Riparian plant species loss alters trophic dynamics in detritus-based stream ecosystems

Riparian vegetation is closely connected to stream food webs through input of leaf detritus as a primary energy supply, and therefore, any alteration of plant diversity may influence aquatic ecosystem functioning. We measured leaf litter breakdown rate and associated biological parameters in mesh bags in eight headwater streams bordered either with mixed deciduous forest or with beech forest. The variety of leaf litter types in mixed forest results in higher food quality for large-particle invertebrate detritivores (‘shredders’) than in beech forest, which is dominated by a single leaf species of low quality. Breakdown rate of low quality (oak) leaf litter in coarse mesh bags was lower in beech forest streams than in mixed forest streams, a consequence of lower shredder biomass. In contrast, high quality (alder) leaf litter broke down at similar rates in both stream categories as a result of similar shredder biomass in coarse mesh bags. Microbial breakdown rate of oak and alder leaves, determined in fine mesh bags, did not differ between the stream categories. We found however aquatic hyphomycete species richness on leaf litter to positively co-vary with riparian plant species richness. Fungal species richness may enhance leaf litter breakdown rate through positive effects on resource quality for shredders. A feeding experiment established a positive relationship between fungal species richness per se and leaf litter consumption rate by an amphipod shredder (Gammarus fossarum). Our results show therefore that plant species richness may indirectly govern ecosystem functioning through complex trophic interactions. Integrating microbial diversity and trophic dynamics would considerably improve the prediction of the consequences of species loss.     

Top-down and bottom-up influences on populations of a stream detritivore

1. Knowledge of the influence of predatory fish in detritus‐based stream food webs is poor. We tested whether larval abundance of the New Zealand leaf‐shredding caddisfly, Zelandopsyche ingens (family Oeconesidae), was affected by the presence of predatory brown trout, Salmo trutta and the abundance of their primary detrital resource (Nothofagus leaves). 2. The density of Z. ingens and the biomass of leaves were determined in seven fishless streams and four trout streams in the Cass region, central South Island, on four occasions spanning 5 years. 3. Physicochemical conditions were similar in trout and fishless streams, but ancova indicated that Z. ingens numbers were positively related to leaf biomass and that caddisfly numbers were significantly greater in fishless streams than trout streams for any given biomass of leaf. The cases of trout stream larvae were also heavier per unit length than those in fishless streams. 4. Our results provide evidence for both top‐down and bottom‐up influences on a detritus‐based stream food web. Although stream detritivores may benefit from a habitat that provides both food and a degree of protection from predators, top‐down effects of predators on detritivore population abundance were still important. Thus, detrital resource availability may determine maximum attainable population size, whereas predation is likely to reduce the population to a level below that.     

Detrital processing in streams exposed to acidic precipitation in the Central Appalachian Mountains

Continuing high rates of acidic deposition in the eastern United States may lead to long-term effects on stream communities, because sensitive catchments are continuing to lose anions and cations. We conducted a two-year study of the effects of pH and associated water chemistry variables on detrital processing in three streams with different bedrock geology in the Monongahela National Forest, West Virginia. We compared leaf pack processing rates and macroinvertebrate colonization and microbial biomass (ATP concentration) on the packs in the three streams. Breakdown rates of red maple and white oak leaf packs were significantly lower in the most acidic stream. The acidic stream also had significantly lower microbial and shredder biomass than two more circumneutral streams. Shredder composition differed among streams; large-particle detritivores dominated the shredder assemblages of the two circumneutral streams, and smaller shredders dominated in the acidic stream. Within streams, processing rates for three leaf species were not significantly different between the two years of the study even though invertebrate and microbial communities were different in the two years. Thus, macroinvertebrate and microbial communities differed both among streams that differed in their capacity to buffer the effects of acidic precipitation and among years in the same stream; these differences in biotic communities were not large enough to affect rates of leaf processing between the two years of the study, but they did significantly affect processing rates between acidic and circumneutral streams.The Unit is jointly sponsored by the National Biological Service, the West Virginian Division of Natural Resources, West Virginia University, and the Wildlife Management Institute.The Unit is jointly sponsored by the National Biological Service, the West Virginian Division of Natural Resources, West Virginia University, and the Wildlife Management Institute.     

