Distribution, abundance, microhabitat use and interspecific relationships among terrestrial salamanders on Vancouver Island, British Columbia

(1)

Distribution, Abundance, Microhabitat Use and Interspecific

Relationships Among Terrestrial Salamanders on Vancouver,

Island, British Columbia

t y

Theodore M. Davis

B.S., Portland State University, 1968 M.Sc., University of Victoria, 1991

A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY in the Department of Biology

We accept this dissertation as conforming

m the required standard

Dr. P. T. Gregory, sjiipervi^r (Department of Biology)

Dr. G. A. Allen, Departmental Member (Department of Biology)

Dr. N_LiMneston,, Departmental Member (Department of Biology)

Dr. E. A. Roth, Outside Member ^Department ofAnthropology)

Dr. N. L. Staub, External Examiner (Biology Department, Gonzaga University)

(2)

Supervisor: Dr. Patrick T. Gregory

Abstract

A fundamental aim of ecology is the study of patterns of distribution and abundance of organisms. These patterns can be influenced by intrinsic responses to environmental conditions, interspecific interactions, or both. If individuals of similar co-existing species use the same limited resources, competition can result in resource partitioning, but this pattern can also be the result of intrinsic differences.

On Vancouver Island, British Columbia, two ecologically similar plethodontid salam anders, P lethodon ve h icu lu m and A n e id e s fe r r e u s , are each common only where the other species is uncommon. I described their distribution and abundance, investigated differential microhabitat use, and evaluated interspecific interactions between them.

At each of nine sites I established arrays of six 0.3 x 2 m artificial cover objects (ACOs). Each ACO consisted of three boards arranged to create multiple microhabitats. ACOs allow sampling without disturbance of natural cover, provide a standard sampling unit, and minimize observer bias. In 1992 and 1993,1 checked 228 ACOs every other week, but less often in 1994. I also searched natural microhabitats and investigated distribution and abundance w ith time-constrained searches at 16 additional sites.

I collected data on 2790 salamanders. At the northern sites, A .

(3)

situation was reversed in the south. I found no differences in site

characteristics that would explain this pattern. Salamander abundance was reduced in clearcuts, but there was no difference among old-growth,

mature, and immature sites.

The density of P. v e h ic u lu m in Coldstream Provincial Park was exceptional. In one area, surface density was 1.8 individuals/m 2, but 200 m away there was <0.03 individuals/m 2. From censuses in fenced plots elsewhere, I estimated that <24% of the P. veh icu lu m population was on the surface at any particular time. Thus, by extrapolation, there are at least 75,000 P. v e h ic u lu m /h a in one area of Goldstream Park.

The density of P. v e h ic u lu m was <0.1 individuals/m 2 across wide areas of forest habitat with occasional patches of higher density. Surface abundance wras correlated with the area of ground covered by coarse woody debris (CWD) and with surface moisture, and abundance varied by a factor of 12 over a distance of 50 m.

I collected microhabitat data o.i 1306 salamanders. Of the A . fe rreu s, 95% were under the bark on logs or within logs. In. contrast, 67% of the P.

v e h ic u lu m were under CWD on the soil and 20% were within logs. A n e id es fe rr e u s used logs in an early stage of decay, and P. veh icu lu m ,

when under bark on logs oi within logs, used logs in a late stage of decay. Under ACOs, 98% of the A . fe rre u s were found between boards, whereas 85% of the P. ve h ic u lu m were found on the soil under boards.

(4)

similar to that of P. veh icu lu m , except that 23% of the T. g ra n u lo sa were on the surface.

In staged encounters, there was no aggression or predation between

A . fe rre u s and P. veh icu lu m . In laboratory and outdoor enclosures,

microhabitat selection was not influenced by the presence of the other species. Thus, differential microhabitat use is due to intrinsic differences, and is not the result of interspecific interactions.

The distribution and abundance of these species is not explained by interspecific interactions or site characteristics as measured in this study. Examination of habitat features at a finer scale might explain differences in distribution and abundance, but the requirements of these species could be so similar or correlated that differences might not be found. Additional sites need to be investigated to determine the detailed pattern of

(5)

Dr. P. T. Gregory, Su^brvisor (Department of Biology)

Dr. G. A. Allen, Departmental Member (Department of Biology)

Dr. N. Lfy/ngston, ^Dep^tmental Member (Department of Biology)

Dr. E. A. Roth, Outside Member (Department of Anthropology)

Dr. N. L. Staub| External Examiner (Biology Department, Gonzaga University)

(6)

List of Tables

Table 1. Salamanders of Vancouver Island British Columbia...5

Table 2. Comparisons of the mean number of salamanders

found per search by site from April 1 to October 1,1992

and 1993... 58

Table 3. Comparisons of the mean number of salamanders

found per search from April 8 to May 15,1993 and 1994...60

Table4. The number of salamanders found under ACOs on 8 searches of primary sites between April 8,1993 and August

12,1993... 61

Table 5. The num ber of salamanders found under ACOs on

matched searches of primary sites...64

Ta b l e 6. Probability of finding at least one individual of a species

on searches of ACO and searches of natural cover... 71

Table 7. Number of individual Taricha g ra n u lo sa found in

fenced and unfenced plots at Lake Cowichan, 1992-1994... 93

Table8. Density of P leth o d o n v e h ic u lu m in Goldstream

Provincial Park, Vancouver Island, B.C. on May 1, 1994... 112

Ta b l e 9a. Location o f sites searched on the first day of 2-hr time

constrained searches (TCSs), in the spring of 1993...119

Table 9b. Location of sites searched on the second day of 2-hr

time constrained searches (TCSs), in the spring of 1993...120

Table 10. Number of salamanders found on time-constrained

searches of 20 -12 sites on southeastern Vancouver Island,

B C ...124 Table 11. Habitat characteristics at forested primary sites... 140-141

Ta b l e 12. Terrestrial microhabitat use by four species of

(11)

List of Figures

Figure 1. Location of study sites O:: iricouver Island, British

Columbia, Canada... 28 Figure 2. Diagram of an artificial cover object (ACO). ... 35 Figure 3. Location of sample plots at Lake Cowichan... 38 Figure 4. Seasonal variation in number of salamanders, including

recaptures, found under artificial cover objects (ACOs) at

Goldstream Provincial Park, 1992-1994... ,...50 Figure 5. Seasonal variation in number of salamanders, including

Lake Cowichan, 1992-1994... ...52 Figure 6. Seasonal variation in number of salamanders, including

the GVW sites, 1992-1994... 54 Figure 7. Seasonal variation in number of salamanders, including

the RMC sites, 1993... 56 Figure 8. Salamander abundance among nine sites on Vancouver

Island, B. C... 62 Figure 9. Proportion of salamander species based on searches of

