Diversity and composition of Amazonian moths in primary, secondary and plantation forests

20 Pages·2009·0.36 MB·English

CORE Metadata, citation and similar papers at core.ac.uk
Provided by Lancaster E-Prints
JournalofTropicalEcology(2009)25:281–300.Copyright©2009CambridgeUniversityPress
doi:10.1017/S0266467409006038 PrintedintheUnitedKingdom
Diversity and composition of Amazonian moths in primary, secondary
and plantation forests
JosephHawes∗1,CatarinadaSilvaMotta†,WilliamL.Overal‡,JosBarlow§,
∗
TobyA.Gardner#andCarlosA.Peres
∗SchoolofEnvironmentalSciences,UniversityofEastAnglia,NorwichNR47TJ,UK
†DepartamentodeEntomologia,InstitutoNacionaldePesquisasdaAmazoˆnia(INPA),CaixaPostal478,Manaus,AM69011-970,Brazil
‡DepartamentodeEntomologia,MuseuParaenseEm´ılioGoeldi(MPEG),Av.Perimetral1901,BairroTerraFirme,Bele´m,PA66077-530,Brazil
§LancasterEnvironmentCentre,LancasterUniversity,LancasterLA14YW,UK
#DepartamentodeBiologia,UniversidadeFederaldeLavras,Lavras,MinasGerais37200-000,Brazil
(Accepted19February2009)
Abstract: Theresponseoftropicalfaunatolandscape-levelhabitatchangeispoorlyunderstood.Increasedconversion
ofnativeprimaryforesttoalternativeland-uses,includingsecondaryforestandexotictreeplantations,highlights
theimportanceofassessingdiversitypatternswithintheseforesttypes.Wesampled1848mothsfrom335speciesof
Arctiidae,SaturniidaeandSphingidae,overatotalof30trap-nights.Samplingwasconductedduringthewetseason
2005,usingthreelight-trapsat15siteswithinareasofprimaryforest,secondaryforestandEucalyptusurograndis
plantationsinnorthernBrazilianAmazonia.TheJaristudyregionprovidesoneofthebestopportunitiestoinvestigate
theecologicalconsequencesofland-usechange,andthisstudyisoneofthefirsttoexaminepatternsofdiversityfora
neotropicalmothassemblageinahuman-dominatedlandscapeinlowlandAmazonia.Wefoundthatthethreemoth
familiesrespondedconsistentlytodisturbanceintermsofabundanceandcommunitystructurebutvariablyinterms
ofspeciesrichness,inamannerapparentlysupportingalife-historyhypothesis.Ourresultssuggestthatsecondary
forestsandEucalyptusplantationscansupportasubstantiallevelofmothdiversitybutalsoshowthattheseforesttypes
holdassemblageswithsignificantlydistinctcommunitystructuresandcompositionfromprimaryforest.Inaddition,
theabilityoftheseconvertedland-usestosupportprimaryforestspeciesmaybeenhancedbyproximitytosurrounding
primaryforest,anissuewhichrequiresconsiderationwhenassessingthediversityandcompositionofmobiletaxain
human-dominatedlandscapes.
KeyWords:Arctiidae,Brazil,human-dominatedlandscapes,land-usechange,Lepidoptera,Saturnidae,Sphingidae
INTRODUCTION attention is now expanding to a wider range of taxa
(Barlow et al. 2007, Gardner et al. 2008). Insects make
The ecological consequences of land-use change vary
an enormous contribution to both tropical diversity
considerablyfordifferenttaxa,asparticularspeciestraits
(Lewinsohn et al. 2005) and ecosystem functioning
interactdifferentlywiththedisturbedenvironment(Daily
(Wilson1987),andmothsareoneofthegroupsplayinga
2001, Koh et al. 2004). A clear understanding of these
central role in numerous ecosystem processes as prey,
consequences is currently lacking for both secondary
herbivores and pollinators (Barlow & Woiwod 1989,
(Brook et al. 2006, Gardner et al. 2007) and plantation
Janzen1987).
forests (Hartley 2002, Lindenmayer & Hobbs 2004),
Relatively few studies of tropical moth faunas have
whicharebothincreasinginimportancewithintropical
been conducted in the neotropics (Brehm et al. 2003,
forest landscapes (Evans & Turnbull 2004, Neeff et al.
