Ecological characterization of carabid species (Coleoptera, Carabidae) in the Netherlands from thirty years of pitfall sampling

26 Pages·1991·13.2 MB·English

by  H Turin

HH.EITJUERRIMNA,NK,.WA.LDLEARASN,EP.&J.ED.EPNENBOTEERR,MSA. NVAN ESSEN,TH.
Dutch CarabidologicalAssociation
ECOLOGICAL CHARACTERIZATION OF CARABID
SPECIES (COLEOPTERA, CARABIDAE) IN THE
NETHERLANDS FROM THIRTY YEARS OF
PITFALL SAMPLING
Turin, H., K. Alders, P. J. den Boer, S. van Essen, Th. Heijerman, W. Laane & E.
Penterman, 1991. Ecologicalcharacterizationofcarabidspecies (Coleoptera,Carabidae)
intheNetherlandsfromthirtyyearsofpitfallsampling.-TijdschriftvoorEntomologie
134: 279-304,figs. 1-20,tabs. 1-10. [isSN0040-7496]. Published 18December 1991.
An ecological characterization of 285 Dutch carabid species based on a large data set
obtainedbypitfalltrappinginvarioushabitatsalloverTheNetherlandsispresented.The
data set contains ca. 1.5 million specimens, collected during thirty years in 1616 year-
samples from 862 localities. Information about habitat, viz. typeofvegetation, type of
soilandhumiditywas recordedforeachofthe862 samplinglocalities.
TwoWay IndicatorSpecies Analysis (twinspan) as wellas DetrendedCorrespondence
Analysis(decorana)wereusedtoclassifythespeciesintohabitatgroups.Sevenprincipal
habitatgroups couldbe recognized. Indices forecologicalamplitude,preferenceforsoil
type and humidity, are formulated and estimated for each species. This resulted in a
description ofthe habitat preferences of the Dutch carabidspecies in fargreaterdetail
thanwaspossiblebefore.Theclassificationiscomparedwithothersfromtheliterature.
Correspondence: H.Turin,Esdoorndreef29,6871 LKRenkum.TheNetherlands.
Keywords. -Ecology,Carabidae,Twinspan,Decorana,pitfall
In 1945, Lindroth published the first volume of land,bypresentinglistsofcharacteristicspeciesof
his majorwork 'Die FennoskandischenCarabidae' these habitat categories. However, his tables only
(Lindroth 1945). In this work he described the show the most abundant species in thatparticular
ecological preferences of the Fennoscandian ca- habitatgroupgiving no information abouttheoc-
rabid species. Since pitfall trapping was not prac- currencesofthese'typical'speciesinothertypesof
ticed until about 1950, Lindroth's ecological char- habitat.
acterizations were based on his own hand- Luffet al. (1989) were the first to classify hab-
collections,supplementedby thoseofothercollec- itats of ground beetles based on a large data set
tors,on laboratoryexperiments andondata taken from nearly 250sites inNorth-EastEngland. The
from literature. This resulted in detailed descrip- carabid samples were grouped, based on pres-
tionsofthespecies'ecologicalpreferences.In1949, ence/absence data, using the Two Way Indicator
Lindroth classified the species into a number of SpeciesAnalysis (twinspan (Hill 1979a) andordi-
ecologicalgroups,recognizingxerophilousspecies, nated by Detrended Correspondence Analysis
mesophilous species (including ubiquists), hygro- (DECORANA (Hill 1979b). Ten primary groups of
philousspecies,arborealspecies,forestspeciesand carabid habitats were recognized, viz.: coastal, up-
synanthropous species (Lindroth 1949). land (dry, wet), woodland, grassland (dry, wet),
From about 1950 onwards, pitfall trapping be- riverside(boulders,shingle,sand) andmarsh.Eyre
cameacommonlyusedtechniqueinEurope.Thiele &Luff(1990a) madeapreliminaryclassificationof
(1977) gave a summaryofmanyecologicalstudies European grassland habitats using carabids. The
basedonpitfall trappingcarriedoutuntilca. 1975. carabid assemblages of 363 pitfalled grasslands
On basis of the results of these studies he charac- throughoutBritain,weredescribedbyEyre& Luff
terized the carabid faunas ofcertain types of hab- (1990b)
itat, such as forests, sandy habitats and cultivated The characterizations of Lindroth (1945, 1949)
279
TijdschriftvoorEntomologie,volume 134, 1991
areusuallyconsideredtobeaccurate,althoughthey enormous amount of data from pitfall trapping
were obtained with a non-quantitative method. was available. Pitfalls were used in The Nether-
Furthermore, Lindroths classification applies to lands already in theearly 1950's byVanderDrift,
theFennoscandiansituation,andspeciespreferen- soon followedbyDenBoer in 1953.
