food webs in forest and pasture streams in the waikato region

New Zealand Journal of Marine and Freshwater Research, 1997, Vol. 31: 651-6640028-8330/97/3105-0651 $7.00/0 © The Royal Society of New Zealand 1997

651

Food webs in forest and pasture streams in the Waikato region,New Zealand: a study based on analyses of stable isotopesof carbon and nitrogen, and fish gut contents

BRENDAN J. HICKS

Department of Biological SciencesUniversity of WaikatoPrivate Bag 3105Hamilton, New Zealand

Abstract Stable isotopes of carbon (C) andnitrogen (N) were studied in 11 stream communitiesin the Waikato region of New Zealand. Fromcomparisons of mean 813C and 815N values, foodwebs in the shaded, forest streams were clearlybased on allochthonous material (conditioned leaflitter and terrestrial invertebrates). Autotrophs inforest streams were not a significant C source forthe food webs. However, the C source of food websin the unshaded pasture streams appeared to be amixture of allochthonous and autochthonousmaterial. Conditioned leaf litter appeared tocontribute to the pasture stream food webs, and the813C and 815N of some samples of epilithic diatomsindicated their consumption by invertebrates inpasture streams. Fish ate a wide range of aquaticinvertebrates; longfinned eels (Anguilladieffenbachii) and banded kokopu (Galaxiasfasciatus) also had a large proportion of terrestrialinvertebrates in their diet. Filamentous green algaewere found only at pasture sites, where they weresometimes abundant. The wide range of 813C valuesof filamentous green algae (-18.8 to -29.7%o)complicated understanding of their role in thestream food webs. The 813C values of Cladophorawere related to water velocity, with more 13C-enriched values in pools than in runs (-23.2%o inpools, mean velocity 0.12 ms" ' ; -28.1%o in runs,mean velocity 0.24 m s"1). Crayfish and thegastropod mollusc Potamopyrgus appeared to bethe only invertebrates to eat filamentous greenalgae.

M97012Received 2 April 1997; accepted 13 October 1997

Keywords stable isotopes; carbon; nitrogen;invertebrates; fish; food webs; land use

INTRODUCTION

In the Waikato region of New Zealand, biomass offish and eel production has been related to land use.Pasture streams had much greater fish biomass andeel production than forested streams (Hicks &McCaughan 1997), and pasture streams also hadthe greater primary production (Quinn et al. 1997a).The aim of this project was to investigate theinfluence of land use on food webs in streams inthe Waikato region. Food webs in streams beginwith inputs of organic and inorganic carbon (C) andlight, and food web components may include algae,leaf litter, non-predatory and predatoryinvertebrates, and fish (e.g., Rounick &Winterbourn 1986). Catchment vegetation,especially in the riparian zone, has a major influenceon food webs in streams, through both shading andlitter inputs (e.g., Vannote et al. 1980; Cummins etal. 1989; Gregory et al. 1991).

Food webs were studied by tracing the fraction-ation of stable isotopes of C (12C and 13C), andnitrogen (N) (14N and 15N), and by examining thegut contents of fish. The basis of the isotopetechnique is that enzymatic and diffusion processesdiscriminate against the heavier isotopes of both Cand N, resulting in a predictable change in isotopicratio in the consumer compared to the source(Griffiths 1991). The use of both C and N isotopestogether improves the resolution of food webscompared to analysis of C isotopes alone (Peterson& Fry 1987; Fry 1991), and overcomes some ofthe criticisms recently aimed at isotope food-webanalyses by France (1995a; 1996a,b,c).

Isotopic ratios of carbonThe importance of allochthonous C compared toautochthonous C as a base for secondary productioncan theoretically be determined from the 813C of

652 New Zealand Journal of Marine and Freshwater Research, 1997, Vol. 31

the consumers. Allochthonous C derived fromterrestrial plants with the C3 photosynthetic pathwaygenerally has a 8I3C value of-26 to -29%o, c. 20%oless than the 813C value of atmospheric CO2 (-8%o;O'Leary 1988). Diffusion and photosynthesisdiscriminate against 13C, the heavier stable isotopeof C. In water, HCO3~ can act as a C source. Sincethe 813C of HCO3" can be 9%o more negative thandissolved CO,, photosynthesis using HCO ~ canresult in aquatic plant tissue that has 813C values ofc. -35%o (e.g., Rau 1980; Rounick & Hicks 1985).

In practice, the S13C values of terrestrial andaquatic is quite variable. Substrate limitation canovercome enzymatic discrimination, resulting inless negative 8I3C values than expected. Carbon inin the leaves of terrestrial plants may varyseasonally in 8I3C by 5%o in response to stomatalclosure at times of water stress (Walcroft et al.1997). High irradiance can cause terrestrial leavesto be less negative in 813C than leavesphotosynthising at low irradiance; leaves in theupper canopy can be 3^4%o less negative than leavesin the lower canopy (Ehleringer et al. 1986;Schlesser 1990, 1992).

Aquatic plants photosynthising in water alsoshow seasonal variability in 813C (Boon & Bunn1994). In lake phytoplankton, high rates ofphotosynthesis lead to substrate limitation and 813Cvalues close to 0%o (McCabe 1985). Despite thisvariability, aquatic plants often have different valuesof 8I3C than land plants (O'Leary 1988), whichenables determination of whether a food web isprincipally based on land plants or in-streamprimary production (e.g., Rounick et al. 1982;Winterbourn et al. 1984; Rounick & Hicks 1985).Where isotopic signatures of C sources do notfollow predictions, the results may still usefullyindicate testable hypotheses concerning ecosystemfunction.

The C ratio of each consumer should reflect itsassimilated C, with an enrichment in 13C causedby the assimilatory and respiratory processes thatuse 12C preferentially. The 813C of the consumercan be 0-1 %o less negative than its food (Peterson& Fry 1987). White muscle of rainbow trout(Oncorhynchus mykiss) aged 0-2 years and fed ontrout pellets was 0.9-1.7%o less negative than theirfood; enrichment declined with age and size(Rounick & Hicks 1985). A similar decline of 13Cenrichment with increasing size was observed inbrook char (Salvelinus fontinalis, Peterson &Howarth 1987).

Isotopic ratios of nitrogenThe ratio of I5N/14N in air is relatively invariant(Mariotti 1983), and is the defined zero point.Leaves of terrestrial plants may range in 815N from-8 to +3%o, and where aquatic plants have differentvalues to those of terrestrial plants that mightcontribute litter to a stream, N isotopes may functionas source markers for tracing the autochthonous orallochthonous origin of N in food webs, augmentinginformation from the ratio of 13C/12C (Peterson &Fry 1987; France 1995b). Nitrogen fixation alsoaffects the 815N of N sources, resulting in organicN that is 815N 0.6-4. l%0 more negative than the Nsource. Values in plant tissues reflect their N source(Handley & Raven 1992; France 1995b).

