The influence of elevated nitrate concentration on rate of leaf decomposition in a stream

Freshwater Biology (1983) 13,177-183

The influence of elevated nitrate concentrationon rate of leaf decomposition in a stream

JUDY L. MEYER and CAROL JOHNSON Zoology Department andInstitute of Ecology, University ot Georgia, Athens. Georgia

SUMMARY. 1. Leaf decomposition was compared in two streams ;tl theCoweeta Hydrologic Laboratory, North Carolina. U.S.A. One streamdrains an undisturbed hardwood watershed, while the other drains asuccessional watershed subject to an insect outbreak. The suecessionalwatershed has elevated nitrate concentrations in the streamwater.

2. Both black locust {Robinia pseudo-acacia) and sweet birch (Retttlaletiia) leaf litter decomposed 2.8 times more rapidly in the stream with highnitrate concentrations.

3. The more rapid decay rates appeared to be partly due to acceleratedmierobial processing in response to nitrate enrichment, beeause mierobialbiomass (as ATP) was higher in the nitrate-enriched stream.

4. At each point in time, nitrogen and phosphorus content of the litterwas lower in the high nitrate stream; however, there was no significantdifference in nitrogen or phosphorus content at the same state of leaf decayin the two streams.

Introduction

Decomposing leaves are an important organiccarbon source iti stnali forested streams (e.g.Anderson <& Scdel!. 1979). In addition luintrinsic differences between leaf species. leafpack decotnposiiion rate is influenced byenvironmental factors such as Ictnpcrature.nutrient content of the water, sediment deposi-tion rate, and densities of ieaf-shredding inver-tebrates (e.g. Meyer. I98(); Elwood «fl/., iy«l;Wallace. Webster &Cuffney. 1982). By alteringthese aspects of the stream environment,watershed disturbances such as clear-cutting canhave an impact on leaf decomposition in streams(Webster & Waide. 1982).

In this study we investigated the effect of anatural watershed perturbation on leaf dccom-

CorrcsptindL-nce: Dr Judy L. Meyer. gDepartment. University of Georgia. Athens. Georgia30602. U.S.A.

position in streams by comparing dccompt>.sitioiirate, mierobial colonization of the leaves, andleaf nutrient content in two streams: GradyBranch and Sawmill Branch. Both streams are inthe Coweeta Hydrologic Laborati>r\', MaconCounty, North Carolina. U.S.A. Gady Branchdrains a watershed (WS18) undisturbed by mansince ai least 1924. Sawtnill Branch drains awatershed (WSd) that was clear-cut, plantedwith grass and fertilized in 1958, treated withherbicide in 1966-67. and allowed to begin oldfield suecesion in I96S. After these disturbances.Sawmill Branch had signifieantly higher nitrateconcentrations than Cirady Braneh. In 1979black locust {Rohinia pseudo-acacia), thedominant tree species on WS6, showed increasedinfestation with locust borers {Megacyllenerohiniae Forst), which killed or weakened manytrees. One correlate of this infestation was thatstreamwater nitrate concentration was furtherelevated (Table 1); a similar increase in nitrate

0046-507(J/83/04(X)-0177 $02.00 © 1983 Blackwell Scientific Publications 177

178 J. L. Mever and C. Johnson

TABLf-; 1. Monihly mean aincentraiion (weigh led tor discharge) nf nuiricntsin streamwater duritig the period ofihis sluUy. Cirady Branch (CiB) drains the reference watershed while Sawmill Branch (SB) drains ihc disturbedwalcr-ihcd. .Ml values arc ;is my/I. Data arc from Ihc Coweeta Hydrologic Laboratory files.

November 1980GBSB

December I9S(IGBSB

.lantiarv I9S1G B 'SB

February 1981G BSB

March 1981G BSB

NO,-N

0.0010.624

0.0090.767

0,01S0,8.17

O.(XI80.8S2

o.ai20.817

Mean over 5 month periodG B O.(XI(iSB 0.773

NH.-N

O.(HWO.OIO

0.002

O.(X)30.002

(I'lxM

0,002

O.(HI2O.tH16

PO,-P Ca

< O.OOI 0.65< O.lHIl I.IO

< 0,1X11 0,52<O.(X)1 0,88

<().OOI 0.51< 0.001 O,*IO

< O.OOI 0.54<O.(H)1 0.91

<.O.(H)1 0.54< (1.0(11 (I.'XI

<O.(HI1 0.56<O.(HI1 0.93

K

0.49(1.76

0.36(t.57

0..130.54

0.35(1.61

0.35(1.56

0,63

Ratio of Sawmill Branch concentration to Gradv Braneh concent rat ion129 3.0 l.('l 1.6 1,6

