Habitat characteristics and macroinvertebrate assemblages in boreal forest streams: relations to catchment silvicultural activities

We compared the stream habitat characteristics and macroinvertebrate assemblages of boreal headwater streams in both the Finnish and the Russian parts of a single river basin, the Koitajoki River. Over the last 50 years, the Finnish side of the catchment has been managed using modern forestry techniques, whereas Russian side has remained nearly unexploited and is near to its natural state. Differences in silvicultural activities were observed to contribute to differences in habitat structure. The channel habitats were in fairly natural state in the Russian reference streams, whereas the impacted Finnish sites were cleared and straightened. In comparison with the impacted channels, the abundance of coarse woody debris (CWD) was 10–100-fold higher in the reference streams. Implications on the forestry-induced deterioration of water quality were also observed. On the contrary, only small differences in macroinvertebrate assemblages were detected. Despite the lower amount of retentive structures (CWD), significantly higher relative abundance of shredders was observed in the forestry-impacted streams. Otherwise the zoobenthic communities were quite similar in the two subcatchments. We suggest that several mechanisms may explain this similarity: (1) community structure is controlled by naturally acidic conditions, (2) the adverse impacts of forestry on habitat structure and water quality of streams may be compensated by increased input of deciduous litter and organic compounds from drained, structurally young riparian forests and (3) macroinvertebrate species have flexible feeding habits and may thus readily adapt to changing conditions.     

Concordance between ecotypes and macroinvertebrate assemblages in Mediterranean streams

1. According to the guidelines of the European Water Framework Directive, assessment of the ecological quality of streams and rivers should be based on ecotype-specific reference conditions. Here, we assess two approaches for establishing a typology for Mediterranean streams: a top-down approach using environmental variables and bottom-up approach using macroinvertebrate assemblages.
2. Classification of 162 sites using environmental variables resulted in five ecotypes: (i) temporary streams; (ii) evaporite calcareous streams at medium altitude; (iii) siliceous headwater streams at high altitude; (iv) calcareous headwater streams at medium to high altitude and (v) large watercourses.
3. Macroinvertebrate communities of minimally disturbed sites ( n  = 105), grouped using UPGMA (unweighted pair-group method using arithmetic averages) on Bray–Curtis similarities, were used to validate four of the five ecotypes obtained using environmental variables; ecotype 5, large watercourses, was not included as this group had no reference sites.
4. Analysis of similarities ( anosim ) showed that macroinvertebrate assemblage composition differed among three of the four ecotypes, resulting in differences between the bottom-up and top-down classification approaches. Siliceous streams were clearly different from the other three ecotypes, evaporite and calcareous ecotypes did not show large differences in macroinvertebrate assemblages and temporary streams formed a very heterogeneous group because of large variability in salinity and hydrology.
5. This study showed that stream classification schemes based on environmental variables need to be validated using biological variables. Furthermore, our findings indicate that special attention should be given to the classification of temporary streams.     

Benthic macroinvertebrate assemblages in north Swedish streams: environmental relationships

An analysis of the relationships between lotic macroinvertebrates and environmental variables was earned out on material from 60 riffle sites in streams in northern Sweden The approach involved the use of TWINSPAN classification and canonical correspondence analysis on presence/absence data from two seasons (spring and autumn) Variables most strongly associated with distribution patterns of assemblages were drainage area, elevation, water quality in terms of alkalinity, colour and phosphate and the presence of macrophytes The significance of affinities of different species to these variables are discussed The eight clusters resulting from the TWINSPAN analysis could biologically be interpreted as classes of taxa related to stream size, chemical conditions and algae A multiple regression analysis for predicting species nchness using three independent variables, viz drainage area amount or organic matter, and discharge was constructed The results of the study could be used as a starting point for further work on the community organization of benthie stream assemblages     

Effects of episodic acidification on macroinvertebrate assemblages in Swiss Alpine streams

