Tropical reptiles in pine forests: Assemblage responses to plantations andplantation management by burning

Beth Mott 1, Ross A. Alford, Lin Schwarzkopf *

School of Marine and Tropical Biology, James Cook University, Townsville, Queensland 4810, Australia

Forest Ecology and Management 259 (2010) 916–925

A R T I C L E I N F O

Article history:

Received 27 August 2009

Received in revised form 24 November 2009

Accepted 24 November 2009

Keywords:

Plantation forestry

Reptile assemblage structure

Tropical reptiles

Forest management

Biodiversity

Pine

Lizards

Snakes

Assemblage composition

Effects of plantations

Exotic timber

Operative environmental temperatures

Solar radiation

A B S T R A C T

Worldwide, the land area devoted to timber plantations is expanding rapidly, especially in the tropics,

where reptile diversity is high. The impacts of plantation forestry and its management on native species

are poorly known, but are important, because plantation management goals often include protecting

biodiversity. We examined the impact of pine (Pinus caribaea) plantations, and their management by fire,

on the abundance and richness of reptiles, a significant proportion of the native biodiversity in tropical

northern Australia, by (i) comparing abundance and diversity of reptiles among pine plantations (on land

cleared specifically for plantation establishment), and two adjacent native forest types, eucalypt and

Melaleuca woodlands, and (ii) comparing reptile abundance and richness in pine forest burnt one year prior

to the study to remove understorey vegetation with pine forest burnt two years prior to the study. We also

examined the influence of fire on reptile assemblages in native vegetation, by comparing eucalypt

woodland burnt two years prior to the study and unburnt for eight years. To quantify mechanisms driving

differences in reptile richness and abundance among forest types and management regimes, we measured

forest structure, the temperatures used by reptiles (operative temperature) and solar radiation, at replicate

sites in all forest types and management regimes. Compared to native forests, pine forests had taller trees,

lower shrub cover in the understorey, more and deeper exotic litter (other than pine), and were cooler and

shadier. Reptile assemblages in pine forests were as rich as those in native forests, but pine assemblages

were composed mainly of species that typically use closed-canopy rainforest and prefer cooler, shadier

habitats. Burning did not appear to influence the assemblage structure of reptiles in native forest, but

burning under pine was associated with increased skink abundance and species richness. Burned pine was

not warmer or sunnier than unburned pine, a common driver of reptile abundance, so the shift in lizard use

after burning may have been driven by structural differences in understorey vegetation, especially amounts

of non-native litter, which were reduced by burning. Thus, burning for management under pine increased

the abundance and richness of lizard assemblages using pine. Pine plantations do not support the snake

diversity common to sclerophyllous native forests, but pine may have the potential to complement

rainforest lizard diversity if appropriately managed.

Crown Copyright � 2009 Published by Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Forest Ecology and Management

journa l homepage: www.e lsevier .com/ locate / foreco

1. Introduction

Presently, both native forests and plantations are used to providethe world’s timber needs, but a growing proportion of the world’stimber is sourced from plantations (FAO, 2007; Hartley, 2002). Therehas been a shift in thinking about the role of plantations globally inthe last 15 years, with many forestry agencies identifyingbiodiversity conservation as a management priority, and movingtowards managing plantation forests to fulfill multiple roles,including timber production and biodiversity conservation (e.g.,

* Corresponding author. Tel.: +61 0417201256; fax: +61 47251570.

E-mail address: lin.schwarzkopf@jcu.edu.au (L. Schwarzkopf).1 Present address: School of Biological Sciences, University of Wollongong,

Northfields Ave., Wollongong, NSW 2522, Australia.

0378-1127/$ – see front matter . Crown Copyright � 2009 Published by Elsevier B.V. A

doi:10.1016/j.foreco.2009.11.031

Hartley, 2002; McDonald and Lane, 2002; Lindenmayer and Hobbs,2004; Scott et al., 2006; The Australian Forestry Standard, 2007). Anunderstanding of biodiversity in timber plantations, and factors thatcontrol it, is critical to this approach.

Pine (Pinus spp.) plantations comprise 15% of the world’splantation timber (FAO, 2007), and nearly 51% of the 1.9 millionhectare plantation estate in Australia (Gavran and Parsons, 2009).There is a significant (albeit declining) rate of native forest clearingcurrently occurring in Australia, with 260,000 hectares of forestcleared yearly between 2000 and 2004 (Montreal ProcessImplementation Group for Australia, 2008). Pine plantationscontinue to expand in Australia, with 6038 new hectares plantedin 2008, 12% of the total area of new plantations established(Gavran and Parsons, 2009). While in Australia new plantations areincreasingly established on previously cleared land, elsewhere in theworld land is cleared specifically for the purpose of plantation

ll rights reserved.

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925 917

development (FAO, 2007), as was usually the case in Australia untilrecently. The Australian forestry future plan is to treble the area ofplantations by 2020 (Mercer and Underwood, 2002). There is muchconservation concern about the value of reforestation withplantations for maintenance of biodiversity, especially in tropicalforests, where much of the world’s biodiversity resides (e.g.,Kanowski et al., 2005; Lindenmayer and Hobbs, 2004). OutsideAustralia, much of this concern centers around reforestation withEucalyptus sp. in areas where they are non-indigenous (Turnbull,1999). Within Australia, there is longstanding concern about theeffects on biodiversity of the historical replacement of native forestwith exotic Pinus sp., and about the conservation value of exoticmonoculture plantations for tropical fauna in the current landscapemosaic (reviewed by Lindenmayer and Hobbs, 2004; Kanowski et al.,2005).

Amazonian and Indo-Malayan rainforests are being destroyedmore quickly than they can regenerate (FAO, 2007), but rainforestsin other, more developed countries, such as Australia, are alsoaffected (Kanowski et al., 2005). In many areas, plantations of non-native trees, particularly Eucalyptus and Pinus species (FAO, 2007),have replaced tropical rainforests (Kanowski et al., 2005), or arecurrently favored as a means of afforesting land adjacent torainforest (Zanne and Chapman, 2001; Ashton et al., 1997).

With the disappearance of tropical forests, or their replacementby plantations, there is an associated loss of faunal biodiversity(Kanowski et al., 2005). Reptiles are an important element of thefauna in tropical forests worldwide, but their responses to tropicalforest management are poorly understood (Alcala et al., 2004;Trainor and Woinarski, 1994; Gardner et al., 2007b; Cunninghamet al., 2007). Reptile populations are especially sensitive tomanipulations of the structure of vegetation communities (Vittet al., 1998; Kavanagh and Stanton, 2005; Gardner et al., 2007a;Brown et al., 2008). Because plantation management stronglyalters vegetation structure (McCullough, 1999), solar radiationlevels, and environmental temperatures (Yirdaw and Luukkanen,2004; Lemenih et al., 2004), the response of reptiles to plantationsmay be driven by their response to these environmental changes(Jaggi and Baur, 1999; Valentine, 2006). Understanding themechanisms driving reptile responses to plantations will improvethe ability of managers to assess the influence of changes tomanagement regimes on reptile populations.

Specific aspects of management (e.g., burning, leaving slash andother coarse woody debris, etc.) are critical to maintaining faunaldiversity in plantations (Wagner et al., 2004; Kanowski et al.,2005). This is particularly true when non-native trees are used forreforestation, because they are even less likely to support diversefaunal assemblages than are native plantations (e.g., Lindenmayerand Hobbs, 2004). Low-intensity burning is commonly used tocontrol fuel load, weeds and understorey vegetation in plantations(Woods, 1990; Richardson, 1993) and native forests (Penman et al.,2007). Fire changes the structure of the vegetation (Gill, 1981;Singh et al., 2002), and can influence the size and dynamics ofassociated animal populations (e.g. Kutt and Woinarski, 2007), soburning for management can strongly affect biodiversity withinplantations. Results from the few studies that have described thespecific effects of burning regimes on ground-associated faunahave identified significant changes in species richness, and species-specific variability in response (Trainor and Woinarski, 1994; Pennet al., 2003). However, research that deals with the response oftropical reptile communities to management-associated controlburning in plantations is scarce.

