reared over 60 species of caterpillars (Lepidoptera) from the Alaka`i Swamp, a high-elevation rain forest on the island of Kauai, that is distant both elevationally and ecologically from agricultural areas where wasps were introduced. Roughly 80% of all species of native Lepidoptera were attacked by alien wasps. The minimum parasitization rate by biological control agents was 6% in 1999 and 11% in 2000. For accidentally introduced wasps it was 2.2% and 2.6% respectively. These data indicate that biocontrol agents, introduced precisely for their aggressive attack rate, are more of a threat to native moths than accidentally introduced wasps. However, all biocontrol agents reared were introduced previous to 1950. This suggests that more recent and presumably more carefully selected biocontrol agents have been unable to invade at least the most remote native habitat on Kauai.

Hess, Steven C., Paul C. Banko, and Jon G. Giffin. Fish & Wildlife Management Program, Department of Ecology, 301 D Lewis Hall, Montana State University, Bozeman, MT 59717, USA (SCH), Pacific Island Ecosystems Research Center, Biological Resources Division/USGS, Kilauea Field Station, P.O. Box 44, Bldg. 344, Hawaii National Park, HI 96718, USA (PCB), Hawaii Department of Land and Natural Resources, Division of Forestry and Wildlife, P.O. Box 4849, 19 E. Kawili St., Hilo, HI 96720, USA (JGG).
Effectiveness of Sheep Reduction Operations in the Mauna Kea Forest Reserve, HAWAII.
Wildlife managers are frequently confronted with situations where reliable population estimates are difficult and expensive to obtain, but nonetheless important for monitoring and sound decisions. In cases where reductions or eradication of alien ungulates are conducted, removal estimators may be used to assess population size. The simplest removal estimator is based on regression of removals per event on cumulative removals. The Zippin estimator is a maximum likelihood form of the regression estimator and is available in capture-recapture programs. Both of these methods, however, may seriously underestimate population size when assumptions of demographic closure are violated. We adapted a simple population model to estimate population size of feral sheep (Ovis aries) and mouflon sheep (O. musimon) in the Mauna Kea Forest Reserve on Hawaii island. There are two forms of the model that may be applied depending on the timing of reductions relative to reproduction. The model is based on projecting the number of removal occasions until the population is effectively eliminated, and employs a polynomial equation with independent estimates of population growth rates to infer the pre-reduction population level, thus explicitly accounting for this source of demographic non-closure. We demonstrate and compare these methods of estimating population size.

HICE, CHRISTINE L. Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409, USA.
MARSUPIAL COMMUNITY PATTERNS IN HUMID TROPICAL FORESTS: THE IMPACT OF HABITAT, DISTURBANCE, AND SEASON.
Marsupials constitute a large and important component of small mammal communities in humid tropical forests. However, their basic natural history and community structure are poorly understood. This paper examines basic natural history and marsupial community structure in the humid tropical forests of northeastern Peru. Comparisons of species abundance distributions indicate significant differences between disturbed and undisturbed habitats, suggesting that anthropogenic disturbance of neotropical forests can dramatically change marsupial species composition. Communities among three types of primary forest also differ in the proportion of species abundances. These changes can be explained by differences in floral and environmental characteristics, and in structural vertical complexity of the habitats. Density of marsupials varies seasonally within these habitats, with higher densities during the rainy season. However, reproductive activity is invariant across seasons. These findings have important conservation implications in the tropics, and demonstrate the need to protect mature forest if the diverse fauna present in neotropical habitats is to be preserved.

HIGGINS, KEVIN, and Michael Lynch. Ecology and Evolution, Univ. of Oregon, Eugene, OR 97403, USA (KH, ML), Dept. of Biological Sciences, Univ. of South Carolina, Columbia, SC 29208, USA (KH).
METAPOPULATION EXTINCTION CAUSED BY MUTATION ACCUMULATION.
Theory suggests that the risk of extinction by mutation accumulation can be comparable to that by environmental stochasticity for an isolated population smaller than a few thousand individuals. Here we show that metapopulation structure, habitat loss or fragmentation and environmental stochasticity can be expected to greatly accelerate the accumulation of mildly deleterious mutations, lowering the genetic effective size to such a degree that even large metapopulations may be at risk of extinction. Because of mutation accumulation, viable metapopulations may need to be far larger and better connected than would be required under just stochastic demography.

HIRSCH, REGINA M., and Stanley A. Temple. University of Wisconsin, Madison, Department of Wildlife Ecology, 1630 Linden Drive, 226 Russell Labs, Madison, WI 53706, USA (RMH, SAT), U.S. EPA, Office of Pesticide Programs, Ariel Rios Building, 1200 Pennsylvania Ave., N.W., Washington, DC 20460, USA (RMH).
BUILDING A CONSORTIUM IN THE FARMING COMMUNITY TO REDUCE EFFECTS OF PESTICIDES ON WILDLIFE.
Farmers make decisions everyday during the growing season on how to control agricultural pests. However, few consider the impacts of their decisions on wildlife. Our project sought to change this by building a consortium, which included a Wisconsin farming community and wildlife ecologists. The farming community consisted of farmers, crop consultants, agriculture extension officers, IPM specialists, and Wisconsin Department of Agriculture biologists. The objectives were to understand current practices and decision-making processes by the farming community for controlling the corn rootworm and to exchange ideas of how to control this pest with minimum impact on wildlife that utilize these agricultural areas. Initial information on current practices and pesticides were obtained through a survey. Then the consortium met for a series of discussions to exchange ideas and information regarding a design and implementation of a study to assess the impacts of the current practices on an avian and amphibian species. Once current practices were assessed the consortium was briefed on the results, and additional discussions centered on alternative pesticides, application methods, and IPM methods to reduce harmful effects on wildlife. As a result of this effort, the farming community successfully implemented the selected practices to decrease pesticide exposure to wildlife.

HOBBS, RICHARD J., and Viki Cramer. School of Environmental Science, Murdoch University, Murdoch, WA 6150, Australia.
KING CANUTE AND CONSERVATION: MANAGING VEGETATION FRAGMENTS IN SALINISING LANSCAPES.
Southwestern Australia is a recognized biodiversity hotspot with high levels of plant diversity and endemism. Extensive vegetation clearance for agriculture has resulted in a highly fragmented landscape and hydrologic changes leading to rising saline watertables. Recent research indicates that remedial actions will not be feasible everywhere and hence that much low-lying remnant vegetation is under severe threat from salinisation. What can we do about this? For high priority areas, it may be possible to implement costly preventative measures, but for many areas we need to contemplate the transition to alternative species and vegetation types. Salinisation does not occur uniformly across the landscape, and small-scale heterogeneity in topography may play an important part in maintaining patches of existing vegetation within salinising areas. We aim to provide options for maintaining existing vegetation where possible and easing the transition to alternative types elsewhere through use of more salt tolerant species. We present an interesting conundrum for conservation: faced with a massive and inevitable threat, what can we do to maintain some conservation value in an already heavily modified landscape?

HODGES, CATHLEEN NATIVIDAD. Haleakala National Park, P.O. Box 369, Makawao, HI 96768, USA.
POSITIVE RESPONSE OF AN ENDANGERED SEABIRD POPULATION TO THE REMOVAL OF EXOTIC ANIMALS FROM HALEAKALA.
Exotic animals such as feral goats, feral pigs, mongooses, feral cats and rats negatively impacted natural and cultural resources at Haleakala National Park on the island of Maui. By 1987, the "crater" portion of the Park was completely enclosed by a fence to keep feral goats and pigs from entering the crater. After completion of the fence, feral goats and pigs were eradicated from the crater. The crater has been goat- and pig-free for at least 10 years. Alien predator populations (mongooses, cats and rats) have been dramatically reduced and are controlled by trapping and toxic baiting. The world's largest known nesting colony of the Hawaiian Dark-rumped petrel (Pterodroma phaepygia sandwichensis, Hawaiian name 'Ua'u) is at Haleakala Crater. This presentation examines the positive response of this endangered seabird population to the removal of these exotic animals. This presentation also examines current challenges of keeping exotic animal populations at manageable levels and ways that these challenges are overcome.

