Herbivory in an acid stream

Freshwater Biology (2000) 43, 545±556 Herbivory in an acid stream MARK E. LEDGER AND ALAN G. HILDREW School of Biological Sciences, Queen Mary and Westfield College, Mile End Road, London, E1 4NS, U.K. SUMMARY 1. Spatial and temporal variation in the distribution and feeding of non-predatory macroinvertebrates was investigated in a first-order, acid stream in the Ashdown Forest, southern England. 2. Stonefly (Nemouridae) and chironomid (Orthocladiinae) larvae were abundant on the upper surfaces of mineral substrata of three sizes (small stones, large stones, bedrock). The density of larvae in each taxonomic group did not vary among substrata of different sizes, although strong seasonal variation existed. 3. Nemourids and chironomids (H. marcidus) collected from the upper surfaces of substrata exhibited generalist feeding habits, consuming algae (diatoms, coccoid and filamentous green algae), detritus (biofilm matrix material and fine particulate organic matter (FPOM)) and inorganic debris. 4. There was spatial variation in the gut contents of nemourids. The proportion of algae in the guts of larvae often increased with the size of the substratum from which they were collected. Strong temporal variation in the composition of the diet also existed. Nemourids ingested a large quantity of attached algae and biofilm matrix from the biofilm in spring and winter, but consumed loose FPOM and associated microflora in summer and autumn. 5. We conclude that, in this acid stream, the trophic linkage between algae and grazers is maintained by `detritivorous' stonefly and chironomid species. The relationship between the feeding habits of these larvae and other life-history attributes, such as mouthpart morphology and mobility, is discussed. Keywords: acid streams, chironomids, herbivory, macroinvertebrates, stoneflies, trophic generalists Introduction Acid streams often lack the characteristic herbivorous invertebrates that graze benthic algae in circumneutral waters (Sutcliffe & Carrick, 1973; Townsend, Hildrew & Francis, 1983; Mulholland et al., 1986; Sutcliffe & Hildrew, 1989; Rosemond et al., 1992). A specialized grazer guild may be absent, either as a result of chemical toxicity (hydrogen ions and/or metals), or as a result of limitation by food quality or quantity. Studies by Sutcliffe (1978), Fromm (1980), Hall et al. (1980) and others support the first of these hypotheses. With regard to the role of food supply, Winterbourn, Hildrew & Orton (1992) reported that Correspondence and present address: Dr M. E. Ledger, Centre for Environmental Research and Training, 13 Pritchatts Road, University of Birmingham, Edgbaston, Birmingham, B15 2TT, U.K. E-mail: [email protected] ã 2000 Blackwell Science Ltd. studies of the distribution of algal biomass in streams suggest that acid streams do not consistently maintain lower standing crops of algae than their circumneutral equivalents. Further, Ledger & Hildrew (1998) showed that variations in light primarily explained spatial and temporal variation in the distribution of algae in one acid stream in southern England. Thus, the present consensus is that grazers are directly acidsensitive and that a suite of physicochemical and biotic factors affect algal production, overriding any general pattern of algal response to acidity. Nevertheless, acid stream food webs are often described as being detritus-based (Hildrew, 1992), since plecopterans, which frequently dominate the non-predatory macroinvertebrate fauna, have been shown to aggregate in leaf packs (Dobson & Hildrew, 1992) and shred leaves (Groom & Hildrew, 1989). Reduced decomposition of leaves by microbes could contribute to the year-round persistence and avail- 545 546 M. E. Ledger and A. G. Hildrew ability of leafy detritus to these detritivorous species, particularly in retentive channels (Townsend & Hildrew, 1988; Dobson & Hildrew, 1992; Griffith & Perry, 1993). This study was conducted in Lone Oak, a small, stony, acidic stream supporting abundant nemourid stoneflies that shred leaves (Dobson & Hildrew, 1992), and detritivorous orthoclad chironomids that ingest FPOM and its associated microflora (Berg, 1995). The geomorphology of the channel, in combination with variable discharge, results in only moderate retention of leafy detritus and a predominantly stony substratum. Dobson & Hildrew (1992) showed that the density of macroinvertebrates, including stoneflies and chironomids, is usually much higher in leaf packs than in the surrounding substratum, and reasoned that detritivores may be food limited in this stream. A number of other studies in low-order streams have shown that detritivores are limited by the quantity and quality of CPOM available (Cummins et al., 1980; Gee, 1988; McArthur et al., 1988; Richardson, 1991). Recent work has confirmed the contention of Cummins (1973) that many species of stream invertebrate are trophic generalists, able to feed and grow on a wide range of food types, and also to switch among resources in space and time (Mihuc & Minshall, 1995; Mihuc, 1997). We hypothesised that the non-predatory invertebrates of acid waters would feed on a broad range of resources, normally partitioned