Landscape resistance to frog movements

Abstract

An animal's capacity to recolonize a patch depends on at least two components: its ability to detect the patch and its ability to reach it. However, the disruption of such processes by anthropic disturbances could explain low animal abundance patterns observed by many investigators in certain landscapes. Through field experiments, we compared the orientation and homing success of northern green frogs (Rana clamitans melanota Rafinesque, 1820) and northern leopard frogs (Rana pipiens Schreber, 1782) translocated across disturbed or undisturbed surfaces. We also monitored the path selected by individuals when presented with a choice between a short distance over a disturbed surface and a longer, undisturbed route. Finally, we measured the water loss and behaviour of frogs on substrates resulting from anthropogenic disturbances and a control. When presented with a choice, 72% of the frogs avoided disturbed surfaces. Although able to orient towards the pond of capture when translocated on disturbed surfaces, frogs had a lower probability of homing successfully to the pond than when translocated at a similar distance on an undisturbed surface. Frogs lost the most water on substrates associated with disturbance and in the absence of cover. Our data illustrate that anthropically disturbed areas devoid of cover, such as mined peatlands and agricultural fields, disrupt the ability of frogs to reach habitat patches and are likely explanations to their reduced abundance patterns in such environments.

Full text

Landscape resistance to frog movements
M.J. Mazerolle and A. Desrochers
Abstract: An animal’s capacity to recolonize a patch depends on at least two components: its ability to detect the
patch and its ability to reach it. However, the disruption of such processes by anthropic disturbances could explain low
animal abundance patterns observed by many investigators in certain landscapes. Through field experiments, we com-
pared the orientation and homing success of northern green frogs (Rana clamitans melanota Rafinesque, 1820) and
northern leopard frogs (Rana pipiens Schreber, 1782) translocated across disturbed or undisturbed surfaces. We also
monitored the path selected by individuals when presented with a choice between a short distance over a disturbed sur-
face and a longer, undisturbed route. Finally, we measured the water loss and behaviour of frogs on substrates resulting
from anthropogenic disturbances and a control. When presented with a choice, 72% of the frogs avoided disturbed sur-
faces. Although able to orient towards the pond of capture when translocated on disturbed surfaces, frogs had a lower
probability of homing successfully to the pond than when translocated at a similar distance on an undisturbed surface.
Frogs lost the most water on substrates associated with disturbance and in the absence of cover. Our data illustrate that
anthropically disturbed areas devoid of cover, such as mined peatlands and agricultural fields, disrupt the ability of
frogs to reach habitat patches and are likely explanations to their reduced abundance patterns in such environments.
Résumé : La capacité d’un animal à recoloniser avec succès une parcelle du paysage dépend d’au moins deux compo-
santes, sa capacité à détecter la parcelle et sa capacité à l’atteindre. Les perturbations de ces processus par certaines
activités anthropiques pourraient expliquer les patrons de faible abondance observés par plusieurs chercheurs dans cer-
tains paysages. Lors d’expériences sur le terrain, nous avons comparé l’orientation et le succès du retour de grenouilles
vertes (Rana clamitans melanota Rafinesque, 1820) et de grenouilles léopards (Rana pipiens Schreber, 1782) déplacées
au delà de surfaces naturelles ou perturbées. Nous avons également évalué, chez des individus relocalisés, la préférence
entre un parcours court et perturbé et un parcours plus long, mais naturel. Finalement, nous avons quantifié la perte en
eau et le comportement chez des grenouilles exposées à différents substrats résultant de perturbations anthropiques et
chez des grenouilles témoins. Lorsqu’une surface naturelle est disponible, 72 % des grenouilles évitent les surfaces per-
turbées lors de leurs déplacements. Bien que capables de s’orienter avec succès vers l’étang d’origine après relocalisa-
tion sur des surfaces perturbées, les grenouilles ont une plus faible probabilité de retour qu’après relocalisation à des
distances semblables sur une surface naturelle. Les grenouilles perdent le plus d’eau sur les substrats associés aux per-
turbations anthropiques et dépourvus de couverture végétale. Nous concluons que les surfaces perturbées par les activi-
tés anthropiques qui résultent en une perte de couverture végétale, telles que les tourbières exploitées et les terres
agricoles, entravent les déplacements des grenouilles vers les parcelles d’habitat et expliquent vraisemblablement leurs
patrons de densité réduite dans ces milieux.
Mazerolle and Desrochers 464
Introduction
Global amphibian declines have been reported during the
last decades (Barinaga 1990; Vitt et al. 1990; Wyman 1990;
Wake 1991; Houlahan et al. 2000). Although investigators
denounce factors such as increased UV radiation, pathogens,
or combinations thereof, habitat loss and fragmentation re-
main the most frequently mentioned agents believed respon-
sible for these declines (Alford and Richards 1999; Davidson
et al. 2002; Johnson et al. 2002). Indeed, it is well estab-
lished that habitat loss and fragmentation, through the estab-
lishment of roads or certain forestry and agricultural prac-
tices, reduce amphibian abundance, species richness, or genetic
diversity (e.g., deMaynadier and Hunter 1995; Hitchings and
Beebee 1998; Vos and Chardon 1998; Kolozsvary and
Swihart 1999; Joly et al. 2001; Scribner et al. 2001). How-
ever, evidence for the processes behind these patterns re-
mains scarce (but see Johnston and Frid 2002; Rothermel
and Semlitsch 2002; Chan-McLeod 2003).
Matrix-dependent mobility is one of the key processes be-
hind population responses to habitat fragmentation. Patch
colonization is assumed to depend mainly on the distance
between patches (Hanski and Gilpin 1991), but there is
growing evidence that the quality of the matrix (e.g., cover,
temperature, humidity) is important for animal movements
(e.g., amphibians: Rothermel and Semlitsch 2002; Chan-
McLeod 2003; mammals: Desrochers et al. 2003). The ca-
pacity of an animal to successfully recolonize a patch will
depend on two major components. First, individuals must be
able to detect the patch through their perceptual range (sensu
Can. J. Zool. 83: 455–464 (2005) doi: 10.1139/Z05-032 © 2005 NRC Canada
455
Received 18 August 2004. Accepted 3 March 2005. Published
on the NRC Research Press Web site at http://cjz.nrc.ca on
18 May 2005.
M. J. Mazerolle1,2 and A. Desrochers. Centre de Recherche
en Biologie Forestière, Pavillon Abitibi-Price, Université
Laval, Québec, QC G1K 7P4, Canada.
1Corresponding author (e-mail: mmazerolle@usgs.gov).
2Present address: USGS Patuxent Wildlife Research Center,
12100 Beech Forest Road, Laurel, MD 20708-4017, USA.

Zollner and Lima 1997). Perceptual range can vary across
species, distance, and weather conditions (Yeomans 1995;
Zollner and Lima 1997, 1999; Gillis and Nams 1998), but its
extent of variation across habitat types remains undeter-
mined. Second, individuals must reach the patches that are
detected, which relates to the landscape’s permeability to
movements, also termed landscape resistance (Ricketts
2001). Impediments to an animal’s mobility include the ma-
trix quality, the distance between patches, as well as in-
creased predation and metabolic risks (Sjögren-Gulve 1994;
Larsen and Boutin 1994; Rothermel and Semlitsch 2002;
Turcotte and Desrochers 2003).
