ISSCA

International Society for the Study

and Conservation of Amphibians

(International Society of Batrachology)

SEAT

Laboratoire des Reptiles et Amphibiens, Muséum national d'Histoire naturelle, 25 rue Cuvier, 75005

Paris, France. — Tel.: (33).(0)1.40.79.34.87. — Fax: (33).(0)1.40.79.34.88. — E-mail: dubois@mnhn.fr.

BOARD

President: C. Kenneth Dopp, Jr. (Gainesville, USA).

General Secretary: Annemarie OHLER (Paris, France).

Deputy Secretarÿ: Alain PAGANO (Angers, France).

Treasurer: Stéphane GROSIEAN (Paris, France).

Deputy Treasurer: Rafael DE SÀ (Richmond, USA).

Councillors: Lauren E. BROWN (Normal, USA); Alain DUBOïs (Paris, France): JIANG Jianping (Chengdu,

ina); Esteban O. LAviLLA (San Miguel de Tucumän, Argentina); Thierry LoDé (Angers, France):

Miguel VENCES (Amsterdam, The Netherlands).

TARIFF FOR 2004

INDIVIDUALS

Regular 2004 subscription to Alytes (volume 22) + ISSCA + Circalytes 50€ ors$

one

Suslent 2004 subscription to Alytes (volume 22) alone

Back issues of Aves: any single issue

Back issues of A/ptes: any double issue

any complete volume (4 issues)

s: complete set of volumes 1 to 21 (1982-2004) 630€ or$

Rae five-year (2004-2008, volumes 22 to 26) individual subscription to Alytes +

ISSCA + Circalytes 2$€0rS

Regular five-year (2004-2008, volumes 22 to 26) individual subscription to Alyres 200 € or $

Sreciat orrer: five-year (2004-2008, volumes 22 to 26) individual subscription to Alpes + ISSCA +

Circalyres, with complete set of back issues of Alyres (1982-2004, volumes 1 to 21) 15€ 0rS

Life individual subscription to Alytes from 2004 on 1080 € orS

Life individual subscription to Alytes + ISSCA + Circalytes from 2004 on 1200 € or$

Patron individual subscription to Alytes from 2004 on 2160 € or $ or more

Patron individual subscription to Alrtes + ISSCA + Circalyres from 2004 on 2400 € or $ or more

Important notice: from 1996 on, any new life or patron individual subscriber to Alytes is offered a free

complete collection of back issues of Alytes from the first issue (February 1982) until the start of her/his

subscription.

INSTITUTIONS

2004 subscription to Ayres (volume 22) + ISSCA + Circalvtes 100€ 0rS

2004 subscription { to Alytes (volume 22) alone 2e €or$

Back issues of A/rres: any single issue ors

Back issues of A/rtes: any double issue €ors

F Alvies: any complete volume (4 issues) 80€ or $

ues of Aves: complete set volumes 1 to 21 1260€0r$

olumes 22 to 26) subscription to Alytes + I! + Circalvte:

(19 olumes 1 to 21) 14:

Circalvres 1 the internal information bulletin of 1SSCA. Back issues of this bulletin are also available:

prices can be provided upon request by our Secretariat

MODES OF PAYMENT

— In Euros, by cheques drawn on a French bank payable 10 “ISSCA”, sent to our secretariat (address

abov

- In Euros, by direct postal transfe

mode of payment, add

In US Dollar:

r Lo our postal account: “ISSCA”, Nr. 1-398-91 L, Paris: if you use this

30 € to your payment for postal charges at our end.

by cheques payable to “ISSCA”, sent to Rafael O. DE S4, Associate Professor,

epariment of Biology, University of Richmond. Richmond, VA 23173, USA (e-mail:

rdesa(@richmond.edu: fax: (804)289-8233).

Source : MNHN, Paris

AIVTES

INTERNATIONAL JOURNAL OF BATRACHOLOGY

May 2005

Volume 22, N° 3-4

Alytes, 2005, 22 (3-4): 65-71.

The United States Geological Survey’s

Amphibian Research and

Monitoring Initiative

Paul Stephen CorN*, Erin MUTHs** , Michael J. ADAMS***

& C. Kenneth Dopp, Jr.****

* US Geological Survey, Aldo Leopold Wilderness Research Institute,

Box 8089, Missoula, Montana 59807, USA

+*_US Geological Survey, Fort Collins Science Center,

2150 Centre Ave, Bldg C, Fort Collins, Colorado 80526, USA

###_US Geological Survey, Forest and Rangeland Ecosystem Science Center,

3200 SW Jefferson Way, Corvallis, Oregon 97331, USA

+##+_ US Geologica

7920 NW 71*S

1 Survey, Florida Integrated Science Center,

Gainesville, Florida 32653, USA

PE.

—

The Amphibian Research and Monitoring Initiative (ARMI) began in 8 =—=

2000 as an attempt by the United States Geological Survey to determine the 2e

status and trends of amphibians on federal lands in the United States and its 2 — ?

ob Fe res S = *

territories. ARMI research focuses on determining causes of declines, if EL —— Ÿ

observed, developing new techniques to sample populations and analyze 5 ==}

data, and disseminating information to scientists and policy makers. Moni- CE

toring is conducted at multiple scales, with an emphasis on an ability to 22 2

draw conclusions about status in well-defined study areas such as national E— À

parks and wildlife refuges. Several papers originally presented at a national A —

symposium in 2004 are published in this special issue of Alutes. 3 —

S—

INTRODUCTION

Amphibian decline achieved recognition as à global issue after the meeting of the First

World Congress of Herpetology in England in 1989. Durir

heensuing decade, considerable

progress was made in documenting the status of populations and in understanding the causes

of some of the declines. However, significant gaps in our knowledge remained, including basic

information on status and life history. Additionally, the occurrence of large numbers of

malformations in some locations in North America in the mid-1990s increased the urgency to

66 ALYTES 22 (3-4)

critically examine the status of anuran populations. To address these needs, the United States

Congress authorized and funded the Amphibian Research and Monitoring Initiative (ARMI)

beginning in October 2000. ARMI is a national program coordinated by the United States

Geological Survey (USGS) and the science and research bureau for the Department of the

Interior (DOI). The goal of ARMI is to better understand the dynamics of amphibian

populations, including causes of declines, so that DOI agencies and other land managers have

the most accurate information from which to develop effective ways to manage and conserve

amphibian populations.

A symposium presenting ARMI monitoring and research results, co-sponsored by the

American Society of Ichthyologists and Herpetologists (ASIH) and the International Society

for the Study and Conservation of Amphibians (ISSCA), was held at the 2004 joint annual

meeting of the three North American herpetological Societies (ASIH, Herpetologists’ League

and Society for the Study of Amphibians and Reptiles) in Norman, Oklahoma. This issue of

Alytes presents a sample (6 of 24 papers presented at the symposium) of this work. Prior to

introducing these papers, we briefly describe the history, objectives, and basic methods

employed by ARMI researchers.

HisToRY OF ARMI

Herpetology in the USGS came into being when the National Biological Service (NBS)

was incorporated into the USGS in 1996. The NBS was a short-lived agency, created only

three years before by combining research scientists from the various DOI agencies with

land-management responsibilities (primarily the United States Fish and Wildlife Service,

National Park Service, and Bureau of Land Management). Several scientists who were

employed by these agencies and who now are involved in ARMI have long histories of

research on amphibian ecology and conservation. For example, BURY et al. (1980) described

the status and conservation issues for a number of amphibians that were either listed as

threatened or endangered or were thought to be in need of conservation research. Other

examples of studies conducted prior to the First World Congress include BURY (1983), CORN

et al. (1989) and DobD (1991, 1992). In the early 1990, BRD herpetologists submitted several

proposals for broad national or regional surveys, but these were not funded, and there was no

coordinated effort among DOI scientists to determine the status and trends of amphibians

nationally.

In 1998, the escalating concern over the status of amphibians and the recent overy of

high incidence of developmental malformations in some populations of ranid frogs in the

upper Midwest (METEYER , 2000; SOUDER, 2000) prompted Bruce Babbitt, then Secretary of

the Interior, to request USGS to prepare a budget request for a national amphibian monitor-

ing and research program. This task was performed by a small group of scientists and

managers from USGS, the National Park Service, Bureau of Land Management, and US

Forest Service during a meeting at Point Reyes National Seashore in June 1998, and funding

for amphibian research and monitoring was included in the USGS budget beginning in Fiscal

Year 2000. Three USGS Disciplines, Biology, Water and Geography, receive funding through

ARMI.

Source : MNHN, Paris

Corx et al. 67

ARMI OBJECTIVES AND METHODS

The goals and methods of ARMI were developed in a series of meetings and workshops

by USGS scientists, including an “Amphibian Leadership Team” composed of scientists and

managers from USGS and other agencies, largely external to ARMI, which conducted a

workshop in Gainesville, Florida in February 2001. The overall goals of ARMI, derived from

these meetings (CorN et al., 2005), are to: (1) establish a network designed to monitor the

status and changes in the distribution and abundance of amphibian species and communities

in the United States; (2) identify environmental conditions known to affect amphibians and

document their differences across the Nation; (3) conduct research that identifies causes of

amphibian population change and malformations; and (4) provide information to managers,

policy makers and the general public in support of amphibian conservation.

The Leadership Team recommended that ARMI adopt a hierarchical approach to

monitoring described by the Committee on the Environment and Natural Resources (ANON-

YMOUS, 1997; BRICKER & RUGGIERO, 1998). This hierarchy can be visualized as a pyramid

(fig. 1). At the base, extensive but necessarily coarse measurements are made at many sites

across the country. At the apex of the pyramid, intensive research and population monitoring

is conducted at a relatively small number of sites throughout the country. At the middle level

of the pyramid, monitoring directed toward detecting change in occurrence and abundance

of species across the landscape is conducted at a moderate number of sites.

Ideally, the ARMI approach would provide unbiased, base of the pyramid estimates of

the status of most amphibians in most habitats across the United States. Realistically, several

constraints prevent this approach. Primary among these constraints is the mandate of USGS

to provide science support for the other DOI agencies, which for ARMI means devoting the

majority of our efforts on lands managed by DOI agencies. Other important constraints are

that there are few species distributed widely across the US, that species richness and habitat

diversity vary widely among geographic regions, and that amphibians display a variety of

reproductive modes and habitat associations. This diversity requires that a variety of sam-

pling methods, rather than a single standardized approach, be used to detect and monitor

amphibians across the country, even within regions (HEYER et al., 1994; Dopp et al., in press).

For example, the USGS coordinates the North American Amphibian Monitoring

Program, an annual volunteer survey of calling frogs in several states in the Midwestern and

Eastern United States (MossMaN et al., 1998). However, the lack of audible calls by many

species, greater aridity of the landscape, sparse road network, and unpredictability of

breeding in desert habitats prevents calling surveys from being widely applicable in most of

the western United States. Even with standardization, the use of frog call surveys has many

limitations associated with sampling representative areas and species detection.

The constraints on collecting base-level data mean that middle-level surveys are the core

of ARMI monitoring efforts and are conducted mainly on large protected areas (national

parks and wildlife refuges) managed by DOI (HALL & LANGTIMM, 2001). At the middle level

of monitoring, ARMI has taken the approach of defining à trend as the change in site

Source : MNHN, Paris

68 ALYTES 22 (3-4)

Apex sites,

population estimates,

demographic studies, detailed

environmental data, long-term research

Core or ARMI monitoring: PAO, basic environmental data,

species richness, screening for potential causes of decline, partnerships

Distribution of species, general inventories, amphibian atlas, integration

of other relevant national databases

Amphibian Environmental Stressors Protocols National Analysisand Partnerships

Monitoring Conditions andCausal Development Databases Reporting

Monitoring Research

Fig. 1.-The conceptual framework of the United States Amphibian Research and Monitoring Initiative

envisioned as a pyramid with three levels. Research and monitoringare integrated across scales, and

the pillars across the bottom indicate what is necessary to support a national assessment of amphi-

bian status. See text and HALL & LANGTIMM (2001) and Corx et al. (2005) for additional details.

occupancy by a given species, as recommended by GR (1997). For example, ARMI

researchers in the mountainous west monitor many lentic-breeding species by documenting

change in the proportion of ponds occupied. Other commonly used methods of trend analysis

are either impossible to implement on a large scale (direct population estimates) or are

unlikely to provide unbiased estimates of change (for example, counts intended to provide an

index to true abundance: ANDERSON, 2001; MACKENZIE & KENDALL, 2002; SCHMIDT, 2003).

Moreover, changes in occupancy are likely to better reflect amphibian status than changes in

abundance for many lentic-breeding species with erratic population dynamics (GREEN, 1997).

Sites are selected for sampling based on a probabilistic scheme to allow inference about

Status and trend for the defined study area. Because absence in a survey may also indicate

failure to detect a species that is actually present, multiple surveys are conducted at sites so

that detection probabilities can be calculated and occupancy adjusted to account for errors in

detection (MACKENZIE et al., 2002). The approach of monitoring changes in site occupancy

of species based on presence/non detection data allows for the estimation of several parame-

ters that can be used to study population and community dynamics, estimate extinction and

colonization probabilities, and test hypotheses concerning how environmental factors affect

population dynamics. This approach also allows for comparable data to be obtained despite

a wide variety of sampling designs. The actual occupancy estimates can only be compared

among middle-level monitoring areas to the extent that sites are defined consistently, but

Source : MNHN, Paris

CorK et al. 69

ARMI researchers can compare unbiased estimates of trends in occupancy across the

country. Whereas inference is limited to the boundaries of the middle-level monitoring areas,

the ARMI approach allows trends to be scaled up to provide regional and national summa-

ries.

Detailed population data are collected by ARMI researchers on a number of species at

relatively few locations (apex sites). Unlike middle-level sites, apex sites are not selected

randomly, but provide locations for determining demographie and life history characteristics

of key species and studying changes in these characteristics over time. Apex monitoring,

coupled with controlled manipulations, can sometimes be used for cause and effect

hypothesis-testing research.

Atall levels of the pyramid, ARMI researchers are encouraged to form partnerships with

other agencies, programs and researchers to broaden the scope of investigation beyond DOI.

One example is a national amphibian atlas, initiated by Michael Lannoo (LANNOO, 2005) and

now hosted by ARMI [http://armi.usgs.gov]. In other cases, middle-level and apex monitoring

sites have been established in partnership with other agencies and organizations.

The causes of amphibian declines are varied and can be complex, and ARMI is

contributing to understanding both direct and subtle interactions through a number of

approaches. Some research is of short-term duration to address a known or suspected

problem, but there are still significant gaps in our knowledge of what is causing the declines of

many species. À major effort of ARMI includes a multidisciplinary approach to determine

environmental factors responsible for the decline or malformation of amphibians.

ARMI sYMPOSIUM

Papers presented at the 2004 symposium and the subset printed in this issue present a

sample of work being conducted by USGS scientists and cooperators. For a more complete

list of published papers, consult the ARMI web site [http://armi.usgs.gov]. As in the sympo-

sium, the papers in this issue reflect a mixture of monitoring and research approaches.

Developing new tools for analysis and refining field methods are ongoing areas of

emphasis in ARMI. JUNG et al. compared capture-recapture and removal methods for

estimating abundance of stream salamanders in the Appa an Mountains in Virginia.

Removal methods usually resulted in higher capture probabilities for most species, but several

sampling episodes are necessary because of high variability among samples.

CorN et al. described a transect of middle-level monitoring sites in the Rocky Mountains

along the Continental Divide that includes several of the premier national parks in the United

States. Status of amphibians in Colorado at the southern end of the transect is apparently

worse than at the northern end in Montana. The southern end of the transect is al o

characterized by much higher human population and use of park lands, suggesting topics

more focused research on causes of declines.

WENTE et al. surveyed known and random localities for two anurans in the Great Basin

in Oregon. Both species were absent from a substantial number of locations where they had

Source : MNHN, Paris

70 ALYTES 22 (3-4)

been recorded previously, and present at few new sites. Despite caveats about the effects of

prolonged drought in the region, they concluded that at least western toads had likely

undergone a recent decline.

In a study related to CorN et al, GREEN & MuTHS surveyed the health status of

amphibians in Colorado in and around Rocky Mountain National Park. They found signifi-

cant levels of infection by chytrid fungus, suggesting the possibility of further declines in this

region.

BRiDGES & LITTLE extracted naturally-occurring compounds from amphibian habitats in

three national parks or wildlife refuges and assessed their toxicity to developing anuran

larvae. The extracts did not cause mortality. However, amphibians reared in extracts had a

lengthened larval period or reduced mass at metamorphosis in at least some of the areas

studied. Extracts from both the air and water at one site lengthened the larval period. These

sublethal effects likely influence life history characteristics which in turn affect population

persistence.

Finally, BATTAGLIN et al. demonstrated an important use of the national amphibian

atlas. They compiled species richness by county and compared the patterns to climate

statistics. As expected, precipitation and temperature were significant variables in explaining

richness in most regions. Trends in climate may provide insight into areas of greater stress on

amphibian populations.

RÉSUMÉ

L'Amphibian Research and Monitoring Initiative (AR MI) a commencé en 2000. II s’agit

d'une tentative de l’United States Geological Survey de déterminer le statut et l’avenir des

amphibiens dans les territoires fédéraux des Etats Unis. Les travaux de l'ARMI sont centrés

sur la recherche des causes des déclins, lorsqu'ils existent, la mise au point de nouvelles

techniques pour échantillonner les populations et analyser les données, et la diffusion de

r information aux scientifiques et aux décideurs. Les travaux sont conduits à diverses échelles,

cent est particulièrement mis sur la possibilité de tirer des conclusions sur le statut des

amphibiens dans des zones d'étude bien définies telles que les parcs nationaux et les réserves

naturelles. Plusieurs communications initialement présentées lors d’un symposium aux Etats

Unis en 2004 sont publiées dans ce numéro spécial d’A/ytes.

ACKNOWLEDGMENTS

We thank the ASIH and ISSCA for sponsoring the symposium at the 2004 meetings, and Deanna

Stouder, Chair of the ASIH Symposium Committee, for assistance in scheduling the papers in the

symposium. Dan James and Rick Kearney, as past and current national coordinators of ARMI, have

provided valuable support and guidance throughout the duration of the initiative. We thank the

anonymous reviewers of these papers for providing thoughtful comments and insights in à most timely

manner. Hanan Enani designed the cover of this special issue under the supervision of Hannah

Hamilton. Funding for this special issue of Aves was provided by ARMI

Source : MNHN, Paris

Corx et al. 71

LITERATURE CITED

ANDERSON, D.R., 2001. — The need to get the basics right in wildlife field studies. Wi/d!. Soc. Bull., 29:

1294-1297.

BRICKER, O. P. & RUGGIERO, M. A., 1998. — Toward a national plan for monitoring environmental

resources. Ecol. Applic., 8: 326-329.

Bury, R. B., 1983. — Differences in amphibian populations in logged and old growth redwood forest.

Northwest Sci., 57: 167-178.

Bury, R. B., Dobp, C. K.., Jr. & FELLERS, G. M., 1980. — Conservation of the Amphibia of the United

States: a review. Resource Publications, Washington, US Fish and Wildlife Service, 134: 1-34.

Cor, P.S., ADAMS, M. J., BATTAGLIN, W. A., GALLANT, A. L., JAMES, D. L., KNUTSON, M., LANGTIMM,

C. A., & SAUER, J. R., 2005. - Amphibian research and monitoring initiative: concepts and

implementation. Scientific Investigations Reports, 2003-5015, Reston, Virginia, US Geological

Survey.

s, Washington, US Fish and Wildlife Service, 80 (40.26): 1-56.

Dopp, CK., Jr. 1991. -The status of the Red Hills salamander Phaeognathus hubrichti, Alabama USA,

1976-1988. Biol. Conserv., 55: 57-75.

Biological diversity of a temporary pond herpetofauna in north Florida sandhills. Biodiv. &

v., 1: 125-142.

Do», C. K., Jr, LOMAN, J., COGALNICEANU, D. & PUKY, M. in press. - Monitoring amphibian

populations. In: H. H. HEATWOLE & J. W. WiLKENSON (ed.), Conservation and decline of amphib-

ians, in: Amphibian biology, Volume 9A, Chipping Norton, New South Wales, Australia, Surrey

Beatty & Sons Pty. Ltd.

Gruex, D. M. 1997. — Perspectives on amphibian population declines: defining the problem and

earching for answers. Zn: D. M. GREEN (ed.), Amphibians in decline — Canadian studies of a global

problem, Herpetological Conservation, St. Louis, Missouri, Society for the Study of Amphibians

and Reptiles, 1: 291-308.

HALL, R. J. & LANGTIMM, C. A. 2001. - The US national amphibian r

and the role of protected areas. George Wright Forum, 18: 14-

HEver, W. R., DONNELLY, M. A., MCDiaRMID, R. W., HAYEK, L. C. & FOSTER, M. S. (ed.), 1994. —

Measuring and monitoring biological diversity: standard methods for amphibians. Washington,

Smithsonian Institution Press.

LANNOO, M. J (ed.), 2005. - Amphibian decline:

University of California Press, in press.

MACKENZIE, D. I. & KENDALL, W. L., 2002. —- How should detection probability be incorporated into

estimates of relative abundan

MACKENZIE, D. L., NICHOLS, J. D., LACHMAN, G. B., DROEGE, S., ROYLE, J. A. & LANGTIMM, C. A., 2002.

Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83:

2248-2255.

Merever. C., 2000. - Field guide to malformations of frogs and toads, with radiographic interpretations.

Biological Science Reports, US Geological Survey, USGS/BRD/BSR-2000-000$: 1-18.

MossMaN, M. J., HARTMAN, L. M. Hay. R., SAUER, JL. R. & DHUEY, B. J., 1998. - Monitoring long-term

population trends in Wisconsin frog and toad populations. Ju: M. J. LANNOO (ed.), Status and

conservation of Midhvestern amphibians, lowa City, University of Towa Press: 169-198.

Scumipr, B. R., 2003. - Count data, detection probabilities, and the demography. dynamics, distribution,

and decline of amphibians. C. r. Biol., 326: S119-S124.

Souber, W. 2000. À plague of frogs: the horrifving true storx. New York, Hyperion Press.

rch and monitoring initiative

the conservation status of United States species. Berkeley,

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 72-84.

Estimation of stream salamander

(Plethodontidae, Desmognathinae

and Plethodontinae) populations

in Shenandoah National Park, Virginia, USA

Robin E. JUNG*, J. Andrew ROYLE*, John R. SAUER*, Christopher ADDISON**,

Rebecca D. RAU***, Jennifer L. SHRk**** & John C. WHissEL*****

al Survey, Patuxent Wildlife Research Center,

12100 Beech Forest Road, Laurel, Maryland 20708, USA

#* 3503 Cascade Loop, Yakima, Washington 98902, USA

*** US Fish and Wildlife Service, Patuxent Research Refuge,

11510 American Holly Drive, Laurel, Maryland 20708, USA

**#** Department of Natural Resources, Fernow Hall,

Cornell University, Ithaca, New York 14853, USA

lle, Virginia 22740, USA

##### PO Box 441, Sperry

Stream salamanders in the family Plethodontidae constitute a large

biomass in and near headwater streams in the eastern United States and are

ising indicators of stream ecosystem health. Many studies of stream

salamanders have relied on population indices based on counts rather than

population estimates based on techniques such as capture-recapture and

removal. Application of estimation procedures allows the calculation of

detection probabilities (the proportion of total animals present that are

detected during a survey) and their associated sampling error, and may be

essential for determining salamander population sizes and trends. In 1999,

we conducted capture-recapture and removal population estimation

methods for Desmognathus salamanders at six streams in Shenandoah

National Park, Virginia, USA. Removal sampling appeared more efficient

and detection probabilities from removal data were higher than those from

capture-recapture. During 2001-2004, we used removal estimation at eight

streai the park to assess the usefulness of this technique for long-term

monitoring of stream salamanders. Removal detection probabilities ranged

from 0.39 to 0.96 for Desmognathus, 0.27 to 0.89 for Eurvcea and 0.27

to 0.75 for northern spring (Gyrinophilus porphriticus) and northern red

(Pseudotriton ruber) salamanders across stream transects. Detection

probabilities did not differ across years for Desmognathus and Eurycea,

but did differ among streams for Desmognathus. Population estimates of

Desmognathus decreased between 2001-2002 and 2003-2004 which may

be related to changes in stream flow conditions. Removal-based procedures

may be a feasible approach for population estimation of salamanders, but

field methods should be designed to meet the assumptions of the sampling

procedures. New approaches to estimating stream salamander populations

are discussed.

Source : MNHN, Paris

JUNG et al. 73

INTRODUCTION

Stream salamanders including Desmognathus, Eurycea, Gyrinophilus and Pseudotriton

species in the family Plethodontidae (lungless salamanders) are found in and near seeps and

streams in the eastern United States. These salamanders play an important role in nutrient

cycling and energy flow and can be top vertebrate predators in fishless headwater streams

(Davic, 2002). In Appalachian old growth forests, plethodontid salamanders are extremely

abundant (up to 18,486 individuals/ha) and constitute a large biomass (16.53 kg/ha) exceeding

that of birds (PETRANKA & MURRAY, 2001). Stream salamander larvae develop in seeps and

streams. After transformation, juveniles and adults spend part of their life in the leaf litter,

rocky substrate and banks along streams, foraging on the surface on wet or humid nights, and

hiding beneath rocks, logs, leaves, moss, bark, and in burrows during the day (PETRANKA,

1998).

Because stream salamanders may serve as indicators of stream ecosystem health (CORN

& BURY, 1989; PETRANKA et al., 1993: WELsH & OLLIVIER, 1998; SOUTHERLAND et al., 2004),

identifying reliable survey methods and population estimation techniques for this group are

important. Many studies of stream salamanders have used population indices such as raw

counts and densities (WELSH & OLLIVIER, 1998: Barr & BaBgiTr, 2002), although studies

using population estimates based on capture-recapture and removal sampling have also been

conducted (BRUCE, 1995; NiHuis & KAPLAN, 1998; PETRANKA & MURRAY, 2001). Population

estimates that include detection probabilities (the proportion of total animals present that are

detected during a survey) and their associated sampling error may be essential in determining

population sizes and trends.

Detection probabilities (5) for salamanders can vary spatially, temporally, and by species

and age class (JUNG et al., 2000; SaLvibio, 2001; BAILEY et al., 2004). If j's differ among study

sites or over time, population indices will not be comparable unless differences in detection are

estimated during sampling. Also, magnitude of detection probability can influence precision

and bias of estimated population sizes; if fs or recapture rates of animals are low, standard

errors of estimated population sizes become large, leading to unreliable or biased estimates.

The purpose of this investigation was two-fold. A study conducted in 1999 was designed

to compare capture-recapture and removal estimation methods for stream salamanders and

determine which method was more efficient and provided higher detection probabilities. A

study conducted from 2001-2004 was meant to assess the usefulness of the preferred tech-

nique (removal estimation) for long-term monitoring of stream salamanders in Shenandoah

National Park, Virginia, USA.

MATERIALS AND METHODS

For all stream salamander surveys in Shenandoah National Park, we recorded

der species, age class (those with gills were recorded as larvae and those without were cla

Source : MNHN, Paris

74 ALYTES 22 (3-4)

as adults), and snout-vent length (SVL) and total length (in mm). Because larvae of some

species were occasionally not distinguished in the field and/or because of low sample sizes, we

combined data for northern dusky (Desmognathus fuscus) and seal (D. monticola) salaman-

ders into Desmognathus and combined data for northern spring (Gyrinophilus porphyriticus)

and northern red (Pseudotriton ruber) salamanders into Gyrinophilus/Pseudotriton for analy-

ses. The northern two-lined salamander (Eurycea bislineata) is found in the northern part of

the park, whereas the southern two-lined salamander (E. cirrigera) is found in the southern

part (GHiTEA & SATLER, 1990). Herein, we treat the two species, which are indistinguishable

morphologically, as one when streams throughout the park are considered. Stream salaman-

der species in the park not detected in our surveys included the long-tailed (E. longicauda) and

three-lined (E. guttolineata) salamanders (Wirr, 1993).

In 1999, we compared capture-recapture and removal methods to estimate stream

salamander populations at six streams in the park. Three of the streams were first order

(Jeremy’s Run, Land’s Run, Piney River) and the other three were second order. Two of four

observers turned over the top layer of rocks and other objects greater than 6.4 em maximum

width or length within two 50 x 1 m transects (one on each side of the stream channel; 100 m°

total area) at each stream, sampling the terrestrial habitat immediately adjacent to the wetted

stream channel. Larvae were captured in the hyporheic zone and comprised 9 + 2.8 % (mean

+ s,) of Desmognathus observations, 35 + 8.7 % of E. bislineata observations, and 68 +

11.6 % of Gyrinophilus/Pseudotriton observations across all surveys.

Approximately weekly from 9 July to 13 August 1999, we captured salamanders during

the day by hand and dipnet and batch marked larvae and adults of or above 25 mm SVL using

visible implant fluorescent elastomer (VIE, Northwest Marine Technologies, Inc.), a biocom-

patible latex-based dye injected just under the skin. We used one of three VIE colors (orange,

red, green) at one of four positions just behind the forelimbs or in front of the hindlimbs and

checked marks under a blanket using an ultraviolet light (JUNG et al., 2000). Other studies

have shown that VIE marks are more permanent and cause salamanders less harm than

toe-clipping (Davis & OvaskA, 2001; MaROLD, 2001). Identities and measurements of

unmarked, marked and recaptured salamanders were recorded at each survey. The percent of

salamanders that escaped per stream across capture-recapture surveys averaged 36 + 2.2 %

for Desmognathus, 38 + 5.5% for E. bislineata, and 43 + 10.3% for Gyrinophilusl

Pseudotriton.

After completing 5-6 capture-recapture surveys (five at Keyser Run, Pass Run Tributary,

Piney River), we conducted temporary removal sampling of stream salamanders from 23 to

28 August 1999 at the same transects. Three passes, at least two hours apart, were made each

day for two consecutive days for a total of six removal passes. We tallied the number of larvae

and adults of each species removed at each pass and kept species and age cl s in Separate

buckets with stream water positioned in the stream in the shade. All salamanders were

released back to transects after the final pass.

During June and July of 2001-2004, we used temporary removal sampling as above

though with 2-3 passes conducted one after the other during the day at 1-2 transects at each of

eight streams in the park. Five of the stre: first order and three were second order or

higher. Surveys were also conducted at a ninth stream, Staunton River, but no salamanders

were detected there, presumably due to residual effects of a large flood along this river in 1995.

Source : MNHN, Paris

JUNG et al. 75

For these surveys, we used a modified transect design: transects were 15 m long and 2 m wide

and located on only one side of the stream, spanning 1 m along the stream bank and 1 min the

stream channel. This design allowed for capture of significantly more larval Eurycea(F= 21.9,

df= 1, P < 0.001) and Gyrinophilus/Pseudotriton (F = 44, df = 1, P = 0.043) salamanders

compared to the 1999 stream transects; larvae comprised 5 + 2.2% of Desmognathus

observations, 75 + 4.1% of Eurycea observations, and 89 + 4.2% of Gyrinophilus/

Pseudotriton observations. Our summer surveys mostly missed Desmognathus larvae, which

typically transform by June-July (PETRANKA, 1998).

For capture-recapture data from 1999, we estimated population sizes (À) and detection

probabilities (5) and their standard errors (s,) using maximum likelihood and Bayesian

estimators. These models assume a closed population and fit a series of models that differ in

their assumptions about variation in ÿ during sampling (Oris et al., 1978; RexSTAD &

BURNHAM, 1991). The test for population closure in program CAPTURE showed that all

populations at each stream were closed (all P > 0.05), so open population models were not

used (REXSTAD & BURNHAM, 1991). At least one assumption of capture-recapture models,

that all animals captured are marked, was not met because we estimated only a subset of the

populations, i.e., individuals of or above 25 mm SVL. The Bayesian estimator (GAZEY &

SraLey, 1986) assumes prior distributions for N and p and estimators are derived based on the

posterior distribution. We used a uniform (0,1) prior distribution on p and a diffuse negative

binomial prior distribution on N (GEORGE & ROBERT, 1992). We fit the Bayesian model using

a Markov chain Monte Carlo technique known as Gibbs sampling, in which the posterior

distributions are estimated by simulation. Bayesian estimators do not rely on the asymptotic

properties of maximum likelihood estimators, and hence are preferred for small sample sizes

(Gazey & SrALEY, 1986).

For removal data from 1999 and 2001-2004, we calculated population estimates using

Zippin model M, (ZiPriN, 1958; Wire et al., 1982), which assumes a behavioral response to

capture. Escaped salamanders were excluded from removal pass counts. Escapes were higher

in 1999 compared to 2001-2004, when the percent of salamanders that escaped per stream

across all removal samples averaged 36 + 3.9 % and 26 + 3.1 % for Desmognathus, 43 +

7.3 % and 15 + 1.9 % for Eurycea, and 44 + 18.7 % and 11 + 3.0 % for Gyrinophilus!

Pseudotriton, respectively. Ba: umptions for removal studies include population closure,

equal sampling effort, equal catchability, and effective reduction of the population after each

search. Unfortunately in the case of stream salamanders, some of these assumptions are

E, 1995).

We used the “closed captures” selection in program MARK (WuiTe & BURNHAM, 1999)

to calculate , s, (Ÿ), j'and s, (ÿ) for capture-recapture model M,,, which assumes a constant

capture probability, and removal model M, For Desmognathus data in 1999, we used paired

1-tests to test for significant differences between the js of the capture-recapture and removal

estimates and used program CONTRAST (SAUER & WILLIAMS, 1989; HINES & SAUER, 1990)

to test whether ps from the capture-recapture and removal data differed among streams. For

the 2001-2004 data, we compared whether p's estimated using removal models differed among

Streams within years and among years within streams for Desmognathus and Eurycea using a

Chi Square test implemented in program CONTRAST. Analyses were conducted within

. then pooled to provide a composite Chi Square test with summed

difficult to meet (BRU

groups (streams or yeë

Source : MNHN, Paris

76 ALYTES 22 (3-4)

degrees of freedom. We also tested for differences among years and streams in Ÿ for

Desmognathus and Eurycea using a two-way ANOVA in SPSS (NorusIs, 1992).

RESULTS

Across the capture-recapture surveys at the six streams in 1999, we marked 180 Desmo-

gnathus, 62 E. bislineata, and 17 Gyrinophilus/Pseudotriton (tab. 1). Recapture rates were fairly

low, ranging from 0 to 33 % across species and stream sites (tab. 1). Because the numbers of

marked and recaptured salamanders were low, the simplest model (model M), which

assumes a constant capture probability, was usually the model of choice in program CAP-

TURE: the results of this model are presented in tab. 2. Unless a population is large or

exhibits high capture probabilities, model selection may not be able to detect a pattern in p's

and will select the default model M; (MENKENS & ANDERSON, 1988). Because capture-

recapture estimates based on maximum-likelihood and Bayesian models were from the same

data set, values for Ÿ and j from these methods were quite similar; we used the estimates based

on maximum-likelihood for analyses. We were only able to calculate capture-recapture

population estimates for Æ. bislineata and Gyrinophilus/Pseudotriton at one stream each,

Jeremy’s Run, where one individual of each species was recaptured (tab. 2). Desmognathus

individuals were recaptured at all 6 streams and capture-recapture fs averaged 0.06 + 0.014

(range: 0.02-0.10) (tab. 2). Capture-recapture f’s differed among streams for Desmognathus

(= 133, df= 5, P = 0.02). Capture-recapture does not perform well unless f's exceed 0.30

(Ware et al., 1982), which was never the case. Because of this, the standard errors of À were

sometimes very large (tab. 2).

Based on the six-pass removal data from 1999, population estimates could be calculated

at all 6 stream transects for Desmognathus, 3 for E. bislineata, and 2 for Gyrinophilus!

Pseudotriton (tab. 2). Removal f's averaged 0.25 + 0.077 (range: 0.08-0.61) for Desmognathus

across stream transects, 0.25 + 0.079 (0.09-0.35) for E. hislineata, and 0.44 + 0.060 (0.38-

0.50) for Gyrinophilus/Pseudotriton. Removal f's differed among streams for Desmognathus

( 27.9, df = 5, P < 0.001). Removal fs were significantly higher than those based on

capture-recapture for Desmognathus (1 = 2.2, df = 5, P = 0.04; tab. 2)

For the 2001-2004 data, we calculated removal population estimates for stream salaman-

ders at stream transects at the eight streams (tab. 3). Estimation was not possible when zero

counts occurred on the second pass when two passes were used or on the second and third

passes when three passes were used, or when counts increased across subsequent passes. We

could calculate population estimates at 55 %, 54 % and 13 % of the total 56 stream transects

surveyed from 2001 to 2004 for Desmognathus, Eurycea and Gyrinophilus/Pseudotriton.

respectively (tab. 3). When estimates were calculable, js averaged 0.66 + 0.025 (range:

0.39-0.96) for Desmognathus acro ream transects, 0.64 + 0.031 (0.27-0.89) for Eurycea,

and 0.62 + 0.065 (0.27-0.75) for Gyrinophilus/Pseudotriton (tab. 3). These detection probabi-

lities were much higher than those found in 1999, Using program CONTRAST, we found that

Ps differed among streams for Desmognathus (7° = 29.7, df = 17, P = 0.028) but not for

Eurycea. Detection probabilities did not differ among years at a stream for either Desmogna-

thus or Eurycea.