Nitrogen in most parts of the biosphere has 815Nvalues ranging from -10 to +10%o, primarilybecause the N uptake by plants and microbes islimited by supply rather than by enzymatic uptake(Peterson & Fry 1987). However, subsequentfractionation is greater for 15N than for 13C,frequently resulting in a c. 3%o increase in 815Nbetween trophic levels (Peterson & Fry 1987;Handley & Raven 1992).

STUDY AREAS

Study sites in the Waikato region were located onstreams draining native forest (three sites), exoticforests of Pinus radiata (four sites), and pasturelands (four sites). The physical characteristics ofthe sites have been previously reported (Hicks &McCaughan 1997). The fish populations of thesestreams were dominated by longfinned eels(Anguilla dieffenbachii), though shortfinned eels {A.australis) were also abundant at pasture sites (Hicks& McCaughan 1997). Banded kokopu (Galaxiasfasciatus) occurred at forested sites, and Cran's andredfinned bullies (Gobiomorphus basalis and G.huttoni) were found at exotic forest and pasturesites. Common smelt {Retropinna retropinna)occurred at one pasture site only. The freshwatercrayfish (Paranephrops planifrons) was commonat all sites.

MATERIALS AND METHODS

The majority of samples were collected in Januaryand early February 1993, and from one site inJanuary 1994. Some additional invertebratessamples used to make up a core group at each site

Hicks—Food webs in pasture and forest streams

were collected late March and early April 1993.Supplementary watercress (Rorippa nasturtium-aquaticum), algal, and molluscan samples werecollected from site PW2 in early March 1995 byS. E. Bunn (Griffith University, Queensland,Australia). On collection, samples were stored at0°C in crushed ice for transport to the laboratorywithin 3 h. Fish guts were removed and preservedin 10% formalin before analysis. White muscle wascut from the fish and oven-dried to a constant weightat 40°C, which usually required 24 h. Muscle fromthe tails of crayfish was removed from theexoskeleton and oven-dried for 24 h at 40°C. Otherinvertebrates and conditioned leaf litter weresimilarly dried whole. Guts were not removed fromthe invertebrates. After drying, leaf material fromwoody plants was ground in a ball mill to a fine,homogeneous powder. Algal and animal tissueswere ground under liquid N2 with a mortar andpestle. The molluscan scrapers Latia andPotamopyrgus were ground whole, with their shells,except for the 1995 sample in which some werefirst acid-washed.

Among the invertebrates, the largest sizedindividuals were selected for stable isotope analysis.Total lengths of fish analysed were 95-242 mm(mean 166 mm, JV= 8) for banded kokopu, 54-92mm(mean71 mm,jV= 17) for Cran's bullies, 178—987 mm (mean 411 mm, N = 79) for longfinnedeels, and 109^98 mm (mean 288 mm, N= 40) forshortfinned eels. Fork length was 79-85 mm (mean83 mm, N = 5) for common smelt. The size rangeof crayfish sampled for stable isotopes was 5.1-36.3 mm OCL (orbit-carapace length; mean 19.4mm, N = 68). These size data were pooled acrossall land-use types.

Isotopic fractionationWater samples were analysed for dissolvedinorganic C (DIC) concentration and 813C using themethods of McCabe (1985). Measurements ofisotopic ratios were made using a VG Micromass602E mass spectrometer with double ion-collectionand inlet systems, equipped for rapid switchingbetween standard and samples for high accuracyand precision. For measuring biotic samples, wherelower accuracy and precision was acceptable, aEuropa Scientific Tracermass mass spectrometerwith a single ion-collector and inlet system wasused. Small subsamples (3-6 mg) of dried, groundmaterial were weighed with a 5-place balance, andoxidised at high temperature in the furnace of aCarlo Erba CHNO analyser gas chromatograph. The

653

resultant CO2 or N2 stream was then analysed withthe Tracermass to a precision of c. 0.1 %o for I3Cand 0.3%o for 15N. Correction for drift was madeautomatically by the Tracermass from a referencesample of known fractionation measured every 40or so samples.

The ratios of 13C/12C and 15N/14N wereexpressed as relative difference per mil (%o) usingthe equation:

L "standard

whereZ= 13C or 15N, andR= 13C/12C or 15N/14N.The ratio of 13C to 12C was compared to the PDBstandard, for which /{standard = 1.1237 atom % 13C(Craig 1957). For 15N/14N, N2 in air was used asthe standard, and standard = 0.3663 atom % 15N(Mariotti 1983).

Data analysisTo find group means for 813C and 815N, the food-web components were assigned a priori to thevarious consumer groups. Assignation to consumergroups was made on the basis of previous studies(Rounick et al. 1982; Winterbourn 1982;Winterbourn et al. 1984; Linklater 1991; Quinn etal. 1992; Linklater & Winterbourn 1993), exceptfor Ichthybotus, which was classified as a predatoron the basis of its large, tusk-like mandibles. Valuesof 813C and 815N were compared among land-usetypes using Kruskal-Wallis nonparametric one-wayanalysis of variance, which reduces to a Mann-Whitney U test for two samples, and Tukey'smultiple comparison test (Zar 1984; Wilkinson etal. 1994).

RESULTS

Stable isotope ratiosLeaf litter was the most ' 5N-depleted food webcomponent in all land-use types (Table 1). Dualisotope plots of 813C and 815N snowed that leaflitter was the primary C source of food webs inforest streams (Fig. 1). Epilithic diatoms were aninsignificant food source in forest streams, as shownby their 15N-enrichment and 13C-depletioncompared to other food-web components. However,in pasture streams the distinction between leaf litterand epilithic diatoms as a C source was less clear(Fig. 1). The wide variation in 813C suggests thatalthough diatoms were probably a food source insome pasture streams, in other pasture streams the


sampled diatoms were too enriched in 13C to be asignificant C source. Filamentous green algae werefound only in pasture streams, and also showed aconsiderable range of SI3C values, but were onaverage too enriched in ' 5N to be a food of the non-predatory invertebrates. However, crayfish inpasture streams appeared likely to have consumedfilamentous algae (Fig. 1). The 8' 3C and ' 5N valuesof the single moss sample from a pasture stream

showed that it was unlikely to have been asignificant C source.

Fish were the top aquatic consumers becausethey had the highest mean 815N values in all landuses (Fig. 1). The mean 513C of fish (all speciescombined) were very similar among land-use types,despite the differences in 813C of the leaf litter ineach land-use type. The range of 813C in all 129fish measured was -22.9 to -28.4%O. Crayfish were

Table 1 Mean 613C and 515N values of putative carbon (C) sourcesforest (NF), exotic forest (EF), and pasture (PA) in the Waikato region

and a priori consumers in streams in nativei, New Zealand, in summer 1993.