Pond-breeding amphibians use several habitats at different
times of the year to complete their life cycles (Sinsch 1990;
Pope et al. 2000). Amphibians can undertake migrations of a
few kilometres, but most move less than 400 m during such
movements (Dodd 1996; Semlitsch and Bodie 2003). In
human-disturbed landscapes, reaching the breeding, sum-
mering, or overwintering habitats often implies crossing hos-
tile environments that are dry or devoid of cover such as
open agricultural fields, forest clearcuts, or peatlands mined
for peat (Bury 1983; Chen et al. 1993; Wheeler and Shaw
1995; Price 1997). For amphibians, this can become a peril-
ous endeavor. With their permeable skin, amphibians require
moist environments, and few venture far from them (Sinsch
1990). Most species do not tolerate water losses exceeding
60% of their body water content (Thorson and Svihla 1943).
Thus, amphibian movements over areas devoid of cover can
be costly, not only because of increased predation risk but
also because of physiological costs.
We addressed issues of patch detection and landscape re-
sistance through a series of field experiments. We predicted
that amphibian movements over anthropogenically disturbed
surfaces are more difficult than on undisturbed surfaces. We
evaluated the ability and costs of moving over hostile areas
for northern green frogs (Rana clamitans melanota Rafinesque,
1820) and northern leopard frogs (Rana pipiens Schreber,
1782), two species commonly found in eastern North Amer-
ica (Wright and Wright 1949; Conant and Collins 1991). Al-
though common in the study area, both species are sensitive
to changes in the landscape resulting from urban develop-
ment, forestry, agriculture, and peat-mining activities (Bonin
et al. 1997; Knutson et al. 1999; Kolozsvary and Swihart
1999; Mazerolle 2001, 2003; Woodford and Meyer 2003). In
addition, these species exhibit strong site fidelity for their
breeding and summering areas (Martof 1953; Dole 1968).
Using frogs translocated from their summering habitat, we
compared the initial orientation and homing success of the
individuals moving across undisturbed and anthropogeni-
cally disturbed surfaces. We also tested whether individuals
avoid moving over a hostile surface, when given a choice
against a safer but longer route. Finally, we determined the
water loss and behaviour (e.g., burrowing, hiding, propped
up above substrate) associated with the exposure to different
types of substrates stemming from different anthropogenic
disturbances. These data will provide useful parameters in
the development of predictive models of the effects of habi-
tat disturbance in the landscape (e.g., spatially explicit mod-
els: Dunning et al. 1995; South 1999; Collingham and
Huntley 2000).
Materials and methods
Study area
All the experiments herein were conducted in eastern New
Brunswick, in the area surrounding Kouchibouguac National
Park. Mixed forest (Picea mariana (P. Mill.) B.S.P., Abies
balsamea (L.) P. Mill., Betula alleghaniensis Britt., Betula
papyrifera Marsh., Acer saccharum Marsh., and Acer ru-
brum L.) and peatlands form most of the landscape. Intense
forestry and peat-mining activities are the main disturbances
in the study area. These severely modify the landscape, leav-
ing bare soil or bare peat as potential barriers to amphibian
movements. Northern green frogs and northern leopard frogs
are common in eastern New Brunswick and occur in various
wetland habitats (Gorham 1970; McAlpine 1997). We used
individuals of both species in the experiments described be-
low in accordance with the Canadian Council on Animal
Care guidelines.
Peat mining
We conducted our study in a peatland-dominated land-
scape (i.e., bogs and fens). These acidic wetlands, mostly
occurring in northern countries (predominantly Canada and
Russia), cover approximately4×10
6km2worldwide
(Maltby and Proctor 1996). However, they have undergone
intensive draining for agriculture, forestry, and urbanization
and few remain unaltered in certain parts of Europe (Poulin
and Pellerin 2001). Attention has recently been directed to
the rapidly growing peat-mining industry, which extracts
large volumes of peat in Europe and eastern North America
for use in horticulture or as fuel (Wheeler and Shaw 1995;
Lavoie and Rochefort 1996). For instance, global peat-
mining production reached 93.7 × 106m3in 1997 (Daigle et
al. 2001).
Peatlands disturbed by peat mining share several charac-
teristics with agricultural fields and lands under certain for-
estry practices, and thus may serve as model systems.
During peat mining, the vegetation is removed, vast net-
works of ditches are established, and surfaces are levelled.
As a result, the peat fields, devoid of live vegetation, are dry
(see details in Mazerolle 2003). Because of the precarious
status of peatlands in many parts of the world, the known
negative impacts on the fauna and flora (Poulin et al. 1999;
Delage et al. 2000; Mazerolle 2003), and the drastic contrast
between mined and undisturbed surfaces, we used these sys-
tems for our landscape-scale experiments.
Orientation and homing success
To test the detection ability of frogs and landscape
resistance, we conducted orientation and homing experiments
on barren and undisturbed surfaces. We selected a pond
(perimeter = 105 m) in an undisturbed portion of Pointe-
Sapin Bog in eastern New Brunswick, Canada (46°57′N,
64°52′W). The pond is 70 m from the edge of barren peat,
and is used mostly as summering habitat by northern green
frogs and northern leopard frogs (M.J. Mazerolle, unpub-
lished data). In late April 2000, we erected a continuous alu-
minum drift fence (40 cm high above ground, 20 cm below
ground) on the pond’s perimeter ca. 2 m from the water’s
edge. We placed 11.4-L pitfall traps at 5-m intervals on both
© 2005 NRC Canada
456 Can. J. Zool. Vol. 83, 2005

sides of the fence (for trap design see Mazerolle 2003), for a
total of 24 traps on each side. Traps were opened from 5 May
2000 – 29 August 2000, and 28 May 2001 – 30 August
2001, when amphibans frequented the site. During the same
trapping periods, we placed 7 minnow traps (for minnow-
trap details see Mazerolle and Cormier 2003) to increase the
trapping effort and capture individuals already within the
fenced area. Traps were checked every day during the peak
of the season, and every other day afterwards. At the end of
summer, pitfall traps were closed with tight-fitting lids, min-
now traps were withdrawn, and parts of the fence removed
to allow individuals to move freely in the pond area between
trapping periods.
Northern green frogs and northern leopard frogs captured
at the pond were measured and marked for individual recog-
nition based on the Donnelly system (Donnelly et al. 1994).
Each individual was then placed for 10 min in a release de-
vice modified from Yeomans (1995) before starting the ex-
periments. The device consisted of an opaque 2-L container
fitted with a lid. We cut a pivoting trap door on the side of
the container on which a string was tied. This allowed the
investigator to open the door from behind without being seen
by the frogs.
Amphibians were put in the release device at the pond and
translocated from the pond according to the following treat-
ments. Each individual was assigned randomly to a distance
treatment (i.e., 35 or 70 m from the pond). Because the pond
is 70 m from the barren surface, all 35-m translocations
were on the undisturbed surface. For the 70-m transloca-
tions, we determined randomly whether the individual would
be placed on the barren or undisturbed surface (i.e., surface-
type treatment). The 70-m translocation on barren peat was
1 m into the peat field, across a large drainage ditch. Previ-
ous experimentations (n= 21 frogs) on the barren surfaces
indicated that northern green frogs and northern leopard
frogs maintained their orientation towards the pond of cap-
ture, regardless of the proximity of a ditch (M.J. Mazerolle,
unpublished data).