Source : MNHN, Paris

Table 1. - Numbers of salamanders summed across 5 to 6 capture-recapture surveys in 1999 that

were too small to mark, escaped capture, or were marked or recaptured (% recaptured in

parentheses).

Species Stream Not marked Escaped Marked Recaptured (%)

Desmognathus Heremy"s Run E F7 56 157

Keyser Run 14 2 35 16)

Land's Run 3 30 28 705)

North Fork Thornton 15 20 34 2(6)

Pass Run Tributary 4 25 18 21)

Piney River û 2 9 1)

Eurycea bislineata eremy's Run 7 14 3 163)

Keyser Run 4 2 2 00)

Land's Run 4 4 5 0(0)

Noah Fork Thoton n 20 3 000)

Pass Run Tributary 1 12 4 00)

Piney River 3 7 3 04)

Gyrinophilus / Pseudotriton | Jeremy's Run 0 7 8 13)

Keyser Run 0 1 1 (0)

Land’s Run 0 3 ll 00)

North Fork Thornton 0 ll 2 00)

Pass Run Tributary 0 o 1 0

Piney River o s 4 00)

Table 2. - Population estimates ( N + standard error, s,), 95 % confidence intervals (CI), detection

probabilities (+ s;) and models used for species encountered during capture-recapture

(CR) and removal (REM) sampling in 1999 at 6 streams in Shenandoah National Park. For

Bayesian results, we present the mean N/mode N.

Species - Seam Method ” Passes N(s:) 95% CI À (6) | Mode

Desmognathus

Jeremy”s Run cR 6 11625) 817% | 010002 | o

Bayesian | 6 124/119 (259) 86185 | 0.10(0.023)

REM 6 |aanasr 97(10.4) si | 025005) | 8

Land's Run cR 6 62(18.8) 4112 | 000%) | o

Bayesian | 6 72/66 (25.6) 41437 | 0.09(0.030)

REM 6 65854 63 7.6) 42180 | 0140069 | 8

Pass Run Tributary CR s 716449) 31242 | 0060038 | o

Bayesian | 5 15/84 (1049) 32392 | 0.06(0.035)

REM 6 1256659 108 (00.0) 52537 | 008008 | 8

Piney River CR s 33294) 13-166 | 006(0.056) | ©

Bayesian | S 95/47 (159) 13494 | 006(0.048)

REM 6 514430 21 (56) 1848 [0231 | 8

Keyser Run cR s 493 (4792) 120-2505 | o02(001 | oO

Bayesian | 5 556419 (461.4) 128-184 | 002(0.015)

REM 6 931,000 23(0.0) 2323 | 061007) | 5

North Fork Thomion CR 6 253 (169.3) 9182 | 0@(016 | 0

Bayesin | 6 346/266 (275.5) 97-1078 | 0.030.016)

REM 6 LRTENIRNS 70(18.3) 53159 |o17@om) | 8

Euryeea bisineata

Jeremy's Run cR 6 407) 320 [os | o

Bayesian | 6 19/7 (60.1) 3110 | 01200095)

REM 6 S0(51.3) 2533 | oœns | 5

Pass Run REM 6 6) 66 0350116 | B

Piney River REM 6 300) 33 030(0.145) | B

Gyrinophilus/Pseudoriton

Jeremy's Run cR 6 27042) nas |ooçosn | 0

Bayesian | 6 8240 (1384) 1143 | 0050.08)

North Fork Thomnion REM 6 ,1,0.0,0, 10) 1 0.500358) | 8

Piney River REM 6 101,100 300) 33 0017) | 8

Source : MNHN, Paris

Table 3. - Population estimates ( N + s,), 95 % confidence intervals (CI) and detection probabilities

(+ ss) for salamanders encountered during removal sampling in Shenandoah National

Park. Analyses were based on two or three diurnal passes conducted consecutively at one

(2001) or two (2002-04) transects (T) at each of eight streams (escaped salamanders

excluded) using model My (Zippin) from program CAPTURE. — indicates a third pass was

not conducted. No data for a particular species for a year, stream or transect indicates that

none were detected or that estimation was not possible.

Taxon - Steam var [r|7ss À os C1 Bus)

L 123

[Desmognainus

Piney River 2 | 2038) 06740272)

Piney Tributary 2 |: 10.54 0.670.192)

2 94186) 05240245)

2 |1 3075) 075217

2 [1 3027) 07540217)

2 3(027) 0750217

13 Creek zoo |1 44(1498) 0450207

2m |: 35(075) 0.90 (0.049)

2 47580) 0.630.129)

2 |1 12685) 0.580207)

2 2 (0.38) 06740272)

2 |1 10 (430) 0.390285)

2 90.69) 0.690.128)

Doyle's River zoo |1 10(035) 091 (0087)

202 |2 12(073) 071411)

2 |1 81.06) 0.620135)

Hawksbil 20 | 2801437) 0.42 (0278)

2e |1 254021) 0.96 (0.038)

2005 |2 34070 0.60(0219)

2 |1 3070 0.60 (0219)

2 800.50) 073 (0.134)

Pass Run 2 |2 940.95) 0.640.128)

2 |1 7(087) 0.610.145)

2 74087) 0.640.145)

Jeremy”s Run zoo |1 BA) 0.680224)

2e | 20415) 0.46 (0177)

2 6 (0.38) 0750153)

2 |1 16(088) 0.84 (0084)

2 15 (6.61) 0.51 (0307)

zoo |1 16(072) 0.73 (0005)

2 10(0.63) 071121)

fEunceu

Piney River 200 | 2 | 1061 170104) 068131)

2003 |2| 110 240.38) 0.66 (0272)

20 |1| 542 REA) 0490215)

Piney Tribu 20 [147 | 2auon 0.60 (0.103)

20 2 {ga | 346405 031 (0189

20 [1 lire. | 247 07311)

2 244677) 0.54 (0228)

2 [1 124073) O7)

2 on 0.62 (0121)

zoo |1 80.40) 0.89 (0.105)

2005 [1 601.05) 07540153)

20 |1 10 (086) 06740122)

2 [1 90143 0.74(0232)

22 [1 221.00) 0850071)

2 | 60.05) 0750153)

2 [1 669) 0.51 (0204)

2 214032) O8 (0073)

Pass Ram 2 |2 2850) 047 (0 146)

20 |1 2740.65 0350 187)

2 [is | aoaris 0360150

Jeremy Run a [1 | 62 K (087) 080127)

2 [ai] sa KG) 0.620135)

2 [ils RU) U65 (133)

2 | s20 T{OAN 0.78 0 139)

Pain Run go [122 | os 27 0 206)

me [1 lies 22 (#27) 053020)

2 fume) ueuse 0x1 out)

ms [2/2 MG 0.640 1501

ECO KE ET cn 0860132)

2 [0 6 0.670 157)

ÉGsrmophans Prentotrnen

IS Creek 2 [2] 42. 64105) 6-0 o7sw1s5

Doyle Rner me [ions | aus wo 0270

2] soi EN 07H15

au [1 | 220 110 5 067 (0 102)

me [ai] 52 S(L20 O7117D

Don Li lisa 1 0) De 040 15m

2 [220 10 14 D 6740 102)

Source : MNHN, Paris

JUNG et al. 79

To analyze population change at a site, at least two years of data are needed and analyses

should rely on population estimates to avoid bias associated with raw counts. For Desmogna-

thus and Eurycea, we had complete sets of À for 2001-2004 at three streams each and À for 3

of the 4 years at another two streams (tab. 3). We found significant differences in À across

years (F = 12.7, df 9, P = 0.001) and streams (F = 9.9, df = 4,9, P = 0.002) as well as a

significant year*stream interaction (F = 4.6, df = 10,9, P = 0.015) for Desmognathus, but no

significant differences for Eurycea. Desmognathus population estimates were significantly

higher in 2001 (24 + 7.8) and 2002 (20 + 5.4) compared to 2003 (9 + 2.7) and 2004 (8 + 1.4).

The park experienced heavy precipitation during the summer of 2003, with average stream

flow rates 19 and 7 times higher than stream flow rates in the summers of 2002 and 2001,

respectively (Shenandoah Watershed Study data, Rick Webb, pers. comm.). Data from July

2004 are not yet available, but 2004 flow rates were most likely intermediate between the 2003

and 2001-2002 flow rates.

DISCUSSION

Long-term monitoring programs require cost-effective and efficient techniques to gather

accurate and precise data. Unfortunately, the spatially variable (i.e., significant differences

among streams) and sometimes low detection probabilities found in this study using capture-

recapture and removal methods reinforce the need for estimating fs as part of stream

salamander abundance estimation studies. Our study also indicates the importance of deve-

loping better methods for estimating stream salamander populations such that estimates are

consistently available on a yearly basis for trend analyses.

We found that removal sampling yielded higher fs for stream salamanders than capture-

recapture sampling. Other capture-recapture surveys of stream salamanders have also shown

low recapture rates and hence detection probabilities (BARTHALMUS & BELLIS, 1972; NuHUIS

& KAPLAN, 1998). Indeed, MAROLD (2001) used VIE to mark 44 E. bislineata and D. fuscus

but did not recapture any in the field. BRUCE (199$) used removal sampling (7 p4 set 2-3

days apart) and found low to moderate standard errors for population estimates of D.

monticola and suggested removal sampling was a promising technique to monitor salamander

demographics. Other factors favoring removal over capture-recapture sampling are that

removal sampling usually requires shorter sampling intervals, reduced field personnel,

and less funding than capture-recapture, and appears to be ideal for amphibians such as

aquatic larvae that are highly detectable and have limited home ranges and mobility (HAYEK,

1994).

If removal sampling is to be used for long-term monitoring, field protocols play an

important role in determining their success. In our removal surveys, effective reduction of

populations sometimes did not oceur even after six passes. This may be due in part to the high

e of salamanders that escaped capture, though if we analyzed removal data includ-

a the p: the same issues would be apparent. Itis important to note that when

ent of escapes were lower as they were in 2001-2004 compared to 1999, the ÿ's for

were higher. With fewer escapes and larger sample sizes, there is potential for better

estimates. Removal estimates using Zippin's method are unreliable if less than half the

percent

Source : MNHN, Paris

80 ALYTES 22 (3-4)

population is removed (BRUCE, 1995), but Wie et al. (1982) considered detection probabil-

ities greater than 0.20 adequate for estimating population abundance in removal experiments.

BRUCE (1995) found that 7 passes probably reduced total D. monticola populations by more

than half at his study sites, but he had difficulties reducing numbers of first year juveniles,

which may have shown “increased surface activity...as the larger salamanders were removed

(i.e., a response to reduced competition or predation)”. SOUTHERLAND et al. (2004) used

two-pass removal sampling and were unable to calculate population estimates for species at an

average of 75 % of the streams surveyed because salamander numbers did not decrease or

were zero in the second pass. Removing salamanders from under the top layer of rocks may

disturb or “unearth” other salamanders deeper in the rock substrate. As we sometimes

observed, this can lead to more salamanders in the surface population during subsequent

passes than in the first pass before disturbance.

Several factors could be changed in our removal protocol to improve N and j estimates.

Conducting surveys on wet or humid nights, when more of the salamander population may be

on the surface foraging, might yield better removal estimates. D. fuscus and E. bislineata

emerge one hour after sunset (HOLOMUZKI, 1980) and D. monticola emerge shortly after dark,

with peak activity occurring around midnight and again at dawn (SHEALY, 1975; HAIRSTON,

1986). However, working at night along rocky streams can be difficult and treacherous.

Another option would be to conduct more removal passes, providing the option to group data

from earlier passes in which no decreases in removals occurred. Pilot studies in which a large

number of removal passes are conducted to determine the appropriate number and grouping

of passes may be useful. Another factor to consider is the size and placement of transects or

plots. In our surveys, we only searched narrow 1- or 2-m bands along and/or in the stream.

Most stream salamanders move between the stream channel, splash zone and bank. Home

ranges of D. fuscus have been shown to vary tremendously, from 1.4 m? in Ohio (ASHTON,

1975) to 25-114 m° in Kentucky (BARBOUR et al., 1969). D. monticola home ranges were

estimated to be 8.4 m° in Kentucky (HARDIN et al., 1969). During warm months, £. bislineata

tagged with radioactive isotopes moved within a 14 m° area (AsHTON & AsHTON, 1978), but in

June some post-breeding migrants moved more than 100 m from a stream (MACCULLOCH &

BibER, 1975), which probably explains the particularly low recaptures we observed for this

species in the 1999 capture-recapture surveys. Surveying a wider area of bank along with the

stream channel to incorporate more of the target species’ individual home ranges may yield

better removal estimates.

Other new approaches may prove to be more useful for stream salamander population

estimation. Our removal estimates were based on populations at single stream transects. New

analytical methods developed by ROYLE (2004a-b), ROYLE et al. (2004) and DorAZI0 et al. (in

press) aggregate information across sample sites such that removal sampling can estimate the

abundance of spatially distinct subpopulations. These models incorporate spatial models of

abundance (e.g., Poisson, negative binomial) with models of detection probability and have

been shown to yield abundance estimates with “similar or better precision than those

computed using the conventional approach of analyzing the removal counts of each subpop-

ulation separately” (DORAZIO et al., in press).

A different approach would be to estimate the proportion of area (in this case, streams)

occupied (PAO) by stream salamanders over time (MACKENZIE et al., 2002). The PAO method

Source : MNHN, Paris

JUNG et al. 81

estimates site occupancy and detectability of species based on presence/absence data recorded

from repeated visits to sites selected using a probabilistic sampling frame within an area of

inference. Stream salamander species that exhibit low detection probabilities and occupy

fewer sites would require more streams and visits per stream for PAO estimation (MACKENZIE

& ROYLE, submitted). Note that repeated visits to streams could be satisfied by surveying

multiple transects along the length of a stream.

Despite the problems evident in this study, population estimation efforts incorporating

detection probabilities may be necessary to assess trends in stream salamander populations.

Better survey methods (e.g., transect designs) and population estimation techniques (e.g.,

aggregated removal or PAO approaches) need to be tested and developed such that reasonably

low bias population estimates can be consistently calculated for sites over time. In addition,

spatial design of sampling associated with hypothesis testing incorporating covariates that

may influence stream salamanders (5, À, site occupancy), such as the percent of impervious

surface in a watershed and stream flow rates, should be incorporated alongside monitoring to

best yield inferences about how changes in stream salamander populations over time are

influenced by environmental factors.

RÉSUMÉ

Dans l’est des États-Unis, les salamandres torrenticoles de la famille des Plethodontidae

représentent une biomasse élevée dans et auprès des ruisseaux issus des sources. Elles peuvent

ainsi constituer d’intéressants indicateurs de la santé de ces écosystèmes. Beaucoup d’études

de ces salamandres se sont appuyées sur des indices démographiques utilisant des décomptes

d'animaux et non pas sur des estimations fondées sur des techniques comme les captures-

recaptures ou le ramassage des individus. L'emploi de procédures d'évaluation permet le

calcul de probabilités de détection (la proportion d'animaux réellement présents détectés lors

d’une étude) et de leur écart-type, et peut permettre de déterminer les tailles et les dynamiques

des populations de salamandres. En 1999, nous avons employé les méthodes de capture-

recapture et de ramassage pour évaluer des populations de salamandres du genre Desmogna-

thus dans six ruisseaux du Pare National de Shenandoah (Virginie, Etats-Unis). La méthode

du ramassage s’est avérée plus efficace: elle a donné des probabilités de détection plus élevées

que celle de capture-recapture. Pendant la période 2001-2004, nous avons employé la méthode

du ramassage dans et auprès de huit ruisseaux du Parc afin d'évaluer la fiabilité de cette

technique pour la surveillance à long terme de ces populations de salamandres. Lors de

transects le long des ruisseaux, nous avons obtenu des probabilités de détection de 0,39 à 0,96

pour Desmognathus, de 0,27 à 0.89 pour Eurycea et de 0,27 à 0,75 pour Gyrinophilus

porphyriticus/Pseudotriton ruber. Les probabilités de détection n’ont pas varié au cours des

années pour Desmognathus et Eurycea, mais ont différé selon les ruisseaux pour Desmogna-

thus. Les évaluations des populations de Desmognathus ont diminué entre 2001-2002 et

2003-2004, ce qui peut être lié à des changements dans le régime hydrique des ruisseaux. Les

procédures de ramassage constituent une méthode fiable pour l'évaluation de populations de

ces salamandres, mais les méthodes de terrain doivent être conçues de manière à remplir les

Source : MNHN, Paris

82 ALYTES 22 (3-4)

conditions statistiques des méthodes d’échantillonnage. De nouvelles méthodes d'estimation

des populations de ces salamandres sont discutées.

ACKNOWLEDGMENTS

We would like to thank James Atkinson at Shenandoah National Park, the 2001-2004 field crews

(Isaac Chellmann, Sara Faust, Lindsay Funk, Evan H. C. Grant, Andrew Mongeon, Priya Nanjappa,

Shecra Schneider, Edward Schwartzman), and Sam Droege, This work was funded by the Park Research

and Intensive Monitoring of Ecosystems Network (PRIMENet), a joint initiative of the National Park

Service and Environmental Protection Agency, and by the US Geological Survey’s Amphibian Research

and Monitoring Initiative (ARMI).

LITERATURE CITED

ASHTON, R. E., Jr., 1975. - À study of movement, home range, and winter behavior of Desmognathus

fuscus (Rafinesque). J Herp., 9: 85-91.

ASHTON, R. E., Jr. & ASHTON, P. S., 1978. - Movements and winter behavior of Eurycea bislineata

(Amphibia, Urodela, Plethodontidae). J Herp., 12: 295-

BAILEY, L. L., SIMONS, T. R. & PoLLOCK, K. H., 2004. - Spatial and temporal variation in detection

probability of Plethodon salamanders using the robust capture-recapture design. J Wildl. Man-

agement, 68: 14-24.

BARBOUR, R. W., HARDIN, J. W., SCHAFER, J. P. & HARVEY, M. J., 1969. - Home range, movements, and

activity of the dusk, lamander, Desmognathus fuscus. Copeia, 1969: 293-297.

Bark, G. E. & Barr, K. J., 2002. - Effects of biotic and abiotic factors on the distribution and

abundance of larval two-lined salamanders (Eurycea bislineata) across spatial scales. Oecologia,

133: 176-185.

BARTHALMUS, G. T. & BELLIS, E. D., 1972. - Home range, homing, and the homing mechanism of the

salamander, Desmognathus fus Copeia, 1972: 632-642,

BRUCE, R. C., 1995. - The use of temporary removal sampling in a study of population dynamics of the

salamander Desmognathus monticola. Austr. J. Ecol., 20: 403-412.

Cor, P. S. & BURY, R. P, 1989. - Logging in western Oregon: responses of headwater habitats and

stream amphibians. Forest Ecol. Management, 29: 39-57.

Davic, R. D. (ed.), 2002. - Field evaluation manual for Ohio's primary headwater habitat streams.

Version 1.0. Columbus, Ohio, USA, Ohio Environmental Protection Agency, Division of Surface

Water: 1-37.

Davis, T. M. & Ovaska, K., 2001. - Individual recognition of amphibians: effects of toe clipping and

fluorescent tagging on the salamander Plethodon vehiculum. J. Herp.. 35: 217-2

Doraz10, R. M. JELKS, H. L. & JORDAN, EF. in press. - Improving removal-b ni dor abundance

by sampling a population of spatially distinet subpopulations. Biometries.

Gazey, W. I. & SraLry, M. J., 1986. - Population estimation from mark-recapture experiments using a

sequential B Igorithm. Ecology, 67: 941-951

GroRGE, E. I. & RoëerT, C. P. 1992. - Capture-recapture estimation via Gibbs sampling. Biometrika, 79:

677-683.

GuiTea, O. & SATTLER, P, 1990. - The distribution and identification of two-lined salamanders in

Virginia, Catesheiana, 10: 11-18.

HARSTON, N, G., Sr., 1986. - Species packing in Desmognathus

tion of predation and competition. Am. Nat., 127: 266-291

HARDIN, JW. SCHAFER, J. P. & BARBOUR, R. W.. 1969. — Observations on the activity of a seal

lamander, Desmognathus monticola. Herpetologica, 25: 150-151.

alamanders: experimental demonstra-

Source : MNHN, Paris

JUNG et al. 83

HAYEK, L. C., 1994. - Removal sampling. Zn: W. R. HEYER, M. A. DONNELLY, R. W. MCDIARMI, L. C.

HaAYEr & M. S. FosTER (ed.), Measuring and monitoring biological diversity: standard methods for

amphibians, Washington, DC, Smithsonian Institution Press: 201-205.

Hines, J. E. & SAUER, J. R., 1990. - Program CONTRAST: a general program for the analysis of several

survival or recovery rate estimates. US Department of the Interior, Fish and Wildlife Service

Technical Reports, 24: 1-7.

HOLOMUZKI, J. R., 1980. - Synchronous foraging and dietary overlap of three species of plethodontid

salamanders. Herpetologica, 36: 109-115.

JuNG, R. E., DROEGE, S., SAUER, J. R. & LANDY, R. B., 2000. - Evaluation of terrestrial and streamside

salamander monitoring techniques at Shenandoah National Park. Environ. Monit. Assessment, 63:

65-79.

MaAccuLLoCH, R. D. & BIDER, J. R., 1975. - Phenology, migrations, circadian rhythm and the effect of

precipitation on the activity of Eurycea b. bislineata in Quebec. Herpetologica, 31: 433-439.

MACKENKIE, D. I., NICHOLS, J. D., LACHMAN, G. B., DROEGE, S., ROYLE, J. A. & LANGTIMM, C. À. 2002.

— Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83:

2248-2255.

MACKENZIE, D. L. & ROYLE, J. A., submitted. - Designing occupancy studies: general advice and tips on

allocation of survey effort. J appl. Ecol.

MaRoOLD, M. R., 2001. - Evaluating visual implant elastomer polymer for marking small, stream-

dwelling salamanders. Herp. Rev., 32: 91-92.

MENKENS, G. E., Jr. & ANDERSON, S. H., 1988. — Estimation of small-mammal population size. Ecology,

69: 1952-1959,

Nuauis, M. J & KaPLaN, R. H., 1998. - Movement patterns and life history characteristics in a

population of the cascade torrent salamander (Rhyacotriton cascadae) in the Columbia River

gorge, Oregon. J Herp., 32: 301-304.

Norusis, M. J., 1992. - SPSS/PC+ Base System User's Guide Version 5.0. Chicago, Il., SPSS Inc.

Oris, D. L., BURNHAM, K. P., WHITE, G. C. & ANDERSON, D. R., 1978. - Statistical inference from capture

data on closed animal populations. Wildlife Monogr. 62: 1-135.

PETRANKA, J. W., 1998. — Salamanders of the United States and Canada. Washington, DC, Smithsonian

Institution Press: 1-587.

PETRANKA, J. W., ETHERIDGE, M. E. & HALEY, K. E., 1993. - Effects of timber harvesting on southern

Appalachian salamanders. Conserv. Biol., 7: 363-370.

PErRANKA, J. W. & MURRAY, S.S., 2001. - Effectiveness of removal sampling for determining salamander

density and biomass: a case study in an Appalachian streamside community. J. Herp., 35: 36-44.

RexsraD, E. & BURNHAM, K., 1991. - User's guide for interactive program CAPTURE: abundance

estimation of closed animal populations. Fort Collins, Colorado, USA, Colorado Cooperative Fish

& Wildlife Research Unit, Colorado State University: 1-29.

ROYLE, J. A., 2004a. - N-mixture models for estimating population size from spatially replicated counts.

Biometrics, 60: 108-115.

Lea 2004b. - Generalized estimators of avian abundance from count survey data. Anim. Biodiv. Conserv.,

: 375-386.

A., Dawson, D. K. & BATES, S., 2004. - Modeling abundance effects in distance sampling.

ology, 85: 1591-1597.

SaLvinio, S., 2001. — Estimating terrestrial salamander abundance in different habitats: efficiency of

temporary removal methods. Herp. Rev., 32: 21-24.

SAUER, JR. & WILLIAMS, B. K., 1989. - Generalized procedures for testing hypotheses about survival or

recovery rates. J Wildl. Management, 137-142.

SHEALY, R. M. 1975. - Factors influencing activity in the salamanders Desmognathus ochrophaeus and D.

monticola (Plethodontidae). Herpetologica, 31: 94-102.

SOUTHERLAND, M. T.. JUNG, R. E., BAXTER, D. P., CHELLMANN, I. C., MERCURIO, G. & VOLSTAD, J. H.,

2004. — Stream salamanders as indicators of stream quality in Maryland, USA. Appl. Herp., 2:

23-46.

WELsH, H.H., Jr. & OLuivier, L. M., 1998. - Stream amphibians as indicators of ecosystem stress: a case

study from California’s Redwoods. Ecol. Appl., 8: 1118-113

Wire, G. C. & BURNHAM, K. P., 1999. — Program MARK: survival estimation from populations of

marked animals. Bird Study, 46: 120-138.

Source : MNHN, Paris

84 ALYTES 22 (3-4)

Ware, G. C., ANDERSON, D. R., BURNHAM, K. P. & Ori, D. L., 1982. - Capture-recapture and removal

methods for sampling closed populations. Los Alamos, New Mexico, USA, Los Alamos National

Laboratory LA-8787-NERP: 1-235.

Wir, W. L., 1993. — Annotated checklist of the amphibians and reptiles of Shenandoah National Park,

Virginia, Catesbeiana, 13: 26-35.

Ziprin, C., 1958. - The removal method of population estimation. J. Wild. Management, 22: 82-90.

Corresponding editor: C. Kenneth DoDp, Jr.

©ISSCA 2005

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 85-94. 85

Status of amphibians

on the Continental Divide:

surveys on a transect from Montana

to Colorado, USA

Paul Stephen CoRN*, Blake R. Hossack*, Erin MUTHS**,

Debra A. PATLA***, Charles R. PETERSON*** & Alisa L. GALLANT***%*

* US Geological Survey, Aldo Leopold Wilderness Research Institute,

Box 8089, Missoula, Montana 59807, USA

** US Geological Survey, Fort Collins Science Center,

2150 Centre Ave., Bldg C, Fort Collins, Colorado 80526, USA

*** Department of Biological Sciences, Idaho State Univers

Campus Box 8007, Pocatello, Idaho 83209, USA

###* US Geological Survey, EROS Data Center,

Sioux Falls, South Dakota 57198, USA

The Rocky Mountain Region of the United States Geological Survey's

Amphibian Research and Monitoring Initiative is conducting monitoring of

the status of amphibians on a transect that extends along the Continental

Divide from Canada to Colorado and comprises four National Parks.

Monitoring uses visual encounter surveys to determine site occupancy, with

its to a subset of sites to estimate detection probabilities for

each species. Detection probabilities were generally high (above 0.65)

among species. There was a gradient in site occupancy, with most species

scarce in the south and relatively common in the north. For example, Bufo

boreas is close to extinction in Rocky Mountain National Park, was found at

fewer than 5 % of sites in Yellowstone and Grand Teton National Parks in

the middle of the transect, but occurs at approximately 10 % of sites in

Glacier National Park. The salamander Ambystoma tigrinum was rare in

Rocky Mountain and occurred at less than 25 % of sites at Yellowstone

and Grand Teton, but A. macrodactylum occurred at more than 50 % of

sites in Glacier. There are numerous differences among parks, such as

latitude, climate, numbers of visitors, and human population density in the

surrounding landscape. The degree to which these factors have influenced

the current distribution and abundance of amphibians is unknown but

should be a focus of additional research.

INTRODUCTION

Amphibian declines have occurred in the western United States at a disproportionate

rate (BRADFORD, 2005). These declines are not restricted to areas where land use has directly

Source : MNHN, Paris

86 ALYTES 22 (3-4)

altered habitat, but have occurred on protected federal lands including national parks and

wilderness areas (DRosT & FELLERS, 1996; KNAPP & MATTHEWS 2000; MATTHEWS et al., 2003,

Murs et al., 2003). To determine the status and trends of amphibian species in the Rocky

Mountains, which contain some of the United States’ most significant protected landscapes,

the Rocky Mountain Region of the United States Geological Survey Amphibian Research

and Monitoring Initiative (ARMI) has established a transect along the Continental Divide

that includes four national parks as middle-level monitoring sites. The middle level refers to a

pyramid of monitoring efforts, starting with base-level inventories over large geographic areas

and ending with focused apex studies at a few sites (HALL & LANGTIMM, 2001; CoRN et al.,

2005). At the middle level, surveys are conducted within a defined study area, such as a

national park, with a sampling design that allows inference about the status and trends of the

amphibians within that area.

The national parks on the Continental Divide are spaced across about 8° of latitude,

from Rocky Mountain in Colorado, through Grand Teton and Yellowstone in northwest

Wyoming (referred to collectively as GYE for brevity and because they compose the center of

the Greater Yellowstone Ecosystem), to Glacier in northwest Montana (fig. 1). The parks

differ in size, climate, and potential degree of anthropogenic influence (tab. 1). All three study

areas are characterized by similar zones of vegetation. At lower elevations, the coniferous

montane forest is dominated by ponderosa pine (Pinus ponderosa), lodgepole pine (P

contorta) or Douglas fir (Pseudotsuga menziesit), with western redcedar (Thuja plicata) and

western larch (Larix occidentalis) in Glacier. The mid-elevation subalpine forests are com-

posed primarily of Engelmann spruce (Picea engelmannii), subalpine fir (Abies lasiocarpa)

and white pines (Pinus flexilis, P. albicaulis). There are also significant alpine habitats above

tree line, which decreases about 100 m in elevation for every 1° north in latitude, from about

3500 m in Rocky Mountain to 2600 m in Glacier (PEET, 1988).

Amphibian declines have occurred to the greatest extent at the southern end of the

transect. In Rocky Mountain National Park, the boreal toad (Bufo boreas) is nearing

extinction (MUTHS et al., 2003) and the northern leopard frog (Rana pipiens) has not been

observed since 1974 (CorN et al., 1997). At the northern end of the transect, surveys in

Glacier have found all of the Park's resident amphibian species, and there is little suggestion

of recent declines (MARNELL, 1997; ADAMS et à 005; PSC & BRH, unpublished). Surveys

in the 1990s and in 2000 and 2001 suggest that amphibian declines in the two parks of the

GYE may be intermediate between Rocky Mountain and Glacier. Rana pipiens has largely

disappeared from the GYE, and B. boreas is present, but at a reduced number of locations

(KocH & PETERSON, 1995; VAN KiRK & PATLA, 2000; PATLA & PETERSON, 2001a-b).

Conducting surveys to determine status and trends of amphibians in all three study areas

(Glacier, GYE, Rocky Mountain) provides the ability to monitor changes over an unprece-

dented latitudinal gradient and the opportunity to compare changes in status of amphibians

to gradients in climate and habitat. Given the erratic nature of the population dynamics of

many amphibian species, ARMI has chosen to follow the advice of GREEN (1997) and

concentrate on detecting change in numbers of populations rather than numbers of individ-

als within populations (CORN et al., 2005). We determin: tus of each species by using the

presence or absence of breeding at individual sites to estimate occupancy, expressed as the

proportion of sites occupied. Trends will be assessed by examining changes in occupancy once

Source : MNHN, Paris

CorK et al. 87

Yellowstone &

Grand Teton

. National Parks

£esper

100 200 km

Rocky Mountain

National Park

Denver

Durango

Fig. L.— The three study areas are distributed on the Continental Divide. The scale bar refers to the four

states figured and not to the inset United States map.

enough data are collected for reliable analyses. However, failure to detect presence of a species

during a visit to a site may not mean that the species is absent. Therefore, we employ multiple

visits to sites and recently-developed statistical tools (MACKENZIE et al., 2002, 2003) that

incorporate detection probabilities to estimate occupancy.

The ultimate goal of Rocky Mountain ARMI is to conduct these surveys over the long

term, in collaboration with the National Park Service. The data and analyses will provide park

managers with valuable information, such as whether declines seen so far in the south are

Source : MNHN, Paris

88 ALYTES 22 (3-4)

Selected statistics describing the location, climate, and human influences on the three study

areas on the Continental Divide. Park areas and use statistics are from ANONYMOUS (2004).

Climate data are the averages of National Climate Data Center reporting stations within or at

the margins of the parks.

Trait Glacier GYE Rocky Mountain

Size (ha) 410,178 1033,389 107,577

Central latitude 4837N 44°22N 40°22N

Central longitude 113°49°W 110933W 105°42°W

2003 recreational visits 1,664,046 5.375.068 3.067.256

2003 backcountry overnight stays 22,958 59,036 36,012

Nearby population! 851,000 528,000 3,925,000

1961-1990 mean annual temperature (°C) 47 27 42

1961-1990 mean annual precipitation (mm) 614 527 457

United States: 2000 Census, sum of county populations where the central coordinate of the county lies within 150 miles

{241 km) of the central coordinate of the study area; Canada: 2001 Census, sum of Census Divisions 02 and 03 in Alberta

and O1 and 03 in British Columbia (<www.statcan.ca>).

moving north, or whether amphibians that have declined are beginning to recover. Our

objectives in this paper are to describe the design of surveys on the transect, report the rates

of site occupancy for each species for the first two years of observations, and provide initial

estimates of trend in site occupancy for B. boreas in Glacier.

METHODS

Our sampling units were a set of hierarchically nested drainage catchments from the

USGS Elevation Derivatives for National Applications Project (Kosr & KELLY, 2001).

Drainages were overlaid onto National Wetlands Inventory (NWI) maps and aggregated to

include 10-50 identifiable water bodies, which resulted in 40 catchments in Rocky Mountain

and 80 in Glacier. There were 1060 catchments in GYE before aggregation. Because of the

large number of catchments in GYE, we did not aggregate them until one was selected for

sampling. We selected survey units by drawing systematic random samples in each study area

to ensure spatial balance. The goal was to survey every accessible water body in selected

catchments at least once, with a subset (Glacier and GYE) or all (Rocky Mountain) sites

surveyed twice or more to estimate detection probabilities. Duplicate surveys were achieved

by independent surveys on different days or on the same day by biologists working indepen-

dently or working together at different times. For example, areas with numerous small water

bodies were sometimes sampled independently by each biologist. Large sites such as wet

meadows or river valleys were sampled by two biologists simultaneously. We used visual

encounter surveys (HEYER et al., 1994) to search for all life stages of amphibians (embryo,

larva and adult) in accessible portions of the water body, using dip nets to sample a with

Source : MNHN, Paris

Corx et al. 89

limited visibility, and recorded site (e.g., pond area, extent of emergent vegetation) and

sampling covariates (e.g., temperature, date) for all surveys (ADaMs et al., 2005). A site was

considered to be occupied only if breeding had occurred there, as indicated by presence of egg

masses, larvae, or recently metamorphosed juveniles. Presence of adults not engaged in

breeding activity was insufficient to consider a site as occupied.

We estimated detection probabilities and occupancy using the program PRESENCE

(MACKENZIE et al., 2002). For this analysis, we used the simplest model, assuming constant

probabilities of detection and site occupancy. Descriptions of the methods are available

elsewhere (MACKENZIE et al., 2002, 2003; MACKENZIE & BAILEY, 2004).

To provide an early assessment of trends in populations of B. horeas in Glacier, we

compared changes in numbers of occupied sites in three catchments that have been monitored

each year from 1999 to 2002 (restrictions due to extensive fires in Glacier in 2003 prevented

access to our sites). However, somewhat differing levels of effort among years made estima-

tion of occupancy problematic, so we computed year-to-year estimates of trend (GREEN,

2003) as: trend = In(N/N,.,), where N = the number of occupied sites, t = the current year, and

n = a previous year (1 to 3, in this case). Only sites that were sampled in both years of the

comparison were included in the calculation of trend. Positive values of trend indicate

increases in site occupancy, and negative values indicate decreases.

