Predator effects on prey population dynamics in open systems

Animal population dynamics in open systems are affected not only by agents of mortality and the influence of species interactions on behavior and life histories, but also by dispersal and recruitment. We used an extensive data set to compare natural loss rates of two mayfly species that co-occur in high-elevation streams varying in predation risk, and experience different abiotic conditions during larval development. Our goals were to generate hypotheses relating predation to variation in prey population dynamics and to evaluate alternative mechanisms to explain such variation. While neither loss rates nor abundance of the species that develops during snowmelt (Baetis bicaudatus) varied systematically with fish, loss rates of the species that develops during baseflow (Baetis B) were higher in streams containing brook trout than streams without fish; and surprisingly, larvae of this species were most abundant in trout streams. This counter-intuitive pattern could not be explained by a trophic cascade, because densities of intermediate predators (stoneflies) did not differ between fish and fishless streams and predation by trout on stoneflies was negligible. A statistical model estimated that higher recruitment and accelerated development enables Baetis B to maintain larger populations in trout streams despite higher mortality from predation. Experimental estimates suggested that predation by trout potentially accounts for natural losses of Baetis B, but not Baetis bicaudatus. Predation by stoneflies on Baetis is negligible in fish streams, but could make an important contribution to observed losses of both species in fishless streams. Non-predatory sources of loss were higher for B. bicaudatus in trout streams, and for Baetis B in fishless streams. We conclude that predation alone cannot explain variation in population dynamics of either species; and the relative importance of predation is species- and environment-specific compared to non-predatory losses, such as other agents of mortality and non-consumptive effects of predators. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Observations on some predators of stream-dwelling triclads

A serological technique was used to demonstrate that one species of trichopteran and three species of plecopteran nymphs were feeding on stream-dwelling triclads. Observations on streams in Caernarvonshire, North Wales, suggested that, in general, these predators were not typical members of the fauna at springheads, where triclads were frequently abundant. In contrast, further downstream the presence of predators was considered to be one of a number of factors which contributed to lower population densities of triclads. However, two small stony streams were encountered where the plecopteran predator Dinocras cephalotes occurred in close proximity to stream-dwelling triclads from the springhead downwards. In one, both predators and prey were found at all stations but highest triclad numbers were recorded where predators were few. In the second stream the population density of the predator was higher, and here, triclads were very scarce on the undersides of stones and were almost entirely confined o t patches of moss. It appears that the predators may be both limiting the micro-distribution of the triclad and also perhaps feeding selectively on larger individuals since mature triclads were rare, except at the springhead.     

Composition of Endemic Fish Assemblages in Relation to Exotic Species and River Regulation in a Temperate Stream

Spatial variation in the fish community of the regulated Raia stream (Portugal) was examined during the summer of 1995. Variation in the native fish community (abundance of species-size combinations) was explicitly related to both the abundance of exotic species and habitat variables. The fish community changed in space from assemblages characterised by the high relative abundance of Leuciscus pyrenaicus, Leuciscus alburnoides and the smaller size classes of Barbus bocagei (all Iberian endemics with total length, TL, < 100mm to assemblages characterised by the high relative abundance of Chondrostoma polylepis (Iberian endemic) and B. bocagei with TL > 200mm and of exotics Micropterus salmoides and Lepomis gibbosus (> 100mm TL). The former assemblages used shallow sites far from downstream dams with some current whereas the latter assemblages used deeper sites closer to downstream dams without current velocity and with abundant floating macrophytes. Both exotic species and habitat variables were significant correlates of endemic assemblage composition in the Raia stream and the abundance of M. salmoides > 150mm in TL was the best biotic predictor of endemic assemblage composition. The total variation in the community of endemic fish was partitioned into four components: (i) associated uniquely with exotic species - 12.6%, (ii) associated uniquely with habitat variables - 27.6%, (iii) associated both with exotic species and habitat variables - 14.5%, and (iv) that unexplained - 45.3%. A significant association of exotic species with the endemic fish community remained after accounting for the selected environmental variables and a strong (habitat) x (exotic species) interaction was indicated.     