1. Despite long‐standing ecotoxicological evidence that episodes of acidification in streams are important biologically, there is still uncertainty about their effects on invertebrate communities. We surveyed 20 streams in an acid sensitive Alpine area (Canton Ticino, Switzerland), where episodes are driven by snowmelt in spring and by rainstorms at other times of the year. Samples of water and macroinvertebrates were collected in pre‐event conditions (winter and summer) and during periods of high flow (spring and autumn). 2. Using pH, [Ca²⁺] and [Alⁿ⁺], streams were clustered into six acid–base groups that were either well buffered (groups 4–6), soft‐water with stable pH (group 3), or poorly buffered with low pH at high flow (groups 1 and 2). 3. Severe episodes occurred during snowmelt, when the group 1 streams became acidic with pH down to 5.0 and [Alⁿ⁺] up to 140 μg L^?1. pH declined to 6.2 in streams of group 2, but remained > 6.6 in groups 3–6. 4. Detrended canonical correspondence analysis showed that the streams sensitive to episodes (groups 1 and 2) had different invertebrate assemblages from well‐buffered sites (groups 4 and 5) or soft‐water stable streams (group 3), with faunal differences largest following spring snowmelt. Empididae, Isoperla rivulorum, Rhithrogena spp. and Baetis spp. were scarce in streams sensitive to episodes (groups 1 and 2). By contrast, Amphinemura sulcicollis was scarcer in hard‐water streams (groups 4–6). Taxonomic richness was lower in the episodic streams of group 1 than in other streams. 5. Together, these results indicate clear biological differences between acid‐sensitive streams with similar low‐flow chemistry but contrasting episode chemistry. Severe episodes of acidification appear to affect macroinvertebrate assemblages in streams in the southern Swiss Alps.     

Longitudinal effects of herbivory on lotic periphyton assemblages

1. The longitudinal effects of herbivory on stream periphyton assemblages were examined in laboratory stream channels, each of which consisted of an upstream chamber, which either contained snail grazers or not, and downstream chambers, none of which contained grazers. Periphyton assemblages of two ages (0–21 days old and 21–42 days old) were sampled in both upstream and downstream chambers to detect proximate (i.e. localized) and longitudinal (i.e. downstream) effects of herbivory. 2. Both proximate and longitudinal effects were detected, although they differed in their impact on the periphyton assemblage. Periphyton biomass and cell accumulation were lower in grazed than in ungrazed upstream chambers throughout the experimental period. Accumulation rates on initially bare tiles were substantially higher downstream of grazed than of ungrazed chambers, but grazing had no effect on cell densities in established (21–42 day old) assemblages downstream. 3. Longitudinal effects of herbivory were not due to quantitative differences in the flux of propagules or nutrients from grazed and ungrazed chambers. Although not tested in this study, it is hypothesized that differences in the physiological condition of exported propagules may have contributed to differences in downstream colonization rates in grazed and ungrazed streams. 4. The magnitude of longitudinal impacts of herbivory and the importance of different causal mechanisms are predicted to vary depending on the standing crop and productive capacity of the periphyton assemblages as well as the consumptive demand of the herbivore guild.     

Spatio-temporal variation in macroinvertebrate assemblages of glacial streams in the Swiss Alps

1. Changes in water chemistry, benthic organic matter (BOM), and macroinvertebrates were examined in four different glacial streams over an annual cycle. The streams experienced strong seasonal changes in water chemistry that reflected temporal changes in the influence from the source glacier, especially in water turbidity, particulate phosphorus and conductivity.
2. Nitrogen concentrations were high (nitrate-N values were 130–274 μg L^–1), especially during spring snowmelt runoff. Benthic organic matter attained >600 g m^–2 dry mass at certain times, peaks being associated with seasonal blooms of the alga Hydrurus foetidus .
3. Macroinvertebrate taxon richness was two to three times higher (also numbers and biomass) in winter than summer suggesting winter may be a more favourable period for these animals. Benthic densities averaged 1140–3820 ind. m^–2, although peaking as high as 9000 ind. m^–2. Average annual biomass ranged from 102 to 721 mg m^–2, and reached >2000 mg m^–2 at one site in autumn.
4. Taxa common to all sites included the dipterans Diamesa spp. and Rhypholophus sp., the plecopterans Leuctra spp. and Rhabdiopteryx alpina , and the ephemeropterans Baetis alpinus and Rhithrogena spp. Principal components analysis clearly separated winter assemblages from those found in summer.     

