Ecotoxicology

, Volume 19, Issue 2, pp 391–404

Polychlorinated biphenyls, dioxins, furans, and organochlorine pesticides in spotted sandpiper eggs from the upper Hudson River basin, New York

Authors

    • U.S. Geological Survey
  • Christine M. Custer
    • U.S. Geological Survey
  • Brian R. Gray
    • U.S. Geological Survey
Article

DOI: 10.1007/s10646-009-0425-z

Cite this article as:
Custer, T.W., Custer, C.M. & Gray, B.R. Ecotoxicology (2010) 19: 391. doi:10.1007/s10646-009-0425-z

Abstract

In 2004, spotted sandpipers (Actitis macularia) were studied on the Hudson River near Fort Edward south to New Baltimore, NY and on two river drainages that flow into the Hudson River. Concentrations of 28 organochlorine pesticides, 160 polychlorinated biphenyl (PCB) congeners, and 17 dioxin and furan (PCDD-F) congeners were quantified in eggs collected on and off the Hudson River. The pattern of organochlorine pesticides and PCDD-F congeners did not differ significantly between eggs collected on and off the Hudson River. In contrast, the pattern of PCB congeners differed significantly between the Hudson River and other rivers. Total PCBs were significantly greater in eggs from the Hudson River (geometric mean = 9.1 μg PCBs/g wet weight) than from the other two rivers (0.6 and 0.6 μg PCBs/g wet weight). Seven of 35 (20%) eggs exceeded 20 μg PCBs/g wet weight, the estimated threshold for reduced hatching in tree swallows (Tachycineta bicolor) and some raptor species; the maximum concentration was 72.3 μg PCBs/g wet weight. Models that predicted nest survival and egg success (the proportion of eggs hatching in a clutch if at least one egg hatched) as functions of contaminant levels were poorly distinguished from models that presumed no such associations. While small sample size could have contributed to the inability to distinguish among contaminant and no toxicant models, we cannot rule out the possibility that contaminant concentrations on the Hudson River were not sufficiently high to demonstrate a relationship between contaminant concentrations and reproductive success.

Keywords

Spotted sandpiperPolychlorinated biphenylsDioxinsHudson RiverOrganochlorine pesticides

Introduction

The Hudson River is one of the most polychlorinated biphenyl (PCB)-contaminated rivers in North America (Secord et al. 1999). The General Electric Company used well over 45 million kg of PCBs at two plants on the Upper Hudson River. It is estimated that between 1 and 5% of those PCBs may have been released into the Hudson River (U.S. EPA 1976).

A number of studies have documented that birds can accumulate PCBs through their diet (Custer et al. 1998, 1999, 2003, 2005; Drouillard and Norstrom 2001; Froese et al. 1998; Secord et al. 1999). Effects of PCBs on birds, and especially on chickens (Gallus domesticus), which are extremely sensitive, have been clearly demonstrated in laboratory studies (Hoffman et al. 1996b). PCB effects in field studies have also been demonstrated (Custer et al. 2003, 2007), but many field studies are problematic because of the multiplicity of chemicals that a species is exposed to and the difficulty of assigning causality. The nature and severity of effects can depend on the dosage, PCB congener composition, and the bird species (Barron et al. 1995; Fernie et al. 2001; Hoffman et al. 1996a, b, 1998). Effects of PCBs in laboratory studies on wild bird reproductive endpoints include increased amounts of time spent in courtship, reduced pair bond formation, delayed egg laying, decreased nest attentiveness, inconsistent incubation causing unusual fluctuations in egg temperatures, eggshell thinning and reduced hatching success (Fernie et al. 2001; Fisher et al. 2006; Haseltine and Prouty 1980; Koval et al. 1987; Peakall and Peakall 1973; Tori and Peterle 1983). Direct embryo toxicity of PCBs, as measured by egg injection studies, has also been demonstrated in the laboratory for wild bird species, but often at concentrations that exceed environmental levels (Hoffman et al. 1998).

There are only a few field studies of the effects of PCBs on Hudson River bird populations. Tree swallows (Tachycineta bicolor) nesting along the Hudson River near Fort Edward had PCB concentrations up to 114 μg/g wet wt. in adult whole bodies (Secord et al. 1999). Mean concentrations of PCBs in eggs ranged from 9.3 to 29.5 μg/g wet wt., while mean concentrations in nestlings ranged from 3.7 to 62.2 μg/g wet wt. (Secord et al. 1999). McCarty and Secord (1999) reported reproductive effects including reduced hatchability due to failure of embryos to develop (presumably infertile), high rates of nest abandonment, and other abnormal parental behavior. In contrast, no association was found between PCB concentrations in eggs and subsequent reproductive success in kingfishers (Ceryle alcyon) nesting on the Hudson River (Custer et al. 2010). Small sample sizes (n = ≤17) could have contributed to the lack of association. However, it is also possible that contaminant concentrations on the Hudson River were not sufficiently high to influence reproductive success in kingfishers.

Spotted sandpipers (Actitis macularia) would seem to be a useful candidate species for ecological risk assessments. They are the most widespread breeding sandpiper in North America and occupy almost all habitats near water (Oring et al. 1997). However, they have not been used in risk assessments. This may be in part due to a perception that because their diet consists mainly of freshwater invertebrates, marine invertebrates, terrestrial invertebrates, and occasionally small fish (Nelson 1939; Rubbelke 1976) they do not accumulate high organic concentrations. Sample size may also be an issue, because they nest solitarily and losses due to predation or flooding can be substantial in some years (Oring et al. 1997; Reed and Oring 1993). Finally, individual females may mate with up to four males, each of which cares for a clutch and brood (Oring et al. 1997). Such associations will typically complicate modeling efforts by inducing correlations among nests from the same female when female-nest associations are not known.

Data on organic contaminant concentrations in spotted sandpipers, although meager, suggest substantial accumulation potential. Two immature spotted sandpipers collected on the Sheboygan River, WI accumulated high concentrations of PCBs (28 and 106 μg/g PCBs in carcasses) and may have been harmed by these levels (Heinz et al. 1984). Polychlorinated biphenyl concentrations were elevated in spotted sandpiper eggs collected on the Hudson River in 2002. Of the eleven species of birds whose eggs were collected in 2002 from Bakers Falls to Lower Schodack Island on the Hudson River, spotted sandpiper eggs had the highest mean PCB concentration (15.2 μg/g wet wt.) and the highest individual egg concentration of PCBs (56.2 μg/g wet wt., Hudson River Natural Resource Trustees 2005).

