Predicting species richness for Australasian freshwater macroinvertebrates: do we want to know?

8 Pages·1997·2.2 MB·English

by  J E Growns

Memoirs ofthe Museum ofVictoria 56(2):483-490 (1997)
PREDICTING SPECIES RICHNESS FOR AUSTRALASIAN FRESHWATER
MACROINVERTEBRATES: DO WE WANT TO KNOW?
J.E. Growns1 and I.O. Growns
Australian WaterTechnologies, PO Box 73, West Ryde, NSW 2114, Australia
lPresent address: Murray-Darling Freshwater Research Centre, PO Box 921, Albury, NSW 2640, Australia
Abstract
Growns, J.E. and Growns, I.O., 1997. Predicting species richness for Australasian fresh-
watermacroinvertebrates: dowewanttoknow?MemoirsoftheMuseum ol Victoria 56(2):
483-490.
The identification offreshwatermacroinvertebratestofamily level isbecomingincreas-
inglypopularforsurveys,predictivemodelsand pollution indicesinAustraliabecauseit is
quickerandcheaperthangenusorspecieslevelidentification,anditrequireslessspecialised
knowledge. Family richness hasbeen used asapredictorofspeciesrichnessforothertaxo-
nomicgroupssuchasvertebrates, antsand plantsandwewereinterested inseeingwhether
this might be a useful method for freshwater macroinvertebrates. Taxon lists from one
Papua New Guinean and 34Australian datasets from lentic and loticwaters were used to
regress the number offamilies against the total number oftaxa (species where possible).
Also, the ability ofthe numberofspeciesand morphospecieswithin someorders(Coleop-
tera, Diptera, Ephemeroptera and Trichoptera) to predict overall species richness was
investigated.Thenumberofall familiesexplained 91%ofthevariationinspeciesrichness.
The number ofspecies within each ofthe orders explained between 60 and 85% of the
variationinoverallspeciesrichness.Weconcludethatitwouldbepossibletopredictspecies
richnessinthisway,particularly ifsamplingtechniquesandsamplingeffortwerestandard-
ised. Theterms'species richness' and'biodiversity' areoften used synonymouslyalthough
the former is only a subset ofthe latter. Some ofthe limitations and dangers ofassessing
species richness instead ofbiodiversity are discussed.
Introduction richnessandconcludedthat 'withcarefulchoice
of higher-taxon rank, it may be possible to re-
Most freshwaterstudiesaddressingbiodiversity
deploy effort from taxonomically intensive to
haveassumedthatbiodiversityandspeciesrich- taxonomically extensive surveys'. Andersen
ness are synonymous (e.g., Lang and Raymond, (1995) used a similar approach to estimate
1993; Allan and Flecker, 1993; but see Collier, speciesrichnessofAustralian antsusinggeneric
1993). Species richness is very expensive to
diversity.
determine using traditional methods and tax- We obtained taxon lists from 34 datasets
onomy. Rapid BiologicalAssessment(RBA)has throughout Australia and one from Papua New
been developed in orderto overcome thesedif- Guinea and used family richness for the whole
ficulties. RBA is based on eitherthe use ofmor- community and the numbers ofmorphospecies
phospecies rather than formally named species
within four orders (Coleoptera, Diptera,
(basic RBA) or on identifying either subsets of Ephemeroptera and Trichoptera) as predictors
communities or only identifying to taxonomic ofcommunity species richness.
