Environmental Toxicology and Chemistry, Vol. 24, No. 3, pp. 617–628, 2005 Printed in the USA 0730-7268/05 $12.00 1 .00 EFFECTS OF CONTAMINANT EXPOSURE ON REPRODUCTIVE SUCCESS OF OSPREYS (PANDION HALIAETUS) NESTING IN DELAWARE RIVER AND BAY, USA PAMELA C. TOSCHIK,† BARNETT A. RATTNER,*‡ PETER C. MCGOWAN,§ MARY C. CHRISTMAN,† DAVID B. CARTER,\ ROBERT C. HALE,# COLE W. MATSON,†† and MARY ANN OTTINGER† †Marine, Estuarine, and Environmental Science Program and Department of Animal and Avian Sciences, University of Maryland, College Park, Maryland 20742, USA ‡U.S. Geological Survey-Patuxent Wildlife Research Center, BARC-East Building 308, 10300 Baltimore Avenue, Beltsville, Maryland 20705 §U.S. Fish and Wildlife Service, Chesapeake Bay Field Office, Annapolis, Maryland 21401 \Delaware Coastal Management Program, State of Delaware, Dover, Delaware 19901, USA #Virginia Institute of Marine Science, Gloucester Point, Virginia 23062, USA ††Department of Wildlife and Fisheries Science, Texas A&M University, College Station, Texas 77843, USA ( Received 17 March 2003; Accepted 24 August 2004) Abstract—Despite serious water-quality problems and pollutant loading and retention, Delaware River and Bay (USA) provide important wildlife habitat. In 2002, we conducted a comprehensive evaluation of contaminant exposure and reproduction of ospreys (Pandion haliaetus) breeding in Delaware River and Bay. Sample eggs were collected from 39 nests and analyzed for organochlorine pesticides, polychlorinated biphenyls (PCBs), and mercury; a subset of 15 eggs was analyzed for perfluorinated compounds and polybrominated diphenyl ethers (PBDEs). The fate of each nest was monitored weekly. Concentrations of 10 organochlorine pesticides or metabolites, total PCBs, and several toxic PCB congeners were greater ( p , 0.05) in eggs collected between the Chesapeake and Delaware Canal (C and D Canal) and Trenton (Delaware River and northern Bay) compared to other sites. Concentrations of p,p9-dichlorodiphenyldichloroethylene ( p,p9-DDE; 0.785–3.84 mg/g wet wt) and total PCBs (5.50–14.5 mg/g wet wt) in eggs collected between the C and D Canal and Trenton were similar to levels recently found in the Chesapeake Bay. In all study segments, at least one young fledged from 66 to 75% of nests. Productivity for Delaware Inland Bays (reference area) and southern Delaware Bay was 1.17 and 1.42 fledglings/active nest, respectively; north of the C and D Canal, productivity was 1.00 fledgling/active nest, which is marginally adequate to maintain the population. Using these data, a logistic regression model found that contaminant concentrations ( p,p9-DDE, heptachlor epoxide, chlordane and metabolites, and total PCBs) were predictive of hatching success. Several perfluorinated compounds and PBDEs were detected in eggs at concentrations approaching 1 mg/g wet weight. These findings provide evidence that contaminants continue to be a significant stressor on osprey productivity in the northern Delaware River and Bay. Keywords—Pandion haliaetus Organochlorine contaminants Polybrominated diphenyl ethers Biomarker Perfluorinated compounds (Hg), lead (Pb), and zinc (Zn) in Delaware River and Bay sediment exceed contamination thresholds and water-quality criteria [2]. A recent sediment study using the human reportergene system bioassay as an indicator of PAH or chlorinated hydrocarbon contamination suggested that parts of the river and bay are degraded following a north-to-south gradient [3], with the greatest responses occurring near the Philadelphia airport (PA, USA) and Dover Air Force Base (DE, USA). A 10-d static bioassay with the amphipod Ampelisca abdita demonstrated high sediment toxicity at scattered locations in the northern bay near Salem Cove (NJ, USA), Little Tinicum (PA, USA), and Cherry Island Flats (NJ, USA) [2]. Finfish consumption advisories are in effect north of the Chesapeake and Delaware Canal (C and D Canal, MD and DE, USA, respectively) to Yardley (PA, USA) because of high concentrations of DDT, dieldrin, chlordane, PCBs, dioxin, As, and Hg [1,4]. Ospreys (Pandion halieatus), a piscivorous bird, were common breeders along the Delaware Bay [5]. However, during the 1950s, coincident with the advent of organochlorine pesticides, the osprey population declined precipitously [5]. During the late 1980s, eggshell thinning (up to 23% in addled eggs) and reproductive impairment caused by dichlorodiphenyldichloroethylene ( p,p9-DDE) and, to a lesser degree, PCBs were reported in Delaware Bay ospreys, peregrine falcons (Falco peregrinus), bald eagles (Haliaeetus leucocephalus), and great blue herons (Ardea herodias) [6–10]. A detailed evaluation of contaminant exposure and productivity in a small INTRODUCTION The Delaware River and Bay (USA) provide important habitat for diverse wildlife, including ospreys. Delaware Bay is part of the National Estuary Program, the Western Hemisphere Shorebird Reserve Network, and a Wetland of International Significance and is on the Last Great Places list of the Nature Conservancy. The Delaware River is the largest undammed river in the eastern United States, draining more than 30,000 km2 of land and providing water for more than 17 million people [1]. The coastal zone in Delaware is highly industrialized, with factories producing steel, manufacturing industrial and commercial chemicals and plastics, and refining petroleum. More than 50 Superfund sites in Pennsylvania (USA) are located within the Delaware River watershed, and at least 40 more Superfund sites are located within Delaware State (USA). The Delaware Bay is the largest oil-transfer port of entry on the east coast, and agricultural lands and intensive poultry farming surround it. The Delaware River does not support drinkingwater use between Camden and Trenton (both NJ, USA). Between Trenton and Wilmington (DE, USA), fluctuations in pH and dissolved oxygen do not always support aquatic life. Concentrations of DDT, dieldrin, chlordane, polychlorinated biphenyls (PCBs), dioxins, polycyclic aromatic hydrocarbons (PAHs), arsenic (As), chromium (Cr), copper (Cu), mercury * To whom correspondence may be addressed (barnettprattner@usgs.gov). 617 618 Environ. Toxicol. Chem. 24, 2005 P.C. Toschik et al. sample of ospreys breeding on the Salem River (NJ, USA), which is below the heavily polluted regions, indicated eggshell thinning and reduced hatching success were still affecting this population in 1989 [7]. A small follow-up collection of osprey eggs from the same study sites in 1998 indicated that concentrations of organochlorine pesticides and PCBs had declined and that eggshell thickness was comparable to that of the pre-DDT era [11]. Before the present study, only one osprey egg had been analyzed for organochlorine contaminant exposure from the Delaware State portion of the Bay (4.6 mg/ g of PCBs and 5 mg/g of p,p9-DDE) [10], and no eggs had been analyzed from the most contaminated regions in northern Delaware and southern Pennsylvania states. Furthermore, other chemicals that bioaccumulate and biomagnify, including perfluorinated compounds and polybrominated diphenyl ethers (PBDEs), have become of concern in high-trophic-level organisms, including raptors and fish-eating birds, because of their potentially toxic effects [12,13]. The osprey population in this region has yet to return to the pre-1950s range after the reproductive impairment and subsequent population decline associated with exposure to DDT and metabolites. Ospreys are currently listed as uncommon breeders in Delaware outside of the Inland Bays [5]. Based on the lack of data for this region and the high concentrations of toxic chemicals found in Delaware River finfish, we conducted the first large-scale ecotoxicological evaluation of ospreys nesting along Delaware River, Bay, and coast. We chose to study this large area to test the hypothesis that breeding success would reflect a gradient of contamination associated with urban and rural areas. In 2002, osprey eggs were collected to determine exposure to organochlorine pollutants and mercury and more contemporary contaminants, including perfluorinated compounds and PBDEs. The fate of the eggs remaining in sampled nests also was monitored through fledging to test the hypothesis that reproductive success may be depressed by elevated contaminant