Particulate and dissolved trace metals in Lake Ontario

Particulate metal concentrations in the nearshore waters of Lake Ontario have been determined to be 690 ng l⁻¹ for Cu; 40 ng l⁻¹ for Cd; 180 ng l⁻¹ for Ni; 1690 ng l⁻¹ for Zn; 2100 ng l⁻¹ for Mn; and 700 μg l⁻¹ for Fe. These values are considerably higher than the particulate metal concentrations in the offshore waters: 130, 8, 34, 230, 110, 260 and 9000 ng l⁻¹ for Cu, Cd, Ni, Zn, Pb, Mn and Fe respectively. In general, 50–80% of the Cu, 10–40% of the Ni, 20–60% of the Cd and >60% of the Pb in the lake water were bound to the suspended particulates. From the standing crop of the particulate metals and the estimated rates of their deposition on the lake bottom, the residence times of the particulate metals in the lake water column have been estimated to be about 0.5 yr. on the average. The suggestion is made that particulate organic matter may be an important vehicle for metal transport to the Lake Ontario sediments.     

Environmental chemistry of copper in Lake Monona, Wisconsin

Lake Monona, located at Madison, Wisconsin, received over 1.5 × 10⁶ pounds of copper sulfate in the past 50 yr to control excessive algal growth. Dissolved copper on Lake Monona epilimnion is inversely related to pH which indicates possible control of dissolved copper by basic copper carbonate. Concentrations as high as about 4 μg Cu l⁻¹ were found in Lake Monona epilimnion, which also contains 3.3 me l⁻¹ (milliequivalents per liter) of alkalinity, mostly bicarbonate. Concentrations of dissolved copper were consistently lower (0.3 μg Cu l⁻¹) in the hypolimnion. Sulfide probably controls dissolved copper in the hypolimnion during anoxic conditions because of sulfide insolubility. Particulate copper concentrations of about 3 μg l⁻¹ increased slightly with depth. The highest concentrations of copper in Lake Monona sediments (650 mg kg⁻¹) were found approximately 60 cm below the current sediment surface. Surface sediments of Lake Monona contained approximately 250 mg Cu kg⁻¹ sediment dry weight.     

Investigation of the Folin-Ciocalteau phenol reagent for the determination of polyphenolic substances in natural waters

The methodology associated with the Folin-Ciocalteau phenol reagent was investigated and the performance characteristics of a method using sodium carbonate as the supporting medium were determined. Calibration curves using phenol, tannic acid, or l-tyrosine were linear up to at least 1000 μg l⁻¹. The limit of detection was 6 μg phenol l⁻¹ and the relative standard deviation at 100 μg phenol l⁻¹ was 5.2% and at 1000 μg phenol l⁻¹ was 4.1%. The absorbances obtained with equal amounts of a range of potential standards showed variations when compared with that of phenol: phenol (100%), l-tyrosine (62%), oak gall tannin (58%), tannic acid (48%), chestnut tannin (26%), oak tannin (24%), fulvic acid (5%). The method was applicable to a wide range of monohydric and polyhydric phenolic substances and interferences from inorganic and non-phenolic organic compounds were examined. Interference would be expected above 30 μg S²⁻ l⁻¹, 300 μg Mn(II) l⁻¹, or 400 μg SO₃²⁻ l⁻¹. Concentrations of iron >2 mg l⁻¹ as Fe(II) or Fe(III) formed the insoluble iron(III) hydroxide which increased the absorbance, but centrifugation could be used to remove this source of interference. Other potential sources of intereference (e.g. reducing agents and certain metabolic products) would be expected to have a negligible effect in unpolluted waters. Methods using diazotised sulphanilic acid or 4-aminoantipyrine (4-AAP) were found to be inferior when applied to natural water samples.     

Chlorophenols in surface waters of the Netherlands (1976–1977)

In this paper an analytical method is described for the capillary gas chromatographic determination, after derivatization, of 19 individual chlorophenols in surface water. The minimum detectable amounts are for monochlorophenols 2 μg l⁻¹, for dichlorophenols 0.05 μg l⁻¹, for trichlorophenols 0.02 μg l⁻¹ and for tetra- and pentachlorophenols 0.01 μg l⁻¹. The results of a monitoring program in the river Rhine and other Dutch surface waters with respect to these compounds are presented. The results cover the period January 1976–December 1977. 2,6-Dichlorophenol, 2,4,6-trichlorophenol, 2,3,4,6-tetrachlorophenol and pentachlorophenol had the highest frequencies of occurrence in the river Rhine and its tributaries. Pentachlorophenol was found in the highest concentrations (up to 11 μg l⁻¹).     

