Environmental variability, historical contingency, and thestructure of regional fish and macroinvertebrate faunas inOuachita Mountain stream systems

Lance R. Williams”, Christopher M. Taylorh, Melvin L. Warren, Jr.’ & J. Alan Clingenpeel““School c$Nuturul Resources, The Ohio Stutc’ Universit.y, 2021 Cc?fle.y Rmd, Columbus, OH 43210, U.S.A.(e-mail: william.s.2323@osu.edu)hlIepartnzcnt of Biologicul Sciences, Mississippi Stutc University, i? 0. Druwer Gx Mississippi State,MS 39762, U.S.A.‘USDA Forest Service, Southern Research Stution, 1000 Front Street , Oxjbrd, MS 386.55, U.S.A.‘USDA Forest Service, Ouuchitu Nutional Forest, Box 1270, Federul Building, Hot Springs, AR 71902, U.S.A.

R e c e i v e d 5 S e p t e m b e r 2 0 0 2 A c c e p t e d 2 5 M a y 2 0 0 3

Key wo&: assemblage structure, biogeography, canonical correspondence analysis, drainage basin,environmental gradients, local versus regional effects, Mantel test, multivariate analysis, variance partitioning

Synopsis

In 1990-1992, the United States Forest Service sampled six hydrologically variable streams paired in three differentdrainage basins in the Ouachita Mountains, Arkansas, U.S.A. Fishes, macroinvertebrates, and stream environmentalvariables were quantified for each stream. We used these data to examine the relationship between regional faunas(based on taxonomy and trophic affi l iat ion of f ishes and macroinvertebrates) and measured environmental variables.Because fishes are constrained to their historically defined drainage basins and many insect taxa are able to crossbasin barriers , we anticipated that both groups would respond differently to environmental variabil i ty. Fishes wereinfluenced more by environmental variability that was unique to their historical drainage basins, but macroinverte-brates were associated more strongly with environmental variabil i ty that was independent of drainage basins. Thus,the individual drainage basins represented a historical constraint on regional patterns of f ish assembly. For both f ishesand macroinvertebrates, groupings based on taxonomy and trophic affi l iat ion showed a similar response to environ-mental variability and there was a high degree of association between taxonomic and trophic correlation matrices.Thus, trophic group structure was highly dependent on the taxonomic make-up of a given assemblage. At the basin-level, fish and macroinvertebrate taxa were associated more strongly with environmental variabi l i ty than the t rophicgroups, and these results have implications for basin-level studies that use trophic groupings as a metric to assessecological patterns. Trophic categories may not be a useful ecological measure for studies at large spatial scales.

Introduction

Quantifying the relationship between stream organ-isms and their environment is an important area ofresearch in aquatic ecology (Gorman & Karr 1978,Tonn et al. 1983, Allan 1995). The importance ofbiot ic and abiotic features in structuring stream fish andmacroinvertebrate assemblages is well known or amplydemonstrated (Gorman & Karr 1978, Lancaster et al.1990, Jackson & Harvey 1993, Richards & Host 1994,

Clenaghan et al. 1998, Vinson & Hawkins 1998).These relationships, however, are highly dependenton the spatial and temporal scales examined (Wienset al. 1986, Lohr & Fausch 1997, Vinson & Hawkins1998, Lammert & Allan 1999). In the past decade,ecologists broadened their perspectives concerningecological processes across large space and time scales(e.g., the macroecological approach). The longstandingparadigm that community-level properties arise fromecological interactions that are resolved quickly and

204

within small areas and habitats has shilted (Schluter &Ricklefs 1993). Practitioners in the field of macro-ecology now embrace a more balanced view that pat-terns of diversity and community structure are causedby a variety of ecological and historical processes.

Biota in aquatic systems pass through a number ofspatial f i l ters; thus, the assembly of regional faunas dif-fers from that of local communities (Tonn 1990). Atlarge space and t ime scales, biota are influenced mostlyby historical events (e.g., geomorphology, biogeo-graphic history) and climatic factors (Hugueny 1997,Lohr & Fausch 1997, Cooper et al. 1998, Vinson &Hawkins 1998, Ricklefs et al. 1999). These large-scale, historical influences act as ‘filters’ for regionalfaunas and limit the total species pool available atsmaller scales (Tonn 1990, Cucnto & de Casenave1999, Ricklcfs et al. 1999). Although regional factorsand historical events undoubtedly influence asscm-blage structure (Matthews 1987, Brazner & Reals 1997,Lohr & Fausch 1997, Cooper et al. 1998, Vinson &Hawkins 1998, Ricklefs et al. 1999), the specificrole of these processes and their interactive effect onthe structure of aquatic faunas are understood poorly(Hildrew & Giller 1994, Vinson & Hawkins 1998).

In this paper, we used a large data set from theUnited States Forest Service to examine the influenceof environmental variability across a biogeographicalscale on fish and macroinvertebrate regional faunas.These data were collected from six small, hydrolog-ically variable streams paired within three drainagebasins of the Ouachita Mountains, Arkansas, U.S.A.Because much of the flow in these streams is dcpen-dent on rainfall, they are exposed to flood events inthe spring and fall, and they dry to isolated pools inthe summer months (Taylor & Warren 2001). Theseperiodic oscillations in flow can have strong effectson stream biota (Delucchi 1988, Boulton et al. 1992,Taylor et al. 1993, Taylor 1997, Taylor & Warren 2001,Williams et al. 2002) and may limit the ability to detectspecies associations (Angermeier & Schlosser 1989)and species-environment relationships (Braaten &Berry 1997). Assemblage structure after a perturbationcvcnt reflects differential dispersal ability and growthrates of colonists or survivors (Power 1992), btlt thespecific response of assemblages to a disturbance ishighly dependent on the extent of the perturbation andthe spatial scale examined by the investigator (Tayloret al. 1993, Williams et al. 2002).

We were particularly interested in how regionalfish and macroinvertebrate faunas were associatedwith environmental variability measured across larger

spatial scales. Most studies of stream fauna arc lim-ited in their taxonomic scope, and spatial and temporalscales (Jackson & Harvey 1993). Our goal was toaddress variability at a much larger spatial scale inorder to examine the influence of historical and ccolog-ical processes on regional fish and macroinvertebratefaunas. Jackson and Harvey ( 1993) found concordantpatterns between fish and invertebrate assemblagesin lakes across a large region in Ontario, althoughboth groups seemed to respond differently to mca-sured environmental variables. Based on their results,and because our system spans three different drainagebasins, we hypothesized that f ishes and macroinverte-brates would respond to the environment in differentways. Specifically, fishes are constrained to their his-torically defined drainage basins and we predicted thatthis historical constraint would strongly limit theirresponse to environmental gradients at the geographicscale. In contrast, many macroinvertebratc taxa emergefrom the aquatic environment as adults and choosefuture breeding sites based on environmental qualityat a stream locali ty (Anderson & Wallace 1984). Manymacroinvertebrates can also or are able to rapidly recol-onize a site after a disturbance event (e.g., flood 01drought; Stanley et al. 1994). Because of their recol-onization potential and ability to cross basin barriers,macroinvertebrate response to environmental variabil-ity should differ from that of fishes (Plafkin et al. 1989,Lammert & Allan 1999); thus, regional macroinverte-brate faunas should be less constrained by historicaleffects. Finally, we were interested in how these pat-terns vary when faunas were classified by taxonomyversus trophic groups. We reasoned that trophic groupswould be less constrained by phylogeny and biogeo-(Traphic history than those organized by taxa. With taxo-anomic groupings, W C predicted endemic species wouldshow a strong historical constraint; however, whenplaced in trophic categories these endemic speciesbecome less unique and should bc associated moreintimately with local environmental features. Thus,we predicted that t rophic groups WOLII~ show strongerassociations with environmental variables than faunasdefined by taxonomy.