The elevated PCB concentrations in spotted sandpiper eggs from the Hudson River were within levels associated with reduced reproductive success in other species. Reduced productivity in bald eagles (Haliaeetus leucocephalus), white-tailed eagles (Haliaeetus albicilla), and peregrine falcons (Falco peregrinus) was associated with 20, 25, and 40 μg/g PCBs wet wt. in eggs, respectively (Henny and Elliott 2007). Additionally, hatching success in tree swallows on the Housatonic River, MA was reduced beginning at about 20 μg/g PCBs (Custer et al. 2003).

The objectives of this study were to document exposure and possible effects of organic compounds on spotted sandpipers nesting in the upper Hudson River basin. Because of the high concentrations of PCBs reported in spotted sandpiper eggs from the Hudson River (Hudson River Natural Resource Trustees 2005) and several studies suggesting adverse effects of PCBs on avian reproduction (Custer et al. 2003, 2007; Henny and Elliott 2007), our working hypothesis was that there would be a negative association between PCB contamination and reproductive success.

Methodology

Field

Beginning in mid-April, 2004 we used canoes or motorized johnboats to identify spotted sandpipers and their nests along a 100 km stretch of the Hudson River from Fort Edward to New Baltimore, NY (Fig. 1). This area of the Hudson River was surveyed weekly through mid-June and then visited irregularly through mid-July. Less intense survey efforts were conducted on the Hoosic River from Eagle Bridge, NY to the confluence of the Hudson River, and at the confluence of the Mohawk River with the Hudson River. Spotted sandpiper presence and nests were noted on field maps.
https://static-content.springer.com/image/art%3A10.1007%2Fs10646-009-0425-z/MediaObjects/10646_2009_425_Fig1_HTML.gif
Fig. 1

Locations of spotted sandpiper nests (n = 36) on the Hudson, Hoosic, and Mohawk rivers, NY, where eggs were collected in 2004 and analyzed for organic contaminants

Nests were marked with 1 m wooden stakes positioned 2 m or more away from the nest to facilitate finding the nest on subsequent visits. The number of eggs was recorded on each visit. If the number of eggs was four or did not increase on a subsequent visit, one egg in the clutch was floated (Alberico 1995) to estimate stage of incubation. One egg was collected late in incubation for contaminant analysis. Starting 2–3 days before the estimated date of hatching, nests were checked every day until all eggs in the clutch hatched or failed to hatch. Eggs were classified as intact, cracked (prior to pipping), pipped, hatched, or missing. Hatching date was estimated by a combination of factors including egg flotation, age of embryo in the sibling collected egg, and nests found with incomplete clutches. Date of first egg laid was estimated by allowing 1 day for each egg laid and 21 days for the incubation period (Oring et al. 1997). The incubation period was estimated from the date the last egg was laid (day 0 of the incubation period) until the day the first egg hatched. Clutch size was considered the largest recorded number of eggs in the clutch.

Eggs collected for contaminant analyses were weighed on an electronic pan balance (0.01 g), measured (length [0.01 mm, two measurements per egg] and width [0.01 mm, two measurements]) with calipers, and the contents emptied into chemically clean jars. The stage of embryonic development and whether the embryo was alive (viable) was noted. Egg samples were then frozen (<−30°C) until chemically analyzed. Unhatched eggs were opened and the stage at which embryo death occurred was noted. Eggs that were addled, i.e. with no obvious embryo visible, were not assigned a stage, but classified with infertile eggs. Eggs were called infertile if an embryo was not visible; the infertile category could include those where the embryo died at a very early age (<a few days old).

We assigned nests to individual females based on the proximity of nests to one another. It is possible that some nests were incorrectly identified to female. More than one female could have been associated with nests suspected to be from only one female because in some habitats females can nest close to one another (Oring et al. 1997). It is also possible that two or more nests from the same female were classified as being from more than one female; females can change breeding sites within a season, apparently in response to predation or mate acquisition (Colwell and Oring 1989; Oring et al. 1994). Without individually marked individuals or genetic testing, the classification of nests to individual females could not be confirmed.

Analytical chemistry

Details of the analytical chemistry are provided in Custer et al. (2010). In summary, egg samples were analyzed by Axys Analytical Services, Ltd., British Columbia, Canada for PCB congeners, polychlorinated dibenzo-p-dioxin and polychlorinated dibenzofuran (PCDD-F) congeners, and chlorinated pesticides. For pesticides and PCB congeners, the tissue was homogenized with anhydrous powdered sodium sulphate and extracted with dichloromethane. The extracts were cleaned with a Biobead SX-3 gel permeation column with 1:1 dichloromethane: hexane and eluted with 1:1 dichloromethane: hexane. A Florisil column cleanup followed. The extract was then split and one split analyzed for pesticides and most PCB congeners (low-resolution gas chromatograph [GC]/mass spectrometer [MS]) while the other split was analyzed for coplanar PCBs with high resolution GC/MS. The split for coplanar PCBs was further cleaned in a 4.5% (0.22 g) carbon AX-21/celite 545 mixture column with hexane rinses and eluted with cyclohexane:dichloromethane (1:1) followed by 10:1 ethyl acetate:toluene (2 mL). For PCDD-F analyses, the procedure was the same through the Biobead gel permeation column. After that the order of columns were layered silica, Florisil, a 4.5% Carbon/Celite, another layered acid/base silica column, and finally an alumina column. All of the target analytes were quantitated using isotope dilution.

The PCB congeners (n = 160) were analyzed following US EPA Method 1668A. Nominal levels of detection for PCB congeners were 0.5 ng/g wet wt. for all congeners except 1.0 ng/g wet wt. for congener #s 40, 41/64/68/71, 43/49, 47/48/75, 52/73, 55, 56/60, 57, 58, 61/74, 66/80, 67, 70/76, 79, 85/120, 89/90/101, 99, 105/127, 106/118, 110, 132/168, 138/163/164, and 153; 0.05 ng/g wet wt. for congener #s 157 and 167; and 0.005 ng/g wet wt. for congener #s 77, 81, 169 and 189. Total PCB concentration was the sum of the PCB congener concentrations. For the calculation of total PCBs, individual PCB congeners below the detection limit were given a value of zero. The PCDD-Fs were analyzed (n = 17) following US EPA Method 1613B using high resolution GC / MS. Nominal limits of detection were 0.001 ng/g wet wt. for all dioxin furan congeners except 0.01 ng/g wet wt. for octachlorodibenzo-p-dioxins (OCDD) and octachlorodibenzofuran (OCDF). The concentration of total PCDD-Fs was the sum of the individual PCDD-F congener concentrations. A PCDD-F congener that was below the detection limit was given a zero value for the calculation of total PCDD-Fs. Nominal levels of detection for 28 chlorinated pesticides were <0.0005 μg/g wet wt. except 0.001 μg/g wet wt. for α-, β-, γ- benzene hexachloride (BHC), heptachlor, and aldrin; 0.05 μg/g wet wt. for oxychlordane, and 0.075 μg/g wet wt. for toxaphene. Recoveries were monitored using 13C labeled analogs, banks/spikes, and Certified Reference Material (CRM). All contaminant concentrations were corrected back to fresh weight (wt.) by multiplying the concentration by the ratio of egg content weight divided by the egg volume (Stickel et al. 1973). Toxic equivalency values (TEQs) were calculated using the World Health Organization (WHO) toxic equivalency factors (TEFs) for birds (Van den Berg et al. 1998). A congener that was below the detection limit was given a zero value for the TEQ calculation.