We randomly allocated the orientation of the opening of
the release device relative to the pond (towards or opposite).
At the start of each trial, the investigator opened the trap
door and retreated at least 10 m from the frog. The investi-
gator crouched down during the trials and remained behind
the frog to minimize disturbance. We recorded the initial ori-
entation of the frog (i.e., first frog movement 1 m from the
release area) relative to the pond. That is, we calculated the
minimum angular deviation between the orientation of the
pond and the frog (i.e., smallest angle between orientation of
the frog and that of the pond). Small pieces of flagging tape
were placed flush to the ground 1 m around the release de-
vice to provide points of reference. The observations were
terminated when the frog reached the perimeter delimited by
the pieces of tape (i.e., 1 m from the release device), or after
30 min.
We conducted the translocations after 1800 to reduce the
disturbance from the peat-mining activities, and to facilitate
frog movements, as individuals are usually most active later
in the day and evening (Oseen and Wassersug 2002). Air
temperature, wind velocity (i.e., low wind or moderate to
strong wind), and percent cloud cover were recorded during
each trial. Frogs were used only once in the experiment.
Translocated frogs recaptured at the pond, either in pitfall
traps or minnow traps, were noted to have successfully
homed to the pond.
We used linear and logistic regressions to model the log
of the angular deviation of frog orientation and the probabil-
ity of homing successfully to the pond, respectively. For
both analyses, the basic model consisted of the intercept and
the explanatory variables year, opening orientation of the re-
lease device, and species. We fitted a series of plausible
models including the variables surface type (i.e., barren vs.
undisturbed), the species × surface-type interaction, distance
(i.e., 35 vs. 70 m), snout–vent length (SVL), air temperature,
wind (i.e., no/low wind vs. high wind), and percent cloud
cover. Model fit was evaluated with the most complex
model. We evaluated the strength of evidence for each model
based on the second-order Akaike’s Information Criterion
adjusted for small sample sizes (AICc), following Burnham
and Anderson (2002). Estimates and standard errors (SEs)
for the parameters of interest were obtained with model-
averaging techniques (Anderson et al. 2000; Burnham and
Anderson 2002).
Avoidance of barren surfaces
As an additional measure of landscape resistance, we
tested whether translocated frogs avoid venturing onto bar-
ren surfaces when given a choice between a short route on a
barren surface and a longer route on an undisturbed surface.
In June 2001, we created a testing arena simulating both un-
disturbed bog vegetation and barren peat surfaces, on an
abandoned part of Pointe-Sapin Bog, bordered by a small
rectangular pond (Fig. 1). The arena was delimited by a
fence 45 cm high made of cloth used in landscaping. Using
spades, we collected the surface vegetation (i.e., sphagnum
(genus Sphagnum L.) moss, ericaceous shrubs, herbs) in-
cluding the roots and peat from the first 10 cm below the
surface of an adjacent undisturbed bog remnant. We then ar-
ranged the blocks of vegetation into two perpendicular corri-
© 2005 NRC Canada
Mazerolle and Desrochers 457
18 m
23·5 m
12 m
80 m
22 m
Pond
Fig. 1. Test arena used to determine the avoidance of barren peat
by amphibians (not drawn to scale). The open circle denotes the
point of release of the frogs Rana clamitans melanota and Rana
pipiens, the shaded surface corresponds to the vegetated corridor,
and the dotted surface represents the barren peat.

dors (2.25m×12mand2.25m×20m)inthearena. The
blocks of vegetation were packed tightly against one another
to reduce dehydration and watered every 48 h. The rest of
the arena was covered with 10 cm of loose peat found on the
barren surface of peat fields.
We captured northern green and northern leopard frogs
with dip nets in the pond and the vicinity for our experi-
ment. Each individual was measured to SVL, marked, and
placed in a release device (described above). The release de-
vice was placed on the corridor 13 m from the pond, with
the trap door either facing the pond (i.e., the risky shortcut)
or the end of the corridor (i.e., the safe detour). The assign-
ment of the treatment was completely randomized. Follow-
ing a 10-min acclimation period, the trap door was opened
and we began observations. The trial was terminated either
after 30 min or when the individual reached the pond, at
which time we noted whether the frog moved over the peat
or not. Trials were conducted after 1800, between 29 June
and 13 August 2001, and on days without precipitation. We
recorded air temperature, wind velocity, and cloud cover.
We evaluated the effect of frog size, species, air tempera-
ture, wind velocity, and cloud cover on the probability of
choosing the barren substrate with logistic regressions. All
models included the intercept and the orientation of the re-
lease device. Each model was ranked based on the AICc.
The estimates and SEs for the parameters of interest were
then computed with model-averaging techniques.
Dehydration
We quantified the physiological costs (i.e., dehydration) of
frogs moving in matrices within landscapes differing in hu-
man disturbance, and used dehydration rates as another mea-
sure of landscape resistance. We captured 126 northern green
frogs for the experiments described below during mornings
at several breeding ponds in the study area. Individuals were
temporarily housed in plastic containers with water for no
more than 24 h before starting the experiments. Each was
used only once. After the experimental trial, the individual
was marked by clipping a single digit and released at its
point of capture at the end of the day. We conducted the ex-
periments between 20 June and 15 August 2002, under simi-
lar meteorological conditions, between 1200 and 1800.
Before the start of each trial, each frog was carefully
cleaned to remove any particles on the skin and blotted dry
with a paper towel. We gently pressed on the abdomen of
the frog to empty the bladder. The frog was then weighed to
the nearest 0.1 g using a portable electronic scale (Acculab,
Huntingdon, Pennsylvania, USA). Following the initial
weighing, the frog was placed in a plastic container (54.5 cm ×
22 cm × 39 cm) that had a layer of 5 cm of one of three sub-
strates. We used a bare-soil substrate (sifted sandy soil) to
simulate conditions encountered during movements over ar-
eas remaining after certain agricultural or forestry practices,
whereas a bare-peat substrate (loose peat) was used to simu-
late movements over bogs undergoing peat mining. A third
substrate, consisting of a living carpet of moist sphagnum
moss taken from a peat bog in the study area acted as a con-
trol treatment. These substrates were likely to be encoun-
tered by frogs during migrations in the study area. During
the experimental trials, a single frog was introduced in each
container. The allocation of individuals to treatments was
completely randomized.
The dehydration experiment was conducted outside in
open mowed fields devoid of any vegetative cover >1 cm.
We covered each container with a nylon window screen fas-
tened with clothespins on the outer rim of the container to
prevent the escape of frogs during the trials. Half of the total
number of containers was placed under an opaque tarpaulin
ca. 1 m above the containers (shade treatment) to simulate
the cover provided by dense vegetation, whereas the other
half was not shaded. Each frog was submitted to a dehydra-
tion period of 2 h, and was weighed at 0, 1, and2hasde
-
scribed above. We calculated the change in mass at each
hour relative to body mass at the previous hour. We selected
a period of2htominimize stress to the frogs and believed
that this would approximate the exposure of frogs when
moving over substrates. Preliminary trials under the same
conditions yielded dehydration rates below the vital limits of
ca. 34% loss in body mass recorded for the species by
Thorson (1955) and Schmid (1965). Consequently, we refer
to frogs having undergone the 2-h dehydration period as
acutely dehydrated frogs.