ACO, 1992-1994...66 Figure 10. Proportion cf salamander species based on searches of

natural cover in ACO plots, 1992-1994...68 Figure 11. Size-frequency histograms by season for P lethodon

v e h ic u lu m found under ACOs at Goldstream Provincial

Park, 1992-94... 73 Figure 12. Size-frequency histograms by season for P lethodon

(12)

Figure 13. Size-frequency histograms for P leth o d o n v e h ic u lu m found under ACOs and natural cover at Lake Cowichan

and the Greater Victoria Watershed (GVW), 1993-1994... 77 Figure 14. Size-frequency histograms for P lethodon v e h ic u lu m

found at Lake Cowichan in fenced ACO plots, 1992-1S94...80 Figure 15. Size-frequency histograms by season for A n e id es fe rreu s

(n=12.3) found under ACOs at the RM.C sites, 1993...82 Figure 16. Size-frequency histograms for Taricha g ra n u lo sa found

under ACOs at Lake Cowichan and the GVW sites... 84 ilgure 17. Mass-length relationships of P lethodon v e h ic u lu m and

A n e id e s fe r r e u s ... ...86

Figure 18. Cumulative number of salamanders caught in unfenced

and fenced plots at Lake Cowichan, 1993-94... 90 Figure 19. Map of Goldstream Provincial Park showing location of

ACO plot and transects for quadrat sampling ...108 Figure 20. Number of salamanders found on time-constrained

searches (TCSs) of secondary sites on Vancouver Island,

B.C 125

Figure 21. Proportion of P lethodon v e h ic u lu m found on time- constrained searches (TCSs) of secondary sites on

Vancouver Island, B.C... 128 Figure 22. Scattergrams showing correlations between salamander

abundance and site characteristics for the forested primary

study sites... 144 Figure 23. Scattergrams showing correlations between the

abundance of P leth o d o n v e h ic u lu m and two plot

characteristics for Lake Cowichan and the Greater Victoria

Watershed (GVW) sites combined... 147 Figure 24. Scattergrams showing correlations between the

abundance of P lethodon v e h ic u lu m and area of CWD...151 Figure 25. Normal temperatures and total precipitation per month

(13)

A n e id e s fe rr e u s , E n sa tin a e sc h sc h o ltzii, P leth o d o n

v e h ic u lu m , and Taricha g r a n u lo s a ... 165

Figure 27. The proportion of logs used by decay class ioxP lethodon

v e h i c u lu m and A n e id e s fe r r e u s ... 168

Figure 28. The proportion of logs by decay class and the proportion of logs used by decay class at Lake Cowichan and Rosewall

Creek...170 Figure 29. Microhabitat use by SVL (mm) for P lethodon

v e h i c u l u m ... 172 Figure 30. ACO microhabita!: use by species... 175 Figure 31. Microhabitat selection in field enclosures by P leth o d o n

v e h ic u lu m and A n e id e s fe r r e u s , singly and with each

(14)

Acknowledgments

I thank Guillaume Audet, Sophie Boizard, Jeannine Caldbeck, Logan Caldbeck, Aziza Cooper, Lisa Crampton, Christian Engelstoft, Trent Garner, Jennifer Harris, Kristina Lauridsen, Lynn Norman, Kristiina Ovaska, and Laura Smith for their assistance in the field. Aziza Cooper helped record much of the field data and I thank her for her dependability and careful attention to detail and accuracy. She also entered data into the computer, and helped with the tedious task of observing staged encounters.

I thank the authorities of the British Columbia Ministry of Environment, Lands and Parks who allowed me access to the site in Goldstream Park, the staff of the Greater Victoria Water District for access to sites under their control, and Douglas Pollard, Valin Marshall, and Tony Trofymow of Forestry Canada for their support and use of equipment.

Patrick Gregory provided helpful advice and encouragement throughout this study, and carefully reviewed an earlier version of this manuscript. I thank my committee members for their valuable comments.

This study was carried out while I held a Natural Sciences and Engineering Research Council (NSERC) Postgraduate Scholarship in 1991- 1992 and 1992-1993. Additional support was provided by the King-Platt Fellowship, 1993-1994, Additional funding was provided by grants to P. T. Gregory from the Forestry Practices Component of Forestry Canada’s Green Plan, and through the British Columbia Ministry of Environment, Lands and Parks, Wildlife Branch.

(15)

Chapter 1: Introduction

A primary goal of ecologists is to identify and explain patterns of distribution, abundance, and diversity of organisms (e.g. Brown 1984; Begon et al. 1990; Ricklefs and Schluter 1993; Krebs 1994). Of particular interest are patterns of distribution and abundance of ecologically similar sympatric species, and the ecological and behavioral processes that produce thjse patterns. Similar species may be found in the same habitats because they use similar resources and their behavioral and physiological responses to environmental conditions are similar. Alternatively, some species may be rare and others common at one location, but the relative abundances may be reversed elsewhere. Such patterns can be the result of intrinsic differences and similarities among species and their responses to particular environmental and resource conditions, but interspecific interactions among similar species, either competition or predation, also can have profound effects on when and where organisms are found, their

abundances, and which species coexist. Individuals of similar species may use the same resources, and if these resources are limited in supply,

competition can result in resource-partitioning or competitive exclusion. Thus, ecologists have long been interested in how similar, coexisting species differ in their use of resources (Darwin 1859; Allee et al. 1949;

(16)

Hutchinson 1959; Mayr 1963; Jaeger 1974; Schoener 1974,1989; Connell 1980; Hairston 1980b; Toft 1985; den Boer 1986).

Although competition for limited resources can result in resource partitioning, other mechanisms can produce the same pattern. Observed differential resource use may be due to either (1) independently derived differences among species, expressed as physiological, morphological or behavioral adaptations, or constraints, (2) interactions among species that happened in the past ("Ghost of Competition Past", Connell 1980), (3) current interactions, or (4) some combination of all these elements. If interspecific interactions are occurring, they may be direct or indirect (Abrams, 1987), but the strongest interactions in amphibians that use similar resources are likely to be direct interference competition or

predation (Toft 1985). Experimental manipulations of natural populations can provide evidence for current competition (Schoener 1983; Connell 1983), but there are no obvious methods that distinguish between independently evolved differences and differences that arose because of interactions in the evolutionary past (Begon et al. 1990).

