Background, aim, and scope Selenium (Se) has been shown to reduce mercury (Hg) bioavailability and trophic transfer in aquatic ecosystems. The study of methylmercury (MeHg) and Se bioaccumulation by plankton is therefore of great significance in order to obtain a better understanding of the estuarine processes concerning Hg and Se accumulation and biomagnification throughout the food web. In the western South Atlantic, few studies have documented trace element and MeHg in fish tissues. No previous study about trace elements and MeHg in plankton has been conducted concerning tropical marine food webs. Se, Hg, and MeHg were determined in two size classes of plankton, microplankton (70-290 μm) and mesoplankton (>=290 μm), and also in muscle tissues and livers of four fish species of different trophic levels (Mugil liza, a planktivorous fish; Bagre spp., an omnivorous fish; Micropogonias furnieri, a benthic carnivorous fish; and Centropomus undecimalis, a pelagic carnivorous fish) from a polluted estuary in the Brazilian Southeast coast, Guanabara Bay. Biological and ecological factors such as body length, feeding habits, and trophic transfer were considered in order to outline the relationships between these two elements. The differences in trace element levels among the different trophic levels were investigated. Materials and methods Fish were collected from July 2004 to August 2005 at Guanabara Bay. Plankton was collected from six locations within the bay in August 2005. Total mercury (THg) was determined by cold vapor atomic absorption spectrometry (CV-AAS) with sodium borohydride as a reducing agent. MeHg analysis was conducted by digesting samples with an alcoholic potassium hydroxide solution followed by dithizone-toluene extraction. MeHg was then identified and quantified in the toluene layer by gas chromatography with an electron capture detector (GC-ECD). Se was determined by AAS using graphite tube with Pin platform and Zeeman background correction. Results and discussion Total mercury, MeHg, and Se increased with plankton size class. THg and Se values were below 2.0 and 4.8 μg g⁻¹ dry wt in microplankton and mesoplankton, respectively. A large excess of molar concentrations of Se in relation to THg was observed in both plankton size class and both fish tissues. Plankton presented the lowest concentrations of this element. In fish, the liver showed the highest THg and Se concentrations. THg and Se in muscle were higher in Centropomus undecimalis (3.4 and 25.5 nmol g⁻¹) than in Micropogonias furnieri (2.9 and 15.3 nmol g⁻¹), Bagre spp (1.3 and 3.4 nmol g⁻¹) and Mugil liza (0.3 and 5.1 nmol g⁻¹), respectively. The trophic transfer of THg and Se was observed between trophic levels from prey (considering microplankton and mesoplankton) to top predator (fish). The top predators in this ecosystem, Centropomus undecimalis and Micropogonias furnieri, presented similar MeHg concentrations in muscles and liver. Microplankton presented lower ratios of methylmercury to total mercury concentration (MeHg/THg) (34%) than those found in mesoplankton (69%) and in the muscle of planktivorous fish, Mugil liza (56%). The other fish species presented similar MeHg/THg in muscle tissue (of around 100%). M. liza showed lower MeHg/THg in the liver than C. undecimalis (35%), M. furnieri (31%) and Bagre spp. (22%). Significant positive linear relationships were observed between the molar concentrations of THg and Se in the muscle tissue of M. furnieri and M. liza. These fish species also showed significant inverse linear relationships between hepatic MeHg and Se, suggesting a strong antagonistic effect of Se on MeHg assimilation and accumulation. Conclusions Differences found among the concentrations THg, MeHg, and Se in microplankton, mesozooplankton, and fishes were probably related to the preferred prey and bioavailability of these elements in the marine environment. The increasing concentration of MeHg and Se at successively higher trophic levels of the food web of Guanabara Bay corresponds to a transfer between trophic levels from the lower trophic level to the top-level predator, suggesting that MeHg and Se were biomagnified throughout the food web. Hg and Se were positively correlated with the fish standard length, suggesting that larger and older fish bioaccumulated more of these trace elements. THg, MeHg, and Se were a function of the plankton size. Recommendations and perspectives There is a need to assess the role of selenium in mercury accumulation in tropical ecosystems. Without further studies of the speciation of selenium in livers of fishes from this region, the precise role of this element, if any, cannot be verified in positively affecting mercury accumulation. Further studies of this element in the study of marine species should include liver samples containing relatively high concentrations of mercury. A basin-wide survey of selenium in fishes is also recommended.

