Archives of Environmental Contamination and Toxicology

, Volume 52, Issue 1, pp 80–89

Contaminants in Fish of the Hackensack Meadowlands, New Jersey: Size, Sex, and Seasonal Relationships as Related to Health Risks


    • Department of RadiologyUniversity of Medicine and Dentistry of New Jersey, N. J. Medical School
  • Jeffrey T. F. Ashley
    • School of Science and HealthPhiladelphia University
    • Patrick Center for Environmental ResearchAcademy of Natural Sciences

DOI: 10.1007/s00244-006-0093-4

Cite this article as:
Weis, P. & Ashley, J.T. Arch Environ Contam Toxicol (2007) 52: 80. doi:10.1007/s00244-006-0093-4


The trace metal content and related safety (health risk) of Hackensack River fish were assessed within the Hackensack Meadowlands of New Jersey, USA. Eight elements were analyzed in the edible portion (i.e., muscle) of species commonly taken by anglers in the area. The white perch collection (Morone americana) was large enough (n = 168) to enable statistically significant inferences, but there were too few brown bullheads and carp to reach definite conclusions. Of the eight elements analyzed, the one that accumulates to the point of being a health risk in white perch is mercury (Hg). Relationships between mercury concentrations and size and with collection season were observed; correlation with lipid content, total polychlorinated biphenyl (PCB) content, or collection site were very weak. Only 18% of the Hg was methylated in October (n = 8), whereas June and July fish (n = 12) had 100% methylation of Hg. White perch should not be considered edible because the Hg level exceeded the “one meal per month” action level of 0.47 μg/g wet weight (ppm) in 32% of our catch and 2.5% exceeded the “no consumption at all” level of 1 μg/g. The larger fish represent greater risk for Hg. Furthermore, the warmer months, when more recreational fishing takes place, might present greater risk. A more significant reason for avoiding white perch is the PCB contamination because 40% of these fish exceeded the US Food and Drug Administration (FDA) action level of 2000 ng/g for PCBs and all white perch exceeded the US Environmental Protection Agency cancer/health guideline (49 ng/g) of no more than one meal/month. In fact, nearly all were 10 times that advisory level. There were differences between male and female white perch PCB levels, with nearly all of those above the US FDA action level being male. Forage fish (mummichogs and Atlantic silversides) were similarly analyzed, but no correlations were found with any other parameters. The relationship of collection site to contaminants cannot be demonstrated because sufficient numbers of game fish could not be collected at many sites at all seasons.

The Hackensack Meadowlands (HM) is an ∼3000-ha estuary 5–12 km west of New York City and draining into Newark Bay. It includes salt marsh, open water (the Hackensack River and its tributaries), and open and closed landfills, as well as a four-century history of residential and commercial development in northeastern New Jersey. It might be the oldest industrial area in North America. Approximately 1200 ha are now protected from further development. Substantial hunting and fishing activity takes place in the HM, although the state of New Jersey has issued advisories related to fish consumption based on data collected in 1986–1987 (Hauge et al. 1990) and commercial fishing and crabbing are prohibited in the HM.

The water quality of the Hackensack River has been improving since the 1970s, due to the US Clean Water Act. This has resulted in an increase in recreational fishing and crabbing within the HM. Although the water quality has improved, river sediments can be a persistent reservoir of particle-bound contaminants and might act as a source of contaminants that, in turn, can cycle through the food web. Recently, an inventory of fish species in the HM was completed ( The sampling allowed not only a species census but also assessment of the state of health of the fish, ecological parameters, and chemical analysis of tissues from selected fish species. Thus, baseline data were provided on the extent to which the fish living in the river are accumulating contaminants. These data could be used to determine whether the level of contamination poses concern for human health and/or ecological risk.


Tissue Analysis

Selection of contaminants of concern

Given the historical anthropogenic impacts to which the Meadowlands has been subjected over the past four centuries and the many studies that have analyzed sediments and surface waters within the HM in the past 20 years, the contaminants of concern (COCs) in the Meadowlands are well known. COCs analyzed in tissue samples include metals (arsenic, cadmium, chromium, copper, lead, mercury, nickel, and zinc) and organic chemicals that are known to bioaccumulate (e.g., chlorinated pesticides and polychlorinated biphenyls [PCBs]). Of those COCs that bioaccumulate, only the organics and mercury (Hg) bioconcentrate (i.e., accumulate to higher levels [typically, an order of magnitude] with each trophic level), making them of special concern.

