The 16 contributors who produced the 15 chapters in Robert Usinger's Aquatic Insects of California (AIC), published by the University of California Press in 1956, included accomplished taxonomists experienced with aquatic insects, others who specialized in nonaquatic insect groups, and still others who were self-trained or worked in nonacademic positions. Richly illustrated, AIC provided species-level keys to adults of most insect groups, a feature not seen in subsequent identification guides to North American aquatic insects. This change might reflect the recent trend of reliance on identifications to higher taxonomic levels in benthic studies, particularly in biomonitoring, and less emphasis on association of adult and immature stages of aquatic insects. Ecological information presented in AIC emphasized mosquito control and anxieties about potential malarial outbreaks and the importance of aquatic insects in sewage treatment. Detailed sampling methods for streams reflected those being developed in California because few descriptions of stream sampling methods were then available. The role of the taxonomist who was self-trained, nontraditionally employed, or had broad-based interests might have parallels today. Identification advances in benthic biology are now being done more often by biomonitoring researchers in government laboratories and consulting firms than in university research programs.

Substrate stability is a key determinant of stream invertebrate community composition, but its measurement can be problematic. Stream ecologists often use different approaches and techniques to quantify bed stability, and this variability makes comparison among studies difficult. We examined the link between 6 reach-scale measures of substrate stability and invertebrate community metrics in 12 New Zealand mountain streams. The strength of the link varied with the method used to define substrate stability. We used morphological budgeting to measure spatial patterns and volumes of scour and fill. We found that as erosion of sediments increased, invertebrate diversity declined exponentially. In particular, increases in the volume of scour reduced taxonomic richness, whereas deposition of coarse sediments was less relevant for invertebrate communities. Overall, the distance travelled by in-situ-marked tracer stones was most strongly linked with all invertebrate community metrics, whereas the bottom component of the Pfankuch Index related very well to diversity. Both metrics showed near-linear declines in diversity with decreasing stability. In contrast, the link between invertebrate communities and the proportion of bed area affected by entrainment was weak. Therefore, we propose tracer-based indices and the Pfankuch bottom component as the most suitable measures for research involving invertebrate–substrate-stability relationships. Measures derived from in-situ-marked tracer stones reflected only entrainment and transport of particles. In contrast, the bottom component of the Pfankuch Index encompassed the widest range of bed-stability characteristics but is prone to observer bias. An objective method that combines the efficiency of the Pfankuch Index with the characteristics measured using tracer stones could serve as a powerful explanatory tool in stream ecology.

Urban land use alters channel morphometry, particle size structure, and sediment-transport dynamics in stream ecosystems, thereby degrading the habitat of aquatic organisms. However, stream form varies substantially among geoclimatic settings, and, thus, the degree of negative effects induced by urbanization may be region-specific. Biota in streams of the Coastal Plain ecoregion of the eastern US consistently show greater tolerance to urban land use than do biota of the adjacent Piedmont, potentially because of a disparity in geomorphic degradation between ecoregions. We quantified channel morphometry, particle mobility, sediment deposition, and floodwater chemistry in similarly sized rural and urban streams of both ecoregions to detect differences in urbanization-induced geomorphic change. Macroinvertebrate rates of recolonization in patches of disturbed benthic habitat also were monitored. No differences in channel morphometry were observed among treatment groups. Riffle particle sizes were significantly larger in urban than in rural Piedmont streams, but a corresponding disparity was absent in Coastal Plain streams. Particle mobility increased in urban settings uniformly between ecoregions. However, transported particles were substantially larger in Piedmont streams. Sediment deposition was higher overall in Coastal Plain streams but more affected by urbanization in Piedmont streams. Macroinvertebrate density in the disturbed habitat rose faster over time in Coastal Plain than in Piedmont streams. Results suggest that geomorphic degradation is greater in Piedmont streams and that organisms may be adapted to benthic instability in Coastal Plain streams. In addition, our findings demonstrate that ecosystem-scale responses of streams to urbanization may vary inherently among geoclimatic settings.

