Multi-biochemical responses of benthic macroinvertebrate species as a complementary tool to diagnose the cause of community impairment in polluted rivers

Biological indexes, based on benthic macroinvertebrate taxa, are currently used worldwide to measure river ecological quality. These indexes assign a global ecological status of the biotic community, but not necessarily may detect specific effects of water pollutants. Conversely a large set of biochemical markers measured in macroinvertebrate benthic species can detect sublethal effects and inform us about additional environmental factors that are impairing benthic communities. This is especially interesting in moderately polluted sites, where other stressors are already affecting communities but not too strongly to be detected by biotic indexes. Up to ten different markers belonging to distinct metabolic paths and 42 contaminants measured in sample collections of the caddis fly Hydropsyche exocellata were assessed across a polluted gradient in the industrialized Mediterranean River basins of Besós and Llobregat (NE, Spain). Twenty four sample collections were selected to include macroinvertebrate communities representing the five impairment degrees defined by the Spanish Environmental authorities using the biotic metrics. Results evidenced a clear deterioration of the ecological water quality parameters and benthic communities towards downstream reaches. Biochemical responses varied significantly across the studied samples and were able to differentiate samples within communities having a good and deteriorated ecological stage. Principal Component Analyses indicated that salinity was one of the major stresses affecting macroinvertebrate assemblages, whereas antioxidant and metabolizing enzymes responded differently and were closely related to high and presumably toxic levels of accumulated organic pollutants. Therefore these results indicate that the use of multiple -markers sensitive to water pollution may provide complementary information to diagnose environmental factors that are impairing macroinvertebrate communities.     

Linking catchment characteristics and water chemistry with the ecological status of Irish rivers

Requirements of the EU Water Framework Directive for the introduction of ecological quality objectives for surface waters and the stipulation that all surface waters in the EU must be of 'good' ecological status by 2015 necessitate a quantitative understanding of the linkages among catchment attributes, water chemistry and the ecological status of aquatic ecosystems. Analysis of lotic ecological status, as indicated by an established biotic index based primarily on benthic macroinvertebrate community structure, of 797 hydrologically independent river sites located throughout Ireland showed highly significant inverse associations between the ecological status of rivers and measures of catchment urbanisation and agricultural intensity, densities of humans and cattle and chemical indicators of water quality. Stepwise logistic regression suggested that urbanisation, arable farming and extent of pasturelands are the principal factors impacting on the ecological status of streams and rivers in Ireland and that the likelihood of a river site complying with the demands of the EU Water Framework Directive, and be of 'good' ecological status, can be predicted with reasonable accuracy using simple models that utilise either widely available landcover data or chemical monitoring data. Non-linear landcover and chemical 'thresholds' derived from these models provide a useful tool in the management of risk in catchments, and suggest strongly that more careful planning of land use in Ireland is essential in order to restore and maintain water quality as required by the Directive.     

Heavy metal bioaccumulation and macroinvertebrate community changes in a Mediterranean stream affected by acid mine drainage and an accidental spill (Guadiamar River, SW Spain)

In the 1998 Aznalcóllar mining accident, 5 million cubic meters of toxic waste were spilled into the Guadiamar River (SW Spain). This paper describes the likely effects of metal pollution on the macroinvertebrate community in the first 15 km of the river adjacent to the mine, 2 years after the spill. The contents of Zn, Cu, Pb, As, Cd, Tl, and Sb in water, sediment, and in the caddisfly Hydropsyche were analysed, together with some physico-chemical parameters. From the results of physico-chemical and community parameters, the studied area was divided into three sections: the first section contained upstream control stations, not affected by current mining activities, and with a diverse macroinvertebrate community structure (mean number of families = 19); the second section, close to the mine, was severely polluted, with more than 50 mg l(-1) of Zn or 1.6 mg l(-1) of Cu in water, a pH of nearly 4, and a very diminished macroinvertebrate community (three families); in the last section, 6-15 km downstream from the mine, the water quality improved, metal concentrations decreased, pH reached neutrality and the macroinvertebrate community comprised eight families. We also determined metal concentrations in the larvae of Hydropsyche and found increases of 3- to 35-fold (for Cu and Cd, respectively) compared to control sites. As far as we are aware, these values are the highest concentrations so far detected in this caddisfly anywhere in the world. Different patterns of metal and pH tolerance were identified in macroinvertebrates. Most of the species were intolerant to present pollution levels and were only present in control sites. Others were tolerant to metal pollution but not to low pH and were present in the main river downstream of the mine, while a very few species managed to live in the most polluted section, with very low pH and high metal concentrations.     