Leaf breakdown, detrital resources, and food webs in streams affected by mine drainage

Breakdown of leaf litter is essential for providing detrital resources for food webs but can be impaired by anthropogenic activities, which may disrupt energy flow to consumers. We investigated the relationship between leaf breakdown and food web structure in 12 streams with or without mining impacts on South Island, New Zealand. Six streams received inputs of acid mine drainage (pH 2.5–4.9), three were naturally acidic (pH ~5.0), and three were circumneutral (pH ~6.8). Streams affected by mining either had highly acidic water (pH <3) or iron precipitates present on substrata. Breakdown rates of leaves were significantly lower in mining-affected streams than circumneutral (by almost 50%) but not naturally acidic streams and were driven primarily by microbial activity, as shredding invertebrates were often absent. Mining-affected stream webs were simplified structures with fewer species and links than those in other streams. With few species to process leaf litter and transfer detrital resources, inputs of AMD disrupted both the mechanisms responsible for breakdown and links for energy flow. While faster breakdown rates were associated with larger food webs, limited function maintained in mining-affected streams was sufficient to support primary consumers and small food webs.     

Leaf breakdown in streams differing in catchment land use

1. The impact of changes in land use on stream ecosystem function is poorly understood. We studied leaf breakdown, a fundamental process of stream ecosystems, in streams that represent a range of catchment land use in the Piedmont physiographic province of the south‐eastern United States. 2. We placed bags of chalk maple (Acer barbatum) leaves in similar‐sized streams in 12 catchments of differing dominant land use: four forested, three agricultural, two suburban and three urban catchments. We measured leaf mass, invertebrate abundance and fungal biomass in leaf bags over time. 3. Leaves decayed significantly faster in agricultural (0.0465 day^?1) and urban (0.0474 day^?1) streams than in suburban (0.0173 day^?1) and forested (0.0100 day^?1) streams. Additionally, breakdown rates in the agricultural and urban streams were among the fastest reported for deciduous leaves in any stream. Nutrient concentrations in agricultural streams were significantly higher than in any other land‐use type. Fungal biomass associated with leaves was significantly lower in urban streams; while shredder abundance in leaf bags was significantly higher in forested and agricultural streams than in suburban and urban streams. Storm runoff was significantly higher in urban and suburban catchments that had higher impervious surface cover than forested or agricultural catchments. 4. We propose that processes accelerating leaf breakdown in agricultural and urban streams were not the same: faster breakdown in agricultural streams was due to increased biological activity as a result of nutrient enrichment, whereas faster breakdown in urban streams was a result of physical fragmentation resulting from higher storm runoff.     

Leaf litter breakdown rates in boreal streams: does shredder species richness matter?

1. Leaf litter breakdown rates were assessed in 23 boreal streams of varying size (first–seventh order) in central and northern Sweden. 2. Shredders were most abundant in small streams, while shredder species richness showed a hump-shaped relationship with stream order, with most species in fourth order streams. 3. In a partial least-squares regression analysis, year, water temperature, shredder species richness and shredder abundance were those factors correlating most strongly with leaf breakdown rates. Shredder species richness was more strongly correlated with leaf litter breakdown rates than shredder abundance, and shredder biomass showed no such correlation. 4. These data suggest that shredder species richness is an important variable in terms of leaf litter dynamics in streams.     