Researchers have reported that certain species adopt spe-
cialized behaviours, such as burrowing in the soil or retreat-
ing to cavities, to reduce water losses under dry conditions
(Bentley 1966; Katz 1989; Schwarzkopf and Alford 1996;
Prather and Briggler 2001; Rohr and Madison 2003). We re-
corded the behaviour of the frogs in the containers at the end
of the first and second hours of the trials. Frogs were ap-
proached slowly to avoid modifying their behaviour and po-
sition. We characterized five types of behaviour: (1) hidden
in cavity (hidden in a small depression in substrate, but not
buried), (2) buried (partially or completely buried in sub-
strate), (3) crouched (head and body against substrate),
(4) head up (head is above substrate, but rest of body against
substrate), and (5) propped up (head and body above sub-
strate, front legs extended). We considered the first three as
behaviours minimizing the surface exposed to evaporative
water loss (i.e., an attempt to reduce dehydration).
The air temperature, percent cloud cover, and wind inten-
sity (low/no wind or moderate to strong wind) were recorded
during the dehydration trials. We took three samples of the
substrate at the start of the trial for each frog of the experi-
ment. The substrate samples were later dried in an oven at
200 °C to determine their percent water content.
We analyzed the change in mass with regression models
using generalized estimating equations (GEEs) (Diggle et al.
1994; Horton and Lipsitz 1999; Stokes et al. 2000) from the
GENMOD procedure in SAS/STAT®version 8.01 (SAS In-
stitute Inc. 1993). GEEs are an extension of generalized lin-
ear models and are specially adapted for repeated measures
(e.g., successively measuring mass in the same individual at
three different periods), yielding robust estimates of parame-
ters and SEs. We used a normal regression for repeated mea-
sures to evaluate the effects of shade, substrate type, wind
speed, cloud cover, and air temperature, on the mass (square
root transformed) lost each hour to dehydration. We ex-
pected a curved response of mass loss with frog size (SVL),
because small frogs have a greater surface to volume ratio
than large frogs, and thus, lose water faster (Thorson 1955;
Schmid 1965). Therefore, we included initial frog mass (i.e.,
© 2005 NRC Canada
458 Can. J. Zool. Vol. 83, 2005

© 2005 NRC Canada
Mazerolle and Desrochers 459
before dehydration) and initial frog mass squared in the
model. We also added the “shade × substrate type” interac-
tion, because these factors were crossed in our experiment.
Similarly, we evaluated the effect of the same variables
mentioned above (except initial frog mass squared) on the
frogs’ probability of minimizing their body surface exposed
during dehydration with a logistic regression for repeated
measures. We built a set of plausible candidate models and
assessed the strength of evidence for each with the AICcto
calculate model-averaged parameters and unconditional SEs.
Results
Orientation and homing success
Regardless of the species, frogs translocated on barren
peat tended to orient accurately towards the pond, as op-
posed to the individuals relocated on undisturbed surfaces
(Table 1). The mean (±SD) angular deviation of frogs trans-
located on the undisturbed surface at 35 and 70 m was 86.2° ±
52.9° and 86.9° ± 54.7°, respectively, whereas that of frogs
70 m on the barren surface was 20.1° ± 41.3°. Frog orienta-
tion did not vary across distance, species, frog size (SVL),
air temperature, wind velocity, or percent cloud cover. Large
frogs were more likely to home successfully than smaller
frogs. The effect of surface type on frog homing success was
less marked, but it suggested that individuals translocated on
the barren surface were less likely to successfully home to
the pond than those translocated on the undisturbed surface
(Table 2). Indeed, the 95% confidence interval for the vari-
able barely included 0, as indicated by the lower confidence
Number of
parameters ∆AICc*Akaike
weight
Model-averaged
parameter ±
unconditional SE
Model
Year, opening, species, air temperature, wind velocity, cloud cover, surface 8 0 0.35
Year, opening, species, surface 5 0.92 0.22
Year, opening, species, snout–vent length (SVL), surface 6 1.89 0.14
Parameter
Species (northern green vs. northern leopard frogs) –0.168±0.357
SVL –0.153±0.129
Distance (35 vs. 70 m) 0.209±0.311
Surface (barren vs. undisturbed) –2.540±0.292
Air temperature –0.033±0.021
Wind velocity 0.496±0.283
Cloud cover 0.003±0.003
Note: R2of the most complex models was 0.66. Estimate in boldface type indicates that 0 is excluded from the 95% confidence interval and that the
variable influences frog orientation. Interaction terms did not influence frog orientation and were not shown for brevity.
*AICcof highest ranked model was 242.05.
Table 1. Highest ranked linear regression models (i.e., change in Akaike’s Information Criterion adjusted for small sample sizes
(∆AICc)≤2) and estimates explaining the initial orientation (i.e., angular deviation) of frogs (Rana clamitans melanota and Rana
pipiens) across undisturbed and disturbed surfaces (n= 76 frogs).
Number of
parameters ∆AICc*Akaike
weight
Model-averaged
parameter ±
unconditional SE
Model
Year, opening, species, SVL, distance, surface 7 0 0.26
Year, opening, species, SVL, surface 6 1.04 0.15
Year, opening, species, distance, surface 6 1.64 0.11
Year, opening, species, SVL 5 1.95 0.10
Parameter
Species (northern green vs. northern leopard frogs) 0.364±0.746
SVL 0.528±0.269
Distance (35 vs. 70 m) –1.044±0.663
Surface (barren vs. undisturbed) –1.386±0.727
Air temperature –0.072±0.056
Wind velocity –1.144±0.665
Cloud cover –0.012±0.008
Note: R2of the most complex models was between 0.12 and 0.13. Estimate in boldface type indicates that 0 is excluded from the 95% confidence in-
terval and that the variable influences frog homing success. Interaction terms did not influence homing and were not shown for brevity.
*AICcof highest ranked model was 98.39.
Table 2. Highest ranked logistic regression models (i.e., ∆AICc≤2) and estimates explaining the probability of homing across undis-
turbed and disturbed surfaces (n= 84 frogs).

limit (i.e., 0.0383). The probability of homing was independ-
ent of distance, species, air temperature, wind velocity, and
cloud cover.
Avoidance of barren surfaces
During the selection experiment, the probability of frogs
moving across the peat did not vary across frog size (SVL),
species, or weather conditions (Table 3). Based on a satu-
rated log-linear model of the frequency of frogs of each spe-
cies venturing on each substrate, more frogs tended to avoid
the barren substrate than venture on it (18 vs. 7 frogs, re-
spectively; type 3 likelihood-ratio statistic = 6.77, df = 1,
P= 0.0093), regardless of species (type 3 likelihood-ratio
statistic = 0.02, df = 1, P= 0.8755). Those that proceeded
across peat moved a mean (±SD) of 8.6 ± 7.6 m over this
surface.