ces may showgeographical variation. In the past, several studies concerning the effi-
ThemajordisadvantageoftheanalysesofThiele ciencyofpitfallsamplinghavebeencarriedout. A
(1977) is, that his compilations are based on data summary of the results of these investigations up
from the British Isles to Russia. Manyofthe spe- to1975waspublishedbyThiele(1977),concluding
cies onlyoccur in a limitedpartofthis area. that pitfall trapping is a suitable technique for
Thepresentstudywascarriedoutbecausethere investigatingcarabidpopulations inaquantitative
is a need for adequate ecological characterizations way. Den Boer (1977, 1990) reviewed investiga-
for the Dutch area with respect to future work, tionscarriedoutbypitfallsampling atthe Biolog-
especially in the field of nature management and icalStationWijster.Itappearedthatso-calledyear-
conservation. The material that is available from catches of most carabid species give reliable rela-
TheNetherlandsallowsustousequantitativeme- tiveestimatesofthemeandensitiesofactiveadult
thods todescribetheecologicalpreferencesof285 specimens around the pitfalls (Baars 1979, Den
species in our area in great detail. The ecological Boer 1979). Ayear-catch isdefinedas thesumma-
characterizations can serve as a basis for future tion per species of all specimens caught in one
descriptions and evaluations of carabid faunas of seriesofpitfallsduringoneyear (oratleastduring
specific areas withinThe Netherlands. the main reproductive period of carabids). These
estimates are relative, meaning that they can be
used only to compare the relative abundances
Materialand methods within a certain species over a series of samples
Material and/oryears, but notbetween species. This is be-
causeeachspecieshascharacteristic"catch-parame-
When preparing the first edition of the Dutch ters', such as activity pattern, way of living and
carabidatlas(Turin,etal. 1977)itappearedthatan catchability (see Luff 1975, Den Boer 1986).
Turinetal.: Carabidbeetles
The dataset for the present study consists of
1616 year-samples from 862 localities in The Ne-
therlands (fig. 1), covering the period 1953-1983.
A list ofall year-samples included in the database
isgiven by Turin & Penterman (1985).
Thenumberofpitfallsusedinapitfallseriesand
the sizes ofthe traps varied considerably between
investigators.Therefore,thenumberofspecimens
in a year-sample were standardized by calculating
thenumberofspecimensperdecimeterpitfalledge
per year (SDY).
Before considering the catches, all sampling lo-
calities were uniformly described, using the 'eco-
code'oftheEuropeanInvertebrateSurveyforThe
Netherlands (VanTol 1979) in aslightlymodified
version (Penterman & Turin 1985). In this code
informationabouttypeofvegetation,soiltype,soil
humidity, size of the locality and type of manage-
ment is recorded. This information allowed us to
recognize33 habitats (table 1,p.292),accordingto
which the 1616year-samples were coded.
Fig. 2. The frequency distribution of the species mean
ln(SDY+l)-vaIues over each of the 33 habitats, taken
fromthedatamatrix.A:theY-axisshowsthenumberof
relativeabundances (seetext) in thedata matrix. B: the
meanln(SDY+l)-values;onlyaverysmallpartoftheX-
axisisshown.Thismeansthatthevaluesareverylowall
overthedatamatrix.Thesefiguresprovedtobeunwor-
kableforTWINSPANclassification.