Realism of model ecosystems: an evaluation of physicochemistry and macroinvertebrate assemblages in artificial streams

Ecological realism is an important yet rarely reported feature of model ecosystems. In this case study, we assess the realism of four outdoor artificial stream mesocosms (4 m²) bordering a chalk river in southern England. Comparisons of physiochemical conditions and benthic macroinvertebrate assemblages were made between the mesocosm units and the parent water body, a side arm of the River Frome. Physicochemistry of the mesocosm replicates was similar to that of the source stream, with congruent temporal variation evident between the real system and each of the models. The high realism of the mesocosms was explained by the outdoor location and close physical proximity of the array to the source stream, and the short mesocosm residence time of water sourced from the parent feeder system. Mesocosms supported a diverse array of benthic macroinvertebrates (60 families from 14 taxonomic orders), including all macroinvertebrate families in the source stream. Individual mesocosms contained a mean of 89% of source stream biota. We conclude that once-through mesocosms can be satisfactory analogues of natural systems, particularly where model and natural scales overlap. Handling editor: D. Dudgeon     

Microhabitat availability in Welsh moorland and forest streams as a determinant of macroinvertebrate distribution

SUMMARY. 1 Eighteen streams in mid-Wales were sampled for macro-invertebrates in both riffles and margins in April 1985–87. Stream macro-flora, substrata and marginal habitats were surveyed in May 1988.
2. TWINSPAN classification of the macroinvertebrate data indicated three major stream groups. One was distinguished by circumneutral pH and had a flora and fauna typical of such conditions. The other two groups consisted of acidic streams with moorland and conifer afforested catchments respectively. The forest streams were the more acidic but the two groups also differed significantly in the composition of their marginal habitats.
3 The acidic moorland streams had more vegetation ('soft' features) in the margins and supported several invertebrate taxa which were relative more abundant there than in the riffles. These taxa may be excluded from forest streams because the margins are 'hard' due to greater erosiveness and shading.
4. In view of the increasing cover by conifer afforestation in Britain, it is clearly necessary to elucidate all its effects on stream ecosystems, which include changes to the physical environment.     

The quantification of local substrate heterogeneity in streams and its significance for macroinvertebrate assemblages

The effect of substrate heterogeneity on the structure of stream macroinvertebrate assemblages (total abundance, taxon richness, and evenness) is still not clear, but this could be due to the lack of standard methods for quantifying substrate heterogeneity. An accurate quantification of substrate heterogeneity was obtained from photographs of sampled areas (each 225 cm²), which were used to create maps that were subsequently digitized and analyzed using image analysis software. These maps allowed the calculation of multiple metrics quantifying two aspects of substrate heterogeneity: composition and spatial configuration of substrate patches. The diversity of substrate types (calculated as the Shannon diversity index), and the heterogeneity of patch compactness (calculated as the coefficient of variation of the relationship between patch dimensions) were the metrics explaining more biotic variance at the sample scale, but at higher scales there were no relationships between assemblage structure and substrate heterogeneity. Most variation in substrate heterogeneity occurred at the sample scale, while some metrics varied significantly at riffle or segment scales; these patterns of variation match those of macroinvertebrate assemblages, which had been previously studied. The importance of quantifying substrate heterogeneity and considering the spatial scales of its study are discussed.     

Drivers of beta diversity of macroinvertebrate communities in tropical forest streams