Selection of target species

Species targeted for tissue analysis included common resident species that are consumed by humans (i.e., “game” fish) and smaller resident species with small home ranges that are consumed by fishes, birds, mammals, and so forth (i.e., “forage” fish). Thus, the target game fish included white perch (Morone americana), carp (Cyprinus carpio), pumpkinseed (Lepomis gibbosus), and brown bullhead (Ictalurus nebulosus). The target forage fish included the mummichog (Fundulus heteroclitus), Atlantic silverside (Menidia menidia), and the inland silverside (Menidia beryllina). Of these species, the only game fish collected in sufficient numbers allowing both seasonal and site comparisons was the white perch, and the only forage fish commonly found were the mummichog and the Atlantic silverside. Other game fish collected and analyzed were the carp and bullhead.

The collection sites, along or near the Hackensack River, are shown in Fig. 1. The collecting methods included gill nets, seining, trawling, and trap nets. These are identified at the sites in Fig. 1 as GN, S, T, and TN, respectively, and they extend from 40°44′35″N × 74°04′41″W (at GN 1) to 40°51′04″N × 74°01′47″W (at TN 6). The gill and trap nets were left in place for ∼24 h. Trawling was done from the shore toward the river channel. A 16-ft otter trawl (constructed using 3/4-in. square body mesh, 5/8-in. square cod-end mesh, with a 1/4-in. mesh cod-end liner) was towed for 3 min. Two duplicate tows were made each time that we sampled a trawl location. Only those trawl sites that were productive are shown in Fig. 1. Collecting was performed at least monthly throughout the year. Data presented here are from specimens collected October 2001 through May 2003.
Fig. 1.

Collecting sites within the New Jersey Meadowlands Commission’s Hackensack Meadowlands District.

Processing and analysis of tissue samples

The New Jersey Meadowland Commission (NJMC) fishery team provided the specimens for analysis. For each collection, they filled out a chain-of-custody form listing the species, collection location, gear type, date, time, length, weight, and any abnormalities observed for each specimen. Specimens (in labeled Ziploc® bags) were placed in ice and brought to the NJMC lab at the end of each collection day. The specimens were kept on ice and dissected at the NJMC lab, providing the tissue necessary for analysis. Standard edible fillets (“skin off”) were cut from the game fish specimens and the remainder of the carcasses were archived at −80°C. The fillets, individually identified, were double-bagged and transferred to the University of Medicine and Dentistry of NJ (UMDNJ) for further processing. These species included the white perch, brown bullhead, and common carp. For purposes of analysis, individual fish were used; it was not necessary to combine or pool specimens. A total of 168 white perch were analyzed for metals. Other parts of some of the fish were used for other studies: edible fillets for organic chemical uptake (n = 30), stomach contents for diet analysis (Weis 2005), and the remainder for histopathology, reproductive status, growth rate, and parasite burdens (Czerwinski et al., companion study, unpublished data). In addition, 29 brown bullheads, and 9 carp were analyzed for metals.

Forage fish were analyzed whole (including stomach contents). In the lab, composite samples were sliced vertically and 7–10 pieces (utilizing sections that included all body parts and organs) from 3–4 fish were combined for analysis. The number of specimens per composite, but not the lengths and weights of the individual fish used to make up each composite sample, were recorded. A total of 30 mummichog and 8 silverside composites were analyzed for metals, and 9 mummichog and 6 silverside composites were analyzed for organics.

Metals analysis

For metals analysis, a sufficient amount of tissue (2.0 ± 0.2 g wet weight, yielding ∼0.4–0.5 g dry weight) was excised (or in the case of homogenized forage fish, combined as described earlier), oven-dried to constant weight (60°C, 48 h), weighed on a calibrated analytical balance to the nearest milligram, and mineralized in 10 mL Trace Metal Grade HNO3 (Fisher Scientific) in Teflon bombs in a MARS-5 programmed microwave digester (CEM Corp., Mathews, NC) at 115 lb/in.2 and 178°C for 30 min. The resultant mineralized solution was boiled off to near dryness, restored to 10 mL volume with 1% HNO3, and divided in half. One half was used by the NJMC laboratory for analysis of Cd, Cr, Cu, Ni, Pb, and Zn by graphite furnace–atomic absorption spectrophotometry (GF-AAS). The other half was used by UMDNJ for total Hg analysis by cold-vapor AAS in a Bacharach MAS-50D mercury analyzer and for As analysis in a Perkin-Elmer 3100Z spectrophotometer by GF-AAS with Zeeman effect. Wet weight metal levels (from which government agencies derive their risk analyses) were back calculated by dividing our dry weight values by the individual moisture contents of 74–78%.

