The Science of the Total Environment 298 (2002) 25–44 Is chemical contamination responsible for the decline of the copper redhorse (Moxostoma hubbsi), an endangered fish species, in Canada? Yves de Lafontainea,*, Nicolas L. Gilberta, Francois Dumouchela, Charles Brochub, ¸ ´ Serge Mooreb, Emilien Pelletierc, Pierre Dumontd, Alain Branchaude a St. Lawrence Centre, Environment Canada, 105 McGill Street, Montreal, Quebec, Canada H2Y 2E7 ` de l’Environnement, 850 Vanier Blvd., Laval, Quebec, Canada H7C 2M7 Direction des Laboratoires, Ministere c ´ ´ des Ursulines, Rimouski, Quebec, Canada G5L 3A1 INRS-Oceanologie, 310 Allee d ´ ´ ´ ´ Direction Regionale de la Monteregie, Faune et Parcs du Quebec, 201 Place Charles-Lemoyne, Longueuil, Quebec, Canada J4K 2T5 e ´ ´ ´ P.O. Box 8888,Centre-Ville Stn., Montreal, Quebec, Departement des Sciences Biologiques, Universite´ du Quebec a` Montreal, Canada H3C 3P8 b Abstract The copper redhorse (Catostomidae: Moxostoma hubbsi) is an endangered fish species whose worldwide distribution is limited to the St. Lawrence River and three of its tributaries, in Canada. Severe reproductive impairment and lack of successful recruitment reported in this species have been hypothetically associated with water pollution. In order to obtain an initial description of contamination levels in copper redhorse, seven accidentally-killed specimens from the Richelieu River were analyzed for trace metals, organochlorine pesticides, chlorobenzenes, PAHs, PCBs, dioxins and furans. Fish varied between 9 and 33 years of age, which corresponds to mature individuals. The levels of contaminants analyzed in different body tissues were close to and often lower than levels reported in other catostomid fish species from nearby locations within the St. Lawrence River basin. Concentrations of total mercury, cadmium and co-planar PCBs increased with fish age. The types and concentrations of contaminants found suggested that the Richelieu River spawning population of copper redhorse would migrate and spend time in the St. Lawrence River. Concentrations of many contaminants were often highest in gonadal tissues, but levels were much lower than reported in the literature as causing reproductive impairment or egg and fry mortality in fish. Further research is needed to assess the potential link between contaminants and reproductive failure in this endangered fish species.
2002 Elsevier Science B.V. All rights reserved. Keywords: Copper redhorse; Moxostoma hubbsi; Endangered species; St. Lawrence River; Chemical contamination; Organochlorines; Trace metals; PAHs; PCBs; Dioxins *Corresponding author. Tel.: q1-514-496-5025; fax: q1-514-496-7398. E-mail address: yves.delafontaine@ec.gc.ca (Y. de Lafontaine). 0048-9697/02/$ - see front matter
2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 0 2 . 0 0 1 6 6 - 3 26 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 1. Introduction The copper redhorse (Catostomidae: Moxostoma hubbsi) is a fish species that was added to the list of threatened species in Canada in 1987 (Mongeau et al., 1988) and in Quebec in 1999. Officially identified as a distinct species in 1952 (Legendre, 1952), the worldwide distribution of this rare species is restricted to the St. Lawrence River and three of its tributaries, the Richelieu, YamaskaNoire and des Milles-Iles rivers in southern Quebec, Canada (Mongeau et al., 1992). This stenophagous benthic fish feeds almost exclusively on molluscs and has the highest fecundity and the fastest growth rate among the five species of redhorses found in Quebec. Historically more abundant, copper redhorse has dramatically declined during the last century (Branchaud and Jenkins, 1999) and estimates of the actual population size do not exceed 2000 individuals. The species is now reported only in the lower Richelieu River, where it spawns in late June and early July (Mongeau et al., 1992). The reasons for its decline are numerous and habitat destruction, recruitment failure and water pollution have been hypothesized as important causes. The reproductive capacity of the copper redhorse is severely impaired at present, and the female maturation process is not always productive (Branchaud and Gendron, 1993). This would result in a reduced production of offspring and subsequent lower recruitment, as suggested by the rarity of juveniles in the present population and the increasing age of spawners in recent years (Branchaud et al., 1995; Vachon, 1999). Fish reproduction and larval survival can be impaired by the presence of toxic chemicals such as trace metals and organochlorine contaminants (Sangalang and O’Halloran, 1980; Giesy et al., 1986; Grimwood and Dobbs, 1995; Kelce et al., 1995; Donohoe and Curtis, 1996; Kavlock et al., 1996). Field studies have reported that fish spawning was less successful in populations heavily contaminated by metals and pesticides (McFarlane and Franzin, 1978; Munkittrick and Dixon, 1988; Spies and Rice, 1988; Hose et al., 1989). However, other laboratory studies using environmentally relevant concentrations of toxic substances have failed to demonstrate a significant effect on fish reproduction or other histopathological indicators (Fonds et al., 1995; Friedmann et al., 1996; Lesko et al., 1996). The levels of contamination in copper redhorse are unknown, and the exposure of this species to various contaminants has never been documented. Specimens may be exposed to contaminants through their benthic diet, consisting mostly of molluscs, including bivalves, which can bioconcentrate various contaminants (Metcalfe and Charlton, 1990; Metcalfe-Smith et al., 1992; de Lafontaine et al., 2000; Regoli et al., 2000). The Richelieu River is affected by various sources of both agricultural and industrial pollution, including inputs of copper, zinc, PCBs, furans and PAHs, mostly in its upper reaches. Although the number of contaminants detected in the water was high, levels were often lower than those measured in other rivers in Quebec (Berryman and Nadeau, 1998). Mercury levels in lower Richelieu River sediments (0.06 mgyg) are within the range of values considered as natural (-0.10 mgyg) (Paul ´ 1989a) and lower than those measand Laliberte, ured in the St. Lawrence River. PCB and DDT levels in sediments of the Richelieu River were also lower than in St. Lawrence River sediments ´ 1989b). The exact migration (Paul and Laliberte, patterns and present distribution of copper redhorse have not been fully described (Mongeau et al., 1992) and it remains to be proven that the Richelieu River spawning population is isolated from the St. Lawrence River. The movement of fish into the St. Lawrence River cannot be ruled out and would therefore increase the risk of their exposure to higher levels of contaminants. Recent efforts to protect and restore the copper redhorse have been impeded by the lack of information on the possible link between contaminant levels and reproductive problems in this species. Its endangered status makes it difficult, if not untenable, to kill specimens for chemical analysis. However, the opportunity of using accidentallykilled specimens allowed for a first description of contamination levels in copper redhorse. The present study was undertaken to determine the type and the levels of trace metals, polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 (PCBs), chlorobenzenes (CBs), organochlorine pesticides (OCPs) and polycyclic aromatic hydrocarbons (PAHs) in body tissues of copper redhorse and to compare them with levels reported in other fish species of the St. Lawrence River. This information is considered a baseline data set for future comparison. 2. Materials and methods 2.1. Sample collection and preparation Seven specimens were used in the present study. They were accidentally killed during field operations to capture individuals for artificial spawning induction. All the fish were caught at the downstream end of the Saint-Ours dam on the Richelieu River (458 519 500 N; 748 89 540 W) between 3 June and 18 June 1993. Total length (cm), total wet weight (g), sex and presence of particular external characteristics were recorded for each individual. The operculum was taken for age determination. Fish were preserved at y20 8C for 9 months prior to their dissection in the laboratory. Pieces of liver, gonad and muscle tissue were taken to be analyzed for contaminants. After thawing, fish liver and gonad tissues were first dissected out, weighed and homogenized with an Omni International 1000 homogenizer equipped with a medium-sized grinder. Muscle tissue ()40 g) was taken from the latero-dorsal and caudal sections of the fish, taking care to exclude any bone structure. After being cut into small pieces with stainless steel scissors, the tissue was homogenized with a Black and Decker HC-200 meat grinder. Homogenates selected for analysis of inorganic contaminants were put into nitric-acid-rinsed polypropylene containers, while those taken for organic contaminants analysis were placed in dark hexane-rinsed glass containers (Environment Canada, 1993). Homogenates were kept frozen at y 20 8C until analysis. Muscle tissues were selected for analysis of mercury (Hg) and other trace metals (As, Cd, Cr, Cu, Pb, Ni and Zn), PCB congeners, dioxins and furans. Depending on the amount of tissue available, both liver and gonad homogenates were analyzed for trace metals (excluding Hg), PCBs, PAHs and 20 OCPs. 27 2.2. Trace metal analyses Metals were analyzed using Environment Canada (1993) standard methods. Mercury was analyzed in 1-g samples by cold-vapor atomic absorption spectrophotometry after mercury reduction and vaporization with stannous chloride (Method CPQ111BO). For As, Cd and Pb, samples (2–5 g) were analyzed by graphite furnace atomic absorption spectrophotometry (GF-AAS) after digestion in concentrated sulphuric and nitric acid for at least 12 h (Method CPQ112BO). Other metals (Cr, Cu, Ni and Zn) were analyzed by atomic emission spectroscopy using induction-coupled argon plasma excitation source (ICP-ES) (Method CPQ110BO). Quality control consisted of method blanks to monitor analytical contamination, certified reference materials to measure precision and accuracy, and the inclusion of sample replicates. Detection limits depended on the sample volumes being analyzed: 30 ngyg for Hg, 80–140 ngyg for As, 4–5 ngyg for Cd, 600–800 ngyg for Cr, 240 ngy g for Cu and Ni, 80–100 ngyg for Pb and 100 ngyg for Zn. Recovery rates were 70% for Hg and ranged from 90 to 99% for other metals. Coefficients of variation (CVs100=S.D.ymean), as a measure of precision, were less than 5%, except for Cr, Pb, As and Cd, with values of 10, 11, 13 and 14%, respectively. All results were corrected for percent recovery. 