We compared the species richness and structure of tropicalAustralian diurnal reptile assemblages between native forests andplantations of Pinus caribaea. In our study, pine plantations weremanaged by low-intensity fuel reduction burning under the trees, apractice that alters understorey vegetation structure in native

forests (Singh et al., 2002). Because reptiles are likely to be sensitiveto such changes, we examined how burning in plantations affectedreptile assemblage structure, and compared it to the influence ofnatural fire in native habitats. Finally, we examined the relationshipsbetween reptile abundance and richness, the amount of sunlightreaching the ground, and available thermal microclimates, in burntand unburnt plantations and natural forest. By comparing thesefactors between plantations and native forests, we aimed to identifywhat effect plantations, and management within these, had on thestructure of a tropical reptile assemblage.

2. Methods

2.1. Study sites

This study was conducted between August 1996 and March1997 at Broadwater, Abergowrie State Forest logging area, in thenorthern Queensland tropics (188280S, 146820E), 47 km west ofIngham. The area is used primarily for the production of Caribbeanpine (P. caribaea), which is grown as a monoculture. Pineplantations were established on land cleared of both rainforestand eucalypt woodland for plantation development, and were 14–15 years old. Native vegetation adjacent to pine plantationsincluded paperbark (Melaleuca viridiflora) and blue gum (Eucalyp-

tus tereticornis) woodlands, as well as riparian lowland rainforestconnected to the southern end of the Wet Tropics bioregion. Allsites established in the study were similar in aspect, slope andelevation. We described the reptile assemblages and vegetationstructure in pine, eucalypt and Melaleuca forests. Rainforest wasexcluded from sampling as patches of this vegetation type wereeither radically different in slope and aspect from the rest of theforest types sampled, or occurred only in riparian strips.

We established two field sites in each of five habitat types: pineburnt 25 months before the study commenced (so-called‘‘unburnt’’ pine sites), pine burnt 13 months before commence-ment of the study (burnt pine sites), eucalypt woodlands unburntfor eight years before our study (unburnt eucalypt sites), eucalyptwoodlands burnt 22 months before the study commenced (burnteucalypt sites), and Melaleuca woodland that had been burnt by ahigh intensity fire approximately 10 years earlier. Fires in eucalyptand pine (controlled burn) sites were of low intensity, with flamesreaching no higher than 20 cms (G. Featonby, personal communi-cation). Sampling areas were approximately 100 ha in size, wereestablished at least 100 m apart, and were surrounded by forest ofthe same origin. Each site was at least 80 m from any forest edge, tominimize edge effects. As we could not orthogonally match firehistories of different forest types, or accurately date the lastburning of the Melaleuca forest, investigations of effects of burningon reptiles and plant assemblage structures were restricted towithin-forest comparisons in pine and eucalypt. Replication offield sites was necessarily limited in this study by the availabilitypine plantations and eucalypt forests in the study area that hadexperienced comparable fire intensity and time since burning.

2.2. Vegetation microhabitat descriptions

To characterize the vegetation structure in different forest typeswith different burning histories, four parallel transects 80-m longand 25-m apart were located at each site. Four 5-m2 quadrats wererandomly sampled from each transect three times during thestudy. Tree height was recorded with a range finder and crownseparation was estimated using the methodology and photographsdescribed by McDonald et al. (1990). One observer (BM) producedvisual estimates of the following ground-level microhabitatfeatures in each quadrat: shrub cover, leaf litter cover and depth,leaf litter composition (% non-native litter excluding pine needles),

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925918

the amount of the quadrat covered by the basal areas of tree trunks,grass cover, log cover, bare ground, stump cover, and understoreyheight. Percentage estimates of microhabitat features wererecorded into one of ten categories to allow repeatability amongquadrats. Each incremental category represented a 10% increase inthe cover of the structural feature being measured.

2.3. Reptile assemblage structure

Reptile diversity and abundance was estimated at each siteusing pitfall traps and diurnal active searches concurrently, to limitthe effect of sampling biases associated with each method(Goldingay et al., 1996). Six pitfall trap arrays were established,each comprising one 1-L trap (PVC pipe 15 cm � 8.75 cm with flyscreen base) and one 20-L trap (a metal bucket 57 cm � 33 cm).Pitfall traps within an array were arranged linearly, and wereconnected by 5 m of 30-cm-high plastic mesh drift-fence, buried5 mm into the soil. Arrays were installed at each site to maximizethe distance between arrays and edges within logging coupes, andso were at least 100-m apart within each site. Four of the sixtrapping arrays at each site were established adjacent to avegetation quadrat to coordinate the spatial locations of speciesand environment data. All reptiles captured in traps were paint-marked (snakes) or toe clipped (lizards) to allow identification ofrecaptures, and released at the site of capture. Paint markspersisted on individuals for the duration of each survey period,confirmed in a pilot study previous to this one (Mott, unpublisheddata). Reptile identification followed Cogger (2000) and Ingramand Covacevich (1989). Fifty days of data collection yielded atrapping effort of 3600 trap-days. To augment pitfall trappingabundance estimates and to sample trap-shy species, reptilediversity was assessed using diurnal active searches. Hour-longsearches were conducted between 06:00 and 09:00 and repeatedbetween 15:00 and 18:30 (Eastern Standard Time) on five rainlessdays in each of the ten field sites between January and March 1997.Timed active searches began at a randomly selected corner of eachsite and were conducted along a randomly selected compassbearing for 500 m. Searches included log rolling and bark removalas well as recording surface-active reptiles sighted within 5 m ofthe transect midline. Searches were conducted at least 80 m fromany trapping array. Species identification was possible for allindividuals sighted. Abundance estimates used in analyses werecalculated from pooled active-search and pitfall trapping data.

2.4. Operative environmental temperatures in different forests

To characterize the operative environmental temperature range(i.e., the range of thermal microhabitats available to diurnalheliothermic reptiles, Bakken et al., 1985) in each forest type,models of lizards were attached to HOBO TempTM miniaturetemperature data loggers and placed in the field. Thermal modelswere built to mimic the thermal properties of a Carlia-sized lizard(the most abundant genus in the study area), and were made of 4-mm diameter clear vinyl tubing, 40 mm long, painted with grey latexpaint (1 black:1 white), and filled with air. Probes attached to loggerswere sealed inside models to prevent convective heat transfer.Experiments showed that these models reached the same equilibri-um temperature as fresh lizard carcasses in identical thermalconditions (unpub. data). In addition, these models were likely toprovide good estimates of operative environmental temperaturesfor all the reptiles at our sites (Shine and Kearney, 2001).

In the field, six models were placed on the ground surface, threein the hottest (full sun) and three in coolest (deep shade)microhabitat available at each site. Models were situated inmicrohabitats where lizards and snakes were observed, to measurethe range of operative environmental temperatures available to

active reptiles. Models, when used this way, measured theextremes of the microclimatic thermal gradient to which reptilescould be exposed, and we assumed that they would use behaviourto achieve any desired body temperature between operativetemperature extremes offered by sun and shade microhabitats(Christian and Weavers, 1996). Loggers recorded temperaturesevery 72 min and were left in place for three days in each allocatedmicrohabitat. Loggers were moved three times to randomlyselected points in the same microhabitat, for two further three-day-long samples, to increase within-microhabitat replication.Each microhabitat was sampled this way three times betweenDecember 1996 and February 1997 resulting in 27 days ofstaggered sampling at each site.