Ecological similarity among species is often, but not always, correlated with their degree of phylogenetic relatedness. Thus, most

experimental studies of competition have focused on closely related species (Connell 1983, Schoener 1983). However, there are many cases of distantly related organisms competing for resources. Classic examples include competition among barnacles, mussels, and algae for space (Lubchenco and

(17)

Menge 1978), among fish and ducks for invertebrate prey (Eriksson 1979), and among rodents and ants for seeds (Brown et al. 1979). Nevertheless, as emphasized by Darwin (1859: chapter 3), competition is generally most intense between closely related species because similar morphology implies similar resource requirements: "As species of the same genus have usually, though by no means invariably, some similarity in habits and constitution, and always in structure, the struggle will generally be more severe between species of the same genus, when they come into competition w ith each other, than between species of distinct genera." Homoplasy is characteristic of terrestrial plethodontid salamanders (Wake 1991) and where they are syntopic they should use similar resources and should compete for those resources w hen those resources are limited in supply.

My overall goal is to determine whether similar, related species of terrestrial plethodontid salamanders on southeastern Vancouver Island differ in their distribution and abundance. If they differ, how do they differ and why? What are the ecological consequences of these differences?

These salamanders are similar in body size and shape, have similar life histories, require similar environmental conditions, and use similar resources. Where such species are in contact and compete for resources, they should either evolve differences in resource use or there should be some shift in resource use by one or both species in response to the

presence of the other. Similar species also may partition resources because of species-specific morphological, physiological, or behavioral constraints.

(18)

Thus, the pattern of differential resource use may be the result of a combination of species-specific traits and interspecific interactions (Toft 1985).

All six species of salamanders found on Vancouver Island are either entirely terrestrial or have an adult terrestrial stage (Table 1). The three plethodontid species (Family Plethodontidae), A n eid es fe rr e u s (Clouded Salam ander), P lethodon v e h ic u lu m (Western Red-backed Salamander), and E n sa tin a esch sch o ltzii (Ensatina Salamander), are entirely terrestrial: they lay their eggs on land, the hatchlings are terrestrial, and they do not require standing or running water during any part or their life cycle. The other three species, T a rich a g ra n u lo sa (Rough-skinned Newt), A m b y s to m a

g ra cile (Northw estern Salamander), and A m b y s to m a m a c ro d a c ty lu m

(Long-toed Salamander), have a terrestrial adult stage, but they m ust return to water to reproduce. These species lay their eggs in temporary or

permanent ponds or creeks. The eggs develop into aquatic larvae that eventually transform into terrestrial salamanders. However, A . gracile are facultatively neotenic (Duellman and Trueb 1986), and in some B. C.

populations most individuals reproduce while retaining the larval external morphology (Eagleson 1976).

A m b y sto m a is the m ost diverse and widespread genus of Family

Ambystomatidae (Nussbaum et a l 1983). This family is morphologically similar to, but distantly related to the Plethodontidae (Duellman and Trueb 1986; Larson and Dimmick 1993). Taricha belongs to the Salamandridae, a

(19)

(1985), and Leonard et al. (1993). SVL = snout-vent length; TL = total length.

Fam ily Species name Size Reproductive m ode/habitat Typical terrestrial habitat SALAMANDRIDAE Rough-skinned Newt,

Taricha granulosa

60-90 mm SVL 120-180 mm TL

Eggs and larvae aquatic/ lakes, ponds, swamps, and slow moving streams.

Under or within logs or leaf litter; noctumally and diumally active on surface. AMBYSTOMAT1DAE Northwestern Salamander, Ambystoma gracile 75-100 mm SVL 150-200 mm TL

Eggs and larvae aquatic / lakes, ponds, and slow moving streams.

Terrestrial habitat poorly known. Underground burrows, occasionally under or within logs.; some

populations may be neotenic. Long-toed Salamander,

A m bystom a macrodactylum

40-60 mm SVL 80-120 mm TL

Eggs and larvae aquatic/ temporary pools, small lakes, ponds, and slow moving streams.

Terrestrial habitat poorly known. Underground burrows, under rocks, bark and logs on soil or within leaf litter.

PLETHODONTIDAE Clouded Salamander, Aneides ferreus

48-66 mm SVL 80-110 mm TL

Eggs terrestrial; direct developm ent/ cavities within logs.

Under bark on logs or within logs.

Ensatina Salamander, Ensatina eschscholtzii

40-60 mm SVL 80-120 mm TL

Eggs terrestrial; direct developm ent/ cavities below surface or under or within logs.

Talus, under rocks, bark and logs on soil or within leaf litter.

Western Red-backed Salamander, Plethodon vehiculum

40-55 mm SVL 80-110 mm TL

Eggs terrestrial; direct developm ent/ in cavities below surface.

Talus, under rocks, bark and logs on soil or within leaf litter and Sword Fem bases.

(20)

family closely related to the Ambystomatidae (Larson and Dimmick 1993). Four of these species, A . fe r r e u s , P. v e h ic u lu m , E . e sc h sc h o ltzii, and

A . m a c ro d a c ty lu m , have very similar adult morphologies. They have

smooth skins, are relatively slender, and there is substantial overlap in the range of adult body sizes (Table 1). The other two species, T. g ra n u lo sa and

A . gracile, are larger and much more robust. The skin of T. g ra n u lo sa may

be dry and granular, depending on the sex and time of year. All these species have wide distributions in similar moist forest habitats and commonly coexist.

Two :necies, A . fe rre u s and P. veh icu lu m , are found throughout Vancouver Island, at least at lower elevations, and can be locally very abundant. Ovaska and Gregory (1989) studied the population ecology of a dense population of P. ve h icu lu m in Goldstream Provincial Park (48°28' N, 123°32' W). I had found A . fe rreu s nearby, but densities appeared to be low, although focused searches were not possible because such searches require the removal of bark from logs, which is rather destructive and also is prohibited in the park. In contrast, I found A . fe rre u s to be abundant farther north (49°27' N, 124°46' W), but P. veh icu lu m appeared to be relatively uncommon (Davis 1991). Casual observation suggested that habitat characteristics and weather conditions were similar among these and other sites. Also, microhabitat differences appeared to exist between these species, and although this question had not been formally addressed at this stage, Corn and Bury (1991) soon reported differences in the use of

(21)

CWD between these species in Oregon. These patterns suggested that interspecific interactions might be occurring between these species, or that the sites differed in conditions or resources important to these species, or both. Of particular interest was whether the presence of one species affects microhabitat selection in the other as it does in some plethodontid

salamanders in eastern N orth America (see Interspecific in tera ctio n s

a m o n g salam anders., below).

Thus, A . fe r r e u s and P. v e h ic u lu m are the species of most interest in this study. The closely related E. eschscholtzii will be considered where numbers captured allow. Taricha g ra n u lo sa , very different in morphology and life history from the plethodontids, will provide an interesting

contrast in certain circumstances. Both species of A m b y s to m a were infrequently captured, so presence/absence data only will be reported.