Hilt et al. 2006, Ricketts et al. 2001) despite higher
2006).
species richness in this region than elsewhere in the
Whilemoststudiestodatehavefocusedoncharismatic
tropics(Hilt&Fiedler2005).Thereisalsoacurrentlack
indicatororflagshipgroupssuchasbirdsandmammals,
of understanding regarding the relative importance of
local forest type versus landscape features of the wider
∗Correspondingauthor.Email:[email protected] countrysideindeterminingpatternsofdiversity(butsee
282 JOSEPHHAWESETAL.
Figure1.MapoftheJarilandholdinginthenorthernBrazilianAmazonandlocationsofthe15siteswithinareasofprimaryforest,secondaryforest
andEucalyptusurograndisplantationswheremothsamplingwascarriedoutbetweenAprilandMay2005.Labelsrefertotheindividualsiteslisted
inTable1.
Ricketts et al. 2001). This study is one of the first to METHODS
examine the diversity patterns of lowland Amazonian
moth assemblages in a human-dominated landscape Studysites
includingprimary,secondaryandplantationforests.
We sampled three families differing in their ecology Sampling was conducted in the 1.7-Mha landholding
and life-histories (Hilt & Fiedler 2006, Janzen 1984): of Jar´ı Celulose S.A., located on the border between
thelarge-bodiedemperormoths(Saturniidae)andhawk the states of Amapa´ and Para´ in northern Brazilian
moths (Sphingidae), as well as the smaller-bodied tiger Amazonia (0◦53(cid:4)S, 52◦36(cid:4)W; Figure 1). The area was
moths (Arctiidae). By assessing changes in abundance, purchasedin1967,andabout10%ofthelandconverted
species richness and community composition, this to exotic tree plantations. Current stands consist of
study examined the various effects of landscape-level Eucalyptusurograndis,whileearlierplantationsofGmelina
disturbanceonmothassemblages.Wetestedtheapriori arborea and Pinus caribaea have mostly been cleared
hypotheses that (1) disturbed forest types (secondary and abandoned. This process has resulted in a complex
forests and Eucalyptus plantations) support distinct and landscape mosaic of Eucalyptus plantations, with large
lessspecies-richmothcommunitiesthanprimaryforest, tracts of regenerating secondary forest and relatively
and(2)thatresponsestodisturbancevarybetweenmoth undisturbed primary forest. We sampled the moth
familiesasaresultofdifferencesinlife-historystrategies community at 15 sites (Figure 1; Table 1), comprising
e.g.mobile,long-livedtaxaarepredictedtopersistbetter five of each of the three forest types: (1) undisturbed
indisturbedforesttypes. primary forest; (2) even-aged native secondary forest
Amazonianmothsinsecondaryandplantationforests 283
Table1.Selecteddetailsofthe15light-trappingsiteswithinprimary(PF)andsecondaryforests(SF),andEucalyptusplantations(EUC)inthe
Jarilandscape.PFsitesarepartoflargecontiguoustractsofrelativelyundisturbedforest.Sitenames:B=Bituba,C=Castanhal,E=Estac¸ao,
P=Pacanari,Q=Quaruba.Light-trapradius=distanceatwhichlight-trapisvisibletothehumaneye,PFin3-km-radiusbuffer=proportionof
a3-km-radiusareaaroundeachsamplingsitethatcontainsPF.Significancewascalculatedwithone-wayANOVAs;F=F-ratio,∗∗=P<0.01,
superscriptlettersdenoteTukey’sHSDsubsets.