Leaf diversity influences in-stream litter decomposition through effects on shredders

1. The functioning of many aquatic ecosystems is controlled by surrounding terrestrial ecosystems. In a view of growing interest in linking biodiversity to ecosystem‐level processes, we examined whether and how leaf diversity influences litter decomposition and consumers in streams. 2. We tested experimentally the hypothesis that the effects of leaf diversity on decomposition are determined by the responses of leaf consumers to resource–habitat heterogeneity. Leaves from three common riparian trees, beech (Fagus sylvatica), hazel (Corylus avellana) and ash (Fraxinus excelsior), were exposed alone and in all possible mixtures of two and three species in a stream. We analysed individual leaf species for decomposition rate, microbial respiration and mycelial biomass, and we determined the species composition, abundance and biomass of shredders in leaf bags. 3. We found that the decomposition of the fastest decomposing leaves (hazel and ash) was substantially stimulated (up to twofold higher than single species leaf packs) in mixtures containing beech leaves, which are refractory. In contrast, the decomposition of beech leaves was not affected by leaf mixing. Such species‐specific behaviour of leaves in species mixtures has been overlooked in previous studies that examined the overall decomposition of litter mixtures. 4. The effects of leaf diversity on decomposition varied with the abundance and biomass of shredders but not with microbial parameters. Beech leaves alone were less attractive to shredders than leaf packs made of hazel, ash or any mixture of species. Moreover, the presence of beech leaves in mixtures led to higher shredder abundance and biomass than we had expected from data from single species exposed alone. Lastly, we found that early instars of the caddisfly Potamophylax (the dominant shredder in terms of biomass) almost exclusively used the toughest material (i.e. beech leaves) to construct their cases. 5. Leaf pack heterogeneity may have altered shredder‐mediated decomposition. Shredders colonising diverse leaf packs benefited from the stable substratum provided by beech leaves, whereas ash and hazel leaves were primarily used as food. Thus, our findings provide strong evidence for an intimate linkage between the diversity of riparian vegetation and aquatic communities.     

Comparison of leaf processing rates under different temperature regimes in three headwater streams

1. We studied the effects of different temperature regimes on leaf litter processing in three forested Appalachian headwater streams of different pH (mean pH = 4.2, 6.5, 7.5).
2. We compared leaf breakdown rates, microbial biomass and macroinvertebrate shredder density and biomass between two 12-week processing periods (October–January and November–February) in each stream. Leaf processing rates were calculated both as k (day^–1) and k _d (degree day^–1).
3. There were no significant differences in processing rates ( k day^–1) between the two study periods for any leaf species in any stream. The average difference in temperature between the two study periods was 175 degree days. Shredder density was significantly higher during the earlier study period on 40% of the sample dates, but shredder biomass was not significantly different between the two study periods. ATP concentration was significantly higher during the early study period for 60% of the sample dates.
4. More significant differences in these variables (shredder density and biomass, ATP concentration) were seen among the three study streams than between the two study periods. This indicates that in this study other factors, particularly stream pH, contributed more to processing rate variation than did differences in thermal regime.     

Comparison of leaf processing rates under different temperature regimes in three headwater streams

1. We studied the effects of different temperature regimes on leaf litter processing in three forested Appalachian headwater streams of different pH (mean pH = 4.2, 6.5, 7.5).
2. We compared leaf breakdown rates, microbial biomass and macroinvertebrate shredder density and biomass between two 12-week processing periods (October–January and November–February) in each stream. Leaf processing rates were calculated both as k (day^–1) and k _d (degree day^–1).
3. There were no significant differences in processing rates ( k day^–1) between the two study periods for any leaf species in any stream. The average difference in temperature between the two study periods was 175 degree days. Shredder density was significantly higher during the earlier study period on 40% of the sample dates, but shredder biomass was not significantly different between the two study periods. ATP concentration was significantly higher during the early study period for 60% of the sample dates.
4. More significant differences in these variables (shredder density and biomass, ATP concentration) were seen among the three study streams than between the two study periods. This indicates that in this study other factors, particularly stream pH, contributed more to processing rate variation than did differences in thermal regime.     