Consumer group

Conditioned leaf litter

Primary producersEpilithic diatoms

Filamentous green algaFilamentous green algaMossFilamentous red alga

Taxon

CladophoraSpirogyra

Batrachospermum

Land-

usetype

NFEFPA

NFEFPAPAPAPAEF

Non-predatory invertebrates: herbivores and detritivoresOligochaete

AmphipodMayfly

Mayfly

Mayfly

MayflyMayfly

Net-spinning caddis

Cased caddisNon-biting midgePulmonate mollusc

Gastopod mollusc

Paracalliope fluviatilisColoburiscus

Deleatidium

Nesameletus

SiphlaenigmaZephlebia

Hydropsychidae

Pycnocentria evectaChironomidLatia

Potamopyrgus

EFPAEFNFEFPANFEFPANFEFPAEFNFEFPANFEFPAPAPANFEFPANFEFPA

N

91214

344

10311

2419

1343445691

3713411118133

13

8l3C(%o)

Mean

-29.7-29.1-27.2

-32.3-31.6-24.6-26.3-20.3-33.0-40.1

-26.1-26.2-27.2-26.9-27.3-26.2-29.1-29.8-27.6-32.7-33.3-25.7-25.6-27.0-26.7-28.7-27.1-27.0-25.9-27.7-29.0-23.4-24.1-23.4-15.9-16.1-18.4

815N(

N

182428

344

1031

2419

133234559137134

1

18133

13

Mean

-1.4-0.2

2.4

10.17.43.86.41.96.9

2.94.3

-0.53.24.25.01.64.84.43.05.95.03.72.04.04.91.85.8

3.9

1.97.95.84.35.44.5


always more enriched in 13C than fish, but wereless enriched in 15N (Fig. 1), suggesting that theyate primarily invertebrates, but that they probablyalso ate a variety of other foods. OCL explained<6% of the variability in 813C and 815N. Theincrease in 813C between leaf litter and fish wasgreater in forest sites than in pasture, and was 4.7%oin native forest (-29.7 to -25.0%o), 4.0%o in exoticforest (-29.1 to -25.1%o), and 2.0%o in pasture(-27.3 to -25.3%o). The average increase in 8I3Cper trophic step across the three steps between leaflitter, non-predatory invertebrates, predatoryinvertebrates, and fish was 1.6%o in native forest,1.3%o in exotic forest, and 0.6%o pasture.

The difference in 815N between predatory andnon-predatory invertebrates other than crayfish

655

suggests some degree of omnivory, or incorrect apriori classification to consumer groups. Theincrease in 815N between leaf litter and fish wassimilar in all land-use types. The increases were7.5%o in native forest (-1.3 to 6.2%o), 8.4%o in exoticforest (-0.2 to 8.2%o), and 7.0%o in pasture (2.4 to9.4%o). The average increase in 815N per trophicstep across the three steps between leaf litter, non-predatory invertebrates, predatory invertebrates,and fish was 2.5%o in native forest, 2.8%o in exoticforest, and 2.3%o pasture.

Conditioned leaf litter and epilithic diatomswere more 13C-enriched at pasture sites than atforested sites (Table 1,2). This trend was also shownby the collector-browsers Deleatidium andNesameletus, and the net-spinning caddises of the

Table 1 (continued).

Consumer group Taxon

Predatory invertebrates: first-level carnivoresMayfly

Burrowing mayfly

Dragonfly

Dobsonfly

Free-living caddis

StoneflyFlatworm

Ameletopsis

Ichthybotus

Antipodochlora

Archichauliodes

Hydrobiosis

StenoperlaTurbellaria

Fish: second-level carnivoresLongfinned eel

Shortfinned eel

Banded kokopu

Cran's bully

Redfinned bullyCommon smeltOmnivoresCrayfish

Anguilla diffenbachii

Anguilla australis

Galaxias fasciatus

Gobiomorphus basalis

Gobiomorphus huttoni

Retropinna retropinna

Paranephrops planifrons

Land-usetype

NFEF

NFEF

EFPA

NFEFPA

NFPA

NF

NFEFPA

NFEFPA

NFEFPA

NFEF

EFPA

EF

PA

NFEFPA

813C(

N

42

24

21

613918

3132

212822

67

21

71

64

1

5

222322

%o)

Mean

-27.7-26.7-26.3-26.6-27.2-29.1-26.2-26.6-25.1-27.4-26.1-26.6-26.6-26.8-26.2

-25.3-25.2-25.6-24.4-23.9-25.2-24.6-26.1-25.6-24.7-25.6-24.3

-24.5-24.5-24.4

815N(

N

42

24

21

5138

17

3

22

212822

67

21

71

64

1

5

222322

:%>)

Mean

4.15.3

3.74.8

4.55.1

3.84.85.7

2.45.9

3.5

6.77.4

6.78.49.4

6.18.68.9

5.06.9

7.68.6

6.3

12.3

4.56.27.8

656

o•

•

•0AD

epilithic diatomsleaf litternon-predatory invertebratespredatory invertebratesfishcrayfishmossfilamentous green algae

New Zealand Journal of Marine and Freshwater Research, 1997, Vol. 31

family Hydropsychidae. In forested streams,Nesameletus was by far the most 13C-depleted ofany invertebrate. This suggests that primary foodof Nesameletus might have been the similarly I3C-depleted epilithic diatom's, but the higher 815Nvalues of the diatoms show that diatoms could nothave been the sole food. Deleatidium, Zephlebia,Latia, and flatworms were not significantly differentin 513C or 515N among land-use types (Table 1,2).

15

10

5

0

15

£ 5

0

15

10

5

0

— a

o

c

o

A

/

k

o

H

/

o

J

L

i (

Nat

>

Exo

/

Pas

y:

vefc

V

ticfc

/

ture

f •• -f

Hl

>rest

) ^r

rest

Ic

.....

b

-34 -32 -30 -28 -26 -24 -22 -20

Table 2 Probabilities (P) for Kruskal-Wallis of signifi-cance of differences in 8'3C and 515N among land-usetypes for trophic groups in forest and pasture streams inthe Waikato region, New Zealand.

Trophic group

Conditioned leaf litterEpilithic diatomsColoburiscusDeleatidiumNesameletusZephlebiaHydropsychidaeLatiaPotamopyrgusA rchichauliodesFlatwormsLongfinned eelsShortfinned eelsCrayfish

513C

0.0010.0250.0270.1920.0030.2200.0320.2850.0230.0020.4390.5390.0070.551

815N

<0.0010.1160.0090.1050.1040.0950.0340.1691.0000.0030.121

O.0010.001

<0.001

Table 3 Dissolved inorganic carbon (DIC) in streamsin exotic forest, native forest, and pasture in the Waikatoregion, New Zealand. (CL = 95% confidence limit.)