Treebasal Mean Mean Mean PFin3-
area understorey canopy light-trap km-radius
Habitattype Sitename Area(ha) Age(y) Altitude(m) (m2ha−1) density(%) cover(%) radius(m) buffer(%)
PF B - - 250–275 31.8 57.0 90.9 72.5 99.8
PF C - - 95 34.0 61.7 94.5 78.0 75.4
PF E - - 90 30.5 54.7 95.2 103.3 58.9
PF P - - 165 28.0 67.0 88.2 68.0 52.8
PF Q - - 100 25.7 66.0 93.2 88.7 55.0
PF Mean - - 142.5 30.0a 61.3b 92.4a 82.1b 68.4a
SF 55 2.9 15 20 7.1 74.0 88.5 44.2 43.9
SF 56 3.2 20 70 9.9 86.3 83.8 65.2 2.5
SF 75 3.0 16 70 21.5 86.0 93.2 66.8 23.7
SF 86 3.7 18 41 19.2 90.7 94.3 38.3 24.0
SF 91 1.1 14 147 7.7 69.7 93.9 48.0 42.2
SF Mean 2.78 16.6 69.6 13.1b 81.3a 90.7a 52.5b 27.3b
EUC 10 1.6 3.9 106 9.0 42.3 61.9 227.5 27.9
EUC 14 1.3 3.7 131 8.8 24.3 63.7 268.0 7.8
EUC 52 4.1 5.2 97 11.4 27.0 71.1 164.2 12.4
EUC 95 0.6 3.9 139 17.7 45.7 75.9 155.0 61.2
EUC 127 1.3 5.1 220 16.7 38.0 75.4 153.3 47.0
EUC Mean 1.78 4.36 138.6 12.7b 35.5c 69.6b 193.6a 31.3b
F 19.7∗∗ 40.0∗∗ 34.1∗∗ 27.4∗∗ 6.5∗∗
(14–20 y since abandonment); and (3) 4–5-y-old Theeffective‘radiusofattraction’oflight-traps(Beck
Eucalyptus plantation stands. Sites were selected to &Linsenmair2006),isdependentuponsexandspecies
minimize age differences within each forest type, and identity(Baker&Sadovy1978)aswellaslightintensity
to maximize their area (mean size of Eucalyptus and andwavelength(Muirhead-Thomson1991).Whilemost
secondary forest blocks was 1687 ha (range = 574– evidence suggests an attraction radius of 50–200 m
3910 ha) and 2682 ha (range = 1079–3508 ha), (Ricketts et al. 2001), Baker & Sadovy (1978) report
respectively) and spatial independence (mean distance distances up to 500 m. To minimize the capture of
between primary, secondary and Eucalyptus sites was vagrants from surrounding forest types, all three light-
30km(range=14–67km),9km(range=4–44km)and trapswerelocatedinthecentreofeachforestpatchandat
11km(7–50km),respectively).Samplingwasconducted least500mfromanyedgewithneighbouringforesttypes.
between 1 April and 18 May 2005, during the wet Theradialextentoflightdiffusionthrougheachstandwas
season (January–June). Average annual rainfall at Jari alsoestimatedforeachtrapbymeasuringthemaximum
is2115mm,andthemeandailyairtemperatureis26◦C distanceintwodiametricallyoppositedirectionsforwhich
(Coutinho&Pires1996). anylightcouldbedetectedbyeye.
Traps were operated from 18h30 to 06h30, and
checked simultaneously every hour by JEH and two
Mothsampling trainedassistants.Arctiidae,SaturniidaeandSphingidae
moths were collected manually from both sides of the
To sample moths we used a 2 × 2-m sheet trap design sheets and the immediately surrounding areas, using a
(Cheyetal.1997),whichhastheadvantagesovermore killing bottle charged with ethyl acetate. The 15 sites
standard light-traps (Intachat & Woiwod 1999) of a weresampledtwiceeach,withonerepeatineachoftwo
selective catch, reduced damage to trapped individuals, rotations (mean interval between consecutive samples
and portability (Axmacher & Fiedler 2004). We placed at the same site ± SD = 27.6 ± 9.9 d, N = 15 sites),
astandardizedsetofthreelight-trapsat200-mintervals resultinginatotalsamplingeffortof30trap-nights(two
alongthelinetransectsofeachsite.A12-WUVblacklight trap-nightsor24trap-hourspersite).Thesamplingorder
tube was used at the central light-trap and a 160-W of sites within each rotation was controlled to account
mercury-vapour light bulb at the two outermost traps. for the fraction of the moon illuminated (http://aa.
For an overview of light-trapping feasibility see Beck & usno.navy.mil/data/docs/MoonPhase.html)andtoavoid
Linsenmair(2006). anysystematicbiasfromthewell-documentedinfluences
284 JOSEPHHAWESETAL.
of weather on light-trap captures (Fry & Waring 2001, semi-supervisedclassificationofa2003Landsat7(30-m
Spalding&Parsons2004,Yela&Holyoak1997).Catches pixel)image.Bufferringswerecreatedaroundthecentral
werealsorestrictedtoperiodswithoutstrongmoonlight point of each moth sampling site before performing an
by avoiding nine nights around the full moon (Yela & intersect overlay with layers containing data on land-
Holyoak 1997), and weather conditions were recorded cover types. Three kilometres was selected as the buffer
every hour during sampling. Bulbs were protected radius as this exceeds the expected attraction of light-
from above but any broken by rain were immediately traps (Baker & Sadovy 1978), yet falls within the flight
replaced. capacity range of large-bodied moths (I. Kitching pers.