Ob je c t iv e s

There are four primary objectives of this study:

1) To determine the general pattern of distribution and abundance

of A n e id e s fe r r e u s and P le th o d o n v e h ic u lu m , and to explore how and why abundance varies among sites. These two species of terrestrial salamanders coexist on southeastern Vancouver H an d , but on preliminary searches, A ,

(22)

v e h ic u lu m dominated in the south. This suggested that their distribution

and abundance was influenced by climatic conditions, interspecific

interactions, or both. Also, abundance appeared :o vary among sites locally at the scale of a km or less, and, w ithin a site, at much smaller scales (<100 m), so I investigated salamander abundance within sites in relation to local structural and biological conditions.

This objective was partly motivated by the issue of declining

amphibian populations (Pechmann and Wilbur 1994; Blaustein and Wake 1995). Although there are no reports of declines of terrestrial salamanders in undisturbed environments, systematic monitoring has been rare, so if declines have occurred, they have not been recorded. There is concern that the fragmentation or loss of temperate old-growth forests might reduce amphibian diversity and abundance in western North America

(Herrington and Larsen 1985; Aubry et a l 1988; Bury and Corn 1988; Gibbons 1988; Raphael 1988; Welsh and Lind 1988; Hansen et al. 1991), of which terrestrial plethodontid salamanders make up a considerable fraction. Because the forests of southeastern Vancouver Island are highly fragmented by logging, I sampled some sites that represent an approximate Ci 'noseries from clearcut to old-growth forest to see w hat effect

disturbance might have on salamander populations. However, the majority of the sites were relatively undisturbed.

2) To document the nature and extent of differential microhabitat use between these species. Time, habitat and food type are considered the

(23)

three most important primary resource dimensions, with habitat and diet being the most important for vertebrates (Pianka 1969; Huey and Pianka 1983; Toft 1985; Schoener, 1974,1989). Toft (1985), in an extensive review of the literature on resource-partitioning in reptiles and amphibians, found that except for amphibian larvae and snakes, habitat is the resource

dimension partitioned first by amphibians and reptiles. Salamanders and lizards tend to be opportunistic feeders, so differences in prey type are generally attributable to habitat (Toft 1985).

Because terrestrial salamanders are euryphagic and there is

considerable overlap in the diet of sympatric species, differential use of prey in similar habitats is primarily related to body and jaw size (e.g. Maiorana 1978; Jaeger 1972; Harestad and Stelmock 1983; Lynch 1985). Thus, if there is differential use of prey, it could be a function of body and jaw size, an artifact of microhabitat partitioning, or both. Also, if food is a limiting resource, competition should occur for high-quality feeding microhabitats as a proximate limiting resource. Alternatively, microhabitat space per se, used as daytime retreats or nesting sites, might be the ultimate resource (Jaeger 1980b; Hairston 1987; Jaeger and Walls 1989; Gabor and Jaeger 1995). Thus, I chose microhabitat use as the most obvious niche dimension to investigate in these species. A fundamental assumption here is that resources are limiting for these species. The basis for this assumption is covered in the next section.

(24)

3) To develop and apply a number of sampling and censusing methods appropriate for terrestrial salamanders. A principal obstacle to accomplishing the above objectives was the lack of unbiased sampling methods. Therefore, a major aim of my work was to develop such methods, especially the use of artificial cover objects. This objective also was partly motivated by the issue of declining amphibian populations. Lack of standardized methodology makes comparison among sites and studies difficult or impossible, so there has been considerable effort in developing standardized methods (Heyer et al. 1994). My work is a contribution to that effort.

4) To evaluate the importance of interspecific interactions between

A . fe rr e u s and P. v eh icu lu m . This was done using a variety of

experimental approaches. A standard method of assessing intra- and interspecific interactions in terrestrial salamanders is to stage encounters between individual salamanders (e.g. Cupp 1980; Jaeger 1984; Keen and Sharp 1984; Nishikawa 1985; Ovaska 1987b, 1993; Davis 1991; Staub 1993), and I did this for these species. Salamander-salamander predation was investigated by offering hatchling P. v e h ic u lu m to adult A . fe rre u s. Finally, I experimentally investigated microhabitat selection in these species in the laboratory and under semi-natural conditions.

(25)

In t e r s p e c if ici n t e r a c t io n s a m o n gs a l a m a n d e r s

Similar species of terrestrial salamanders are able to coexist if they feed at different times, feed on different size prey, or use different

microhabitats (Burton 1976; Fraser 1976a). For example, in Oregon, three syntopic species of terrestrial salamanders (A n eid es fe rreu s, E nsatina

e sc h sc h o ltzii, and B atrachoseps w r ig h ti) differ in their use of different

decay classes of coarse woody debris (CWD; Bury and Corn 1988). Similarly, in the Oregon coast range mountains, Corn and Bury (1991) found that

P leth o d o n v e h ic u lu m , E. e sch sc h o ltzii, and A . fe rr e u s differ in their use of

CWD by decay class, or by their specific locations within or under CWD, or both. Such microhabitat partitioning may result from each species being better able to exploit critical resources in slightly different microhabitats, or because one species, through interspecific competition, forces another into a different microhabitat (Schoener 1974,1982,1983; Hairston 1987). These factors may operate concurrently, but where differences in resource use among species are intrinsic, competition for resources or interspecific predation will be rare or absent.

Some species of terrestrial salamanders use the same or similar microhabitats, but differ in the use of other resources. Bury and Martin

(1973) found that A n e id e s lu g u b r is , A . fe rre u s, E, eschscholtzii, and

B atrachoseps a tte n u a tu s consume different sizes, types, and numbers of

(26)

primarily in open, non-forested areas (74.1% of captures), occasionally in second-growth forests (19.3% of captures), and never in old-growth forests. The other species were relatively rare in the open areas, but common in both second- and old-growth forests. Microhabitats used by all species were similar, and individuals of different species were sometimes found under the same cover object. Differential use of prey may be related to

morphological and behavioral differences among these species, but it is not known if competition for cover objects occurs among them.

Several species of terrestrial salamanders in eastern North America are known to partition microhabitats through interspecific interference competition. For example, P lethodon cinereus excludes P. Shenandoah from areas of deep, moist soil through interspecific aggression (Jaeger 1970; Jaeger 1971a,b; Jaeger 1972; Jaeger and Gergits 1979; Wrobel et al. 1980). As a result, P. Shenandoah is confined to islands of xeric talus that are

physiologically unsuitable for P. cinereus (Jaeger 1971b).

Similarly, the semiaquatic D esm o g n a th u s fu s c u s moves to

microhabitats significantly farther from a stream when in the presence of

D esm o g n a th u s m o n tico la (Keen, 1982), but juvenile D. m o n tic o la shift

from rocks to wood in the presence of adult D. fu s c u s (Southerland 1986). Keen and Sharp (1984) showed that resident D. m onticola were more aggressive toward D. fu s c u s intruders than toward conspecific intruders.