The validity of the Gammarus:Asellus ratio as an index of organic pollution: abiotic and biotic influences.

In freshwaters. Gammarus spp. are more sensitive to organic pollution than Asellus spp. and the relative abundance of the two taxa has been proposed as a pollution index. We tested the validity of this by examining the relationship between the Gammarus: Asellus (G : A) ratio and (1) a suite of physico-chemical variables. (2) established biotic (average score per taxon, ASPT) and richness (species richness (S) and Ephemeroptera, Plecoptera and Trichoptera families richness (EPT family richness)) indices generated from the macroinvertebrate community. In addition, we investigated a suspected biotic interaction, predation, between Gammarus and Asellus. Both univariate and multivariate analyses showed that the G: A ratio was sometimes responsive to changes in parameters linked to organic pollution, such as BOD5 and nitrate levels. However, the G : A ratio also appeared responsive to variables not directly linked to organic pollution, such as conductivity and distance from source. There were significant positive correlations among the G : A ratio and the ASPT, S and EPT, indicating that changes in the relative abundances of Gammarus and Asellus were reflected in changes in the pollution sensitivity and richness of the wider macroinvertebrate community. A laboratory experiment revealed significant predation of Asellus aquaticus juveniles by Gammarus duebeni celticus adults, but no reciprocal predation. We propose that the G: A ratio may be useful as a crude measure of organic pollution that could supplement more complex indices in a multimetric approach to pollution monitoring or be used for monitoring individual sites, where a simple technique is required for monitoring purposes over a period of time. Also, we urge recognition of the possible role of biotic interactions among taxa used in the generation of pollution indices.     

A comparison between various criteria for the interpretation of biological data in the analysis of the quality of running waters

In this work it is proposed to apply the most common biological criteria for the assessment of water quality to the macroinvertebrate populations in the Torrente Parma. The data concerning both the structure and the dynamics of the macroinvertebrate community over 1 yr have been analysed through the species diversity indexes, the saprobien index and the biotic indexes of Woodiviss and Verneaux-Tuffery. The values obtained by applying the various indexes have been compared. A tentative conversion scale has been developed between the indexes which have given the highest degree of correlation.Furthermore, a synthetic representation of the annual species distribution has been obtained using cluster analysis and Fager's affinity index.     

Preliminary observations on the benthic macroinvertebrate communities of Ross Bay,a polluted area of lough leane,south-west Ireland

The physical, chemical and biological water quality characteristics of Ross Bay, monitored during the period July 1974–May 1975, are described. The benthic macroinvertebrate standing crop of the littoral zone was sampled quantitatively during July 1974. The results are described in terms of biomass distribution and the distribution of taxa. The sampling stations are placed in an ordinal series based on their respective pollutional status. Changes in community structure and taxonomic composition along the series are described in terms of species diversity and the distribution of “indicator” organisms. Based on this distribution, a scheme of biotic indices is devised in order to reduce complex community composition to simple numerical terms. For this purpose, individual organisms are given a water quality rating which is dependent on their relative tolerance or intolerance to pollution. Significant correlation is shown between the ordinal series, water quality, Diversity Index ($\text{d}\text{?}$) values, Community Similarity (τ) values and Biotic Quality (qb) values. The biological evaluation of water quality in lakes is discussed and it is suggested that the littoral benthos may serve in any future assessment of organic pollution and enrichment in Lough Leane.     

Use of the aquatic protozoa to formulate a community biotic index for an urban water system

Protozoan were collected from 16 stations in water system of Changde City (China) using the PFU method. Sampling programs were conduced on a yearly basis, with seasonal frequency at diverse sites in the water system and 488 species of protozoa was identified. At the same time, Water sampling from these stations was conducted and various water chemical parameters, including DO, COD, BOD(5), NH(3), TP, and Volatile Phenol, were analyzed. The aim of the research was, on one hand, using chemical method to take an investigation to the water pollution status of Changde City; on the other hand, using protozoan to make an evaluation to the water quality. With the chemical water parameters and protozoa data, a biotic index was derived for the investigated region. The species pollution value (SPV) of 469 protozoa species was established, and the community pollution value (CPV) calculated from SPV was used to evaluate water quality. The method of the biotic index was tested and the result showed that CPV calculated from SPV had a close correlation with the degree of water pollution (p<0.00001). This indicated that the method of the biotic index is reliable. The water quality degrees divided by CPV were suggested.     