Concentrations of heavy metals in small Norwegian lakes

As part of regional surveys of lakes in Norway the concentrations of Zn, Pb, Cu and Cd were measured in surface- and bottom-water samples collected from representative, small, pristine lakes (136 in southern Norway sampled in October 1974, 58 resampled in March 1975, and 77 in northern Norway sampled in March 1975). The lakes, a statistically representative sample of small lakes in Norway, were chosen such that their watersheds are undisturbed. Heavy-metal concentrations in these lakes thus reflect only natural inputs and anthropogenic inputs via the atmosphere.The generally low concentrations (Zn 0.5–12.0 μg l⁻¹; Pb 0–2.0 μg l⁻¹; Cu 0–2.0 μg l⁻¹; Cd 0.1-0.5 μg l⁻¹) measured in lakes in central and northern Norway provide estimates of natural “background” levels. These estimates may be too high because they include the global-scale deposition of heavy metals from the atmosphere which has increased as a result of industrial activities.Concentrations of Zn and Pb in lakes in southernmost and southeastern Norway lie above these “background” levels, apparently because of atmospheric deposition associated with the acidic precipitation that falls over southern Scandinavia. Increased heavy-metal concentrations in acid lakes may also be due to increased mobilization of metals due to acidification of soil- and surface-waters.     

Complexation analysis of fresh waters by equilibrium diafiltration

Equilibrium ultrafiltration has been employed to determine the extent of Cu, Ni, Co, Fe, Mn and Zu chelation in oxic waters from mineralized terrains. Concentrations of Cu, Ni and Co were determined in different organic fractions separated by various Amicon ultrafilters. Titrations of fresh water samples with metal ions were made, and “free” metal concentration measured by equilibrium diafiltration. Experiments were made at concentrations and pH at which metals and ligands occur in natural waters. Conditional stability functions ranged from 10^6.72 at pH 6.5 to 10^7.07 at pH 7.6 for Ni and 10^6.85 at pH 6.8 to 10^6.97 at pH 7.6 for Co. At low metal concentrations (8 × 10⁻⁷ M) only one complexing class of major importance for natural water environments was observed. The order of metal complexing ability was found to be Cu > Ni > Co > Zn > Mn with Cu showing a preference towards the higher MW organic fractions.     

Effects of aluminum and pH on the early life stages of smallmouth bass (Micropterus dolomieui)

The sensitivity of smallmouth bass Micropterus dolomieui to acidified conditions was examined by exposing recently-hatched fish to pH levels ranging from 5.1 to 7.5 and aluminum concentrations ranging from 32 to 1000 μg l⁻¹. The range of pH and aluminum concentrations included those found in the northern part of the species' range. Acute bioassays (96 h) conducted at a pH of 5.1 and aluminum concentrations 180 μgl⁻¹ resulted in total mortality. The LC₅₀ calculated for this species was 130 μg l⁻¹. At pH values of 6.1 and 7.5, mortality was low ( 20%) regardless of aluminum concentrations. A 30-day chronic toxicity test was conducted at three pH levels (low 5.1, intermediate 5.5–5.7 and high 7.3), each with two aluminum concentrations (approx. 0 and 200 μg l⁻¹). Survival was significantly lower in the test at pH 5.1 with aluminum, and at pH 5.7 with aluminum treatments than in the other treatments. Fish in the pH 5.1 without aluminum treatment had intermediate survival, while fish exposed to pH 5.7 without aluminum, pH 7.3 without aluminum and pH 7.3 with aluminum had high, and similar, survival. Sublethal effects on fish exposed to low pH and aluminum included deformities, reduced activity and abnormal swimming behavior. We conclude that the sensitivity of smallmouth bass to low pH and aluminum concentrations corroborates field investigations linking acidification and aluminum mobilization with depletion of smallmouth bass populations.     