among a greater number of taxa in more species-rich circumneutral waters. Earlier work on the diets of nemourid stoneflies (Hynes, 1976; Henderson, Hildrew & Townsend, 1990), plus our own observations that nemourids and orthoclad chironomids are commonly found on the upper surfaces of biofilm-coated stones in streams of the Ashdown Forest, southern England, suggests that they may be supplementing their detritivorous diet with algae obtained by grazing biofilms. In a previous paper (Ledger & Hildrew, 1998), we showed that the quantity and quality of the biofilm coating stones in Lone Oak stream varied temporally, but also often increased with substratum particle size. Our main objective here was to show whether generalist feeding maintains the algae-grazer link in an acid stream. Specifically, we address the distribution and feeding of the `detritivorous' invertebrates found on these biofilm-coated surfaces, and test whether their density depends on variation in food availability. Methods Site Description Lone Oak is a first order, acidic (mean annual pH about 5.0) stream situated in the Ashdown Forest in southern England (NGR TQ475331). The underlying geology of the catchment is Ashdown Sandstone and soils are typically podsolic. The stream primarily drains acid heathland, although deciduous woodland dominated by oak (Quercus robur L.) and holly (Ilex aquifolium L.) forms the riparian zone. The stream water is clear, oligotrophic (mean annual conductivity about 80 ms cm±1) and cool. Over the course of our study, water temperature varied between 4 and 15 °C, with minima in January and maxima in August (Ledger & Hildrew, 1998). The stream is relatively shallow at summer baseflow (10±30 cm deep) and 2± 4 m wide, with a bed consisting of sandstone pebbles, cobbles, small boulders and patches of bedrock. The macroinvertebrate community is depauperate, and dominated by non-predatory stoneflies (Nemouridae and Leuctridae), orthoclad chironomids and predatory polycentropodid caddisfly larvae. Fish are absent (Ledger, 1997). Sampling design The study was conducted in a 100-m section of Lone Oak stream, dominated by stony riffles, between February 1994 and March 1995. To determine the overall abundance of potential grazers in the stream, 10 Surber sample units (0.0625 m2, mesh size 330 mm) were taken at random each month from riffles and preserved in the field in 70% alcohol. Samples were subsequently sorted and invertebrates identified to species where possible. In order to compare the numbers of potential grazers resident on biofilm patches of variable quality/quantity, and to provide specimens for gut contents analysis, we also sampled invertebrates exposed on the upper surfaces of three size classes of mineral substratum: small stones (upper surface area 100 cm2), large stones (250 cm2) and bedrock. Each month, seven small stones and seven large stones were carefully removed at random from the stream bed. All invertebrates were picked from the upper surfaces with forceps and immediately frozen on dry-ice. The upper surface area of each stone ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Herbivory in an acid stream sampled was traced using an acetate sheet, and subsequently estimated (cm2) from the mass (g) of the tracings. Each week, invertebrates exposed on a fixed area of bedrock were counted using a glass viewer anchored to metal poles fitted into holes drilled in the substratum. The base of the viewer comprised a 14 ´ 8 grid of 4 cm2 quadrats. Larvae on three contiguous patches were counted, resulting in the same area of 0.7 m2 being examined on each occasion. Invertebrates on the upper surfaces of all substrata consisted of nemourid stoneflies (Nemurella pictetii KlapaÂlek, Nemoura cinerea Retzius, Nemoura cambrica Stephens) and larvae of the orthoclad chironomid Heterotrissocladius marcidus (Walker). It was not possible to distinguish between the three species of nemourid on bedrock in the field using our viewer, so quantitative counts of stoneflies were assessed as total Nemouridae. Stoneflies collected for gut analysis were identified to species. Gut contents analysis The fluorochromatic stain 496 diamidino-2-phenylindole (DAPI) and epifluorescence microscopy (Walker et al., 1988), in combination with light microscopy, were used to identify particles ingested by the larvae. Where possible, 15±20 individuals belonging to each taxonomic group (i.e. each of the three nemourid species and H. marcidus) were taken from small stones, large stones and bedrock each month and their gut contents examined. Some species were not found on all substrata each month, however. The gut of each larva was removed under a dissecting microscope and food particles from the foregut were teased from the peritrophic membrane into a few drops of distilled water, transferred to an Eppendorf tube and stained with DAPI. After 20 min, the suspended particles were filtered through an Isopore black polycarbonate membrane filter (25 mm diameter, 0.2 mm pore size). The damp filter was removed and cleared on a slide smeared with a film of immersion oil. A drop of oil was then placed on the filter and a coverslip added. Slides were stored horizontally in the dark at 5 °C. All particles in fifteen fields of view were counted and assigned to one of several food categories (see below). The area of particles in each category was measured using an eyepiece graticule and expressed as a percentage of the total particle area. ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 547 When illuminated with UV light, DAPI-stained gut particles can be identified by fluorescence, colour and shape (Walker, Olds & Merritt, 1988). Bacterial and algal nuclei fluoresce bright blue and other particles autofluoresce diagnostically: algae appear dull orange and plant detritus dull yellow. By further illuminating the prepared material with white light, we were able to distinguish between two types of detritus ingested by stonefly larvae: (1) biofilm matrix material and (2) fine particulate organic matter (FPOM). This was possible because the structure of the biofilm coating mineral substrata in this oligotrophic and naturally acid stream was sufficiently different from that of FPOM to allow distinction to be made reliably. We sampled biofilm on the upper surfaces of stones from riffles in six streams in the Ashdown Forest, all of which received inorganic acids from the catchment, and found that at all sites small coccoid green algae and adnate diatoms (mainly Eunotia sp.) were clearly embedded in a very thin and discrete brown film of matrix material that had a porous and granular structure. Relatively large fragments of this detritus were therefore readily identified. This fine structure was not observed in any of the circumneutral streams in the Ashdown Forest, and we believe a distinction between the biofilm matrix and FPOM in those systems would be impossible. Loose particles of FPOM, collected from between stones, had a very different appearance under the microscope: they were darker, fluoresced less brightly under UV light, and lacked the obvious porosity of the biofilm matrix material (Ledger, 1997). Some small detrital particles could not always be safely classified into these two groups and were assigned to an `unidentified particles' category, along with nondetrital particles of unknown origin. In the case of the chironomids, detritus in the gut could not reliably be identified as `biofilm matrix' or `FPOM' due to the small size of the particles, and these categories were therefore merged to form a single `detritus' category. We were thus able to assign particles from the guts of stoneflies to one of seven categories: (i) diatoms (ii) filamentous green algae (iii) coccoid green algae (iv) biofilm matrix (v) fine particulate organic matter (FPOM) (vi) inorganic debris and (vii) unidentified particles. Six particle types were identifiable in chironomids guts, since categories (iv) and (v) were merged. 548 M. E. Ledger and A. G. Hildrew Results The invertebrate fauna Fig. 1 Temporal changes in the mean (‹ 1 SE, n = 10) benthic density of the chironomid Heterotrissocladius marcidus, and nemourid stoneflies in Surber samples. Data analysis Data were log-transformed prior to analysis, except for percentage data which were arcsin transformed. A two-way ANOVA was performed to determine the overall effect of month and substratum particle size on invertebrate density, and to determine any interaction between these two factors. To test the effect of substratum particle size on the proportion of algae in gut contents, a one-way ANOVA was performed on the data each month. A pairwise comparison of means (Tukey test) was then used to determine which substratum particle sizes were significantly different from each other. Nemourid stoneflies and the orthoclad chironomid H. marcidus made up more than 35% by numbers of the macroinvertebrate fauna in the Surber sample units each month. No Emphemeroptera or Gastropoda, taxa usually associated with algal grazing, were found in the benthos. Three species of Nemouridae (N. pictetii, N. cinerea and N. cambrica) were found in the benthos. N. pictetii was most abundant throughout the year, particularly in summer and autumn following peak recruitment (Fig. 1). Conversely, N. cinerea and N. cambrica were more abundant in the benthos in spring and early summer than in autumn. H. marcidus was most numerous in mid-summer (Fig. 1). Relative proportions of the three nemourid stoneflies on the upper surfaces of mineral substrata were similar to those in the benthos and, thus, were dominated mainly by N. pictetii (Fig. 2). The density of nemourids in the benthos was often higher than that on upper surfaces overall (e.g. in August 1994, mean densities on small stones, large stones, bedrock and in the benthos were 297, 324, 56 and 629 m±2, respectively) although the temporal pattern, particularly on stones, was similar, with highest densities through late summer and autumn (Fig. 2). Seasonal differences in density on the upper surfaces were highly significant, but there were no significant differences among the substrata (Table 1). As in the benthic samples, H. marcidus larvae were most abundant on the upper surfaces of substrata in summer (Fig. 3). In spring and summer, larvae were much more abundant on upper surfaces than in the benthos overall and their numbers rose earlier in the year (Fig. 