Dehydration
At the end of the 2-h dehydration period, the amount of
water lost by northern green frogs on the sphagnum moss
substrate was almost half of that lost on either the soil or the
peat substrates (Fig. 2). The sphagnum moss, soil, and peat
substrates had a water content of 91.3% ± 1.1% (mean ±
SE), 10.1% ± 0.5%, and 63.1% ± 1.1%, respectively. The
mass lost by northern green frogs was greatest on the soil
out of the shade (shade × substrate interaction in Fig. 2 and
Table 4). We detected a curvilinear response of water loss
with the frog mass before dehydration. Weather conditions
also influenced frog dehydration. Although water loss de-
creased with cloud cover and was greatest under windy con-
ditions, it was independent of air temperature.
During the first and second hours of dehydration, 53% of
frogs exhibited postures minimizing their body surface to
evaporative water loss. When out of the shade, northern
green frogs tended to minimize the body surface exposed to
evaporative water loss (Table 5). The small frogs reduced the
surface exposed to evaporation more often than larger frogs.
Substrate type, air temperature, wind speed, and cloud cover
did not influence frog behaviour.
Discussion
The results of the homing, barren-surface avoidance, and
dehydration experiments consistently indicate that barren
surfaces devoid of cover, following anthropogenic distur-
bances such as peat mining, are resistant to amphibian
movements. Northern green frogs and northern leopard frogs
avoided barren surfaces when offered a choice between
moving on the undisturbed and barren surfaces. For the first
time, we provide evidence that patterns of abundance in hos-
tile environments are the result of amphibian behaviour and
physiology. Indeed, in previous studies exclusively based on
trap rates in different environments, researchers concluded
that amphibians avoided open habitats without substantial
evidence. For instance, deMaynadier and Hunter (1999) and
Rothermel and Semlitsch (2002) reported that juvenile wood
frogs (Rana sylvatica LeConte, 1825), American toads (Bufo
americanus Holbrook, 1836), and spotted salamanders
(Ambystoma maculatum (Shaw, 1802)) avoided open-canopy
habitats, as fewer individuals were captured in traps in these
habitats. Similarly, Gibbs (1998) reported that certain adult
amphibians also seem to avoid other environments devoid of
© 2005 NRC Canada
460 Can. J. Zool. Vol. 83, 2005
Number of
parameters ∆AICc*Akaike
weight
Model-average
parameter ±
unconditional SE
Model
Opening, species 3 0 0.27
Opening, SVL, species 4 0.89 0.17
Opening, SVL 3 1.03 0.16
Parameter
Species (northern green vs. northern leopard frogs) –1.460±1.083
SVL 0.671±0.655
Air temperature –0.087±0.128
Wind velocity –0.628±1.042
Cloud cover 0.011±0.015
Note: R2of the most complex models was between 0.04 and 0.17. The 95% confidence intervals for all the estimates included 0,
indicating that the probability of frogs jumping on peat was independent of the variables in the model.
*AICcof highest ranked model was 32.89.
Table 3. Highest ranked logistic regression models (i.e., ∆AICc≤2) and estimates explaining the probability of
frogs jumping on the barren peat surface when given a choice between short route on barren peat surface and lon-
ger route on the undisturbed surface (n= 25 frogs).
0
0.5
1
1.5
2
2.5
3
PEAT SOIL SPHAGNUM
Mean mass lost (g/h)
No shade
Shade
Fig. 2. Mean (±1 SD) mass of water lost (g) per hour for north-
ern green frogs exposed to three different substrates and two
shade treatments.

cover, such as forest–road edges. These results likely stem
from the effects of dehydration observed in dry environ-
ments such as clearcuts, mined bogs, or open fields (Bury
1983; Chen et al. 1993; Wheeler and Shaw 1995; Price
1997).
For the first time in a field experiment, we contrasted am-
phibian dehydration rates and behaviour between substrates
associated with different degrees of human disturbance. We
found that northern green frogs lost the most water on the
soil substrate without shade. Consequently, cover such as
that provided by vegetation can greatly reduce amphibian
dehydration rates on dry substrates. Vegetative cover proba-
bly reduces evaporative water loss by providing shade and
shelter from the wind. Unshaded frogs had a greater ten-
dency to minimize their exposed body surface, whereas the
substrate type did not influence their behaviour. Such behav-
iours to minimize water loss are consistent with other re-
ports of a more fundamental nature than our study (Thorson
© 2005 NRC Canada
Mazerolle and Desrochers 461
Number of
parameters ∆AICc*Akaike
weight
Model-averaged
parameter ±
unconditional SE
Model
Shade, substrate, mass, mass squared, air temperature,
wind velocity, cloud cover, shade × substrate
11 0 0.96
Parameter
Interaction
Shade × peat 0.134±0.114
Shade × soil 0.395±0.129
Mass 0.064±0.007
Mass squared –0.001±0.0002
Air temperature 0.008±0.006
Wind velocity 0.211±0.052
Cloud cover –0.003±0.001
Note: R2of the global model was 0.69. Mass was the body mass before dehydration; mass squared was the squared body mass
before dehydration. Shade and sphagnum moss were the reference levels for shade and substrate treatments, respectively. Estimates
in boldface type indicate that 0 is excluded from the 95% confidence interval and that the variable influences the loss of frog mass.
*AICcof highest ranked model was 249.88.
Table 4. Highest ranked normal regression models for repeated measures (i.e., ∆AICc≤2) and estimates explaining
the mass of water lost (square root transformed) during dehydration of northern green frogs on different substrates
(n= 121 frogs).
Number of
parameters ∆AICc*Akaike
weight
Model-averaged
parameter ±
unconditional SE
Model
Shade, mass 3 0 0.27
Shade, substrate, mass, air temperature, shade × substrate 8 1.38 0.14
Shade, substrate, mass, air temperature 6 1.5 0.13
Shade, substrate, mass, shade × substrate 7 1.84 0.11
Shade, substrate, mass 5 1.85 0.11
Parameter
Shade 1.0662±0.475
Substrate
Peat 0.432±0.469
Soil 0.300±0.513
Mass –0.054±0.012
Air temperature 0.056±0.042
Wind velocity 0.027±0.431
Cloud cover 0.172±0.272
Note: R2of the global model was 0.18. Mass was the body mass before dehydration. Shade and sphagnum moss were the refer-
ence levels for shade and substrate treatments, respectively. Estimates in boldface type indicate that 0 was excluded from the 95%
confidence interval and that the variable influences frog behaviour. Interaction terms did not influence frog behaviour and were not
shown for brevity.
*AICcof highest ranked model was 288.83.
Table 5. Highest ranked logistic regression models for repeated measures (i.e., ∆AICc≤2) and estimates explaining
the probability of northern green frogs minimizing the surface exposed to evaporative water loss on different sub-
strates (n= 118 frogs).

and Svihla 1943; Packer 1963; Dole 1967; Parris 1998), and
may be an efficient adaptation to reduce water losses in un-
disturbed environments. However, in regularly disturbed en-
vironments, such as mined peat fields (i.e., harrowed several
times a day), these behaviours inevitably lead to death or se-
rious injury for animals seeking refuge in the substrate. In
other cases, amphibians may refrain altogether from burrow-
ing in substrates associated with human disturbances, even
under dry conditions (Jansen et al. 2001). These results sug-
gest that surfaces devoid of cover jeopardize the survival of
individuals attempting to cross them, and constitute barriers
to frog movements.
