effect of low dissolved oxygen on survival, emergence, …€¦ · 251 j. n. am. benthol. soc.,...

251

J. N. Am. Benthol. Soc., 2004, 23(2):251–270� 2004 by The North American Benthological Society

Effect of low dissolved oxygen on survival, emergence, and drift oftropical stream macroinvertebrates

N. M. CONNOLLY1

Rainforest Cooperative Research Centre, Australian Centre for Tropical Freshwater Research, and School ofTropical Biology, James Cook University, Townsville, Queensland 4811, Australia

M. R. CROSSLAND2

Cooperative Research Centre for Coastal Zone Estuary and Waterway Management, and Australian Centrefor Tropical Freshwater Research, James Cook University, Townsville, Queensland 4811, Australia

R. G. PEARSON3

Rainforest Cooperative Research Centre, Australian Centre for Tropical Freshwater Research, and School ofTropical Biology, James Cook University, Townsville, Queensland 4811, Australia

Abstract. The effect of different dissolved oxygen (DO) concentrations on the macroinvertebrateassemblages from 2 Australian tropical streams (1 upland, 1 lowland) was measured using artificialstream mesocosms. Responses to 5-d exposures were tested. Both the upland and lowland assem-blages showed a similar response. Most taxa tolerated all but very low DO levels (�10% saturation),although a reduction in emergence of insect taxa at intermediate levels (25–35% and 10–20% satu-ration) was observed. Mayflies showed the highest sensitivity to low oxygen conditions, and lethaleffects were observed at DO levels �20% saturation for several upland and lowland species. For othertaxa, including several Chironomidae, mortality was observed when oxygen concentrations were be-low 8% saturation. A drift response was observed only when oxygen concentrations reached nearlethal levels (�10% saturation). The lack of a drift response at DO concentrations of 25 to 35% and10 to 20% saturation indicates that, in moderately poor oxygen conditions, macroinvertebrates willremain at a location and, hence, experience sublethal effects such as suppressed emergence. It is clearthat these animals can persist in hypoxic conditions in the short term. However, because of sublethaleffects, understanding how low DO concentrations affect natural assemblages of aquatic macroinver-tebrates may require studies of populations over several generations.

Key words: dissolved oxygen concentration, hypoxia, lethal effects, mesocosm, sublethal effects,tropics.

Oxygen availability is a widely recognizedfactor influencing the composition of freshwatercommunities because it critically affects the dis-tribution of many species (Hynes 1960, Gillerand Malmqvist 1998, Dodds 2002). Dissolvedoxygen (DO) concentrations can vary spatiallyand temporally because of respiration by organ-isms, photosynthesis by plants, atmosphericlosses and gains, changes in pressure and tem-perature, and groundwater inflow (Hynes 1970,Allan 1995, Dodds 2002). Anthropogenic im-pacts have increased the frequency, duration,and intensity of hypoxia in many aquatic sys-tems, resulting in changes in community com-position and often a loss of diversity (Hynes

1 E-mail addresses: [email protected] [email protected] [email protected]

1960, Pearson and Penridge 1987). Changes inthe composition of aquatic assemblages havelong been used to indicate environmental con-ditions (e.g., Hynes 1960, Hellawell 1986, Rosen-berg and Resh 1993) because different specieshave different water-quality requirements andtolerances, such as minima in DO concentra-tions. Surprisingly, there are few studies on thehypoxia tolerance of freshwater macroinverte-brates given that 1) DO concentration is a keywater-quality indicator, 2) macroinvertebratesare used in bioassessment, and 3) the occur-rence of anthropogenically induced hypoxia iswidespread.

Aquatic macroinvertebrates possess a diversearray of structural and behavioral respiratoryadaptations (Eriksen et al. 1984), suggesting thatdifferent taxa differ in their oxygen require-

Composition03

252 [Volume 23N. M. CONNOLLY ET AL.

ments and tolerance to hypoxia. Some aquaticinsects (and early instars of others) respire bydiffusion through the cuticle, whereas some taxa(and later instars of others) with a lower surfaceto volume ratio require tracheal or spiraculargills for respiration. Other groups, such as adultelmid beetles, use a plastron, an air bubble thatfunctions as a physical gill that allows diffusionof oxygen into a tracheal system. Further, someaquatic insects have respiratory pigments (e.g.,haemoglobin in some Chironomidae and Noto-nectidae). These differences are reflected in spe-cific physiological responses to changes in DOconcentration: some species are oxygen con-formers (their internal oxygen concentrations re-flect the external environment, e.g., the mayfliesBaetis sp. and Ephemera vulgata, Fox et al. 1937,and the chironomid Tanytarsus brunnipes, Wal-she 1948), whereas others are oxygen regulators(their internal oxygen concentrations are largelyindependent of the external environment, e.g.,the mayflies Cloeon dipterum and Leptophlebiamarginata, Fox et al. 1937, and the chironomidChironomus longistylus, Walshe 1948). Therefore,we expected that these physiological factorswould interact with biophysical variables to de-termine the response and, ultimately, the distri-bution of species (Prosser and Brown 1962, Mer-ritt and Cummins 1984).

To predict how DO concentrations influencelotic macroinvertebrate distributions requiresknowledge of the specific environmental re-quirements of macroinvertebrates from a varietyof streams across geographic regions. A reviewof the literature (Table 1) showed that tests ofDO tolerance have been conducted for few taxa(mainly Ephemeroptera) and only in temperateregions. These studies were based primarily onacute, short-term experiments. They were usu-ally conducted with late-instar larvae and sug-gested a wide range of tolerance to hypoxia. Itwas evident that some macroinvertebrates couldtolerate exposure to short periods of very lowDO concentrations, perhaps to cope with ex-treme diel fluctuations in oxygen and other nat-ural processes in particular habitats (e.g., Erik-sen 1963, Dean and Richardson 1999).

We investigated the response of macroinver-tebrates from an upland and lowland stream inthe Wet Tropics World Heritage Area in north-ern Australia to low DO concentrations. Weused upland and lowland assemblages becausewe expected that they would respond different-

ly to hypoxia because macroinvertebrates inwarm, lowland streams typically are exposed toextreme diurnal cycles in DO and, late in thedry season, to deteriorating, eutrophic condi-tions exacerbated by the input of agriculturalcontaminants. Consequently, we believed thatmacroinvertebrates from lowland streamswould be more tolerant to hypoxia than thosefrom upland streams in this tropical region.

We tested respective assemblages under sim-ulated stream conditions using artificial streammesocosms, aiming to identify tolerant and sen-sitive taxa (if differences existed) and determinegeneral assemblage responses. We chose riffleassemblages for testing because we expected rif-fles to be well aerated and to contain more hyp-oxia-intolerant species than poorly aerated hab-itats, as has been shown for mayflies (Fox et al.1937) and chironomids (Walshe 1948). We ex-amined 5-d exposure to establish acute toler-ances and identify nonlethal effects. We mea-sured survival and drift. Drift is a key behav-ioral response of lotic macroinvertebrates usedas an avoidance mechanism to poor environ-mental conditions (Brittain and Eikeland 1988).We also collected emergent insects to measurethe effect of different degrees of hypoxia on de-velopmental success.

Methods

Site descriptions

The upland experiments were conducted atthe James Cook University Field Station at Pal-uma village (lat 19�00�S, long 146�11�E; altitude910 m asl). The climate at Paluma is seasonal,typically with dry winters (approximately Mayto November) and warm, wet summers (Decem-ber to April). Macroinvertebrates used in theupland experiments (Experiments 1 and 2) werecollected from a riffle in Camp Creek �5 kmfrom Paluma. It is a 2nd-order stream at �820m asl and is surrounded by tropical rainforest(Simple Notophyll Vine Forest, Tracey 1982).The stream consists of alternating riffles andpools, mostly with granitic substrata of cobbles,larger rocks, and sand. Flow is permanent butcan be low in the dry season, whereas monsoonand cyclonic rains cause flooding in the wet sea-son. Rosser and Pearson (1995) give a more de-tailed description of streams at Paluma.

