J. N. Am. Benthol. Soc., 2008, 27(2):321–331 Ó 2008 by The North American Benthological Society DOI: 10.1899/07-054.1 Published online: 25 March 2008
Leaf decomposition and invertebrate colonization responses to manipulated litter quantity in streams
S. D. Tiegs1
F. D. Peter2
Department of Aquatic Ecology, Eawag: Swiss Federal Institute of Aquatic Science and Technology, and Institute of Integrative Biology (IBZ), ETH Zurich, 6047 Kastanienbaum, Switzerland
C. T. Robinson3
Department of Aquatic Ecology, Eawag: Swiss Federal Institute of Aquatic Science and Technology, and Institute of Integrative Biology (IBZ), ETH Zurich, 8600 Du¨bendorf, Switzerland
M. O. Gessner5 Department of Aquatic Ecology, Eawag: Swiss Federal Institute of Aquatic Science and Technology, and Institute of Integrative Biology (IBZ), ETH Zurich, 6047 Kastanienbaum, Switzerland
Abstract. Resource availability is an important ecosystem attribute that can influence species distributions and ecosystem processes. We manipulated the quantity of leaf litter, a critical resource in streams, in a replicated field experiment to test whether: 1) greater litter quantity promotes microbial leaf decomposition (through greater microbial inoculum potential), and 2) reduced litter quantity enhances decomposition by leaf-shredding invertebrates (because shredders aggregate on rare resource patches). In each of 3 streams, we identified reaches in which litter quantity was either: 1) augmented, 2) depleted, or 3) left unchanged. We determined decomposition rates and macroinvertebrate colonization of alder leaves placed in coarse- and fine-mesh litter bags, an approach intended to allow or prevent access to leaves by leaf-shredding macroinvertebrates. Responses to litter manipulations were complex. In 2 streams, litter quantities differed among treatments, but high quantities of litter in the control reach of the 3rd stream produced an overall variable pattern. Microbial decomposition was similar across litter treatments. In contrast, in the 2 streams where litter manipulation was successful, decomposition in coarse-mesh bags tended to be faster where litter was scarce than where it was abundant. Abundances of total and leaf-shredding macroinvertebrates in litter bags did not differ among litter manipulations in these 2 streams. However, a litter-consuming amphipod (Gammarus fossarum) tended to be most abundant in bags placed in litter-depleted reaches in the 2 streams, indicating that this large and highly mobile shredder might have been instrumental in causing differences in decomposition in response to litter manipulations. Overall, the effects caused by alteration of litter quantities on leaf decomposition and macroinvertebrate colonization were relatively weak. Nevertheless, results from 2 of the 3 streams where litter manipulation was successful were consistent with the hypothesis that short-term changes in resource availability might influence ecosystem processes by determining the spatial distribution of key consumers. Key words: leaf retention, invertebrate aggregation, benthic organic matter, leaf breakdown, shredders, Gammarus, ecosystem process.
The quantity of resources available in ecosystems is a key factor that determines the spatial distribution of organisms, which, in turn, might govern ecosystem processes. A fundamental resource in many streams is leaf litter derived from terrestrial vegetation (Wallace et al. 1999). Leaves in the canopies of streamside trees cast shade that strongly limits instream primary
Current address: Department of Biological Sciences, University of Notre Dame, Notre Dame, Indiana 46556 USA. E-mail: [email protected]
2 E-mail addresses: [email protected]
3 [email protected]
4 [email protected]
5 [email protected]
S. D. TIEGS
production (Sabater et al. 2000). When shed and retained in stream channels, these leaves provide habitat and food or substrate to benthic invertebrates and microorganisms (Baldy et al. 2007, Greenwood et al. 2007). As a result, terrestrially derived leaves constitute the major basal resource of food webs in small forest streams (Wallace et al. 1999), and the quantity of leaf litter present is likely to be a critical factor governing stream ecosystem structure and function. Quantities of litter present at a given time and location hinge on 3 processes: input, retention, and decomposition. Litter input depends on the density, composition, and productivity of riparian vegetation (Benfield 1997) and on the hillslope transport (i.e., lateral input) that delivers material to the stream from the forest floor (Webster et al. 1995). Retention of litter in streams is determined by interactions between hydrologic and geomorphic features, such as channel depth, grain size of the substratum, and abundance of large wood (Jones 1997, Hoover et al. 2006). Because these factors vary markedly across time and space, litter retention is also highly variable (Webster et al. 1999, Larranaga et al. 2003). Litter retention might be as important as input in determining quantities of benthic litter. For example, in a multiple regression analysis with data from 19 streams located throughout the USA, variables related to channel retentiveness explained more variability in benthic litter quantities across streams than did variables related to input (Jones 1997). When retained in stream channels, litter is colonized and used by detritivorous macroinvertebrates (shredders) and by microbial decomposers, particularly fungi (Anderson and Sedell 1979, Hieber and Gessner 2002). The interplay of these organisms determines the biological decomposition of leaf litter in streams (Gessner et al. 1999). Experimental manipulation of litter quantity in streamside channels has shown that shredders can track litter resource patches. Such resource tracking can lead to aggregation of shredders on leaf packs and consequent acceleration of decomposition (Rowe and Richardson 2001). Results from this small-scale and short-term experiment have been corroborated by data from resource-depleted streams, such as those that drain clear-cut catchments (Benfield et al. 2001) or those located above tree line (Robinson et al. 1998). In this and other situations where litter resources in streams are rare, aggregation of shredders on experimental leaf packs can be massive and can dramatically accelerate decomposition beyond the rates caused by microbial activity alone (e.g., Baldy and Gessner 1997, Robinson et al. 1998). Effects of shredder aggregation on decomposition