2.5. Amounts of solar radiation reaching the forest floor

Hemispherical photography was used to characterize canopystructure in pine and native forests by measuring the penetrationof radiant energy through the canopy (Messier and Puttonen,1995). A Canon AE camera was fitted with a 1808 F1-series fisheyelens, loaded with high-contrast black and white film, and fixed to atripod 1 m above the ground. The top of the camera was oriented tomagnetic north, and set level, lens pointing up and parallel to theground. One photograph was taken every 10 m along four parallel50-m transects spaced 80 m apart, providing a total of 20photographs per site for each of the five habitat types. Transectswere located such that they encompassed one vegetation structuretransect, to achieve congruency between the spatial locations ofthe three data sets within each site, and to ensure that the radiationdata collected included microhabitats that reptiles used. Thefifteen clearest photographs from each site were digitised at aresolution of 2025 dpi using a Polaroid Sprintscan 35 scanner. Allphotographs were taken under uniformly cloudy skies, to avoidoverexposing the region of the photograph around the sun(Trichon et al., 1998). Amounts of direct solar radiation belowthe canopy, were estimated for each day of the year using the‘‘Winphot’’ canopy analysis program (ter Steege, 1996). Thisprogram uses hemispherical photographs and solar geometry topredict amounts of photosynthetically active radiant energyreaching the ground at the point sampled, throughout the courseof a year. This type of radiant energy is thermally important toreptiles (e.g., Langkilde et al., 2003), in contrast to diffuse orscattered light below the canopy, which provides too little radiantenergy to allow effective thermoregulation.

2.6. Data analysis

2.6.1. Relating reptile assemblage and vegetation structure

We used principal component analyses (PCA) to explore thestructure of the data, and tested for differences betweentreatments using one-way multivariate randomization tests(Manly, 1991). These tests have high power to detect relationshipsand are robust to low degrees of freedom (Manly, 1991). The largemultivariate dimensions of the vegetation and reptile abundancedata sets, and low replication at the level of site, reduced theusefulness of parametric multivariate analysis of variance (MAN-OVA) to test for differences in vegetation structure and reptileabundances associated with different forest types, because of itsrestrictive assumptions about data distribution. Therefore, thestatistic chosen for randomization testing was the sum of squaredEuclidean distances between data points and their treatmentcentroids, which is less sensitive to data outliers than is Wilks’Lambda, the statistic often calculated in parametric MANOVAs(Manly, 1991).

When the results of a randomization test were significant, theoverall test was followed by a series of pairwise comparisons

Fig. 1. Associations among vegetation variables, as identified by principal

components analysis, at sites characterized by different forest types (plantation

pine, eucalypt woodland, and Melaleuca woodland), and showing fire history within

forest types. Symbols are centroids of sites. Plantation pine sites separate strongly

from native forest types, whereas the two native vegetation types separate less

strongly. Fire histories within eucalypt and pine forest also separate clearly.

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925 919

testing for differences between individual treatment levels toidentify where differences lay. All tests used 20,000 randomiza-tions unless otherwise specified. Each set of pairwise comparisonswas performed at an unadjusted significance level of a = 0.05, asthe initial overall test ensured that the experimentwise type I errorrate was maintained at 0.05 (Hochberg and Tamhane, 1987).

We used randomization tests to determine whether ourmeasures of vegetation structure differed among forest types,using values of each vegetation variable from individual quadrats.We used separate randomization tests for eucalypt and pine foreststo determine whether vegetation structure differed between burntand unburnt areas. We analyzed pooled abundances from pitfalltrapping and active search data to determine whether reptileassemblage structure differed among forest types. Hypothesis testson species richness and abundance were performed using Kruskal–Wallis (K–W) tests or Mann–Whitney U-tests, with significancetesting using Monte Carlo methods with 20,000 iterations. Weused two K–W tests to determine whether species richness andabundance, pooled across fire histories, differed among eucalypt,Melaleuca and pine forests. We could not compare reptileabundance and richness between burnt and unburnt sites withinhabitat types using multivariate randomization tests, because non-parametric tests cannot return a significant result for data withtwo classes and two replicates per class. Instead we usedqualitative description of abundance data and examined separa-tion on PCA biplots. Further, to verify that any difference inabundance of reptiles between burnt and unburnt pine was not anartifact of greater detection probability during visual surveys in themore open burnt treatment, we compared estimates of abundancegenerated only from pitfall trapping data, which does not rely onsightings.

Relationships between vegetation structure, reptile abun-dances and forest types identified by our multivariate randomi-zation tests were explored using PCA. The vegetation variableswere z-score transformed to standardise variables, and thusresultant correlation matrices, before graphing. Reptile abun-dances were log(x + 1) transformed, then column-centered tonormalise the data and generate optimally scaled covariancematrices. A biplot overlaying loadings of reptile abundance dataand centroids of means of vegetation data on a reduced space mapwas generated to illustrate the relationships suggested by therandomization tests.

2.7. Describing operative environmental temperatures

Only daytime temperatures were used in analyses, as more thaneighty percent of the reptiles we captured or saw were diurnal. Wetested for differences in operative environmental temperaturesamong forest types by using multivariate randomization tests toconduct the non-parametric equivalent of multivariate repeated-measures ANOVAs on means of data generated by the biophysicallizard models. Time of day was categorized into fourteen, 1-hperiods extending from 06:00 to 20:00 and pooled across dayswithin trips. Our first analysis tested for overall differencesbetween sun and shade temperatures, using data pooled across allburning regimes and forest types. Because this test detectedsignificant differences between sun and shade microhabitats whenall three forest types were pooled, we included forest type as afactor in further analyses. We used four multivariate randomiza-tion tests to compare the daily profiles of sun and shadetemperatures between burnt and unburnt eucalypt and pine sites.

2.8. Describing solar radiation reaching the forest floor

To test the hypothesis that solar radiation reaching the forestfloor was homogenous across forest types, direct, below-canopy

radiant energy (photosynthetic photon flux density—PPFD) wascompared among pine, eucalypt and Melaleuca forests using nestedrepeated-measures analysis of variance. Data reduction wasrequired, as the Winphot program estimates daily levels of PPFDat a sampled point over the course of a year, and thus necessarilygenerates a repeated-measures data structure. Data were pooledacross transects within sites and averaged over months to produce24 samples for each forest type. Repeated-measures ANOVA wasthen used to test for differences in mean monthly radiationpenetration in eucalypt, Melaleuca and pine forests. In our design,forest type was a fixed factor, and site was a nested, random factor.Within-subjects tests were Greenhouse–Geisser adjusted F-tests,as too few degrees of freedom were available to test for constantcorrelation between sampling units through time (Jongman et al.,1997). To test for effects of fire history on mean solar radiationreaching the ground, burnt and unburnt eucalypt and pine siteswere compared by repeating the analyses, but modifying themodel to include fire history as a fixed factor crossed with foresttype, and excluding the Melaleuca habitat.

Statistical analyses reported in this study were performed usingstatistical programs SPSS V.13.0, RAMAN V.1.73 (Alford, 1999) andS-PLUS (2000). P values <0.05 were considered significant.

3. Results

3.1. Vegetation and reptile assemblage structure

Vegetation structure differed strongly among forest types (SSEuclidean distances among vegetation types = 1789.4, s d.f. = 1, md.f. = 6.5, n d.f. = 53, P = 0.0002). The first principal component(explaining 63% of the variation) of the PCA examining thesedifferences indicated that high tree and understorey height, as wellas low shrub cover and large amounts of deep and exotic litter,were characteristic of pine (Fig. 1). Eucalypt and Melaleuca forestsseparated on the second principal axis (12% of the variation), due to

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925920

high cover and depth of litter, and large numbers of saplings andstumps, in the eucalypt sites, and low crown separation, largenumbers of logs, bare ground and grass, and low foliage cover inthe Melaleuca sites. The relatively low separation between eucalyptand Melaleuca forests suggested they were structurally moresimilar to each other than either was to pine. If reptile assemblagestructure was determined by vegetation structure, this would leadto a greater degree of species overlap between eucalypt andMelaleuca forests than between either of these forests and pine.