DIC(millimol litre1)

Land-use type N Mean Lower CL Upper CL

DIC concentrationNative forest 5 0.21Exotic forest 6 0.16Pasture 8 0.22

0.150.130.21

813C

0.270.190.24

Stable carbon isotopic composition of the DICNative forest 6 -14.2 -15.3 -13.0Exotic forest 6 -13.0 -15.5 -10.5Pasture 7 -16.5 -19.0 -13.9

•< Fig. 1 Mean fractionation of stable carbon (C) andnitrogen (N) isotopes by presumed consumer categoriesin streams grouped by land-use type in forest and pasturein the Waikato region of New Zealand. Vertical andhorizontal bars are 95% confidence intervals. Ellipse foreach land use defines the main path of isotopicfractionation.


However, Potamopyrgus was more negative in 813Cat pasture sites than at forested sites, reflecting thelower 813C of DIC at pasture sites that wasincorporated into the non-dietary C in its shell(Table 3B). The 13C-enrichment of Latia alsoshowed contamination with non-dietary C.

Archichauliodes, Ichthybotus, Stenoperla, andthe flatworms (Turbellaria) were clearly predators,whereas Ameletopsis and Antipodochlora aremore likely to be omnivores. The predator

Table 4 Variability of 513C and 515N values ofCladophora and watercress at a pasture site (PW2) inlate summer (9 March 1995; S. E. Bunn unpubl. data).

Autotroph

CladophoraCladophoraCladophora

CladophoraCladophoraCladophora

WatercressWatercressWatercressWatercressWatercress

Habitat

poolpoolpool

runrunrun

513C

-22.7-23.2-23.6

-27.3-27.4-29.7

-27.8-28.3-28.6-28.9-29.3

515N

3.422.832.89

3.142.202.57

4.015.024.535.525.67

657

Archichauliodes had less negative 813C and 815Nvalues at pasture than at forest sites (Table 1,2).Fish were clearly second-level carnivores. Amongfish and crayfish there was a strong trend ofincreasing 815N values from native forest to exoticforest to pasture (Table 1). Common smelt had muchhigher 815N than other fish at site PT2, probablyreflecting their up-stream migration from the seaand lower river to their stream spawning sites.

The mean amount of DIC was greater in pasturethan in exotic forest (Table 3A; Kruskal-Wallis P =0.018; Tukey multiple range test P = 0.007).However, the 813C of the DIC was not differentamong land-use types (Table 3B; Kruskal-Wallis P= 0.068), and ranged widely from -10.3 to -19.5%o.Assuming that the DIC was the main source of Cfor aquatic photosynthesis, the mean shift in 813Ccaused by photosynthesis in the epilithic diatomswas 18.1%o at native forest sites (-14.2 decreasedto -32.2%o), 18.6%o at exotic forest sites (-13.0decreased to -31.6%o), but only 8.1%o at pasturesites (-16.5 decreased to -24.6%o). By visualexamination it was found that all thin (<0.5 mm)epilithic films in this study were dominated bydiatoms. Filamentous green algae at pasture siteshad a very wide range of 813C values (-18.8 to-28.9%o). Batrachospermum, a filamentous red alga

Table 5 Occurrence of aquatic and terrestrial prey items in the guts of fish instreams in native and exotic forest, and in pasture, in the Waikato region, NewZealand.

Land-usetype

Longfinned eelsNative forestExotic forestPastureShortfinned eelsNative forestExotic forestPastureBanded kokopuNative forestExotic forestCran's bullyExotic forestPasture

Redfinned bullyExotic forestCommon smeltPasture

Total no.offish

examined

222823

77

21

71

64

1

5

Total no.of taxa

324231

71127

191

99

5

8

%aquatic

taxa

566258

7110078

53100

100100

100

100

%terrestrial

taxa

443842

290

22

470

00

0

0


that was found at one exotic forest site, had a verylow 513C value (-40.1%o; Table 1).

Localised differences in water velocity appearedto have affected the 13C depletion of filamentousgreen algae, and might account for the wide rangeof 813C values that were found. Cladophora andSpirogyra were the commonly occurring genera offilamentous green algae, and formed substantialaccumulations in slower flowing microsites in thepasture streams in January and February 1993. The8'3C of Cladophora was considerably less negativein pools (mean -23.2%o) than in runs (mean-28.1%o; Table 4). Pools in similar sized Waikatostreams had a mean velocity of 0.12 m s"' (N =95), but at the margins, where the greatestabundance of filamentous algae occurred, watervelocities were close to zero. Runs had a meanvelocity of 0.24 m s"1 (N= 95; B. J. Hicks unpubl.data). Mean S15N for filamentous algae was not

different among pool and run habitats, and the meanfor both habitats was 2.8%o (95% confidenceinterval = 0.4%o, N = 6). Watercress was too 13C-depleted, and too 15N-enriched to have contributedsignificantly to the food webs of pasture streams(Table 4).

The gastropod scrapers Latia and Potamopyrguscollected in 1993 and 1994 had low 813C valuesbecause they were processed with their shells, sowere contaminated with non-dietary C. To test theeffect of this contamination, the 1995 samples fromPW2 were acid-washed to remove the CaCO3 inthe shells. An acid-washed sample of Potamopyrgushad a 813C of-25. l%o, compared to -20.l%o for asample that was not acid-washed. The 815N valuewas unaffected by acid washing; the acid-washedPotamopyrgus sample had a 815N- of 5.0%o,compared to 4.7%o for the sample that was not acid-washed (S. E. Bunn unpubl. data).

Table 6 Food items in fish guts in streams in forest and pasture in the Waikato region, New Zealand. Aquaticinsects are larval or nymphal stages unless otherwise stated. Only items occurring in >10% of the guts examinedare listed. (Origin: t = terrestrial; a = aquatic.)