Statistical analyses: descriptive and multivariate statistics

For descriptive and multivariate statistics comparisons of contaminants among locations, one egg was selected from each of the 18 females. For females with multiple nests, one nest was randomly selected and that egg used in the analysis.

Analysis of similarity (ANOSIM; Clarke and Warwick 2001) was used to compare the pattern of organochlorine pesticides, PCB congeners, and PCDD-F congeners among the Hudson, Hoosic, and Mohawk rivers. The test statistic, “R,” may vary from −1 to +1. An R value close to +1 indicates that there are very clear differences in patterns among the groups being tested. A value near zero means that the distribution of patterns is as similar among the groups as within the groups. An R is considered statistically significant based on its P value (e.g. P < 0.05). Because R can be significantly different from zero yet inconsequentially small, the size of R indicates the degree of difference. Clear differences in patterns are evident when R is ≥0.4. There is some support for pattern differences when R is ≥0.3 to <0.4, and patterns are viewed as barely differing when R is <0.3 (adapted from Clarke and Warwick 2001). Non-metric multi-dimensional scaling plots (Kruskal 1964) were constructed to display patterns among the Hudson, Hoosic, and Mohawk rivers.

Congeners were included in the multivariate analyses if ≥50% of samples had detected values; one-half the detection limit was assigned to samples below the detection limit. Contaminant data were log-transformed prior to ANOSIM analysis for the concentrations analyses and standardized (converted to percent of total by sample) for composition analyses. Bray-Curtis resemblance matrices were used. When pattern differences were identified, the similarity percentage (SIMPER) subroutine was used to identify which congeners contributed most to the observed differences.

Concentrations of organochlorine pesticides, total PCBs, total PCDD-Fs, PCB TEQs, and PCDD-F TEQs in spotted sandpiper eggs were compared among the Hudson, Hoosic, and Mohawk rivers with 1-way analysis of variance (ANOVA) when suggested by the multivariate analysis. These analyses were done to provide descriptive information about exposure. Concentration data were log-transformed prior to statistical analysis to more closely meet the homogeneity of variance assumption of ANOVA tests. Geometric means, 95% confidence intervals (CI), and ranges are presented in tables and text. Pearson’s correlation method was used to assess the correlations among total PCBs, total PCDD-Fs, PCB TEQs and PCDD-F TEQs.

Statistical analyses: information criterion

Associations between reproductive success and chemical contaminants in eggs were evaluated using Akaike Information Criterion (AIC; Burnham and Anderson 2002). We used the small sample variant of the AIC (AICc; Anderson et al. 2001; Burnham and Anderson 2002) because it is recommended for studies when sample sizes are small relative to the number of parameters being estimated. The AICc is defined as −2 log likelihood + 2 × (number estimated parameters, or K) × (small sample correction factor), where the correction factor = n/(nK−1) and n = sample size (Hurvich and Tsai 1989). AICc statistics were used to rank the candidate hypotheses. The model with the smallest AICc value is given a delta (Δ) ΑΙCc value of zero. Support for other models is quantified by the differences in AICc values between each model and the ΑΙCc best model. Models with Δ values within 2, 2–4, and >10 units are considered as having equivalent support, some support, and essentially no support from the data (Burnham and Anderson 2002; Anderson et al. 2001). If contaminant associations are to be demonstrated, then models with explanatory factors should be supported by the data better than the base or intercept models, which hypothesize no covariate, i.e., no contaminant, associations.

Akaike weights (wi), also called model probabilities, provide a measure of the strength of the evidence in favor of each model. The wi is the probability that a given model is the best among the models considered. The wi is calculated as exp(−ΔAICc/2) and is standardized by dividing by the sum of the exp(−ΔAICc/2) values for all models. Weights specific to a particular toxicant class (i.e., PCBs, dioxins and furans, DDE and none) were calculated by adjusting for the number of models within each class (2 1/2, 2 1/2, 1 and variable, respectively).

Statistical analyses: reproductive effects

All nests found with eggs (n = 42) were included in the analysis of reproductive success regardless of whether an egg was collected for contaminant concentrations. This sample included 24 nests (from 13 females), 8 nests (5 females), and 10 nests (3 females) from the Hudson River, Hoosic River, and Mohawk River, respectively.

Two measures of reproductive success, nest survival and egg success, were used to assess the relationship between toxicants and reproductive success. For these measures, we used the sample egg method where eggs are randomly collected from each nest and contaminant concentrations in these sample eggs are compared to the success of the remaining eggs in the clutch (Blus et al. 1974). One egg was analyzed for chemical analysis from each of the 36 nests; 23 eggs (13 females), 4 eggs (2 females), and 9 eggs (3 females) from the Hudson River, Hoosic River, and Mohawk River, respectively. Of the 36 eggs, 4 were excluded for the analysis of nest failure. Reasons for exclusion included depredation (n = 2), flooding (n = 1), and rejection the measurements of OCDD and OCDF by the laboratory (n = 1). An additional three abandoned nests were excluded for the analysis of egg success in relation to PCBs, PCDD-Fs, and DDE.

The daily probability (Pij) of nest failure was estimated over the period from when the nest was found with eggs through hatching. Nest failure was estimated using complementary log–log models (CLL, log[−log(1−Pij)], Allison 1999). Adjustment for varying observation intervals was performed by setting an offset equal to loge (interval) (McCullagh and Nelder 1989). This model is related to proportional hazard models, where hazard is defined as the number of nest failures per day. Survival at the daily scale was estimated from the CLL model as exp(-exp(η)), where η is on the log hazard scale and includes both survival predictors and the log offset. The hazard or expected number of nest failures per day is obtained as exp(η), while the change in the hazard with an increase in a covariate is obtained as exp(β1), where β1 denotes the parameter estimate associated with the covariate. Random effects of nest or female on nest survival were not addressed because nest failures were infrequent (three total).