2.3. Dioxins, furans and coplanar PCBs Sample homogenates (2–10 g) were dried with sodium sulfate (125 g), blended and spiked prior to extraction with a mixture of nine 13C12-labelled 2,3,7,8-substituted PCDDyPCDF, the three 13C12 coplanar PCBs (IUPAC No. 77, 126, 169) and six 13 C12 ortho-substituted PCBs. Sample extractions were performed in a glass column with 650 ml of hexaneymethylene chloride (50:50). Fractionation of PCBs, planarqmono-ortho PCBs and PCDDy PCDF was performed on alumina columns as described by Sakai et al. (1993). An additional clean-up of the PCDDyPCDF fraction was performed over a carbon column according to the method of Smith et al. (1984). PCDDyPCDF and 28 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 planar PCB fractions were evaporated to dryness and 50 ml of an internal standard solution was added prior to analysis. Analyses were done by HRGCyHRMS on a Hewlett-Packard 5890 GC interfaced to a VG AutospecQ (VG AnalyticalyFisons Manchester, UK). Detection limits were 0.1–0.2 pgyg and the recovery rate of labelled substitutes varied between 60 and 100%. All data were corrected for percent recovery and expressed in pgyg of wet weight. 99% for OCPs. All results are expressed in ngyg of wet weight and are corrected for percent recovery. PAHs were analyzed by the method described in Wise et al. (1991). Briefly, homogenates were blended with sodium sulfate, extracted in dichloromethane and fractionated on silicayalumina columns. The PAH-containing fraction was separated by HPLC and analyzed by GCyMS. 2.5. Data analysis 2.4. PCBs, OCPs, CBs and PAHs Analytes (2–10 g) were extracted with an acetoneyhexane (1:4) solution. Extracts were dried with sodium sulfate, filtered and concentrated to 10 ml by rotary evaporation. Two millilitres of extract was used for lipid content determination by a gravimetric method. The remainder of the extract was concentrated, purified by gel permeation chromatography and fractionated on a Florisil 60–100 column with three eluents: hexane, ether– hexane (5:95), and ether–hexane (50:50). Fractions were analyzed using a HP-5890 Series II capillary column gas chromatography (GC) with a splitless injector and electron capture detection and equipped with two fused silica columns (DB5 and DB-1780, 30 m=0.32 mm I.D.). The PCB congeners analyzed were: IUPAC Nos. 101, 105, 118, 128, 138, 153, 156, 167, 170, 180, 183, 189 and 194. In addition to the three coplanars determined with dioxins and furans (see above), the list of PCB congeners analyzed included those belonging to priority groups Ia, Ib and II as defined by McFarland and Clarke (1989), and considered most environmentally harmful. Quality control consisted of method blanks to monitor contamination, method spikes to monitor accuracy (% recovery), and replicated measurements of certified reference materials to measure precision. Spiked samples included PCB congeners, OCPs and CBs. Detection limits were 0.2– 0.3 ngyg for PCB congeners, 0.2–0.5 ngyg for OCPs, 0.2 ngyg for hexachlorobenzene, 0.3 ngyg for pentachlorobenzene, 1 ngyg for tetrachlorobenzenes, and 50 ngyg for tri- and di-chlorobenzenes. Percent recoveries were greater than 75% for PCB congeners, from 60 to 80% for CBs, and 70 to Three physiological indices were calculated from morphological variables: Fulton’s condition factor K s 100000•W L3 Gonado-somatic index GSI s 100 • Wg WyWg Hepato-somatic index HSI s 100 • Wl WyŽWgqWl. where W, Wg, Wl and L are, respectively, total weight (in g), gonad weight (in g), liver weight (in g), and total length (in mm). Concentrations of separate organic congeners were summed to determine total contaminant levels: SPCB, SPCDD, SPCDF and SPAH. In the case of OCPs, the summed groups were: SHCH: a-HCH, b-HCH, g-HCH (lindane); SDDT: p,pDDE, o,p-DDD, p,p-DDD, o,p-DDT, p,p-DDT; SChlordane: cis-, trans- and oxy-chlordane. Values below the detection limit were set to equal zero for calculation of summation, but were assigned to one-half of the detection limit for the statistical treatment of individual compounds. Descriptive statistics were calculated for each parameter when more than 50% of values were above the detection limit. Because of the small sample size, normal distribution of data could not be confirmed and consequently all reported averages are geometric means. Inter-individual variability was provided by the coefficient of variation (CVs100=S.D.yarithmetic mean). Relationships between contaminant tissular levels and fish characteristics were evaluated by Spearman rank correlation analyses (Conover, 1980). Comparisons between tissues were Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 tested by the non-parametric Kruskal–Wallis test (Conover, 1980) or by ANOVA using log-transformed data when sample size permitted. In the case of OCPs, the relationship between contaminant levels and lipid percentage was first examined by visual inspection and by correlation analysis, and comparison between tissues was based upon the use of an analysis of covariance (ANCOVA), taking percent lipid as a covariate, as recommended by Hebert and Keenleyside (1995). Model-I regression (Sokal and Rohlf, 1981) was used to develop a predictive linear relationship between contaminant concentration and a given dependent variable. All statistical analyses were performed using a SAS software program. As there was no reference site or control population to compare our results with, data were compared to other fish species of the St. Lawrence River (Ion and de Lafontaine, 1998; de Lafontaine ´ 1998) et al., 1999), the Richelieu River (Piche, and nearby water bodies (Metcalfe-Smith et al., 1995). Whenever possible, comparisons were made to white sucker (Catostomus commersoni), a closely related benthic freshwater species belonging to the same family (Catostomidae). 3. Results 3.1. Fish characteristics Specimens ranged from 53.7 to 69 cm in length (means59.4; CVs8%) and from 1810 and 5360 g in weight (means3605; CVs30%). Fish age was more variable, ranging from 9 to 33 years (means18.8; CVs39%). These values are within the range reported for this species in the Richelieu River in recent years (Mongeau et al., 1986). Weight and length were positively correlated (rs 0.90, P-0.01) and both significantly increased with fish age (rs0.94 for weight and rs0.98 for length; P-0.001). Except for one low value (1.169) observed in the youngest fish, the condition index (K) varied between 1.63 and 2.05 (means1.72; CVs17%). These values are slightly higher than the average K values of 1.47 and 1.58 calculated for males and females, respectively, by Mongeau et al. (1986). K was not significantly related to other parameters. 29 Three females and three males were collected; one fish could not be sexed. The Gonado-somatic Index (GSI) was highly variable between individuals, ranging from 0.32 to 8.2, with only two values)5 (means1.62, CVs104%). These estimates are somewhat lower than those of prespawning individuals, which generally have GSI) 5 in June (Mongeau et al., 1992). This would indicate that our specimens were not quite ready to spawn when captured or that reproduction may have been impaired. HSI values varied between 0.80 and 7.03, with a mean of 2.75 (CVs87%), but were not significantly related to other fish characteristics. Lipid content differed between tissues and was higher in gonads (ranges4.3–23.0%) than in liver (1.1–5.0%) or muscle (0.85–4.2%). 