Streams that drain agricultural watersheds in the midwestern US deliver large N loads to downstream water bodies. Denitrification is a potential sink for N in streams, but its importance in agricultural streams is unclear. Denitrification was examined in Big Ditch, a NO₃-rich tributary to the Sangamon River in east-central Illinois. Denitrification associated with benthic sediments and floating mats of algae and macrophytes was measured from May to November 2002. Daily NO₃-N loads were calculated for the 2002 calendar year to provide a context for the denitrification rates. Four other streams were sampled less intensively and the results indicated that Big Ditch was typical of agricultural streams in the region. During the growing season, plant biomass in Big Ditch was occasionally >200 g dry mass (DM)/m², but biomass declined sharply following scouring spates in June and August. Maximum rates of plant-associated and sediment denitrification were similar on a DM basis (4.2 and 3.7 μg N₂O [g DM]⁻¹ h⁻¹, respectively). However, denitrification rates in the sediments were more than an order of magnitude greater than the rates associated with plant material on an areal basis. Large floating mats of algae and macrophytes often covered much of Big Ditch, but were not major sites for denitrification. Denitrification rates in the sediments generally were higher than those reported from other streams, with a maximum value of 15.8 mg N m⁻² h⁻¹, but daily NO₃-N loads in Big Ditch often were >5 Mg N/d (Mg = 10⁶ g). Denitrification rates were high at times, but instream denitrification appeared not to substantially affect N losses from this agricultural watershed.

Litterfall is often an important source of organic matter in streams, and biofilms play an essential role in the decomposition and cycling of organic matter in leaves. Biofilms form rapidly on the surfaces of decomposing leaves, but little is known about the taxonomic composition of the bacterial community in leaf-surface biofilms. In addition, it is not known if bacteria in the biofilm colonize leaves from stream water, or if the biofilm forms from growth of bacteria already on the leaf surface. The purposes of our study were to examine the contribution of major taxa of freshwater bacteria to maple-leaf biofilms and to determine if these taxa were present in the biofilm because of colonization from the stream water. Exposed and dialysis-bag-enclosed maple leaves were incubated in a stream in northeastern Ohio during autumn 1998 and 1999. Leaves were collected, the biofilm was removed, and bacteria were counted using 4′,6-diamidino-2-phenylindole (DAPI) staining or fluorescence in situ hybridization (FISH) with probes targeted to Domain Bacteria and to the α-, β-, and γ-Proteobacteria. DAPI-stained cells and Bacteria were more abundant in 1999 than 1998. Differences in abundance between years may have been related to water velocity, which was much lower in 1999 than 1998. The α-Proteobacteria were the most abundant group in the biofilm (up to 40% of Domain Bacteria), whereas the γ-Proteobacteria were the least abundant (generally <10%). Percentages of β-Proteobacteria were much greater in 1999 than 1998 (26% and 15%, respectively). Abundances of biofilm α- and β-Proteobacteria were much greater on exposed leaves than on dialysis-bag-enclosed leaves, suggesting that large numbers of cells may colonize leaf surfaces from stream water. In conclusion, Proteobacteria made up most bacteria on the surface of decomposing maple leaves. Erosion of cells from the biofilm and colonization from stream water may have been important factors controlling biofilm development.

Seasonal changes in the abundance of bacteria belonging to different phylogenetic groups in epilithic biofilms from a northeastern Ohio (USA) stream were examined using fluorescent in situ hybridization (FISH). Changes over a 13-mo period were observed in the biofilm assemblages, dominated by α- and β-proteobacteria. Numbers of β-proteobacteria and Cytophaga–Flavobacterium peaked during the winter months and coincided with increased NO₃ concentration. Actinobacteria (Gram-Positive bacteria with high guanine and cytosine [GC] content) had no relationship with any measured environmental variable and accounted for <3% of the overall bacterial assemblage. Abundance of 3 bacterial species examined, Acinetobacter calcoaceticus, Burkholderia cepacia, and Pseudomonas putida, was similar, except in the summer when numbers of B. cepacia were higher than the other 2 species. Detrended correspondence analysis extracted 2 factors that explained 69.2% of the total variation. β-proteobacteria and Cytophaga–Flavobacterium clustered with conductivity and concentrations of NO₃, dissolved organic C, and soluble reactive P in the 1st factor, while A. calcoaceticus, B. cepacia, and P. putida clustered with temperature and turbidity in the 2nd factor. Our study revealed large seasonal fluctuations in the abundance of the different bacterial taxa examined in biofilms, and also demonstrated the potential influences of various environmental variables on microbial community composition in aquatic systems.