Na

(1.96

l.(H)

0.901.04

(1.831.(13

0.S71,10

0.S7l.lHt

0.901.(14

1.2

C\

0.611.41

0.521.03

0.54

0.621.00

0.510.S4

(!.571.08

1.9

Mg

0.3(111.60

0.240.54

0.270.55

0.280.55

0.27(1.53

0.280.56

2.(1

SO,

0.81O.Sl

0.750.70

11.710.64

0.640.70

0.520.57

0,66(1.69

1 (!

SiO,

7,626.IX!

7.276.52

7..•166.64

7. OS6.58

7.296.59

7.316.67

(1.9

concentration has been observed in otherCoweeta streams asociatcd with insect intesta-tionsand defoliation (Swank et al., 1981).

We- examined the effeet of this watershedperturbation on litter deeompositicin rate in thestream using leaves from two abtindant treespecies: black locust and sweet bireh {Betulaletita L.). Black locust litter has considerablyhigher nitrogen content than sweet birch litter.Hence it was also possible to determine it' litternitrogen content affected the miinner in whichdecomposition rate responded to the nitrateenrichment.

Methods

Leaves were colleeted just prior to abscissionIVom black locust and sweet birch trees and airdried at room temperature for 2 weeks. Leafletswithout rachises were used from black locust. 5gof leaves were put into bags (17 x 35em. 5mmmesh opening), and twenty-five bags «f leavesfrom each species were placed in similar rifflehabitats in Grady Branch and Sawmill Branch on13 November IWd,

On each samplint: date (10. M). 5S. M7 and 115days after placement in the stream). li\u'

randomly selected bags of each species wereremoved from each stream. Leaf bags wererefrigerated until invertebrates could beremoved from the bags (usually within 24h ofcollection) and stnreil in y5''f ethanol for lateridentification. The remaining leaf material waswashed to remove excess sediment, and thewashings were passed through a mesh bag toavoid loss of leaf particles > 5 mm. Leaves werethen dried to constant weight at >0"C Sampleswere ground in a Wiley mill (4(1 mesh), anti sub-samples were ashed (45(rC' for 6 h) to determineash-free dry weight (AFDW). Other subsampleswere frozen lor later N and P analyses using a[nodified Kjehldahl procedure (TeehnieonAutiiAnalyzerlL 1^77).

Within 1 h ot" removal from the stream, a disc(1.5 cm diameter) was cut from a leaf in each bag.After ATP extraction, the discs were dried andashed for weight measurements. Discs werealways taken from an area of the leaf equidistantfrom the midrib and leaf edge. ATP was extractedfrom each disc by 5 min exposure to 5 ml boilingNaHCO, ((1.1M) (Bancroft. Paul & Wiebe.1476). The extract was ihen added to 5 ml frozenTris buffer for rapid cooling and immediatelyIro/.en, Ihe A"[P content nf eaeh extract was

Leaf decay in a high NO^ stream 179

analysed in triplicate at a later date (Bancroft etal., 1976). Each sample was checked tor inter-ference Irom fulvic acids by spiking with a knownamount of ATP (Cunningham & Wetzel, iy78).

Results

For litter from both species, decay rates were 2,8times greater in the disturbed stream (SawmillBrunch) than in the relerenee stream (GradyBranch) (Table 2). 1 he temporal pattern of eachspecies' weight loss was signiticantly different inthe two streams (analysis of variance, /*<().01).1 n both streams locust leaflets decayed about 1.5times more rapidly than birch leaves (Table 2),

Although locust leaflets had a higher '̂ ^̂ N (mgtotal KJchidahl N/mg AFDW) than birch leavesat all times, both species showed a decline in %Nover time in the disturbed stream (Fig. 1). In thereference stream, locust leaflets showed a slightincrease in Or N over the first 2 months, followedby a decline, while birch leaves showed acontinuous slow decline in %N {Fig. 1). Valuesot f̂-N for leaves were significantly different inthe two streams for each species over time | two-way analysisof viiriance. /^'(I.O?). Because ' > Nvaried as the leaves decomposed, and becausethey decomposed at different rates in the twostreams, a comparison <if TrN over time in thetwo streams docs mn answer the question ofwhether ''rN in leaves at the same state of decaywas the same in the two streams. To answer thisquestion. "/rN was plotted as a function of % oforiginal AFDW remaining (Fig. 2). For bothspecies, leaves from the reference streamappeared to have a slightly higher N content atthe same state of decay; however, thesedifferences were not significant (analysis ofvariance. P > 0.05).