Methods

The six streams in this study are located in forestedwatersheds of the Ouachita National Forest within the

basins of the Saline, Cossatot, and Petit Jean riversin the Ouachita Mountains physiographic province(Table 1, Figure 1). The Ouachitas are a series ofeast-west oriented mountains located in southeastern

T~rhle 1. Streams sampled by the USDA Forest Servicein 1990-1992. All streams are locnted in the OuachitaNational Forest. Arkansas, U.S.A.

S t r e a m D r a i n a g e b a s i n Area (ha)

South Alum Creek S a l i n e R i v e r IS33

Bread Creek S a l i n e R i v e r 1517Caney Creek C o s s a t o t R i v e r 2518

Brushy Creek C o s s a t o t R i v e r 3 4 2 8Dry Creek Peti t Jean 2170

J a c k s C r e e k Petit J e a n 2 9 3 8

20s

Oklahoma and southwestern Arkansas. The moun-tains are composed of Paleozoic sedimentary rock, andstreams are dominated by bedrock, boulder, and cob-ble substrata with some finer sediment interspersed(Robison 1986). The steep gradient of these streamsproduces high-velocity f loods during periods of heavyrainfall. Most streams within the Ouachitas are main-tained primarily by these rainfall events, and theymay become intermittent during dry periods (Taylor &Warren 200 1).

Each pair of study streams is located in geograph-ically disparate river systems, which in turn, sup-port assemblages with distinctive characteristics. TheSaline River is a tributary of the Ouachita R., whichultimately empties into the Red River in east-centralLouisiana. The Cossatot River is part of the Little

Bread Ck.u \

Figrrre I. Location of the six study streams within the Ouachita National Forest, Arkansas, U.S.A. Watersheds of each stream archighlighted, and major drainage basins within the state are outlined. Jacks and Dry Creeks are tributaries of the Petit Jean/ArkansasRivers. South Alum and Bread Creeks drain into the Saline/Ouachita/Red Rivers, and Brushy and Caney Creeks are tributaries of the

L i t t l e / R e d R i v e r s .

206

River system, which also joins the Red River, butdoes so in southwestern Arkansas over 400-river kmupstream of the mouth of the Ouachita River. ThePetit Jean River is a tributary of the Arkansas River(Mississippi River basin) rather than the Red River. Theaquatic fauna1 assemblages of the Ouachita, Litt le, andPeti t Jean river systems differ in several characterist ics,including species composition (Matthews & Robison1988, Poulton & Stewart 1991, Moulton & Stewart1996), species richness, species density, and endemic-ity (Mayden 1985, Crump & Warren 1999, Warren &Hlass 1999, Warren et al. 1999). As all six streamsoccurred in the headwaters with no reservoirs in closeproximity, there was no evidence of species stockingsor introduct ions in these systems (Nice & Fuller 1999).The Ouachita Mountains are bordered to the east andsouth by the Gulf Coastal Plain physiographic provinceand to the north by the wide, low-gradient floodplainof the Arkansas River (Robison 1986). These likelyrepresent barriers to dispersal for many upland aquaticspecies, particularly endemics, effectively isolatingassemblages in tr ibutary systems, such as the Ouachita,Lit t le , and Peti t Jean r ivers . Isolat ion in upland streamsand headwater fauna1 exchange are hypothesized toexplain patterns of genetic diversi ty, shared fauna, andendemism for many aquatic species in this endemic-and species-rich region (Echelle et al. 1975, 1999,Mayden 1985, Cross et al. 1986, Matthews & Robison1988, Poulton & Stewart 1991, Moulton & Stewart1996). Regardless of the mechanism, the streams inthis s tudy are geographical ly isolated and differ in theirfaunas so that consideration of historical constraint asa major component affecting observed assemblages iswarranted.

All streams were sampled in late May to early Augustfrom 1990 to 1992 by Forest Service inventory teams(led by JAC). As streams in the southern portion ofthe Ouachitas typically were more perennial, samplingproceeded from north to south. Streams within the samedrainage were sampled in the same week. Jacks andDry Creeks generally were sampled in June, followedby Bread and South Alum Creeks in July and Caneyand Brushy Creeks in July/August . A mesohabitat clas-sification system proposed by McCain et al. (1990)was used to divide each stream into habitat units fromtheir headwaters to downstream reaches (Will iams et al .2002).

Physical stream features were measured within eachclassified habitat unit. Stream width, mean depth, andthalweg depth were estimated along a transect at themidpoint of each habitat unit. Substrate compositionwas estimated visually as the percentage of six size-classes (i.e., bedrock, boulder, cobble, gravel, sand, andclay/fine sediments). Embeddedness was estimated asthe average percent of cobble surrounded by fine sed-iments. Instream cover factors (i.e., undercut banks,large and small woody debris, whitewater, boulderledges, terrestrial vegetation overhanging the stream,vegetat ion cl inging to substrate, and rooted vegetat ion)were estimated visually as the percentage of area occu-pied within a habitat unit. A clinometer was used to esti-mate bank angle, and bank stability was estimated asthe percentage of the bank that was not eroded. Canopyclosure was determined from a spherical densiometerread at the center of each habitat unit .

Water samples were collected in 10% of each typeof habitat unit (e.g., if 300 mid-channel pools werepresent in a stream, 30 were sampled). These col-lections were stratified along the length of streams(Clingenpeel & Cochran 1992). Water samples fromeach site were analyzed for conductivity, pH, bromide,nitrate, phosphorus, manganese, magnesium, sodium,cobalt, calcium, and sulfate. Analyses were conductedat Berea, Kentucky, U.S.A. with 1983 EPA Methodsfor Chemical Analysis of Water and Wastes. Dissolvedoxygen and temperature were measured in the heldwith adigital meter when water samples were collected.