The lowland experiments were conducted at

2004] 253MACROINVERTEBRATE RESPONSE TO LOW DOT

AB

LE

1.Su

mm

ary

oflit

erat

ure

ond

isso

lved

oxyg

en(D

O)

tole

ranc

eof

loti

cm

acro

inve

rteb

rate

s.M

ajor

taxa

abbr

evia

ted

inp

aren

thes

es:T

�T

rich

opte

ra,E

�E

phem

erop

tera

,P�

Ple

cop

tera

,Ch

�C

hiro

nom

idae

,Cr

�C

rust

acea

.All

DO

conc

entr

atio

nsco

nver

ted

to%

satu

rati

on,a

ssu

min

g0.

5�

S/cm

spec

ific

cond

ucta

nce.

LC50

�le

thal

conc

entr

atio

nto

kill

50%

ofth

ep

opu

lati

on.

Taxo

nTe

mp

(�C

)D

ura

tion

(d)

Let

hal

DO

(mg/

L)

Est

imat

edle

thal

DO

(%sa

tura

tion

)A

utho

ran

dre

gion

Ana

bolia

nerv

osa

(T)

Sten

ophy

lax

stel

latu

s(T

)R

hyac

ophi

lado

rsal

is(T

)H

ydro

psyc

hein

stab

ilis

(T)

Wor

mal

dia

subn

igri

a(T

)Po

lyce

ntro

pus

flavo

mac

ulat

us(T

)

16 16 16 16 16 16

Unt

ilce

ased

mov

ing

0.0

0.7

2.0

(0.7

stir

red

)3.

3(0

.9st

irre

d)

4.0

(0.4

stir

red

)0.

0

0.0

7.1

20.3

(7.1

stir

red

)33

.6(9

.2st

irre

d)

40.7

(4.1

stir

red

)0.

0

Phi

llip

son

1954

,UK

Phi

llip

son

1954

,UK

Phi

llip

son

1954

,UK

Phi

llip

son

1954

,UK

Phi

llip

son

1954

,UK

Phi

llip

son

1954

,UK

Ephe

mer

asi

mul

ans

(E)

Hex

agen

ialim

bata

(E)

13 132 2

0.44

0.27

4.2

2.6

Eri

ksen

1963

,USA

Eri

ksen

1963

,USA

Ase

llus

inte

rmed

ius

(Cr)

Hya

lella

azte

ca(C

r)

Gam

mar

usfa

scia

tus

(Cr)

Gam

mar

usps

eudo

limna

eus

(Cr)

10 20 30 10 20 30 20 10 20 30

�7

�7 6 7 7 7 0.

125

0.12

51

h�

15m

in

�0.

5�

0.5

0.5

(LC

50)

�0.

5(L

C50

)�

1.0

(LC

50)

�1.

7(L

C50

)0.

06(L

C50

)0.

06–0

.09

(LC

50)

0.06

–0.0

9(L

C50

)0.

06–0

.09

(LC

50)

�4.

4�

5.5

6.6

4.4

11.1

22.6 0.7

0.5–

0.8

0.7–

1.0

0.8–

1.2

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Spra

gue

1963

,Can

ada

Ephe

mer

ella

rotu

nda

(E)

8 200.

330.

33�

2.1

�4.

4�

17.8

�48

.6H

all

1969

,USA

Hal

l19

69,U

SAP

tero

narc

ysdo

rsat

a(P

)18

.54 30

2.2

(LC

50)

4.8

(LC

50)

23.6

63.7

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Acr

oneu

ria

lyco

rias

(P)

Hex

agen

ialim

bata

(E)

Bae

tisc

ala

uren

tina

(E)

Lept

ophl

ebia

nebu

losa

(E)

14 18.5

18.5

18.5

4 4 4 30 4

3.6

(LC

50)

1.4

(LC

50)

3.5

(LC

50)

5.0

(LC

50)

2.2

(LC

50)

35.0

15.0

37.5

53.6

23.6

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Ephe

mer

ella

subv

aria

(E)

Ephe

mer

asi

mul

ans

(E)

Tany

tars

usdi

ssim

ilis

(Ch)

18.5

18.5

18.5

4 30 4 30

3.9

(LC

50)

4.5

(LC

50)

�0.

6(L

C50

)�

0.6

(LC

50)

41.8

48.2

�6.

4�

6.4

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Hyd

rops

yche

bett

eni

(T)

21 18.5

17 10

4 4 4 4

2.9

(LC

50)

2.6

(LC

50)

2.3

(LC

50)

1.0

(LC

50)

32.7

27.9

23.9 8.9

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Neb

eker

1972

,USA

Hex

agen

ialim

bata

(E)

4–20

212.

0(8

0%m

orta

lity)

15.3

–22.

1W

inte

ret

al.1

996,

Can

ada

Para

tya

curv

iros

tris

(Cr)

152

1.0

(27%

mor

talit

y)9.

9D

ean

and

Ric

hard

son

1999

,N

ewZ

eala

ndB

aeti

str

icau

datu

s(E

)4.

514

5.0

(60–

90%

mor

talit

y)38

.7L

owel

lan

dC

ulp

1999

,C

anad

a


FIG. 1. Recirculating mesocosm used in the experiments (without the polyethylene seal).

James Cook University, Townsville (lat 19�20�S,long 146�45�E, altitude 25 m asl). Townsville hasa tropical climate with warm, dry winters andhot, moist summers. Macroinvertebrates used inthe lowland experiments (Experiments 3 and 4)were collected from a riffle in Hawkins Creek(lat 18�35�S, long 146�04�E, altitude 60 m asl)�120 km from Townsville. Hawkins Creek is a3rd-order stream with a catchment and riparianzone dominated by Mesophyll Rainforest (Tra-cey 1982). It is similar to Camp Creek in termsof instream substrata, gradient (�1%), and can-opy cover (�75%).

Stream mesocosms

Experiments were conducted in recirculatingartificial stream mesocosms made of polyethyl-ene oval channels, 350 mm wide and 400 mmdeep (Fig. 1). Sand and cobbles from each re-spective stream were washed vigorously in tapwater before being used to construct the base of2 riffle and 2 pool habitats within each meso-cosm (Fig. 1). Ten cobbles (10–12 cm diameter)freshly collected from a riffle from each respec-tive stream were transported in aerated waterand placed evenly over the top of each riffle area

within each mesocosm. These cobbles were notcleaned and provided habitat for macroinverte-brates. Water completely covered the cobbles ata depth of 80 mm above the riffle sand base.Two Aquaclear� Powerhead 802 submersibleaquarium pumps controlled water flow. Onewas positioned at the head of each riffle, main-taining a water velocity of 0.05 to 0.10 m/s overthe riffles. A venturi pipe (a narrow opening inthe water outlet that draws gas into the outletby producing a partial vacuum) sucked in air aswater passed through the pumps and aeratedthe water to �100% saturation. Mesocosms wereshaded using 75% shade cloth.