Among several species of D esm o g n a th u s, territorial aggression can grade into predation (Keen and Sharp 1984; Southerland 1986).

(27)

Interspecific predation is known to affect microhabitat use and abundance among several species in this genus (Southerland 1986; Hairston 1986, 1987). For example, D. quadram aculatus and the ecologically similar

G y r in o p h ilu s p o r p h y ritic u s may affect distribution and microhabitat use by D e sm o g n a th u s ochrophaeus (Formanowicz and Brodie 1993).

If the body sizes of individuals are sufficiently different, interspecific predation also might occur among terrestrial salamanders. For example, aggressive behavior by A m b y sto m a m a c u la tu m toward P. cinereus can lead to predation or interference competition, and may affect the distribution of

P. cinereus among microhabitats on the forest floor (Ducey et al. 1994). In

general, adult terrestrial salamanders might prey on hatchling or juvenile salamanders of their own or other species, but very few species have been examined with this in mind. Powders (1973) described an apparent case of cannibalism in P lethodon teyahalee (=P. g lu tin o su s), but such predation must be rare because there are no reports of vertebrate parts in the stomach contents of terrestrial salamanders (e.g. Dumas 1956; Storm and Aller 1947; Altig and Brodie 1971; Jaeger 1972; Bury and Martin 1973; Lee and Norden 1973; Burton 1976; Stelmock and Harestad 1979; Lynch 1985; Whitaker et al. 1986).

Although direct interspecific interactions generally favor one species over another, in some cases a more balanced relationship is possible. For example, in the Black Mountains of North Carolina and the Great Smoky Mountains of N orth Carolina and Tennessee, Plethodon jo rd a n i and P.

(28)

teyahalee compete for space strongly enough that their distributions

overlap only along a narrow elevational band (Hairston 1951,1980a, 1987). Presumably, the altitude at which each species occurs depends on intrinsic physiological and behavioral responses to physical conditions that vary with altitude. P lethodon jo rd a n i is favored in the cooler, moister

microhabitats that occur at higher elevations as well as in deep ravines and north-facing slopes. P lethodon teyahalee is favored in the warmer, dryer microhabitats that occur at lower elevations (Hairston 1949; Hairston 1980a). However, in the Balsam Mountains of North Carolina, which are geographically between the other two ranges, there is a broad band of overlap between the two species (Hairston 1951; Hairston 1980a). In field experiments, Hairston (1980a,b) was able to show that interspecific

competition is m uch stronger where the overlap is narrow than where the overlap is broad. Also, individuals from the Great Smoky Mountains, where competition is strong, are more aggressive than individuals from ihe Balsam Mountains, where competition is weak (Nishikawa 1985). Other experiments showed that neither prey nor foraging microhabitats are the object of competition, but that space, perhaps for nesting sites, is the most likely limiting resource (Nishikawa 1985; Hairston et al. 1986; Hairston 1987).

Such a relatively symmetrical relationship can also be the result of a shift in competitive abilities at different life stages. Wilbur (1980) suggested that two species may coexist if one species is competitively superior in the

(29)

larval stage while the other has an advantage in the adult stage. Such an ontogenetic switch in competitive abilities occurs in A m b y s to m a

ta lp o id e u m and A . m a cu la tu m . These species have an aquatic larval stage,

but become terrestrial at metamorphosis. In the larval stage, A .

ta lp o id e u m is aggressively superior to A . m a cu la tu m (Walls and Jaeger

1987), b u t the relationship is reversed at metamorphosis (Walls 1990). It is not know n if competition between these species in the terrestrial stage causes a shift in resource use.

Aggressive behavior has been observed between other terrestrial salamanders. Grant (1955) observed territorial defense in captive Eurycea

b islin ea ta and H e m id a c ty liu m s c u ta tu m . Thurow (1975) observed

aggressive interactions among several species of P lethodon. Ovaska (1993) reported that P. d u n n i shows aggressive behavior toward P. veh icu lu m , and Smith and Pough (1994) demonstrated that D. ochrophaeus displaces P.

cinereus from cover objects. However, the ecological effects of these

interactions in nature, if any, are unknown.

Although interspecific interactions among terrestrial salamanders can profoundly affect local abundance and diversity, few studies have been carried out for species in western North America. Maiorana (1978) found that when large prey were scarce, there was substantial overlap in prey size between A . lu g u b ris and B. a tte n u a tu s, but she concluded that prey were less important in limiting populations than the availability of burrows of different sizes. However, the importance of interspecific competition

(30)

between these species is unclear, and there are no experiments that

demonstrate competition for burrows or territoriality. Lynch (1985) found that four species of terrestrial salamanders, A . lugubris, A . fla v ip u n c ta tu s ,

B. a tten u a tu s, and E. eschscholtzii, despite considerable dietary overlap,

exhibited significant differences in the size of prey taken. Competition for food or space among these species is plausible, but again, no experimental work has been done. Dumas (1956) studied P lethodon d u n n i and P.

veh icu lu m where their ranges overlap in western Oregon. He found slight

differences in microhabitat use and in the variety of prey consumed by these species, but there was much overlap. Ovaska and Davis (1992) found that these species recognize and display toward each others' fecal pellets (which are used in chemical communication), and Ovaska (1993) observed attacks and agonistic display behavior by P. d u n n i toward P. veh icu lu m . However, much work needs to be done to clarify the extent and

significance of these interspecific interactions in nature.

Most of the work on intra- and interspecific interactions among terrestrial salamanders has focused on species from the central

Appalachian Mountains (Hairston 1987). However, life history patterns, including the nature and extent of interspecific interactions, may be different elsewhere. For example, because of climatic differences between the eastern and western United States, the timing of surface activity, courtship, and oviposition by salamanders in the two regions are very

(31)

different (Houck, 1977). Similarly, patterns of resource use and community structure may differ between the two regions.

Such differences may also exist between northern and southern regions. For example, warm, dry summers may limit surface activity for several months in California, but in the relatively wet Pacific Northwest and on Vancouver Island, salamanders may be able to remain on the surface for longer periods. Finally, because there are fewer species dividing up the available resources, the northern species may be less subject to interspecific competition and predation than southern species. It is notable that neither Ovaska (1987b) nor I (Davis 1991) found territorial behavior in

P. v e h ic u lu m or A . fe rr e u s , respectively, but similar species in the eastern

United States [e.g. P. cinereus (Jaeger et a l 1982) and A n e id e s aeneus (Cupp 1980)] are territorial. Also, in Washington State, P. d u n n i, which is

restricted to rocky splash zones near streams, is more aggressive than the wider ranging plethodontid salamanders with which it coexists (Ovaska 1993). This suggests that Canadian populations of P. cinereus might have very different selection pressures than populations of P. cinereus in the central Appalachians. The specific prediction is that individuals from northern populations would be less aggressive and less territorial than individuals from southern populations.