Background Due to the requirements of the EU Water Framework Directive and other legal regulations (e.g., national laws like the German Federal Soil Protection Act), but also due to the implementation of the new EU chemicals management system REACH, environmental monitoring will gain increasing importance for the surveillance of environmental quality as well as for the assessment of chemicals. Against this background, the Work Group on 'Environmental Monitoring' of the Division of 'Environmental Chemistry and Ecotoxicology' within the German Chemical Society has compiled a position paper on substance-related environmental monitoring. Scope Core elements of this position paper are the definitions of important terms like monitoring, exposure monitoring, effect monitoring, and integrated monitoring. Moreover, temporal and spatial aspects (monitoring of spatial distributions, trend monitoring, and retrospective monitoring) and their applications are discussed. The main focus of this position paper is the coverage of aspects which have to be observed for the preparation and implementation of a monitoring program. Essential is the clear specification of the targets of the monitoring which determine the development of the monitoring concept and its realization, e.g., if environmental media (compartments) or organisms are most appropriate as samples for the aim of the study. Of course, also the properties of the target substances are important (e.g., lipophilicity/bioaccumulation as pre-requisite for an exposure monitoring with organisms). Finally, the monitoring phases of sample planning, sampling, storage and transport of samples, selection of analytical methods, quality assurance measures as well as reporting are discussed. Perspectives An important issue for the future is to link the quantification of chemicals in environmental compartments (exposure and pollution monitoring) more closely to the study of biological effects (effect monitoring) than has been the case up to now. Furthermore, by inclusion of a spatial differentiation, a comprehensive evaluation of the state of an ecosystem can be obtained and the relevance of the results improved. Practical examples of monitoring studies which illustrate various aspects covered in this position paper will be presented in a series of publications by members of the Work Group in the following issues of this journal.

Background, aim, and scope The reclamation of nonferrous metal-polluted soil by phytoremediation requires an overall and permanent plant cover. To select the most suitable plant species, it is necessary to study metal effects on plants over the time, thereby checking that metals remain stored in root systems and not transferred to aerial parts. In this purpose, the seasonal and annual variations of metal bioaccumulation, transfer, and phytotoxicity in Trifolium repens and Lolium perenne grown in a Cd-, Pb-, and Zn-contaminated soil were also studied. Materials and methods The experimental site was located near a closed smelter. In spring 2004, two areas were sown with T. repens and L. perenne, respectively. Thereafter, the samplings of plant roots and shoots and surrounding soils were realized in autumn 2004 and spring and autumn 2005. The soil agronomic characteristics, the Cd, Pb, and Zn concentrations in the surrounded soils and plant organs, as well as the oxidative alterations (superoxide dismutase [SOD], malondialdehyde [MDA], and 8-hydroxy-2'-deoxyguanosine [8-OHdG]) in plant organs were carried out. Results Whatever the sampling period, metal concentrations in soils and plants were higher than background values. Contrary to the soils, the fluctuations of metal concentrations were observed in plant organs over the time. Bioaccumulation and transfer factors confirmed that metals were preferentially accumulated in the roots as follows: Cd>Zn>Pb, and their transfer to shoots was limited. Foliar metal deposition was also observed. The results showed that there were seasonal and annual variations of metal accumulation in the two studied plant species. These variations differed according to the organs and followed nearly the same pattern for the two species. Oxidative alterations were observed in plant organs with regard to SOD antioxidant activities, MDA, and 8-OHdG concentrations. These alterations vary according to the temporal variations of metal concentrations. Discussion Metal concentrations in surrounded soils and plant organs showed the effective contamination by industrial dust emissions. Metals absorbed by plants were mainly stored in the roots. With regard to this storage, the plants seemed to limit the metal transfer to their aerial parts over the time, thereby indicating their availability for metal phytostabilization. Aerial deposition was another source of plant exposure to nonferrous metals. Despite the occurrence of metal-induced oxidative alterations in plant organs, both plant species seemed to tolerate a high metal concentration in soils. Conclusions Taken together, these results indicated that T. repens and L. perenne were able to form a plant cover on highly Cd-, Pb-, and Zn-polluted soils, to limit the metal transfer to their aerial parts and were relatively metal-tolerant. All these characteristics made them suitable for phytostabilization on metal-contaminated soils. These findings also highlighted the necessity to take into account seasonal and annual variations for a future phytomanagement. Recommendations and perspectives In this work, the behavior of plant species grown in metal-polluted soil has been studied during 2 years. Obviously, this time is too short to ensure that metals remain accumulated in the root system and few are transferred in aerial parts over the time. It is why regular monitoring should be achieved during more than a decade after the settlement of the plant cover. This work will be completed by the study of the T. repens and L. perenne effects on mobility of metals in order to evaluate the quantities of pollutants which could be absorbed by the biota and transferred to groundwater. Bioaccessibility tests could be also realized on polluted soils in order to evaluate the phytostabilization impacts on the exposition risks for humans.