have been difficult to demonstrate by manipulating litter quantity in whole-stream experiments. Reice (1991) altered litter quantities in a series of 30-m stream reaches and found no evidence for changes in decomposition rates in either litter-augmented or litter-depleted reaches. Leaf decomposition rate did not differ between a headwater stream from which litter was experimentally excluded and a reference stream during the 1st year after exclusion (Eggert and Wallace 2003). The absence of immediate responses to litter manipulations in these 2 studies is surprising in view of numerous field observations of shredder aggregation, corresponding rapid decomposition, and clear effects on decomposition in experimental stream channels. Thus, the extent to which shredder aggregation responses to altered quantities of benthic litter determine leaf litter colonization and decomposition is unclear. Microbial decomposers also might respond to changes in litter availability, which would have consequences for decomposition. Experimentally introduced litter might immobilize nutrients and slow microbial decomposition when dissolved nutrient concentrations in stream water are low (Eggert and Wallace 2003). In contrast, abundant benthic litter might promote, rather than curb, microbial decomposition when nutrient supply is less critical. Such an effect could arise as a consequence of the life-cycle characteristics of aquatic hyphomycete fungi, the key microbial decomposers of leaf litter in streams (Gessner et al. 2007). Aquatic hyphomycetes are characterized by rapid and dense sporulation following initial establishment and growth in decomposing litter (Gessner and Chauvet 1994). As a result, spore densities in stream water should be greater when benthic litter is abundant than when it is sparse, and the difference is likely to be large, as indicated by often-dramatic increases in spore concentrations following autumn leaffall (Ba¨rlocher 2000) and greater spore concentrations following experimental enhancement of litter retention (Laitung et al. 2002). Given that leaf decomposition in microcosms is considerably faster when spore densities are high than when they are low (Treton et al. 2004), greater fungal colonization and decomposition of leaves would be expected when quantities of benthic litter are high. We present results from an experimental manipulation of benthic litter quantity at the scale of the stream reach (sensu Bisson and Montogomery 2006) in 3 streams. We designed this replicated experiment to test 2 hypotheses related to the effect of benthic litter quantity on macroinvertebrates colonization and decomposition. First, we hypothesized that, in the absence of pronounced nutrient limitation, augmenta-
TABLE 1. Physical and chemical characteristics of the 3 streams at low-flow conditions. Temperature data refer to daily means during the study period. Other measurements were taken on a single sampling date during the study. Values in parentheses are standard deviations; n/a indicates data were not available. Stream
Elevation (m asl)
Andelsbach Fohrenbach Mu¨hlebach
530 740 420
2.3 (0.53) 3.8 (0.71) 2.9 (0.65)
30 65 33
5.5 (1.5) 5.3 (1.6) 6.9 (1.2)
2.3 n/a 1.3
0.67 0.47 1.08
108 115 151
3.5 81.1 5.3
5.0 4.3 3.0
961 649 2500
tion of benthic litter would increase microbial decomposition relative to controls. We addressed this hypothesis by testing whether rates of leaf decomposition in fine-mesh litter bags, which restrict shredder access, varied across stream reaches in response to experimental augmentation or depletion of the quantity of benthic litter. Second, we hypothesized that shredding invertebrates would promote decomposition, particularly in stream reaches where benthic litter was scarce. We tested this hypothesis by comparing shredder colonization and decomposition of leaf packs in coarse-mesh litter bags placed in reaches with experimentally augmented, depleted, or unmanipulated litter. Methods Study sites Three streams (Andelsbach, Fohrenbach, Mu¨hlebach) were selected in the southern Black Forest, Germany. Streams had similar watershed geology (granitic), land cover (primarily managed forests), and size (3rd-order). Channels were characterized by pool and riffle sequences, substrata were dominated by coarse gravel and small cobble (sensu Wentworth 1922), and similar volumes of large wood were present, as judged by visual inspection along the study sections. Riparian vegetation consisted of a closed-canopy mixed community, including alder, maple, ash, and beech. Table 1 summarizes the major chemical and physical attributes of the streams. Experimental design Three reaches were delineated in each stream. Reaches were 40 m long and homogeneous in terms of morphological characteristics and riparian vegetation. One of 3 treatments was assigned to each reach in each stream: 1) litter augmentation above background levels, 2) litter depletion, and 3) unmanipulated control. Potentially confounding upstream–downstream effects were mitigated by constraining random assignment of treatments to reaches using the criterion that each treatment was replicated in an upstream,
middle, and downstream reach, and each treatment was replicated only once within a stream. Reaches were separated by .100 m to minimize influences of upstream treatments on treatments in middle or downstream reaches. Thus, the result was a constrained complete block design using the stream as the blocking factor (Fig. 1). Litter augmentation was achieved by means of litter traps (Dobson and Hildrew 1992, Dobson 2005). Traps consisted of plastic mesh (20 3 20 cm, 1-cm mesh size) held vertically in the stream by 2 rebars hammered into the streambed and oriented perpendicular to stream flow. In each augmented reach, 105 to 140 traps were added (average trap density ¼ 1/m2) just before leaffall (Dobson 2005). Litter depletion was accomplished by removing all visible leaf material from the reach by hand at weekly intervals. Each handful of leaf material was rinsed gently in the stream to minimize removal of resident invertebrates. The first depletion campaign occurred the same day that litter traps were installed. Litter volume removed during each depletion campaign was quantified by placing the collected leaf material into a rigid plastic bin (volume ¼ 0.20 m3) and measuring the height of leaf material in the bin. Numerous holes (2cm diameter) drilled into the walls of the bin allowed water to drain. The content of the bin was mixed thoroughly by hand, and three 1.6-L subsamples were taken to the laboratory where they were oven-dried (1058C, 5 d) and weighed. Dry mass of subsamples was extrapolated to the volume of leaf material in the bin to estimate the total mass of litter collected in reaches during each removal campaign. Effectiveness of litter augmentation and depletion in modifying the quantity of benthic litter on the streambed was determined with a cylindrical sampler (area ¼ 0.071 m2) 58 d after initiation of the experiment. The sampler was placed at 20 locations that were randomly selected within a grid along each stream reach. The enclosed litter on the streambed was gathered, placed in plastic bags, transported to the laboratory, oven-dried (1058C, 5 d), and weighed.