HOEBEE, SUSAN E., and Andrew G. Young. Centre for Plant Biodiversity Research, CSIRO Plant Industry, GPO Box 1600, Canberra, ACT 2601, Australia and The Department of Forestry, Australian National University, Canberra, ACT 0200, Australia (SHE, AGY).
CONSERVING AN ENDANGERED GREVILLEA - AN EXAMINATION OF FINE-SCALE MATING PATTERNS.
Grevillea iaspicula (Proteaceae) is an endangered, self-incompatible, hermaphrodite plant that is restricted to south-eastern Australia. Initial allozyme studies suggest that the populations maintain moderately high levels of genetic diversity but that paternal diversity within seed arrays is low. This work also indicates that gene flow among the populations is limited and that the populations are substantially differentiated. Microsatellite markers have allowed us to explore more directly the fine-scale mating patterns operating in these populations. Again, high levels of genetic diversity were found (H_o = 0.40-0.79). Paternity analyses suggest that mating may not be as restricted as previously thought and that immigration could be as high as 18%. In smaller populations, plant size correlates with paternal reproductive success and, as a result, larger plants dominate within population paternity. In larger populations, more paternal plants contribute to individual seed arrays. This may be a function of stand age structure. There is some evidence of inter-specific hybridization that could threaten the integrity of the species. Given that clones have been re-introduced into some populations, these results have important implications for effective conservation of the species.

hoekstra, jonathan M., Bill Fagan, and Jeff Bradley. Department of Zoology, University of Washington, Box 351800, Seattle, WA 98195, USA (JMH), Department of Biology, Arizona State University, Tempe, AZ 85287, USA (BF), School of Forestry, University of Washington, Seattle, WA 98195, USA (JB).
Is critical habitat really critical for endangered species recovery?
The U. S. Fish and Wildlife Service has suspended all new and proposed listing actions under the Endangered Species Act (ESA) in order to dedicate staff and funding to critical habitat designations required by recent court orders and settlement agreements. This begs the question: does critical habitat really benefit endangered species? In particular, do critical habitat designations influence recovery plans and promote improvements in species' status? No! We analyzed an extensive database on the attributes and content of 181 recovery plans, and found that recovery plans for species with designated critical habitat did not contain more detailed information about habitat requirements, were not more likely to include habitat acquisition or management among recovery tasks, and were not even more likely to include habitat considerations among criteria for measuring recovery. Furthermore, species with critical habitat were not more likely to show an improving status trend. To make designations more useful, we recommend that critical habitat be defined as a set of biological and ecological standards against which the value and restoration potential of all available habitats can be measured.

HOFFMAN, NANCY J., and Keith W. Larson. Midway Atoll National Wildlife Refuge, P.O.Box 29460, Honolulu, HI 96820, USA.
SPATIAL AND TEMPORAL ANTHROPOGENIC EFFECTS ON ALBATROSS POPULATIONS AT MIDWAY ATOLL.
The impact of humans on insular avian populations within the Northwestern Hawaiian Islands has been documented since the late 1800's. Beginning in 1991, annual surveys of breeding albatross on Midway Atoll have been derived from a combination of estimates and counts. The number of Laysan Albatross nesting pairs has decreased from 429,308 (based on counts and estimates) in 1991 to 284, 604 (based on counts) in 2000. The number of nesting Black-footed Albatross has shown a downward fluctuation during the past ten years. Using these nesting pair counts to document change in albatross populations, the spatial and temporal impacts of human development and restoration efforts on Midway Atoll can be quantified.

Holland, Brenden S., and Michael G. Hadfield. Kewalo Marine Laboratory, Pacific Biomedical Research Center, University of Hawaii, 41 Ahui Street, Honolulu, HI 96816, USA.
Islands within an Island: Conservation Genetics of an Endangered Hawaiian Tree Snail.
The objective of this investigation was to provide resource managers with relevant data to guide conservation strategy for this endemic species. Mitochondrial DNA sequences were used to estimate genetic divergence within and among populations of Achatinella mustelina from throughout its range in the Wai'anae Mountains of O'ahu. Molecular genetic data were used to define evolutionarily significant units (ESUs). Hierarchical F-statistics (F_ST), effective number of migrants per generation (N_m), number of fixed differences, and uncorrected pairwise genetic distances were used and maximum parsimony and neighbor-joining trees were inferred to characterize genetic variation among 17 tree snail populations. These genetic data support a pattern of strong population level subdivision wherein maximum genetic distances are correlated with deep, arid canyons and mountain peaks, independent of geographic distance. ESU designation was based on mean intra-population genetic divergence of less than 1.0%, a value that also corresponds to the genetic variation along ridge crests, such that populations distributed along ridges appear genetically as panmictic units. Therefore each set of populations with genetic divergence values of 1% or less was combined into a single ESU, such that the 17 populations collapsed to eight ESUs for conservation efforts including in situ protection and laboratory captive propagation.

HOLMES, ELIZABETH, and Anne York. REUT Division, Northwest Fisheries Science Center, 2725 Montlake Blvd. Seattle, WA 98112, USA (EH), National Marine Mammal Laboratory, 7600 Sand Point Way N.E. F/AKC4, Seattle, WA 98115, USA (AY).
Monitoring the effect of marine protected areas on long-lived species, a case study with steller sea lions.
Slow responsiveness of population size to perturbation is a characteristic of long-lived species and presents one of the major challenges for demonstrating positive effects of management actions such as implementation of marine protected areas. In contrast, age structure can exhibit rapid shifts, and matrix models suggest that transient spikes in age-structure may be a useful tool for quickly detecting the effects of management actions. This idea was tested using Steller sea lions, a long-lived endangered otariid found throughout the north Pacific Rim. A matrix model for Steller sea lions was used to develop simple stage-ratio metrics that would be sensitive but also practical to measure in the field. Then using aerial photographs taken during population censuses between 1970-1998, these metrics were tested for their ability to detect survivorships and/or fecundity changes in the 1980s when no-entry buffer zones around sea lion rookeries were instituted and incidental takes sharply curtailed. Large transitory spikes in simple stage-ratios were detected 2-4 years after the management actions as predicted by the matrix model. This study suggests that age-structure changes can be a powerful and practical tool for monitoring the effectiveness of management actions such marine protected areas on large long-lived marine mammals.

HOOKER, SASCHA K., Hal Whitehead, and Shannon Gowans. British Antarctic Survey, Madingley Road, Cambridge, CB3 0ET, UK (SKH), Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada (HW, SG).
THE USE OF HIGHER PREDATORS AS INDICATOR SPECIES: THE GULLY, A CASE STUDY.
The Gully submarine canyon off eastern Canada has recently been designated as a marine protected area (MPA), primarily due to the relatively high diversity and abundance of its cetacean fauna. Here I discuss the use of these cetaceans as indicators aiding the identification and design of boundaries for protection in this area. The use of higher predators as indicators for marine protection is attractive due to their visibility at the water's surface and thus increased accessibility to study. After correcting for search effort, cetacean distribution within the Gully was assessed relative to several spatial and temporal parameters: depth, slope, sea surface temperature and month. Species' distributions were most strongly correlated with depth suggesting that protection should be established based on bathymetric contours. However, the use of indicator species to define protected areas should not be undertaken without reference to the whole ecosystem, as boundaries defined by top predator distribution may not incorporate the spatial needs of lower trophic levels. Analysis of energy flow through the Gully ecosystem suggests that this system receives substantial spatial subsidy at lower levels. Conservation priorities for this MPA should take account of this, primarily by providing additional spatial protection in the form of buffer zones.

HOSSAIN, AMZAD. Institute for Sustainability and Technology Policy (ISTP), Murdoch University, Western Australia 6150.
LOCAL KNOWLEDGE AND POPULATION GROWTH IN BANGLADESH: AN OVERVIEW.
To an outsider, over population of Bangladesh has caused poverty, environmental degradation and scarcity of resources (PEDASOR). The country statistics show that Bangladesh has reduced its population growth from 3.8% to 1.8% over the last three decades. The World Bank finds that the country is still in poverty, including 30 million hard-core poor and its PEDASOR is increasing. Why? Local knowledge that includes the wisdom of the Baul philosophers suggests that Bangladesh may not yet be over populated, but the present population control policy is likely to make it so. It gives a way out to fix the population problem and PEDASOR. Based on local knowledge, this paper asserts that the impact of the current population control policy coupled with acculturation is leading the country to intellectual poverty in the future. Four socio-economic studies in village Bangladesh conducted by the author since 1987 suggest that the existing population control policy and socio-economic culture needs reformation in order to reverse the current situation with PEDASOR. Accordingly, a three-pronged population control approach has been developed to help maintain a sustainable population growth amongst the various occupational groups. The approach includes self-reliant livelihood, optimisation of renewable resources and valuing traditional culture.