1600
TijdschriftvoorEntomologie,volume 134, i99i
where PRESj represents the ecological amplitude species.Asimilarproblemexistswhereitconcerns
of the j-th species, Oj the number of habitats in the species Pterostichus nigrita and Asaphidion
which this species is present, and n the total flavipes. Pterostichus rhaeticus was not separated
number of habitats recognized (33). PRES will from P. nigrita (see: Koch 1984), andAsaphidion
range from to 1. curtum as well asA. stierlininot from A.flavipes
For the second 'eurytopy' estimate (SIM) the (see: Lohse 1983, Lompe 1989, Schweiger 1975).
index of Simpson (D) is used (Simpson 1949), The full names of all species are given in the ap-
which characterizes the distribution of the abun- pendix.
dances over the habitats in the formof: 1-D.This
may thus bewritten as: Results
Classification ofhabitats
' - 2 (A-)2 ^^- Figs.4-11presenttheresultsfromtheTWINSPAN
classificationoftherelativeoccurrencesoverthe33
habitats from table 1. Seven main habitat groups
where aj is a measureofthe relative abundanceof could be recognized (fig. 4):
lans(pSeDcYie+sl)ijn,hwahbiitlaet group j and is equal to mean 1-5G.roAupfuIr.t-hePreastubadnidvihseiaonthovfegtehtiastieonnds,-ghraobuiptatiss
presented in fig. 5.
n Group II. - Poor grassland and dune habitats,
a tot = 2= l in (SDY + 1), habGirtaotusp6I-I1I.1-(Csuulbtdiivvaitseidonlasnedeafingd. o6)p.enconiferous
j
plantations,habitats 12-15 (subdivisionseefig. 7).
Thevalue ofSIM may range from to 1. GroupIV. -Matureforests,habitats 16-20 (sub-
division see fig. 8).
Group V. - Moist/wet forests (forests in water
Soiltypeand humidity
meadows; brook forests) and ruderal grass locali-
When characterizing the sampling localities, ties, including limestonegrassland, habitats 21-25
three humidity classes were recognized, viz. 'dry', (subdivision see fig. 9).
'moist' and 'wet'. The humidity preference of a Group VI. - Moist habitats overgrown with
species was estimated at an ordinal scale from 1 weeds,polder-(colonization-)habitats,habitats26-
(very xerophilous) to 5 (very hygrophilous). Spe- 30 (subdivision see fig. 10).
cies only caught in 'wet' localities score 5, species GroupVII.-Wethabitats/shores,habitat31-33
from'wet'aswellasfrom'moist'localitiesscore4, (subdivision see fig. 11).
speciesonly from 'moist' localities score 3,species
from 'moist' as wellas from 'dry' localities score2
andspecies only from 'dry' localities score 1.
Each sampling locality was also classified into
one of seven soil types: limestone (li), loam (lo),
loamysand/sandyclay (Is), riverclay (re),seaclay
(sc), peat moor (pm) and sand (s). A species is
regardedtoshowapreferenceforacertainkindof
soilwhentheln(SDY+l),averagedoveralllocal-
ities with this specific soil type, is at least 2 times
the sum of the mean ln(SDY+l) values of the
localities from theothersoiltypes takentogether.
Fig. 4. Dendrogram of the main division into seven
groups according toTWiNSPANclassification ofthe spe-
Nomenclatureand taxonomy cies'relativeoccurrences. I-VII=mainhabitatgroups:I
The nomenclature followsTurin (1990),except I=I=PeaPtooarndgrhaesastlhalnadndanvdegdetuanteiohnasbi(thaatbsit(ahtab1i-t5a,tse6e-1f1i,g.s5e)e,
for three Calathus species of the melanocephalus fig.6),III=Cultivatedlandandopenconiferousplanta-
gmraoteurpi,aflorfwrhoimcphitifsarlelfterrarepdpitnogApurkeesmenat(e1d99h0e)r.eTuhne- t(ihoanbsita(tha1b6i-t2a0t,1s2e-e15f,ig.se8)e,fVig.=7)M,oiIsVt=/ wMeattufroreesftosreasntds
dkmeierxmtatuhrIee99onOfa).mC.eOnmloefylaiCnn.otcmheeeplhmaaonlrouecserpaehncadelnutC.sscaiimnnpcltfiaucsntg(iA(scua-a. wrs(euihedatbehfiritgaw.alet1e0g2d)r1s,a-,s2Vs5Ip,Iolls=odeceeaWrlf-eiit(gti.ceo9hs)l,a,obnViiitInzaca=tltsui/Mdosionhni-gos)rtelsshiiatm(beehisstaabt(tiohstan,abetiotv3ga1ert-ar3s2g3s6r,l-oa3swn0end,e
after 1980),C.cinctuswasrecognizedasaseparate fig. 11).