Native forest

Item

Longflnned eelsCicadaLeptophlebiidaeCrayfishHarvestmanSpiderZephlebiaDeleatidiumGordian wormGreen beetleOligochaete

Shortfinned eelsLeptophlebiidaeCrayfishDeleatidiumAntBeetlePolyplectropusZephlebia

%

36363227231814141414

71572914141414

Origin

taattaaata

aaattaa

Exotic

Item

ZephlebiaOligochaeteHarvestmanIchthybotusCrayfishMillipedeCicadaAoteapsycheCaterpillarDeleatidiumSpiderAntHydrobiosis

ZephlebiaCrayfishIchthybotusOligochaeteAmeletopsisAmphipodArchichauliodesCeratopogindaeColoburiscusDeleatidiumLeptophlebiidae

forest

%

50292525252521181818141111

5743292914141414141414

Origin

aataattatatta

aaaaaaaaaaa

Pasture

Item

HydrobiosisPotamopyrgusHarvestmanCrayfishLeptophlebiidaeAmphipodAoteapsycheBrown beetleChironomidDeleatidium

HydrobiosisChironomidaeLeptophlebiidaeOxyethiraAmphipodCased caddisHarvestmanOligochaetePotamopyrgusCaterpillarDeleatidiumSpiderTipulid

%

70483530262217131313

52332424141414141410101010

Origin

aataaaataa

aaaaaataatata


Gut contents of fishOut of 132 fish guts examined, only two werecompletely empty. It was necessary to enumerateitems in the whole gut, rather than just the stomach,as the hindgut of many fish was much fuller thanthe stomach. Most prey items were identifiable,though because of the advanced stage of digestion,some items could be recognised only by smallfragments. For this reason volumetric and massanalyses were impractical. Aquatic prey were foundin all fish with any food in their gut, but terrestrialprey were also important for longfinned eels in allland-use types, and for banded kokopu (Table 5).

659

Terrestrial prey had a lower percentage occurrencein the guts of shortfinned eels than in longfinnedeels, and were not eaten at all by Cran's bully,common smelt, or the one redfinned bully that wasfound.

For longfinned eels in forested streams, cicadas{Amphisalta zelandica) were a common food item,as were the leptophlebiid mayfly nymphs includingZephlebia and Deleatidiwn, and crayfish (Table 6).In exotic forest streams, oligochaetes were acommon food. In pasture streams, Hydrobiosis andPotamopyrgus were the most frequent food itemsfor longfinned eels, whereas terrestrial harvestmen

Table 6 (continued).

Native

Item

Banded kokopuAntCicadaZephlebiaHarvestmanBeetleHydraenid adultAoteapsycheBullyCaddisflyCentipedeDeleatidiumElmid adultGreen beetleHydrobiosisIsopodLeptophlebiidaeMillipedePotamoprygusSpiderCran's bullies

Common smelt

forest

%

86575743292914141414141414141414141414

Origin

ttattaaaataatatatat

Exotic

Item

Deleatidium

DeleatidiumLeptophlebiidaeAoteapsycheColoburiscusConoesucidaeIchthybotusPolycentropidaeZephlebia

forest

%

100

3333171717171717

Origin

a

aaaaaaaa

Pasture

Item

ChironomidOxyethiraCandonocyprisHydrobiosisAustrosimuliumDeleatidiumDipteranEmpididae adult

OxyethiraChironomidAmphipodCaddisflyElmid adultDeleatidiumHydrobiosisLeptophlebiidae

%

10075505025252525

10060404040202020

Origin

aaaaaaaa

aaaaaaaa


(Order Opiliones) were frequent dietary items inall land-use types. For shortfinned eels, harvestmenwere the only important terrestrial prey in pasturestreams (Table 6). Leptophlebiid mayflies andcrayfish and were the most frequent prey items ofshortfinned eels in forested streams, whereasHydrobiosis and chironomid larvae were the mostfrequent prey items in pasture streams. Bandedkokopu in native forest streams ate mainly ants(Family Formicidae), cicadas, Zephlebia, andharvestmen (Table 6). In Cran's bullies in exoticforest, Deleatidium and other leptophlebiid mayflieswere the dominant dietary items, whereas in pasture,chironomids, the caddises Oxyethira andHydrobiosis, and the ostracod Candonocypris werethe principal foods. Common smelt ate mainlyOxyethira larvae, chironomids, amphipods,unidentified caddis larvae, and elmid beetles.

Among the less common items eaten by fish(i.e., < 10% occurrence) there was a wide variety ofaquatic and terrestrial taxa. Aquatic taxa were morenumerous (N = 43) than were terrestrial taxa (N =23). The intestinal parasite Eustrongilydes, orshagworm, was found in 86% of longfinned eels innative forest, 82% in exotic forest, and 74% inpasture. In shortfinned eels, shagworms occurredin 29% of eels in native forest, but not in the otherland-use types. In banded kokopu, shagwormsoccurred in one out of the seven fish found in nativeforest, but not in the one fish found in exotic forest.There was no difference in 813C between fish thathad a large proportion of terrestrial taxa (>40% offood items) and those that had none in theirstomachs when sampled (Mann-Whitney [/test; P= 0.390).

DISCUSSION

Stable isotope ratios and food web analysisFood webs in these forested Waikato streams werebased on allochthonous material. Epilithic diatomsappeared to be unimportant C sources for allinvertebrates except for the l3C-depleted collector-browser Nesameletus. At some stream sites on theWest Coast, South Island, Nesameletus wassimilarly 13C-depleted (Winterbourn et al. 1984).Patterns of isotopic fractionation among the food-web components were consistent with the highlyshaded nature of the forested sites, wherephotosynthesis was limited, epilithon biomass waslow (Quinn et al. 1997a), and the riparian treesensured a supply of allochthonous material.

Food webs in the open, pasture streams in thisstudy were more 13C-enriched than were the shaded,forested streams, contrary to some previous NewZealand studies in which food webs at open siteswere more 13C-depleted than at shaded sites(Rounick et al. 1982; Winterbourn et al. 1984).Food-web 13C-enrichment in Waikato streamsseems to be attributable to the autotrophs. Epilithicdiatoms were significantly more 13C-enriched atpasture sites than at forested sites (Table 1), incontrast to Wellington streams where stone surfacescrapings at open sites were highly 13C-depleted(-34.3 to-35.8%o; Rounick & Hicks 1985). Thoughl3C-enriched food webs occurred in Wellingtonstreams, this enrichment at both open and shadedsites was based on the 13C-enrichment of fineparticulate organic matter rather than epilithon(Rounick & Hicks 1985). However, at open, pasturesites in the Waikato, the epilithic diatomsthemselves were l3C-enriched compared to shaded,forest sites.

The wide range of 813C and 15N values for theautotrophs in pasture streams made the extent oftheir contribution to food webs difficult to assess.The high rates of photosynthesis at these unshadedsites, combined with high temperatures (Quinn etal. 1997a), suggest that autotrophic production wasresponsible for the high eel biomass and product-ivity found by Hicks & McCaughan (1997). Thepresent study lacked a truly representative samplingof autotrophs that might have shown the quantitativecontribution of primary producers to the pasturestream food webs. Also, microsite water velocityinfluenced the 8I3C value of filamentous greenalgae in the 1995 sampling (Table 4). Water velocitywas not recorded in the original 1993/94 sampling.The link between high rates of photosynthesis, lowwater velocities, and 13C-enrichment of aquaticautotrophs requires further investigation.