HOWALD, GREGG R., Bernie Tershy, Bradford Keitt, Steve Ortega, Don Croll, and Kate Faulkner. Island Conservation and Ecology Group, P.O. Box 141, Davenport, CA, 95017, USA (GH, BT, BK, DC), Channel Islands National Park, 1901 Spinnaker Drive, Ventura, CA, 93001, USA (SO, KF).
PROGRESS IN SHIP RAT ERADICATION FROM ANACAPA ISLAND, CHANNEL ISLANDS NATIONAL PARK, CALIFORNIA.
Rattus spp. have been introduced to 80% of island groups and are responsible for >40% of recent bird and reptile extinctions. The 13 island rat eradications in North America used a grid of bait stations refilled regularly with rodenticide bait for 1-2 years. However, on many islands, such as Anacapa Island (296 ha) where introduced R. rattus threaten native species including an endemic deer mouse, bait stations cannot be used due to steep topography and sensitive species. On these islands, rats can be eradicated by aerial broadcasting bait. Anacapa will be the first North American island where rats will be eradicated using an aerial broadcast. Consequently, one year was required for environmental compliance and EPA registration. In fall 2000 we did a trial aerial broadcast of 25 ppm brodifacoum spread at 15 kg/ha on a 2.5 ha test plot. All 10 radio-collared rats died and none of the 10 ear-tagged rats were recaptured post broadcast. None of the 74 ear tagged endemic deer mice were recaptured, but deer mice recolonized the test plot within 6 weeks. Eradication of rats is scheduled for Fall 2001. Planned mitigation for deer mice includes captive holding with timed release and follow up monitoring.

HOWELL, CHRISTINE A., Bette A. Loiselle, Catherine Graham, Thomas Brooks, and Paul H. Williams. Dept. of Biology, University of Missouri-Saint Louis, 8001 Natural Bridge Rd, St. Louis, Missouri 63121, USA (CAH, BAL), Museum of Vertebrate Zoology, Berkeley, CA 94970, USA (CG), Conservation International, 2501 M Street NW, Washington DC 20037, USA (TB), Natural History Museum, Cromwell Rd, London, SW7 5BD, UK (PHW).
IDENTIFYING CONSERVATION PRIORITIES: SENSITIVITIES TO MODEL SELECTION.
Conservation biologists are pressed to make recommendations about which are the most important sites to protect based on limited species distribution and abundance data. Consequently, the use of spatial models to predict potential species distributions are becoming regular tools in the conservationist's tool-box. Using eleven types of spatial models, we constructed potential distribution models for 11 species of Cotingidae (all birds of conservation conern or globally threatened) in the Atlantic forests of Brazil (one of the most endangered regions of the world). We examined the consequences of identifying priorities area (reserve network selection) based on the 11 different distribution models using complementarity rules in the WORLDMAP spatial software program. We present the data in 2 ways: near minimum set required to capture all species and near minimum set required to provide multiple representation of all taxa. From two to four areas are necessary to represent all 11 bird species in the Atlantic forests of Brazil depending on the species distribution model employed. Moreover, models differ in the number of sites available and the geographic extent from which the reserve network was selected. Thus different species distribution models provide different answers to decision-makers regarding priorities for reserve area selection.

HOYT, REGINALD A. Philadelphia Zoo, 3400 West Girard Avenue, Philadelphia, PA 19104-1196, USA.
National and Regional Instability: Impacts on Conservation in Liberia, West Africa.
Liberia is facing economic and social crises that have had and continue to have major impacts upon biodiversity. Although the civil war (1989-1997) resulted in an apparent decline in bushmeat hunting, timber harvesting served as a major source of cash for the purchase of arms. The post-war economic crisis has left the country further dependent upon its natural resources and created a "gold-rush" climate for transnational firms, with little political will to enforce existing regulations. A recent United Nations report implicates President Charles Taylor in supporting rebel activities that are destabilizing the region. He is accused of illegal diamond trading and utilizing timber exportation by a transnational firm as a means of purchasing and importing arms for delivery to rebel forces. The involvement of the international community in the regional strife and the threat of sanctions against Liberia provide new opportunities for conservation action. Sanctions or the threat of sanctions place "the eyes of the world" on Liberia. The potential economic impact of a timber embargo provides incentives for governmental and international actions that are more related to political rather than environmental agendas, but that nonetheless may benefit biodiversity conservation in Liberia.

HU, CHENG-HENG. Conservation Biology Program, Univ. of Minnesota, 180 McNeal Hall, 1985 Buford Ave., St. Paul, MN 55108, USA.
ENDANGERED SPECIES IN A BIODIVERSE WEB: CONSERVATION OF GREEN SEA TURTLE IMPACTS NATIVE EGG-EATING SNAKE.
My research focuses on the impact of conservation practices for green sea turtle (Chelonia mydas) and its native egg-predator snake (Oligodon formosanus) in Lanyu Island, Taiwan. These lycodontine snakes are specialized for eggs of squamate reptiles by using blade-like posterior maxillary teeth to open leathery eggshell, and consume yolk. Between June and September 1997, a total of 97 snakes were recorded preying on eggs in 13 nests laid by 8 female sea turtles. The snakes were satiated before early August while the egg-laying of sea turtles peaked in late August, resulting in later broods being more successful. Conservation practices for sea turtles, such as fence protection and nest translocation, were effective in preventing nest-depredation by snakes. In addition, such nest protections also led to frequent fights while snakes competed for decreasing food supplies. A total of 16 females and 7 males were recorded with broken tails, an indication of severe physical confrontation. The breakage of tail tip peaked first in early June due to competition for scarce turtle nests, and peaked again in early August due to overwhelming competitors. Their strong competition implied that this nonlethal anti-predator project might have unforeseen ecological consequence on the snakes and other squamate species.

HUBBARD, J. ANDREW, Evan Gwilliam, Nels Barrett, and William A. Patterson III. Inventory and Monitoring Program, Cape Cod National Seashore, 99 Marconi Site Rd., Wellfleet, MA 02667, USA (JAH, EG), USDA-Natural Resources Conservation Service, 344 Merrow Rd., Suite A, Tolland, CT 06084-3917, USA (NB), Holdsworth Natural Resources Center, University of Massachusetts, Box 34210, Amherst, MA 01003, USA (WAP).
VEGETATION DYNAMICS OF COASTAL HEATHLANDS AT CAPE COD NATIONAL SEASHORE: "NATURAL" VS. CULTURAL.
Coastal heathlands are diverse, globally-rare ecosystems dominated by low-growing ericaceous shrubs. Though heathlands represent only a small proportion of coastal uplands, they are of particular interest ecologically as they contain many habitat restricted plant and animal species. In contrast to the intensively managed European Colluna heathlands, Corema heathlands in North America are rapidly disappearing; since Cape Cod National Seashore (CCNS) was established (1961), heathland acreage has diminished ca. 60% with concomitant declines in associated rare native plants. Recently, the Inventory and Monitoring Program at CCNS established permanent plots on previously sampled heathland sites, yielding a 10-year perspective on vegetation dynamics. Our surveys identified naturally-occurring heathlands as well as heathlands resulting from documented severe disturbance by humans (e.g., recent fire, grazing or land clearing), with distinct differences in species composition and abundance. Naturally-occurring heathlands were arranged along stable dunes with exposure to wind and salt spray, whereas anthropogenic heathlands were typically located inland in a patchwork of Pinus rigida forest. Naturally-occurring heathlands were generally stable during the 10-year interval between surveys, while succession towards forest was well under way on most anthropogenic heathlands. Our results suggest the dichotomy between "natural" vs. "cultural" landscapes may need reevaluation for realistic management of heathlands.

HUERTA EDMUNDO, and Medellín Rodrigo A. Lab. de Eclogía y Conservación de Vertebrados, Instituto de Ecología, Universidad Nacional Autónoma de México, Apartado Postal 70 - 275, Ciudad Universitaria, UNAM 04510, México.
BAT AND BIRD SEED RAIN REGENERATION ON LACANDONIA TROPICAL RAIN FOREST, MEXICO.
The Lacandonia region in the state of Chiapas, Mexico, has a serious deforestation problem and the research about the regeneration of disturbed habitats like abandoned corn fields and cattle pastures is very important for tropical rain forest conservation. It is well know the role of bats and birds seed rain in regeneration of disturbed habitats. We made an analysis of the bat and bird seed rain with seed traps placed to different distance from a forest edge, for one year we measure the seed rain abundance and diversity, and temporal patterns. We found that the distance from the forest edge has a differential effect with the habitat and disperser. We did not find differences between the number of seed dispersed by bats and birds, but we found differences in habitat and disperser on the seed diversity through Shannon Wiener Index and Greig-Smith Similarity Index. We found some differences between temporal patterns in habitat and disperser kind. Finally we found 169 species dispersed by bats, birds and wind. Bats and Birds have an important role on natural regeneration disturbed habitats of the neotropical rain forest, dispersing important pioneer species for the vegetal succession process.