3). The density of H. marcidus on upper surfaces varied with month but not with substratum particle size (Table 1). There was a strongly significant interaction term between substratum category and month. Gut contents of nemourid stoneflies Gut contents of all three species of nemourid stoneflies consisted of algae (diatoms, coccoid green algae and filamentous green algae), inorganic debris, a large ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Herbivory in an acid stream 549 Fig. 3 Mean (‹ 1 SE) monthly (stones) and weekly (bedrock) density of Heterotrissocladius marcidus on the upper surfaces of small stones, large stones and bedrock. Fig. 2 Mean (‹ 1 SE) monthly (stones) and weekly (bedrock) densities of nemourid stoneflies on the upper surfaces of small stones, large stones and bedrock. proportion of detritus (biofilm matrix material and FPOM), and unidentified particles (Figs 4 and 5). Heterotrophic bacteria were relatively common in the guts, but these were not counted since their total area was < 1% of the total particle area. No significant difference was found among the diets of the three nemourid species so data were pooled to test, monthby-month, whether the proportion of algae in the guts differed among the three substrata. Algae were abundant in the diets of nemourid stoneflies all year round (Fig. 6) and comprised three Table 1 Results of two-way ANOVAs comparing the density (log x + 1 transformed) of nemourid stoneflies and the chironomid Heterotrissocladius marcidus, on the upper surfaces of substrata of different sizes over months. ***P < 0.001, nsP > 0.05 Determinant Factor d.f. Nemouridae density Particle size 2/112 Month 13/112 Month ´ particle size 26/112 H. marcidus density Particle size 2/112 Month 13/112 Month ´ particle size 26/112 F 3.261ns 3.442*** 3.460*** 2.154ns 16.886*** 4.474*** ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 groups. Filamentous green algae (mainly Mougeotia sp.) were often most abundant (10±60% of the total particle area), particularly on bedrock in spring 1994 and winter 1994±spring 1995. Diatoms, almost exclusively Eunotia exigua and Eunotia rhomboidea, were also common in gut contents (< 10% total particle area), while coccoid greens were usually rare (only < 1% of total particle area). The proportion of algae in the gut contents varied from month-to-month but displayed no apparent temporal pattern. However, a significant increase in the proportion of algae was found with increasing particle size in seven of the 14 months sampled (Fig. 6; Table 2). Two types of detritus were identified in the gut contents of Nemouridae, biofilm matrix material and fine particulate organic matter (FPOM) (Fig. 7). Throughout spring 1994, when discharge was sufficiently high to keep mineral substrata clear, stoneflies appeared to graze exclusively on the biofilm matrix. In summer and autumn when discharge declined and FPOM accumulated on stones, this material began to appear in the diet. A resumption of grazing on the biofilm matrix coincided with a series of high flow events which scoured the stream bed in late October (see discharge hydrograph in Fig. 7), and continued 550 M. E. Ledger and A. G. Hildrew Fig. 4 Photomicrographs of DAPI-stained particles from guts of (A±C) Nemurella pictetii and (D) Heterotrissocladius marcidus illuminated under UV light. Filamentous green algae (A), coccoid green algae (B), the diatom Eunotia exigua (Eu), biofilm matrix (BM) and inorganic particles (In) (C) were commonly found. Detritus was abundant in the diet of Heterotrissocladius marcidus (D). Under UV light, red, blue and yellow coloration represents autofluorescence of algal chlorophyll, algal cell walls and membranes, and plant detritus, respectively. Scale Bar = 30 mm. throughout winter and spring 1995. In addition to biofilm and FPOM, nemourids ingested large quantities of inorganic debris and also unidentified particles (Appendix 1). Gut contents of chironomid larvae Particles in the guts of H. marcidus were assigned to one of six categories (Fig. 5; Appendix 1). In the case of H. marcidus, we could not distinguish between particles originating from the biofilm matrix and FPOM. Algal cells were found in the guts of larvae taken from all three substrata (Appendix 1), although they were usually less abundant than in nemourid stoneflies, and comprised < 10% of total particle area (except in July on bedrock). Three groups of algae (diatoms, coccoid greens and filamentous greens) were found in the guts of H. marcidus, with diatoms ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Herbivory in an acid stream 551 Detritus usually made up > 80% of total particle area in the guts of H. marcidus (Fig. 4; Appendix 1). Large quantities of inorganic debris (mineral particles) were ingested by larvae and comprised up to 50% of total particle area (Appendix 1), but unidentified particles were relatively uncommon (usually < 10%). The diet of H. marcidus showed little obvious seasonal pattern and few differences were found among substrata. Discussion Fig. 5 Mean annual composition of gut contents of three species of nemourid stoneflies and Heterotrissocladius marcidus, collected from the upper surfaces of mineral substrata. The detritus category comprises FPOM and biofilm matrix combined. being most numerous (< 9% total particle area). Coccoid green algae and filamentous green algae were found only occasionally. Fig. 6 Mean monthly (‹ 1 SE) percentage of algae in the diet of nemourid stoneflies collected from the upper surfaces of small stones, large stones and bedrock. `X' indicates that no larvae were available for examination. ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 In this study, we set out to establish whether `detritivorous' stonefly and chironomid larvae grazed biofilm from the upper surfaces of substrata in a stony riffle. We found that macroinvertebrates collected from three size classes of mineral substratum consumed algal biofilm throughout the year. Thus, the algae-grazer pathway was not qualitatively broken in this acid stream but was maintained, in the absence of specialist grazers, by dietary generalists. Ledger & Hildrew (1998) found that the quantity and quality of biofilm varied spatially and temporally in Lone Oak, its highest potential nutritional value (as evidenced by the protein, lipid and carbohydrate content) being on bedrock in spring. We hypothesised that the density and/or the gut contents of larvae would reflect this variation in biofilm state. Our current research has shown, however, that there was no response in the density either of the nemourids or of H. marcidus to this variation in food availability. In contrast to H. marcidus, however, the proportion of algae in the guts of stonefly larvae increased with particle size of the substratum from which they were collected. On some occasions when algal filaments were abundant, particularly on bedrock, they accounted for the entire gut contents. It appears therefore that the composition of gut contents reflected the composition of benthic material available, such that the intensity of grazing at Lone Oak, as indicated by gut content analysis, increased with substratum particle size. There was also temporal variation in the composition of the diet: nemourids ingested a large quantity of algae and biofilm matrix from the biofilm in spring and winter when substrata were cleared by flushing flows, but consumed FPOM and associated microflora in late summer and autumn when discharge fell and these materials accumulated. Thus, nemourids exhibited strongly opportunistic feeding, with a switch in 552 M. E. Ledger and A. G. Hildrew Fig. 7 Discharge hydrograph for Lone Oak, and mean monthly (‹ 1 SE) percentages of two types of detritus, biofilm matrix (shaded bars) and FPOM (black bars), in the guts of nemourid stoneflies collected from substrata of three sizes. `X' indicates that no larvae were available. diet according to the availability of attached (biofilm) and loose (FPOM) resources in Lone Oak over the study period. Temporal variation in the food consumed by invertebrates was also observed by Chapman & Demory (1963), working in two small Oregon streams. They found that several taxa (Baetis, Paraleptophlebia, Cynigmula, Epeorus, stone-cased Limnephilidae and Hydrobaenidae) showed seasonal changes in the consumption of algae and detritus, according to their availability. Further, they observed that facultative feeders tended to consume algae in winter and spring, and detritus in summer and autumn; patterns resembling those observed in this study. Similarly, significant differences in the composition of gut contents over time were shown in abandoned beaver ponds (Hall & Pritchard, 1975), where larvae of Tipula sacra Alexander grazed proportionally more algae in spring than summer in response to the alga's abundance. Many nemourid stoneflies have been described as shredders (Merritt & Cummins, 1984), a description supported by studies conducted in Broadstone Stream, an acidic and retentive iron-rich stream in southern England characterised by a low algal standing crop and abundant leaf litter. Gut contents of N. pictetii in that stream consisted of leaf fragments, detrital particles, iron bacteria and a small amount of algae, fungi and pollen (Henderson et al., 1990). Since larvae were over-represented on allochthonous litter, it is probable that they rasped the surface of leaves and ingested microbes as a result (Groom & Hildrew, 1989). However, Winterbourn et al. (1992) found a positive association between numbers of N. pictetii and algal biomass on nutrient-diffusing substrata in Broadstone Stream, suggesting that nemourids also fed on algae and thus exhibited dietary generalism in that system. Leaf litter is retained less efficiently in Lone Oak than Broadstone Stream, and biofilms are better developed on stone surfaces (Ledger, 1997). Dobson & Hildrew (1992) found that densities of nemourids in accumulations of leaf litter in Lone Oak were higher than on the surrounding stony substratum and reasoned that larvae could be food limited by the quantity of leafy detritus retained in this stony stream, and further, may prefer leafy detritus to biofilm as a source of food. We have shown here that biofilm from the upper surfaces of stones forms a substantial part of the diet of nemourids collected from stony substrata, and it could be that the poor food