Cascading effects of predatory fish on the composition of benthic algae in high‐altitude streams

Cascading effects of predators can affect ecosystem properties by changing plant biomass, distribution and assemblage composition. Using data from field surveys and whole‐stream experiments we tested the hypothesis that predatory trout change assemblage composition of benthic algae in high‐elevation streams mediated by grazer behavior. Field surveys revealed that the taxonomic composition of algal assemblages differed significantly between streams that contained trout and those that were fishless; but comparisons of palatable versus unpalatable algal taxa between fish and fishless streams were equivocal because of high natural variability. Therefore, we tested for a behavioral (non‐consumptive) trophic cascade experimentally by adding brook trout chemical cues to six naturally fishless streams for 25 days and compared responses of grazers and algae to six reference streams without fish cues added. Algal response variables included rates of change in the abundance of three physiognomic categories, from most palatable (attached erect and prostrate diatoms) to least palatable (non‐diatoms), as determined from food selectivity analyses of the most common grazers (mayflies and caddisflies). Fish cues did not affect the mean densities or changes in densities of total grazers or any individual grazer species. However, in streams where fish cues were added, rates of accrual of attached erect diatoms, which was the preferred algal type for the grazer most vulnerable to trout predation (Baetis), were higher and their densities increased significantly faster with increasing densities of this grazer species than in reference streams. Results of his experiment support the hypothesis that predator induced suppression of grazer foraging behavior, rather than cascading effects of top predators on grazer density, may contribute to variation in the composition of algal assemblages among streams by allowing proliferation of most palatable algal species.     

Differential effects of largemouth and smallmouth bass on habitat use by stoneroller minnows in stream pools

Previous investigations of within-reach distribution patterns of the stoneroller minnow, Campostoma anomalum Rafinesque, suggest that this species behavioural response to piscivorous bass differs among streams with different bass species. This study compares the responses of Camposloma to two common piscivorous fishes, largemouth bass, Micropterus salmoides Lacepede, and smallmouth bass, M. dolomieui Lacepede. Field experiments in Brier Creek, Oklahoma, U.S.A. showed that Campostoma responded to the presence of adult largemouth bass by shifting to shallow water habitats (commonly < 20 cm), but responded weakly or not at all to adult smallmouth bass. The magnitude of the response of Campostoma was positively related to activity level of the predators. Differences in behaviour between these two predator species may contribute to the differences in spatial distribution patterns of Micropterus and Campostoma observed in earlier investigations, but differences in stream channel morphology and temperature regimes among streams may also be important.     

Helminth communities of native and introduced fishes in Lake Pátzcuaro, Michoacán, México

In Lake Pátzcuaro in the Mesa Central of México, a total of 19 species of helminths was found in 598 fishes and comprised five digeneans, two monogeneans, four cestodes, one acanthocephalan and seven nematodes, of which ten species were represented by larval or immature states. The richest and most diverse helminth communities were found in the native carnivorous goodeid Alloophorus robustus. In general, the helminth communities in the different fish species were not particularly species rich and the parasite assemblages were numerically dominated by larvae of the bird trematode, Posthodiplostomum minimum. Patterns of helminth community richness and diversity were similar to those previously observed in north-temperate freshwater fishes. Most enteric helminths occurred with low abundance and only a small proportion of the gut helminth communities was numerically dominated by any one species. Helminths dominating their enteric communities showed some level of host specificity. Helminth communities in carnivorous fish species were generally richer than those in herbivores and detritivores, with the exception of the predominantly herbivorous Goodea atripinnis. The helminth fauna of introduced fishes, Cyprinus carpio, Micropterus salmoides and Oreochromis niloticus , consisted of either few or no host-specific adult helminth(s) translocated from their original geographical areas and by larval stages of helminths of piscivorous birds. Based on the geological history of the area and the biogeography of the endemic fish fauna, it is hypothesized that host-switching and relationships with the nearctic fauna have been fundamental in determining the helminth fauna of the endemic fish hosts.     

Life-history traits of invasive fish in small Mediterranean streams

We compared the life-history traits of native and invasive fish species from Catalan streams in order to identify the characters of successful invasive fish species. Most of the exotic fish species were characterized by large size, long longevity, late maturity, high fecundity, few spawnings per year, and short reproductive span, whereas Iberian native species exhibited predominantly the opposite suite of traits. Species native to the southeastern Pyrenees watershed were also significantly different from species native to the rest of the Iberian Peninsula but not native to this watershed. Iberian exotic species come predominantly from large river basins, whereas Catalan streams (and other small, coastal river basins) correspond to basins and streams of a smaller size and different hydrology, with differences in species composition and life-history traits of fish. The occurrence and spread of invasive species was not significantly related to life-history traits but to introduction date. The successful prediction of future invasive species is limited due to small differences in life-history and ecological traits between native and exotic species. Fecundity, age at maturity, water quality flexibility, tolerance to pollution and habitat seem the most discriminating life-history variables. This revised version was published online in July 2006 with corrections to the Cover Date.     

