Skeletal Pathology in White Storks ( Ciconia ciconia ) Associated With Heavy Metal Contamination in Southwestern Spain

Abstract

In 1998, a mine tailings dyke in southwestern Spain broke, flooding the Agrio-Guadiamar river system with acid tailings up to the borders of one of the largest breeding colony of white storks in the western Palearctic, Dehesa de Abajo. Over the following years, a high proportion of nestlings developed leg defects, prompting this study. Ten fledglings with leg deformities from the spill area were compared with 11 normal storks of the same year class from another region far from the spill. However, metals were analyzed as a continuum rather than by site, as reference birds also contained high levels of metals. Gross pathology of the legs was supported by histopathology, which showed that bone remodeling activity was greater in the deformed storks, which also had more irregular subperiosteal bone, and tended to have higher residual islets of cartilage in their metaphyses, which, in turn were related to metal contaminant residues. Both Ca and P in bone were affected independently by metals. Deformed birds had lower serum bone alkaline phosphatase. Bone malformations, measured by leg asymmetry, was only partially explained by bone metals, indicating that a combination of factors was involved with the abnormal development in these young storks.

Introduction

In April 1998, an environmental disaster struck southwestern Spain. The tailings dyke from an iron pyrite mine in Aznalcóllar (Sevilla) broke open, flooding the valleys of the Agrio and Guadiamar rivers up to the border of the Doñana National Park, a critical breeding and wintering site for many species of threatened and endangered animals (Grimalt et al., 1999). The slurry of toxic acidic mud and water covered an area 40 km long and 0.5 km wide, was approximately 2 × 10⁶ m³ in volume and contained potentially toxic elements such as Pb, Zn, As, Cu, Cd, and other sulphide-related elements (Grimalt et al., 1999). Of particular concern was the proximity of the inundation to Dehesa de Abajo, the location of one of Western Europe’s largest breeding colony of white storks (Ciconia ciconia). The tailings flooded the river valley and marshes as close as 1 km from the colony. The chemical nature of the tailings, plus hydrological, geomorphological details, as well as effects on soil, groundwater, plants, invertebrates, fish and birds in 1998, were detailed in an issue of Science of the Total Environment (1999) devoted to examining this accident.

In the years since the spill, researchers have documented different problems in the nestling storks. Genotoxic damage has been tracked in the colony for several years (Pastor et al., 2004). Since the spill, a variable but substantial proportion of the nestlings were discovered to have deformed legs and bills that had not been documented previously in this or any other stork colony in Spain. The peak in deformities occurred in 2001, with nearly 10% of the nestlings being affected to some degree (R. Baos, unpublished). As detailed below, legs were consistently affected just distal to the tarsal joint, with thickening, local malformation, and lateral or medial deviations of the lower leg. In some nestlings, the bill, especially the maxilla, was curved dorsally or laterally.

Our objective was to conduct a comprehensive examination characterizing the nature and extent of physiological and anatomical aberrations in young storks in relation to metal contaminant loads in tissues. To this end, deformed juveniles from Dehesa de Abajo (DdA) were compared with another small group of storks of the same age collected from different breeding sites in Cáceres province, an area over 200 Km away from the spill . Pathological plus other morphological and physiological endpoints were compared between the groups of birds.

Materials and Methods

Animals

In 2001, 10 nestling storks with obviously deformed tarsometatarsal (MT) bones were collected from the DdA colony adjacent to the spill site just prior to reaching fledging age (birds ranged from 40–55 days old), and were housed in the Acebuche wildlife recuperation center (Doñana National Park, Huelva) for 3 months. Eleven storks, also young of the year and apparently healthy (except they had fallen from their nests or suffered some other misadventure) were collected from a wildlife recovery center “Sierra de Fuentes” in the neighbouring area of Extremadura. Birds were housed in groups of 5 or 6 colony mates, in 4 open pens. Pens (4w × 6d × 3h m) were enclosed with solid mesh walls, open wire fronts, and roofs with partial cover. Birds had access to small ponds within their enclosures, and were maintained on a standard diet of various species of marine fish purchased from a fish-processing factory, which produces fish for human consumption. Any fish that were too small or otherwise unsuitable for the human food market, made up the diet of the storks. Their diet was not analyzed for metals, but was assumed to meet human food quality criteria. Once the birds arrived at the wildlife recovery centre, their legs were radiographed. The storks remained in the rehabilitation centre for 3 months before it was logistically possible to conduct the necropsy examinations. All work was conducted under the proper Animal Care Protocols.