(32)

Na t u r a l His t o r y

Plethodontid salamanders lack lungs so the exchange of respiratory gases occurs across the highly vascularized skin and buccal cavity. As a requirement of transcutaneous gas exchange, the skin is very permeable to water, and the body size is small, making these salamanders especially susceptible to desiccation and restricting them to moist microhabitats (Shoemaker e t al. 1992). Nevertheless, many plethodontid salamanders, including the three species on Vancouver Island, are completely terrestrial during their entire life cycle. Females lay and brood small clutches of eggs in the early summer. The larval stage is passed in the egg, and young hatch in the autum n as miniature replicas of the adults (McKenzie 1970; Ovaska and Gregory 1989; Davis 1991). Amphibians are ectothermic and heat is obtained mainly from the external environment (Hutchison and Duprd 1992). Thus, their surface activity is constrained by moisture and

tem perature.

O n Vancouver Island, these salamanders can be found at the surface fairly easily throughout most of the year, b ut disappear during the coldest part of the winter when beinw-freezing temperatures force them

underground or deep inside logs. Dry conditions late in the summer and in early autum n can have a similar effect. Peak abundance at the surface is usually in the spring or early summer (Ovaska and Gregory 1989, Davis 1991; Davis and Gregory 1993), but surface abundance can increase

(33)

dramatically at any time of the year if conditions are favorable. Resources such as mates and prey are obtained primarily at the surface and little or no feeding is thought to take place underground (Jaeger 1972,1980c; Fraser 1976a,b; Maiorana 1976). Thus, their ability to grow and reproduce is directly related to the length of time that they spend near the surface (Houck 1977; Jaeger 1980b; Semlitsch and West 1983). They are opportunistic predators, feeding on small terrestrial invertebrates.

C l o u d e d S a l a m a n d e r ( A n e i d e s f e r r e u s) - The geographical

distribution of this species is strikingly disjunct: Vancouver Island, British Columbia, and south of the Columbia River in Western Oregon and northwestern California (Wake 1965). In British Columbia, A. ferreu s are found on Vancouver Island at altitudes less than about 600 m, and are well established on many of the smaller islands nearby (Davis and Gregory 1993). They can be found in moist terrestrial habitats such as under exfoliating bark and in cracks and cavities of decomposing logs, stumps, and snags, in talus, and occasionally in trees (Nussbaum et al. 1983;

Stebbins 1985; Davis and Gregory 1991; Leonard et al. 1993). They are very site-specific and most movements of individuals are less than 2 m between captures that may be many months apart (Davis 1991). Peak abundance at the surface occurs in June (Davis 1991). Courtship and mating take place in the spring, and females lay small clutches of eggs in cavities within

(34)

1993). Prey consists of small terrestrial arthropods, especially insects (Storm and Aller 1947; Bury and Martin 1973; Stelmock and Harestad 1979;

Whitaker et al. 1986).

W e s t e r n R e d - b a c k e d S a l a m a n d e r ( Pl e t h o d o n v e h i c u l u m) - This

species is found from southern Oregon to southern British Columbia, west of the Cascade and Coast mountains (Stebbins 1985; Leonard e t al. 1993). In British Columbia, it is found throughout Vancouver Island and on the mainland in the Fraser Valley as far east as Hope (Green and Campbell 1984). Curiously, this salamander has not been recorded from any of the Gulf Islands and is absent from m ost of the other islands surrounding Vancouver Island. This contrasts dramatically w ith A . fe rre u s (see above).

P le th o d o n v e h ic u lu m are found w ithin leaf litter and Sword Fern (P o ly stic h u m m u n itu m ) bases, under moss, rocks or CWD on the forest

floor, and under or among rocks on talus and rock outcrops (Bury et al. 1991; Corn and Bury 1991; Leonard et a l 1993). They favor damp, but not wet, shady areas of the forest and can be very abundant (Ovaska and

Gregory 1989). Cvaska (1988b) reported a high degree of site-specificity and small home ranges. Most movements of individuals between captures over a two-year period were less than 3 m. On Vancouver Island,

salamanders move from underground retreats during w arm wet weather, so peak surface abundance occurs in the spring and autumn, depending on recent weather conditions (Ovaska and Gregory 1989). Courtship and

(35)

mating occur mainly in October and November, but eggs are laid the following summer (Ovaska and Gregory 1989). Eggs and nests are not well documented, probably because eggs are laid beneath the surface and are difficult to locate (Leonard et al. 1993). Hatchlings appear in the autumn and take two to three years to reach sexual maturity (Ovaska and Gregory 1989). Prey consists of a variety of terrestrial invertebrates (Dumas 1956).

E n s a t i n a S a l a m a n d e r ( En s a t i n a e s c h s c h o l t z i i) - This species is found

from extreme northwestern Baja California to southern British Columbia, west of the Sierra Nevada mountains in California (but absent from the Great Valley of California), west of the Cascade crest in Oregon and Washington, and on eastern Vancouver Island, the adjacent mainland, and up the Fraser valley to Boston Bar in British Columbia (Green and Campbell 1984; Leonard et al. 1993). These salamanders can be found in damp microhabitats under rocks or CWD on the forest floor, at the entrance of rodent burrows, under or among rocks on talus, and

particularly within and under bark piles at the base of snags and stumps, but are almost never found in perpetually wet areas (Corn and Bury 1991; Leonard et al. 1993). They are difficult to find during the day, but may be abundant on the surface at night during warm wet weather, and are sometimes observed on paved roads when conditions are favorable. This species has not been studied in British Columbia, but its ecology and

(36)

genetics have been studied in California (Stebbins 1954; Wake and Yanev 1986; Jackman and Wake 1994).

If disturbed, individuals may show a defensive display and produce a milky poison on the dorsal side of the tail. Also, the tail is easily

autotomized, usually at the basal constriction, leaving a predator w ith a writhing noxious tail while the salamander escapes (Green and Campbell 1984; Leonard et al. 1993).

R o u g h - s k i n n e d N e w t (Ta r i c h a g r a n u l o s a) - This species is found in

the hum id coastal forests from southeast Alaska to Northern California, primarily west of the Cascade and Coast Range mountains (Leonard et al. 1993; Stebbins 1995). It is much larger than the plethodontid salamanders described above, uses lungs in gas exchange and has a relatively thick skin. Thus, it is probably less subject to desiccation than the plethodontid species. In the water, adult T. gra n u lo sa prey on a variety of invertebrates,

especially amphipods and insects, and on frog tadpoles (Chandler 1918; Efford and Tsumura 1973; Lefcort and Eiger 1993). In the terrestrial phase, these salamanders eat a wide variety of invertebrates including

earthworms, snails, spiders, mites, springtails, and a variety of adult and larval insects (Chandler 1918).