Toxicity of sodium selenite to rainbow trout fry

In a study designed to examine the long-term effects of inorganic selenium (IV) on early life stages of rainbow trout (Salmo gairdneri), survival was significantly reduced at selenium concentrations of 47 and 100 μg l⁻¹ after 90 days of exposure. Length and weight were significantly reduced after 90 days of exposure to 100 μg l⁻¹. Whole-body residues of selenium increased with increasing exposure concentrations but appeared to decline between 30 and 90 days of exposure. Analyses of trout backbone indicated little change in bone development with exposure to selenium (IV) with one exception; calcium concentrations were significantly decreased in fish exposed to 12 μg l⁻¹ of selenium. Results of our study indicates that a recommended safe level of 10 μg l⁻¹ for inorganic selenium would not significantly affect growth and survival of rainbow trout; however, concentrations of selenium near this level can reduce the levels of calcium in the backbones of trout.     

Speciation (including adsorbed species) of copper, lead, nickel and zinc in the Meuse River: Observed results compared to values calculated with a chemical equilibrium computer program

The adsorption of trace metals on sediments of the Meuse River was interpreted in terms of competition between metals and protons for surface sites. Surface constants (^*β₁^surf) were determined for Cu, Zn and Cd (10^−1.8, 10^−3.6 and 10^−3.7). The constants for Pb, Ni, Ca and Mg (10^−1.7, 10^−3.8, 10^−6.5 and 10^−5.2) were estimated using a correlation between hydrolysis and surface constants. A chemical equilibrium computer program in which surface sites (for adsorption reactions) are treated as conventional ligands was used to calculate the speciation of Cu, Pb, Ni and Zn in the Meuse River. Calculated values of the adsorbed/dissolved distribution agreed well with observed values, after some realistic data manipulation. This work indicates that dissolved trace metal concentrations in the Meuse River are controlled by adsorption and not by precipitation mechanisms. The relationship between organic matter and suspended matter greatly influences the adsorption of metals like Cu and Pb.     

Chronic toxicity of water-borne and dietary lead to rainbow trout (Salmo Gairdneri) in lake Ontario water

Rainbow trout (Salmo gairdneri) exposed to lead in Lake Ontario water demonstrated a 21-day LC₅₀ of 2.4 mg l⁻¹ lead. At lead concentrations ranging from 3 to 120 μg l⁻¹, log₁₀ of lead concentrations in most tissues of exposed fish appeared linearily related to log₁₀ of lead concentrations in water. Highest concentrations occurred in opercular bone followed by gill and kidney. Lead accumulation by brain was not clearly demonstrated. Exposure to lead in water at concentrations as low as 13 μg l⁻¹ caused significant increases in red blood cell (RBC) numbers, decreases in RBC volumes, decreases in RBC cellular iron content and decreases in RBC δ-amino levulinic acid dehydratase activity. No changes were observed in hematocrit or whole blood iron content. The changes indicated increased erythropoiesis to compensate for inhibition of hemoglobin production and increased mortality of mature red blood cells. After 32 weeks exposure to 120 μg l⁻¹ lead in water, 30% of remaining fish exhibited black tails, an early indication of spinal deformities. Lead added to food was not available for lead uptake by fish. Lead content of fish exposed to dietary lead was not elevated above control levels and the majority of lead consumed could be accounted for in the faeces. Dietary lead may have slightly inhibited uptake of dietary iron.     

Acute and chronic toxicity of lead to rainbow trout salmo gairdneri, in hard and soft water

Lead was found to be highly toxic to rainbow trout in both hard water (hardness 353 mg l⁻¹ as CaCO₃) and soft water (hardness 28 mg l⁻¹. Analytical results differ greatly with methods of analysis when measuring concentrations of lead in the two types of water. This is exemplified in LC50's and maximum acceptable toxicant concentrations (MATC's) obtained when reported as dissolved lead vs total lead added in hard water. Two static bioassays in hard water gave 96-h LC50's of 1.32 and 1.47 mg l⁻¹ dissolved lead vs total lead LC50's of 542 and 471 mg l⁻¹, respectively. In a flow-through bioassay in soft water a 96-h LC50 of 1.17 mg l⁻¹, expressed as either dissolved or total lead, was obtained. From chronic bioassays, MATC's of lead for rainbow trout in hard water were between 18.2 and 31.7 μg l⁻¹ dissolved lead vs 120–360 μg l⁻¹ total lead. In soft water, where exposure to lead was initiated at the eyed egg stage of development, the MATC was between 4.1 and 7.6 μg l⁻¹. With exposure to lead beginning after hatching and swim-up of fry, the MATC was between 7.2 and 14.6 μg l⁻¹. Therefore, fish were more sensitive to the effects of lead when exposed as eggs.     