Fishes and macroinvertebrates were also collectedin 10% of habitat units arranged longitudinally ineach stream. Fishes were collected with multiple-passelectrofishing and block nets (Van Deventor & Platts1985, Clingenpeel & Cochran 1992) and preservedfor identification. Macroinvertebrates were collectedat each site with a kick-net and substrate washing(5 min each). These samples were pooled and preservedfor later identification (Clingenpeel Cyr Cochran 1992).Macroinvertebrates (predominantly insect larvae) wereident i f ied to genus in most cases .

We summarized stream habitat and water chemistrycharacteristics for each stream by year and computedmeans for all measured variables. We calculated anoverall mean and coefticient of variation for depthmeasurements (transect and thalweg). We used 39physico-chemical variables in a principal components

207

analysis (PCA) to reduce the dimensionality of thesedata and applied the broken-st ick model (Jackson 1993)to evaluate axis strength. Eigenvalues exceeded thoseof the broken-stick model for the first five axes, whichaccounted for 72% of the total variance. We did notattempt to interpret PCA axes; rather, we used PCA toobject ively select a small subset ofthe original data thatrepresented the major gradients in the study streams.We retained six variables (conductivity, percent canopycover, percent boulder substrate. CV depth, percentcover of rooted vegetation, and bank stability) for fur-ther analyses because they had the highest respectivecomponent loadings for the first five axes. Conductiv-ity and canopy cover had equally high loadings on thefirst axis, so we retained both variables.

We also summarized fish and macroinvertebrate datafor each stream by year. For taxonomic analyses, weused fish species and macroinvertebrate genera (ororder/family when the genus was not determined). Wedid not distinguish between juveniles and adults forfishes, but most individuals observed were adults . Weclassified fishes and macroinvertebrates into trophicgroups according to Horwitz (I 978) and Merritt &Cummins (19X4), after Allan (1995).

We used canonical correspondence analysis (CCA;ter Braak 1986, ter Braak & Smilauer 1998) to describethe relationships between regional f ish and macroinver-tebrate faunas (taxonomic and trophic) and the reducedset of environmental variables. We then used partialCCA to partition fauna1 variability (Borcard et al. 1992,Magnan et al. 1994, Aude & Lawesson 1998) into threevariance components: ecological effects of the six envi-ronmental variables, effects related to the historicallydefined drainage basins, and their shared variation thatcould not be partitioned into pure effects (i.e., envi-ronmental variat ion unique to the drainage basins). Weused Monte Carlo tes ts (1000 permutat ions) to es t imatethe significance of each model derived from CCA.

We used Mantel tests (Mantel 1967, Fortin &Gurevitch 1993) to examine the association betweenregional faunas and measured environmental variables.We were specifically interested in the amount of vari-ation in taxonomic and trophic matrices that could beaccounted for by the environmental matrix, and howdependent trophic matrices were on the taxonomicmakeup of the regional faunas. For the Mantel test,we used rectangular n x n matrices (e.g., fish taxaby region) to construct triangular correlation matricesacross all pairwise regional comparisons. The Manteltest determines the association between two of thesetriangular correlat ion matrices at a t ime, test ing the null

hypothesis that there is no associat ion within elementsof the matrices. WC tested the ability of the environ-mental matrices to explain variation in the taxonomicand trophic matrices. We also examined the strength ofthe correlation between trophic and taxonomic matri-ces. We expected some correlation because the trophicmatrices were built from taxonomic matrices. Becausethe trophic matrices were constructed from taxonomicmatrices and are not independent, we focused only onthe strengths of matrix correlations (r) and not associ-ated p-values. We also performed partial Mantel tests,which are comparable in function and interpretation topartial correlation analyses (Zar 1996, Manly 1997).We conducted all Mantel tests separately for f ishes andmacroinvertebrates.

Results

A total of 30 species (9 families) of f ishes (Appendix A)and I.52 genera (65 families, 20 orders) of macroinver-tebrates (Appendix B) were collected over the 3-yearstudy period. Examples of all major trophic groups(Horwitz 1978, Merritt and Cummins 1984, Ailan1995) were represented in these data.

For fishes (grouped by taxonomy and trophic affili-ation), streams within individual drainage basins wereassociated closely and separated from other basins inmultivariate space (Figure 2a and b). The six environ-mental variables we used in the analyses successfullyseparated both fish species and trophic groups intotheir respective streams according to drainage basins.In general, streams in the Saline basin had greater aerialcanopy cover and more rooted vegetation. Streams inthe Arkansas basin were characterized by more boul-der substrate and greater bank stability, and streamsin the Cossatot system had higher conductivity andgreater variability in stream depth. Monte Carlo testswere significant for CCA axes 1 and 2 (Figure 2a, p =0.01 and 0.02; Figure 2b, p = 0.01 and 0.05, for axis1 and 2, respectively). Species with negative associa-tion along axis I included: Lepornis hybrids, Noturuslmhneri (endemic to the Saline and Ouachita River sys-tems), Ethcostonw collettei, Micropteru.s ,sulmoidrs,and Aphredoderu,s sayanus. These species were morecommon in tributaries of the Saline River, exceptM. .sdnwicle.s, which was most common in tributariesof the Arkansas River. Species with high positiverelationship with axis 2, indicating strong associa-tion with tributaries of the Arkansas River, included:E. spectcdd~, N. exilis, M. srrlmoides, and Pinwphales

(4 Fish species 0 CossatotW Sal ine

Rooted Veg.

::0.8

0.6

0.4

0.2

0.0

-0.2

-0.4

-0.6

-0.8

1 I

04 Fish trophic groups /il Boulder

iooted Vea.

Bank StabilityA

A

nll Conductivity

40 0.8

0.6

0.4

0.2

0.0

-0.2

-0.4

-0.6

-0.8 i- r---7-2 -1 0 1 2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

CCA 1 CCA 1

Fi’~~ure 2. The relationship between environmental gradients F~,~uw 3. The relationship between environmental gradients(vectors), streams by ycur, and lish taxa (a) and trophic groups (veclors), streams by year, and macroinveilcbrate taxa (a) and

(b) l’or six streams located in the Ouachita Mountains. Shown trophic groups (b) for six Ouachita Mountain streams. Shown ;ire

are the first (x) and second (y) ordinalion axes derived from the first (x) and second (y) ordination axes derived from CCA.

CCA. Axes I and 2 were significant (Monte Carlo rantlomiz;ltioll Both nxes I and 2 were significant (Monte Carlo randomi/.ation

test) t’or 1~0th fish taxa (p = 0.01 and 0.02) and trophic groups test) for taxa (p = 0.01 ) . and the lirst axis was significant for

(p = 0.01 and 0.05). trophic groups (p = 0.0 I ) .

rzotutus. Species located near the center of the plotwere more common in tr ibutaries of the Cossatot River,and included: Cmrpostori~ri ~non1~lun2, E. rdiosurn,F‘rmlul~~ c’utmatus, Awc~iutxr t ~ r h u l o s ~ ~ . r , Notropishoq)s, and Lythrwxs srwlsmi (endemic to the LittleRiver system).