The dual isotope plot suggests thatallochthonous material also contributed to the foodwebs in pasture streams (Fig. 1). The 13C-enrichment of conditioned leaf litter in pasturestreams compared to forested streams (Table 1) mayhave been caused by diatoms growing onsubmerged leaf surfaces. In pasture streams, theinvertebrates Deleatidium, Nesameletus, and thehydropsychid caddises also showed ' ^-enrich-ment, consistent the 813C of the potential foodsepilithic diatoms and FPOM derived fromconditioned leaf litter. The l3C-enrichment ofArchichauliodes suggested that it preyed primarilyon some or all of these invertebrate taxa.


Filamentous green algae appear to have beenunavailable to invertebrates other thanPotamopyrgus and crayfish; Bunn et al. (1997)similarly concluded that Spirogyra in a tropicalstream had limited availability to invertebrateconsumers. Watercress was also not a significant Csource in the pasture streams in which it occurred.

Fish in the Waikato had similar 813C values inall land-use types (means -23.9 to -25.6%o). Thishomogeneity may have resulted partly from therelatively large sample size (N= 129) in this studycompared to previous studies. Fish in streams inthe Wellington region had 513C values between-19.5 and -24.9%o in a mixture of open and shadedsites (Rounick & Hicks 1985). In streams undernative forest on the West Coast of New Zealand,banded kokopu had 813C values between -22 and-25%o (Main & Lyon 1988).

Terrestrial taxa were significant in the diet oflongfinned eels and banded kokopu in the Waikato,but the diets of shortfinned eels, Cran's bullies, andcommon smelt were dominated by aquatic taxa(Table 5). Main & Lyon (1988) also reported thatbanded kokopu ate terrestrial prey. In the Waikato,fish feeding to a large extent (>40%) on terrestrialprey could not be distinguished isotopically fromfish that ate exclusively aquatic prey. Terrestrialinvertebrates as a C source might have confusedthe food web analysis, particularly since the isotopicratios of terrestrial taxa were not measured in thisstudy. Main & Lyon found that the 8I3C of terrestrialinvertebrates ranged from -21 to -28%o, with mostvalues close to -25%o. Another New Zealand studyfound that the 513C for terrestrial invertebrates was-24.4 ± 1.4%o (mean± 95% confidence interval;N= 10: McDowall et al. 1996). Without corres-ponding 815N values, these values cannot bedistinguished from those of aquatic invertebrates.

Potamopyrgus appears to have eaten fila-mentous green algae. The 1995 mean 813C offilamentous green algae in runs (-28.1%o) and itsmean 815N (2.8%o) suggest that it could have beena food for Potamopyrgus (acid-washed 813C =-25.1%o and 815N = 5.0%o). Further sampling isrequired to improve our knowledge of therelationship between molluscs and filamentousalgae. The contamination of the Potamopyrgussamples by non-dietary C (e.g., Tanaka et al. 1986)has been confirmed in this study, and should beavoided in future.

Crayfish in Waikato streams appear from their813C and 815N values to be primarily predators,deriving most of their C from invertebrates. Though

661

detritivory has been demonstrated in New Zealandcrayfish (Parkyn et al. 1997), assimilation rates ofplant material are likely to be low. In streams inMissouri, United States, the crayfish Orconectesassimilated only 14% of its detrital food comparedto 92% of invertebrate food. These crayfish alsoate filamentous algae, with an assimilationefficiency of 39% (Whitledge 1996). Theassimilation efficiency of Paranephropszealandicus for fresh and decayed oxygen weed(Elodea canadensis) averaged 21% (Musgrove1988). Crayfish in the present study (mean 813C-24.4 to -24.5%o) were the most 13C-enriched ofany consumer except for the gastropod scrapers,whose analyses were contaminated with non-dietarycarbonates. Main & Lyon (1988) also found that/;planifrons was the most highly 13C-enrichedinvertebrate (-24.6%o) in streams in SouthWestland, New Zealand. One explanation for the13C-enrichment is that food in the crayfish gut mightbe assimilated by micro-organisms, with subsequentassimilation of these micro-organisms by thecrayfish. Cellulase activity, which is normallyabsent from higher eukaryote organisms, has beendetected in the gut of P. zealandicus (Musgrove1988). This activity was associated with bacteriain the gut, but the extent of the contribution tocrayfish nutrition by these micro-organisms was notestablished, and thus the expected isotopicenrichment is unknown. Filamentous green algaeare a possible food source of crayfish in pasturestreams in the Waikato, as the fractionation of Cand N isotopes is in the right direction, and of aboutthe right magnitude. In forested streams, wherefilamentous algae were generally absent, the 8I3C-enrichment of crayfish suggests that theirpredominant food was invertebrates.

Analyses of 815N augmented the 813C analysesin this study, providing a second dimension that wasuseful for resolving trophic links. The use of bothC and N isotopes overcame some of the criticismsrecently aimed at stable C isotope studies by France(1995a; 1996a,b,c). There was a progressiveenrichment in 815N in equivalent trophic levelsbetween native forest and exotic forest, and betweenexotic forest and pasture. Invertebrates, eels, andcrayfish in pasture streams were 2-3%o greater in815N than they were in native forest streams (Table1). Contributions of ruminant excreta or nitrogenousfertilisers to ground water might have caused thel5N-enrichment of pasture streams. The exoticforest sites were in pasture until c. 15 years ago,which might account for their intermediate

662 New Zealand Journal of Marine and Freshwater Research, 1997, Vol. 3115N-enrichment. The difference in 815N valuesbetween leaf litter and top consumers in this study(7.0-8.4%o) was similar to forested streams in thePacific northwest of the United States (Fry 1991),though top consumers were considerably moreenriched in 13C in the United States' studies. Theaverage size of trophic steps was 2-3%o for 815N,and 0.7-1.6%o for 513C, close to expected values(Rounick & Hicks 1985; Peterson & Fry 1987;Handley & Raven 1992). The 815N of the migratorycommon smelt (12.3%o) was twice that of any otherfish species, possibly reflecting consumption of15N-enriched estuarine or marine plankton(Peterson & Howarth 1987).

Variability of autotrophic 513CVariability in DIC did not appear to account for thevariable S13C of the autotrophs in this study. DIC8I3C means (-13.0 to -16.5%o) were similar in allland uses, and were also similar to the rangepreviously reported in New Zealand streams (-11.1to -15.5%o; Rounick et al. 1982; Rounick & Hicks1985). In forested Waikato streams, epilithicdiatoms (mean 5I3C c. -32%o) were 18-19%o morenegative than the DIC, similar to predicted 13C-depletion by photosynthesis. At pasture sites,assuming that DIC was the C source for photo-synthesis, epilithic diatoms were only c. 8%o morenegative than the DIC. 13C-enriched periphytonhave been found at other stream sites in NewZealand; 813C values of-10 to -15%o were foundat open sites in the Styx River (Lester et al. 1995).Filamentous algae from open sites in theManganuiateao River, central North Island, also had13C-enriched values (813C -12.5 to-16.8%>; Collier&Lyon 1991).