Frogs translocated on the barren surface were less likely
to home to the pond than those translocated on the undis-
turbed surface. This suggests that barren surfaces impede
frog movements. Habitat loss and fragmentation are known
to disrupt the movements of certain taxa, either directly, by
lack of cover (amphibians: Rothermel and Semlitsch 2002;
Johnston and Frid 2002; birds: St. Clair et al. 1998; Bélisle
et al. 2001; Bélisle and Desrochers 2002; mammals: Diffen-
dorfer et al. 1995; reptiles: Stanley 1998), or, less intuitively,
by restricting movement activity to optimal weather con-
ditions (Mazerolle 2001; Johnston and Frid 2002; Chan-
McLeod 2003). Furthermore, individuals are susceptible to
predation during movements (Larsen and Boutin 1994; Bon-
net et al. 1999) and are presumably obvious to predators in
areas without cover. This might also have decreased the
homing probability of frogs translocated on barren surfaces,
although predators were rarely seen on mined surfaces.
Nonetheless, based on our homing experiment, for routes of
equal distances, we should expect individuals moving under
cover to have the best chances of reaching a distant habitat
patch.
Small frogs had a particularly low probability of homing
successfully. This may be the result of a higher evaporative
water loss than in large individuals, such as that observed in
our dehydration experiment. Weather conditions limit the ac-
tivity patterns of amphibians (e.g., Mazerolle 2001; Johnston
and Frid 2002; Chan-McLeod 2003), as well as their poten-
tial to move across the landscape (Preest and Pough 1989).
For instance, Preest and Pough (1989) observed that the
most dehydrated American toads travelled the farthest at
intermediate temperatures. This relationship is potentially
greater for small individuals and requires further investiga-
tion. Regardless, it is essential to assess the distance thresholds
below which amphibians cross both hostile and favourable
environments successfully. This will yield pond-isolation
measures based on amphibian movements and physiology,
and considerably improve predictive models of amphibian
pond recolonization in complex landscapes.
Frogs translocated on barren peat oriented and moved to-
wards the pond on the undisturbed surface, whereas those
translocated on the undisturbed surface did not have a spe-
cific orientation. This may stem from an urgency to flee the
hostile conditions on the peat, whereas moisture and cover
are high on the undisturbed bog surface and render it suit-
able for foraging. At the distances we tested, substrate type
did not influence the frogs’ perceptual range. Small mam-
mals are generally capable of orienting relative to forest hab-
itat, when relocated in fields <30 m from the forest edge
(Zollner and Lima 1997; Gillis and Nams 1998). Schooley
and Wiens (2003) also recently reported directional
movements of an arthropod in an unsuitable matrix. In our
experiments, frogs were 70 m from the pond when trans-
located on the barren surface, but the undisturbed surface
was ca. 3 m from the point of release. We are conscious that
this may have helped frog orientation, but maintain that
comparison with individuals translocated on the undisturbed
surface is warranted. Indeed, additional translocations at
greater distances into mined surfaces yielded similar results
(M.J. Mazerolle, unpublished data).
Amphibians and landscape resistance
Based on the behaviours we observed in our experiments,
anthropogenically disturbed areas devoid of cover and espe-
cially those that offer dry substrates, such as barren peat, ag-
ricultural land, or recently cut stands, increase the resistance
of the landscape to amphibian movements. This is consistent
with low abundances of amphibians in cut forests or mined
peat bogs (e.g., deMaynadier and Hunter 1995; Grialou et al.
2000; Mazerolle 2003). These patterns are likely the result
of lack of moisture followed by direct mortality (from desic-
cation or predation), emigration, or subsequent avoidance of
such areas by amphibians.
We have shown, for the first time, that frogs can success-
fully orient in disturbed environments, and when a choice is
given, avoid them. Although certain individuals do venture
on hostile surfaces, their chance of moving successfully over
such areas is lower than for those moving on undisturbed
surfaces. Small individuals are least likely to home success-
fully. This has direct implications for connectivity and the
persistence of amphibians in the landscape, as recruitment
will be low in disturbed environments. Furthermore, our data
yield important movement parameters (i.e., orientation and
probability of homing successfully) to ameliorate simulation
models and strengthen predictions on the effects of habitat
disturbance in the landscape on amphibian populations (e.g.,
spatially explicit models: Dunning et al. 1995; South 1999;
Collingham and Huntley 2000).
Acknowledgements
We thank M. Cormier, M. Huot, A. Tousignant, C. Drolet,
and S. Boudreault for their assistance in the field. É.
Tremblay provided invaluable logistic support. Comments
from J. Bourque, C. Girard, and Y. Turcotte improved the
manuscript. Financial support was provided by the Natural
Sciences and Engineering Research Council of Canada
(NSERC), Fonds pour la Formation de Chercheurs et l’Aide
à la Recherche (FCAR), Quebec, and the New Brunswick
Wildlife Trust Fund to M.J. Mazerolle, and NSERC and
FCAR to A. Desrochers and L. Rochefort.
References
Alford, R.A., and Richards, S.J. 1999. Global amphibian declines:
A problem in applied ecology. Annu. Rev. Ecol. Syst. 30: 133–
165.
Anderson, D.R., Burnham, K.P., and Thompson, W.L. 2000. Null
hypothesis testing: problems, prevalence, and an alternative. J.
Wildl. Manag. 64: 912–923.
Barinaga, M. 1990. Where have all the froggies gone? Science
(Wash., D.C.), 247: 1033–1034.
© 2005 NRC Canada
462 Can. J. Zool. Vol. 83, 2005

© 2005 NRC Canada
Mazerolle and Desrochers 463
Bélisle, M., and Desrochers, A. 2002. Gap-crossing decisions by
forest birds: an empirical basis for parameterizing spatially-
explicit, individual-based models. Landsc. Ecol. 17: 219–231.
Bélisle, M., Desrochers, A., and Fortin, M.-J. 2001. Influence of
forest cover on the movements of forest birds: a homing experi-
ment. Ecology, 82: 1893–1904.
Bentley, P.J. 1966. Adaptations of Amphibia to arid environments.
Science (Wash., D.C.), 152: 619–623.
Bonin, J., Desgranges, J.-L., Rodrigue, J., and Ouellet, M. 1997.
Anuran species richness in agricultural landscapes of Québec:
foreseeing long-term results of road call surveys. In Amphibians
in decline: Canadian studies of a global problem. Edited by
D.M. Green. Society for the Study of Amphibians and Reptiles,
Saint Louis, Mo. pp. 141–149.
Bonnet, X., Naulleau, G., and Shine, R. 1999. The dangers of leav-
ing home: dispersal and mortality in snakes. Biol. Conserv. 89:
39–50.
Burnham, K.P., and Anderson, D.R. 2002. Model selection and
multimodel inference: a practical information–theoretic ap-
proach. 2nd ed. Springer-Verlag, New York.
Bury, R.B. 1983. Differences in amphibian populations in logged
and old growth redwood forest. Northwest Sci. 57: 167–178.