3.2. Trace metals Mercury, cadmium, copper and zinc were commonly detected in all tissues analyzed. Levels of Hg in muscle varied between 86 and 359 ngyg (geometric means241 ngyg, CVs38%; Table 1) and were positively correlated with fish age (r 2s 0.77, Ps0.03; Fig. 1) as well as length, albeit to a lesser extent (r 2s0.60, Ps0.07). No significant correlation was found with other fish characteristics. Cadmium levels were significantly higher in liver than in muscle or gonads (ANOVA Fs17.1, Ps0.0025). The variation between individuals, however, was relatively high (CVs150%), and largely due to the very high concentrations found in all tissues of the oldest fish (33 years old). The correlation between Cd and fish age or length was significant (P-0.05), but not when the oldest fish was removed from the statistical analysis. Cadmium levels were not related to other fish characteristics. Levels of Cu and Zn in muscle varied little between individuals and were significantly lower than in liver or gonads (Table 1). Zinc levels were highest in ovaries (3 times higher than in testes). Concentrations of Cu and Zn were not significantly correlated (P-0.05) with fish age, size or condition factor, but levels in liver varied inversely with liver weight (Spearman correlation rssy0.9, P0.001). The liverymuscle ratio of concentrations 30 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 was relatively constant for Zn (1.5–3.0), but much more variable for Cu (1–405). Levels of Hg and Cd in muscle were positively correlated (rss0.98, P-0.05), while concentrations of Cd, Zn and Cu in liver were also related to each other (rss0.9, P-0.05). Lead (Pb) was frequently detected in muscle, but only once in liver and never in gonads (Table 1). Levels in muscle were, however, low and close to detection limits and the difference between tissues may be related to variable detection limits. Levels of As, Cr and Ni were always below detection limits in all tissues. The detection limits of Cr and Ni were probably too high to allow for adequate quantification of these metals in fish tissues. Fig. 1. Mercury contamination (ngyg) in copper redhorse muscle tissue as a function of fish age. 3.3. PAHs 3.4. Dioxins and furans PAHs were rarely detected except for one liver sample with total PAHs exceeding 38 000 ngyg, and of which naphthalene contributed 89%. Neff and Burns (1996) have warned about naphthalene being a common laboratory contaminant which can appear randomly in tissue samples and, therefore, recommended to exclude it when calculating total PAHs. In following this recommendation, the levels of total PAH changed, ranging from below detection limits (-45 ngyg) to 4000 ngyg. Acenaphthene, anthracene, phenanthrene, benzo(a)pyrene and dibenzo(ah)anthracene were detected in all liver samples, but concentrations did not exceed 300 ngyg. Traces (-0.5 pgyg) of dioxins were detected in only one muscle sample. Among furans, only T4CDFs were detected and levels in muscle ranged from 1.5 to 4.6 pgyg (means2.6 pgyg, CVs 45%) (Table 2) and were not related to any fish characteristic. The 2,3,7,8 isomer accounted for ) 85% of total T4CDFs. In one specimen, T4CDF levels in muscle (1.6 pgyg), liver (2.4 pgyg) and ovaries (2.2 pgyg) were very similar, with an overall CV of 20%. The variability between the three tissues was even lower (CVs8.5%) when concentrations were standardized on a lipid weight (LW) basis (muscles48.5 pgyg LW; livers48.0 Table 1 Geometric mean and range (in parentheses) of trace metal concentrations (ngyg) in body tissues of copper redhorse Cd Cu Hg Pb Zn As Cr Ni a b Muscle (ns6)a Liver (ns5)a Testes (ns2)b Ovaries (ns2)b 4.44 (-4–15) 597 (543–672) 241 (86–359) 103 (-70–206) 6780 (5480–7910) -140 -700 -240 122 (22–2040) 2520 (510–20220) NA. -80 (-80–184) 15050 (10100–25600) -140 -700 -240 -8–318 1237–1267 NA. -100 11400–13600 -200 -700 -240 -4–8 905–2285 NA. -100 30400–44400 -140 -700 -240 Geometric mean (range); NA, not analyzed. Range. Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 31 pgyg LW; ovariess41.5 pgyg LW). The betweentissue (within one individual) variability was thus lower than the between-individual variability for muscle tissue. 3.5. Planar PCBs PCB congener 77 was the predominant coplanar PCB and was 7.5 times more concentrated than congener 126 (Table 2). Congener 169 was never detected. Congeners 77 and 126 were strongly correlated to each other (Model-I regression: PCB126sy0.0117q0.1763 PCB77, r 2s0.94, ns6, P-0.05). PCB77 in muscle varied between 110 and 1100 pgyg (means377 pgyg, CVs 75%), the maximum being measured in the oldest fish (Fig. 2). Variability in PCB77 between tissues within one fish was equal to that measured between fish (CVs24%) and was reduced (CVs 12%) when levels were standardized for lipid weight (3.33, 3.6 and 2.83 ngyg LW for muscle, liver and ovaries, respectively). Using the toxic equivalents factors (TEF) provided by Walker and Peterson (1991) for congeners 77 and 126 and by Zabel et al. (1995) for congener 169, mean concentrations of 2,3,7,8 T4CDF and coplanar PCBs in muscle of copper redhorse were calculated as equivalent to 0.38 pgy g of 2,3,7,8 T4CDD (Table 2). Congener 126 accounted for 65% of these toxic equivalents. Values ranged from 0.14 to 1.2 pgyg and the oldest Fig. 2. Concentration (pgyg) of PCB 77 in muscle tissue of copper redhorse as a function of age. fish had the highest toxic burden equivalent. An examination of the data indicated that toxic equivalent levels tend to be higher in male fish, but the low sample size precluded statistical testing. 3.6. Mono and di-ortho PCBs Total PCBs (Aroclor 1254) in muscle varied by a factor of 4.4, ranging from 86 ngyg to 380 ngy g among individuals with a geometric mean of 212 ngyg (Table 3). Levels were slightly lower in liver and higher in gonads, but variability in these two tissues was higher than in muscle. Although Table 2 Geometric mean and range (in parentheses) of PCDF and planar PCB concentrations expressed in pgyg wet weight and in pgyg lipid weight (LW), and 2,3,7,8-TCDD toxic equivalents (TEQs) in copper redhorse muscle. TEF is the toxic equivalent factor Concentration in muscle TEF TEQ (pgyg 2,3,7,8-TCDD) (pgyg) (pgyg LW) PCDFs Total T4CDFs 2,3,7,8 T4CDF 2.6 (1.5–4.6) 2.3 (1.0–4.3) 119.4 (48.5–192.3) 103.9 (30.3–192.3) 0.028 0.063 (0.028–0.12) Coplanar PCBs 77 126 169 377 (110–1100) 50 (18–200) NC (-0.1–1.4) 17397.6 (3333–84615) 2312 (545–15385) NC 0.00016 0.005 0.000041 0.06 (0.018–0.176) 0.25 (0.09–1.0) NC Total NC, not calculated (F50% values above limit of detection). 0.38 (0.14–1.2) 32 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 Table 3 Geometric mean and range (in parentheses) of PCB congeners (ngyg), Aroclor-1254 (ngyg) and lipid content (% wet weight) in body tissues of copper redhorse PCB Muscle (ns6) Liver (ns6) Testes (ns3) Ovaries (ns3) 101 105 114 118 123 128 138 153 156 157 167 170 180 183 189 194 Scongeners Aroclor 1254 Lipids (%) 8.4 (4.3–20.8) 4.0 (-1.4-18.5) NC (-0.3–0.8) 10.4 (3.9–43.4) 0.9 (-0.3–6.9) 1.4 (-0.6–6.3) 9.7 (3.6–33.7) 18.5 (7.1–58.4) 1.7 (-0.6–7.4) NC (-0.3–0.6) 1.0 (-0.6–4.7) 5.0 (1.5–16.7) 5.3 (1.3–14.3) 2.9 (1.0–9.2) -0.3 0.6 (-0.3–1.2) 73.1 (29.1–242.4) 212 (86–380) 3.3 (0.9–18.5) 2.0 (0.5–23.0) -0.3 3.9 (0.7–37.1) -0.3 0.7 (-0.3–7.4) 6.9 (1.5–60.2) 6.0 (1.3–53.2) NC (-0.3–4.1) -0.3 NC (-0.3–2.7) 1.9 (0.5–16.8) 3.5 (0.8–32.1) 1.0 (-.3–7.7) -0.3 0.5 (-0.2–4.6) 31.6 (7.6–268.0) 87 (20–690) 38.1 (18.7–56.0) 34.2 (11.1–74.3) -0.3 61.9 (22.8–113.4) -0.3 12.1 (3.9–30.9) 83.5 (30.0–159.1) 68.4 (21.2–169.4) 8.4 (3.4–16.0) -0.3 5.3 (1.6–12.0) 25 (7.6–55.1) 47.4 (13.3–114.3) 14.5 (4.3–31.0) NC (-0.3–1.3) 7.0 (1.7–18.6) 410.6 (139.8–848.4) 1237 (430–2200) 8.0 (3.5–25.3) 5.8 (2.4–18.2) -0.3 10.2 (3.4–35.6) -0.3 2.4 (1.0–6.3) 16.5 (7.5–61.1) 15.5 (9.7–37.4) NC (-0.3–0.7) -0.3 0.8 (-0.3-4.0) 4.5 (1.8–19.5) 8.6 (3.6–32.2) 2.2 (0.9–9.9) -0.3 1.0 (0.4–4.3) 80.3 (39.2–254.1) 218 (100–740) 2.2 (0.85–4.2) 1.9 (1.1–5.0) 4.3–23 7.2–13 NC, not calculated (F50% values above limit of detection). PCBs were highest in the oldest fish, levels were not significantly related to fish characteristics, except for a weakly significant negative correlation (Spearman rssy0.83, Ps0.045) between liver concentrations and liver weight. Pooling data of all tissues showed total PCBs to be significantly correlated with lipid content (Spearman rss0.56, Ps0.016). Results of ANCOVA indicated that the difference in concentrations between tissues was mainly due to the variability in lipid content of each tissue and fish age. ANCOVA-adjusted means of PCBs in each tissue showed that levels of Aroclor 1254 in testes (14 566 ngyg LW) were not significantly different than in other tissues (ovariess9448 ngyg LW, livers9062 ngyg LW, muscles11498 ngyg LW). Non-parametric tests performed on PCB levels standardized for lipid content produced similar results, indicating no significant difference between tissues (Kruskal– Wallis test, P-0.05). The relative proportion of mono- and di-ortho PCB congeners varied significantly between tissues (Kolmogorov–Smirnov test, P)0.05). PCB153 was the dominant congener in muscle tissue and was almost twice as high as the next three most important congeners (118, 138 and 101). In liver and gonads, congener 138 ranked first but was only slightly more concentrated than congener 153, the latter being 1.5–1.7 times more abundant than congeners 118 and 180. 3.7. Organochlorine pesticides DDT and its derivatives, dieldrin, cis-chlordane, and mirex were detected in all liver and gonad samples (Table 4). Total DDTs were the most concentrated organochlorine pesticides and averaged 16.6 ngyg, 42.8 ngyg and 303.8 ngyg in liver, ovaries and testes, respectively. Endrin, aand g-HCH (lindane), heptachlor epoxide, transand oxy-chlordane, photomirex and octachlorostyrene were detected in varying proportions and more frequently measured in gonads than in liver. Their levels rarely exceeded 1 ngyg. Residues of aldrin, heptachlor, methoxychlor, bHCH and a- and b-endosulfan were never found. Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 33 Table 4 Frequency of detection ()DL), geometric mean and range of organochlorine pesticides (ngyg) and chlorobenzenes in copper redhorse Liver (ns6) Testes (ns3) Ovaries (ns3) mean (min–max) )DL mean (min–max) )DL mean (min–max) Organochlorine pesticides a-HCH 3 b-HCH 0 g-HCH (Lindane) 0 0.27 (-0.3–0.75) -0.5 -0.3 3 0 2 0.79 (0.37–1.9) -0.5 0.49 (0.37–1.7) 3 0 2 0.56 (0.36–1.3) -0.5 0.41 (-0.3–0.97) cis-Chlordane trans-Chlordane Oxychlordane Heptachlor epoxide SChlordane Ratio cisytrans 6 5 6 1 1.9 (0.47–12.5) 0.48 (0.13–2.2) 0.77 (0.21–3.5) NC (-0.3–0.39) 3.2 (0.77–16.7) 6.6 (1.3–17.8) 3 3 3 3 28.7 (14.1–43.6) 4.6 (2.0–10.0) 8.1 (4.1–13.3) 1.3 (0.4–4.0) 44.2 (23.5–71.0) 2.9–19 3 3 3 2 4.5 (1.6–13.2) 0.98 (0.46–1.8) 1.3 (0.44–5.2) 0.46 (-0.3–1.6) 7.2 (2.5–21.9) 3.5–7.2 o,p9-DDT p,p9-DDT o,p9-DDD p,p9-DDD p,p9-DDE SDDT ratio DDEySDDT 5 6 3 6 6 1.0 (0.27–5.5) 3.1 (0.83-9.4) 0.72 (.33-1.5) 3.9 (1.9-20.7) 7.1 (1.2-51.8) 16.6 (4.1-88.9) 0.46 (0.31-0.72) 3 3 3 3 3 21.2 (6.9–54.1) 52.2 (14.3-107.0) 2.3 (1.6-3.3) 46.7 (22.5-98.0) 174.4 (90-287.5) 303.8 (136-535.7) 0.57-0.71 3 3 1 3 3 2.4 (0.71–9.6) 7.9 (3.2-25) NC (-0.7-4.2) 5.8 (0.83-37.7) 25.4 (6.9-103.7) 42.8 (11.6-180.4) 0.62-0.65 Dieldrin Endrin Mirex Photomirex Octachlorostyrene 6 3 6 1 1 1.3 (0.47-2.8) NC (-0.4-0.99) 0.52 (0.16-3.7) NC (-0.5-1.2) NC (-0.5-0.71) 3 3 3 2 2 6.6 (2.0-22.7) 1.45 (0.65-2.5) 5.0 (0.8-12.8) 2.3 (-0.5-6.6) 1.23 (-0.4-2.4) 3 3 2 0 1 3.7 (1.6-17.3) 0.6 (0.4-1.0) 0.4 (-0.2-2.5) -0.5 NC (-0.3-0.7) Chlorobenzenes Pentachlorobenzene Hexachlorobenzene Lipids (%) 1 5 0.27 (-0.2-0.81) 0.33 (-0.2-0.91) 1.9 (1.2-5.0) 3 3 0.63 (0.46-0.95) 1.74 (1.1-3.7) 4.3-23 3 3 0.60 (0.33-1.27) 0.91 (0.32-1.86) 5.3-13 )DL NC, not calculated. Due to differences in lipid content among tissues, levels of pesticides were always higher in gonads than in liver. Testes were generally more contaminated than ovaries. The differences between tissues were greatly reduced when accounting for lipid percent variation (ANCOVA, P-0.05), but levels of chlordane and DDT products in testes appeared significantly higher than in other tissues. Except for photomirex, all organochlorines were strongly correlated to each other (Spearman rss 0.55–0.94, P-0.05). Although maximum concentrations were often measured in the oldest specimen, correlations between OCP levels and fish age or other fish characteristic were not significant (Spearman correlation, P)0.05). The relative proportion of DDT, DDD and DDE was, on average, 0.26, 0.22 and 0.52, respectively, and did not vary among tissues. The p,p-DDE contributed for 31–72% of the total DDT (means 52%). The DDEyDDT ratio was also independent of fish characteristics. Cis-chlordane contributed 41–76% of total chlordane and the relative proportion increased with fish age. The cis-ytransratio in chlordane varied between 1.3 and 18.9 (means6.6) and tended to increase with fish age or length (Spearman rss0.66 and 0.74, P-0.05). This mean value is very close to the ratio of 8.2 observed in a laboratory study of the bioaccumulation of chlordane in shorthead redhorse (Moxotoma macrolepidotum) (Roberts et al., 1977). The 34 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 a-yg-ratio for HCH in gonad samples ranged from 0.91 to 2.31, averaging 1.45 (CVs38%). 3.8. Chlorobenzenes Among CBs, pentachlorobenzene (PenCB) and hexachlorobenzene (HCB) were always detected and significantly (ANOVA, P-0.05) more concentrated in gonads than in liver, where detection frequency was 1 out of 6 and 5 out of 6, respectively (Table 4). Concentrations were slightly higher in male gonads. The difference between tissues was mainly related to the difference in lipid content. Levels of PenCb and HCB were positively related to the lipid content of the samples (Spearman rss0.80 and 0.75, respectively, P-0.05) and results of ANCOVA adjusting for lipid content revealed no significant difference (P)0.05) in HCB and PenCB concentrations between tissues (HCB: livers16.9 ngyg LW; gonadss15.8 ngyg LW; PenCB: livers14.1 ngyg LW; gonadss7.8 ngyg LW). Neither of these two compounds was correlated with any fish characteristic or morphological parameter (Spearman correlation, P) 0.05). Di-, tri- and tetra-chlorobenzenes were never detected in our samples. This may be due to the high detection limits used in the present analysis andyor the high water solubility (low octanol– water coefficient Kow) and the low bioaccumulation potential of these compounds in fish tissues (Axelman et al., 1995; Legierse et al., 1998). 4. Discussion Despite their relatively old age, specimens of copper redhorse from the Richelieu River did not appear to be highly contaminated by bioaccumulative toxic substances. Indeed, their contaminant levels were similar to and even lower than those reported in younger specimens of other catostomid species, including other Moxostoma species from the Yamaska River basin (adjacent to the Richelieu River) (Metcalfe-Smith et al., 1995) and white sucker (Catostomus commersoni) from the Riche´ 1998) and the St. Lawrence lieu River (Piche, River (Ion and de Lafontaine, 1998; de Lafontaine et al., 1999) (Tables 5 and 6). The positive relationship between mercury concentration in muscle and age of copper redhorse (Fig. 1) is consistent with similar observations in white sucker, yellow perch and pike (Harrison and Klaverkamp, 1990; Berninger and Pennanen, 1995; Ion et al., 1997; de Lafontaine et al., 1999), providing evidence of mercury bioaccumulation in fish muscle tissue. The moderately high Hg levels in copper redhorse were mainly due to the older age of these fish relative to other fish from the St. Lawrence River. Using the predictive relationship derived from Fig. 1, we extrapolated Hg levels for fish between 5 and 10 years old to be 100–150 ngyg, which corresponds to half of that measured in white sucker of similar age from the St. Lawrence River (Table 5). This would suggest that copper redhorse were characterized by a lower bioaccumulation rate andyor a weaker exposure to Hg than fish from the St. Lawrence River. Levels of Cd, Cu and Zn in copper redhorse were always higher in liver and gonads than in muscle, as usually observed in fish bioaccumulation studies (McFarlane and Franzin, 1980; Miller et al., 1992; Mersch et al., 1993; Berninger and Pennanen, 1995; Goldstein and De Weese, 1999). For Cd, De Conto et al. (1997) showed that longterm exposure of carp (Cyprinus carpio) to high concentrations of dissolved cadmium increased the concentration in liver up to 10 times that in muscle. These authors concluded that Cd accumulation in muscle would be stimulated only when the storage capacity limits of the liver (;20 000–25 000 ngy g in carp) and kidney are reached. McFarlane and Franzin (1980) had previously shown that Cd levels in liver tissue of white suckers increase with age, but the slope of the relationship varies between sampling locations in response to the pH of the water. Our results, Cd levels being 27 times higher in liver than in muscle and concentrations in liver increasing with age in copper redhorse, are consistent with the results of previous studies. Liver is the preferred tissue for concentrating Cu and Zn, two essential metals subject to physiological regulation in fish (Mersch et al., 1993; Blanchard et al., 1999). In their analysis of eight tissues of white suckers from different lakes, Miller et al. (1992) concluded that muscle was the poorest indicator of Cu and Zn exposure while liver Table 5 Comparison of trace metal concentrations between copper redhorse and other species of sucker (Catostomidae) Location Year Age Hg (muscle) Cd (liver) Cu (muscle) Cu (liver) Zn (muscle) Zn (liver) Reference Copper redhorse Richelieu River 1993 9–33 241 122 597 2520 6780 15050 This study White sucker Richelieu River (upper reaches) Richelieu River (lower reaches) St. Lawrence River (Quebec City) St. Lawrence River (Lake St. Francois) St. Lawrence River (Lake St. Louis) Ontario lakes (polluted lakes) Ontario lakes (control lakes) Flin Flon (smelter) Manitoba (moderately contaminated) Saskatchewan (slightly contaminated) Prairie lakes Assiniboine River 1995 ? 160 ´ 1998 Piche, 1995 ? 250 ´ 1998 Piche, 1995 3–11 239 1989 5–10 172 459 3606 Ion and de Lafontaine, 1998 1989 2–10 312 275 3467 Ion and de Lafontaine, 1998 1988 9 600–1000 18800–19600 3800–5800 48000–55000 Miller et al., 1992 1988 6–8 600–800 9200–10200 4200–4400 24000–24600 Miller et al., 1992 1982 1982 7–11 5–12 20 30 280 30 220 250 20200 4300 5500 3000 43000 19000 Harrison and Klaverkamp, 1990 Harrison and Klaverkamp, 1990 1982 7–10 60 150 340 10200 3500 23000 Harrison and Klaverkamp, 1990 Shorthead sucker Prairie lakes Assiniboine River Longnose sucker Northern Quebec lakes 195 310 6290 4400 21100 de Lafontaine et al., 1999 180 240 370 220 4220 3900 Green and Beck, 1995 Green and Beck, 1995 140 390 280 330 4590 4560 Green and Beck, 1995 Green and Beck, 1995 200–210 2550–3410 SOMER, 1993 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 Species 35 36 Species Site Year Fish age Lipids (%) PCB congeners 105 118 138 153 Photo mirex Dieldrin SDDT Ratio DDEy SDDT 87 (20–690) 0.52 (0.16–3.7) -0.5–1.2 1.3 16.6 0.46 This study 100.1 0.72 Metcalfe-Smith et al., 1995 Richelieu River 1993 9–33 1.9 (1.2–5.0) 2.0 3.9 Redhorse sp. Yamaska River 1986 – n.d. 19.9 35.0 130.1a (114–146) Redhorse sp. Noire River Richelieu River 1986 4 – (2–6) 1995 5–10 5.1 2.0 (2.8–10.8) White sucker St. Lawrence River (Cornwall) 1994 3–11 White sucker St. Lawrence River (Quebec City) 1995 3–11 6.8 4.8 (1.7–17.0) White sucker Lake Ontario 1988 Large White sucker a Geometric mean calculated from the sum of 56 PCB congeners 6.0 Mirex Copper redhorse 9 (3–19) 6.9 Aroclor 1254 13.8 16.8 23.2 135.5a (49–263) 16.7 7.2 14.3 414 (40–1750) 260–290 2.0 1.7–2.1 10.2 13.6 18.3 304 1.7 (47–2500) (-1.3–9) 12–41 2.6 2.6 (-1.4–20) 8–44 Metcalfe-Smith et al., 1995 209.7 0.91 (82–509) 102 0.76 ´ 1998 Piche, 17–21 Ridgway et al., 1999 33.5 0.93 (3.4–283) de Lafontaine et al., 1999 Sergeant et al., 1993 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 Table 6 PCB and organochlorine pesticide concentrations (ngyg) in catostomid livers Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 concentrations appeared to be a reliable indicator of chronic exposure. Data from Miller et al. (1992) showed that the liverymuscle ratio ranged from levels of 12.7 to 31.3 and from 5.7 to 12.6 for Cu and Zn, respectively, with the highest values being associated with more contaminated lakes (see our Table 5). Ratios (Cus20.3; Zns4.8) for white sucker from the St. Lawrence River near Quebec City were within that range (Table 5). In contrast, ratios calculated for copper redhorse (Cus4.2; Zns2.2) were slightly lower than those for white suckers (Table 5), and may be indicative of low contaminant exposure. As reported in white sucker (Miller et al., 1992), concentrations of Cu and Zn were higher in the ovaries than in the testes of copper redhorse. Zinc concentrations in ovaries were even higher than in the liver (Table 1), a situation also reported in white sucker from unpolluted lakes (Miller et al., 1992). The relatively similar pattern in the tissue concentrations of these trace metals between copper redhorse and white sucker strongly supports the argument that hepatic concentrations of Cd, Cu and Zn are better indicators of chronic exposure and environmental contamination than muscle concentrations in fish (Miller et al., 1992). In white sucker from heavy-metal-contaminated lakes, liver concentrations of Cu and Zn were positively correlated with sediment contamination, while muscle concentrations were not (Harrison and Klaverkamp, 1990). High concentrations of Cd, Cu and Zn would therefore be expected in the liver tissues of relatively old and large fish (Evans et al., 1993), if they had been highly exposed to these metals. The evidence for low exposure is consistent with the results of Berryman and Nadeau (1998) who, as part of a monitoring program using aquatic mosses as tracers of contaminants, reported low levels for both Cu and Zn but did not detect the presence of Cd in the Richelieu River. Mean liver concentrations of these metals in copper redhorse, however, were lower than levels measured in St. Lawrence River white sucker and much less than fish from highly polluted water bodies (Table 5). Levels in copper redhorse can be considered more typical of low or slightly contaminated environments. We conclude that the exposure of copper redhorse to Cd, Cu and Zn was probably weak 37 and certainly less than that of relatively younger white sucker from the St. Lawrence River in recent years. Levels of PAHs in copper redhorse were low and typical of fish from other areas (Hellou and Upshall, 1995; McDonald et al., 1995; Neff and Burns, 1996; Fayad et al., 1996; Pelletier et al., 1999), adding to the evidence that PAHs do not bioaccumulate in fish muscle because of an extremely efficient detoxifying mechanism in the liver (D’Adamo et al., 1997). PAHs were found in the liver of copper redhorse and their composition wdominated by phenanthrene, anthracene, and benzo(a)pyrenex was very similar to that reported in white sucker from an upstream area of the St. Lawrence River, near Cornwall, Ontario (Ridgway et al., 1999). Because of their rapid biodegradation in fish, low levels of PAHs are not necessarily indicative of low exposure or low potential toxicity impact. Monteverdi and Di Giulio (2000) demonstrated that benzo(a)pyrene can be rapidly (in less than 48 h) transferred and accumulated in maturing oocytes of gravid fish. The non-detection of PAHs in gravid ovaries therefore indicates that the copper redhorse had not been recently exposed to such compounds prior to their capture. Levels of TCDF in copper redhorse were very low and comparable to levels reported in fish from areas of very low contamination (Hodson et al., 1992; Muir et al., 1995; Brochu et al., 1995a). Levels in liver and gonads were slightly higher than in muscle tissue as a result of higher lipid content. A similar pattern was also observed in carp (Cyprinus carpio) (Wu et al., 2001) and in mummichog (Fundulus heteroclitus) (Monteverdi and Di Giulio, 2000). The mean level of TCDF in muscle tissue of copper redhorse (2.6 ngyg) was similar to that measured in white sucker (geometric mean 3.5 pgyg) collected upstream of pulp and paper mills along the Saint-Maurice River, in Quebec (Hodson et al., 1992). In contrast to those suckers, the tissues of our copper redhorse were not contaminated by PCDDs at detectable levels. Dioxins and furans detected in the water of the Richelieu River were OCDD, OCDF and TCDF, to a lesser extent, but TCDD was not reported (Berryman and Nadeau, 1998). There are no available data on PCDD contamination in white sucker 38 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 from the St. Lawrence River or the Richelieu River. Levels of TCDD and TCDF in copper redhorse were relatively similar to those reported in brown trout (Salmo trutta) and rainbow trout (Oncorhynchus mykiss) from the St. Lawrence River, but much lower than in chinook salmon (Oncorhynchus tshawytscha) (19 ngyg TCDD) or American eel (Anguilla rostrata) (31.9 ngyg TCDF) also caught in the St. Lawrence River (Brochu et al., 1995b). This discrepancy can be explained by the fact that the first two species are resident fish of the St. Lawrence River while the other two are vagrant or migratory fish originating in Lake Ontario, where high levels of TCDD and TCDF have been recorded in various fish species (Reiner et al., 1995). These data indicate that copper redhorse were exposed to very low concentrations of PCDDs and PCDFs and that their contamination expressed as 2,3,7,8-TCDD toxic equivalents was much lower than for white sucker from the Saint-Maurice River (Hodson et al., 1992). Levels transferred to gonads were well below those potentially causing toxic effects in fish embryos (Olivieri and Cooper, 1997). The ratio of PCB 77 over PCB 126 in copper redhorse was 7.5. This was close to the values calculated for brown trout (6.8) and rainbow trout (7.1) from the St. Lawrence River, but quite different from those of 0.86 and 0.11 found in American eel and Chinook salmon, respectively (Brochu et al., 1995b). This variability does not relate to species characteristics but can rather be explained by the origin of fish and the different PCB pattern between St. Lawrence River and Lake Ontario fish. In addition, levels of coplanar PCBs in copper redhorse were comparable to those noted in brown trout and rainbow trout and were considerably lower than those in chinook salmon and eel (Brochu et al., 1995b), suggesting that the source of coplanar PCBs in copper redhorse was weak and apparently similar to the one to which resident fish from the St. Lawrence River were exposed. The low concentrations and low proportion of the most toxic congener (126) contributed to yielding a very low toxicity index (TEQ) for copper redhorse (Table 2). Congeners 138 and 153 were the dominant PCB components in copper redhorse, as classically reported in many other fish species from the St. Lawrence River (Gagnon et al., 1990; de Lafontaine et al., 1999) and elsewhere in the world ´ (McFarland and Clarke, 1989; Sanchez et al., 1993; Blanchard et al., 1997; Leah et al., 1997; Fisk and Johnston, 1998). The relative proportion of the various congeners in copper redhorse, however, differed from that of redhorse species from the Yamaska River or white sucker from the Richelieu River and the St. Lawrence River. It is, therefore, difficult to come to any conclusions about whether the different PCB patterns were due to the species-specific metabolic rates of the various congeners or to the relative abundance of these congeners in the different environments from which the fish were caught (Gagnon et al., 1990; Blanchard et al., 1997). Levels of individual congeners and total PCBs in the liver of copper redhorse were generally lower than in white sucker from the St. Lawrence River or redhorse species (Moxostoma sp.) from the Yamaska and Noire rivers (Table 6), the two rivers adjacent to the Richelieu River. Levels were much more similar, however, when expressed on a lipid weight basis, indicating that the apparent lower contaminant load of copper redhorse was largely due to their lower lipid content. Levels of PCB Aroclor in copper redhorse and in St. Lawrence white sucker were practically identical (;4500 ngyg LW) and approximately half that measured in white sucker (8100 ngyg LW) from the Richelieu River in 1995 ´ 1998). (Piche, The lower contaminant levels measured in copper redhorse relative to white sucker from the St. Lawrence River or fish from the Yamaska River were also generalized for all organochlorine pesticides and chlorobenzenes (Table 6). As shown with PCBs, the difference was greatly reduced when accounting for the lower lipid content in the liver of copper redhorse. Total DDT was the most concentrated pesticide and consisted predominantly of p,p9-DDE as often reported in fish studies. The DDEySDDT ratio, which varied between 0.31 and 0.72, depending on fish tissues, was slightly less than that calculated for redhorse from the Yamaska River or white sucker from the St. Lawrence