1. There has recently been increasing interest in patterns of beta diversity but we still lack a comprehensive understanding of these patterns in various regions (e.g. the tropics), ecosystems (e.g. streams) and organism groups (e.g. invertebrates). 2. Our aim was to investigate the patterns of beta diversity of stream macroinvertebrates in relation to key environmental (i.e. stream size, pH and habitat degradation) and geographical variables (i.e. latitude, longitude, altitude) in a tropical region. We surveyed a total of 8–10 riffle sites in each of 34 streams (altogether 337 riffle sites were sampled) in Peninsular Malaysia to examine variation in macroinvertebrate community composition at within‐stream and among‐stream scales. 3. Based on test of homogeneity of dispersion, we found that the streams studied differed significantly in within‐stream variation in community composition (i.e. among‐site variation of within stream beta diversity). The patterns were similar based on Bray–Curtis coefficient on abundance data, Sorensen coefficient on presence–absence data and Simpson coefficient on presence–absence data. We also found that within‐stream beta diversity was significantly related to stream size, pH and latitude, with each of these variables individually accounting for around 20% of the variation in beta diversity in simple regressions, while the total variation explained by the three significant variables amounted to around 50% in multiple regressions. By contrast, habitat degradation, longitude and altitude were not significantly related to beta diversity. We also found that the factor drainage basin accounted for much of the variation in beta diversity in general linear models, suppressing the effects of environmental variables. 4. We concluded that within‐stream beta diversity is mainly related to a combination of the identity of a drainage basin and stream environmental factors. Our findings provide important background for stream environmental assessment and conservation planning by emphasising that (i) macroinvertebrate communities within streams are not homogeneous, but show considerable beta diversity, (ii) streams differ in their degree of within‐stream beta diversity, (iii) stream size and water pH should be considered in applied contexts related to within‐stream beta diversity and (iv) historical effects may be different in different drainage basins and may affect present‐day patterns of within‐stream beta diversity.     

Some effects of current velocity on periphyton communities in laboratory streams

Summary Some effects of current at velocities of 9 and 38 cm/sec on periphyton communities have been determined in laboratory streams.The diatom community that developed in the faster current formed a dense, felt-like growth on the gravel and rubble substrate and usually appeared dark green or brownish in color. At the slower current velocity, the community was dominated by species of Stigeoclonium, Oedogonium, and Tribonema which formed long, loose oscillating filaments on the substrate and resembled the aggregations of green filamentous algae often observed in ponds.Although the accumulation of biomass on the gravel and rubble was much more rapid in fast current than slow current, by the end of the experiment, the organic matter per unit area of substrate was approximately the same at both velocities. The export of biomass was consistently greater from the community subjected to the faster current, and at a near steady-state or constant standing crop, the highest productivity was maintained at the faster velocity.Dr. Harry K. Phinney has provided helpful advice during this work. The author is also indebted to Dr. Charles E. Warren and Dr. Peter Doudoroff for their invaluable administrative assistance.This paper is a contribution of the Pacific Cooperative Water Pollution and Fisheries Research Laboratories, Oregon State University, and U. S. Public Health Service Cooperating. This investigation was supported in part by National Science Foundation Research Grant GB 467.Technical Paper 1918, Oregon Agricultural Experiment Station.

---

Zusammenfassung Effekte von Strömungsgeschwindigkeiten von 9 and 38 cm/Sek auf Aufwuchsgesellschaften wurden unter Laboratoriumsbedingungen untersucht.Die Diatomeengesellschaft, die sich in der schnelleren Strömung entwickelte, zeigte dichten filzartigen Wuchs auf dem Kies- und Geröllsubstrat und war gewöhnlich dunkelgrün oder braun gefärbt. Die langsamere Strömungsgesch windigkeit führte ein Vorherrschen von Stigeoclonium, Oedogonium und Tribonema Arten herbei, die lange und lockere, oszillierende Filamente auf dem Substrat bildeten, welche den Ansammlungen von grünen Fadenalgen ähnelten, die man häufig in Teichen beobachtet.Die Mengen organischer Substanz pro Einheitsfläche des Substrates waren nahezu gleich am Ende des Experiments für beide Strömungsgeschwindigkeiten, obwohl die Ansammlung von Biomasse auf Kies and Geröll viel rascher in der schnellen als in der langsamen Strömung vor sich ging. Der Abtransport von Biomasse war durchwegs grösser in der schnelleren Strömung und hier wurde nach Erreichen des Gleichgewichtzustandes oder stetigen Ertragzustandes auch die höchste Produktivität verzeichnet.