HUGHES, R. FLINT, Julie Denslow, Susan Cordell, and Roddy Nagata. Institute of Pacific Islands Forestry, USDA Forest Service, 23 E. Kawili Street, Hilo, HI 96720, USA.
FACILITATION OF CLIDEMIA HIRTA INVASION BY AN N-FIXING TREE (PARASERIANTHES FALCATARIA) IN LOWLAND FORESTS OF HAWAII.
Clidemia hirta is a noxious weed in Hawaii capable of rapid spread and establishment in wet forests. We investigated constraints to C. hirta invasion on relatively young, nitrogen-poor lava flows (200-750 YBP) in wet lowlands of eastern Hawaii and the ways in which the N-fixing, alien tree Paraserianthes falcataria facilitates invasion of C. hirta. Where the native tree Metrosideros polymorpha and fern Dicranopteris linearis dominate these systems, C. hirta is infrequent. During the past 50 years, however, P. falcataria has spread throughout this area. Where this large tree occurs, the understory is often dominated by dense stands of C. hirta. Greenhouse experiments demonstrate that C. hirta growth is limited on young volcanic soils. Net CO₂ assimilation rates were substantially higher in fertilized versus unfertilized seedlings, as were mean relative growth rates (10.5 vs. 3.1 mm/mm/day). Leaf areas of fertilized seedlings were nearly three times higher than those of unfertilized seedlings. In addition, foliar N concentrations of C. hirta growing under P. falcataria were 36% higher than those growing under Metrosideros canopies. Results suggest that by increasing N inputs into this previously N limited system, P. falcataria may facilitate the spread of C. hirta and perhaps other invasive species as well.

HUNTER, CHRISTINE, David Fletcher, and Hal Caswell. Department of Zoology, University of Otago, PO Box 56, Dunedin, New Zealand (CH), Department of Mathematics and Statistics, University of Otago, PO Box 56, Dunedin, New Zealand (DF), Biology Department MS-34, Woods Hole Oceanographic Institution, Woods Hole MA 02543, USA (HC).
ELASTICITY ANALYSIS OF SEABIRD POPULATIONS: EFFECTS OF SIMPLIFICATION OF MODEL STRUCTURE.
Procelariiform seabirds are threatened by habitat destruction, introduced species, and incidental mortality in long-line fisheries. Elasticities of population growth rate can be used to compare management tactics for these populations, but must often be calculated from simplified models based on fragmentary data. Here, we report the effects on elasticity of simplifying stage-structured matrix models for the Short-tailed Shearwater (Puffinus tenuirostris) and the Wandering Albatross (Diomedea exulans). Beginning with detailed models, we progressively reduced 1) the number of pre-breeding age classes; 2) the number of breeding age classes; or 3) both. We computed population growth rate its sensitivity and elasticity from the resulting matrices. Simplification changed population growth rates only slightly; simplifying pre-breeder structure tended to increase growth rate. The results of the perturbation analyses were strikingly similar among all models; simplification of model structure did not alter rankings of the sensitivities or elasticities of population growth rate to transition elements or lower-level parameters. Our conclusion---that population growth rate and perturbation analyses of deterministic stage-structured models are robust to simplification of the number of model stages---probably also applies to other long-lived species with similar life histories (high adult survival and low fecundity).

HUNTER, MALCOLM L, JR., Jack W. Witham, and Alan S.White. Department of Wildlife Ecology, University of Maine, Orono, Maine 04469-5755, USA (MLH), Holt Research Forest, Box 309, Arrowsic, Maine 04530, USA (JWW), and Dept. of Forest Ecosystem Science, University of Maine, Orono, Maine 04469-5755, USA (ASW).
DESIGNING COARSE-FILTER METHODOLOGIES: VEGETATION VS SOILS AS PREDICTORS OF PLANT SPECIES COMPOSITION.
Maintaining biodiversity by protecting a representative array of ecosystems (often called a coarse-filter approach) requires a system of classifying ecosystems or communities that reflects species distributions. However, community composition is not stable through time because species respond to climate change individualistically. Because of this it has been suggested that the coarse-filter approach should be designed around physical environments--the long-term "arenas" of biodiversity--rather than ephemeral communities. To explore this idea further we sought to ascertain whether soils or overstory vegetation would be a better predictor of understory species composition in an oak-pine forest in Maine with 198 vascular species. We used simple classifications of soil drainage (poor, well, excessively) and overstory vegetation cover (mixed, coniferous, deciduous) to characterize 243 0.625 ha plots with known understory floras. Canonical correspondence analysis indicated that the soil drainage classification was a better predictor of understory composition as it explained 23.1% of the total variation while overstory vegetation explained only 15.4%.

HUSAK, JERRY F., and Emily N. Ackland. Department of Zoology, Oklahoma State University, Stillwater, OK 74078, USA.
BEHAVIOR PATTERNS OF THE RETICULATE COLLARED LIZARD (Crotaphytus reticulatus) IN SOUTH TEXAS.
The reticulate collared lizard (Crotaphytus reticulatus) is a Texas threatened species that is geographically restricted to the extremely disturbed thornscrub ecosystem of southern Texas and northeastern Mexico. Very little information exists about the basic biology of this species. Many sources anecdotally state that C. reticulatus is less dependent on rocky habitat than other species in the genus Crotaphytus, instead utilizing more brushland areas characterized by gravelly soils. The purpose of this study was to determine habitat use and behavior patterns for the species. Adult and juvenile lizards were observed in Starr County, Texas. Transects were walked during the lizards' daily activity period and microhabitat and macrohabitat use was recorded. Twenty-minute focal observations were also conducted throughout the day, recording all behavior patterns. Lizards inhabited rocky areas similar to congeners, but also utilized other microhabitats as well. They spent the majority of their active time basking for thermoregulatory purposes. A large portion of south Texas is utilized for cattle ranching, and the rocky habitat that these lizards seem to prefer is useless for grazing and often destroyed to make more suitable land for cattle. This has great conservation implications for any future management plans of this species.

HYRENBACH, K. DAVID, Karin A. Forney and Paul K. Dayton. PRBO - Marine Science Division, 4990 Shoreline Hwy, Stinson Beach, CA 94970, USA (KDH), Scripps Institution of Oceanography, La Jolla, CA 92093, USA (PKD), Southwest Fisheries Science Center, NMFS, 110 Shaffer Road, Santa Cruz, CA 95060, USA (KAF).
PELAGIC MPA DESIGN CRITERIA: OCEANOGRAPHIC AND NATURAL HISTORY CONSIDERATIONS.
Little consideration has been given to the design and implementation of open ocean reserves due to the dynamic nature of pelagic systems. Nevertheless, because many pelagic species use predictable habitats to breed and forage, Marine Protected Areas (MPAs) could be designed to protect foraging aggregations and breeding grounds. We offer a review of the major issues involved in pelagic reserve design in hopes of stimulating future discussions, research, and management actions. Reserve designs must be guided by an understanding of natural history and habitat variability. In particular, effective MPA designs must take into account the differences in scale and predictability that set aside highly dynamic pelagic systems from terrestrial and nearshore ecosystems. While traditional reserve designs may suffice to protect sessile species associated with bathymetric features, many pelagic habitats are neither fixed nor predictable. Therefore, pelagic MPAs will require new designs including dynamic boundaries defined by the extent of large-scale oceanographic features. Moreover, an understanding of the physical mechanisms that influence the formation and persistence of marine habitats will be essential to implement protected areas in the pelagic environment. In this review, we consider the following oceanographic habitats for MPA implementation: (1) static, (2) persistent, and (3) ephemeral hydrographic features.

IBARRA, MA. E., G. Espinosa , R. Díaz, J. Angulo, A. Nodarse, J. Azanza, and F. Hernández. Colectivo de Estudiantes de la Facultad de Biología, J. Pacheco, F. Moncada. Centro de Investigaciones Marinas (MEI, GE, RD, JA, AN, JA) Facultad de Ciencias de Biología (FH,CEFC), Univ. de La Habana, Calle 16 #14, e/ 1m y 3m, Playa, Cd. Habana; Delgación Provincial CITMA Pinar del Río (JP), Centro de Investigaciones Pesqueras, Ministerio de la Industria Pesquera (FM).
UNIVERSITY PROJECTS FOR RESEARCH AND CONSERVATION OF MARINE TURTLES IN CUBA.
Since 1997 the Center for Marine Research and the Faculty of Biology at the University of Havana have been working on a research project aimed to study and protect marine turtles in Cuba. The selected study area was the "Peninsula de Guanahacabibes" Biosphere Reserve located in the province of Pinar del Rio. Collaboration has been received from several provincial institutions, such as the Comisión para el Desarrollo Integral de Guanahacabibes, CITMA, Poder Popular Provincial, Municipio Sandino and the Ministry of Education. Up to date three working periods, from May to September, have been completed. Field trips last 15 days and students gather scientific data, such as morphometric measurements, number of nests and eggs from nesting turtles. Environmental parameters and genetic diversity have also been studied. This research has been carried out in six beaches: Antonio, el Resguardo, Perjuicio, La Barca, El Holandés and Caleta de los Piojos. Last year work on environmental education was begun in elementary schools in the study area. This project has not only increased our knowledge about turtle nesting populations, but at the same time, has elevated conscience for the conservation of this species in schools and local communities.