Invasive species may threaten the fundamental role played by native macroinvertebrate shredders in determining energy flow and the trophic dynamics of freshwater ecosystems. Functionally, amphipods have long been regarded as mainly shredders, but they are increasingly recognized as major predators of other macroinvertebrate taxa. Furthermore, intraguild predation (IGP) between native and invasive amphipods underlies many species displacements. We used laboratory mesocosms to investigate what might happen to shredders and leaf-litter processing in water bodies invaded by the highly predatory Ponto-Caspian amphipod Dikerogammarus villosus, which is spreading rapidly throughout Europe and may soon invade the North American Great Lakes. The leaf-shredding efficiency of D. villosus was significantly lower than that of 3 Gammarus species (2 native and 1 invasive) that D. villosus has either already displaced or may be currently displacing in The Netherlands. In addition, D. villosus was a major predator of all of these native and invasive amphipod shredders and of a common isopod shredder Asellus aquaticus. Leaf processing in Gammarus and Asellus mesocosms declined rapidly in the presence of D. villosus and ceased altogether within 4 d because by then, all potential shredders had been killed and consumed. Furthermore, the shredding efficiency of surviving amphipods and isopods declined significantly within 2 d of the release of D. villosus, a result indicating that predator-avoidance behavior may override leaf processing. We discuss the implications of these direct and indirect effects of D. villosus invasions and species displacements on community structure and litter processing in aquatic ecosystems.

We addressed how timber harvest can interact with salmon (Oncorhynchus spp.) spawning activities to influence benthic macroinvertebrate communities in streams on Prince of Wales Island, Alaska. We predicted that spawning salmon would cause greater disturbance to macroinvertebrates in streams in watersheds with high than with low timber-harvest intensity because finer sediments that accumulate in streams after timber harvesting would be readily dislodged by spawning salmon and lower volumes of large wood in the streams would decrease availability of macroinvertebrate refugia from salmon activity. We used a modified Hess sampler to collect benthic macroinvertebrates from 6 riffles in each of 7 streams before and during the annual salmon run. Diptera biomass was lower and Plecoptera biomass was higher during the salmon run than before it. During the run, insect biomass, scraper biomass, and Ephemeroptera biomass were higher in streams with low than with high timber-harvest intensity, possibly because the finer sediments in the high timber-harvest intensity streams were more readily dislodged by salmon. Macroinvertebrate community structure (density and biomass) differed between before and during the run. Epeorus longimanus, Baetis, Seratella tibialis, Suwallia, Chironomidae, and Simuliidae were significant indicators of before-salmon benthic communities, whereas Sweltsa and Zapada cinctipes typified during-salmon communities. Some taxa probably are more tolerant of salmon disturbance than others, and their life histories might be adapted to the autumn salmon run. Overall, our results indicate that strong interactive effects can occur between anthropogenic activities and natural disturbance and that timber-harvest activity can intensify the effects of spawning-salmon disturbance on macroinvertebrates.

Freshwater unionid mussels are among the most endangered groups of organisms in the world. They develop indirectly via a host, usually a fish, and this dependence appears to limit the reproduction and distribution of freshwater mussels. Epioblasma torulosa rangiana and Lampsilis fasciola are 2 endangered species in Canada. Epioblasma t. rangiana has a low abundance and limited distribution, whereas L. fasciola has a higher abundance and less-constrained distribution. Three known host species were examined for each mussel species. Results were that: 1) E. t. rangiana glochidia had significantly higher metamorphosis rates (i.e., proportion of attached glochidia that successfully metamorphosed to the juvenile mussel stage) on Etheostoma exile (mean ± SE: 44 ± 9%) and Cottus bairdi (42 ± 6%) than on Etheostoma nigrum (10 ± 3%) and 2) L. fasciola glochidia had significantly higher metamorphosis rates on Micropterus dolomieu (82 ± 2%) and Micropterus salmoides (63 ± 8%) than on C. bairdi (37 ± 7%). Variation in the co-occurrence of mussels and their primary vs marginal host species (i.e., high vs low infestation and metamorphosis rates, respectively) appears to explain the distributions and abundances of these 2 endangered mussels in Canada. An understanding of the quality of different host fishes of endangered mussel species is needed to facilitate effective conservation strategies.