The relative influences of exotic species and environmental factors

Spatial variation in the native fish community of the lower Guadiana basin (southern Iberia) was related to both biotic (abundance of exotic species) and environmental factors using canonical correspondence analysis, CCA. After choosing the best predictors among the environmental and biotic variables, we partitioned the total variation in native species abundances (18 species-size combinations by 44 sampling locations matrix) into that accounted for (1) solely by selected environmental variables, (2) solely by selected biotic variables, (3) by both environmental and biotic variables together, and (4) that unexplained. Of the 22 variables initially considered in the environmental CCA, only RIVER (dummy variable codifying sampling locations belonging to the main river), SALT (dummy variable codifying sampling locations belonging to tributaries that discharge to the brackish Guadiana), substrate heterogeneity, and macrobenthos abundance were selected as best predictors. In the biotic CCA, three species-size combinations were selected from the eight considered: Micropterus salmoides > 150 mm in total length (TL), Lepomis gibbosus > 100 mm in TL and Gambusia holbrooki. The total variation in the native fish community was partitioned into the following components: pure environmental (24.4%), pure biotic (12.4%), shared (9.2%) and unexplained (54.0%). There was a significant influence of exotic fish on native community variation after accounting for the effect of environmental factors. This revised version was published online in July 2006 with corrections to the Cover Date.     

Do predators limit the abundance of alternative prey? Experiments with vole‐eating avian and mammalian predators

Predation has been invoked as a factor synchronizing the population oscillations of sympatric prey species, either because predators kill prey unselectively (the Shared Predation Hypothesis; hereafter SPH), or because predators switch to alternative prey after a density decline in their main prey (the Alternative Prey Hypothesis; APH). A basic assumption of the APH is that the impact of predators on alternative prey depends more on the density of main prey than on the predator/alternative prey ratio. Both SPH and APH assume that the impact of predators on alternative prey is at least periodically strong enough to depress prey populations. To examine these assumptions, we utilized data from replicated field experiments in large areas where we reduced the breeding densities of avian predators during three years and the numbers of least weasels (Mustela nivalis) in two years when vole populations declined. In addition, we reduced the breeding densities of avian predators in two years when vole populations were high. The reduction of least weasels increased the abundance of their alternative prey, small birds breeding on the ground, but did not affect the abundance of common shrews (Sorex araneus). In years when vole populations declined, the reduction of avian predators increased the abundance of their alternative prey, common shrews and small birds. Therefore, vole‐eating predators do at least periodically depress the abundance of their alternative prey. At high vole densities, the reduction of avian predators did not increase the abundance of common shrews, although the ratio of avian predators to alternative prey was similar to years when vole populations declined, which supported APH. In contrast, the abundance of small birds increased after the reduction of avian predators also at high vole densities, which supported SPH. The manipulations had no obvious effect on the number of game birds, which are only occasionally killed by these small‐sized predators. We conclude that in communities where most predators are small or specialize on a single prey type, the synchronizing impact of predation is restricted to a few similar‐sized species.     

Seasonal changes in fish community structure of intermittent streams in the middle reaches of the Guadiana basin, Portugal

Seven sites on three tributaries with intermittent flow regimes in the middle reaches of the Guadiana, i.e. Xévora, Caia and Degebe, were sampled approximately every 3 months from April 1995 to April 1997. The density and biomass of fish were greater in downstream rivers and sites. Along the Xévora River, a series of point abundance samples showed evidence of fish zonation along the river. Canonical correspondence analysis (CCA) suggested a simple relationship between the different species and their preferred habitat. The small-sized Rutilus alburnoides and Leuciscus pyrenaicus used most or all of the habitats with some cover and flowing water. Greater concentrations of larger fish were found in downstream reaches which generally had greater depths as stream width increased: large Barbus microcephalus , B. steindachneri and B. comiza preferred such sites on the Degebe River. Droughts in seasonal Mediterranean streams such as the Guadiana River are a predictable natural disturbance. In the summer, flow ceases and some stretches of river consist of isolated pools. This results in intense aggregations of fish and possible competition for food and/or space. Although some species are well adapted to natural droughts, habitat degradation and possibly the introduction of exotic species contribute to marked variability in species composition.     