S. D. TIEGS
FIG. 1. Schematic diagram of the randomized block design used to test the effect of benthic litter quantity on leaf decomposition and macroinvertebrate abundance. Litter quantity was manipulated in 3 reaches (litter depletion, litter augmentation, and unmanipulated control) in each of 3 study streams. Random assignment of litter manipulations to reaches was constrained to ensure that each treatment was replicated in an upstream, middle, and downstream reach, and each treatment was replicated only once within a stream. Stream-flow direction is from the top of the figure downward. Minimum distance between reaches was 100 m.
Litter-bag preparation, installation, and sample processing A litter-bag approach was used to determine decomposition rates. Bags were constructed of either coarse-mesh (10-mm mesh size) or fine-mesh (0.5-mm mesh size) plastic netting to allow or prevent, respectively, macroinvertebrate access to enclosed leaves. Recently senesced leaves of alder (Alnus glutinosa [L.] Gaertn.), a common riparian species throughout most of Europe, were collected, air-dried, and weighed into batches of 5.00 6 0.25 g. After weighing, each batch was remoistened to render the leaves pliant, and the leaves were placed into the mesh bags. Five randomly selected litter bags were used to estimate initial leaf moisture content by drying (1058C, 24 h) and reweighing the material. Litter bags were taken to the field the next day, and 6 coarse-mesh and 6 fine-mesh bags were placed in the middle of each of the 9 study reaches. Steel rods were hammered into the streambed, and a coarse- and finemesh bag was attached to each rod with nylon cord. Flat cobbles were placed on the cord immediately
upstream of each bag to prevent bags from moving in the current and to ensure contact with the sediment. After 41 d of exposure in the streams, all leaf bags were retrieved and placed in plastic bags. Bags were returned to the laboratory in a cooler and frozen for later processing. In the laboratory, the contents of each plastic bag were emptied into a shallow tray with a small amount of water and allowed to thaw. Each leaf was cleaned individually with a soft-bristled paint brush to remove adhering debris and macroinvertebrates, placed in an aluminum tray, oven-dried (1058C, 48 h), and weighed to the nearest 0.01 g. Invertebrates were collected on a 500-lm mesh screen, preserved in 70% ethanol, identified to the lowest practicable taxon, counted, and assigned to functional feeding groups (Gessner and Dobson 1993, Merritt and Cummins 1996). Statistical analysis Differences in log10(x þ 1)-transformed quantities of benthic litter on the streambed were tested with
TABLE 2. Results of analysis of variance for the effect of litter manipulation and stream on benthic litter quantity. Stream is treated as a blocking factor according to Newman et al. (1997). – indicates F was not calculated because the stream 3 litter manipulation term was significant.
Source of variation Stream Litter manipulation Litter manipulation 3 stream Error FIG. 2. Mean (61 SE) quantity of benthic litter on the streambed in litter-depleted, unmanipulated control, and litter-augmented reaches in 3 streams; n ¼ 20 samples in each reach. Note that the y-axis begins at 0.1 g.
analysis of variance (ANOVA) using the stream as a blocking factor. When the effect of litter manipulation was not consistent across streams, the litter manipulation 3 stream interaction term was included in the ANOVA model, following the rationale of Newman et al. (1997) and Quinn and Keough (2002). Differences in the percentage of leaf dry mass remaining in litter bags also were tested with ANOVA (using litter manipulation and mesh size as the main factors of interest, with stream as a blocking factor), and when streams differed in their responses, the interaction terms with stream were included in the model as well (Newman et al. 1997, Quinn and Keough 2002). Differences in log10(x þ 1)-transformed invertebrate abundance in coarse-mesh litter bags were tested with ANOVA (using litter manipulation as the main factor of interest and stream as a blocking factor). In this case, the litter manipulation 3 stream interaction was never significant, so all p-values reported for invertebrates refer to ANOVA models without this interaction term. When significant (i.e., p , 0.05) differences were observed among litter manipulations, Tukey’s post-hoc tests were used to identify the means that differed. Results
Sum of squares (%)
indicating an inconsistent effect of litter manipulation among streams. This effect was caused by large quantities of naturally accumulated benthic litter in the control reach of Mu¨ hlebach (Fig. 2). When Mu¨hlebach was excluded from analysis, the litter manipulation 3 stream interaction term was not significant (F2,114 ¼ 1.08, p ¼ 0.34), litter manipulation had a marginally significant effect on the quantity of benthic litter on the streambed (F2,2 ¼ 16.9, p ¼ 0.056), and the quantity of benthic litter was greater in augmented than in control and depleted reaches (Tukey’s post-hoc comparisons, p , 0.001) but did not differ significantly between control and depleted reaches (p ¼ 0.08). Benthic litter removed from depletion reaches The quantity of benthic litter removed from depleted reaches varied among streams. A total of 7.2 kg dry mass was collected in Andelsbach, 7.7 kg in Fohrenbach, and 21.8 kg in Mu¨hlebach, corresponding to a depletion of 72, 51, and 189 g/m2 in each stream channel. The mass of benthic litter removed from depleted reaches declined over the first month after litter traps were installed and subsequently remained low (Fig. 3), except in Mu¨hlebach in late November, when the quantity of benthic litter removed increased at a time that coincided with a minor rainfall and stream-flow event that occurred in this watershed but not the 2 others.