Study areas

The colony referred to as DdA, was located in a natural area (Puebla del Río, province of Sevilla, 6°10′16″ W: 37°12′33″N), while the Cáceres birds came from a variety of habitats in the Extremadura region of southwestern Spain. The linear distance between the origins of these 2 groups of birds was greater than 200 km. The DdA colony was far from urban environments, close to a permanent water source (within 2 km of a marsh in Doñana National Park which is a very productive ecosystem), with nests in wild olive trees. Details of the origins of the birds that came from the Cáceres rehabilitation center were not available.

Pathology

Storks were humanely euthanized by overdosing with Halothane inhalation anaesthetic. Their heads were placed into a plastic bag containing halothane soaked cotton swabs, until respiratory and cardiac arrest within 30 to 45 seconds. Body mass was determined immediately. A complete necropsy examination was conducted. Because of the leg deformities in DdA birds, length of both right and left MT bones (i.e., area of leg spanning the tarsometatarsal bone) were measured to the nearest mm and the difference between the 2 was used to describe the degree of leg asymmetry. All joints, musculoskeletal system and viscera were examined, and the mass of the liver, bursa of Fabricius, and spleen were recorded. Organ mass was expressed as “somatic indices” (SI) (e.g., (liver SI = (liver mass/(body mass − liver mass)) × 100). Samples of liver and kidney were wrapped in xylene rinsed foil and frozen in sample bags at −20°C. Tarsometatarsal bones were immersed in 10% neutral buffered formalin.

From the proximal epiphysis, extending 2 to 3 cm distally along the shaft, to encompass the area of the deformations, the MT bones were sectioned midsagitally. The bones were then placed in decalcification solution (20% formic acid in water) until they were demineralized sufficiently for proper histological preparation. To avoid information bias, the evaluation of the bone sections was carried out by the pathologist unaware of the animals’ identification or origin. The H&E-stained bone sections were examined microscopically for evidence of pathology. After consultation with a colleague/osteopathologist, the bones were assessed according to the following features. The ostealperiosteal surface was described as smooth or irregular (Figure 1a and b). After examination of 6 high power fields (HPF: 200 × magnification), the occurrence of irregularly shaped islets of cartilage in the diaphyseal or cortical bone around the growth plate was scored as “mild” if there were 0–1 islets per HPF, moderate if there were 2–4 per HPF, and marked, if there were greater than or equal to 5 per HPF. The remodeling activity in the bone was determined based upon the presence of an open, haphazard pattern of newly formed bone with remaining collagen and occasional round osteoblasts visible (Figure 2a). Remodeling was most easily determined using a polarizing filter to distinguish it from the well organized, lamellar, parallel pattern of older bone (Figure 2b). It was categorized as mild if < 10% of each field showed remodeling, moderate if 10–50% showed newly forming bone, and marked if > 50% showed remodeling activity. When remodeling activity changed along the bone section, the most commonly noted pattern was recorded for that case.

Clinical pathology

Prior to euthanasia, a blood sample was taken from each bird. Because of the important role of Ca and P in bone development, levels of both minerals were assessed in serum using an automated instrument (Hitachi 912 automatic analyzer, Montreal, Canada). In veterinary medicine, serum levels reported as total alkaline phosphatase (AP) are the sum of AP isoenzymes from the liver, intestine, and bone. Because of the specific role of bone AP, which is produced at the highest rate by osteoblasts during bone growth and development (Hoffman and Solter, 1999), we measured the total, hepatic and bone AP, to determine if there was a detectable change in bone AP activity associated with the deformed long bones. Serum levels of bone AP were determined using an adaptation of the mammalian assay by Hoffman et al. (1994).

Mineral and metal analysis

Liver and kidney samples from these storks were placed into a drying oven in covered porcelain dishes for several days until all moisture was gone from the tissues. A 2 cm slice of each MT bone was sectioned 3 cm distal to the proximal epiphysis, individually bagged and submitted to the analytical laboratory at the Department of Geological Sciences, University of Saskatchewan. The dried liver and kidney samples were then ground and homogenized using a corundum grinder. Aliquots of the powdered tissues (0.1 g) were placed into Teflon vials with 2 ml double distilled, ultrapure HNO₃ and a few drops of H₂O₂. Vials were left open in a clean fume hood until the reaction subsided. Vials were then capped and placed on a hot plate at 100°C until samples were completely digested and a clear solution was formed.