levels higherthan species(ordinal RBA: Beattie
et al., 1993). We present an evaluation of how Methods
effective ordinal RBA is likely to be for fresh-
water invertebrate communities. The number of families, total number of taxa
Ordinal RBA assumesthat higher-taxon rich- and the numbers of coleopteran, ephemerop-
nessorthe numberofspecies intaxonomic sub- teran, trichopteran and dipteran taxa were
setsarecloselyrelatedtooverallspeciesrichness determined for 34 taxon lists from surveys
but this has rarely been tested. Williams and throughout Australia and 1 from Papua New
Gaston (1994) used family richness to predict Guinea (Table 1). The Australian studies were
species richness ofBritish ferns, British butter- from all states and one territory: Western Aus-
flies, Australian passerine birds and North and tralia, 10; New South Wales, 9; Victoria, 5;
CentralAmericanbats.Theyfoundthatforeach Queensland, 3; Northern Territory, 3, South
of these groups of organisms, family richness Australia, 2, Tasmania, 2. The types of fresh-
explainedat least 79%ofthevariationin species water system were also very varied and eight of
483
— 5
i i
484 J.E. GROWNS AND I.O. GROWNS
£ a c £ c £ OO C>U
cQu. *c- "tS^j Et"•S"? c^-c"^e»--o.cuoc.,u u u o o u tu o *c-ul fll -ct_u* -ct_u» o o u o .y
—_£o< .Cc5U .cC5U —o*h -Cc5U —o* o _o -4-» _o ccu .rmC5iUt alOinTi lotic.rmc5iut rmciut lotic lotic —O' S=c3aiOCca O '-—§ IZJ
.
G-CO^U y.y, < GZO GzO GzOakJ>> > HZ
z
ca
cu
S3
-eoa 3 X3) -aca 0O3 C -co c
X) •C cu
-1C) 0U3 T.«a033.£tC*U U uc3a o-u.- Qccau TCi3£>Caw>s 0CcC1o3J)o Ec0cCccaauu0 «Ec0cac3aua0*^,Tv3> > 3Socio^—nW«e)a5So>i Ec c 0u>4 X0o>)4 >> N
XQc>a 2O-sd5X>C)UJT«u"DJ5^oZIh"2jC«U?> ztSc1-u).-0e3 scHC003330 soo0ca0 H«eO<&u .^ C D3 Oo3o.hosga.2 ooo (0CU3 0u>3,.Tp>5a3..>03a<3 (SBJma3J,\J^3ooOu,3S*-'"(<«-C2>U0 —^oat1—,31 X^03)3i CWxto3O:: JJSCt^3u
Hca cIaD OOONOnOXOOnOMOONO(—JniOMVnDOOTnNtO^antOOCsnJ OONN OO-ann- OiONnN OoONoO OcO«NNi qO(5q^£I^rTOnOntNONO'nO*OOtnONOOtnNNOOTNN)OO-NNOOOOnO0OOOnOtOONn'OOaOn-O oOTjon- <0ON0n mOOnn
tu
o
>."=cu a T3 T3
O OTCC033O3 o0oUtCC33u0-<~u- Ot^n OCcou^a 'CcIOa 'I.. ^TW0oa>33 SooccIa (o•Js ^TP0Oa323S =C-cOoQalS ^Tatc3u .aE2 3Coc3c3/u.3 3Cnc3t3Ou. TZOCo3l OTC03O33 03 •ca
<3 ol -J T^3-" Di—.-^c o 1Cooao .0S0 cw UC oea o03 ccuu oicc_ua TC033 cu
CScCtcOu3QaPCo3O3s!i3ODa3Ow3C-QcC3oC-O-QCSU_SUUe<cE§frUtt'S.o33233"uc^t«Eyu} "=o1°Xc<1Sa> XJHUVIUw^Qc«Oa>wfSji Dc>c3aa Cc>eQaa Qlccuaa3Qj<tc«duuj QoS3aS3ogo3.3Cgthu3PccUu!Sica •_occno-u.ox*ccoc-:aa• ZV•oi-—)ZiiVo--Iii^.a,s"0~—0
•
«Mr<i\t'nvOr^oo ^-cnNor-~oooNO — ridfi
NO r~ oo
Q03 <N tN cm
PREDICTING SPECIES RICHNESS 485
c
u
a
O o o 4-*
S
a Xrie o +o3 Bl
u
c
H < < < ^ o
Z
T3
c
_
u X!