Macroinvertebrates were collected from eachstudy stream to seed the mesocosms using arandomized procedure. Cobbles, haphazardlyselected throughout a stream riffle, were liftedfrom the streambed and gently washed into acontainer using a squirt bottle. Collecting wascarried out for 20 min per container, resultingin a large number of animals with no apparentdamage or mortality. The containers were aer-ated and transported to the mesocosms. Onecontainer was randomly selected to seed eachmesocosm. This procedure was repeated dailyover 4 d prior to each experiment, resulting in

2004] 255MACROINVERTEBRATE RESPONSE TO LOW DO

each mesocosm being seeded with 4 randomlyallocated containers of invertebrates. It was as-sumed that this resulted in similar numbers ofinvertebrates in each mesocosm. No attemptwas made to measure numbers prior to each ex-periment to minimize further handling stress.The seeding of the mesocosms was randomizedto avoid systematic differences between treat-ments. Controls were used for comparisons andas a measure of variation between mesocosms.Two liters of stream leaf litter then were addedto each mesocosm to provide a source of organicmaterial (for food, fine particulate organic mat-ter, etc.). This leaf material accumulated at thehead of each riffle section and on several rockswithin each riffle in the mesocosms.

Two days after the addition of macroinverte-brates (day 6), each mesocosm was tightly cov-ered with clear plastic film. At the start of eachexperiment, the DO concentration was adjustedover a 2-h period to the selected treatment level.Each experiment ran for 5 d. DO concentrationswere manipulated by supplying air or N gasthrough the pump venturi, enabling rapid de-oxygenation by N stripping or rapid reaeration,if required, to maintain treatment oxygen levels.Oxygen stripping using N gas is an establishedmethod to reduce DO concentrations withoutany confounding toxic effects on biota (e.g.,Gaufin and Gaufin 1961, Barnhart 1995, Rich-ardson et al. 2001). CO2 was bubbled throughthe venturi when necessary to minimize pHfluctuations caused by expulsion of dissolvedCO2 by N stripping. N and CO2 gases were sup-plied to mesocosms through high-pressure reg-ulators connected to a polypropylene manifold,from which 4-mm plastic hoses were connectedto each pump. Valves on each hose were usedto control the amount of N, CO2, or air enteringeach pump, maintaining selected DO concentra-tions in each mesocosm. DO, temperature, andpH levels were monitored throughout each ex-periment using a Hydrolab� Datasonde 3 mul-tiprobe and a WTW� Multiline Meter. Treat-ment and control mesocosms used in each ex-periment were selected randomly.

At the end of each experiment, the plastic filmwas removed and the macroinvertebrates fromeach riffle were collected by washing the 10 cob-bles from each riffle through a 63-�m meshsieve, and the collected macroinvertebrates werefixed in 70% ethanol. Thus, the sampling unitwas 20 cobbles for each mesocosm. The macro-

invertebrate samples from all experiments weresorted, identified, and counted in the laboratory.

Upland stream experiments

Experiments 1 and 2 examined survival re-sponse of the upland macroinvertebrate assem-blage exposed to hypoxia. In Experiment 2, thenumber of successfully emerged insects wasalso measured to assess the effect on develop-ment. The upland experiments were run duringMay 1998 (Experiment 1) and March 1999 (Ex-periment 2). For each experiment, 6 mesocosmswere established as described above. Water tem-perature and pH were consistent between me-socosms but differed between experiments be-cause of different ambient conditions. In Exper-iment 1, water temperature ranged from 14.9 to18.3�C, and pH between 6.4 and 7.0. In Experi-ment 2, water temperature ranged from 20.2 to22.3�C, and pH between 6.2 and 7.1. These val-ues are within the range found normally inCamp Creek (RGP and NMC, unpublisheddata).

Experiments 1 (100% vs 25–35%) and 2 (100%vs 10–20%): survival and emergence. In Experi-ments 1 and 2, 3 control mesocosms were main-tained at 100% DO saturation (Experiment 1:9.4–10.1 mg/L; Experiment 2: 8.6–9.0 mg/L). Inthe other 3 mesocosms, DO concentrations werereduced to a band between 25 and 35% satu-ration (2.5–3.3 mg/L) in Experiment 1, and be-tween 10 and 20% (1.0–2.0 mg/L) in Experi-ment 2. The treatment levels were chosen fol-lowing a review of the literature (see Table 1)and based on previous experiments (Pearsonand Penridge 1992). Nebeker (1972), for exam-ple, recorded lethal effects for several ephem-eropteran species between 15% and 48.2% DOsaturation in experiments of similar duration toours (Table 1). Emerged insects were collectedfrom the underside of the plastic film and frominternal tank surfaces at the end of Experiment2 and fixed in 70% ethanol. Emerged insectswere not collected during Experiment 1.

Lowland stream experiments

Experiment 3 examined survival responseand emergence in the lowland stream macro-invertebrate assemblage. Experiment 4 exam-ined behavioral responses (i.e., drift rate) in thelowland stream assemblage. The lowland ex-


periments were run during May 1999 (Experi-ment 3) and August 1999 (Experiment 4). Thelarger facilities at James Cook University en-abled us to run 12 mesocosms simultaneouslyfor each experiment. The distance from the me-socosm facility to Hawkins Creek precluded theuse of Hawkins Creek water in these experi-ments. Instead, either dechlorinated municipaltown water (Experiment 3) or water collectedfrom a tributary in an undisturbed catchmentupstream of the facility (Experiment 4) wasused. Water temperatures and pH were consis-tent between mesocosms during each experi-ment (Experiment 3: 17.6–29.0�C, pH 6.4–7.4;Experiment 4: 21.1–27.5�C, pH 6.7–7.6), similarto natural conditions in Hawkins Creek (RGPand MRC, unpublished data).

Experiment 3 (100% vs 25–35%, 10–20%, and2–8%): survival and emergence. In Experiment 3,12 mesocosms were established as describedabove and were distributed among the controls(3 replicates), with DO maintained at �100%saturation, and 3 low DO treatments (3 repli-cates each). These treatments were: 25 to 35%(1.54–2.91 mg/L), 10 to 20% (0.78–1.60 mg/L),and 2 to 8% (0.19–0.74 mg/L). Treatment levelsfor Experiment 3 were chosen to match thoseused in the upland experiments (Experiments 1and 2), with the addition of a lower saturationtreatment (2–8% saturation) because of the lackof significant mortality observed in Experiments1 and 2. Emerged insects were collected fromthe underside of the plastic film and from in-ternal tank surfaces and fixed in 70% ethanol.

Experiment 4 (100% vs 50–60%, 25–35%, and�10%): drift response. In Experiment 4, 12 me-socosms were established as in Experiment 3.The 12 mesocosms were distributed among thecontrols (3 replicates) with DO maintained at�100% saturation, and 3 low DO treatments (3replicates each). The treatments were: 50 to 60%(3.99–5.15 mg/L), 25 to 35% (2.02–2.84 mg/L),and �10% (�0.87 mg/L). These treatment levelswere just above the treatment levels used in Ex-periment 3 to determine whether a drift re-sponse would be detectable before effects onbenthic densities were apparent. During this ex-periment, drifting macroinvertebrates were col-lected at 6-h intervals on day 1 and day 3 and12-h intervals on days 2, 4, and 5 with a driftnet (12.5 cm x 10.5 cm x 10.5 cm opening, 63-�m mesh) placed at the downstream end ofeach riffle. Drift nets were removed and re-

placed through a small sealable opening in theplastic film. This procedure was done quicklyand did not affect DO concentrations in the me-socosm water. Macroinvertebrates were exam-ined for movement to ensure that they werealive when collected and then fixed in 70% eth-anol.