The vial was then opened so the liquid could evaporate at 70°C until it reached ambient dryness, 1 ml HNO₃ was added to the residue and the vials capped for 1 hour until all residues were dissolved, the solution was diluted with deionized water to 100 ml, and stored at 4°C until analysis. For the bone samples, the sections were crushed with an agate pestle, ground and homogenized with corundum grinder, 0.1 g of the powder was used for digestion following the procedure described above. Analysis for metals was performed using a Platform collision cell inductively coupled plasma mass spectrometer (ICP-MS, Micromass, Manchester, UK). A low flow Meinhard concentric nebulizer (Glass Expansion, Hawthorne, Australia) was employed. A water-cooled (4°C) Scott type double-pass spray chamber was also used to generate homogeneous sample aerosols. Calibration was achieved using a series of multi-element standards (1, 5, 10, 20 ppb), and a calibration curve was established. Rhodium was used as an internal standard to correct for sample matrix effects and instrument drift. Three standard reference materials were used for the data quality control; a river water standard SLRS-4, a dogfish liver standard DOLT-2 and a dog-fish muscle standard DORM-2 (National Research Council of Canada, Ottawa, Canada).

Statistical analyses

Although we present the metal residues in liver and kidney as well as bone to allow comparison with findings by other researchers, we limit all statistical analyses to metal levels in the bones. The rationale for the latter is that all the birds had been on the same standard fish diet at the rehabilitation centre for 3 months, allowing differences that might have originally existed in the liver and kidney to disappear. Bone metal concentrations were assumed to be unchanged from the time the fledglings were taken from their original colonies. Bone elements are quite stable in full-sized animals compared with rapidly growing nestlings (Scheuhammer, 1987 and references therein).

The full biological impact of metals was unlikely to be revealed by repeated testing of individual elements, as there are numerous interactions and correlations among the metals themselves. Therefore, we used principal component analysis (PCA) to derive fewer and independent variables. In 1 analysis we took this approach for 9 elements in the bone that we interpret as contaminants (Pb, Co, Cr, Ti, Zn, Sn, V, Ba, Sr). For a second analysis we used 5 normal, physiologically important elements in the bone (Ca, P, Mg, Na, S). For these PCAs, we chose elements that were consistently detectable and that were, or could be transformed to be, normally distributed. The PCA of contaminants resulted in three components accounting for 63.7% of the total variability: PC1 explained 29.7%, and PC2 explained 21.7%, and PC3 explained 12.3% of the variance in the data set. Examination of the relative loading of the variables on the 3 components, and subsequent correlation analyses, yielded the following interpretation. PC1 was largely an axis of V, Ba, Sr, and Cr all varying positively. PC2 was Zn and Ti (positively), and Co, Sn, and Cr (negatively). PC3 was largely influenced by a positive relationship with Pb and Co. For the PCA of the normal bone elements 2 components explained 91.0% of the variance: (PC1 70.6%, PC2 20.4%). Ca, P, Mg, and Na were all highly, and positively associated with PC1, whereas PC2 was primarily influenced by S.

We used ANOVA where nonsignificant interactions and variables were removed iteratively, to obtain the most parsimonious model that explained variation in the dependent variable. Two-tailed tests were performed using SPSS (Norušis, 1993), and we considered results significant at the 0.05 level.

For the analysis of leg asymmetry, it was not valid to simply compare birds on the basis of location since all deformed birds came from the same location. However, it was clear that individuals varied in the degree of leg deformity and so we derived the variable, absolute leg asymmetry, i.e., difference in length between the right and left metatarsi. In the ANOVAs we thus used this variable, and location as a factor.

Results

Metal contamination by location

Table 1 shows clearly that neither birds from the reference location nor from the spill site were free of metal contamination. In fact, for some of the metals, concentrations were higher in the “reference” birds. For the PCA analysis of bone metals, location was not significant for PC1 (F_1,19 = 1.672, p = 0.211), marginally nonsignificant for PC2 (F_1,19 = 3.837, p = 0.065) and significant for PC3 (F_1,19 = 5.436, p = 0.031) in which the difference was driven by the higher Pb in bones of the Cáceres birds.