VVL- used nM A FP/g AFDW as an indicator ofmierobial biomass on ihc leaves and observed a

TABLE 2. Exponcniiiil decay cocllkicnts {k) forlocust and birch leaves in two strciims. Values arc theneg;itivo slopes ± 1 SR nt the regression (if the naturallogarithm of ('.' AFDW remaining) v. number nf daysleaves were in ihe streams.

Locust

k (day"' I InSawmill Branch

k {(Jay'') inGrudy Branch

Q

Black lucustSweet birch

0 40 80DAYS

FIG. I, Nitrogen content (HHI x mgN/mg Al-DW)of locust and hirch leaves as a function of number ofdays they were in Ihe refereticc (•) or disturbed (O)stream. Values are the means from live leal packs withone standard error iiidicaled.

general trend of increasing mierobial biomas.swith titne (Fig. 3) and wiih increasing leaf decay.Mierobial biomass on leaves in the two streamswas compared for each species over time: thedisturbed stream had signilicantty greatermierobial biomass (two-\\ ay analysis <»f variance,/'• (1.05). We also compared mierobial biomasson the two leaf species. In the reference stream,hirch leaves had significantly lower mierobialhiomass than locust leaves sampled at llic sametime (analysis of variance, P- 0.05). whereas inthe disturbed stream, there was no signifieantdifference in microbiai biomass between species.

As discussed above for nitrogen, these com-parisons do not reveal whether microbiaibiomass at the same state of leaf decay wassimilar in the two streams. A plot of microbiaibiomass v. % original AFDW remainingrevealed considerable overlap betwcert the twostreams for both species, and an analysis ofvariance showed no signilicant ilifference(P">().05) in microbiai bioniass at similar statesof leaf decavin the two streams.

J. L. Mever and C. Johnson

o

4-

3=-

1-

Locust

0

oo

o

c

c-^ 9

0

^O Q

. . . ^

• ^ o -̂o °o

7-

5-

3-

1-

Birch

o

0

i j

oo

•

• ; • •

o

0

Q

r>^ , O 0 ,

0

oQ

0 20 40 60 80 100

% of ORIGINAL AFDW REMAININGFIG. 2. Nitrogen content ( KKI x mgN/mg AFDW) of lucust and hirch leaves as a tuncilnn of state of decay ol theleaves, i.e. percentage of original AFDW remaining in the leaf pack. Data from both reference (•) and disturbed( o) streams are shown. The solid line indiciiiestjrieinal nitrogen eontcni of the leaves.

Locust

40 80DAYS

120

KIO. .1, Microbiai biomass measured as ATP on ioeustand birch teave>i as a fuiiciion of number of days theywere in the reference (•) or disuirbed (C) siream.Values arc the means fn)ni live leaf packs vvith onestandard error itidieated.

Four genera of Icaf-shrcdding insects werecommon in the leaf packs: Peltoperla, Nemnura,Pvcnopsvche and Tipula. l iiere was no signifi-cant difference between absolute abundances ofshredders found in leaf packs from the twostreams. However, when abundance wasexpressed per g of leaf material, the relativeabundance of shredders was slightly greater inthe disturbed stream (Fig. 4, analysisof variance,P<0X)5),

Discussion

The two streams studied drain watersheds thathave very different treatment histories, and thatdiffer in area by 4iV-^ and in discharge over theperiod of the experiment by 27%; SawmillBranch is the smaller stream. The two streamshave slightly different temperature regimes;Sawmill Branch was about I T warmer in thewinter (J. O'Hop, pers. comm.}. There are alsodifferences in the cationie composition ofstreamwater (Table I). The leaf packs from thedisturbed stream had a significantly higher

