1Non-directional asymmetries (fluctuating asymmetry, FA) from bilateral symmetry in morphological traits have been used as an indicator of environmental stress and may become an important diagnostic tool in environmental sciences, such as ecotoxicology.
2In this study the relationship was examined between wing feather asymmetry, measured as the difference between the length of the third primary on the left and right wing, and blood concentration of selected persistent organochlorines (OCs) in Arctic breeding Glaucous Gulls (Larus hyperboreus, Gunnerus).
3There was a positive relationship between primary asymmetry and blood concentrations of two PCB (polychlorinated biphenyl) congeners (P < 0·05), oxychlordane (P < 0·05), DDE (p′p′-dichlorodiphenyldichloroethylene) (P < 0·05), and especially HCB (hexachlorbenzene) (P < 0·001). At HCB levels above 30 ng g−1 (wet mass) there was a 60% probability that the birds had asymmetric wing feathers.
4This study indicated that the present levels of organochlorines in the European Arctic are stressors for Glaucous Gulls, and that asymmetry measurements of wing feathers may be used as an indicator of both exposure and effects of such contaminants. FA may also be a promising measurement for monitoring the early effects of organochlorine pollution on bird populations.
Fluctuating asymmetry (FA) refers to non-directional deviations from bilateral symmetry in morphological traits. Because both sides of a bilateral trait are produced from the same genome, the degree of FA is supposed to indicate an individual's developmental stability in the face of environmental and genetic stress, e.g. nutritional deprivation, changes in thermal and light conditions, parasitic infections and mutations (Van Valen 1962; Palmer 1994; Møller & Swaddle 1997; Bjorksten, Fowler & Pomiankowski 2000). It has therefore been suggested that FA is a good early indicator of environmental quality (Clarke 1995). FA is being established as a useful diagnostic tool for environmental sciences, such as ecotoxicology. Studies in various animal groups such as fish, insects, birds and mammals have demonstrated increased FA in various traits, owing to pollution exposure (e.g. Valentine, Soulé & Samallow 1973; Panakakoski, Koivisto & Hyarinen 1992; Graham, Emlen & Freeman 1993; Eeva et al. 2000). However, most studies have compared mean FA at the population level. Yet, populations differ in features other than specific environmental stress, and a more powerful way to examine effects of environmental stress would be to compare FA among individuals from the same population, exposed to different levels of stress.
Among birds, a major stress factor thought to influence both populations and individuals is persistent organochlorine pollutants. Among the adverse effects of such compounds are impaired embryonic development, immune suppression and hormonal dysfunction (Colborn, vom Saal & Soto 1993; Hoffman, Rice & Kubiak 1996). Some organochlorines (OCs), e.g. polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT), are global contaminants that biomagnify in food chains. Hence, top predators often carry high burdens of such compounds. For example, in the Glaucous Gull (Larus hyperboreus, Gunnerus), which is a top predator in Arctic ecosystems, high concentrations of PCBs and other OCs have been found repeatedly (Bourne & Bogan 1972; Cleeman et al. 2000). At Bear Island in the European Arctic, disrupted behaviours and increased mortality among Glaucous Gulls have been linked to high levels of OCs (Bourne & Bogan 1972; Gabrielsen et al. 1995; Bustnes et al. 2001a).
Many studies of FA have focused on morphological traits that individuals grow once, such as bones or fins. In birds, however, feathers are moulted and grown annually and it has been shown that various stresses may lead to asymmetry in wing feathers (Swaddle & Witter 1994; Witter & Lee 1995; Carrascal et al. 1998). Wing asymmetry increases the energetic costs of flight (Thomas 1993), and in passerines it has been shown that asymmetry in primaries increases with proximity to heavy metal pollution sources (Eeva et al. 2000). However, so far no study has tried to link the concentrations of organochlorines in individual birds to FA in wings. In this study we examined whether wing feather asymmetry (third primary) could be linked to circulating levels of various organochlorine compounds in individual Glaucous Gulls.