     

Seasonal fluctuations in soil phenolics of a coniferous forest: effects on seed germination of different coniferous species

This study examined seasonal variations in the content of water-soluble phenolics extracted from litter and Ah horizon of forest soil under Pinus laricio Poiret trees. After having identified and quantified different phenolics, the effects of seasonal phenolic extracts, single phenolic acids and synthetic phenol mixtures on seed germination of Pinus laricio Poiret, Pinus pinaster Aiton and Pinus halepensis Mill. were evaluated, and the activity of phenolics on the main enzymes involved in the glyoxylate cycle was tested. The results indicated a seasonal variation of phenol content in soil, with largest concentrations of water-soluble phenolic acids in autumn and smallest concentrations in summer. Addition of phenol extracts to germination medium reduced seed germination showing phytotoxic effects which differed, depending on the species and the fractions tested. Phenols extracted from litter and Ah horizon in autumn and winter, vanillic acid, and synthetic mixtures reproducing phenols extracted from soil in winter inhibited seed germination of Pinus laricio, pinaster and halepensis more than other treatments. A greater inhibitory effect was observed on seed germination of Pinus pinaster and halepensis compared to Pinus laricio. The phenols also had the greatest inhibitory effect on glyoxylic enzyme activities. Section Editor: H. Lambers     

Parallels and contrasts in the effects of drought on stream macroinvertebrate assemblages

• 1 It is axiomatic that unusually long dry periods (droughts) adversely affect aquatic biota. Recovery after drought is rapid by macroinvertebrates that possess strategies to survive drying or are highly mobile but other taxa take longer to recolonise depending on the timing, intensity, and duration of the dry phase.
• 2 Although drought acts as a sustained ‘ramp’ disturbance, impacts may be disproportionately severe when certain critical thresholds are exceeded. For example, ecological changes may be gradual while a riffle dries but cessation of flow causes abrupt loss of a specific habitat, alteration of physicochemical conditions in pools downstream, and fragmentation of the river ecosystem. Many ecological responses to drought within these habitats apparently depend on the timing and rapidity of hydrological transitions across these thresholds, exhibiting a ‘stepped’ response alternating between gradual change while a threshold is approached followed by a swift transition when a habitat disappears or is fragmented.
• 3 In two Australian intermittent streams, drought conditions eliminated or decimated several groups of macroinvertebrates, including atyid shrimps, stoneflies and free‐living caddisflies. These taxa persisted during the early stages of the drought but did not recruit successfully the following year, despite a return to higher‐than‐baseflow conditions. This ‘lag effect’ in response to drought emphasises the value of long‐term survey data. Although changes in faunal composition were inconsistent among sites, marked shifts in taxa richness, abundance and trophic organisation after the riffle habitat dried provide evidence for a stepped response.
• 4 Responses by macroinvertebrate assemblages to droughts of differing severity in English chalk streams were variable. The prolonged 1988–92 drought had a greater impact than shorter droughts in the early 1970s but recovery over the next 3 years was swift. Effects of the 1995 summer drought were buffered by sustained groundwater discharge from the previous winter. These droughts tended to reduce available riverine habitats, especially via siltation, but few taxa were eliminated because they could recolonise from perennial sections of the chalk streams.
• 5 In the contrasting environments of the intermittent streams studied in England and Australia, there are parallels in the rapid rates of recolonisation. However, recruitment by taxa that lack desiccation‐resistant stages or have limited mobility is delayed. Currently, long‐term data on these systems may be insufficient to indicate persistent effects of droughts or predict the impacts of excessive surface or groundwater abstraction or the increased frequency and duration of droughts expected with global climate change.

     

The effect of periphyton stoichiometry and light on biological phosphorus immobilization and release in streams