282
Turinetal.: Carabidbeetles
1
TijdschriftvoorEntomologie,volume 134, 1991
accordingtothepercentualscaling,whereforeach
speciesthehighestmeanln(SDY+l) inoneofthe
33habitats,isputat100%.Inthebodyofthetables
thenumbersindicate: l=5-15%;2= 15-25%etc.,
9 = 85-95% and * = >95% respectively. Points
indicate the habitats where the species is present,
but where the relative occurrences are below 5%.
Specieswereordinatedbydecorana,according
totheirrelativeoccurrencesperhabitat.Infigs. 14-
20 the position of the species belonging to the
groupsA-Honthefirsttwoaxesoftheordination
is given. The first axis shows a dry-wet gradient:
1000, 1001, 1010, 1011 species preferring dry habitats (group A) on the
left (fig. 14),andspeciesfromwethabitats (group
1100, 1101, 1110, 1111 H)ontheright(fig. 19).Thesecondaxisisrelated
with amountofshade: speciesofexposedhabitats
Fig. 13. Dendrogramofthespeciesgroups according to (group A) at the bottom (figs. 14, 15), and forest
theTWINSPANclassification.ThelettersA-Hindicatethe dwellingspecies (groupD) atthetopofthefigure
TWINSPAN end-groups, treated in the respective species (fig. 17).ThehighlyeurytopicspeciesofgroupEU,
glervoeulps.ofTthheetnwuimnbseprasn0c0l0as0s-i1f1ic1a1tioinnd(iccoatmepatrheetfaibrlset2f-o7u)r. take a central position according to that of the
otherspeciesgroups,resemblingthatoftheinter-
mediarygroups E, F andG (fig. 18).
SpeciesgroupC(table4): Speciesofopenareas,
dune localities as well as colonization sites and
Ecologicalamplitude, soil typeandhumidity
shores, mainly in habitat-groups II, III and/or VI,
VII. The indices for the ecological amplitude of the
Species group D (table 5): Species of forests species (Pres, Sim), for soil preference (So) and
and/or raderai places, including limestone grass- humidity preference (Hu) are given in the right
land; occurrences mainly inIV and/orV. columnsofthetables2-9.Alsothenumberofyear-
SpeciesgroupE,F,G(table6):Thesegroupsare samples (Sa), the numberofindividuals (N).
not treated separately, because the separate end-
groups are rather heterogeneous and less typical
Discussion ofthe species groups
foracertain habitatgroup than theotherspecies-
groups.However,manyspeciesoccurinmoistand Inthischapterwewillgiveashortdiscussionper
shaded localities, such as moist forests, scrubs and species group. Analyses more into detail can be
reedland;havinganaccentinhabitatgroupsVand expectedinfurtherpaperswhere thespeciescom-
VI,andothersshownospecialpreferenceformoist position ofseveral habitats or habitat-groups will
habitats. be treated.