Periphyton response to riparian canopy opening and salmon carcass addition in coastal streams of northern California was evaluated in a manipulative field experiment. The experiment followed a split-plot design, with streams as whole plots and two 100-m reaches in each of 6 streams as subplots. At the subplot level, riparian hardwoods were removed from one reach in each stream. At the whole-plot level, carcasses were added to both open- and closed-canopy reaches of 3 of the streams. Thus, treatments consisted of reaches with open or closed canopies, in the presence and absence of carcasses. Nutrient limitation of the periphyton was assessed in 2 streams (1 with carcasses and 1 without carcasses) using nutrient-diffusing clay saucers (N-enriched, P-enriched, N P-enriched, or unenriched control) incubated in open- and closed-canopy reaches in the streams. Canopy and carcass treatments did not affect gross primary productivity or periphyton biomass on natural substrates. The periphyton assemblage consisted primarily of diatoms in all reaches on all dates. N amendment of agar in nutrient-diffusing, clay saucers and canopy removal increased biofilm ash-free dry mass on the saucers, but carcass introduction did not. Failure of periphyton to respond to carcass addition may have reflected overriding light limitation, inadequate within-stream retention of carcass nutrients, and/or limitations of the study design.

Podostemum ceratophyllum Michx. has been associated with extremely high secondary production of benthic macroinvertebrates in open-canopy rapids. We conducted an experiment in the 7^th-order Little Tennessee River, North Carolina, to test whether varying amounts of Podostemum influenced macroinvertebrate abundance, biomass, community composition, and functional feeding group structure. The experiment consisted of 3 treatments in which P. ceratophyllum was completely, partially, or not removed from portions of 4 bedrock outcrops at 2 sites. Macroinvertebrates were sampled at 0, 3, and 6 wk post treatment. Complete removal of P. ceratophyllum greatly reduced overall macroinvertebrate abundance and biomass and altered assemblage structure, but had relatively little effect on functional structure. The lack of change in functional feeding group structure was probably a result of the importance of P. ceratophyllum as a substrate for epiphytic algae, and the availability of nearby colonists in undisturbed habitats. We found a strong positive relationship between surface area of Podostemum and total macroinvertebrate abundance and biomass. We estimated that P. ceratophyllum increased surface area by 3 to 4 times over bare bedrock. Podostemum ceratophyllum in the Little Tennessee River serves as an important habitat supporting high abundance and biomass of macroinvertebrates.

A food web based on the gut contents of consumers (invertebrates and fishes) in pools of Tai Po Kau Forest Stream (TPKFS), Hong Kong (southern China), indicated the importance of periphyton and fine organic particles; coarse particulate organic matter was a less important food source despite its higher relative abundance in this shaded hillstream. Stable isotope analysis of consumer tissues was undertaken to confirm this result. IsoSource software was used to model n-isotope and >n 1-sources, so that the relative contribution of the potential food sources could be determined. Results of an IsoSource mixed model of δ¹⁵N and δ¹³C stable isotope signatures of primary consumers generally supported evidence from gut content analyses about the importance of autochthonous resources. Inconsistencies between the results of gut content analysis and isotope signatures of consumer tissues occurred in a few cases but could be explained either by small sample size or the wide range of feasible solutions provided by the isotopic mixed model. Both techniques were needed to resolve the trophic position of omnivores. For instance, the gut contents of balitorid loaches indicated that they were primarily herbivorous but their stable isotope signatures revealed a significant dependence on animal food.

Considerable overlap in the diets of predatory fishes and invertebrates (odonates, perlid stoneflies, palaemonid shrimps) was confirmed by both gut contents and stable isotope analyses. This finding, along with a lack of intraguild predation, resulted in a short mean and maximum foodchain length, high links per species, and high connectance for the TPKFS food web when compared with literature reports of other stream food webs. Periods of spate-induced disturbance during the wet season and limited algal productivity in TPKFS might also have contributed to the short food chains. Inconsistent levels of resolution for different taxonomic groups within the food web may have generated artefacts of low linkage complexity, high predator–prey ratio, and a small number of basal and intermediate species, a pattern that has been confirmed for stream food webs elsewhere. Our study is the first example of a food web based on complementary analyses of gut contents and stable isotope signatures for any tropical stream. This combined approach is recommended for future studies of food webs, especially in habitats where omnivores are an important component of the community.