<L> re
J3 re
re <cu
'3
JuS O
o c
c Z<U31)
00 ^ T-t ' ' re
••s^l§ X>
3
«3go 00 ais &
S 3 D,
Oh <L)
O —
OS m<N rn m*f
r*l <«">' rn §8
486 J.E. GROWNS AND I.O. CROWNS
the studies had at least some sites that were in the calculation of the regression equation.
polluted (Table 1). Percentage error in prediction of species rich-
The total number oftaxa for each study was nesswascalculatedbysubtractingthenumberof
taken as the number of recognised taxonomic predicted species from the actual number of
units (RTUs) for that study, e.g., where chiron- species, sothatanegativeerrormeansthatmore
omid larvae were not identified beyond family, species were predicted than were actually
this was counted as one RTU. Immature and recorded. Percentage error was plotted against
unidentified taxa were not included unless they number of families to identify any systematic
were the only RTU in that family or genus. errors.
Coleopteran adults and larvae were considered
to be different RTU's unless they were both
identified to the same published species name. Results
Where identification wasnottaken eventofam- Allregressionswerehighlysignificant.Thenum-
ily level, eachgroupwastakenasone RTU, e.g., beroffamiliesexplained 91%ofthevariation in
Hydracarina. the number oftotal taxa (Table 2). The number
Plots ofthe numbers oftotal taxa against the of taxa from the taxonomic sub-groups were
numbers offamilies and coleopteran, dipteran, poorer predictors oftotal numbers oftaxa with
ephemeropteran and trichopteran taxa were the numbers ofcoleopteran taxagiving the best
examined for outliers. The state in which each result (Table 2).
study occurred was superimposed on theplot of For most states, not enough datasets were
numbers of families, in order to look for any available to assess regional variation. However,
regional variation. the plot suggested that, for the datasets avail-
Linear regression was used to examine the able, lotic systems in south-west Western Aus-
relationships between the total numbers oftaxa traliahad lowspeciesrichness, whereasthose in
and thenumbersoffamiliesand thenumbersof NSWhadconsistentlyhighspeciesrichness(Fig.
coleopteran, ephemeropteran, trichopteran and 1). There was no evidence that the different
dipteran taxa. All data were log| (x+l) trans- states showed different relationships between
formed and residuals were examined for nor- species and family richness, i.e. the intercepts
mality using Normal Scores plots. and slopes ofthe lines for each state appeared
To test the predictive value ofthe regression very similar (Table 2).
of number of families versus total numbers of The plot oftotal number oftaxa against the
taxa,dataonnumbersoffamiliesandspeciesfor number of coleopteran taxa (Fig. 2a) showed
single sites were used from Marchant et al. that Carey Brook (study 21) and the Serpentine
(1995). The data from the lower La Trobe and River(14) had lower proportions ofbeetle taxa
Thomson Rivers and the upper La Trobe River than would have been expected and the Collie
werenot used forthistest asthey had been used River (31) had a higher proportion than would
Table 2. Regression results ofpredictions oftotal numbers oftaxa using log (x- 1) transformed
data.
Independent variable Intercept Slope r1
Number offamilies:
AllNdSatWasets 35 -0.488 1.530 0.91
9 -0.730 1.671 0.98
VIC
6 0.062 1.298 0.80
WA
10 -0.046 1.260 0.83
Number ofcoleopteran taxa 35 1.174 0.713 0.85
Number ofdipteran taxa 35 0.981 0.732 0.70
Number oftrichopteran taxa 35 1.356 0.606 0.68
Number ofephemeropteran taxa 35 1.573 0.632 0.60
"
PREDICTING SPECIES RICHNESS 487
Tasmania o NorthernTerritory larvae. This plot also showed a marked differ-
(J Victoria Queensland
* SouthAustralia NewSouthWales ence in the proportions ofDiptera between the
• W9stAustralia ft PapuaNewGuinea Williams River (study 7) and the River Murray
andtributaries(study 1),althoughtheybothhad
similar total numbers of taxa. The plot for
Ephemeroptera (Figure 2c) indicated that the
River Murray and tributaries, the Perth wet-
lands (study 13) and the Werribee and Lerder-
dergRivershad lownumberscompared totheir
total numbers of taxa. The trichopteran plot
(Figure 2d) indicated thattheRiverMurrayand
tributaries had low numbers ofcaddis whereas
40 50 60 70 BO 90 1OO 11O the Werribee and Lerderderg Rivers had high
Numberoffamilies numbers compared to their total numbers of
Figure 1. Plotoftotal numberoftaxaagainst number taxa.