Statistical analysis

Counts of macroinvertebrates were trans-formed by log (x � 1) prior to analysis to nor-malize data distributions (Zar 1999). Whentransformed data met assumptions of normalityand homogeneity of variances, analyses weredone using t-tests or analysis of variance (AN-OVA) as appropriate. When transformed datafailed to meet these assumptions, the nonpara-metric Kruskal–Wallis ANOVA was used on rawdata. Significance was set at p � 0.05. Bonferronicorrections were not used in these analyses be-cause a separate hypothesis was tested for eachtaxon, and the same mean was not used morethan once (unlike multiple pairwise compari-sons where a Bonferroni correction would be re-quired). Post-hoc tests were unnecessary toidentify significant treatment effects because ofobvious differences among treatments. It wasexpected that treatment differences would notbe detectable for rare taxa but these taxa wereincluded to indicate overall trends and to de-scribe the macroinvertebrate assemblage beingtested.

Results

Upland vs lowland assemblages

A diverse macroinvertebrate fauna was pres-ent in the mesocosms in both the upland andlowland experiments, comprising several abun-dant taxa and many less numerous taxa (Tables2, 3). The 2 assemblages were similar to eachother and to those found in rainforest streamsat Paluma (Rosser 1999, Rosser and Pearson1995, Pearson and Connolly 2000) and in Haw-kins Creek (RGP and MRC, unpublished data).The 2 assemblages had 38 taxa in common outof a combined 62. Twelve taxa occurred only inthe upland assemblage and 12 taxa occurredonly in the lowland assemblage. At the familylevel, the trichopterans were the most differentwith the upland containing 8 unique families.


At the genus and species levels, the mayfly as-semblages were identical except for the occur-rence of Neboissophlebia sp. ‘NQ1’ in the low-lands (it was previously collected in the uplandsbut occurred in low abundance; F. Christidis,James Cook University, personal communica-tion), the Genus WT sp. 1 in the uplands (onlyknown from Paluma; F. Christidis, personalcommunication), and the Genus WT sp. 2 in thelowlands (not present in the Paluma area; F.Christidis, personal communication). The chi-ronomids were more diverse in the lowland as-semblage but still had many species in commonwith the upland assemblage. The upland assem-blage was dominated by Hydracarina, Baetidae,the chironomids Echinocladius martini and Nilo-tanypus sp., and larval elmids. The lowland as-semblage was dominated by Simuliidae, the chi-ronomids Thienemaniella ‘alpha’, T. ‘delta’, andRheotanytarsus ‘alpha’, with the Hydracarina andAustrophlebioides sp. common. Atyids occurred inthe lowland but not the upland assemblages.

The response to comparable DO treatmentswas similar in the upland and lowland assem-blages, with some taxa increasing in benthicdensities because of suppressed emergence inthe 25 to 35% and 10 to 20% saturations. Forexample, densities of the mayfly Nousia sp.‘NQ2’ increased in the 25 to 35% saturationtreatments and then declined sharply in the 10to 20% saturation treatments in both the uplandand lowland experiments (Tables 2, 3). Similarly,when densities of chironomids were pooled theyincreased in the 25 to 35% and 10 to 20% satu-ration treatments in the upland and lowland ex-periments.

Upland stream experiments

Experiment 1 (100% saturation vs 25–35% sat-uration): survival. A total of 37 taxa was col-lected from control and treatment mesocosms inExperiment 1. Control mesocosms had an av-erage of �22 taxa each compared to treatmentmesocosms with an average of 24 taxa each (Ta-ble 2). Most taxa showed similar mean numbersbetween the control and treatment mesocosmsin Experiment 1 (Table 2), indicating that an ox-ygen saturation of 25 to 35% had little effect onsurvival. However, 4 taxa (Hydracarina, Nousiasp. ‘NQ2’, Thienemanniella ‘alpha’, and Elmidaeadults) showed significantly higher densities inthe low DO treatment. The higher densities of

these taxa contributed to a higher mean totaldensity in the 25 to 35% saturation DO treat-ment (Table 2, Fig. 2); however, this differencewas not statistically significant.

Experiment 2 (100% vs 10–20%): survival andemergence. A total of 42 taxa was collected fromcontrol and treatment mesocosms in Experi-ment 2. Both control and treatment mesocosmshad an average of �26 taxa (Table 2). The lowDO treatment (10–20% saturation) in Experi-ment 2 affected several taxa. Five dipterans (in-determinate small instars of Chironomidae wereregarded as a taxon) showed significant increas-es in density, whereas 2 ephemeropteran taxashowed significant declines in the low DO treat-ment. Most other taxa were more numerous intreatment than control samples. The mean totaldensity of individuals was also higher in thetreatment than control samples (although thedifference was not statistically significant) (Table2, Fig. 2). However, counts of emerged insectsshowed an �30-fold decrease in abundance intreatment mesocosms relative to controls (Fig.2).

Lowland stream experiments

Experiment 3 (100% vs 25–35%, 10–20% and 2–8%): survival and emergence. A macroinverte-brate fauna similar to that in Hawkins Creek(RGP and MRC, unpublished data) was estab-lished in the mesocosms in the lowland exper-iments. Approximately 30 taxa were collectedfrom each of the 95 to 100%, 25 to 35% and 10to 20% saturation mesocosms, compared to only15 taxa being collected from the 2 to 8% satu-ration mesocosms (Table 3). The responses ofseveral taxa were statistically significant (Table3, Fig. 3). Effects were of 4 main types: 1) agradual decline with decreasing DO (e.g., Pse-phenidae, Fig. 3A); 2) no change until the 2 to8% treatment (e.g., Atyidae and Platyhelmin-thes, Fig. 3B) or the 10 to 20% treatment (e.g.,WT sp. 2, Fig. 3C); 3) an increase in the 25 to35% treatment, followed by a decline in the 10to 20% treatment (e.g., the ephemeropterans,Austrophlebioides sp., Nousia sp. ‘NQ2’, and Caen-idae, Fig. 3C) or the 2 to 8% treatment (e.g., thechironomids, Thienemanniella ‘alpha’, T. ‘delta’,and Rheocricotopus ‘alpha’, Fig. 3D); and 4) nosignificant change in numbers with changes inoxygen saturation (Table 3). No significantchange in numbers with changes in oxygen sat-


TABLE 2. Density of macroinvertebrates (no./20 cobbles SE) surviving in upland mesocosms after 5 d indifferent dissolved oxygen (DO) treatments (indicated by % saturation). 100% saturation was the control. – �no individuals were present in the sample. p-values �0.05 are highlighted in bold. Taxonomic name codes referto the Australian National Insect Collection (Chironomidae—P. Cranston, University of California, Davis, Cal-ifornia, personal communication) and the Museum of Victoria Collection (Leptophlebiidae—J. Dean, Environ-ment Protection Agency, Victoria, Australia, personal communication).

DO saturation

Taxon

Experiment 1

100% 25–35% p

Experiment 2

100% 10–20% p

TurbellariaOligochaetaCopepoda

CyclopodaHydracarina

–11.0 10.5

–12.6 3.1

–8.6 2.9

–43.6 16.7

–0.841

–0.038

–10.0 3.6

–83.6 46.8

0.6 0.620.0 2.6

0.3 0.3100.6 57.7

0.3740.089

0.3740.830

PlecopteraGripopterygidaeEustheniidae

14.6 7.1–

9.6 2.0–

0.539–

17.0 8.67.0 6.5

19.3 7.20.3 0.3

0.8460.364

EphemeropteraBaetidaeCaenidaeLeptophlebiidae

Atalomicria sexfasciataAustrophlebioides sp.Genus K sp. ‘AV2’Nousia sp. ‘NQ1’Nousia sp. ‘NQ2’‘WT sp. 1’