Natural and constrainment-induced factors influencing the breakdown of dogwood and oak leaves

Breakdown rates and microbial colonization patterns of dogwood and oak leaves were measured between November and June of 1987–88 and 1988–89. Leaves were placed in artificial streams loose (unconstrained), in bags, or in packs. Discharge was maintained at approximately 0.25 s^–1, and no shredders were present in the streams. Average microbial biomass as ATP, for all species and treatments, increased from near 0 mg g^–1 AFDW in November to over 8 mg g^–1 AFDW in June. Microbial respiration increased from about 0.01 µg glucose respired hr-g^–1 AFDW in November to about 0.03 µg hr-g^–1 AFDW in June. Microbial biomass and activity were significantly greater on dogwood leaves than on oak leaves. Dogwood and oak leaf breakdown rates were fastest when unconstrained, –0.0034 and –0.0027 degree-day^–1 respectively. Breakdown rates of dogwood leaves were faster in bags (–0.0025 degree-day^–1) than in packs (–0.0015 degree-day^–1) while rates of oak leaves were not significantly different between bags and packs (–0.0014 and –0.0018 degree-day^–1 respectively). Breakdown rates of dogwood and oak leaves obtained in this study were much slower than those obtained by other investigators either in the presence or absence of shredders. A comparison of results from this study with results from other studies revealed that dogwood leaves may be affected more by turbulence, while oak leaves may be influenced more by shredder activity.     

The importance of terrestrial subsidies in stream food webs varies along a stream size gradient

Cross system subsidies of energy and materials can be a substantial fraction of food web fluxes in ecosystems, especially when autochthonous production is strongly limited by light or nutrients. We explored whether assimilation of terrestrial energy varied in specific consumer taxa collected from streams of different sizes and resource availabilities. Since headwater streams are often unproductive, we expected that inputs from surrounding terrestrial systems (i.e. leaf litter, terrestrial invertebrates) would be a more important food source for consumers than in mid‐size rivers that have more open canopies and higher amounts of primary production available for consumers. We collected basal resources, invertebrates, and fish along a gradient in stream size in the Adirondack Mountains (NY, USA) and in Trinidad and Tobago and analyzed all samples for hydrogen isotopes as a means of differentiating biomass derived from allochthonous versus autochthonous sources. We found significant differences in allochthonous energy use within individual consumer taxa, showing that some taxa range from being entirely allochthonous to entirely autochthonous depending on where they were collected on the stream size gradient (grazers and collector–gatherer functional feeding groups), while other taxa are relatively fixed in the source of energy they assimilate (shredder and predator functional feeding groups). Consistent with expectations, allochthonous energy use was positively correlated with canopy cover in both regions for most feeding groups, with individuals from small, shaded streams having a more pronounced allochthonous signal than individuals collected from larger streams with less canopy cover. However, consumers in the shredder/detritivore feeding group did not vary among sites in their allochthonous energy use, and had a mostly allochthonous signal regardless of canopy cover and algal biomass. Our results demonstrate that the importance of energy from terrestrial subsidies can vary markedly but are similar in both temperate and tropical streams, suggesting a widely consistent pattern.     

Effects of a forest disturbance on shredder production in southern Appalachian headwater streams

• 1 The effects of a forest disturbance were investigated by comparing production of leaf-shredding aquatic insects in three streams draining a mature hardwood forest and three streams draining an 11-year-old, cable-logged clearcut.
• 2 Reference streams contained significantly greater mean annual standing crop of leaf material and significantly more slow-processing leaf material than disturbed streams. Disturbed streams had a significantly higher mean annual standing crop of fast-processing leaf material than the reference streams.
• 3 Leaf-shredding cranefly (Tipula abdominalis), caddisfly (Pycnopsyche gentilis), and stonefly (Tallaperla maria) larvae comprised over 95% of shredder biomass in all streams. Total shredder production was significantly greater (P<0.05) in disturbed versus reference streams, but individual production rates were not significantly different between stream types.
• 4 Pycnopsyche gentilis larvae were present at higher densities and achieved significantly greater annual biomass in disturbed versus references streams, Biomass of P. gentilis was significantly correlated with the standing crop of fast-processing, early successional leaf material in samples, whereas biomass of other shredders was correlated significantly with medium or slow-processing leaf species characteristic of later stages of forest succession.