Associations between toxicant levels and egg success in successful nests, the percent of eggs that hatched in nests that hatched at least one egg, were analyzed using logistic models (Allison 1999). Note that, because we estimated egg success only from successful nests, egg success did not require adjustment for observational period. Egg success-toxicant associations were expressed as the odds of an egg hatching with a unit increase in toxicant level relative to the odds of it hatching without the increase. Extra-binomial variation among nests and females was addressed using random effects on the logit scale. Including these random effects allowed for a more realistic assessment of contaminant associations.

Models were fitted using SAS®’ nonlinear mixed modeling procedure (PROC NLMIXED; SAS 2003), and were evaluated using AICc and the associated Akaike weight, wi (Burnham and Anderson 2002) as explained above. Numbers of observations, for AICc calculations, were defined as number of nests. The 95% CIs around the estimated effect levels were used to assess the magnitude and directionality of effects, e.g. a positive association with toxicant level was indicated when the interval was above the reference level of 1. Egg and nest failures resulting from depredation or flooding were presumed to be unrelated to the effects of the toxicants, so were excluded from the analysis of nest survival and egg success.

We hypothesized that three toxicant classes (PCBs, PCDD-Fs, and DDE) might adversely affect reproductive success of spotted sandpipers (Custer et al. 1999, 2003, 2005). We estimated the effects of PCBs and PCDD-Fs using measured concentrations as well as TEQs. A fourth set of models contained no explanatory factors and were termed no-covariate or intercept models. An additive 2-variable model included both PCB TEQs and PCDD-F TEQs. Each of these models represented a different hypothesis to explain why reproductive success might differ among nests.

Results

Chemical analysis: frequency of chemical detection

Aldrin, endosulfan I & II, endrin aldehyde, endrin ketone, α-, β-, and γ-BHC, methoxychlor, o,p′-dichlorodiphenyldichloroethane (o,p′-DDD), o,p′-dichlorodiphenyldichloroethylene (o,p′-DDE), o,p′-dichlorodiphenyltrichloroethane (DDT), and toxaphene were below the detection limit in all 18 sandpiper eggs, one egg per nesting female. Endrin, γ-chlordane and heptachlor were detected in one egg; p,p′-DDD, oxy-chlordane, and p,p′-DDT were detected in 7, 5, and 6 eggs respectively. Of the pesticides detected in >50% of samples, mirex, α-chlordane, and endosulfan sulphate were detected in 13, 13, and 17 eggs. p,p′-DDE, dieldrin, heptachlor epoxide, hexachlorobenzene (HCB), and cis- and trans-nonachlor were detected in all eggs.

Seven of the PCDD-Fs were detected in ≥50% of samples (Table 1): 1,2,3,7,8-pentachlorodibenzo-p-dioxin (PeCDD) (n = 14); 1,2,3,4,7,8-hexachlorodibenzo-p-dioxin (HxCDD (9); 1,2,3,6,7,8-HxCDD (18); 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (HpCDD) (10); 1,2,3,7,8-pentachlorodibenzofuran (PeCDF) (17); 1,2,3,4,7,8-hexachlorodibenzofuran (HxCDF) (10); and 1,2,3,4,6,7,8-heptachlorodibenzofuran (HpCDF) (11). 1,2,3,7,8-PeCDF, 1,2,3,7,8,9-HxCDF, 1,2,3,4,7,8,9-HpCDF and OCDF were not detected in any eggs. The remaining PCDD-Fs were detected as follows: 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (n = 5); 1,2,3,7,8,9-HxCDD (2); OCDD (4); 2,3,7,8-tetrachlorodibenzofuran (TCDF) (6); 1,2,3,6,7,8-HxCDF (3); and 2,3,4,6,7,8-HxCDF (3).
Table 1

Concentrations of polychlorinated biphenyl (PCBs), dioxin and furan congeners, and organochlorine pesticides in eggs of spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY in 2004

Analyte

Geometric mean concentration (ng/g wet wt.)/(95% CI)/{range}

Hudson (n = 13 eggs)

Hoosic (n = 2 eggs)

Mohawk (n = 3 eggs)

Total PCBs

9128 Aa

589 B

606 B

 

(4309–19337)

(547–634)

(359–1021)

 

{862–72318}

{567–611}

{368–914}

PCB TEQs

0.728 Ab

0.044 B

0.012 B

 

(0.342–1.55)

(0.031–0.062)

(0.007–020)

 

{0.074–6.53}

{0.037–0.052}

{0.008–0.019}

Total Dioxins and Furans

0.009

0.007

0.006

 

(0.005–0.017)

(0.006–0.007)

(0.001–0.029)

 

{0.001–0.068}

{0.006–0.007}

{0.001–0.025}

Dioxin and Furan TEQs

0.002

0.003

0.001

 

(0.001–0.003)

(0.003–0.004)

(0.0004–0.003)

 

{0.0005–0.006}

{0.003–0.004}

{0.0004–0.002}

Total TEQs

0.732

0.047

0.013

 

(0.344–1.55)

(0.034–0.066)

(0.008–0.022)

 

{0.075–6.54}

{0.040–0.056}

{0.008–0.021}

1,2,3,7,8-PeCDD

0.0006

0.0009

0.0005

 

(0.0004–0.0008)

(0.0009–0.0009)

(0.0002–0.002)

 

{3NDc–0.001}

{0.0009–0.0009}

{1ND–0.001}

1,2,3,4,7,8-HxCDD

0.0003

0.0005

0.0003

 

(0.0003–0.0005)

(0.0005–0.0006)

(0.0001–0.0007)

 

{7ND–0.001}

{0.0005–0.0006}

{2ND–0.0007}

1,2,3,6,7,8-HxCDD

0.001

0.002

0.001

 

(0.001–0.002)

(0.002–0.002)

(0.0004–0.003)

 

{0.0006–0.005}

{0.002–0.002}

{0.0004–0.002}

1,2,3,4,6,7,8-HpCDD

0.0007

NCd

0.0015

 

(0.0004–0.001)

NC

(0.0005–0.004)

 

{6ND–0.005}

{2ND}

{0.0007–0.004}

2,3,4,7,8-PeCDF

0.001

0.001

0.0006

 

(0.0008–0.002)

(0.001–0.001)

(0.0003–0.0009)

 

{1ND–0.004}

{0.001–0.001}

{0.0003–0.0008}

1,2,3,4,7,8-HxCDF

0.0004

0.0003

NC

 