52.6 6.8–

–0.3 0.3

––

5.0 0.0–

78.3 8.61.3 0.8

––––

12.6 2.70.3 0.3

0.0810.205

–0.374

––

0.0480.374

154.6 34.011.0 6.1

1.0 0.5–

1.3 0.80.3 0.3

22.3 0.60.3 0.3

14.0 7.312.3 2.6

––––

3.0 1.10.3 0.3

0.0160.851

0.158–

0.2050.3740.0001.000

OdonataAeshnidaeAmphipterygidaeLibellulidae/Cordulidae

1.3 0.80.3 0.3

–

1.0 0.51.3 0.8

–

0.7680.349

–

–0.3 0.30.6 0.3

–0.3 0.30.3 0.3

–1.0000.519

MegalopteraCorydalidae

LepidopteraPyralidae

–

1.0 1.0

–

–

–

0.374

–

–

0.3 0.3

–

0.946

–Trichoptera

AntipodoeciidaeCalamoceratidaeCalocidae/HelicocidaeConoesucidaeEcnomidaeGlossosomatidaeHelicopsychidae

11.0 4.90.6 0.3

59.6 16.60.6 0.3

–2.0 1.5

–

9.6 1.70.6 0.6

57.6 3.5––

2.0 0.5–

0.8121.0000.9120.116

–1.000

–

5.6 3.7–

1.0 0.5–––

48.6 12.4

15.6 7.0–––

0.6 0.6–

82.0 19.8

0.278–

0.158–

0.374–

0.228HydrobiosidaeHydropsychidaeHydroptilidaeLeptoceridaeOdontoceridaePhilorheithridaePolycentropodidae

1.3 1.30.3 0.31.3 1.3

––

0.3 0.34.6 3.2

0.6 0.60.3 0.33.0 1.7

––

1.3 0.34.3 3.8

0.6781.0000.488

––

0.1010.951

0.3 0.3–

8.0 2.60.3 0.30.3 0.31.3 0.84.3 0.6

––

10.6 0.81.6 1.2

–2.3 1.23.6 1.4

0.374–

0.3930.3450.3740.5390.698

DipteraChironomidae

Corynoneura sp.Cricotopus sp.Dicrotendipes ‘alpha’Echinocladius martini

3.4 1.923.3 8.9

4.5 4.525.3 10.1

3.7 1.315.2 10.7

2.2 1.124.6 5.2

0.8990.5920.6420.956

–7.3 2.2

12.6 1.378.1 16.1

–27.2 4.58.8 4.3

171.1 33.2

–0.0180.4400.066


TABLE 2. Continued.

DO saturation

Taxon

Experiment 1

100% 25–35% p

Experiment 2

100% 10–20% p

Nilotanypus sp.Orthoclad ‘beta’Rheotanytarsus sp.Riethia sp.Tanytarsus sp.Thienemanniella ‘alpha’Indeterminate small

instars

37.3 12.61.5 1.5

14.4 3.3–

0.3 0.36.2 1.3

36.3 14.5

43.8 11.62.3 1.27.8 1.50.4 0.40.7 0.3

15.5 2.5

19.0 2.6

0.7250.6880.1440.3740.5450.033

0.305

42.9 4.76.5 3.1

13.9 4.6–

5.8 1.216.2 16.2

7.0 1.5

84.4 10.331.5 0.425.7 7.3

–9.7 2.5

–

57.0 8.6

0.0220.0020.247

–0.2410.374

0.005EmpididaeSimuliidaeTipulidae

0.3 0.315.0 4.1

–

–27.3 5.1

–

0.3740.137

–

–6.0 3.00.3 0.3

–16.6 4.92.0 0.0

–0.1370.007

ColeopteraElmidae (larvae)Elmidae (adults)HydrophilidaePsephenidaeScirtidae

29.0 10.411.0 3.0

–––

33.6 2.321.6 1.8

–0.3 0.3

–

0.6850.041

–0.374

–

69.0 5.233.6 3.3

––

0.3 0.3

75.3 14.848.0 18.50.3 0.31.0 0.50.6 0.3

0.7090.4900.3740.1580.519

Mean number of taxa 21.7 1.4 24.0 0.6 0.210 26.3 1.9 26.3 0.3 1.000Mean total density of

individuals 400.0 82.3 465.7 53.8 0.541 706.7 50.0 874.0 106.8 0.229

uration was detected for many of the less-nu-merous taxa, but it is unclear what this responserepresents. However, many of these uncommontaxa were absent in the 2 to 8% treatment. Noneof the more numerous taxa showed a type 4 re-sponse.

The response in mean total density followedthe 2nd part of the type 3 pattern describedabove because the high numbers of Thieneman-niella ‘alpha’ and T. ‘delta’ dominated total ben-thic densities (Table 3, Fig. 4). Although benthicdensities were higher in the 25 to 35% and 10to 20% treatments compared to the control, thenumber of emerged insects was much lower inthese treatments (Fig. 4). Benthic densities werelower in the control mesocosms partly becauseof a higher emergence rate in the control me-socosms compared to treatment mesocosms.Consequently, more animals had remained inthe 25 to 35% and 10 to 20% treatment meso-cosms, resulting in higher benthic densities. Asubstantial decline in benthic densities in the 2to 8% treatment indicated that mortality oc-curred, resulting in few individuals emerging.

Experiment 4 (100% vs 50–60%, 25–35%, and�10%): drift response. There was no significant

difference in total drift over the duration of theexperiment between control mesocosms and the50 to 60% or 25 to 35% oxygen saturation treat-ments. However, ANOVA indicated that driftwas enhanced in the �10% saturation treatment(F3,8 � 15.91, p � 0.001; Fig. 5). Four taxashowed increases in drift rate in the �10% treat-ment (Acarina: F3,8 � 12.69, p � 0.002; Caenidae:F3,8 � 67.59, p � 0.001; Chironomidae: F3,8 �23.84, p � 0.001; and Leptophlebiidae: F3,8 �12.56, p � 0.002; Fig. 6). Chironomidae and Lep-tophlebiidae had high drift on the first day (Fig.6A, B), whereas drift accumulation was moregradual for the Caenidae (Fig. 6C) and, to a less-er extent, the Acarina (Fig. 6D). Drift remainedvery low in the control and intermediate oxygentreatments throughout Experiment 4 (Fig. 5).

Discussion

Survival

There were strong similarities between theupland and lowland macroinvertebrate assem-blages, with some differences in presence/ab-sence of taxa, or of relative abundances of


TABLE 3. Density of macroinvertebrates (no./20 cobbles SE) surviving in lowland mesocosms after 5 din different dissolved oxygen (DO) treatments (indicated by % saturation). 100% saturation was the control. –� no individuals were present in the sample. p values �0.05 are highlighted in bold. Taxonomic name codesrefer to the Australian National Insect Collection (Chironomidae—P. Cranston, University of California, Davis,California, personal communication) and the Museum of Victoria Collection (Leptophlebiidae—J. Dean, Envi-ronment Protection Agency, Victoria, Australia, personal communication).

Taxon

Experiment 3

95–100% 25–35% 10–20% 2–8% p

TurbellariaOligochaetaCrustacea

AtyidaeCaridina sp.