RESULTS

We surveyed 140 (2002) and 79 (2003) sites in 15 catchments in Rocky Mountain, 183

(2002) and 189 (2003) sites in 17 catchments in GYE, and 325 (2002) and 116 (2003) sites in 17

catchments in Glacier. Species detected included B. boreas, wood frog (R. sylvatica) and

boreal chorus frog (Pseudacris maculata) in Rocky Mountain, B. boreas, P. maculata, Giger

salamander { Ambystoma tigrinum) and Columbia spotted frog (R. luteiventris) in GYE, and

B. boreas, R. luteiventris and long-toed salamander (4. macrodactylum) in Glacier. We

detected À. rigrinum at Rocky Mountain, but sampling was inadequate and we do not report

these data here. We also observed P maculata, the Pacific treefrog (P regilla) and the Rocky

Mountain tailed frog (Ascaphus montanus) in Glacier, but these species either have ranges that

terminate near the border of the Park and have been found in less than 10 sites each (the

hylids) or occur in habitats that we did not sample in these surveys (4. montanus inhabits

headwater streams). Our observations of P maculata, however, do represent an addition to

the fauna occurring in Glacier (HOSSACK & YALE, 2002).

Detection probabilities were high for all species (tab. 2), with the exception of one

anomalous value for P maculata. We are unsure of the cause of this low value, but we consider

it to be an outlier. Anurans tended to have higher detection probabilities (mean 0.84, not

including P maculata in 2003 at Rocky Mountain) than did salamanders (mean 0.68).

Estimated occupancy for each species varied between 0 and 0.57 and decreased from north to

south (tab. 2). The highest occupancy observed in Rocky Mountain was 0.10 for R. sylvarica

in 2003, but this species has a restricted distribution in the Park. In contrast, the only

occupancy below 0.10 in Glacier was for B. boreas in 2002, and À. macrodactylum occurred at

Source : MNHN, Paris

90 ALYTES 22 (3-4)

Table 2. — Site occupaney and detection probabilities of amphibians in the three study areas. Dashes

indicate parameters that could not be estimated due to sparse data.

70 ET

Study areu/ Species —

Gisir

Anbystoms mocrodacptan 038 068 03 | oos | ow o7s os | 00

Bu borees 006 090 007 | oo | où o7 os | oo

Rare huentris ou a o7 | os | 02 02 ox | oo

ave

Aystome iron on o61 oi | ons | ou ve ox | oo

Bu borees 005 o0s | oow6 | oo x 002 oo

Pseudacrs acute 038 os 04 005 030 os ox | oo

Rama huehenris 020 075 02% | os | ou 09 on | oo

Rocky Mountain

Bufoboreas oo! - 5 : o - ; ;

Paendacrs maculote 002 on 02 | oo | 0 02 os | ous

Rama sharica 00 u 003 | oo | ou os ow | ooss

less than 50 % of sites. Occurrence of amphibian species in GYE was intermediate between

Rocky Mountain and Glacier.

There was no apparent trend in numbers of breeding sites of B. boreas in Glacier between

1999 and 2002 (tab. 3). There was no trend between 1999 and 2000, an increase in site

occupancy between 2000 and 2001, followed by a decrease in occupancy of the same

magnitude between 2001 and 2002. Trend over greater intervals shows less variability.

Comparing 2001 to 1999, there was only a slight increase in site occupancy, and there was no

trend between 1999 and 2002.

DISCUSSION

Results of our surveys in 2002-2003 confirm data from earlier studies that amphibian

occurrence is greatly reduced in the southern Rocky Mountains. The cause of the decline of B.

boreas in Rocky Mountain and elsewhere in the southern Rocky Mountains is thought to be

the result of infection by the fungus Batrachochytrium dendrobatidis (MUTHs et al., 2003;

CaREY et al., 2005). Although B. dendrobatidis has been detected in R. sylvatica and P.

maculata in Rocky Mountain (RITTMANN et al., 2003; GREEN & MuTHs, 2005), we do not

know how or whether these species have been affected.

There has been relatively little direct physical alteration of amphibian habitats in any of

the parks. However, the high visitor use and the greater size of the nearby human population

at Rocky Mountain suggests that it is reasonable to hypothesize that anthropogenic influences

have contributed directly or indirectly to the low occupancy of amphibians there. For

example, DAviDsON et al. (2002) found an anthropogenic influence (amount of agriculture

Source : MNHN, Paris

Corx et al. 91

Table 3. — Annual trends in populations of 8. boreas in three catchments in Glacier National Park. The

highlighted diagonal shows the comparisons between each year (columns) and the previous

year (rows). Values above the diagonal show comparisons between sites sampled both in the

year of each column and 2 or 3 years before.

Vear-to-Year trend

Year Sites surveyed _ |Sites with breeding a et D

1999 E 8 0.12 000

2000 18 2 0.08

2001 28 23 3 024

2002 41 28 - = -

upwind) was correlated with reduced occurrence of four species of amphibians in California.

Adjusted for area, Rocky Mountain receives about five times the visitor use, both in the

backcountry and by visitors touring the parks by automobile, than do Glacier or the parks in

the GYE (tab. 1). Rocky Mountain is the smallest of the parks in the transect, but has the

greatest number of people living nearby and may be less buffered from outside influences. For

example, land use change (urbanization and agricultural development) on the Colorado

piedmont east of the park has resulted in cooler and wetter summers in Rocky Mountain

(STOHLGREN et al., 1998), and increased nitrogen deposition from urban and agricultural

sources has altered the diatom communities of Alpine lakes (WoLre et al., 2001). We can offer

no direct evidence that human influences have affected the occurrence of amphibians in

Rocky Mountain, but this topic deserves further research.

Differences in climate among study areas could influence the patterns we observed.

Rocky Mountain is warmer than the GYE, but receives the least precipitation of the three

study areas. Climate and extreme weather events can directly affect amphibian populations

(Cai & ALEXANDER, 2003), and the interactions between climate and anthropogenic

influences have received little study.

Whether the levels of habitat occupancy seen in GYE and Glacier represent declines of

any species is unknown. It is suspected that B. boreas should occur at a greater number of

potential breeding sites than has been observed recently (KoCH & PETERSON, 1995; MAXELL et

al., 2003). Alternatively, the Rocky Mountains have low diversity of amphibians, and it may

be that low species diversity and low habitat occupancy are linked. Unfortunately we lack

knowledge of historical habitat occupancy rates for any species. The recent data from Glacier

indicate no trend since 1999, but this is still 00 brief a time series to draw firm conclusions.

The goal of Rocky Mountain ARMI is to collect data over the long term, so that we will

be able to draw conclusions about trend for each species within the national parks, and

Patterns across the region. Longer time series will allow the use of more sophisticated models

to estimate detection probabilities and occupancy, including rates of population turnover

(MACKENZIE et al., 2003). During the first years of monitoring, the effort in each study area

was constrained somewhat by funding, resulting in uneven allocation of effort based on size

of study areas. Beginning in 2005, we will use size of the areas inallocating samples, which will

Source : MNHN, Paris

92 ALYTES 22 (3-4)

result in increased effort at GYE. Future analyses will also incorporate covariates in the

estimation of detection probabilities and site occupancy. Site-specific covariates do not

change across sampling occasions within years (e.g., many habitat variables), whereas sam-

pling covariates (e.g., date or weather) may change for each sampling occasion. Both site and

sampling covariates can influence detection probabilities, but only site covariates are used to

model occupancy probabilities. Analyses of trend also will need to include covariates. For

example, annual precipitation may determine the number of available breeding sites. Paradox-

ically, some species may have higher occupancy in dry years, because there are fewer sites

available. The increase in occupancy by B. boreas in Glacier in 2001, a year with extremely low

winter snowpack, may be an illustration of this phenomenon. We need to be able to

distinguish changes in occupancy caused by variation in natural conditions, such as habitat

changes due to drought and wildfire or outbreaks of infectious disease, from changes caused

by anthropogenic influences, such as introduction of a novel pathogen.

RÉSUMÉ

Les chercheurs de la Rocky Mountain Region de l’United States Geological Survey’s

Amphibian Research and Monitoring Initiative ont effectué une étude sur le statut des

amphibiens le long d’un transect qui suit la ligne de partage des eaux du Canada au Colorado

et qui comprend quatre parcs nationaux. L'étude s’est appuyée sur des inventaires par contact

visuel pour déterminer l'occupation des sites, avec des visites multiples à un sous-ensemble de

sites pour estimer les probabilités de détection pour chaque espèce. Pour chaque espèce, les

probabilités de détection ont été généralement élevées (plus de 65 %). Un gradient d’occupa-

tion des sites a été mis en évidence, la plupart des espèces étant rares dans le sud et relativement

communes dans le nord. Par exemple, Bufo boreas est proche de l'extinction dans le Rocky

Mountain National Park, présent dans moins de 5 % des sites dans les parcs nationaux de

Yellowstone et du Grand Teton au milieu du transect, mais présent dans environ 10 % des sites

dans le Glacier National Park. La salamandre Ambystoma tigrinum est rare à Rocky Moun-

tain et présente dans moins de 25 % des sites à Yellowstone et au Grand Teton, mais 4.

macrodactylum a été observée dans plus de 50 % des sites à Glacier. Il y a de nombreu

différences entre les parcs, telles que la latitude, le climat, l'abondance des visiteurs et la densité

de la population humaine dans la région avoisinante. Le degré auquel ces facteurs ont

influencé la distribution et l'abondance actuelles des amphibiens est inconnu et devrait faire

l’objet de recherches complémentaires.

ACKNOWLEDGMENTS

Surveys were funded by the US Geological Survey and the National Park Service. We thank the

administrative, natural resource management, and ranger staff of G Grand Teton, Rocky Moun-

tain, and Yellowstone national parks for their cooperation and assistance.

Source : MNHN, Paris

Corx et al. 93

LITERATURE CITED

ANONYMOUS [National Park Service], 2004. — Sratistical abstract 2003. Denver, Colorado, National Park

Service Social Science Program, Public Use Statistics Office: 1-74.

ADAMS, M. J., HossaCK, B. R., KNAPP, R. A., CoRN, P.S., DIAMOND, $. A., TRENHAM, P. C. & FAGRE, D.,

2005. — Distribution patterns of lentic breeding amphibians in relation to ultraviolet radiation

exposure in western North America. Ecospstems, in press.

BRADFORD, D. F., 2005. — Factors implicated in amphibian population declines in the United States. Zn:

M. J. LANNOO, (ed.), Amphibian decline: the conservation status of United States species, Berkeley,

University of California Press: in press.

Carey, C. & ALEXANDER, M. A., 2003. - Climate change and amphibian declines: is there a link? Diversity

and Distributions, 9: 111-121.

CaREY, C., CoRN, P.S., JONES, M.S., Livo, L. J., MuTHS, E. & LOEFFLER, C. W., 2005. - Environmental

and life history factors that limit recovery in southern Rocky Mountain populations of boreal

toads (Bufo b. boreas). In: M. J. LANNOO, (ed.), Amphibian decline: the conservation status of United

States species, Berkeley, University of California Press: in press.

Cow, P.S., ADAMS, M. J., BATTAGLIN, W. A., GALLANT, A. L., JAMES, D. L., KNUTSON, M., LANGTIMM,

C. A., & SAUER, J. R., 2005. - Amphibian research and monitoring initiative: concepts and

implementation. Scientific Investigations Report, 2003-5015, Reston, Virginia, US Geological

Survey.

Cor, P. S., JENNINGS, M. L. & MUTHS, E., 1997. - Survey and

Rocky Mountain National Park. Northwestern Naturalist, 78: 34-55.

DavipsoN, C., SHAFFER, H. B. & JENNINGS, M. R., 2002. — Spa tests of the pesticide drift, habitat

destruction, UV-B, and climate-change hypotheses for California amphibian declines. Conserva-

tion Biology, 16: 1588-1601.

Drosr, C. A. & FELLERS, G. M., 1996. - Collapse of a regional frog fauna in the Yosemite area of the

California Sierra Nevada, USA. Conservation Biology, 10: 414-425.

GREEN, D. E. & MuTHs, E., 2005. - Health evaluation of amphibians in and near Rocky Mountain

National Park (Colorado, USA). Alytes, 22 (3-4): 109-129.

GREEN, D. M., 1997. — Perspectives on amphibian population declines: defining the problem and

searching for answers. Jr: D. M. GREEN (ed.), Amphibians in decline — Canadian studies of a global

problem, Herpetological Conservation, St, Louis, Society for the Study of Amphibians and Rep-

tiles, 1: 291-308.

- 2003. — The ecology of extinction: population fluctuation and decline in amphibians. Biological

Conservation, 1: 331-343.

HaLz, R. J. & LANGTIMM, C. A. 2001. - The US national amphibian research and monitoring initiative

and the role of protected areas. George Wright Forum, 18: 14-25.

HEyer, W. R., DonNeLLY, M. A., MCDiarMiD, R. W., HAYEK, L. C. & Foster, M. S. (ed.), 1994. -

Measuring and monitoring biological div standard methods for amphibians. Washington,

Smithsonian Institution Press: 1-364.

Hossack, B. R. & YALE, K. J., 2002. - Geographic distribution. Pseudacris maculata. Herpetological

Review, 33: 63.

KNaPP, R. A. & MATTHEWS, K. R., 2000. - Nonnative fish introductions and the decline of the mountain

yellow-legged frog from within protected areas. Conservation Biology, 14: 428-438.

KocH, E. D. & PETERSO) R., 1995. — Amphibians and reptiles of Yellowstone and Grand Teton National

Parks. Salt Lake City, University of Utah Press: 1-188.

KosT, J. R. & KELLY, G. G., 2001. - Watershed delineation using the National Elevation dataset and

semiautomated techniques. Proceedings, 2001 ESRI International User Conference, Redlands,

California, Environmental Systems Research Institute, Inc. [Available at http///gis.esri.com/

library/userconf/proc0l/professional/papers/papH21/p421_ hum]

MaCKenzir, D. L. & BAILEY, L. L the fit of site occupancy models. Journal of

Agricultural, Biological, and E vironmental atisties, 9: 300-318.

ssessment of amphibian populations in

Source : MNHN, Paris

94 ALYTES 22 (3-4)

MACKENZIE, D. I, NicHoLs, J. D., HINES, J. E., KNUTSON, M. G. & FRANKLIN, A. B., 2003. - Estimating

site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology.

84: 2200-2207.

MACKENZIE, D. I., NiCHOLs, J. D., LACHMAN, G. B., DROFGE, S., ROYLE, J. A. & LANGTIMM, C. A.,

2002. - Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83:

2248-2255.

L, L. F, 1997. - Herpetofauna of Glacier National Park. Northwestern Naturalist, 78: 17-33.

K. R., POPE, K. L., PREISLER, H. K. & KNaPp, R. A. 2001. - Effects of nonnative trout on

Pacific treefrogs (Hyla regilla) in the Sierra Nevada. Copeia, 2001: 1130-1137.

MaxeLL, B. A. WERNER, J. K., HENDRICKS, P. & FLATH D. L., 2003. - Herpetology in Montana.

Northwest Fauna, 5: 1-135.

Murs, E., CoRN, P.S., PESSIER, A. P. & GREEN, D. E., 2003. - Evidence for disease-related amphibian

decline in Colorado. Biological Conservation, 110: 357-365.

PATLA, D. A. & PETERSON, C. R., 2001a. — Status and trends of amphibian populations in the Greater

Yellowstone Ecosystem. Pocatello, Idaho State University Division of Biological Sciences, Pro-

gress Report.

un 2001. — Summary of GYE-ARMI Activities. Pocatello, Idaho State University Division of

Biological Sciences, Progress Report.

PEer, R. K.., 1988. — Forests of the Rocky Mountains. /n: M. G. BARBOUR & W. D. BILLINGS (ed.), North

American terrestrial vegetation, Cambridge, Cambridge University Press: 63-101.

RITTMANN, S., MUTHS, E. & GREEN, D. E., 2003. - Pseudacris triseriata (western chorus frog) and Rana

svlvatica (wood frog). Chytridiomycosis. Herpetological Review 34: 53.

STOHLGREN, T. J., CHASE, T. N., PIELKE, R. A., SR. KITTEL, T. G. F. & BARON, JS., 1998. - Evidence that

local land use practices influence regional climate, vegetation, and stream flow patterns in adjacent

natural areas. Global Change Biology, 4: 493-504.

VAN KIRK, R., BENJAMIN, L. & PATLA, D. A., 2000. — Riparian area assessment and amphibian status in the

watersheds of the Greater Yellowstone Ecosystem. Bozeman, Montana, Greater Yellowstone

Coalition, Final report.

WoLrr, À. P., BARON, J. $. & CORNETT, R. J., 2001. - Anthropogenic nitrogen deposition induces rapid

ecological changes in alpine lakes of the Colorado Front Range (USA). Journal of Paleolimnology,

25: 1-7.

Corresponding editor: C. Kenneth Dop», Jr.

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 95-108. 95

Evidence of decline for Bufo boreas

and Rana luteiventris

in and around the northern Great Basin,

western USA

Wendy H. WENTE ! , Michael J. ADAMS & C. A. PEARL

st and Rangeland Ecosystem Science Center

3200 SW Jefferson Way, Corvallis, Oregon 97219, USA

Revisiting historical sites for patch-associated fauna can provide a

relatively fast assessment of species status in a region. Between 2000 and

2003 we repeatedly surveyed historically documented sites for two amphib-

ian species of concern in the northern Great Basin: the western toad (Bufo

boreas) and the Columbia spotted frog (Rana luteiventris). We estil

that B. boreas occupies 49.5 % (34.1-65.0 %) of 34 Historical

that R. luteiventris occupies 52.9 % (43.5-62.3 %) of its 30

sites. B. boreas was more likely to be detected at Historical sites that

were human-altered whereas R. luteiventris was more likely to be detected

at sites that had deeper water than other B.

whereas R. luteiventris was detected at six of 16 (37.5 %) Proxi

B. boreas was detected at one of 187 Randomly selected sites, and

R. luteiventris at three sites. Given that the number of Historical

available for resurveys was small, our results should be interpreted

caution. Moreover, a species can shift its distribution away from Historical

sites due to habitat succession or metapopulation dynamics without neces-

sarily declining. We suggest that there is sufficient evidence to conclude

that B. boreas and R. luteiventris have declined in and around the

northern Great Basin. Comparatively high occupancy of Proximal Sites by

R. luteiventris suggests this species may not have declined as much as has

B. boreas in our study region.

INTRODUCTION

Declines in amphibian populations are a global problem and their causes are often

complex (ADAMS, 1999; KiEsECKER et al., 2001). An essential step in understanding the causes

of decline and, perhaps, slowing or halting this process is to document the regional extent of

declines. Historical site records, such as those stored in museums or documented in the

published literature, are a valuable resource for gauging changes in occupancy patterns of

fauna in general (REZNICK et al., 1994) and have often been used to assess the status of

1. Corresponding author: <wwente@hotmail.com>

Source : MNHN, Paris

96 ALYTES 22 (3-4)

amphibians (Corx et al., 1989; Ross et al., 1995; DrosT & FELLERS, 1996; FISHER & SHAFFER,

1996; Davipson et al., 2001; SkELLY et al., 2002).

The Pacific Northwest region of the United States has been a focal area of research on

amphibian decline; however, broad-scale studies of amphibian species status remain rare in

the region. The scope of most research completed in the region has focused on suspected

causal agents such as UVB, disease, and introduced species (e.g. BLAUSTEIN et al., 1994a-b;

ADAMS et al., 2001; PEARL et al., 2004). Declines of several anurans in the west have been

attributed to changes in land use patterns, specifically, habitat destruction and the use of

agrochemicals (DAvViDsoN et al., 2001; Davipson et al., 2002). The western toad (Bufo boreas)

has experienced dramatic declines in other parts of its range (CORN et al., 1989; Ross et al.,

1995; Drosr & FELLERS, 1996; VERTUCCI & CORN, 1996; FISHER & SHAFFER, 1996), but its

status is unknown in the northern Great Basin, an arid region in the intermountain northwest

of the United States with large public land holdings. Likewise, the Columbia spotted frog

(Rana luteiventris) is thought to be declining in portions of its range (PATLA, 1997; REASER,

1997) and is a Candidate Species under the US Endangered Species Act (ANONYMOUS, 1997).

To determine the status of these amphibians in and around the northern Great Basin, we

documented the occupancy of three types of sites by B. boreas and R. luteiventris: historically

occupied sites (Historical sites), sites located near the Historical sites (Proximal sites), and

randomly selected sites from across the region (Random sites). We also analyzed how habitat

characteristics and certain potential stressors affect amphibian occurrence at Historical

sites.

MATERIALS AND METHODS

We developed a list of Historical sites from museum records, published accounts, and

field notes of local biologists. Records ranged in date from 1939 to 1999 but any record with

an observation prior to the years of this study could be included as a Historical site. Most

Historical sites we selected (r = 34 for B. horeas; n = 18 for R. luteiventris) were located in

Oregon east of the Cascade Range but we also included a small number of R. luteiventris

Historical sites (n = 12) in northeastern Nevada. The overall study area including both

Oregon and Nevada ranged from 45°57°35"N, 117°28°27" W in the north to 40°19°16"N,

116°0507"W in the south, and from 42°41°14"N, 117°01°29° W in the east to 43°54317N,

121°45°37" W in the west. Sites ranged in elevation from 406 to 2,256 meters. We gave priority

to Historical sites that had records of breeding (i.e., eggs, larvae or newly metamorphosed

individuals) but we also included some adult-only sites to increase our sample size. Aside from

status at known Historical locations, to determine how rare each specit in the Upper Great

Basin habitat of southeastern Oregon, we also surveyed sites in the v ty of the Historical

sites (usually within 25 km). When possible, these Proximal sites were located within the same

watershed or were physically similar to the Historical sites. Finally, occurrence data from 187

Randomly selected sites in the southeastern quarter of Oregon were also included in the study.

Random sites were drawn using a spatially balanced design (general randomized stratified

design with unequal probability sampling; S ss & O: 2004) that provided a widely

spaced random selection of fifth field hydrologic units (HUCS). Within the selected HUCSs, we

Source : MNHN, Paris

WENTE et al. 97

surveyed all accessible ponds and streams located on US Department of the Interior lands.

We visited sites indicated on GIS maps developed by the US Bureau of Land Management

(BLM) and 7.5 minute US Geological Survey topographie maps as well as sites incidentally

encountered in the field (less than 5 % of total number visited).

The region covered by our surveys roughly coincides with, but in some cases exceeds, the

northern Great Basin. Some Oregon Historical and Proximal sites were located on the east

slope of the Cascades and in the Blue Mountains ecoregion around the margins of the

northern Great Basin. Our surveys focused on lands managed by the BLM and therefore

excluded higher elevation lands managed by the US Forest Service (USFS) in northern and

central Nevada where REASER (1997) has previously documented a decline of R. luteiventris.

In 2000-2003, we surveyed sites between April and September when amphibians are most

likely to be active. All sites in Oregon (7 = 52) were surveyed in more than one year and a few

(1 = 14) were surveyed more than once during a single field season. Sites in Nevada (7 = 12)

were surveyed once. We excluded data from sites that were dry. We used a standard visual

encounter survey method including wading and dipnetting to detect amphibians (OLSON et

al, 1997). At each site, we recorded all amphibian taxa and life history stages as well as 10 site

and habitat characteristics. Our protocol called for two workers to slowly search all habitats

within one meter of standing water. The workers turned any cover objects by hand and used

a long-handled dip net to sweep through open water, aquatic vegetation, or along the

substrate. At larger lentic sites only a partial survey was completed. If the historical record

included specific locality details about a large site, that portion was searched. Otherwise, we

searched areas with appropriate pool and vegetative cover characteristic for amphibian

habitat (expert opinion).

STATISTICAL ANALYSIS

Because detection of amphibians at ponds is expected to be less than 100 % (OLSON et al.,

1997), we used the program PRESENCE (MACKENZIE et al., 2002) to estimate the proportion

of Historical sites that are currently occupied in addition to reporting the proportion of sites

where we detected the species (the naïve occupancy rate). PRESENCE is designed to estimate

the proportion of sites occupied (4°) by a species of interest while incorporating an estimate of

species detectability (p). There are two possible explanations of a non-detection result at a

surveyed site: either the species was not present at the site or we failed to detect it even though

it was present. By visiting a site multiple times, we were able to estimate species detectability

and produce an unbiased estimate of occupancy. We treated all searches as if they were

conducted during a single season. This violates an assumption of PRESENCE that occu-

pancy does not change during the study. This violation may have caused us to underestimate

detectability by treating some true absences as false absences. Underestimating detectability

would cause us to overestimate occupancy so any declines may be more severe than our

estimates indicate.

PRESENCE also allows for the inclusion of covariate data. We constructed models to

estimate occupancy using the response variable Detection along with one survey covariate

and nine site covariates that were selected from the larger suite of characteristics collected

Source : MNHN, Paris

98 ALYTES 22 (3-4)

Table 1. — Variables used in models to describe detection data for Bufo boreas and Rana

luteiventris. Class refers to the variable type (response variable, survey or site covariate) as

well as the category it represents (habitat or stressor).

Variable Class Description

Detection response |binomial: presence or non-detection of species at site

Season survey [binomial: indicates if detection in spring (April-May) differs from

summer (June-September).

Record Type site |binomial: historical evidence of breeding or not

Elevation site/habitat _ |continuous: elevation of site in meters

Cover site/habitat continuous: per cent emergent vegetation at site averaged across visits

Shallows site/habitat continuous: per cent shallows (< 0.5m) at site averaged across visits

Wetland Type site/habitat _|binomial: site was pond/lake/spring or river/stream

Origin site/habitat _|binomial: naturally occurring or human-altered/human-made

Grazing site/stressor | binomial: evidence of cattle grazing at site or not

Distance from Road | site/stressor |binomial: site was less than or greater than 100 m from nearest road

Fish site/stressor |binomial: fish of any species detected at site or not

during each Historical site survey. The response variable in all cases was coded 1 if the species

was detected during the given survey and 0 if it was not. We limited the suite of covariates to

10 due to our small sample sizes (tab. 1). The survey covariate Season indicated whether the

site was surveyed in spring (April-May) or summer (June-September). This was included

because we suspected that detectability was lower in the summer than in the spring. One site

characteristic, Record Type, indicated whether the site historically had a record of breeding or

non-breeding. Besides the survey covariate and Record Type, we included variables that, as

indicated by previous work (HADDEN & WESTBROOKE, 1996; AKER, 1998; Vos & CHARDON,

1998; SEMLITSCH, 2000; PiLLIOD & PETERSON, 2001), might be related to amphibian decline:

Grazing (whether or not the site was used by cattle), Distance to Road (whether or not the site

was less than 100 m from the nearest road) and Fish presence (whether or not fish were

detected). We also included variables that are often important to amphibian biology: Eleva-

tion, Cover (percentage of the site covered with emergent vegetation), Shallows (percentage

of the site with water less than 0.5 m in depth), Origin (whether the site appeared to be natural

or if it was human-altered/human-made) and Wetland Type (lentic or lotic). We screened all

10 of the predictor variables for collinearity. If a pair of variables had a Pearson correlation

above 0.70, one must be eliminated (NasH & BRADFORD, 2001); however, none of our variable

pairs broke this rule. We also screened for multiple collinearity by building a simple linear

regression model with detection as the response variable. In the collinearity diagnostics, a

variance inflation factor above 10 for any of the covariates would have indicated a strong

linear relationship with one or more of the other covariates and that variable would be

dropped (SPSS 9.0, SPSS Inc., Chicago, IL, USA). None of the covariates needed to be

dropped. We then built a model predicting occupancy for each species using PRESENCE to

conduct a forward stepwise procedure. At each step of the stepwise selection process, we

added the variable that resulted in the best model based on the small sample version of

Source : MNHN, Paris

WENTE et al. 99

Akaike’s Information Criterion (AIC.). AIC, is a measure of the information content of a

model relative to the number of parameters in the model (BURNHAM & ANDERSON, 2002). We

stopped adding covariates when none decreased the AIC, score by greater than 2.0 following

BURNHAM & ANDERSON (2002). Our estimate of site occupancy was based on the model that

resulted from this procedure. We report the estimated proportion of sites occupied (y) and a

95 % confidence interval.

As we progressed through the stepwise regression, each model was tested for

goodness-of-fit using the Pearson chi-square test statistic and a parametric bootstrap

method available in PRESENCE. If a model had failed the goodness-of-fit test (none failed)

we would have adjusted our model selection procedures following MACKENZIE & BAILEY (in

press).

Finally, we present raw detection data for Proximal and Random sites to provide a more

complete depiction of B. boreas and R. luteiventris status in the region.

RESULTS

Bufo boreas

We detected B. boreas at 11 of 34 Historical sites yielding a naïve occupancy rate of

32.4 % (fig. 1-2). Using PRESENCE, we estimated the proportion of sites occupied to be

49.5 % (34.1-65.0 %) and the detection probability for a single survey to be 0.3204. Nineteen

of our 34 Historical records were pre-1990; we found B. boreas at four of these sites (21.1 %).

At 15 more recently recorded sites (1990-1999), occupancy was higher (46.7 %). In eastern

Oregon, we detected B. horeas at three of 41 (7.3 %) Proximal sites and at one of 187 Random

sites (0.5 %).

Origin was the only covariate selected using stepwise regression (tab. 2). The odds of B.

boreas occupying an altered site were 9.2 times greater than for naturally occurring sites.

Season did not affect detection probability for this species.

Rana luteiventris

We detected R. luteiventris at 12 of 18 (66.7 %) Historical sites in Oregon (fig. 3-4). In

Nevada, we found R. luteiventris at only one of 12 (8 %) Historical sites that had water (six

sites were dry) but sites in Nevada were only visited once. Overall, we detected À luteiventris

at 43.3 % of Historical sites. Using PRESENCE, we estimated the proportion of sites

occupied to be 52.9 % (43.5-62.3 %) and the detection probability for a single survey to be

0.73. Because sites in Nevada were only visited once, we had to assume that detection

probabilities were similar in the two states. We found R. lureiventris at four of the eight (50 %)

Historical sites that were documented before 1990. At 22 more recently recorded sites

(1990-1999), the rate of detection was slightly higher (59.1 %). We detected À luteiventris at

six of 16 Proximal sites (37.5 %) and three of 187 Random sites (1.6 %).

Source : MNHN, Paris

100 ALYTES 22 (3-4)

60 O0 60 Kilometers

(ER Ga pee

Fig. 1. - Historical sites surveyed for Bufo boreas (n = 34) in 2000-2003. Circles indicate Historical sites.

Triangles represent Proximal Sites (filled: present; empty: B. horeas not detected).

100

$ 70

8 60

8 50 43.8

£ 40] 34

5 5 22.2

56 7.3

0.5

Bee HE 5? 3 8

S É $ Ex: Do — DE

BE ÊSS je €S 6

Tue +5 5 5 ÊS $=

< ” » $ S &

«if à Œ

Record Type

Fig. 2. — Occupancy data for Bufo boreas at sites in eastern Oregon. In addition to Proximal and Random

sites, the data are separated into 3 categories based on evidence of breeding from the Historical

record.

Source : MNHN, Paris

WENTE et al. 101

Table 2. - Models evaluated for best fit to the Bufo boreas detection data using the single season

model of PRESENCE. #: number of parameters; n: sample size (number of sites). (.)

indicates the null condition where no covariates are included. Site covariates are associated

with the occupancy of B. boreas at the site (W). Survey covariates are associated with B.

boreas detectability (p). w-hat is the estimate of site occupancy for B. horeas under the

tested model along with standard error. The best model for each step is indicated with bold

type.

Model kln AIC AICe w-hat (SE)

Step 1

vOPO 2 | 34 | 90.446 | 908337 | 04672 (0.1434)

VO) p(Season) 3 | 34 | 917882 | 925882 | 04659 (0.1436)

34 90.984 91.784 0.4606(0.1406)

34 92.3288 93.1288 04714 (0.1473)

34 91.5037 92.3037 0.4564 (0.1349)

91.3205 92.1205 04711 (01454)

34 92.2078 93.0078 0.4615 (0.1399)

34 91.4929 92.2929 0.4760(0.1474)

34 92.411 93.211 0.4690 (0.1448)

88.3268 89.1268 0.4953 (0.1542)

92.4284 049698 (0.1460)

y (Record Type) p(.)

Y (Elevation) p(.)

Y (Shallows) p(.)

4 (Cover) p(.)

y (Wetland Type) p(.)

V (Grazing) p()

V (Distance from Road) p{.)

Y (Origin) p(.)

Y (Fish) p(.)

w w & & & W w

Ê£

€

In the stepwise regression analysis, Shallows and Season were the covariates identified as

components of the best model describing site occupancy by À. luteiventris (tab. 3). For each

increase in Shallows of 10 %, the odds of R. luteiventris occurring at a site decreased by a

factor of 1.8. The detection probability of R. luteiventris was higher in spring (April and

May).

DISCUSSION

Large-scale, multi-year studies of amphibian status are time and resource intensive. As a

result, few have been completed (Corn et al., 1989; FISHER & SHAFFER, 1996; DAVIDSON et al.,

2001; SKELLY et al., 2002). Our study was designed to offer a first ssment of species status

for B. boreas and R. luteiventris at sites spread over a large geographic area (eastern Oregon

and northeastern Nevada). We found evidence that both B. boreas and R. luteiventris have

declined in eastern Oregon. We failed to detect both species at a large proportion of older

Historical records, which might be expected due to a greater likelihood of habitat change at a

particular site. We also failed to detect both species at many of the Historical sites known to

be occupied within the last 10 years.

Source : MNHN, Paris

102 ALYTES 22 (3-4)

Table 3. - Models evaluated for best fit to the Rana luteiventris detection data using the single

season model of PRESENCE. 4: number of parameters; 7: sample size (number of sites). (.)

indicates the null condition where no covariates are included. Site covariates are associated

with the occupancy of R. luteiventris at the site (y). Survey covariates are associated with À.

luteiventris detectability (p). y-hat is the estimate of site occupancy for R. luteiventris under

the tested model along with standard error. The best model for each step is indicated with

bold type.

Model kon AIC AICe V-hat (SE)

Step

vOPO 30 | 881578 | 8860224 | 0.6254(0.1088)

y (Record Type) p(.) 30 89.8828 90.80588 | 0.5107(0.1100)

V (Elevation) p(.) 30 | 88.546 | 89.51768 | 0.5094 (0.1069)

(Shallows) p(.) 3 | 30 | 846093 | 85.53238 | 0.5186(0.0935)

y (Cover) p{.) 3 | 30 | 901527 | 9107578 | 0.5229(0.1114)

Y() p(Season) 3 30 83.9262 84.84928 0.5444 (0.1147)

W (Wetland Type) p(.) 30 | 882465 | 8916958 | 0.5167(0.1052)

V (Grazing) p) 30 | 890036 | 8992668 | 0.5292 (0.108)

Y (Distance from Road) p() 3 | 30 | 859227 | 86.84578 | 0.5049 (0.0975)

v (Origin) p(.) 3 | 30 89.7491 90.67218 | 0.5262 (0.114)

V (Fish) p() 3 | 30 | 878085 | 88.82158 | 0.4967(0.1012)

Step2

Y (Shallows) p(Season)

y (Record Type) p(Season)

30 80.5017 82.1017 0.5293 (0.0941)

30 85.8223 87.4223 0.5332 (0.1167)

30 84.8601 86.4601 0.5219 (0.1102)

30 85.9163 87.5163 0.5444 (0.1147)

84.0302 85.6302 0.5364 (0.1078)

30 84.7084 86.3084 0.5506 (0.1126)

30 81.8528 83 8 0.5205 (0.0996)

30 85.4464 87.0464 0.5493 (0.1146)

30 84.1889 85.7889 0.5075 (0.1045)

W (Elevation) p(Season)

W (Cover) p(Season)

(Wetland Type) p(S

V (Grazing) p(Season)

‘ason)

W (Distance from Road) p(Season)

BRSEERESS SES

Although these results sem to demonstrate declines, several factors complicate their

interpretation. First, the period of time when our visits were made (2000-2003) coincided with

a drought that progressed in severity from moderate to extreme across the region (National

Oceanic and Atmospherie Administration National Drought Mitigation Center, unpublished

data). This drought was more extreme in the northern Great Basin region of our study and we

noted a lower B. boreas naïve occupancy rate in this region (25 %) than at sites overall (32 %).

Individuals in breeding populations of B. boreas occasionally skip a year or two of breeding,

possibly due to a decrease in foraging opportunities that impact egg production (OLSON,

1991). In several amphibian species, skipping of the breeding season has been attributed to

Source : MNHN, Paris

WENTE et al. 103

100 0 100 200 Kilometers

mea car

Fig. 3. - Historical sites surveyed for Rana luteiventris (n = 30) in 2000-2003. Circles indicate Historical

sites. Triangles represent Proximal Sites (filled: present; empty: R. luteiventris not detected).