(0.0003–0.0006)

(0.0002–0.0007)

NC

 

{5ND–0.001}

{1ND–0.0005}

{2ND–0.0004}

1,2,3,4,6,7,8-HpCDF

0.0005

0.0005

NC

 

(0.0003–0.0006)

(0.0001–0.002)

NC

 

{4ND–0.001}

{1ND–0.001}

{2ND–0.0004}

P,p′-DDE

57.8

50.4

56.3

 

(37.3–89.4)

(49.3–51.4)

(25.5–124.2)

 

{20.0–604}

{49.8–50.9}

{30.9–121.3}

α-Chlordane

0.374

0.393

1.05

 

(0.250–0.558)

(0.125–1.24)

(0.649–1.70)

 

{4ND–2.36}

{1ND–0.706}

{0.767–1.71}

Dieldrin

2.38

2.14

4.45

 

(1.76–3.23)

(1.0–4.60)

(2.21–8.96)

 

{1.28–10.3}

{1.45–3.16}

{2.58–8.71}

Endosulphan sulphate

0.197

0.204

0.209

 

(0.140–0.276)

(0.065–0.64)

(0.156–0.281)

 

{1ND–0.656}

{0.11–0.367}

{0.180–0.268}

Heptachlor epoxide

1.69

1.11

1.19

 

(0.932–3.05)

(0.929–1.33)

(0.535–2.64)

 

{0.564–15.8}

{1.01–1.22}

{0.539–2.09}

Hexachlorobenzene

2.89

1.20

1.90

 

(2.03–4.12)

(0.974–1.47)

(0.948–3.82)

 

{0.794–6.32}

{1.08–1.33}

{1.05–3.60}

Mirex

0.261

0.183

0.754

 

(0.212–0.322)

(0.119–0.283)

(0.167–3.39)

 

{4ND–0.586}

{0.147–0.229}

{1ND–1.68}

Cis-nonachlor

1.17

1.23

2.14

 

(0.765–1.80)

(0.574–2.65)

(1.31–3.49)

 

{0.416–9.06}

{0.835–1.82}

{1.58–3.51}

Trans-nonachlor

4.78

5.33

7.97

 

(3.08–7.43)

(1.62–17.6)

(5.24–12.1)

 

{1.86–45.8}

{2.90–9.80}

{5.70–11.9}

The data include contaminant concentrations in one egg each from 18 female spotted sandpipers

aMean not sharing the same letter between rivers are significantly different (F = 8.62; df = 2,15; P = 0.003)

bMean not sharing the same letter between rivers are significantly different (F = 15.52; df = 2,15; P = 0.0002)

cThe number before ND (not detected) is the number not detected

dBecause more than half of the samples were not detected, the geometric mean and confidence intervals were not calculated (NC)

Forty of 160 (25%) PCB congeners were detected in <50% of samples. Those included: congener # 1, 2, 3, 4/10, 6, 7/9, 11, 12/13, 14, 18, 19, 23/34, 29, 30, 35, 36, 38, 39, 46, 50, 54, 57, 58, 62/65, 78, 79, 83/108, 96, 104, 112, 122, 125, 134/143, 145, 150, 152, 161, 173, 186, and 204. Seventy-four percent of the remaining 120 congeners were detected in ≥90% of samples. The congeners for which there are WHO TEFs (congener # 77, 81, 105, 114, 118, 123, 126, 156, 157, 167, and 189) were detected in all 18 samples except for congener # 169 which was detected in 14 of 18 samples.

Contaminant concentrations

The pattern of PCDD-F congeners and organochlorine pesticides based on concentration or based on composition did not differ significantly among rivers (Table 2). In contrast, the pattern of PCB congeners differed significantly both by concentration and composition between the Hudson River samples and those from the Hoosic and Mohawk rivers; the Hoosic and the Mohawk rivers did not differ statistically from one another (Fig. 2; Table 2). The PCB congeners that contributed up to 50% of the dissimilarity in composition between the Hudson and Mohawk rivers included a higher percentage of lower numbered congeners (PCB 66/80, PCB 47/48/75, PCB 28, PCB 61/74) and a lower percentage of higher numbered congeners (PCB 153, PCB 138/163/164, PCB 180) on the Hudson River than on the Mohawk River (Fig. 3, SIMPER subroutine). A higher percentage of lower numbered congeners (PCB 66/80, PCB 47/48/75, PCB 28, PCB 61/74) and a lower percentage of higher numbered congeners (PCB 153, PCB 138/163/164, PCB 106/118) on the Hudson River than on the Hoosic River contributed to more than 50% of the dissimilarity in PCB congener composition (Fig. 3, SIMPER subroutine).
Table 2

Results of analysis of similarity for polychlorinated biphenyl (PCBs) congeners, dioxin and furan (PCDD-F) congeners, and organochlorine (OC) pesticides in spotted sandpiper eggs from the Hudson (n = 13 eggs), Hoosic (n = 2 eggs), and Mohawk (n = 3 eggs) rivers in 2004

Comparison

Concentration

Composition

R-statistic

P

R-statistic

P

PCB congenersa

0.57

0.001

0.97

0.001

 Hudson, Mohawk

0.65

0.002

1.00

0.002

 Hudson, Hoosic

0.62

0.01

0.99

0.01

 Mohawk, Hoosic

0.42

0.20

1.00

0.10

PCDD-F congenersb

0.003

0.46

0.18

0.14

 Hudson, Mohawk

0.11

0.26

0.07

0.32

 Hudson, Hoosic

−0.16

0.72

0.23

0.14

 Mohawk, Hoosic

−0.08

0.70

0.92

0.10

OC Pesticidesc

0.06

0.36

0.17

0.16

 Hudson, Mohawk

0.19

0.19

0.23

0.15

 Hudson, Hoosic

−0.10

0.61

0.12

0.24

 Mohawk, Hoosic

0.00

0.50

−0.08

0.60

The data include contaminant concentrations in one egg each from 18 female spotted sandpipers

aIncludes 120 of 160 PCB congeners detected in ≥50% of samples

bIncludes seven PCDD-F congeners detected in ≥50% of samples (see Table 1)

cIncludes nine organochlorine pesticides detected in ≥50% of samples (see Table 1)

https://static-content.springer.com/image/art%3A10.1007%2Fs10646-009-0425-z/MediaObjects/10646_2009_425_Fig2_HTML.gif
Fig. 2