Background, aim, and scope At tropical latitudes, and especially on the semi-arid coasts of the Brazilian Northeast, the rainfall regime governs the water quality of estuaries due to the pronounced difference between the rainy and dry seasons. These changes may be responsible for seasonal changes in bioavailability of mercury (Hg) and other pollutants to the estuarine and coastal biota. Mercury bioaccumulates along estuarine-marine food chains usually result in higher concentrations in tissues of top predators and posing a risk to both marine mammals and humans alike. The Goiana River Estuary (7.5° S) is a typical estuary of the semi-arid tropical regions and supports traditional communities with fisheries (mollusks, fish, and crustacean). It is also responsible for an important part of the biological production of the adjacent coastal waters. Materials and methods Trichiurus lepturus (Actinopterygii: Perciformes) is a pscivorous marine straggler. Fish from this species (n = 104) were captured in a trapping barrier used by the local traditional population and using an otter trawl net along the main channel of the low estuary during two dry seasons (D1 = November, December 2005, January 2006; D2 = November, December 2006, January 2007) and the end of a rainy season (R = August, September, October 2006). Fish muscle samples were preserved cold and then freeze-dried prior to analysis of its total mercury (Hg-T) contents. Total mercury was determined by cold vapor atomic absorption spectrometry (CV-AAS) with sodium borohydride as a reducing agent. Results The studied individuals (n = 104) were sub-adult (30-70 cm, 71 ind.) and adult fish (>70 cm, 33 ind.). Weight (W) (204.1 ± 97.9 g, total biomass = 21,229.7 g) and total length (TL) (63.1 ± 10.1 cm, range 29.5-89.0 cm) presented a significant (p < 0.05) correlation. Two-way ANOVA (n = 81) showed that TL and W had significant differences (p < 0.05) among seasons, being higher in D1 than in D2 and R, respectively. Moreover, season vs. month interaction were detected for the variables length and weight. For the variable weight was detected significant difference for the factor month (p < 0.05). It suggests that the fish enter the estuary at the end of the rainy season and increase in length and weight during the time they spend in the estuary. Fish from this estuary are shown to be fit for human consumption (125.3 ± 61.9 μgHg-T kg⁻¹ w.wt.; n = 104). Fish mercury contents increased with size and weight. Correlations between TL and Hg-T (r = 0.37286) and between W and Hg-T (r = 0.38212) were significant (p < 0.05). Dryer months showed higher mercury concentrations in fish (D1 773.4 ± 207.5 μgHg-T kg⁻¹ d.wt., n = 27; D2 370.1 ± 78.8 μgHg-T kg⁻¹ d.wt., n = 27; R 331.2 ± 138.5 μgHg-T kg⁻¹ d.wt., n = 27). The variable mercury concentration showed differences in relation to the factor season (p < 0.05), where fish captured during the first dry season showed the highest concentration of mercury. The correlation between Hg-T and rainfall (Rf) showed a negative correlation (r = -0.56; p < 0.05). Discussion The main likely source of mercury to this estuary is diffuse continental run off, including urban and industrial effluents. Since concentration of mercury in fish tissue is negatively correlated to rainfall, but positively correlated with fish length and weight, it suggests that fish growth in this estuary results in mercury uptake and concentration on the fish tissue. In the dry season of 2005-2006, when rainfall remained below the historic average, fish bioaccumulated significantly more mercury than in the dry season 2006-2007, when rainfall was within the predictable historic average. It is suggested that less rainfall, and consequently less particulate matter and less primary production in the estuary, make mercury more available to the higher levels of the estuarine food chain. In the case of higher rainfall, when river flow increases and water quality in the estuary is reduced, mercury probably is quickly exported associated to the particulate matter to the adjacent coastal waters where it then disperses. This species is a potential routine bioindicator for mercury contamination of the biota, but so far was used only with a limited number of individuals and contexts. Conclusions Fish from the Goiana River estuary can still be safely consumed by the local population. However, any further contamination of this resource might lead to total mercury levels above the recommended limits for pregnant women and small children. The proposed heavy dependency of total mercury levels in fish on water quality indicates that land use and water quality standards must be more closely watched in order to