H
Speciesgroup (table7):Speciesofmoreopen
moist-wethabitats,reedlandandshores;occurren- Species-groupA (table 2 p. 294)
ces mainly in groups VI and/or VII, (twinspan Mostspecies listedin table2 showclearprefer-
species end-groups 1100, 1101, 1110 and 1111). encesforpeatysoilsand/orsandysoils.Whenthey
Eurytopic species (EU) (table 8): These species have no preference for any soil type, they are
are placed in a separate group on basis of their mainlyrestrictedtooligotrophichabitats. Thehu-
ecologicalamplitude: PRES>0.75 orSIM>0.85. midity preference varies: most species that also
The species are arranged according to their eury- occurintheCorynephoretum,suchasPterostichus
topyestimates(Pres,Sim).Theend-groupnumber lepidus,BradycellusruficollisandAmarainfima,do
indicates fromwhichofthespeciesgroups menti- not occur in humid habitats. Other species have a
onedabove (A-H) theyoriginate. ratherhighpreferenceforhumidsites(humidity4-
Rare species (R) (table 9): Species have been 5),suchasPterostichusaterrimus,Agonumericeti,
placed in thisgroup when the numberofsamples andAnisodactylusnemorivagus,which areknown
(Sa)<6andwhenthenumberofindividuals<50. from literature as true peat-moor dwellers (Lin-
Ifthespeciesismerelypresentin3-5year-samples droth 1945, Mossakowski 1970a, 1970b). The re-
andthese samples belong to the same habitat, the sultsagreewiththoseofinvestigationsinGerman
species was notplaced intogroup H. heath-moor complexes (Horion & Hoch 1954,
Grossecappenberg et al. 1978). According to the
In tables 2-9 relative occurrences arepresented TWiNSPAN-classification (figs. 4-5), peat moor has
284
Turinetal.: Carabidbeetles
B^
TijdschriftvoorEntomologie,volume 134, i99i
SpeciesgroupD (table 5,p. 296) century, has increased this area to about 6% in
Three subgroups were grouped together based 1950and8-9% atthemoment.Thus,manyDutch
ontheiroccurrencesinthemainhabitatgroupsIV forestsareratheryoung.Themajorityofthisfore-
and V: Dl - Stenotopic species of cultivated and sted area consists of coniferous and exotic trees.
raderai sites such as poor unfertilized fields, Apartfromthelightoak-birchforests,whichform
gardens and limestone grasslands, mainly on li- themorenaturalclimaxvegetationonsandysoils,
mestone soil in the southern partofthe province and the riverine forests, the more rich types of
ofLimburg;D2-Speciesofmoreorlessmoistand deciduousforest,e.g.Querco-Carpinetum,canonly
shadedplacespartlyonlimestonesoil; D3 -Eury- be found in restricted areas. Species characteristic
topicandstenotopic forest species. (Thiele 1977) of these kind of forests in Central
Belonging to the first subgroup are several Europe, e.g. Abax ovalis (not in pitfalls), Carabus
heath-preferring species (xerotherm species: auronitens, Molops piceus, Pterostichus cristatus
Becker1975,Lindroth1949)whichinTheNether- andTrichotichnusnitens,arerestrictedtothevery
lands are exclusively confined to limestone soil: eastern and southern parts of The Netherlands.
Parophonusmaculicornis.Amaranitida,Brachinus Thiele (1977) reviews studies on the most abund-
crepitans,Harpalusdimidiatus,andOphonusmel- antspeciesofthe(sub)centreuropicforests,includ-
leti.Otherthermophilousspeciescanalsobefound ing many investigations from Germany, such as
on localities outside the limestone area of The those of Heckendorf et al. (1986), Kolbe (1968,
Netherlands, on open sun-exposed, xerotherm 1970), Lauterbach (1964), Rabeier (1957, 1962,
sites such as slopes ofriverdikes with a southern 1963, 1967 and 1969) and Thiele (1956). Also li-
exposition (Turin 1983,Turinetal. 1977): Amara terature concerning The Netherlands was taken
montivaga, Ophonus puncticeps, Harpalus ru- into account by Thiele, e.g. Van der Drift (1959)
bripes, andLebia chlorocephala. Ophonus rufibar- and Den Boer (1965). Recently, Turin & Heijer-
bis is the least thermophilous and most eurytopic man(1988)carriedoutafirstsurveyonthepresent
Ophonus species. It can be found on more or less data in a more simple way by only examining the
shady sites (Lindroth 1974, 1986). species having thehighest numbersofindividuals
ThespeciesoftheD2 subgrouparerestrictedto andpresences inalltypesofforestinTheNether-
the southern part of the province of Limburg or lands together, and arranging them according to
have a more or less fluviatile distribution in The thefractionofspecimenscaughtinforests,ascom-
Netherlands. Most species prefer chalky, clayish pared with those in other habitats. The terms
soil or loam, some of them inhabiting shaded 'stenotopic/eurytopicforestspecies'theyuse,only
(wooded) sites: Stomis pumicatus, Bradycellus apply to the Dutch area. Especially the species of
sharpi, Carabus coriaceus, and Pterostichus madi- mountainous localities are missing (highest alti-
dus. These species also occur in limestone grass- tude in The Netherlands is below 400 m). Most
land at slopes with a northern exposition (Turin stenotopic forest species mentioned by Turin &
1983). Heijermanare now listedin table 5. Heijerman&
SubgroupD3 consists ofthe true forest species. Turin(1989)foundgreatdifferencesinfaunacom-
The first six species are rather eurytopic, which position between forests indifferent parts ofThe
alsoappliestoLeistusrufomarginatusandCarabus Netherlands.Thiscanpossiblybeexplainedbythe
problematicus.Someofthemorestenotopicforest factthatmostoftheforestedareasareratheryoung
speciesaremoreorlessrestrictedtotheoak-horn- and have beencolonized recently.