offamilies for the 35 studies. The state in which the Thepredictionofspeciesrichnessfromfamily
study was done is overlaid. richness for 19 single sites(data from Marchant
et al., 1995) gave a mean error of —7.7% (stan-
darderror,2.5)withamaximumerrorof—31%.
have been expected. All these 3 areas are in Theplotofpercenterroragainstnumberoffam-
south-westWesternAustralia.Thesameplotfor ilies showed that roughly halfofthe sites were
dipteran taxa (Figure 2b) showed that Carey predicted to within 10% of the actual species
Brookhadahigh proportion ofDipterawhereas richness (Fig. 3). Also, 14 out ofthe 19 predic-
the Werribee (study 2) and Lerderderg rivers tions were for more species than actually
(study 3), which are both ephemeral rivers in occurred.
Victoria, had low proportions of dipteran
(a) 600 500" WibmsRiver—
i
O 400 ton -
8 RiverMurrayand
300 300 -WerribeeRiver LerderdergRiver tributaries
I CaieyBtook \y-
£ SerpentineRivef "
3 200 r, 200- ... <-CareyBrook
O :
O
100.
100 •
—
H
1 1 1
60 100 150
40 60
Numberofcoleopleiantaxa Numberofdipterantaxa
(C) 500T bull- —
-RiverMurrayandtributaries RiverMurrayand p.
tributaries
9 400 40U-
o Perthwetlands Perthwetlands
5 300 LerderdergRiver 300-
£>
E
B 200- _,WerribeeRlvei 200- * \
• .* t LerderdergRiver
100- " WerribeeRiver
*'
10 -2+0- 3+0- 0- 1 410 601- -810 1100
Numberofephemeropterantaxa Numberoftrichopterantaxa
Figure2.Plotsoftotalnumberoftaxaagainstnumbersoftaxafor(a)Coleoptera,(b)Diptera,(c)Ephemeroptera
and (d) Trichoptera.
488 J.E. GROWNS AND I.O. GROWNS
20 t
10
©
o>
O
3Di 40 50 60 70
c
OCD Numberoffamilies
-10
I
-20
-
-30
-40
Figure 3. Plot ofpercentageerror in prediction ofspecies richnessagainst numberoffamiliesforthe 19 single
sites from Marchant et al. (1995).
Discussion observed for tabanid diptera and odonates, as
wellasforothercommunities,onthenorthcoast
The numberoffamilieswas a verygood predic- of NSW, which has been designated the
tor of community species richness (i.e., total Macpherson-Macleay overlap (CSIRO, 1991).
number oftaxa) whereas the taxon richness for MacArthur (1972) observed that lotic invert-
the four orders were not such good predictors; ebrates are the principal exception to the rule
beetles were the best ofthe four. This suggests that the tropics have greater biodiversity then
that analysing whole communities to coarse temperate areas. However, Lake et al. (1994)
taxonomic levels may be more reliable than found that two Queensland creeks had signifi-
species level identification ofindicator groups, cantly higher species richness than streams of
if estimates of whole community species rich- similar stream orders in Victoria. Our data did
ness are wanted. not show higher species richness for tropical
There was no evidence that there were differ- areasbutthismaybeduetothesmallnumberof
ent relationships between numbers of families datasets from the tropics in this study.
and community species richness in different Several ofthe species lists showed that where
regions ofAustralia. However, there was some communities are low in species numbers ofone
NSW
evidence that lotic systems in had very taxonomic group, they are high for another
high species richness whereas lotic systems in group. The intermittent Werribee and Lerder-
south-west Western Australia had low species derg Rivers in Victoria had low numbers of
richness. Bunnand Davies(1990)suggestedthat dipteran larvae and mayfly nymphs but high
the low species richness ofsouth-west Western numbers ofcaddis fly larvae. Carey Brook, a in
Australian loticsystemsmaybeduetothearea's south-west Western Australia, had low numbers
isolation, historicalaridityandlowprimarypro- ofbeetles but high numbers ofdipteran larvae.
ductivity. Although low in numbers of species There are doubtless historical and evolutionary
for many groups, this area has a very high pro- reasons forthese patterns but it is interesting to
portion ofendemics among its flora and fauna, note that theoverall relationship between num-
including the lotic macroinvertebrates (CSIRO, bersoffamiliesandcommunityspeciesrichness
1992; Christensen, 1992). In contrast, high was the same for these areas as for all the other
levels of species richness have previously been studies.