Periphyton stoichiometry can vary substantially as a result of differences in stream nutrient availability. A decrease in the periphyton carbon to phosphorus (C:P) ratio should decrease the demand for new P to be immobilized from stream water, but no studies to our knowledge have explored the relationship between periphyton stoichiometry and net P immobilization and release by periphyton. We sought to model biological P immobilization and release (flux) in streams by measuring periphyton stoichiometry and light availability. We measured P flux to and from intact periphyton on stream cobbles (20–100 mm diameter) in 1 L microcosms incubated with streamwater under variable light conditions. Net P immobilization occurred in 75% of microcosms, net P release occurred in only 5% of microcosms, and 20% of microcosms had neither net immobilization nor net release. When normalized to stream conditions, net P immobilization was highest when light availability was high (<60% canopy attenuation) and the periphyton C:P ratio was also high. In contrast, net P release occurred only when light availability was low (>60% canopy attenuation) and the periphyton C:P ratio was also low. A multiple regression model that included both periphyton stoichiometry and light availability from the growing season only, and the interaction term of these two variables, explained 99% of the variation in daily periphyton P flux observed in the study. These results indicate that in order to predict periphyton P immobilization, periphyton stoichiometry and light availability should be considered together. Furthermore, the results indicate that net P immobilization occurs even in very P-rich periphyton, which can act as a P sink when light availability is high.     

Land use and the ecology of benthic macroinvertebrate assemblages of high-altitude rainforest streams in Uganda

1. In sub‐Saharan Africa, tropical forests are increasingly threatened by accelerating rates of forest conversion and degradation. In East Africa, the larger tracts of intact rainforest lie largely in protected areas surrounded by converted landscape. Thus, there is critical need to understand the functional links between large‐scale land use and changes in river conditions, and the implications of park boundaries on catchment integrity. 2. The objective of this study was to use the mosaic of heavily converted land and pristine forest created by the protection of the high‐altitude rainforest in Bwindi Impenetrable National Park, Uganda to explore effects of deforestation on aquatic systems and the value of forest in buffering effects of adjacent land conversion. A set of 16 sites was selected over four drainages to include four categories of deforestation: agricultural land, deforested upstream (of the park boundary), forest edge (park boundary) and forest. We predicted that forest buffer (downstream or on the edge) would moderate effects of deforestation. To address this prediction, we quantified relationships between disturbance level and both physicochemical characters and traits of the macroinvertebrate assemblages during six sampling periods (February 2003 and June 2004). 3. Results of both principal components analysis and cluster analyses indicated differences in limnological variables among deforestation categories. PC1 described a gradient from deforested sites with poor water quality to pristine forested sites with relatively good water quality. Agricultural sites and deforested upstream sites generally had the highest turbidity, total dissolved solids (TDS), and conductivity values and low transparency values. Forest sites and boundary site groups generally exhibited low turbidity, TDS, and conductivity values and high water transparency values. Sites also clustered according to deforestation categories; forest and forested edge sites formed a cluster independent of both agricultural land and deforested‐upstream. 4. Water transparency, water temperature, and pH were the most important factors predicting benthic macroinvertebrate assemblages. Sensitive invertebrate families of Trichoptera, Ephemeroptera, Plecoptera, and Odonata dominated forested sites with high water transparency, low water temperature, and low pH while the tolerant families of Ephemeroptera, Diptera, Hemiptera, and Coleoptera were abundant in agriculturally impacted sites with low water transparency, high water temperature, and high pH. 5. This study provides support for the importance of riparian buffers in moderating effects of deforestation. Forest and forested edge sites were more similar in both limnological and macroinvertebrate assemblage structure than sites within or downstream from agricultural lands. If the protected area cannot encompass the catchment, the use of rivers as park boundaries may help to maintain the biological integrity of the rivers by buffering one side of the watercourse.     

Similar breakdown rates and benthic macroinvertebrate assemblages on native and Eucalyptus globulus leaf litter in Californian streams