A total of 728 individuals comprising 17 species of lizard (699individuals) and 7 species of snake (29 individuals) were capturedin this study (Table 1). Litter-dwelling skinks were numericallydominant, and comprised 95% and 75% of the total number ofindividuals and species captured, and 98% of the reptile faunapresent in pine. Total abundance and species richness differedstrongly among forest types when pooling data across fire histories(abundance: K–W X2 = 5.6, d.f. = 2, Monte Carlo P � 0.04; richness:K–W X2 = 5.7, d.f. = 2, Monte Carlo P � 0.04), as did the compositionof the reptile assemblage (SS Euclidean distances = 84.500, sd.f. = 2, m d.f. = 9.5, n d.f. = �8.0, P � 0.001, Table 1, Fig. 2), withpairwise comparisons identifying that the assemblage in pine wasdifferent from that in eucalypt forests (SS Euclidean dis-tances = 32.116, s d.f. = 1, m d.f. = 9.0, n d.f. = �7.5, P � 0.029),while there was no significant difference between Melaleuca andeither eucalypt (SS Euclidean distances = 23.941, s d.f. = 1, md.f. = 8.0, n d.f. = �7.5, P � 0.068), or pine (SS Euclidean dis-tances = 20.247, s d.f. = 1, m d.f. = 8.5, n d.f. = �8.0, P � 0.066)forests. PCA identified that 68% of the variation in the reptile data

Table 1Abundance and species richness of reptiles in two Australian tropical native forest types

areas of eucalypt and pine forests.

Melaleuca Numb

Eucaly

Unbur

Scincidae

Carlia foliorum 4 0

Carlia pectoralis 56 60

Carlia rostralis 21 46

Carlia rubrigularis 0 0

Carlia storri 75 25

Cryptoblepharus virgatus 1 1

Ctenotus robustus 8 2

Ctenotus taeniolatus 15 19

Cylodomorphus gerrardi 0 0

Egernia frerei 0 0

Glaphyromorphus punctulatus 0 2

Lampropholis delicata 0 0

Saproscincus basiliscus 0 0

Agamidae

Chlamydosaurus kingii 0 1

Diporiphora australis 6 5

Pygopodidae

Lialis burtonis 0 1

Varanidae

Varanus scalaris 1 1

Colubridae

Boiga irregularis 0 0

Dendrelaphis punctulata 6 1

Stegonotus cucullatus 1 1

Elapidae

Acanthophis praelongus 0 0

Demansia psammophis 5 2

Oxyuranus scutellatus 2 0

Pseudonaja textilis 3 1

Total abundance 204 168

Total richness 14 15

set was attributed to the difference in assemblage compositionbetween pine and native forests (Fig. 2). As with the vegetation,reptile species associated with eucalypt and Melaleuca forestsseparated along the second principal axis, which accounted for 14%of the variation in the data (Fig. 2). The congruence between thepatterns we found using separate PCAs on the vegetation structureand reptile species assemblages suggests that reptile assemblagestructure was strongly correlated with species-specific ties tostructural aspects of the vegetation.

The reptile assemblages identified from pine, eucalypt andMelaleuca forests in this study differed among forest types (Table 1,Fig. 2). The pine assemblage was characterized by an almostcomplete absence of snakes, a high abundance of Carlia rubrigularis

(a species which did not occur in any other habitat) and thepresence of rarer lizard species such as Egernia freri, Lampropholis

delicata and Saproscincus spectabilis. Although total reptileabundances did not differ significantly between Melaleuca andeucalypt forests (richness: Melaleuca vs Eucalypt Mann-WhitneyU = 5, d.f. = 1, Monte Carlo P = 0.73, abundance: Melaleuca vsEucalypt Mann–Whitney U = 5, d.f. = 1, Monte Carlo P = 0.8), thedistribution of abundances of shared species were quite differentbetween these forest types (Table 1, Fig. 2). The eucalypt forest wascharacterized by high abundances of Carlia pectoralis and Carlia

rostralis, and by a suite of snakes including Demansia psammophis,

Boiga irregularis and Acanthophis praelongus, while the highabundance of Carlia storri and the presence of Carlia foliorum

and more elapid snakes, including Oxyuranus scutellatus, definedthe Melaleuca forest.

(Melaleuca and eucalypt woodland), and pine plantations, and in burnt and unburnt

er of individuals

pt Pine

nt Burnt Unburnt Burnt

0 0 0

122 4 57

24 0 1

0 5 39

19 1 2

2 0 0

7 0 0

22 0 1

0 0 2

0 0 3

1 4 7

0 2 6

0 0 2

2 0 1

12 0 0

1 0 0

0 0 0

2 0 0

2 0 0

0 1 0

1 0 0

2 0 0

0 0 0

1 0 0

220 17 123

15 6 12

Fig. 2. Associations among Australian tropical reptile species abundance (as

measured by pitfall trapping and active surveys) in three forest types (eucalypt

forests and woodland, Melaleuca woodland, and plantation pine). Eucalypt

woodland and pine had sites with different fire histories. Symbols are centroids

for sites. Unburnt pine differed strongly from burnt pine, native forest types differed

strongly from pine, but there was no separation between burnt and unburnt

eucalypt sites.

Fig. 3. Mean operative environmental temperatures (�SE) available in burnt and

unburnt native Eucalypt woodlands, native Melaleuca woodlands of unknown fire

history, and burnt and unburnt plantation pine. Temperatures of models were recorded

in full sun and deep shade at locations where reptiles were sighted, over 30 days in each

site. All temperatures were recorded during the wet season, December–February.

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925 921

3.2. Fire and reptile assemblage structure

Disturbance due to fire had a significant effect on vegetationstructure within both eucalypt (SS Euclidean distances = 1343.116,s d.f. = 1, m d.f. = 36.5, n d.f. = 27, P � 0.0002) and pine forests (SSEuclidean distances = 1275.613, s d.f. = 1, m d.f. = 40.7, n d.f. = 27,P � 0.0002). In both forest types, burnt sites had shallower litterdepth and more saplings than unburnt sites (Fig. 1). However,reptile assemblages responded differently to these changes in eachforest. Burnt and unburnt sites in eucalypt woodland were similarin species composition (Table 1) and in species richness, althoughsome species did appear to differ between burnt and unburntnative vegetation sites: C. pectoralis was more abundant in burntsites, and C. rostralis had higher abundance in unburnt sites (Table1). Burnt sites were also characterized by slightly higher richnessand abundance of snakes, while the lizards Glaphyromorphus

punctulatus and Varanus scalaris distinguished unburnt sites (Table1). In pine, however, elapid snakes were absent, and the totalabundance and richness of lizards was much lower in the absenceof fire. The total abundance of all lizard species was seven timesgreater in burnt than in unburnt sites (Table 1). Examination of thepitfall trapping data alone supports this separation between burntand unburnt pine, and revealed that lizard abundance was fourtimes greater in burnt than in unburnt pine. The large separationbetween burnt and unburnt pine sites (Fig. 2) was driven by theextremely low abundances of all lizard species in unburnt pine, andthe lower species richness in unburnt pine. Burnt pine sites werecharacterized by the six-fold greater abundance of C. rubrigularis,and C. pectoralis, and the occurrence of E. freri, and Saproscincus

basiliscus in low numbers, species that did not occur in unburntpine sites at all (Table 1).