On Vancouver Island, T. granulosa lay a series of single eggs in ponds from April to July (Oliver and McCurdy 1974). Larvae develop over the summer and transform by the end of August. At higher altitudes,

(37)

some larvae m ay overwinter and transform the following summer (Chandler 1918). After metamorphosis, the young leave the water until they reach their 4th or 5th year when they become sexually mature and return to the pond to breed. On southern Vancouver Island, Oliver and McCurdy (1974) found that adult males normally remain permanently aquatic, but adult females migrate from breeding ponds to overwinter cn land. However, at Marion Lake, B.C., which is on the mainland, Efford and Mathias (1969) reported that males as well as females left the water by mid- October and returned early in the spring. This is the same pattern reported by Chandler (1918) and Pimentel (1960) near Corvallis, Oregon. Mass migrations of newts to breeding ponds are limited to females only

(Pimentel 1960). Details of the activities and natural history of T. granulosa during the terrestrial phase are unknown.

Breeding males develop a swollen vent, high tail crest, smooth skin, and cornified, melanized nuptial pads (Oliver 1974). Nonbreeding males have a granulated skin as do females at all times. Sexes can be

distinguished by anatomical details of the cloaca (Stebbins 1954:45). The skin of T. gra n u lo sa contains high concentrations of

tetrodotoxin (TTX), a neurotoxin, which functions as a defense against predators (Brodie 1968; Brodie and Brodie 1990). Thus, they are virtually immune to predation by fish, and unlike many other aquatic amphibians, are able to coexist with fresh water salmonids (Efford and Mathias 1969; Efford and Tsumura 1973; Taylor 1984). However, Brodie and Brodie (1991)

(38)

found that skin extracts of newts from Reed Island (an island adjacent to Vancouver Island) were at least three orders of magnitude less toxic than skin extracts of newts from the Willamette Valley, Oregon. They

concluded that Reed Island T. g ranulosa have little or no TTX in their skins, and assumed that this applies to all Vancouver Island T. granulosa. However, Macartney and Gregory (1981) found that garter snakes

( T h a m n o p h is o rd in o id es and T. e ltg a n s ) that were force fed T. g ra n u lo sa

(39)

Chapter 2: Methods

In this chapter, I provide an overview and justification of the use of artificial cover objects (ACOs), details of the main site locations, the

location of plots within sites, and a description of ACOs and their

arrangement within plots. I also describe fences that were used t o enclose some ACO plots as well as sampling frequency, searches of natural cover in ACO plots, the handling and measurement of salamanders and other general methodological details. Detailed methodology relevant to particular chapters only is described in those chapters.

M e t h o d o l o g ic a l Ov e r v ie w

A r tific ia l co ver objects (A C O s) - Estimation of temporal and spatial

variation in abundance is of fundamental importance in ecology and conservation biology. Estimates of relative abundance of terrestrial

amphibians in space and time are often made using unit-effort searches of natural cover objects and microhabitats (Corn and Bury 1990). Methods include time- or area-constrained searches, surveys of coarse woody debris (CWD), and quadrat, transect and patch sampling (Corn and Bury 1990;

(40)

Heyer et al. 1994). However, because of disturbance of the natural habitat, these methods may be unsuitable where repeated searches of the same area are needed, or where disturbance of the natural habitat is unacceptable or prohibited. For example, A n e id e s fe rr e u s is typically found under bark on logs or within logs (Da vis 1991), and one thorough search of this

microhabitat can be very destructive. Such destructive sampling is unacceptable where more than one sample is needed and may be

prohibited in parks and reserves. Also, searches of natural cover among sites that differ in the amount and type of CWD may not be comparable because some types of cover may be difficult to search efficiently, resulting in unequal search effort. Finally, search effort may vary among individual searchers (Heyer et al. 1994) or through time with the same individual (T. Davis, unpublished data; see Chapter 9).

To overcome some of these difficulties, unit-effort pitfall trapping may be used to estimate relative abundance. This has the advantage that the amount and type of CWD and individual effort are independent of the trapping effort. However, the capture rate varies widely among species (Buhlmann et al. 1988; Bury and Corn 1987; Corn and Bury 1990; Welsh 1990; Welsh and Lind 1988), and many species, probably because they are relatively site-tenacious, are not readily trapped (Welsh and Lind 1988). Also, the traps m ust be checked frequently to prevent accidental mortality of the various vertebrates that may be caught.

(41)

In principle, ACOs can overcome many of these difficulties (Grant et

al. 1992; Heyer et a l , 1994). They are especially useful when repeated

sampling is necessary, are relatively easy to sample, result in little or no damage to the natural habitat, and can attract species that are difficult to trap in pitfall traps. Unlike sampling with pitfall traps, sampling with ACOs can be opportunistic with no risk of mortality from failure to check the traps frequently (Grant et al. 1992). Because ACOs can be checked repeatedly over long periods, rare species can eventually be detected without damage to the natural habitat that could result from repeated searches of natural cover. In conjunction with mark-recapture methods, ACOs can be used to monitor individual movements and to estimate life- history parameters and abundance. Also, they can be used to investigate the relationship between habitat characteristics and abundance, and differential microhabitat use among species. Because ACOs can be of a standard size and number, and are independent of the amount and type of CWD and individual searching effort, they should give more reliable estimates of relative abundance across sites that differ in structure than do searches of natural cover.

G eneral m eth o d o lo g y - The study sites were limited to southeastern

Vancouver Island, from Goldstrean1 Provincial Park, north to Rosewall Provincial Park, and west to Lake Cowichan (Fig. 1), I selected nine main (primary) study sites to monitor with ACOs. Within these sites, I

(42)

Figure 1. Location of study sites on Vancouver Island, British Columbia, Canada. Primary study sites with ACOs are indicated by large black dots and the following notation: RMC = Rosewall, McNaughton and Cook Creeks; LC = Lake Cowichan; GVW = Greater Victoria Watershed (4 sites); GS = Goldstream Provincial Park. Black squares and open squares indicate secondary sites, with the open squares representing sites that were not searched on the second and third searches (Chapter 5). Other locations (Chapter 9) are indicated by black dots and place names.

(43)

Miracle Beach Provincial Park Denman Is. RMC Port Albemi Cleland Is. Tofino

i® Portland Is.