Leaf decay in a hifih .VO, stream 181

6-

o

toQ:UJQQLJCCX

en

4-

0-0 20 40 60 80 100 120

DAYSFIG. 4. Mean number (tf shredders per g AFDW oflitter on cnch siimpling diiy. Squares and circlesrepresent locust and hirch leaf packs respectively;closed symbols and siilid lines rcprcsenl the referencestream, while open symbols and dashed lines representthe disturhed siream.

ash content (analysis (if variance. /'<(I.O3),indicating a higher sediment load in that stream.Preliminary seston data also showed higherseston eoncentrations there (J. R. Webster,pers. comm.). The current velocity in similarhabitat types and the statiding stoek of linedetritus are not signitieantly different in the twostreams(Haefner& Wallace. 19SI).

Other studies have det7ions[rated significiintlylower shredder densities in the disturbed stream(Woodall & Wallace. 1972: Haefncr & Wallace.1981). Our data (Fig. 4) indicated slightly highernumbers of shredders/g leaf in the disturbedstrenm. It is likely that our leaf packs wereislands of a resource that was in limited ahun-danee elsewhere in the habitat, since thestanding stoek of leaf detritus was half that in theundisturbed stream (Haefner& Wallace. I9KI).

Higher shredder densities in the leaf packs andthe warmer water tctnperattire in the disturbedstream would accelerate leaf decay in thisstream. The temperature differenee amountedto about 115 degree-days over the course of thisstudy. If IO()1) degree-days arc neeessary for90%of weight loss tooceur (B. J. Hanson and K. W.Cummins, pers. comm.). then the differenee inaccumulated degree days is insufficient toaccount for the nearly threefold increase indecay rates observed. The most striking

difference between the two streams is in nitrateconcetitration. which is 129 times higher in thedisturbed stream (Table I). Hence, thedifferences observed in leaf decomposition rateappear to be primarily a consequence of Nenrichment in the disturbed stream. A higherleafdecay rate hasalso been observed in a streamdraining a clear-cut watershed at Coweeta,although factors in addition to elevated nitratecoticentrations (temperature and availability ofleaves to shredders) wen? ptobably important inthat study (Webster & Waide, 1982).

Other studies do not consistently show astimulatory effect of N enrichment on leafdecay(e.g. Kaushik & Hynes. 1971; Triska &. Sedell,1976). Some of the inconsistency may be due tothe leaf species involved. Cellulose degradationappears to be enhanced by N enrichment, whilelignin decay is not (Egglishaw. i'J72; Federle &Vestal, 1980). Hence a species with a high lignincontent might be expected to show less responseto N enriehtiient. Another more importantfactor leading to inconsistent results in fieldstudies is that streams may differ in ways otherthan just nitrate concentration. In laboratory orcontrolled stream studies where other factorsinlltiencing leaf decay were held constant, thestimulatory effect of N enrichment was apparentwhen adctitiately high concentrations (>ll.2mgNOrN/l) were used (e.g. Howarth Ik Fisher,1976).

In the disturbed stream, both species of leaflitter sht>wed a general decline in ':̂ N over time;in the reference streatn. birch litter showed adecline, while locust litter initially increased in%N, even though its original N content wasabout 3 times that of bireh. These results areperplexing. Mtist studies of leaf deeompositionin streams have reported an increase in %Nduring at least the initial stages of leaf deeay (e.g.Triska. Sedell & Buekley, 1975; Suberkropp.Godshalk & Klug. 1976: Triska & Sedell. 1976;Triska & Buckley. I97K) although some workershave noted little change in '̂ ^N with leafdecay(e.g. Hart &. Howmiller. 1975: Howarth &Fisher. 1976), and others have observed adecline in T-f N (fJart & Howmiller. 1975). Theehanges we observed in ' tV with leafdecay werevery similar to those observed for '̂ ^N (Fig. .>),although another study reported an increase in Pcontent rather than the decline reported here(Meyer. 1980). The inconsistency between pub-lished results with respect to the N and P content

1S2 J. L. Mexer and C. Johnson

10-

05-

Q

Q_

.05-

Locust

Birch

80 120

DAYSFIG. 5. Phosphorus content ( KHI x mgP/mg AFDW)(if loL-usi and htrch leaves as a function of nutnher ofdays the leaves were in the reference (•) or di.'.turbcd(OI stream. Values ure means from five leaf packs wiihone standard crrnr indicated.

of leaves during deeomposition implies thatthere are both species and environmentally-caused (e.g. microbiai flora in the habitat)differenees in the behaviour of the nutrientctintent of leaves during the deeay process.

