Content uploaded by Stephen C Sillett
Author content
All content in this area was uploaded by Stephen C Sillett
Content may be subject to copyright.
Herpetological Conservation and Biology 1(1):16-26
Submitted: June 15, 2006; Accepted: July 19, 2006
16
INTRODUCTION
The temperate salamanders of North America are primarily
terrestrial and fossorial, except some species in the family
Plethodontidae that have been reported to occupy moist vertical
rock faces (genus Desmognathus) and several species (genus
Aneides) that climb into trees at least seasonally (Petranka 1998;
Waldron and Humphries 2005). However, there has been no
conclusive evidence of a temperate zone salamander species
completing its entire life cycle in an arboreal environment. This
report documents the year-round residency of the wandering
salamander, Aneides vagrans, in the canopy of old-growth
redwood forest in northwestern California.
Recent genetic evidence (Jackman 1998) indicated that the
clouded salamander (Aneides ferreus) consisted of two separate
species. A new species, the wandering salamander (A. vagrans),
was proposed for populations south of the south fork of the
Smith River in northwestern California. This species occurs
primarily in northern California with disjunct populations that
were introduced to Vancouver Island, British Columbia where
they are abundant in terrestrial habitats (Jackman 1998; Davis
2002b). The name A. ferreus was retained for populations that
occur primarily in western Oregon.
Aneides vagrans has a prehensile tail that it uses to assist in
climbing vertical surfaces (Petranka 1998; Spickler and Sillett,
pers. obs.) and long limbs with slender digits bearing sub-
terminal toe pads (Petranka 1998). This species has previously
been described as a primarily terrestrial salamander that is also
found on logs, in trees, and on shrubs. It occupies moist
terrestrial habitats, especially under exfoliating bark and in
cracks and cavities of decomposing logs, stumps, snags, and
talus (Davis 2002a; Stebbins 2003). Similarly, A. ferreus has
climbing ability with individuals found as high as 6.5 m in trees
and, in the laboratory, will leap from the hand to nearby objects,
clinging with great tenacity, even to vertical surfaces (Nussbaum et al.
1983). The arboreal salamander (A. lugubris) has been found in trees
over 18 m above ground, and may deposit eggs in decay holes in live
oak trees up to 9 m above ground (Staub and Wake 2005).
The first evidence that A. vagrans might reside in the temperate forest
canopies of the redwood region was the discovery of a clutch of eggs
(later hatched in the lab) inside a leatherleaf fern (Polypodium scouleri
Hook. & Grev.) mat that had been dislodged from high in the crown of
a redwood being felled for lumber (Welsh and Wilson 1995). Soon
after the first in situ scientific investigations of old-growth redwood
forest canopies began in 1996, we observed the arboreal presence of A.
vagrans (Sillett 1999). All observations were made of individuals and
pairs occupying tunnels and cavities in large epiphytic fern mats in
trees, except one observation (SCS) of a mummified adult found in a
shallow trunk cavity located 88 m above the ground in a large redwood
tree.
Our objective was to study A. vagrans inhabiting an old-growth
forest canopy in Prairie Creek Redwoods State Park, Humboldt County,
California, including several trees whose crowns have been explored by
two of us (JCS & SCS) since 1996. In particular, we investigated
habitat use, activity patterns, and movements in the crowns of five large
redwood trees to glean new information on the ecology of A. vagrans in
trees.
The Redwood Forest Canopy Environment.—Old-growth forests
dominated by Sequoia sempervirens (D. Don) Endl. (hereafter
‘redwood’) are home to some of the world’s tallest and largest trees.
Individuals can exceed 112 m in height, 7 m in diameter, and have
wood volumes over 1,000 m3 (Sawyer et al. 2000). Old-growth
redwood forests contain some of the oldest and most structurally
complex trees on the planet. These trees often live over 1000 years and
develop highly individualized crowns shaped by natural forces
EVIDENCE OF A NEW NICHE FOR A NORTH AMERICAN SALAMANDER:
ANEIDES VAGRANS RESIDING IN THE CANOPY OF OLD-GROWTH
REDWOOD FOREST
JAMES C. SPICKLER1, STEPHEN C. SILLETT1, SHARYN B. MARKS1,
AND HARTWELL H. WELSH, JR.2,3
1Department of Biological Sciences, Humboldt State University, Arcata, CA 95521, USA
2Redwood Sciences Laboratory, USDA Forest Service, 1700 Bayview Drive, Arcata, CA 95521, USA
3Corresponding Author, email: hwelsh@fs.fed.us
Abstract.—We investigated habitat use and movements of the wandering salamander, Aneides vagrans, in an old-growth forest
canopy. We conducted a mark-recapture study of salamanders in the crowns of five large redwoods (Sequoia sempervirens) in
Prairie Creek Redwoods State Park, California. This represented a first attempt to document the residency and behavior of A.
vagrans in a canopy environment. We placed litter bags on 65 fern (Polypodium scouleri) mats, covering 10% of their total
surface area in each tree. Also, we set cover boards on one fern mat in each of two trees. We checked cover objects 2–4 times per
month during fall and winter seasons. We marked 40 individuals with elastomer tags and recaptured 13. Only one recaptured
salamander moved (vertically 7 m) from its original point of capture. We compared habitats associated with salamander
captures using correlation analysis and stepwise regression. At the tree-level, the best predictor of salamander abundance was
water storage by fern mats. At the fern mat-level, the presence of cover boards accounted for 85% of the variability observed in
captures. Population estimates indicated that individual trees had up to 29 salamanders. Large fern mats have high water-
holding capacities, which likely enable year-round occupation of the canopy by A. vagrans. Other observations indicate that A.
vagrans and its close relative A. ferreus also occupy additional habitats in forest canopies, especially moist cavities inside decaying
wood.
Key Words.—Aneides vagrans, A. ferreus, Sequoia sempervirens, forest canopy, arboreal habitat use, salamander
Herpetological Conservation and Biology 1(1):16-26
17
(Van Pelt 2001). Disturbances (e.g., windfall, crown fires) that
increase light availability within tree crowns stimulate new
growth from damaged trunks and branches. In redwood, this
new growth can be in the form of either horizontal branches or
vertical trunks (hereafter reiterated trunks), each with its own set
of branches (Sillett 1999). Reiterated trunks can originate from
other trunks or from branches. When a trunk arises from a
branch, the branch thickens in response to the added weight and
hydraulic demand of the trunk, creating a “limb.” Trunks,
limbs, and branches also become fused with each other during
crown development (Sillett and Van Pelt 2001). The highly
individualized crowns of complex redwoods offer a myriad of
substrates and habitats for epiphytic plants and other arboreal
organisms (Williams 2006).