Non-metric multi-dimensional scaling plot of the pattern of the concentration (upper panel) and composition (lower panel) of 120 polychlorinated biphenyl (PCB) congeners in spotted sandpiper eggs on the Hudson, Hoosic, and Mohawk rivers, NY in 2004. Note that the axes of non-metric multi-dimensional scaling plots are without units

https://static-content.springer.com/image/art%3A10.1007%2Fs10646-009-0425-z/MediaObjects/10646_2009_425_Fig3_HTML.gif
Fig. 3

Mean concentrations of polychlorinated biphenyl (PCBs) congeners in spotted sandpiper eggs from the Hudson (upper panel), Hoosic, (middle panel), and Mohawk rivers, NY, (lower panel). The right vertical axis indicates the percent contribution of each congener to total PCB concentrations

As with the larger set of PCB congeners, the composition of the 12 dioxin-like PCB congeners differed significantly among rivers based on their contribution to total PCBs (R = 0.79, P = 0.001, upper panel, Fig. 4) and PCB TEQs (R = 0.70, P = 0.001, lower panel, Fig. 4). For the composition of PCB TEQs, the Hudson River differed significantly from the Mohawk River (R = 0.79, P = 0.004) and the Hoosic River (R = 0.69, P = 0.038) but the Mohawk and Hoosic rivers did not differ from one another (R = 1.0, P = 0.10). The PCB congeners that contributed more than 50% of the dissimilarity in composition of TEQs between the Hudson and Hoosic rivers included a higher percentage of PCB 77 and a lower percentage of PCB 126 on the Hudson River (SIMPER subroutine); the same congeners accounted for more than 50% of the dissimilarity between the Hudson and Mohawk rivers.
https://static-content.springer.com/image/art%3A10.1007%2Fs10646-009-0425-z/MediaObjects/10646_2009_425_Fig4_HTML.gif
Fig. 4

Percent contribution to total polychlorinated biphenyl (PCB) concentration of 12 AH-active congeners in spotted sandpiper eggs (upper panel) from 13 females on the Hudson River (open bars), two females on the Hoosic River (gray bars), and three females on the Mohawk River, NY (black bars). Percent contribution to total toxic equivalent (TEQ) concentrations of 12 AH-active congeners in spotted sandpiper eggs (lower panel)

Even though the multivariate results indicated no significant differences in organochlorine pesticide concentrations and PCDD-F congener concentrations among locations, summary information is provided for organochlorine pesticides and PCDD-Fs among the Hudson River, Hoosic River, and Mohawk River (Table 1). Because the multivariate results suggested differences in PCB congeners and PCB TEQs, total PCBs and PCB TEQs were presented and compared among rivers as well. Total PCBs and PCB TEQs were significantly greater in eggs from the Hudson River than from the other two rivers (Table 1). Geometric mean concentrations of total PCBs and PCB TEQs were approximately 15 times higher on the Hudson than on the other two rivers (Table 1). Seven of 35 (20%) eggs exceeded 20 μg PCBs/g wet wt., the estimated threshold for effects for tree swallows (Custer et al. 2003); the maximum concentration was 72.3 μg/g wet wt. Arithmetic mean concentrations for the Hudson, Hoosic, and Mohawk rivers were 22, 0.7, and 0.8 μg/g wet wt., respectively. The estimated correlation between total PCBs, PCB TEQs, and total TEQs were significant and approached or equaled 1.0 (Table 3). However, correlations among other toxicants were not significant.
Table 3

Estimated correlations among concentrations of polychlorinated biphenyl (PCB), total dioxins and furans (PCDD-F), PCB toxic equivalents (TEQs), PCDD-F TEQs, and total TEQs (PCB TEQs + PCDD-F TEQs) in one egg each from 18 female spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY in 2004

 

Pearson correlation coefficient/P values (in parentheses)

PCBs

PCDD-Fs

PCB TEQs

PCDD-F TEQs

PCDD-Fs

0.12 (0.63)

   

PCB TEQs

0.99 (<0.0001)

0.04 (0.87)

  

PCDD-F TEQs

0.34 (0.17)

0.08 (0.76)

0.38 (0.12)

 

Total TEQs

0.99 (<0.0001)

0.04 (0.87)

1 (<0.0001)

0.38 (0.12)

Reproductive success

Forty-two nests associated with 21 female spotted sandpipers were found with eggs on the Hudson, Mohawk, and Hoosic rivers (Fig. 1; Table 4); one nest was found with chicks. The number of eggs per clutch varied from 3 (one nest) to 4 (38 nests) and averaged 3.97 (SE = 0.26) eggs per clutch; the clutch size of four nests was uncertain. Clutches (n = 34) were initiated as early as 13 May and as late as 24 June; median clutch initiation was 2–5 June (Fig. 5); nest initiation was uncertain for eight nests.
Table 4

Nest and egg success of spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY in 2004

Reproductive parameter

River

Hudson

Hoosic

Mohawk

Incubation period

Number of nests (number of females)

24 (13)

8 (5)

10 (3)

 Number successfula

19 (79%)

4 (50%)

7 (70%)

 Number failedb

5 (21%)

4 (50%)

3 (30%)

  Number depredated

3

2

1

  Number flooded

1

2

 

  Number abandoned

1

 

2

Number of eggs

94

32

36

 Number collected

22

6

7

 Number damaged

1

0

0

 Number incubated

71

26

29

  Number depredated

9

6

4

  Number flooded

3

8

0

  Number abandoned

3

0

4

  Number dead embryos

2

0

1

  Number missing

1

0

0

  Number infertile

1

1

0

 Number hatched

52/71 (73.2%)

11/26 (42.3%)

20/29 (69.0%)

 Number hatched (not including depredation, and flooded)

52/59 (88.1%)

11/12 (91.7%)

20/25 (80%)

 Number hatched in successful nests

52/56 (92.9%)

11/12 (91.7%)

20/21 (95.2%)

aAt least one egg hatched

bNo eggs hatched

https://static-content.springer.com/image/art%3A10.1007%2Fs10646-009-0425-z/MediaObjects/10646_2009_425_Fig5_HTML.gif
Fig. 5

Date of first egg laid for spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY, in 2004

Twelve of 42 nests were not successful (Table 4); six were depredated, three were flooded and three were abandoned. Of 162 eggs laid, 36 were collected. Of the remaining 126 eggs, 83 (66%) hatched. When losses due to depredation (19 eggs), flooding (11 eggs), and abandonment (7 eggs) were excluded, 83 of 89 eggs (93%) hatched. Of 45 embryos examined, one was deformed. The embryo was found dead with one eye missing and an obviously crossed bill. This egg was from a nest located on the Mohawk River near the confluence with the Hudson River.