Hydracarina

3.7 0.30.3 0.3

28.0 1.229.3 6.9

3.0 0.6–

26.0 2.022.3 2.6

3.7 1.81.3 1.3

25.7 1.918.7 0.3

–3.0 1.2

–2.7 0.7

0.0030.113

�0.001�0.001

PlecopteraGripopterygidae 1.7 0.9 1.3 0.3 – – 0.064

EphemeropteraBaetidaeCaenidae

8.3 0.9–

6.3 0.95.7 0.9

4.3 3.03.3 2.3

1.3 0.70.7 0.3

0.1020.040

LeptophlebiidaeAtalomicria sexfasciataAustrophlebioides sp.Genus K sp. ‘AV2’Neboissophlebia sp. ‘NQ1’Nousia sp. ‘NQ1’Nousia sp. ‘NQ2’‘WT sp. 2’

3.0 1.520.7 3.2

––

1.3 1.36.3 2.3

14.7 4.2

2.0 1.222.7 3.8

–0.7 0.7

–11.0 5.013.3 0.9

1.7 0.72.3 1.5

–––

3.0 0.68.0 4.0

––

0.3 0.3––

0.7 0.3–

0.111�0.001

0.3920.3920.3920.0150.029

OdonataAeshnidaeDiphlebiidaeLibellulidae/Cordulidae

0.7 0.71.0 1.00.7 0.3

0.3 0.30.7 0.30.3 0.3

0.3 0.31.0 0.6

–

–––

0.7370.4410.214

MegalopteraCorydalidae 1.3 1.3 – 1.0 0.6 – 0.269

LepidopteraPyralidae 1.0 1.0 0.7 0.7 1.7 0.3 – 0.236

TrichopteraHelicopsychidaeHydropsychidaeHydroptilidaeLeptoceridaeOdontoceridaePhilopotamidaePolycentropodidaeIndeterminate small instars

0.7 0.314.3 6.3

1.0 0.00.3 0.31.0 0.6

–––

0.7 0.314.3 7.5

1.3 0.90.7 0.35.3 4.43.3 1.80.7 0.34.7 4.2

–15.3 5.92.3 0.92.3 1.32.3 1.51.7 1.7

–1.0 0.6

–0.7 0.7

–0.3 0.30.3 0.3

–––

0.1390.1120.0580.1810.5410.2240.0860.153

DipteraChironomidae

?Apsectrotanypus ‘alpha’Corynoneura ‘alpha’Cricotopus ? brevicornisDicrotendipes ‘alpha’Echinocladius martiniEukiefferiella ‘alpha’Nanocladius ‘alpha’Nilotanypus ‘alpha’Orthoclad ‘beta’?Paratanytarsus ‘alpha’

–10.0 2.5

3.7 1.81.3 0.90.3 0.30.7 0.31.3 0.95.7 2.0

––

–13.7 6.73.7 2.73.7 2.30.3 0.3

–3.7 1.56.7 3.7

–0.3 0.3

–32.7 5.49.7 0.95.7 2.0

–0.7 0.3

16.0 3.03.7 0.90.3 0.3

–

0.3 0.32.3 0.32.0 0.60.3 0.3

–0.3 0.3

–0.3 0.30.3 0.3

–

0.3920.0010.1640.1470.5320.3260.0010.0310.5320.392


TABLE 3. Continued.

Taxon

Experiment 3

95–100% 25–35% 10–20% 2–8% p

Polypedilum ? oresitrophusProcladius ‘alpha’Rheocricotopus ‘alpha’Rheotanytarsus ‘alpha’Riethia ‘beta’Tanytarsus ‘epsilon’Thienemanniella ‘alpha’Thienemanniella ‘delta’

1.7 0.3–

14.7 2.282.0 9.2

–0.3 0.3

69.7 33.297.3 33.2

0.3 0.30.3 0.3

39.7 9.297.7 10.5

––

232.3 16.2220.7 79.0

4.3 1.9–

78.0 1.5108.7 7.1

1.3 1.30.3 0.3

284.7 20.2230.3 21.3

1.3 0.9–

3.3 1.89.7 3.5

––

8.3 2.98.0 3.2

0.0640.3920.016

�0.0010.3920.532

�0.001�0.001

SimuliidaeTipulidae

220.0 78.58.0 3.6

395.3 133.44.7 1.8

356.7 122.112.7 8.8

14.7 8.70.7 0.7

0.0030.134

ColeopteraElmidae (larvae)Elmidae (adults)HydrophilidaePsephenidaeScirtidae

5.7 0.70.7 0.3

–7.7 1.53.7 1.2

1.7 1.20.3 0.3

–5.0 1.22.7 2.7

7.3 3.52.7 2.20.3 0.31.7 1.26.7 4.3

–––––

0.0450.3060.3920.0020.133

Mean number of taxaMean total density of

individuals30.7 1.7

704.0 127.830.0 1.2

1254.3 206.032.0 2.5

1353.3 206.814.7 2.169.0 13.7

0.0000.002

FIG. 2. Mean (� SE) total density of benthic macroinvertebrates remaining in, and mean (� SE) number ofinsects emerging from, upland mesocosms (Experiments 1 and 2) after 5 d in different dissolved oxygen (DO)treatments (indicated by % saturation). 100% saturation was the control. Emergent insects were not measuredin Experiment 1.


FIG. 3. Mean ( SE) total density of macroinvertebrates for different taxonomic groups in different dissolvedoxygen (DO) treatments in lowland Experiment 3 (only taxa that had a statistically significant response areplotted). 100% saturation was the control. A.—Coleoptera taxa. B.—Hydracarina, Atyidae, and Turbellaria. C.—Ephemeroptera taxa. D.—Chironomidae taxa.

shared taxa. The similarities in the assemblagesprobably reflects the similarity between sites inmorphology and shade, the relatively small dif-ference in elevation between upland and low-

land sites, and the close proximity of the low-land site to the base of the range. The macro-invertebrate assemblages in our study showedhigh tolerance to moderate hypoxia, at least over


FIG. 3. Continued.

the 5-d duration of our experiments. However,all taxa were intolerant to DO saturation �10%,indicating there was a clear threshold of toler-ance for most taxa. The common taxa madecomparisons between upland and lowland as-semblages possible at the species level, and itwas found that the treatment response was

largely consistent between the 2 assemblages.Mayflies showed the highest sensitivity to lowoxygen concentrations; for example, lethal ef-fects occurred at DO levels �20% saturation forthe leptophlebiid Nousia ‘NQ2’ and baetids inthe upland experiments, and for the leptophle-biids Austrophlebioides sp., Nousia ‘NQ2’, and


FIG. 4. Mean (� SE) total density of benthic macroinvertebrates remaining in, and mean (� SE) number ofinsects emerging from, lowland mesocosms (Experiment 3) after 5 d in different dissolved oxygen (DO) treat-ments (indicated by % saturation). 95 to 100% saturation was the control.

possibly WT sp. 2 in the lowland experiments.For most other taxa, including the Chironomi-dae, distinct reductions in numbers were not ob-served in the 25 to 35% or 10 to 20% DO treat-ments in the upland or lowland experiments butwere evident in the 2 to 8% DO treatment in thelowland experiment.

Emergence suppression

Suppressed emergence was observed formany taxa at 25 to 35% saturation (lowland) and10 to 20% saturation (lowland and upland). Aninitial unexpected response was higher densitiesin benthic samples for several taxa in the 25 to35% and 10 to 20% oxygen treatments in bothupland and lowland experiments. The moststriking example of this response was observedfor the chironomids Thienemanniella ‘alpha’ andT. ‘delta’, and for several Ephemeroptera in thelowland experiment. This response correspond-ed to reduced numbers of emergent animals inthese treatments, indicating that the intermedi-ate oxygen treatments imposed a sublethalstress on these animals by suppressing their de-

velopment. The increases in benthic abundanceindicate that this effect was largely caused by adelay in pupation rather than mortality of in-dividuals. The suppression of emergence in-creased with further reductions in DO, and ahigh number of animals perished in the 2 to 8%treatment.