Materials and Methods
The study was conducted at Bear Island (74°30′ N, 19°01′ E) where about 2000 pairs of Glaucous Gulls breed (see Bustnes et al. 2000 for details about the study area).
In late May and early June 2000, breeding Glaucous Gulls were caught on their nests during incubation, using a nest trap (Bustnes et al. 2001a), and blood was sampled from the wing vein (c. 10 ml) with a syringe. Bill length, bill height (±0·1 mm), skull length (head + bill), tarsus (±0·5 mm) and wing length (±1 mm) were measured, and a principal component analysis (PCA) was performed on these measurements. The first principal component (PC1) was used as a composite measurement of body size. PC1 explained 84% of the variation in the PCA model. The birds were weighted to the nearest 10 g and body mass was used as an index of body condition, controlling for body size (PC1) in the statistical models (García-Berthou 2001). Sex was determined by size, males being larger than females (see Bustnes et al. 2001a for details). Since changes in body condition of Glaucous Gulls may influence the blood concentration of various OCs (Bustnes et al. 2001b), the relationship between body condition and blood concentration for all measured contaminants was examined, by multiple regression analysis, where OCs were the dependent variables and body mass, body size (PC1) and sex, were the independent variables. No significant relationships were found (all R2 values < 0·15 and all P values > 0·30).
Since the data were collected in one season, the present circulating blood concentrations of OCs were related to FA of primaries grown in the preceding year. However, blood levels of OCs in individual Glaucous Gulls during breeding are very stable in our study population, also between years, and 70% of the variation in persistent OCs could be predicted from the level in the preceding year (Bustnes et al. 2001b). In our study population, individual Glaucous Gulls are thus exposed to similar OC levels during feather growth in several years.
The deviation between the web lengths (the distance between the start of the growth zone of the web and the feather tip) of the third primary on each wing was used as a measure of FA. The web length was used because distinct colour patterns (white vs. grey) could be identified at the lower base of the primary shaft, which was suitable for measuring with a ruler (nearest mm). To test the repeatability of the primary measurements feathers were measured twice on a separate sample of birds (N = 30). The deviation was regressed between right and left primary during the second measurement and the same deviation during the first measurement. The repeatability was high (R2 = 0·94, F1,29 = 402·3, P < 0·0001). The distributions of right minus left (R −L) values did not differ significantly from normality with a mean of zero (Kolmogorov–Smirnov, D = 0·10), showing that asymmetries were not directional. Since the range in asymmetry was only between 0 and 4 mm, and the steps were discrete (ordinal), the data were analysed with logistic regression (SAS 1990) with binary response variables (1 and 0). Birds with a deviation between the primaries of 2 mm or more were considered asymmetric while birds with deviations of 0 or 1 mm were considered symmetric. The division at 1 mm was made to avoid small measurement inaccuracies affecting the result. Logistic regression is a less sensitive statistical procedure than regression on continuous variables because only part of the available information is used, and stronger associations are needed to demonstrate statistically significant relationships. The best statistical models were selected by backward selection, where FA was the dependent variable (0 and 1) and OC concentration, sex, body mass, body size (PC1), laying date and the interaction OC × sex, were independent variables. We started with a full model, removing independent variables one at a time if they were not significant (P > 0·05). Body condition and egg-laying date were controlled for, because these variables often reflect nutrition deprivation and other stresses that may confound the effect of organochlorines. The organochlorine concentrations were log10-transformed to meet the assumptions of normal distribution. Concentrations of the different contaminants were obtained from the same blood samples. Following Rothman (1990), we did not adjust for multiple comparisons when testing the predictions about the relationship between OCs and FA, and our results are reported without Bonferroni or similar adjustments. However, P values from two-tailed tests are reported.
The organochlorine analyses were performed at the Environmental Toxicology Laboratory at the Norwegian School of Veterinary Science, Oslo. The laboratory is accredited as a testing laboratory for OCs according to the requirements of NS-EN 45001 (1989) and ISO/IEC Guide 25 (1990). The whole blood samples (~8 g) were weighed and internal standards (PCB-29, 112, 207) were added. The samples were extracted twice with cyclohexane and acetone and the percentage extractable fat was determined gravimetrically. The cyclohexane fraction was washed with ultra-pure sulphuric acid 96%, according to Brevik (1978).