concentrations. MATERIALS AND METHODS Study area and sample collection The present study spanned from the Atlantic Ocean and the tidal areas in the shallow Inland Bays (DE, USA) to the freshwater, nontidal Delaware River near East Stroudsburg (PA, USA). The Delaware Estuary is a ‘‘drowned-riverbed’’ estuary characteristic of mid-Atlantic coastal habitat. For the present study, the Delaware River and Bay region was divided into four segments (Fig. 1) based on water-quality data collected by the Delaware River Basin Commission and the U.S. Environmental Protection Agency [1,14]. Study segment 1 (river) in the Easton–East Stroudsburg region (PA, USA), including part of the Delaware Water Gap, supported all uses of water and was not classified as threatened [1]. Study segment 2 (north) included the upper tidal portion of the Delaware River from Trenton, 90 km south to the C and D Canal, in Delaware Bay; most of this area has serious water-quality problems, has a high vulnerability to pollution, and in some locations, does not support aquatic life criteria [1]. Study segment 3 (central) included central and southern Delaware Bay from the C and D Canal, 130 km south to Cape Henlopen (DE, USA); this area has serious water-quality problems and high vulnerability to pollution [1]. The central segment supports aquatic life but is threatened in some areas [1]. Study segment 4 (south) encompassed Rehoboth Bay and Indian River Bay (i.e., Delaware Inland Bays) along the Atlantic coast of Delaware; this study Fig. 1. Study segments and osprey nests sampled in the Delaware River and Bay (USA) region. Contaminant exposure and reproduction of Delaware Bay osprey segment has less serious water-quality problems but is highly vulnerable to pollution. Locations of osprey nests were identified by a fixed-wing aircraft survey conducted by the U.S. Geological Survey and U.S. Fish and Wildlife Service (U.S. FWS) in May 2001, boat and ground searches in 2002, and communication with state and federal agencies and conservation groups. Before initiation of sampling, a power analysis was conducted to determine the minimum number of eggs necessary to collect per study segment (a 5 0.05, b 5 0.80, D 5 50%). The analysis was done using contaminant concentrations from osprey eggs collected in the Chesapeake Bay during 2000 and 2001 [15]. It was determined that a collection of 12 eggs per segment would allow us to accurately detect significant differences in contaminant concentrations among regions (only three eggs were collected in the river segment because of the small breeding population and inaccessible nesting structures). All nests were intensively monitored (7- to 14-d intervals) by climbing its supporting structure or by use of a pole-mounted mirror. On completion of a clutch, a single egg was collected at random from each nest for contaminant analysis. Numerous studies have demonstrated that concentrations of organochlorine contaminants in the sample egg are quite representative of concentrations in remaining eggs of the clutch [10,16,17]. The remaining eggs in each nest were monitored to determine hatching success. Eggs remaining in nests 10 d beyond the expected hatch date (i.e., addled; n 5 6) also were collected. Survival of all nestlings was monitored until fledging (days 50–60). Evidence of egg and nestling loss because of predation, adverse weather, and other such disturbances was noted. At approximately 40 d of age, a randomly selected nestling was briefly (,10 min) removed from each nest. Culmen length (precision, 0.1 mm; Scienceware Dial Caliper model 5921; Bel-Art Products, Pequannock, NJ, USA) and body weight (precision, 10 g; spring scale; Douglas Homs Corp, Belmont, CA, USA) were measured, general condition of the nestling noted, and a 5- to 7-ml brachial blood sample obtained in a heparinized syringe. A 2-ml aliquot of whole blood for genetic damage assessment was placed in a cryotube, immediately frozen in liquid nitrogen, and transferred to a 2808C freezer within 24 h, and the remainder of the blood sample was frozen (2208C) for metal analyses (unpublished data). Analytical methods Each egg was weighed (precision, 0.01 g; electric balance; model V-200; PGC Scientific, Frederick, MD, USA), measured for length and diameter (precision, 0.1 mm; Scienceware Dial Caliper), and its contents examined to determine the condition and approximate age of the embryo and then transferred to a chemically clean jar (ICHEM Research, New Castle, DE, USA) and stored at 2208C (n 5 39 fresh eggs 1 6 addled eggs). These samples were chemically analyzed for 25 organochlorine pesticides or metabolites, total PCBs (sum of all congeners), arylhydrocarbon (Ah) receptor–active PCB congeners, and Hg [18–21] by contract laboratories (Geochemical and Environmental Research Group and Trace Element Research Laboratory, College Station, TX, USA) under the rigorous quality-assurance and quality-control guidelines of the U.S. FWS Patuxent Analytical Control Facility (n 5 4 procedural blanks, spikes, and duplicates for pesticides and PCBs; n 5 2 method blanks, reference material blanks, spikes, and duplicates). For the organochlorine pesticides or metabolites and PCBs, egg contents were homogenized, anhydrous sodium sulfate was added, and the mixture was extracted with di Environ. Toxicol. Chem. 24, 2005 619 chloromethane. The extract was evaporated, concentrated, and then dissolved in hexane. Alumina/silica gel chromatography was used to fractionate this solution, followed by high-resolution gas chromatography with electron capture detection. For Hg analysis, egg homogenates were freeze-dried and analyzed by cold-vapor atomic absorption spectroscopy [22] (http://www.epa.gov/cgibin/claritgw?op-Display&document5clserv:ORD:0209;&rank5 4&template5epa). The minimum detection limit for organic chemicals ranged from 0.0015 to 0.025 mg/g; the minimum detection limit for Hg was approximately 0.005 mg/g. A subset of 15 freeze-dried egg homogenates (selected from the 39 fresh eggs collected), selected to represent the entire spatial extent of the study area, was analyzed for perfluorinated compounds (Exygen Research, State College, PA, USA) and PBDEs. Perfluorosulfonates and perfluorocarboxylic acids (C8–C12) were extracted with methanol, followed by addition of activated carbon and filtration. Extracts were analyzed using high-pressure liquid chromatography followed by mass spectrometry/mass spectrometry electrospray on a Hewlett-Packard 1100 (Avondale, PA, USA) with a Jones Chromatography Genesis C-8 column (50 mm length 3 2.1 internal diameter width 3 4 mm particle packing size; Foster City, CA, USA) interfaced to a Micromass Quattro Ultima (Beverly, MA, USA). Laboratory control spikes (n 5 5), matrix spikes (n 5 3), and injection duplicates (all samples) were used for quality control. The limit of detection was approximately 0.5 ng/g dry weight, and the limit of quantification was 10 ng/g dry weight. Lyophilized eggs were initially spiked with a surrogate standard, PCB 204 (Ultra Scientific, North Kingstown, RI, USA). Blanks were run concurrently with the samples to assess possible laboratory contamination. Eggs were subjected to enhanced solvent extraction (Dionex ASE 200, Sunnyvale, CA, USA) with methylene chloride. Large–molecular-weight-biogenic compounds were separated from the PBDEs by chromatography of the extracts on an Envirosep size-exclusion column (350 mm length 3 21.2 mm diameter with 60 3 21.1 mm guard column; Phenomenex, Torrance, CA, USA). The resulting fraction of interest was further purified on 2,000-mg, silica gel, solid-phase extraction columns (EnviroPrep; Burdick and Jackson, Muskegon, MI, USA). Following solvent exchange to hexane, the PBDEs in the purified extracts were separated on a gas chromatograph (Varian 3400, Sugar Land, TX, USA) equipped with a DB-5 column (60 m length 3 0.32 mm inner diameter 3 0.25 m film thickness; J&W Scientific, Folsom, CA, USA). The carrier gas was He. Injections were made in the splitless mode with pentachlorobenzene as the internal standard. Data were corrected based on recovery of PCB 204 in each sample. Identification was accomplished by mass spectrometry in the fullscan, electron-ionization mode (Varian 4D ion trap). Quantitation was done by comparison of the sum of the areas of the three major ions of each polybrominated diphenyl ether (BDE) congener (BDE 47, 49, 99, 100, 153, and 154; Cambridge Isotope Laboratories, Andover, MA, USA) versus that of the internal standard. Several additional PBDE congeners for which standards were not available were identified using mass spectrometry by their degree of bromination. These were quantitated by assuming a response similar to standards with similar degrees of halogenation. Because the stage or state of egg samples (e.g., freshly laid, embryonated, addled or failed to hatch) varied, the volume of each egg was estimated and the concentrations of contaminants in eggs were adjusted to account for moisture loss occurring during incubation [23]. All contaminant concentrations are presented on a wet-weight basis. 