Dynamics of benthic macroinvertebrates in created wetlands receiving wastewater

This paper presents results of a study on the status and dynamics of benthic macroinvertebrates in a wetland constructed for wastewater treatment in Nairobi (Kenya), conducted for 16 weeks during the period between December 1995 and March 1996. Seventeen species of macroinvertebrates were recorded in the wetland. Species richness in the wetland was 9.6 ± 1.3 while species diversity index was 0.466. Mean pH and dissolved oxygen increased gradually from the intake point to the outlet while the BOD₅, suspended solids and free and saline ammonia decreased as the wastewater flowed throughout the wetland. The benthic macroinvertebrate community was poorly developed while microbial community appeared to be well developed. It was evident that the community structure improved with the purification of wastewater along the wetland. As discussed in the paper, aquatic benthos are crucial to wetlands ecosystems function and they are potential indicators of health of these ecosystems.     

Using silicone passive samplers to detect polycyclic aromatic hydrocarbons from wildfires in streams and potential acute effects for invertebrate communities

Silicone rubber passive samplers spiked with 4 deuterated performance reference compounds were deployed for 29-33 days to estimate the concentrations of 16 polycyclic aromatic hydrocarbons (PAHs) in 9 streams in Victoria, Australia, following a wildfire. Silicone rubber strips of 2 thicknesses were used to obtain information on the status of uptake of the chemicals of interest at retrieval. In addition, we monitored the stream macroinvertebrate community for potential effects of PAHs or other fire organics. All selected PAHs were detected in the passive samplers and the sampling rates ranged from 0.5 to 50 L/day significantly varying between sites but not compounds, presumably due to differences in current velocity. The estimated water concentrations were 0.1-10 ng/L for total PAHs with phenanthrene, pyrene and fluoranthene accounting for 91% of the total concentration. All PAHs were a factor of 1000 or more below the reported 48-h median lethal concentrations (48-h LC50) for Daphnia magna. Two sites located closest to the fires exhibited elevated concentrations compared to the other sites and the passive samplers in these sites remained in the integrative uptake regime for all compounds, suggesting precipitation-associated PAH input. No acute toxic effects of PAHs or other fire organics on the invertebrate community were detected using a biotic index for organic toxicants (SPEAR), whereas a non-specific biotic index (SIGNAL) decreased in two sites indicating impacts from changes in other environmental parameters. We conclude (1) that silicone-based passive samplers with two different area-to-volume ratios represent a promising tool for determining organic toxicants and (2) that PAHs from wildfires are unlikely to be a common main cause for fire-related ecological effects in streams adjacent to burnt regions.     

Effects of limestone strip mining on benthic macroinvertebrate communities

The changes in water quality and benthic macroinvertebrate community structure and composition related to limestone strip mining in western Pennsylvania are recorded and compared to the recorded effects of coal strip mining. In contrast to coal strip mining characterized by high total acidity and iron concentrations with decreased total alkalinity and pH, limestone strip mining showed only a slight increase in total acidity and bicarbonate alkalinity, and no increase in iron concentration due to the formation of heavy deposits of ferric hydroxide. The sulphate concentration and conductance increased with both coal and limestone strip mining. Limestone strip mining did not result in the extreme deterioration of benthic community structure and composition reported with coal strip mining. The decreased species diversity recorded was directly correlated with the sulphate concentrations and resulted from changes in species abundance rather than species composition. The changes in species abundance appear to be related to changes in food resource availability due to the deposition of ferric hydroxide on the detritus.     

Quality assessment of bed sediments of the Po River (Italy)

