Eagles and Lead

Ingestion of lead-contaminated waterfowl carcasses by bald eagles was a primary consideration for the 1991 ban on lead shot for waterfowl hunting in the United States (USFWS, 1986). A 16-year retrospective study from 1980 to 1995 determined that average concentrations of lead in bald and golden eagle blood samples had decreased after 1991; however, the prevalence of lead poisoning in eagles did not change (Kramer and Redig, 1997). Thus, despite the 1991 ban on lead shot for hunting waterfowl, lead exposure and poisoning in eagles has been an ongoing concern (Cruz-Martinez and others, 2012; Russell and Franson, 2014).

Before and after the 1991 ban on lead shot for waterfowl hunting, several studies documented lead toxicity in bald and golden eagles. The U.S. Geological Survey National Wildlife Health Center (NWHC) diagnosed lead poisoning in 16 percent of 484 bald eagles and 5 percent of 68 golden eagles examined 1975–2013 (Russell and Franson, 2014). Examination of bald eagle carcasses from three upper midwestern States (Iowa, Minnesota, and Wisconsin) during a single winter (2011–12) revealed lethal lead concentrations (in other words, greater than [>] 6 milligrams per kilogram [mg/kg] wet weight [ww]) in 40 percent of 58 carcasses and detectable lead concentrations in another 20 percent (Warner and others, 2014). In another report, 14 percent of 127 bald eagles from New England had concentrations of lead indicative of lead poisoning (in other words, >30 mg/kg dry weight [dw]; Mierzykowski and others, 2013). A more recent analysis reports high concentrations of chronic (as much as 47 percent) and acute (as much as 35 percent) lead poisoning in both eagle species across the United States (Slabe and others, 2022), with the potential for long-term population suppression (Hanley and others, 2022; Slabe and others, 2022). Lead bullets used for big game hunting may be a primary exposure pathway for dietary lead in eagles (Stauber and others, 2010; Bedrosian and others, 2012) and the use of nonlead ammunition can reduce lead exposure in these species (Bedrosian and others, 2012); however, recreational and pest management shooting of Cynomys spp. (prairie dogs) and Sciuridae spp. (ground squirrels) with lead ammunition is common across the western United States and exposure to lead through scavenging of these carcasses also poses a potential exposure pathway (Pauli and Buskirk, 2007; Herring and others, 2016). McTee and others (2017) determined that 66 percent of Urocitellus columbianus (Columbian ground squirrels), a common prey species for golden eagles, contained lead fragments after being shot with lead bullets. The shot carcasses of Spermophilus richardsonii (Richardson’s ground squirrels), another common prey species for golden eagles, were also determined to be a source of lead exposure, meaning that raptors scavenging these carcasses may be at high risk of lead poisoning (Knopper and others, 2006).

Eagles and Mercury

Mercury is a naturally occurring element that can be highly toxic and mobile in the environment. Inorganic mercury is emitted more by anthropogenic than natural sources (Nriagu, 1989 as cited by Shore and others, 2011), largely because of coal-fired electricity generation (Streets and others, 2009; Dziok and others, 2021). Human activities have increased atmospheric concentrations of mercury by 3–5 fold during the last 150 years (Eagles-Smith and others, 2014). When inorganic mercury enters an aquatic ecosystem, it can undergo a process known as methylation resulting in the formation of methylmercury (MeHg). Methylation of mercury within the water column and sediments of surface waters is the most important biogeochemical process from an ecotoxicological perspective (Wiener and others, 2003, Scheuhammer and others, 2008). The methylated form of mercury is typically the most stable, bioavailable, and toxic form of mercury to wildlife (Shore and others, 2011). Mercury accumulated in fish and at higher trophic levels of aquatic food webs is almost entirely MeHg and is readily bioaccumulated and transferred in food webs and can biomagnify to high concentrations in predatory fish and wildlife (Wiener and others, 2003); thus, eagles may be exposed to mostly MeHg when they consume prey. The liver is where mercury is demethlylated, and the liver of fish and mammalian prey can accumulate methylmercury to harmful concentrations. Usually, MeHg is more than 50 percent of the total mercury measured in bald eagle livers (Weech and others, 2003;Mierzykowski and others, 2013); however, observations of elevated concentrations of total mercury in the liver are not sufficient for diagnosing methylmercury intoxication because of demethylation and the subsequent formation of MeHg complexes with selenium that have relatively low toxicity (Wiener and others, 2003).