Gross pathology

Radiographic evidence of leg deformities are shown in Figure 3. The deformed, healed bones were locally thicker with a triangular, rather than round, cross section. In all affected birds, the cortical bone just distal to the tibiotarsal joint was abnormally thickened with callus formation and varying severity of valgus deviation of the distal portion of the limbs. There was no radiographic evidence of decreased cortical thickness or density, or of any generalized increase in lucency of the cortices over the length of the bones. No fracture lines were visible.

Histopathology

A Chi-square test revealed remodeling bone score to be dependent on location, as birds in Cáceres had minimal to moderate bone remodeling whereas those from DdA had moderate to active remodeling activity (Pearson Chi-Square Test, p = 0.015). The nature of the surface of the osteum/periosteum was similarly dependent on location (Fisher’s Exact Test, p = 0.001). There was no statistically significant effect of location on scores for the cartilage islets in the bone, although this may be due to the small sample size, as the trend was for 70% of birds in Cáceres, vs only 33% in DdA, to have low amounts of residual cartilage in their metaphyses. When contaminant levels, as described using PCA, were examined in relation to histological scores, there was no difference among scores for bone remodeling or surface of the osteum/periosteum; however, PC2 values increased as the occurrence of residual islets of cartilage increased (F_2,16 = 5.412, p = 0.016) (Figure 4).

Leg asymmetry, associated pathology, and clinical biochemistry

Legs of DdA birds were more asymmetrical than the reference birds (F_1,17 = 10.123, p < 0.005). Leg deformities, as measured on a continuous scale of leg asymmetry, were associated with a number of other physiologically relevant variables. When the histological appearance of the osteal/periosteal surface was smooth, leg asymmetry was significantly less than if the birds had irregular bone surfaces (F_1,17 = 6.819, p = 0.018) (Figure 5). Considering the bone remodeling activity, the leg asymmetry increased significantly with increasing evidence of remodeling (F_2,14 = 4.981, p = 0.023) (Figure 6). The presence and density of cartilaginous islets in the proximal tarsometarsal sections of bone was independent of the symmetry of the leg bones (F_1,15 = 0.787, p = 0.473).

In the ANOVAs, body mass and sex were never found to be significantly related to leg asymmetry and so were always dropped from models. Using asymmetry as the dependent variable, both effect of location and levels of bone alkaline phosphatase (BAP) levels in plasma (F_1,17 = 4.483, p < 0.05) were significantly different between DdA and Caceres birds. For birds in both locales, increasingly asymmetrical legs had decreasing amounts of the bone enzyme BAP (Figure 7). We could not detect an association between asymmetry and either PC1 (p > 0.13) or PC2 (p > 0.71) of the normal bone minerals. Of the organs measured, only liver showed any relationship with limb pathology. As liver somatic index (liver size relative to body mass) increased, leg asymmetry also increased (F_1,17 = 8.805, p = 0.009), while again location explained considerable variance (F_1,17 = 14.088, p = 0.002).

The degree of leg asymmetry was also examined with respect to the 3 PCAs based on metal contaminants. Although location was considered as well, it was dropped from the final model, and PC3 remained significant (F_1,18 = 10.385, p = 0.005). However, the relationship was unexpected, as bones with low PC3 scores (i.e., driven by low Pb levels) were most asymmetrical.

Body and organ mass

Independent ANOVAs were conducted for the dependent variables of liver, bursa and spleen somatic indices, as well as body mass. The initial models used sex, location, and the 3 PCA contaminant-independent variables. For body mass, sex (F_1,17 = 11.717, p = 0.003) and PC3(F_1,17 = 5.406, p = 0.033) were significant, where males were heavier (as is known for this species), as were birds with higher PC3 levels. For the liver somatic index, PC3 was again a significant covariate (F_1,18 = 5.102, p = 0.037). Neither the splenic nor bursal indices were significantly associated with the PCA for metals.