620 Environ. Toxicol. Chem. 24, 2005 Toxic equivalents The toxicity of Ah receptor–active PCB congeners in each sample was estimated by summing the products of the congener concentrations and toxic equivalency factors [24]. Biomarkers of exposure and effects Eggs (n 5 45) were cut in half to empty the contents, and the eggshells were rinsed with distilled water (membrane left intact) and dried at room temperature. Eggshell thickness, a biomarker for p,p9-DDE exposure, was determined as the average of three measurements around the equator using a dialgauge micrometer accurate to 0.01 mm (model 1010M; L.S. Starrett, Athol, MA, USA) [10]. Genetic damage was assessed in nestling blood samples (n 5 29) based on the coefficient of variation of cellular DNA content using flow cytometric methods [25,26]. Samples were randomized before processing to avoid experimental bias. Nuclear suspensions from blood samples were treated with RNase, and the remaining DNA was stained with propidium iodide. A Coulter Elite flow cytometer (Beckman Coulter, Fullerton, CA, USA) with a coherent laser at 514 nm and 500 mW of power was used to quantify fluorescence from nuclei. Cells were gated on side scatter, forward scatter, and the ratio of peak to integrated fluorescence. Ten thousand nuclei, which satisfied all gating parameters, were measured from each sample, and the intercellular variation in DNA content was reported as the half-peak coefficient of variation (CV) or the full-peak CV. Samples that did not have 10,000 nuclei counted by the end of the 4-min run were excluded from subsequent statistical analyses. High coefficients of variation are indicative of genetic damage, possibly because of exposure to PAHs or metals [26,27]. Statistical methods Continuously distributed variables, including contaminant concentrations, culmen length, body weight, full-peak DNA CV, and eggshell thickness, were tested for homogeneity of variance and log transformed to stabilize variances when appropriate. Differences in these variables among study areas were determined using analysis of variance and Tukey’s honestly significant difference method of multiple comparisons or, when appropriate, the Savage score test and the Wilcoxon two-sample t test with a Bonferroni correction (SASt Ver. 8, SAS Institute, Cary, NC, USA). Reproductive success was determined using standard methods (number of occupied nests, successful pairs, young fledged per successful nest, young per occupied nest) and by the Mayfield method [28,29]. Productivity differences among regions were determined using Fisher’s exact test with a Bonferroni correction when appropriate [30]. Linear regression was used to examine contaminant concentrations for geographic trends. The relation between contaminant concentrations in the sample egg and the fate of the remaining eggs in that nest was examined using logistic regression [31]. All organochlorine contaminants in the model ( p,p9-DDE, chlordane and metabolites, heptachlor epoxide, and total PCBs) were correlated, so a principle components analysis was used. These particular contaminants were selected for modeling, because they were found at highest concentrations in eggs relative to their known effect levels. RESULTS Contaminant exposure Osprey eggs from the north (study segment 2) had greater concentrations of organochlorine pesticides or metabolites P.C. Toschik et al. than eggs from other segments (Table 1). Nine of 12 eggs collected from the north had p,p9-DDE concentrations within the 95% confidence interval associated with 10 to 15% eggshell thinning [10]. The following compounds were detected only in eggs collected from the north, at concentrations less than 0.01 mg/g: o,p9-DDE (n 5 2), g-chlordane (n 5 12), and pentachloroanisole (n 5 1). Other organochlorine pesticides or metabolites (aldrin, heptachlor, a-hexachlorocyclohexane [HCH]), D-HCH, g-HCH, and toxaphene) were not detected. Mercury was detected in all eggs at concentrations less than 0.3 mg/g (Table 1). Total PCBs, predominantly Aroclors 1254 and 1260 (Monsanto, St. Louis, MO, USA), also were greatest in the north (study segment 2), with the minimum northern concentration exceeding the maximum concentrations from all other study segments (Table 2). Aroclor 1242 was not detected in any eggs; Aroclor 1248 was detected in one egg from the river (study segment 1) and one egg from the north at concentrations less than 0.5 mg/g. Exposure to total PCBs, Aroclor 1254, Aroclor 1260, and PCB congeners 105, 118, 126, 128, 156, 166, 138/ 160, and 170/190 in the north was greater than at all other sites excluding the river; concentrations of PCB congeners 158, 167, and 189 in eggs from ospreys in the north were significantly greater than those for all other regions. Total PCBs, Aroclors, and coplanar PCB congener concentrations in eggs from the river segment were most similar to those from the northern and central segments. A significant north-to-south gradient in PCB concentrations in eggs was detected (latitude vs total PCBs, Pearson correlation, r2 5 0.471, p , 0.0001, n 5 39) (Fig. 2a). Many Ah receptor–active PCB congeners were found in higher concentrations in eggs from the north, but several of the most potent PCB congeners (77, 81, and 169) and toxic equivalents did not differ among study segments. The predominant perfluorinated compounds detected in osprey eggs were perfluorooctanesulfonate and perfluoroundecanoic acid (Table 3). Perfluorodecanoic acid, perfluoroundecanoic acid, perfluorododecanoic acid, and perfluorodecanesulfonate generally were greater in the northern segment compared to other study segments. Concentrations of perfluorodecanesulfonate exhibited a north-to-south gradient, with the highest concentrations occurring in the north and decreasing toward the south (Spearman rank-order correlation, r2 5 0.692, p 5 0.0001, n 5 15). Trace quantities (,10 ng/g wet wt) of perfluorooctanoic acid and perfluorononanesulfonate were detected in all eggs with the exception of those from the river segment. Perfluorundecanesulfonate and perfluorododecanesulfonate were detected in trace quantities in a few eggs from the northern and central segments. Total PBDEs; congeners 47, 99, 100, 153, and 154; and the unidentified congener hexa-a were detected in all eggs analyzed (n 5 15). However, the unidentified congener hexa-b was not detected in any sample, and mixture hexa-c was detected only in one of one egg from the river segment, four of five eggs from the northern segment, and four of six eggs from the central segment (Table 4). The only egg containing quantifiable concentrations of the unidentified congener penta-a was collected from a nest located on a channel marker at the mouth of Little Creek, which receives drainage from parts of Dover Air Force Base. Total PBDEs exhibited a latitudinal trend, with the highest concentrations occurring in the north and decreasing toward the south (Spearman rank-order correlation, r2 5 0.842, p , 0.0001, n 5 15) (Fig. 2b). All congeners and total PBDEs were significantly ( p , 0.05) higher in eggs collected Environ. Toxicol. Chem. 24, 2005 Contaminant exposure and reproduction of Delaware Bay osprey 621 Table 1. Wet weight concentrations (mg/g) of organochlorine pesticides or metabolites, mercury, and biomarker responses in samples from Delaware River and Bay area (USA)a Matrix Analyte (mg/g) Egg p,p9-Dichlorodiphenyldichloroethylene p,p9-Dichlorodiphenyldichloroethane p,p9-DDT o,p9-Dichlorodiphenyldichloroethylene o,p9-Dichlorodephenyltrichloroethane