Collectedmothswereoven-driedandidentifiedatthe comm.).
Entomology Department of the Instituto Nacional de
PesquisasdaAmazoˆnia(INPA)inManaususingtheINPA
referencecollectionandavailableguides(d’Abrera1995, Statisticalanalyses
1998; Kitching & Cadiou 2000, Lemaire 1988, Pin˜as-
Rubio&Pesa´ntez2000,Pin˜as-Rubioetal.2000,http:// Total moth abundances per trap-night were compared
www.inra.fr/Internet/Produits/PAPILLON/arct_guy/ar between each forest type using one-way ANOVAs
ct_guy.htm). Morphospecies were identified by INPA with Tukey’s post-hoc test. To highlight the variable
staffusinganatomicalfeaturesandwingpatterns,with effectiveness of light-traps in different vegetation types
caretakentominimizeover-splittingasaresultofsexual (e.g.thegreaterlightpenetrationdistanceinEucalyptus
dimorphismornaturalvariation.Specimensofallspecies plantations compared with the dense undergrowth of
andmorphospeciesweresubsequentlydepositedatINPA. secondary forests) we repeated these analyses with
abundancepertrap-nightdividedbytheareaeffectively
surveyed by each light-trap (calculated from our trap-
Vegetationsampling radiusmeasurements).
There is currently no general consensus on an
Trees and woody lianas were sampled along the same optimal method to deal with the difficult problem of
transect lines in each of the 15 sites. We measured all variableattractionradiiforlight-trapsindifferentforest
standingtrees≥10cmindiameteratbreastheight(dbh) types (Beck & Linsenmair 2006). Human perception of
andlianas≥5cmina10×1000-mplotestablishedat trap-attraction differs from moths so ideally UV light
eachofthe10primaryandsecondaryforestsites.Basal penetrationwouldbemeasured,Mark-Release-Recapture
area in plantations was estimated from 23 10-m radial experiments would be performed and standardization
plotspersite(7226m2)andconvertedtobasalareaper wouldalsoaccountforthethree-dimensionalcatchment
hectare.Densityofsaplings(tallerthan1mand<10cm of each trap. However, we feel that the simple
dbh), and lianas (< 5 cm dbh) were determined by standardization we performed indicates the possible
recordingallstemswithinthree2.5×2.5-m(6.25m2) impacts of variable vegetation density and emphasizes
subplots placed at 23 locations every 50 m along each that crude abundances should be interpreted with
transect(totalof69subplotspersite). caution.
Canopycoverandunderstoreydensityweremeasured Forassessingspeciesrichnessandalphadiversity,the
at each light trap location following the methodology raw catch data were pooled from the three individual
of Barlow et al. (2002). A reading with a spherical trapsfrombothnightsatagivensite,assamplesizeswere
densiometer (Lemmon 1957) was taken in each of the insufficient to compare the relative attractions of lamps
fourcompassdirectionsandaveragedbeforeconverting with different spectral emissions. The observed number
toapercentagecanopycover.Similarly,a2.5-mgraded of species per site gives a poor and often misleading
pole was used to estimate understorey density. In each indication of total richness because of the virtual
of the four compass directions, the number of 10-cm impossibilityofobtainingacompleteinventoryofspecies-
sections visible from a distance of 15 m, were recorded richtropicalinvertebratecommunities(Priceetal.1995).
andconvertedtoapercentagedensity. Moresuitableestimatesaregivenbytheextrapolationof
speciesaccumulationcurvesortheshapeofthespecies-
abundance distribution (Magurran 2004). Rarefaction
Land-coveranalysis and non-parametric estimators also provide powerful
approachestoestimatespeciesrichness(Gotelli&Colwell
Ageographicinformationsystem(GIS)wasemployedto 2001). Sample-based rarefaction curves were therefore
measure the relative extent of different forest types in produced for the three forest types, and an average
the immediate surroundings of any given site. A land- of three abundance based estimators (Chao1, Jack1
cover classification was developed from a combination and ACE) was calculated for each site using EstimateS
of land-use data provided by Jari Celulose S.A and a 7.5 (R. K. Colwell, http://purl.oclc.org/estimates).