3.3. Operative environmental temperatures

When data were pooled across fire histories in the pine andeucalypt forests, mean operative environmental temperaturesdiffered among pine, eucalypt and Melaleuca woodlands in the sun(SS Euclidean distances over all 3 forest types in the sun 96.8, sd.f. = 2, m d.f. = 2, n d.f. = 4, P = 0.02). This was driven by significantdifferences between the two native forest types and the pine (SSEuclidean distances eucalyptus vs pine 83.4, s d.f. = 1, m d.f. = 2.5, nd.f. = 3, P = 0.05; SS Euclidean distances Melaleuca vs pine 53.7, sd.f. = 1, m d.f. = 2.5, n d.f. = 0.5, P = 0.03), rather than by differencesbetween the two native forest types (SS Euclidean distanceseucalyptus vs Melaleuca 56.4, s d.f. = 1, m d.f. = 2.5, n d.f. = 0.5,P = 0.21). There were, however, no significant differences in meantemperature in the shade among forest types (SS Euclideandistances over all 3 forest types in the shade 119.1, s d.f. = 2, md.f. = 2, n d.f. = 4.5, P = 0.21) (Fig. 3). Mean operative environmentaltemperatures did not differ between burnt and unburnt pine in thesun or shade (SS Euclidean distances for burnt and unburnt pine insun 88.2, s d.f. = 3, m d.f. = 1.5, n d.f. = 3, P = 0.9; SS Euclideandistances pine in shade 94.4.2, s d.f. = 3, m d.f. = 1.5, n d.f. = 2,P = 0.6), or between burnt and unburnt eucalypt woodlands in thesun, although this difference approached significance (SS Euclide-an distances in burnt and unburnt eucalypt in sun 83.6, s d.f. = 1, md.f. = 2, n d.f. = 2, P = 0.06). However, operative temperatures in theshade were significantly higher in burnt eucalypt woodlands thanin unburnt eucalypt woodlands (SS Euclidean distances for burntand unburnt eucalypt woodlands in the shade 81.5, s d.f. = 3, md.f. = 1.5, n d.f. = 2.5, P = 0.005, Fig. 3).

Operative environmental temperatures in Melaleuca woodlandwere consistently higher than in the other forest types (Fig. 3), withtemperatures in the sun above 40 8C from 07:30 until 15:00. Incontrast, operative environmental temperatures in full sun in pinewere only above 40 8C for 1 h in the middle of the day, and the pinewas the coolest microhabitat. In eucalypt woodland, operativetemperatures in the sun microhabitat were between theseextremes, above 40 8C between 10:00 and 15:00.

Fig. 4. (A) Monthly means (�SE) of radiant energy useable by reptiles (measured as photosynthetic photon flux density in W/m2) penetrating below the canopy in Eucalyptus,

Melaleuca and pine forests in tropical Australia. (B) Estimates of mean (�SE) radiant energy penetration for burnt and unburnt habitats in the pine and eucalpt forests, showing

means for Melaleuca forests for comparison.

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925922

3.4. Radiant energy

There were significant average differences in penetration ofradiant energy over time among forests (F2,6 = 105.7, P < 0.001).Pine sites (mean incident solar radiation = 14.96 � 1.90 W/m2)consistently received less solar energy at the ground than eithereucalypt (mean = 28.26 � 1.90 W/m2) or Melaleuca (mean = 29.27 �1.90 W/m2) forests (F5.2,6 = 378.06, P < 0.001) (Fig. 4a). Penetrationpeaked between April and September, but during this time maximumlevels of radiant energy in pine were 15 W/m2 lower than in eucalyptand Melaleuca sites, and only just attained the minimum levels ofpenetration experienced by the native forests outside the peak periodof the year (Fig. 4a).

Burning significantly affected amounts of radiant energyreaching the forest floor in eucalypt forests (F1,8 = 232.2,P < 0.001), but not in pine (Fig. 4b). The interaction betweenforest type and fire (i.e., the fact that burned sites) averaged overtime was significant (F1,8 = 6.4, P = 0.035) (Fig. 4b).

4. Discussion

In our study area, historical replacement of the nativevegetation with non-native Pinus plantations has strongly affectedreptile assemblage structure. There were strong associationsbetween reptile assemblage structure and vegetation structurespecific to each forest type. Pine plantations contained structurallysimpler understorey vegetation, were cool for more of the day, andreceived less radiant energy at the ground than did native forests.These physical attributes attracted a reptile assemblage distinctfrom those using the sunnier and more complex adjacent nativeforests. In eucalypt forests, fire altered understorey vegetationstructure, operative environmental temperatures and radiantenergy penetration to the forest floor, but it did not appear tosignificantly alter reptile assemblage structure. On the other hand,in pine, burning strongly affected the species richness and relativeabundance of reptiles; reptile assemblages using unburnt pinewere depauperate in terms of snakes, and lizard abundance wasreduced in comparison to recently burnt sites. Burning in pineaffected understory vegetation structure, removing weedy leaflitter and larger woody weeds. Burning in pine did not, however,affect operative environmental temperatures or radiant energypenetration, which was very low in both burned and unburnedpine.

Introduced plantation pine is usually resource-poor withdepauperate reptile faunas (Kanowski et al., 2005; Gardner etal., 2007b). However, in our study, overall reptile richness was

broadly similar in burnt pine and native forests. Pine altered lizardassemblage structure by excluding species from several generacommon in eucalypt and Melaleuca forests, including most Carlia,

Ctenotus, Diporiphora and Cryptoblepharus, and by attractingspecies usually associated with cool thermal environments(Wilson and Swan, 2003); commonly found in closed-canopyrainforest and complex vine-thickets in the region, including pink-tongued lizards (Cyclodomorphus gerrardi), Egernia frerei, L. delicata

and S. basiliscus. In an unreplicated, but suggestive, studycomparing African pine and native forests, Vonesh (2001) alsofound that pine strongly altered the reptile assemblage usingforests, and enhanced habitats for species normally found inrainforest. Because pine offers a cool thermal environment, andbecause the lizard species using pine preferred cool microhabitats(Ehmann, 1992), it is likely that, in the tropics, pine can act as botha habitat and a thermal refuge for lizards dispersing from nativeforests with closed canopies and cool thermal environments. Assuch, pine may act as an important element of the landscapemosaic in areas where closed-canopy forests (e.g., lowlandrainforests) are disappearing. On the other hand, snakes andseveral thermophilic lizard species were excluded from pineplantations in this study. Thus, although pine plantations may actto complement native forest in a landscape mosaic by providinghabitat for some species, they are unlikely to replace native forestas habitat for snakes and many other lizard species. It may beuseful to think of tropical plantations as a refuge for a select groupof reptile species, because they can support high reptile diversity incomparison to cleared land (Borsboom et al., 2002). Thus,understanding the conservation value of tropical plantations willrequire attention to the overall role of plantations in the landscapemosaic.

Reptiles often associate with particular habitats on the basis offine-scale habitat features (e.g. Brown and Nelson, 1993; Green-berg et al., 1994; Lindenmayer et al., 2001; Fischer et al., 2004).Structurally complex vegetation may increase reptile speciesrichness and abundance within a habitat by providing a greaterdiversity of niches (e.g., Jones, 1981; Hadden and Westbrooke,1996). Pine plantations are structurally simple (Friend, 1980) incomparison to native forest, due to a planting pattern thatpromotes closed-canopy structure designed to maximize produc-tion and shade out competitive tree species, and to maintenancepractices (frequent burning) intended to remove understoreyregrowth. Reptile abundances may have been lower in pine than innative forests in this study because of reduced structuralcomplexity compared to native vegetation. Reduced lizardabundance could be caused by different litter structure in pine

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925 923

and native forests, and concomitant altered retreat site availabilityand opportunity for camouflage (Valentine et al., 2007). Changes tosoil chemistry associated with decomposing pine litter may havealso made plantations unattractive to reptiles, either by alteringinvertebrate food availability (e.g., Finch and Szumelda, 2007;Robson et al., 2009) or influencing other factors such as nestingsuccess (e.g., Marco et al., 2005), but we have no evidence for this.Future studies might explore the influence of soil chemistry onreptiles in this system.