Quantity of benthic litter on the streambed
The quantity of benthic litter on the streambed differed significantly among streams (F2,175 ¼ 8.6, p , 0.001) and among litter manipulations (F2,175 ¼ 8.0, p , 0.001). Depleted reaches had consistently less benthic litter than control or augmented reaches (Fig. 2). However, the litter manipulation 3 stream interaction term was significant and was included in the ANOVA model (model 1 of Newman et al. 1997; Table 2),
Leaf decomposition was significantly faster in coarse-mesh than in fine-mesh litter bags (F1,100 ¼ 31.4, p , 0.001) and differed significantly among streams (F2,100 ¼ 8.1, p , 0.001; Fig. 4A, B). In contrast, litter manipulation had no significant effect on leaf decomposition (F2,100 ¼ 1.9, p ¼ 0.15), nor was the litter manipulation 3 mesh size interaction term significant (F2,100 ¼ 0.74, p ¼ 0.48). Leaf decomposition was more
S. D. TIEGS
FIG. 3. Dry mass of benthic litter removed from each litter-depleted reach through time. The abrupt increase observed in Mu¨hlebach on 25 November coincided with a rainfall event in this watershed and not the others.
variable across streams and litter manipulations in coarse-mesh (Fig. 4A) than in fine-mesh litter bags (Fig. 4B). Patterns of leaf decomposition were similar among litter manipulations in 2 streams (Andelsbach and Fohrenbach), whereas the pattern in Mu¨hlebach differed (Fig. 4A). When Mu¨hlebach was excluded from the analysis, the litter manipulation did not significantly affect leaf decomposition (F2,65 ¼ 2.6, p ¼ 0.08), but the litter manipulation 3 mesh size interaction term was significant (F2,65 ¼ 3.2, p ¼ 0.045). In Andelsbach and Fohrenbach, leaf decomposition rate in coarse-mesh bags decreased with the quantity of benthic litter in the reach (augmented , control , depleted; Fig. 4A). Leaf decomposition in fine-mesh bags varied little overall (Fig. 4B); the difference between mean leaf decomposition across all streams and litter manipulations was ,4% (Fig. 4B). Macroinvertebrate colonization Thirty-two taxa of macroinvertebrates were identified from coarse-mesh bags. Shredders accounted for 46% of all macroinvertebrates and consisted of stoneflies (Amphinemura, Leuctra, Nemoura, Protonemura, Taeniopteryx), limnephilid caddisflies, and the amphipod Gammarus fossarum. Stonefly shredders accounted for 77%, and the genus Nemoura accounted for 60% of all shredders. Large macroinvertebrates were almost never encountered in fine-mesh bags, indicating that fine-mesh bags effectively excluded the most important shredders. Very small nemourids were found in fine-mesh bags (449 individuals in fine-mesh bags; 1732 individuals in coarse- and fine-mesh bags combined). This suggests that an appreciable fraction of nemourids in
FIG. 4. Mean (61 SE) leaf dry mass remaining in coarsemesh (A) and fine-mesh (B) litter bags after 41 d of decomposition in litter-depleted, litter-augmented, and unmanipulated control reaches of three streams; n ¼ 6 litter bags for each mesh size in each reach. Note that the y-axes begin at 40%.
fine-mesh bags and, by inference, also in coarse-mesh bags, were very early instars, small enough to pass through a 0.5-mm mesh. In contrast, only 2 individuals of Gammarus (,0.5% of the total number) were observed in fine-mesh bags, indicating that Gammarus were typically larger than other abundant shredders. Most of the remaining individuals in fine-mesh bags (55% of all macroinvertebrates) were chironomids. Numbers of other invertebrates (collector–gatherers, collector–filterers, scrapers, and nonchironomid predators) were consistently low across streams and litter manipulations. The total number of invertebrates per coarse-mesh litter bag and number of shredders per coarse-mesh litter bag did not differ significantly among streams (total invertebrates: F2,49 ¼ 0.73, p ¼ 0.49; shredders: F2,49 ¼ 0.25, p ¼ 0.78) or litter manipulations (total invertebrates: F2,49 ¼ 0.24, p ¼ 0.79; shredders: F2,49 ¼ 0.20, p ¼ 0.81) (Fig. 5A, B). Numbers of individuals in coarse-mesh litter bags from other functional feeding groups did not differ significantly among litter
litter manipulations were not significant (F2,49 ¼ 2.86, p ¼ 0.067). However, when data from Mu¨hlebach (where only a few Gammarus colonized litter bags) were excluded from the analysis, the number of Gammarus per coarse-mesh litter bag also differed significantly among litter manipulations (F2,32 ¼ 3.47, p ¼ 0.043; Fig. 5D). In Andelsbach and Fohrenbach, the number of Gammarus per coarse-mesh litter bag was significantly lower in augmented than in depleted litter manipulations (Tukey’s post-hoc comparison, p ¼ 0.033). Discussion Test of the shredder aggregation hypothesis
FIG. 5. Mean (61 SE) number of total macroinvertebrates (A), total shredders (B), nemourid stoneflies (C), and Gammarus (D) in coarse-mesh litter bags after 41 d of leaf decomposition in litter-depleted, litter-augmented, and unmanipulated control reaches of 3 streams; n ¼ 6 coarsemesh litter bags in each reach.