o,p9-Dichlorodiphenyldichloroethane Dieldrin Endrin Heptachlor epoxide a-chlordane cis-Nonachlor trans-Nonachlor Oxychlordane Mirex Hexachlorobenzene Mercury Shell thickness (mm, arithmetic mean 6 standard error) % Difference from pre-DDT era Blood Biomarker - full-peak coefficient of variation (arithmetic mean 6 standard error) n a South (n 5 12) Central (n 5 12) 0.352C 0.172–1.20 12 0.017C 0.006–0.038 12 0.003 ND–0.004 4 — ND 0 0.007B 0.002–0.011 12 0.012C 0.005–0.018 12 0.004C 0.002–0.012 12 — ND 0 0.004B 0.001–0.013 12 0.004AB ND–0.006 3 0.004C 0.002–0.011 12 0.002C 0.001–0.011 12 0.005B 0.001–0.010 12 0.006 0.001–0.058 12 — ND 0 0.05 0.020–0.14 12 0.852B 0.349–4.61 12 0.057B 0.022–0.216 12 0.003 0.001–0.013 12 — ND 0 0.010B 0.005–0.019 12 0.020B 0.011–0.039 12 0.008B 0.003–0.026 12 0.001 0.001–0.005 12 0.007B 0.003–0.015 12 0.003B 0.001–0.010 12 0.010B 0.003–0.023 12 0.005AB 0.002–0.009 12 0.008B 0.004–0.016 12 0.005 0.002–0.013 12 — 0.002–0.002 2 0.09 0.02–0.27 12 0.468 6 0.009BC 29.3 4.05 6 0.12 9 0.501 6 0.011AB 20.8 3.98 6 0.075 9 North (n 5 12) 1.77A 0.785–3.84 12 0.182A 0.089–0.323 12 0.004 0.002–0.016 12 — ND–0.003 2 0.028A 0.018–0.052 12 0.038A 0.024–0.066 12 0.041A 0.026–0.084 12 0.002 0.001–0.005 12 0.021A 0.014–0.036 12 0.007A 0.002–0.018 12 0.027A 0.014–0.052 12 0.010A 0.003–0.030 12 0.025A 0.016–0.046 12 0.008 0.004–0.014 12 0.005 0.004–0.014 12 0.06 0.03–0.23 12 0.454 6 0.010C 210.1 3.98 6 0.065 9 River (n 5 3) 0.656ABC 0.394–1.05 3 0.031BC 0.027–0.036 3 0.003 ND–0.003 2 — ND 0 0.009B 0.003–0.016 3 0.017BC 0.015–0.019 3 0.012BC ND–0.025 2 — ND 0 0.007B 0.004–0.012 3 0.004AB ND–0.005 2 0.005BC 0.004–0.005 3 0.003BC 0.002–0.005 3 0.010B 0.006–0.014 3 0.004 0.003–0.006 3 — ND 0 0.04 0.030–0.061 3 0.527 6 0.008A 4.4 3.90 6 0.10 2 The first entry for an analyte represents the geometric mean, the second the extremes, and the third the n quantifiable. Means with different capital letters are significantly different by Tukey’s honestly significant difference method of multiple comparison ( p , 0.05). ND 5 not detected; — 5 no mean value calculated because contaminant was detected in fewer than half the samples. from nests in the northern segment compared to those collected from the central segment. The fraction of total PBDEs contributed by BDE 47 exhibited a distinct latitudinal trend (Spearman rank-order correlation, r2 5 0.322, p 5 0.027, n 5 15), with the proportion increasing from north to south (minimum, 31%; maximum, 69%) (Fig. 2b). Reproductive success Nest success and productivity in the Delaware River and Bay study regions generally were similar and fell within the range estimated to maintain a stable osprey population [32]. Although number of eggs laid, percentage of eggs hatched, percentage of eggs lost, number of nestlings lost, percentage 622 Environ. Toxicol. Chem. 24, 2005 P.C. Toschik et al. Table 2. Wet-weight concentrations of total polychlorinated biphenyls (PCBs) and arylhydrocarbon (Ah) receptor-active PCB congeners in osprey eggs collected in the Delaware River and Bay area (USA)a Analyte Total PCBs (mg/g) Aroclor 1254 (mg/g) Aroclor 1260 (mg/g) Ratio of Aroclors 1254 to 1260 PCB 77 (pg/g) PCB 81 (pg/g) PCB 105 (ng/g) PCB 118 (ng/g) PCB 126 (pg/g) PCB 128 (ng/g) PCB 156 (ng/g) PCB 158 (ng/g) PCB 166 (pg/g) PCB 167 (ng/g) PCB 169 (pg/g) PCB 189 (ng/g) PCB 138/160 (ng/g) PCB 170/190 (ng/g) Toxic equivalents (pg/g) a South (n 5 12) Central (n 5 12) North (n 5 12) River (n 5 3) 1.44C 0.469–2.43 12 0.551C 0.212–1.20 12 0.876C 0.259–1.44 12 0.63 0.427–1.00 12 97 36–366 12 698 170–2,634 12 17.1B 5.44–34.2 12 72.9B 25.4–143 12 200B 103–383 12 22.2C 7.49–42.9 12 16.6C 4.95–26.3 12 15.7C 5.01–28.1 12 189C 44–3,351 12 11.9B 3.63–20.4 12 — ND–103 1 2.00C 0.618–3.55 12 163C 48.3–332 12 32.7C 9.68–55.0 12 108 45.5–299 12 2.44B 1.23–5.15 12 0.829BC 0.470–1.80 12 1.55B 0.676–4.35 12 0.53 0.111–0.819 12 135 37–263 12 482 142–7,616 12 24.2B 11.4–43.9 12 114B 48.3–205 12 228B 103–449 12 33.6BC 14.4–63.1 12 25.0BC 9.31–46.5 12 28.1B 11.6–61.1 12 552B 118–793 12 18.3B 8.04–34.9 12 — ND–111 3 3.95B 1.65–9.39 12 240BC 100–489 12 59.9B 25.1–154 12 90.3 35.2–828 12 8.68A 5.50–14.5 12 3.52A 2.23–6.53 12 5.01A 2.48–8.29 12 0.70 0.428–1.37 12 238 42–752 12 585 190–2,050 12 96.4A 58.9–179.9 12 411A 262–701 12 557A 284–964 12 117A 73.5–179 12 116A 62.2–251 12 139A 78.6–237 12 3,999A 2,265–9,203 12 57.4A 38.5–87.6 12 75 39–133 12 13.6A 7.60–21.4 12 769A 449–1,303 12 239A 139–408 12 160 68.4–341 12 4.38AB 4.17–4.70 3 1.67AB 0.835–2.36 3 2.38AB 1.72–3.34 3 0.70 0.250–1.38 3 195 105–292 3 357 344–370 3 57.5A 36.5–111 3 259A 201–343 3 367AB 287–504 3 68.1AB 55.7–97.3 3 53.7AB 39.8–89.6 3 60.3B 41.7–92.2 3 4,773A 1,999–8,887 3 21.6B 16.8–27.7 3 — ND 0 3.91BC 2.94–5.07 3 475AB 428–505 3 110AB 73.9–142 3 97 78.6–126 3 The first entry for an analyte represents the geometric mean, the second the extremes, and the third the n quantifiable. Means with different capital letters are significantly different by Tukey’s honestly significant difference method of multiple comparison ( p , 0.05). ND 5 not detected; — 5 no mean value calculated because contaminant was detected in fewer than half the samples. of nestlings fledged, and number of nests failed did not differ among the segments ( p . 0.05, Fisher’s exact test) (Table 5), significantly more eggs were lost in the northern than in the central segment ( p , 0.05). Fledglings produced per active nest ranged from 1.00 in the northern and river segments to 1.47 in the central segment. Survival probabilities during the egg-laying period and incubation were high and similar among study segments. Productivity estimates derived using the Mayfield method were nearly identical to observed productivity. Environ. Toxicol. Chem. 24, 2005 Contaminant exposure and reproduction of Delaware Bay osprey 623 Fig. 2. Relative concentrations of total polychlorinated biphenyls (PCBs; a) and polybrominated diphenyl ethers (PBDEs; b) congeners in sample osprey eggs. Note that the relative contribution of BDE 47 to total PBDEs increases from north to south. Table 3. Wet-weight concentrations (ng/g) of perfluorinated compounds in osprey eggs collected in the Delaware River and Bay area (USA)a Analyte (ng/g) Perfluorononanoic acid Perfluorodecanoic acid Perfluoroundecanoic acid Perfluorododecanoic acid Perfluoroctanesulfonate Perfluorodecanesulfonate a South (n 5 2) Central (n 5 6) North (n 5 6) River (n 5 1) — 2.38 1 2.66B 1.63–4.33 2 7.00B 4.91–9.98 2 1.89B 1.58–2.67 2 37.8B 33.8–42.3 2 — 1.24 1 10.0 3.30–29.1 6 7.43B 5.23–12.6 6 31.8B 13.9–86.8 6 4.69B 2.69–7.28 6 96.9AB 37.4–370 6 4.96B 2.20–11.6 6 27.8 4.49–50.1 6 29.8A 9.54–69.5 6 121A 66.0–221 6 31.8A 10.8–72.7 6 293A 127–799 6 26.8A 8.99–52.4 6 — NQ 0 — 4.74 1 — 12.8 1 — 3.69 1 — 122 1 — 8.98 1 A subset of 15 samples were selected for these analyses based on geographical distribution, rather than equal representation of study segments. Means with different capital letters are significantly different by Tukey’s honestly significant difference method of multiple comparison ( p , 0.05). The first entry for an analyte is the geometric mean, the second the extremes, and the third the n quantifiable. NQ 5 not quantifiable; — 5 no mean value calculated because contaminant was found in fewer than half the samples. 