Organic analysis

In this study, 110 individual PCB congeners, DDXs, and chlordanes were quantified in 9 mummichog, 6 silverside, and 30 white perch samples at the Academy of Natural Sciences’ Patrick Center for Environmental Research. DDXs are comprised of the two isomers (p,p and o,p) of DDT [1,1,1-trichloro-2,2-bis-(p-chlorophenyl)ethane], the parent compound widely used to control insect pests on agricultural crops and those carrying infectious diseases, and the two isomers (p,p and o,p) of each of its metabolites, DDE [1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene] and DDD [1,1-dichloro-2, 2-bis(p-chlorophenyl) ethane]. Like DDXs, values for “chlordanes” were calculated as the summation of the concentrations of heptachlor, heptachlor expoxide, oxychlordane, gamma chlordane, alpha chlordane, cis-nonachlor, and trans-nonachlor.

Preparation of fish subsamples followed previously published methods (e.g.,Ashley et al. 2000). Briefly, ∼2 g of each fish homogenate was used. Sodium sulfate (Na2SO4) was added to eliminate water. Each dried sample was extracted using a Soxhlet apparatus with dichloromethane for a minimum of 16 h. Lipids were removed from fish extracts by gel permeation chromatography (GPC) using dichloromethane as the mobile phase. The collected fraction containing the organic analytes was concentrated by roto-evaporation and an N2 stream. Solid–liquid chromatography using Florisil® was performed as an additional cleanup. Using this technique, PCBs and DDEs were eluted from the Florosil column using petroleum ether (F1 fraction). The remaining analytes were eluted using 50:50 petroleum ether and dichloromethane (F2 fraction).

Congener-specific PCBs, DDXs, and chlordanes were analyzed using a Hewlett-Packard 5890 gas chromatograph equipped with a 63Ni electron capture detector and a 5% phenylmethyl silicon capillary column. The identification and quantification of PCB congeners followed the “610 Method” (Swackhamer 1987), in which the identities and concentrations of each congener in a mixed Aroclor standard (25:18:18 mixture of Aroclors 1232, 1248, and 1262) were determined by calibration with individual PCB congener standards. Congener identities in the sample extracts were based on their chromatographic retention times relative to the internal standards added. In cases in which two or more congeners could not be chromatographically resolved, the combined concentrations were reported. DDXs and chlordanes were identified and quantified based on comparisons (retention times and peak areas) with a known calibration standard prepared from individual compounds.

Lipid analysis

Lipid normalization for organic contaminants was calculated on a gravimetric basis at the Academy of Natural Sciences using aliquots of the dichloromethane/Soxhlet elutriates produced for organic analyses.

Methylmercury analysis

In addition to the above analytical activities, eight subsamples of white perch muscle tissue for methylmercury (meHg) were analyzed at the University of Georgia’s Skidaway Institute (second year only) by cold vapor atomic fluorescence detection (Bloom 1992). An additional 12 white perch muscle samples were sent to Flett Research Ltd., Winnipeg for similar meHg analysis (when the Skidaway Institute was no longer performing meHg analysis).

Quality assurance/quality control

Quality control for the analysis of fish included the following: chain-of-custody documentation of all materials selected for analysis and archiving; the use of carbon-steel dissection instruments to avoid chromium contamination from stainless steel; the use of deionized/distilled water; acid-washing and triple rinsing of glassware; use of an analytical balance calibrated with both internal and external standards; for metal analysis, inclusion of the NRC-Canada certified reference material (CRM) dogfish liver tissue (DOLT-2) and method blanks (1 CRM and 1 blank with each 12 unknowns). An acceptable run was one in which the CRM data were within the published 95% confidence interval (CI). (An exception to this was arsenic analysis, for which we were consistently at 75% of the published value) Minimum detection levels were defined as three times the standard deviation of the blanks. For the meHg analysis, both contractors used DORM-2 (dogfish muscle tissue) for a CRM as part of their quality assurance/quality control.

For organic contaminant analyses, analyte loss through analytical manipulations was assessed by the addition of surrogate PCB congeners 14, 65, and 166 prior to extraction by Soxhlet apparatus. These surrogates were not industrially prepared and, therefore, are not present in the environment. Average recoveries of congeners 14, 65, and 166 were 105 ± 12%, 89 ± 7%, and 97 ± 9%. Due to the relatively high surrogate recoveries and the low standard deviations, all reported values for organic analytes were not corrected for analyte loss. Matrix blanks (six) were generated to monitor possible laboratory contamination and to calculate the detection limits for organic analytes. Chromatograms of most blanks were void of significant peaks, suggesting that little contamination through laboratory exposure occurred. The detection limits were calculated as the average (or mean) mass plus three times the standard deviation of the mass. The matrix-blank-based detection limits for individual organic analytes ranged from 0.01 to 1.0 ng/g wet weight. Based on six matrix blanks, the detection limit for total PCBs (t-PCBs, the sum of all quantified PCB congeners) was 11 ng/g wet weight. National Institute for Standards and Technology (NIST) standard reference material (SRM 1974B—Organics in Mussel Tissue was used to evaluate extraction efficiency and analytical accuracy. The average percent recovery for NIST-reported analytes was 88 ± 36%. To assess precision of the organic contaminant analyses, sample duplicates of randomly selected samples were performed at a frequency of 10%. The mean relative percent difference (RPD) for t-PCBs in duplicates was 4 ± 4. The mean RPD for t-DDXs and chlordanes in duplicates were 7 ± 8 and 9 ± 6. Duplicate analyses revealed exceptional precision.