Crown-level complexity in redwoods promotes accumulation
of organic material, including epiphytic plants, on tree surfaces
(Sillett and Bailey 2003). Crotches between the trunks, the
upper surfaces of limbs and branches, and the tops of snapped
trunks provide platforms for debris accumulation. Vertical and
horizontal sections of dead wood also provide substrates for
fungal decomposition. Over time, this debris develops into soil
as organic materials decompose into humus, which provides a
rooting medium for vascular plants. The most abundant
vascular epiphyte in redwood rain forests is the evergreen fern,
P. scouleri (Sillett 1999), with individual trees supporting up to
742 kg dry mass of these ferns and their associated soils
(hereafter ‘fern mats,’ Sillett and Bailey 2003). As fern mats
grow in size and number, their effects on within-crown
microclimates become pronounced. Like a sponge, large fern
mats store water within the crown, increasing the humidity
(Ambrose 2004) and providing refuge for desiccation-sensitive
species, including mollusks, earthworms, and a wide variety of
arthropods (Sillett 1999; Jones 2005). Large fern mats also tend to be
internally complex, with tunnels and cavities between the rhizomes and
dense roots as well as interstitial space around embedded sticks
(Stephen Sillett, pers. obs.).
MATERIALS AND METHODS
Study Area.—We studied A. vagrans in five redwood trees located in
Prairie Creek Redwoods State Park (PCRSP), Humboldt County,
California within an old-growth redwood forest. Mean annual rainfall
in the study area was 1.67 m, with summer temperatures ranging from
7°–31° C and winter temperatures ranging from 1°–23°C during 2002–
2004. Trees were selected from a 1-ha permanent reference stand that
is 50 m elevation and 7 km from the Pacific Ocean. Within the
reference stand, redwood accounts for 95.8% of the trunk basal area
with the remainder consisting of Douglas-fir (Pseudotsuga menziesii
[Mirb.] Franco), hemlock (Tsuga heterophylla [Raf.] Sarg.), and a few
hardwoods.
We selected study trees (Fig. 1) on the basis of size, structural
complexity, and epiphyte abundance. Trees 1 (‘Kronos’) and 2 (‘Rhea’)
have interdigitating sections of their crowns, where fern-covered
branches and limbs allow the possibility of salamander movement from
tree-to-tree without going to the ground. Tree 3 (‘Demeter’) stands 16
m from Kronos and Rhea. Its crown does not interact with these trees,
so movement of a salamander between them would require ground
contact. Trees 4 (‘Prometheus’) and 5 (‘Iluvatar’) stand over 50 m from
each other and the other trees; they were selected because of their high
crown-level structural complexity and epiphyte loads.
Tree access.—We achieved access to tree crowns by using a high-
powered compound bow mounted to an open-face fishing reel. A
rubber-tipped arrow trailing fishing filament was shot over branches
FIGURE 1. A two-dimensional display (view angle = 120°) of the three-dimensional crown structure of five redwood trees surveyed in this study. Main
trunks and reiterated trunks are shaded gray. Limbs are indicated by thin, black lines. No branches are shown. Locations of Polypodium scouleri fern
mats and Aneides vagrans captures are shown according to the legend. Note that “floating” symbols indicate locations on branches. “Sampled mats” are
fern mats that were selected for placement of cover objects.
Spickler et al.—Aneides vagrans in Redwood Canopies September 2006
18
high in the crown, and a nylon cord was then reeled back over
the branches and used to haul a 10 mm diameter static
kernmantle climbing rope into the crown and back to the
ground. One end of the climbing rope was then anchored at
ground level, and the other end was climbed via single rope
technique (Moffett and Lowman 1995). We had access to the
rest of the crown via arborist-style rope techniques (Jepson
2000; Fig. 2). The climbing rope was threaded through a pulley
hung from a sturdy branch near the treetop. The rope could be
easily replaced with nylon cord when the tree was not being
climbed.
Tree crown mapping.—We described tree crowns by
measuring dimensions of the main trunk and all reiterated trunks
with a basal diameter over 5 cm. We measured trunk diameters
at 5 m height intervals. For reiterations arising from the main
trunk or other reiterated trunks, we recorded: top height, base
height, basal diameter, and distance and azimuth (i.e., compass
direction) of base and top from center of main trunk. For
reiterations arising from limbs we recorded the following
additional measurements: limb basal diameter, diameter of limb
at the base of the reiteration, and limb height of origin. Thus, the
XYZ coordinates and architectural context of every measured
diameter could be determined for use in 3-dimensional mapping.
Total tree height was determined by dropping a tape from the
uppermost foliage to average ground level.
We derived three structural variables and three fern mat variables
from the mapping data, including total fern mat mass (kg), fern mat
mass in crotches, proportion of fern mass in crotches, main trunk
volume (m3), reiterated trunk volume, and limb volume. Volumes of
main trunks, reiterated trunks, and limbs were estimated by applying the
equation for a regular conical frustum to the diameter data (Table 1)
such as:
Volume = Length × π/3 × (lower radius2 + lower radius × upper radius
+ upper radius2).
In each tree, we also determined the XYZ coordinates of all P. scouleri
fern mats by measuring their heights above ground as well as their
distances and azimuths from the main trunk. Fern mat size was
quantified by the following measurements: mat length, mat width,
average soil depth (calculated from multiple measurements with a metal
probe), and maximum frond length. We calculated surface areas of fern
mats by applying the equation for an ellipse:
Area = π × 0.5(mat length) × 0.5(mat width).
Surface area was multiplied by average soil depth to calculate fern
mat volume. Dry masses of all mats were estimated by applying the
following model equation (n = 18, R2 = 0.995; unpubl. data of Sillett
FIGURE 2. Left, a climber (Steve Sillett) ascends a large redwood in Prairie Creek Redwoods State Park, California, in search of Aneides vagrans.
(Photograph by Marie Antoine, compliments of Steve Sillett). Right, Humboldt State University students Naomi Withers and Cameron Williams search a
complex redwood crown for salamanders. This search area is 60 meters above the ground. (Photographed by James C. Spickler).
Herpetological Conservation and Biology 1(1):16-26
19
and Van Pelt):
Total mass (kg) = 32.912 × mat volume + 0.0250 × maximum
frond length.
To better visualize individual tree crown complexity, we
generated three-dimensional models of tree crowns using
Microsoft Excel and the crown structure data (Sillett and Van
Pelt, unpubl. data). We overlaid locations of fern mats and
salamander captures on the crown models via their XYZ
coordinates (Fig. 1). We used this information to quantify
movements of salamanders captured more than once during the
study.
Capturing salamanders.—To locate A. vagrans without
destructive sampling, we placed cover objects on fern mats
within each tree crown. We constructed cover objects from gray
fiberglass screening. We cut and folded materials to produce
flat envelope-like bags (hereafter ‘litter bags’) that were filled
with decomposing leaf litter and soil, producing both small (25
× 20 cm) and large (25 × 40 cm) bags. To limit introduction of
foreign materials to the canopy, only litter and soil from each
selected site were used to fill the bags.