Fish faunal ‘breaks’ and stream order in the eastern and central United States

Synopsis The distribution of fishes within 23 streams in the eastern and central United States was analyzed for evidence of distinct longitudinal faunal breaks, and to compare location of the breaks to stream order and longitudinal position of stations within watersheds. On the basis of presence-absence data for fish species, distinct faunal breaks were found in only 8 of the 23 streams. Qualitative faunal breaks were not significantly more common in any particular reach of streams, and the breaks did not correspond to changes in stream order. In 11 of 14 streams for which data on fish abundance were available, quantitative faunal breaks were found, but the locations of these breaks were unrelated to stream order or longitudinal position of stations in streams. Earlier studies have postulated that stream orders may represent distinct biological units for fishes. In contrast, this analysis shows that local fish faunas differ as much within as between adjacent stream orders, and suggests that stream order alone is not a pervasive organizer of lotic fish communities.     

Organization patterns of headwater-stream fish communities in the Upper Paraguay–Paraná basins

We compared environmental influences on the assemblages of stream fishes of contiguous but biogeographically distinct parts of the Upper Paraguay and Paraná basins in Brazil, aiming to describe the main distribution patterns of fish species in the headwater streams. We analyzed bimonthly samples in 10 streams in each basin from January to November 2004. Sixty fish species were collected, including 40 species in streams of the Paraguay Basin and 42 species in streams of the Paraná Basin. The species abundance rank did not differ between the basins. We found a clear differentiation between the Paraguay and Paraná stream fish assemblages. There were significant differences in species composition among streams, but no seasonal differences in species composition. Connectivity between streams explains the differences in species composition for the Paraguay streams, but not for the Paraná streams. Hydrological characteristics were the main factors determining species distribution in both basins, suggesting that the migratory capacity of each species and biogeographical barriers may act to filter the local species composition from the total species pool of colonizers. Handling editor: S. M. Thomaz     

Predators reduce abundance and species richness of coral reef fish recruits via non-selective predation

Predators have important effects on coral reef fish populations, but their effects on community structure have only recently been investigated and are not yet well understood. Here, the effect of predation on the diversity and abundance of young coral reef fishes was experimentally examined in Moorea, French Polynesia. Effects of predators were quantified by monitoring recruitment of fishes onto standardized patch reefs in predator-exclosure cages or uncaged reefs. At the end of the 54-day experiment, recruits were 74% less abundant on reefs exposed to predators than on caged ones, and species richness was 42% lower on reefs exposed to predators. Effects of predators varied somewhat among families, however, rarefaction analysis indicated that predators foraged non-selectively among species. These results indicate that predation can alter diversity of reef fish communities by indiscriminately reducing the abundance of fishes soon after settlement, thereby reducing the number of species present on reefs.     

Effects of Flow Restoration and Exotic Species Removal on Recovery of Native Fish: Lessons from a Dam Decommissioning

Flow diversion and invasive species are two major threats to freshwater ecosystems, threats that restoration efforts attempt to redress. Yet, few restoration projects monitor whether removal of these threats improve target characteristics of the ecosystem. Fewer still have an appropriate experimental design from which causal inferences can be drawn as to the relative merits of removing exotic fish, restoring flow, or both. We used a dam decommissioning in Fossil Creek, Arizona, to compare responses of native fish to exotic fish removal and flow restoration, using a before‐after‐control‐impact design with three impact treatments: flow restoration alone where exotics had not been present, flow restoration and exotic fish removal, and flow restoration where exotics remain and a control reach that was unaffected by restoration actions. We show that removal of exotic fish dramatically increased native fish abundance. Flow restoration also increased native fish abundance, but the effect was smaller than that from removing exotics. Flow restoration had no effect where exotic fish remained, although it may have had other benefits to the ecosystem. The cost to restore flow ($12 million) was considerably higher than that to eradicate exotics ($1.1 million). The long‐term influence of flow restoration could increase, as travertine dams grow and re‐shape the creek increasing habitat for native fish. But in the 2‐year period considered here, the return on investment for extirpating exotics far exceeded that from flow restoration. Projects aimed to restore native fish by restoring flow should also consider the additional investment required to eradicate exotic fish.     