624 Environ. Toxicol. Chem. 24, 2005 P.C. Toschik et al. Table 4. Wet-weight concentrations (ng/g) of polybrominated diphenyl ethers (PBDEs) in osprey eggs collected in the Delaware River and Bay Region area (USA)a Analyte Congener 47 Congener 99 Congener 100 Congener 153 Congener 154 Hexa-a Hexa-c Total PBDEs a South (n 5 2) Central (n 5 6) North (n 5 6) River (n 5 1) 46.6 43.0–50.1 2 7.93 6.20–9.66 2 12.40 10.5–14.3 2 5.32 2.80–7.83 2 7.87 3.44–12.3 2 2.08 1.43–2.72 2 — ND 0 82.2AB 70.9–93.5 2 124B 90.7–231 6 18.0 8.61–52.8 6 34.5 22.2–85.1 6 10.8 5.93–28.2 6 12.5 6.92–21.0 6 3.32 1.73–5.28 6 2.15 ND–2.65 4 206B 141–429 6 276A 223–453 6 111 45.8–228 6 99.3 62.3–155 6 40.1 11.1–93.3 6 36.6 12.9–68.5 6 2.99 1.97–3.96 5 2.59 ND–3.21 4 572A 442–820 6 — 227 1 — 140 1 — 82.0 1 — 61.2 1 — 43 1 — 2.27 1 — 2.50 1 — 557 1 A subset of 15 samples were selected for these analyses based on geographical distribution, rather than equal representation of study segments. Hexa-a and hexa-c are hexa-brominated compounds that have not yet been specifically identified; concentrations are corrected for recovery of polychlorinated biphenyl 204 surrogate standard and loss of moisture during incubation. Means with different capital letters are significantly different by the Savage score test and Wilcoxon two-sample t test ( p , 0.05). The first entry for an analyte is the mean, the second the extremes, and the third the n quantifiable. ND 5 not detected; — 5 no mean value calculated because contaminant was found in fewer than half the samples. Morphological and biochemical endpoints All nestlings appeared to be in good health; no external lesions or other abnormalities were observed. Body weight and culmen length at the time of blood collection were greater in the northern segment (mean 6 standard error; 1,631 6 38.7 g and 30.3 6 0.46 mm, respectively) than in the southern segment (1,357 6 65.2 g and 27.9 6 0.61 mm, respectively; p , 0.05). Because both parameters were greater in the north, this probably indicates that nestlings from southern Delaware were either slightly younger at the time of measurement or that the birds actually were smaller in the south because of latitudinal variation in body size, as recently noted in Chesapeake Bay ospreys [15]. One nestling in the Inland Bays appeared to develop more slowly than other nestlings (based on weight and age), although this may have resulted from inexperienced parents. Fishing line was observed in several nests, and one nestling in Indian River Bay had to be cut free from fishing line entangling its legs and wings on two consecutive nest visits. Eggshell thickness was significantly greater in the river segment than in the north or south; however, this may reflect small sample size from the river segment rather than a true difference (Table 1). Eggshells from the northern segment exhibited 10.1% thinning compared to pre-DDT era eggs and were thinner ( p , 0.05) than those from the central segment, which was consistent with the difference in p,p9-DDE con- centrations. However, eggshell thickness was not correlated with p,p9-DDE concentration (Pearson correlation, r2 5 0.17, p 5 0.23, n 5 39). The CV in blood cell DNA content, as a biomarker of genetic damage, is presented as full rather than half peak, because these data had a relatively strong signal to noise ratio. Full-peak CVs did not differ among study segments, suggesting that cell genetic damage did not differ among segments. Relation of contaminants and reproduction Contaminants with the highest concentrations relative to their known toxicity ( p,p9-DDE, chlordane and metabolites, heptachlor epoxide, and total PCBs) were examined further for relationships with hatching and fledging success. Individually, the aforementioned contaminants were associated with hatching success (logistic regression model, Pearson correlation, r2 5 0.053–0.141, p , 0.1, n 5 39) but not with fledging success. However, because these four compounds were highly correlated with one another (Pearson correlation, p , 0.001), a principal components analysis was used to clarify their association with hatching success. Hatching success could be explained by contaminants in either of two ways: A simple multiplicative model, or a model using the principal components analysis. The logistic regression model revealed that hatching success was a function of the product of ( p,p9-DDE)(total PCBs)(total chlordane metabolites)(heptachlor epoxide), which is the same as the sum of the log-transformed concentrations of these contaminants (r2 5 0.104, p , 0.05, n 5 39). The principal components analysis weighted the four chemical groups almost equally within the first component, which explained 91.8% of the variation. The equation for the equivalent logistic regression model predicting hatching success was ln[ p/(1 2 p)] 5 0.301 1 0.478[log10( p,p9DDE) 1 log10(total chlordane metabolites) 1 log10(heptachlor epoxide) 1 log10(total PCBs)], which is effectively the same equation to describe hatching success as the multiplicative relationship. The predicted probability of hatching success versus contaminant concentrations is presented in Figure 3. Perfluorinated compounds, total PBDEs, and toxic equivalents did not show a significant relationship with the number of fledglings produced per active nest (perfluorinated compounds and total PBDEs, Spearman correlation, p . 0.05; toxic equivalents, Pearson correlation, p . 0.05), although the patterns in the data (especially perfluorodecanesulfonate, perflurodecanoic acid, perfluorononoic acid, and perfluorooctanesulfonate) indicate that further research (with a larger sample size) might be warranted. DISCUSSION The present findings provide evidence that environmental contaminants continue to be one of several significant stressors on osprey productivity in the northern Delaware River and Bay. Our principal hypotheses were not rejected, because we found that contaminant concentrations in osprey eggs in the present study did exhibit a north-to-south gradient and that the observed low hatching success in some areas could be predicted by elevated contaminant concentrations detected in eggs. Environmental contaminant exposure Most organochlorine pesticide and PCB concentrations in osprey eggs exhibited a north-to-south gradient, as one would expect based on sediment, water, and fish contaminant data. The lack of significant differences between the river segment and other study segments likely resulted from the small sample Environ. Toxicol. Chem. 24, 2005 Contaminant exposure and reproduction of Delaware Bay osprey 625 Table 5. Productivity of sampled osprey nests in Delaware River and Bay area (USA) in 2002 South Active nests sampled Eggs laid Eggs naturally incubated Fate of eggs Lost–unknowna Lost–depredated Lost–weather Failed to hatch Hatched Hatchabilityb Fate of nestlings Depredated Found out of nest Disappeared Fledged (%) Successful nests Fledglings/active nest Mayfield method estimates Egg laying and incubation periodc Daily survival rate 6 standard error Survival rate to hatching (A) Nestling period Daily survival rate 6 standard error Survival rate to fledging (B) Nest success (A 3 B) Probability of an egg hatching given that the nest is successful (C) Probability of young living to 53 d given that the nest is successful (D) Egg success (A 3 B 3 C 3 D) Mean clutch size (E) Mean number of young surviving to 53 d (A 3 B 3 C 3 D 3 E ) Mean number of young surviving to 53 d, less sample egg (A 3 B 3 C 3 D 3 (E 2 1)) Central North River 12 38 26 12 40 28 12 39 27 3 9 6 4 2 0 0 20 (77%) 20/20 (100%) 1 1 2 2 22 (79%) 22/24 (92%) 11 2 0 2 12 (44%) 12/14 (86%) 0 0 0 1 5 (83%) 5/6 (83%) 4 1 1 14 (70.0%) 9 of 12 1.17 0 0 5 17 (77%) 9 of 12 1.42 0 0 0 12 (100%) 9 of 12 1.00 0 0 2 3 (60%) 2 of 3 1.00 n 5 12 0.998 6 0.070 0.914 n 5 11 0.996 6 0.068 0.822 0.751 n 5 12 0.995 6 0.084 0.807 n 5 10 0.998 6 0.082 0.899 0.725 n 5 12 0.992 6 0.099 0.740 n58 1 1.00 0.740 n53 1 1.000 n53 0.987 6 0.166 0.503 0.503 0.950 0.857 0.842 1.00 0.737 0.526 3.17 0.944 0.587 3.33 0.750 0.467 3.25 1.00 0.503 3.00 1.66 1.96 1.52 1.51 1.14 1.37 1.05 1.01 Unknowns were considered to be lost at the egg stage (n 5 3 south, n 5 4 north); total eggs lost per region was significantly greater in the northern than in the central segment ( p , 0.05). Hatchability is the number or percentage of eggs that remain in nests throughout incubation that hatch. c A 5 daily survival rate to the 39th power accounting for a 35- to 43-d incubation period; B 5 daily survival rate to the 53rd power accounting for a 50- to 55-d nestling period; C 5 number of eggs that hatched in successful nests divided by the total number of eggs in successful nests; D 5 number of nestlings that fledged in successful nests divided by the total number of nestlings in successful nests; E 5 clutch size (arithmetic mean). a b size (n 5 3) from that region. However, the eggs from the river segment were slightly more contaminated than expected. These data provide further evidence that the osprey is a sentinel of local contamination, because the latitudinal trends seen in egg contaminant exposure are unlikely to result from contaminant exposure on the wintering grounds. For most organochlorine pesticides or metabolites and PCBs, concentrations in eggs from the river segment were similar to