The effects of phenol and some alkyl phenolics on batch anaerobic methanogenesis

Phenol and seven alkylphenols (o-, m- and p-cresol, 2.5-, 2.6-, 3.4- and 3,5-dimethylphenol) were added at various concentrations to aliquots of domestic anaerobic sludge in Hungate serum bottles and these were incubated at 37°C. The concentration of methane in the headspace gas was monitored to determine if the phenolics were fermented to methane or if they inhibited the anaerobic process. Only phenol and p-cresol were fermented to methane. At 500 mg l⁻¹ (but not at 300 mg l⁻¹) 2,5-, 3,4- and 3,5-dimethylphenol reduced the rate and the amount of methane produced. The cresols were inhibitory at 1000 mg l⁻¹ but not at 400 mg l⁻¹.In cultures supplemented with acetate and propionate (VOA), and in unsupplemented cultures, phenol at concentrations up to 500 mg l⁻¹ was fermented to methane. Between 800 and 1200 mg l⁻¹ phenol, methane production was neither enhanced nor inhibited relative to control cultures containing no phenol. Inhibition of methane production was evident when phenol was present at 2000 mg l⁻¹. Thus the methanogens are less susceptible to phenol inhibition than are the phenol-degrading acid formers. In similar experiments with p-cresol: enhanced methane production was observed at concentrations of 400 mg l⁻¹; no enhancement or inhibition was observed at 600 mg l⁻¹; and inhibition was noted when p-cresol was present at 1000 mg l⁻¹.     

Trace metals in the marine bivalve Macoma balthica in the westerschelde estuary, The Nertherlands. Part 3: Variability of the role of cytosol in metal uptake by the clams

Tellinid clams Macoma balthica were sampled every 2 months for 2.5 years at two locations from the Westerschelde estuary (The Netherlands) and submitted to in vitro short-term exposure to Cd, Cu and Zn. Total and heat-stable cytosolic metal concentrations have been measured before and after exposure to study the effects of the sampling season on metal uptake by the bivalve. We observed much higher uptakes of Cd and, to a lesser extent, Cu in winter than in summer, while Zn appears to be constantly regulated. These phenomena are amplified in the cytosol, especially for Cd, a non-essential element, for which the concentrations after exposure can reach 35–45 μg·g⁻¹ (dry wt.) in winter, but only a maximum of 0.5 μg·g⁻¹ in summer, all natural values being between 0.01 and 0.1 μg·g⁻¹ on average. The contents (μg) of the various elements, either at the total or at the cytosolic level, are not constant, hence showing that the seasonal body weight fluctuation of the clams (dilution or concentration effect) is not the only parameter responsible for the metal uptake variability. Furthermore, we have shown that the proportion of cytosolic metal is not constant, but increases with the absolute cytosolic concentration up to 35–40. Therefore, the cytosolic phase of the cells plays a growing role in Cd, Cu and Zn storage as long as their total concentration increases.     

Trace metals: Cd, Cu, Pb and Zn in gelatinous macroplankton from the Northwestern Mediterranean

Gelatinous macroplankton organisms were collected in May 1984 in Villefranche-sur-Mer Bay and analysed for cadmium, copper, lead and zinc. Analyses were carried out by polarography for Cd, Cu and Pb and by flame atomic absorption for Zn. Phosphorus was also measured in the samples as a biomass parameter due to difficulties inherent in measuring dry weight of gelatinous organisms. The samples belong to the Tunicates, the Cnidarians (Hydromedusae, Siphonophores and Scyphomedusae), the Ctenophores and the Molluscs. Crustaceans living on some Tunicates were also sampled.As regards cadmium, copper and lead, mean concentrations did not show significant differences among the phyla studied: especially for Tunicates with mean values of 0.1 ng Cd μg P⁻¹, 2.0 ng Cu μg P⁻¹ and 0.9 ng Pb μg P⁻¹ and for Cnidarians with mean values of 0.5 ng Cd μg P⁻¹, 2.0 ng Cu μg P⁻¹ and 0.9 ng Pb μg P⁻¹ and for Cnidarians with mean values of 0.5 ng Cd μg P⁻¹, 2.0 ng Cu μg P⁻¹, 1.0 ng Pb μg P⁻¹. On the other hand, mean zinc concentrations were significantly lower in Tunicates (7.9 ng Zn μg P⁻¹) than in Cnidarians (36.8 ng Zn μg P⁻¹).Zinc seems to be preferentially concentrated in organisms which are rich in collagen, constituting the mesoglea, such as the Cnidarians, the Ctenophore and the gelatinous Mollusc studied, rather than in organisms rich in tunicin such as the Tunicates.     