Stream metabolism at both ecosystem and functional-compartment scales was measured in a low-order Pampean stream (La Choza) over a 3-wk period to characterize metabolic rates and discern the contribution of each functional compartment (submerged macrophytes, benthos, floating macroalgae, water column, and hyporheic zone) to ecosystem metabolism. La Choza stream is an autotrophic ecosystem during low flows and has gross primary production rates of up to 22 g O₂ m⁻²d⁻¹, which are among the highest reported in the literature and set an upper bound on how productive streams can be in the absence of light and nutrient limitations. Floating macroalgae provided most of the primary production (30–90%), whereas the hyporheic zone provided most of the ecosystem respiration (40–80%). The differential effects of high flows on the different functional compartments depressed the production∶respiration ratio, suggesting a strong relationship between flow and metabolism. Thus, low flows enhanced primary production and led to diel dissolved O₂ concentration oscillations between 0 and 25 g O₂/m³. In contrast, high flow depressed primary production by an order of magnitude and increased ecosystem respiration. High production rates during the low-flow period and extreme physicochemical conditions (anoxia for 7–8 h on a daily basis) may be typical in this type of ecosystem during extended low-flow periods.

Consumer nutrient cycling supplies limiting elements to autotrophic and heterotrophic organisms in aquatic systems. However, the role of consumers in supplying nutrients may change depending on their diet and their own stoichiometry. We evaluated the stoichiometry, N and P excretion, and diets of the dominant macroinvertebrates and fish at 6 stream sites to determine if the nutritional composition of food alters nutrient excretion. We used Sterner's (1990) nutrient homeostasis model as a reference to gauge whether consumer nutrient excretion is influenced by diet. Body stoichiometry explained 61% of the variation in N∶P excretion by macroinvertebrates but only 11% of the variation for fish. In both cases, the relationship was driven by 2 P-rich end-members, crayfish and mottled sculpin. Results of Akaike Information Criterion (AIC) analysis showed that family alone explained 71% of the variation in N∶P excretion in macroinvertebrates and 31% of the variation in fish. Diet explained only 8% of the variation in both cases. Most consumers (9 of 11) had N∶P excretion values that were well below predictions of Sterner's model. Two taxa, crayfish and sculpin, had N∶P excretion that overlapped the model's predictions. Our results suggest that crayfish and sculpin may display strict homeostasis with respect to N and P and that their growth might be P-limited. Other consumers may be more flexible in their stoichiometry and not P-limited. We speculate that the extremely low excretion N∶P measured for many consumers might have been the result of semiflexible homeostasis, inaccuracies in our assessment of dietary nutrients, growth-limiting nutrients other than N or P, or lack of egestion data. Our results suggest that crayfish and sculpin may alter N and P dynamics in streams by excreting low amounts of P relative to N compared to what is generally available in the water column.

Water-body size is one of the most important factors affecting the structure and function of aquatic ecosystems. The categorical variable, Strahler stream order, is commonly used as a surrogate for stream size, perhaps because stream size is a multidimensional attribute that defies simple definition. Some stream-size attributes, including continuous variables, such as catchment area, distance to source, and model estimates of discharge also are available as geographic information system (GIS)-derived or modeled variables. These GIS measures are commonly used by stream ecologists along with field-derived attributes, such as discharge, stream cross-sectional area, width, and depth, which are more direct measures of stream size as experienced by aquatic organisms. Our objective was to quantify how well some commonly used stream-size attributes are predicted from Strahler order in the US as a whole and within major ecoregions and hydrologic landscape regions. We based our analysis on field-channel-survey and digital-stream-trace data (NHD-Plus) from 2162 US stream sites, ranging from 1^st to 8^th order (at 1∶100,000 scale). Strahler order provided a surprisingly useful approximation of the ranges of catchment size, distance to source, modeled mean annual discharge, and field-based low-flow and bankfull channel dimensions for most streams within a given Strahler order. However, even within geoclimatically and ecologically similar regions, site-specific predictions of stream size from Strahler order can have large errors. Correlations between Strahler order and the size measures considered here varied widely (r  =  0.48–0.91). Within individual Strahler orders, the alternative size measures varied by 5 and 4 orders of magnitude at national and regional scales, respectively. The same size-measure value could occur in 1 to 7 different stream orders at the national scale and in some regions, with generally good agreement in mountains and poor agreement in plains. Therefore, we conclude that Strahler order is useful for relating information about stream size, but that researchers should base analyses on multiple, continuous measures of stream size and should communicate stream-size results or associations based on the size-related measurements. Two characteristics of Strahler order make it useful for selecting sites across the range of stream sizes encountered in regional and national surveys, as long as limitations are explicitly recognized. First, the number of Strahler orders is limited. Second, Strahler order is easy to extract from stream networks constructed from digital elevation data and from national hydrographic data sets.