61.5

43.3

40 35.3 ne

Percent sites occupied

(617)

(13/30)

All Historical

Records

Historical

larvae/eggs

breeding (8/13) À

Random Sites

(8/187)

Historical non-

Proximal Sites

(6/16)

Record Type |

Fig. 4. - Occupancy data for Rana luteiventris at sites in eastern Oregon and northeastern Nevada. The

larvac/eggs category represents Historical sites where evidence of breeding was found. The

non-breeding category represents Historical sites where only adults or juveniles were recorded.

Source : MNHN, Paris

104 ALYTES 22 (3-4)

drought (PECHMANN et al., 1991; BABBITT & TANNER, 2000; TRENHAM et al., 2000; MARSH &

TRENHAM, 2001). However, the fact that most of our sites were visited for three seasons

suggests that B. boreas might be permanently absent. Because we excluded Historical sites

that were dry from our analysis, our evidence of decline is conservative. Including those sites

would have caused our results to indicate a more severe decline.

A second factor that could complicate interpretation of these data is the variation in the

type of historical record that we used. Although many of our sites had multiple B. boreas

records, half were based on adult and/or juvenile records rather than direct evidence of

breeding. Record Type was not selected in the final model for either species using a stepwise

process so we conclude that the stage of the specimen was not strongly related to the odds of

current occurrence.

A final complicating factor is that historical site revisitation studies are inherently biased

toward detecting declines rather than expansions (SKELLY et al., 2002). In some cases, a

population lost from a historical site may be replaced by a new population due to metapopu-

lation dynamics, successional changes in the suitability of other sites on the landscape, or

other processes (SKELLY et al., 1999). We have no way to estimate how often population

replacement might have occurred. However, available information suggests that both species

have high site fidelity (OLSON, 1991; REASER, 1996; ENGLE, 2001; PizLiop et al., 2002).

Observations of colonization are rare but such events may be episodic. For example, following

a large wildfire in Glacier National Park, small numbers of B. boreas colonized twelve

previously unused breeding sites within two years (Hossack & Corn, unpublished data) and

both species can colonize newly constructed ponds (C. Pearl, unpublished data). B. boreas

extensively colonized new ponds on Mount St. Helens following the 1980 eruption (LOVETT,

2000). The fact that a high proportion of R. luteiventris Proximal sites were occupied might

suggest that declines in this species are less severe than in B. boreas despite similar estimates of

occupancy for Historical sites.

Whereas our analysis indicates that some B. boreas population losses have occurred,

these declines do not appear to be as severe as declines described in other regions in the

western USA. For example, CoRN et al. (1989) documented B. boreas at 10 of 59 (17 %)

historical sites in Wyoming and Colorado. DRosr & FELLERS (1996) found B. boreas at only

one of eight (12.5 %) historical sites in and around Yosemite National Park, California. These

results can be compared with our naïve estimate of 32 % occupancy based on multiple surveys

over multiple years. CorN et al. (1989) also visited most sites over multiple years but DROST &

FELLERS (1996) only visited sites during one year. The higher occupancy rate in our study may

indicate that losses around the Great Basin have not been as severe as those in other parts of

B. boreas’ range. Comparable data on R. luteiventris status are available from at least three

sources. PATLA (1997) reported a dramatic decline in a single large population at Yellowstone

National Park, Wyoming, that occurred between the 1950s and 1990s. In southwestern Idaho,

R. luteiventris was missing from five of 16 (31 %) known sites in 2001 (ENGLE, 2002). REASER

(1997) documented R. luteiventris at only 50 % of the historical sites she surveyed in Nevada.

The low R. luteiventris occupancy rate (1 of 12 wet sites) at our BLM sites in northeastern

Nevada suggests declines might be more pronounced in lower-elevation habitats than in the

higher elevation sites surveyed by REASER (1997) in Nevada. We recommend return surveys at

Nevada sites we were only able to survey once.

Source : MNHN, Paris

WENTE et al. 105

Potential causes of decline of B. boreas in our region could be related to natural processes

of succession or human induced habitat change. Grazing by cattle is one of the most extensive

uses of public lands in the western United States (BELSKY et al., 1999) and many reservoirs

were built on public lands as an improvement specifically for grazing in the northern Great

Basin. Amphibians do use these man-made reservoirs but the specific amphibian habitat use

patterns across the region are not known prior to the building of the reservoirs. The fact that

our final model showed a tendency for B. boreas to occupy human-altered sites at least

indicates that cattle reservoirs and other human-altered sites provide some suitable habitat for

the persistence of the species.

Rana luteiventris may be experiencing a less severe decline than B. boreas. We base this

conclusion on the fact that this species was present at 37.5 % of the Proximal sites. This is a

high rate of occupancy for sites that were only selected for surveys based on their proximity

and similarity to Historical sites. Declines appeared more severe in Nevada but the lack of

repeat surveys and Proximal sites limits our ability to interpret these data. The significance of

the covariate Shallows is likely related to R. luteiventris’ need for permanent water. The

inclusion of the covariate Season can probably be attributed to the Nevada data because these

sites were all visited in the summer and we detected R. luteiventris at only one site. R.

luteiventris in nearby southwestern Idaho populations has been infected with a chytrid fungus

and other disease organisms in recent years that might be in part responsible for population

declines observed at sites in that state (DREW, 2003). In addition, non-native fish have a strong

negative impact on site occupancy by R. luteiventris in eastern Washington (AKER, 1998) and

Idaho (PiLL10D & PETERSON, 2001).

In summary, our study provides evidence of at least a moderate decline in B. boreas and

R. luteiventris in and around the northern Great Basin. Though none of the variables we

examined clearly explained the potential declines of these species the overall patterns we

observed warrant further research.

RÉSUMÉ

Le réexamen de sites dont la faune a été étudiée dans le passé (sites historiques) peut

fournir relativement rapidement des informations sur le statut des populations animales dans

une région. Entre 2000 et 2003 nous avons à plusieurs reprises étudié les populations

d'amphibiens de sites historiques. Les deux espèces étudiées, Bufo boreas et Rana luteiventris,

sont menacées dans la région du nord du Grand Bassin. Nous estimons que B. boreas occupe

49.5 % (34.1-65.0 %) de 34 sites historiques et R. luteiventris 52.9 % (43.5-62.3 %) de 30 sites

historiques. B. boreas s’est avéré plus présent dans les sites historiques altérés par l'homme,

tandis que R. luteiventris était plus abondante dans les sites à eau profonde que dans les autres

sites étudiés. B. boreas a été trouvé dans trois sites voisins des sites historiques sur 41 (7.3 %);

tandis que R. luteiventris a été trouvée dans six sites de ce type sur 16 (37.5 %). B. boreas a été

trouvé dans un site sur 187 sélectionnés au hasard; et R. luteiventris dans trois d'entre eux. Nos

résultats doivent être interprétés avec prudence car le nombre de sites historiques disponibles

pour réexamen était faible. De plus, la répartition d’une espèce peut s'éloigner des sites

historiques en raison de changements dans les habitats ou de la dynamique de ses métapopu-

Source : MNHN, Paris

106 ALYTES 22 (3-4)

lations sans pour autant décliner. Nos résultats suggèrent que B. boreas et R. luteiventris ont

décliné dans la partie nord du Grand Bassin et alentours. La relative abondance de R.

luteiventris dans les sites voisins des sites historiques semble indiquer que cette espèce pourrait

ne pas avoir autant décliné que B. boreas dans la région d’étude.

ACKNOWLEDGMENTS

This study was funded by the US Geological Survey Amphibian Research and Monitoring Initiative

(ARMD) and the US Bureau of Land Management. D. Jarkowsky did much of the work assembling

historical records for Nevada. We are also grateful to C. Tait, M. Obradovich, R. Demmer and R. MeNatt

of the Bureau of Land Management, M. Laws and M. Bray of the US Fish and Wildlife Service, W. Amy

of the US Forest Service, A. Cook of the Nevada Department of Wildlife as well as J. Bowerman for their

help with the project. Other sources of historical site information included the California Academy of

Science Collection, Oregon State University, University of Michigan Museum of Zoology, United States

National Museum, University of California Davis, Los Angeles County Museum Herpetological Col-

lection, Marjorie Barrick Museum of Natural History at the University of Nevada Las Vegas, and the

Berkeley Museum of Vertebrate Zoology. We thank À. Olsen for his help with statistical design and L.

Bailey for her advice on the statistical analysis. Work was completed under Oregon Scientific Taking

Permit # 116-03, Nevada Scientific Collection Permit # 823390 and Oregon State University Institutional

Animal Care and Use Permit LAR-ID#2459.

LITERATURE CITED

ANONYMOUS, 1997. — Endangered and Threatened Species: Review of Plant and Animal Taxa: Proposed

Rule. Federal Register. United States Fish & Wildlife Service Department of the Interior, 62 (182):

50 CFR Part 17.

ADAMS, M. J., 1999. — Correlated factors in amphibian decline: Exotic species and habitat change in

western Washington. Z Wild. Man., 63: 1162-1171.

ADAMS, M. J., SCHINDLER, D. E. & BuRY, R. B., 2001. - Association of amphibians with attenuation of

ultraviolet-b radiation in montane ponds. Oecologia, 128: 519-525.

Aker, L. M. 1998. — The effect of native and non-native fish on amphibian populations in northeast

Washington. MS Thesis, Cheney, Eastern Washington University: 1-21.

Bagrrr, K. J. & TANNER, G. W., 2000. — Use of temporary wetlands by anurans in a hydrologically

modified landscape. Werlani 20: 313-322.

BELSKY, A. J., MATZKE, A. & USELMAN, S., 1999. — Survey of livestock influences on stream and riparian

ecosystems in the western United States. J. Soil Water Conserv., 54: 419-431.

BLAUSTEIN, A. R., HOFFMAN, P., HOKIT, D., KIESECKER, J., WALLS, S. & HAYS, 994a.— UV repair and

ance to solar UV-B in amphibian eggs: a link to population declines? Proc. nain. Acad. Sci.,

1791-1795.

BLAUSTEIN, A. R., HOkiT, D. G. & O'HaRA, R. K., 1994b. — Pathogenic fungus contributes to amphibian

losses in the Pacific norhtwest. Biol. Conserv., 67: 251-254.

BURNHAM, K. P. & ANDERSON, D. R., 2002. - Model selection and multimodel inference: a practical

information-theoretic approach. Second edition. New York, Springer.

Corn, P. $., SrouzenBurG, W. & BURY, R. B., 1989. — Acid precipitation studies in Colorado and

Wyoming: interim report of surveys of montane amphibians and water chemistry. US Fish & Wildlife

Fort Collir olorado: 1-57.

SON, C., SHAFFER, H. B. & JENNINGS, M. R.

climate, UV-B, habitat, and pesticides hypothes

001. - Declines of the California red-legged frog:

. Ecol. app, 1: 464-479.

Source : MNHN, Paris

WENTE et al. 107

en 2002. - Spatial tests of the pesticide drift, habitat destruction, UV-B, and climate-change hypotheses

for California amphibian declines. Conserv. Biol., 16: 1588-1601.

Drew, M. L., 2003. - Laboratory report. Spotted frogs in southwester Idaho. Disease investigation and

water quality analysis. Wildlife Health Laboratory, Idaho Department of Fish & Game, Idaho

State Department of Agriculture.

Drosr, C. A. & FELLERS, G. M., 1996. - Collapse of a regional frog fauna in the Yosemite area of the

California Sierra Nevada, USA. Conserv. Biol., 10:414-425.

ENGLE, J. C., 2001. - Population biology and natural history of Columbia spotted frogs (Rana luteiventris)

inthe Owyhee Uplands of southwest Idaho: implications for monitoring and management. MS Thesis,

Boise, Boise State University: 1-66.

ss 2002. - Columbia spotted frog Great Basin population (Owyhee subpopulation) long-term monitoring

plan. Year 2001 results. Ydaho Department of Fish & Game Report, Boise, Idaho: 1-64.

FisHER, R. N. & SHAFFER, H. B., 1996. - The decline of amphibians in California’s great central valley.

Conserv. Biol., 10: 1387-1397.

HADDEN, S. A. & WESTBROOKE, M. E., 1996. — Habitat relationships of the herpetofauna of remnant

buloke woodlands of the Wimmera Plains, Victoria. Wild. Res., 23: 363-372.

KIESECKER, J. M., BLAUSTEIN, A. R. & Mier, C. L., 2001. — Potential mechanisms underlying the

displacement of native red-legged frogs by introduced bullfrogs. Ecology, 82: 1964-1970.

Loverr, R. A., 2000. - Mount St. Helens, revisited. Science, 288: 1578-1579.

MACKENZIE, D. L., NICHOLS, J. D., LACHMAN, G. B., DROFGE, S., ROYLE, A. & LANGTIMM., C. A. 2002.

— Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83:

2248-2255.

MACKENZIE, D. L. & BaILey, L. L., in press. - Assessing the fit of site occupancy models. J Agri. Biol.

Ecol. Stat.

Mars, D. M. & TRENHAM, P. C., 2001. - Metapopulation dynamics and amphibian conservation.

Conserv. Biol., 15: 40-49.

NASH, M. S. & BRADFORD, D. F., 2001. - Parametric and nonparametric logistic regressions for prediction

of presencelabsence of an amphibian. United States Environmental Protection Agency, EPA/600/R-

01/081, Washington, DC: 1-40.

OLsox, D. H., 1991. - Ecological susceptibility of amphibians to population declines. /n: Symposium on

Biodiversity of Northwestern California, Santa Rosa, CA: 55-62.

OLsoN, D. H., LEONARD, W. P. & BURY, R. B., 1997. — Sampling amphibians in lentic habitats. First

edition. Olympia, Washington, Society for Northwestern Vertebrate Biology: i-x + 1-134.

ParLA, D. A., 1997. - Changes in a population of spotted frogs in Yellowstone National Park between 1953

and 1995: the effects of habitat modification. MS Thesis, Pocatello, Idaho State University.

PEARL, C. A., ADAMS, M. J., BURY, R. B. & MCCREARY, B., 2004. - Asymmetrical effect of introduced

bullfrogs (Rana catesbeiana) on native ranid frogs in Oregon. Copeia, 2004: 11-20.

PECHMANN, J. H. K., SCOTT, D. E., SEMLrTSCH, R. D., CALDWELL, J. P., Virr L.J. & G1BBONS, J. W., 1991.

— Declining amphibian populations: the problem of separating human impacts from natural

fluctuations. Science, 253: 892-895.

Piu1op, D. S. & PETERSON, C. R.., 2001. - Local and landscape effects of introduced trout on amphibians

in historically fishless watersheds. Ecosystems, 4: 322-333.

Pizu1op, D. S., PETERSON, C. R. & RiTsON, P. L., 2002. - Seasonal migration of Columbia spotted frogs

(Rana luteiventris) among complementary resources in a high mountain basin. Can. J. Zool., 80:

1849-1862.

R, J. K., 1996. — Rana pretiosa (Spotted frog) vagility. Herp. Rev., 27: 196-197.

1997. — Amphibian declines: conservation science and adaptive management. PhD Thesis, Stanford.

REZNICK, D., BAXTER, R. J. & ENDLER, J., 1994. - Long-term studies of tropical stream fish communities.

The use of field notes and museum collections to reconstruct communities of the past. Am. Zool.,

34: 452-462.

Ross, D. A., ESQUE, T. C., FRIDELL, R. A. & HOVINGH, P., 1995. — Historical distribution, current status,

and a range extension of Bufo boreas in Utah. Herp. Rev., 26: 187-189.

SemLrTscH, R. D., 2000. - Principles for management of aquatic-breeding amphibians. / Wild. Man. 64:

615-631.

WERNER, E. E. & CORTWRIGHT, S. A., 1999. - Long-term distributional dynamics of à

n amphibian assemblage. Ecology, 80: 2326-2337.

Source : MNHN, Paris

108 ALYTES 22 (3-4)

SKELLY, D. K., YUREwWICZ, K. L., WERNER, E. E. & RELYEA, R. A., 2002. — Estimating decline and

distributional change in amphibians. Conserv. Biol., 17: 744-751.

STEVENS, D. L. & OLSEN, A. R., 2004. - Spatially balanced sampling of natural resources. J. am. Stat.

Assoc., 99: 262-278.

TRENHAM, P. C., SHAFFER, H. B., KOENIG, W. D. & STROMBERG., M. R., 2000. — Life history and

demographic variation in the California Tiger Salamander (Ambystoma californiense). Copeia,

2000: 365-377.

VERTUCCI, F. A. & CoRN, P.S., 1996. - Evaluation of episodic acidification and amphibian declines in the

Rocky Mountains. Ecol. appl., 6: 449-457.

Vos, C. C. & CHARDON, J. P., 1998. — Effects of habitat fragmentation and road density on the

distribution pattern of the moor frog Rana arvalis. J. appl. Ecol., 35: 44-56.

Corresponding editor: C. Kenneth Dop, Jr.

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 109-129. 109

Health evaluation of amphibians in and near

Rocky Mountain National Park

(Colorado, USA)

D. Earl GREEN* & Erin Muras**

* US Geological Survey, National Wildlife Health Center,

6006 Schroeder Road, Madison, Wisconsin 53711, USA

‘US Geological Survey, Fort Collins Science Center,

2150 Centre Avenue, Building C, Fort Collins, Colorado 80526, USA

We conducted a health survey of amphibians in and adjacent to Rocky

Mountain National Park (RMNP) to document current disease presence

inside RMNP and identify disease outside RMNP with the potential to spread

to the Park’s amphibians. Amphibians from five sites within RMNP and

seven sites within 60 km of Park boundaries were collected and examined.

Necropsies (n = 238), virus isolation, bacterial and fungal cultures, and

histological examinations were carried out on amphibian egg masses

(outside RMNP/within RMNP: 26/22), larvae (30/42), imagos (recently

metamorphosed individuals) (0/3) and adults (61/67) of five species.

Marked infections by a pathogenic chytrid fungus (chyridiomycosis), Batra-

chochytrium dendrobatidis, were detected in three species (Bufo boreas,

Pseudacris maculata and Rana sylvatica) from three of five sites within

RMNP and in one of three species (P. maculata) from three sites outside

RMNP. Of the fully metamorphosed individuals tested (B. boreas, P.

maculata and R. sylvatica), chytridiomycosis was found in 60 % (n = 3),

46 % (n = 37) and 54 % (n = 7), respectively. Chytridiomycosis was the

principal lethal pathogenic infectious disease detected in three amphibian

species within or adjacent to RMNP. Higher fungi were isolated from the

cloaca and skin of all five amphibian species. Watermolds (Oomycetes) were

isolated from amphibian eggs or skin of all five species. No evidence of

Ranavirus was found in cultures and histological examinations of 176 and

142 amphibians, respectively. Fifteen genera of bacteria were identified in

larval and just metamorphosed amphibians, and a potentially pathogenic

lungworm, Rhabdias sp, was identified in 61.1 % (n = 11) of B. woodhou-

sii outside RMNP, but in only 2 (15.4 %) R. sylvatica within the Park.

INTRODUCTION

Boreal toads (Bufo boreas) currently exist as remnant populations in Rocky Mountain

National Park (RMNP) (a roughly rectangular 107,625 hectares park in northern Colorado,

USA; elevation range: 2,440 to 4,345 m:; latitude and longitude at approximate center of park:

40°40°N, 105°60°W) where their historic range was once more extensive (CORN et al., 1997).

Recent precipitous declines in two of three populations of boreal toads within RMNP

Source : MNHN, Paris

110 ALYTES 22 (3-4)

(Murs et al., 2003) have put this toad in danger of local extinction. The third population

where toads were observed in the recent past, is now thought to be extirpated. Fortunately,

breeding toads have been observed at two new additional sites in the Park since the conclusion

of our study (Muths, unpublished data).

Boreal toads and northern leopard frogs (Rana pipiens) have declined severely through-

out the southern Rocky Mountain region in the last 20 years, and many populations have

been extirpated (CORN, 2000; Corn & FOGELMAN, 1984; Murs et al., 2003). Recent studies

have implicated infections by the pathogenic chytrid fungus, Batrachochytrium dendrobatidis,

in the decline of toad populations in RMNP (Murs et al., 2003), and by ranaviruses in mass

mortality events in tiger salamanders (4Ambystoma tigrinum) in the western United States

(JaNcovicx et al., 1997; DocxerTY et al., 2003). Basidiobolus ranarum, another fungus, was

implicated in the decline of Wyoming toads (Bufo baxteri) (TAYLOR et al., 1999a-b) but the

diagnosis has been revised to indicate that B. dendrobatidis was the pathogen (CAREY et al.,

2003).

The long-term goals of our research were to determine additional threats to boreal toads

within RMNP and to determine basic health parameters in amphibians that have lived

sympatrically with this species historically. Baseline biomedical information for most amphib-

ian populations is lacking. For example, the prevalence of bacterial pathogens and normal gut

and skin flora for most amphibian populations is unknown, although amphibians may harbor

Salmonella spp. and Leptospira-like and chlamydial organisms (TAYLOR et al., 2001; O’SHEA

et al., 1990; BERGER et al., 1999; R&ED et al., 2000). Ranavirus, a distinct genus of the family

Iridoviridae, has been identified as the causative agent in anuran and urodelan mortality

events in adjacent states (JANCOVICH et al., 1997; DocHerTY et al., 2003), but has not been

detected in cultures and tissue sections of amphibians within and adjacent to RMNP.

Our immediate objectives were threefold, namely: (1) to develop baseline biomedical

standards for amphibians in this area; (2) to determine pathogens present in amphibians in

RMNP and surrounding areas; and (3) to examine the potential for spread of diseases from

outside RMNP to amphibians in RMNP. Based on previous work (MUTHS et al., 2003;

RiTTMANN et al., 2003), we expected to find B. dendrobatidis and possibly ranavirus. No other

lethal amphibian diseases have been documented previously in RMNP.

MATERIALS AND METHODS

Within RMNP, all extant populations of boreal toads were sampled (n = 2 sites). Other

sites were selected by: (1) current presence of one or more of the three other species extant in

RMNP (HAMMERSON, 1999); (2) ease of access; and (3) spatial coverage of the Park (fig. 1,

tab.1). Sites outside the Park were selected by: (1) proximity to RMNP (within 60 km); (2)

nce of B. boreas (n = 1, Twin Lakes Reservoir); (3) presence of amphibians; and (4) ease

ss. We collected data on boreal toads, chorus frogs (Pseudacris maculata), tiger

salamanders and wood frogs (Rana sylvatica) in the Park, and boreal toads, Woodhouse’s

toads (Bufo woodhousii) and chorus frogs outside of the Park. Amphibians were collected

from June 2000 through September 2002

Source : MNHN, Paris

GREEN & MUTHS 111

* Collection Sites

Larimer |

Colorado

Fig. 1. - Location of Rocky Mountain National Park and surrounding federal and private lands.

FIELD COLLECTION

Adults, imagos (just metamorphosed specimens), larvae and portions of egg masses

(approximately 25 eggs per egg mass) were captured by hand or by dipnet. We used disposable

latex gloves to handle each animal. AI animals were held temporarily and mailed alive in

Separate containers (toads) or 2-8 animals per container (chorus frogs, wood frogs and tiger

salamanders) according to protocols of the United States Geological Survery, National

Wildlife Health Center (NWHOC) [http:/www.nwhe.usgs.gov/research/amph_dc/amph_sop.

htm]. Adult boreal toads were sampled non-lethally because they are an endangered species

in the State of Colorado and have undergone declines statewide (JUNGWIRTH, 2004). Boreal

toad populations in the Park are currently monitored using capture-recapture methods

Source : MNHN, Paris

112 ALYTES 22 (3-4)

Table 1. - Location of sites and distance from RMNP boundary. Negative distances indicate that

site is within RMNP. Easting and northing coordinates are North American Datum of 1927

(NAD27) of the Universal Transverse Mercator system, grid zone 13 (UTM13).

Kette Tam Larimer | 455090 us] 6 22 aa

Spruce Lake Larimer | 441689 mes] 10 28 #10

Horseshoe Park Larimer | 445950 wmns] ET EE

Gaskil Ponds Larimer 126505 ae] 0 2686 47

Timber Creek Larimer | 42573 46852] 0 ans 145

Tin Lake Reservoir | Larimer | 451014 CIO) EE] 289 3300

Lily Pond Larimer | 428649 ET EE) 296 BST NW)

Horsetooth Reservoir | Larimer | 486682 aa8606s] 0 1616 3032)

Fort Collins Larimer | 492719 3366] 0 154 37380NE)

Windsor Wed 506360 use] 0 1464 48.58 CE)

Pennock Pass Larimer | 459612 EE NE] 2538 TATINE)

Riverbend Ponds Larimer | 498097 ETTIE] I] 14% 2228(NE)

with passive integrated transponder (PIT) tags to identify individuals. We used a dilute bath

of benzocaine (0.2 % solution, Sigma Chemical Co., Saint Louis, Missouri) to sedate each

toad individually. When toads were sedated fully (after approximately 5-10 min), they were

rinsed in fresh water. The cloaca and oral cavity of adult toads were swabbed twice using

Mini-Tip Culturettes (Becton-Dickinson, Sparks, Maryland). Swabs were submitted for virus

isolation and bacterial and fungal cultures. Blood (0.5 ml) was collected from anesthetized

adult toads (more than 10 g) via heart puncture (WRIGHT, 1995) with single-use, disposable

25 gauge needles and 1 ml tuberculin syringes. Blood was placed immediately into plain

hematocrit tubes and sealed with wax. Samples were shipped to NWHC within 48 hours

of collection. At NWHC, capillary tubes were centrifuged, hematocrit was determined,

and serum in capillary tubes was archived (- 70°C). Toads were allowed to recover under

observation in the field (30-45 min). In addition to the non-lethal sampling, five boreal toads

were found dead and one abnormal live adult toad was collected. The live toad was mailed

with ice packs and dead toads were promptly fixed in the field by emersion in 10 % formalin.

LABORATORY PROCEDURES

Necropsy

Amphibians that were dead on arrival at NWHC were necropsied the same day as they

were received. Live larvae and just metamorphosed frogs were euthanized in 1:500 solution of

MS222 (methanesulfonate salt, Sigma Chemical Co., St. Louis, Missouri); adult toads and

tiger salamanders were euthanized by applying 2-3 em of 20 % benzocaine ointment (Orasol

gel, Clay-Park Labs Inc, Bronx, New York) to the dorsal midline of head and thorax. External

and internal examinations were performed using a dissecting microscope equipped with a

35 mm camera.

Source : MNHN, Paris

GREEN & MUTHS 113

Hematology

Blood was collected into plain capillary tubes and onto Nobuto blood filter strips

(Advantec MFS, Inc., Pleasanton, California) for determination of hematocrit and archiving

of sera, respectively, from each metamorphosed amphibian.

Virus isolation

Samples of the liver, mesonephros (“kidney”) and spleen were pooled for virus cultures

and isolations were attempted on fathead minnow cell lines (DOCHERTY et al., 2003).

Bacterial and fungal cultures

Samples of liver, urine, mesonephros, bile, spleen or lung were submitted for aerobic

bacterial cultures. A 2 mm x 3 mm segment of cloaca and a 2-4 mm segment of distal toe were

submitted for fungal cultures. Tissues and body fluids for routine aerobic bacterial cultures

(approximately 1 mm) were placed directly into vials of 2 ml tryptie soy broth with glycerine

(TSB) and incubated at room temperature (25-27°C). Cultures for Salmonella spp. were done

in Rappaport-Vassiliadis R10 broth (Becton, Dickinson & Co., Cockeysville, Maryland).

Subcultures were performed on 5 % sheep blood agar plates and cosin methylene blue plates.

Biochemical identifications of bacterial isolates were performed using the Biolog MicroSta-

tion Microbial Identification System (Hayward, California).

Fungal cultures were performed on Sabouraud dextrose agar plates with chlorampheni-

col and tetracycline (Hardy Diagnostics, Santa Maria, California). Fungal isolates were

identified morphologically by features of their hyphae and spores.

Parasitology

Parasites were identified to phylum during necropsies by a pathologist. Some helminths

and insects were archived in hot buffered formalin or 70 % ethanol. Identifications to genus

were based on external morphology of the live helminths at a dissecting microscope, tissue

location in the host and histological features. Representative insects and helminths were

identified by parasitologists and aquatic ecologists.

Histology

Portions of ventral skin, digits, heart, liver, lung, spleen, mesonephros, stomach, intes-

tine, pancreas, urinary bladder and gonads were fixed in 10 % buffered neutral form:

processed routinely, sectioned at 5 microns, and stained with hematoxylin and eosin. Portions

of liver, ventral skin, muscle, lung and mesonephros were placed in 1.8 ml cryovials and

archived at -70°C at NWHC (Madison, Wisconsin USA).

Source : MNHN, Paris

114 ALYTES 22 (3-4)

Table 2. - Number and stage of specimens collected from outside (7 sites) and within (6 sites)

RMNP. Absent: not detected and not expected to be at site (HAMMERSON, 1999); —: not

detected or not collected.

Number of (metamorphosed/larvae/eggs) collected

Location

Bufo boreas Bufo woodhousii | Pseudacris maculata| … Rana svlvatica | Ambystoma tigrinum

Vithin RMNP T

Ketlle Tarn (n= 1) 40/1 Absent Absent Absent Absent

Spruce Lake (n= 1) 11/62 Absent Absent Absent Absent

Horseshoe Park (nr = 2) Absent Absent 12/6/6 Absent 61152

Gaskil (n= 1) Absent Absent 16/0/1 10/0/4 _-

Timber Creek (= 1) Absent Absent 1202 302 <

Outside RMNP

Lily Pond Absent Absent 1585 005 Absent

Twin Lakes Reservoir 1/10/0 Absent 1922/5 Absent _

Horsetooth Reservoir 2/00 _ Absent _-

Pennock Pass Absent Absent 2/06 Absent _

Riverbend Ponds Absent 3/04 _ Absent _

Fort Coltins Absent - cos Absent ë

Windsor Absent 1300 L Absent L

RESULTS

One-hundred-twenty-one amphibians from 5 sites within the Park and 117 amphibians

(5 species) from 7 sites outside of Rocky Mountain National Park were sampled or necropsied

(tab. 2).

NECROPSY AND PARASITOLOGY

oma tigrinum. — Twenty-three individuals were examined from one site within

RMNP. Two egg masses were considered normal and free of watermolds. Adult and larval

tiger salamanders had spargana (unidentified encysted immature cestodes) within muscles (7

of 15 larvae) and unidentified adult cestodes within gut lumina (7 of 21 larvae and adults).

Adult trematodes were present in the urinary bladders of 2 of 6 adult specimens. Encysted

metacercariae occurred in the mesonephroi of 14 of 15 larvae. Amputations of extremities

(gill and tail tips) were evident in two larvae and one adult but malformations were not found.

Bufo boreas. — Five adults, one imago, six larvae (Gosner stages 40-44) and two egg

masses from two sites within RMNP and one site outside the Park were examined. Two adult

boreal toad carcasses were desiccated, two were severely autolyzed and one was submitted in

formalin. One live adult toad was submitted because of its moribund state at capture and

small red blisters were present in its ventral skin. Unidentified adult trematodes were found in

Source : MNHN, Paris

GREEN & MUTHS 115

the urinary bladder of one specimen; no other helminths were detected probably because of

poor post mortem condition of four carcasses. Deformities were limited to one short hindlimb

digit (brachyphalangy) in one toad. One tadpole had mild scoliosis of the tail and all had

marked depletion of fat bodies.

Bufo woodhousii. — Four partial egg masses and 18 adults from three sites outside RMNP

were examined. Egg strings appeared normal and free of watermolds. External abnormalities

in adult Woodhouse’s toads were a focal ulcer in one tubercle and bilateral hypomelanism of

tubercles in a second toad. Internally, one specimen had miliary white hepatic foci and three

of 18 toads had mild effusions in the lymphatic sacs. Minute larval nematodes were found in

the body cavities of two toads and adult Rhabdias sp. (1 to 35 per toad) were found in the lungs

of 12 of 18 specimens. One toad had adult trematodes in one lung consistent with Haemato-

loechus sp. Additional helminths were found in the small intestines of nine toads; these

included unidentified adult cestodes in the duodenums of nine individuals, unidentified

nematodes in the cloacae of six individuals and adult trematodes in the mid-intestine of one

individual. Three toads had pale gastric erosions or ulcers.

Pseudacris maculata. - Twenty-eight eggmasses (9 within the Park from 3 sites; 19 from

outside the Park, tab. 2) were examined. Twelve eggmasses were considered normal; 11

eggmasses contained 1.4 to 83 % moldy eggs and another five contained 11 to 100 % dead,

mold-free eggs. From some sites, minute unidentified pyriform protozoa were visible in the

capsules and vitelline spaces of eggs. Red larval insects (Ablabesmyia sp.) were present

between eggs of 8 eggmasses within and outside of RMNP. Nineteen of 36 tadpoles were

normal but five were dead on arrival. Six larvae had deformities: five cases of domed skull

(fig. 2) from one site and one case of forked tail tip. One tadpole had oral saprolegniasis and

another had non-specific fraying of lower toothrows and lower jaw sheath. Helminthic

parasites were observed in four tadpoles, including pinworms (Gyrinicola sp.) in one, renal

metacercariae (consistent with Echinostoma sp.) in three, and unidentified encysted metacer-

cariae within the body cavities of two.

Eighty-two adult chorus frogs were examined. Externally, three adults showed abnormal

molts (dys-ecdysis); one had a single minute red ventral skin ulcer. Five chorus frogs had 1-3

short toes (brachyphalangy) and one had a fractured femur. One frog had unilateral

microphthalmia. An unidentified beetle was found in the dorsal lymphatic sac overlying the

urostyle of one specimen (fig. 3). Internally, two chorus frogs had herniation of viscera

through the abdominal wall into lymphatic sacs. Mildly enlarged livers or spleens occurred in

three specimens and one adult male had unilateral atrophy of a testis. Four chorus frogs had

adult helminths in the intestine. Encysted renal metacercariae and adult trematodes in the

urinary bladder were found in seven specimens.

Rana sylvatica. — Ten adults, 3 imagos and nine partial egg masses were examined from

three sites, two within and one outside RMNP. AI eggs were considered normal and free of

Watermolds. Two eggmasses had Ablabesmyia-like larvae burrowing between eggs. Two wood

frogs had Rhabdias sp. in their lungs. Two imagos (Gosner stage 46) and one adult from the

same site had encysted renal metacercariae consistent with Echinostoma sp. Two adults had

mildly reddened ventral and digital skin and one had brachyphalangy of two digits. One wood

Source : MNHN, Paris

116 ALYTES 22 (3-4)

Fig. 2. - Deformity in larval boreal chorus frog, lateral view of head and body. Prominent raised,

dome-shaped dorsal skull of unknown etiology occurred in 5 of 36 larvae from one site outside of

RMNP. Snout-body length of this larva was 9.7 mm.

Fig. 3. - Parasitism in an adult male boreal chorus frog. (A) Dorsal view showing markedly raised and

mildly discolored ovoid patch of skin to left of urostyle (arrow). (B) Close-up of dorsal region

of urostyle with skin partially reflected to show unidentified adult beetle within lymphatic sac

Source : MNHN, Paris

GREEN & MUTHS 117

frog had multiple internal abnormalities suggestive of crushing injury, due possibly to

attempted predation or capture. Four of 10 adults had mildly enlarged livers.

HEMATOLOGY

AII anesthetized adult boreal toads recovered within 60 min. Eight of 10 anesthetized

and heart-punctured boreal toads were recaptured in subsequent years (verified by individual

Passive Integrated Transponder [PIT] tag numbers), but recaptured toads were not resam-

pled.

Hematocrits (packed cell volumes, PCV) were determined on 70 adult amphibians and

two larval À. tigrinum. Mean (number of animals; median; range) PCV for each of the five

endemic adult amphibians were: tiger salamanders, 54.6 % (n = 6; 58.6 %; 28.9-65.1 %);

boreal toads, 39.9 % (n = 11; 39.7 %; 33.2-44.0 %); Woodhouse’s toads, 34.0 % (n = 14;

30.1 %; 17.1-46.4 %); chorus frogs, 32.8 % (7 = 34; 32.4 %; 17.7-65.3 %); and wood frogs,

39.8 % (n = 5; 34.4 %; 29.9-54.4 %). Two larval tiger salamanders had PCVSs of 34.9 % and

36.8 %. Seven of 34 adult chorus frogs had epidermal chytridiomycosis; these 7 chytrid-

positive specimens had a mean PCV of 40.3 % (median: 38.2 %; range: 27.6-65.3 %) while the

chytrid-free specimens (7 = 27) had a mean PCV of 30.8 % (median: 30.8 %; range: 12.9-

49.7 %).