Macroinvertebrate taxa showed a similar pattern tofish species, but the separation of slreams into theirrespective basins was not as evident for macroin-vertebrate trophic groups (Figure 31 and b). Thesame environmental gradients that separated basinsfor fishes also differentiated macroinvertebrate taxaamong basins, and Monte Carlo tests were significant

1.5

1 .o

0.5

0.0

-0.5

-1.0

-1.5

(4 Macroinvertebrate taxa l cossat0tA

L

M S a l i n eA A r k a n s a s

A

Boulder A A l8�Bank Stability:\ ---;-~ CV Depth

A

Canopy ’

qc”

m \b

l Conductivity

c Rooted Veg.=m

cv -2 -1 0 1 2

(b) Macroinvertebrate trophic groupsRooted Veg.,$

I C V D e p t h

l

n �A A

nv 0

ACanopy kYyT Condu;vity

A -. lA lBank Stability A

Boulder

for both axes (p = 0.01 for both). Several taxawere associated with only one drainage basin. Gastro-pods and bivalves were only found in tributariesof the Saline River, as were the families Saididae,Curculionidae, Culicidae and Phoridae. Tributaries ofthe Cossatot River had the most unique taxa, includ-ing: Nematomorpha, Astacidae, Ephemeridae, Isony-chiidae, Calopterygidae, Capniidac, Mesoveliidae,Helicopsychidae and Psychomyiidae. The fami-lies Ephemerellidae, Nemouridae, Glossosomatidae,Tortricidae and Muscidae were unique to tributariesof’ the Arkansas River. For macroinvertebrate trophicgroups, streams were not clearly separated by drainage

209

I E n v i r o n m e n tM E n v i r o n m e n t * B a s i nW Basin

F;,qur~ 4. Amount of variation in fish and macroinvertebrate taxaa n d t r o p h i c g r o u p s e x p l a i n e d b y d r a i n a g e b a s i n s ( b a s i n ) , e n v i r o n -m e n t a l gradienls i n d e p e n d e n t 0T d r a i n a g e b a s i n s ( e n v i r o n m e n t ) ,

a n d their shared var iance components ( i . e . , spa t ia l l y s t ruc turedenv i ronmenta l f ea tures ; environment*basin).

basin (Figure 3b), and Monte Carlo tests were signifi-cant only for the first axis (p = 0.01).

Variance part i t ioning with part ial CCA indicated thatfishes and macroinvertebrates responded differentlyto historical and environmental variance components.Most of the variation for fishes was explained bybasin-level differences in the environmental variables(environmentcbasin; Figure 4). In contrast, environ-mental variation that was independent of drainagebasins explained most of the variation for macroinvcr-tebrates (environment; Figure 4). This was part icularlytrue for macroinvertebrate trophic groups, with morethan 40% of‘thcir variat ion being explained by environ-mental variabil i ty independent of historical , basin-levelconstraint. Drainage basins alone explained only aminor portion (basin; Figure 4) of the variability foreither fishes or macroinvertebrates. This factor repre-sents a purely historical component ( i .e. , drainage basinisolat ion effects) of the analyses.

Fish Matrices

Taxonomic

/0.49 0.70 (0.67)

\.‘.“‘.r

Environment 0.34 (0.01) . Trophic

Macroinvertebrate Matrices

Taxonomic

/0.46 \

0.66(0.66)

k

Environment __~~ ~~~~O.Zl(O.13) + Trophic

Fi,qure 5. R e s u l t s f r o m M a n t e l t e s t s s h o w i n g t h e r e l a t i o n s h i p s ( r )b e t w e e n t a x o n o m i c a n d t r o p h i c m a t r i c e s and la rge-sca le env i ron-menta l va r i ab i l i t y . Resu l ts f rom par t i a l Mante l t es ts a re ind ica ted

in parentheses.

Mantel tests indicated that fish and macroinverte-brate trophic groups were associated with environ-mental variables and taxonomic matrices but wereassociated most strongly with taxonomic groups. Thestrongest correlations were between trophic and taxo-nomic matrices for both fishes and macroinvertebrates(Figure 5). Correlations between trophic groups andenvironmental variabil i ty were not as strong, and whenthe effect of taxonomy was removed (partial Mantel;For-tin & Gurevitch 1993), there was little correlationleft between trophic groups and environmental vari-ables (Figure 5). Thus, the correlat ions between trophicand environmental matrices were confounded by thestrong relationships between trophic and taxonomicmatrices. For fishes particularly, these results are con-gruent with the s imilar pat terns observed for taxonomicand trophic groups when variance was partitionedamong effects (Figure 4). Perhaps the most interest ingpattern to emerge from these analyses was that f ish andmacroinvertebrate taxa were associated more stronglywith environmental variables than were trophic groups(Figure 5). This pattern was the opposite of our origi-nal predict ion that t rophic groups would be associatedmore s trongly with environmental var iabi l i ty than taxa.

Discussion

Streams can be extremely complex and variableenvironments, and this variability contributes tothe dynamic nature of their fauna1 assemblages(Lancaster et al. 1990). At the stream reach scale,

210

environmental variability is important in structuringfish and macroinvertebrate assemblages (Gorman &Karr 1978, Matthews et al. 1988, Lancaster et al.1990, Richards & Host 1994, Clenaghan et al. 1998,Lonzarich et al. 1998, Taylor & Warren 2001). How-ever, species assembly at smaller spatial scales isalso influenced by regional and historical processes(Hugueny 1997, Ricklefs et al. 1999), and knowledgeof these regional and historical influences are criticalfor rehiring hypotheses of community assembly andorganization for stream biota (Brooks & McLennan1993, Caley & Schluter 1997, Angermeier & Winston1998). Large-scale influences provide a physical andbiogeographic framework for stream systems withinwhich local processes must operate (Tabacchi et al.1998). Thus, it is important to understand the role ofhistory and large-scale environmental effects on theabundance and distr ibution of regional faunas.

Because most fish species are constrained to theirhistorically defined drainage basins and many insecttaxa (after emergence) are free to cross basin barri-ers, we anticipated that each taxonomic group wouldrespond differently to the environmental variables thatwere examined. Fish species were influenced moreby spatially structured environmental features (i.e.,environmental variabil i ty associated with an individualdrainage basin) than variability that was independentof the drainage basins. Environmental variabil i ty asso-ciated with drainage basins likely reflects historicaldifferences in geology and climate, and the resultingdifferences in hydrology and other physical streamfeatures. This type of variability could also representdifferences in land use among basins, but Will iams et al .(2002) used these same data to address land use impactsand found no significant impacts of silvicultural activ-ities on regional fish or macroinvertebrate faunas inthese systems. Thus, as we predicted, historical con-straint was an important l imitat ion on regional fish fau-nas. History, in this case, was operating as a constrainton local habitat condit ions, which, in turn, affected thetypes of assemblages that would occupy those streams.