1. Eucalyptus globulus, a tree species planted worldwide in many riparian zones, has been reported to affect benthic macroinvertebrates negatively. Although there is no consensus about the effects of Eucalyptus on aquatic macrobenthos, its removal is sometimes proposed as a means of ecological restoration. 2. We combined the sampling of macroinvertebrates with measurement of the colonisation of leaf packs in mesh bags, to examine the effects of riparian Eucalyptus and its litter on benthic macroinvertebrates in three small streams in California, U.S.A. Each stream included one reach bordered by Eucalyptus (E‐site) and a second bordered by native vegetation (N‐site). 3. The macrobenthos was sampled and two sets of litter bags were deployed at each site: one set with Eucalyptus litter (Euc‐bags) and one with mixed native tree litter (Nat‐bags) containing Quercus, Umbellularia, Acer and Alnus. Bags were exposed for 28, 56 and 90 days and this experiment was repeated in the autumn, winter and spring to account for effects of changing stream flow and insect phenology. 4. Litter input (average dry mass: 950 g m^?2 year^?1 in E‐sites versus 669 g m^?2 year^?1 in N‐sites) was similar, although in‐stream litter composition differed between E‐ and N‐sites. Litter broke down at similar rates in Euc‐bags and Nat‐bags (0.0193 day^?1 versus 0.0134 day^?1), perhaps reflecting the refractory nature of some of the leaves of the native trees (Quercus agrifolia). 5. Summary metrics for macroinvertebrates (taxon richness, Shannon diversity, pollution tolerance index) did not differ significantly between the E and N sites, or between Euc‐bags and Nat‐bags. No effect of exposure time or site was detected by ordination of the taxa sampled. However, distinct seasonal ordination clusters were observed in winter, spring and autumn, and one of the three streams formed a separate cluster. 6. The presence of Eucalyptus was less important in explaining the taxonomic composition of the macrobenthos than either ‘season’ or ‘stream’. Similarly, these same two factors (but not litter species) also helped explain the variation in leaf breakdown. We conclude that patches of riparian Eucalyptus and its litter have little effect on stream macrobenthos in this region.     

Spatial variability in macroinvertebrate assemblages along and among neighbouring equatorial glacier‐fed streams

1. During the past two decades, understanding of the structure and function of glacier‐fed stream ecosystems at temperate latitudes has increased substantially. In contrast, information on their tropical counterparts is very limited. We studied three neighbouring glacier‐fed streams in the tropical Andes of Ecuador. Our main goals were (i) to determine overall longitudinal patterns in density, taxon richness and the composition of macroinvertebrate assemblages and driving factors in equatorial glacial streams and (ii) to examine variability among replicate streams in faunal metrics and assemblages, and stream‐specific effects of supposed environmental key factors. 2. We measured four geographical and 17 environmental factors and collected five Surber samples (500 cm²) of macroinvertebrates at each of nine sites, three sites along three streams. The streams were located 1–5 km apart. In each stream, the three sites were placed at comparable distances from the glacier and were grouped as ‘upper’ (50–200 m), ‘middle’ (1.5 km) and ‘lower’ sites (3.5–5.6 km). 3. In total, 2200 individuals (64% chironomids) were collected and 47 taxa (30 dipterans, 18 of these Chironomidae) identified. Density ranged from 176 to 372 ind. m^?2, and the number of taxa ranged from 2 to 6 at the upper sites and 868–3044 ind. m^?2 and 21–27 taxa at the lower sites. Density, number of taxa, rarefied richness and axis‐1 coordinates from a MDS ordination increased logarithmically with distance from the glacier. These faunal metrics were equally related to altitude and glacier per cent of catchment and correlated with maximum conductivity, mean temperature, mean daily maximum temperature and a channel stability index. As expected, the mean difference in distance decay in similarity was higher at the upper (47% km^?1) than at the lower reaches (20% km^?1) of the streams. 4. The number of taxa varied among sites within the upper and middle groups, but not among the lower sites. In contrast, but in accordance with our expectation, assemblage composition did not differ among upper sites but did so at middle and lower sites, following a supposed decrease in environmental harshness along the streams. Relationships between faunal metrics and the four environmental variables mean temperature, the stability index, chlorophyll a and coarse particulate organic matter also varied among the three streams. Generalised linear model analyses revealed that temperature interacted with stream on macroinvertebrate density, while chlorophyll a had a significant effect on the number of taxa in interaction with stream and stability. 5. The basic predictions of the Milner et al. (2001a) , model regarding longitudinal faunal patterns and temperature and stability as main driving factors were met by our three replicate equatorial glacial streams. Qualitative departures from the model were mainly because of zoogeographical differences. We demonstrated that variability in assemblages between comparable sites in closely situated streams was considerable, and the effect of key environmental factors varied among streams and interacted with other factors. Quantifying spatial variation in benthic assemblages may help us foresee possible consequences for biodiversity as a result of glacial retreat.