Placement of litter bags was determined randomly. The total
surface area of a tree’s fern mats was calculated by summing the
surface areas of all the mats on the tree. Ten percent of the mat
area on each tree was covered such that half was covered by
each type of litter bag. The probability of an individual fern mat
being randomly selected for a given litter bag was proportional
to its surface area. Thus, some fern mats, especially large ones,
received multiple litter bags while others, especially small ones,
received none. The placement of individual litter bags on
selected fern mats was not done randomly. Instead, we spaced
the bags across the mats in an attempt to minimize the likelihood
of their being blown from the crown during storms. This
involved nestling the bags into relatively flat regions of the
mats. Wooden sticks were placed underneath each litter bag to
maintain crawl spaces for salamanders.
Besides litter bags we deployed cover boards, which were
crafted from pairs of 2-cm-thick boards cut into 25 x 25 cm
sections (Davis 1991). We placed boards together but
separated by parallel 1-cm-thick strips of wood that created a
crawl space for salamanders. Our cover boards were designed
to simulate preferred terrestrial habitats of A. vagrans: 6 mm
spaces between bark and heartwood with a smooth firm
surface (Davis 1991). This species is often found under the
splintered wood of recently fallen trees or exfoliating bark
(Davis 2002b; Stebbins 2003). We limited use of cover boards
for fear of causing injury to climbers and tourists visiting the
grove if the boards happened to fall from the trees. However,
we left two cover boards on a large fern mat in Prometheus
and one on a large fern mat in Iluvatar. These locations seemed stable
enough to prevent loss of the boards during storms. As an extra
precaution we equipped the boards with small lengths of cord anchored
to the tree.
Access restrictions.—Summer and spring observations were not
possible due to climbing restrictions to protect the nesting habitat of two
threatened species in the area: the Marbled Murrelet (Brachyramphus
marmoratus) and Northern Spotted Owl (Strix occidentalis caurina).
Thus, our field season was limited to the fall (late September) through
winter (end of January), during three field seasons from 2000 to 2002.
During these periods, we checked our cover objects 2-4 times per
month, weather permitting. We also made weekly checks of litter bags
and cover boards in Prometheus during the 2002-2003 field season, and
made one visit to Iluvatar during this time. During each visit, all cover
objects were checked. A description of any salamander activity, time
and location of each capture were recorded.
Marking salamanders.—We anesthetized captured A. vagrans using
a pH neutral solution of MS-222 (3-aminobenzoic acid ethyl ester)
achieved by combining 1.0 g MS-222 + 2.4 g sodium bicarbonate
dissolved in 500 ml distilled water. Once salamanders were immobile,
they were permanently and uniquely marked under anesthesia with 1 x
2 mm fluorescent alpha-numeric tags (Northwest Marine Technologies,
Inc., Seattle, Washington, USA) injected subcutaneously on the ventral
side of the tail immediately posterior to the vent. Photographs of dorsal
patterns were taken of salamanders too small to be injected with tags.
Marked animals were returned to their point of capture once fully
recovered from the MS 222. We recorded snout to vent length (i.e.,
from tip of snout to anterior margin of vent), total length, number of
costal folds between adpressed limbs, weight (to the nearest 0.1 g), sex
if recognizable by secondary sexual characteristics (e.g., shape of head,
presence of mental glands, cirri, eggs in oviducts), and any injuries or
other identifying marks.
Data analyses.—We used stepwise multiple regression analysis to
evaluate potential effects of individual fern mat characteristics on
TABLE 1. Summary of tree size, Polypodium scouleri fern mats, soil water storage, and salamander abundance in five redwood trees from Prairie Creek
Redwoods State Park, California. Soil water storage values are whole-tree annual averages derived from a model (Sillett and Van Pelt unpublished).
Salamander abundance is the number of Aneides vagrans captured on fern mats in each tree, excluding those captured with cover boards.
Tree: Rhea Demeter Kronos Iluvatar Prometheus
Height (m) 95.5 97.5 91.6 91.5 97.4
DBH (cm) 405 434 428 614 559
Main trunk volume (m3) 359.3 389.7 335.4 874.0 598.5
Reiterated trunk volume (m3) 1.2 20.2 30.5 162.5 63.1
Limb volume (m3) 1.5 6.4 14.5 24.6 3.2
Fern mat dry mass (kg) 205 39 275 249 352
Fern mat dry mass in crotches (kg) 8 6 18 97 249
Soil water storage (l) 1003 437 1561 1908 4416
Fern mat salamander abundance 2 3 8 7 14
TABLE 2. Estimated sizes of Aneides vagrans populations on large redwood trees
over two years derived from mark-recapture data using the Chapman (1951)
method (see Chao and Huggins 2005). Numbers in parentheses are one standard
error. Estimates are only for the portion of the arboreal population using fern
mats.
Salamander Abundance
Tree January 2002 January 2003
Prometheus 11 (0) 29 (8)
Iluvatar 11 (2) 20 (11)
Kronos 8 (4) –
Five trees
combined 54 (15) –
Spickler et al.—Aneides vagrans in Redwood Canopies September 2006
20
salamander abundance in those mats with cover objects (n = 65).
The following independent variables were included: percentage
of surface covered by litter bags, total area covered by litter
bags, total surface area, dry mass, and height above ground. The
number of cover boards on each fern mat (0, 1, or 2) was also
used as an independent variable to account for the potential
effects of this sampling technique. The dependent variable was
the number of salamander captures per mat.
We evaluated potential effects of fern mats and tree structure
on A. vagrans abundance using correlation analysis. Tree-level
independent variables (n = 5) included total fern mat mass (kg),
mass of fern mats in crotches, and the average amount of water
stored (l) in each tree’s fern mats throughout the year. This last
variable was derived from a canopy soil hydrology model
developed for the permanent reference stand that includes all of
the trees in this study (Sillett and Van Pelt, unpubl. data).
Structure variables included volumes (m3) of each tree’s main
trunk, reiterated trunks, and limbs. The dependent variable was
the number of marked animals per tree. We corrected for
sampling effort by dividing the actual number of visitations per
tree (n = 27–33) by the highest number of visitations for any
tree. We eliminated the potentially confounding effects of cover
boards by removing those two mats from the data set prior to the
analysis.
Tree-level salamander abundance was estimated with the
Chapman (1951) method (see Chao and Huggins 2005). We
used the unbiased estimator for population size (N):
1
1)1)(1( −
+
++
=
R
CM
N
where M = number of individuals marked in the first sample, C
= total number of individuals captured in the second sample, and
R = number of marked individuals recaptured in the second
sample. For this analysis, we made the following assumptions:
1) sampling was random; 2) the population was closed (i.e., no
immigration, emigration, birth, or death) within each field
season; 3) all animals had the same chance of being caught in
the first sample; 4) marking individuals did not affect their
catchability; 5) animals did not lose marks between sampling
intervals; and 6) all marks were reported on discovery in the
second sample. We recognize that there are limitations to this
method (see Pollack et al. 1990) but our small samples did not
permit a more sophisticated approach. As a consequence we
consider these estimations only as first approximations of
salamander abundance in fern mats.