In contrast, macroinvertebratc taxa were infu-enced more strongly by environmental variabilitythat was independent of drainage basins, indicatingless historical constraint than for fishes. In short,macroinvertebrates were associated with large-scaleenvironmental gradients regardless of basin. Insectsassociated with both terrestr ial and aquatic systems atdifferent life-history stages can choose breeding sitesbased on environmental condit ions at a s tream local i ty(Anderson & Wallace 1984) and have the potential

to cross drainage basin boundaries. These same basinboundaries represent historical barriers to fish disper-sal. In general, macroinvertebrates seem to respondmore to local environmental condit ions that vary acrossa larger geographic region than do fishes, which aremore affected by environmental factors unique to thedrainage basins that contain them (Lammert & Allan1999).

Fish taxa and trophic groups showed a similarresponse to the measured environmental variablesand also similar degrees of historical constraint. Formacroinvertebrates, trophic and taxonomic groups alsoseemed to show similar patterns, although the trophicgroups were less constrained by basin-level propert ies.These results were surprising, particularly for macro-invertebrates, considering that the trophic groupingswe used reflect their functional feeding role in streams(Merritt & Cummins 1984, Allan 1995). We did notexpect such a similar pattern between the taxonomicand trophic groups for fishes or macroinvertebrates.We predicted that trophic groups would be stronglyassociated with measured environmental variables(independent of drainage basins) and less influencedby historical, basin-level processes than taxonomicgroupings. This was generally the case for macroin-vertebrates, but for tishes, both t rophic and taxonomicgroups separated along the same mult ivariate gradients .

Trophic groupings for fishes and macroinvertebrateswere associated with both environmental variables andtaxa, with the strongest correlations between trophicand taxonomic matrices. In contrast to our originalprediction, we found that taxa were associated moreclosely with the environmental variables than trophicgroups, at least at the drainage basin scale. Origi-nally, we predicted that the trophic groups would havea stronger association with large-scale environmen-tal variability because of their relative freedom fromphylogenet ic and his tor ical constraints , but this was notthe case. Although the causal mechanisms responsiblefor this pattern are beyond the scope of our analyses,we propose two potential explanations. The first is anissue of scale, in that these patterns may be affected bythe basin-level sampling design. We were essentiallycomparing the ‘entire’ species pool of one stream withthe ‘entire’ species pool of other streams, rather thancomparing discrete samples from discrete habitat types.At this broad spatial scale, t rophic groups may be toocoarse of a classilication system to elucidate meaning-ful relationships with environmental variability. Thesecond explanation is related to the trophic categoriesthemselves. Our results are highly dependent on the

211

trophic classification system we used (see Horwitz1978, Merritt & Cummins 1984, Allan 1995). The pat-terns we found may differ with more detailed trophicgroupings. If trophic groups are to be used as an eco-logical measure in basin-level studies, a more thoroughanalysis of the relat ionships between different t rophicclassification schemes and their taxonomic make-upshould be attempted. Perhaps better trophic classi-fication systems are needed for stream organisms,particularly for fishes where most schemes are basedon diets rather than functional feeding roles (Matthews1998). Because many govermnent agencies are recom-mending and adopting these types of basin-level s tudiesto manage aquatic systems (Ponce et al. 1982, Hankin1984, Hankin & Reeves 1988, Clingenpeel & Cochran1992, Williams et al. 2002), the limitations of somemeasures (i.e., trophic groupings) in elucidating eco-logical relat ionships need to be addressed. As modeled,trophic groups had a comparatively weak relationshipwith environmental variability at the regional scale.

One of the primary goals of ecology is to quantifythe relat ionship between species and their environment(Krebs 2001), but understanding these relationshipsis complicated by their strong dependence on spa-tial and temporal scale (Wiens et al. 1986, Lohr &Fausch 1997). In stream systems, factors like history,climate, and hydrology are important in determiningthe structure of assemblages at larger scales. As the spa-tial scale is reduced, local habitat variability becomesmore important (Poff & Ward 1989, Wiley et al. 1997,Cooper et al. 1998, Pusey et al. 1998). As shownhere, the structure of regional fish (and to a lesserdegree macroinvertebrate) faunas was dependent inlarge part on biogeographical history and the degreeof spat ial constraint operat ing on environmental gradi-ents . Although most s tudies of s treams are conductedat smaller scales, large-scale and historical factors mustbe considered to fully understand the structure andfunction of local assemblages (Brooks & McLennan1993, Angermeier & Winston 1998, Caley & Schluter1997). If limitations (e.g., interpretation of causal rela-tionships) are understood, multivariate analyses canbe a useful tool to assess these relationships at differ-ent scales and contribute to our ability to understand,

predict , and effectively manage aquatic systems (Tonnet al. 1983, Williams et al. 2002).

Acknowledgements

Financial support for th is work was prov ided

by the USDA Forest Service, Southern Research

Station, Ouachita Mountains Ecosystem ManagementResearch Project, Ouachita National Forest, and the

Department of Biological Sciences at Mississippi StateUniversi ty. We thank personnel provided by the ForestService, Northeast Louisiana University, and ArkansasState University for their assistance. We also thankR. Altig, E. Dibble, F. Vilella, M. Williams and D. Wisefor helpful comments on the manuscript.

References

Allan, J.D. 1995. Stream Ecology: Structure and Function ofRunning Waters. Chapman & Hall, London, 388 pp.

Anderson. N.H. & J.B. Wallace. lYX4. Habitat, life history,and behavioral adaptations of aquatic insects. pp. 3X-58. 111:R.W. Merritt & K.W. Cummins (ed.) An Introduction 10 theAquatic Insects, 2nd edn, Kendall/Hunt Publishing Company,D u b u q u e , I o w a .

A n g e r m e i e r , P.L. & 1.J. S c h l o s s e r . 1 9 8 9 . S p e c i e s - a r e a r e l a t i o n -s h i p s f o r s t r e a m tishes. E c o l o g y 7 0 : 1 4 5 0 - 1 4 6 2 .

Angermeier, P.L. & M.R. Winston. 1998. Local vecs~s regionelinfluences on local diversity in stream fish communities ofVirginia. Ecology 79: Y I I-927.

Aude, E. & J.E. Lawesson. 1908. Vegetation in Danish beechf o r e s t s : T h e i m p o r t a n c e o f s o i l , m i c r o c l i m a t e ~rnd m a n a g e m e n tfactors, evaluated by variation partitioning. Plant Ecol. 134:53-65.

Borcard, D., P. Legendrc & P. Drapeau. 1992. Partialling outt h e s p a t i a l c o m p o n e n t o f e c o l o g i c a l v a r i a t i o n . E c o l o g y 7 3 :1045-1055.