One mechanism by which autotrophs canbecome 13C-enriched is through C limitation ofphotosynthesis, as competition for C substratecauses the usual discrimination against 13C to breakdown. Low water velocity can cause increases inautotrophic 8l3C through restricted mixing. Thoughsubstrate limitation of photosynthesis is unexpectedin a well-mixed, temperate, lotic system, algae andcyanobacteria with 13C-enriched signatures arecommon in lentic environments (McCabe 1985;Bunn & Boon 1993; Boon et al. 1994). Substratelimitation in lakes can result in phytoplankton with8'3C values close to 0%o (McCabe 1985). In streamswith high algal biomass, C limitation ofphotosynthesis might be sufficient to cause 13Cenrichment of the algae at low-velocitymicrosites.

Biomass of epilithon, including filamentousalgae, was higher at pasture sites than at forestedsites (Quinn et al. 1997a). The mean values atpasture sites in November (3.6 g AFDM m~2) werenot high compared to previously recorded valuesin New Zealand streams, and to values recorded inunshaded channels at a pasture site in summer(maximum 12 g C m~2, = c. 24 g AFDM nr2; Quinnet al. 1997b). Cladophora-dominated algalcommunities can have biomasses in excess of 100g AFDM nr 2 (Biggs & Price 1987). However, inmidsummer at my sites, substantial abundance offilamentous green algae was observed. Algalaccumulation depends on the length of period offlow recession as well as on nutrient status of thewater (Biggs 1995). Maximum accumulation wouldbe expected to occur around midsummer, when thesamples for this study were taken, rather than inNovember when Quinn et al. (1997a) sampled.

Thus it is conceivable that in summer atmicrosites with high temperature, high accum-ulation of algal biomass, and low water velocity,the C substrate for photosynthesis in might becomelimiting in streams. Variable 813C values inautotrophs could have resulted from sampling avariety of microhabitats with different watervelocities. Alternatively, microbial respirationunder low velocity conditions might have resultedin an undiscovered pool of dissolved inorganiccarbon that was used by the autotrophs in Waikatostreams.

CONCLUSIONS

Stable C and N isotope analyses combined with fishgut analyses have helped elucidate the structure offood webs in Waikato streams under different landuses. Within each of the three land uses, streamecosystems behaved consistently. The food websin forested streams followed the conventional small-stream, allochthonous-dominated model, whereasfood webs in pasture streams had a mixture ofallochthonous and 13C-enriched autochthonousinputs. For pasture streams, key areas for moredetailed analysis are suggested. Questionsremaining after this study include: (1) the cause ofthe variability in 813C of the autotrophs, includingseasonal variation; (2) the 813C and 8'5N signaturesof the contributing ground waters, and the extentto which the stream food web reflects these; (3) therole of dissolved organic carbon in stream foodwebs; and (4) the quantitative importance ofterrestrial invertebrates. To improve our


understanding of stream food webs through analysisof stable C and N isotopes, more attention to the Cand N dynamics of epilithic films, and to theirrelation to groundwater inputs is necessary.

ACKNOWLEDGMENTS

I thank Lee Laboyrie, Barry Strong, and John Hicks forassistance with the field sampling. Lee Laboyrie alsoassisted with sample processing in the laboratory. Stableisotope ratios were determined by Anjana Rajendram ofthe Waikato Stable Isotope Unit, School of Science andTechnology, University of Waikato. Chris Hendy of theDepartment of Chemistry, University of Waikato, carriedout the DIC determinations. Stuart Bunn (GriffithUniversity, Queensland, Australia) collected samples forfurther analysis of algae and Potamopyrgus. Criticalcomment was provided by Mike Winterbourn, Universityof Canterbury, and Bryce Cooper, John Quinn, and KevinCollier of NIWA, Hamilton, and two anonymousreferees. This project was entirely funded by a grant fromthe University of Waikato Research Committee.

REFERENCES

Biggs, B. 1995: Low flows and algal growth in streams.Water and atmosphere 3(1): 15-16.

Biggs, B. J. R; Price, G. M. 1987: A survey of filamentousalgae proliferations in New Zealand rivers. NewZealand journal of marine and freshwaterresearch 2J: 175-191.

Boon, P. I.; Bunn, S. E. 1994: Variation in the stableisotope composition of aquatic plants and theirimplications for food web analysis. Aquaticbotany 48: 99-108.

Bunn, S. E.; Davies, P. M.; Kellaway, D. M. 1997:Contributions of sugar cane and invasive pasturegrass to the aquatic food web of a tropicallowland stream. Marine and freshwater research48: 173-179.

Collier, K. J.; Lyon, G. L. 1991: Trophic pathways ofblue duck (Hymenolaimus malacorhynchos) onManganuiateao River: a stable carbon isotopestudy. New Zealand journal of marine andfreshwater research 25: 181-186.

Craig, H. 1957: Isotopic standards for carbon and oxygenand correction factors for mass spectrometricanalysis of carbon dioxide. Geochimicacosmochimica acta 12: 133-149.

Cummins, K. W.; Wilzbach, M. A.; Gates, D. M.; Perry,J. B.; Taliaferro, W. B. 1989: Shredders andriparian vegetation. Leaf litter that falls intostreams influences communities of streaminvertebrates. Bioscience 39: 24-30.

Ehleringer, J. R., Field, C. B.; Zhi-fang Lin; Chun-yenKuo. 1986: Leaf carbon isotope and mineralcomposition in subtropical plants along anirradiance cline. Oecologia 70: 520-526.

663

France, R. 1995a: Critical examination of stable isotopeanalysis as a means for tracing carbon pathwaysin stream ecosystems. Canadian journal offisheries and aquatic sciences 52: 651-656.

France, R. L. 1995b: Source variability in 515N ofautotrophs as a potential aid in measuringallochthony in freshwaters. Ecography 18:318-320.

France, R. L. 1996a: Carbon-13 conundrums: limitationsand cautions in the use of stable isotope analysisin stream ecological research. Canadian journalof fisheries and aquatic sciences 53: 1916-1919.

France, R. L. 1996b: Scope of use of stable carbonisotopes in discerning the incorporation forestdetritus into aquatic food webs. Hydrobiologia325: 219-222.

France, R. L. 1996c: Absence or masking of metabolicfractionations of I3C in a freshwater benthic foodweb. Freshwater biology 36: 1-6.

Fry, B. 1991: Stable isotope diagrams of freshwater foodwebs. Ecology 72: 2293-2297.

Gregory, S. V; Swanson, F. J.; McKee, W. A. 1991: Anecosystem perspective of riparian zones.Biosciences 41: 540-551.

Griffiths, H. 1991: Applications of stable isotopetechnology in physiological ecology. Functionalecology 5: 254-269.