Statistical Procedures

Regression analyses, t-tests, one-way analysis of variance (ANOVA), Bartlett’s F-test for homogeneity, Kruskall–Wallace test, and Dunnett’s multiple comparison test were calculated within the graphics program GraphPad Prism® 4.0.



A total of 168 white perch were caught, representing virtually all locations and all seasons. This is significant because this species is the one most often sought and caught by recreational anglers. However, many season/site combinations lacked sufficient numbers of specimens for individual statistical validity. In addition, 30 pooled mummichog samples were found year-round at 5 locations, and 8 pooled Atlantic silverside samples were found at several seasons at the three seining locations. Although not relevant to human health, these two species are important in the ecology of the HM, as they represent the most abundant forage fish in this system.

The results for overall metal burdens are summarized in Table 1. As stated earlier, the white perch are of special interest because of their frequency and desirability on the part of fishers. The amounts of metals in the white perch fillets were as follows:
  1. 1.

    Arsenic: ranged from <MDL (minimum detection level) (0.09 μg/g) to 1.01 μg/g dry weight, equivalent to 0.02–0.25 μg/g wet weight.

  2. 2.

    Cadmium: ranged from <MDL (0.09 μg/g) to 0.94 μg/g dry weight, equivalent to 0.02–0.23 μg/g wet weight.

  3. 3.

    Chromium: ranged from 0.03 to 1.28 μg/g dry weight, equivalent to 0.01–0.32 μg/g wet weight.

  4. 4.

    Copper: ranged from 0.86 to 6.48 μg/g dry weight, equivalent to 0.22–1.62 μg/g wet weight. (Two outliers were omitted because they were one and two orders of magnitude higher than the other 166 data points; white perch are notorious for accumulating Cu in liver [mg/g, rather than μg/g; Bunton et al. 1987], and liver tissue might have contaminated these two muscle samples.)

  5. 5.

    Mercury: ranged from 0.07 to 5.47 μg/g dry weight, equivalent to 0.02–1.17 μg/g wet weight.

  6. 6.

    Nickel: ranged from 0.24 to 7.27 μg/g dry weight, equivalent to 0.06–1.82 μg/g wet weight.

  7. 7.

    Lead: - ranged from <MDL (0.21 μg/g) to 4.25 μg/g dry weight, equivalent to 0.05–1.06 μg/g wet weight.

  8. 8.

    Zinc: ranged from <MDL (0.05 μg/g) to 23.1 μg/g dry weight, equivalent to 0.13–5.75 μg/g wet weight.

Table 1.

Metal burdens in fish (μg/g dry weight, means and standard deviaton











White perch


0.19 (0.24)

0.13 (0.15)

0.24 (0.19)

2.49 (1.58)

1.80 (0.99)

1.94 (1.60)

0.70 (0.98)

9.95 (11.31)

Brown bullhead


0.50 (1.27)

0.11 (0.08)

0.24 (0.17)

2.42 (0.70)

0.67 (0.59)

0.94 (0.74)

0.69 (0.62)

24.7 (3.36)



0.06 (0.07)

0.08 (0.07)

0.22 (0.08)

2.57 (1.38)

1.20 (0.71)

1.40 (1.05)

0.24 (0.44)




0.14 (0.17)

0.07 (0.05)

1.12 (1.24)

11.76 (7.49)

0.25 (0.16)

4.94 (6.28)

0.93 (1.21)

11.99 (6.23)

Atlantic silversides


0.53 (0.36)

0.09 (0.07)

0.83 (1.05)

3.79 (1.45)

0.48 (0.21)

2.13 (2.05)

0.77 (0.74)

3.73 (2.51)

a Composites of two to five fish each.

b Data not available.


Of the metals that we analyzed, the one that is of greatest concern is Hg. It is the only metal known with certainty to biomagnify, becoming an order of magnitude higher with each trophic level. The reason for this is that the most likely form to be found in fish, monomethylmercury, an especially toxic organic form, is taken up by organisms in a manner similar to organic compounds. The traditional action level for Hg in fish is 1 ppm (μg/g) wet weight (US FDA 1993). This was exceeded by 4 of the 168 white perch, and these were all larger specimens. However, there is an Environmental Protection Agency (EPA) guideline for fish consumption vis-á-vis Hg. This agency has recommended no more than one meal per month of 0.47–0.94 ppm Hg in fish (US EPA 1999c, 2000, 2001). Thus, there is a moderately high probability of catching a “risky” white perch, as 53 of 168 (32%) exceeded the 0.47-ppm level.