The almost total absence of snakes in pine plantations that weobserved accords with other studies that have identified managedpine as less speciose than native forests, at both stand (Loehle et al.,2005), and landscape (Gardner et al., 2007a), levels. Radiationpenetration through the canopy is particularly important forthermoregulation in sub-tropical snakes that are not shuttlingheliotherms, or that bask in retreats (e.g. Pringle et al., 2003), andeffective thermoregulation positively effects reproduction andactivity in tropical snakes (Liuiselli and Akani, 2002). In our study itis probable that the low level of incident radiation penetrating thepine canopy generated a thermally hostile habitat for most snakes(e.g. Owens et al., 2008). Snakes present in our study area hadpreferred temperature ranges between 27 and 39 8C (Greer, 2005)that would have been unachievable in pine for the majority of theday. Further, other research has demonstrated that snake diversityin pine relies heavily on the presence of complex retreat sites onthe forest floor, such as coarse woody debris (e.g. Owens et al.,2008), or soil cracks, neither of which were present in our pineplantations due to site establishment procedures. A thirdexplanation for the lack of snakes in pine may have been a lackfood resources, as many Australian pine forests support very lowabundances of small mammals (e.g. Lindenmayer et al., 1999),although whether this was the case in our study area would requirea more mammal-focused assessment of diversity.

4.1. The effect of fire

Reptile species using fire-prone native forests, such as eucalyptforests in Australia and pitch pine (Pinus rigidia) forests in America,are adapted to the periodic burning that is typical of these habitats(Letnic et al., 2004). Because of this adaptation, reptile assemblagesfrom such habitats may be more resilient to changes following firethan those from other habitat types (e.g. Ford et al., 1999). In fire-prone habitats, individual reptile species may show decreasedabundance post-fire due to higher predation rates (Webb andShine, 2008) or decreased prey availability (Friend, 1993).However, many species are either unaffected (Singh et al.,2002), or only negatively affected in specific sub-groups withinan assemblage (e.g., litter dwellers, Penn et al., 2003).

Although burning altered vegetation structure in pine andeucalypt forests in our study, only in pine forests were thesechanges associated with changes in reptile assemblage structure.This difference may be have been influenced by the disparity intime since burning between the treatments. Because burnteucalypt sites were burnt two years before sampling, whereasburnt pine sites were burnt only one year before sampling, theeucalypt vegetation and lizard assemblage had more time toregenerate, so our ability to identify short-term differences drivenby fire in eucalypt woodlands was reduced. However, even if thefire history of the two forests had been more similar, it is likely thatburning would still have more strongly influenced reptileassemblage structure in pine, because burning reduced the amountof unfavorable exotic leaf litter present in the pine understorey,and exotic litter reduces the abundance of reptiles (Valentine et al.,2007).

The understorey vegetation in pine consisted of exoticperennial herbs, vines and forbs, which are much more fire

sensitive than the native vegetation understorey (DiTomaso et al.,2006). In the pine understorey, recent burning produced plantassemblages structurally dissimilar to those present prior to themost recent burn. Before burning, pine sites were characterized bythe low vine Desmodium and large stands of woody Lantana,whereas after burning pine sites were dominated by a herbaceousweed (Ageratum) and stands of Imperata grass. The differences inreptile assemblage structure between burnt and unburnt pinewere probably driven by the relationship of the species using pinewith available structural microenvironments.

Analysis of estimates of diversity and abundance generatedusing pitfall trapping data (and therefore not influenced byobservation bias) confirmed that low-intensity burning enhancedreptile species abundance and richness in the pine. Burning in pineremoved introduced, broadleaved weeds and non-native pine-needle litter. Australian tropical leaf-litter skinks avoid weeds andnon-native leaf litter, because of reduced food availability andreduced opportunities for camouflage in these habitats (Valentineet al., 2007). Thus, reduced litter depth caused by burning in pinemay enhance the habitat for skinks (by removing exotic or weedylitter), whereas reduced litter depth in native habitats afterburning reduces the attractiveness of the habitat, because itremoves a favored habitat feature.

The increased richness and abundance of lizards in burnt pineforests was not caused by increased opportunities for thermoreg-ulation. Our data demonstrated that removal of the understoreyvegetation in the pine did not increase operative environmentaltemperatures or alter the extremely low levels of solar radiation atthe ground. Controlled burning did not directly affect the treecanopy, which remained closed.

4.2. Implications of the study

We acknowledge some constraints in our design. In particular,low replication at the level of site, and sampling fauna in only oneyear could have reduced our ability to detect the effects of foresttype and treatment on reptile abundance and richness. In spite ofthis, we found that characteristics of sites within treatments weresimilar, and strong influences of these forest types and burningregimes on abundance and assemblage composition could bedetected. Increasing replication, and so necessarily confoundingfire history and plantation age, would have reduced our ability todistinguish the effects of burning. Studies such as ours areimportant because they suggest a starting point from whichfurther, more detailed exploration of responses to disturbance canbe undertaken.

Native fauna of a variety of taxa are typically less abundantwithin exotic plantations, but are also often less species rich(reviewed by Hartley, 2002). Often, fauna using plantations are asubset of the fauna that use adjacent native forests, typicallygeneralist species that can use a range of habitats (e.g. Kavanaghet al., 2001; Borsboom et al., 2002). Our results differed from thispattern. Like Vonesh (2001), we found that the reptiles within pineplantations were not a subset of adjacent native forest fauna, butwere characterized by several species that were typical of locallowland rainforest (Kanowski et al., 2006). We also found thatcontrolled burning of pine plantations greatly increased overallspecies richness, primarily by increasing the abundance of litterskink species (C. pectoralis, C. rubrigularis).

These results have three important management implications.First, at least in the tropical plantations we studied, pine did notconstitute a ‘‘biological desert’’ (Allen et al., 1995) for diurnalreptiles, but supported reptile communities nearly as rich as thosein native forests (Vonesh, 2001). The plantations we studied werecharacterized by high levels of retained native vegetation alongcreeks and in conservation areas. The occurrence of rainforest

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925924

reptiles in other habitat types is partially dependent on theproximity of rainforest from which to recolonise (Kanowski et al.,2006), so distance to native vegetation should be taken intoaccount when interpreting our results: our sites were close toretained rainforest patches. Other pine plantations, isolated fromnative forests, or which do not possess retained native vegetation,may be far less species diverse than those in our study. However, itis important that the ability of pine to support many reptile speciesbe recognized in conservation planning, particularly where pine isdestined to be adjacent to cleared land, as is the case in many newplantations due to current moves towards afforesting reclaimedland. Finally, it is important to acknowledge that while pine doessupport reptiles, it represents a radically different environmentfrom that available in local schlerophyllous forests, and, in ourstudy, did not support lizards requiring high operative tempera-tures (i.e., the majority of species we captured, or snakes).

Second, the species supported by the cool, shady pineenvironment in this study were typical of rainforest species foundin the region (see species reported in Kanowski et al. (2005)),suggesting that pine may serve as a refuge habitat for somelowland rainforest species, if recolonisation is possible (Kanowskiet al., 2006). Because we did not measure the richness orabundance of species within rainforest, determining the relativeefficacy of pine as a refuge requires further research.

Third, the management of pine (burning to remove understoreyvegetation) had a positive impact on the species richness andabundances of reptiles using pine. This was probably caused bychanges in the structure of the pine understorey, particularly theremoval of non-native litter, which allowed invasion of this habitatby species that were otherwise excluded. Thus, while pine is not asubstitute for schlerophyll forest, if appropriately managed it hasthe potential to positively complement a proportion of the reptilediversity of native forest at a landscape scale. To advocate burningas standard management practice requires more detailed study,but our results suggest that burning does enhance reptile richnessin pine. This finding should be considered where increasing faunaldiversity in pine is a management goal, as current trends inplantation management in Australia are to exclude fire due to itsnegative effects on productivity (Woods, 1990—Pinus radiata). Intropical plantations, the choice of whether or not to burn should bemade considering the tradeoff between maximizing profit andconserving plantation faunal diversity.