Several studies have documented partial mor-tality or sublethal effects for macroinvertebratesexposed to low DO concentrations (Eriksen1963, Nebeker 1972, Nebeker et al. 1992, 1996,Winter et al. 1996, Dean and Richardson 1999).Nebeker (1972) found that emergence of may-flies was inhibited at DO concentrations consid-erably greater than 96-h LC50 (lethal concentra-tions required to kill 50% of the population) val-ues; concentrations of �80% saturation resultedin a significant reduction in emergence. In con-trast, Chironomidae larvae emerged and repro-duced successfully when exposed to oxygenconcentrations of �7% saturation. In the longterm, a delay in pupation and emergence prob-ably influences reproductive success, productiv-ity and, ultimately, persistence at a location. Forexample, sublethal effects such as reductions in


FIG. 5. Mean (� SE) total number of macroinvertebrates drifting (n � 3) over 5 d in mesocosms subjectedto different dissolved oxygen (DO) treatments (Experiment 4). 95 to 100% saturation was the control.

rates of feeding and growth in response to low-ered DO have been documented for temperatebiota (Winter et al. 1996, Lowell and Culp 1999).The overall impact clearly will depend on theduration of exposure to the oxygen stress.

The number of emergent insects indicatedsublethal effects in semiaquatic taxa but it wasnot obvious whether fully aquatic taxa also ex-perienced these effects. The Atyidae and theHydracarina, persisted until the 2 to 8% oxygensaturation treatment and then declined dramat-ically (Experiment 3), suggesting a lethalthreshold at this level. However, the Hydracar-ina and Elmidae adults had significantly highernumbers in the low oxygen treatments in theupland experiments (Experiments 1 and 2). It ispossible that these animals had moved fromother parts of the mesocosms and aggregated inthe faster current of the riffles to compensate forthe detrimental effects of the low oxygen con-ditions. Such movement also may have contrib-uted to the high numbers of insects sampled inthe 25 to 35% and 10 to 20% saturation treat-ments in both upland and lowland experiments.For example, Wiley and Koler (1980) found thatmayfly larvae responded to low oxygen condi-

tions by moving to more current-exposed posi-tions, and Lowell and Culp (1999) found thatexposing the mayfly Baetis tricaudatus to 5 mgO2/L for 14 d at 4.5�C (�40% saturation) re-sulted in 2 to 3x more larvae moving into re-gions of high current velocity compared withlarvae exposed to 11 mg O2/L (�100% satura-tion).

Drift response

An increased drift rate was observed for alltaxa only at the lowest DO treatment (�10% sat-uration). Many macroinvertebrates respond todeteriorating water quality (e.g., acidification,sedimentation, and pollutants) by increasingtheir drift rate (Brittain and Eikeland 1988). Thelack of a significant drift response to interme-diate DO concentrations, despite evidence ofsublethal effects, suggests that animals wait,perhaps at a reduced metabolic rate (Eriksen1963, Kapoor and Griffiths 1975), for conditionsto improve. However, the delay in drift responsemay further disadvantage these animals in thatthey will remain at a location in moderatelypoor conditions, hence increasing their exposure


FIG. 6. Cumulative mean ( SE) number of individuals of 4 taxa drifting in mesocosms from Experiment4 during 5-d exposures to oxygen saturation and various degrees of hypoxia. A.—Chironomidae. B.—Lepto-phlebiidae. C.—Hydracarina. D.—Caenidae. 95 to 100% saturation was the control. Dark bars represent nighttime. 95–100%: �—�; 50–60%:▫—▫; 25–35%: —; and �10% saturation: ●—●.


FIG. 6. Continued.

and suffering nonlethal effects if conditions donot improve. Few data are available regardingthe drift response of macroinvertebrates to lowlevels of DO. Wiley and Kohler (1980) found thatmayfly larvae exposed to 15-min episodes of

low DO showed considerable interspecific vari-ation in response. The most sensitive species ac-tively entered the drift at �60% saturation,whereas the most tolerant species did not re-spond until �10% saturation.


Temperate and tropical comparisons

It might be postulated that tropical stream in-vertebrates would be more susceptible to hyp-oxia than those from cooler climates because ofpotentially higher water temperatures and ratesof metabolism. For example, Hall (1969) and Ne-beker (1972) showed that a much higher DO sat-uration is required to maintain survival at hightemperatures than would be explained by oxy-gen solubility alone for the ephemeropteranEphemerella rotunda and the trichopteran Hydrop-syche betteni (Table 1). However, this suggestionis not supported by our results because thetropical assemblages in our experiments wereresistant to hypoxic stress. This finding wasconsistent from uplands to lowlands even in asingle species; for example, the response of Nou-sia sp. ‘NQ2’ was similar in upland and lowlandexperiments carried out at different tempera-tures (14.9–18.3�C and 20.2–22.3�C in uplandExperiments 1 and 2, respectively; 17.6–29.0�Cin lowland Experiment 3). Acclimatization to aparticular temperature before the experimentmay have a strong effect on tolerance (Nagelland Fagerstrom 1978), perhaps explaining someof the response differences observed by Hall(1969) and Nebeker (1972) and the similarity inresponse of Nousia sp. ‘NQ2’ in upland and low-land experiments. This hypothesis may be sup-ported by Walshe’s (1948) findings that chiron-omid larvae become more independent of en-vironmental oxygen after some hours of adap-tation to low DO in glass tubes.

In conclusion, studies from a variety of loca-tions and on a variety of taxa all show that se-vere reductions in DO saturation are requiredto cause mortality in the short term (Table 1).However, comparisons are hampered by the dif-ferent conditions of each study (e.g., tempera-ture, current velocity, and duration). For exam-ple, it is difficult to determine whether the am-phipods tested by Sprague (1963) are more tol-erant of hypoxia than the aquatic insects testedby Nebeker (1972) because Nebeker’s (1972) ex-perimental system provided some flow in testchambers, whereas Sprague’s (1963) did not.Only Phillipson (1954) repeated tests with andwithout current, finding several trichopteransable to tolerate much lower DO saturation whenwater current was provided.

If we compare our results to experiments list-ed in Table 1 that provided flow (Phillipson

1954, Hall 1969, Nebeker 1972, Winter et al.1996, Dean and Richardson 1999, and Lowelland Culp 1999), the response of the assemblagesin our experiments are consistent. For example,the levels observed by Nebeker (1972) to inducemortality in several mayflies species (15–48%saturation) and the chironomid Tanytarsus dissi-milis, (�6.43% saturation) are consistent withour results for both the upland and lowlandstream assemblages. The hypothesis that themacroinvertebrates from the lowland streamwould be more tolerant of hypoxia was not sup-ported by our results because both assemblagesshowed a similar response. We detected suble-thal effects for mayflies in the 25 to 35% satu-ration treatments. Mortality occurred in the 10to 20% saturation treatments for the mayfliesbut mortality did not occur in the chironomidsuntil the 2 to 8% saturation treatment. The sup-pression of emergence by insect taxa suggestedthat development was impeded at reduced DOconcentrations well before lethal concentrationswere reached. Further, the lack of drift responseat DO concentrations of 25 to 35% and 10 to 20%saturation indicates that in moderately poor ox-ygen conditions these animals will probably re-main at a location and hence experience suble-thal effects that are likely to be detrimental overthe longer term. Therefore, although it is clearthat these animals can persist in hypoxic con-ditions in the short term, to understand the ef-fects of low DO concentrations on natural as-semblages of aquatic macroinvertebrates mayrequire studies of populations over several gen-erations.

Acknowledgements

We thank Linda Davis, Joanne Burden, andGordon Kovacs for providing field and labora-tory assistance, including macroinvertebratesorting and identification; and Brendan McKieand Faye Christidis for confirmation of chiron-omid and leptophlebiid identifications, respec-tively. Barry Butler helped with our understand-ing of water-quality issues. We also thank Chris-topher Robinson and 2 anonymous reviewersfor constructive suggestions for improvement ofthe paper, and Sean Connolly, Anthony Dell,Mia Hoogenboom, and Maria Dornelas for ad-ditional comments. Financial support was pro-vided by the Rainforest Cooperative Research


Centre, Land and Water Australia, and the Sug-ar Research and Development Corporation.