manipulations (F2,49 , 1.29, p . 0.28). The number of nemourid stoneflies per coarse-mesh litter bag did not differ significantly among streams (F2,49 ¼ 2.38, p ¼ 0.10) or litter manipulations (F2,49 ¼ 1.78, p ¼ 0.18), even though nemourids were rare in Fohrenbach (Fig. 5C). In contrast, the number of Gammarus individuals per coarse-mesh litter bag differed strongly among streams (F2,49 ¼ 35.3, p , 0.001), although differences among
Previous studies on ecosystem effects of litter depletion on decomposition have yielded equivocal responses. Results from observational field studies and small-scale experiments in streamside channels have indicated an acceleration of decomposition when litter is scarce (e.g., Benfield et al. 1991, 2001, Robinson et al. 1998, Rowe and Richardson 2001), but this response has not been evident in field experiments (e.g., Reice 1991). The overall effects of benthic litter manipulations in our experiment were subtle; however, results from coarse-mesh bags in 2 of the 3 study streams were consistent with the predicted pattern: the leaf decomposition rate tended to be fastest in depleted reaches and slowest in augmented reaches. Differences in shredder colonization of experimental litter bags have been proposed as the mechanism underlying varying rates of leaf decomposition in response to litter quantity. This hypothesis proposes that shredders would aggregate most in litter bags exposed in resource-depleted environments (e.g., Benfield et al. 2001, Rowe and Richardson 2001), whereas shredders would be distributed across a larger number of resource islands in control reaches and, especially, in resource-augmented reaches. Data consistent with these hypotheses must show that: 1) shredders cause significant litter mass loss, and 2) shredders aggregate in experimental leaf bags in reaches where benthic litter is scarce, and they do not aggregate in experimental litter bags in reaches where benthic litter is abundant. The first requirement was met in our study streams. Leaf decomposition was significantly faster in coarse-mesh than in fine-mesh bags. More rapid decomposition in coarse-mesh than in fine-mesh litter bags can be caused by factors other than shedders (Boulton and Boon 1991), but such effects were unlikely to be important in our study because no indications of mechanical fragmentation in coarse-mesh bags or O2 depletion in fine-mesh bags were observed. Moreover, controlled experiments under various hydraulic conditions in experimental
S. D. TIEGS
stream channels have shown that litter mass loss does not differ between coarse-mesh and fine-mesh litter bags when shredders are absent, but it does differ significantly between the 2 types of litter bags when shredders are present (Ferreira et al. 2006). The 2nd requirement, aggregation of shredders in coarse-mesh bags placed in litter-depleted reaches, was less well met in our study because total numbers of macroinvertebrates, shredders, and nemourid stoneflies did not follow the predicted pattern. However, the distribution of Gammarus across reaches supports the idea—this shredder did converge on resource islands provided by our experimental litter bags. Specifically, Gammarus was more abundant in litter bags in depleted reaches and rarer in litter bags in augmented reaches. This pattern matched the patterns in benthic litter quantities and leaf decomposition rate in the 2 study streams where litter manipulations were successful (i.e., in Andelsbach and Fohrenbach). Gammarus are very effective leaf shredders (e.g., Groom and Hildrew 1989, Baldy and Gessner 1997, Dangles et al. 2004) that feed very selectively (Ba¨rlocher and Kendrick 1973, Arsuffi and Suberkropp 1989, Grac¸a et al. 2001). Moreover, they are extremely mobile relative to other invertebrates in our study, and, thus, they are capable of seeking out and making use of resource islands. Gammarus were almost never encountered in fine-mesh litter bags (0.5-mm mesh size), suggesting that specimens were larger and more effective at consuming leaves than other shredder species in our study streams. For example, the numerically abundant nemourid stoneflies often were found in fine-mesh litter bags, indicating that a large proportion of them were early instars with low biomass (2-mm length, ,0.05 mg). These early instars would have had a limited shredding capacity, even if leaves were their main diet. Mass loss also was significantly faster in coarsemesh litter bags than in fine-mesh bags in the stream where Gammarus was rare. This result suggests that other shredders contributed to litter mass loss as well. However, the lower mobility of those taxa might require greater differences in benthic litter quantities than those achieved in our study to demonstrate aggregation effects on leaf decomposition in field situations. Collectively, this evidence indicates that aggregation in experimental litter bags of highly mobile Gammarus, but not other shredders, could have been instrumental in causing the decomposition patterns observed across litter-manipulated stream reaches. The general implication is that resource availability might influence ecosystem functioning by modulating aggregation of key consumer species. In the long term, shredder-mediated changes of
decomposition rates also might occur as a result of changes in shredder production, an idea that is supported by data from a whole-stream litter-exclusion experiment (Eggert and Wallace 2003). No immediate response to litter exclusion was observed in that study, but decomposition of red maple leaves 1 and 2 y later was much slower in the litter-exclusion stream than in the reference stream, which received normal litter inputs (k ’ 0.010/d compared to 0.017/d). The suggested mechanism was severe food limitation, which restrained recruitment of shredders in the years following litter depletion. This mechanism is in accordance with data from another study, which showed that production of some large shredder taxa was markedly lower in the litter-exclusion stream than in the reference stream 1 y after litter inputs were prevented (Wallace et al. 1999, Eggert and Wallace 2003). Such an effect, while possible, would not have been captured by our single-season experiment. In contrast to litter-depletion experiments done at a large scale, litter-augmentation experiments typically have been conducted at smaller spatial scales. Furthermore, most litter-augmentation experiments have considered macroinvertebrate responses, rather than responses of litter decomposition or other