Other salamander observations.—The inaccessibility of
study trees during the spring and summer greatly limited our
ability to make year-around observations of arboreal A. vagrans
activity. However, several relevant observations were made by
forest activists participating in “tree-sits” at other nearby
locations, and by scientists working in the canopy on research
unrelated to this study. We include a summary of these
anecdotal observations with our results because these accounts
fill gaps in our temporal record and provide documentation of
salamander presence in the canopy throughout the entire year.
RESULTS
Tree-level population estimates.—A total of 55 captures were
made of 42 individual A. vagrans, including 13 recaptures. One
individual was captured five times, two individuals were
captured four times, three individuals were captured twice, and 36
individuals were captured only once. Captured individuals ranged from
1.3–7.1 cm in SVL, 2.4–14.7 cm in total length, and 0.1–5.9 g in mass.
Salamanders were found in all five study trees with the most captures in
Prometheus (n = 28) and the least in Rhea (n = 2). Small sample sizes
forced us to use entire field seasons as sampling intervals to make
population estimates for each tree. Thus, A. vagrans abundance was
estimated once for three trees (Prometheus, Iluvatar, Kronos) in January
2002 for animals marked in the first field season and marked or
recaptured in the second field season (8–11 individuals per tree), and
again for two trees (Prometheus and Iluvatar) in January 2003 for
animals marked in the second field season and marked or recaptured in
the third field season (20–29 individuals per tree, Table 2). There were
insufficient data to make any tree-level population estimates for two of
the trees (Demeter and Rhea). However, we combined data from all
five trees to calculate an estimate of 54 salamanders for these five tree
crowns collectively in January 2002 based on animals marked in the
first field season and marked or recaptured in the second field season
(Table 2).
Tree-level effects on salamander abundance.—Based on correlation
analyses at the tree-level, there were two significant predictors of
salamander abundance per tree: average water storage by fern mats (r =
0.930, P = 0.022) and mass of fern mats in crotches (r = 0.885, P =
0.046). Our small sample size (n = 5 trees) prohibited further analyses
of tree-level effects for other fern mat variables (total fern mat mass,
proportion of total fern mat mass in crotches), and three structural
variables (main trunk volume, reiterated trunk volume, and limb
volume).
Effects of fern mat characteristics on salamander captures.—Fern
mat-level effects on A. vagrans captures and recaptures were evaluated
separately for a total of 65 fern mats (i.e., only those with cover objects)
in five trees using regression analysis. Total number of A. vagrans
captured, including recaptures, was positively associated with number
of cover boards (R2 = 0.85, P < 0.0001), area covered by litter bags (R2
= 0.38, P < 0.0001), fern mat mass (R2 = 0.28, P <0.0001), and fern mat
area (R2 = 0.22, P < 0.0001). No associations were found between
captures and either the percentage of fern mat surface area covered by
litter bags (R2 = 0.002, P = 0.70) or height (R2 = 0.004, P = 0.62).
Stepwise multiple regression analysis revealed that number of cover
boards (adjusted R2 = 0.85, P < 0.00001), fern mat mass (cumulative R2
= 0.90, P < 0.00001), and height of fern mat (cumulative R2 = 0.91, P <
0.03) all accounted for significant amounts of variation in the number of
salamander captures.
The strongest variable affecting the number of A. vagrans captured
was not a physical characteristic of the fern mats, but was an artifact of
our sampling technique. Significantly more salamanders were captured
on fern mats with cover boards than on mats with only litter bags. In
Prometheus, the total number of captures on one fern mat was 15,
representing 5 individuals. All of the captures were made in two cover
boards, although 8 litter bags occurred in close proximity to the cover
boards. Nine of the 15 captures were recaptures, including four of a
single large male who had apparently taken up residence in an area that
included both of the cover boards, which were located < 0.5 m apart.
He was captured during all 3 years of the study, and on several
occasions he was found with other salamanders. On one fern mat in
Iluvatar, there were 9 captures representing seven individuals. Seven of
these were made in a cover board, while the remaining two were made
under a litter bag located 75 cm away.
Movement of recaptured salamanders.—We found no evidence of
among-tree movements of marked salamanders, via interacting crowns
Herpetological Conservation and Biology 1(1):16-26
21
or the ground. Of the 13 recaptures, 12 were of individuals
found in the same locations as their initial captures. The single
exception was a juvenile A. vagrans (1.2 g, SVL= 4.35 cm)
found under a litter bag (first capture) and then recaptured a
week later on the surface of a fern mat 7.5 m higher in the tree.
Seasonal activity.—Our limited field season precluded
observations of seasonal differences in movement and habitat
use, but based on our findings and several anecdotal
observations made outside of our field seasons (see below), it
appears that at least some individual A. vagrans occupy the
forest canopy throughout the year.
Other observations.—The few spring and summer
observations were often made while canopy researchers were
conducting surveys for protected species (Marbled Murrelet and
Spotted Owl). Also, salamander observations were made by
non-scientists illegally occupying trees to protect them from
logging. It is understandable that the protection of threatened
species takes priority over new research dealing with a
salamander that appears to be abundant, at least in terrestrial
habitats, but the lack of data for these seasons left us with
several unknowns concerning the life history and ecology of A.
vagrans in redwood forests. The following observations may
help us to understand A. vagrans behavior during these periods.
The willingness of tree sitters to stay aloft for extended
periods enables them to make observations that scientists
working under research permits cannot afford to do. In the
spring of 2002, an activist designated as Remedy began a tree-sit
on private timber lands. Remedy, along with other activists,
established sleeping platforms in several large redwoods near
Freshwater, in coastal Humboldt County, California. Remedy
remained aloft for nearly a year before being forcibly removed
and arrested for trespassing. In that time period she made
numerous observations of a pair of wandering salamanders.
On seven occasions from April to September, Remedy
observed the “same pair” of wandering salamanders moving
within an area around a small cavity located 3 m from her living
platform. The original leader of the tree had broken at an
approximate height of 40 m; the living platform was located a
few meters below the break. The loss of the leader occurred at
least 100 years before, and two reiterated trunks had replaced it.
A zone of decaying wood that had formed around the break
created the cavity that the salamanders occupied. The same
cavity was also shared by a small “tree squirrel,” probably a
Douglas’ Tree Squirrel (Tamiasciurus douglasii) or a Northern
Flying Squirrel (Glaucomys sabrinus). A P. scouleri fern mat
occupied the top of the broken trunk.
Remedy often observed the salamanders moving in close
proximity to each other, but they appeared to be “moving
independently as if unaware of each other.” Most of the
salamander activity was limited to the area on and around the
fern mat, but on two occasions a salamander moved out along
branches and continued to the outer crown where it could no
longer be seen. All observations were made during early
evening and under similar microclimatic conditions: dry
substrate with elevated air humidity. Conditions were described
as “warm and muggy, perfect weather for flying insects.” One
stated impression was that the salamanders were more affected
by temperature than by moisture as no animals were observed
moving during the rain or immediately thereafter. There was
limited flying insect activity during and immediately after rain
storms. Observations were always made during calm
conditions with little or no wind. The two salamanders were
observed throughout the spring and summer with the last observation
occurring on 21 September 2002, when “evenings became too cold for
foraging.”