Boulton, A.J., C.G. Peterson, N.B. Grimm & S.G. Fisher. 1992.S t a b i l i t y o f a n a q u a t i c m a c r o i n v e r t e b r a t e a s s e m b l a g e c o m m u -n i t y i n il m u l t i y e a r h y d r o l o g i c d i s t u r b a n c e r e g i m e . E c o l o g y 7 3 :2 192-2207.

Braaten, P.J. 81 C.R. Berry. 1997. Fish associations with four hahi-tat types in a South Dakotn prairie stream. J. Freshwater Ecol.12 : 4 7 7 1 1 8 9 .

Bmzner, J.C. & E.W. Deals. 1997. Patterns in lish asscmhlagcsf r o m c o a s t a l welland a n d b e a c h h a b i t a t s i n G r e e n B a y , LakeM i c h i g a n : A m u l t i v a r i a t e a n a l y s i s o f abiotic a n d b i o t i c f o r c i n gfactors. Can. J. Fish. Aquat. Sci. 54: 1743-1761.

Brooks. D.R. & D.A. McLennan. 1993. Historical ecology:E x a m i n i n g p h y l o g e n e t i c c o m p o n e n t s ofcommuniry e v o l u t i o n .p p . 267-280. In: R . E . R i c k l c f s & D . Schlutcr (ed.)Species Diversity in Ecological Communities: Historicalrind Geographical Perspectives, University of Chicago Press.C h i c a g o , I l l i n o i s .

Caley, M . J . & D . S c h l u t e r . 1 9 9 7 . The r e l a t i o n s h i p b e t w e e n l o c a la n d regional d i v e r s i t y . E c o l o g y 7 X : 70-80.

Clenqhan, C.. P.S. Giller. J. O’Halloran & R. Hcrnan. 199X.S t r e a m m a c r o i n v e r t e b r a t e c o m m u n i t i e s i n a c o n i f e r - a f f o r e s t e dc a t c h m e n t i n I r e l a n d : R e l a t i o n s h i p s t o p h y s i c o - c h e m i c a l a n db i o t i c f a c t o r s . F r e s h w a t e r B i o l . 4 0 : I7S-lY3.

Clingenpecl, J.A. & B.G. Cochmn. 1992. l ising physical, chen-ical and biological indicators to assess water quality on theOuachita National Forest using basin xcil stream survey

212

methods. Proceedings of the Arkansas Academy of Scicnccs46: 33-35.

Coopcr, S.D., S. Diehl, K. Kratz & 0. Sarnellc. 1998. Implicationsof scale for patterns and processes in stream ecology. Aust. J.Ecol. 23: 27-40.

Cross. F.B., R.L. Mayden & J.D. Stewart. 19X6. Fishes in the west-cm Mississippi Basin (Missouri, Arkansas, and Red Rivers).pp. 363%412. I/7: C.H. Hocutt & E.O. Wiley (ed.) The Zoo-geography of North American Freshwater Fishes. John Wiley &Sons. New York.

Grump, B.G. Kc M.L. Warren, Jr. 1999. Diversity of aquaticinsects. pp. 140-145. In: U.S. Department of Agriculture.Forest Service. Ozark-Ouachita Highlands asscssmcnt: aquaticconditions. Report 3 of 5, Southern Research Station, GeneralTechnical Report SRS-33. Asheville, North Carolina.

Cuento, V.R. & J.L. de Cnscnave. 1999. Determinants of birdspecies richness: role of climate and vegetation structure at aregional scale. J. Biogeogra. 26: 487-492.

Delucchi, CM. 198X. Comparison of community structure amongstreams with different temporal How regimes. Can. J. Zool. 66:579-586.

Echelle, A.A., A.F. Echelle, M.H. Smith & L.G. Hill. 1975. Elcc-trophoretic analysis of genie continuity in a headwater fish,Etheostotncr ,rrdio.smr. Copeia 1975: 197-204.

Echelle, A.F., A.A. Echclle, L.R. Williams, C.S. Toepfer &W.L. Fisher. 1999. Allozyme perspective on genetic varia-tion in a threatened percid fish, the leopard darter (Percinu

prrrzrhe~inu). Am. Midland Naturalist 142: 393400.Fortin, M. & J. Gurevitch. 1993. Mantel tests: Spatial struc-

ture in field experiments. pp. 342-359. In: S.M. Scheiner &J. Gurevitch (ed.) Design and Analysis of EcologicalExperiments, Chapman and Hall, New York.

Gorman, O.T. & J.R. Karr. 1978. Habitat structure and stream fishcommunities. Ecology 59: 507-S IS.

Hnnkin, D.G. 1984. Multistage sampling design in fisheries:Applications in s~nall streams. Can. J. Fish. Aqua. Sci. 41:1575-1591.

Hankin, D.G. & G.H. Reeves. 1988. Estimating total fish abun-dance and total habitat area in small streams based on visualestimation methods. Can. J. Fish. Aqua. Sci. 45: X34-844.

Hildrew, A.G. &P.S. Giller. 1994. Patchiness, species interactionsand disturbance in the stream henthos. pp. 2 l-62. In: P.S. Giller,A.G. Hildrew & D.G. Raffaelli (ed.) Aquatic Ecology-scale,Pattern, and Process, Blackwell Science, Oxford.

Horwitz, R.J. 1978. Temporal variability patterns and the distri-butional patterns of stream fishes. Ecological Monographs 4X:307-32 I.

Hugueny, B. 1997. The relationship hctwcen local and regionalspecies richness: Comparing biotas with different evolutionaryhistories. Oikos 80: 5X3-587.

Jackson, D.A. 1993. Stopping rules in principal components anal-ysis: A comparison of heuristical and statistical approaches.Ecology 74: 2204-22 14.

Jackson, D.A. & H.H. Harvey. 1993. Fish and benthic inverte-brates: Community concordance and community-environmentrelationships. Can. J. Fish. Aqua. Sci. SO: 2641-265 I.

Krebs, C.J. 2001. Ecology, the Experimental Analysis ofDistribution and Abundance, 5th edn, Benjamin Cummings,San Francisco, California, 695 pp.

Lammert, M. Kr J.D. Allan. 1999. Assessing the biotic integrityof streams: Effects of scale in measuring the inRuence of landuse/cover and habitat structure on tish and rrtacroinvertehrates.Environ. Manag. 23: 257-270.

Lancaster. J., A.G. Hildrew & C.R. Townsend. 1990. Streamflow and predation effects on the spatial dynamics of benthicinvertebrates. Hydrobiologia 203: 177-190.

Lohr, S.C. & K.D. Fausch. 1997. Multiscale analysis of naturalvariability in stream fish assemblages of a western Great Plainswatershed. Cop& 1997: 706-724.

Lonzarich, D.G.. M.L. Warren, Jr. & M.R.E. Lonlarich. 1998.Effects of habitat isolation on the recovery of fish assemblagesin experimentally defaunated stream pools in Arkansas. Can.J. Fish. Aquat. Sci. 55: 2141-2149.