U ^ S * f v . Carmanah Valiey Provincial Park 1 26 1 2 5 _Victoria Jordan River K> SO

(44)

established two to six circular plots, each approximately 10 m in diameter. Within each plot, I haphazardly arranged six 0.3 x 2 m ACOs. Each ACO consisted of three boards arranged to create multiple microhabitats. I periodically searched for salamanders under the ACOs, and this is the primary source of my field data.

For two field seasons, I checked 228 ACOs at approximately two week intervals. In the third field season, ACOs were checked less often. After recording the snout-vent length (SVL), weight and sex, and marking the salamanders by toe-clipping, I released the salamanders at the same spot where they were captured. At one site (Lake Cowichan), some plots were fenced to restrict the movement of salamanders in order to take a census of them. Additional sampling was done by unit-effort searches of surface microhabitats. These included time-constrained searches (TCSs; time- limited equal-effort searches of all available microhabitats) and area- constrained searches (ACSs; area-limited equal-effort searches of all available microhabitats). Detailed within-site sampling was done with randomly placed l x l m quadrats along equally spaced parallel transects at one site (Goidstream), and by within-site comparison of the 10-m-diameter plots at other sites (Lake Cowichan, Greater Victoria Watershed sites). To assess the general distribution and abundance of salamanders on southern Vancouver Island, an additional 16 secondary sites plus four of the primary sites were searched with TCSs.

(45)

I collected data on the general habitat features of the primary sites within 12 x 12 m plots centered on the 10-m-diameter ACO plots. Within these plots, I recorded the amount and type of all coarse woody debris (CWD), and measured the diameter at breast height (DBH) of each tree and identified the species. I also estimated the percentage of bare ground, the percentage area covered by understory plants, and measured and counted Sword Ferns (P o ly s tic h u m m u n itu m ) , as these were thought to provide shelter for salamanders.

I recorded the microhabitat in which each salamander was found. The ACOs contained two microhabitats: 1) under wood on soil, and 2) under wood on wood. Natural microhabitats searched on TCSs and ACSs included, in part: under logs or other CWD on the ground, under bark on logs and within logs, within Sword Fern bases, in moss, and under rocks. I found some salamanders on the surface when conditions were favorable.

I noted weather conditions and temperature at the time of each search, and supplemented these data with observations obtained from Environment Canada weather stations.

I investigated microhabitat use in the laboratory and in outdoor enclosures, and interspecific interactions in staged encounters between individual salamanders in the laboratory.

(46)

De t a il e d Me t h o d o l o g y

S tu d y sites - There were nine primary study sites (Fig. 1). Four sites

were in the Greater Victoria Watershed (GVW; 48°34’ N, 123°39' W, 200- 250 m), three near Rosewall Creek Provincial Park (RMC; Rosewall Creek: 49°27' N, 124°46' W, <50 m; McNaughton Creek: 49°27’ N, 124°46' W, <50 m; Cook Creek: 49°27' N, 124°45' W, <50 m), and one each at Goidstream Provincial Park (48°28‘ N, 123°32’ W, <50 m), and the University of Victoria research property (Jeannie Simpson Resource Centre) on Marble Bay, Lake Cowichan (48°50’ N, 124°10' W, Lot 29,163 m). The GVW sites can be found w ithin the following polygon numbers from the Greater Victoria Water District Forest Cover Maps: GVW clear-cut: 753; GVW immature: 640; GVW mature: 699/V; GVW old-growth: 648 and 650. Polygon num bers from Forest Cover Map 92F.47 (Ministry of Forests, British Columbia) for the Rosewall, McNaughton, and Cook Creek sites (RMC) are 205 and 207,202, and 194, respectively.

The GVW clear-cut site was logged and burned in 1985 and forest seedlings were planted in 1986. The McNaughton Creek site was logged in 1977 and planted with forest seedlings in 1988. All the other sites were forested, but had been selectively logged (not clear-cut) or otherwise disturbed within the last 100 years. Nevertheless, all the forested sites contained at least a few very old trees (estimated at >200 years) and a similar suite of dominant tree species and understory plants. The

(47)

Hemlock (T su g a heterophylla), but Western Red Cedar (T h u ja plicata), Red Alder (A l n u s ru b ra ), and Broadleaf Maple (A ce r m a c ro p h y llu m ) were common as well. Little light penetrated the forest canopy, ground

vegetation was generally sparse, and some areas were virtually devoid of undergrowth. At all the forested sites, Sword Fern {P o ly stic h u m

m u n itu m ) was usually the dom inant ground species, but in some areas

Salal (G a u lth eria sh a llo n ) was dominant. Logs and woody debris, in various states of decomposition, were common. The GVW sites were selected by Forestry Canada to study the effect of forestry practices on carbon and nutrient dynamics and biodiversity in Douglas-fir and Western

Hemlock coastal forests, and approximately represent a chronoseries [GVW old-growth (> 150 years), GVW clear-cut (10-20 years), GVW immature (40- 60 years), and GVW mature (80-100 years)].

Sixteen secondary sites and four primary sites were used to assess the general distribution and abundance of salamanders on southeastern

Vancouver Island. Details of secondary site selection are in Chapter 5.

A rtific ia l cover objects (A C O s). - An individual artificial cover object

(ACO) consisted of a baseboard and two top boards. A large and massive cover object was needed to resist moisture loss and temperature changes; I judged that a 1.8 m long x 30.5 cm wide x 5 cm thick board to be about the largest baseboard that could be reasonably carried by one person into the

(48)

study sites. The space beneath the baseboard was cleared of vegetation, and the board was placed flat on the soil surface. Two 1.8 m long x 15.3 cm wide x 2.5 cm thick top boards were placed on top of the baseboard. Strips of cedar lath nailed to the baseboard separated the top boards from the

baseboard in such a way as to create a wedge-shaped space between the top boards and the baseboard (Fig. 2). Rain water dripped through the crack between the two top boards into this space. This created a complex microhabitat so that a salamander could be found on the soil under the baseboard, or between the baseboard and the top boards. Commercial lumber is often treated w ith fungicides that may harm amphibians, so untreated full dimensional lumber was used. Each baseboard was

individually labeled at one end w ith both a metal tag and a marking pen. The ACOs contained two basic microhabitats: on soil under the baseboard and on top of the baseboard under the top boards. Although the spaces under the top boards were wedge-shaped in cross section, because of the arrangement of the cedar lath spacers, the height of the space was different for each of the two boards (Fig. 2b). The edge of one top board rested directly on the baseboard and the other edge was raised above the baseboard by the cedar lath spacer in the center of the baseboard by

approximately 7 mm. The other board rested on the same spacers in the center of the baseboard (7 mm high), but the outer edge rested on a double layer of cedar lath spacers (14 mm high).

Distribution, abundance, microhabitat use and interspecific relationships among terrestrial salamanders on Vancouver Island, British Columbia