quality of leaves and biofilm in acid streams favours species with broad trophic requirements. Two factors that permit nemourids to exploit a wide variety of resources are mobility and mouthpart morphology. Recent field experiments in Broadstone Stream (Winterbottom et al., 1997) showed that nemourid larvae are very active, and colonize new substrata rapidly. `Continuous redistribution' (sensu Townsend & Hildrew, 1976) of this long legged, rapid ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Herbivory in an acid stream 553 Table 2 Results of monthly one-way ANOVAs comparing the percentage of algae (% total particle area) in gut contents of nemourid stoneflies from the upper surfaces of small stones (Sm), large stones (Lg) and bedrock (Bed). ***P < 0.001, **P < 0.01, *P < 0.05, ns P > 0.05 Month Test between d.f. Mar 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan 1995 Feb Mar Sm, Lg, Bed Sm, Lg Sm, Bed Lg, Bed Sm, Lg Sm, Lg, Bed Sm, Lg, Bed Sm, Lg, Bed Sm, Lg, Bed Lg, Bed Sm, Lg, Bed Lg, Bed Lg, Bed 2 2 1 1 1 2 2 2 2 1 2 1 1 crawler is likely to foster movement among microhabitats such as leaf packs, mineral substrata and even patches of filamentous algae. A generalist feeding strategy is also dependent on suitable mouthpart morphology. Observations of larvae in the laboratory and field suggest that N. pictetii larvae use their biting-mouthparts to rasp at the surface of substrata (Groom & Hildrew, 1989; Ledger, 1997). Thus, on conditioned leaves, larvae ingest a mixture of leaf debris and attached microbes, whereas on mineral substrata, larvae consume algae, bacteria and biofilm matrix as well as inorganic debris. Clearly, mouthpart morphology does not denote an obligate trophic status for the nemourid stoneflies. Chironomidae Larvae of H. marcidus were very abundant on the upper surfaces of stones and bedrock for a relatively short period in summer, when discharge was low and stable, and there was no significant variation in their numbers among substrata of different sizes. The diet of H. marcidus included a smaller proportion of algae than that of nemourid stoneflies (mean annual percentage 3.4 compared with 17.5), indicating that this chironomid was less important in maintaining the algae-grazer link. Although the relatively small size of detrital particles in chironomid guts made separation of biofilm matrix and FPOM impossible in this study, it is likely that the `detritus' fraction of the diet was composed mainly of FPOM, since increases in ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 F 16.195*** 4.198* 5.554* 0.078ns 1.015ns 1.827ns 0.033ns 1.865ns 22.398*** 169.473*** 3.282ns 5.067* 9.216** Post-hoc test (Tukey) Sm < Lg, Bed Sm < Lg, Bed Sm < Bed Sm, Lg < Bed Lg < Bed Lg < Bed Lg < Bed abundance of chironomid larvae occurred over the summer months when FPOM was accumulating. We found little temporal or spatial variation in the diet of H. marcidus in Lone Oak. This contrasts with the considerable variation observed in the diets of nemourids, and can probably be explained by differences in the mobility and feeding methods of the two groups. Thus, the nemourids are highly mobile and therefore encounter variable patches of biofilm and FPOM, whereas H. marcidus larvae are relatively sessile tube-dwellers that feed in a limited area immediately adjacent to their tubes. Little work has been done on the feeding biology of H. marcidus, although Winterbourn, Hildrew & Box (1985) reported that larvae could be reared from egg to third or fourth instar on epilithon from Broadstone Stream, and were also able to grow when fed loose FPOM. Trophic generalism The nemourid stoneflies (N. pictetii, N. cinerea, N. cambrica) and, to a lesser extent, the orthoclad chironomid H. marcidus exhibited generalist feeding habits. Such plasticity in feeding has been shown for various species of Amphipoda (Friberg & Jacobsen, 1994), Diptera (Hall & Pritchard, 1975; Berg, 1995; references therein), Trichoptera (Rhame & Stewart, 1976; Friberg & Jacobsen, 1994; Snyder & Hendricks, 1995) and Plecoptera (Mihuc & Mihuc, 1995; Mihuc & Minshall, 1995). Furthermore, Winterbourn et al. (1985) showed 554 M. E. Ledger and A. G. Hildrew that several acid-tolerant invertebrates, including nemourids, which shred leafy detritus, could also grow on biofilms from acid streams. The prevalence of opportunistic feeding, evidenced by these studies and our own, substantiates the contention of Mihuc (1997), that the assignment of taxa to functional feeding groups according to their mouthpart morphology does not equate to an obligate consumption of a single food type. Moreover, variation in feeding, particularly between streams, strongly suggests that a reliance on the literature as a source of information regarding invertebrate diet is inappropriate, and no replacement for an empirical analysis of feeding. Depending, as it does, on gut contents analysis, our study cannot reliably ascribe detrital carbon to autochthonous or allochthonous sources. We have revealed, however, that algae were prominent in the diets and our major conclusion is that these purportedly detritivorous species can and do maintain an important food web linkage with algae in acid streams. It