The application of deoxyribonucleic acid (DNA) barcoding in benthic ecology is proceeding rapidly, and a special session held at the 2009 North American Benthological Society Annual Meeting in Grand Rapids, Michigan, USA, brought together benthic scientists working in areas varying from taxonomy to biomonitoring to present the state of the science. The papers arising from this meeting are grouped in this special series of papers. Here, we present a brief introduction to this emerging area of benthic science, highlight the contributions of previous researchers, and provide an overview of these latest contributions to this highly active field. Paper topics include the development of DNA libraries for identification of material from field studies, application of DNA-based taxonomy in the discovery of new faunas, and the use of DNA-based identification in field biomonitoring studies, present and future. In addition, a critique is presented of the use of DNA barcoding in benthology and the potential pitfalls facing researchers who seek to use this method in their future research.

As part of its ongoing work in biomonitoring, Environment Canada's Canadian Aquatic Biomonitoring Network (CABIN) program has assembled an expert-verified reference collection of 3864 specimens of 604 species of Canadian freshwater macroinvertebrates. Such collections are a key resource for developing a deoxyribonucleic acid (DNA) barcode library to facilitate molecular identification of biomonitoring samples. We examined the problems encountered in using such legacy material to obtain reference barcodes. We focused on the influence of specimen age and preservation history. To supplement work on the reference collection, we determined the time-dependent effects of formalin preservation on DNA-barcode integrity in 4 common arthropod taxa by controlled exposure of fresh material obtained from laboratory cultures. Specimens in the reference collection were preserved with short-term fixation in formalin followed by prolonged preservation in 70% ethanol. Only 19 caddisfly larval specimens out of the total of 650 analyzed returned full-length sequences. In contrast, formalin preservation of freshly collected material for up to 20 d yielded good sequencing success and high-quality sequences. Freshly collected material clearly provides the best basis for the future development of DNA-barcode libraries, and formalin preservation should be avoided where possible to ensure that DNA integrity is maximized.

Deoxyribonucleic acid (DNA) barcoding is an effective tool for species identification and life-stage association in a wide range of animal taxa. We developed a strategy for rapid construction of a regional DNA-barcode reference library and used the caddisflies (Trichoptera) of the Great Smoky Mountains National Park (GSMNP) as a model. Nearly 1000 cytochrome c oxidase subunit I (COI) sequences, representing 209 caddisfly species previously recorded from GSMNP, were obtained from the global Trichoptera Barcode of Life campaign. Most of these sequences were collected from outside the GSMNP area. Another 645 COI sequences, representing 80 species, were obtained from specimens collected in a 3-d bioblitz (short-term, intense sampling program) in GSMNP. The joint collections provided barcode coverage for 212 species, 91% of the GSMNP fauna. Inclusion of samples from other localities greatly expedited construction of the regional DNA-barcode reference library. This strategy increased intraspecific divergence and decreased average distances to nearest neighboring species, but the DNA-barcode library was able to differentiate 93% of the GSMNP Trichoptera species examined. Global barcoding projects will aid construction of regional DNA-barcode libraries, but local surveys make crucial contributions to progress by contributing rare or endemic species and full-length barcodes generated from high-quality DNA. DNA taxonomy is not a goal of our present work, but the investigation of COI divergence patterns in caddisflies is providing new insights into broader biodiversity patterns in this group and has directed attention to various issues, ranging from the need to re-evaluate species taxonomy with integrated morphological and molecular evidence to the necessity of an appropriate interpretation of barcode analyses and its implications in understanding species diversity (in contrast to a simple claim for barcoding failure).