Based on comparisons of information criteria values, nest survival models were poorly distinguished from one another. The AICc values for all single-covariate toxicant models fell within 1.6 units of each other and within 2.2 units of the best model, which was the intercept model (Table 5). Correspondingly, the 95% CI for the models with contaminant covariates all included 1.0, the null value. The 2-variable model was within 2.3 units of the intercept model indicating a lack of an additive effect of the two classes of TEQs. The sum of adjusted model weights per toxicant class (adjusted for number of models per class—see “Methodology”) were ordered as no toxicants (wi = 0.48), PCDD-Fs (wi = 0.23), PCBs (Σwi = 0.16), and DDE (wi = 0.16). The evidence ratio for PCB models to no toxicant model was 0.16/0.46 = 0.35 (with rounding error) suggesting little evidence of PCB effects relative to the models that postulated no toxicant effects.
Table 5

Model results for nest failure (n = 32 nests) analyses for spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY in 2004

Model

# Model parameters (K)

AICca

ΔAICc

Model weights (wi)

Hazard ratio (95% confidence interval)b

Interceptc

1

35.5

0.0

0.29

NA

PCBs

2

37.6

2.1

0.10

1.015 (0.961–1.072)

PCB TEQs

2

37.7

2.2

0.10

1.147 (0.594–2.213)

PCDD-Fs

2

37.6

2.1

0.10

1.015 (0.951–1.083)

PCDD-F TEQs

2

36.2

0.7

0.21

0.564 (0.196–1.618)

DDE

2

37.7

2.2

0.10

1.010 (0.963–1.060)

PCDD-F TEQs + PCB TEQs

3

37.8

2.3

0.10

0.455 (0.120–1.717)

1.425 (0.725–2.801)

The nests selected represent a subset of all nests where a sample egg was collected during incubation and where the nests were not lost to flooding or depredation

aSmall sample variant of Akaike Information Criterion

bHazard ratio is the predicted failure rate per unit increase of contaminant. PCB units were μg/g wet wt., DDE and PCB TEQ units were ng/g wet wt., and PCDD-Fs, and PCDD-F TEQ units were pg/g wet wt. Reference or null value = 1.0

cIntercept model contains no contaminant terms

NA not applicable

Egg success models were also poorly distinguished from one another (Table 6). The sum of adjusted model weights per toxicant class (adjusted for number of models per class—see “Methodology”) were DDE (wi = 0.54), PCBs (Σwi = 0.17), no toxicants (Σwi = 0.17), and PCDD-Fs (Σwi = 0.13). The evidence ratio for PCB models to no toxicants models was 0.17/0.16 = 1.1. Odds ratios for all toxicant covariates included 1, the null value.
Table 6

Model results for egg success (n = 29 nests) for spotted sandpipers nesting on the Hudson, Hoosic, and Mohawk rivers, NY in 2004

ΔAICc modela

Number of model parameters (K)

AICcb

ΔAICc

Model weights (wi)

Odds ratio of success (95% confidence interval)c

Interceptd

 Fixed

1

32.9

0.7

0.19

NA

 Random nest

2

35.2

3.0

0.06

NA

 Random nest, female

3

37.7

5.5

0.02

NA

 Random female

2

35.2

3.0

0.06

NA

PCBs

2

34.2

2.0

0.10

0.984 (0.944–1.025)

PCB TEQs

2

34.2

2.0

0.10

0.829 (0.528–1.302)

PCDD-Fs

2

34.8

2.6

0.08

1.003 (0.941–1.068)

PCDD-F TEQs

2

34.8

2.6

0.08

1.046 (0.610–1.793)

DDE

2

32.2

0.0

0.28

1.041 (0.981–1.104)

PCDD-F TEQs + PCB TEQs

3

36.4

4.2

0.03

1.142 (0.645–2.022)

0.795 (0.495–1.278)

The nests selected represent a subset of all nests where a sample egg was collected during incubation and where the at least one of the remaining eggs hatched

aAll toxicant models presume a fixed intercept; the AICc values for the random intercept models were larger than that for the fixed intercept model

bSmall sample variant of Akaike Information Criterion

cPredicted change in odds of success per unit increase of contaminant. Units of PCBs and DDE were μg/g wet wt. and units of PCB TEQs, PCDD-Fs, and PCDD-F TEQs were pg/g wet wt. Reference or null value = 1.0

dIntercept models contain no contaminant terms

NA not applicable

Including only nests with contaminant information and excluding nests that were depredated or flooded, nest survival per day with 95% CIs was estimated as 0.988 (0.950–0.997) on the Hudson River and 0.997 (0.974–1.000) off the Hudson River. Estimated nest survival for the entire 21 day incubation period (Oring et al. 1997) was 77.6% on the Hudson River and 93.9% off the Hudson River. Hatching success for the Hudson and non-Hudson Rivers was 93% (76–98%) and 93% (82–97%).

Discussion

Contaminant exposure

Based on PCB concentration and composition, differences between eggs on the Hudson River compared to the Hoosic or Mohawk rivers suggested that the sources of PCBs differed among rivers. In the multivariate assessment, the pattern of concentrations of 115 congeners that were detected in >50% of samples almost completely separated the three rivers. The composition of PCB congeners, irrespective of the concentration, also differed substantially. The PCB congeners that primarily accounted for the separation of the Hoosic and Mohawk rivers from the Hudson River by composition were higher proportions of 66/80, 47/48/75, 28, and 61/74 and lower proportions of 153, 138/163/64, 180, and 106/118. The higher representation of the lower-numbered congeners listed above and a lower representation of the higher-numbered congeners is consistent with Aroclor 1016 and 1242 being the source of the PCBs (Frame et al. 1996; Schmidt 2001) on the Hudson River. The source of PCBs on the other two rivers is unknown, but the much lower concentrations of PCBs are indicative of background concentrations (Custer et al. 1998, 1999, 2003, 2005).

The PCB congener pattern in spotted sandpiper eggs was very similar to that reported in belted kingfisher eggs collected on the Hudson River in the same year (Custer et al. 2010). In both species, PCB congener concentrations and composition were significantly different on the Hudson River than nearby rivers. Additionally, seven of the eight PCB congeners that separated spotted sandpiper eggs between on and off the Hudson River were included in the PCB congener list that separated belted kingfisher eggs on and off the Hudson River (Custer et al. 2010).

Other aspects of the pattern of organic contamination in spotted sandpipers were also similar to that reported for belted kingfisher eggs collected in the same year on the Hudson River (Custer et al. 2010). In both species, total PCBs and PCB TEQs were significantly higher on the Hudson River. Also, in both species, the concentration and composition patterns of PCDD-F congeners and organochlorine pesticides did not differ between the Hudson River and nearby rivers.