beam forests ofthe southernpartofLimburg and
the easternmost part of The Netherlands, e.g. Speciesgroup E, F, G (table6,p. 297)
Abaxparallelus,Molopspiceus,andTrichotichnus Thisgroupispolytheticandinfactacompilation
nitens. Amara praetermissa can be found in the ofseven smalltwinspanend-groups. The species
Corynephoretum,butalsoatacidsitesontopofthe predominantlyoccurinmoistandshadedhabitats.
limestone hills in Limburg with a more heather- Most species are hygrophilous, reaching high
like {Calluna) vegetation,whichexplainsthepres- numbers in wet grassy forests, moist grasslands
enceofthis species in the D-group. and reedland: El - speciesofmoist forests,some-
Therearefarmoreinvestigations into thecom- times indunevalleys andin moistgrassland; Fl -
positionofthecarabidfaunaofforeststhaninthat two species of rather open and dynamic habitats;
of heathland vegetations. Nowadays heathland is F2 - species of moist and shaded sites; Gl - only
rapidly disappearing from The Netherlands, but a single species; G2 - species with high numbers
about a hundred years ago it was one of the most in reedland and in moist grassland, Oodes helo-
prominent components of the landscape. In 1850 pioides lives at wet and shaded sites; G3 - only
the areacoveredwith forest was below 3%. Affo- Bembidion obtusum, which more or less prefers
restation, which started in the first part of this cultivatedcountry;G4- mainlyhygrophilous spe-
286
Turinetal.: Carabidbeetles
cies ofwet meadows, marshes and reedland. Most (1987) andthey mentionseveralspecies as typical
species ingroup G4 inhabit ratherwet and dense, for this habitat. The species occurring in saline
shaded vegetations. Exceptions are: Amara ovata habitatsinTheNetherlandsarealltreatedinTurin
andy^./aweZ/CA,whichbothUveinmoreopensites. (1991).
Investigations into the fauna of moist and wet Manyofthecharacteristicandabundantspecies
habitats arefor reedland,swampsandlakeshores: of inland shores and river banks cannot be found
David & Marchai (1963), Dawson (1965),Jarmer in Group H, since pitfalls do not function well in
(1971), Koch (1977), Obrtel (1972), Renkonen shore habitats. Especially focused on the riparian
(1944) and Wasner (1977). The present results faunaofrunningwaterarethestudiesof:Andersen
largely agree with the studies mentioned. (1970, 1982, 1983), Krogerus (1948), Lehmann
(1965) andPlachter (1986).Species livingsoclose
Speciesgroup H (table 7, p. 298) to the water that they hardly can be trapped with
The species belonging to group H reach high pitfalls areforinstance: Agonum micans,Bembid-
numbersinthemoistorwethabitatgroupsVIand iondoris, B. obliquum, Dyschiriusaeneusand Ne-
VII. Subdivisions based on the twinspan classifi- bria livida. The species Bembidionarticulatum, B.
cationare: HI - asingle rathereurytopic,coloniz- punctulatum, B. semipunctatum andB. testaceum.,
ing species: Amara sim-ilata; H2 - mostly hygro- all from river banks and often abundant in The
philous speciesofyounghabitats inpoldersandin Netherlands,arenotpresentinthepitfallmaterial.
othercolonizationsites; H3 - fourspeciesofopen Someshorespecies thatalsocanbefoundatsome
sites near water; H4 - Species ofshore habitats, a distance of the water, are listed among the rare
numberofthemconfinedtosalinelocalities. Pter- species (table 9).