Magmm, P., M.A. Rodriguez, P. Legendrc Cyr S. Lacasse. 1994.Dietary variation in a freshwater lish species: Relative contribu-tions ofbiotic interactions, abiotic factors, and spatial structure.Can. J. Fish. Aquat. Sci. 5 I: 2X56-2865.

Manly, B.F.J. 1997. Randomiaition, Bootstrap and Monte CarloMethods in Biology, 2nd edn, Chapman and Hall, London,399 pp.

Mantel, N. 1967. The detection of disease clustering and ageneralized regression approach. Cancer Res. 27: 209-220.

Matthews, W.J. 1987. Physicochemical tolerance and selcctiv-ity of stream fishes as related to geographic ranges and localdistributions. pp. II l-120. 1/l: W.J. Matthews & D.C. Heins(ed.) Community and Evolutionary Ecology of North AmericanStream Fishes, University of Oklahoma Press, Norman,Oklahoma.

Matthews, W.J., R.C. Cnshner & F.P. Gelwick. 1988. Stability andpersistence offish faunas and assemblages in three midwesternstreams. Copeia 1988: 945-955.

Matthews, W.J. & H.W. Robison. 1988. The distribution offishesin Arkansas: a multivariatc analysis. Copeia 1988: 358-174.

Matthews, W.J. 1998. Patterns in Freshwater Fish Ecology,Chapman & Hall, New York, 756 pp.

Mayden, R.L. 1985. Biogeography of Ouachita highland fishes.Southwestern Naturalist 30: 195-2 I 1.

McCain, M., D. Fuller, L. Decker & K. Overton. 1990. Streamhabitat classification and inventory procedures for northernCalifornia. USDA Forest Service, Pacific Southwest Region.FHR Currents I: l-16.

Merritt, R.W. & K.W. Cummins. 1984. An introduction to theaquatic insects of North America. 2nd edn, Kendall/HuntPublishing Company, Dubuque, Iowa, 722 pp.

Moulton, S.R., II & K.W. Stewart. 1996. Caddisflies (Trichoptera)of the Interior Highlands of North America. Memoirsof the American Entomological Institute 56, AmericanEntomological Institute, Gainesville, Florida. 33 I3 pp.

Nice, L.G. & P.L. Fuller. 1999. Spatial and temporal patterns ofnonindigenous fish introductions in the United States. Fisheries24: 16-27.

PI&in, J . L . , M . T . Barbour, K.D. Por ter , S .K. Gross &R.M. Hughes. 1989. Rapid Bioassessment Protocols for Usein Stream and Rivers: Benthic Macroinvertebrates and Fish.EPA/444/4-X9-01 I. U.S. Environmental Protection Agency,Washington, D.C.

Poff, N.L. & J.V. Ward. 1989. Implications ofstreamllow variabil-ity and predictability for lotic community structure: A regionalanalysis of strcamfiow patterns. Can. J. Fish. Aquut. Sci. 46:1805-1818.

Ponce, S.L., D.W. &hinder & R.C. Avertt. 1982. The use of thepairctl-basin technique in few-related wildland water-qualitystudies. USDA Forest Service, Watershed Systems Devcl-opmcnt Group, WSDG-TP-00004. Fort Collins. Colorado.21 pp.

Poulton, B.C. & K.W. Stewart. I99 I. The stoneflies of theOzark and Ouachita Mountains (Plecopter-a). Memoirs of theAmerican Entomological Society Number 38, American Ento-

mological Society, Academy ofNatural Sciences. Philadelphia.Pennsylvania, I I6 pp.

Power. M.E. 1992. Top-down and botton-up l’orces in food webs:

Do plants have primacy? Ecology 73: 733-746.Pusey, B.J.. A.H. Arthington Xc M.G. Read. 1098. Freshwatcl

fishes of the Burtlekin River, Australia: Biogeography, historyand spatial variation in community structure. Environ. Bid.

Fish. 53: 303---31X.Richards, C. & G. Host. 1994. Examining land use inllucnces on

stream habitats and macroinvertchratcs: a GIS approach. WaterResources Bull. 30: 729-738.

Ricklcfs. R.E., R.E. Latham & H. Qian. IYYY. Global pat-terns ol‘tree species richness in moist I’orcsts: Distinguishingecological influences and historical contingency. Oikos 86:36Y-373.

R o b i s o n . H . W . I986 Zoogeographic imp l i ca t i ons o f t heMississippi River basin. pp. 267-286. In: C.H. Hocutt &50. Wiley (ed.) The Zoogeography of North American Fresh-

water Fishes, John Wiley and Sons, New York.Schluter, D. & R.E. Ricklefs. 1993. Species diversity. An

introduction to the problem. pp. I-IO. In: R.E. Ricklefs &

D. Schluter (etl.) Species Diversity in Ecological Communi-ties, Historical and Geographical Perspectives, The Universityof Chicago Press, Chicago, Illinois.

Stanley, E.H., D.L. Buschman, A.J. Boulton, N.B. Grimm &

S.G. Fisher. 1994. Invertebrate resistance and resilience tointcrmittency in a desert stream. Am. Midland Naturalist I3 I :2X8-300.

Tabacchi, E., D.L. Correll, R. Hauer, G. Pinny, A. Planty-

Tabacchi & R.C. Wissmar. 1998. Development. maintenanceand role of riparian vegetation in the river landscape.Freshwater Bid. 40: 4Y7---5 I(,.

Taylor, C.M.. M.R. Winston & W.J. Matthews. I Y93. Fish specics-

environment and abundance relationships in a Great Plains riversystem. Ecography 16: 16-23.

Taylor, C.M. I9Y7. Fish species richness and incidence patterns

in isolated and connected stream pools: Elf‘ects of pool volumeand spatial position. Oecologia I IO: 560-566.

213

Taylor, C.M. & M.I.. Warren, Jr. 2001. Dynamics in species com-position of stream fish assemblages: Environmental variabilityand neskd subsets. Ecology X2: 2320-2330.

ter Braak, C.J.F. 1986. Canonical correspondence analysis: A new

eigenvcctor technique fitr multivariate direct gradient analysisin ecology. Ecology 67: I l67- I 179.

ter Braak, C.J.F. & P. Smilauer. 10%. CANOCO Rctercnce Man-ual and User’s Guide to Canoco for Windows: Software h

Canonical Community Ordination (version 4). MicrocomputerPower, Ithaca, 352 pp.

Tonn, W.M., J.J. Magnuson & A.M. Forbcs. lY83. Community

analysis in fishery management: An application with northernWisconsin lakcs. Transactions of the AIN. Fish. Sot. 112:368-377.

Tonn, W.M. 1990. Climate change and fish communities: Aconceptual framework. Trans. Am. Fish. Sot. I 19: 337-352.