     

How biological diversity influences ecosystem function: a test with a tropical stream detritivore guild

We investigated the relationship between diversity and ecosystem function, which is controversial and has rarely been examined for consumer assemblages, for the process of leaf breakdown by the shredder guild in a tropical stream. We manipulated species richness, evenness and identity of four macroinvertebrate shredder species (three caddisflies and one mayfly) in microcosms and tested their effect on leaf breakdown rates measured as leaf mass loss per capita and per milligram of animal. Species richness, evenness and species identity all affected leaf breakdown rates. Breakdown rates tended to increase with higher richness, but only for the three caddisflies, probably through a release of intraspecific interference, although other mechanisms such as niche complementarity or facilitation cannot be discarded. Leaf breakdown by the caddisflies was reduced in the presence of the mayfly, possibly because of its mode of movement by swimming instead of crawling and its similarity to some predators that are common in leaf litter. Species identity was more important than species richness in determining leaf breakdown rates, indicating that some species within the shredder guild are not redundant, and suggesting important consequences of particular species loss for the functioning of the ecosystem.     

Indirect effects of predatory trout on organic matter processing in detritus-based stream food webs

Indirect effects of predators on basal resources in allochthonous-based food webs are poorly understood. We investigated indirect effects of predatory brown trout ( Salmo trutta ) on detritus dynamics in southern beech ( Nothofagus spp.) forest streams in New Zealand through predation on the obligate detritivore, Zelandopsyche ingens (Trichoptera, Oeconesidae). Trout presence/absence and Z. ingens density were manipulated in flow-through tanks to investigate the lethal and sub-lethal effects of trout on litter processing by Z. ingens . An experiment that allowed trout access to Z. ingens showed trout predation reduced densities of Z. ingens resulting in slower breakdown of coarse particulate organic matter (CPOM) and reduced production of fine particulate organic matter (FPOM). A second experiment that prevented trout access to Z. ingens , but allowed the transmission of trout cues, resulted in no change in litter processing rates in the presence of trout. Litter processing rates were higher in high Z. ingens density treatments compared to low density treatments. Thus, trout effects on litter processing were due to reduced Z. ingens densities, not trout-induced modifications to Z. ingens feeding behaviour. Field assays of litter processing rates using artificial leaf packs in natural streams showed significant reductions in CPOM loss in trout streams compared to fishless streams. Z. ingens dominated biomass in fishless stream leaf packs, but a facultative shredder, Olinga feredayi , dominated trout stream leaf packs. Thus, the absence of Z. ingens drove differences in processing rates between trout and fishless streams and the indirect effects of trout on litter processing observed in mesocosms were evident in complex, natural food webs. Overall our study provides evidence that predators can influence resource dynamics in donor-controlled food webs through their effects on consumers.     

Habitat structure and fish predation: effects on invertebrate colonisation and breakdown of stream leaf packs

1. We investigated the effects of two features of leaf‐pack habitat structure (i.e. mass of a leaf pack and surface area of leaves comprising a leaf pack) and fish predation on colonisation of shredders and leaf breakdown rates in a coldwater stream. Packs were constructed of red maple (Acer rubrum) leaves. 2. A 2 × 3 × 3 factorial experiment was used to manipulate fish predation (exclusion and control cage), leaf‐pack mass (1, 3 and 5 g dry mass) and leaf surface area (small: approx. 17.9 cm², medium: approx. 34.6 cm², large: approx. 65.6 cm²). Exclusion cages had mesh on all sides, whereas control cages lacked mesh on two sides to provide access to fish. 3. Common shredders were Gammarus pseudolimnaeus, Pycnopsyche and Lepidostoma. Shredder biomass per leaf pack increased with the mass of a leaf pack (P < 0.001), but biomass per unit mass of leaf pack did not differ with leaf‐pack mass (P = 0.506). Shredder densities did not respond to the exclusion of fish (P > 0.7) or leaf surface area (P > 0.7), and interactions among treatment factors were not significant (P > 0.2). 4. Breakdown rates were lower for leaf packs comprised of small leaves (P < 0.001) and leaf packs with high mass (P = 0.001). Excluding fish did not significantly affect leaf breakdown rates (P = 0.293), and interactions among treatment factors were not significant (P > 0.3). Breakdown rates were highest when packs consisted of few leaves (i.e. leaf packs with large leaves and low mass) and were colonised by many shredders. 5. Fish predation was not an important factor controlling shredder densities in leaf packs over the spatiotemporal scale of our experiment. Nevertheless, we found shredder colonisation was proportional to leaf‐pack mass and breakdown rates were affected by leaf‐pack size (i.e. number of leaves in a pack). We suspect that fragmentation is the primary mechanism causing the breakdown rates to be dependent on leaf‐pack size.     