ostichus cupreus is similar to Am,ara similata in
living in moist grassland and arable land, both Eurytopicspecies (table 8, p. 299)
species being rather eurytopic. Bem-bidion bruxel- A group of 53 eurytopic species and ubiquists,
lense, Chlaenius nigricornis, Agonum albipes, extracted from all habitatgroups onbasis oftheir
Elaphruscupreus andE. ripariusarecharacteristic ecological amplitude. Most species in this group
ofriver meadows andreedland.Amaraconvexius- show high occurrences and abundances in one of
cula is a typical species of sea clay soil and found the habitatgroups III,V and/orVI. Apart from a
mostlyalongthecoast,butitoccurs,justasAcupal- fewspecies,theydonotclearlypreferaspecialsoil
pusexiguus,LasiotrechusdiscusandAmaramajus- type or humidity class. Some of them have high
cula,alsoinhigh numbers intheDutchIJsselmeer tolerances with respecttofertilizers andintensive
polders (Turinetal. 1977). For largeareas suchas agricultural managementandcan thereforebenu-
the Dutch polders, only constantly macropterous merous in arable land and in pastures. Thiele
species and full-wingedspecimensofwing-dimor- (1977),whocomparedthecarabidfaunasofarable
phic species are capable of colonizing such areas land, meadows / pastures and clover / alfalfa, in
(Haeck 1971, Meijer 1973, 1974, 1980). In colon- fact gives incomplete information of the species
ization of small sites, such as burnings and clear- preferences. The species listed in Thiele's compi-
ings, also big wingless species may be able to im- lationarenotcharacteristicforcultivatedareas,but
migrate (Szyszko 1986), but in general, pioneer most of them simply are highly eurytopic species
communities can be characterized by their high (see table 8). A similar survey was published by
numbers of winged species. An example of long- Basedow et al. (1976) and the results resemble
termcolonizationandsuccessionofcarabidsalong thoseofThiele.
rather homogenous road verges on clay soil in Notiophilus biguttatus and Nebria brevicollis
three Dutch polders of different age is given by areeurytopicforestspeciesthatalsooccurindense,
Haeck et al. (1980). It illustrates that the carabid shadygrasslands andinothershadedsites,such as
faunaoftheDutch polders keeps acoastalcharac- gardens and orchards (Turin & Heijerman 1988).
ter for many decades.
MostspeciesoftheH4-groupliveclosetowater Rare species (table9,p. 300-301)
in open habitats. Several species ofthis group are Agroupof59rarespecies.Justasintheprevious
restricted to the salt marshes of the Waddensea groups, the original twinspan end-groups (GR)
area and the islands of the province of Zeeland: arementionedafterthespeciesnames.Twenty-six
Dicheirotrichus gustavi, D. obsoletus, Pogonus ofthespeciesonlyoccurinonekindofhabitat.For
chalceus,Bembidion norm,annum,B.pallidipenne, some species a lack ofdata is the main cause that
Dyschirius salinus, D. obscurus, and Bradycellus the information on the species preferences is in-
distinctus.Aboutsalinehabitatsseveralstudiesare complete. However, for many species that have a
available: Heydemann (1962, 1967),Mossakowski highly restricteddistribution in The Netherlands,
(1971), Niemela (1988) and Rueda & Montes especially those living near the fringes of their
287
,
TijdschriftvoorEntomologie,volume 134, 1991
ranges, the relative occurrences shown in table 9 studyarea,bylisting thedataofhundredsofyear-
willgivea reliablepictureofthespeciespreferen- catchesinmanykindsofhabitatintheprovinceof
cesintheDutchterritory.Thespeciesofthisgroup Drenthe (The Netherlands). However, his study
willbe treated in a separate study. areaisratherlimitedandmanyDutchhabitatsand
soil types are absent from his study (e. g. coastal
dunes,claysoilandlimestonegrasslands). A com-
Discussion
parisonwithDenBoer's results isnotappropriate
Only a few studies concerning classification of here,becausehisdataformasignificantpartofour
carabidspeciesbasedonalargedatasetareknown data set.