Amazonianmothsinsecondaryandplantationforests 285
Species-abundance relationships were examined using primary forests or Eucalyptus plantations. With the
standardized Whittaker plots. To assess alpha diversity overall abundance of trap-night samples standardized
we calculated Fisher’s alpha of the logarithmic series in relation to the light-trap areas there were still
distribution(Fisheretal.1943),whichhasbeenwidely significantdifferencesamongforesttypes(Figure2;F
2,87
used in tropical moth diversity studies and is relatively =8.6,P<0.001)butthisstandardizedcatchwaslower
independentofsamplesize(Magurran2004). in Eucalyptus plantations than in either primary and
Patterns of community structure and composition secondaryforests.Again,thispatternwassimilarforeach
among different sites and between forest types were familyseparately,althoughdifferencesamongforesttypes
visualized using non-metric multidimensional scaling werenotasclearfortheSphingidae.
(NMDS) of a similarity matrix based on the Bray– Rarefaction curves for all three families combined
Curtisindex(standardizedandsquareroot-transformed). showed slightly higher levels of species richness in
The analyses were performed on both abundance secondary forest and Eucalyptus plantations than in
(quantitative) and presence/absence (qualitative) data. primaryforest,althoughthedifferencesbetweenallthree
Abundance data reveal patterns based primarily on the foresttypeswereonlymarginallysignificant(Figure3).
common species (i.e. community structure), whereas This pattern was driven mainly by the pattern within
presence/absence data give more weight to the Arctiidae, while Saturniidae showed a lower species
distributionofrarespecies(i.e.communitycomposition). richness in Eucalyptus plantations than primary and
Differences between forest types were assessed using secondary forest. There were no differences between
an analysis of similarities (ANOSIM), and the identity foresttypesforSphingidae,wherethesamplesizeswere
of species contributing most to any differences was smallest.
determined using an analysis of percentage similarities Comparingobservedspeciesrichnesswiththeaverage
(SIMPER)(Clarke&Warwick2001). richness estimate shows that a complete inventory was
Theinfluenceoftheforesttypesurroundingstudysites not achieved for any family in any of the three forest
was investigated using a Spearman’s correlation of the types (Table 2). Values for Fisher’s alpha did not differ
amount of primary forest within a 3-km buffer against significantly between forest types for the three families
rarefied species richness. The effects of surrounding combinedorwithineachfamily(Table2).Nodifference
primary forest plus other environmental parameters between forest types was evident in the proportions of
(lunar phase, weather conditions and forest structure), local singletons captured at a site for the three families
as well as the geographic distance between sites on combined(F =2.2,P=0.152).
2,12
community structure were assessed using the BIOENV
and RELATE (analogous to a Mantel test) functions
respectively.Communityanalyseswereconductedusing Communitystructureandcomposition
Primer5(PRIMER-ELtd.,Plymouth,UK).
Levels of dominance were similar in the three forest
types with long tails of rare species in each case.
RESULTS When examining only the 25 most abundant species
in each forest type (Figure 4), secondary forests appear
Mothabundanceandspeciesrichness to exhibit the lowest levels of dominance, with most of
the dominance in primary forest accounted for by just
Atotalof1848mothindividualswerecollectedduringthe one species, Evius albicoxae (108 individuals from 677
30trap-nightsat15sites.Thesecomprised974Arctiidae primary forest captures). However, few of the 25 most
individualsof231species,772Saturniidaeindividualsof abundant species in primary forest were also the most
65species,and102Sphingidaeindividualsof39species. abundant species in secondary forests or Eucalyptus
Ofthese,452Arctiidaeand11Saturniidaewereidentified plantations, indicating a high level of community
to160andsixmorphospeciesrespectively.Samplesizes turnover.
in this short-term study were therefore relatively small, This is supported by the two-dimensional NMDS
particularlyfortheSphingidae.Forafullspecieslistsee ordination plot based on abundance data for the three
Appendix1(supplementarymaterial). families combined, which shows a distinct grouping of
Total abundance of the three moth families ranged sitesintothethreeforesttypes(Figure5;GlobalANOSIM:
from 42 to 264 individuals per site. The abundance R=0.75,P=0.001)andsignificantdifferencesbetween
patterns for each family closely resemble the pattern each pair of forest types (Pairwise ANOSIM between
forthethreefamiliescombinedwheremeanabundance primary (PF), secondary forest (SF) and Eucalyptus
per trap-night differed significantly among forest types plantation (EUC): PF-SF (i.e. between primary and sec-
(Figure 2; ANOVA: F =5.8, P=0.004), with fewer ondaryforests):R=0.64,P=0.008;PF-EUC:R=0.96,
2,87
moths captured in secondary forests than in either P=0.008; SF-EUC: R=0.58, P=0.008). An NMDS