Host specificity and population structure of two exotic helminths,Camallanus cotti (Nematoda) and Bothriocephalus acheilognathi (Cestoda), parasitizing exotic fishes in Waianu Stream,O'ahu,Hawai'i

Introduction of exotic fishes into Hawai'ian streams has resulted in the simultaneous introduction of exotic parasites. Camallanus corti (Nematoda) and Bothricephalus acheilognathi (Cestoda) are the most prevalent and abundant fish helminths in Hawai'ian streams. The population structure and host specificity of C. cotti and B. acheilognathi in exotic poeciliids were examined during May and June 1995 in Waianu Stream, O'ahu, Hawai'i. Prevalence and mean abundance of C. cotti were significantly different among Poecilia reticulata (65.2%, 1.05), Poecilia mexicana (49.0%, 0.87), and Xiphophorus helleri (32.3%, 0.44). Prevalence of B. acheilognathi was significantly higher in P. mexicana (6.1%) than in P. reticulata (2.1%) and X. helleri (1.6%). However, tapeworm differences in mean abundance were not significant among P. mexicana (0.08), P. reticulata (0.04), or X. helleri (0.03). Nematode and tapeworm prevalence and mean abundance were not significantly different with regard to fish sex. Camallanus cotti prevalence and mean abundance increased as P. mexicana body size increased (r2 = 0.84 and r2 = 0.72, respectively), whereas B. acheilognathi displayed no significant trend with respect to poeciliid body size.     

Predatory fish loss affects the structure and functioning of a model marine food web

The rate of species loss is increasing on a global scale and predators are most at risk from human-induced extinction. The effects of losing predators are difficult to predict, even with experimental single species removals, because different combinations of species interact in unpredictable ways. We tested the effects of the loss of groups of common predators on herbivore and algal assemblages in a model benthic marine system. The predator groups were fish, shrimp and crabs. Each group was represented by at least two characteristic species based on data collected at local field sites. We examined the effects of the loss of predators while controlling for the loss of predator biomass. The identity, not the number of predator groups, affected herbivore abundance and assemblage structure. Removing fish led to a large increase in the abundance of dominant herbivores, such as Ampithoids and Caprellids. Predator identity also affected algal assemblage structure. It did not, however, affect total algal mass. Removing fish led to an increase in the final biomass of the least common taxa (red algae) and reduced the mass of the dominant taxa (brown algae). This compensatory shift in the algal assemblage appeared to facilitate the maintenance of a constant total algal biomass. In the absence of fish, shrimp at higher than ambient densities had a similar effect on herbivore abundance, showing that other groups could partially compensate for the loss of dominant predators. Crabs had no effect on herbivore or algal populations, possibly because they were not at carrying capacity in our experimental system. These findings show that contrary to the assumptions of many food web models, predators cannot be classified into a single functional group and their role in food webs depends on their identity and density in 'real' systems and carrying capacities.     

Effects of coalbed natural gas development on fish assemblages in tributary streams of the Powder and Tongue rivers

1. Extraction of coalbed natural gas (CBNG) often results in disposal of large quantities of CBNG product water, which may affect aquatic ecosystems. We evaluated the effects of CBNG development on fish assemblages in tributary streams of the Powder and Tongue rivers. We used treatment and control, impact versus reference sites comparisons, surveys of CBNG product‐water streams and in situ fish survival approaches to determine if CBNG development affected fish assemblages. 2. Several of our results suggested that CBNG development did not affect fish assemblages. Species richness and index of biotic integrity (IBI) scores were similar in streams with and streams without CBNG development, and overall biotic integrity was not related to the number or density of CBNG wells. Fish occurred in one stream that was composed largely or entirely of CBNG product water. Sentinel fish survived in cages at treatment sites where no or few fish were captured, suggesting that factors such as lack of stream connectivity rather than water quality limited fish abundance at these sites. Fish species richness did not differ significantly from 1994 to 2006 in comparisons of CBNG‐developed and undeveloped streams. Biotic integrity declined from 1994 to 2006; however, declines occurred at both impact and reference sites, possibly because of long‐term drought. 3. Some evidence suggested that CBNG development negatively affected fish assemblages, or may do so over time. Specific conductivity was on average higher in treatment streams and was negatively related to biotic integrity. Four IBI species richness metrics were negatively correlated with the number or density of CBNG wells in the catchment above sampling sites. Bicarbonate, one of the primary ions in product water, was significantly higher in developed streams and may have limited abundance of longnose dace (Rhinichthys cataractae). Total dissolved solids, alkalinity, magnesium and sulphate were significantly higher in developed streams. 4. Biological monitoring conducted before the development of CBNG, and continuing through the life of development and reclamation, together with data on the quantity, quality and fate of CBNG product water will allow robust assessment of potential effects of future CBNG development worldwide.     