Comprehensive and contemporary evaluations of physical, chemical and toxicological endpoints have been performed on bed sediments of the Po River, the major Italian watercourse. Two extensive sampling campaigns were conducted in summer and winter low-flow conditions. Composite sediment samples were collected from ten reaches of the main river: the first was located in the upper region (ambient control), and the others downstream of the confluences of nine principal tributaries. The two sampling programs were paralleled by contemporary investigations on the macroinvertebrate community. The particle-size composition along the Po River showed a relatively uniform distribution of fine sand, a progressive downstream decrease of coarse sands and a corresponding increase of fine materials. The levels of polychlorinated biphenyls (PCB), polycyclic aromatic hydrocarbons (PAH), extractable organo halides (EOX), Cd, Cr, Cu, Hg, Ni, Pb and Zn were determined in sediment fine particles (< 63 microm), and showed marked changes across the ten river reaches. Their longitudinal trends, as those of organic carbon and total nitrogen, were very similar and largely independent of the survey season. Sediment quality benchmarks were used to evaluate sediment chemistry, and, although the overall level of contamination was from moderate to low, the reaches located downstream of the tributaries Dora Riparia, Dora Baltea, Lambro and Oglio were considered to be at risk. Sediments were tested for toxicity on Oncorhynchus mykiss, Ceriodaphnia dubia, Raphidocelis subcapitata and Vibrio fischeri. The toxicity tests were conducted both with sediment extracts and whole samples. Sediment extracts showed toxic potentials that were consistent with the spatial distribution of contaminants. Whole-sediment toxicity showed moderate/low effects which also included false positives and negatives. Alterations of the macroinvertebrate community were found for many kilometers downstream of Dora Riparia, and with a seasonal dependence, also in other reaches of the Italian river. Principal component analysis (PCA) was used to describe the longitudinal and temporal changes of the Po River, and allowed the selection of the most useful and discriminating indicators.     

Genotoxicity in water and sediment extracts from the St Lawrence river system, using the SOS chromotest

Surface water and sediments from the St Lawrence River system (Québec region) were analysed for genotoxicity using the SOS Chromotest, as well as for their chemical concentrations of polycyclic aromatic hydrocarbons and heavy metals. Additionally, chlorobenzenes, polychlorinated biphenyls, organochlorinated pesticides, ammonia and nitrite concentrations in sediments were determined. Water and sediments sampled from twenty-five sites were initially partitioned into their aqueous and particulate phases by tangential flow filtration and centrifugation, respectively. Organic contaminants were extracted from the fractions with dichloromethane. For surface water, fifteen extracts of filtered water and seven of particulates, and for sediments one extract of pore water and three of particulates proved to be weakly genotoxic. All but one of the genotoxic responses observed in the surface water were obtained from samples taken from the highly industrial portion of the St Lawrence River system, with the strongest responses observed in Lake St Louis. Surface water genotoxicants partitioning favours the particulate fraction. Bottom particulates genotoxicity was 1000-fold weaker than suspended particulates, on a per unit weight basis. Additionally, whole sediments were extracted with a 10% dimethylsulfoxide-saline solution. The observed distributions of genotoxicity values did not correlate with observed concentrations of demonstrated SOS inducers, mutagens and/or carcinogens, nor with the presence of other toxic chemicals.     

基于昌黎七里海潟湖藻类群落结构的水质评价

潟湖的藻类群落对水环境变化非常敏感,但是关于其藻类群落结构变化能否有效指示水质变化过程,目前开展的相关研究较少,这严重制约着滨海湿地水环境的可持续发展。因此,本文在河北昌黎七里海潟湖湿地开展了为期6个月的水化学、浮游植物、底栖藻类监测与水质评价工作,探讨了七里海潟湖湿地水体、水生生物群落结构及生物量的动态变化。同时,有别于传统的水化学水质评价方法,本研究将水生生物学评价方法与水化学指标有机融合,利用综合污染指数评价法评估七里海潟湖的水质变化。研究结果证明,潟湖的藻类群落结构能够灵敏地指示水质变化,同时可以弥补传统的水化学水质评价方法的缺陷。本研究可为七里海潟湖湿地水环境的治理与保护提供较为完整的数据支撑和相对完善的技术方案。     

Comparative assessment of genotoxicity of mineral water packed in polyethylene terephthalate (PET) and glass bottles

The potential migration of genotoxic compounds into mineral water stored in polyethylene terephthalate (PET) bottles was evaluated by an integrated chemical/biological approach using short-term toxicity/genotoxicity tests and chemical analysis. Six commercial brands of still and carbonated mineral water bottled in PET and in glass were stored at 40 °C for 10 days in a stove according to the standard EEC total migration test (82/711/EEC), or at room temperature in the dark. After treatment, the samples were analysed using gas-chromatography/mass spectrometry (GC/MS) to detect volatile and non-volatile compounds, the Microtox^® test to evaluate potential toxicity of the samples, and three mutagenicity tests - Tradescantia and Allium cepa micronucleus tests and the Comet assay on human leukocytes - to detect their genotoxic activity.GC/MS analysis did not detect phthalates or acetaldehyde in the water samples. The Microtox^® test found no toxic effects. Mutagenicity tests detected genotoxic properties of some samples in both PET and glass bottles. Statistical analyses showed a positive association between mineral content and mutagenicity (micronuclei in A. cepa and DNA damage in human leukocytes). No clear effect of treatment and PET bottle was found. These results suggest the absence of toxic compounds migrating from PET regardless of time and conditions of storage. In conclusion, bottle material and stove treatment were not associated with the genotoxic properties of the water; the genotoxic effects detected in bottled water may be related to the characteristics of the water (minerals and CO₂ content).     