guarantee that best possible practices are in place to prevent bioaccumulation of mercury and its transfer along the food chain. Human interventions and climatic events which affect river water flow are also playing a role in the mercury cycle at tropical semi-arid estuaries. Recommendations and perspectives T. lepturus is largely consumed by coastal populations of tropical and sub-tropical countries all over the world. It is also consumed by a number of marine mammals over which we have a strong conservation interest. This species is also a link among different ecosystems along the estuarine ecocline. Therefore, knowledge of its degree of contamination might contribute to public health issues as well as marine conservation actions. Studies on mercury and other contaminants using this species as bioindicator (cosmopolitan, readily available) could help elucidating mechanisms through which pollutants are being transferred not only through the food chain, but also from estuarine-coastal-open waters. In addition, using the same species in marine pollution studies, especially as part of a mosaic of species, allows for wide range comparisons of marine food chain contamination.

Background, aim and scope Estrogenic and non-estrogenic chemicals typically co-occur in the environment. Interference by non-estrogenic chemicals may confound the assessment of the actual estrogenic activity of complex environmental samples. The aim of the present study was to investigate whether, in which way and how seriously the estrogenic activity of single estrogens and the observed and predicted joint action of estrogenic mixtures is influenced by toxic masking and synergistic modulation caused by non-estrogenic chemical confounders. Materials and methods The yeast estrogen screen (YES) was adapted so that toxicity and estrogenicity could be quantified simultaneously in one experimental run. Mercury, two organic solvents (dimethyl sulfoxide (DMSO) and 2,4-dinitroaniline), a surfactant (LAS-12) and the antibiotic cycloheximide were selected as toxic but non-estrogenic test chemicals. The confounding impact of selected concentrations of these toxicants on the estrogenic activity of the hormone 17ß-estradiol was determined by co-incubation experiments. In a second step, the impact of toxic masking and synergistic modulation on the predictability of the joint action of 17ß-estradiol, estrone and estriol mixtures by concentration addition was analysed. Results Each of the non-estrogenic chemicals reduced the apparent estrogenicity of both single estrogens and their mixtures if applied at high, toxic concentrations. Besides this common pattern, a highly substance- and concentration-dependent impact of the non-estrogenic toxicants was observable. The activity of 17ß-estradiol was still reduced in the presence of only low or non-toxic concentrations of 2,4-dinitroaniline and cycloheximide, which was not the case for mercury and DMSO. A clear synergistic modulation, i.e. an enhanced estrogenic activity, was induced by the presence of slightly toxic concentrations of LAS-12. The joint estrogenic activity of the mixture of estrogens was affected by toxic masking and synergistic modulation in direct proportion to the single estrogens, which allowed for an adequate adaptation of concentration addition and thus unaffected predictability of the joint estrogenicity in the presence of non-estrogenic confounders. Discussion The modified YES proved to be a reliable system for the simultaneous quantification of yeast toxicity and estrogen receptor activation. Experimental results substantiate the available evidence for toxic masking as a relevant phenomenon in estrogenicity assessment of complex environmental samples. Synergistic modulation of estrogenic activity by non-estrogenic confounders might be of lower importance. The concept of concentration addition is discussed as a valuable tool for estrogenicity assessment of complex mixtures, with deviations of the measured joint estrogenicity from predictions indicating the need for refined analyses. Conclusions Two major challenges are to be considered simultaneously for a reliable analysis of the estrogenic activity of complex mixtures: the identification of known and suspected estrogenic compounds in the sample as well as the substance- and effect-level-dependent confounding impact of non-estrogenic toxicants. Recommendations and perspectives The application of screening assays such as the YES to complex mixtures should be accompanied by measures that safeguard against false negative results which may be caused by non-estrogenic but toxic confounders. Simultaneous assessments of estrogenicity and toxicity are generally advisable.