HisroLoGy

Tissue sections were examined from eggs, embryos, larvae and metamorphosed speci-

mens (7 = 179) of all species.

Ambystoma tigrinum.— Minimal to moderate intestinal coccidiosis was detected in eight

tiger salamanders. Coccidial oocysts within mucosal epithelial cells contained about 8 sporo-

zoites. An unidentified systemic protozoal infection was detected in the liver, spleen, heart,

pancreas or mesonephros of two adult and 3 larval salamanders (fig. 4) from one site in

RMNP; protozoal cysts were haemogregarine-like, intracellular, small schizonts 15-25

microns in diameter. Seven larval salamanders had encysted spargana within axial muscles

that were 100-700 microns in diameter. Adult cestodes were present in the intestinal lumina of

four larval and three adult salamanders. Adult trematodes were present in the urinary

bladders of three adult specimens. Nematodes were detected in the intestine of one larva.

Encysted metacercariae with encircling granulomatous inflammation were present in the

mesonephros of 10 larvae and two adults; presence of small eosinophilic spines in some

metacercariae identified the trematodes as Echinostoma sp.

Bufo boreas. - Larvae (n = 6) and fully metamorphosed specimens (7 = 4) from two sites

within RMNP were examined histologically. Two tadpoles had non-specific minimal focal

lymphocytic hepatitis and one had minimal bacterial hepatitis. Three of four metamorphosed

specimens had mild to severe proliferative (acanthotic and hyperkeratotic) mycotic epidermi-

tis of the ventral and digital skin (fig. 5) typical of chytridiomycosis (BERGER et al, 1998;

Source : MNHN, Paris

Fig. 4. - Protozoïasis of liver of an adult male tiger salamander (24 g, SVL 97 mm). Multiple intracellular

protozoal schizonts are present within liver cells or sinusoidal macrophages. Hematoxylin and

eosin stain, *X 1000.

Fig. 5. Chytridiomycosis of ventral skin of adult male boreal chorus frog. The section shows numerous

black, spherical to ovoid chytridial thalli within superficial skin cells; a few thalli have minute thin

elongate root-like projections called rhizoids. Warthin-Starry stain, x 1000.

Source : MNHN, Paris

GREEN & MUTHS 119

Murs et al, 2003). Chytrid thalli were not detected in the keratinized structures of the oral

disc of the six tadpoles.

Bufo woodhousii. — Seventeen adult Woodhouse’s toads from three sites outside RMNP

were examined histologically. Chytridiomycosis was not detected in any specimen. Histologi-

cal findings included ova within seminiferous tubules (separate from Bidder’s organs) in 3 of

9 adult males, yolk deposition in ova within Bidder’s organs of 1 of 9 males, nematodal

pneumonia in 9 of 17 specimens consistent with infection by Rhabdias sp., mild intestinal

coccidiosis in 1 of 17 toads, unidentified adult tapeworms in intestines of 2 of 17 specimens,

non-specific minimal acute liver necrosis in 2 of 17 toads, and unilateral cataract of one lens

in a female toad. The grossly observed miliary white liver foci in one toad were attributed to

granulomatous nodules associated with larval nematodes, whereas the gastric ulcers showed

necrosis and sloughing of mucosal cells with no inflammatory cells or organisms.

Pseudacris maculata. — Thirty-two larvae and 82 adults (26 female, 56 male) from three

sites within and four sites outside RMNP were examined histologically. Three of five tadpoles

from one site within RMNP had encysted renal metacercariae consistent with Echinostoma

sp. Oral saprolegniasis in one tadpole was characterized by acute necrosis of one upper

toothrow with clustered watermold filaments. One of 22 larval chorus frog from a site outside

RMNP had a marked intracellular protozoal infection of the liver only; the protozoa were not

identified. Intestinal pinworms (Gyrinicola sp.) were detected in one tadpole.

Adult chorus frogs with chytridiomycosis were found at two of three sites within RMNP

and three of four sites outside of Park boundaries. Chytrid fungal infections ranged from

minimal to marked in 37 of 82 adult chorus frogs. Within RMNP, 14 of 40 specimens had

chytridiomycosis, and outside Park boundaries, 23 of 42 specimens were infected. Minimal

infections were characterized by chytridial thalli in superficial skin cells with no host reaction;

advanced infections showed acanthosis, hyperkeratosis, dysecdysis and large numbers of

immature, sporulated and empty thalli. Some infections were accompanied by infiltrates of

bacteria into the epidermis and into empty chytrid thalli; no fungal hyphae were seen in the

skin.

Few additional histological abnormalities were detected in adult chorus frogs. One adult

each had a para-hepatic xanthomatous nodule, embedded stomach larval nematodes, eosi-

nophilic cytoplasmic inclusions in duodenal epithelium, acute focal necrotizing mycotic

pneumonia, encysted renal metacercariae typical of Echinostoma sp., and one intersex frog

(fovotestis”).

Rana sylvatica. —- Thirteen metamorphosed wood frogs from two sites within RMNP

were examined histologically. Seven (54 %) had minimal to moderate epidermal chytridiomy-

cosis similar in distribution and extent to infections in chorus frogs and boreal toads. Other

histological findings were yolk-induced inflammation in the coelom, pneumonia due to

Rhabdias sp. in two frogs, encysted renal metacercariae due to Echinostoma sp. in three

specimens, and a displaced (ectopic) nodule of liver tissue in thigh muscles. The four

specimens with hepatomegaly had histologically normal livers.

Source : MNHN, Paris

120

ALYTES 22 (3-4)

Table 3. - Bacteria cultured from individual amphibians. Numbers indicate numbers of individuals

from each in which each bacterium was isolated. AMTI: Ambystoma tigrinum; BUBO: Bufo

boreas, BUWO: B. woodhousit, PSMA: Pseudacris maculata; RASY: Rana sylvatica;

RMNP: Rocky Mountain National Park; SQ fluid: fluid from Îymphatic sacs. Sixty-five

organs had no growth; these included 36 livers, 7 spleens, 7 kidneys, 7 eggs, 5 urines, 2 SQ

fluids, and one fat body. AIl isolates are from adults, except: * eggs; ** larvae.

Fr es (Outside of RMNP Within RMNP Haras

Aeromonas enchelela Clones T e

Skin ulcer 1

Aeromonas hydrophila Closca 1 Te ï 2 7

Urine 1

Bacillus spp. Closca 0 T .

SQ nuid 1

Curobacter freundit Cloaca 2 z

Enterobacter Sp. Mouth ï 1

Clonca 1 2 .

Urine 2

SQ fuid 2

Enierococeus Spp. Closca 1 0 :

SQ fuid 1

Escherichia coli Cloaca 2

Urine 1 4

SQ fuïd 1

Escherichia vulneris Liver 1 1

Hana aheï Moutk 1

Cloaca 1 5 1 1 _

SQ fluid 1

Liver 1

Klebsiella spp. Cloaca 1 1

Pantoea agglomerans Close 1 1

Pseudomonas spp Est eg a 5

Skin ulcer 1

Mouth ! 1

Cloaca 3 2 1 1

Urine 2 1 >

SQ fluid 4

Liver 3

Skin vesicles 2

Serratia fonticola Cloaca 1 R

SQ fluid 1

Sphngomonas paucimobilis | Mouth D 1

Staphylococeus spp. SQ fuid ï 1

[Srreprococcus spp. Cloaca ü 1 2

Lagococcus sp Cloaca 1 1

BACTERIAL, FUNGAL AND WATERMOLD CULTURES

Aerobic bacteriological cultures (oral cavity, cloaca, liver, spleen and kidney) from 24

amphibians (boreal toads, Woodhouse’s toads, chorus frogs and tiger salamanders) yielded

18 bacteria of 14 genera; none were known human pathogens (tab. 3). One boreal toad

(Spruce Lake), one Woodhouse’s toad (outside the Park) and one tiger salamander (Horse-

shoe Park) tested positive for Aeromonas hydrophila. Select cultures of 71 cloacae, intestines

and livers of larval and adult amphibians were negative for Sulmonella spp.

Source : MNHN, Paris

GREEN & MUTHS 121

Sixty-two fungal cultures were attempted on cloacae, intestines, mouths and hindlimb

digits of 42 larval and metamorphosed amphibians. Special cultures for watermolds (Oomy-

cetes) were attempted on moldy eggs from three egg masses of chorus frogs. Saprolegnia

diclina was isolated from two dead moldy eggs from two egg masses of chorus frogs, and

Saprolegnia sp. was isolated in routine fungal cultures of the digits and cloacae of one adult

boreal toad, two adult Woodhouse’s toads and one adult wood frog. Twenty isolants of

Aspergillus candidus, Basidiobolus spp., Cladosporium spp., Fusarium poae, Mucor sp., Peni-

cillium spp. and Rhizopus sp. were obtained from the cloacae of 18 larvae and adults of all

species. Nineteen isolants of Aspergillus niger, Basidiobolus spp., Cladosporium tenuissimum,

Cladosporium spp. and Penicillium spp. were identified from the digits of 17 adult amphibians

of all five endemic amphibians. An unidentified yeast and unidentified fungus of the taxon

Zygomycetes were isolated from the cloacae of an adult chorus frog and tiger salamander,

respectively. Basidiobolus spp. were isolated from 8 of 62 (13 %) amphibians and Cladospo-

rium spp. were isolated from 14 of 62 (23 %) larval and adult amphibians. No fungi were

isolated from the mouths, cloacae and digits of 33 amphibians.

VIRUS ISOLATION

Cultures were completed on 176 amphibian egg masses, larvae and fully metamorphosed

specimens. Organs (lung, liver, spleen and kidney from 164 amphibians of all five species), and

12 sets of oral and cloacal swabs from adult boreal and Woodhouse”’s toads failed to produce

cytopathic effect in fathead minnow cell lines.

DISCUSSION

Whereas the role of emerging infectious diseases in amphibian declines has been exam-

ined (e.g.: DASzAK et al., 1999; Carey, 2000; CAREY et al., 2003), little is known about baseline

health of amphibians (but see GLORIOSO et al., 1974; Hip et al., 1981). Much information

published previously on amphibian baseline health and morbidity and mortality events is

confounded by recent taxonomic splitting of the presumptive agent of red-leg disease,

Aeromonas hydrophilia, into over 15 species (JosepH & CARNAHAN, 1994), discovery of new

pathogens (e.g.: LONGCORE et al., 1999, JANCOvICH et al., 1997; DOCHERTY et al., 2003) and

inadequacies in specimen preservation and length of time between death and necropsy

(TAYLOR et al., 2001).

In our study, Aeromonas Spp. was found in 9.1 %, 33.3 %, 33.3 % and 66.7 % of live

free-living boreal toads, Woodhouse’s toads, tiger salamanders and chorus frogs, respectively.

These data mirror findings by Hirp et al. (1981), who found Aeromonas sp. in 32 % (94 of 294)

of northern leopard frogs from Minnesota and concluded that presence of this genus of

bacterium was not the cause of disease or population declines. The most commonly isolated

gut bacteria were Hafnia alvei, Pseudomonas spp. and Enterococcus spp.; these bacteria are

considered widespread and innocuous genera in the amphibian digestive tracts and probably

reflect water microbiology, invertebrate prey and other environmental features of the amphib-

Source : MNHN, Paris

122 ALYTES 22 (3-4)

ian’s habitat (Waaur et al., 1974). Other bacteria from the amphibian’s digestive tracts, such

as Sphingomonas sp, Citrobacter sp. and Klebsiella sp. also are common flora of aquatic

environments and insects, and likely reflect water quality and prey items (Waau et al., 1974).

The mammalian enteric bacterium, Escherichia coli, was isolated from three toads captured at

onesite in an agricultural area near a residence (formerly a farm house) with agricultural fields

on two sides. We suggest that coliform bacteria are uncommon in the digestive tracts of

amphibians from remote or nearly pristine sites (e.g.. RMNP) but may be acquired in

amphibians associated with human activities or livestock. Similarly, salmonellae were not

isolated from any amphibians in this study; EVERARD et al. (1979) suggested that Sa/monella

spp. are more common in tropical amphibians in close association with humans. Other studies

of toads (Bufo spp.) in urban and tropical regions found 55.6 %, 36.7 % and 12.7 % to be

carriers of salmonellae in Surinam (BooL & KAMPELMACHER, 1958), India (SHARMA et al.,

1977) and eastern Australia (O’SHEA et al., 1990), respectively. TAYLOR et al. (2001) concluded

“that most, if not all, amphibians carry one or more Salmonella sp”. Our findings refute this

statement and provide evidence that toads and other amphibians from temperate zones and

high altitudes are seldom carriers of salmonellae.

Basidiobolus spp. are problematic Zygomycetes that are isolated commonly from gut

contents of insectivorous amphibians (GUGNANI & OKAFOR, 1980; OKAFOR et al., 1984) but

also have been implicated as a primary epidermal pathogen of amphibians (GROFF et al.,

1991; TayLor et al., 1999a-b; TayLOoR & MiiLs, 1999). Purported basidiobolomycosis of

amphibians is histologically indistinguishable from chytridiomycosis, but basidiobolomycosis

in all other vertebrate classes is noteworthy for the presence of fungal hyphae, intense

inflammatory cell response, and invasion of non-keratinized tissues (GUGNANI, 1999). Chy-

tridiomycosis of amphibians produces no hyphae, only a slight or no inflammatory cell

response and is an intracellular infection of cutaneous keratinized cells only (BERGER et al.,

1998). Some published cases of basidiobolomycosis in amphibians are now believed to have

been chytridiomycosis (CAREY et al., 2003; MuTHs et al., 2003). Only 1 of 47 amphibians with

chytridiomycosis in this study had fungal hyphae in tissue sections, and the hyphae were

observed in the lung. Seven of 8 isolants of Basidiobolus spp. in this study were from cloacae

of adult amphibians. AIl three isolants of Basidiobolus spp. from Woodhouse’s toads were

cloacal, and all were negative for chytridiomycosis by histology. Fungal cultures of the skin

were attempted on 18 frogs (chorus frogs and wood frogs) with histological chytridiomycosis

and an additional 21 chytrid-negative amphibians; Basidiobolus sp. was isolated from 1 of 39

(2.6%) skin samples from a chytrid-infected wood frog. We conclude that, in RMNP,

Basidiobolus spp. are common non-pathogenie fungi in the alimentary tract of amphibians,

which supports reports by GUGNANI & OKAFOR (1980) and OKAFOR et al. (1984). The low

isolation rate of Basidiobolus Spp. from amphibian skin suggests that the organism is rare on

the epidermis of free-living amphibians, and many skin isolants may be due to fecal contam-

ination of the skin.

A few common and usually innocuous protozoan and helminthic infections were found

in some species and populations. Intestinal coccidiosis and cestodiasis were detected in

Ambystoma tigrinum. Whereas most coccidial protozoa of vertebral wildlife are host-sp

parasites, life-threatening infections may occur in immature individuals. Because all coccidial

infections were considered mild, and because all sympatric chorus frogs (7 = 18) from the

same site were free of coccidia, we conclude that this parasite is not à threat to anuran

Source : MNHN, Paris

GREEN & MUTHS 123

populations. However, another unidentified non-enteric systemic protozoal infection was

detected in three tiger salamanders from one site within the Park and one adult chorus frog

outside of the Park. No morbidity or mortality was associated with this unidentified systemic

protozoan infection. Because of low prevalences of these protozoa and absence of infections

in sympatric amphibians at each site (which suggests host specificity), we suggest the systemic

protozoal infections may be an endemic parasite. The single incidence of a beetle found

embedded in the dorsum of a chorus frog is not necessarily a significant finding regarding the

parasite load of this species in the Park but is unusual. The frog hosting the beetle was received

at NWHC alive, euthanized and dissected immediately indicating that the beetle was not a

post-mortem invader.

The only major lethal pathogen associated previously with amphibian mortality events

and population declines identified in this study was Batrachochytrium dendrobatidis. Other

potential pathogens associated with infrequent morbidities and mortalities were intestinal

coccidiosis in tiger salamanders, heavy parasitic infections of Woodhouse’s toad by the

amphibian lungworm Rhabdias sp., hepatic or systemic protozoiasis by unidentified proto-

zoa, and saprolegniasis of eggs by S. diclina. Virus cultures were negative in all amphibians of

all life stages and there was no cultural or histological evidence of bacterial septicemias (‘red

leg” syndrome).

Chytridiomycosis in boreal toads in RMNP was associated with severe population

declines and mortality events in 1998-2000 (Murus et al., 2003). This study confirms contin-

ued mortality in Bufo boreas due to chytridiomycosis within RMNP. In addition, chytridio-

mycosis was identified in two new amphibian hosts, Pseudacris maculata and Rana sylvatica,

in a total of 44 animals (many more individuals than reported initially by RITTMANN et al.,

2003). Histological examinations suggest that the intensity of infections by B. dendrobatidis in

some amphibians of each species was sufficient to have caused morbidity and mortality. The

prevalences of chytridiomycosis in fully metamorphosed specimens of each host species were

similar: 60 % in B. boreas, 45 % in P maculata and 54 % in R. sylvatica. The high prevalences

of chytridiomycosis in chorus and wood frogs are worrisome and are equivalent to preva-

lences in boreal toads; population declines of B. boreas have occurred throughout the

southern Rocky Mountains (Cor et al., 1997: Murns et al., 2003; JUNGWIRTH 2004), but

population data for sympatric frogs are unavailable. Monitoring anuran populations in

Colorado and Wyoming as well as landscape scale assessments of the number of populations

extant in the region are warranted to determine if disease-related declines are occurring.

Additionally, experiments to fulfill Koch’s postulates using chorus and wood frogs are

necessary to verify pathogenicity of B. dendrobatidis and determine mortality rates in imagos

and adults.

The mechanism of lethality of B. dendrobatidis infections in amphibians remains un-

known, Our hematological findings support the hypothesis that skin infections by & dendro-

batidis disrupt essential functions of the amphibian epidermis. Acanthosis, hyperkeratosis

and dysecdysis of the epidermis are associated with advanced 8. dendrobatidis infections and

may impair essential water absorption through the skin and disrupt osmoregulation. The

mean hematocrit of infected chorus frogs was 40.3 % whereas non-infected frogs had a mean

hematocrit of 30.8 %. There was no difference between the hematocrit values of infected and

non-infected animals (ANOVA, F= 1.73, P = 0.20, df = 1). Whereas the mean values were not

Source : MNHN, Paris

124 ALYTES 22 (3-4)

statistically different, the possibility remains that impaired osmoregulation (BERGER et al,

1998; DaszaK et al., 1999) and elevated hematocrits (usually indicative of dehydration) occur

in some anurans with epidermal chytridiomycosis; additional hematological studies are

needed.

‘Watermold infections (saprolegniasis) in eggs and embryos, some of which were identi-

fied as S. diclina, affected 25 % (11 of 44) of anuran egg clutches, but in all egg clutches, some

live, non-infected embryos were present. Mass mortality of anuran eggs in the Cascade

Mountains has been associated previously only with Saprolegnia ferax (KIESECKER & BLAUS-

TEIN, 1995, 1997), but whether S. diclina is a primary pathogen, a secondary invader of

abnormal eggs, or a saprobe on infertile eggs or eggs killed by other agents could not be

determined in this study.

Lungworms of two taxa (Rhabdias sp. and Haematoloechus sp.) and intestinal coccidio-

sis were found in Bufo woodhousii. Immature, infective stages of the lungworm Rhabdias sp.

have killed experimental juvenile Bufo marinus (WiLLiaMs, 1960). Rhabdias spp. have a direct

life cycle (ï.e., without intermediate hosts) and can infect a range of amphibian hosts (FLYNN,

1973), suggesting that this lungworm may infect amphibians in RMNP in situations of

crowding, fecal contamination or interspecies contact accompanied by appropriate tempera-

tures and humidity.

Amphibian diversity in RMNP is naturally depauperate, including only five amphibian

species (HAMMERSON, 1999). Of these, Rana pipiens has been extirpated recently (CORN et al.,

1997), B. boreas has declined precipitously, and populations of R. sylvatica occur only on the

west side of the continental divide in RMNP. The latter populations are part of a relictual

(meta)population in Colorado and Wyoming that is isolated from other populations in North

America (HAMMERSON, 1999). This situation (small, isolated, relict populations) leaves R.

sylvatica at risk for disease-related extirpation with the subsequent potential loss of genetic

diversity.

Given the state of the existing boreal toad populations in RMNP (small isolated

populations, continued declines associated with chytridiomycosis), the arrival of another

infectious disease could be disastrous. There is potential for natural immigration of healthy

(or unhealthy) animals into RMNP, but it is limited. Twin Lakes Reservoir is less than 8 km

north of the nearest B. boreas breeding site on the north edge of RMNP, and B. boreas and P.

maculata currently are found there. Based on habitat and elevation, boreal toads are expected

to be present at Pennock Pass (2552 m), a site less than 10 km north of the Park boundary, but

have not been documented there. Lily Pond is less than 9 km northeast of the Park boundary

but more than 20 km from the nearest B. boreas breeding sites on the north edge of RMNP.B.

boreas was found at Lily Pond historically (late 1960s) (A. Spencer and P. S. Corn, personal

communication) and P maculata and R. sylvatica currently reside there. Specimens are not

available to test for disease, so it remains unknown why boreal toads disappeared from this

site. However, 6 of 15 (40 %) adult chorus frogs captured at Lily Pond in 2001 and 2002 had

chytridiomycosis.

There is patchy, appropriate habitat between the locations discussed above and the B.

boreas sites in RMNP (Arapahoe Roosevelt National Forest Service, Comanche Peak Wil-

derness Area and National Park). À putative migration route into the Park would include

significant elevation gain to the top of Stormy Peaks pass (764 m, Twin Lakes Reservoir;

Source : MNHN, Paris

GREEN & MUTHS 125

652 m, Lily Pond; and 1031 m, Pennock Pass), including several kilometers of high alpine

habitat (over 3500 m). Boreal toads have been observed in multiple years at elevations of

3383 m (Lake Husted) (CoRN et al., 1997) and are capable of moving over substantial

distances. We have documented individual toads moving 5 km between breeding sites (over

multiple years) (Muths & Corn, unpublished data) such that it would not be surprising to find

toads moving from Twin Lakes Reservoir into the northeast portion of RMNP where

populations of B. boreas currently are found. Migration out of the Park into surrounding

areas would be equally likely.

Chytridiomycosis has been detected in all four anuran species within and adjacent to

RMNP. B. boreas has suffered severe state-wide declines with chytridiomycosis playing a

significant role. Whereas the status of populations of chorus and wood frogs are largely

unknown, the prevalence of chytridiomycosis (45 % and 54 %, respectively) is high and

equivalent to its prevalence in declining boreal toads in this region (Green, personal obser-

vation). The cause of the extirpation of R. pipiens in RMNP remains unexplained, although

chytridiomycosis was detected in museum specimens that were captured less than 100 km

from RMNP in the 1970's (CAREY et al., 2003).

Diseases in B. woodhousii were of special interest because chytridiomycosis and other

diseases have been reported in several species of Bufo in the western United States in recent

years (KIESECKER and BLAUSTEIN, 1997; TAYLOR et al, 19994; GREEN & KAGARISE SHERMAN,

2001). 8. woodhousii occurs only at the periphery of RMNP and although Woodhouse’s and

boreal toads are thought to be allopatric in Colorado, potential for sympatry exists in

Archuleta County (HAMMERSON, 1999). If a change in climate occurs (e.g., prolonged

drought), boreal or Woodhouse’s toads may begin to use habitat below or above previous

elevational ranges. Lungworms of two taxa, in particular Rhabdias sp., and intestinal cocci-

diosis were found in B. woodhousii. If the toads become sympatric, Rhabdias sp. and coccidial

parasites could be transmitted directly to B. boreas.

The etiologies of musculo-skeletal deformities and gonadal deformities in anurans were

not determined. Individuals with displaced or ectopic ova within seminiferous tubules were

diagnosed as intersexes (ovotestes) in 3 male Woodhouse’s toads and one chorus frog; similar

abnormal gonads have been reported in Acris crepitans in Illinois (REEDER et al., 1998).

Abnormal testes may be due to environmental contaminants (e.g., estrogen-mimicking

chemicals), or they may be variations of normal anatomy and development. However, one

adult male B. woodhousii from a partially urbanized site (outside of the Park) had a mild

accumulation of yolk (vitellogenin) in ova of Bidder’s organs: the production of yolk in a

male toad is considered evidence of recent exposure to an estrogen-mimicking chemical

(PaLMER & SELCER, 1996). The etiology and precise nature of the domed skulls in one group

of larval chorus frogs was not determined; additional studies are in progress.

CONCLUSION

Although our study failed to detect any novel diseases associated with high mortality

rates (GREEN et al., 2002) in amphibian populations outside the park, there are potential

Source : MNHN, Paris

126 ALYTES 22 (3-4)

routes into the park if such diseases are detected in the future. RMNP is surrounded almost

completely by National Forest and Wilderness Areas which include appropriate habitat for

amphibians. Whereas surrounding public lands may be useful habitats facilitating immigra-

tion and dispersal, these areas also could facilitate the transmission of disease into the park

since chytridiomycosis has been found in frogs at 3 of 7 sites outside the park. Back country

use by sportsmen and domestic animals, including outfitters with pack animals, hikers with

dogs and sportsmen using live bait for fishing, could increase the potential of mechanical

transmission of pathogens, especially at water sources which may be breeding areas for

amphibians. Although the RMNP specifically bans dogs and limits where pack animals can

be tethered and on what trails they can use, the Forest Service does not have the same

restrictions. None of these potential vectors are proven methods of transmission of amphib-

ian pathogens or are linked directly to the decline of amphibians; the mode of transmission

of most major amphibian pathogens remains unknown (GREEN et al., 2002; CAREY et al.,

2003). However, anthropogenic movement of nonindigenous species and pathogens are well

documented, and at least one major amphibian disease (chytridiomycosis) may be linked to

human-mediated transmission (MOREHOUSE et al., 2003).

Of the diseases that have impacts on amphibian populations (chytrid fungus, ranavi-

ruses, and a new mesomycetozoa-like organism; GREEN et al., 2002), only B. dendrobatidis was

found within and outside RMNP. Ranaviruses and mesomycetozoan-like organisms were not

found in any amphibians in this study. We suggest that the principal hazard to anuran

populations within and adjacent to RMNP is chytridiomycosis and that chytridiomycosis is

the only major lethal infectious disease of multiple species of amphibians in and around

RMNP. Our data are preliminary and limited by sample size and numbers of sites, although

amphibians were collected during the months of peak disease activity (GREEN et al., 2002).

The biomedical data presented here provide information on common bacteria and fungi

of free-living amphibians, disease characteristics in populations of amphibians found at

higher elevations, and a baseline from which to continue to monitor amphibian health and

population declines in the Rocky Mountains.

RÉSUMÉ

Nous avons étudié les caractéristiques sanitaires des amphibiens au sein du Rocky

Mountain National Park et aux alentours pour établir la présence actuelle de maladies dans

le parc et en dehors de celui-ci mais risquant de contaminer les amphibiens du parc. Des

amphibiens ont été récoltés et examinés dans cinq sites du parc et sept sites à moins de 60 km

des limites de celui-ci. Nous avons effectué des autopsies (7 = 238), des isolements de virus, des

cultures bactériennes et fongiques, et des examens histologiques sur des masses d'œufs

d'amphibiens (hors du parc/dans le parc: 26/22), des larves (30/42), des imagos (individus

récemment métamorphosés) (0/3) et des adultes (61/67) de cinq espèces. Des infections

importantes par le champignon chytride pathogène Batrachochytrium dendrobatidis (chyri-

diomycose) ont été détectées chez trois espèces (Bufo boreas, Pseudacris maculata et Rana

sylvatica) de trois des cinq sites au sein du par et une des trois espèces (P maculata) de trois

sites extérieurs au parc. Chez les animaux métamorphosés des trois espèces (B. boreas, P.

Source : MNHN, Paris

GREEN & MUTHS 127

maculata and R. sylvatica), la chytridiomycose a été trouvée respectivement chez 60 % (7 = 3),

46% (n = 37) et 54 % (n = 7) des individus examinés. La chytridiomycose était la maladie

infectieuse létale principale détectée chez les trois espèces d'amphibiens au sein du parc ou

près de celui-ci. Des champignons supérieurs ont été isolés à partir du cloaque et de la peau

des cinq espèces d’amphibiens. Des Oomycètes ont été isolés à partir d'œufs ou de la peau des

cinq espèces. Aucun Ranavirus n’a été trouvé dans les cultures ni par l'examen histologique de

176 et 142 amphibiens des deux provenances. Quinze genres de bactéries ont été identifiés chez

des amphibiens larvaires et métamorphosés, et un nématode parasite potentiellement patho-

gène, Rhabdias sp, a été trouvé chez 61.1 % (n = 11) des B. woodhousii en dehors du pare, mais

seulement chez 2 (15.4 %) R. sylvatica au sein du parc.

ACKNOWLEDGMENTS

We thank Rocky Mountain National Park and the Department of Interior’s Amphibian Research

and Monitoring Initiative for funding this project. We are indebted to S. Rittmann for an extraordinary

amount of field- and lab-work and care in animal collection and transport. Thanks to J. &S. Chamberlain

and to M. Brannon and W. Wolf for access to their property. Thanks also to D. Docherty, R. Long and

B. Berlowski (NWHC) for extensive assistance with cultures. We also thank two anonymous reviewers for

constructive comments that improved the manuscript.

LITERATURE CI

BERGER, L., SPEARE, R., DASZAR, P., GREEN, D. E., CUNNINGHAM, À. À., GOGGIN, C. L., SLOCOMBE, R.,

4 M. A. HyaïT, A. D., McDonaLp, K. R., HINES, H. B., Lirs, K. R., MARANTELLI,

H., 1998. — Chytridiomycosis causes amphibian mortality associated with population

declines in the rainforests of Australia and Central America. Proc. natl. Acad. Sci. USA, 95:

9031-9036.

BerGER, L., VoLr, K., MATTHEWS, $., SPEARE,

free-ranging giant barred frog (Mivoph)

2378-2380.

Book, P. H. & KAMPELMACHER, E. H., 1958. - Some data on the occurrence of Sulmonella in animals in

R. & Timms, P., 1999. — Chlamydia pneumoniae in à

s tus) from Australia. Z clinical Microbiol., 37 (7):

Surinam. Anronie van Leeuvenhoek, 24: 76-80.

Carey, C., 2000. - Infectious disease and worldwide declines of amphibian populations, with comments

on emerging diseases in coral reef organisms and in humans. Environ, Health Persp., 108, suppl. 1:

143-150.

Carey, C.. BRADrORD, D. F., BRUNNER, J. L., COLLINS, J. P,, DAVIDSON, E. W., LONGCORE, J. E.. OUELLET,

M.. Pessier, A. P. & ScHock, D. 2003. - Chapter 6. Biotic factors in amphibian population

declines. LiNDER, S. K. Kkesr & D. W. SPARLING (ed.), Amphibian decline: an integrated

analysis of multiple stressor effects, Pensacola, Florida, SETAC Press: 153-208.

000. - Amphibian declines: review of some current hypotheses. Ju: D. W. SPARLING, G.

LinDe & C. A. Bishop (ed.), Ecoroxicology of amphibians and reptiles, Pensacola, Florida, SETAC

Press: 663-696.

Cox, PS. & FOGELMAN, J. C., 1984. - Extinction of montane populations of the northern leopard frog

CRana pipiens) in Colorado. J. Herp.. 18 (2): 147-152.

Corn, PS. JENNINGS, M. L. & MUTuS, E., 1997. - Survey and assessment of amphibian populations in

Rocky Mountain National Park. Northwest. Nat. 78 (1)

Source : MNHN, Paris

128 ALYTES 22 (3-4)

DaszaK, P., BERGER, L., CUNNINGHAM, A. A., HYATT, A. D., GREEN, D. E. & SPEARE, R., 1999. —

Emerging infectious diseases and amphibian population declines. Emerging infectious Diseases, 5

(6): 735-748.

DocEeRTY, D. E., METEYER, C. U., WANG, J., M0, J., CASE, S. T. & CHINCHAR, V. G., 2003. - Diagnostic

and molecular evaluation of three iridovirus-associated salamander mortality events. J_ Wildlife

Diseases, 39 (3): 556-566.

EveRARD, C. O. R., TorA, B., BASsETT, D. & ALt, C., 1979. — Salmonella in wildlife from Trinidad and

Grenada, WI. J Wildlife Diseases, 15 (2): 213-219.

FLYNN, R. J., 1973. — Parasites of laboratory animals. Ames, Iowa, USA, Iowa State University Press:

505-642.

GLORIOSO, J. C., AMBORSKI, R. L., AMBORSKI, G. F. & CuLLey, D. C., 1974. - Microbiological studies on

septicemic bullfrogs (Rana catesbeiana). Am. J. ver. Res., 35 (9): 1241-1245.

Green, D. E., CONVERSE, K. A. & SCHRADER, A. K. 2002. - Epizootiology of sixty-four amphibian

morbidity and mortality events in the USA, 1996-2001. Ann. New York Acad. Sci., 969: 323-339.

GROFF J. M., MUGHANNAM, A., MCDOWELL, T. S., WONG, A., DYKSTRA, M. J., FRYE, F L. & HEDRICK,

R. P, 1991. - An epizootic of cutaneous zygomycosis in cultured dwarf African clawed frogs

(Hymenochirus curtipes) due to Basidiobolus ranarum. J. med. vet. Mycol., 29: 215-23.

Guanani, H. C., 1999. - A review of zygomycosis due to Basidiobolus ranarum. Eur. J. Epidemiol., 15:

923-929.

GuGnaNt, H. C. & OKAFOR, J. I., 1980. - Mycotic flora of the intestine and other internal organs of

certain reptiles and amphibians with special reference to characterization of Basidiobolus isolates.

Mykosen, 23: 26-268.

HAMMERSON, G. A., 1999. — Amphibians and reptiles in Colorado. 2" edition. Boulder, University Press of

Colorado and Colorado Division of Wildlife: i-xxii + 1- 484.

Hp, D. W. DrrscH, S. L., MCKINNELL, R. G., GorHAM, E., MARTIN, F. B., KURTZ, S. W. &

DuBROvOLNY, C., 1981. — Aeromonas hydrophila in wild-caught frogs and tadpoles (Rana pipiens)

in Minnesota. Lab. Anim. Sci., 31: 166-169.

JaNCOvICH, J. K, DavipsON, E. W., MORADO, J. F., JacoBs, B. L. & COLLINS, J. P., 1997. — Isolation of

a lethal virus from the endangered tiger salamander Ambystoma tigrinum stebbinsi. Dis. aquat.

Organ., 31: 161-167.

Josepx, S. W. & CARNAHAN, A., 1994. — The isolation, identification, and systematics of the motile

Aeromonas species. Annual Rev. Fish Dis., 4: 315-343.

JUNGwWIRTH, T. (ed.), 2004. - Report on the status and conservation of the boreal toad (Bufo boreas boreas)

in the southern Rocky Mountains 2003. Colorado Division of Wildlife, 6060 Broadway, Denver,

CO 80216.

KIESECKER, J. M. & BLAUSTEIN A. R., 1995. - Synergism between UV-B radiation and a pathogen

magnifies amphibian embryo mortality in nature. Proc. natl. Acad. Sci. USA, 92: 11049-11052.

KER, J. M. & BLAUSTEIN, A. R., 1997. — Influences of egg laying behavior on pathogenic infection

of amphibian eggs. Consers. Biol., 1 (1): 214-220.

LONGCORE, J. E., PESSIER, A. P. & NICHOLS, D. K., 1999. - Barachochytrium dendrobatidis gen. et sp. nov.,

a chytrid pathogenic to amphibians. Mycologia, 91: 219-227.

MOREHOUSE, E. A., JAMES, T. Y., GANLEY, A. R. D., VILGALYS, R., BE L., Murray, JP. &

LONGCORE, JL E., 2003. - Multilocus sequence typing suggests the chytrid pathogen of amphibians

is a recently emerged clone. Mol. Ecol., 12: 395-403.

Murs, E., CORN, P. S., PESSIER A. P. & GREEN, D. E., 2003. - Evidence for disease related amphibian

decline in Colorado. Biol. Conserv., 110: 357-365.

OkAFOR, J. L., TESTRAKE, D., MUSHINSKY, H. R., & YANGCO, B. G., 1984. - À Basidiobolu

association with reptiles and amphibians in southern Florida. J. med. ver. Mycol.,

O'SHEA, P., SPEARE, R. & THOMAS, A. D., 1990. -Salmonellas from the cane toad, Bufo marinus. Aust. vet.

J., 67 (8): 310.