Chan-McLeod, A.C.A. 2003. Factors affecting the permeability of
clearcuts to red-legged frogs. J. Wildl. Manag. 67: 663–671.
Chen, J., Franklin, J.F., and Spies, T.A. 1993. Contrasting microcli-
mates among clearcut, edge, and interior of old-growth Douglas-
fir forest. Agric. For. Meteorol. 63: 219–237.
Collingham, Y., and Huntley, B. 2000. Impacts of habitat fragmen-
tation and patch size upon migration rates. Ecol. Appl. 10: 131–
144.
Conant, R., and Collins, J.T. 1991. A field guide to reptiles and
amphibians: eastern and central North America. Houghton
Mifflin Company, New York.
Daigle, J.-Y., Gautreau-Daigle, H., and Keys, D. 2001. Canadian
peat harvesting and the environment. Sustaining Wetlands Issues
Pap. No. 2001-1, North American Wetlands Conservation Coun-
cil (Canada), Ottawa.
Davidson, C., Shaffer, H.B., and Jennings, M.R. 2002. Spatial tests
of the pesticide drift, habitat destruction, UV-B, and climate-
change hypotheses for California amphibian declines. Conserv.
Biol. 16: 1588–1601.
Delage, V., Fortin, M.-J., and Desrochers, A. 2000. Effets de lisière
et d’isolement des habitats d’oiseaux chanteurs dans les
tourbières exploitées. Ecoscience, 7: 149–156.
deMaynadier, P.G., and Hunter, M.L., Jr. 1995. The relationship
between forest management and amphibian ecology: a review of
the North American literature. Environ. Rev. 3: 230–261.
deMaynadier, P.G., and Hunter, M.L., Jr. 1999. Forest canopy clo-
sure and juvenile emigration by pool-breeding amphibians in
Maine. J. Wildl. Manag. 63: 441–450.
Desrochers, A., Hanski, I.K., and Selonen, V. 2003. Siberian flying
squirrel responses to high- and low-contrast forest edges.
Landsc. Ecol. 18: 543–552.
Diffendorfer, J.E., Gaines, M.S., and Holt, R.D. 1995. Habitat frag-
mentation and movements of three small mammals (Sigmodon,
Microtus, and Peromyscus). Ecology, 76: 827–839.
Diggle, P.J., Liang, K.-Y., and Zeger, S.L. 1994. Analysis of longi-
tudinal data. Oxford University Press, Oxford.
Dodd, C.K., Jr. 1996. Use of terrestrial habitats by amphibians in
the sandhill uplands of north-central Florida. Alytes (Paris), 14:
42–52.
Dole, J.W. 1967. The role of substrate moisture and dew in the wa-
ter economy of leopard frogs, Rana pipiens. Copeia, 1967: 141–
149.
Dole, J.W. 1968. Homing in leopard frogs, Rana pipiens. Ecology,
49: 386–399.
Donnelly, M.A., Guyer, C., Juterbock, J.E., and Alford, R.A. 1994.
Techniques for marking amphibians. In Measuring and monitor-
ing biological diversity: standard methods for amphibians.
Edited by W.R. Heyer, M.A. Donnelly, R.W. McDiarmid,
L.-A.C. Hayek, and M.S. Foster. Smithsonian Institution Press,
Washington, D.C. pp. 277–284.
Dunning, J.B., Jr., Stewart, D.J., Danielson, B.J., Noon, B.R., Root,
T.L., Lamberson, R.H., and Stevens, E.E. 1995. Spatially ex-
plicit population models: current forms and future uses. Ecol.
Appl. 5: 3–11.
Gibbs, J.P. 1998. Amphibian movements in response to forest
edges, roads, and streambeds in southern New England. J.
Wildl. Manag. 62: 584–589.
Gillis, E.A., and Nams, V.O. 1998. How red-backed voles find hab-
itat patches. Can. J. Zool. 76: 791–794.
Gorham, S.W. 1970. The amphibians and reptiles of New Bruns-
wick. New Brunswick Museum, Saint John.
Grialou, J.A., West, S.D., and Wilkins, R.N. 2000. The effects of
forest clearcut harvesting and thinning on terrestrial salaman-
ders. J. Wildl. Manag. 64: 105–113.
Hanski, I., and Gilpin, M. 1991. Metapopulation dynamics: brief
history and conceptual domain. Biol. J. Linn. Soc. 42: 3–16.
Hitchings, S.P., and Beebee, T.J.C. 1998. Loss of genetic diversity
and fitness in common toad (Bufo bufo) populations isolated by
inimical habitat. J. Evol. Biol. 11: 269–283.
Horton, N.J., and Lipsitz, S.R. 1999. Review of software to fit gen-
eralized estimating equation regression models. Am. Stat. 53:
160–169.
Houlahan, J.E., Findlay, C.S., Schmidt, B.R., Meyer, A.H., and
Kuzmin, S.L. 2000. Quantitative evidence for global amphibian
population declines. Nature (Lond.), 404: 752–755.
Jansen, K.P., Summers, A.P., and Delis, P.R. 2001. Spadefoot toads
(Scaphiopus holbrookii holbrookii) in an urban landscape: ef-
fects of nonnatural substrates on burrowing in adults and juve-
niles. J. Herpetol. 35: 141–145.
Johnson, P.T.J., Lunde, K.B., Thurman, E.M., Ritchie, E.G., Wray,
S.N., Sutherland, D.R., Kapfer, J.M., Frest, T.J., Bowerman, J.,
and Blaustein, A.R. 2002. Parasite (Ribeiroia ondatrae) infec-
tion linked to amphibian malformations in the western United
States. Ecol. Monogr. 72: 151–168.
Johnston, B., and Frid, L. 2002. Clearcut logging restricts the
movements of terrestrial Pacific giant salamanders (Dicampto-
don tenebrosus Good). Can. J. Zool. 80: 2170–2177.
Joly, P., Miaud, C., Lehmann, A., and Grolet, O. 2001. Habitat ma-
trix effects on pond occupancy in newts. Conserv. Biol. 15:
239–248.
Katz, U. 1989. Strategies of adaptation to osmotic stress in anuran
amphibia under salt and burrowing conditions. Comp. Biochem.
Physiol. A, 93: 499–503.
Knutson, M.G., Sauer, J.R., Olsen, D.A., Mossman, M.J.,
Hemesath, L.M., and Lannoo, M.J. 1999. Effects of landscape
composition and wetland fragmentation on frog and toad abun-
dance and species richness in Iowa and Wisconsin, U.S.A.
Conserv. Biol. 13: 1437–1446.
Kolozsvary, M.B., and Swihart, R.K. 1999. Habitat fragmentation
and the distribution of amphibians: patch and landscape corre-
lates in farmland. Can. J. Zool. 77: 1288–1299.
Larsen, K.W., and Boutin, S. 1994. Movements, survival, and set-
tlement of red squirrel (Tamiasciurus hudsonicus) offspring.
Ecology, 75: 214–223.
Lavoie, C., and Rochefort, L. 1996. The natural revegetation of a

harvested peatland in southern Québec: a spatial and
dendroecological analysis. Ecoscience, 3: 101–111.
Maltby, E., and Proctor, M.C.F. 1996. Peatlands: their nature and
role in the biosphere. In Global peat resources. Edited by E.