Concentrations of mercury above toxicity thresholds have been reported in bald eagle blood (Carlson and others, 2012), feather (Bechard and others, 2009), and liver samples (Weech and others, 2003; Mierzykowski and others, 2013). Although tissue-based effect thresholds have not been derived for eagles, a synopsis of data available for other avian species indicates that liver concentrations >2 (experimental) and >20 (field) mg/kg ww may be indicative of adverse reproductive effects and death (Shore and others, 2011). A study that tested blood samples from 64 nestling bald eagles in Montana and Wyoming reported that mercury contamination had become more widespread from 1991 to 2007, whereas the prevalence of lead contamination during that period was unchanged (Harmata and Oakleaf, 1992; Carlson and others, 2012).

Eagles and Rodenticides

Several types of ARs have been used in the United States to control rodents including prairie dogs and ground squirrels in western States. Poisoned animals can take days to weeks to become moribund and are more susceptible to predation (Cox and Smith, 1992; C.A. Yoder, U.S. Department of Agriculture, unpub. data, 2008; these data are not currently available from the U.S. Department of Agriculture. Please contact the U.S. Department of Agriculture for further information). Large field applications of ARs have been known to attract raptors to the easy-to-capture prey (Vyas and others, 2017).

In 2012, two first generation anticoagulant rodenticides (FGARs), trade names Rozol TM (active ingredient chlorphacinone) and Kaput TM (active ingredient diphacinone), were registered to poison Cynomys ludovicianus (black-tailed prairie dogs), resulting in large scale field applications that may individually cover thousands of acres of grassland/prairie dog habitat (U.S. Environmental Protection Agency [EPA], 2016). Studies have recently demonstrated that FGARs are more toxic than the EPA’s “moderately toxic” classification (Rattner and others, 2012;Vyas, 2013) and raptors may be especially sensitive to these compounds when multiple doses are consumed on multiple days (Rattner and Mastrota, 2018). Acute diphacinone toxicity tests indicate that Falco sparverius (American kestrels) are more than 20 times more sensitive than Colinus virginianus (northern bobwhite) and more than 30 times more sensitive than Anas platyrhynchos (mallards), two test species required by the EPA for pesticide registration (Rattner and others, 2010, 2011). Little is known about golden eagle sensitivity to ARs, although golden eagles fed muscle from diphacinone-treated sheep exhibited extreme weakness, ataxia (lack of muscle control), and hemorrhages (Savarie and others, 1979).

In addition to FGAR prairie dog baits, second generation anticoagulant rodenticides (SGARs) are commonly used to poison small mammals, including Mus spp. (mice) and Rattus spp. (rats) around buildings (Kelly Registration Systems, 2019). Compared to FGARs, the SGARs are more persistent in tissues (Erickson and Urban, 2004) and are also highly toxic to animals that consume poisoned prey; however, like FGARs, it can take several days for rodents to die after they consume a lethal dose (Cox and Smith, 1992). Furthermore, unlike FGARs, residues of SGARs in tissues can persist for months, increasing the risk of multiple feedings and secondary toxicity to predators (Rattner and Mastrota, 2018). Concentrations of the SGAR brodifacoum in Mustela erminea (stoats) were greater during the three months after the removal of baits than during the poison operation (Murphy and others, 1998); thus, SGARs appear to pose the greatest risk among ARs to predatory and scavenging birds (Erickson and Urban, 2004;Rattner and Mastrota, 2018).

Accipitridae (hawks) and Strigidae (owls) have succumbed to SGAR poisoning (Hegdal and Colvin, 1988; Murray and Tseng, 2008; Albert and others 2010) as these species are known to forage on the rodents that SGARs target. Concentrations of brodifacoum greater than 100–200 micrograms per kilogram (µg/kg) in Tyto alba (barn owls) have been associated with clinical signs of rodenticide poisoning and increased mortality (Shore and others, 2002); however, liver brodifacoum concentrations associated with toxicity can differ among species, and it is unclear if these criteria would apply to other raptor species (Thomas and others, 2011; Christensen and others, 2012;Nakayama and others, 2019).