Food webs in tropical Australian streams: shredders are not scarce

1. Macroinvertebrates were collected in dry and wet seasons from riffles and pools in two streams in tropical north Queensland. Total biomass, abundance and species richness were higher in riffles than in pools but did not differ between streams or seasons. 2. Gut contents of all species were identified. Cluster analysis based on gut contents identified five dietary groups: I, generalist collectors; II, generalist shredders and generalist predators; III, generalist scrapers; IV, specialist shredders; and V, specialist predators. Species were allocated to functional feeding groups (FFGs) based on these dietary groups. 3. Many species were generalist in their diets, but specialist predators and shredders were particularly prominent components of the invertebrate assemblages in terms of biomass and species richness. 4. Community composition (proportions of biomass, abundance and species richness of the different FFGs) varied between habitat types, but not between streams or seasons, although differences between riffles and pools varied with season. 5. Comparison of the fauna of 20 streams showed that our study sites were similar to, or not atypical of, low‐order streams in the Queensland wet tropics.     

Spatial variation in relative abundance of a widespread, numerically dominant fish species and its effect on fish assemblage structure

Collections of fish assemblages from streams in the midwestern United States were used to examine assemblage-level effects of spatial variation in relative abundance of the red shiner, Cyprinella lutrensis, a widespread and highly abundant minnow species. This species has been widely introduced outside its native range and is suspected to have impacted local assemblages where it has become established. Given its overall dominance of midwest fish assemblages, and its suspected impact on assemblage structure, we asked if structure of the residual fish assemblages (red shiners excluded) was a function of the relative abundance of red shiners throughout the native range of C. lutrensis in the USA. Although red shiner ranked first in abundance in half of the assemblages and numerically dominated 28% of the assemblages, red shiner relative abundance in an assemblage had no detectable effect on richness, diversity, evenness, or complexity of other (residual) species in the assemblage. Relative abundance of red shiners did have a positive effect on the abundance of benthic minnows in the residual assemblage, but not on water column minnows that are ecologically most like red shiners. Environmental factors did not explain a significant amount of the variation in relative abundance of red shiners, but did explain some variation in residual assemblage structure. Although widespread and numerically dominant at many localities, red shiners do not appear to have a strong impact on local fish assemblage structure within their native range. This is in sharp contrast to the reported negative effects of red shiners on fish assemblages where they have been introduced outside their native range.     

Resource partitioning among the fishes of rainforest streams in Sri Lanka

The fish assemblage typical of small rainforest streams of Sri Lanka was investigated to see if the fishes were characterized by a high degree of specialization, expected of equilibrium communities, or if they were relatively unspecialized, expected of more stochastic (non-equilibrium) communities. Morphological features, habitats, microhabitats and diets of the 20 most abundant fish species were measured. The fishes included 11 Cyprinidae (including seven Barbus spp. and two Rasbara spp.), two Gobiidae, two Channidae, and one species each in the Cobitidae, Cyprinodontidae, Belontidae, Belonidae, and Mastacembelidae. Microhabitats were determined by measuring (for 6760 fish) water column depth, distance of fish from bottom, mean water column velocity, water velocity at fish, and substrate. Morphologically, the species showed a high degree of specialization, especially in structures related to feeding. Although several species were habitat generalists, most species occurred in distinct habitats. Within habitats, microhabitat overlap among co-occuring species was low, particularly in relation to position in the water column. Principal component analysis of the microhabitat measurements produced three new variables. Most species not clearly segregated by the overlap analysis showed segregation on the new variables. Fishes not segregated by habitat or microhabitat tended to show low dietary overlaps. Specialization in feeding habits was more pronounced among the Sri Lankan fishes than noted for fishes in small rainforest streams elsewhere, in part because of the greater reliance of the Sri Lanka fishes on autochthonous foods. Overall, the fish assemblage had the characteristics expected of an equilibrium (deterministic) assemblage.