Background, aim, and scope Decabromodiphenyl ether (DecaBDE) is used as an additive flame retardant in polymers. It has become a ubiquitous environmental contaminant, particularly abundant in abiotic media, such as sediments, air, and dust, and also present in wildlife and in humans. The main DecaBDE constituent, perbrominated diphenyl ether (BDE-209), is susceptible to transformations as observed in experimental work. This work is aimed at identifying and assessing the relative amounts of products formed after UV irradiation of BDE-209. Materials and methods BDE-209, dissolved in tetrahydrofuran (THF), methanol, or combinations of methanol/water, was exposed to UV light for 100 or 200 min. Samples were analyzed by gas chromatography/mass spectrometry (electron ionization) for polybrominated diphenyl ethers (PBDEs), dibenzofurans (PBDFs), methoxylated PBDEs, and phenolic PBDE products. Results The products formed were hexaBDEs to nonaBDEs, monoBDFs to pentaBDFs, and methoxylated tetraBDFs to pentaBDFs. The products found in the fraction containing halogenated phenols were assigned to be pentabromophenol, dihydroxytetrabromobenzene, dihydroxydibromodibenzofuran, dihydroxytribromodibenzofuran, and dihydroxytetrabromodibenzofuran. The PBDEs accounted for approximately 90% of the total amount of substances in each sample and the PBDFs for about 10%. Discussion BDE-209 is a source of PBDEs primarily present in OctaBDEs but also to some extent in PentaBDEs, both being commercial products now banned within the EU and in several states within the USA. It is notable that OH-PBDFs have not been identified or indicated in any of the photolysis studies performed to date. Formation of OH-PBDFs, however, may occur as pure radical reactions in the atmosphere. Conclusions Photolysis of decaBDE yields a wide span of products, from nonaBDEs to hydroxylated bromobenzenes. It is evident that irradiation of decaBDE in water and methanol yields OH-PBDFs and MeO-PBDFs, respectively. BDE-202 (2,2',3,3',5,5',6,6'-octabromodiphenyl ether) is identified as a marker of BDE-209 photolysis. Recommendations and perspectives BDE-209, the main constituent of DecaBDE, is primarily forming debrominated diphenyl ethers with higher persistence which are more bioaccumulative than the starting material when subjected to UV light. Hence, DecaBDE should be considered as a source of these PBDE congeners in the environment.

The Institute of Environmental Engineering at the University of Pannonia has undertaken the challenge to develop an online environmental information system. This system is able to receive and process the collected environmental data via Internet. The authors have attached importance to the presentation of the data and have included other comprehensible information for laymen as well in order to work out visualisation techniques that are expressive and attract attention for environmental questions through the developed information system. The ways of visualizing physical and chemical parameters of surface water and the effects of motorway construction were examined.

Background, aim, and scope The need for global and integrated approaches to water resources management, both from the quantitative and the qualitative point of view, has long been recognized. Water quality management is a major issue for sustainable development and a mandatory task with respect to the implementation of the European Water Framework Directive as well as the Swiss legislation. However, data modelling to develop relational databases and subsequent geographic information system (GIS)-based water management instruments are a rather recent and not that widespread trend. The publication of overall guidelines for data modelling along with the EU Water Framework Directive is an important milestone in this area. Improving overall water quality requires better and more easily accessible data, but also the possibility to link data to simulation models. Models are to be used to derive indicators that will in turn support decision-making processes. For this whole chain to become effective at a river basin scale, all its components have to become part of the current daily practice of the local water administration. Any system, tool, or instrument that is not designed to meet, first of all, the fundamental needs of its primary end-users has almost no chance to be successful in the longer term. Materials and methods Although based on a pre-existing water resources management system developed in Switzerland, the methodological approach applied to develop a GIS-based water quality management system adapted to the Romanian context followed a set of well-defined steps: the first and very important step is the assessment of needs (on the basis of a careful analysis of the various activities and missions of the water administration and other relevant stakeholders in water management related issues). On that basis, a conceptual data model (CDM) can be developed, to be later on turned into a physical database. Finally, the specifically requested additional functionalities (i.e. functionalities not provided by classical commercial GIS software), also identified during the assessment of needs, are developed. This methodology was applied, on an experimental basin, in the Ialomita River basin. Results The results obtained from this action-research project consist of a set of tangible elements, among which (1) a conceptual