Understanding of colonization and breakdown of leaves by invertebrates in a tropical stream is enhanced by using biomass as well as count data

We hypothesized that (i) the importance of shredders for leaf breakdown is more evident in terms of their biomass than their abundance, due to the large bodies and high-feeding efficiencies of some typical shredders; (ii) non-shredder invertebrates select more refractory leaves because these are a more stable substrate for colonization or to obtain other forms of food. To test these hypotheses, we performed a decomposition experiment with leaves of contrasting chemical composition in a tropical stream, and determined the changes in the ash-free dry mass (AFDM) of the litter, and the invertebrate abundance and biomass during a 44-day period. The biomass of shredders showed a positive relationship with AFDM remaining, whereas their abundance was unrelated to AFDM. While shredder abundance represented only 4–12% of total invertebrate abundance, shredder biomass constituted 19–36% of total invertebrate biomass. We conclude that (i) shredder biomass expresses better than abundance the role of this guild in the decomposition of leaf detritus, demonstrating that they are important for the functioning of tropical streams; (ii) incubation time rather than stability of leaf litter as a substrate influences colonization by non-shredder invertebrates.     

Colonization and processing of leaf litter by macroinvertebrate shredders in streams of contrasting pH

• 1 Leaf litter processing rates and macroinvertebrate shredder assemblages in leaf packs were compared in four streams on the Allegheny plateau in the central Appalachian Mountains, U.S.A.; these streams were characterized by different bedrock geology and streamwater pH.
• 2 Leaf litter processing rates were fastest in the neutral streams, slowest in the acidic stream, and intermediate in the most alkaline stream.
• 3 Slower processing rates in the acidic stream were associated with lower total shredder biomass, made up predominantly by small leuctrid and nemourid stoneflies.
• 4 The differences in processing rates between the more alkaline stream and the neutral streams were not associated with differences in shredder biomass, but appeared to be related to taxonomic differences in the shredder assembiages. Insects were dominant in the neutral streams, and amphipods were dominant in the more alkaline stream.

     

Effects of crayfish on leaf processing and invertebrate colonisation of leaves in a headwater stream: decoupling of a trophic cascade

I performed a field experiment to test the hypotheses that omnivorous crayfish both promote breakdown of leaves (basal resources) and decouple any potential trophic cascade by simultaneously affecting intermediate consumers as well as their basal resource. Leaf packs were placed inside in situ artificial channels, which excluded or allowed access to crayfish. During a 4-week period, crayfish greatly promoted leaf processing, with decomposition rates among the fastest ever recorded from temperate streams. Crayfish also affected invertebrate abundance in the leaf packs. As a result of resource consumption, predation and bioturbation, crayfish treatments contained significantly lower densities of invertebrates. In contrast, exclusion of crayfish did not promote leaf decay via increased colonisation by detritivores, primarily because of the conspicuous lack of shredder insects in New Zealand streams. The results support the hypothesis that omnivorous top consumers decouple cascading chains through simultaneous direct and indirect effects on intermediate consumers and basal resources. Decapod consumers, which have been largely ignored in leaf decomposition studies, can be key leaf processors in temperate streams where shredder insects are poorly represented. Received: 8 February 2000 / Accepted: 14 April 2000     