Net-spinning caddisflies are an important group of filter feeders in many streams, but the mechanisms by which their nets capture particles are poorly understood. Our objective was to determine the effects of particle size on the efficiency of particle capture by 2 species of net-spinning caddisflies, Ceratopsyche morosa and C. sparna. We measured particle capture by nets and ingestion by larvae. We used a mathematical model that incorporated drag and the presence of a near-bed boundary layer to predict reduction of flow passing through a caddisfly net. We evaluated model predictions by measuring the velocity of flow through a scaled model of a C. sparna net in a laboratory flume. Velocity of water through the net was reduced 45% at free-stream velocities of 71 cm/s and 36% at 99 cm/s. We released known quantities of 3 sized particles (Artemia nauplii, 528 μm; corn pollen, 85 μm; paper mulberry pollen, 12 μm) in several stream riffles to determine whether size-selective capture and ingestion occurred. Both C. morosa and C. sparna nets captured particles primarily by sieving; particles smaller than the mesh size (paper mulberry pollen) were caught at 28× lower efficiencies than sievable particles by C. morosa nets and 2.2× lower efficiencies by C. sparna nets. Larvae ingested 73 to 83% of the largest particles captured by nets, compared to 4 to 6% for the smallest particles. The overall probability of particle capture and ingestion of the largest, sievable particles was ∼400× greater than for the smallest, nonsievable particles, with the difference much greater for C. morosa than for C. sparna. We compared actual rates of particle capture to theoretical predictions. Ceratopsyche morosa captured sievable particles at 34% of the predicted rate, whereas nonsievable particles were captured at 0.4% of the predicted rate. Ceratopsyche sparna showed less difference between capture mechanisms, and captured sievable particles at 5% and nonsievable particles at 1.1% of predicted rates. Natural seston concentrations were skewed in the direction of smaller particles; the smallest size class, 0.5 to 30 μm, made up 65% of the total ash-free dry mass of the seston. Large particles (>65 μm) were captured at higher efficiencies by nets and ingested more readily by larvae; therefore, we predict that they will be the primary food items of 5^th-instar larvae of both species studied despite their rarity in the water column.

Most suspension-feeding trichopterans spin a fine-silk capture net that is used to remove suspended matter from the water. The efficiency of these nets has previously been studied by considering the geometry of the web structure but the material from which the nets is constructed has received little attention. We report measurements of the tensile strength and extensibility of net silk from Hydropsyche siltalai. These measurements place caddisfly silk as one of the weakest natural silks so far reported, with a mean tensile strength of 221 ± 22 megaNewtons (MN)/m². We also show that H. siltalai silk can more than double in length before catastrophic breakage, and that the silk is at least 2 orders of magnitude stronger than the maximum force estimated to act upon it in situ. Possible reasons for this disparity include constraints of evolutionary history and safety margins to prevent net failure or performance reduction.

Experimental removal of early instar black fly larvae from 3 oligotrophic lake-outlet streams in southern Quebec showed that larvae reduced phytoplankton biomass, but had little effect on bacterioplankton biomass. The rates of decline in bacterioplankton and phytoplankton biomass with distance downstream from 6 lake outlets were determined before the insecticide Bacillus thuringiensis var. israelensis was used to remove black fly larvae from 3 of the streams. The rates of decline in bacterioplankton and phytoplankton biomass were measured again after the larvae were removed, and the differences in the rates of decline between before and after were used as indicators of the rates of biomass ingestion by the black fly larvae. The proportion of bacterioplankton ingested/m stream length was <0.081%/m in all streams, but the proportion of phytoplankton ingested/m stream length was 0.35%/m and 0.24%/m in 2 of the 3 manipulated streams. Downstream declines in both bacterioplankton and phytoplankton persisted in the absence of black fly larvae. Factors other than black fly feeding (such as biofilm adhesion and ingestion by other filter-feeders) were responsible for 50 to 90% of the total downstream loss rate of bacteria, whereas factors other than black fly feeding were responsible for 0 to 55% of the total downstream loss rate of phytoplankton. Apparently, free-living bacteria were not ingested to a great extent by black fly larvae in these oligotrophic lake-outlet streams, and the link between bacteria in the plankton and their potential black fly predators seemed weak. The relative importance of algivory over bacterivory is expected to be greater in eutrophic lake-outlet streams, where the ratio of bacterial C to phytoplankton C is smaller than in oligotrophic lake-outlet streams, and in systems with a greater abundance of later-instar larvae, which are less efficient at capturing bacteria than early instar larvae.