Literature Cited

ALLAN, J. D. 1995. Stream ecology: structure and func-tion of running waters. Kluwer Academic Pub-lishers, London, UK.

BARNHART, M. C. 1995. An improved gas-strippingcolumn for deoxygenating water. Journal of theNorth American Benthological Society 14:347–350.

BRITTAIN, J. E., AND T. J. EIKELAND. 1988. Invertebratedrift—A review. Hydrobiologia 166:77–93.

DEAN, T. L., AND J. RICHARDSON. 1999. Responses ofseven species of native freshwater fish and ashrimp to low levels of dissolved oxygen. NewZealand Journal of Marine and Freshwater Re-search 33:99–106.

DODDS, W. K. 2002. Freshwater ecology: concepts andenvironmental applications. Academic Press, SanDiego, California.

ERIKSEN, C. H. 1963. Respiratory regulation in Ephem-era simulans Walker and Hexagenia limbata (Ser-ville) (Ephemeroptera). Journal of ExperimentalBiology 40:455–467.

ERIKSEN, C. H., V. H. RESH, S. S. BALLING, AND G. A.LAMBERTI. 1984. Aquatic insect respiration. Pages27–37 in R. W. Merritt and K. W. Cummins (edi-tors). An introduction to the aquatic insects ofNorth America. 2nd edition. Kendall/Hunt Pub-lishing Company, Dubuque, Iowa.

FOX, H. M., C. A. WINGFIELD, AND B. G. SIMMONS.1937. The oxygen consumption of ephemeridnymphs from flowing and from still waters in re-lation to the concentration of oxygen in the water.Journal of Experimental Biology 14:210–218.

GAUFIN, R. F., AND A. R. GAUFIN. 1961. The effect oflow oxygen concentrations on stoneflies. UtahAcademy Proceedings 38:57–64.

GILLER, P. S., AND B. MALMQVIST. 1998. The biology ofstreams and rivers. Oxford University Press, Ox-ford, UK.

HALL, C. A. S. 1969. Mortality of the mayfly nymph,Ephemerella rotunda (Morgan), at various low dis-solved oxygen concentrations. Journal of the Eli-sha Mitchell Scientific Society 85:34–39.

HELLAWELL, J. M. 1986. Biological indicators of fresh-water pollution and environmental management.Elsevier Applied Science Publishers, London, UK.

HYNES, H. B. N. 1960. The biology of polluted waters.Liverpool University Press, Liverpool, UK.

HYNES, H. B. N. 1970. The ecology of running waters.University of Toronto Press, Toronto, Ontario.

KAPOOR, N. N., AND W. GRIFFITHS. 1975. Oxygen con-sumption of nymphs of Pasganophora capitata (Pic-tet) (Plecoptera) with respect to body weight and

oxygen concentrations. Canadian Journal of Zo-ology 53:1089–1092.

LOWELL, R. B., AND J. M. CULP. 1999. Cumulative ef-fects of multiple effluent and low dissolved oxy-gen stressors on mayflies at cold temperatures.Canadian Journal of Fisheries and Aquatic Scienc-es 56:1624–1630.

MERRITT, R. W., AND K. W. CUMMINS (EDITORS). 1984.An introduction to the aquatic insects of NorthAmerica. 2nd edition. Kendall/Hunt PublishingCompany, Dubuque, Iowa.

NAGELL, B., AND T. FAGERSTROM. 1978. Adaptationsand resistance to anoxia in Cloeon dipterum(Ephemeroptera) and Nemoura cinerea (Plecop-tera). Oikos 30:95–99.

NEBEKER, A. V. 1972. Effect of low oxygen concentra-tion on survival and emergence of aquatic insects.Transactions of the American Fisheries Society 4:675–679.

NEBEKER, A. V., S. T. ONJUKKA, D. G. STEVENS, AND G.A. CHAPMAN. 1996. Effect of low dissolved oxy-gen on aquatic life stages of the caddisfly Clisto-ronia magnifica (Limnephilidae). Archives of En-vironmental Contamination and Toxicology 31:453–458.

NEBEKER, A. V., S. T. ONJUKKA, D. G. STEVENS, G. A.CHAPMAN, AND S. E. DOMINGUEZ. 1992. Effect oflow dissolved oxygen on survival, growth and re-production of Daphnia, Hyalella and Gammarus.Environmental Toxicology and Chemistry 11:373–379.

PEARSON, R. G., AND N. M. CONNOLLY. 2000. Nutrientenhancement, food quality and community dy-namics in a tropical rainforest stream. FreshwaterBiology 43:31–42.

PEARSON, R. G., AND L. K. PENRIDGE. 1987. The effectsof pollution by organic sugar mill effluent on themacro-invertebrates of a stream in tropicalQueensland, Australia. Journal of EnvironmentalManagement 24:205–215.

PEARSON, R. G., AND L. K. PENRIDGE. 1992. An ecolog-ical survey of selected rivers in Queensland withparticular reference to the effects of sugar milleffluents. Report No. 92/02. Australian Centre forTropical Freshwater Research, James Cook Uni-versity, Townsville, Australia.

PHILLIPSON, G. N. 1954. The effect of water flow andoxygen concentration on six species of caddis fly(Trichoptera) larvae. Proceedings of the Zoologi-cal Society of London 124:547–564.

PROSSER, C. L., AND F. A BROWN. 1962. Comparativeanimal physiology. W. B. Saunders Company,Philadelphia, Pennsylvania.

RICHARDSON, J., E. K. WILLIAMS, AND C. W. HICKEY.2001. Avoidance behaviour of freshwater fish andshrimp exposed to ammonia and low dissolvedoxygen separately and in combination. New Zea-


land Journal of Marine and Freshwater Research35:625–633.

ROSENBERG, D. M., AND V. H. RESH (EDITORS). 1993.Freshwater biomonitoring and benthic macroin-vertebrates. Chapman and Hall, London, UK.

ROSSER, Z. C. 1999. The role of disturbance in com-munity dynamics and structuring of tropicalrainforest stream communities. PhD Thesis,James Cook University, Townsville, Australia.

ROSSER, Z. C., AND R. G. PEARSON. 1995. Responses ofrock fauna to physical disturbance in two Austra-lian tropical rainforest streams. Journal of theNorth American Benthological Society 14:183–196.

SPRAGUE, J. B. 1963. Resistance of four freshwatercrustaceans to lethal high temperature and lowoxygen. Journal of the Fisheries Research Boardof Canada 20:387–415.

TRACEY, J. G. 1982. The vegetation of the humid trop-ical region of North Queensland. Commonwealth

Scientific and Industrial Research OrganisationPublishing, Melbourne, Australia.

WALSHE, B. M. 1948. The oxygen requirements andthermal resistance of chironomid larvae fromflowing and from still waters. Journal of Experi-mental Biology 25:35–44.

WILEY, M. J., AND S. L. KOHLER. 1980. Positioningchanges of mayfly nymphs due to behaviouralregulation of oxygen consumption. CanadianJournal of Zoology 58:618–622.

WINTER, A., J. J. H. CIBOROWSKI, AND T. B. REYNOLD-SON. 1996. Effects of chronic hypoxia and reducedtemperature on survival and growth of burrow-ing mayflies, Hexagenia limbata (Ephemeroptera:Ephemeridae). Canadian Journal of Fisheries andAquatic Sciences 53:1565–1571.

ZAR, J. H. 1999. Biostatistical analysis. 4th edition.Prentice Hall, Upper Saddle River, New Jersey.

Received: 2 June 2003Accepted: 9 March 2004

effect of low dissolved oxygen on survival, emergence, …€¦ · 251 j. n. am. benthol. soc.,...

Documents