Unlike the numerous reports on lead and mercury exposure, few reports have evaluated eagle exposure to ARs in the United States. In one report assessing liver AR residues and related mortality in 116 bald and 17 golden eagles, detectable sublethal concentrations were reported in 83 percent of bald eagles and 77 percent of golden eagles assessed and AR toxicosis was determined to be the cause of death in 4 percent of eagles (Niedringhaus and others, 2021). Similarly, a review of AR poisoning of nontarget animals by Nakayama and others (2019) determined that 80 percent of 25 bald eagles tested had detectible concentrations of ARs, with 19 of those detects being SGARs. Bald eagle mortality from ingesting chlorophacinone-poisoned prairie dogs has been reported in Nebraska and South Dakota (EPA, 2016). Although less is known regarding prevalence and effects of AR exposure in golden eagles, the availability of poisoned prey may exacerbate local golden eagle population declines (Golden and Gober, 2010).

Sample Criteria and Cause of Death Determination

Necropsies were done either at NWHC or, for specimens originating in Colorado, by Colorado Parks and Wildlife using similar criteria. A subset of carcasses meeting NWHC diagnostic case submission criteria (which includes birds that had available field history described previously, had not been frozen >6 months, and had not spent >72 hours in rehabilitation) were handled as diagnostic service cases and received comprehensive postmortem examinations done by board-certified veterinary pathologists and ancillary laboratory analysis to determine the cause of death (COD). If carcasses were in good postmortem condition, frozen for >6 months, had at least county-level location and the date of field recovery known, they were sent to NWHC but handled as special study cases, receiving gross postmortem examination and tissue collection for contaminant analysis for this study. COD was attempted on these cases based on gross postmortem examination only. In a few cases, no postmortem exam was performed, and tissues were collected in the field by USFWS staff for contaminant analysis. Carcasses in this group were placed into appropriate COD categories after contaminant analysis was completed, if possible, based on field history or other information; for example, if the COD was suspected or confirmed as gunshot in the field, it was placed in the trauma category or after contaminant analysis confirmed lead poisoning (>10 mg/kg ww) it was placed in the lead poisoning category.

For this study, COD is defined as the most important factor contributing to death based on a combination of the field history, gross and (or) microscopic assessment, and ancillary test results as available. Testing of lead, mercury, and rodenticides specifically for this study were not used to make a COD determination for diagnostic cases. Results of lead, mercury, and AR testing done at Michigan State University (MSU) Veterinary Diagnostic Laboratory were considered in COD determination for special study cases and field sampled specimens, as no ancillary testing was otherwise performed. For diagnostic cases submitted to the NWHC or Colorado Parks and Wildlife where lead poisoning was suspected as the COD, liver lead testing was done at the NWHC or Colorado State University, respectively. Rodenticide testing was also performed on two diagnostic cases as part of the COD determination process. In both cases, separate samples of liver were obtained for inclusion in this study. For the purpose of this study, diagnostic, special study cases, and field sampled specimens were divided into six simplified COD categories for analysis: (1) “trauma” for eagles with COD confirmed or suspected as a result of trauma including vehicle collision, intraspecies aggression, entrapment, gunshot, blunt force, or euthaniasia or death in captivity because of severity of trauma; (2) “electrocution” for eagles with COD confirmed or suspected as a result of electrocution; (3) “disease” for eagles with COD confirmed or suspected as a result of disease or other toxin not part of this study; (4) “lead toxicity” for eagles with COD confirmed or suspected of lead toxicity based on a combination of liver lead concentrations, postmortem lesions, and field history (Franson and Pain, 2011); (5) “other” for eagles that had other causes of death such as drowning, neoplasia, or emaciation/starvation, not fitting into the other categories; and (6) “undetermined” for eagles when a COD could not be identified, such as when necropsy results did not indicate a specific COD or if carcasses were unsuitable for necropsy because of postmortem autolysis. Age for each eagle was categorized as either adult (including subadult) or juvenile based on plumage, gonad development, and presence or absence of thymus and bursa of Fabricius. Sex was determined from gonad morphology and categorized as male, female, or unknown. At necropsy, a minimum of 5 grams (g) of liver was collected, placed into whirl-pak bag, and archived frozen for AR analysis; another 5 g of liver was collected for lead and mercury analysis at MSU. The number of carcasses collected per State varied depending on availability of suitable specimens.