Monitoring water quality with aquatic insects as sentinels requires taxonomic knowledge of adult and immature life stages that is not available in many parts of the world. We used deoxyribonucleic acid (DNA) barcoding to expedite identification of larval caddisflies from 20 sites in the headwaters of the Tigris River in northern Iraq by comparing their mitochondrial cytochrome c oxidase subunit I (COI) sequences to a global reference library (the Trichoptera Barcode of Life). We obtained full-length DNA barcodes for 16 COI haplogroups from 11 genera in 9 Trichoptera families. The most haplogroups and genera were recorded from Sulaimani Province. Two distinct COI haplogroups were found for the genus Psychomyia, and 5 haplogroups were found for Hydropsyche. The Hydropsyche COI haplogroups do not form a monophyletic clade with reference to the world fauna, but 4 out of 5 haplogroups are related to other Palearctic species. Three larval Rhyacophila specimens in a single COI haplogroup are closely related to specimens of Rhyacophila nubila Zetterstedt and Rhyacophila dorsalis (Curtis) from Europe, but adults from Iraq are needed to confirm their species identity.

A taxonomic perspective of deoxyribonucleic acid (DNA) barcoding is presented here with respect to its use within the freshwater science community. The goals and methods of DNA barcoding and the criticisms leveled at the procedure by the taxonomic community are explained. The major goal of DNA-barcoding efforts is to aid identification of specimens by matching sequences to a sequence library. This goal is achievable, but barcoding efforts must be coupled with a much wider sampling regime and should be done with the full cooperation and collaboration of the taxonomic community. Through collaborative efforts, taxonomists and barcode advocates can build capacity for taxonomy and sequence the vast number of specimens needed to build a truly comprehensive barcode library for use by benthic scientists. Other barcoding goals, such as species diagnosis and biodiversity surveys, are more problematic, but solutions are possible.

We developed a multilocus deoxyribonucleic acid (DNA)-barcoding approach to identify newly transformed juvenile mussels collected from naturally infested fishes in a federally protected waterway that is home to a diverse mussel community, the St Croix River (Minnesota/Wisconsin, USA). We used new and publicly available data downloaded from GenBank to build reference databases for identified adult mussels. We assessed the efficacy of the mitochondrial loci cytochrome oxidase c subunit I (COI) and nicotinamide adenine dinucleotide dehydrogenase subunit 1 (ND1) for DNA barcoding. We concluded that the barcoding gap between average intra- and interspecific genetic distances is wider for ND1 than for COI, but both loci perform well for species identification in character-based phylogenetic analyses. Almost every species formed a monospecific clade with high bootstrap and posterior-probability support. We obtained newly transformed juvenile mussels by collecting individuals of 3 different fish species that were infested with unionid larvae. We held the fish in aquaria until the mussels emerged naturally. We then extracted DNA and sequenced our loci of interest. When sequences from the juveniles were included in phylogenetic analyses, they grouped with single species (or, in one case, a pair of closely related species) with high bootstrap and posterior-probability support. Identifying juveniles using morphology alone is difficult and, in some cases, impossible. Therefore, our approach will be useful to researchers interested in the relationship between unionid mussels and their fish hosts.

Four approaches to or levels of identifying macroinvertebrates (amateur/family, expert entomologist/genus, expert entomologist/genus–species, and DNA barcoding/species) were used to assess community structure and water quality in White Clay Creek, Pennsylvania, USA. Macroinvertebrates were collected in March 2008 from 2 riffle sites 3.9 km apart on the same stream. The downstream site was known to be degraded by land and water use. About 98% of the 1617 specimens used for analysis, including small, immature, and damaged specimens, were successfully barcoded (sequenced) for the mitochondrial cytochrome c oxidase subunit I gene. A criterion of 2 to 4% genetic divergence provided good separation of presumptive species. Barcodes increased the taxonomic inventory across the 2 sites by 475% (124 taxa) relative to the amateur level, and 125% (83 taxa) and 70% (62 taxa) relative to the expert genus and species levels, respectively. Barcoding revealed species not currently described in larval taxonomic keys, including multiple (2–11) coexisting congeneric species. That 150 species were revealed by barcoding samples collected on the same date and in the same habitat was unprecedented, as was the fact that 60 cm² of stream bottom supported an average of 248 to 347 individuals representing 55 to 68 species. Most barcode species were rare, with 42% represented by ≤2 individuals. Across all species, 43 of 89 barcode species were unique to upstream site 11 and 60 of 107 were unique to downstream site 12. In terms of water-quality assessment, most of the 17 metrics studied changed significantly (α  =  0.05) when taxonomy changed from family to genus–species (79% and 93% for sites 11 and 12, respectively), and many also changed when taxonomy changed from genus to species (59 and 65% for sites 11 and 12, respectively). The proportion of metrics able to detect a difference (α  =  0.05) between sites 11 and 12 increased with improved taxonomic resolution (36, 47, 65, and 76% for family, genus, genus–species, and barcode, respectively). The results revealed a pollution-tolerance gap because barcoding pushed larval taxonomy beyond the available pollution-tolerance data. Regardless, the combined morphological and molecular approach provides a finer resolution for evaluating environmental change associated with both natural and anthropogenic processes. The ability to distinguish larvae at the species level through barcoding finally puts biodiversity assessments for aquatic communities in terms comparable to those used for terrestrial ecosystems where estimates of biodiversity for plants and animals are never quantified at the level of genus or family. We conclude that DNA barcodes of stream macroinvertebrates will improve descriptions of community structure and water quality for both ecological and bioassessment purposes.