Of the other game species, only 1 of 29 brown bullheads and none of 9 carp exceeded that 0.47-ppm one-meal-per-month risk level.

Mercury versus fish size
The Hg/size correlation that typically exists in fish is demonstrated in our white perch data (Fig. 2). Although strong correlations for both length and weight were not observed, they were significant. We can conclude from Fig. 2A that a white perch longer than ∼230 mm will have a 50% probability of being a risky meal and this length typifies the minimum-sized “keeper” for a recreational angler.
Fig. 2.

Size relationships for mercury uptake in white perch, by both length (A) and weight (B). Slight but significant relationships are demonstrated, using data from all sites and all seasons.

Mercury–seasonal relationships
The white perch data were analyzed for seasonal trends. The higher Hg levels tended to occur in warmer weather, when more people are fishing. Thus, the probability of catching the more contaminated white perch would be increased in the summer (Fig. 3). (To aid in interpreting these Hg uptake graphs, recall that these are data for dry weight. Because the tissues were 74–78% moisture, the 0.47-μg/g consumption advisory level, which is based on wet weight, would be about 2 μg/g on these graphs.)
Fig. 3.

The relationship of season to mercury uptake; it is seen to be higher in the summer than in the cooler months (days 130–240 are mid-May–September). The curve represents the predictable level and was calculated as y = a + bx + cx2. For this curve, r2 = 0.211, p < 0.01.

Mercury–site relationships
Collection sites were correlated with white perch Hg concentrations, after averaging all sizes and seasons for each site (Fig. 4). The geographic distribution of Hg shows that the highest levels were in TN3. However, there were no significant correlations between fish Hg and the sites’ sediment contaminant levels. (Among all the metals analyzed, there was a significant correlation with collection site sediment only for Pb [ E. Konsevick, personal communication].)
Fig. 4.

The north–south distribution of Hg burdens is demonstrated in this box and whiskers plot. The total area covered by this graph is from Gill Net 1 at River Mile (RM) 2.7 to Trap Net 6 at RM 12.5. Data without boxes are those for which only a few fish were found at each of those sites (thus, not enough data were available to generate all of the predictions available in this type of graph).


Eight samples of white perch from a mid-October collection were sent to the University of Georgia’s Skidaway Institute for analysis. Only small fractions of Hg in the tissues were found to be methylated. The meHg levels were 0.048 ± 0.027 μg/g. These were 16% of the total Hg encountered in this species. Typically, meHg as a percentage of total Hg is close to 100% in fish (Bloom 1992), including HM mummichogs (Weis et al. 1986), so this was surprising. Therefore, this issue was revisited. Twelve additional white perch muscle samples from June and July collections were sent to Flett Research, Inc., Winnepeg, which found that 110 ± 13.2% of the Hg was meHg. Both contract laboratories used atomic fluorescence spectrophotometry (Bloom 1992) and both, using the same CRM, demonstrated acceptable accuracy. What differed was the season during which the fish were collected.

Organic Contaminants

Comparisons by wet weight

The concentrations of t-PCBs, t-DDXs, and t-chlordanes is illustrated in Figure 5. t-PCB uptake was substantially greater than the other two classes of compounds. There were no significant differences among the three fish species for any of the three classes of compounds. The size and seasonal relationships for t-PCB levels and the relationship to Hg levels in the same fish are illustrated in Figure 6. There were no significant correlations between PCB concentrations and length, weight, or Hg burden. The seasonal relationship, although qualitatively similar to that for Hg, was not significant, likely due to the limited number of samples analyzed.
Fig. 5.

Organic contaminants in fish by species: A concentrations by wet weight; B concentrations by lipid weight. *Statistically significantly different from mummichog.
Fig. 6.

The correlations of total PCBs to length (A), weight (B), mercury burdens (C), and season (D) are shown here. None of the correlations are significant.

Concentrations of t-PCBs for white perch were higher than for silverside and mummichog, which had similar body burdens, expressed on either a wet tissue weight basis or by lipid normalization, but the interspecies differences were not significant by wet weight comparisons. This might relate to the small differences in trophic status and feeding among these species. This trend was also observed for DDXs, but it was not as strong for chlordanes, in which body burdens among the three organic contaminant classes was similar.