Both leaf species showed consistently lower'^'fH and %V in the disturbed, high nitratestream. This is surprising beeause most workershave reported either higher %N or no significantdifference in ''̂ •N with N enrichment (e.g.Kaushik & Hynes. 1971: Triska & Sedell, 1976).The differences between streams in %N overtime appear to be a eonsequence of differences indecay rate in the two streams, because %N issimilar at the same state of decay in both streams(Fig. 2). Due to the accelerated leafdecay rates,leaf material that had been in the disturbedstream for I month had a N content comparableto leaves that had been in the reference streamfor 4 months.

1 he more rapid leaf decay in the disturbedstream appeared to be partly due to more rapidmicrobiai processing of the leaves. Otherworkers have noted greater respiration rates ondecaying leaves in N-enriched systems (Alma/an& Boyd, 1978; Anderson, I97S; Durbin, Nixon

&. Oviatt. 1^79). We noted greater microbiaibiomass. Microbiai biotiiass accumulationappeared to be sensitive to N availability. IfN stimulated microbiai growth, one wouldexpect lo see a greater effect of high ambientnitrate concentrations in the species with lowesttissue N. That is preeisel) what was observed. Inthe reference stream, birch leaves had signifi-cantly lower microbiai biomass than locustleaves: but wiih the higher ambient nitrateeoncentrations in the disturbed stream,microbiai biomass on the two species was notsignificantly different. Greater ambient N com-pensated lor lower tissue N in birch leaves: thishas also beet! observed in deeomposing maero-phytes (Carpenter & Adams. I979)and has beensuggested as an explanation for the accelerateddecay of rhododendron litter after elear-cutting(Webster& Waide. I9S2).

Microbiai biomass appears to be a goodindicator of detrital food quality (Ward &Cutnmins. 1979) and therefore the highermicrobiai biomass on leaves in the disturbedstreatn has inieresiing consequences for leaf-shredding insects. Pelioperla inaria Needhamand Smith, an abundant shredder speeies, islarger and has a shorter generation time in thedisturbed stream (J. O'Hop & J. U. Wallace,pers. comm.). 1 his may be parily due to higherquality food in that stream, since the quantity offood available to shredders is lower (Haefner &Wallace. I9H1).

The data presented here indicate that naturalwatershed disturbances sueh as insect infesta-tions or anthropogenic disturbances such as clearcutting, by increasing the nitrate eontent ofstream water, can alter rates of detrital pro-cessing in streams. In a detritus-based ecosystemlike a stream, a change in detrital processingrates could profoundly affect stream ecosystemfunction.

Acknowledgments

We thank L. R. Pomeroy for the use of his ATPphotometer and E. SchuUz for assistance withthe chemical analyses. L, Boring, D. A.Crossley. .Ir. B. Haines, W. T. Swank, J. B.WiLllacc and j . R. Webster provided usefulci>mments on an earlier version of thismanuscript. This research was funded by grantS(J-!2093 from the Ecosystem Studies Programof NSF.

Leaf decay in a high NO, stream 183

References

Alma/iin G. & Boyd C.E. (1978) Effects iif nitrogenlevels on rales of oxygen consumplion duringdL'cay of aquatic planls. Aquatic Botanv, 5, 119-12ft.

Andersen F.O. (1978) Eflects of nuirieiH level on ihedocomposilion of I'hnigmires comnni/u.'i Trin,Archil fiir Hydrohiolofiic, S4, 42-54.

Anderson N.H. & Sodeil. J.R. (iy79) Delriuis pro-cessing by macroinvertebrates in stream eco-systems. Anmtat Review of Entoniologv. 24, 351-377.

Bancroft K., Paul E.A. & Wiebe W.J. (iy7f.) 'i"hcextraction and measurcmcni of ATP from marinesedimenis. Linutology and Ocea/io^rapliv, 21,473-480.

Carpenter S.R. & Adams M.S. (1979) Effects ofnutrients and temperature on decomposition i)fMyriophyllum spicatum L. in a iiard-watereutrophic lake. Limnology and Oceanography, 24.52l»-528.