data model adapted to the Romanian specificities regarding water resources management (needs, data availability, etc.), (2) a related spatial relational database (objects and attributes in tables, links, etc.), that can be used to store the data collected, among others, by the water administration, and later on exploited with geographical information systems, (3) a toolbar (in the ESRI environment) offering the requested data processing and visualizing functionalities. Lessons learned from this whole process can be considered as additional, although less tangible, results. Discussion The applied methodology is fairly classical and did not come up with revolutionary results. Actually, the interesting aspects of this work are, on the one hand, and obviously, the fact that it produced tools matching the needs of the local (if not national) water administration (i.e. with a good chance of being effectively used in the day-to-day practice), and, on the other hand, the adaptations and adjustments that were needed both at the staff level and in technical terms. Conclusions This research showed that a GIS-based water management system needs to be backed by some basic data management tools that form the necessary support upon which a GIS can be deployed. The main lesson gained is that technology transfer has to pay much attention to the differences in existing situations and backgrounds in general, and therefore must be able to show much flexibility. The fact that the original objectives could be adapted to meet the real needs of the local end-users is considered as a major aspect in achieving a successful adaptation and development of water resources management tools. Time needed to setup things in real life was probably the most underestimated aspect in this technology transfer process. Recommendations and perspectives The whole material produced (conceptual data model, database and GIS tools) was disseminated among all river basin authorities in Romania on the behalf of the national water administration (ANAR). The fact that further developments, for example, to address water quantity issues more precisely, as envisaged by ANAR, can be seen as an indication that this project succeeded in providing an appropriate input to improve water quality in Romania on the long term.

Background, aim, and scope Pharmaceuticals and personal care products (PPCPs) including antibiotics, endocrine-disrupting chemicals, and veterinary pharmaceuticals are emerging pollutants, and their environmental risk was not emphasized until a decade ago. These compounds have been reported to cause adverse impacts on wildlife and human. However, compared to the studies on hydrophobic organic contaminants (HOCs) whose sorption characteristics is reviewed in Part IV of this review series, information on PPCPs is very limited. Thus, a summary of recent research progress on PPCP sorption in soils or sediments is necessary to clarify research requirements and directions. Main features We reviewed the research progress on PPCP sorption in soils or sediments highlighting PPCP sorption different from that of HOCs. Special function of humic substances (HSs) on PPCP behavior is summarized according to several features of PPCP-soil or sediment interaction. In addition, we discussed the behavior of xenobiotic chemicals in a three-phase system (dissolved organic matter (DOM)-mineral-water). The complexity of three-phase systems was also discussed. Results Nonideal sorption of PPCPs in soils or sediments is generally reported, and PPCP sorption behavior is relatively a more complicated process compared to HOC sorption, such as the contribution of inorganic fractions, fast degradation and metabolite sorption, and species-specific sorption mechanism. Thus, mechanistic studies are urgently needed for a better understanding of their environmental risk and for pollution control. Discussion Recent research progress on nonideal sorption has not been incorporated into fate modeling of xenobiotic chemicals. A major reason is the complexity of the three-phase system. First of all, lack of knowledge in describing DOM fractionation after adsorption by mineral particles is one of the major restrictions for an accurate prediction of xenobiotic chemical behavior in the presence of DOM. Secondly, no explicit mathematical relationship between HS chemical-physical properties, and their sorption characteristics has been proposed. Last but not least, nonlinear interactions could exponentially increase the complexity and uncertainties of environmental fate models for xenobiotics. Discussion on proper simplification of fate modeling in the framework of nonlinear interactions is still unavailable. Conclusions Although the methodologies and concepts for studying HOC environmental fate could be adopted for PPCP study, their differences should be highly understood. Prediction of PPCP environmental behavior needs to combine contributions from various fractions of soils or sediments and the sorption of their metabolites and different species. Recommendations and perspectives More detailed studies on PPCP sorption in separated soil or sediment fractions are needed in order to propose a model predicting PPCP sorption in soils or sediments based on soil or sediment properties. The information on sorption of PPCP metabolites and species and the competition between them is still not enough to be incorporated into any predictive models.