The Food and Drug Administration (FDA) action level for t-PCBs in food is 2000 ng/g wet weight (US FDA 1993). This was met or exceeded in 12 of the 30 white perch (40%). The distribution of these 12 white perch covers the entire length of that part of the Hackensack River that was surveyed. None of the forage fish exceeded the action level for PCBs. The US EPA one-meal-per-month (cancer health) guideline for PCBs is 23–47 ng/g (US EPA 1999b, 2000), and this was exceeded by all fish species. The lowest measured t-PCB value was a white perch at 243 ng/g wet weight.

The action level for DDXs is 5000 ng/g wet weight. None of the white perch met or even approached that level. The highest was 2159 ng/g, and most were less than one-tenth of the action level. None of the forage fish exceeded that level. The US EPA one-meal-per-month (cancer health) guideline for DDXs is 140–280 ng/g (US EPA 2000), and this was met or exceeded by 20 of the 30 white perch. (Although this guideline is irrelevant for forage fish, it was noted that seven of the nine mummichog composites and all of the silversides composites exceeded this level as well.)

The action level for total chlordanes is 300 ng/g wet weight. One of the white perch exceeded that with 304 ng/g. Most were substantially lower, even into the single digits. None of the forage fish exceeded that level. The US EPA one-meal-per-month (cancer health) guideline for chlordanes is 130–270 ng/g (US EPA 2000), and this was met or exceeded by 13 of 30 white perch (as well as 3 of 9 mummichog and none of the silversides.)

Lipid Relationship

When analyzing organic contaminants in tissues, it can be useful to correlate the chemical burdens with the total lipid in the tissues. Variability in hydrophobic organic contaminant concentrations like PCBs and DDXs, as well as organic Hg, might often be a result of variations in lipid content (e.g., Hebert and Keenleyside 1995; Stow et al. 1997). This normalization procedure serves as a standardization. There is a strong correlation with lipids for t-PCBs but none for Hg (Fig. 7).
Fig. 7.

Correlation of PCB (left) and Hg (right) burdens in white perch with the total lipid content of the muscle tissue.

Upon normalization to lipid of the wet weight concentrations of the three classes of organic contaminants, variability among species was dampened, but not eliminated, suggesting that lipid content, in part, was one determinant of contaminant levels (Fig. 5B). White perch t-PCB and t-DDX levels remained higher than those found for mummichog or silversides upon lipid normalization. These differences are significant between white perch and mummichogs (the Kruskall–Wallace statistic for t-PCBs was 10.79, p = 0.0045, and for t-DDXs, it was 8.441, p = 0.0147; for the differences between white perch and mummichogs, p < 0.01 for t-PCBs and p < 0.05 for t-DDXs by Dunnett’s multiple comparison test. No other comparisons were significant. These nonparametric tests were used because Bartlett’s test for homogeneity of variances showed all the organic uptakes to have unequal variances, precluding the use of ANOVA.)

Sex Differences

Because female fish lay a large amount of eggs, a lipid-rich tissue, it is generally considered that fat-soluble contaminants are depurated as a result of spawning. We separated the data by sex and found that PCBs, but not Hg, are higher in males than in females (Fig. 8). Because only 12 females and 18 males were analyzed, the interesting different trends that appear here are not statistically significant (other than the overall PCB difference in Fig. 8A). Because the males had much higher PCB levels than the females, the relationship to size was recalculated with separation of the sexes. As seen in Fig. 8B, there were very different trends between the sexes, with the males continuing to bioaccumulate over time (as shown with increasing fish length) and the females maintaining a steady state, owing probably to the depuration with the eggs. This can be contrasted with Fig. 6A. (Note also that the shortest fish in Fig. 8B is a female outlier. Because this individual was too small to be sexually mature, it has not yet ovulated. This might be why it was the highest in PCBs of all the females; when removed from the equation, the regression line for females is nearly flat, suggesting a “steady state” in PCBs for the female white perch.) Unlike PCBs, Hg does not differ between the sexes (Fig. 8C), nor does it correlate with PCBs in male white perch (Fig. 8D).
Fig. 8.

Sex differences and relationships in white perch for PCB (A, B) and Hg (C, D) uptake. For regression lines, solid lines are males and dashed lines are females.


The levels of arsenic reported here are typical of seafood in general and are not considered of significance. Most arsenic in seafood is in the form of organoarsenicals, most of which are poorly absorbed and metabolized by consumers. Thus, they are not considered to be toxic (Irgolic 1992). The EPA guidelines for arsenic in seafood of 2.8–5.6 μg/g wet weight was not exceeded by any of our fish. The FDA criterion for arsenic is even higher: 76 μg/g (US FDA 1993).