Handley, L. L.; Raven, J. A. 1992: The use of naturalabundance of nitrogen isotopes in plantphysiology and ecology. Plant, cell, andenvironment 15: 965-985.

Hicks, B. J.; McCaughan, H. M. C. 1997: Land use,associated eel production, and abundance offishand crayfish in streams in Waikato, New Zealand.New Zealand journal of marine and freshwaterresearch 31: 635-650.

Lester, P. J.; Mitchell, S. R; Scott, D.; Lyon, G. L. 1995:Utilisation of willow leaves, grass and periphytonby stream macroinvertebrates: a study usingstable carbon isotopes. Archiv fur Hydrobiologie133: 149-159.

Linklater, W. 1995: Breakdown and detritivorecolonisation of leaves in three New Zealandstreams. Hydrobiologia 306: 241-250.

Linklater, W.; Winterbourn, M. J. 1993: Life historiesand production of two trichopteran shredders inNew Zealand streams with different riparianvegetation. New Zealand journal of marine andfreshwater research 27: 61-70.

Main, M. L.; Lyon, G. L. 1988: Contribution of terrestrialprey to the diet of banded kokopu (Galaxiasfasciatus Gray) (Pisces: Galaxiidae) in SouthWestland, New Zealand. VereinigungInternationale fiir Theoretische undAngewandteLimnologie Verhandlungen 23: 1785-1789.


Mariotti, A. 1983: Atmospheric nitrogen is a reliablestandard for natural 15N abundance measure-ments. Nature 303: 685-687.

McCabe, B. 1985: The dynamics of 13C in several NewZealand lakes. Unpublished DPhil thesis,University of Waikato, Hamilton, New Zealand.278 p.

McDowall, R. M; Main, M. R.; West, D. W.; Lyon, G.L. 1996: Terrestrial and benthic foods in the dietof the shortjawed kokopu, Galaxias postvectisClarke (Teleostei: Galaxiidae). New Zealandjournal of marine and freshwater research 30:257-269.'

Musgrove, R. J. 1988: Digestive ability of the freshwatercrayfish Paranephrops zealandicus (White)(Parastacidae) and the role of microbial enzymes.Freshwater biology 20: 305-314.

O'Leary, M. H. 1988: Carbon isotopes in photosynthesis.Biosciences 38: 328-335.

Parkyn, S. M.; Rabeni, C. R; Collier, K. J. 1997: Effectsof crayfish (Paranephrops planifrons:Parastacidae) on in-stream processes and benthicfaunas: a density manipulation experiment. NewZealand journal of marine and freshwaterresearch 31: 685-692.

Peterson, B. J.; Fry, B. 1987: Stable isotopes in ecosystemstudies. Annual review of ecology and svstematics18: 293-320.

Peterson, B. J.; Howarth, R. W. 1987: Sulfur, carbon,and nitrogen isotopes used to trace organic matterflow in the salt-marsh estuaries of Sapelo Island,Georgia. Limnology and oceanography 32:1195-1213.

Quinn, J. M.; Cooper, A. B.; Davies-Colley, R. J.;Rutherford, J. C ; Williamson, R. B. 1997a: Land-use effects on habitat, water quality, periphyton,and benthic invertebrates in Waikato, NewZealand, hill-country streams. New Zealandjournal of marine and freshwater research 31:579-597.'

Quinn, J. M.; Cooper, A. B.; Stroud, M. J; Burrell, G. P.1997b: Shade effects on periphyton andinvertebrates: an experiment in streamsidechannels. New Zealand journal of marine andfreshwater research 31: 665-683.

Quinn, J. M; Williamson, R. B.; Smith, R. K.; Vickers,M. L. 1992: Effects of riparian grazing andchannelisation on streams in Southland, NewZealand. 2. Benthic invertebrates. New Zealandjournal of marine and freshwater research 26:259-273.

Rau, G. H. 1980: Carbon-13/carbon-12 variation insubalpine lake aquatic insects: food sourceimplications. Canadian journal of fisheries andaquatic sciences 37: 742-745.

Rounick, J. S.; Hicks, B. J. 1985: The stable carbonisotope ratios of fish and their invertebrate preyin four New Zealand rivers. Freshwater biology15:201-214.

Rounick, J. S.; Winterbourn, M. J. 1983: The formation,structure and utilisation of stone surface organiclayers in two New Zealand streams. Freshwaterbiology 13: 57-72.

Rounick, J. S.; Winterbourn, M. J. 1986: Stable carbonisotopes and carbon flow in ecosystems.Biosciences 36: 171-177.

Rounick, J. S.; Winterbourn, M. J.; Lyon, G. L. 1982:Differential utilisation of allochthonous andautochthonous inputs by aquatic invertebrates insome New Zealand streams: a stable carbonisotope study. Oikos 39: 191-198.

Schlesser, G. H. 1990: Investigations of the 813C patternin leaves of Fagus sylvatica L. Journal ofexperimental botany 41: 565-572.

Schlesser, G. H. 1992: 813C pattern in a forest tree as anindicator of carbon transfer in trees. Ecology 73:1922-1925.

Tanaka, N.; Monaghan, M. C ; Rye, D. M. 1986:Contribution of metabolic carbon to mollusc andbarnacle shell carbonate. Nature 320: 520-523.

Vannote, R. L.; Minshall, G. W.; Cummins, K. W. 1980:The river continuum concept. Canadian journalof fisheries and aquatic sciences 37: 130-137.

Walcroft, A. S.; Silvester, W. B.; Whitehead, D.; Kelliher,R M. 1997: Seasonal changes in stable carbonisotope ratios within annual rings of Pinusradiata reflect environmental regulation ofgrowth processes. Australian journal of plantphysiology 24: 57-68.

Whitledge, G. 1996: Trophic ecology of crayfish in theJacks Fork River, Missouri. Unpublished MScthesis, University of Missouri-Columbia, UnitedStates. 81 p.

Wilkinson, L.; Hill, M.; Howe, P.; Birkenbeuel, G.; Beck,J.; Liu, J. 1994: SYSTAT for DOS: advancedapplications, version 6 edition. SYSTAT Inc.,Evanston, Illinois. 902 p.

Winterbourn, M. J. 1982: The invertebrate fauna of aforest stream and its association with fineparticulate matter. New Zealand journal ofmarine and freshwater research 16: 271-281.

Winterbourn, M. J.; Cowie, B., ; Rounick, J. S. 1984:Food resources and ingestion patterns of insectsalong a West Coast, South Island, river system.New Zealand journal of marine and freshwaterresearch 18: 379-388.'

Zar, J. H. 1984: Biostatistical analysis. Second edition.Englewood Cliffs, New Jersey, Prentice-Hall.718 p.

food webs in forest and pasture streams in the waikato region

Documents