from western and north-western Europe. The Severalextensive studies onclassificationofca-
most important are chronologically: Lindroth rabidspecies havebeenpublishedrecently. Luffet
(1945, 1949),Thiele(1977),DenBoer(1977),Luff al. (1989) is related to theclassificationofcarabid
et al. (1989) andEyre& Luff (1990a, 1990b). habitatsinnorth-eastEngland,whileanotherdeals
Lindroth (1945, 1949) indicated species prefer- with the ground beetles of grassland habitats in
ences for humidity, and certain vegetation types, Europe (Eyre & Luff 1990a). Finally, Eyre & Luff
such as woodland. Although his conclusions were (1990b)presentedaclassificationofgroundbeetles
based mainly on material from Fennoscandia, his oftheBritishgrasslands in moredetail. Inall stu-
classification proved to be valuable for a much diesDECORANA andTWINSPANwereusedasstatis-
larger area. It was used in many carabidological ticalmethods.Forthefirststudy (Luffetal. 1989),
studies all over northern and western Europe. very heterogenous material from 248 sites in
When comparing the results of the present study North-East England was used, primarilycollected
with Lindroth's classification (table 10), one can foruseintheBritishmappingscheme.Tenhabitat
conclude thattheycorrelateratherwell. Thexero- groups were recognized based on the distribution
philous species (X) in groups Al, Bl, B2, CI and of the species presence over the sites. A method
Dl (table 10), the hygrophilous species (H) in wasdevelopedtofitinnewsites. Afirstlimitation
groups Al,G4,H2,H3 andH4 (table 10) andthe of this method is that new sites can only be fit in
(HW+W+WA)
forest species of Lindroth are using species that already contributed to the orig-
found ingroups D2, D3 andEl. Lindroth's meso- inal ordination. Secondly, that only pres-
philousandmoreorlessubiquitousspecies(N)are ence/absence data can be used. The original data
scattered over a wide range ofour species-groups setconsistsbothofdatafrompitfalls,andsampling
with the highest numbers in the eurytopic EU- by hand. The possibility to include also samples
group (table 10). Fifty-nine species mentioned by fromlocalities wherepitfallsdonotfunction,such
Lindroth were caught in pitfalls in too low as shore habitats, certainly is an advantage when
numbersforacharacterization(groupR)and85of usingonlythespeciespresence. A maindisadvan-
theDutchspeciesmentionedbyLindroth,werenot tageisthatthespecies'relativeoccurrencescannot
caught in pitfalls at all inThe Netherlands (table betakenintoaccount.ThestudyofLuffetal. is,as
10 group O). From Lindroth's characterizations theyalsonoticeintheirdiscussion,inthefirstplace
(table 10) we can conclude that these species that a methodologicaloneand theresultsdo notgivea
could not be characterized in the present study, definitiveandcompletepictureofcarabidhabitats
belongtotheinhabitantsofshoresandriverbanks of the British Isles. Many habitats, for instance
(80species,table 10). Others areveryrareoreven woodland,willhavetobeinvestigatedmoreinten-
extinctinTheNetherlands,e.g. manyxerophilous sively.ThreeofthetenhabitatsdonotoccurinThe
species (see Desender & Turin 1989,Turin 1990). Netherlands (uplanddry,uplandwetandboulder).
Thiele (1977) attempted to characterize the The remaining habitat groups are rather broadly
carabid faunas of certain habitat groups such as defined: coastal, woodland, dry grassland, wet
forests,cultivatedlandandsandyhabitats,compar- grassland, wet running, wet still - silt, wet still -
ing them on basis of species presence. He mainly sand. It is rather difficult to compare the species
usedliteraturesources from theentire northwest- preferences for British habitats with the present
ern part of Europe. Although he included several results,althoughmanyspeciespreferencesseemto
lowlandandmountainforestsforthecharacteriza- correlatewell,e.g.thoseofCalathuserratus,Amara
tion of the forest fauna, we feel that this method bifrons, Calathus rotundicollis and those of many
gives an incorrect picture of the species preferen- eurytopicspeciessuchasCalathusmelanocephalus
ces,especiallybecausetherelativeoccurrencesand Loricera pilicornis, Pterostichus nigrita and Tre-
abundances inother localities were not taken into chusobtusus.Fromthepresentclassificationitcan
account. belearnedthatthemostimportantenvironmental
DenBoer (1977) gives afarmorecompletepic- factorinfluencingtheordinationofcarabidspecies
ture for 75 of the most abundant species in his is soil moisture (fig. 14). This is in supportofthe
288