Lead and Mercury Analysis

Liver samples collected from carcasses examined during 2014 were analyzed at NWHC using microwave digestion and flame atomic absorption spectroscopy as part of the COD determination. Briefly, liver samples were prepared and digested using a MARS 6 microwave digestion system (CEM Corporation, Matthews, North Carolina) in accordance with manufacturer’s instructions. Once samples were digested, they were transferred to the atomic absorption spectrometer (Thermo Solaar M6, Thermo Scientific, Wilmington, Delaware) and run according to manufacturer instructions. The limit of detection (LOD) was 0.25 mg/kg ww. This same process was followed for liver samples collected for COD determination on diagnostic cases in which lead was suspected during gross examination from 2015 to 2017. This was done to provide a timely COD diagnosis to the natural resources managing agency that submitted the carcass, and the NWHC lead results were not used for analysis in this study unless no MSU results were available.

Liver samples collected from carcasses examined from 2015 to 2017 were sent to the U.S. Geological Survey Forest and Rangeland Ecosystem Science Center (FRESC) or MSU, where they were weighed wet, dried at 75 °C overnight in a drying oven, and then weighed dry (less than or equal to [≤] 0.2 g). Lead and mercury residue analysis was done on all dried samples at MSU, by inductively coupled plasma mass spectrometry (Agilent 7900 ICP–MS, Agilent Technologies, Santa Clara, California) in accordance with manufacturer’s instructions. Dry samples were then digested with 1–2 milliliter (mL) concentrated 69–70 percent nitric acid (Avantor Performance Materials, Center Valley Pennsylvania, type J.T. Baker American Chemical Society grade) reagent in 15-mL digestion vessels (Savillex, Eden Prairie, Minnesota) in a 95-°C oven overnight. The amount of acid added was relative to the weight of sample. Overnight digests were then diluted 1:100 in Millipore Filter (Burlington, Massachusetts) deionized water before analysis. Standard reference materials (SRMs) used as quality control for this process included National Institute of Standards and Technology (Gaithersburg, Maryland) SRM 1577c Bovine Liver, National Institute of Standards and Technology SRM 2976 Mussel, and an in-house maintained tissue spike control and a digest blank. Lead ww sample concentrations are calculated by multiplying the dw concentration in micrograms per gram by the dw to ww ratio for individual samples. The LOD for liver samples analyzed by MSU varies depending on sample mass but is generally less than (<) 0.5 mg/kg dw and <0.1 mg/kg ww for mercury and lead, respectively. Results for lead are reported as milligrams per kilogram ww and categorized following Franson and Pain (2011) as background (<2.0 mg/kg), subclinical poisoning (2.0 to <6.0 mg/kg), clinical poisoning (6 to 10 mg/kg), and severe clinical poisoning (>10 mg/kg). Unless otherwise noted, concentrations of mercury are reported herein as dw and ARs are reported as ww. Concentrations of mercury in liver are categorized as low (0 to <20 mg/kg dw), moderate (20 to <80 mg/kg dw), and high (>80 mg/kg dw) following Weech and others (2003).

Anticoagulant Rodenticide Analysis

For AR testing, liver tissue was analyzed at MSU using a panel that includes FGARs chlorophacinone, diphacinone, and warfarin and SGARs brodifacoum, bromadiolone, difenacoum, and difethialone. Anticoagulants were identified and quantitated in tissues by gradient high performance liquid chromatography (HPLC) with ultraviolet photodiode array (PDA) and fluorescence (FL) detection. Brodifacoum and difenacoum were screened and quantitated by FL detection in a basic mobile phase system and confirmed by PDA in an acidic mobile phase system. Chlorophacinone, difethialone, and diphacinone were screened, confirmed, and quantitated by PDA in basic mobile phase. Bromodiolone and warfarin were screened and quantitated by PDA in an acidic mobile phase system; bromodiolone was confirmed by FL detection in a basic mobile phase system, whereas warfarin was confirmed by HPLC-electrospray quadra-pole mass spectrometry. This method was also available to confirm the remaining analytes in difficult situations that may arise with interfering chromatographic peaks. Gas chromatography mass spectrometry was also available for warfarin confirmation if needed. The LODs for AR analytes are as follows: brodifacoum 0.002 mg/kg ww; bromadiolone 0.02 mg/kg ww; chlorophacinone 0.2 mg/kg ww; difethialone 0.07 mg/kg ww; diphacinone 0.2 mg/kg ww; difenacoum 0.02 mg/kg ww; and warfarin 0.02 mg/kg ww. Concentrations of ARs were reported below LOD when positive values within the instrument's limit of quantification were confirmed by the methods described previously (for example, difenacoum confirmend by mobile acidic phase HPLC with PDA). Concentrations of ARs in eagles are reported as presence/absence because no toxic threshold concentrations are available for these species; however, we noted AR concentrations in eagles that were above a potentially lethal range for barn owls (in other words, AR >100–200 µg/kg; Shore and others, 2002).