Lappalainen. International Peat Society, Jyskä, Finland. pp. 11–
19.
Martof, B. 1953. Home range and movements of the green frog,
Rana clamitans. Ecology, 34: 529–543.
Mazerolle, M.J. 2001. Amphibian activity, movement patterns, and
body size in fragmented peat bogs. J. Herpetol. 35: 13–20.
Mazerolle, M.J. 2003. Detrimental effects of peat mining on am-
phibian abundance and species richness in bogs. Biol. Conserv.
113: 215–223.
Mazerolle, M.J., and Cormier, M. 2003. Effects of peat mining in-
tensity on green frog (Rana clamitans) occurrence in bog ponds.
Wetlands, 23: 709–716.
McAlpine, D.F. 1997. Historical evidence does not suggest New
Brunswick amphibians have declined. In Amphibians in decline:
Canadian studies of a global problem. Edited by D.M. Green.
Society for the Study of Amphibians and Reptiles, Saint Louis,
Mo. pp. 117–127.
Oseen, K.L., and Wassersug, R.J. 2002. Environmental factors in-
fluencing calling in sympatric anurans. Oecologia (Berl.), 133:
616–625.
Packer, W.C. 1963. Dehydration, hydration, and burrowing behav-
ior in Heleioporus eyrei (Gray) (Leptodactylidae). Ecology, 44:
643–651.
Parris, M.J. 1998. Terrestrial burrowing ecology of newly meta-
morphosed frogs (Rana pipiens complex). Can. J. Zool. 76:
2124–2129.
Pope, S.E., Fahrig, L., and Merriam, H.G. 2000. Landscape
complementation and metapopulation effects on leopard frog
populations. Ecology, 81: 2498–2508.
Poulin, M., and Pellerin, S. 2001. La conservation des tourbières:
le contexte international, canadien et québécois. In Écologie
des tourbières du Québec–Labrador: une perspective nord-
américaine. Edited by S. Payette and L. Rochefort. Presses de
l’Université Laval, Laval, Que. pp. 503–518.
Poulin, M., Rochefort, L., and Desrochers, A. 1999. Conservation
of bog plant species assemblages: assessing the role of natural
remnants in mined sites. Appl. Veg. Sci. 2: 169–180.
Prather, J.W., and Briggler, J.T. 2001. Use of small caves by anu-
rans during a drought period in the Arkansas Ozarks. J.
Herpetol. 35: 675–678.
Preest, M.R., and Pough, F.H. 1989. Interaction of temperature and
hydration on locomotion of toads. Funct. Ecol. 3: 693–699.
Price, J. 1997. Soil moisture, water tension, and water table rela-
tionships in a managed cutover bog. J. Hydrol. 202: 21–32.
Ricketts, T.H. 2001. The matrix matters: effective isolation in frag-
mented landscapes. Am. Nat. 158: 87–99.
Rohr, J.R., and Madison, D.M. 2003. Dryness increases predation
risk in efts: support for an amphibian decline hypothesis.
Oecologia (Berl.), 135: 657–664.
Rothermel, B.B., and Semlitsch, R.D. 2002. An experimental in-
vestigation of landscape resistance of forest versus old-field
habitats to emigrating juvenile amphibians. Conserv. Biol. 16:
1324–1332.
SAS Institute Inc. 1993. SAS/STAT®software: The GENMOD
procedure. SAS Tech. Rep. P-243, SAS Institute Inc., Cary, N.C.
Schmid, W.D. 1965. Some aspects of the water economies of nine
species of amphibians. Ecology, 46: 261–269.
Schooley, R.L., and Wiens, J.A. 2003. Finding habitat patches and
directional connectivity. Oikos, 102: 559–570.
Schwarzkopf, L., and Alford, R.A. 1996. Desiccation and shelter-
site use in a tropical amphibian: comparing toads with physical
models. Funct. Ecol. 10: 193–200.
Scribner, K.T., Arntzen, J.W., Cruddace, N., Oldham, R.S., and
Burke, T. 2001. Environmental correlates of toad abundance and
population genetic diversity. Biol. Conserv. 98: 201–210.
Semlitsch, R.D., and Bodie, J.R. 2003. Biological criteria for
buffer zones around wetlands and riparian habitats for amphibi-
ans and reptiles. Conserv. Biol. 17: 1219–1228.
Sinsch, U. 1990. Migration and orientation in anuran amphibians.
Ethol. Ecol. Evol. 2: 65–79.
Sjögren-Gulve, P. 1994. Distribution and extinction patterns within
a northern metapopulation of the pool frog, Rana lessonae.
Ecology, 75: 1357–1367.
South, A. 1999. Dispersal in spatially explicit population models.
Conserv. Biol. 13: 1039–1046.
Stanley, M.C. 1998. Homing in the skink, Oligosoma grande,
within a fragmented habitat. J. Herpetol. 32: 461–464.
St. Clair, C.C., Bélisle, M., Desrochers, A., and Hannon, S.J. 1998.
Winter response of forest birds to habitat corridors and
gaps. Conserv. Ecol. [online], 2: 13. Available from http://
www.consecol.org/vol2/iss2/art13 [cited 15 August 2004].
Stokes, M.E., Davis, C.S., and Koch, G.G. 2000. Categorical data
analysis using the SAS system. 2nd ed. SAS Institute Inc., Cary,
N.C.
Thorson, T., and Svihla, A. 1943. Correlation of the habitats of am-
phibians with their ability to survive the loss of body water.
Ecology, 24: 374–381.
Thorson, T.B. 1955. The relationship of water economy to
terrestrialism in amphibians. Ecology, 36: 100–116.
Turcotte, Y., and Desrochers, A. 2003. Landscape-dependent re-
sponse to predation risk by forest birds. Oikos, 100: 614–618.
Vitt, L.J., Caldwell, J.P., Wilbur, H.M., and Smith, D.C. 1990. Am-
phibians as harbingers of decay. Bioscience, 40: 418.
Vos, C.C., and 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.
Wake, D.B. 1991. Declining amphibian populations. Science
(Wash., D.C.), 253: 860.
Wheeler, B.D., and Shaw, S.C. 1995. Restoration of damaged
peatlands. Her Majesty’s Stationery Office, London.
Woodford, J.E., and Meyer, M.W. 2003. Impact of lakeshore devel-
opment on green frog abundance. Biol. Conserv. 110: 277–284.
Wright, A.H., and Wright, A.A. 1949. Handbook of frogs and
toads of the United States and Canada. Comstock Publishing
Company, Ithaca, N.Y.
Wyman, R.L. 1990. What’s happening to the amphibians? Conserv.
Biol. 4: 350–352.
Yeomans, S.R. 1995. Water-finding in adult turtles: random search
or oriented behaviour? Anim. Behav. 49: 977–987.
Zollner, P.A., and Lima, S.L. 1997. Landscape-level perceptual
abilities in white-footed mice: perceptual range and the detec-
tion of forest habitat. Oikos, 80: 51–60.
Zollner, P.A., and Lima, S.L. 1999. Illumination and the perception
of remote habitat patches by white-footed mice. Anim. Behav.
58: 489–500.
© 2005 NRC Canada
464 Can. J. Zool. Vol. 83, 2005