Studies on the toxicity of copper sulphate to stone loach Noemacheilus barbatulus (L.) in hard water

The toxicity of copper sulphate to Noemacheilus barbatulus was studied for 64 days in a water of total hardness 249 mg l⁻¹ as CaCO₃. The 63-day lc₅₀ was approximately 0.25 mg Cu l⁻¹. Larger fish survived longer, and at concentrations greater than 0.29 mg l⁻¹ fish hid less during daylight. Noemacheilus surviving 0.12 mg Cu l⁻¹ for 64 days shed copper when placed in clean water for 7 days: gill, muscle, eye and vertebrae lost significant amounts of copper during this period. The opportunity to shed copper during a short period when the poison supply to their tank failed, may have allowed fish exposed to 0.49 mg l⁻¹ to live 12 days longer than expected. The sensitivity of Noemacheilus to copper, cadmium and zinc was compared with that of rainbow trout. Salmo gairdneri.     

Application de la spectrometrie d'absorption atomique zeeman aux dosages de huit elements traces (Ag, Cd, Cr, Cu, Mn, Ni, Pb et Se) dans des matrices biologiques solides

The use of biological indicators in studies of aquatic pollution (in fresh, estuarine or sea waters), as well as research about the metal transfers in food chains, need a great accuracy of the trace element determination. Therefore, as shown by the results of international intercalibration exercises, the mastery of analytical techniques is far from being perfect in all the laboratories.One of the main sources of error in atomic absorption results from non-specific absorptions due to the presence of important organic and mineral matrixes in biological materials and especially in aquatic and marine organisms. In this case, the correction of unspecific absorption by using deuterium lamp was insufficient and the determination of trace elements had to be preceded by a pre-instrumental stage which allowed the elimination of the organic matter by mineralization and of a large part of the mineral matrix by extraction. The previous separation was long and induced contamination risks. The use of the Zeeman effect background correction allows the transfer of most processes from the pre-instrumental to the instrumental stage. Moreover, the Zeeman effect has three advantages: (1) the background correction is effective up to 2 units of absorbance; (2) the correction is effective from 190 to 900 nm; (3) the method of the double beam is optimalized.The aim of this study was to apply the Zeeman effect to the determination of eight trace elements (Ag, Cd, Cr, Cu, Mn, Ni, Pb, Se) in three different biological solid samples, two originated from the marine environment (lobster hepatopancreas TORT-1, standard reference material from the National Research Council Canada; oyster tissues SRM 1566 from the US National Bureau of Standards) and one from the vegetable kingdom (tomato leaves, SRM 1573 from the US NBS).The experimental procedure is reduced to a minimum since it consists in the digestion of an aliquot of 100 mg of the powdered sample with 1 ml of concentrated nitric acid at 95°C for 1 h. Then the volume is adjusted to 4 ml with deionized water. The metal analysis is carried out using a graphite furnace coated with tantalum carbide.This analysis is achieved according to the method of standard addition. The three added concentrations used for each element are listed in Table 1. The analytical conditions and graphite atomizer program are indicated in Table 2. The temperature program has to be modified according to the type of equipment.The internal quality control of the suggested method related to four criteria: sensitivity, repeatability, accuracy, practicability. The results are shown in Table 3. The threshold of sensitivity (3 times the SD of a series of eight results obtained for a blank of digestion) are low: < 1 μg kg⁻¹ for Ag, Cd and Mn; 1 μg kg⁻¹ for Cr and Pb; 5 μg kg⁻¹ for Cu and Ni and 15 μg kg⁻¹ for Se. The variation coefficients, calculated for both two series of six determinations each, are generally included between 5 and 10%. The trace element concentrations determined by using this method are in perfect agreement with the certified values of the US NBS and NRC Canada (Table 3.)The quality of the results establishes the possibility of using a very easy and fast method to determine the level of eight trace elements in materials with high mineral and organic matter content.     