286 JOSEPHHAWESETAL.
Figure2.Abundance(mean±SE)ofArctiidae,Saturniidae,Sphingidaeandallfamiliescombined,capturedinprimary(PF)andsecondaryforests
(SF),andEucalyptusplantations(EUC)pertrap-night(a)andstandardizedbytrap-areasurveyedpertrap-night(b).LettersabovebarsdenoteTukey’s
HSDsubsets.
plot based on presence/absence data (i.e. examining EUC: R=0.51, P=0.008) and Saturniidae (PF-SF:
community composition) shows the same grouping R=0.51, P=0.016; PF-EUC: R=0.88, P=0.008; SF-
into forest types (Global ANOSIM: R=0.64, P=0.001; EUC: R=0.51, P=0.016), but for Sphingidae only
Pairwise ANOSIM: PF-SF: R=0.60, P=0.008; PF- primary forest and Eucalyptus plantation held distinct
EUC:R=0.84,P=0.008;SF-EUC:R=0.44,P=0.008). communities (PF-SF: R=0.10, P=0.206; PF-EUC:
Examining community structure for the three families R=0.54,P=0.008;SF-EUC:R=0.13,P=0.206).
separately (Figure 5) shows significant differences Thesamplingsessionhadasignificantinfluenceonthe
between all forest types for both Arctiidae (PF-SF: overall pattern of community structure (ANOSIM: R =
R=0.32, P=0.032; PF-EUC: R=0.82, P=0.008; SF- 0.24,P=0.004)butsignificantdifferencesamongforest
Amazonianmothsinsecondaryandplantationforests 287
Figure3.Sample-basedrarefactioncurvesforArctiidae,Saturniidae,Sphingidaeandallfamiliescombined,sampledinprimary(PF)andsecondary
forests(SF),andEucalyptusplantations(EUC).X-axisrescaledtoshowmeannumberofindividualspersample.Dottedlinesshow95%confidence
intervals.
types remained consistent when analysing the data for Responsestoforesttypeandlandscapestructure
eachsessionindependently.Thiswasalsoconfirmedby
a two-way crossed analysis of similarities that shows a The BIO-ENV analysis identified canopy cover as the
difference in community structure between forest types bestsinglepredictorofmothcommunitystructurewhen
whilstaveragingacrosssamplingrounds(ANOSIM:R= examining all forest types combined (Table 4). Basal
0.49,P=0.001). area of lianas and basal area of live trees were the best
Beta diversity (Bray–Curtis dissimilarity) values predictorsinprimaryandsecondaryforestsrespectively,
were highest between primary forest and Eucalyptus whereas the moth community structure in plantation
plantations (84%) but all forest types supported largely sites was strongly correlated to the attraction radius
unique communities (Figure 6), with five species of the light-traps, which was significantly negatively
contributing 9–11% of the total dissimilarity in each correlated with both understorey density (r = –0.862,
pairwiseforestcomparison(Table3).Primaryforestsites P<0.001) and canopy cover (r = –0.921, P<0.001).
weredistinctlargelydueto(1)anabundanceofspecies The amount of primary forest within 3 km was not
rarelyfoundindisturbedforesttypes(Eviusalbicoxaeand identified as a strong predictor in the BIO-ENV analysis
Adeloneivia subangulata) and (2) lower relative densities andgeographicdistancebetweensiteswasalsonotrelated
or absence of species common in secondary forests topatternsofcommunitydissimilarity(RELATE:Rho=–
(e.g. Dirphia tarquina and Periga cynira) and Eucalyptus 0.09, P=0.738). There was no difference in the lunar
plantations (e.g. Automeris liberia and Idalus admirabilis) phase across forest types (mean fraction of the moon
(Table2). visiblepernight;F =0.564,P=0.583).
2,30
288 JOSEPHHAWESETAL.
Table2.DetailedresultsofthesamplingforArctiidae,SaturniidaeandSphingidaeatthe15light-trappingsiteswithinprimary(PF)andsecondary
forests(SF),andEucalyptusplantations(EUC).Sitenames:B=Bituba,C=Castanhal,E=Estac¸ao,P=Pacanari,Q=Quaruba.N=numberof
individualscaptured,Sobs=observedspeciesrichness,Richnessestimate=averageofthreeabundance-basedrichnessestimators(Chao1,Jack
1andACE).Significancewascalculatedwithone-wayANOVAs,F=F-ratio.