processes, to altered litter availability. Shredders have shown positive responses when benthic litter quantities have been experimentally elevated (Dobson and Hildrew 1992). For example, in streamside experimental channels with varied quantities of benthic leaf litter, density and biomass of shredders increased in response to greater litter availability (Richardson 1991). Furthermore, abundances of invertebrates increased relative to controls when boulders or litter traps installed in stream channels increased litter quantity relative to controls (Dobson and Hildrew 1992, Negishi and Richardson 2003). However, shredder abundance remained unchanged in a similar boulder-introduction experiment that enhanced litter retention (Lepori et al. 2005, see also Wallace et al. 1995), and leaf decomposition also remained unchanged (Lepori et al. 2005). The results of these augmentation studies lend indirect support to the hypothesis that shredders accelerate leaf decomposition in resource-limited environments. Test of the microbial decomposition hypothesis Our hypothesis that larger quantities of benthic litter would lead to faster microbial decomposition in augmented than in control or depleted reaches was not supported by our results. The rationale behind this hypothesis was that larger quantities of decomposing benthic litter should lead to higher concentrations of fungal spores in stream water. A greater fungal
inoculum would accelerate fungal colonization of fresh litter and, therefore, microbial decomposition. Such an outcome has been observed in a microcosm experiment (Treton et al. 2004). In one field study, fungal spore concentrations in stream water were greater in stream reaches to which logs or litter traps (identical to those used in our study) had been added as retention structures (Laitung et al. 2002). Some of the control and depleted reaches in our study could have been exposed to elevated spore concentrations from augmented upstream reaches, but our experiment was designed to prevent the same systematic upstream– downstream effect in all 3 streams (Fig. 1). Furthermore, even if spore concentrations in our experimental reaches were influenced by upstream litter manipulations, the effect on decomposition was negligible because mass loss of leaves in fine-mesh litter bags was highly consistent across litter manipulations in all 3 streams. Effect of differences among streams on results of litter manipulations Macroinvertebrate responses to litter manipulation were different in Mu¨hlebach than in the other 2 streams. The control reach of this stream flowed along the base of a steep hillslope that delivered large lateral inputs of litter to the channel, so quantities of benthic litter were higher in the control than in the augmented reach. As a consequence, relative quantities of benthic litter in the 3 reaches of this stream differed from those intended by our manipulations. However, even if a less extreme control reach had been chosen, the response pattern to our litter manipulation probably would have been different from the responses in the other 2 streams because Gammarus was rare in all 3 reaches of Mu¨hlebach. If our conclusion regarding the critical role of Gammarus in leaf decomposition is correct, then the low abundance of this species in the Mu¨hlebach probably explains why decomposition patterns across reaches did not reflect shredder abundances. The deviating pattern among streams in our study illustrates the importance of replicating manipulative ecosystem experiments and of exercising great care when extrapolating results to other ecosystems, even when they appear to be similar. Whole-ecosystem manipulations are among the best means to assess effects of environmental or biotic factors on ecosystem processes and properties (Carpenter et al. 1989, 1995). A drawback of the approach is that practical constraints often preclude replication of treatments (sensu Hurlbert 1984). A suite of methods has been proposed to alleviate this difficulty (e.g., Carpenter et al. 1989,
1995, Wallace et al. 1999), but none of the methods fully resolves the problem (e.g., Murtaugh 2002). However, although not always practical, use of replicated designs often is possible, even in manipulative ecosystem experiments (e.g., Maron et al. 2006, Entrekin et al. 2008). Headwater streams are prime candidates for this approach because of their relatively small size. In summary, we observed that the quantity of leaf litter in stream channels influenced the colonization of experimental litter bags by Gammarus, and the abundance of Gammarus in litter bags, in turn, might have influenced leaf decomposition rate. In litterdepleted reaches, Gammarus appeared to aggregate in litter bags and to accelerate decomposition. However, these results were not consistent among all streams examined, which illustrates the importance of treatment replication when conducting manipulative ecosystem experiments. Acknowledgements We thank Markus Schindler, Simone Graute, Torsten Diem, Catherine Hoyle, Lucia Klauser, Caroline Joris, Angelika Rohrbacher, Michael Siegrist, and Michael Vock for their help in the field. Andrew Boulton, Alan Covich, Pamela Silver, and anonymous referees provided many useful comments on previous drafts of the paper. This research was funded by the Swiss State Secretariat of Education and Research (SBF No. 01.0087) as part of the European Union project RivFunction (contract no. EVK1-CT-2001–00088). Literature Cited ANDERSON, N. H., AND J. R. SEDELL. 1979. Detritus processing by macroinvertebrates in stream ecosystems. Annual Review of Entomology 24:351–377. ARSUFFI, T. L., AND K. SUBERKROPP. 1989. Selective feeding by shredders on leaf-colonizing stream fungi: comparison of macroinvertebrate taxa. Oecologia (Berlin) 79:30–37. BALDY, V., AND M. O. GESSNER. 1997. Towards a budget of leaf litter decomposition in a first-order woodland stream. Comptes Rendus de l’Acade´mie des Sciences Se´rie III 320:747–758. BALDY, V., V. GOBERT, F. GUEROLD, E. CHAUVET, D. LAMBRIGOT, AND J.-Y. CHARCOSSET. 2007. Leaf litter breakdown budgets in streams of various trophic status: effects of dissolved inorganic nutrients on microorganisms and invertebrates. Freshwater Biology 52:1322–1335. BA¨RLOCHER, F. 2000. Water-borne conidia of aquatic hyphomycetes: seasonal and yearly patterns in Catamaran Brook, New Brunswick, Canada. Canadian Journal of Botany 78:157–167. BA¨RLOCHER, F., AND B. KENDRICK. 1973. Fungi and food preferences of Gammarus pseudolimnaeus. Archiv fu¨r Hydrobiologie 72:501–516.