Polychlorinated biphenyl concentrations in this study (arithmetic mean = 22 μg/g wet wt.) were comparable to a collection of spotted sandpiper eggs from the Hudson River in 2002 (Hudson River Natural Resource Trustees 2005). Total PCB concentrations in those 13 spotted sandpiper eggs averaged 15.2 μg/g wet wt.

Clutch size, nest initiation, and hatching success

Clutch size (mean = 3.97 eggs), nest initiation dates (range 13 May–24 June, median 2–5 June), and hatching success (92–95% eggs hatched in successful nests) of spotted sandpipers reported here were comparable to those reported earlier. Spotted sandpipers are determinate layers usually laying four eggs in a clutch; occasionally clutches contain three eggs (Oring et al. 1997). Median clutch initiation in north-central Minnesota over 9 years was 22 May–2 June (Lank et al. 1985). Annual hatching success in north-central Minnesota from 1983 to 1990 varied from 22.1 to 90.7% (Reed and Oring 1993).

Reproductive effects

There was no evidence for a contaminant effect for either nest survival or egg success. For both nest survival and egg success, models with contaminant covariates were indistinguishable from intercept models based on the AICc model criterion. In this data set, there were very few nest or egg failures even though PCB concentrations ranged up to 72.3 μg/g wet wt. The small sample could have contributed to the inability to distinguish among the contaminant models and the no toxicant models, however there were few failures even with substantial PCB contamination. The number of nests not flooded or depredated and for which contaminant data were available was 32 nests for 17 females of which 20 nests and 12 females were from the Hudson River. Similar data collected on belted kingfishers nesting on the Hudson River in 2004 also did not demonstrate an association between contaminants and nest survival or egg success (Custer et al. 2010). The sample size for belted kingfishers was even smaller than that of spotted sandpipers; 17 nests were suitable for an evaluation of nest survival and 15 nests suitable for an evaluation of egg success. Other studies that demonstrated a relationship between contaminants and success have had larger sample sizes. For example, a significant correlation between DDE in double-crested cormorant (Phalacrocorax auritus) eggs and hatching success was based on 75 eggs (Custer et al. 1999), and in tree swallows a significant correlation between PCB concentrations and hatching success was found with >100 nests (Custer et al. 2003). Based on a sample of 53 eggs, a significant relationship was found between lower hatching success and DDE concentrations in black skimmers (Rhynchops niger; Custer and Mitchell 1987). The minimum number of nests needed to demonstrate a relationship with high probability, will be a function of the minimum effect you want to observe, species sensitivity, and the distribution of contaminant concentrations among nests.

It is also possible that reproductive success in spotted sandpipers is not as sensitive an endpoint for PCBs as it is with some other avian species. Because PCB concentrations in spotted sandpiper eggs from the Hudson River were within toxic thresholds suggested for other avian species, we might have expected to see an association between reproductive success and increasing PCB concentrations, if such an association was present at the observed concentrations. Reduced productivity in tree swallows, bald eagles (Haliaeetus leucocephalus), white-tailed eagles (Haliaeetus albicilla), and peregrine falcons (Falco peregrinus) was associated with 20, 20, 25, and 40 μg PCBs/g in eggs, respectively (Custer et al. 2003; Henny and Elliott 2007). Arithmetic mean PCB concentrations in spotted sandpiper eggs from the Hudson River (22 μg/g) were comparable to these threshold levels and 7 of 35 (20%) eggs exceeded 20 μg/g wet wt.; the maximum PCB concentration was 72 μg/g.

Low concentrations of p,p′-DDE in spotted sandpiper eggs probably contributed to the lack of a DDE effect on reproductive success. p,p′-DDE concentrations were not even close to the toxic threshold of 3 μg/g for one of the most sensitive avian species (Blus et al. 1974) and averaged <0.06 μg/g. In contrast, a relationship between p,p′-DDE concentrations in eggs and hatching success was reported for double-crested cormorants (Custer et al. 1999) where p,p′-DDE concentrations averaged 3.9 μg/g and the highest concentration exceeded 10 μg/g. Reduced productivity in bald eagles (Haliaeetus leucocephalus), white-tailed eagles (Haliaeetus albicilla), and peregrine falcons (Falco peregrinus) was associated with 12, 6, and >15 μg DDE/g in eggs, respectively (Henny and Elliott 2007).

The usefulness of the spotted sandpiper as an indicator of toxicant effects may be limited. As mentioned above, sample sizes will tend to be low because spotted sandpiper nest densities are often low due to territoriality and lack of available nesting habitat. Spotted sandpipers also place their nests close to the water and so are prone to flooding, which can necessitate re-finding and re-sampling nests. Because females are polyandrous (Oring et al. 1997) and hence clutches may not be independent, pseudoreplication may need to be addressed. Finally, the geographic extent of contamination may often only encompass a few spotted sandpiper territories, so that highly contaminated individuals may be few in number.

Conclusions

Based on concentrations in eggs, spotted sandpipers nesting on the Hudson River, NY were exposed to elevated PCB concentrations; egg concentrations averaged 22 μg/g (geometric mean = 9.1 μg/g) and ranged up to 72 μg/g. The pattern of PCB congeners, a higher prevalence of lower-numbered congeners and lower prevalence of the higher-numbered congeners, was consistent with Aroclor 1016 and 1242 being the source of the PCBs on the Hudson River. Concentrations of PCBs, PCDD-Fs, and organochlorine pesticides were not associated with reduced nest survival or egg success of spotted sandpipers nesting on the Hudson River, NY. The lack of clear response to these toxicants may have resulted from small sample sizes or because this species was not sensitive to these contaminants.

Acknowledgments

The authors thank the Hudson River Trustees for funding the study; Paul M. Dummer, Steven C. Houdek, Anna Karolyshyn, Christopher D. Pollentier, Ryan Pottinger, Kathryn J. Schneider, Amanda J. Stein, Matthew Stuber, and Will Yandik for field assistance; New York Department of Environmental Conservation and U.S. Fish and Wildlife Service for scientific collecting permits; Dennis Heisey and Terry Shaffer for helpful comments regarding survival analysis; Oliver Schabenberger for code to fit the multilevel SHNONE3 sandpiper model; and Gary Heinz and John Elliott for comments on earlier drafts of the manuscript. Any opinions presented are those of the principal investigators and not the position of the United States or the Hudson River Trustees. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Copyright information

© Springer Science+Business Media, LLC 2009