Organic matter breakdown as a measure of stream health in New Zealand streams affected by acid mine drainage

Functional indicators of stream health have the potential to provide insights into stream condition that cannot be gained by traditional structural indices. We examined breakdown of leaves, wood, and cotton cloth strips at 18 sites along a gradient of effects of drainage from coal mines in New Zealand to determine the usefulness of these methods as functional indicators of stream health. The pH varied from 2.7 to neutral across the streams, and the more acidic streams typically had higher concentrations of aluminum, iron, zinc, and other metal ions. Precipitates of metal (mainly iron) hydroxides were present in most streams affected by mine drainage, especially in those with a pH of 4–5. Breakdown rates of all organic matter types were highest in several reference streams with neutral pH and lowest in sites with high rates of metal hydroxide deposition. Breakdown was relatively fast in the most acidic streams (pH < 3), in some cases as fast as at reference sites; these sites also had elevated nutrient concentrations. Shredding invertebrates were absent in litterbags from acidic streams and common at only 2 reference sites; their presence contributed to fast breakdown of leaves in the field and in lab microcosms. Microbial respiration was closely related to breakdown rates of leaves and wood; it was high at neutral and highly acidic streams, but lower at sites with pH 4–5, where metal hydroxides were precipitating onto solid surfaces. In these metal hydroxide-stressed streams, leaf and wood breakdown was slower, and associated biota, including microbes, were more affected than by water chemistry stressors (pH, dissolved metals) associated with mine drainage. Litter breakdown and microbial respiration provide insight into the functioning of streams, yielding different responses than traditional structural measures based on macroinvertebrates, which did not accurately distinguish impacts from acid mine drainage.     

The importance of crayfish in the breakdown of rhododendron leaf litter

1. Rhododendron (Rhododendron maximum) is a common evergreen shrub in riparian areas of the southern Appalachians, where its leaves can comprise a large proportion of leaf litter in streams. However, they are relatively refractory and generally considered a low quality food resource for detritivores. 2. Our objective was to assess whether macroconsumers [primarily crayfish (Cambarus bartonii)] influence rhododendron leaf breakdown in a forested southern Appalachian stream in both summer (when leaves other than rhododendron are relatively scarce) and autumn (when other leaves are relatively abundant). We conducted two leaf decay experiments, one in summer and one in autumn, using pre‐conditioned leaves. Macroconsumers were excluded from the benthos of a fourth‐order stream using electric ‘fences’; we predicted that excluding macroconsumers would reduce the decay rate of rhododendron leaves in both summer and autumn. 3. In both experiments, breakdown rate was lower in exclusion treatments. Macroconsumers accounted for approximately 33 and 54% of rhododendron decay in summer and autumn, respectively. We attribute this effect to direct shredding of rhododendron by crayfish. Biomass of insect shredders, insect predators and fungi did not differ between control and exclusion treatments, indicating that insectivorous sculpins (Cottus bairdi) had no effect on rhododendron decay and that omnivorous crayfish did not exert an indirect effect via alteration of insect or fungal biomass. 4. The influence of shredding insects varied between summer and autumn. In summer, when other, more palatable leaf types were not available, rhododendron leaf packs appeared to provide ‘resource islands’ for insect shredders. There was a significant inverse relationship between insect shredders and leaf pack mass in the summer exclusion treatment: insects were the only organisms eating leaves in this treatment and, as shredder biomass increased, remaining leaf pack mass decreased. In the control treatment, however, we did not see this relationship; here, the effect of insect shredders was presumably swamped by the impact of crayfish. In autumn, when other leaves were abundant, insect shredder biomass in rhododendron leaf packs was less than one‐third of summer values. 5. Even at low density (approximately 2 m^–2) crayfish were able to influence an ecosystem process such as leaf decay in both summer and autumn. Given the threatened status of many crayfish species in the United States, this finding is especially relevant. Even small alterations in crayfish assemblages, whether via loss of native species and/or introduction of exotic species, may have significant repercussions for ecosystem function.