Background, aim and scope Glass wools are man-made vitreous fibres, which consist principally of sodium, calcium and magnesium silicates, but may contain smaller amounts of other elements, including boron. The boron contents originate from the use of borates in the glass melting process as a glass former and a flux agent. During the production and application of glass wool insulation products, workers may legally be exposed to glass fibre up to the occupational limit value, commonly of 1 fibre/cm³. However, in practice, the fibre exposure will be at least ten times lower. Boron is a non-metallic element widely distributed in nature, where it occurs as boric acid, borates and borosilicates. Humans are mainly exposed to boron via vegetarian food and drinking water, mineral supplements and various consumer products. Boron is an essential element for plant growth, but the essentiality for humans is not proven, although intakes of trace amounts of the element seem to be useful for bone health and proper brain function; higher concentrations of boron, however, may be toxic. In relation to the European Union legislation on dangerous substances, an EU Expert Group has recommended classifying boric acid and borates with risk phrases for reproductive toxicity. The aim of this paper is to assess whether the new EU hazard classification of boron compounds should imply that glass wool products used for building insulation in the future should be labelled, “may impair fertility and cause harm to the unborn child”, because of the low boron content. Materials and methods Boron intakes are estimated in a worst-case occupational situation with human exposure to glass wool fibres at the occupational limit of 1 fibre/cm³ by calculation of the mass of the amount of fibres inhaled during an 8-h work day. Fibres are supposed to be cylinders of glass with a length of 30 μm, an average diameter of 1.5-2 μm and containing either 1.5% or 3.5% boron. As a worst-case scenario, the density of the fibres is set to 2,700 kg/m³. The inhalation rate of the individuals at moderate work load was set to 2 m³/h. A worst-case scenario also corresponds to 100% retention and to 100% solubility of the retained fibres in the lungs. Results With the normal boron content of 1.5% in glass wool fibres for building insulation, the extra daily occupational boron intake/uptake will be 0.03-0.06 mg B for 5 days a week. For more uncommon glass wool with maximum boron content of 3.5%, the worst-case daily boron intake/uptake will be 0.08-0.16 mg B. The main boron exposure in the general population is from vegetarian food, and the average daily dietary intake with food is estimated to 1.2-1.5 mg B/day. In addition, significant intakes may come with drinking water, especially from mineral water. In some instances, exposure from mineral supplements, cosmetics and other consumer products may be significant. For example, individuals taking mineral supplements, e.g. for bodybuilding, may have an additional intake to that of 1-10 mg/day. During the years, various organisations have recommended safe intake values for boron. Recently, the Scientific Panel on Dietetic Products, Nutrition and Allergies of the European Food Safety Authority (EFSA) has established the 'Tolerable Upper Intake Level' (UL) for the intake of boron (boric acid and borates) at 0.16 mg B/kg body weight per day or about 10 mg B/day for an adult. Discussion The calculated, worst-case exposure scenario during an 8-h work day will result in an extra daily boron intake that only corresponds to about 10% of the average daily adult boron intakes through food and drinks of about 1.5 mg. The inter-individual variations in boron intakes from foods, water and supplements will be much greater than an eventual, very worst-case, additional intake of boron from inhalation of glass wool fibres. In addition, the combined intakes are far lower than the 'Tolerable Upper Intake Level' of 10 mg B/day for a person weighing 60 kg, as recommended by the European Food Safety Agency. The potential boron intake from inhalation of glass wool fibres is also much lower than boron intakes by workers in the boron industry, who at the present occupational limit value will be exposed to 50 mg of boron 5 days a week, or 100 times more than the worst case for glass wool fibres. Furthermore, in practice, exposure levels will mostly be ten to 100 times lower than the occupational limit used here as a worst case. Conclusions The estimated boron intake from inhalation of glass wool fibres in occupational settings will be insignificant and without any health risks, even in the case of non-compliance with the occupational limit value. Any proposal requiring hazard labels on commercial glass wool products for building insulation, because of the boron content, is not supported by the present scientific knowledge. Recommendations and perspectives The European Commission should ensure that the new EU hazard classification of boron compounds is not applied to commercial glass wool products for building insulation having a low content of boron.