The cadmium, chromium, copper, lead, and zinc levels in the fish were not particularly high either. They do not bioconcentrate (Rhinefelder et al. 1998; Suedel et al. 1994). The chromium level of concern for consumption of crustacean shellfish is 22 μg/g (US FDA 1993), two orders of magnitude higher than the means found in our Hackensack game fish species. The US FDA criteria for cadmium (3 μg/g) and for lead (1.5 μg/g) were not exceeded either. There are not, to our knowledge, US agency criteria for copper or zinc. There are European guidelines for maximum permissible levels of several trace metals, including Cd (0.06 mg/day/person), Cu (2–3 mg/day/person), and Zn (15 mg/day/person) (European Communities 2001). These translate to 0.3 μg/g Cd, 10–15 μg/g Cu; and 75 μg/g Zn, all for a 200-g meal, if eaten daily. By these European guidelines, the HM white perch do not pose a risk for these metals.

Of the other metals, only Hg was found to be of concern, and that was only in white perch. The Hg in carp, although not of concern, was about three times higher than the average of 0.11 μg/g wet weight for this species in the northeast part of North America (US EPA 1999b).


Mercury versus fish size

Of the several metals measured in this study, only Hg biomagnifies, thus increasing with size and with trophic level (Rhinefelder et al. 1998; Suedel et al. 1994). Considering the amount of Hg circulating in the HM, the amount in white perch was unexpectedly low. The answer can be found in the white perch’s dietary habits. The stomach content analysis done for a companion study shows that they rarely eat fish; they eat mostly small crustaceans. This low trophic level was verified by δ15N stable isotope analysis (Weis 2005).

Mercury–seasonal relationships

The reason for higher mercury in fish during the warmer months might be the higher food intake at this time. Depuration occurs at a rate that, like any physiological activity or chemical reaction, is temperature dependent. Nevertheless, it still continues during the winter, a time when food is scarce to nonexistent, so that a fish will show a net loss of Hg during that time. Changes in Hg burden in relation to season were previously reported in HM mummichogs (Weis et al. 1986).

Mercury–site relationships

The geographic distribution of Hg shows that the highest amounts were from TN3, but this interpretation might be spurious because most of these fish were caught in June, near the height of the curve shown in Figure 3. The site expected to be highest was the Berry’s Creek Canal (T7), because this drains an area with three Superfund sites, one of which is infamous for Hg contamination. Unfortunately, only three fish were caught there, so that result is inconclusive. Conversely, the relatively low levels at TN4 might be similarly biased because these were all caught in the colder part of the year (October 25, December 11, and March 6). It would have been appropriate to have fish from each sampling site year-round, but the fish were not always trapped at each site. It is not known how much “site fidelity” (i.e., staying in one area) white perch have or to what extent they migrate up or down river other than their spawning runs into fresher water in the spring, although it was determined in a Chesapeake study with 15 tagged fish that they tend to remain in a 0.0128-km2 area (McGrath 2005). It might be that the inability to have a more representative collection is because our sampling methods were less than 100% successful. What is conclusive from our data is that a fish with an unacceptable Hg level can be caught anywhere on the river.


Kannan et al. (1998) studied total mercury and meHg in water, sediment, and fish from South Florida estuaries and found that, among the many fish species studied, the percentage of methylated Hg varied from 20% to 100%. Mason (2004) analyzed several game fish species from numerous sites within the Chesapeake Bay system and found that white perch (n = 6) averaged 28% methylated form of Hg; no reporting of collection dates was made, however.

Methylmercury is bioaccumulated mostly (>85%) from the diet (Hall et al. 1997). Furthermore, meHg is associated more with the hypoxic interface of estuarine systems (Mason et al. 1993). Hypoxia is typically a warm-weather event. These two relationships (diet and hypoxia, both of which are greater in the warmer months) would explain the greater uptake of Hg in the summer and why white perch Hg is much more in methylated form at this time.


King et al. (2004) measured PCB levels in fish from 14 tributaries in the Chesapeake Bay watershed and correlated their findings with the amount, type, and distribution of developed land in each location. White perch t-PCBs were greatest in relation to the most intense development. In more highly developed areas, these researchers found that the farther the development was from the water, the lower the contaminant levels were in the sampled fish. An index of development inversely proportional to distance from the shoreline gave the highest correlation to t-PCBs in white perch (r2 = 0.99). King et al. concluded that >4% total residential/commercial development in the watershed predicts exceeding the US EPA (1999b, 2000) guideline for more than one meal per month (47 ng/g t-PCBs). New Jersey has the highest population density in the United States, and this is in large part due to the northeast part of the state, the location of the HM. Thus, high PCBs are to be expected in the HM, considering the high proportion and long history of development in the area. All of our white perch exceeded the US EPA consumption guideline for fish (>47 ng/g t-PCBs), the lowest being 242 ng/g.