IRWIN, MITCHELL T., Steig E. Johnson, and Patricia C. Wright. Interdepartmental Doctoral Program in Anthropological Sciences, Dept. of Anthropology, SUNY at Stony Brook, Stony Brook, NY 11794-4364, USA (MTI), Dept. of Anthropology, University of Texas at Austin, Austin TX 78712, USA (SEJ), and Dept. of Anthropology, SUNY at Stony Brook, Stony Brook, NY 11794-4364, USA (PCW).
LEMUR CONSERVATION IN SOUTHEASTERN MADAGASCAR: POPULATION ASSESSMENTS USING SURVEYS, SATELLITE IMAGERY AND G.I.S.
Successful management of animal populations is often hindered by incomplete knowledge of such basic statistics as distribution, density, and population size. The lemurs of Madagascar, despite the international attention they have received, remain poorly known outside protected areas. We use a synthetic approach combining satellite imagery and ground surveys to rank the lemurs of southeastern Madagascar based on their population size and extinction risk. First, we present new data on geographic range limits and population densities for fourteen lemur taxa derived from line-transect censuses at 15 localities. Second, satellite images are used to assess existing (1999-2000) forest cover and identify areas with high recent deforestation rates. Using GIS to overlay range limits on forest cover, a maximum extent of occurrence is calculated for each taxon. Finally, data from surveys and satellite images are combined to yield population size estimates for each taxon using two scenarios: (1) all available forest area is occupied, and (2) only protected areas are occupied. Our results suggest that four taxa exist at extremely low population sizes: Hapalemur aureus, H. simus, and Varecia variegata variegata due to patchy distribution, and Eulemur fulvus albocollaris due to an extremely restricted range.

ISMAIL, MOUTASIM GAMAR ELDIN, Mohamed D., Mohamed Z. H., and Mohamed Z. B. Department of System Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia, P.O.Box 328; UPM, Faculty of Forestry (MZH).
PREDICTION OF THE EFFECT OF SELECTIVE LOGGING ON THE SAMBAR DEER USING SENSITIVITY ANALYSIS
The extent to which the Sambar deer (Cervus unicolor) population could be maintained in selectively logged forest is of primary importance. A study was conducted at Sungai Lalang Forest Reserve, Malaysia from October 1998 to May 1999, to determine the relative abundance of sambar deer in primary and logged forest and its relation to microclimate. Three compartments (C) with different logging histories (C33 and C18) and a primary forest (C24) were selected within the forest reserve. Track Identification Method was used to survey the deer along four transects at each compartment. Statistical analysis indicated that the relative abundance (146.53 km^-2 ) of sambar deer was significantly higher at the primary forest than that at the logged compartments (38.5 km^-2 and 29.24 km^-2 for C33 and C18, respectively). Sensitivity analysis showed that the number of footprint was inversely proportional to relative humidity and directly proportional to light intensity. The study concludes that keeping big patches of primary forest within the logged forest may contribute to maintain the relative abundance of sambar deer which appeared to be sensitive to light intensity and relative humidity.

IWAI, YUKINO. Asian and African Area Studies, Kyoto University, 606-8501, Japan.
LOCAL HUNTING AND "SCIENTIFIC" CONSERVATION IN SERENGETI: CHANGES IN WILDLIFE USE AND INTERACTION OF THE IKOMA.
The conservation policy in Tanzania has put much effort to restrict local hunting activity since colonial period until today. Some researches tried to seek the possibility to legalize local hunting through evaluating local hunting methods in terms of game selectivity and killing reliability. Other researches tried to estimate wildebeest utilization quota based on population dynamics influenced by rainfall. Tanzanian conservation policy still employs these "scientific" results rather than local reality in spite of its community-based advocacy. I present the change of hunting method and the human-wildlife interaction in the Ikoma society in Serengeti to examine the gap between "scientific" logic and local one. The Ikoma traditionally hunted wildlife for both subsistence and commercial use, and continue to do with innovating new hunting tools and methods even after hunting regulations strictly limited their activity in 1970's. Although those regulations resulted in people getting apart from wildlife, the wildebeest migration, passing through the Ikoma villages around Protected Areas every year, make villagers possible to experience strong human-wildlife interaction still now. This paper suggests that conservation policy makers should take account of local reliability and wildebeest migration ecology when they regulate hunting method and quota.

JACOBI, JAMES D. U.S. Geological Survey, Pacific Island Ecosystems Research Center, Kilauea Field Station, P.O. Box 44, Hawaii National Park, HI 96718, USA.
OVERVIEW AND CONSERVATION STATUS OF HAWAIIAN ECOSYSTEMS.
The extreme isolation of the Hawaiian archipelago has fostered the development of a highly endemic flora and fauna with ancestral connections to all of the major biomes found around the Pacific rim. For nearly 70 million years the vast expanse of the Pacific Ocean kept certain biological groups from becoming established here. Most notably absent from the native biota are terrestrial mammals, reptiles, and amphibians, as well as ants, aphids, mosquitoes, conifers, figs, and bamboos. The size and altitudinal range (over 4200 m for the highest mountains) of the islands, provide a wide variety of ecological zones, including alpine to coastal lowland habitats, across a moisture gradient that ranges from 250 to over 10,000 mm of annual rainfall. The colonization of the Hawaiian Islands, starting with the first Polynesian settlers over 1400 years ago, resulted in major changes to the native ecosystems and their associated species. Today, only about 47% of the landscape is dominated by native plant communities, and approximately 75% of the endemic terrestrial bird species and 53% of the vascular plants are extinct, endangered, or declining. Conservation efforts are focused on habitat protection, alien species prevention and control, species and habitat restoration, and public education.

JACOBSON, SUSAN K., Sampreethi Aipunjiguly, and Richard Flamm. Program for Studies in Tropical Conservation, Dept. Wildlife Ecology and Conservation, University of Florida, P.O. Box 110430, Gainesville, FL 32611-0430, USA (SKJ and SA), and Florida Marine Research Institute, Florida Fish and Wildlife Conservation Commission, 100 Eighth Ave. SE., St. Petersburg, FL 33701-5095, USA (RF).
CONSERVING MANATEES: KNOWLEDGE, ATTITUDES AND BEHAVIORAL INTENTIONS OF FLORIDA BOATERS.
Collisions with boats account for about 25% of endangered Florida manatee (Trichechus manatus latirostris) deaths each year, emphasizing the need to understand the human dimensions context of manatee conservation. We surveyed 504 Florida boaters to examine their knowledge, attitudes, and behavioral intentions regarding manatee protection and management. Boaters indicated greater support for public education and speed reduction, than for more stringent regulations setting no-entry areas or wake limits. A majority of respondents believed that about a quarter of the boaters violated speed zones and entered closed areas, and perceived that almost none harassed manatees. Greater knowledge about manatees was correlated with support for their conservation. Knowledge and support were positively related to perceptions of speeding, no-entry and harassment violations. An analysis of trends in the last ten years compared our results with previous public surveys, and indicated a potential decrease in knowledge about manatees and their conservation. Based on the survey results, public communication needs were identified, to (1) fill the gaps in public knowledge about manatees and their conservation, (2) clarify speed zone regulations and harassment behaviors (3) increase public awareness of manatee conservation efforts, and (4) use mass media more effectively to increase compliance with law enforcement.