4
PREDICTING SPECIES RICHNESS 489
The predictions ofspecies richness for single phenomenon, such as evolution (see Richard-
sites were not good. Only about halfofthe pre- son, this volume). It is obviously impossible for
dictions for single sites fell within 10% of the all of this information to be obtained for all
actual species richness and errors of up to 31% species, let alone in the time frames required by
occurred. The majority of the predictions for managersandlegislators. However,thisdoesnot
single sites were for higher numbers of species mean that species richness alone should be used
than actually occurred. This may in part be asasurrogateforbiodiversity. Scientistsneedto
because Marchant et al. (1995) included oligo- reach a consensus on what aspects of biodiver-
chaetes, triclads and mites as single taxa. How- sity need tobeconsidered forconservation pur-
ever, several ofthe 35 studies used to calculate poses and then communicate this to the wider
the regression equation also did this. The con- community.
sistently high predicted species numbers are
more likely tobebecausethepredictionsarefor Acknowledgements
soifnmglueltsiiptelsewshieterseainsatnhearreeag.reTshsieonlowwaesrosfasmtpuldiinegs TRhicahnakrsdaPreeardsuoentaonLdoBrenlainCdhaarRlotbosn,onSfaomrtLaakkien,g
eoffbaftmoairitlnyfaocrlootmwhpeearrspeirndgolpteoorstimitueolsntwiopofluselpdesciitbeeessewtxiuptdeihecistn.eedFaocthro tlhisehetdimdeattaospetrso.viBdreucsepeCchieessslimsatsnfmraodmeunuspeufbu-l
comments on the manuscript. State Forests of
example,atmostsitesyouwouldbelikelytofind NSW, HunterWaterandthe HunterCatchment
beetles from the family Dytiscidae. However, ManagementTrust are gratefully acknowledged
there are likely to be different species of dytis-
cidsatdifferentsites. Soasurveyofseveral sites for allowing unpublished data to be used.
would find more species ofdytiscids than a sur-
veyofonlyonesite,whereasbothsurveyswould References
recordonlyonefamilyforthedifferentnumbers Allan, J.D and Flecker, A.S., 1993. Biodiversity con-
ofspecies. This highlights the care with which servation in running waters. Bioscience 43: 32-
this type of predictive technique must be 43.
used. Anderson, A.A., 1995. Measuring more ofbiodiver-
Our results should be treated with some cau- sity: genus richness as a surrogate for species
tion asthe samplingintensity, methodsand tax- richness in Australian ant faunas. Biological
odevinesort,rmiywbuevtairboeinleidoefavemtohtenhagsttsuwtdueideihseaswvaeansdshuthoneewvngeenot.ghartaHpothwhii-cs BeattiCPnior:eno,scAeeRJrea.vdp,aiiMtndaigjoscnbri7,oo3fdJ:i.vD3et.9rh-seai4nt3dy.BiOloaidsviseveres,rssImi.e.tny1t9:9A3s.asPepsr.se4vmi-ee1nwt.
approach would be highly successful in predict- Workshop. Macquarie University.
ing the species richness offreshwater macroin- Bennison, G.L., Hillman. T.J. and Suter, P.J.. 1989.
vertebrates for an area. It might also show Macrotnvertebrates ofthe River Murray (survey
interesting patterns, such as the apparent high and monitoring: 1980-1985). Murray Darling
levels of species richness in northern NSW. In Basin Commission: Canberra.
addition, ourobservation that where one group Boulton. A.J., 1988. Composition and dynamics of
orifchonregsasn,iasnmosthienratacxoomnmoumniictgyrohuaps ilnowthsepsecaimees Boulttmoeannct,rosAit.nrJve.eamrastn.edbPrhLalt.ocDy.dc,othmeLms.uiNsn.,i,tMi1oe9ns9a1.sinhMtaUwcnorioviienrntsveiertrmyti.ct--
community may have high species richness, has brate assemblages in floodplain habitats of the
interesting evolutionary implications. lower River Murray, South Australia. Regulated
However, we are concerned that the almost Rivers: Research andManagement 6: 183-201.