Arctiidae Saturniidae Sphingidae
Habitat Richness Richness Richness
type Sitename N Sobs estimate Fisher’sα N Sobs estimate Fisher’sα N Sobs estimate Fisher’sα
PF B 56 19 47.9 10.1 68 14 17.1 5.4 1 1 1.3 -
PF C 37 25 72.7 33.8 40 20 37.1 15.9 3 2 3.1 2.6
PF E 78 48 210.2 53.1 44 18 26.1 11.4 2 2 3.2 -
PF P 149 37 109.2 15.8 115 27 36.3 11.1 0 0 0.0 -
PF Q 28 13 44.4 9.4 54 17 28.4 8.5 1 1 1.3 -
PF Mean 69.6 28.4 96.9 24.5 64.2 19.2 29.0 10.5 1.4 1.2 1.8 2.6
SF 55 58 36 196.7 40.5 48 17 24.3 9.4 2 2 3.2 -
SF 56 43 32 85.6 56.7 49 19 54.8 11.4 1 1 1.3 -
SF 75 35 18 52.8 14.9 45 21 36.9 15.3 5 5 13.1 -
SF 86 17 15 49.5 61.2 23 10 12.1 6.7 2 1 1.3 0.8
SF 91 33 22 37.8 28.8 21 11 16.7 9.3 2 2 3.2 -
SF Mean 37.2 24.6 84.5 40.4 37.2 15.6 28.9 10.4 2.4 2.2 4.4 0.8
EUC 10 70 22 43.4 11.0 27 4 4.5 1.3 3 1 1.0 0.5
EUC 14 80 33 62.7 21.0 29 8 13.9 3.7 9 5 7.9 4.6
EUC 52 54 24 55.9 16.6 86 9 14.5 2.5 1 1 1.3 -
EUC 95 118 67 149.0 64.3 48 22 52.5 15.7 56 24 46.2 15.9
EUC 127 118 52 232.7 35.5 74 19 29.5 8.3 14 8 13.5 7.8
EUC Mean 88 39.6 108.7 29.7 52.8 12.4 23.0 6.3 16.6 7.8 14.0 7.2
F 2.9 1.4 0.1 0.8 1.5 1.6 0.3 1.4 2.1 2.0 1.6 -
Themaindifferenceinland-coverwithin3kmofsample diversity within either secondary forests or Eucalyptus
sitesinsecondaryforestandEucalyptusplantationswas plantations.Nevertheless,theareassurroundingthetwo
the proportion of primary forest, varying from 2.5% to Eucalyptus sites containing the highest levels of alpha
61.2%.However,therewerenosignificantrelationships diversity(Areas95and127)hadthehighestproportion
between the amount of primary forest and moth alpha ofprimaryforest(Table1).
Table3.Breakdownofaveragecommunitydissimilarity(diss.)betweenprimary(PF)andsecondaryforests(SF),
andEucalyptusplantations(EUC)intopercentagecontributionsfromeachspecies.Thefivespecieswiththegreatest
contributionareshownforeachhabitat-paircomparisoninorderofdecreasingcontribution.Arc=Arctiidae,
Sat=Saturniidae,Diss./SD=theratioofdissimilaritytothestandarddeviationofdissimilarity:highervalues
indicateamoreconsistentcontributiontodifferencesbetweencommunities.