those from the central and southern segments. In the present study, concentrations of p,p9-DDE, dieldrin, and total PCBs in eggs were up to twofold greater than those reported by Clark et al. [11] in their 1998 collection from comparable locations. Furthermore, geometric mean and maximum values of these compounds from our northern segment were up to fourfold greater than those reported by Clark et al. but were not unlike the values reported by Steidl et al. [7] from their 1989 collection. Two of the eggs sampled in the present study were from nests on Prime Hook National Wildlife Refuge (central segment); they contained p,p9-DDE concentrations of 0.69 and 0.57 mg/g wet weight. This is considerably less than the 5 mg/g of p,p9-DDE in the only osprey egg analyzed from the state of Delaware before the present study, which coincidentally was collected from the same refuge in 1974 [10]. In contrast, concentrations of p,p9-DDE in eggs from the northern part of the Delaware River and Bay (geometric mean, 1.77 mg/g; range, 0.785–3.84 mg/g) were similar to those observed in eggs from Wisconsin and Michigan (USA) and the Canadian Great Lakes between 1980 and 1996 [33– 35], similar to those of fresh eggs collected in the New Jersey portion of the Delaware Bay in 1989 [7], and greater than those observed in 1998 [11] (the maximum p,p9-DDE concentration observed in 1998 was 1.61 mg/g). Mirex concentrations in the osprey eggs were well below concentrations expected to have effects on avian species [36]. However, the maximum concentrations of mirex in eggs from the northern, central, and southern segments were greater than the 0.01 mg/g dietary threshold considered to be protective for mammalian wildlife [36], including mink (Mustela vison) [37]. Because osprey eggs are depredated by avian and mammalian species, this may present some cause for concern. Although not observed in the present study, other studies have detected mirex and dieldrin more frequently in addled eggs than in randomly sampled eggs [7]. The presence of low concentrations of endrin in all eggs from the central and northern segments suggests a historical source in the main Delaware Bay area, which is consistent with the extensive agriculture in this area. Mercury was detected in all eggs at moderately low concentrations (Table 1), in the same range as concentrations 626 Environ. Toxicol. Chem. 24, 2005 Fig. 3. Relationship of the sum of the log-transformed organochlorine contaminants (OCs; p,p9 -dichlorodiphenyldichloroethylene [ p,p9DDE], chlordane and metabolites, heptachlor epoxide, and total polychlorinated biphenyls [PCBs]) to the probability of egg loss. The equation (diamonds) and 95% confidence intervals (dashes) for the equivalent logistic regression model predicting that the probability of egg loss was ln[ p/(1 2 p)] 5 0.301 1 0.478[log10( p,p9-DDE) 1 log10(total chlordane metabolites) 1 log10(heptachlor epoxide) 1 log10(total PCBs)] (r2 5 0.104, p , 0.05, n 5 39). in osprey eggs from the Chesapeake Bay in 2000 and 2001 [15] and from the Delaware Bay in 1998 [11]. Based on observations in the present study, ospreys in the southern Delaware River and northern Delaware Bay are being exposed to high concentrations of PCBs, which is in contrast to recent findings in the central part of the bay [11]. Polychlorinated biphenyl concentrations in the northern and river segments remain in the range of 1989 samples from the Delaware Bay [7]. Despite the significant differences among regions for total PCB concentrations, values were within the range to cause cytochrome P450 induction or immunotoxicity in some species of wild birds [38]. All eggs in the northern segment and one egg from the central segment had total PCB concentrations of greater than 2 mg/g (range, 5.15–14.5 mg/ g), which is the maximum recommended concentration in fish in the United States for human consumption [39]. Notably, the greatest PCB concentration detected in an egg from the central segment (5.15 mg/g) was from the northernmost nest in the segment, located on a channel marker at the mouth of the Leipsic River adjacent to Bombay Hook National Wildlife Refuge (DE, USA). Similar to our observations for p,p9-DDE, the two eggs sampled from nests on Prime Hook National Wildlife Refuge (central segment) contained concentrations of 2.36 and 1.31 mg/g of PCBs, or approximately half the value observed in 1974 (PCBs, 4.6 mg/g) [10]. In contrast, concentrations of PCBs in eggs in the northern segment were similar to those in eggs from Wisconsin and Michigan between 1980 and 1996 [33,34], similar to those of fresh eggs collected in the New Jersey portion of the Delaware Bay in 1989 [7], and greater than those observed in 1998 [11]. In fact, the lowest PCB concentration in eggs from 2002 in the northern segment (5.50 mg/g) was greater than the maximum concentration seen in eggs collected in 1998 from the New Jersey portion of the bay (4.45 mg/g) [11]. Perfluorinated acid compounds were detected more often, P.C. Toschik et al. whereas sulfonate compounds were detected less often, in eggs collected from Delaware Bay, which is in contrast to recent observations in Chesapeake Bay [15]. Concentrations of perfluorooctanesulfonate in the osprey eggs were similar to those seen in egg yolks and blood serum and plasma in other fisheating water birds in the United States, Japan, and Korea [13,40]. Despite widespread exposure of wildlife to perfluorinated compounds [13,15,40,41], little is known about their toxicity in birds. Recent acute and dietary toxicity studies in northern bobwhites (Colinus virginianus) and mallards (Anas platyrhynchos) suggest that perfluorooctanesulfonate may be moderately to highly toxic (J.P. Giesy, Michigan State University, East Lansing, MI, USA, personal communication). Studies of laboratory rodents indicate that these compounds are hepatotoxic, induce the CYP4A isoenzymes, and can reduce fetal viability [42–44]. Controlled laboratory studies regarding the effects of perfluorinated compounds on birds (e.g., fetal viability, immunotoxicity, growth, and survival of hatchlings) would be useful for interpreting these exposure data. Concentrations of total PBDEs and individual congeners in Delaware osprey eggs were similar to those observed in Chesapeake Bay osprey eggs [15] and within the range of values observed in herring gull (Larus argentatus) eggs collected from the Great Lakes region in 2000 [45]. Osprey muscle samples collected in Sweden from birds that were found dead contained 2,140 ng/g lipid weight of PBDEs, which is somewhat lower than our egg values (when converted to a lipid-wt basis) [46,47]. The latitudinal trends detected in total PBDEs and congeners in eggs collected from the Delaware River and Bay may reflect distance from the source and differences in congener mobility. The predominance of BDE 47 in the southern segment may be attributable to its being more volatile and water soluble and, hence, more mobile and recalcitrant. The lipophilicity, estrogenic activity, effects on thyroid hormone transport, cytochrome P450, behavior, and immune function of some PBDE congeners, as well as the dramatic increase in their concentration in the environment, have garnered a great deal of attention in the scientific community [12,48]. The BDE 47 appears to be one of the more toxicologically potent congeners and the most abundant congener in the Delaware osprey eggs (31–69% of total PBDEs) as well as the most bioavailable congener [12]. Remarkably, to our knowledge, no information exists regarding the toxicity of these compounds to avian species. Productivity and effects of contaminants Ospreys were once common breeders throughout coastal Delaware, although by the mid-1950s, coincident with the advent of organochlorine pesticides, the population declined precipitously [5]. By the mid-1980s, the number of osprey nests had increased, but to date, the population in the Delaware Bay has not returned to pre-1950s levels. Overall productivity for the Delaware River and Bay was within or above the range estimated to maintain a population for ospreys on the east coast of the United States (0.80–1.15 fledglings/active nest) [32]. Productivity generally was greater than that in 1989 (,0.80 fledglings/active nest) [6], and that in the southern and central segments was