Logistic regression was used to assess associations among detection versus nondetection of lead, mercury, and brodifacoum with COD, age, and sex. All analyses were done in a Bayesian framework R (R Core Team, 2018) using the library “rjags” (Plummer, 2018). For the linear regression, values below the LOD were input in jags as 0–LOD using the “dinterval” function (Plummer, 2018). For logistic and linear models, “State” (Wyoming, Montana, South Dakota, etc.) was included as a random effect, and evidence of spatial autocorrelation was tested using Moran's I (Cliff and Ord, 1981). Small sample sizes (in other words, typically n<10) precluded significance testing for differences in COD among the States where eagles were collected.

Cause of Death

Postmortem examination to evaluate COD was attempted on 295 carcasses. A COD was suspected or determined for 269 carcasses (91 percent), including 152 (88 percent) of 172 bald eagles and 117 (82 percent) of 142 golden eagles (tables 2 and 3); COD for the remainder of the birds was undetermined. Of those carcasses where COD was determined, trauma was the leading cause of mortality in both species, accounting for 138 (51 percent) of 269 eagle deaths, including 70 (46 percent) of 152 bald eagles and 68 (58 percent) of 117 golden eagles (tables 2 and 3). Electrocution was the second leading COD, affecting 56 (21 percent) of 269 birds (24 [16 percent] and 32 [27 percent] of 152 bald and 117 golden eagles, respectively).

Lead toxicity was the third leading COD overall, with 55 (20 percent) of 269 birds affected; however, it was the second leading COD (47 [31 percent]) after trauma in 152 bald eagles (table 2). Only 8 (7 percent) of 117 golden eagles examined were determined to have died of lead poisoning (table 3). Disease and toxins (other than lead) caused the death of 10 eagles (6 bald, 4 golden). Ten eagles (5 bald, 5 golden) had other causes of mortality such as drowning, emaciation, starvation, and neoplasia. Cause of death could not be determined for 45 eagles (20 bald, 25 golden). Eagles that had a postmortem examination attempted to determine COD and at least county-concentration location are summarized in figure 1.

Lead Results

A total of 299 of 314 eagle carcasses collected had liver available to measure concentrations of lead (table 4). Overall, 75 (25 percent) of 299 eagles tested had some concentration of lead poisoning (in other words, liver lead >2 mg/kg ww; Franson and Pain, 2011). When compared by species, 60 (36 percent) of 166 bald eagles had evidence of lead poisoning compared to 15 (11 percent) of 133 golden eagles. Eagles that had liver analyzed for lead exposure and at least county-concentration location are summarized in figure 2.

Lead concentrations were indicative of severe clinical lead toxicity (>10 mg/kg ww; Franson and Pain, 2011) in 56 (19 percent) of eagles tested: 50 (30 percent) of 166 bald eagles and 6 (4.5 percent) of 133 golden eagles (table 4). The highest lead concentration was 95.57 mg/kg ww, which was detected in an adult female bald eagle collected in Nebraska. Only three eagles (1 percent) had liver lead concentrations within the range indicative of clinical lead poisoning (in other words, 6–10 mg/kg ww; Franson and Pain, 2011). Causes of death determined for these 3 birds were trauma (2) and lead toxicity (1).

Subclinical lead toxicity concentrations (2.0 to <6.0 mg/kg ww) were detected in 16 (5 percent) of 299 eagles. Causes of death assigned to eagles in this category were trauma (9), undetermined (3), other (2), electrocution (1), and lead toxicity (1). Of the remaining 224 eagles, lead was detected at background concentrations (<2.0 mg/kg) in 125 eagles, and 99 eagles had no lead detected to the LOD (<0.25 mg/kg). Of the 138 birds that died because of trauma, 130 were tested for lead, and 117 of these eagles had concentrations of lead in the liver in the background range (<2.0 mg/kg ww). For eagles with trauma as the primary COD, five bald eagles and four golden eagles had lead concentrations in the subclinical range. In the clinical range there were four bald eagles with trauma as the primary COD. Three of 56 eagles that died because of electrocution also had toxic liver lead concentrations indicating severe clinical lead toxicity (>10 mg/kg ww).