Patterning and predicting aquatic macroinvertebrate diversities using artificial neural network

A counterpropagation neural network (CPN) was applied to predict species richness (SR) and Shannon diversity index (SH) of benthic macroinvertebrate communities using 34 environmental variables. The data were collected at 664 sites at 23 different water types such as springs, streams, rivers, canals, ditches, lakes, and pools in The Netherlands. By training the CPN, the sampling sites were classified into five groups and the classification was mainly related to pollution status and habitat type of the sampling sites. By visualizing environmental variables and diversity indices on the map of the trained model, the relationships between variables were evaluated. The trained CPN serves as a 'look-up table' for finding the corresponding values between environmental variables and community indices. The output of the model fitted SH and SR well showing a high accuracy of the prediction (r>0.90 and 0.67 for learning and testing process, respectively) for both SH and SR. Finally, the results of this study, which uses the capability of the CPN for patterning and predicting ecological data, suggest that the CPN can be effectively used as a tool for assessing ecological status and predicting water quality of target ecosystems.     

Copper and zinc water quality standards under the EU Water Framework Directive: the use of a tiered approach to estimate the levels of failure

Environmental quality standards are an important tool for assessing the chemical quality of water bodies under the Water Framework Directive. However, there must be confidence in assessments of any failure to avoid disproportionate investment in unnecessary risk reduction. Metals present a number of unique challenges for environmental regulators in that they are naturally occurring and their ecotoxicology is driven, in part, by the physico-chemical conditions of the water body in which they are present. This paper describes the use of a tiered approach that could be adopted to assess compliance with any future environmental quality standards for metals under the Water Framework Directive. Through this approach, the use of background concentrations is considered and also bioavailability via the use of biotic ligand models. This assessment is based on an analysis of routine Environment Agency chemical monitoring data combined with biological indices to support results of the approach. Using copper and zinc as examples, it is shown that it is important to take account of background concentrations and the bioavailability of metals, otherwise the risk of impact from metals may be significantly overestimated. The approach presented here provides a methodology by which regulators and the regulated community may implement surface water standards for metals under the Water Framework Directive.     

Using a battery of bioassays, benthic phytoplankton and the AUSRIVAS method to monitor long-term coal tar contaminated sediment in the Cache la Poudre River, Colorado

This survey provides information on sediment toxicity and structural characteristics of the macroinvertebrates and benthic phytoplankton at 10 locations in the Cache la Poudre River after long-term exposure to coal tar residue. The application of the Australian river bioassessment system (AUSRIVAS) as well as a biotest battery was used to evaluate the river ‘health’ condition. Coal tar is a dense nonaqueous-phase liquid of significant environmental concern due to its toxicity and persistence in the subsurface. Organisms like Selenastrum capricornutum, Daphnia magna and Chironomus tentans, representing different complexities in the biosphere, were selected as test systems for ecotoxicological studies. The results obtained in this study indicate that a biotest battery, macroinvertebrate and benthic phytoplankton communities are in principle suitable biological tools for evaluation of toxic oil and coal-derived substances in long-term contaminated river sediment.     

Modelling aquatic exposure and effects of insecticides--application to south-eastern Australia

Agricultural pesticides are widely used and can affect freshwater organisms. We applied a spatially explicit exposure model, validated for central Europe, to estimate exposure to insecticides through runoff for streams in south-eastern Australia. The model allows the identification of streams potentially affected by insecticide runoff located in 10 × 10 km grid cells. The computation of runoff relies on key environmental factors such as land use, soil texture, slope and precipitation. Additionally, the model predicted the ecological effect of insecticides on the macroinvertebrate community. We predicted insecticide surface runoff that results in a moderate to poor ecological quality for streams in half of the grid cells containing agricultural land. These results are in good accordance with the results obtained by estimating pesticide stress with a biotic index (SPEAR_pesticides) based on macroinvertebrate monitoring data. We conclude that the exposure and effect model can act as an effective and cost-saving tool to identify high risk areas of insecticide exposure and to support stream management.     