could be that acidified freshwaters particularly favour trophic generalists (Sutcliffe & Hildrew, 1989). Species richness is limited in many acid stream systems and this could favour an increase in the feeding niches of the few remaining tolerant species. Alternatively, many of the insect taxa found in acid streams may be feeding generalists wherever they occur, and simply be outcompeted by specialists in more species-rich, circumneutral waters. It is also possible that the absence of predatory fish, such as trout, in acid streams enables primary consumers to exploit algal resources on the upper surfaces of stones where, in circumneutral waters, they would be vulnerable to predation. Whatever the mechanism, there is some evidence of ecological release in abundance of acid tolerant stoneflies in species-poor acidified streams (Hildrew, Townsend & Francis, 1984). Acknowledgments We thank Professor Jeffrey Duckett for his assistance with epifluorescence microscopy, Ms Jan Mckenzie for producing Fig. 4, and Drs Jill Lancaster and Peter Schmid, who identified stonefly larvae and chironomids, respectively. We are grateful to Charlotte Gjerlov, Simon Harrison, Helen Wilcock, Julie Win- terbottom, Guy Woodward and, particularly, Stuart Orton for help in the field. Prof. Mike Winterbourn, and two anonymous referees, made constructive comments on the manuscript. M.E.L was in receipt of a Natural Environment Research Council Studentship. References Berg M.B. (1995) Larval food and feeding behaviour. The Chironomidae: the Biology and Ecology of Non-Biting Midges. (Eds P. Armitage P. S. Cranston & L. C. V. Pinder), pp. 136±167. Chapman & Hall. Chapman D.W. & Demory R.L. (1963) Seasonal changes in the food ingested by aquatic insect larvae and larvae in two Oregon streams. Ecology, 44, 141±146. Cummins K.W. (1973). Trophic relations of aquatic insects. Annual Review of Entomology, 18, 183±206. Cummins K.W., Spengler G.L., Ward G.M., Speaker R.M., Ovink R.W., Mahan D.C. & Mattingly K.L. (1980) Processing of confined and naturally entrained leaf litter in a woodland stream ecosystem. Limnology and Oceanography, 25, 952±957. Dobson M. & Hildrew A.G. (1992) A test of resource limitation among shredding detritivores in low order streams in southern England. Journal of Animal Ecology, 61, 69±71. Friberg N. & Jacobsen D. (1994) Feeding plasticity of two detritivore±shredders. Freshwater Biology, 32, 133±142. Fromm P.O. (1980) A review of some physiological and toxicological responses of freshwater fish to acid streams. Environmental Biology of Fish, 5, 79±93. Gee J.H.R. (1988) Population dynamics and morphometrics of Gammarus pulex L. evidence of seasonal food limitation in a freshwater detritivore. Freshwater Biology, 19, 333±343. Griffith M.B. & Perry S.A. (1993) Colonization and processing of leaf litter by macroinvertebrate shredders in streams of contrasting pH. Freshwater Biology, 30, 93± 103. Groom A.P. & Hildrew A.G. (1989) Food quality for detritivores in streams of contrasting pH. Journal of Animal Ecology, 58, 863±881. Hall R.J., Likens G.E., Fiance S.B. & Hendry G.R. (1980) Experimental acidification of a stream in the Hubbard Brook Experimental Forest, New Hampshire. Ecology, 61, 976±989. Hall H.A. & Pritchard G. (1975) The food of larvae of Tipula sacra Alexander in a series of abandoned beaver ponds (Diptera: Tipulidae). Journal of Animal Ecology, 44, 55±66. ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Herbivory in an acid stream Henderson J., Hildrew A.G. & Townsend C.R. (1990) Detritivorous stoneflies of an iron- rich stream: food and feeding. Mayflies and Stoneflies (Ed. I. C. Campbell), pp. 249±254. Kluwer, Dordrecht. Hildrew A.G. (1992) Food webs and species interactions. The Rivers Handbook (Eds P. Calow and G. E. Petts), Vol. 1. Blackwell Scientific Publications, Oxford. Hildrew A.G., Townsend C.R. & Francis J.E. (1984) Community structure in some southern English streams: the influence of species interactions. Freshwater Biology, 14, 297±310. Hynes H.B.N. (1976) The biology of Plecoptera. Annual Review of Entomology, 21, 135±153. Ledger M.E. (1997) Grazing of algae by invertebrates in streams of contrasting pH. PhD Thesis, University of London. Ledger M.E. & Hildrew A.G. (1998) Temporal and spatial variation in the epilithic biofilm of an acid stream. Freshwater Biology, 40, 655±670. McArthur J.V., Barnes J.R., Hanson B.J. & Leff L.G. (1988) Seasonal dynamics of leaf litter breakdown in a Utah alpine stream. Journal of the North American Benthological Society, 7, 44±50. Merritt R.W. & Cummins K.W., (Eds) (1984) An introduction to the aquatic insects of North America. Kendall/Hunt, Dubuque, Iowa, USA. Mihuc T.B. (1997) The functional trophic role of lotic primary consumer: Generalist versus specialist strategies. Freshwater Biology, 37, 455±462. Mihuc T.B. & Mihuc J.R. (1995) Trophic ecology of five shredders in a Rocky Mountain stream. Journal of Freshwater Ecology, 10, 209±217. Mihuc T.B. & Minshall G.W. (1995) Trophic generalists vs. trophic specialists: Implications for food web dynamics in post-fire streams. Ecology, 76, 2361±2372. Mulholland P.J., Elwood J.W., Palumbo A.V. & Stevenson R.J. (1986) Effect of acidification