Regression analyses indicated that adult bald eagles had liver lead concentrations that were about 52 percent higher than juveniles, and male bald eagles had lead concentrations that were about 28 percent lower than females (table 5); however, differences among adults and juveniles and males and females were not statistically significant (all estimated differences overlapped zero; table 5). Lead concentrations did not differ among bald eagles diagnosed with trauma, electrocution, infectious disease, or other COD (except lead poisoning; table 5), indicating no association between these causes of death and liver lead concentrations. For golden eagles, adult lead concentrations were about 81 percent higher than juveniles (table 5). Male golden eagles also had liver lead concentrations about 18 percent lower than females; however, this difference was not statistically significant. Golden eagles dying of causes other than lead toxicity had lead concentrations approximately 92–94 percent lower than eagles dying of lead poisoning.

For bald and golden eagles combined, odds of an adult eagle dying of lead poisoning were 11 times greater than a juvenile (table 6). Odds of a male eagle dying of lead poisoning are about half that of a female (table 6). Bald eagles were nine times more likely to die of lead poisoning than golden eagles (table 6). These differences were statistically significant (in other words, odds ratios did not overlap one). Results for whether eagles had high concentrations of lead >10 mg/kg ww versus below 10 mg/kg ww were similar (table 6).

Mercury Results

Liver mercury concentrations were available for 277 eagles in the study: 159 bald eagles and 118 golden eagles (table 7). For bald eagles, three (2 percent) birds had high concentrations of mercury (in other words, >80 mg/kg dw). Eleven (7 percent) bald eagles had moderate mercury concentrations that ranged from 20.85 to 78.16 mg/kg. Low concentrations (<20 mg/kg) were determined in the remaining 145 bald eagles. No golden eagles had high liver mercury concentrations. The remaining 117 (99 percent) golden eagles had low mercury exposure (<20 mg/kg). Eagles that had liver analyzed for mercury exposure and at least county-concentration location are summarized in figure 3.

Overall, adult eagles were more likely to have detections of mercury than juveniles, and female eagles were more likely to have higher concentrations than males (table 8). Adult eagles had overall higher mercury concentrations than juvenile eagles, and bald eagles had overall higher concentrations than golden eagles (table 8).

Anticoagulant Rodenticide Results

A total of 298 eagles (165 bald eagles and 133 golden eagles) had liver samples tested for ARs. Of the two FGARs included in the panel, chlorophacinone was not detected in any eagles tested; diphacinone was detected once, in a juvenile golden eagle from Wyoming. Warfarin was not detected in liver tissues of birds tested.

Unlike FGARs, SGARs (mostly brodifacoum) were frequently detected in liver tissue of bald and golden eagles. Brodifacoum was detected in 166 (56 percent) of 298 eagles tested: 114 (70 percent) bald eagles and 52 (39 percent) golden eagles (table 9). Eagles that had liver analyzed for lead exposure and at least county-concentration location are summarized in figure 4. Concentrations of brodifacoum ranged from 0.002 to 2.64 mg/kg ww. Of the leading causes of mortality, 54 percent of trauma, 46 percent of electrocution, and 89 percent of lead toxicity cases had evidence of exposure to brodifacoum. Kansas had the highest proportion of bald eagles (15 of 16) with detectable concentrations of brodifacoum.

In bald and golden eagles, the odds of brodifacoum detections and the concentration of brodifacoum in liver were statistically higher in adult eagles versus juvenile eagles and in bald eagles versus golden eagles (table 10). Moran’s I and spatial variograms indicated no spatial patterns in brodifacoum concentrations.

Other less common SGAR detections included difenacoum (number of eagles [n]=3), difethialone (n=1), and bromadiolone (n=1). The four eagles with detectable difenacoum or bromadiolone also had detectable concentrations of brodifacoum.

In total, 9 (5 percent) of 165 bald eagles and 3 (2 percent) of 133 golden eagles were determined to have liver concentrations of brodifacoum in a potentially lethal range (>0.2 mg/kg ww) reported for barn owls (Shore and others, 2002); however, 10 of these eagles had CODs other than AR poisoning and two had undetermined CODs.