I surveyed treeholes in central Pennsylvania for 7 mo in 1995 to investigate relationships among insect communities and water and leaf litter resources. I used water volume, essential to growth of treehole larvae, as an indicator of habitat size. Leaf litter is the basal food resource in treehole communities, and litter volume is related to the amount of energy available. Insect species richness and larval mosquito (Ochlerotatus triseriatus) and ceratopogonid midge (Culicoides guttipennis) densities were higher in treeholes that maintained high water volumes than those with low volumes throughout the study. Treeholes with high litter volume, irrespective of water volume, had the highest densities of C. guttipennis, and higher total larval densities than those with lower litter volumes; however, litter volume did not affect insect species richness. Scirtid beetle larvae were common in all treeholes, although their densities were unaffected by both litter and water volume. Insect species richness was not related to litter volume, unlike some other studies of phytotelmata. However, my study supported earlier conclusions that both habitat size and resource availability influence structure of treehole communities and maintain species populations. The influence of each resource on particular species presence and density may be related to physical aspects of treeholes, biotic interactions, and natural history of individual species.

Population and production dynamics of the chydorid Eurycercus vernalis were studied in the laboratory and in a small wetland during a 2-y period. Laboratory growth studies were conducted to measure the effects of temperature on E. vernalis growth and reproduction and to develop a multiple-regression equation that used temperature and mass-specific growth rates to estimate secondary production in the field population. Density, biomass, and production were estimated from benthic, water-column, and Nymphaea odorata leaf habitats within vegetated (Nymphaea) and nonvegetated (open-water) areas using monthly samples from the wetland. Eurycercus vernalis exhibited optimal growth, reproductive output, and net reproductive rate when reared at temperatures between 15 and 20°C. Reproductive age and egg development time decreased with increasing temperature, and these decreases led to increased innate capacity of population increase and decreased generation time with increasing temperature. In the field studies, annual density, biomass, and production were significantly higher in the Nymphaea zone than the open-water zone during both years. Eurycercus vernalis populations developed during the fall from oversummering resting eggs and reached maximum density, biomass, and daily production in mid to late spring when water temperatures reached 18 to 21°C. Mean annual density and biomass for the wetland pond were 2134/m² and 42.2 mg dry mass (DM)/m² in year 1 (1993), and 1122/m² and 19.0 mg DM/m² in year 2 (1994–1995). Annual production and the production/biomass ratio were 2138 mg DM m⁻² y⁻¹ and 50.7/y in year 1, and 1111 mg DM m⁻² y⁻¹ and 58.5/y in year 2. High production values suggested that E. vernalis is an important component of the microcrustacean community in the Nymphaea zone, especially during winter when production of most microcrustacean species is low.

Long-distance dispersal by aquatic insects can be difficult to detect because direct measurement methods are expensive and inefficient. When dispersal results in gene flow, signs of that dispersal can be detected in the pattern of genetic variation within and between populations. Four hundred seventy-five base pairs of the mitochondrial gene, cytochrome b, were examined to investigate the pattern of genetic variation in populations of the stonefly Pteronarcys californica and to determine if long-distance dispersal could have contributed to this pattern. Sequences were obtained from 235 individuals from 31 different populations in the western United States. Sequences also were obtained for Pteronarcella badia, Pteronarcys dorsata, Pteronarcys princeps, Pteronarcys proteus, and Pteronarcys biloba. Phylogenies were constructed using all of the samples. Nested clade analysis on the P. californica sequence data was used to infer the processes that have generated the observed patterns of genetic variation. An eastern North American origin and 2 distinct genetic lineages of P. californica could be inferred from the analysis. Most of the current population structure in both lineages was explained by a pattern of restricted gene flow with isolation by distance (presumably the result of dispersal via connected streams and rivers), but our analyses also suggested that long-distance, overland dispersal has contributed to the observed pattern of genetic variation.