PALMER, B. D. & SELCER, K. W., 1996. — Vitellogenin as a biomarker for xenobiotic

EL (ed.), Environmental toxicology and risk ass

‘onshohocken, Pennsylvania, ASTM: 3-22.

RITTMANN, S., MUTHS, E. & G: . E., 2003. — Pseudacris triseriata (western chorus frog) and Rana

sylratica (wood frog). Chytridiomycosis. Herp. Rev, 34 (1): 53.

KIESI

rogens: a review. /n:

ssment: biomarkers

Source : MNHN, Paris

GREEN & MUTHS 129

REED, K. D., RUTH, G. R., MEYER, J. A. & SHUKLA, S. K., 2000. - Chlamydia pneumoniae infection in a

breeding colony of African clawed frogs (Xenopus tropicalis). Emerging infectious Diseases, 6 (2):

196-199.

REEDER, A. L., FOLEY, G. L., NicHoLs, D. K., HANSEN, L. G., WiKOFF, B., FAEH, S., EISOLD, J., WHEELER,

M.B., WaRNER, R., MURPHY, J. E. & BEASLEY, V. R., 1998. — Forms and prevalence of intersexu-

ality and effect of environmental contaminants on sexuality in cricket frog (Acris crepitans).

Environ. Health Persp., 106 (5): 261-266.

SHARMA, V. K., ROHDE, R., GARG, D. N. & KUMAR, A., 1977. - Toads as natural reservoir of Salmonella.

Zbl. Bakt. Parasit., 239: 172-177.

TAYLOR, S. & MiLs, K. W., 1999. - Mortality of captive Canadian toads from Basidiobolus ranarum

mycotic dermatitis. J Wildlife Diseases, 35 (1): 64-69.

TAYLOR, S., WILLIAMS, E. S. & Müics, K. W., 1999b. — Experimental exposure of Canadian toads to

Basidiobolus ranarum. J. Wildlife Diseases, 35 (1): 58-63.

TAYLOR, S., WILLIAMS, E.S., THORNE, E. T., MiLLs, K. W., WirHErs, D. I. & PIER, A. C., 1999a.- Causes

of mortality in Wyoming toads. J Wildlife. Diseases, 35 (1): 49-57.

TAYLOR, S. K., GREEN, D. E., WRIGAT, K. M. & WHiTAKER, B. R., 2001. - Chapter 13. Bacterial diseases.

. M. WRIGHT & B. R. WHITAKER (ed.), Amphibian medicine and captive husbandry, Malabar,

Florida, Krieger Publishing Comp. 159-191.

WaAU, D. VAN DER, COHEN, J. B. & NACE, G. W., 1974. - Colonization patterns of aerobic gram-negative

bacteria in the cloaca of Rana pipiens. Lab. Anim. Sci., 24 (2): 307-317.

WILLIAMS, R. W., 1960. - Observations on the life history of Rhabdias sphaerocephala Goodey 1924 from

Bufo marinus L., in the Bermuda Islands. J Helminthol., 34: 93-98.

WRIGHT, K., 1995. - Blood collection and hematological techniques in amphibians. Bull. Assoc. Rept.

Amphib. Ver. 5 (2): 8-10.

Corresponding editor: C. Kenneth Dopp, Jr.

SCA 2005

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 130-145.

Toxicity to amphibians of environmental

extracts from natural waters

in National Parks and

Fish and Wildlife Refuges

Christine M. BRIDGES & Edward E. LITTLE

US Geological Survey, Columbia Environmental Research Center,

4200 New Haven Road, Columbia, Missouri 65201, USA

Amphibian population declines are not limited to overtly degraded

habitats, but often occur in relatively pristine environments such as national

parks or wildlife refuges, thus forcing biologists to examine less obvious

causes for declines such as the presence of contaminants. The objective of

our study was to extract naturally-occurring compounds from amphibian

habitats (using semipermeable membrane devices) in three national parks

or wildlife refuges (two sites within Sequoia Kings Canyon National Park,

Big Bend National Park, and Kenai National Wildlife Refuge), and assess

their toxicity to developing larvae using bioassays. Extracts did not cause

mortality, so all effects observed were sublethal, influencing life history

characteristics. In all three areas studied, amphibians reared in extracts

from at least one of the two sites exhibited either a lengthened larval period

or reduced mass at metamorphosis. Extracts from both the air and water at

one site in Sequoia Kings Canyon National Park lengthened the larval

period, which is in agreement with studies showing elevated levels of

aerially transported contaminants at sites such as this within the park.

Ultraviolet radiation, which is also suspected of having caused amphibian

declines and was included as a factor in our study, did not act alone or alter

the toxicity of the extracts.

INTRODUCTION

Reports of amphibian population declines in seemingly pristine environments, particu-

larly in the western United States, have become increasingly common (CAREY, 1993; FELLERS

& Drosr, 1993: BRADFORD et al., 1994; CoRN, 1994; JENNINGS & HAYES, 1994; KNapp &

MaATTHEWs, 2000). Although in some areas of the west the status of amphibian populations

remain largely unstudied (e.g., along the Rio Grande river; JUNG et al., 2002), declines have

been documented throughout the Sierra Nevada mountains (BRADFORD et al., 1993), in

western Colorado (CAREY, 1993) and in the Cascade mountains (MCALLISTER et al., 1993),

among other locations. Although few declines have been definitively linked with causes, others

cases have been more clear (KNAPP & MATTHEWS, 2000).

Source : MNHN, Paris

BRIDGES & LITTLE 131

Environmental contamination has been proffered as having contributed to some of these

declines. For instance, DAvipsoN et al. (2001, 2002) found a strong positive relationship

between observed patterns of declines in a number of species in the Sierra Nevada mountains

and upwind agricultural land usage in the Central Valley of California. SPARLING et al. (2001)

correlated a reduction in cholinesterase levels (which may indicate exposure to certain types

of toxicants) with areas of population decline in the Sierra Nevadas. It is often difficult,

however, to determine what specific contaminants are present at a site or in what concentra-

tions, which often hinders determining the influence of exposure to contaminants on the

populations of organisms inhabiting the area. Furthermore, several environmental variables

(e.g., pH, Dissolved Oxygen Content, predators) may be acting alone or in combination with

the contaminants, contaminants may be present in complex mixtures, and/or the identity and

concentrations of specific substances and their metabolites within the composition of such

mixtures may be unknown or difficult to determine.

One of our goals was to expose developing amphibian larvae to composite samples of

waterborne contaminants in various sites within national parks and fish and wildlife refuges,

using in-laboratory bioassays. To accomplish this, we deployed semipermeable membrane

devices. These are integrative samplers containing lipid that accumulates lipophilic organic

compounds from the environment in a manner that is similar to aquatic organisms (HUCKINS

et al., 2002). The uptake rates of fat-soluble compounds by semipermeable membrane devices

can be used to define the approximate daily exposure to lipophilic compounds by aquatic

organisms (HUCKINS et al., 2002). Thus, the composite waterborne environmental contami-

nantexposure of organisms to the bioavailable lipophilie compounds present at the site can be

accomplished by using extracts from semipermeable membrane devices as toxicant solutions

for bioassays (Perry et al., 2000; HuCkINS et al., 2002) based on the duration of deployment

in the environment. Using semipermeable membrane device extracts in bioassays has

been shown to be an effective way to assess the toxicity and teratogenicity of naturally

occurring environmental compounds to larval anurans (BRIDGES & LiTrLe, 2003; BRIDGES et

al., 2004).

Ultraviolet (UV) radiation has also been suspected of causing declines among amphib-

ian populations. Exposure to UVB radiation at larval and egg stages has been suspected of

negatively impacting some amphibian species directly by reducing hatching success and

increasing rates of embryonic malformation (BLAUSTEIN et al., 1998; LiZANA & PEDRAZA,

1998; BROOMHALL et al., 2000; STARNES et al., 2000) as well as generating larval malformations

(ANKLEY et al., 2000). And although being a lesser factor than contaminants, DAviDsON et al.

(2002) suspected UV as a likely contributor to amphibian declines in the Sierra Nevada

mountains. However, because no environmental factor acts in isolation, it has become clear

that single-factor explanations simply may not be sufficient to explain this widespread

phenomenon of population declines. For example, UV radiation is also known to increase the

toxicity and/or teratogenicity to amphibians of various compounds in aquatic habitats (ZAGA

et al., 1998; Lrrrie et al., 2000; BRIDGES et al., 2004).

One of our objectives was to investigate whether exposure to lipophilic compounds

extracted from natural environments located in conventionally pristine areas (i.e., Sequoia

Kings Canyon National Park, Big Bend National Park, Kenai National Wildlife Refuge) can

negatively affect developing amphibian larvae. Furthermore, because UV radiation can also

Source : MNHN, Paris

132 ALYTES 22 (3-4)

impact amphibians (in and of itself as well as when in combination with environmental

contaminants), the effects of UV radiation were also explored. In this study we examined the

effects of these two factors, and their interaction, on amphibian larval survival, length of the

larval period, and size at metamorphosis for various anuran species.

MATERIALS AND METHODS

SEMIPERMEABLE MEMBRANE DEVICE EXPOSURE AND EXTRACTION

To determine whether contaminants present at our study sites can cause mortality or

malformations in larval amphibians, we deployed semipermeable membrane devices at two

sites within each refuge or national park. Five standard semipermeable membrane devices

(described in HUCxINSs et al., 2002) were placed in each of three replicate stainless-steel

canisters attached to a steel cable anchored to the shore. The conditions of each deployment

are given below.

At the time of deployment a metal can containing semipermeable membrane devices

(hereafter, field blank) was opened and exposed to the air at each site. These provided a control

for any airborne contaminants present while the semipermeable membrane devices were

exposed to the air (i.e., before being placed into the water). Once the semipermeable mem-

brane devices were in the water, the cans containing the field blanks were sealed and stored at

0-4°C for the time between semipermeable membrane device deployment and retrieval to halt

contaminant uptake (from the site air that had filled the can). During retrieval, field blanks

were once again exposed to the air at the site for as long as it took to remove the semiperme-

able membrane devices from the water and seal them in cans. After retrieval, both semiper-

meable membrane devices and field blanks were shipped on ice and remained frozen until they

were processed at the Columbia Environmental Research Center (Columbia, Missouri) using

techniques outlined in Perry et al. (2000).

AIT semipermeable membrane device extracts from each site were pooled into a single

composite sample. Field exposure and control extracts (1.e., field blanks) were dissolved into

sterile dimethylsulfoxide (DMSO) by solvent exchange. Semipermeable membrane device

extracts were added to 90 ml of sterile DMSO so that each ml contained the approximate

equivalent of a 1-d exposure (1.e., representing the amount of bioavailable residues extracted

from site water by a standard semipermeable membrane devices in one days’ time) when

added to 1 lof water.

Assuming lentic conditions, an ambient temperature between 10 and 18°C, and minimal

biofouling, 1-gram equivalent of a standard 1-ml triolein semipermeable membrane devices

will clear or extract hydrophobic organic contaminants (e.8., PAH, PCB, organochlorines,

pyrethroids) from about 0.01 to 2.0 l/g of water daily (HuCKINS et al., 2002). Although much

greater variability exists in the uptake rate constants of aquatic organisms for the same

chemicals, the values for invertebrates and fishes generally range from 0.03 to 8.0 Ld''g!

(Mackay et al., 1991; 1992; 1997). Thus, it is reasonable to expect that aliquots of semiper-

meable membrane device extracts, which represent the daily volume of water extracted by a

Source : MNHN, Paris

BRIDGES & LITTLE 133

whole standard semipermeable membrane devices, are representative of the amounts of

chemicals to which aquatic organisms (e.g., tadpoles) are exposed daily.

COLLECTION SITES AND CONDITIONS

Because our semipermeable membrane devices were deployed during various times of

the year, the extracts may not be composed of the same compounds that might be available

during the spring, when many amphibian species breed. However, semipermeable membrane

devices are designed to sample very persistent contaminants (PETrY et al., 2000), and whereas

the concentrations of the waterborne lipophilic contaminants may change with increased

runoff, etc., the composition of these contaminants would likely change very little. Conse-

quently, sampling the water in the late summer/early fall likely represents a conservative effects

assessment. Regardless of when our semipermeable membrane devices were deployed, the

time of deployment coincides with at least partial development of most species in these

habitats.

We attempted to choose sites within each park or national refuge that were disparate in

their habitats (e.g., one high and one low elevation), initially wishing to select one site that

would be more contaminated than the other. Species used within two of the three tests are

species found within our sampling area. In one case (1.e., Alaskan sites), however, this was not

possible and we used a surrogate species.

Sequoia and Kings Canyon National Park

The first site within Sequoia and Kings Canyon National Park was at Yucca Creek, just

east of the western border of the park. Itis a mostly shaded, slowly flowing, clear stream with

a rocky bottom. Semipermeable membrane devices were deployed at approximately 1 m

underwater. Water temperature was 16°C when the devices were deployed on 12 September

2000 and when they were retrieved on 11 October 2000. Upon retrieval there was no evidence

that the devices had been disturbed.

The second site within Sequoia and Kings Canyon National Park was in Aster Lake, a

clear, lentic alpine lake at about 2800 m elevation approximately 8 km west of the Wolverton

trail head within the park. The site is not shaded and the semipermeable membrane devices

rested 1.5 m underwater. The water temperature was 10°C when the devices were deployed on

13 September 2000 and 12°C upon retrieval on 12 October 2000. Upon retrieval there was no

evidence that the devices had been disturbed.

Big Bend National Park

The first site within Big Bend National Park was in the Rio Grande River approximately

three miles from Castalon. The water, which was unshaded, muddy and had a moderate flow,

°C upon deployment of the devices on 4 April 2001. We are uncertain as to the exact

depth of deployment but anticipate it was at least 2 m. Upon deployment, the amount of UVA

transmitted through the water column at a depth of 10 em was 9,700 uW/cm° and the UVB

was 10.3 uW/em° when measured with a Macam Photometrics broadband UV meter. Upon

Source : MNHN, Paris

134 ALYTES 22 (3-4)

retrieval on 10 May 2001 the water temperature was 24°C and the devices had a substantial

amount of mud and grime (ï.e., biofouling) coating them, but there was no evidence of them

having been disturbed.

The second site within Big Bend National Park was at Lower Cattail Falls, at the end of

a trailhead, the entrance of which is across the street from the Sam Nail Ranch. This site is a

shaded clear flowing stream and the device was anchored in 1 m of water. The water

temperature was 19.8°C when the devices were deployed on 5 April 2001 and was 21°C when

the devices were retrieved on 10 May 2001. Upon deployment, the amount of UVA transmit-

ted through the water column at a depth of 10 cm was 1,182 uW/cm? and the UVB was 40.4

uW/cm? when measured with a Macam Photometrics broadband UV meter. There was no

evidence of biofouling or disturbance.

Kenai National Wildlife Refuge

Both sites within the Kenai National Wildlife Refuge are located within the Swanson

River unit. Both are boggy, unshaded wetlands adjacent to oil fields. All devices were

deployed on 11 July 2001 and retrieved on 8 August 2001. At Oil Field 1 (PARSONS, 2001) the

water temperature was 16°C when the devices were deployed and 21°C when they were

retrieved. Upon deployment, the amount of UVA transmitted through the water column at a

depth of 10 em was 2,190 uW/cm? and the UVB was 22.6 uW/cm° when measured with a

Macam Photometrics broadband UV meter. At our second site, Oil Field 3 (PARSONS, 2001),

the water temperature was 18°C at deployment and 19°C at retrieval. Upon deployment, the

amount of UVA transmitted through the water column at a depth of 10 cm was 2,940 uW/cm°?

and the UVB was 46.2 uW/cm°. When each set of devices was collected, no biofouling had

occurred and there was no evidence of tampering.

EXPOSURE TO SEMIPERMEABLE MEMBRANE DEVICE EXTRACTS

Exposures were carried out over two years due to space and time limitations. In June

2001, Pacific treefrog (Pseudacris regilla) tadpoles from three egg masses were received from

Sunriver, Oregon to be used in tests with extracts from Sequoia and Kings Canyon National

Park. The exposure began on 8 June 2001 and was completed 40 d later when the last tadpole

reached metamorphosis. In April 2002, three spring peeper (Pseudacris crucifer) egg masses

were collected from a farm pond in Boone County, Missouri to use in the tests of extracts

from Kenai National Wildlife Refuge. This exposure began on 20 April 2002 and ended 51 d

later when the last tadpole reached metamorphosis. We selected the spring peeper as a

surrogate test species for the native wood frog because of its availability from an uncontami-

nated habitat, its comparatively short developmental time to metamorphosis, and its high

sensitivity to environmental contaminants. BIRGE et al. (2000) ranked the spring peeper as

highly sensitive based upon acute toxicity tests of 34 inorganic elements and 27 organic

chemicals. In these tests the spring peepers consistently exhibited lower tolerance for such

exposures than any of the ranid species tested. Thus, any results obtained using this species

should yield a conservative estimate of effects that wood frogs would experience. In June 2002,

canyon treefrog (Hyla arenicolor) tadpoles were collected from Big Bend National Park to be

Source : MNHN, Paris

BRIDGES & LITTLE 135

used in exposures to extracts from that park. The exposure began on 21 June 2002 and was

completed 45 d later when the final tadpole reached metamorphosis.

The experiment was designed as a 2 X 5 complete factorial: tadpoles were exposed to one

of two UV light treatments and to one of five semipermeable membrane device treatments

(see below), replicated three times.

Test solutions were created by filling 3-1 beakers with 2 1 of well water, and adding 1.0 ml

of the appropriate semipermeable membrane device extract treatment (hereafter, semiper-

meable membrane devices treatment) or field blank semipermeable membrane devices. We

also used a well water (pH 7.8; hardness 286 mg/l CaCO;; alkalinity 258 mg/l CaCO;) control.

Our high UV treatment (16.2 :W/em? UVB) was achieved by wrapping the sides of each

chamber in polycarbonate plastic (0.030 inch thickness, Cope Plastics, Inc., St. Louis,

Missouri) and covering the top of the chamber with two pieces of cellulose acetate (0.015 inch

thickness, Cope Plastics, Inc., St. Louis, Missouri) and one piece of and shade cloth (50 %

shading, Lowe’s Home Improvement Center). The low UV treatment (<1 :W/em? UVB) was

created by wrapping the sides of the chambers with Mylar D (0.005 inch thickness, Cope

Plastics, Inc., St. Louis, Missouri) and covering the tops of the chambers with two pieces of

polycarbonate plastic and one piece of shade cloth.

Prior to testing, the irradiance level of each treatment was measured with a spectrora-

diometer at a 10-em depth in a solar-simulating chamber, as described in LITTLE & FABACHER

(1996). The high irradiance value we used throughout the experiment in the solar simulator

fell below the mean value we measured in the field at a 10 cm depth. This corresponds to the

depth at which many egg masses are found and where they can be found thermoregulating

and feeding. The simulator was programmed on a 16L:8D photoperiod. UVB lights were

activated five hours into the light cycle for five hours to simulate midday solar intensity. This

lighting regime provides for the induction of cellular photorepair functions and prevents the

over-estimation of UV-induced injuries.

Groups of three tadpoles were placed in the 3-1 jars randomly arranged under a solar

simulator in a 20°C flow-through water bath. Tadpoles were exposed to one of five semiper-

meable membrane device treatments: (1) a well water control; (2) semipermeable membrane

device extract from the first site in the refuge/national park; (3) field blank extract from the

first site; (4) semipermeable membrane device extract from the second site within the

refuge/national park; and (5) field blank extract from the second site. Each time water was

changed (i.e., every third day), 1 ml of the appropriate extract (the equivalent to a 2-d dose)

was added to the jars. Tadpoles were fed ground fish flakes (Tetra-Min” brand) ad libitum at

each test water change until metamorphosis, which was defined as the emergence of at least

one forelimb (stage 42, GOsnER, 1960). At metamorphosis individuals were removed from the

experimental chamber and housed in the laboratory until tail resorption (about 4 d, or stage

46, GOSNER, 1960), when they were weighed to the nearest 0.1 mg.

STATISTICAL ANALYSES

Values for the three tadpoles in each replicate chamber were pooled to attain a single data

point. For each location we used analysis of variance (ANOVA) to determine whether mass at

Source : MNHN, Paris

136 ALYTES 22 (3-4)

metamorphosis and the length of the larval period were dependent on semipermeable

membrane device extract treatment, UV treatment, or the interaction of these factors. Both

endpoints were log-transformed to increase normality. In analyses of mass at metamorphosis,

length of the larval period was used as a covariate and vice versa because these two variables

can be correlated. In the instances where the type I sums of squares of these covariates were

not significant, they were removed from the model. When there were significant differences

among semipermeable membrane device extract treatments, Bonferroni multiple comparison

tests were conducted to discern differences among specific treatments at # = 0.05. Differences

in survival were minimal (i.e., nearly all animals survived through metamorphosis), so no

analyses were conducted on this endpoint.

RESULTS

In Sequoia and Kings Canyon National Park, the effect of extract treatment on the size

at metamorphosis of tadpoles was marginally significant (tab. 1). Individuals reared in the

water controls were significantly larger at metamorphosis than tadpoles in extracts from

Yucca Creek (fig. 1). However, Yucca Creek tadpoles were not significantly smaller than

tadpoles reared in extracts from Aster Lake, or in either of the field blank treatments. The

effect of extract treatment on the length of the larval period was not significant. However,

when the effects of UV (which were not significant) were removed from the model, the main

effect of semipermeable membrane device extract treatment became marginally significant

(E,24 = 2,58; P = 0.0634). Tadpoles reared in SPMD extracts from Aster Lake and in extracts

from the Aster Lake field blank took longer to reach metamorphosis than those from any

other treatment (fig. 2). Neither UV treatment alone nor the interaction between the UV

treatment and the semipermeable membrane device extract treatment significantly affected

the length of the larval period (tab. 1) or the mass at metamorphosis.

In Big Bend National Park, semipermeable membrane device extract treatment signifi-

cantly affected the number of days it took for tadpoles to reach metamorphosis (tab. 1).

Individuals reared in water containing extracts from the Rio Grande River took significantly

longer to transform than individuals reared in field blanks from either site, or controls (fig. 3).

The length of the larval period in tadpoles reared in water from Lower Cattail Falls was not

significantly different than those from the Rio Grande River, and did not differ from controls

or field blanks from either site (fig. 3).

At Kenai National Wildlife Refuge, semipermeable membrane device extract treatment

significantly affected the length of the larval period (tab. 1). Individuals reared in semiper-

meable membrane device extracts from Oil Field 3 took significantly longer to reach meta-

morphosis than individuals reared with Oil Field 3 field blanks, Oil Field 1 semipermeable

membrane device extracts, or the control (fig. 4). The length of the larval period was not

significantly affected by UV treatment or the interaction between semipermeable membrane

device treatment and UV. Semipermeable membrane device extr: from Kenai National

Wildlife Refuge also significantly affected the size at metamorphosis for tadpoles. Individuals

reared in extracts from Oil Field 3 were significantly smaller than those reared in any other

Source : MNHN, Paris

BRIDGES & LITTLE

137

Table 1. — Analyses of variance on the effects of SPMD extract (extract), UV radiation (UV) and

their interaction on the length of the larval period (days) and the mass at time to

metamorphosis (mass) for amphibian larvae. Type III mean squares (MS) are reported.

BBNP, Big Bend National Park; KNWR, Kenaï National Wildlife Refuge; SEKI, Sequoia

Kings Canyon National Park.

Response variable Source df MS F P

Days (SEKI) Extract 4 0.0053 1.80 0.1705

uv 1 0.0013 047 0.5032

Extract x UV 4 0.0010 0.35 0.8403

Error 19 0.0029

Mass (SEKI) Extract 4 0.0509 2.67 0.0637

UV : 0.0174 0.92 0.3505

Extract x UV 4 0.0086 0.45 0.7690

Error 19 | 00190

Days (BBNP) Mass 1 0.0058 2.02 01718

Extract 4 0.0122 425 0.0126

Uv 1 0.0022 0.76 03930

Extract x UV 4 0.0014 0.48 0.7481

Error 19 | 0.0029

Mass (BBNP) Days 1 0.0295 1.72 02053

Extract 4 0.0217 1.57 0.2223

UV ll 0.0035 0.21 0.6549

Extract * UV 4 0.0019 0.12 0.9751

Error 19 0.0172

Days (KNWR) Mass 1 32.39 6.15 0.0227

Extract 4 15.61 2.96 0.0464

UV 1 1.19 0.23 0.6394

Extract x UV 4 7.99 1.52 02371

Error 19 5.26

Mass (KNWR) | Days 1 0.08 631 0.0212

Extract 4 0.05 3.72 0.0212

UV ll 0.00 0.07 0.7887

Extract x UV 4 0.03 2.25 0.1020

Error 19 0.01

Source : MNHN, Paris

138 ALYTES 22 (3-4)

0.50 -

0.45 | L

0.40 -

0.35 - B B

——

0.30 -

0.25 -

0.20 :

Mass at metamorphosis (g)

0.15 5 1

WC YC YCFB AL ALFB

Site

Fig. 1. — Mass at metamorphosis for Pseudacris regilla tadpoles reared in each treatment for Sequoia

Kings Canyon National Park. AL, Aster Lake; ALFB, Aster Lake Field Blank; WC, water control;

YC, Yucca Creek; YCFB, Yucca Creek Field Blank. Vertical lines represent £ 1 5,.

37 - |

35 -

33 -

30 1

WC YC YCFB AL ALFB

Site

Fig. 2. - Time to metamorphosis in days for Pseudacris regilla tadpoles reared in each treatment for

quoia Kings Canyon National Park. AL, Aster Lake: ALFB, Aster Lake Field Blank: WC, water

control; YC, Yucca Cr YCFB, Yucca Creek Field Blank. Bars with the same letters do not differ

significantly from one another. Vertical lines represent + 1 5,

Time to metamorphosis (days)

n 4

Source : MNHN, Paris

BRIDGES & LITTLE 139

& 35

TZ 34

a 33 an

5: 3

5 31 à

É . 8 I 8

à i

£ 29

£ 28

Ê 27

F 264 1

WC CF CFFB RG RGFB

Site

Fig. 3. — Time to metamorphosis in days for Hyla arenicolor tadpoles reared in each treatment for Big

Bend National Park. CF, Cattail Falls; CFFB, Cattail Falls Field Blank; RG, Rio Grande; RGFB,

Rio Grande Field Blank; WC, water control. Bars with the same letters do not differ significantly

from one another. Vertical lines represent + 1 5,.

ma 41 B

LL]

Z 40

ü 39

É AB

2 38 A

Ê 37 A

[0]

£ 36 |

FF 34 à

WC OF1 OF1FB OF3 OF3FB

Site

Fig. 4. Time to metamorphosis in days for Hyla arenicolor tadpoles reared in each treatment for Kenai

National Wildlife Refuge. OFI, Oil Field 1; OFIFB, Oil Field 1 Field Blank: OF3, Oil Field 3;

OF3FB, Oil Field 3 Field Blank; WC, water control. Bars with the same letters do not differ

significantly from one another. Vertical lines represent + 15,

Source : MNHN, Paris

140 ALYTES 22 (3-4)

se 0.25 -

S “ A

[2] A

2 0.23 - |

S A

6 0.21 - Î

S |

Ê 0.19 - i

[4]

0.15 1

WC OF1 OF1FB OF3 OF3FB

Fig. 5. - Mass at metamorphosis in days for Pseudacris crucifer tadpoles reared in each treatment for

Kenai National Wildlife Refuge. OF001, Oil Field 1; OF001FB, Oil Field 1 Field Blank; OF003,

Oil Field 3; OF003FB, Oil Field 3 Field Blank; WC, water control. Bars with the same letters do not

differ significantly from one another. Vertical lines represent + 1 s,.

treatment (fig. 4). There was no effect of UV radiation or the interaction between UV

radiation and semipermeable membrane device extract treatment on the size at metamor-

phosis.

DISCUSSION

Utilizing semipermeable membrane device extracts in amphibian bioassays has been

shown to be an effective tool in determining whether lipophilic compounds found in

aquatic amphibian habitats are toxic (BRIDGES et al., 2004) without actually having to

determine which compounds are present. Importantly, using semipermeable membrane

devices rules out detrimental effects attributable to parasites or pathogens because the pore

size of the membrane is too small (1.e., 10 À) to allow passage of bacterial, viral or fungal

cells. Although we did not attempt to determine which compounds were present in the

extracts used in our experiments, once the presence of contaminants in the environment are

confirmed in a bioassay it would be possible to resample the sites and generate a contaminant

profile.

Despite the fact that Sequoia and Kings Canyon National Park is a protected, natural

area, the presence of contaminants has been revealed throughout the park. Surveys of the

Source : MNHN, Paris

BRIDGES & LITTLE 141

Sierra Nevada mountains around Sequoia and Kings Canyon National Park indicate the

presence of chlorpyrifos, malathion, diazanon and DDT in water (ZABIk et al, 1993;

MCCOoNKELL et al., 1998) and in frog tissue (Cory et al., 1970) even at high altitudes within the

park. SPARLING et al. (2001) revealed reduced cholinesterase levels of native amphibians in

Sequoia Kings Canyon National Park, suggesting the presence of cholinesterase-inhibiting

pesticides. DAvipsoN et al (2001, 2002) correlated the amount of downwind pesticide use in

California’s Central Valley with declines in amphibian populations within the park. LENOIR

et al. (1999) have shown the presence of diazanon and chlorpyrifos in atmospheric transport

at elevations where native amphibians are showing the greatest declines.

Our data using Pacific treefrogs (2 regilla, which are native to the Sequoia and Kings

Canyon National Park region) suggests the presence of a compound in the water at Yucca

Creek that generates a smaller size at metamorphosis, which can have a number of fitness

consequences (ALTWEGG & REYER, 2003). Individuals that are large at metamorphosis can

realize higher overwintering success, greater survival to first reproduction, and earlier repro-

duction (SMITH, 1987; SEMLrTSCH et al., 1988). Further, larger females can carry more eggs,

and larger males often gain access to a greater number of females during breeding, leading

to increased reproductive success (BERVEN, 1982). The fact that there were no significant

differences when comparing Yucca Creek with the other treatments may suggest the presence

of a compound in all of our semipermeable membrane device extracts and field blanks that

depresses growth (although not detectable significantly in the other treatments).

There was a trend for frogs reared in extracts from Aster Lake and the Aster Lake field

blank to take longer to reach metamorphosis. This effect was marginally significant only after

the removal of UV effects, perhaps because our statistical power was low. That reductions

were observed both in the field blank and the waters extracts from this high elevation site

suggests the presence of a compound that is airborne and/or waterborne. This is in agreement

with data that indicate the presence of contaminant compounds at high elevations in the

Sierra Nevadas suspected of causing declines (LENOIR et al., 1999). Frogs delaying metamor-

phosis can suffer some of the same fitness detriments as animals undergoing metamorphosis

at a small size.

In 2002, Big Bend National Park was designated an endangered park primarily due to its

increasingly high levels of human impacts over the past 15 years. Poor air quality attributable

to drift from coal-fired power plants is common. Increasing aquatic pollution and water

diversion upstream has caused the water quality to decline in the Rio Grande River, which is

the park’s most prominent source of water. In fact, low water levels lead to concentration of

aquatic pollutant from upstream municipal and agricultural sources in the US and Mexico,

further worsening the quality of aquatic habitat in this river.

Metamorphosis was delayed among canyon treefrog (H. arenicolor) tadpoles, which are

native to Big Bend National Park, developing in water containing extracts from the Rio

Grande River. Levels of lipophilic contaminants (including DDT) have been detected at

sampling sites upstream and downstream of the park (ScHMirr et al., 2004) and in biota

feeding in this stretch of the river (Mora et al., 2002). TI uggests that amphibian larvae

developing in the Rio Grande, or in pools formed by its floodwaters, are at a potential

disadvantage which may result in decreased fitness. Tadpoles reared in extracts from Cattail

Falls did not take significantly longer to develop than tadpoles in the Rio Grande, but also did

Source : MNHN, Paris

142 ALYTES 22 (3-4)

not differ from the controls. No effects were seen due to extracts from the field blanks,

indicating that the contaminants sampled at each site were waterborne.

The Kenai National Wildlife Refuge covers nearly 2 million square acres on Alaska’s

Kenai Peninsula, south of Anchorage. In recent years, activities have taken place on the Kenai

National Wildlife Refuge that have increased chemical contamination on this site including

oil and gas development, pesticide application, military and recreational uses, mining, and use

of fire retardants (PARSONS, 1991). Many of these sites were impacted with little or no post-

contamination remediation efforts undertaken, primarily because of Alaska’s remoteness. We

selected two ponds on the Kenai National Wildlife Refuge adjacent to oil drilling operations

and that also contain populations of wood frogs (Rana sylvatica).

At Kenai National Wildlife Refuge, frogs exposed to semipermeable membrane device

extract from the pond at Oil Field 3 were smaller at metamorphosis and took longer to reach

metamorphosis than frogs exposed to any other treatment. Why there were effects at Oil Field

3 and not Oil Field 1, which are both adjacent to oil drilling operations, is unclear. The fact

that the growth and development of the frogs reared in extract from the Oil Field 3 field

blank did not differ from the water control indicates that the contaminant in question is

aquatic and not airborne. Whereas a number of petroleum hydrocarbon spills have been

documented in the Swanson River Oil Field (PARSONS, 2001), where our two sites were

situated, it is unclear whether any of these historic spills occurred near our study sites.

Without chemical analysis of the semipermeable membrane device extracts, it is not known

what chemicals may have elicited this biological response, nor whether this phenomenon is

related to oil field operations. Hydrocarbons, among other chemicals, have been shown to be

made more toxic in the presence of UV radiation (ZAGA et al., 1998; Lrrrie et al., 2000).

Because there was not a significant interaction between the semipermeable membrane device

extracts and UV radiation, it is likely that hydrocarbons are not a significant source of

contamination at these two sites.

We did not observe any effects of UV radiation alone within our study, suggesting that

the intensities of UV which were measured in the field are not high enough to singly harm

amphibians, There was also no interaction between UV and semipermeable membrane device

extracts from any site. This indicates that the compounds present in these waters are not

broken down or photoactivated to be made less or more toxic, respectively.

CONCLUSIONS

Habitat decimation has been cited as being the major contributor of amphibian declines,

but when declines are occurring in protected areas we are forced to consider other, more

insidious, factors such as contamination. Often, habitats that appear pristine can be unfavor-

able for amphibians and other aquatic organisms. In our study, alterations in amphibian life

history characteristics were evident within all three of the parks and wildlife refuges exam-

ined. Even if contaminants do not cause outright mortality, alterations in life history

characteristics have the potential to alter population structure over time. For example,

if postponing metamorphosis delays age at first reproduction, population growth rates

Source : MNHN, Paris

BRIDGES & LITTLE 143

potentially decrease (STEARNS, 1992). If in a contaminated environment there were a shift

toward later reproduction in an amphibian population due to delayed maturity, the

demographic structure of the population may be altered, resulting in a gradual decline in size

over time. Utilizing semipermeable membrane devices offers a way to sample aquatic habitats

for contaminants in a non-destructive manner, and can be used to help assess the health of

aquatic amphibian habitats.

RÉSUMÉ

Les déclins des populations d'amphibiens ne sont pas limités aux habitats ouvertement

dégradés, mais s’observent souvent dans des environnements relativement intacts tels que des

parcs nationaux ou des réserves naturelles, forçant ainsi les biologistes à examiner des causes

de déclin moins évidentes a priori, comme la présence de contaminants. L'objectif de notre

étude était d'extraire, au moyen de membranes semi-perméables, des composés présents dans

des habitats naturels d'amphibiens dans trois pares nationaux ou réserves naturelles (Sequoia

Kings Canyon National Park, Big Bend National Park et Kenai National Wildlife Refuge;

deux sites étudiés pour chacun), et d'évaluer leur toxicité par des tests biologiques sur des

larves. Les extraits ne causèrent pas de mortalité, et tous les effets observés furent sublétaux,

modifiant des caractéristiques du déverloppement. Dans les trois régions étudiées, les amphi-

biens élevés avec des extraits provenant d'au moins un des deux sites ont manifesté soit un

allongement de la période larvaire soit une réduction de la masse à la métamorphose. Les

provenant de l’air et de l’eau d’un site du Sequoia Kings Canyon National Park

allongèrent la période larvaire, ce qui est en accord avec les résultats d’études montrant des

niveaux élevés de contaminants transportés par l'air dans de tels sites au sein du parc. Les

radiations ultraviolettes, qui ont également été soupçonnées d’avoir causé des déclins

d'amphibiens, ne se sont pas avérées agir seules ou altérer la toxicité des extraits.

extrail

ACKNOWLEDGMENTS

Our thanks go to H. Werner, R. Skiles, J. Goldstein, $. Olkson, D. Hughes, R. Calfee, J. Connor and

JL. Bridges, for assisting in deploying and/or retrieving devices. J. Bowerman, G. Dayton and T. Timock

helped in collecting and/or shipping tadpoles 10 be used in studies. R. Clark, W. Cranor, J. Huckins and

1 Petty assisted us with assembling and processing the semipermeable membrane devices. $. Saura Mas,

1 Little, JL Wells and M. Boone helped with the set-up and maintenance of the tests. This manuscript was

improved with the thoughtful comments of $. James.