Toxicological effects of dehydroabietic acid (DHAA) on the trout, Salmo gairdneri Richardson, in fresh water

Toxicological and physiological effects of dehydroabietic acid (DHAA), a major poison to fishes in pulp and paper mill effluents, were studied by two experiments with rainbow trout, Salmo gairdneri Richardson: in the first, fish were acutely exposed for 4 days to an average DHAA concentration of 1.2 mg l⁻¹ (Exp. I) and in the second for 30 days to an average of 20 μg DHAA l⁻¹ (Exp. II).Compared to the controls, fish of Exp. I displayed a decreased relative weight of liver, an increased blood haematocrit, and increased haemoglobin as well as plasma protein concentrations. The aspartate aminotransferase activity of heart muscle was significantly elevated, as was also the lactate dehydrogenase (LDH) of white muscle tissue. In the blood plasma, the proportion of muscle type LDH activity was simultaneously increased. UDP-glucuronyl-transferase activities of liver and kidney were strongly decreased. Results suggest an increased and altered use of body energy reserves, decreased plasma volume and impaired liver function.Fish of Exp. II showed an increased relative weight of spleen. In addition, liver and gill LDH shifted towards heart-type. We conclude that 20 μg l⁻¹ is close to the “minimum effective concentration” of DHAA to rainbow trout.     

Cardiovascular-respiratory responses of rainbow trout (Salmo gairdneri) during chronic exposure to sublethal concentrations of cadmium

The effects of exposure to 3.6 and 6.4 μg l⁻¹ cadmium for periods up to 178 days on cardiac and ventilatory rates, hematocrit, hemoglobin concentration and erythrocyte adenosine triphosphate concentration in adult rainbow trout, Salmo gairdneri, were investigated. Except for slight transitory responses, 3.6 μg l⁻¹ cadmium had no effect on any of the cardiovascular/respiratory parameters. Significant increases in cardiac and ventilatory rates, blood hematocrit and hemoglobin were observed in fish exposed to 6.4 μg l⁻¹ Cd over the entire exposure period while erythrocyte ATP concentration declined during the last stages of exposure. Further experiments on the responses of fish exposed to 6.4 μg l⁻¹ Cd for 30 days demonstrated an impairment of oxygen transfer across the gill. The results are discussed in terms of possible gill impairment and hyperactivity as toxic responses to cadmium.     

Mercury fluxes in a natural forested Amazonian catchment (Serra do Navio, Amapá State, Brazil)

Mercury (Hg total) fluxes were calculated for rainwater, throughfall and stream water in a small catchment located in the northeastern region of the Brazilian Amazon (Serra do Navio, Amapá State), whose upper part is covered by a natural rainforest and lower part was altered due to deforestation and activities related to manganese mining. The catchment area is 200 km from the nearest gold mining (garimpo). Minimum and maximum Hg concentrations were measured monthly from October 1996 to September 1997 and were 3.5–23.4 ng l⁻¹ for rainwater, 16.5–82.7 ng l⁻¹ for throughfall (March–August 1997) and 1.2–6.1 and 4.2–18.8 ng l⁻¹ for stream water, in natural and disturbed areas, respectively. In the natural area, the inputs were 18.2 μg m⁻² year⁻¹ in rainwater and 72 μg m⁻² year⁻¹ in throughfall. This enrichment was attributed to dry deposition. The stream output of 2.9 μg m⁻² year⁻¹ indicates that Hg is being recycled within the forest as other chemical species or is being retained by the soil system, as confirmed by the cumulative Hg burden in the 0–10 cm surface layer, which was 36 480 μg m⁻². When the disturbed area of the catchment was included, the stream output was 9.3 μg m⁻², clearly indicating the impact of the deforestation of the lower part of the basin on the release of mercury. The Hg burden in the disturbed area was 7560 μg m⁻² for the 0–10 cm surface layer.     

Distribution of halomethanes in potable waters of Kuwait

The distribution of bromine containing trihalomethanes in the water distribution system of Kuwait has been studied. Total halomethanes in the drinking water averaged 25.6 ± 9.1 μg l⁻¹ with a maximum of 50.5 μg l⁻¹. Average concentrations (μg l⁻¹) of individual compounds were: CHBr₃, 13.6 ± 4.6; CHBr₂Cl, 8.8 ± 3.7; CHCl₂Br, 3.3 ± 1.5. Water from roof top storage tanks contained significantly less halomethanes than that from underground reservoirs.