Seasonal periodicity of plankton and benthic fauna community structure and diversity in a small North African reservoir

Aquatic biota in terms of phytoplankton, zooplankton and benthic fauna were qualitatively and quantitatively followed seasonally during 2007 in a North African small reservoir, Aswan Reservoir. Simultaneously, seasonal variations in water quality were investigated in order to determine how far the aquatic biota could be altered by physicochemical and biotic factors. In phytoplankton of the investigated reservoir, 75 species belonging to four systematic groups were recorded, comprising 11 Cyanoprokaryota, 24 Bacillariophyceae, 2 Dinophyceae and 38 Chlorophyceae. Most numerous zooplankton were Rotifera (23 species). Cladocera and Copepoda were less numerous, with only seven and three species, respectively. Early development stages of copepods, nauplii and copepodites were sometimes recorded in considerable numbers. Twenty‐two invertebrate species related to five animal groups were exclusively benthic fauna: three Oligochaeta, nine Mollusca, seven larvae of Chironomidae, one Platyhelminthes, one Hirudinea and one Decapoda. Generally, the estimated values of phytoplankton species richness were relatively higher than those of zooplankton or benthic fauna. In particular, phytoplankton was influenced by physicochemical factors – water transparency, salinity and soluble reactive silica contents – as well as biotic factors – zooplankton and benthic fauna. Levels of pH value, Mg⁺², NO₃‐N and chlorophyll‐a contents appeared to be the most influential parameters dictating the pattern of seasonal periodicity of benthic fauna. Numerically, the phytoplankton was somewhat more abundant in winter and autumn than in spring or summer. The peak of zooplankton population densities was observed in spring season. Benthic fauna tended to be abundant in summer and autumn seasons, showing a reverse seasonal trend from that of zooplankton. Cyanoprokaryota and Bacillariophyceae were the most important groups that determined the seasonal fluctuation pattern of phytoplankton. Rotifera and Copepoda were the most predominant groups influencing the seasonal periodicity of zooplankton. Oligochaeta and Mollusca alternated the dominancy of benthic fauna community in Aswan Reservoir.     

Water quality and macroinvertebrate community response following pesticide applications in a banana plantation, Limon, Costa Rica

Pesticides used in banana production may enter watercourses and pose ecological risks for aquatic ecosystems. The occurrence and effects of pesticides in a stream draining a banana plantation was evaluated using chemical characterization, toxicity testing and macrobenthic community composition. All nematicides studied were detected in the surface waters of the banana plantation during application periods, with peak concentrations following applications. Toxicity tests were limited to the carbofuran application and no toxicity was observed with the acute tests used. However, since pesticide concentrations were generally below the lowest LC50 value for crustaceans but above calculated aquatic quality criteria, there remains a risk of chronic toxicity. Accurate ecological assessments of pesticide use in banana plantations are currently limited by the lack of local short-term chronic toxicity tests and tests using sensitive native species. Relatively constant levels of four pesticides (imazalil, thiabendazole, chlorpyrifos and propiconazole), which had toxic effects according to the 96h hydra and 21d daphnia chronic test, were recorded in the effluent of the packing plant throughout the study, indicating that the solid waste trap used in this facility was not effective in eliminating toxic chemicals. Certain taxa, such as Heterelmis sp. (Elmidae), Heteragrion sp. (Megapodagrionidae, Odonata), Caenis sp. (Caenidae, Ephemerotera), and Smicridea sp. (Hidropsychidae, Trichoptera), were more abundant at reference sites than in the banana farm waters, and may be good candidates for toxicity testing. Multivariate analyses of the macroinvertebrate communities clearly showed that the banana plantation sites were significantly different from the reference sites. Moreover, following the pesticide applications, all the banana plantation sites showed significant changes in community composition, with the same genera being affected at all sites and for all pesticides (terbufos, cadusafos and carbofuran). Consequently, the results presented here show that multivariate analysis of community composition was more sensitive in distinguishing pesticide effects than the toxicity tests and richness and composition measures used. We conclude that monitoring macroinvertebrate communities can be a powerful tool in the assessment of ecological effects of banana production.