S. D. TIEGS
BENFIELD, E. F. 1997. Comparison of litterfall input to streams. Journal of the North American Benthological Society 16: 104–108. BENFIELD, E. F., J. R. WEBSTER, S. W. GOLLADAY, G. T. PETERS, AND B. M. STOUT. 1991. Effects of forest disturbance on leaf breakdown in southern Appalachian streams. Verhandlungen der Internationalen Vereinigung fu¨r theoretische und angewandte Limnologie 24:1687–1690. BENFIELD, E. F., J. R. WEBSTER, J. L. TANK, AND J. J. HUTCHENS. 2001. Long-term patterns in leaf breakdown in streams in response to watershed logging. International Review of Hydrobiology 86:467–474. BISSON, R. A., AND D. R. MONTOGOMERY. 2006. Valley segments, stream reaches, and channel units. Pages 23–49 in G. A. Lamberti and R. F. Hauer (editors). Methods in stream ecology. Academic Press, Amsterdam, The Netherlands. BOULTON, A. J., AND P. I. BOON. 1991. A review of methodology used to measure leaf litter decomposition in lotic environments: time to turn over an old leaf? Australian Journal of Marine and Freshwater Research 42:1–43. CARPENTER, S. R., S. W. CHISHOLM, C. J. KREBS, D. W. SCHINDLER, AND R. F. WRIGHT. 1995. Ecosystem experiments. Science 269:324–327. CARPENTER, S. R., T. M. FROST, D. HEISEY, AND T. K. KRATZ. 1989. Randomized intervention analysis and the interpretation of whole-ecosystem experiments. Ecology 70:1142–1152. DANGLES, O., M. O. GESSNER, F. GUE´ROLD, AND E. CHAUVET. 2004. Impacts of stream acidification on litter breakdown: implications for assessing ecosystem functioning. Journal of Applied Ecology 41:365–378. DOBSON, M. 2005. Manipulation of stream retentiveness. Pages 19–24 in M. A. S. Grac¸a, F. Ba¨rlocher, and M. O. Gessner (editors). Methods to study litter decomposition: a practical guide. Springer, Dordrecht, The Netherlands. DOBSON, M., AND A. G. HILDREW. 1992. A test of resource limitation among shredding detritivores in low order streams in southern England. Journal of Animal Ecology 61:69–77. EGGERT, S. L., AND J. B. WALLACE. 2003. Litter breakdown and invertebrate detritivores in a resource-depleted Appalachian stream. Archiv fu¨r Hydrobiologie 156:315–338. ENTREKIN, S. A., J. L. TANK, E. J. ROSI-MARSHALL, T. J. HOELLEIN, AND G. A. LAMBERTI. 2008. Responses in organic matter accumulation and processing to an experimental wood addition in three headwater streams. Freshwater Biology (in press). FERREIRA, V., M. A. S. GRAC¸A, J. L. M. P. DE LIMA, AND R. GOMES. 2006. Role of physical fragmentation and invertebrate activity in the breakdown rate of leaves. Archiv fu¨r Hydrobiologie 165:493–513. GESSNER, M. O., AND E. CHAUVET. 1994. Importance of stream microfungi in controlling breakdown rates of leaf litter. Ecology 75:1807–1817. GESSNER, M. O., E. CHAUVET, AND M. DOBSON. 1999. A perspective on leaf litter breakdown in streams. Oikos 85:377–384. GESSNER, M. O., AND M. DOBSON. 1993. Colonization of fresh and dried leaf-litter by lotic macroinvertebrates. Archiv fu¨r Hydrobiologie 127:141–149.