Van Dcventor. J.S. & W.S. Platts. IYXS. A computer software

system for entering, managing, and analyzing lish capture data

f’rom streams. USDA Forest Service, Intermountain RcscarchStation, Ogden. Utah, I2 pp.

Vinson, M.R. & C.P. Hawkins. IYYX. Biodivcrsity of stream

insects: Variation at local, basin, and regional scales. Annu.Rev. Entomol. 43: 27 l-293.

War ren , M .L . , J r . & I . . Hlass. 1999. D i v e r s i t y OF lishcs.pp. 100-l 14. In: U.S. Department of Agriculture, Forest

Service. Orark-Ouachita Highlands assessment: Aquatic con-ditions. Report 3 of 5, Southern Research Station, GcncralTechnical Report SRS-33. Asheville, North Cnrolina.

Warren, M.L., Jr., H.W. Robison & K. Tinkle. 1999. Diversity

oI’ crayfishes. pp. 132-l 39. 1,~: IJ.S. Department ol‘ Agricul-ture, Forest Service. Ozark-Ouachita Highlands assessment:Aquatic conditions. Report 3 of 5, Southern Research Station,

General Technical Report SRS-33, Asheville, North Carolina.Wiens, J.A., J.F. Addicott, T.J. Case & J. Diamond. 1986.

Overview: The importance ofspatial and temporal scdc in cco-

logical investigations. pp. 145-l 53. 111: J. Diamond 6t T.J. Case(ed.) Community Ecology, Harper & Row, New York.

Wiley, M.J., S.L. Kohler & P.W. Seelbach. 1997. Reconcilinglandscape and local views of aquatic communities: Lessonsfrom Michigan trout streams. Freshwater Bid. 37: 133--14X.

Williams, L.R., C.M. Taylor, M.L. Warren Jr. Xr J.A. Clingcnpeel.2002. Large-scale effects of timber harvesting on stream sys-tems in the Ouachita Mountains, Arkansas, U.S.A. Environ.Manag. 2Y: 76-87.

Zar, J.H. 1996. Biostatistical Analysis, 3rd edn, Prentice Hall,U p p e r Saddle R i v e r , N e w J e r s e y , 662 p p .

Appmdi.x A. Fish species collected by IJ.S. Forest Service inventory tea~ns in the six study streams(grouped by basin) and their assigned trophic groups.* Presence in a drainage basin indicated by ‘I’.

Trophic group Saline

ALG IGI 0

WCI 0WC1 I

C0ssat01 Arkansas

I II 0I 0I I

214

Trophic group Saline COSSatot Arkansas

WC1OMNGI 1

OMN

OMNOMNB IB l

00

OMN I

SlSl

00

WC1 0 0

GIClGIGlGIPISPI.7PIS

10I00000

00

00

BIBlBIBIBl

100

0 0II

I 0 0

*Assigned trophic groups are as follows: ALG = a&ore, BI = benthic insectivore, GI = generalinsectivore, OMN = omnivore, PIS = piscivore, Sl = surfzace insectivore, and WC1 = water-columninsectivore.

Appcnc1i.r B. Macroinvertebmtc taxa collected by U.S. Forest Service inventory teams in the sixstudy streams (grouped by basin) and their assigned trophic groups*. Presence in a drainage basinindicated by ‘ 1’.

Order/family (no. genera) Trophic group Saline Cossatot Arkansas

GastropodaAncylidae (1)Planorbidae ( I )

HivalviaSphacriidae

NcmatomorphaOligochactaBranchiohdellidaHirudinen

Sl1S H

CFNFCGc oP A

I 0 0I 0 0

I 0 00 I 01 I II I II I 0

215

Order/family (no. genera) Trophic g r o u p S a l i n e C o s s a t o t Arkansas

AmphipodaCrangonyctidae (2)Talitridae (1)

Isopoda

Asellidae (2)Decapoda

AstacidaeCambaridac (2)

H y d r a c h n i d aHydrachnidae

Collembola

EphcmcropteraRaetidae (3)C a e n i d a e ( I )

Ephemerellidae (2)Ephemeridae (2)Heptageniidae (4)L e p t o p h l c b i i d a c ( 3 )

lsonychiidac (I)Lcptohyphidae (I)

OdonataAeshnidae (I)

Calopterygidae (2)C o e n a g r i o n i d a e ( 2 )Corduliidae (3)Gomphidae (3)

Libellulidae (I)Plecoptera

Capniidae ( I )

Leuct r idaeNcmouridae (I)Per l idae (5)

Hemiptera

Corixidae (2)Gerridac (4)Mesoveliidae (I )N o t o n e c t i d a e ( I)

Saldidae (I)Veliidae (2)

MegaloptcrnCorydalidae (2)

Sialidae (I)Trichoptera

Glossosomatidae ( I )

Helicopsychidae (I)Hydropsychidae (4)Lepidostomatidae ( I )Leptoccr idae (2 )

Limnephilidae (2)Philopotamidae (2)Polycentropodidae ( I )Psychomyiidae (I )

R h y u c o p h i l i d a e (I)

CC ICG I

CG I

CG 0CG I

P R ICG I

CG IC G 1

CG 0CG 0S C 1

CG I

CF 0CG I

PR

P RP RP RP R

P R

SH

S WS HP R

I

0I

II

I

0

I01

P IP IP IP I

P IPI

P R

P R

S C

S CCFS H

CG. PRSC, SHCFP R

S CP R

I

I

0

0II1

III0

I

0I

I

I

I

II

II

0II

III

IIII

II

I

I0

I

0II

I0

I

II

0

II

I

I0

I

III

II

I

0I

I0

II

I0

II

00

I

0I0

I

0

0

III

1I

0

I0

I

I

I

I

0II

0III

0I

2 1 6

Order/family (no. genera) Trophic g r o u p S a l i n e Cossalot Arkansas

L e p i d o p t e r aP y r a l i d a c ( I )

Tortricidac (1)Coleoptera

Curculionidne ( I )

Dryopidac (1)Dytiscidae (7)Elmidae (5)Gyrinidae (I)

Hydrophilidae (4)Psephen idae (2 )S t a p h y l i n i d a e (I)

D i p t e r a

C e r a t o p o g o n i d a e ( 2 )Chironomidae (35)Culicidae (1)

Dixidae (2)Empididae (I)Muscidae (I)P h o r i d a e

Simuliidae (2)Tnbanidae (4)Tipulidae (5)

S HSH

S H

S HPI, PRCGP R

PI, PRS CP R

P RCG, PR, SH. CF, SC

CFCCP RP I

P RC PCG, PI

PR, SH, Cc;

I10

II

0

0

I

0

0

II

I

0

I

0

0I

I

‘Assigned trophic groups are its follows: CF = collector-tilterer, CG = collector-gatherer, CO =

commensal, NF = non-feeding, PI = predator-piercer, PR = predator-engulf&, SC = scraper, andSH = shredder.