JAMES, E. A., S. Akiyama, C. Knight. Royal Botanic Gardens Melbourne, Birdwood Avenue, South Yarra, Victoria 3141, Australia (EAJ), Melbourne Zoological Parks and Gardens, Parkville, Victoria, Australia (SA & CK).
Building a liferaft for Diuris fragrantissima (Orchidaceae) in Australia.
Diuris fragrantissima is an Australian terrestrial orchid species that is critically endangered because of changed land-use patterns. Only three plants remain in the wild. A co-ordinated conservation program is underway to re-introduce the species to its former range. Ex-situ plants derived from plants collected in the wild during the 1970s and 1980s form the basis of the re-introduction material. Propagation had been ad hoc with no records kept of individual crosses. Genetic and morphological analyses of ex-situ collections have identified a reasonable level of genetic diversity and the information for individual plants has been used to devise a breeding program to increase plant numbers while minimising inbreeding. Seedlings propagated in vitro are being acclimatised under nursery conditions and will be introduced into secure sites in its previous range starting in 2002. Reintroduction attempts from the 1980s were thought to have failed but a flowering plant was found in October 2000, 15 years after the reintroduction attempt and highlights the importance of longterm monitoring. It is hoped that the genetic diversity, given the limitations of the starting material, will be sufficient to ensure that a number of individuals are able to thrive and develop into sustainable populations in perpetuity.

JAMES, HELEN F. Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560, USA.
THE PALEOECOLOGY OF ENDANGERED HAWAIIAN BIRDS.
Many of Hawaii's native birds are threatened with extinction. Two thousand years ago, these birds shared their habitats with a surprising variety of other avian species, most of which have already become extinct. Their habitats have also been depleted by population declines and extinctions of native snails, land crabs, insects, and plants. Because the ecological deterioration began prehistorically, management decisions for Hawaii's endangered birds can benefit from insights gained through paleontological investigations. Paleodistributions provide insights into the types of habitat occupied by endangered species when they were more widespread and abundant, as well as prior range justifications for reintroductions. Stable isotopes in bone collagen and analyses of ancient feces provide insights into former diets. With ancient DNA from subfossil bones, changes in genetic diversity that may have accompanied population decline can be traced. Paleoecology is helpful when planning habitat restoration, because it can reveal many aspects of ecosystem structure before the period of ecological deterioration. Examples of these various types of paleontological insights will be given for such endangered species as the Nene (Branta sandvicensis), Laysan Duck (Anas laysanensis), `Alala (Corvus hawaiiensis), Palila (Loxioides bailleui), and Po`o-uli (Melamprosops phaeosoma).

JANSEN, SANDY. Harold L. Lyon Arboretum-University of Hawaii at Mänoa, 3860 Manoa Road, Honolulu, HI 96822, USA.
EX SITU CONSERVATION AT LYON ARBORETUM.
One of the main purposes of a Botanical Garden is to preserve biological diversity. Lyon Arboretum is helping to conserve species that are in danger of extinction by preserving and propagating germplasm of endangered plant species, especially those endemic to Hawaii. As a tropical rain forest, Lyon Arboretum offers a suitable environment for endangered species from the wet tropics, worldwide. Within this tropical rainforest, there are approximately 16,600 living accessions of plants embracing over 5,000 taxa and cultivars including endangered species represented in Amaranthaceae, Asteraceae, Camanulaceae, Fabaceae, Heliconiaceae, Malvaceae, Orchidaceae, Palmae, Sapindaceae, and Theaceae. For example, Lyon Arboretum houses one of the world's largest palm collections, some which are extremely rare in their native habitat. Lyon Arboretum also has an extensive Heliconia collection, which includes several endangered species, and has been designated a germplasm repository by the Heliconia Society International. Lyon Arboretum houses these collections on the grounds, as well as in the greenhouse and in tissue culture.

JARVI, SUSAN I., Carter.T. Atkinson, Robert C. Fleischer, and Paul C. Banko. Department of Biology, University of Hawaii at Hilo, Hilo, HI 96720, USA (SIJ), USGS Pacific Island Ecosystem Research Center, Hawaii National Park, HI 96718, USA (CTA, PCB), Molecular Genetics, National Zoological Park, Smithsonian Institution, Washington DC 20008 (RCF).
MALARIA AND HAWAIIAN HONEYCREEPERS: DIAGNOSTICS AND DIVERSITY.
Introduced disease (Plasmodium relictum) has significantly impacted the survival and distribution of many species of Hawaiian honeycreepers. To understand why and how some species and individuals within species survive epidemics, while others succumb, requires knowledge of the complex interactions between malarial parasites and the avian immune system. Diversity of the parasite is being evaluated by SSCP analyses of rRNA and TRAP genes of the parasite. Current data suggests the presence of at least three P. relictum variants in the Islands. Knowledge of parasite diversity may reduce the risk of unintentional introduction of a new variant during geographic translocations or captive releases. We are evaluating host diversity at loci important in both cellular and humoral immunity (major histocompatibility complex, Mhc) and at the genomic level by AFLP. We have found that 'Amakihi, with a mortality rate of 60%, appear more genetically diverse than I'iwi, which have a mortality rate of >90%. Additionally, accurate diagnostics are important for the detection of (especially)chronically-infected individuals since they may have some genetic resistance to the parasite. We have developed nested-PCR diagnostic tests based on rRNA and TRAP genes, and have compared the sensitivity of these tests to that of blood smears and serological methods.

JOHNS, DAVID M. Portland State University and The Wildlands Project, PO Box 725, McMinnville, OR 97128, USA.
OUR REAL CHALLENGE: MANAGING OURSELVES INSTEAD OF NATURE.
The stories we tell have a powerful influence on how we behave. Most stories that attempt to explain, criticize or justify the human destruction of the natural world have two things in common: our self-importance and a superficial and simplified image of who we are, thus obscuring the root causes of the current extinction crisis. Evolutionarily we arejust another animal, yet our impact on the world seems undeniably disproportionate. Four factors underlie this disproportionate impact: 1) We can change our behavior and social organization to accommodate larger numbers and other changed circumstances; 2) We can develop tools that enable us to capture greater amounts of energy to feed, clothe, house increased numbers and to feed desires to accumulate wealth; 3) The biopsychological attributes that make adaptation through social and technological change possible are double-edged; and 4) Feedback loops that might alert us to the consequences of our actions are often too long for us to learn quickly enough; or those making decisions are insulated from the feed back. New stories are needed and available, but making them compelling to key constituencies had not been adequately addressed. I offer an analysis of the key elements in making new stories compelling.

Johnson, A., Madhu Rao, MICHAEL HEDEMARK, and Troy Hansel. Wildlife Conservation Society, International Programs, 185th Street and Southern Boulevard, Bronx, NY 10460, USA.
CHALLENGES IN ADAPTIVE MANAGEMENT OF CONSERVATION PROGRAMS: A CASE STUDY IN LAO P.D.R.
While engaged in efforts to practice adaptive management, conservation programs in developing countries are frequently implemented under conditions of limited national planning and coordination, vague and conflicting natural resource policy and unstable political environments. We conducted a case study of a conservation program in Lao P.D.R. and address the following questions: (1) How can we improve conservation success under existing sociopolitical conditions and (2) How can program evaluations be appropriately designed to maximize their effectiveness. We used the recommendations of an external evaluation and analyzed the extent to which the program was able to implement recommendations and adapt program design. Each recommendation was assigned an indicator and the presence/absence of that indicator was noted. Where absent, assumptions underlying the recommendations were tested using the constraints found to have precluded implementation. Lessons learned include the following: (1) Recognition of the overwhelming development mandate of growing economies demands flexibility in program design to achieve conservation objectives. (2) Definitions of key terms such as wildlife conservation, rural development and technical assistance diverge between host country governments and international agencies. (3) Evaluations should take into account sociopolitical realities by identifying why problems exist in program design and determine the feasibility of overcoming those problems.

JOHNSON, ALAN R., and Sandra J. Turner. Department of Environmental Toxicology, Clemson University, P.O. Box 709, Pendleton, SC 29670-0709, USA (ARJ), The Department of Biological Sciences, St. Cloud State University, St. Cloud, MN 56301, USA (SJT).
MODELING POLLINATOR-LIMITED PLANT METAPOPULATIONS: THE CASE OF THE FRINGED ORCHIDS.
Fringed orchids, a set of morphologically similar species in the genus Platanthera, typically exist as metapopulations associated with fragmented habitat. Reproduction depends upon pollination by specific animal vectors (usually lepidopterans), and pollinator visitation is probably often the factor limiting seed production. Within the US, several species of fringed orchids are listed as threatened or endangered at either the Federal or State level. Conservation of these species depends upon an understanding of the combined effects of fragmentation and variable reduced pollinator availability on the dynamics of local populations and regional metapopulations. We present results of a simulation model developed to represent the life history and demography of a typical fringed orchid. Pollinator limitation is modeled as an Allee effect at the level of the local population. We explore metapopulation dynamics and quantify extinction risk as a function of pollinator availability and demographic variables. Sensitivity analysis provides a measure of the relative influence of the Allee and other demographic parameters on the predicted extinction risk. We propose our modeling approach as a useful framework for evaluating the potential effects of habitat fragmentation and reduced pollinator availability in plants with specialized pollination syndromes.