Acknowledgments

Funding and logistic support for this study was provided by theDept of Natural Resources, Queensland. Funding for a part of thisresearch was generously provided by the Linnaean Society of NewSouth Wales. The manuscript was improved by comments fromanonymous reviewers. Work in this study follows the AustralianCode of Practice for the Care and Use of Animals for ScientificPurposes and was reviewed by the James Cook University EthicsCommittee. A Queensland National Parks permit to Interfere withWildlife and a State Forestry Permit to Traverse supported this study.

References

Alcala, E.L., Alcala, A.C., Dolino, C.N., 2004. Amphibians and reptiles in tropicalrainforest fragments on Negros Island, the Philippines. Environmental Conser-vation 31, 254–261.

Alford, R.A., 1999. RAMAN manual. http://www.pibweb.com/ross/RAMAN/raman.pdf.Allen, R.B., Platt, K.H., Coker, R.E.J., 1995. Understorey species composition patterns

in a Pinus radiata D. Don plantation on the central North Island volcanic plateau.New Zealand Journal of Forestry Science 25, 301–317.

Ashton, P.M.S., Gamage, S., Gunatilleke, I.A.U.N., Gunatilleke, C.V.S., 1997. Restora-tion of a Sri Lankan rainforest: using Caribbean pine Pinus caribaea as a nurse forestablishing late-successional tree species. Journal of Applied Ecology 34, 915–925.

Bakken, G.S., Santee, W.R., Erskine, D.J., 1985. Operative and standard operativetemperature: tools for thermal energetics studies. American Zoologist 25, 933–943.

Borsboom, A.C., Wang, J., Lees, N., Mathieson, M., Hogan, L., 2002. Measurement andIntegration of Fauna Biodiversity Values in Queensland Agroforestry Systems.Natural Heritage Trust RIRDC Publication No. 02/044, Rural Industries Researchand Development Corporation.

Brown, G.W., Nelson, J.L., 1993. Influence of successional stage of Eucalyptus regnans(mountain ash) on habitat use by reptiles in the Central Highlands, Victoria.Austral Ecology 18, 405–417.

Brown, G.W., Bennett, A.F., Potts, J.M., 2008. Regional faunal decline—reptile occur-rence in fragmented rural landscapes of south-eastern Australia. WildlifeResearch 35, 8–18.

Cogger, H.G., 2000. Reptiles and Amphibians of Australia. Reed New Holland, Sydney.Cunningham, R.B., Lindenmayer, D.B., Crane, M., Michael, D., MacGregor, C., 2007.

Reptile and arboreal marsupial response to replanted vegetation in agriculturallandscapes. Ecological Applications 17, 609–619.

Christian, K.A., Weavers, B.W., 1996. Thermoregulation of monitor lizards in Aus-tralia: an evaluation of methods in thermal biology. Ecological Monographs 66,139–157.

DiTomaso, J.M., Brooks, M.L., Allen, E.B., Minnich, R., Rice, P.M., Kyser, G.B., 2006.Control of invasive weeds with prescribed burning. Weed Technology 20, 535–548.

Ehmann, H., 1992. Encyclopedia of Australian Animals: Reptiles. Harper Collins,Sydney.

Finch, O.D., Szumelda, A., 2007. Introduction of Douglas fir (Pseudotsuga menziesii(Mirb.) Franco) into Western Europe: epigaeic arthropods in intermediate-agedpure stands in northwestern Germany. Forest Ecology and Management 242,260–272.

Food and Agriculture Organisation, 2007. The State of the World’s Forests (SOFO).Food and Agriculture Organisation of the United Nations, Rome.

Ford, W.M., Menzel, M.A., McGill, D.W., Laerm, J., McCay, T.S., 1999. Effects of acommunity restoration fire on small mammals and herpetofauna in the south-ern Appalachians. Forest Ecology and Management 114, 233–243.

Fischer, J., Lindenmayer, D.B., Cowling, A., 2004. The challenge of managing multiplespecies at multiple scales: reptiles in an Australian grazing landscape. Journal ofApplied Ecology 41, 32–44.

Friend, G.R., 1980. Wildlife conservation and softwood forestry in Australia: someconsiderations. Australian Forestry 43, 217–224.

Friend, G.R., 1993. Impact of fire on small vertebrates in Mallee woodlands andheathlands of temperate Australia: a review. Biological Conservation 65, 99–114.

Gardner, T.A., Barlow, J.A., Peres, C.A., 2007a. Paradox, presumption and pitfalls inconservation biology: the importance of habitat change for amphibians andreptiles. Biological Conservation 138, 166–179.

Gardner, T.A., Ribeiro Jr., M.A., Barlow, J.A., vila-Pires, T.A.S., Hoogmoed, M.S., Peres,C.A., 2007b. The biodiversity value of primary, secondary and plantation forestsfor a neotropical herpetofauna. Conservation Biology 21, 775–787.

Gavran, M., Parsons, M., 2009. Australia’s Plantations 2009 Inventory Update.Bureau of Rural Sciences, Canberra.

Gill, A.M., 1981. Adaptive responses of Australian vascular plant species to fire. In:Gill, A.M., Groves, R.H., Noble, I.R. (Eds.), Fire and the Australian Biota. AustralianAcademy of Science, Canberra, pp. 234–269.

Goldingay, R., Daly, G., Lemckert, F., 1996. Assessing the impacts of logging onreptiles and frogs in the Montane forests of southern New South Wales. WildlifeResearch 23, 495–510.

Greenberg, C.H., Neary, D.G., Harris, L.D., 1994. Effect of high-intensity wildfire andsilvicultural treatments on herpetofaunal communities in sand-pine scrub.Conservation Biology 8, 1047–1057.

Greer, A.E., 2005. In: Greer, A.E. (Ed.), Encyclopedia of Australian Reptiles. AustralianMuseum Online, Sydney.

Hadden, S.A., Westbrooke, M.E., 1996. Habitat relationships of the herpetofauna ofremnant buloke woodlands of the Wimmera Plains, Victoria. Wildlife Research23, 363–372.

Hartley, M.J., 2002. Rationale and methods for conserving biodiversity in plantationforests. Forest Ecology and Management 155, 81–95.

Hochberg, Y., Tamhane, A.C., 1987. Multiple Comparison Procedures. John Wiley &Sons, New York.

Ingram, G., Covacevich, J., 1989. Revision of the genus Carlia (Reptilia, Scincidae) inAustralia with comments on Carlia bicarinata of New Guinea. Memoirs of theQueensland Museum 27, 443–490.

Jaggi, C., Baur, B., 1999. Overgrowing forest as a possible cause for the localextinction of Vipera aspis in the northern Swiss Jura mountains. Amphibia-Reptilia 20, 25–34.

Jones, B.K., 1981. Effects of grazing on lizard abundance and diversity in westernArizona. Southwestern Naturalist 26, 107–115.

Jongman, R.H.G., ter Braak, C.J.F., van Tongeren, O.F.R., 1997. Data Analysis inCommunity and Landscape Ecology. Cambridge University Press, Cambridge.

Kanowski, J., Catterall, C.P., Wardell-Johnson, G.W., 2005. Consequences of broad-scale timber plantations for biodiversity in cleared rainforest landscapes oftropical and subtropical Australia. Forest Ecology and Management 208, 359–372.

Kanowski, J., Reis, T.M., Catterall, C.P., Piper, S.D., 2006. Factors effecting the use ofreforested sites by reptiles in cleared landscapes in tropical and subtropicalAustralia. Restoration Ecology 14, 67–76.

Kavanagh, R.P., Law, B., Lemkert, F., Stanton, M., Chidel, M., Towerton, A., 2001.Birds, mammals, reptiles and amphibians in eucalypt plantations near

B. Mott et al. / Forest Ecology and Management 259 (2010) 916–925 925

Albury–Wodonga: a pilot study of variables influencing biodiversity. StateForest of New South Wales on progress under action 2.5 of the NSW SalinityStrategy.