Taxonomic identification of benthic macroinvertebrates is critical to protocols used to assess the biological integrity of aquatic ecosystems. The time, expense, and inherent error rate of species-level morphological identifications has necessitated use of genus- or family-level identifications in most large, statewide bioassessment programs. Use of coarse-scale taxonomy can obscure signal about biological condition, particularly if the range of species tolerances is large within genera or families. We hypothesized that integration of deoxyribonucleic acid (DNA) barcodes (partial cytochrome c oxidase subunit I sequences) into bioassessment protocols would provide greater discriminatory ability than genus-level identifications and that this increased specificity could lead to more sensitive assessments of water quality and habitat. Analysis of DNA barcodes from larval specimens of Ephemeroptera, Plecoptera, and Trichoptera (EPT) taxa collected as part of Maryland's Biological Stream Survey (MBSS) revealed ∼2 to 3× as many DNA-barcode groups or molecular operational taxonomic units (mOTUs) as morphologically identified genera. As expected, geographic distributions for several mOTUs were tighter than for the parent genus, but few mOTUs showed closer associations with water-quality variables or physical-habitat features than did the genus in which they belonged. The need for improved protocols for the consistent generation of DNA barcodes is discussed.

Understanding the mechanisms that create spatial and temporal patterns of functional diversity in stream networks is a goal of basic research and has implications for effective conservation of freshwater ecosystems. These patterns are likely to be influenced by the combination of temporally variable environmental conditions, movement constraints imposed by network structure, and the trait composition of local communities. We developed a simplified metacommunity model to investigate complex interactions among these factors under lottery competition for local resources, such as establishment sites. We used this model to examine how local and regional community composition varied in 3 scenarios: a null implementation involving only spatial effects, an implementation that combined network constraints with dispersal-trait variation, and an implementation in which a trade-off between multiple functional traits was paired with varying levels of temporal autocorrelation in the intensity of mortality. These simulations clarified the conditions that allow a single functional strategy to exclude others in a dendritic network and demonstrated 2 distinct modes of regional partitioning that can support the persistence of multiple functional strategies within such networks. The results suggested that the emergence of watershed or headwater–outlet partitioning depends on the functional dispersal differences present in the metacommunity and that autocorrelated mortality levels can collapse these regional divisions when they depend on a trade-off between dispersal ability and mortality resistance. We discuss the need to confront the complexity of interacting controls on community composition in rivers and streams and suggest opportunities to move beyond the basic framework we present.

Spatial complexity in metacommunities can be separated into 3 main components: size (i.e., number of habitat patches), spatial arrangement of habitat patches (network topology), and diversity of habitat patch types. Much attention has been paid to lattice-type networks, such as patch-based metapopulations, but interest in understanding ecological networks of alternative geometries is building. Dendritic ecological networks (DENs) include some increasingly threatened ecological systems, such as caves and streams. The restrictive architecture of dendritic ecological networks might have overriding implications for species persistence. I used a modeling approach to investigate how number and spatial arrangement of habitat patches influence metapopulation extinction risk in 2 DENs of different size and topology. Metapopulation persistence was higher in larger networks, but this relationship was mediated by network topology and the dispersal pathways used to navigate the network. Larger networks, especially those with greater topological complexity, generally had lower extinction risk than smaller and less-complex networks, but dispersal bias and magnitude affected the shape of this relationship. Applying these general results to real systems will require empirical data on the movement behavior of organisms and will improve our understanding of the implications of network complexity on population and community patterns and processes.