Cunningham H.W. & Weizei R.G, (1^78) Fulvic acidinterferences on ATP determination in sediments.IJmriolo^y and Oceimof'raphv. 23. 1W)-173.

Diirbin A.G.. Nixon S.W, & Oviati C.A. (1979)Effects of the spawning migration of the aiewifc.Alosti p.ti'udoluirrn^u.'i, on freshwater ecosystems.l:l gy

Egglishaw H.J. (1972) An experimental study of thebreakdown of cellulose in fast-flowing streams.Memorif deli Instituto ludiano di Idrohiohgia(Supplement). 29. 41}K-128.

Elwood J.W.. Newbold J.D.. Trimble A.F. & StarkR.W. (1981) The limiting role of phosphorus in awoodland stream ecosystem; effects of P enrich-ment on leaf decomposition and primaryproducers. Ecology. 62, 14C>-I58.

Fedcrie T.W. & Vestal J.R. (I9S()) Lignocelluloscmineralization by arctic lake sediments in responseto nutrient manipulation. Applied tmd Environ-int-nttd Microliiologv. 4\\, 32-39.

Haefner J.D. & Wallace J.B. (1981) Shifts in aquaticinsect populations in a first-order southernAppalachian stream following a decade of old fieldsuccession. Carttidiim Journal of Fisheries andAquatic Sciences, 38, .35A-359.

Hart S.D. & Howmiller R.P. (1975) Studies on thedecomposition of alliK-hthonous detritus in twosouthern California streams. Proceedings of theInternutiomd .Society for Theoreticiil and AppliedLimnology. 19. 1665-1674.

Howarth R.W. &. Fisher S.G. (1976) Carbon.nitrogen, and phosphorus dynamics during leaf

decay in nutrient enriched stream micro-ecosystems, l-'reshwatcr Hiohgw 6, 221-22S.

Kaushik N.K. & Hynes H.B.N. (1971) The fate ofthedead leaves that fall inU) streams. Archiv fiirHydrohiolof-ie, 68, 4(i5-.'' 15.

Meyer J.L, (msO) Dynamics of phosphorus andorganic matter during leaf decomposition in aforcsl sircam. Oikos. 34, 44-53.

Suherkropp K.. Godshalk G.L. ik Klug M.J. (1976)Changes in the chcmitral composition of leavesduring processing in a woodland stream. Ecology,S7 inin

Swank W.T.. Waide J.B.. C rosslcy D.A.. Jr & ToddR.L. (19SI) hiseci defoliation enhant:e.s nitrateexport from forest ecosystems. Oecologia. 51,297-299.

Technicon AuioAnalyzer II (1977) Modilication ofKjchidahl digestion for tota! I'and KjchldahlN forplani samples. Industrial Method No. .329-47 W/B.Technicon Industrial Systems, larryunvn. NewYork.

Triska F.J. & Buckley B.M. (197X) Patterns ofnitrogen uptake and loss in relation to litterdisappearance and associated invertebratebiomass in sis streams of ihe Pacific Northwest,LI.S.A, Proceedings of the Internaliomil Societyfor Theoreticiil and Applied Limnologv, 20, 1324-1332.

Triska F.J. & Sedell J.R. (1976) Decomposition offour species of leaf litter in response to nitratemanipulalion. Ecoloiiy. 57, 7S3-792.

Triska F.J.. Sedell J.R. & Buckley B. (1975) Theprocessing ol' conifer and hardwood leaves in twoconiferous forest streams. II. Biochemical andnutrient changes. Proceedings of the InternaiionalSocU'tv for Theoretical and Applied Umnnli)i>\',19.162H^1639.

Wallace J.B.. Webster J.R. & Cutfney T. (1982)Stream detritus dynamics: regulation by inverte-brate consumers. Onolofiin, 53, 197-21)1),

WardG.M. &. Cummins K.W. (1979) Effects of foodquality on growth of a stream detritivore, I'ara-temlipes albimamis (Meigen) (Diptera:Chiro-nomidae). Ecology, 60.57-64.

Webster J.R. & Waide J.B. (1982) Effects of forestctearcutting on leaf breakdown in a southernAppalachian slream. Freshwater Biologv, 12,331-344.

Woodall W.R.. Jr & Wallace J.B. (1972) The benthicfauna in four small southern Appalachian streams.American Midland Nutumli.'it. 88, 393—K)7.

f Manuscript accepted 14 September 1982)