exclusive focus on species richness in biodiver- Bunn.S.E.and Davies, P.M., 1990. Whyisthestream
sityassessment isunwarranted. Biodiversityisa fauna of south-western Australia so impover-
difficult concept to define as it comprises many ished? Ilydrobiologia 194: 169-176.
odinfef.erWenet nideeeads,toofkwnhoiwchwshpaetcitehsersipcehcnieesssairseo:nalny BunnSt,peabS.trEia.at,leEafdnawduantraedm,opfoDr.saHtl.revaaanmrdsiaLtooifnonethrieangtnahoner,tmNha.ecRrr.no,ijn1av9re8rr6a.-h
abislsietysstmoeenxttionfctthieoinraennddedmiisctirtiyb,utriaornitcy,ansutshceenptbie- fForreessth.waWteesrteBironloAguystr1a6l:ia6:7-c9o1.mmunity structure.
made, orat least attempted. It is also important Chessman. B.C. and Growns, J.E.. 1994. Williams
toknowwheretheyfit intotheircommunity, i.e. Riveraquaticmacroinvertehratestudy. Australian
whetherthey are needed forthe continuingsur- WaterTechnologies, EnSight Report no. 94/84.
vival ofother organisms, and whether they are Chessman. B.C., Growns, J.E., Hardwick, R. and
an example of a scientifically important Hollelcy. D.. 1994. Warung Forest Management
490 GROWNS AND GROWNS
J.E. l.O.
Area: aquatic fauna report. Australian Water Journal of the North American Bentholot>tcal
Technologies, EnSight Report no. 94/187. Society 13: 417-438.
Chessman, B.C., Growns, J.E., Hardwick, R., Lake, P.S.,Schreiber. E.S.G., Milne.B.J.andPearson,
Holleley, D., Jackson, J.E. and McEvoy, P.K., R.G., 1994. Species richness in streams: patterns
1994. Dorrigo three-year environmental impact over time, with stream size and with latitude.
statement area: aquaticfauna report. Australian VerhandlungenderInternationale Vereinigungfur
Water Technologies, Science and Environment Theoretische und Angewandte Limnologie 24:
Report no. 93/123. 1822-1826.
Chessman,B.C..O'Connor,R.andHolleley,D., 1995. MacArthur,R.H., 1972.Geographicalecology. Harper
Fenterfield Forest Management Area: aquatit- and Row: New York.
environment report. Australian Water Marchant, R., 1982. Seasonal variation in the
Technologies. EnSight Report no. 95/32. macroinvertebrate fauna of billabongs along
Chnstensen. P.E.S., 1992. TheKarriForest.Itsconser- Magela Creek, Northern Territory. Australian
vation significance and management. CALM. Journal ofMarine and Freshwater Research 33:
Western Australia. 329-42.
Collier, K„ 1993. Review ofthe status, distribution, Marchant, R., Barmuta, L.A. and Chessman, B.C.,
NanedwcZoenaslearnvda.tioNnewofZferaelsahnwadteJrouirnnvaelrtoefbrMaatersinien t1a9x95o.nomIinfcluernecseoluotfiosnamopnle tqhueantoirfdiicnaattioinonanodf
andFreshwaterResearch 21: 339-356. macroinvertebrate communities from running
Cosser, P.R., 1988. Macroinvertebrate community waters in Victoria, Australia. Marine andFresh-
structureandchemistryofanorganicallypolluted waterResearch 46: 501-506.
JcoreuerknalinofMsaourtihn-eeaastndQFureeesnhswlaatnerd.ResAeuasrtcrahli3a9-n Marcthrainbtu,tiRo.n,lMiisttcfhoerllt,heP.aaqnuadtNiocrriinsv,erRt.e,br1a9t8e4s.iAndtihse-
671-683.