Dispersal rates and the diversity of the regional species pool strongly affect community assembly in habitat patches. Incorporating these elements mechanistically into a model of community assembly requires adoption of a metacommunity paradigm. We developed a hierarchical model of community assembly for stream insects that incorporates regional effects (distance to and generic richness of other stream reaches) and local effects (water quality and community composition). We tested our model with a unique data set detailing changes in stream-insect community composition over 6 sampling periods across a 27-y period of watershed recovery from anthropogenic effects. α and γ richness increased greatly over the time period, whereas β richness declined strongly. Generic richness of individual stream reaches was significantly related to dispersal distance and generic richness of surrounding immigrant pools in preceding years. However, the strength of the relationship declined over time indicating that distance to potential colonists played a major role only early in community assembly. Water quality, characterized by an ordination of pH, temperature, conductivity, dissolved O₂, NO₃, NH₄, and orthophosphate, was correlated with generic richness at all time periods during the community-assembly sequence. The functional diversity (diversity of functional attributes present in an assemblage of species) of entire communities was lower than expected from random simulations in all sampling years. However, functional diversity of individual functional feeding groups varied through time and amongst themselves. Our results suggest that both deterministic and random processes are important in metacommunity assembly, and their relative strengths vary throughout the assembly process.

Given the unique spatial structure of stream habitat and the stochasticity characteristic of lotic ecosystems, metacommunity approaches hold much promise in the field of stream ecology. We take advantage of the tight parallels between neutral theories of molecular evolution and community assembly and present a novel conceptual approach to evaluating the role of stochasticity and dispersal limitation in structuring stream metacommunities. The analogous neutral frameworks generate similar expectations for the spatial structure of species comprising communities and genes comprising populations, particularly among patchily distributed, isolated local habitats (e.g., headwater streams) where among-stream dispersal is expected to be limited. Our emphasis is on the conceptual value of this approach, but we also used a limited data set composed of intensively sampled community and population-genetic data collected from 5 to 7 Rocky Mountain alpine streams for illustrative purposes. We characterized neutral population genetic structure as mitochondrial haplotype distributions for the blackfly Prosimulium neomacropyga, which shows strong genetic isolation by distance across the study region. We assessed community structure as the relative frequencies of ecologically similar species in the family Chironomidae collected annually over 2000–2003. Species richness was positively correlated with haplotype richness across streams (r  =  0.55), and analogous methods revealed significant β diversity at both the population-genetic (F_ST  =  0.23) and the community (F_ST  =  0.16) levels. However, a more spatially explicit comparison of pairwise community dissimilarity and genetic distance revealed no evidence of correlation (Mantel test, r  =  −0.18, p  =  0.54), a result suggesting no effect of spatial distance on community similarity at the relatively fine spatial scale of our study. We discuss 2 key arguments to explain the lack of community distance decay: 1) unaccounted environmental differences among streams, and 2) varying spatial and temporal scales of neutral processes (including genetic drift and ecological drift; i.e., stochastic temporal change) between population-genetic and community levels. Evidence for both selective and neutral explanations is compelling in this system. The explanations cannot be disentangled with the current data, but the study has strong heuristic value that we hope will stimulate future efforts in stream metacommunity analysis.

