1. Emergence and inland dispersal of adult stoneflies (Plecoptera) and caddisflies (Trichoptera) from Broadstone Stream, an acidic and iron‐rich stream in southern England, were studied over 10 months in 1996–1997. Fifteen pyramidal emergence traps were placed randomly in a 200‐m stretch. Three Malaise traps were placed above the stream and six more on each side (one wooded, one open) along a transect at distances of 1, 15, 30, 45, 60 and 75 m from the channel. 2. More than 16 000 stoneflies, belonging to 11 species, and just under 400 caddisflies (22 species) were caught. Four dominant stoneflies (Leuctra fusca, Leuctra nigra, Leuctra hippopus and Nemurella pictetii) accounted for 96% and 95% of the catches in the emergence and Malaise traps, respectively. Two caddisflies (Plectrocnemia conspersa and Potamophylax cingulatus) accounted for 63% of the catch in the Malaise traps. Few caddisflies were taken in emergence traps. 3. The emergence periods of L. fusca, L. nigra and L. hippopus were well‐defined and unimodal, whereas that of N. pictetii was prolonged and erratic. Overall, more females (1285) emerged than males (740). 4. Female stoneflies and caddisflies were in the majority in the Malaise traps above the stream. On land, significantly more females than males of L. fusca, L. nigra and P. cingulatus were caught. The sex ratio of the remaining species did not deviate significantly from 1:1. 5. The three Malaise traps placed above the stream caught most of the stoneflies though there was also dispersal away from the channel, the numbers caught declining with distance. Exponential models explained between 67% and 99% of the variation in numbers of individuals with distance from the channel in the four common stoneflies. Half the individuals went less than 11–16 m from the stream, while 90% travelled less than 51 m. Significantly more L. nigra and N. pictetii were caught in the woodland than on the open side, whereas L. hippopus showed no overall preference for either side.

Seasonal food webs were constructed for the whole invertebrate assemblage (meio‐ and macrofauna) inhabiting Broadstone Stream (southeast England). High and uniform taxonomic resolution was applied in a dietary analysis, by resolving the complete benthic community to species, including the meiofauna, protozoa, and algae. Meiofauna accounted for 70% of all species in the summary web and for 73% and 63% of those in the summer/autumn and spring webs, respectively. The web structure changed between summer/autumn and winter/spring, due to differences in species composition. Many stream invertebrates fed on meiofauna and organic matter. Addition of meiofauna to the Broadstone web increased the percentage of intermediate species. Seasonal webs contained between 54 (spring 1997) and 86 (autumn 1996) interactive taxa and 229–378 trophic links. Marked differences in web complexity were found between the summer/autumn and winter/spring periods. Meiofauna accounted for most of the links in the web with a high proportion of intermediate–intermediate links in summer and autumn (0.421–0.440) and also of intermediate–basal links during winter and spring (0.509–0.628). In general, the summary web showed that intermediate species and basal resources were numerically dominant components in this stream. Web connectance rose slightly between summer (0.052) and winter (0.061) and increased further in spring (0.079), coinciding with a reduction in species number. A high fraction of detritivores was combined with omnivorous predators, many of which supplemented their diets with organic matter and, depending on season, with algae and invertebrate eggs. In addition, a wide range of feeding modes was found among meiofaunal species. The diversity of the Broadstone community suggests that the impact of top predators tends to dissipate. A low proportion of top predators in the web was combined with a low mean number of prey items, other than detritus, in their guts (large predators, 1.08–1.26 prey/individual gut; small‐sized tanypods, 2.15–2.32 prey/individual gut). Dietary similarity was highest in autumn and winter 1996, and observed feeding links of the most common predatory species showed low overlap in their diets. The web architecture of this stream is reticulate and complex, and the patterns observed in these seasonal webs differed from previous stream webs, resulting in low connectance, high linkage density, long food chains, and a high proportion of intermediate species and of intermediate–intermediate links. The food web derived from Broadstone Stream clearly demonstrates that the meiofauna increases web complexity and thus, taking into account their functional diversity, may be crucial to the understanding of food web properties and ecosystems processes in streams.

No abstract

1. Field experiments were carried out to determine whether animals move into areas of low shear stress during periods of peak flow and therefore avoid hydraulic disturbance. 2. Flow at the scale of 0.05 m2 patches was reduced experimentally by creating artificial ‘refugia’. Invertebrate colonization of cages with 1.1 mm mesh sides, which provided such potential refugia by reducing hydraulic forces within the cage, was compared with colonization of cages with coarser (15 mm) mesh which did not restrict the flow. 3. Colonization of these cages was tested over a series of weekly periods in two different streams. Nine trials were completed in a stream with abundant natural flow refugia (Broadstone Stream, SE England), and during three of these trials strong spates occurred. Six trials were completed in a stream with comparatively few natural flow refugia (Dargall Lane, SW Scotland), and peak flows were relatively less. 4. In Broadstone Stream, the relative colonization of refugium cages was greatest during periods of peak flow, suggesting animals had used these low‐flow areas as refugia during spates. Use of the artificial refugia did not occur in Dargall Lane, at least at the flows achieved during the trials. 5. Our results are consistent with the hypothesis that stream invertebrates accumulate in refugia during high flow disturbances. Whether their distribution among patches is altered by active or passive means remains unclear.

Summary1. The properties of food webs are important both in theoretical ecology and environmental management, yet remain elusive. Here, we examined 12 new stream food webs of higher taxonomic resolution and completeness than any previously published data set and combined them with other 10 published stream webs. 2. Compared with most previously published food webs, these stream communities (containing between 22 and 212 species) had more feeding links per species, a higher fraction of intermediate species (0·78-0·88) and lower fractions of top (0·04-0·09) and 'basal' species (0·07-0·15). 3. The exponent of the log link-log species relationship ( S 1·3 ) differed significantly (  : P < 0·001) from the link-species 'law' ( S 1 ) and the constant connectance hypothesis ( S 2 ). Based on these results, the link scaling law and the constant connectance hypothesis must be rejected for food webs in stream systems. 4. Connectance was markedly lower than previously reported and decreased with web size. We attribute this to the body-size disparity of organisms at the top and the bottom of the web, flow disturbance and to the high physical heterogeneity and complexity of the sedimentary habitat. All these may reduce the fraction of possible feeding links that are realized.