Remedy reported an A. vagrans eating while in the canopy. One
evening, she noted an insect, a “winged termite,” alight on a small
branch approximately 30 cm from the salamander. The salamander
then rapidly moved to the insect, which it ate without hesitation. After
a moment, the salamander continued moving along the branch to the
outer crown.
Similar observations were made by another activist, Raven,
participating in a tree-sit in the Van Duzen watershed in Humboldt
County. Raven made several observations of a pair of A. vagrans
foraging near his sleeping platform. He also described how A. vagrans
activity decreased along with decreasing nighttime temperatures as
autumn and winter approached. On 2 February 2003, he observed a pair
of A. vagrans move on to his platform. He watched them for several
minutes before they continued off into the darkness.
On 17 September 2002 at 0800 hrs, one of us (Stephen Sillett) and his
graduate student (A. Ambrose) observed an adult A. vagrans while
crown-mapping a large redwood in Humboldt Redwoods State Park.
The observation was made during warm conditions with high air
humidity and low cloud cover; the tree’s bark was dry. We observed a
single adult A. vagrans moving vertically along the trunk at a height of
93 m above ground. The salamander’s path was exposed with no soil or
obvious cover nearby. The nearest area of apparent cover was in a
cavity of dead wood located 100.6 m above the ground, but the surface
of this site was also exposed and dry. Obvious fissures and crevices in
the decaying wood, however, likely allow such animals to enter and
retire within damp cavities.
We also have made incidental observations of A. ferreus, a close
relative of A. vagrans. On three separate occasions in 2002, one of us
(James Spickler) and N. Bowman observed adult A. ferreus while
studying the nesting behavior of the red-tree vole (Arborimus pomo) in
old-growth Douglas-fir forests of coastal Oregon (BLM forest lands,
Salem and Eugene Districts). Observations were made in the summer
(July-August), midday during periods with high humidity and on moist
substrates. In all cases, salamanders were inactive and hidden within
the stick nests of a western grey squirrel (Sciurus griseus). Two of
these salamanders were found in an active nest containing fresh feces
and elevated temperatures from the recently departed rodent’s body.
In 1993, Stephen Sillett observed an A. ferreus while conducting
canopy research in a 700-year-old Douglas-fir forest (Middle Santiam
Wilderness Area, Willamette National Forest, Willamette County,
Oregon; see Sillett 1995). While climbing in a large Douglas-fir tree
adjacent to a 30-year-old clearcut, he found an adult salamander under
moss (Antitrichia curtipendula) on a large branch approximately 30 m
above the ground. After being disturbed, the salamander moved
horizontally across the branch and retreated under a bark flake on the
tree trunk. The observation was made midday during the dry season
(early autumn), and the moss mat was “merely damp.”
DISCUSSION
Plethodontid salamanders are unique in that they are the only
salamander family to have invaded the tropics, where many species
occupy arboreal niches (Lynch and Wake 1996). However, in spite of
the high number of species displaying arboreal habits in tropical forests,
little is known about this phenomenon beyond a few anecdotal accounts
(e.g., Good and Wake 1993; McCranie and Wilson 1993). Our results
here provide information on a new niche dimension for a North
American temperate zone plethodontid salamander, the resident use of
arboreal habitats in redwood forest canopies by Aneides vagrans.
Like other plethodontid salamanders, A. vagrans is lungless and
respires exclusively through its skin and buccopharynx. Presumably,
Spickler et al.—Aneides vagrans in Redwood Canopies September 2006
22
this requires the maintenance of skin moisture to facilitate
respiratory gas exchange (Shoemaker et al. 1992). The skin of
most amphibians is highly permeable to liquid and gas, allowing
for moisture exchange rates similar to those of standing water
(Spotila and Berman 1976). To avoid fatal desiccation,
amphibians have developed a variety of behavioral and
physiological means by which to control water loss (Shoemaker
et al. 1992). Plethodontid salamanders select habitats with
suitable microsites that retain relatively high moisture contents
as the macrosite begins to dry (Thorson 1955; Cunningham
1969; Ovaska 1988; Cree 1989; Shoemaker et al. 1992). This
desiccation-avoidance behavior has been observed in terrestrial
A. vagrans (Davis 2002b).
Our correlation analysis of tree-level effects on salamander
abundance highlights the importance of water storage in soils
beneath epiphytes and location of this material within the crown
(e.g., in crotches). Soils on limbs drain faster than those in
crotches (Ambrose 2004; Enloe et al. 2006) and thus may
become too dry for perennial occupancy by salamanders.
Microclimate data from fern mats show that crotches have more
stable moisture and temperature regimes than branches or limbs
(Ambrose 2004; Sillett and Van Pelt, unpubl. data). Compared
to those on branches or limbs, fern mats in crotches hold more
water per unit mass and store water longer (Sillett and Van Pelt,
unpubl. data). Furthermore, soils in crotches have higher bulk
densities and lower hydraulic conductances than soils on
branches or limbs (Enloe et al. 2006), providing relatively stable
refugia from desiccation during the dry season. Trees with soil
in deep crotches likely provide suitably moist arboreal habitats
for the year-round occupancy of old-growth redwood forest
canopies by A. vagrans, enabling this salamander to breed and
potentially live its entire life within tree crowns.
The effects of fern mat size and height on salamander
captures are ecologically interpretable. The positive correlation
between fern mat size and A. vagrans abundance can be
attributed to the larger surface area available for foraging, higher
water-holding capacity, and greater internal complexity of larger
fern mats. Although the arboreal feeding habits of A. vagrans
have not been studied, the salamanders probably take prey from
fern mats. Fern mat surfaces (at least seasonally) have more
invertebrate biomass than other surfaces (e.g., bark and foliage)
in redwood crowns (Jones 2005). In fact, the mites and
collembolans inhabiting fern mats experience population
explosions during the wet season, and have densities similar to
those observed in terrestrial habitats under similar conditions
(Jones 2005).
Larger, deeper fern mats have greater water storage and
slower rates of desiccation than smaller mats (Ambrose 2004),
thus providing more stable, moist microclimates conducive to A.
vagrans habitation. As a fern mat increases in size, new roots
and rhizomes grow to replace the old ones, which subsequently
decay. Although debris from litter fall, especially tree foliage, is
a major component of the P. scouleri fern mats, the majority of
organic material in these mats comes from P. scouleri itself,
especially humus derived from decaying roots and rhizomes
(Sillett and Bailey 2003). Dead, decomposing rhizomes leave
behind “tunnels” in the soil. Larger debris (e.g., branches) that
falls onto fern mats can also create tunnels and internal cavities
as it is covered by other debris and begins to decompose. On
three occasions, Sillett and Bailey (2003) found A. vagrans
occupying interstitial spaces in P. scouleri mats (mats were
being harvested for the development of equations to predict fern
mass). Also, an egg cluster of A. vagrans was found within a P.
scouleri mat on a freshly fallen old-growth redwood (Welsh and Wilson
1995). These observations suggest that the tunnels and cavities in fern
mats are used by A. vagrans, and it is likely that they are important
refugia, but the fragile nature of the substrate makes searching the
tunnels nearly impossible without permanently altering the habitat.