on periphyton composition, chlorophyll and productivity. Canadian Journal of Fisheries and Aquatic Sciences, 43, 1846±1858. Rhame R.E. & Stewart K.W. (1976) Life cycles and food habits of three Hydropsychidae (Trichoptera) species in the Brazos River, Texas. Transactions of the American Entomological Society, 102, 65±99. Richardson J.S. (1991) Seasonal food limitation of detritivores in a montane stream: an experimental test. Ecology, 72, 873±887. Rosemond A.D., Reice S.R., Elwood J.W. & Mulholland P.J. (1992) The effects of stream acidity on benthic ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 555 invertebrate communities in the south-eastern United States. Freshwater Biology, 27, 193±209. Snyder C.D. & Hendricks A.C. (1995) Effect of seasonally changing feeding habits on whole-animal mercury concentrations in Hydropsyche morosa (Trichoptera: Hydropsychidae), Hydrobiologia, 299, 115±123. Sutcliffe D.W. (1978) Water chemistry and osmoregulation in some arthropods, especially Malacostraca. Freshwater Biological Association Annual Report, 46, 57± 69. Sutcliffe D.W. & Carrick T.R. (1973) Studies on mountain streams in the English Lake District. Freshwater Biology, 3, 437±462. Sutcliffe D.W. & Hildrew A.G. (1989) Invertebrate communities in acid streams. Acid Toxicity and Aquatic Animals (Eds R. Morris E. W. Taylor D. J. A. Brown and J. A. Brown), pp. 13±29. Seminar Series of the Society for Experimental Biology. Cambridge University Press, Cambridge. Townsend C.R. & Hildrew A.G. (1976) Field experiments on the drifting, colonization and continuous redistribution of stream benthos. Journal of Animal Ecology, 45, 759±772. Townsend C.R. & Hildrew A.G. (1988) Pattern and process in low-order acid streams. Verhandlungen der Internationalen Vereinigung fuÈr Theoretische und Angewandte Limnologie, 23, 1267±1271. Townsend C.R., Hildrew A.G. & Francis J. (1983) Community structure in some southern English streams: the influence of physicochemical factors. Freshwater Biology, 13, 521±544. Walker E.D., Olds E.J. & Merritt R.W. (1988) Gut content analysis of mosquito larvae (Diptera: Culicidae) using DAPI stain and epifluorescence microscopy. Journal of Medical Entomology, 25, 551±554. Winterbottom J.H., Orton S. & Hildrew A.G. (1997) Field experiments on the mobility of benthic macroinvertebrates in a Southern English stream. Freshwater Biology, 38, 37±47. Winterbourn M.J., Hildrew A.G. & Box A. (1985) Structure and grazing of stone surface organic layers in some acid streams of southern England. Freshwater Biology, 15, 363±374. Winterbourn M.J., Hildrew A.G. & Orton S. (1992) Nutrients, algae and grazers in some British streams of contrasting pH. Freshwater Biology, 28, 173±182. (Manuscript accepted 23 July 1999) 556 Heterotrissocladius marcidus Nemouridae Inorganic debris Unidentified Total algae Month Sm Lg Bed Sm Lg Bed Sm Mar 1994 37.5 (4.8) 27.3 (9.8) 45.3 (10.9) ± 9.5 (4.7) 30.9 (7.1) ± 37.5 (4.8) ± 5.0 (0.4) 21.7 (6.5) 11.1 (4.7) ± 2.1 (0.8) 8.5 (2.6) ± 5.4 (0.1) ± ± Apr May Jun Jul Aug ã 2000 Blackwell Science Ltd, Freshwater Biology, 43, 545±556 Sep Oct Nov Dec Jan 1995 4.4 (0.5) 26.5 (12.1) 12.3 (7.0) 0 0 11.0 (5.1) ± Feb 57.1 (6.6) ± Mar ± 0 0 3.3 (1.0) 3.9 3.8) 12.4 (3.7) 3.8 (0.5) 8.9 (1.0) 23.1 (10.0) 41.5 (22.8) 8.7 (5.1) 38.3 (2.3) 42.8 (9.6) 14.4 (5.5) ± 7.2 (2.7) 28.9 (12.2) 1.9 (1.9) 22.9 (4.2) 23.0 (9.0) 11.1 (8.1) 24.3 (6.4) 7.9 (3.8) 6.0 (2.0) 6.3 (0.5) 3.2 (2.3) 4.1 (2.1) 7.0 (5.0) ± 27.8 (3.7) ± ± 10.9 (2.2) 5.2 (0.5) 5.3 (0.1) 5.6 (0.3) 5.0 (0.1) 6.2 (0.1) 2.8 (0.5) 12.3 (2.1) 6.4 (4.1) 40.3 (5.6) 25.1 (6.7) 3.6 (0.5) ± 2.9 (0.8) 1.2 (0.2) 1.5 (0.4) 0 0 1.8 (0.5) 0 0 8.1 (3.4) 1.3 (0.5) 2.7 (1.2) 6.4 (3.3) 5.2 (2.4) 8.4 (2.1) 3.2 (1.4) 4.9 (1.4) 2.0 (0.4) ± ± ± ± ± Lg 2.0 (0.4) ± 3.1 (1.3) 2.3 (0.9) 2.5 (0.4) 2.3 (0.4) 4.5 (1.9) ± ± 4.1 (2.2) 2.6 (0.5) 1.9 (0.3) 1.2 (0.5) Detritus Bed Sm ± ± ± 9.5 (1..0) 3.2 (1.8) 50.3 (10.1) 1.2 (0.3) 9.0 (4.9) 2.1 (0.4) ± 65.2 (5.4) 31.9 (7.2) 89.5 (4.4) 35.8 (10.2) 75.2 (5.0) 72.2 (10.6) 59.1 (7.3) ± ± ± ± ± ± ± ± ± Inorganic debris Lg 95.8 (15.1) ± 83.9 (10.1) 97.6 (2.3) 90.0 (5.5) 85.5 (20.3) 94.2 (5.8) ± ± 72.7 (5.4) 89.3 (12.6) 86.2 (6.5) 97.1 (6.0) Bed Sm ± ± ± 52.4 (10.6) 80.8 (10.5) 34.9 (5.8) 48.2 (7.6) 38.9 (13.8) 77.8 (11.1) ± 30.4 (5.6) 45.8 (10.4) 20.1 (4.8) 51.2 (9.9) 30.0 (5.9) 25.3 (8.1) 34.4 (5.4) ± ± ± ± ± ± ± ± ± Lg 1.2 (0.4) ± 4.5 (1.0) 2.5 (1.1) 3.0 (1.2) 5.3 (2.2) 7.2 (2.5) ± ± 26.1 (4.7) 8.0 (2.0) 13.7 (4.1) 2.4 (2.1) Unidentified Bed Sm Lg ± ± ± 5.2 (2.1) 15.9 (7.7) 2.3 (2.1) 4.6 (2.1) 0.8 (0.2) 2.5 (1.5) 8.2 (2.5) ± 2.2 (0.6) ± ± 11.2 (4.9) 16.4 (8.2) 12.5 (6.1) 50.9 (7.2) 42.8 (11.3) 18.6 (6.3) ± ± ± ± ± ± ± ± ± 5.5 (2.3) 1.5 (0.9) 0.5 (0.4) 0.7 (0.3) 0.9 (0.4) ± ± 1.1 (0.7) 0.2 (0.01) 0.2 (0.01) 0.3 (0.1) Bed ± ± 0.5 (0.2) 0.3 (0.1) 2.3 (1.0) 0.5 (0.1) 9.2 (4.2) 1.6 (0.5) ± ± ± ± ± M. E. Ledger and A. G. Hildrew Appendix 1 Mean monthly percentage area (SE in brackets below) of particles in the gut contents of the Nemouridae (inorganic and unidentified particles only) and of Heterotrissocladius marcidus collected from small stones (Sm), large stones (Lg) and bedrock (Bed) over 13 months. `±' denotes that no larvae were available on upper surfaces