GESSNER, M. O., V. GULIS, K. A. KUEHN, E. CHAUVET, AND K. SUBERKROPP. 2007. Fungal decomposers of plant litter in aquatic ecosystems. Pages 301–324 in C. P. Kubicek and I. S. Druzhinina (editors). The Mycota. Volume IV: microbial and environmental relationships. 2 nd edition. Springer-Verlag, Berlin, Germany. GRAC¸A, M. A. S., C. CRESSA, M. O. GESSNER, M. J. FEIO, K. A. CALLIES, AND C. BARRIOS. 2001. Food quality, feeding preferences, survival and growth of shredders from temperate and tropical streams. Freshwater Biology 46: 947–957. GREENWOOD, J. L., A. D. ROSEMOND, J. B. WALLACE, W. F. CROSS, AND H. S. W EYERS . 2007. Nutrients stimulate leaf breakdown rates and detritivore biomass: bottom-up effects via heterotrophic pathways. Oecologia (Berlin) 151:637–649. GROOM, A. P., AND A. G. HILDREW. 1989. Food quality for detritivores in streams of contrasting pH. Journal of Animal Ecology 58:863–881. HIEBER, M., AND M. O. GESSNER. 2002. Contribution of stream detritivores, fungi, and bacteria to leaf breakdown based on biomass estimates. Ecology 83:1026–1038. HOOVER, T. M., J. S. RICHARDSON, AND N. YONEMITSU. 2006. Flow–substrate interactions create and mediate leaf litter resource patches in streams. Freshwater Biology 51:435– 447. HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54: 187–211. JONES, J. B. 1997. Benthic organic matter storage in streams: influence of detrital import and export, retention mechanisms, and climate. Journal of the North American Benthological Society 16:109–119. LAITUNG, B., J. L. PRETTY, E. CHAUVET, AND M. DOBSON. 2002. Response of aquatic hyphomycete communities to enhanced stream retention in areas impacted by commercial forestry. Freshwater Biology 47:313–323. LARRANAGA, S., J. R. DIEZ, A. ELOSEGI, AND J. POZO. 2003. Leaf retention in streams of the Ague¨ra basin (northern Spain). Aquatic Sciences 65:158–166. LEPORI, F., D. PALM, AND B. MALMQVIST. 2005. Effects of stream restoration on ecosystem functioning: detritus retentiveness and decomposition. Journal of Applied Ecology 42: 228–238. MARON, J. L., J. A. ESTES, D. A. CROLL, E. M. DANNER, S. C. ELMENDORF, AND S. L. BUCKELEW. 2006. An introduced predator alters Aleutian Island plant communities by thwarting nutrient subsidies. Ecological Monographs 76: 3–24. MERRITT, R. W., AND K. W. CUMMINS (EDITORS). 1996. An introduction to the aquatic insects of North America. 3rd edition. Kendall/Hunt, Dubuque, Iowa. MURTAUGH, P. A. 2002. On rejection rates of paired intervention analysis. Ecology 83:1752–1761. NEGISHI, J. N., AND J. S. RICHARDSON. 2003. Responses of organic matter and macroinvertebrates to placements of boulder clusters in a small stream of southwestern British Columbia, Canada. Canadian Journal of Fisheries and Aquatic Sciences 60:247–258.
NEWMAN, J. A., J. BERGELSON, AND A. GRAFEN. 1997. Blocking factors and hypothesis tests in ecology: is your statistics text wrong? Ecology 78:1312–1320. QUINN, G., AND M. KEOUGH. 2002. Experimental design and data analysis for biologists. Cambridge University Press, Cambridge, UK. REICE, S. R. 1991. Effects of detritus loading and fish predation on leafpack breakdown and benthic macroinvertebrates in a woodland stream. Journal of the North American Benthological Society 10:42–56. RICHARDSON, J. S. 1991. Seasonal food limitation of detritivores in a montane stream: an experimental test. Ecology 72:873–887. ROBINSON, C. T., M. O. GESSNER, AND J. V. WARD. 1998. Leaf breakdown and associated macroinvertebrates in alpine glacial streams. Freshwater Biology 40:215–228. ROWE, L., AND J. S. RICHARDSON. 2001. Community responses to experimental food depletion: resource tracking by stream invertebrates. Oecologia (Berlin) 129:473–480. SABATER, F., A. BUTTURINI, E. MARTI´, I. MUN˜OZ, A. ROMANI´, J. WRAY, AND S. SABATER. 2000. Effects of riparian vegetation removal on nutrient retention in a Mediterranean stream. Journal of the North American Benthological Society 19:609–620. TRETON, C., E. CHAUVET, AND J. Y. CHARCOSSET. 2004. Competitive interaction between two aquatic hyphomycete
species and increase in leaf litter breakdown. Microbial Ecology 48:439–446. WALLACE, J. B., S. L. EGGERT, J. L. MEYER, AND J. R. WEBSTER. 1999. Effects of resource limitation on a detrital-based ecosystem. Ecological Monographs 69:409–442. WALLACE, J. B., J. R. WEBSTER, AND J. L. MEYER. 1995. Influence of log additions on physical and biotic characteristics of a mountain stream. Canadian Journal of Fisheries and Aquatic Sciences 52:2120–2137. WEBSTER, J. R., E. F. BENFIELD, T. P. EHRMAN, M. A. SCHAEFFER, J. L. TANK, J. J. HUTCHENS, AND D. J. D’ANGELO. 1999. What happens to allochthonous material that falls into streams? A synthesis of new and published information from Coweeta. Freshwater Biology 41:687–705. WEBSTER, J. R., J. B. WALLACE, AND E. F. BENFIELD. 1995. Organic processes in streams of the eastern United States. Pages 117–187 in C. E. Cushing, K. W. Cummins, and G. W. Minshall (editors). Ecosystems of the world. Volume 22: river and stream ecosystems. Elsevier, Amsterdam, The Netherlands. WENTWORTH, C. K. 1922. A scale of grade and class terms for clastic sediments. Journal of Geology 30:377–392. Received: 11 June 2007 Accepted: 5 February 2008