Lipid Relationship

Variability in hydrophobic organic contaminant concentrations like PCBs and DDXs, as well as organic Hg, migh often be a result of variations in lipid content (e.g., Hebert and Keenleyside 1995; Stow et al. 1997). Upon normalization to lipid of PCB wet weight concentrations, variability between species was dampened, but not eliminated, suggesting that lipid content, in part, was one determinant of contaminant levels. Although lipid content might determine contaminant accumulation to a degree in fish, other factors such as trophic position and resident versus migratory behavior are also important. For example, Rasmussen et al. (1990) found food chain structure to be a prime determinant of PCB levels in lake trout and other pelagic fish. Ashley et al. (2000, 2003) found that resident fish with limited home ranges were more reflective of their habitats, whereas those with wider home ranges or those undergoing annual migrations were not. For chlordanes, all fishes had similar lipid-normalized concentrations, suggesting that lipid content might be a larger determinant for bioaccumulation of these contaminants or that chlordane contamination is equally dispersed over the study area. White perch t-PCB levels remained higher than those found for mummichog or silversides upon lipid normalization, suggesting biomagnification of contaminants from prey items to predator (white perch). The species’ differences were significant only between mummichog and white perch for t-PCBs and t-DDXs. However, as noted earlier, white perch probably does not occupy a higher trophic level than mummichog (Weis 2005).

Unlike PCBs, Hg did not correlate with lipid content. This suggests that the small proportion found to be methylated during the colder seasons might obfuscate the potential relationship, because meHg is lipid soluble, whereas inorganic Hg is not. However, the 12 data points that we have for the summer do not show a lipid relationship (data not shown). Also, meHg is not nearly as hydrophobic as the organic contaminants studied here; the octanol-water partition coefficients (log Kow) for various meHg species predicted for the pH and salinity conditions of the Hackensack River are <2 (Faust 1992), whereas those for PCBs are >6 (Linkov et al. 2005). For mummichog and silverside, the low number of samples analyzed and the narrower range of lipid values spanned by these fish likely explains the relatively low degrees of correlation found for organic contaminants. However, weak but positive correlations do suggest that lipid content is at least one driving factor in the accumulation of these lipophilic contaminants.

Sex Differences

Because female fish lay a large amount of eggs, a lipid-rich tissue, it is generally considered that fat-soluble contaminants are depurated as a result of spawning. This is probably responsible for the sex differences in PCB uptake shown in Fig. 8B. That there is near-identity in the Hg burdens of males and females (Fig. 8C) also reinforces the finding of relatively little meHg, at least for part of the year, and there is no sex difference as found for the PCBs. This does not explain, however, what processes are occurring in warmer weather when methylation is high. Separation of sexes allows demonstrating a possible relationship between PCBs and Hg, not evident in Fig. 6C, in males only. However, more analyses would have to be performed before more definitive conclusions can be drawn, as with the data in Fig. 8B.


  1. 1.

    The trace metal and chlorinated hydrocarbon content and safety (health risk) of HM fish has been assessed. There were too few carp to reach logical conclusions. The white perch collection, on the other hand, was large enough to enable valid conclusions.

  2. 2.

    It is suggested that HM white perch should be considered not edible. The Hg level exceeded the “one meal per month” action level of 0.47 μg/g wet weight (ppm) in 32% of our catch and 2.5% exceeded the “no consumption at all” level of 1 μg/g. The larger fish represent greater risk Furthermore, the warmer months, when more recreational fishing takes place, might present the greater risk for Hg, as well, because there is more Hg in the muscle tissue, and it is all methylated at this time. A greater reason for not consuming white perch is the PCB contamination, as 40% of these fish exceeded the FDA action level for this class of compound and all exceeded the US EPA guideline of no more than one meal/month (US EPA 1999b, 2000). In fact, nearly all were 10 times that advisory level.

  3. 3.

    The relationship of sampling site to mercury cannot be demonstrated because of the inability to obtain sufficient numbers of game fish at many sites at all seasons.

  4. 4.

    Brown bullheads and, possibly carp, might be safe to eat in relation to metals.



The assistance of many people is gratefully acknowledged: at the NJMC, Brett Bragin, who led the collection team, Yefim Lewinsky, who performed much of the atomic absorption spectrophotometry, and Edward Konsevick, who performed data management; at UMDNJ, Theodore Proctor, who has been an invaluable lab assistant for many years, for tissue processing and atomic absorption spectrophotometry; at the Academy of Natural Sciences, Ms. Linda Zaoudeh for organic contaminant analyses. This project was part of a multifaceted, multi-institutional study funded by the Meadowlands Environmental Research Institute.

Copyright information

© Springer Science+Business Media, Inc. 2006