The negative effect of fern mat height on salamander captures can be
attributed to the varying microclimates at different heights within a
forest canopy. During periods with no precipitation, the upper canopy
receives more light and wind, and the air is less humid compared to the
lower canopy (Parker 1995, Sillett and Van Pelt, unpubl. data).
Therefore, fern mats in the upper canopy, regardless of size, are
subjected to more frequent and severe periods of desiccation than those
in the lower canopy. In redwood forest canopies this effect can be seen
in P. scouleri itself. Although fern mat size is not correlated with
height, the size and shape of fronds become progressively smaller with
increasing height in the forest (Sillett and Bailey 2003). The negative
effect of height on number of A. vagrans captured can be attributed to
the less stable microclimate of upper canopy fern mats compared to
those in the lower canopy. Fern mats higher in a tree may be important
for salamanders foraging during wet periods, but the prolonged
occupation of these sites may be risky during dry periods. This idea is
supported by our discovery of two mummified individuals near the tops
of two trees over 90 m tall (see also Maiorana 1977).
Dead wood may represent another important habitat for arboreal
salamanders in redwood forests. At the forest level, the average water
storage in dead wood (16,500 l ha-1) rivals the amount stored in canopy
soil (19,700 l ha-1), and seasonal variation in dead wood water storage is
less than that in soils on branches and limbs (Sillett and Van Pelt,
unpubl. data). Even though we did not quantify salamander abundance
in dead wood habitats, a number of anecdotal observations suggest that
A. vagrans use dead wood and hollow cavities. The highest observation
of this species ever made (93 m) was of a salamander climbing upwards
on a late summer morning towards the dead, broken top of a large
redwood nearly lacking vascular epiphytes and soil. It is likely that
large populations of A. vagrans reside within hollow trunks of standing
redwoods in old-growth forests.
Movement and territoriality.—If a salamander finds a habitat that
has a favorable moisture regime and sufficient prey availability, it
would be advantageous for the animal to stay in that habitat or return to
it frequently (Jaeger 1980). Terrestrial A. vagrans move only short
distances, are site-tenacious, and return periodically to particular
habitats within their home range (Davis 2002a). Our canopy findings
parallel these terrestrial observations.
On 6 occasions we captured more than one salamander on a fern mat.
Twice we found two males in a cover board with a single female. We
also found two females together with no male present and two males
together with no female present. Twice we found a pair of salamanders
on the same fern mat but not within the same cover board: a male with
a female and a male with another male. Males did not appear to be
defending females from other males, and neither sex appeared to be
defending a particular site, both of which are major components of
territorial behavior (Brown and Orians 1970; Jaeger et al. 1982; Mathis
et al. 1995). Similar behaviors were observed in terrestrial A. vagrans
on Vancouver Island, British Columbia (Davis 2002a). Although
arboreal A. vagrans in redwood forests appear to be acting similarly to
terrestrial individuals in British Columbia, we did not sample during
either the breeding season (presumably spring) or the summer.
Arboreal A. vagrans may behave differently during certain times of the
year if resources, such as nest sites, prey items, or moist habitats,
become limited.
Herpetological Conservation and Biology 1(1):16-26
23
Seasonality.—The seasonal restrictions on canopy research in
old-growth redwood forests prevented us from sampling
salamanders for eight months of the year, including the summer
dry season. When considering the hydric constraints of
plethodontid salamanders, it is likely that A. vagrans would be
less active on forest canopy surfaces during the dry season.
Anecdotal observations of A. vagrans, as well as observations of
other species of Aneides, suggest otherwise however. Arboreal
Aneides may be most active during the drier and warmer spring
and summer given the strong marine influences that contribute
to mild temperatures and high relative humidity during these
seasons in the coastal redwood forest (Sawyer et al. 2000).
Green salamanders (Aneides aeneus) use arboreal habitats
seasonally (Waldron and Humphries 2005). These animals
over-winter in rock outcrops and migrate into woody or arboreal
habitats (primarily hardwoods) during the onset of spring. They
remain in these habitats throughout the summer and breeding
season before returning to rock outcrops sometime in October
and November. Green Salamanders prefer larger (in diameter)
and more complex trees having a variety of visible cavities. On
dry days individuals are often found under the flaky and
furrowed bark of several different tree species. The maximum
height above the ground where Green Salamanders have been
observed is 21 m at the mouth of a tree hollow (Waldron and
Humphries 2005). Gordon (1952) noted a decrease in Green
Salamander abundance during the dry season, but Waldron and
Humphries (2005) demonstrated that the salamanders may be
climbing into the canopy where they remain throughout the
summer and cannot be easily detected from the ground.
It is unlikely that A. vagrans utilizes arboreal habitats
seasonally like A. aeneus. Anecdotal observations do not
support such a scenario but instead suggest year-round
residency. At most, A. vagrans may shift its use of particular
microhabitats within the canopy, but it remains unclear which
suitably moist locations might be preferred during different
seasons.
Effects of tree- and fern mat-level variables on
salamanders.—Our mat-level analysis indicated an effect of
cover boards on number of salamanders. Nearly half of the A.
vagrans captured during our study were found in cover boards
even though their use was quite limited. By placing cover
boards on top of fern mats we may have created a preferred
habitat type. This assertion is supported by observations that
dead wood substrates were favored by terrestrial A. vagrans
populations in Vancouver Island, British Columbia (Davis
2002b). It is unclear whether the salamanders captured in our
cover boards were residents of the fern mat on which the boards
were placed, originally residing in the tunnels and other
complexities of the fern mat, or if placing the cover boards on
the fern mat created a habitat allowing foraging individuals from
other parts of the tree the opportunity to stay and take up
residence. The paucity of recaptures under litter bags suggests
that A. vagrans prefers crevices but will use litter bags
opportunistically for cover while foraging.
Future research.—Future studies should examine how A.
vagrans uses other habitats besides fern mats, since the
preponderance of cover board captures as well as anecdotal
observations suggest that A. vagrans inhabits crevices, cavities,
and lodged woody debris at least as much as it does soil beneath
ferns (P. scouleri). Crevices and cavities are difficult to search
manually in a non-destructive fashion, and we discourage this
activity. Placing cover boards adjacent to these sites would
allow capture of salamanders coming out to forage on other tree
surfaces without permanently altering the habitat. Cover boards also
create new habitats within tree crowns for salamanders and their prey.
The entrance to natural and artificial crevices and cavities could be
monitored continuously (even during summer months when canopy
access is restricted) via motion-sensitive, infrared video cameras.