Normal bone elements

In the multivariate analysis of normal bone constituents, we considered length of right metatarsus as a potential covariate. There was no significant effect of the three PCA contaminant variables on PC1 of the normal bone minerals. However, for PC2 of bone minerals, there were significant effects of metatarsal length (F_1,16 = 8.752, p = 0.009), and PC1 of metal contaminants (F_1,16 = 5.099, p = 0.038). Because Ca and P are the 2 major elements comprising bone, they were examined individually in relation to the PCA of contaminant metals. In ANOVAs we used bone Ca (and subsequently P) as the dependent variable, and considered sex and location as factors. We considered body mass, P (or subsequently Ca) in bone and the 3 components of the PCA of contaminants as potential covariates. In the final significant model, Ca was explained by body mass (F_1,16 = 9.748, p = 0.007), P in bone (F_1,16 = 369.785, p < 0.001), and PC1 metals (F_1,16 = 6.188, p = 0.024). Similar analysis of bone P levels showed significant effects of body mass (F_1,15 = 11.701, p = 0.004, Ca in bone (F_1,15 = 441.029, p < 0.001), location (F_1,15 = 6.347, p = 0.024) and PC2 metals (F_1,15 = 4.722, p = 0.046). In addition, PC1 proved, predictably, to be significantly correlated with Ca:P ratio (r_s = 0.644, p = 0.002), but PC2 and PC3 were nonsignificant.

Discussion

In this study we examined physiological and anatomical aberrations in juvenile storks in relation to metal residues in bones. Fledglings were collected from the DdA colony adjacent to the inundated area because of having deformed tarsometarsal bones. Through pathology, biochemistry, and tissue residue analyses, we describe the toxicology and explore the nature of the physiological disturbances that have resulted in bone problems.

Leg asymmetry was an important feature related to bone contaminant levels. The leg deformities were most likely the result of fractures which occurred at an early stage during the nestling period, and, because nestlings do not bear weight on their legs until about 45 to 50 days of age, the fractures could heal through apositional bone production. The consistent location of the leg pathology at the proximal end of the tarsometatarsus indicates that physical stress occurs at this site. Radiographs taken after fledging when the birds would have reached maximum skeletal size, did not show deficient cortical bone compared with the legs of the reference birds, although the bones would have gone through substantial changes during the nestling period, readily masking or correcting early defects in mineralization. Metals, once bound in bone matrix, are considered non-mobilizable (Klaassen, 2001). The metals were not different between the birds from DdA and Cáceres, so it is possible that the rate or timing of exposure to those metals would have played an important role in expression of toxicity in young, quickly developing birds. Maybe very young storks are fed on more contaminated, smaller, filter-feeders so they get higher doses of mud-associated contaminants early in life. Field observations indicate that one important prey item for DdA nestlings is the red swamp crayfish Procambarus clarkii (Negro et al., 2000), which can accumulate metals in their tissues. Crayfish captured in the spill area in 2000 had metal levels exceeding that allowed for human consumption (Sánchez-López et al., 2003).

The limb deformities were originally suspected to be due to toxicity by metals associated with the spill, that could directly disturb bone formation. Cadmium and Pb as well as Al are known to interfere with bone turnover and Ca metabolism (Goyer et al., 1994; Berglund et al., 2000). As it turned out, the levels of toxic metals were often lower in the affected birds than those from the reference site distant from the tailings accident. The Pb in tissues from the DdA storks was confirmed as coming from the Aznalcollar mine spill through isotopic identification (Meharg et al., 2002). Of the metals used for PCA analyses, only Pb, Co, and Zn were identified in the mud slurry. The remaining metals included in the analyses (V, Ba, Sr, Cr, Ti) were identified as coming from the natural river sediments unaffected by the spill (Alastuey et al., 1999).

By using multivariate analyses, relationships between toxic metal concentrations and essential bone components became evident. Using a PCA allowed us to recognize physiological costs associated with exposure to a complex mixture of metals in an acid matrix. Sulfur levels in the bones had a significant relationship with contaminant metals possibly implying that this essential mineral was being compromised in birds with higher metal contamination. Calcium and optimal Ca:P ratios, which are critical for normal bone development and remodeling, were linked to the combination of toxic metals in the bone. Because blood Ca is tightly regulated in the body through hormonal and renal controls, and the Ca:P ratio was abnormal, it stands to reason that the P, which was associated with metal concentrations, and which was different between the 2 populations, was not being properly regulated in the DdA birds.