Kavanagh, R., Stanton, M.A., 2005. Vertebrate species assemblages and speciessensitivity to logging in the forests of north-eastern New South Wales. ForestEcology and Management 209, 309–341.

Kutt, A.L.S., Woinarski, J.C.Z., 2007. The effects of grazing and fire on vegetation andthe vertebrate assemblage in a tropical Savanna woodland in north-easternAustralia. Journal of Tropical Ecology 23, 95–106.

Langkilde, T., O’Connor, D., Shine, R., 2003. Shelter-site use by five species ofMontane scincid lizards in south-eastern Australia. Australian Journal of Zool-ogy 51, 175–186.

Lemenih, M., Gidyelew, T., Teketay, D., 2004. Effects of canopy cover and under-storey environment of tree plantations on richness, density and size of colo-nizing woody species in southern Ethiopia. Forest Ecology and Management194, 1–10.

Letnic, M., Dickman, C.R., Tischler, M., Tamayo, B., Beh, C.-L., 2004. The responses ofsmall mammals and lizards to fire and rainfall in arid Australia. Journal of AridEnvironments 59, 85–114.

Lindenmayer, D.B., Cunningham, R.B., Pope, M.L., Donnelly, C.F., 1999. A large-scaleexperiment to examine the effects of landscape context and habitat fragmen-tation on mammals. Biological Conservation 88, 387–403.

Lindenmayer, D., Cunningham, R.B., MacGregor, C., Tribolet, C., Donnelly, C.F., 2001.A prospective longitudinal study of landscape matrix effects on fauna inwoodland remnants: experimental design and baseline data. Biological Con-servation 101, 157–169.

Lindenmayer, D.B., Hobbs, R.J., 2004. Fauna conservation in Australian plantationforests: a review. Biological Conservation 119, 151–168.

Loehle, C., Wigley, T.B., Shipman, P.A., Fox, S.F., Rutzmoser, S., Thill, R.E., Melchiors,M.A., 2005. Herpetofaunal species richness responses to forest landscapestructure in Arkansas. Forest Ecology and Management 209, 293–308.

Liuiselli, L., Akani, G.C., 2002. Is thermoregulation really unimportant for tropicalreptiles? Comparative study of four sympatric snake species from Africa. ActaOecol. -Int. J. Ecol. 23, 59–68.

Manly, B.F.J., 1991. Randomization, Bootstrap and Monte Carlo Methods in Biology.Chapman & Hall, Oxford.

Marco, A., Lopez-Vincente, M.L., Perez-Mallado, V., 2005. Soil acidification nega-tively effects embryonic development of flexible-shelled lizard eggs. The Her-petological Journal 15, 107–111.

McCullough, R.B., 1999. Four common myths about plantation forestry. New Forests17, 111–118.

McDonald, G.T., Lane, M.B., 2002. Converging global indicators for sustainable forestmanagement. Forest Policy and Economics 6, 63–70.

McDonald, R.C., Isbell, R.F., Speight, J.G., Walker, J., Hopkins, M.S., 1990. AustralianSoil and Land Survey Field Handbook. Inkata Press, Sydney.

Mercer, D., Underwood, A., 2002. Australian timber plantations: national vision,local response. Land Use Policy 19, 107–122.

Messier, C., Puttonen, P., 1995. Spatial and temporal variation in the lightenvironment of developing Scots pine stands: the basis for a quick andefficient method of characterizing light. Canadian Journal of Forest Research25, 343–354.

Montreal Process Implementation Group for Australia, 2008. Australia’s State of theForests Report 2008. Bureau of Rural Sciences, Canberra.

Owens, A.K., Moseley, K.R., McCay, T.S., Castleberry, S.B., Kilgo, J.C., Ford, W.M., 2008.Amphibian and reptile community response to coarse woody debris manipula-tions in upland loblolly pine (Pinus taeda) forests. Forest Ecology and Manage-ment 256, 2078–2083.

Penman, T.D., Kavanagh, R.P., Binns, D.L., Melick, D.R., 2007. Patchiness of prescribedburns in dry sclerophyll eucalypt forests in South-eastern Australia. ForestEcology and Management 252, 24–32.

Penn, A.M., Sherman, W.B., Lunney, D., Banks, P.B., 2003. The effects of a low-intensity fire on small mammals and lizards in a logged, burnt forest. WildlifeResearch 2003, 477–486.

Pringle, R.M., Webb, J.K., Shine, R., 2003. Canopy structure, microclimate, andhabitat selection by a nocturnal snake. Ecology 84, 2668–2679.

Richardson, B., 1993. Vegetation management-practices in plantation forests ofAustralia and New Zealand. Canadian Journal of Forest Research 23, 1989–2005.

Robson, T.C., Baker, A.C., Murray, B.R., 2009. Differences in leaf litter invertebrateassemblages between radiata pine plantations and neighbouring native euca-lypt woodland. Austral Ecology 34, 368–376.

S-PLUS, 2000. Professional Edition for Windows, Release 3.Scott, D.A., Burger, J.A., Crane, B., 2006. Expanding site productivity research to

sustain non-timber forest functions. Forest Ecology and Management 227,185–192.

Shine, R., Kearney, M., 2001. Field studies of thermoregulation: how well dophysical models predict operative temperatures? Functional Ecology 15,282–288.

Singh, S., Smyth, A.K., Blomberg, S.P., 2002. Effect of a control burn on lizards andtheir structural environment in a eucalypt open-forest. Wildlife Research 29,447–454.

ter Steege, H., 1996. Winphot 5: a programme to analyze vegetation indices, lightand light quality from hemispherical photographs. Tropenbos-Guyana Reports95, Tropenbos-Guyana Programme, Georgetown, Guyana.

The Australian Forestry Standard, 2007. Australian Forestry Standard Limited. ACT,Australia.

Trainor, C.R., Woinarski, J.C.Z., 1994. Responses of lizards to three experimental firesin the savanna forests of Kakadu National Park. Wildlife Research 21, 131–148.

Trichon, V., Walter, J.M.N., Laumonier, Y., 1998. Identifying spatial patterns in thetropical rain forest structure using hemispherical photographs. Plant Ecology137, 227–244.

Turnbull, J.W., 1999. Eucalypt plantations. New Forests 17, 37–52.Valentine, L.E., 2006. Habitat avoidance of an introduced weed by native lizards.

Austral Ecology 31, 732–735.Valentine, L.E., Roberts, B., Schwarzkopf, L., 2007. Mechanisms driving avoidance of

non-native plants by lizards. Journal of Applied Ecology 44, 228–237.Vitt, L.J., Avila-Pires, T.C.S., Caldwell, J.P., Oliveira, V.R.L., 1998. The impact of

individual tree harvesting on thermal environments of lizards in Amazonianrain forest. Conservation Biology 12, 654–664.

Vonesh, J.R., 2001. Patterns of richness and abundance in a tropical African leaf-litter reptiles. Biotropica 33, 502–510.

Wagner, R.G., Newton, M., Cole, E.C., Miller, J.H., Shiver, B.D., 2004. The role ofherbicides for enhancing forest productivity and conserving land for biodiver-sity in North America. Wildlife Society Bulletin 32, 1028–1041.

Webb, J.K., Shine, R., 2008. Differential effects of an intense wildfire on survival ofsympatric snakes. Journal of Wildlife Management 72, 1394–1398.

Wilson, S., Swan, G., 2003. Reptiles of Australia. Princeton University Press, Oxford.Woods, R.V., 1990. 2nd rotation decline in Pinus radiata plantations in South

Australia has been corrected. Water Air and Soil Pollution 54, 607–619.Yirdaw, E., Luukkanen, O., 2004. Photosynthetically active radiation transmittance

of forest plantation canopies in the Ethopian highlands. Forest Ecology andManagement 188, 17–24.

Zanne, A.E., Chapman, C.A., 2001. Expediting reforestation in tropical grasslands:distance and isolation from seed sources in plantations. Ecological Applications11, 1610–1621.