JOHNSON, ARLYNE, Robert Bino, and Paul Igag. Wildlife Conservation Society, Box 6712, Vientiane, Lao PDR, Research and Conservation Foundation of Papua New Guinea (RB, PI).
monitoring the sustainability of village capture and trade of dwarf cassowary in Papua New Guinea.
Wildlife capture and trade for cultural ceremony and exchange in Papua New Guinea has led to extirpation of dwarf cassowary Casuarius bennetti in some areas and increasing pressure for trade from areas where they remain. We report on a village-based monitoring program to evaluate sustainability of wildlife capture and trade by households in the Crater Mountain Wildlife Management Area. We monitor type and number of wildlife species captured and their eventual use. For dwarf cassowary, we compared estimated annual harvest to maximum sustainable harvest, and compared change in hunter effort and catch over time. We found that capture rates of 0.06-0.07 per km² were unsustainable in one catchment and on the threshold of unsustainable use in another. Hunters from a densely populated highland sites were travelling significantly farther to encounter cassowary than hunters from lowland sites. Findings suggest that extirpation of cassowary in the management area may be avoided by restricting the method or quantity of the harvest and improving husbandry of captive birds. Results indicate that household monitoring of wildlife capture and trade is a feasible protocol for villages to implement and should be encouraged as means for communities to monitor and evaluate the sustainability of cassowary harvest.

JOHNSON, CATHERINE M., Lucinda B. Johnson, Carl Richards, Joseph Murphy, and Val Beasley. Natural Resources Research Institute, University of Minnesota, 5013 Miller Trunk Hwy, Duluth, MN, 55811, USA (CMJ, LBJ); Minnesota Sea Grant College Program, University of Minnesota, 2305 East Fifth Street, Duluth, MN 55812, USA (CR) and Department of Veterinary Biosciences, University of Illinois, Urbana, IL, USA (JM, VB).
ANURANS COMMUNITIES AS INDICATORS OF STRESSORS IN AGRICULTURAL LANDSCAPES.
Much attention has been focused on potential amphibian population declines over the last decade and possible links to habitat loss and fragmentation. Relationships between anuran communities and wetland characteristics associated with particular land uses (e.g., in agricultural landscapes) and landscape metrics could lend insight into causes for or amelioration of potential stressors. We examined the relationships between amphibian communities, wetland characteristics and landscape metrics for 36 Minnesota wetlands. Anuran, vegetative and macroinvertebrate communities were surveyed at all wetlands from 1998-2000, water and soil samples also were taken for laboratory analysis. Landscape attributes were determined for concentric buffers (10 m - 100 m) from each wetland using aerial photos; broader land cover data (to 10 km) was based on satellite imagery. Anuran species richness was negatively correlated with percent of agricultural land, as was macroinvertebrate species richness. Relationships also were noted for individual anuran species and several landscape metrics. Despite negative associations between anuran species richness and wetlands in agricultural landscapes, the presence of some natural land cover, especially woodland, in the immediate vicinity of such wetlands appeared to increase their habitat value.

JOHNSON, ELIZABETH, Christina Kennedy, Elizabeth Gordon, and Jessica Wilkinson. Center for Biodiversity and Conservation, American Museum of Natural History, Central Park West at 79^th Street, New York, NY 10024, USA (EJ), Environmental Law Institute, 1616 P Street, NW Suite 200, Washington, DC 20036, USA (CK, EG, JW).
NEEDS ASSESSMENT OF BIODIVERSITY INFORMATION USERS IN NEW YORK STATE.
Many scientists optimistically think that information about biodiversity is available and accessible to local and state decision makers and readily incorporated into decision-making. Yet this is often not the case. The American Museum of Natural History's Center for Biodiversity and Conservation initiated the New York State Biodiversity Project in part to address this issue in New York. Working in partnership with the project steering committee, the Environmental Law Institute conducted a needs assessment of decision makers in New York to identify what information is being used, its perceived adequacies, whether additional information is needed and, if so, in what format. Phone interviews were conducted with representatives from federal, state, and local agencies, the business and environmental communities, planning organizations, among others. Results indicate that most decision makers rely primarily on information from the Natural Heritage Program and state agencies, and would welcome additional data, especially if site-specific, GIS-compatible and presented electronically. The assessment highlights the need for the development of a statewide biodiversity inventory and assessment, education and outreach programs, and a central clearinghouse for biodiversity information and technical assistance. Enhancing the effectiveness of New York's biodiversity laws and policies is also important to conservation in the state.

JOHNSON, JERALD B., and Mark C. Belk. Conservation Biology Division, NMFS - Northwest Fisheries Science Center, 2725 Montlake Blvd. East, Seattle, WA, 98117, USA (JBJ), Department of Zoology, Brigham Young University, Provo, UT 84602, USA (MCB).
SPECIES CONCEPTS AND RARE DESERT FISHES: DETECTING A CRYPTIC SPECIES OF LEATHERSIDE CHUB.
Delineating meaningful taxonomic units remains a vexing problem in evolutionary biology with important implications for the conservation of rare species. In this study, we examine evolutionary divergence in leatherside chub (Gila copei), a threatened fish endemic to the Bonneville Basin and upper Snake River drainages of western North America. Using morphological, reproductive, and phylogenetic species criteria, we find that leatherside chub is actually an ancient two-species cryptic complex geographically divided into a Bonneville Basin species in the south and a Snake River / Bear River species in the north. Within each species, geographic fragmentation in the late Pleistocene accounts for genetic subdivision among populations. Phylogenetic analyses (based on mtDNA) further suggest that these two leatherside chub species are paraphyletic with respect to the geographically distant endangered Colorado River spinedaces (Lepidomeda species). Populations of leatherside from both species appear to be in general decline. However, our results now render the newly defined northern clade extremely rare, composed only of four known populations. Hence, immediate conservation efforts should focus on the northern species.

JOHNSON, LUANNE, Paul C. Banko, and Richard J. Camp. Pacific Island Ecosystems Research Center, U.S. Geological Survey-Biological Resources Division, Kilauea Field Station, PO Box 44, Hawaii National Park, HI 96718, USA.
ANNUAL POPULATION ESTIMATES OF A TROPICAL BIRD SPECIES: LESSONS IN CONFOUNDING VARIABLES.
Obtaining reliable population estimates is crucial to monitoring endangered species and developing recovery strategies. The palila (Loxioides bailleui) is a finch-billed Hawaiian honeycreeper found only on Mauna Kea on the island of Hawaii. In the last 20 years over 90% of the population has become increasingly concentrated into <30 km² of sub-alpine, dry forest on the western slope. Large annual fluctuations in population estimates have been documented from surveys conducted during the non-breeding season (January-March) using the Variable Circular Plot method. While we believe that some of the variability in the estimates reflects real changes in the population, we show that much of the variation in the estimates can be linked to a breeding season that varies in length from 2-8 months. We recommend that researchers working with tropical bird species, whose breeding seasons can be highly variable, carefully select the timing of annual censuses to coincide with months that reflect the most constant range of vocalization in the target species.

JOHNSON, MATTHEW D., and Thomas W. Sherry. Department of Wildlife, Humboldt State University, Arcata, CA 95521, USA (MDJ), Department of Ecology and Evolutionary Biology, Tulane University, New Orleans, LA 70118, USA (TWS).
MEASURING HABITAT QUALITY FOR A MIGRATORY WARBLER IN NATURAL AND AGRICULTURAL AREAS IN JAMAICA.
As tropical forests are cleared, a greater proportion of migratory songbirds are forced to winter in agricultural and disturbed habitats, which, if poorer in quality than natural forests, could contribute to their population declines. We document demographic indicators of habitat quality for American Redstarts (Setophaga ruticilla) wintering in Jamaican citrus orchards and shade-coffee plantations, and we make comparisons with published data from four natural Jamaican habitats: mangrove, coastal scrub, coastal palm, and dry limestone forests. Although redstart density, survival (f, estimated by program MARK), and overwinter body mass retention varied significantly among habitat types, citrus and coffee values were relatively high, lying between the highest (mangrove) and lowest (dry limestone) values in natural habitats. These data suggest that citrus orchards and coffee plantations lie within the variation found among natural forests. Changes in body mass in citrus, coffee, and dry limestone forest corresponded with overwinter changes in arthropod biomass, suggesting that food availability may ultimately determine habitat quality. Density, especially in spring, was well correlated with other demographic variables, suggesting that measures of abundance may be useful indicators of habitat quality for wintering warblers.