Bone alkaline phosphatase is normally produced during active growth and remodeling in bone. It should be present in serum at relatively higher levels in animals that are growing quickly (Hoffmann and Salter, 1999). In other studies from a larger group of nestling storks, those which had slightly deformed limbs had significantly higher BAP than either normal or severely deformed chicks (Smits, unpublished), perhaps indicating that they were mounting a successful attempt to remodel the damaged bones. However, in the current study with full sized birds, decreased BAP levels were associated with more severe leg asymmetry, although histologically those limbs appeared to be at an active stage of remodelling. The lower BAP activity indicated defective or static development in bone that had never properly matured. According to recent studies in passerines, increased serum BAP is seen during the final maturation and mineralization of bone (Tilgar et al., 2004).

Increased liver mass is a nonspecific response associated with induction of enzymes involved in biotransformation of environmental contaminants to enhance their polarity and facilitate their excretion (Klaassen, 2001). The relatively larger livers seen in the storks with more severely affected legs, although a nonspecific response, is consistent with an increased detoxification effort by animals dealing with contaminant exposure. Because many metals were higher in the physically normal birds from the reference area, other sources and types of xenobiotics must be acknowledged as potential contributors to the hepatic differences in these birds. There is no indication that other factors such as trauma, or other physical or genetic factors may have been responsible for the observed limb defects. In small mammals (Mus spretus) collected in the spill area, biochemical responses indicate the simultaneous occurrence of organic pollutants as well as the contaminant metals (Bonilla-Valverde et al., 2004).

In DdA, storks generally forage in nearby marshes and rice paddies, feeding their nestlings crayfish, fish, insects, and other invertebrates (Negro et al., 2000). Both the DdA population (Jovani and Tella, unpublished) and other populations of storks are known to forage in garbage dumps (Tortosa and Caballero, 2002), which likely explains the chicken offal that has been identified in the regurgitation of nestling storks along with the more normal dietary items.

The toxic mud resulting from the breached tailings dyke was removed within several months (Hudson-Edwards et al., 2003). Besides the residual mine contaminants, the DdA stork population is also close (within 3 km) to a vast delta area of rice production, which, during the breeding season of the storks, receives frequent applications of agrochemicals to support crop production and control pests. It was not possible to obtain reliable information about the compounds being used. However, from speaking with local agronomists, or by finding recently emptied pesticide containers, it was evident that both carbamate (carbaryl) and organophosphate compounds (malathion, fenitrothion, and triclorfon), along with several types of fungicides were being used in the delta region. Cholinesterase-inhibiting compounds undergo extensive biotransformation in all forms of life, with very species specific routes and rates of metabolism (Ecobichon, 2001). Organophosphorous (OP) esters stimulate the cytochrome P450 isoenzyme system. The normal role of P450 enzymes is to enhance eventual excretion of metabolic breakdown products. One theory that could be explored is that in cases of low grade, subchronic exposure to OPs, other phosphoric acids may be metabolized and excreted at an increased rate because of the overall upregulation of this enzyme system that is relatively indiscriminant, thus causing a disturbance of normal P regulation.

Lower BAP was seen in the deformed birds, which was indicative of retarded or arrested bone maturation (Figure 2a, b). PC3 associated most with Pb levels, was making a negative contribution to bone maturation also. The metals appear to be at least in part responsible for increasing hepatic detoxification efforts (and thus higher liver SI) assuming that metalothionine synthesis (the major protein produced to protect the body against toxic metals) is one contributing factor, along with induced biotransformation activity within the liver, plus hepatic storage of metals. One could speculate on a potential role of Se, which was higher in DdA birds, that could offer some protection against the metal residues. Selenium has the capacity to form stable associations with some metals, blocking them from forming ligands with the sulfhydryl groups of essential enzyme systems (Sugiura et al., 1976).

We have shown with this study, that the relationship between bone malformations and metal contamination occurs in a complex manner which cannot be explained by one or two metals. The juvenile storks with leg deformities were affected in part by toxic elements, but the metals alone do not explain the pathology that developed. Physiological costs of exposure to toxic metals have been exacerbated by unknown factors that may be related to the intense agricultural activities occurring in proximity to the affected colony. Rice production is one conspicuous source of xenobiotic inputs that is not occurring in the reference area. We know that the DdA population had higher genotoxic damage than did storks sampled from distant locations (Pastor et al., 2004).

Although we could not prove a direct link with any particular metal, something about the contaminant burdens was affecting bone development. Clearly, a combination of factors is involved with the abnormal development seen in these young storks that requires further investigation.