Microclimate data could also be compared to videos to determine
preferred conditions for foraging and also to document salamander
behavior throughout the annual cycle. Identification of salamanders via
videos would be possible using a visual implant fluorescent elastomer
marking technique and by marking individuals on their dorsal surfaces.
Hopefully our discovery of resident arboreality in Aneides vagrans
will trigger a renewed interest in studying adaptations of plethodontid
salamanders to the use of arboreal habitats. For example, the recently
described modification (Sapp 2002) of the typical plethodontid
courtship tail-straddling walk (Houck and Arnold 2003) from linear to
circular in the genus Aneides may be an adaptation to arboreality.
Conservation implications.—Resident populations of arboreal
salamanders in old-growth forests may be top predators of diverse,
heterotrophic communities fueled by the productivity of epiphytes and
the trees themselves. Such ecosystems appear to be lacking in younger
forests regenerating on logged-over land. Only 4% of the original old-
growth redwood forests remain, and second-growth forests originating
before 1930 are scarce (Noss 2000). Nearly all regenerating redwood
forests are younger than rotation age (~ 50 years) and consisting of trees
less than 40 m tall with small branches. As a consequence of their
simple structure, the biological diversity of young redwood forests is
low, and many old-growth-associated plants and animals are now
restricted to a few National and State Parks (Sawyer et al. 2000;
Cooperrider et al. 2000). Redwood forests capable of supporting
arboreal salamanders are rare outside of these Parks. Even within the
Parks, epiphytic vascular plants and associated soil communities do not
occur in the crowns of small trees despite their proximity to large trees,
because the structural complexity (e.g., large limbs and reiterated
trunks) necessary to support these organisms develops very slowly.
Thus, the arboreality of A. vagrans in redwood forests may be a
phenomenon restricted to a tiny portion of the landscape. Further
investigations to establish a basis for the conservation of A. vagrans and
associated organisms in protected forests are warranted.
Acknowledgments.—We thank the U.S. Forest Service’s Pacific
Southwest Research Station, Save-the-Redwoods League, and Global
Forest Science (GF-18-1999-65) for grants supporting this research.
We also thank Cameron Williams, Marie Antoine, and Nolan Bowman
for assistance with fieldwork. We appreciate the co-operation of the
California Department of Fish and Game (permit SCP#003419) and the
Redwood National and State Parks (permit REDW1999Sillett) in
support of this research. This project was conducted under Humboldt
State University IACUC permit number 00/01.B.31.
LITERATURE CITED
Ambrose, A. 2004. Water-holding capacity of canopy soil mats and
effects on microclimates in an old-growth redwood forest. M.A.
Thesis, Humboldt State University, Arcata, California, USA.
Brown, J.L., and G.H. Orians. 1970. Spacing patterns in mobile
animals. Annual Review of Ecology and Systematics 1:239-262.
Chao, A., and R.M. Huggins. 2005. Classical closed-population
capture-recapture models. Pp. 22-35 In Amstrup, S.C., T.L.
McDonald, and B.F.J. Manly (Eds.). Handbook of Capture-Recapture
Analysis. Princeton University Press, Princeton, New Jersey, USA.
Spickler et al.—Aneides vagrans in Redwood Canopies September 2006
24
Chapman, D.H. 1951. Some properties of the hypergeometric
distribution with applications to zoological censuses.
University of California Publications in Statistics 1:131-160.
Cooperrider, A., R.F. Noss, H.H. Welsh Jr., C. Carroll, W.
Zielinski, D. Olson, S.K. Nelson, and B.G. Marcot. 2000.
Terrestrial fauna of redwood forests. Pp. 119-164 In Noss,
R.F. (Ed.). The Redwood Forest: History, Ecology, and
Conservation of the Coast Redwoods. Island Press,
Washington, D.C., USA.
Cree, A. 1989. Relationship between environmental conditions
and nocturnal activity of the terrestrial frog, Leiopelma
archeyi. Journal of Herpetology 23:61-68.
Cunningham, J.D. 1969. Aspects of the ecology of the Pacific
Slender Salamander, Batrachoseps pacificus, in southern
California. Ecology 41:88-89.
Davis, T.M. 1991. Natural history and behaviour of the Clouded
Salamander, Aneides ferreus. M.Sc. Thesis, University of
Victoria, Victoria, British Columbia, Canada.
_____. 2002a. An ethogram of intraspecific agonistic and
display behavior for the Wandering Salamander, Aneides
vagrans. Herpetologica 58:371-382.
_____. 2002b. Microhabitat use and movements of the
Wandering Salamander, Aneides vagrans, on Vancouver
Island, British Columbia, Canada. Journal of Herpetology
36:699-703.
Enloe, H.A., R.C. Graham, and S.C. Sillett. 2006. Arboreal
histosols in old-growth redwood forest canopies, northern
California. Soil Science Society of America Journal 70:408-
418.
Good, D.A., and D.B. Wake. 1993. Systematic studies of the
Costa Rican moss salamander, genus Nototriton, with
descriptions of three new species. Herpetological Monographs
6: 131-159.
Gordon, R.E. 1952. A contribution to the life history and
ecology of the plethodontid salamander Aneides aeneus.
American Midland Naturalist 47:666-701.
Houck, L.D., and S.J. Arnold. 2003. Courtship and mating. Pp.
383-424 In Sever, D.M. (Ed.). Phylogeny and Reproductive
Biology of Urodela. Science Publishers, Enfield, New
Hampshire, USA.
Jackman, T.R. 1998. Molecular and historical evidence for the
introduction of Clouded Salamanders (genus Aneides) to
Vancouver Island, British Columbia, Canada, from California.
Canadian Journal of Zoology 76:1-11.
Jaeger, R.G. 1980. Fluctuation in prey availability and food
limitation for terrestrial salamanders. Oecologica 44:335-341.
_____, D. Kalvarsky, and N. Shimizu. 1982. Territorial behavior
of the Redbacked Salamander: expulsion of intruders. Animal
Behavior 30:490-496.
Jepson, J. 2000. The Tree Climber’s Companion. Beaver Tree
Publishing, Longville, Minnesota, USA.
Jones, C.B. 2005. Arthropods inhabiting epiphyte mats in an
old-growth redwood forest canopy. M.A. Thesis, Humboldt
State University, Arcata, California, USA.
Lynch, J.F., and D.B. Wake. 1996. The distribution, ecology,
and evolutionary history of plethodontid salamanders in
tropical America. Natural History Museum of Los Angeles
County Science Bulletin 25:1-65.
Maiorana, V.C. 1977. Observations of salamanders (Amphibia,
Urodela, Plethodontidae) dying in the field. Journal of
Herpetology 11:1-5.
Mathis, A., R.G. Jaeger, W.H. Keen, P.K. Ducey, S.C. Wallace,
and B.W. Buchanan. 1995. Aggression and territoriality by
salamanders and a comparison of the territorial behavior of frogs. Pp.