River (Table 6) or the averaged ratio of 0.71 for different fish species in Quebec (Paul and Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 ´ 1989b). The DDEySDDT ratio may Laliberte, represent an indicator of the extent of DDT degradation where higher values would indicate expo´ sure to more weathered or distant sources (Sanchez et al., 1993), but differences between species may be due to variations in DDT metabolic transformation in biota relative to degradation processes in the environment. The fact that p,p9-DDE levels in Yamaska River redhorse were higher by more than one order of magnitude than copper redhorse from the Richelieu River, while p,p9-DDT levels are almost equal, would indicate that past DDT input was probably more significant in the Yamaska River basin than in the Richelieu River basin, while recent input (mainly due to atmospheric deposition) was similar. DDT concentrations in water samples from the Yamaska and Richelieu rivers in 1991–1992 were not statistically different (Pham et al., 1993) and the DDEySDDT ratio in water samples varied seasonally, being as low as ;0.20 during winter and peaking at ;0.85 in the summertime (Pham et al., 1996). The presence of mirex in biota has been used as a distinctive tracer for biological populations from Lake Ontario and the St. Lawrence River (Castonguay et al., 1989). From two major industrial sources located in Lake Ontario, mirex bioaccumulated in the entire food chain of the Lake Ontario–St. Lawrence River basin from the 1940s until its ban in 1976 (Comba et al., 1993). Mirex disappears by photodegradation into photomirex, which also persists in sediments (Sergeant et al., 1993). Kaiser et al. (1990) reported a decrease in mirex by a factor of ;5 between the Lake Ontario outlet and the St. Lawrence River at Quebec City (;400 km distance apart). Levels of mirex and photomirex in white sucker from the St. Lawrence River, near Quebec City (de Lafontaine et al., 1999) were 10 times lower than those recorded in specimens from Lake Ontario (Sergeant et al., 1993). The ratio of mirex over photomirex was generally )1 in Lake Ontario fish but -1 in St. Lawrence fish. Interestingly, mirex was not detected (-0.2 ngyg) in various cyprinid and catostomid fish species from the Noire River and the Yamaska River in 1986, two rivers adjacent to the Richelieu River (Metcalfe-Smith et al., 1995). The fact that residues of mirex and photomirex were 39 present in copper redhorse from the Richelieu River but not detected in resident fish species from two adjacent rivers leads us to hypothesize that copper redhorse were exposed to mirex via the St. Lawrence food chain, probably by migrating and living in the St. Lawrence River for some part of their life cycle. The low levels of mirex in this long-lived fish were nonetheless indicative of weak exposure to this contaminant and to most organochlorine pesticides and chlorobenzenes. Overall, the relatively low levels of most contaminants analyzed and the observed contamination profile strongly suggest that copper redhorse have been subjected to low exposure to these bioaccumulative substances. For some contaminants (i.e. Hg, TCDD), the low concentrations measured in copper redhorse are consistent with previous reports indicating low levels of these products in the Richelieu River (Paul and Laliber´ 1989a; Berryman and Nadeau, 1998; Piche, ´ te, 1998). On the other hand, the presence and the relative concentrations of other contaminants (i.e. cadmium, mirex, ratio PCB 77 over 126) more typically associated with the St. Lawrence River water masses would suggest that the spatial distribution of the copper redhorse is not solely restricted to the Richelieu River, but would extend into the St. Lawrence River. The copper redhorse is found in the Richelieu River during summertime for spawning, but its wintertime distribution remains virtually undocumented (LaHaye and Huot, 1995). The short, unrestricted distance (;20 km) between the Saint-Ours dam in the Richelieu River (where our specimens were captured) and the St. Lawrence River may allow copper redhorse to migrate between these two systems and therefore become exposed to different contaminants. The slightly lower levels of contaminants measured in copper redhorse relative to St. Lawrence fish may result from the short exposure time in the St. Lawrence River and possibly the slower metabolic activity during wintertime exposure. This hypothesis would be best tested by means of a fish tagging experiment. Given the indication that the copper redhorse distribution extends into the St. Lawrence River, habitats used by the species in the St. Lawrence River should be 40 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 identified and protected as a strategy to save and re-establish this endangered species. The finding that specimens of copper redhorse, despite their old age, were characterized by low levels of all bioaccumulative contaminants, was somewhat unexpected. Given the positive relationship established between fish age and concentrations of many contaminants (i.e. Hg, Cd, co-planar PCBs), one could speculate that the mean levels of many compounds would have been smaller if analyses had been performed on younger specimens, as in the case of other fish species from the ´ 1998) or the St. Lawrence Richelieu River (Piche, River (de Lafontaine et al., 1999). This adds weight to our interpretation that copper redhorse tend to bioaccumulate lower burdens of contaminants as a result of short toxic exposure times relative to St. Lawrence fish. Monitoring studies of riverine fish have found a sharp decline in residual levels following the ban on organochlorine contaminants (Loganathan and Kanna, 1994). Assuming that concentrations have now stabilized in the environment (Stow et al., 1995), increments in residual levels of organochlorine compounds in copper redhorse seem highly improbable in the coming years. Levels were probably higher in both male and female gonadal tissues because of lipid transfer during the maturation process (Niimi, 1983; Fisk and Johnston, 1998). Although this transfer can contribute to lowering the body burden of many lipophilic toxic substances through the annual release of spawning products, it might increase the toxicity for early life stages, even if contamination did not affect adult fish. Levels of all organochlorine contaminants analyzed here were much lower than those previously reported to cause reproductive impairment or egg and fry mortality in fish (Monod, 1985; Hose et al., 1989; Monosson et al., 1994; Fitzsimons, 1995). Based upon a review of laboratory results, Fitzsimons (1995) concluded that the relationship between parental contaminant burdens, physiology and egg viability is not clear, while Smith (1998) provided evidence of strong fish recruitment in heavily contaminated wild fish populations. It is not impossible that copper redhorse may be more sensitive and less tolerant than other species to toxicants affecting reproduc- tion. This species becomes sexually mature at the age of ;10 (Mongeau et al., 1992), and is, therefore, subject to a long-term chronic exposure period before first spawning. Although we do not entirely rule out contaminants as a possible cause of reproductive failure in copper redhorse, it seems improbable that problems arose from the suite of contaminants analyzed here. Levels of these toxicants would not be a sole and sufficient explanation for the failure of copper redhorse recruitment in recent years. It should be recalled that spawning and recruitment failure can also be related to several factors other than chemical contamination, such as spawning ground degradation and early life habitat loss. On the other hand, the Richelieu River basin supports important agricultural activities using large quantities of non-bioaccumulative pesticides during summertime (Rondeau, 1996; Berryman and Nadeau, 1998), at times of peak spawning in copper redhorse (Mongeau et al., 1986). The toxic impact of these chemicals on natural fish populations and on copper redhorse specifically deserves more detailed investigation. Artificial spawning attempts in copper redhorse, and especially observations of fry survival and development, might bring additional information about the exact nature of copper redhorse reproductive abnormalities (Gendron and Branchaud, 1997). Acknowledgments ´ We wish to thank Chantal Menard, Isabelle ´ Menard and Jean Leclerc for fish dissection and sample preparation work, Patrice Turcotte for trace-metal analyses, and Andre´ Fouquet for supervising the analyses of organochlorine pesticides. This study was funded by the St. Lawrence Vision 2000 Action Plan and Environment Canada, Quebec Region. References ¨ C, Pettersen H. Compound Axelman J, Broman D, Naf dependence of the relationship log Kow and log BCF: a comparison between chlorobenzenes (CBs) for rainbow trout and polycyclic aromatic hydrocarbons (PAHs) for Daphnia. Environ Sci Pollut 1995;2:33 –36. Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 Berninger K, Pennanen J. Heavy metals in perch (Perca flavescens) from two acidified lakes in the Salpausselkä esker area in Finland. Water Air Soil Pollut 1995;81:283 – 294. ` Richelieu: Berryman D, Nadeau A. Le bassin de la riviere ´ contamination de l’eau par des meatux et certaines substances organiques toxiques. In: Le Bassin versant de la ` Richelieu: etat ´ ´ ` riviere de l’ecosysteme aquatique—1995. ` de l’Environnement et de la Faune (ed.), Direction Ministere ´ ` ´ des ecosystemes aquatiques, Quebec, Rapport EA-13, Section 2 1998. Blanchard M, Teil MJ, Carru AM, Chesterikoff A, Chevreuil M. Organochlorine distribution and mono-orthosubstituted PCB pattern in the roach (Rutilus rutilus) from the River Seine. Wat Res 1997;31:1455 –1461. Blanchard M, Teil MJ, Carru AM, Garban B, Ollivon D, Chesterikoff A, Chevreuil M. Biota contamination by PCBs and trace metals in the freshwater estuary of the River Seine (France). Hydrobiologia 1999;400:149 –154. Branchaud A, Gendron AD. Artificial spawning and rearing of the copper redhorse, Moxostoma hubbsi (Teleostei: Catostomidae). Can Field Nat 1993;107:279 –282. ˆ ´ L, Dumont P, Fortin R. Branchaud A, Hatin D, Cayer P, Cote ´ ´ MinReproduction artificielle et elevage du Suceur cuivre. ` de l’Environnement et de la Faune (ed.), Service de istere ´ ´ l’amenagement et de l’exploitation de la faune, Montreal, Rapport 06-34, 49 pp., 1995. Branchaud A, Jenkins RE. Pierre Fortin (1823–1888) et la ` description scientifique du chevalier cuivre, ´ Moxpremiere ostoma hubbsi. Can Field Nat 1999;113:345 –358. Brochu C, Moore S, Pelletier E. Polychlorinated dibenzo-pdioxins and dibenzofurans in sediments and biota of the Saguenay Fjord and the St. Lawrence Estuary. Mar Poll Bull 1995a;30:515 –523. Brochu C, Hamelin G, Moore S, Laliberte´ D, de Lafontaine Y. Contribution of PCDDyPCDFs and ortho-substituted PCB congeners to the total TCDD-equivalent concentration in fish fillet from the St. Lawrence River. Organohalogen Compounds 1995b;26:293 –298. Castonguay M, Dutil JD, Desjardins C. Distinction between American eels of different geographic origins on the basis of their organochlorine contaminant levels. Can J Fish Aquat Sci 1989;46:836 –843. Comba ME, Norstrom RJ, Macdonald CR, Kaiser KLE. A Lake Ontario–Gulf of St. Lawrence dynamic mass budget for mirex. Environ Sci Technol 1993;27:2198 –2206. Conover WJ. Practical nonparametric statistics. 2nd ed. New York: John Wiley & Sons, 1980. (493 pp) D’Adamo R, Pelosi S, Trotta P, Sansone G. Bioaccumulation and biomagnification of polycyclic aromatic hydrocarbons in aquatic organisms. Mar Chem 1997;56:45 –49. De Conto C, Petit-Ramel M, Faure R, Garin D. Cadmium bioaccumulation in carp (Cyprinus carpio) tissues during long-term exposure. Analysis by induced coupled plasmamass spectrometry. Ecotox Environ Safety 1997;38:137 – 143. 41 De Lafontaine Y, Gagne´ F, Blaise C, Costan G, Gagnon P, Chan HM. Biomarkers in zebra mussels (Dreissena polymorpha) for the assessment and monitoring of water quality of the St. Lawrence River (Canada). Aquat Toxicol 2000;50:51 –71. ´ De Lafontaine Y, Gauthier J, Menard C. Suivi de la contami` nation chimique de six especes de poissons a` un site de ´ ´ reference du fleuve Saint-Laurent. Environment Canada– Quebec Region, Environmental Conservation, St. Lawrence Centre, Report ST-211, 1999, 79 pp. Donohoe RM, Curtis LR. Estrogenic activity of chlordecone, o,p9-DDt and o,p9-DDE in juvenile rainbow trout: Induction of vitellogenesis and interaction with hepatic estrogen binding sites. Aquat Toxicol 1996;36:31 –52. Evans DW, Dodoo DK, Hanson PJ. Trace element concentrations in fish livers: implications of variations with fish size in pollution monitoring. Mar Poll Bull 1993;26:329 –334. ´ Environment Canada. Manuel des Methodes d’Analyse. Environment Canada, St. Lawrence Centre, Ecotoxicology and Environmental Chemistry Section. Montreal, 1993, 100 pp. Fayad NM, El-Mubarak AH, Edora RL. Fate of oil hydrocarbons in fish and shrimp after major oil spills in the Arabian Gulf. Bull Environ Contam Toxicol 1996;56:475 –482. Fisk AT, Johnston TA. Maternal transfer of organochlorines to eggs of walleye (Stizosdion vitreum) in Lake Manitoba and western Lake Superior. J Great Lakes Res 1998;24:917 – 928. Fitzsimons JD. A critical review of the effects of contaminants on early life stage (ELS): mortality of lake trout in the Great Lakes. J Great Lakes Res 1995;21:267 –276. Fonds M, Casal E, Schweizer D, Boon JP, van der Veer H. Effects of PCB contamination on the reproduction of the dab Limanda limanda L. under laboratory conditions. Neth J Sea Res 1995;34:71 –79. Friedmann AS, Watzin MC, Leiter JC, Brinck-Johnsen T. Effect of environmental mercury on gonadal function in Lake Champlain Northern Pike (Esox lucius). Bull Environ Contam Toxicol 1996;56 :486 –492. Gagnon MM, Dodson JJ, Comba ME, Kaiser KLE. Congenerspecific analysis of the accumulation of polychlorinated biphenyls (OCBs) by aquatic organisms in the maximum turbidity zone of the St. Lawrence Estuary, Quebec, Canada. Sci Total Environ 1990;97y98:739 –759. Gendron AD, Branchaud A. Impact potentiel de la contamination du milieu aquatique sur la reproduction du Suceur ` des connaissances. cuivre´ (Moxostoma hubbsi): Synthese ` de l’Environnement et de la Faune (ed.), Direction Ministere ´ ´ ´ ´ ´ generale de la Monteregie, Service de l’amenagement et de ´ l’exploitation de la faune, Longueuil, Quebec, Rapport 1602, 1997, 160 pp. Giesy JP, Newsted J, Garling DL. Relationships between chlorinated hydrocarbon concentrations and rearing mortality of chinook salmon (Oncorhynchus tshawytscha) eggs from Lake Michigan. J Great Lakes Res 1986;12:82 –98. Goldstein RM, De Weese LR. Comparison of trace element concentrations in tissue of common carp and implications 42 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 for monitoring. J Am Water Res Ass 1999;35(5):1133 – 1140. Green DJ, Beck AE. Mercury and other metal residues in fish from the Assiniboine River, Manitoba, Canada, 1978–1992. Manitoba Environment, Water Quality Management Section, Report No. 95-07, 1995. Grimwood MJ, Dobbs TJ. A review of the aquatic ecotoxicology of polychlorinated dibenzo-p-dioxins and dibenzofurans. Environ Toxicol Water Qual 1995;10 :57 –75. Harrison SE, Klaverkamp JF. Metal contamination in liver and muscle of northern pike (Esox lucius) and white sucker (Catostomus commersoni) and in sediments from lakes near the smelter at Flin Flon, Manitoba. Environ Toxicol Chem 1990;9:941 –956. Hebert CE, Keenleyside KA. To normalize or not to normalize? Fat is the question. Environ Toxicol Chem 1995;14(5):801 – 807. Hellou J, Upshall C. Monocyclic aromatic hydrocarbons in bile of flounder exposed to a petroleum oil. Int J Environ Anal Chem 1995;60:101 –111. Hodson PV, McWhirter M, Ralph K, Gray B, Thivierge D, Carey JH, van der Kraak G, Whittle DM, Levesque MC. Effects of bleached kraft mill effluents on fish in the St. Maurice River, Quebec. Environ Toxicol Chem 1992;11:1635 –1651. Hose JE, Cross JN, Smith SG, Diehl D. Reproductive impairment in a fish inhabiting a contaminated coastal environment off southern California. Environ Pollut 1989;57:139 –148. Ion J, de Lafontaine Y. Spatial variation of contaminant levels in six species of fish of the St. Lawrence River. Environment Canada–Quebec Region, Environmental Conservation Branch, St. Lawrence Centre, Scientific and Technical Report ST-165E, 1998, 152 pp. Ion J, de Lafontaine Y, Dumont P, Lapierre L. Contaminant levels in St. Lawrence River yellow perch (Perca flavescens): spatial variation and implications for monitoring. Can J Fish Aquat Sci 1997;54:2930 –2946. Kaiser KLE, Lum KR, Comba ME, Palabrica VS. Organic trace contaminants in the St. Lawrence River water and suspended sediments, 1985–1987. Sci Total Environ 1990;97y98:23 –40. Kavlock RJ, Daston GP, DeRosa C, Fenner-Crisp P, et al. Research needs for the risk assessment of health and environmental effects of endocrine disruptors: a report of the U.S. EPA-sponsored workshop. Environ Health Persp 1996;104:715 –740. Kelce WR, Stone CR, Laws SC, Gray LE, Kempainnen JA, Wilson EM. Persistent DDT metabolite p,p9-DDE is a potent androgen receptor antagonist. Nature 1995;375:581 –585. LaHaye M, Huot, M. Situation du suceur cuivre´ (Moxostoma ´ ` susceptible d’etre ˆ designee ´ ´ menhubbsi) au Quebec: espece ´ ou vulnerable. ´ acee Rapport du groupe de recherche SEEQ ´ ´ au Ministere ` de l’Environnement et de la Faune, presente ` ´ Direction de la gestion des especes et des habitats. Quebec, 1995, 39 pp. Leah RT, Johnson MS, Connor L, Levene C. Polychlorinated biphenyls in fish and shellfish from the Mersey estuary and Liverpool Bay. Mar Environ Res 1997;43:345 –358. ´ ` un siecle ` Legendre V. Redecouverte apres et reclassification ` ´ d’une espece de Catostomides. Naturaliste Can. 1952;69:227 –233. Legierse KCHM, Sijm DTHM, van Leeuwen CJ, Seinen W, Hermens JLM. Bioconcentration kinetics of chlorobenzenes and the organophosphorus pesticide chlorthion in the pond snail Lymnaea stagnalis: A comparison with the guppy Poecilia reticulata. Aquat Toxicol 1998;41:301 –323. Lesko LT, Smith SB, Blouin MA. The effect of contaminated sediments on fecundity of the brown bullhead in three Lake Erie tributaries. J Great Lakes Res 1996;22:830 –837. Loganathan BG, Kanna K. Globalorganochlorine contamination trends: an overview. Ambio 1994;23 :187 –191. McDonald SJ, Kennicutt MC, Liu H, Safe SH. Assessing aromatic hydrocarbon exposure in Antarctic fish captured near Palmer and McMurdo stations, Antarctica. Arch Environ Contam Toxicol 1995;29:232 –240. McFarland VA, Clarke JU. Environmental occurrence, abundance, and potential toxicity of polychlorinated biphenyl congeners: considerations for a congener-specific analysis. Environ Health Persp 1989;81:225 –239. McFarlane GA, Franzin WG. An examination of Cd, Cu, and Hg concentrations in livers of northern pike, Esox lucius, and white sucker, Catostomus