Contribution
Species Family Averageabundance Averagediss. Diss./SD todiss.(%)
PF-EUCAveragediss.=84.2 PF EUC
Automerisliberia Sat 0.0 12.4 2.3 1.8 2.8
Idalusadmirabilis Arc 0.2 9.0 1.9 2.7 2.3
Adeloneivaiasubangulata Sat 7.4 0.0 1.9 1.6 2.3
Eviusalbicoxae Arc 21.6 3.6 1.7 1.4 2.1
Periphobaarcaei Sat 0.6 7.6 1.6 1.3 1.9
PF-SFAveragediss.=79.2 PF SF
Eviusalbicoxae Arc 21.6 1.8 1.7 1.4 2.1
Adeloneivaiasubangulata Sat 7.2 2.4 1.6 1.4 2.0
Dirphiatarquinia Sat 0.2 3.6 1.4 1.6 1.8
Perigacynira Sat 1.2 3.8 1.3 1.5 1.7
Automerismidea Sat 6.8 0.2 1.3 1.4 1.6
SF-EUCAveragediss.=79.0 SF EUC
Perigacynira Sat 3.8 0.6 1.8 1.8 2.3
Idalusadmirabilis Arc 3.0 9.0 1.7 2.5 2.2
Automerisliberia Sat 2.2 12.4 1.6 1.2 2.0
Eaclespenelope Sat 1.2 14.8 1.6 1.2 2.0
Dirphiatarquinia Sat 3.6 0.0 1.5 1.6 1.9
Amazonianmothsinsecondaryandplantationforests 289
Figure4.Whittakerplotsformothspeciessampledinprimaryforest(a),secondaryforest(b)andEucalyptusplantation(c).Steeperplotsindicate
higherdominance.Openbarsin(b)and(c)representspeciessharedwithprimaryforest,whicharelabelledwithlettercodescorrespondingtothose
in(a)andthespecieslistinAppendix1.
Table4.ResultsofBIO-ENVanalysisshowingtherelationshipsofmoth DISCUSSION
communitystructureandweather,habitatandlandscapevariablesin
primary(PF)andsecondaryforests(SF),Eucalyptusplantations(EUC), This is the first study to quantitatively evaluate
and all habitats combined. CC = canopy cover, UD = understorey the diversity patterns and community structure of
density,LR=light-trapradius,BAL=basalareaoflianas≥5cmdbh,
moth assemblages in a human-dominated landscape
BAT=basalareaoflivetrees≥10cmdbh,S=numberofsaplings
<10cmdbh,L=numberoflianas<5cmdbh,WPCA1=weather in lowland Amazonia. Light-trapping for nocturnal
PCA1score(degreeofcloudcover).Othervariablesanalysed:moon Arctiidae, Saturniidae and Sphingidae in the Jari
phase,proportionofprimaryforestina3-kmradius.ρW=weighted landscape of Brazilian Amazonia revealed three major
Spearman’s. patterns: (1) undisturbed primary forests were not
Best 2ndbest Best distinctly richer or more diverse than secondary forests
Habitat variable ρW variable ρW combination ρW
orEucalyptusplantations,buteachforesttypeexhibited
PF BAL 0.64 WPCA1 0.31 BAL 0.64 a distinct community in terms of both structure and
SF BAT 0.71 CC 0.42 CC,BAT 0.84
composition; (2) species turnover was highest between
EUC LR 0.88 S 0.67 CC,S,L 0.95
primaryforestsandEucalyptusplantations(highestbeta
All CC 0.40 LR 0.38 LR,UD,S,L 0.56
diversity),withsecondaryforestsitesbeingintermediate;
290 JOSEPHHAWESETAL.
Figure5.BioticNMDSordinationfromspeciesabundancesofArctiidae,Saturniidae,Sphingidaeandallfamiliescombined,from15sitesinprimary
(PF)andsecondaryforest(SF),andEucalyptusplantation(EUC).
and(3)thethreemothfamiliesvariedintheirresponse notbesensitivetodisturbance(Schulze&Fiedler2003,cf.
to disturbance in terms of species richness but changes Becketal.2006).TheArctiidaesampledcontainveryfew
inabundanceandcommunitystructurewererelatively fromsubfamiliesCtenuchinaeorLithosiinaebuttheseare
consistent. likelytobeincludedinthelargenumberidentifiedonlyto
morphospecies.
The lower overall abundance of all three families
Abundanceandspeciesrichnessinprimary,secondaryand in secondary forests could be largely attributed to the
plantationforests dilution effect on the effectiveness of light-traps as a
resultofdenservegetation(Hilt&Fiedler2005,Ricketts
TotalcapturesofSphingidaeweremuchlowerthanthose et al. 2001). In contrast, light-traps appear especially
intheotherfamiliessampledinthisstudy(Arctiidaeand effective in the relatively open Eucalyptus plantations
Saturniidae),andtheycontributedlittletoanyobserved in attracting moths from a larger surrounding area
community differences between forest types. However, and,whenstandardizedbytrap-area,abundanceswere
Sphingidae are not expected to be common in the dark actually much lower in Eucalyptus plantations than
primaryforestunderstorey(Schulzeetal.2001)andmay secondaryorprimaryforests.