similar to the 1994-to-1998 productivity for ospreys in the Salem and Maurice River areas of New Jersey (1.10 fledglings/active nest) [11]. However, productivity in the northern and river study segments in 2002 was only 1.00 fledgling per active nest, which is marginal to maintain the population. Contaminants may be one of the causal factors in the slow recovery of the ospreys. In support of this contention, Contaminant exposure and reproduction of Delaware Bay osprey p,p9-DDE concentrations in the central and northern segments were in the range that causes 10 to 15% eggshell thinning in ospreys. Concentrations of organochlorine contaminants, perfluorinated compounds, and PBDEs were greatest in these regions. Furthermore, using these data, a logistic regression model (Fig. 3) indicated that concentrations of p,p9-DDE, chlordane and metabolites, heptachlor epoxide, and total PCBs were predictive ( p , 0.05) of hatching success. However, this should be viewed not as a cause–effect relationship but, merely, as an indication that these and, presumably, other lipophilic contaminants may act synergistically to impair hatching success. Bald eagles residing in New Jersey tributaries of the northern Delaware Bay between 1988 and 1997 had poor productivity (0–0.14 fledglings/nest), which has been attributed to high p,p9-DDE and, possibly, PCB concentrations in eggs [49]. Low bald eagle productivity (0.75 fledglings/nest) also has been observed in northern Delaware (K. Heckscher, Delaware Natural Heritage Program, Smyrna, DE, USA, unpublished data). In the present study, reproductive problems were most pronounced during incubation and hatching. Surprisingly, p,p9DDE and eggshell thickness were not correlated, which may be attributable to the limited number of samples having p,p9DDE concentrations greater than 2 mg/g or greater than 10% eggshell thinning. Contaminant concentrations apparently were below thresholds that adversely affect fledging success, and no evidence of impaired growth was found. Additionally, no evidence of chromosomal damage in nestlings was found. The major contaminants found in the eggs are not thought to be highly clastogenic. However, fishing line as well as other plastics and debris commonly found in osprey nests can entangle nestlings and may be a looming problem in some portions of the Delaware and Chesapeake bays [15]. Education of anglers regarding proper disposal of fishing equipment and enforcement of such disposal are especially important in areas with high densities of both ospreys and anglers, such as the Inland Bays and Lewes (DE, USA). The results of the logistic regression model make sense considering observations of individual nests. The three nests that failed in the southern segment all exhibited evidence of predation, and all had low concentrations of p,p9-DDE and total PCBs. In the central segment, one of the nests that failed apparently was blown off the channel marker; all nestlings disappeared from a second nest, which is suggestive of predation. The third nest that failed in the central segment was located at the mouth of Little Creek, a tributary that received drainage from part of Dover Air Force Base. The sample egg and an addled egg from that nest contained high concentrations of p,p9-DDE (4.3 and 4.6 mg/g, respectively); these p,p9-DDE concentrations exceeded the 4.2 mg/g threshold associated with 15%, respectively, eggshell thinning [10]. Total PCB values in these eggs were moderate (4.4 and 4.8 mg/g, respectively). In the northern segment, cormorants (Phalcrocorax auritus) predated one nest. However, high concentrations of p,p9-DDE and total PCBs were found in eggs from two nests that failed near New Castle and Trenton (3.6 and 3.8 mg/g of p,p9-DDE, respectively, and 14.5 and 11.3 mg/g of PCBs, respectively). Incidentally, an addled egg from the nest near Trenton contained 4.6 mg/g p,p9-DDE and 12.6 mg/g of PCBs. The sample egg from the nest that failed in the riverine segment had low p,p9-DDE (0.39 mg/g) and moderate PCBs (4.29 mg/g). In summary, of the nests that failed, six exhibited evidence of predation or weather-related loss, three had high p,p9-DDE Environ. Toxicol. Chem. 24, 2005 627 concentrations in the sample egg (two of these three nests also contained addled eggs), and one was lost for unknown reasons. Further examination of individual nest failures indicates that most result from depredation, with contaminants contributing significantly in the more highly polluted areas. Competition for nesting structures in the northern segment also may be a contributing factor to the slow recovery of ospreys in the northern bay area. Cormorants were observed breeding on an osprey nest and on all levels of the tiered channel marker that held the original nest. The osprey eggs were found broken on this platform, presumably predated by the cormorants during a hostile nest-site takeover. Cormorants also occupied several other channel markers in the northern segment near New Castle that were of suitable design for osprey nesting. The larger number of eggs lost in the northern segment may reflect greater predation pressure, competition between ospreys and cormorants for nesting structures, or even stress from multiple factors, including contaminant exposure. CONCLUSION The outlook for ospreys in Delaware River and Bay is improving, but the population still faces challenges in its recovery from the past population decline. Concentrations of many chemicals in osprey eggs have decreased during the last few decades (e.g., p,p9-DDE). However, human impacts on this population are still measurable, and chemicals known to be harmful to laboratory animals consistently are detected in the osprey eggs. Further research to develop effect concentrations of perfluorinated chemicals and PBDEs in birds should be a priority for comparison with levels in this region and other urban wildlife habitats. Future monitoring of the nesting activity, breeding success, and contaminant exposure (i.e., organochlorine contaminants, perfluorinated chemicals, and PBDEs) of ospreys throughout the Delaware River and Bay should provide wildlife managers with an excellent sentinel species and bioindicator for the health of the coastal habitat. Acknowledgement—Funding was provided by the U.S. Geological Survey–Biomonitoring of Environmental Status and Trends Program, U.S. FWS, and the Delaware Coastal Programs. Cole Matson was supported by the National Institute of Environmental Health Sciences Grant ES04917. We thank E. Broderick, C. Koppie, B. Lantz, T. Lucas, J. Miller, and R. Scarborough for assistance in field collections. We acknowledge Bombay Hook and Prime Hook National Wildlife Refuges, Conectiv, Delaware Coastal Programs, Delaware Natural Heritage Program, Delaware Fish and Wildlife Service Police, Delaware State Parks, Eastern Industries, Neshaminy State Park (PA, USA), University of Delaware at Lewes, U.S. Coast Guard, and others for granting permission and facilitating access to some osprey nests. We thank J. Hatfield for advice on the power analysis. REFERENCES 1. Kausch RC, Collier CR, Featherstone JP, Fikslin TJ. 2000. Delaware River and Bay water quality assessment. 2000 305 (b) Technical Report. Delaware River Basin Commission, West Trenton, NJ, USA. 2. National Oceanic and Atmospheric Administration. 2001. Magnitude and extent of contaminated sediment and toxicity in Delaware Bay. NOS ORCA 148 Technical Report. Centers for Coastal Monitoring and Assessment, National Centers for Coastal Ocean Science, Silver Spring, MD, USA. 3. McCoy DL, Jones JM, Anderson JW, Harmon M, Hartwell I, Hameedi J. 2002. Distribution of cytochrome P4501A1-inducing chemicals in sediments of the Delaware river-bay system, USA. Environ Toxicol Chem 21:1618–1627. 4. Sutton CC, O’Herron JC II, Zappalorti RT. 1996. The scientific characterization of the Delaware Estuary. DRBC Project 321; HA File 93.21 Technical Report. The Delaware Estuary Program, West Trenton, NJ, USA. 628 Environ. Toxicol. Chem. 24, 2005 5. Hess GK, West RL, Barnhill MV III, Fleming LM. 2000. Birds of Delaware. University of Pittsburgh Press, Pittsburgh, PA, USA. 6. Steidl RJ, Griffin CR, Niles LJ. 1991. Differential reproductive success of ospreys in New Jersey (USA). J Wildl Manag 55:266–272. 