LITERATURE CITED

ALTWEGG, R. & REYER, H-U., 2003. - Patterns of natural selection on size at metamorphosis in water

frogs. Evolution, 57 (4): 872-882

Source : MNHN, Paris

144 ALYTES 22 (3-4)

ANKLEY, G. T., TIETGE, J. E., HOLCOMBE, G. W., DEFOE, D. L., DIAMOND, S. A. JENSEN, K.. M. & DEGITZ,

S. J., 2000. — Effects of laboratory ultraviolet radiation and natural sunlight on survival and

development of Rana pipiens. Can. J. Zool., 78: 1092-1100.

BERVEN, K. A., 1982. - The genetic basis of altitudinal variation in the wood frog Rana sylvatica. 1. An

experimental analysis of life history traits. Evolution, 36: 962-983.

BIRGE, W. J., WESTERMAN, A. G. & SPOMBERG, J. A., 2000. - Comparative toxicology and risk assessment

of amphibians. /n: D. W. SPARLING, G. LINDER & C. A. Bishop (ed.), Ecotoxicology of Amphibians

and Reptiles, Pensacola, Florida, SETAC Press: 727-792.

BLAUSTEIN, A. R., KIESECKER, J. M., CHIvERS, D. P., Hokir, D. G., MARCO, A., BELDEN L. K. & HATCH,

A., 1998. — Effects of ultraviolet radiation on amphibians: field experiments. Am. Zool., 38:

799-812.

BraDroRD, D. F., GRABER, D. M. & TaBaTABAI, F., 1994. - Population declines of the native frog, Rana

muscosa, in Sequoia and Kings Canyon National Parks, California. Southwest. Nat., 39: 323-327.

BRADFORD, D. F., TABATABAI, F. & GRABER, D. M., 1993. — Isolation of remaining populations of the

native frog, Rana muscosa, by introduced fishes in Sequoia and Kings Canyon National Parks,

California. Conserv. Biol., 7 (4): 882-888.

BRIDGES, C. M. & LiTTLr, E. E., 2003. - Using semipermeable membrane devices (SMDPs) to assess the

toxicity and teratogenicity of aquatic amphibian habitats. /n: G. LINDER, S. KREST, D. SPARLING &

E. LITTLE (ed.), Multiple stressor effects in relation to declining amphibian populations: ASTM STP

1443, West Conshohocken, Pennsylvania, ASTM International.

BRIDGES, C. M., LITTLE, E. E., GARDINER, D. M., PETTY, J. D. & HUCKINS, J. N., 2004. - Assessing the

toxicity and teratogenicity of pond water in Minnesota to native amphibian larvae. Environ. Sci.

Poll. Res., 11: 233-239.

BROOMHALL, S. D., OSBORNE, W. S. & CUNNINGHAM, R. B., 200

ultraviolet-B radiation on two sympatric species of Australian frogs. Conserv. Biol, 14: 420-427.

CAREY, C., 1993. - Hypothesis concerning the causes of the disappearance of boreal toads from the

mountains of Colorado. Conserv. Biol., 7: 355-362.

Cor, P. S., 1994. - What we know and don't know about amphibian declines in the west. Jr: W. W.

CoviNGToN & L. F. DEBANO (ed.), Sustainable ecological systems: implementing an ecological

approach to land management, USDA Forest Service Technical Report RM-247: 59-67.

Cory, L., FIELD, P. & SERAT, W., 1970. — Distribution patterns of DDT residues in the Sierra Nevada

Mountains. Pest. Mon. J., 3: 204-211.

DavipsON, C., SHAFFER, H. B., & JENNINGS, M. R., 2001. - Declines of the California red-legged frog:

climate, habitat, and pesticide hypotheses. Ecol. Appl., 11: 464-479.

2002. — Spatial tests of the pesticide drift, habitat destruction, UV-SS, and climate-change hypotheses

for California amphibian declines. Conserv. Biol., 16: 1588-1601.

FELLERS, G. M. & Drosr, C. A., 1993. - Disappearance of the Cascades frogs, Rana cascadae, at the

southern end of its range. Biol. Conserv., 65: 177-181.

Gosner, K. L., 1960. — A simplified table for staging anuran embryos and larvae with notes on

identification. Herpetologica, 16: 183-190.

HUCKINS, J. N., PETTY, J. D., PREST, H. F., CLARK, R. C., ALVAREZ, D. A., ORazio, C. E., LEBO, J. A.,

CRANOR, W. L. & JOHNSON, B. T., 2002. — À guide for the use of semipermeable membrane devices

({SMDPSs) as samplers of waterborne hydrophobic organic contaminants, Report prepared for the

American Petroleum Institute (API), API Nr. 4690: 1-186.

JENNINGS, M. R. & HAYES, M. P., 1994. — Amphibian and reptile species of special concern in California.

Report prepared for California Department of Fish and Game, Inland Fisheries Division, Rancho

Cordova, California.

JUNG, R. E., BONINE, K. E., ROSENSHIELD, M. L., DE LA REZA, A., RAIMONDO, $. & DROGE, S., 2002. —

Evaluation of canoe surveys for anurans along the Rio Grande in Big Bend National Park, Texas.

J. Herp., 36 (3): 390-397.

KNaPP, R. A. & MATTHEWS, K. R., 2000. - Non-native fish introductions and the decline of mountain

yellow-legged frogs from within protected areas. Conser. Biol., 14: 428-438.

LENOIR, J. S.. MCCONNELL, L. L., FELLERS, M. G., CaHILL, T. M. & SFIBER, J. N., 1999. - Summertime

transport of current-use pesticides from California’s Central Valley to the Sierra Nevada mountain

range, USA. Environ. toxicol. Chem., 18: 2715-2722.

— Comparative effects of ambient

Source : MNHN, Paris

BRIDGES & LITTLE 145

Lrrree, E.E., CALFEE, R., CLEVELAND, L., SKINKER, R., ZAGA-PARKHURST, À. & BARRON, M. G., 2000.—

Photo-enhanced toxicity in amphibians: synergistic interactions of solar ultraviolet radiation and

aquatic contaminants. J. Jowa Acad. Sci., 107: 767-771.

Lrrree, E. E. & FABACHER, D. L., 1996. - Exposure of freshwater fish to simulated solar UVB radiation.

In: G. OSTRANDER (ed.), Techniques in aquatic toxicology, New York, New York, Lewis Publishers:

141-158.

LizanA, M. & PEDRAZA, E. M., 1998. - The effects of UV-B radiation on toad mortality in mountainous

areas of central Spain. Conserv. Biol., 12: 703-707.

Mackay, D., Smiu, W.-Y. & MA, K.-C., 1991. - Hlustrated handbook of physical-chemical properties and

environmental fate for organic chemicals. Volume 1. Chelsea, Michigan, Lewis Publishers: 1-704.

_S 1992. — Iustrated handbook of physical-chemical properties and environmental fate for organic

chemicals. Volume 2. Chelsea, Michigan, Lewis Publishers: 1-608.

2h 1997. — Justrated handbook of physical-chemical properties and environmental fate for organic

chemicals. Volume 5. Chelsea, Michigan, Lewis Publishers: 1-832.

MCALLISTER, K. R., LEONARD, W. P. & SroRM, R. M., 1993. - Spotted frog (Rana pretiosa) surveys in the

Puget trough of Washington, 1989-1991. Northwest. Nat., 74: 10-15.

McConxeLL, L. L., LENOIR, J. $., DaTTA, S. & SeIBER, J. N., 1998. - Wet deposition of current-use

pesticides in the Sierra Nevada mountain range. Environ. toxicol. Chem., 17: 1908-1916.

MorA, M. A. SKILES, R., MCKINNEY, R., PAREDES, M., BUCKLER, D. & PAPOULIAS, D. K. D., 2002. -

Environmental contaminants in prey and tissues of the peregrine falcon in the Big Bend Region,

Texas, USA. Environ. Poll., 116: 169-176.

PARSONS, T. A. S., 2001. - Kenai National Wildlife Refuge Contaminant Assessment. US Fish and Wildlife

Service unpublished report: 1-123.

Perry, J. D., JONES, S. B., HUCKINS, J. N., CRANOR, W. L., PARRIS, J. T., MCTAGUE, T. B. & BOYLE, T. P.,

2000. - An approach for assessment of water quality using semipermeable membrane devices

(SMDPs) as bioindicator tests. Chemosphere, 41: 311-321

ScHmrrT, C.J., DETHLOFF, G. M., HINCK, J. E., BARTISH, T. M., BLAZER, V. S., COYLE, J. J., DENSLOW,

N. D. & Tiurrr, D.E., 2004. - Biomonitoring of environmental status and trends (BEST) program:

environmental contaminants and their effects on fish in the Rio Grande Basin, US Geological Survey,

Scientific Investigations Report 2004-5108: 1-117.

SEMLITSCH, R. D., SCOTT, D. E. & PECHMANN, J. H. K., 1988. - Time and size at metamorphosis related

10 adult fitness in Ambystoma talpoideum. Ecology, 69: 184-192.

SmirH, D. C., 1987. - Adult recruitment in chorus frogs: effects of size and date at metamorphosis.

Ecology, 68: 344-350.

SParLING, D. W., FELLERS, G. M. & MCcConneLu, L. L., 2001. — Pesticides and amphibian population

declines in California, USA. Environ. toxicol. Chem., 20: 1591-1595.

STARNES, S. M., KENNEDY, C. A. & PETRANKA, J. W., 2000. — Sensitivity of embryos of southern

Appalachian amphibians to ambient solar UV-SS] radiation. Conserv. Biol, 14: 277-282.

STEARNS, S. C., 1992. — The evolution of life histories. New York, New York, Oxford University Press:

1-249.

Zaik, J. M. & SeiBR, J. N., 1993. — Atmospheric transport of organophosphate pesticides from

California’s Central Valley to the Sierra Nevada Mountains. J. environ. Qual., 22: 80-90.

ZAGA, A., Lire, E. E., RABENI, C. F. & ELLERSIECK, M. R., 1998. — Photoenhanced toxicity of a

carbamate insecticide to early life stage anuran amphibians. Environ. toxicol. Chem., 17: 2543-

2553.

Corresponding editor: C. Kenneth Dobp, Jr.

Source : MNHN, Paris

Alytes, 2005, 22 (3-4): 146-167.

Climate patterns as predictors

of amphibian species richness

and indicators of potential stress

William BATTAGLIN*, Lauren HAy**, Greg MCCABE**,

Priya NANJAPPA#*## & Alisa GALLANT**#*

* US Geological Survey, Box 25046, MS 415, Denver Federal Center, Lakewood, Colorado 8022:

**_ US Geological Survey, Box 25046, MS 412, Denver Federal Center, Lakewood, Colorado 80225, USA

*#* US Geological Survey, 12100 Beech Forest Road, Laurel, Maryland 20708, USA

#*** US Geological Survey, 47914 252" Street, Sioux Falls, South Dakota 57198, USA

Amphibians occupy a range of habitats throughout the world, but

species richness is greatest in regions with moist, warm climates. We

modeled the statistical relations of anuran and urodele species richness with

mean annual climate for the conterminous United States, and compared the

strength of these relations at national and regional levels. Model variables

were calculated for county and subcounty mapping units, and included

40-year (1960-1999) annual mean and mean annual climate statistics,

mapping unit average elevation, mapping unit land area, and estimates of

anuran and urodele species richness. Climate data were derived from more

than 7,500 first-order and cooperative meteorological stations and were

interpolated to the mapping units using multiple linear regression models.

Anuran and urodele species richness were calculated from the United States

Geological Survey’s Amphibian Research and Monitoring Initiative (ARMI)

National Atlas for Amphibian Distributions. The national multivariate linear

regression (MLR) model of anuran species richness had an adjusted coeffi-

cient of determination (R?) value of 0.64 and the national MLR model for

urodele species richness had an R? value of 0.45. Stratifving the United

States by coarse-resolution ecological regions provided models for anurans

that ranged in R? values from 0.15 to 0.78. Regional models for urodeles

had R? values ranging from 0.27 to 0.74. In general, regional models for

anurans were more strongly influenced by temperature variables, whereas

precipitation variables had a larger influence on urodele models.

INTRODUCTION

Amphibian populations appear to be declining worldwide (HOULAHAN et al. 2000:

CaREyY et al., 2001; YOUNG et al., 2004). À number of possible causes of decline have been

proposed including changes in climate ( PounDs & CRUMP, 1994: DONNELLY & CRUMP,

1998: Pounps et al., 1999), increased UV radiation (e.g., ALFORD & RiCHARDS, 1999:

BLAUSTEIN et al., 2003), habitat loss/fragmentation/alteration (e.g., FAHRIG et al., 1995:

Source : MNHN, Paris

BATTAGLIN et al. 147

DEMAYNADIER & HUNTER, 1998; KRZYSIK, 1998; COLLINS & STORFER, 2003), introduction of

nonindigenous competitive species (e.g., HAYES & JENNINGS, 1986; ROSEN & SCHWALBE, 1995:

FIsHEr & SHAFFER, 1996; KIESECKER & BLAUSTEIN, 1998; LAWLER et al., 1999), occurrence of

contaminants (e.g., BERRILL et al., 1993; BONIN et al., 1997; DAvipson et al., 2002; HAYES et

al., 2002), exposure to pathogens (e.g., LAURANCE et al., 1996; KIESECKER & BLAUSTEIN, 1999)

and over-harvesting (KOONTZ, 1992; LANNOO, 1996). Many herpetologists believe that com-

binations of stresses are being placed on amphibian populations (GREEN, 1997; BRITSON &

THRELKELD, 2000; COLLINS & STORFER, 2003; LANNOO et al., 2003: LITTLE et al., 2003).

There is widespread acknowledgment that the global climate is changing (HOUGHTON et

al., 2001). Changes in land cover may also affect climate by altering the physical properties of

the land surface (HAYDEN, 1998; PIELKE et al., 1999, 2002). Short- and long-term changes in

climate have the potential to affect the ranges of individual amphibian species and hence

species richness in any given locality (THoMas et al., 2004). Climatic conditions not only

directly stress amphibian populations (Dopp, 1997; Pouxps et al., 1999; Corn & MUTHS,

2002), but also may influence their resistance to disease or their ability to withstand attacks by

environmental pathogens (CAREY & ALEXANDER, 2003). Water availability, air temperature

and relative humidity can influence amphibian breeding, development, foraging, mobility,

calling, immune response and habitat availability (DONNELLY & CRumP, 1998; GisBs &

BReIsCH, 2001). Climate also can influence the spread of amphibian pathogens (DASZAK et al,

2003; JOHNSON & CHASE, 2004).

Amphibians have a long (-350 million years) history of survival under extremes in global

climate (CAREY & ALEXANDER, 2003), yet their life histories (DUELLMAN, 19994) suggest that

individual amphibian populations may be vulnerable to short-term variations in climate.

Amphibians occupy a range of habitats throughout the world, but species richness is greatest

in regions with moist, warm climates (DUELLMAN, 1999h). Other natural factors that can

affect amphibian species richness include historical lineages, barriers to migration, interspe-

cies competition and the availability of food, shelter and breeding sites.

OBJECTIVES AND SCOPE

This research assesses the degree to which average climatic conditions in the contermi-

nous United States over the last four decades explain historical patterns of amphibian species

richness. The primary objectives of the research were to model the statistical relations of

anuran and urodele species richness with mean annual climate for the conterminous United

States, and to compare the strength of these relations at national and regional levels. Trends

in climatic conditions during this period were also evaluated to determine if they might be

leading towards more stressful conditions for amphibians (e.g., decreases in available breeding

habitat, shortening of breeding season).

There were limitations or biases implicit in the datasets used for these analyses. First, the

species occurrence records incorporated into the ARMI National Atlas for Amphibian

Distributions were not associated with an explicit time period and are best described as an

historic compilation of occurrence records. Second, the species richness estimates were

Source : MNHN, Paris

148 ALYTES 22 (3-4)

compared to climate statistics averaged for 1960-1999. Climate from this time period may not

match longer-term averages or averages from other time periods. Third, the species richness

estimates [http:/www.mp2-pwrc.usgs.gov/armiatlas/] were based on mapping units

(counties/subcounties) that are not of uniform size, resulting in the potential for an inherent

bias towards a larger number of species occurring in larger counties. Fourth, neither the

weather stations nor the spatial variability of weather were uniformly distributed across the

United States, so the quality of information varied from mapping unit to mapping unit. Fifth,

the mean elevation was used for each mapping unit, but the concept of average elevation may

not be useful in mountainous areas. No attempt was made to account for the patchiness of the

distribution of species within a county/subcounty (KiEsTER, 1971), because such data do not

exist for most of the United States, and no attempt was made to account for the effect of

climate variation on species that prey upon amphibians or that amphibians consume. Despite

these limitations, and because strong relationships between climate and ecosystem develop-

ment are widely recognized (WALTER, 1973; FORMAN & GoDRON, 1986; MONSERUD &

LEEMANS, 1992), it was appropriate to expect that relations between climate and amphibian

species richness would emerge from the analysis.

METHODS

SOURCE AND PROCESSING OF AMPHIBIAN DATA

Species richness estimates for anurans and urodeles were derived from the ARMI

National Atlas for Amphibian Distributions (hereafter called “atlas”) [http:/vww.mp2-

pwrc.usgs.gov/armiatlas/], which uses a combination of counties and subcounties as a spatial

framework for documenting the geographic occurrence of the nearly 300 species of amphi-

bians currently recognized in the United States (LANNOO et al., 2005). Counties are used as

mapping units for all but five Western states (Arizona, California, Nevada, Oregon, and

Washington), for which subcounties are used to help overcome the wide disparity in county

sizes across the nation. The atlas is a compilation of both current and historic records of

amphibian occurrences, bounded by no explicit time period. The records are from peer-

reviewed scientific literature, museum vouchers, state and regional herpetological atlases, and

other confirmed and validated observations. Data sources vary by state and are not standard-

ized in their geographic precision. Thus, some records in the atlas may represent assumed

presence, as from a range map, whereas other records represent vouchered specimens with

specific location information. Because the atlas database incorporates a county/subcounty

coding system that follows Federal Information Processing Standards, a geographic informa-

tion system (GIS) was used to link species occurrence records with a digital map of county

and subcounty polygons [http://www.census.gov/geo/www/cob/scale.html]. Species richness

was calculated for anurans and urodeles by tallying the number of species recorded as

occurring Within each map unit.

Source : MNHN, Paris

BATTAGLIN et al. 149

SOURCE AND PROCESSING OF CLIMATE DATA

Estimates of 1960-1999 mean annual and annual mean climate statistics were calculated

from approximately 7,500 National Weather Service first-order and cooperative temperature

stations, and 11,500 National Weather Service first-order and cooperative precipitation

stations. First-order stations are operated by professional staff and report a comprehensive

array of weather variables each hour. Cooperative sites are more numerous, but generally only

make once-daily observations of a few weather variables (e.g., minimum and maximum daily

temperature and precipitation). These data were extracted from the National Climate Data

Center Summary of the Day Dataset and have been quality controlled by the National

Climate Data Center (EisCHEID et al., 2000; CLARK et al., 2004). Estimates of mean annual

and annual mean climate statistics (tab. 1) for each county/subcounty were calculated using

multiple linear regression (MLR) models. The MLR method was used to distribute the

climate statistics (dependent variable) calculated at each station to each county/subcounty

based on the “XYZ” value (longitude X, latitude Y, and mean elevation Z, respectively) of the

county/subcounty polygon centroid (Hay et al., 2000; Hay & MCcaBE, 2002; Hay & CLARK,

2003). The MLR equation [1] was developed for each dependent variable (climate statistic,

“CS”) using the independent XYZ variables from a set of National Weather Service climate

stations:

CS=bix+b,y+b;z+bo[I]

The MLR equations were computed to determine the regression surface that described

the spatial relations between the dependent CS and the independent XYZ variables. Equation

[1] describes a plane in three-dimensional space with slopes b,, b, and b, intersecting the CS

axis at b,. The best MLR equation for each CS did not always include all the independent

variables.

To estimate the climate statistics for each county/subcounty (CNTY), the following

procedures were followed: first, mean daily CS and corresponding mean XYZ values from a

set of stations (STAMEAN) were used with the slopes of the MLR from equation [1] to

estimate a unique y-intercept (bçest, see equation [2]), and second, equation [3] was solved

using the coefficients (b,, b, and b;) from equation [1], best from equation [2], and the XYZ

values of the CNTY.

byest = CS(STAMEAN) - (b, x(STAMEAN) + b, YSTAMEAN) + b, ASTAMEAN)) P]

CS(CNTY) = bgest + b, x(CNTY) + b, YICNTY) + b; Z(CNTY) [3]

The set of stations comprising the STAMEAN in each calculation were chosen from the

20 closest stations to the CNT Y. Outliers (i.e., stations determined to be too far away from the

data site or residing in another physiographic region) were not used in the STAMEAN

calculation. The same MLR equations are used but the time series of mean daily CS and their

corresponding mean XYZ values are obtained from station data to estimate a unique best.

Thus, the slope of the MER for the CS remained constant, but the y-intercept changes based

on the mean CS and XYZ values.

Trends in climate were calculated by comparing, through regression analysis, the annual

mean CS in each county/subcounty against time (year). When the annual mean CS values

Source : MNHN, Paris

150 ALYTES 22 (3-4)

Table 1. — 1960-99 Mean annual climate statistics and other independent variables used for this

study.

Climate statistic or other variable (and definition) Unit Variable name

Mean annual precipitation intensity

(average for all days)

Mean annual precipitation minus mean annual potential evapotranspiration

(average for all years)

millimeters per day PRE

millimeters per year | PRE-PET

Mean annual minimum temperature degrees Celsius TMN

{average for all days)

Mean annual mean temperature degrees Celsius TME

(average for all days)

Mean annual maximum temperature degrees Celsius TMX

(average for all days)

Mean annual number of wet days

WDAY

Gays with measured precipitation) is Loire

Mean annual number of dry days see der DDAY

(days without measured précipitation)

Mean annual number of cold days Gays pere CDAY

(days with minimum temperatures below 0°C)

Mean annual number of hot days 7e HDAY

(days with maximum temperatures above 35°C) Hyper ent

Mean annual solar radiation Sunièe pe RAD

average for all days)

Mean annual total winter degree days ({Tue - Tac} were PA Le ÿoD

Tin = 3°C and Tave = {Time + Tai} / 2, Z8F0 if negative)

Mean annual total summer degree days ({Taue— Te} Where

SDD

Trase = 3°C and Tiye = {Tax + Tin} / 2, 2670 if negative)

Mean elevation meters ELEV

County area square Kilometers AREA

{otal land area of county)

were missing or Zero, no trend was calculated and a zero trend value was assigned to the

county/subcounty. Simulated CS in each county/subcounty for the years 1960 and 1999 were

calculated using the trend regressions and the mean annual CS. Hence, the differences

between the simulated CS values for the two years represent the magnitude of the trend over

the 40 year time period and not the differences between any two years of actual CS data.

SOURCE AND PROCESSING OF ELEVATION AND AREA DATA

Two additional variables used to augment the climate information for each

county/subcounty were average elevation and total land area. Elevation data were obtained

from the USGS National Elevation Dataset [http://ede.usgs.gov/products/elevation/ned.html]

and were projected from geographic coordinates referenced to the World Geodetic Survey of

1984 to an Albers equal-area conic projection using a bilinear interpolation, 1000-meter cell

resolution, and the following parameters: ellipsoid = World Geodetic Survey of 1984, 1*

Source : MNHN, Paris

BATTAGLIN et al. 151

standard parallel = 29.5°, 2" standard parallel = 45.5°, central meridian = -96.0°, latitude of

origin = 23.0°, and no false easting or northing. Average elevation was calculated as the mean

of all cells within each mapping unit. Polygons for map units [http://www.census.gov/geo/

www/cob/scale.html] were represented in an Albers equal-area conic projection using the

same parameters as for the elevation data. Total mapping unit areas were determined from the

county/subcounty polygons.

STATISTICAL METHODS

Multiple linear regression (MLR) models (HELSEL & HirsCH, 1992) were developed

using the SAS statistical software system (ANONYMOUS, 1990) to relate amphibian species

richness to climate and location. Dependent and independent model variables were standard-

ized by subtracting the respective mean and dividing by the respective standard deviation.

Standardized variables have equal weights in regression models, and the resulting model

coefficients are proportional to their explanatory power in the models. “Best” and “stepwise”

SAS regression procedures were used to screen potential models; however, neither method

prevents correlated independent variables from entering the models. Multicollinearity among

independent variables, as indicated by variance inflation factor (VIF) values greater than 10,

can cause MLR model coefficients to be unrealistic in sign or magnitude (HELSEL & HIRSCH,

1992). When an MLR model contained an independent variable with a VIF value greater than

3, the independent variable was not used.

Two sets of regression models were developed: one set for the entire conterminous

United States (including one model for anurans and one for urodeles), and one set for each of

10 coarse-scale ecological regions (ANONYMOUS, 1997) (fig. 1). Because the primary objective

of this research was to determine the degree to which climate explains patterns of amphibian

species richness, model selection was manually supervised to favor climatic terms and prevent

highly correlated independent variables from entering the same model. The adjusted coeffi-

cient of determination (R°) and root mean square error (RMSE) statistics were used to

evaluate the predictive skill of the models for a particular region or the nation (ANONYMOUS,

1990; HELseL & HirsCH, 1992). The residuals between model and atlas estimates of species

richness are used to compare the predictive capabilities of the national and regional models.

Box plots are used to show the distributions of these residuals. The box plots show high and

low outliers as cireles. The central box extends from the 25 to 75! percentile of the data, and

the box whiskers extend to the 5" and 95" percentiles.

RESULTS

NATIONAL REGRESSION MODELS

Anuran species richness ranged from a maximum of 26 to a minimum of 1 (fig. 2a). The

R° for the national anuran model (fig. 3a) was 0.64 (tab. 2), with an RMSE of 3.07 species.

Mean annual temperature and mean annual precipitation (fig. 1b) accounted for the largest

Source : MNHN, Paris

152 ALYTES 22 (3-4)

Ecological

region

Easter Temperate Forests

Great Plains

North American Deserts

Northwestern Forested Mountains ° 800 MILES

Northern Forests

Mediterranean California

Marine West Coast Forests

Temperate Sierras

Southern Semi-Arid Highlands

Tropical Wet Forests

800 KILOMETERS

Mean annual

tation,

JM 165 than 1.0

[104015

1.610 2.0

2.11t02.5

2.6 10 3.0

3.1103.5

EX more than 3.5

300 MILES

300 KILOMETERS

Fig. 1.— Maps showing in the conterminous United States (a) coarse-level ecological regions and State,

county and subcounty boundaries, and (b) 1960-1999 mean annual precipitation.

Source : MNHN, Paris

BATTAGLIN et al. 153

300 MILES

300 KILOMETERS

Number of

KE el

urodele species A]

none reported

1104

5108

91012

13 10 16

17 to 20

211030

300 MILES

0 dookLoMETERS

Fig. 2.- showing in the conterminous United States (a) anuran species richness and (b) urodele

species richness, both from the Amphibian Research and Monitoring Initiative (AR MI) National

as for Amphibian Distributions.

Source : MNHN, Paris

154 ALYTES 22 (3-4)

Number of

anuran species

none reported

| 1104

5108

9 t0 12 300 MILES

1310 16

17 10 20

300 KILOMETERS

more than 20

Number of NS

urodele species

[ none reported

1104

5t08

910 12 o 300 MILES

1310 16 LRU

more than 20

Fig. 3. - Maps showing in the conterminous United States the national regr

anuran species richness and (b) urodele species richness.

sion model estimates of (a)

Source : MNHN, Paris

BATTAGLIN et al.

155

Tab. 2. - Best-fitting national and ecological region standardized regression models of amphibian

species richness and adjusted coefficient of determination (R°). NA. North American; NW,

Northwestern. See tab.

1 for other abbreviations.

National or ecological Region Regression model La

Anurans

National 0.57*PRE + 0.56*TME - 0.49*PRE-PET + 0.07*ELEV 0.64

Eastern Temperate Forests 0.71*TMX —-0.21*ELEV + 0.10*WDAY + 0.07*AREA 0.63

Great Plains 0.78*TME + 0.18*PRE + 0.08* AREA — 0.07*ELEV 0.78

NA Deserts 0.79*TMX + 0.36*PRE + 0.22*AREA + 0.16*ELEV 047

NW Forested Mountains -0.31*WDD - 0.27*ELEV + 0.25*AREA 0.23

Northern Forests -0.35*ELEV + 0.34*CDAY + 0.33*PRE 0.15

Mediterranean California 0.45*ELEV —0.41*CDAY + 0.39*RAD + 0.21*PRE 0.34

Marine West Coast Forests 0.34*TME + 0.22*ELEV + 0.14*AREA 0.25

Temperate Sierras No statistically significant model. :

Southern Semi-Arid Highlands |0.80*ELEV + 0.68*RAD -— 0.68*CDAY 0.68

Tropical Wet Forests oo few mapping units (5) to develop a model. -

Urodeles

National 0.70*PRE - 0.24*PRE-PET + 0.11*TME - 0.05*ELEV 0.45

Eastern Temperate Forests -0.65*WDD + 0.32*ELEV + 0.29*WDAY + 0.15*PRE-PET 0.50

Great Plains 0.55*PRE + 0.20* TME + 0.16*PRE-PET + 0.15*AREA 0.49

NA Deserts -0.57*ELEV + 0.37*PRE + 0.20*AREA — 0.19*TME 0.27

NW Forested Mountains 0.55*PRE + 0.20*TME — 0.18*ELEV 0.60

Northern Forests -0.67*WDD + 0.35*EL +0.33*PRE 0.74

Mediterranean California 0.47*PRE — 0.43*WDD — 0.34*TMX + 0.26*ELEV 0.50

Marine West Coast Forests 0.33*PRE + 031*TME + 0.23*ELEV + 0.22*RAD 0.40

Temperate Sierras: nificant model :

Southern Semi-Arid Highlands [No statistically -

Tropical Wet Forests Too few mapping units (5) to develop a model. -

proportion of the v

iation (because they have the largest model coefficients; tab. 2), and were

both positively associated with species richness. Mean annual precipitation minus mean

annual potential evapotranspiration also accounted for a substantial proportion of the

variation and was inversely associated with species richness. Mean mapping unit elevation

accounted for a small proportion of the variation and was positively associated with anuran

species richness. The national regression model overestimated anuran species richness along

the Mississippi embayment and in parts of California, Florida, and Oregon. The model

underestimated anuran species richness along the Atlantic coastal plain and in parts of Maine

and Texas (fig. 2a, 3a).

Urodele species richness ranged from a maximum of 30 to a minimum of 0 (fig. 2). The

R° for the national urodele model (fig. 3b) was 0.45, with an RMSE of 4.54 species. The

Source : MNHN, Paris

156 ALYTES 22 (3-4)

national urodele model (tab. 2) used the same CS as the national anuran model, but the

coefficient values were appreciably different. Mean annual precipitation accounted for the

largest proportion of the variation and was positively associated with species richness. Mean

annual precipitation minus mean annual potential evapotranspiration accounted for a smaller

proportion of the variation and was inversely associated with species richness. The national

regressions model overestimated urodele species richness in the central United States and in

parts of Florida and Washington; and underestimated species richness in most of the Eastern

United States except for Florida and Maine (fig. 2b, 3b).

REGIONAL REGRESSION MODELS

Separate regression models were developed for each coarse-resolution ecological region

(fig. la) to evaluate regional differences in the strength of climate as a predictor of amphibian

species richness. No models were developed for the Tropical Wet Forest ecological region,

which was predominant only in five mapping units, and represented less than 0.3 % of the

conterminous United States. The mean of anuran and urodele species richness for these five

mapping units was used in place of a model.

Eastern Temperate Forests

The Eastern Temperate Forests ecological region was predominant in 1,789 mapping

units, representing 31.8 % of the conterminous United States (fig. la). Anuran species

richness in this ecological region ranged from 2 to 26, and urodele species richness ranged

from 1 to 30. The R? for the Eastern Temperate Forest ecological region anuran model (fig. 4a)

was 0.63, and the RMSE was 2.91 species. Mean annual maximum temperature accounted for

the largest proportion of the variation and was positively associated with species richness. The

Eastern Temperate Forest model overestimated anuran species richness in parts of Arkansas,

Florida and Louisiana, and underestimated anuran species richness along the Atlantic

Coastal Plain and in parts of Alabama and Indiana. The residuals (model estimate minus

atlas estimate) for the Eastern Temperate Forest anuran model were much smaller than the

residuals between national model and atlas estimates of anuran species richness in those same

mapping units (fig. 5a; gray box plots are residuals from regional models and black box plots

are residuals from national model in the same mapping units). The R? for the Eastern

Temperate Forest ecological region urodele model (fig. 4b) was 0.50, and the RMSE was 3.78

species. The total of mean annual winter degree days accounted for the largest proportion of

the variation, and was inversely associated with species richness. The Eastern Temperate

Forest model overestimated urodele species richness in parts of Arkansas, Florida, Illinois,

Louisiana and Texas: and underestimated urodele species richness along the Eastern coastal

and inland plains and in parts of Alabama, Indiana and Kentucky. The residuals for the

Eastern Temperate Forest urodele model were smaller than the residuals between the national

model and atlas estimates of urodele species richness (fig. 5b).

Source : MNHN, Paris

BATTAGLIN et al. 157

800 MILES

300 KILOMETERS

Number of

urodele species

L__! lessthan 1

1104

5t08

91012 300 MILES

131016

1710 20 300 KILOMETERS

more than 20

Fig. 4. — Maps showing in the conterminous United States a compilation of regional regression models

stimates of (a) anuran species richness and (b) urodele species richness.

Source : MNHN, Paris:

158 ALYTES 22 (3-4)

DE Ti) 20

8 E ()

8 15 = 15

Le]

£ 10 $ o 8 = 10

5 ° °

£ Se DE h ae ; pe

5 8 De le 8

3 0 ÿ F5 e FE ge [6 d en FA

: EL à

2 -5- ë 8 o — -5

£ $ 2

2 -10- o o — -10

8 SE oo o = -15

—20 + L ste 1 PERRET 1 1 k + —20

Nation ETF GP NAD NFM NF MC MWCF SSH

Region

20 — = 20

(b)

— 08

CG)

n Number of Caudate Species

Tri

oonm-{ 1} 0 0

cam {1}

0m—{T+—6b@æwe® 0o

œ-— {1

oow{[}-—æ 0

o0m—{1+-

o LA

-5 = -5

—10- o 10

15 — 15

Nation ETF GP NAD NFM NF MC MWCF SSH

Region

Fig. 5. — Box plots showing residuals (in species) between national regression model and atlas (black) and

regional regressions models and atlas (gray) estimates of (a) anuran species richness and (b)

urodele species richness. ETF, Stern Temperate Forest; GP, Great Plains, NAD, North

American Deserts; NFM, Northwestern Forestal Mountains: NF, Northern Forests: MC,

Mediterranean California: MWCF, Marine West Coast Forests: 4, Southern Semi-Arid

Highlands.

Source : MNHN, Paris

BATTAGLIN et al. 159

GREAT PLAINS

The Great Plains ecological region was predominant in 837 mapping units, representing

28.9 % of the conterminous United States (fig. la). Anuran species richness in this ecological

region ranged from 1 to 23, and urodele species richness ranged from 0 to 13. The R? for the

Great Plains ecological region anuran model (fig. 4a) was 0.78, and the RMSE was 2.12

species. Mean annual temperature accounted for the largest proportion of the variation and

was positively associated with species richness. The Great Plains model overestimated anuran

species richness in parts of Kansas, Missouri and Nebraska; and underestimated anuran

species richness in parts of North Dakota, Oklahoma and Texas. The residuals for the Great

Plains anuran model were much smaller than the residuals between the national model and

atlas estimates of anuran species richness (fig. 5a). The R? for the Great Plains ecological

region urodele model was 0.49, and the RMSE was 1.10 species. mean annual precipitation

accounted for the largest proportion of the variation and was positively associated with

species richness. The Great Plains model overestimated urodele species richness in parts of

Iowa, Kansas and Oklahoma; and underestimated urodele species richness in parts of North

Dakota and Texas. The residuals for the Great Plains model were much smaller than the

residuals between the national model and atlas estimates of urodele species richness (fig. 5b).