CSIRUOn,ive1r9s9i1t.y PFrheess:inMseecltbsouorfnAeu.stralia. Melbourne 6lO3co-cw7al9sa.inodnarlegPiaopnerosfftrhoemLtaheTrMoubseeuRimveorf,ViVcitcotroiraia1.:
Davis, J.A., Barmuta, L.A and Balla, S.A., 1988a. Metzeling, L., Graesser, A., Suter, P. and Marchant,
Serpentine River and Gooralong Brook aquatic R., 1984.Thedistributionofaquaticmacroinver-
fauna study. Murdoch University, WA. tebrates in the upper catchment ofthe La Trobe
Davis, J.A., Barmuta, L.A. and Balla, S.A., 1988b. River, Victoria. Occasional Papers from the
DirkWBrAo.okaquaticfaunastudy. Murdoch Univer- Museum ofVictoria 1: 1-62.
sity, Norris,R.H., Lake,P.S.andSwain, R„ 1982.Ecologi-
Davis, J.A., Harrington, S.A. and Friend, J.A., 1993. cal effects of mine effluents on the South Esk
Invertebratecommunitiesofrelict streams in the River, north-eastern Tasmania III. Benthic
arid zone: the George Gill Range, Central Aus- macroinvertebrates.AustralianJournalofMarine
tralia. Australian Journal ofMarine and Fresh- andFreshwaterResearch 33: 789-809.
waterResearch 44: 483-505. Norris,R.H.,Moore,J.L., Maher,W.A. andWensing,
Davis, J.A., Rosich, R.S., Bradley, J.S., Growns,J.E., L., 1993. Limnological characteristics of two
tShcehmSiwdatn,cLo.aGs.taalnpdlaiCnh.eaVlo,lFu.m,e169.93W.etWleatnldancldasssoi-f Acuosatsrtaalliad.unAeustlraakelsi,anJeJroviusrnaBlay,ofsoMuatrhineeastaenrdn
fication on the basis ofwater quality and invert- FreshwaterResearch 44: 437-458.
ebrate community data. Water Authority of Outridge, P.M., 1987. Possiblecauses ofhigh species
Western Australia/Environmental Protection diversityintropicalAustralianfreshwatermacro-
De DAeuctkhkoerri,tyP.oafnWdesWtielrlinamAuss,trWa.lDi.a., 1982. Chemical b1e0n7.thic communities. Hydrobiologia 150' 95-
and biological features ofTasmanian salt lakes. Richardson, A.M.M., Swain, R. and Wong, V.E.,
Australian Journal of Marine and Freshwater 1996.ThecrustaceanandmolluscanfaunaofTas-
DoegfR.aeusT.ne.aaro1cf9h8t43h.3e:RM1ei1st2pt7oa-n1Ms1ie3t2to.afRtihveerm.acVriocitnorviear,tetboratthee smViiatcnitoionarainla 5hs6aa.blittmaatr.shMese:mociornsserovfattihoen Moufseaumtrano-f
construction and operation ofDartmouth Dam. Towns, D.R., 1985. Limnological characteristics ofa
GrowM2n.us,IsrerIiu.gmOa.t,ioofn1V9ri9ec2lt.eoarsiMeaa.cIrO:ocic1na0vs1ei-ro1tn0ea8bl.rPaatpeercsomfmruonmitthye CwSraoetueetkrh.RAAeuussstetrararalclihiaan3n6i:nJto8eu2rr1mn-ia8tlt3e7on.ftMsatrreianme, BarnodwFnreHsihl-l
structure in thestreams ofthesouthern forests of Williams, P.H. and Gaston, K.J., 1994. Measuring
Western Australia: the influence of seasonality, more of biodiversity: can higher-taxon richness
lDo.ngtihetsuidsi,naMlugrrdaodciehntUsniavnedrsfiotrye.stry activities. Ph. Cproendsiecrtvatwihoonle6s7a:le21s1p-e2c1i7e.s richness? Biological
Growns, l.O. and Davis, J.A., 1994. Longitudinal Yule, CM., 1995. Benthic invertebrate fauna of an
changes in near-bed flowsand macroinvertebrate aseasonal tropical mountain stream on Bougain-
communities in a Western Australian stream. ville Island, Papua New Guinea. Marine and
Freshwater Research 46: 507-518.