7. Steidl RJ, Griffin CR, Niles LJ. 1991. Contaminant levels of osprey eggs and prey reflect regional differences in reproductive success. J Wildl Manag 55:601–608. 8. Steidl RJ, Griffin CR, Niles LJ, Clark KE. 1991. Reproductive success and eggshell thinning of a reestablished peregrine falcon population. J Wildl Manag 55:294–299. 9. U.S. Fish and Wildlife Service. 1995. Evaluation of environmental contaminant residues in Delaware Bay bald eagle nestlings. Technical Report. New Jersey Field Office, Pleasantville, NJ. 10. Wiemeyer SN, Bunck CM, Krynitsky AJ. 1988. Organochlorine pesticides polychlorinated biphenyls and mercury in osprey eggs 1970–1979 and their relationships to shell thinning and productivity. Arch Environ Contam Toxicol 17:767–788. 11. Clark KE, Stansley W, Niles LJ. 2001. Changes in contaminant levels in New Jersey osprey eggs and prey, 1989 to 1998. Arch Environ Contam Toxicol 40:277–284. 12. de Witt C. 2002. An overview of brominated flame retardants in the environment. Chemosphere 46:583–624. 13. Kannan K, Franson JC, Bowerman WW, Hansen KJ, Jones PD, Giesy JP. 2001. Perfluorooctanesulfonate in fish-eating water birds including bald eagles and albatrosses. Environ Sci Technol 35: 3065–3070. 14. U.S. Environmental Protection Agency. 1997. The index of watershed indicators. EPA-841-R-97-010. Office of Wetlands, Oceans, and Watersheds, Washington, DC. 15. Rattner BA, McGowan PC, Golden NH, Hatfield JS, Toschik PC, Lukei RF Jr, Hale RC, Schmitz-Afonso I, Rice CP. 2004. Contaminant exposure and reproductive success of ospreys ( Pandion haliaetus) nesting in Chesapeake Bay regions of concern. Arch Environ Contam Toxicol 47:126–140. 16. Spitzer PR, Risebrough RW, Walker W II, Hernandez R, Poole AF, Puleston D, Nisbet ICT. 1978. Productivity of ospreys in Connecticut-Long Island increases as DDE residues decline. Science 202:333–335. 17. Custer TW, Pendleton G, Ohlendorf HM. 1990. Within- and among-clutch variation of organochlorine residues in eggs of black-crowned night herons. Environ Monit Assess 15:83–89. 18. Qian Y, Sericano JL, Wade TL. 1998. Extraction of biological tissues for trace organic analysis. NOAA NOS ORCA 130 Technical Report. National Oceanic and Atmospheric Administration, Silver Spring, MD, USA. 19. Sericano JL, Gardinali P, Wade TL. 1998. Quantitative determination of chlorinated hydrocarbons. NOAA NOS ORCA 130 Technical Report. National Oceanic and Atmospheric Administration, Silver Spring, MD, USA. 20. Cromartie EW, Reichel WL, Locke LN, Belisle AA, Kaiser TE, Mulhern BM, Prouty RM, Swineford DM. 1975. Residues of organochlorine pesticides and polychlorinated biphenyls and autopsy data for bald eagles, 1971–1972. Pestic Monit J 9:11–14. 21. Wade TL, Atlas EL, Broks JM, Kennicutt MC, Fox RG, Sericano J, Garcia-Romero B, De Freitas D. 1988. NOAA Gulf of Mexico Status and Trends Program: Trace organic contaminant distribution in sediments and oysters. Estuaries 3:171–179. 22. U.S. Environmental Protection Agency. 1991. Method 245. 6. Determination of mercury in tissues by cold vapor atomic absorption spectrometry, Rev. 2.3. In Methods for the Determination of Metals in Environmental Samples. Technical Report EPA/600/4-91-010. Washington, DC. 23. Stickel LF, Wiemeyer SN, Blus LJ. 1973. Pesticide residues in eggs of wild birds: Adjustment for loss of moisture and lipid. Bull Environ Contam Toxicol 9:193–196. 24. Van den Berg ML, Birnbaum L, Bosveld ATC, Brunstrom B, Cook P, Feeley M, Giesy JP, Hasegawa R, Kennedy SW, Kubiak T, Larsen JC, Liem AK, Nolt C, Peterson RE, Poellinger L, Safe S, Schrenk D, Tillitt D, Tysklind M, Younes M, Waern F, Zacharewski T. 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, and PCDFs for humans and wildlife. Environ Health Perspect 106:775–792. 25. Custer TW, Bickham JW, Lyne TB, Lewis T, Ruedas LA, Custer CM, Melancon MJ. 1994. Flow cytometry for monitoring contaminant exposure in black-crowned night-herons. Arch Environ Contam Toxicol 27:176–179. 26. Theodorakis CW, Swartzm CD, Rogers WJ, Bickham JW, Donnelly KC, Adams SM. 2000. Relationship between genotoxicity, P.C. Toschik et al. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. mutagenicity, and fish community structure in a contaminated stream. J Aquat Ecosys Stress Recovery 7:131–143. Custer TW, Custer CM, Hines RK, Sparks DW, Melancon MJ, Hoffman DJ, Wickliffe JK. 2000. Mixed-function oxygenases, oxidative stress, and chromosomal damage measured in lesser scaup wintering on the Indiana Harbor Canal. Arch Environ Contam Toxicol 38:522–529. Mayfield HF. 1961. Nesting success calculated from exposure. Wilson Bull 73:255–261. Bart J, Robson DS. 1982. Estimating survivorship when the subjects are visited periodically. Ecology 63:1078–1090. Everitt BS. 1977. The Analysis of Contingency Tables. Chapman & Hall, New York, NY, USA. Hosmer DW Jr, Lemeshow S. 1989. Applied Logistic Regression. John Wiley & Sons, New York, NY, USA. Poole AF. 1989. Ospreys: A Natural and Unnatural History. Cambridge University Press, New York, NY, USA. Karasov WH, Meyer MW. 2000. Testing the role of contaminants in depressing avian numbers. Revista Chilena de Historia Natural 73:461–471. Ewins PJ, Postupalsky S, Hughes KD, Weseloh DV. 1999. Organochlorine contaminant residues and shell thickness of eggs from known-age female ospreys ( Pandion haliaetus) in Michigan during the 1980s. Environ Pollut 104:295–304. Martin PA, de Solla SR, Ewins P. 2003. Chlorinated hydrocarbon contamination in osprey eggs and nestlings from the Canadian Great Lakes Basin, 1991–1995. Ecotoxicology 12:209–224. Eisler R. 2000. Handbook of Chemical Risk Assessment: Health Hazards to Humans, Plants, and Animals, Vol 2—Organics. Lewis, Boca Raton, FL, USA. Heaton SA, Bursian SJ, Giesy JP, Tillitt DE, Render JA, Jones PD, Verbrugge DA, Kubiak TJ, Aulerich RJ. 2004. Dietary exposure of mink to carp from Saginaw Bay, Michigan. 1. Effects on reproduction and survival, and the potential risks to wild mink populations. Arch Environ Contam Toxicol 28:334–343. Rice CP, O’Keefe PW, Kubiak TJ. 2003. Sources, pathways and effects of PCBs, dioxins, and dibenzonfurans. In Hoffman DJ, Rattner BA, Burton GA Jr, Cairns J Jr, eds, Handbook of Ecotoxicology. Lewis, Boca Raton, FL, USA, pp 501–573. U.S. Environmental Protection Agency. 1992. National study of chemical residues in fish. Vol II. 823-R-92-008b Technical Report. Washington, DC. Kannan K, Choi J-W, Iseki N, Senthilkumar K, Kim DH, Masunaga S, Giesy JP. 2002. Concentrations of perfluorinated acids in livers of birds from Japan and Korea. Chemosphere 49:225–231. Kannan K, Newstead J, Halbrook RS, Giesy JP. 2002. Perfluorooctanesulfonate and related fluorinated hydrocarbons in mink and river otters from the United States. Environ Sci Technol 36: 2566–2571. Adinehzadeh M, Reo NV, Jarnot BM, Taylor CA, Mattie DR. 1999. Dose–response hepatotoxicity of the peroxisome proliferator, perfluorodecanoic acid and the relationship to phospholipid metabolism in rats. Toxicology 134:179–195. Chinje E, Kentish P, Jarnot B, George M, Gibson G. 1994. Induction of the CYP4A subfamily by perfluorodecanoic acid: The rat and the guinea pig as susceptible and nonsusceptible species. Toxicol Lett 71:69–75. Harris M, Birnbaum L. 1989. Developmental toxicity of perfluorodecanoic acid in C57BL/6N mice. Fund Appl Toxicol 12:442–448. Norstrom RJ, Simon M, Moisey J, Wakeford B, Weseloh DV. 2002. Geographical distribution (2000) and temporal trends (1981–2000) of brominated diphenyl ethers in Great Lakes herring gull eggs. Environ Sci Technol 36:4783–4789. Jansson B, Andersson R, Asplund L, Litzén K, Nylund K, Sellström U, Uvemo U-B, Wahlberg C, Wideqvist U, Odsjö T, Olsson M. 1993. Chlorinated and brominated persistent organic compounds in biological samples from the environment. Environ Toxicol Chem 12:1163–1174. Sellström U, Jansson B, Kierkegaard A, de Witt C. 1993. Polybrominated diphenyl ethers (PBDE) in biological samples from the Swedish environment. Chemosphere 26:1703–1718. Alaee M, Wenning RJ. 2002. The significance of brominated flame retardants in the environment: current understanding, issues, and challenges. Chemosphere 46:579–582. Clark KE, Niles LJ, Stansley W. 1998. Environmental contaminants associated with reproductive failure in bald eagle ( Haliaeetus leucocephalus) eggs in New Jersey. Bull Environ Contam Toxicol 61:247–254.