NORTH AMERICAN DESERTS

The North American Deserts ecological region was predominant in 349 mapping units

representing 19.8 % of the conterminous United States (fig. la). Anuran species richness in

this ecological region ranged from 1 to 16, and urodele species richness ranged from 0 to 4.

The À? for the North American Deserts ecological region anuran model (fig. 4a) was 0.47, and

the RMSE was 2.05 species. Mean annual maximum temperature accounted for the largest

proportion of the variation and was positively associated with species richness. The North

American Deserts model overestimated anuran species richness in parts of California and

Utah, and underestimated anuran species richness in parts of Arizona and Texas. The

residuals for the North American Deserts anuran model were much smaller than the residuals

between the national model and atlas estimates of anuran species richness (fig. 5a). The À? for

the North American Deserts ecological region urodele model (fig. 4b) was 0.27, and RMSE

was 0.65 species. Mean mapping unit elevation accounted for the largest proportion of the

variation and was inversely associated with species richness. The North American Deserts

model overestimated urodele species richness in parts of California and Nevada; and under-

estimated urodele species richness in parts of California. The residuals for the North Ameri-

can Deserts urodele model were smaller than the residuals between the national model and

atlas estimates of urodele species richness (fig. 5b).

NORTHWESTERN FORESTED MOUNTAINS

The Northwestern Forested Mountains ecological region was predominant in 289

mapping units, representing 9.1 % of the conterminous United States (fig. la). Anuran species

Source : MNHN, Paris

160 ALYTES 22 (3-4)

richness in this ecological region ranged from 1 to 8, and urodele species richness ranged from

Oto 12. The R° for the Northwestern Forested Mountains ecological region anuran model (fig.

4a) was 0.23, and the RMSE was 1.20 species. The total of mean annual winter degree days

accounted for the largest proportion of the variation and was inversely associated with species

richness. The Northwestern Forested Mountains model overestimated anuran species rich-

ness in parts of Colorado and Idaho, and underestimated anuran species richness in parts of

Oregon. The residuals for the Northwestern Forested Mountains anuran model were much

smaller than the residuals between the national model and atlas estimates of anuran species

richness (fig. 5a). The R° for the Northwestern Forested Mountains ecological region urodele

model (fig. 4b) was 0.60, and the RMSE was 1.77 species. mean annual precipitation

accounted for the largest proportion of the variation and was positively associated with

species richness. The Northwestern Forested Mountains model overestimated urodele species

richness in parts of Washington, and underestimated urodele species richness in parts of

Colorado and Oregon. The residuals for the Northwestern Forested Mountains urodele

model were much smaller than the residuals between the national model and atlas estimates of

urodele species richness (fig. 5b).

NORTHERN FORESTS

The Northern Forests ecological region was predominant in 134 mapping units, repre-

senting 5.2 % of the conterminous United States (fig. la). Anuran species richness in this

ecological region ranged from 5 to 10, and urodele species richness ranged from 1 to 15. The

R° for the Northern Forests ecological region anuran model was 0.15, and the RMSE was 0.92

species. Mean mapping unit elevation accounted for the largest proportion of the variation

and was inversely associated with species richness. The residuals for the Northern Forests

anuran model were much smaller than the residuals between the national model and atlas

estimates of anuran species richness (fig. 5a). The R° for the Northern Forests ecological

region urodele model was 0.74, and the RMSE was 1.60 species. The total of mean annual

winter degree days accounted for the largest proportion of the variation and was inversely

associated with species richness. The residuals for the Northern Forests urodele model were

slightly smaller than the residuals between the national model and atlas estimates of urodele

species richness (fig. 5b).

MÉDITERRANEAN CALIFORNIA

The Mediterranean California ecological region was predominant in 277 mapping units,

representing 2.1 % of the conterminous United States (fig. la). Anuran species richness in this

ecological region ranged from 2 to 9, and urodele species richness ranged from 0 to 10. The R°?

for the Mediterranean California ecological region anuran model was 0.34, and the RMSE

was 1.04 species. Mean mapping unit elevation accounted for the largest proportion of the

variation and was positively associated with species richness. The residuals for the Mediter-

ranean California anuran model were much smaller than the residuals between the national

model and atlas estimates of anuran species richness (fig. 5a). The R? for the Mediterranean

California ecological region urodele model was 0.50, and the RMSE was 1.46 species. Mean

Source : MNHN, Paris

BATTAGLIN et al. 161

annual précipitation accounted for the largest proportion of the variation and was positively

associated with species richness. The residuals for the Mediterranean California urodele

model were smaller than the residuals between the national model and atlas estimates of

urodele species richness (fig. 5b).

MARINE WEST COAST FORESTS

The Marine West Coast Forests ecological region was predominant in 219 mapping

units, representing 1.1 % of the conterminous United States (fig. la). Anuran species richness

in this ecological region ranged from 3 to 6, and urodele species richness ranged from 2 to 10.

The R? for the Marine West Coast Forests ecological region anuran model was 0.25, and the

RMSE was 0.67 species. Mean annual temperature accounted for the largest proportion of

the variation and was positively associated with species richness. The residuals for the Marine

West Coast Forests anuran model were much smaller than the residuals between the national

model and atlas estimates of anuran species richness (fig. Sa). The R? for the Marine West

Coast Forests ecological region urodele model was 0.40, and the RMSE was 1.55 species.

Mean annual precipitation accounted for the largest proportion of the variation and was

positively associated with species richness. The residuals for the Marine West Coast Forests

urodele model were much smaller than the residuals between the national model and atlas

estimates of urodele species richness (fig. 5b).

TEMPERATE SIERRAS

The Temperate Sierras ecological region was predominant in 18 mapping units represen-

ting 1.1% of the conterminous United States (fig. la). Anuran species richness in this

ecological region ranged from 8 to 14, and urodele species richness was always 1. No

statistically significant model of anuran species richness could be developed from the available

independent variables (tab. 2). No model of urodele species richness was attempted since there

was no variation in the dependent variable.

SOUTHERN SEMI-ARID HIGHLANDS

The Southern Semi-Arid Highlands ecological region was predominant in 17 mapping

units representing 0.6 % of the conterminous United States (fig. la). Anuran species richness

in this ecological region ranged from 9 to 15, and urodele species richness was either 0 or 1.

The À? for the Southern Semi-Arid Highlands ecological region anuran model was 0.68, and

the RMSE was 0.88 species. Mean mapping unit elevation accounted for the largest propor-

tion of the variation and was positively associated with species richness. The residuals for the

Southern Semi-Arid Highlands anuran model were slightly larger than the residuals between

the national model and atlas estimates of anuran species richness (fig. Sa). No statistically

significant model of urodele species richness could be developed from the available indepen-

dent variables due to the limited variation in the dependent variable.

Source : MNHN, Paris

162 ALYTES 22 (3-4)

CLIMATE TRENDS

Amphibian species richness was strongly associated with several of the mean annual

climate variables, and mean annual precipitation and mean annual temperature were statisti-

cally significant variables in 12 and 8 models, respectively (tab. 2). Increasing trends in annual

mean temperature and precipitation were prevalent across much of the conterminous United

States between 1960 and 1999 (fig. 6). Exceptions include decreasing mean annual tempera-

ture in parts of the Great Plains ecological region, and decreasing mean annual precipitation

in the southeastern part of the Eastern Temperate Forests ecological region.

DISCUSSION

At the national level, the model for anurans performed better than that for urodeles (fig.

5). Both models included mean annual precipitation as a strong variable for predicting

patterns of richness. KIESTER (1971) and DUELLMAN & SWEET (1999) previously noted a

strong correlation in the conterminous United States between amphibian species richness and

mean annual rainfall. Partitioning the country by coarse-resolution ecological regions resul-

ted in improved models for both anurans and urodeles. The residuals (model estimate minus

atlas estimate) for the compilation of regional anuran and urodele models were much smaller

than the residuals between national model and atlas estimates of anuran and urodele species

richness in all mapping units (fig. 5). In several cases the R? for the regional models were less

than that of the national model, but the residuals were also smaller. In general, temperature

variables (mean annual mean and mean annual maximum) figured more strongly in anuran

models, whereas precipitation (mean annual precipitation intensity) had greater explanatory

value in urodele models. This makes sense from the perspective that there is no urodele

counterpart to toads; hence, anurans are less restricted by arid conditions than are

urodeles.

In general, trends in climate during 1960-1999 were toward wetter, warmer conditions for

most of the conterminous United States. This could have provided more surface moisture

availability for breeding habitat, and air and soil temperatures more amenable to regulating

amphibian body temperatures throughout the year. Trends toward drier conditions in part of

the southeastern United States and southwest Oregon may have resulted in reduced availabil-

ity of breeding habitat in those areas.

This effort to model the relations between anuran and urodele species richness and mean

annual climate in the United States capitalized on the strong dependence of amphibians on

their external environment for internal hydrothermal regulation. A limitation of the approach

was that it assumed that the climate experienced by amphibians was reflected by long-term

climate statistics summarized at the county/subcounty level. In fact, amphibians interact with

climate at multiple scales, and alter their behaviors in concert with mierohabitat features (sun

flecks, burrows, duff, vegetation cover, wetlands, etc.) to modify the effects of the broader-

scale conditions. Therefore, the conditions represented by the data in this study likely

addressed only the broadest effects of climate. For this reason, the statistical models pre

Source : MNHN, Paris

BATTAGLIN et al. 163

Millimeters x

per vear

more than 4.0

| soo nes

300 KILOMETERS

less than -4.0

no trend data

1/100 degrees

centigrade

= more than 2.5

& less than 2.5

no trend data

Fig. 6. — Maps showing in the conterminous United States the trend for 1960-1999 in (a) annual mean

precipitation and (b) annual mean temperature.

o 300 MILES

O 300 KILOMETERS

Source : MNHN, Paris

164 ALYTES 22 (3-4)

here were aimed at the general level of anuran and urodele richness, and were not aimed at

predicting the fate of particular species.

Other explanatory variables may improve the ability to explain patterns of amphibian

richness. For example, seasonal climate statistics may be more informative than annual

statistics for certain measures, and additional landscape factors (e.g., current and historic land

cover/use, hydrology, glaciation) and information such as evolutionary lineage could be very

useful. Additionally, models could be developed at the family level, or for groups of species

having similar life history or developmental characteristics.

The coarse-level ecological regions used for this study were somewhat problematic.

Highly discontinuous mountain regions in the West often did not align well with

county/subcounty units, so not all discontinuous portions of these regions were represented in

the models. The largest regions (Eastern Temperate Forests, Great Plains and North Ameri-

can Deserts) included a lot of variation in temperature and moisture gradients. A finer level of

regionalization (Level II regions defined in ANONYMOUS, 1997) may have been more appro-

priate, as it would have subdivided the largest regions, while leaving the smaller regions intact;

however, the number of map units per region may have been insufficient for developing

models for several of the regions at this level.

RÉSUMÉ

Les amphibiens occupent une grande diversité d’habitats sur la planète, mais leur

richesse spécifique est plus élevée dans les régions aux climats humides et chauds. Nous avons

modélisé les relations statistiques entre la richesse spécifique en anoures et urodèles et le climat

annuel des Etats Unis continentaux, et comparé ces relations aux niveaux national et régional.

Les variables modélisées ont été calculées pour des unités cartographiques correspondant aux

contés ou aux sous-contés, et se sont appuyées sur des statistiques climatiques annuelles

moyennes recueillies sur une période de 40 années (1960-1999), l'altitude moyenne et la

surface des unités cartographiques, et des estimations de la richesse spécifique en anoures et

urodèles. Les données climatiques ont été obtenues à partir de plus de 7500 stations metéo-

rologiques et ont été incorporées dans les données concernant les unités cartographiques au

moyen de modèles de régression linéaire multiple. Les richesses spécifiques en anoures et

urodèles ont été calculées à partir de l’atlas national de distribution des amphibiens préparé

par l’'Amphibian Research and Monitoring Initiative (ARMI) de l'United States Geological

Survey. Le modèle de régression linéaire multivariée (MLR) national pour la richesse spéci-

fique en anoures à un coefficient de détermination ajusté (R?) de 0,64 et celui concernant les

urodèles un R°? de 0,45. Lorsque les Etats Unis sont divisés en régions écologiques grossières,

on obtient des modèles pour les anoures dont les R° se répartissent entre 0,15 et 0,78 pour les

anoures, et entre 0,27 et 0,74 pour les urodèles. En général, les modèles régionaux pour les

anoures se sont avérés plus fortement influencés par des variables de température, tandis

que les variables liées à la précipitation avaient plus d'influence sur les modèles pour les

urodèles.

Source : MNHN, Paris

BATTAGLIN et al. 165

ACKNOWLEDGMENTS

The United States Geological Survey’s Toxies Program and Amphibian Research and Monitoring

Initiative provided funding to support this study. We thank S. Char for support with GIS analysis. We are

grateful to B. Moring, USGS, Texas, and B. Klaver, USGS, South Dakota, and two anonymous reviewers

for comments on earlier versions of this manuscript.

LITERATURE CITED

ANONYMOUS [SAS INSTITUTE], 1990. — SAS/STAT user's guide, Version 6, 4" edition, Cary, North

Carolina, 1,686 pp.

ANONYMOUS [Commission for Environmental Cooperation], 1997. - Ecological regions of North

America. Towards a common perspective. Montreal, Canada, CEC: 1-71.

ALFORD, R. À. & RICHARDS, S. J., 1999. - Global amphibian declines: a problem in applied ecology. Ann.

Re. Ecol. Syst., 30: 133-165.

BERRILL, M., BERTRAM, S., WILSON, A., LOUIS, S., BRIGHAM, D. & STROMBERG, C., 1993. — Lethal and

sublethal impacts of pyrethroid insecticides on amphibian embryos and tadpoles. Environ. Toxicol.

Chem., 12: 525-539.

BLAUSTEIN, A. R., ROMANSIC, I. M., KIESECKER, J. M. & HATCH, A. C., 2003. - Ultraviolet radiation, toxic

chemicals and amphibian population declines. Diver. & Distrib., 9: 123-140.

BoNIN, J., OUELI M., RODRIGUE, J., DESGRANGES, J.-L., GAGNE, F., SHARBEL, T. F. & LOWCoCK, L.

A. 1997. - Measuring the health of frogs in agricultural habitats subjected to pesticides. /n: D. M.

GREEN (ed.), Amphibians in decline: Canadian studies of a global problem, Herpetological Conser-

vation, St. Louis, Missouri, Society for the Study of Amphibians and Reptiles, 1: 246-257.

BrirsoN, C. A. & THRELKELD, S. T., 2000. — Interactive effects of anthropogenic, environmental, and

biotic stressors on multiple endpoints in Hyla chrysoscelis. J. Iowa Acad. Sci., 107: 61-66.

CaREy, C. & ALEXANDER, M. À., 2003. - Climate change and amphibian declines: is there a link? Diver.

& Distrib., 9: 111-121.

C., HEYER, W. R., WiLKINSON, J., ALFORD, R. A., ARNTZEN, J. W., HALLIDAY, T., HUNGERFORD,

L., Lips, K. R., MIDDLETON, E. M., ORCHARD, S. A. & RAND, A. S., 2001. - Amphibian declines

and environmental change: use of remote-sensing data to identify environmental correlates.

Conserv. Biol., 15: 2001.

CLark, M. Hay, L., RAJOGOPALAN, B. & Wicgy, R., 2004. — The Schaake shuffle: a method for

reconstructin: ce-time variability in forecasted precipitation and temperature fields. J Hydro-

meteorol., 5: 243

CoLuis, J. P. & Srorrer, A., 2003. - Global amphibian declines: sorting the hypotheses. Diver &

Distrib., 9: 89-98.

Cor, P. S. & Murns, E., 2002. — Variable breeding phenology affects the exposure of amphibian

embryos to ultraviolet radiation. Ecology, 83: 2958-2963.

Daszak, P., CUNNINGHAM, A. À. & HYATT, A. D., 2003. — Infectious disease and amphibian population

declines. Diver. & Distrib., 9: 141-150.

DavipsoN, C., SHAFI

destruction, UV-B, and climate-change hypoth

Biol., 16: 1588-1601

DEMAYNADIER, P. G. & HUNTER, M. L., Jr., 1998. - Effects of silvicultural edges on the distribution and

abundance of amphibians in Maine. Consen : 340-352.

Do», C.K.. Jr., 1997. - Imperiled amphibians: a perspective. In: G. BENZ & D. COLLINS (ed.),

Aquatic fauna in peril: the southeastern perspective, Special Publication 1, Decatur, Georgia,

Southeast Aquatic Research Institute, Lenz Design & Communications: 165-200.

CARE

, M. R., 2002. - Spatial tests of the pesticide drift, habitat

for California amphibian declines. Conserv.

Source : MNHN, Paris

166 ALYTES 22 (3-4)

DonxELLY, M. A. & CRUMP, M. L., 1998. — Potential effects of climate change on two neotropical

amphibian assemblages. Clim. Change, 39: 541-561.

DuELLMAN, W. E., 1999a. — Patterns of distribution of amphibians: a global perspective. Baltimore,

Maryland, The John Hopkins University Press: 1-633.

ee 1999b. — Global distribution of amphibians: patterns, conservation, and future challenges. /n:

DUELLMAN (19994): 1-30.

DuELLMAN, W. E. & SWEET, S. S., 1999. — Distribution patterns of amphibians in the nearctic region of

North America. /n: DUELLMAN (1999a): 31-110.

EisCHEID, J. K.., PASTERIS, P. A., Diaz, H. F., PLANTICO, M. S. & Lorr, N. J., 2000. - Creating a serially

complete, national daily time series of temperature and precipitation for the western United States.

J. appl. Meteorol., 39: 1580-1591.

FABRIG, L., PEDLAR, J. H., POPE, S. E., TAYLOR, P. D. & WEGNER, JF, 1995. - Effect of road traffic on

amphibian density. Biol. Conserv., 73: 177-182.

FIsHER, R. N. & SHAFFER, H. B., 1996. - The decline of amphibians in California’s Great Central Valley.

Conserv. Biol., 10: 1387-1397.

FoRMAN, R.T. T. & GoprON, M. 1986. - Landscape ecology. New York, New York, John Wiley & Sons:

1-620.

Games, J. P. & BRrISCH, A. R., 2001. - Climate warming and calling phenology of frogs near Ithaca, New

York, 1990-1999. Conserv. Biol., 15: 1175-1178.

GReeN, D. M., 1997. — Perspectives on amphibian population declines: defining the problem and

searching for answers. In: D. M. GREEN (ed.), Amphibians in decline: Canadian studies of a global

problem, Herpetological Conservation, St. Louis, Missouri, Society for the Study of Amphibians

and Reptiles, 1: 291-308.

Hay, L. E. & CLaRk, M. P, 2003. - Use of statistically and dynamically downscaled atmospherie model

output for hydrologic simulations in three mountainous basins in the western United States. J.

Hydrol., 282: 56-75.

Hay, L. E. & MCCaRE, G. J., 2002. — Spatial variability in water-balance model performance in the

conterminous United States. J Am. Water Resourc. Assoc., 38: 847-860.

Hay, L. E., WizBy, R. L. & LEAVESLEY, G. H., 2000. — A comparison of delta change and downscaled

GCM scenarios for three mountainous basins in the United States. J Am. Water Resource. Assoc.,

36: 387-397.

HAYDEN, B. P., 1998. - Ecosystem feedbacks on climate at the landscape scale. Phil. Trans. r. Soc. London,

Biol. Sci., 353: 5-18.

Hayes, M. P. & JENNINGS, M. R., 1986. - Decline of ranid frog species in western North America: are

bullfrogs (Rana catesbeiana) responsible? J. Herp., 20: 490-509.

Hayes, T. B., COLLINS, A., LEE, M., MENDOZA, M., NORIEGA, N., STUART, A. A. & VONK, A., 2002. -

Hermaphroditic, demasculinized frogs after exposure to the herbicide atrazine and low ecologi-

cally relevant doses. Proc. natl. Acad. Sci. USA, 99: 5476-5480.

Hecsez, D. R. & Hir R. M., 1992. — Statistical methods in water resources. Amsterdam, The

Netherlands, Elsevier: 1-55:

HOUGHTON, J. T., DING, Y.., GRiGGs, D. J., NOGUER, M., VAN DER LINDEN, P., DAI, X.. MASKELL, K. &

JOHNSON, C. I. (ed.), 2001. — Climate change 2001: the scientific basis. Cambridge University,

Cambridge, UK, Contribution of Working Group 1 to the Third ment Report of the

Intergovernmental Panel on Climate Change.

HOULAHAN, J. E., FINDLAY, C. CHMIDT, B. R., MEYER, À. H. & KUZMIN, S. L., 2000. - Quantitative

evidence for global amphibian population declines. Nature, 404: 752-755.

JoHNsON, P.T. J. & CHASE, J. M., 2004. - Parasites in the food web: linking amphibian malformations and

aquatic eutrophication. Ecol. Let. 521-526.

KIESECKER, J. M. & BLAUSTEIN, A. R., 1998. - Effects of introduced bullfrogs and smallmouth bass on

microhabitat use, growth, and survival of native Red-legged frogs (Rana aurora). Conserv. Biol., 12:

776-787.

is 1999. - Pathogen reverses competition between larval amphibians. Ecology, 80: 2442-2448.

KIESTER, A. R., 1971. - Species density of North American amphibians and reptiles. Syst. Zool., 20:

127-137.

Source : MNHN, Paris

BATTAGLIN et al. 167

KoowZ, W., 1992. - Amphibians in Manitoba. /n: C. A. Bisnop & K.. E. PErriT (ed.), Declines in Canadian

amphibian populations: designing a national monitoring strategy, Canadian Wildlife Service Occa-

sional Paper Nr. 76: 19-22.

KkzYSIK, A. J,, 1998. - Amphibians, ecosystems, and landscapes. /n: M. J. LANNOO (ed.), Status &

conservation of midwestern amphibians, lowa City, Iowa, University of Iowa Press: 31-41.

LaNNOO, M. J,, 1996. — Okoboji wetlands: a lesson in natural history. owa City, Iowa, University of Iowa

Press: 1-156.

LANNOO, M. J., GALLANT, A. L., NANJAPPA, P., BLACKBURN, L. & HENDRICKS, R.., 2005. - Ecological and

ecoregional analyses of amphibian distributions as an approach to amphibian conservation. /n:

M. I. LANNOO (ed.), Declining amphibians: a United States response to the global problem, Berkeley,

California, University of California Press, in press.

LaNNOO, M. J., SUTHERLAND, D. R., JONES, P., ROSENBERRY, D., KLAVER, R. W., HOPrr, D. M.

JOHNSON, P. T. J., LUNDE, K. B., FACEMIRE, C. & KAPFER, J. M., 2003. - Multiple causes for the

malformed frog phenomenon. /n: G. LINDER, S. KREST, D. SPARLING & E. LITTLE (ed.), Multiple

stressor effects in relation to declining amphibian populations, ASTM Stock Nr. STP1443: 233-262.

LAURANCE, W. F., MCDONALD, K. R. & SPEARE, R.., 1996. - Epidemic disease and the catastrophic decline

of Australian rain forest frogs. Conserv. Biol., 10: 406-413.

LAWLER, S. P., Drirz, D., STRANGE, T. & HoLyoAKk, M., 1999. - Effects of introduced mosquito fish and

bullfrogs on the threatened California red-legged frog. Conserr. Biol., 13: 613-622.

Lorie, E. E., BRIDGES, C. M., LINDER, G. & BOONE, M., 2003. - Establishing causality in the decline and

deformity of amphibians: the amphibian research and monitoring initiative model. /n: G. LiNDER,

S. KREST, D. SPARLING & E. LITTLE (ed.), Multiple stressor effects in relation to declining amphibian

populations, ASTM Stock Nr. STP1443: 263-277.

MONSERUD, R. A. & LEEMANS, R., 1992. - Comparing global vegetation maps with the kappa statistic.

Ecol. Modell., 62: 275-293.

PieLkE, R.A., Sr, MARLAND, G., BETTS, R. A., CHASE, T. N., EASTMAN, J. L., NILES, J. O., NiyoGt, D. D.

S. & RUNNING, S. W., 2002. - The influence of land-use change and landscape dynamics on the

climate system: relevance to climate-change policy beyond the radiative effect of greenhouse gases.

Phil. Trans. r. Soc. London, 360: 1705-1719.

PIELKE, R. A., Sr., WALKO, R. L., STEYAERT, L. T., VIDALE, P. L., LISTON, G. Lyoxs, W. A. & CHASE,

T.N., 1999. - The influence of anthropogenic landscape changes on weather in South Florida.

Mon. Weather Rev., 127: 1663-1673.

Pounps, J. A. & CRUMP, M. C., 1994. — Amphibian declines and climate disturbance: the case of the

golden toad and the harlequin frog. Conserv. Biol., 11: 1307-1322.

Pounps, J. A., FOGDEN, M. P. L. & CAMPBELL, J. H., 1999. — Biological response to climate change on a

tropical mountain. Nature, 398: 611-615.

RosEN, P. C. & SCHWALBE, C. R., 1995. — Bullfrogs: introduced predators in southwestern wetlands. /n:

. T. LAROE, G. S. FARRIS, C. E. PUCKETT, P. D. DoraN & M. J. MAC (ed.), Our living resources:

a report to the nation on the distribution, abundance, and health of US plants, animals, and

ecosystems, Washington, DC, US Department of the Interior, National Biological Service: 452-

454.

THOMAS, C. D., CAMERON, A., GREEN, R., BAKKENES, M., BEAUMONT, L., COLLINGHAM, Y., ERASMUS, B.,

FERREIRA DE SIQUEIRA, M., GRAINGER, A., HANNAH, L., HUGHES, L., HUNTL + VAN JAARS-

El . G., Mis, L., ORTEGA-HUERTA, M., PETERSON, A., PHILLIPS, O. & WILLIAMS,

004. - Extinction risk from climate change. Nature, 427: 145-148.

H., 1973. Vegetation of the earth. New York, New York, Springer-Verlag: 1-237.

YouxG, B. E., STUART, S. N., CHANSON, JL $., COX, N. A. & BOUCHER, T. M. 2004. - Disappearing jewels:

the status of new world amphibians, Arlington, Virginia, NatureServe: 1-55.

Corresponding editor: C. Kenneth Dopp, Jr.

Source : MNHN, Paris

Alytes, 2005, 22 (3): 108. Book reviews

New books on frogs and salamanders

Annemarie OHLER

Museum national d'Histoire naturelle

Departement de Systématique et Evolution

USM 602 Taxinomie et Colletion — Reptiles & Amphibiens

25 rue Cuvier, 75005 Paris, France

e-mail: ohler@mnhn.fr

Sergius L. KUZMIN, 1995. - The clawed salamanders of Asia. Genus Onychodactylus. Biology; distribution,

and conservation. Magdeburg, Westarp Wissenschaften, Die Neue Brehm Bücherei, Vol. 622,

1995: 1-12. ISBN 3-80432-495-2.

This short book gives a presentation of various traits of the two species of clawed salamanders. This

genus has a large distribution in Asia including Russia, China, Korea and Japon. Therefore most of the

studies available have been published in the languages of these four countries and are not easily available

for many Western researchers. In this book all this information is now presented in English. The book

introduces to the history of the studies of this interesting group of salamanders. The systematics and

distribution of the family Hynobiidae are presented in order to explain the taxonomic position of the

genus Onychodactylus. For both species, data on systematics, morphology, karyology, ontogeny, anatomy,

ecology and behaviour are given. The content is thoroughly illustrated by numerous black and white

figures and drawings, and two colour plates. This is a complete book about this small genus of

salamanders. It will be of interest to all students of biology and evolution of amphibians.

Sally HAINES, 2000. — Sith toves. Hustrated classic herpetological books at the University of Kansas in

pictures and conservation. Society for the Study of Amphibians and Reptiles & The Kenneth

Spencer Research Library at the University of Kansas, Contributions to Herpetology, 16: i-vii +

1-184, ISBN 0-916984-53-2.

Sitting late in an autumn afternoon in a old fashioned library and turning the pages of those

historical books that spoke for the first time of herps leaves unforgettable souvenirs of a fascinating world

of images and stories. Sometimes one is stroke by the fantastic world that cohabits with detailed precise

observation, like in the works of Osterdam, a student of Linnaeus, where the sketch of a mairmaid gives

even more strength to the morphological study of the Siren. One is surprised to find all the details of

giving birth of a viper in a book dated 1589 at a period where other authors just gave badly outlined

figures of quite unidentifiable things. 1 always was fascinated by the chronological and descriptive details

on ontogeny of European frogs that can be found in the book of Roesel of Rosenhof (1758) — a

Linnean work according to nomenclature — that is merely known for the high quality of the figures.

ss author not only was a precise observer of the traits of the species but had also detailed insight in their

life history. Beside the strong and intense feeling that fieldwork and observation of animals in their habitat

gives 10 the herpetologist, the discovery of the books and the authors that made up our knowledge is one

of the major pleasures of our herpetological work

very ni s

these mysteries with the open eyes of a child and to be able to make abstre sputes

that govern our daily work. Sally presents several books, choosing a single plate of each and putting it in

its historical frame. In the introduction some explanation on the methodology of printing figures gives

technical clues to understand the works. À list of herpetological works containing figures is added and

presented both in alphabetic and chronological way.

This book gives a introduction to the fascinating world preserved in the libraries. It allows you to take

home a souvenir of those autumn afternoons. It will surely make you return to discover more of these

treasures and it will make us attentive to the thoughts and observations of those who preceded us.

Source : MNHN, Paris

AIVTES

International Journal of Batrachology

published by ISSCA

EDITORIAL BOARD

Chief Editor: Alain Dugois (Laboratoire des Reptiles et Amphibiens, Muséum national d'Histoire naturelle,

25 rue Cuvier, 75005 Paris, France; <adubois@mnhn.fr>).

Deputy Editor: Thierry LODÉ (Laboratoire d'Ecologie animale, Université d'Angers, 2 boulevard Lavoisier,

9045 Angers Cedex, France; <thierry.lode@univ-angers.fr>).

Conservation Editor: Stephen J. RICHARDS (Vertebrates Department, South Australian Museum, North Terrace,

Adelaide, S.A. 5000, Australia; <Richards.Steve@saugov.sa.gov.au>).

Editorial Board: Franco ANDREONE (Torino, Italy); Lauren E. BROWN (Normal, USA); Janalee P. CALDWELL

(Norman, USA); Ulisses CARAMASCHI (Rio de Janeiro, Brazil); Claude Comes (Perpignan, France) ; C.

Kenneth Dopp, Jr. (Gainesville, USA) ; Günter GOLLMANN (Wien, Austria); Heinz GRILLITSCH (Wien,

Austria); Tim HAaLLIDAY (Milton Keynes, United Kingdom); W. Ronald Herr (Washington, USA):

Esteban O. LAViLLA (Tucumän, Argentina); Masafumi MATsuI (Kyoto, Japan): Alain PAGANO (Angers,

France): John C. POYNTON (London, England): Miguel VENCEs (Konstanz, Germany).

Technical Editorial Team (Paris, France): Alain Dugois (texts); Roger Bour (tables); Annemarie OHLER (figures).

Book Review Editor: Annemarie OHLER (Paris, France).

SHORT GUIDE FOR AUTHORS

(for more detailed Instructions 10 Authors, see Alytes, 1997, 14: 175-200)

Alytes publishes original papers in English, French or Spanish, in any discipline dealing with amphibians.

Beside articles and notes reporting results of original research, consideration is given for publication to synthetic

review articles, book reviews, comments and replies, and to papers based upon original high quality illustrations

(such as colour or black and white photographs), showing beautiful or rare species, interesting behaviours, etc.

The title should be followed by the name(s) and address(es) of the author(s). The text should be typewritten

or printed double-spaced on one side of the paper. The manuscript should be organized as follows: English

abstract, introduction, material and methods, results, discussion, conclusion, French or Spanish abstract,

acknowledgements, literature cited, appendix.

Figures and tables should be mentioned in the text as follows: fig. 4 or tab. 4. Figures should not exceed 16 x

24 em. The size of the lettering should ensure its legibility after reduction. The legends of figures and tables

should be assembled on a separate sheet. Each figure should be numbered using a pencil.

References in the text are to be written in capital letters (BOURRET, 1942; GRAF & POLLS PELAZ, 1989; INGER

et al., 1974). References in the Literature Cited section should be presented as follows:

Bourker, R., 1942. - Les batraciens de l'Indochine. Hanoi, Institut Océanographique de l'Indochine: i-x + 1-547,

1. 114.

GRAF, JD. & PoLLs PELAZ, M., 1989. - Evolutionary genetics of the Rana esculenta complex. In: R. M. DAWLEY

& JP. Bogart (ed.), Évolution and ecology of unisexual vertebrates, Albany, The New York State Museum:

289-302.

IxGER, R. E., Voris, H. K. & Vois, H. H., 1974. - Genetic variation and population ecology of some Southeast

Asian frogs of the genera Bufo and Rana. Biochem. Genet., 12: 121-145.

Manuscripts should be submitted in triplicate either to Alain DuBois (address above) if dealing with

amphibian morphology, anatomy, systematics, biogeography, evolution, genetics, anomalies or developmental

biology, or to Thierry LODÉ (address above) if dealing with amphibian population genetics, ecology, ethology or

life history, or to Stephen J. RICHARDS (address above) if dealing with conservation biology, including declining

amphibian populations or pathology. Acceptance for publication will be decided by the editors following review

by at least two referees.

If possible, after acceptance, a copy of the final manuscript on a floppy disk (3 14 or 5 ) should be sent to

the Chief Editor, We welcome the following formats of text processing: (1) preferably, MS Word (1.1 to 6.0, DOS

or Windows), WordPerfect (4.1 to 5.1, DOS or Windows) or WordStar (3.3 to 7.0); (2) less preferably, formated

DOS (ASCII) or DOS-formated MS Word for the Macintosh (on a 3 %4 high density 1.44 Mo floppy disk only).

Page charges are requested only from authors having institutional support for this purpose. The publication

of colour photographs is charged. For each published paper, 25 free reprints are offered by ISSCA to the

author(s). Additional reprints may be purchased.

Published with the support of AALRAM

(Association des Amis du Laboratoire des Reptiles et Amphibiens

du Muséum National d'Histoire Naturelle, Paris, France)

Directeur de la Publication: Alain Duvois,

Numéro de Commission Paritaire: 64851

Source : MNHN, Paris:

Alytes, 2005, 22 (3-4) : 65-168

The Amphibian Research and Monitoring Initiative

Proceedings of a Symposium held in

Norman, Oklahoma, USA, 2004

Edited by C. Kenneth Dopp, gr.

Contents

P. Stephen CorN, Erin MuTHS, Michael ApaMs & C. Kenneth Dopp, Jr.

The United States Geological Survey’s Amphibian Research and Monitoring Initiative 65-71

Robin E. JUNG, J. A. ROYLE, John R. SAUER, C. ADDIsON, R. D. Rau, J. L. SHirk & J. C. WHIssEL

Estimation of stream salamander (Plethodontidae, Desmognathinae and Plethodontinae)

populations in Shenandoah National Park, Virginia, USA .................. 72-84

P. StephenCorN, B. R. HossaK, Erin MUTHS, D. A. PATLA, Charles R. PETERSON & Alisa

L. GALLANT

Status of amphibians on the Continental Divide:

surveys on a transect from Montana to Colorado, USA .................... 85-94

Wendy H. WENTE, Michael J. ADAMS & C. A. PEARL

Evidence of decline for Bufo boreas and Rana luteiventris in and around

the northern Great Basin, western USA :.: 2... us. dute.uise 95-108

D. Earl GREEN, & Erin MUTHS

Health evaluation of amphibians in and near Rocky Mountain, National Park,

(CORSA EU A A dd race 109-129

Christine M. BRIDGES & Edward E. LITTLE

Toxicity to amphibians of environmental extracts from natural waters in National

Parks and Eish'and Wildlife Refuges #..:......1..2..."thu.. 4. 130-145

William BATTAGLIN, Lauren HAY, Greg MC CABE, Priya NANJAPPA & Alisa GALLANT

Climate patterns as predictors of amphibian species richness and indicators of

FO L De c Sonsuur es USE 146-167

Book REVIEWS

Annemarie OHLER

New books on frogs and salamanders . . 168

Alytes is printed on acid-free paper.

Alytes is indexed in Biosis, Cambridge Scientific Abstracts, Current Awareness in Bilogical Sciences,

Pascal, Referativny Zhurnal and The Zoological Record.

Imprimerie F. Paillart, Abbeville, France.

Dépôt légal : 2° trimestre 2005.

Source : MNHN, Paris