JOHNSON, NATHAN. Commonwealth of the Northern Mariana Islands, Division of Fish and Wildlife, P.O. Box 10007, Lower Base, Saipan, MP 96950.
THE POPULATION STATUS OF THE MARIANA FRUIT BAT (PTEROPUS MARIANNUS) IN THE NORTHERN ISLANDS OF THE MARIANA ARCHIPELAGO.
Several of the remote northern islands in the Mariana archipelago were surveyed for the endemic Mariana fruit bat (Pteropus mariannus) between June-August 2000. Many of these islands are subjected to severe feral ungulate damage and fruit bat poaching pressure. Surveys were conducted using direct colony, evening station and evening departure counts. Fruit bat populations have declined dramatically since 1983 on the islands of Pagan and Anatahan. Evidence of fruit bat poaching was prevalent on Pagan and Anatahan. Increased feral ungulate forest degeneration was also documented. It is believed that the cumulative effects of poaching and feral ungulate destruction are largely responsible for the decline of the Mariana fruit bat population.

JOHNSON, ORLAY, Kathleen Neely, and Robin Waples. NMFS, Northwest Fisheries Science Center, Conservation Biology Divis, 2725 Montlake Blvd. East, Seattle, WA 98112, USA.
USE OF FLUCTUATING ASYMMETRY TO ESTIMATE DEVELOPMENTAL STABILITY IN SNAKE RIVER SALMONID POPULATIONS.
Fluctuating asymmetry of meristic characters was used to evaluate possible long term effects of outplanting hatchery-reared fish on natural and wild populations. Juvenile chinook salmon and steelhead were collected in 1989 and throughout the 1990s from wild, natural, and hatchery populations in Snake River basin of Idaho, Oregon, and Washington. The experimental design involves monitoring genetic and meristic characters in yearly samples. Indices of asymmetry were developed from paired meristic characters and evaluated as indicators of developmental stability, which has been shown to be correlated with levels of genetic/environmental variability. Lower levels of asymmetry were detected in some drainages than in others and levels were lower in steelhead than in chinook salmon. Also, some hatchery populations of chinook salmon had high levels of directional asymmetry but hatchery steelhead did not. Comparisons of these values in the two species reveal differences that may be linked to differences in hatchery management or life history. In 1994-1995 salmonids in the Snake River basin experienced a severe population crash and comparison of asymmetry levels in samples taken before and after the crash may provide insight into the genetic consequences of the bottleneck.

JONES, KENNETH L., Lindy Rodwell, Kevin McCann, Gerhard Verdoorn, and Mary V. Ashley. University of Illinois at Chicago, Department of Biological Sciences (M\C 066), 845 W. Taylor St., Chicago, IL 60607-7060, USA (KLJ, MVA), South African Crane Working Group, Private Bag X11, Parkview, 2122, South Africa (LR, GV), Eskom/EWT NCCP Project, P.O. Box 1047, Mooi River, 3300, South Africa (KM).
CONSERVATION GENETICS OF WATTLED CRANES.
The Wattled Crane (Bugeranus carunculatus) is the largest and rarest of the six crane species in Africa. Its once vast range now consists of three disjunct populations. Most of the remaining birds occur in south-central Africa, with one small, disjunct populations in Ehiopia and another in South Africa. The objective of this study was to characterize the genetic structure of Wattled Cranes across south central Africa and assess the genetic status of the South African population. Samples from South Africa (11), Zimbabwe (12), and Botswana (3) were analyzed for variation at ten microsatellite loci and sequences of the mitochondrial DNA (mtDNA) control region. Results indicate that the Zimbabwean and Botswanan cranes are genetically similar to each other but are distinct from the South African population. Several microsatellite alleles unique to either the South African or Zimbabwe/Botswana populations were found, and of 10 mtDNA haplotypes found, only 1 was shared. Thus gene flow between these regions may have been historically limited and the populations should be considered distinct for management purposes. Further, the South African cranes show no loss of genetic diversity and translocation of new breeders into South Africa is not warranted from a genetic standpoint.

JONES, MALCOLM T., Gerald J. Niemi, and JoAnn Hanowski. Natural Resources Research Institute, University of Minnesota-Duluth, 5013 Miller Trunk Hwy, Duluth, MN 558110, USA.
LINKING HABITAT RELATIONSHIPS TO A LANDSCAPE LEVEL FOREST SIMULATION TO PREDICT FUTURE CHANGE OF FOREST BREEDING BIRDS IN THE WESTERN GREAT LAKES.
The Minnesota Forest Bird Diversity Initiative was started in 1991 to monitor populations and assess habitat relationships of forest breeding birds in northern Minnesota and northwestern Wisconsin. Here we present the initial results of our effort to develop a stochastic simulation that links the output of a landscape-level forest growth and succession model, LANDIS to habitat relationships developed for 150 species of breeding birds. We developed a set of a priori probabilities of the occurrence of a set of abundance categories based on our forest bird monitoring data. We tested this modeling approach in a forested area of Pine County in east-central Minnesota where we have observed a total of 75 species between 1994 and 1996. Using a 300 year simulation, the predicted number of unique species was 68, and seven of these species have never been observed during our monitoring, having range limits to either the north (6 species) or to the south (1 species). Therefore, we under-predicted the number of unique species by 12. Although there are limits to all predictive models, we believe our approach provides an unique method to explore the consequences of alternative forest management strategies on future bird communities.

JONTZ, JIM, and David Johns. American Lands, 5825 N. Greeley Ave., Portland, OR 97217, USA (JJ), Portland State University and The Wildlands Project, PO Box 725, McMinnville, OR 97128, USA (DJ).
IT TAKES POWER TO INFLUENCE POWER: EFFECTIVE TARGETING IN THE POLICY PROCESS.
With the election of a President not noted for support of conservation it will be more important than ever for conservation scientists and advocates to be strategic in trying to influence elected and appointed policy makers. Too frequently conservationists focus on organizing groups that share their views most closely, regardless of whether or not these groups have much political clout. To maximize our ability to influence policy we need to focus on mobilizing groups that are not only likely to act in support of conservation, but who have the political resources at their behest that will make policy makers pay attention. Each situation is somewhat different, but all elected and appointed officials have constituencies that are of particular import to them. A successful strategy must incorporate how to identify those constituent groups relevant to decision makers and how to work in relation to them. By way of example, our paper will include working through a policy issue and will make specific recommendations for the most effective approach.

JUREK, RONALD M. Species Conservation and Recovery Program, Habitat Conservation Planning Branch, California Department of Fish and Game, 1416 Ninth Street, Sacramento, CA 95814, USA.
SOCIAL AND POLITICAL ASPECTS OF EXOTIC SPECIES REMOVAL PROGRAMS.
Programs involving removal of nonnative animals from natural habitats in California frequently are challenged by local citizens or animal rights and welfare organizations. Protectors of some targeted introduced mammals--such as red foxes, feral cats, and feral goats--often appeal to news media, government officials, and the courts to leave introduced animals undisturbed on the land, or to ensure that animals would be humanely trapped and relocated to other areas or to captivity. Delay or cessation of removal projects results in increased costs of conservation programs and continued damage to native species and ecosystems. Adding to problems land managers face are mandated or illegal rescues and relocations of exotic animals, destruction of trapping equipment, intimidation of trappers, and distortion of information about the nature of the exotic animals, the seriousness of threats, and the motives of conservation groups and agencies. Endangered species in coastal habitats survive typically in remnant areas of near-natural habitat fragments that are like habitat "islands" in "seas" of developed lands. Lessons learned in recent decades from the control of nonnative animals to protect California's coastal endangered species in such habitat fragments may be helpful to managers when planning removals of exotics from islands.

KAAUMOANA, MAKAALA, and Jan M. Surface. Hanalei Heritage River, P.O. Box 1285, Hanalei, HI 96714, USA.
HANALEI RIVER HUI: A COMMUNITY MODEL FOR ASSESSING TODAY'S AHUPUA'A.
The community of Hanalei, through a series of public meetings, established the Hanalei River Hui (Hui) to coordinate local activities undertaken through the American Heritage River Initiative. The Hui strives to malama the ahupua'a (watershed) of Hanalei, guided by the Hawaiian principals of malama'aina (sustainability and stewardship), pono (integrity and balance), laulima (cooperation) and aloha, especially as it applies to cultural equity and respect. The Hui developed a list of concerns and issues within the ahupua'a and developed a watershed action plan. Keeping Hanalei River "swimable and fishable " was prioritized, A long-term ecological monitoring program