commersoni, from five lakes near a base metal smelter at Flin Flon, Manitoba. Can J Fish Aquat Sci 1980;37:1573 –1578. McFarlane GA, Franzin WG. Elevated heavy metals: a stress on a population of white suckers (Catostomus commersoni) in Hamell Lake, Saskatchewan. J Fish Res Board Can 1978;35:963 –970. Mersch J, Dubost N, Pihan JC. Comparaison de plusieurs ´ supports analytiques inertes et biologiques pour apprecier la ´ ´ contamination metallique de la retenue du site electronu´ cleaire de Cattenom, France. Annals Limnol 1993;29:325 – 337. Metcalfe JL, Charlton MN. Freshwater mussels as biomonitors for organic industrial contaminants and pesticides in the St. Lawrence River. Sci Total Environ 1990;97y98:595 –615. Metcalfe-Smith JL, Merriman JC, Batchelor SP. Relationship between concentrations of metals in sediment and two species of freshwater mussels in the Ottawa River. Water Poll Res J Can 1992;4:845 –869. Metcalfe-Smith JL, Maguire RJ, Batchelor SP. Polychlorinated biphenyl congeners and chlorinated pesticides in fish from the Yamaska River basin, Quebec. Water Qual Res J Canada 1995;30:179 –204. Miller PA, Munkittrick KR, Dixon DG. Relationship between concentrations of copper and zinc in water, sediment, benthic invertebrates, and tissues of white sucker (Catostomus commersoni) at metal-contaminated sites. Can J Fish Aquat Sci 1992;49(5):978 –984. Mongeau JR, Dumont P, Cloutier L. La biologie de suceur ´ Moxostoma hubbsi, une espece rare et endemique a cuivre, la region de Montreal, Quebec, Canada. Ministere du Loisir, Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 de la Chasse et de la Peche (ed.), Service de l’amenagement et de l’exploitation de la faune, Montreal, Rapport technique 06-39, 1986. Mongeau JR, Dumont P, Cloutier L. La biologie du suceur ´ a` celle de quatre autres cuivre´ (Moxostoma hubbsi) cmparee ` especes de Moxostoma (M. anisurum, M. carinatum, M. ´ macrolepidotum et M. valenciennesi). Can J Zool 1992;70:1354 –1363. ´ Mongeau JR, Dumont P, Cloutier L, Clement AM. Le statut ´ Moxostoma hubbsi, au Canada. Can Field du suceur cuivre, Nat 1988;102:132 –139. Monod G. Egg mortality of Lake Geneva charr (Salvelinus alpinus L.) contaminated by PCB and DDT derivatives. Bull Environ Contam Toxicol 1985;35:531 –536. Monosson E, Fleming WJ, Sullivan CV. Effects of planar PCB 3,3’,4,4’-tetrachlorobiphenyl (TCB) on ovarian development, plasma levels of sex steroid hormones and vitellogenin, and progeny survival in the white perch (Morone americana). Aquat Toxicol 1994;29:1 –19. Monteverdi GH, Di Giulio R. Oocytic accumulation and tissue distribution of 2,3,7,8-tetrachlorodibenzo-p-dioxin and benzo-a-pyrene in gravid Fundulus heteroclitus. Environ Toxicol Chem 2000;19:2512 –2518. Muir DCG, Pastershank GM, Crosley R, Noton L, Ramamoorthy S, Brownlee B. Pathways of accumulation and temporal trends of PCDDyFs in fishes downstream from a bleached kraft pulp mill on the Athabasca River (Alberta). Organohalogen Compounds 1995;24:463 –468. Munkittrick KR, Dixon DG. Growth, fecundity, and energy stores of white sucker (Catostomus commersoni) from lakes containing elevated levels of copper and zinc. Can J Fish Aquat Sci 1988;45:1355 –1365. Neff JM, Burns WA. Estimation of polycyclic aromatic hydrocarbon concentrations in the water column based on tissue residues in mussels and salmon: an equilibrium partitioning approach. Environ Toxicol Chem 1996;15(12):2240 –2253. Niimi AJ. Biological and toxicological effects of environmental contaminants in fish and their eggs. Can J Fish Aquat Sci 1983;40:306 –312. Olivieri CE, Cooper KR. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in embryos and larvae of the fathead minnow (Pimephales promelas). Chemosphere 1997;34:1139 –1149. ´ Paul M, Laliberte´ D. Teneurs en mercure des sediments et des ` poissons des rivieres L’Assomption, Richelieu, Yamaska, ` de ¸ Saint-Francois et du Lac Saint-Pierre en 1986. Ministere ´ ´ ´ l’Environnement du Quebec, Direction generale de l’assainissement des eaux, Rapport QE 89-1, 1989a. Paul M, Laliberte´ D. Teneurs en BPC, HAP et pesticides ´ dans les sediments ´ ` organochlores et les poissons des rivieres ¸ L’Assomption, Richelieu, Yamaska, Saint-Francois et du ` de l’Environnement du Lac Saint-Pierre en 1986. Ministere ´ ´ ´ Quebec, Direction generale de l’assainissement des eaux, Rapport QE 89-2, 1989b. Pelletier E, Canuel G, Padros J, Clermont Y, Gobeil C. Concentrations d’hydrocarbures aromatiques polycycliques ` ` chez quatre especes de poissons de fond et deux especes de 43 ´ de l’estuaire et du golfe du Saint-Laurent et du crustaces fjord su Saguenay. Rapp Stat Can Sci Halieut Aquat 1052, 1999, 20 pp. Pham T, Lum KR, Lemieux C. The occurrence, distribution ´ and sources of DDT in the St. Lawrence River, Quebec (Canada). Chemosphere 1993;26:1595 –1606. Pham T, Lum KR, Lemieux C. Seasonal variation of DDT and its metabolites in the St. Lawrence River (Canada) and four of its tributaries. Sci Total Environ 1996;179:17 –26. ` Richelieu: contamination du Piche´ I. Le bassin de la riviere ´ poissons par des metaux et certaines substances organiques ` Richelieu : etat ´ toxiques. In : Le bassin versant de la riviere ´ ` ` de l’ecosysteme aquatique—1995. Ministere de ´ l’environnement et de la faune (ed.), Direction des ecosys` ´ temes aquatiques, Quebec, 1998, Rapport EA-13, section 3. Regoli L, Chan HM, de Lafontaine Y. Organotins in zebra mussels (Dreissena polymorpha) in the St. Lawrence River. J Great Lakes Res 2000;25:839 –846. Reiner EJ, MacPherson KA, Kolic TM, Hayton A, Clement RE. Characteristic levels of chlorinated dibenzo-p-dioxins and dibenzofurans in fish from Ontario Great Lakes. Organohalogen Compounds 1995;24:379 –384. Ridgway LL, Chapleau F, Comba ME, Backus SM. Population characteristics and contaminant burdens of the white sucker (Catostomus commersoni) from the St. Lawrence River near Cornwall, Ontario and Massena, New York. J Great Lakes Res 1999;25:567 –582. Roberts JR, DeFrietas ASW, Gidney MAJ. Influence of lipid pool size on the bioaccumulation of the pesticide chlordane by northern redhorse suckers (Moxostoma macrolepidotum). J Fish Res Board Can 1977;34:89 –97. Rondeau B. Pesticides dans les tributaires du fleuve SaintLaurent 1989–1991. Environment Canada–Quebec Region, St. Lawrence Centre, Scientific and Technical Report ST62, 1996, 58 pp. Sakai S, Hiroaka M, Takeda N, Shiozaki K. Coplanar PCBs and PCDDsyPCDFs in municipal waste incineration. Chemosphere 1993;27:233 –240. Sangalang GB, O’Halloran MJ. Cadmium-induced testicular injury and alterations of androgen synthesis in brook trout. Nature 1980;240:470 –471. ´ Sanchez J, Sole´ M, Albaiges J. A comparison of distribution of PCB congeners and other chlorinated compounds in fishes from Spanish coastal areas and remote lakes. Int J Environ Anal Chem 1993;50:269 –284. Sergeant DB, Munawar M, Hodson PV, Bennie DT, Huestis SY. Mirex in the North American Great Lakes: New detections and their confirmation. J Great Lakes Res 1993;19:145 –157. Smith M, Stalling DL, Johnson JL. Determination of part-pertrillion levels of polychorinated dibenzofurans and dioxins in environmental samples. Anal Chem 1984;56 :1830 –1842. Smith DW. Evidence of strong lake trout recruitment despite high PCB (total Aroclors) concentrations. J Great Lakes Res 1998;24:974 –979. Sokal RR, Rohlf FJ. Biometry. 2nd ed. San Francisco: W.H. Freeman & Co., 859 pp, 1981. 44 Y. de Lafontaine et al. / The Science of the Total Environment 298 (2002) 25–44 Spies RB, Rice DW. Effects of organic contaminants on reproduction of starry flounder, Platichthys stellatus, in San Francisco Bay, California, USA. II. Reproductive success of fish captured in San Francisco Bay and spawned in the laboratory. Mar Biol 1988;98:191 –200. ´ ´ SOMER. Etude comparative des teneurs en metaux dans la ´ chair de poissons provenant du reservoir de La Grande 2 et ´ ´ de lacs naturels. Hydro-Quebec, Vice-Presidence Environnement, 1993, 30 pp. Stow CA, Carpenter SR, Eby LA, Amrhein JF, Hesselberg RJ. Evidence that PCBs are approaching stable concentrations in Lake Michigan fishes. Ecol Applic 1995;5:248 –260. ´ ´ Vachon N. Ecologie des juveniles 0q et 1q de chevalier ` menacee, ´ comparee ´ cuivre´ (Moxostoma hubbsi), une espece ` a` celle des quatre autres especes de Moxostoma (M. anisurum, M. carinatum, M. macrolepidotum, M. valenciennesi) ` ` Richelieu. Societe ´ ´ de la Faune dans le systeme de la riviere ´ ´ et des Parcs du Quebec, Direction de l’amenagement de la ´ ´ faune de la Monteregie, Longueuil, rapport technique 1606, 175 pp. 1999. Walker MK, Peterson RE. Potencies of polychlorinated dibenzo-p-dioxin, dibenzofuran, and biphenyl congeners, relative to 2,3,7,8-tetrachlorodibenzo-p-dioxin, for producing early life stage mortality in rainbow trout (Oncorhynchus mykiss). Aquat Toxicol 1991;21:219 –238. Wise SA, Berner BA, Cristensen RG, Koster BJ, Kurz J, Schantz MM, Zeisler R. Preparation and analysis of a frozen mussel tissue reference material for the determination of trace organic constituents. Environ Sci Technol 1991;25:1695 –1703. Wu WZ, Schramm KW, Xu Y, Kettrup A. Accumulation and partition of polychorinated dibenzo-p-dioxins and dibenzofurans (PCDDyF) in the muscle and liver of fish. Chemosphere 2001;43:633 –641. Zabel EW, Cook PM, Peterson RE. Toxic equivalency factors of polychlorinated dibenzo-p-dioxin, dibenzofuran and biphenyl congeners based on early life stage mortality in rainbow trout (Oncorhynchus mykiss). Aquat Toxicol 1995;31:315 –328.