Measures of species diversity are valuable tools for assessing ecosystem health. However, most assessments have addressed individual sites or regional taxon pools, with few comparisons of differences in assemblage composition within or among regions. We examined the effects of anthropogenic disturbance on local richness (α diversity) and species turnover (β diversity) of benthic macroinvertebrates in small streams within and between 2 ecoregions (Northern Piedmont vs Southeastern Plains ecoregions) of the Patuxent River basin (Maryland, USA). Regional species pools did not differ between ecoregions (Piedmont  =  166 taxa, Plains  =  162 taxa); however, local richness was lower in the Plains (mean  =  17.4 taxa/stream) compared to the Piedmont (mean  =  22.2 taxa/stream). When streams were categorized into disturbance classes (low, medium, high), local richness did not differ among categories for either region. However, at the entire Patuxent scale, local richness tended to decrease with % impervious cover in a watershed. Variation in species composition, analyzed with nonmetric multidimensional scaling (nMDS), differed significantly between Piedmont and Plains streams, and Plains streams had higher β diversity than Piedmont streams. When partitioned by disturbance category and region, β diversity differed only between the low-disturbance sites (Plains > Piedmont). Relationships between β diversity and environmental variables varied by region. β diversity was weakly negatively related to % row-crop cover in a watershed at the entire Patuxent scale. For the Piedmont region, β diversity tended to decrease with % forest, % pasture, and % row-crop cover in a watershed. Such negative relationships between β diversity and landuse variables indicate a possible homogenization of the assemblage. The incongruence between diversity measures and composition measures, together with differing effects of anthropogenic land use on β diversity in the 2 regions, emphasizes the need to incorporate both α and β diversity and regional environmental factors in conservation/land management studies.

If organisms move beyond the boundaries of local sampling units, regional metacommunity dynamics could undermine the ability of bioassessment studies to characterize local environmental quality. We tested the prediction that fish dispersal influences local fish community structure and bioassessment metrics as a function of site position within stream networks. We evaluated fish community data from the US Environmental Protection Agency's Regional Environmental Monitoring and Assessment Program in West Virginia, USA, to compare the influences of stream network position, ecoregion, basin, and stream size on local fish community composition. We assigned sites to 1 of 3 stream network positions: 1) main channels (MC, n  =  12) encompassed streams with upstream catchment areas >200 km², 2) mainstem tributaries (MT, n  =  43) flowed into MC-sized confluences within 15 fluvial km, 3) headwater tributaries (HT, n  =  31) lacked such riverine confluences within 15 fluvial km. MT and HT sites had similar upstream catchment sizes and landuse gradients, but species richness was greater in MT sites than HT sites, whereas MT and MC sites were not different in this regard. Three bioassessment metrics were greater in MT sites than HT sites (intolerant species richness, cyprinid species richness, benthic species richness), but a multimetric index of biotic integrity did not differ among stream network positions. Ordinations revealed that fish community composition was organized primarily by zoogeographic basin (Monongahela River basin, New River basin, Ohio River basin), ecoregion (Central Appalachian Plateau, Western Appalachian Plateau, Ridge and Valley), and stream size. Riverine specialists were more abundant in MT than HT sites and were more abundant in basins connected to the Ohio River than in basins isolated from the Ohio River by a large waterfall (New River). Our results suggest that contemporary dispersal among streams influences fish community composition over small spatial scales (10¹ km), historical dispersal constrained by zoogeographic barriers influences community structure over larger areas (10² km), and contemporary dispersal by fishes influences certain metrics commonly used in bioassessment programs.

Explaining the mechanisms underlying patterns of species diversity and composition in riverine networks is challenging. Historically, community ecologists have conceived of communities as largely isolated entities and have focused on local environmental factors and interspecific interactions as the major forces determining species composition. However, stream ecologists have long embraced a multiscale approach to studying riverine ecosystems and have studied both local factors and larger-scale regional factors, such as dispersal and disturbance. River networks exhibit a dendritic spatial structure that can constrain aquatic organisms when their dispersal is influenced by or confined to the river network. We contend that the principles of metacommunity theory would help stream ecologists to understand how the complex spatial structure of river networks mediates the relative influences of local and regional control on species composition. From a basic ecological perspective, the concept is attractive because new evidence suggests that the importance of regional processes (dispersal) depends on spatial structure of habitat and on connection to the regional species pool. The role of local factors relative to regional factors will vary with spatial position in a river network. From an applied perspective, the long-standing view in ecology that local community composition is an indicator of habitat quality may not be uniformly applicable across a river network, but the strength of such bioassessment approaches probably will depend on spatial position in the network. The principles of metacommunity theory are broadly applicable across taxa and systems but seem of particular consequence to stream ecology given the unique spatial structure of riverine systems. By explicitly embracing processes at multiple spatial scales, metacommunity theory provides a foundation on which to build a richer understanding of stream communities.