633-676 In Heatwole, H., and B. K. Sullivan (Eds.). Amphibian
Biology. Volume 2. Social Behavior. Surry Beatty and Sons,
Chipping Norton, New South Wales, Australia.
McCranie, J.R., and L.D. Wilson. 1993. Life history notes: Nototriton
barbor reproduction. Herpetological Review 23:115-116.
Moffett, M.W., and M.D. Lowman. 1995. Canopy access techniques.
Pp. 3-26 In Lowman, M.D., and N.M. Nadkarni (Eds.). Forest
Canopies. Academic Press, San Diego, California, USA.
Noss, R.F. (Ed.). 2000. The Redwood Forest: History, Ecology, and
Conservation of the Coast Redwoods. Island Press, Washington,
D.C., USA.
Nussbaum, R., E.D. Brodie, Jr., and R.M. Storm. 1983. Amphibians
and Reptiles of the Pacific Northwest. University of Idaho Press,
Moscow, Idaho, USA.
Ovaska, K. 1988. Spacing and movement of the salamander Plethodon
vehiculum. Herpetologica 44:377-386.
Parker, G.G. 1995. Structure and microclimate of forest canopies. Pp.
73-106 In Lowman, M.D., and N.M. Nadkarni (Eds.). Forest
Canopies. Academic Press, San Diego, California, USA.
Petranka, J.W. 1998. Salamanders of the United States and Canada.
Smithsonian Institute Press, Washington, D.C., USA.
Pollock, K.H., J.D. Nichols, C. Brownie, and J.E. Hines. 1990.
Statistical inference for capture-recapture experiments. Wildlife
Monographs 107:1-97.
Sapp, J. 2002. Courtship behaviors in the salamander genus Aneides.
M.Sc. Thesis, Oregon State University, Corvallis, Oregon, USA.
Sawyer, J.O., S.C. Sillett, W.J. Libby, T.E. Dawson, J.H. Popenoe, D.L.
Largent, R. Van Pelt, S.D. Veirs Jr., R.F. Noss, D.A. Thornburgh,
and P. Del Tredici. 2000. Redwood trees, communities, and
ecosystems: a closer look. Pp. 81-118 In Noss, R.F. (Ed.). The
Redwood Forest: History, Ecology, and Conservation of the Coast
Redwoods. Island Press, Washington, D.C., USA.
Shoemaker, V.H., S.S. Hillman, S.D. Hillyard, D.C. Jackson, L.L.
McClanahan, P.C. Withers, and M.L. Wygoda. 1992. Exchange of
water, ions, and respiratory gases in terrestrial amphibians. Pp. 125-
150 In Feder, M.E., and W.W. Burggren (Eds.). Environmental
Physiology of the Amphibians. University of Chicago Press,
Chicago, Illinois, USA.
Sillett, S.C. 1995. Branch epiphytes assemblages in the forest interior
and on the clearcut edge of a 700-year-old Douglas-fir canopy in
western Oregon. Bryologist 98:301-312.
_____. 1999. The crown structure and vascular epiphyte distribution in
Sequoia sempervirens rain forest canopies. Selbyana 20:76-97.
_____, and M.G. Bailey. 2003. Effects of tree crown structure on
biomass of the epiphytic fern Polypodium scouleri (Polypodiaceae) in
redwood forests. American Journal of Botany 90:255-261.
_____, and R. Van Pelt. 2001. A redwood whose crown may be the
most complex on Earth. Pp. 11-18 In Labrecque, M. (Ed.). L’Arbre
2000. Isabelle Quentin, Montréal, Québec, Canada.
Spotila, J.R., and E.N. Berman. 1976. Determination of skin resistance
and the role of the skin in controlling water loss in amphibians and
reptiles. Comparative Biochemistry and Physiology 55:407-411.
Staub, N.L., and D.B. Wake. 2005. Aneides lugubris Hallowell, 1949.
Pp. 662-664 In Lannoo, M. (Ed.). Amphibian Declines: the
Conservation Status of United States Species. University of
California Press, Berkeley, California, USA.
Stebbins, R.C. 2003. A Field Guide to Western Reptiles and
Amphibians. 3rd Edition. Houghton Mifflin Company, Boston,
Massachusetts, USA.
Thorson, T.B. 1955. The relationship of water economy to terrestrialism
in amphibians. Ecology 36:100-116.
Herpetological Conservation and Biology 1(1):16-26
25
Van Pelt, R. 2001. Forest Giants of the Pacific Coast. University
of Washington and Global Forest Press, Seattle, Washington,
USA.
Waldron, J.L., and W.J. Humphries. 2005. Arboreal habitat use
by the Green Salamander, Aneides aeneus, in South Carolina.
Journal of Herpetology 39:486-492.
Welsh, H.H., and R.A. Wilson. 1995. Aneides ferreus
(Clouded Salamander) reproduction. Herpetological
Review 26:196-197.
Williams, C.B. 2006. Epiphyte communities on Redwood (Sequoia
sempervirens) in northwestern California. M.A. Thesis, Humboldt
State University, Arcata, California, USA.
A Wandering Salamander (Aneides vagrans) from Humboldt County, California (United States). (Photograph by: William Leonard, ©2004. All
Rights Reserved.)
Spickler et al.—Aneides vagrans in Redwood Canopies September 2006
2
JAMES C. SPICKLER is the senior biologist for Eco-Ascension Research
and Consulting (eco-ascension.com). He received his B.S. in Wildlife,
and M.A. in Biology from Humboldt State University, California. His
research interests include herpetology and temperate/tropical forest
canopy ecology. This publication represents research conducted at
Humboldt State University for his M.A. degree in Biology.
STEPHEN C. SILLETT is a Professor of Botany and the Kenneth L. Fisher Chair in
Redwood Forest Ecology at Humboldt State University. He received his B.A. in
Biology at Reed College, M.Sc. in Botany at University of Florida, and Ph.D. in
Botany and Plant Pathology at Oregon State University. Steve's research focuses
on the linkages between structure and biodiversity in forest canopies as well as
the biophysical and ecological constraints on maximum tree height.
SHARYN B. MARKS is a Professor of Zoology at Humboldt State
University. She received her B.A. in Biological Sciences at University
of Chicago, and her Ph.D. in Integrative Biology at the University of
California at Berkeley. Sharyn’s research focuses on integrating studies
of amphibian ecology with the management of natural resources to
promote amphibian population viability.
HARTWELL H. WELSH, JR. is a research wildlife ecologist with the Pacific Southwest
Research Station of the U. S. Forest Service. He is stationed at the Redwood Sciences
Laboratory in Arcata, California. He received a B.S. in Zoology from the University of
California at Berkeley, M.S. in Wildlife Biology from Humboldt State University, and
Ph.D. In Wildlife Ecology from University of California at Berkeley. His current
research interests include: (1) the relationships of forest structure and riparian attributes
to the distribution and abundance of forest herpetofauna; (2) the use of amphibians as
indicators of ecosystem health and integrity; and (3) the mechanisms of amphibian
declines.
26