New York State's fauna is exceptionally rich in odonates (dragonflies and damselflies), whose lengthy aquatic larval phases render them susceptible to effects of lake acidification, including the loss of fish. We used a collection of benthic macroinvertebrate samples taken by the Adirondack Lakes Survey Corporation to compare odonate communities in 460 lakes. Half were from the Adirondack Mountains, where acid neutralizing capacity (ANC) is low (mean ANC = 108.0 μg/L) and Al concentrations are high (mean Al = 111.61 μg/L), and half were from the Lower Hudson Valley, where ANC is significantly higher (mean ANC = 554.6 μg/L) and Al is significantly lower (mean Al = 0.049 μg/L). Many more lakes in the Adirondack lakes were fishless (52) compared to the lower Hudson (3), and the pH in Adirondack fishless lakes was an order of magnitude lower than the pH of Adirondack lakes with fish. Ninety-nine odonate taxa were identified (86 to species). In Adirondack samples, co-occurrence patterns were correlated with presence or absence of insectivorous fish and with acidic waters. Similar patterns were not apparent in Lower Hudson Valley samples. In Adirondack samples, richness of common taxa (found in ≥20 lakes) was higher in lakes with fish than in lakes without fish, regardless of pH. Loss of fish may enhance the top predator role of large larval dragonflies, causing change in odonate community structure, an interpretation consistent with previous research. Acidification of Adirondack lakes appears to promote a nonrandom subset of possible odonate communities, with negative implications for regional diversity.

Spatial and temporal patterns of transported organic matter (seston) and macroinvertebrates (drift) and benthic macroinvertebrate densities were examined before, during, and shortly after each of a series of scheduled, experimental floods in a flow-regulated river in the Swiss National Park. Temporal patterns in the lateral transfer of seston, drift, and benthic macroinvertebrates were evaluated in the flooded riparian area during 3 to 4 separate floods of different magnitude. No clear spatial pattern was found in the lateral transfer of seston, drift, or benthic macroinvertebrates, but the concentrations of seston and the densities of macroinvertebrates in the drift usually were lower in samples collected farthest from the main channel. Seston and drift increased significantly (from <1 g to 4–20 g ash-free dry mass/m³ and <10 to 250–1300 ind./m³, respectively) in the initial stages of each flood, but decreased to baseflow levels after ∼2 to 3 h. Macroinvertebrates responded passively to the floods, and their densities followed the hysteresis pattern of sediment and organic particles entrained during the course of each flood. The total number of macroinvertebrates drifting during each flood ranged from 33 × 10⁶ to >300 × 10⁶ individuals. The average density of macroinvertebrates stranded in the riparian area after each flood ranged from ∼6000 to 22,000 ind./m². Benthic macroinvertebrates were collected from pool, run, bedrock, and riffle habitats in the main channel the day before and the morning after 5 floods to test whether specific habitats provided flow refugia for macroinvertebrates. Floods reduced macroinvertebrate densities by 14% to 92%, averaged across habitat types, and the % reduction was related to flood magnitude. Fewer organisms were lost from bedrock habitats (43%) than from the other habitat types, and the most macroinvertebrates typically were lost from pools (>90%). Macroinvertebrate responses (e.g., recovery patterns) changed significantly between early floods and sequentially later floods, reflecting temporal changes in assemblage composition and abundance.

We investigated the importance of shrimps (Atyidae and Palaemonidae) and ephemeropterans (Baetidae) in the removal of periphyton and sediments in a Neotropical stream. The experimental site was open with homogeneous bedrock, shallow depth (8–15 cm), and intermediate water current velocity (0.2–0.3 m/s). We used 2 intensities of electrical current to exclude both shrimps and ephemeropterans (high-intensity treatment) or only shrimps (low-intensity treatment) from fixed areas (180 × 30 cm) of bedrock. When both ephemeropterans and shrimps were excluded in 2 experiments, matter accumulated on the bedrock to 5 and 20× the level in controls; when only shrimps were excluded, no accumulation was observed. Chlorophyll a increased significantly in the high-intensity exclusion, but most of the accumulation was fine organic and inorganic matter. In experiment 1, benthic matter and chlorophyll a decreased by an order of magnitude in areas in which ephemeropterans increased ∼40×. The increase in ephemeropterans was associated with a reduction in shrimp activity. In experiment 2, no increase in ephemeropterans was observed. Palaemonid shrimps (Macrobrachium olfersi) were more common than atyid shrimps (Potimirim glabra) in the study area. Therefore, we assumed that Macrobrachium, rather than Potimirim, interacted negatively with ephemeropterans and that Macrobrachium did not remove periphyton. We concluded that baetid ephemeropterans (particularly Americabaetis sp.) were the most important grazers and removers of benthic matter in this system.