Aliquots of the lipid-free organochlorine extracts were injected automatically on a high-resolution gas chromatograph (Agilent 6890 Series GC system), equipped with an ECD detector (ECD-63N), 300°. Separation was performed on a capillary column: SPB-5, 60 m, 0·25 mm ID, and 0·25 µm film layer (Supelco Inc., Belleponte, PA). Details of extraction, clean-up and chromatographic separation are described by Bernhoft et al. (1997).
The following PCB congeners were determined (IUPAC nos, Ballschmiter & Zell 1980): 99, 118, 138, 153, 170 and 180. Other compounds analysed included: HCB (hexachlorbenzene), DDE (p′p′-dichlorodiphenyldichloroethylene) and oxychlordane. The reasons for using these compounds were their high persistence and high levels in top predators, such as the Glaucous Gull (Bustnes et al. 2001a).
Detection limits for individual PCB congeners were determined as three times the noise level and were between 0·004 and 0·026 ng g−1 wet mass. All calculations were done within the linear range of the detector's five-level calibration curve. The reproducibility was tested continuously by analysing the PCB levels in the laboratory's own reference sample (seal blubber), which was within the mean coefficient of variance for year 2000 (8·7%). The repeatability of the gas chromatography performance was tested by repeated injection of standard components at regular intervals.
The quantification was performed using PCB-29 and PCB-207 as internal standards in each sample. Percentage recoveries and coefficient of variance (CV) of OCs in spiked sheep blood varied from 83 to 119% and 3·4 to 18·1, respectively (n = 8). Blank samples were included in each series to test for interference. The present analysis precision, linearity and sensitivity were within the laboratory's accredited requirements.
The concentration of OCs in blood wet mass were used as a measure of OC burden, since wet mass is usually considered most relevant for potential toxic effects (for discussions see Klaasen & Eaton 1991; Bignert et al. 1993; Henriksen, Gabrielsen & Skaare 1996). Other studies suggest a relatively high short- and long-term stability of the blood levels of OCs in incubating Glaucous Gulls under stable conditions such as at Bear Island, indicating that blood concentration is a reliable, relative (compared to other individuals) measurement for body burden of an individual (Henriksen et al. 1998; Bustnes et al. 2001b).
The dominant OCs were PCB-153, PCB-138 and DDE while other contaminants were found in lower concentrations (Table 1). The mean web length of the third primary was 270 mm (range 257–287) in females and 281 mm (range 263–295) in males (Table 1). In both females (N = 44) and males (N = 40), 20% of the birds were classified as asymmetric (deviation of 2 mm or more between the two primaries).
Table 1. Concentrations of selected organochlorines in the blood (ng g−1, wet mass), and length (mm) of third primaries in Glaucous Gull. Data from Bear Island, 2000
Sex (P values between 0·35 and 0·75), the interaction sex × OCs (P values between 0·13 and 0·73), body mass (P values between 0·63 and 0·84), body size (P values between 0·56 and 0·70), and egg-laying date (P values between 0·63 and 0·94) did not have any significant effect on the probability of a bird showing FA in any of the statistical models with different OCs as dependent variables. These variables were therefore removed from further analyses by backward selection. The relationship between FA and OC concentration was positively significant (P < 0·05) for PCB-99, PCB-118, oxychlordane and DDE, while it was highly significant (P < 0·001) for HCB (Table 2, Fig. 1). For example, when HCB levels increased from 3 to 30 ng g−1, the probability that the birds had asymmetric wings increased from nearly 0 to 60% (Fig. 1). For the other compounds the relationship between FA and OC concentration was close to significant (P < 0·1) (Table 2). This suggests that organochlorines in general, and HCB in particular, are stressors for this population of Glaucous Gulls, influencing their ability to grow symmetrical wings.
Table 2. Association between various organochlorine pollutants (OCs) and fluctuating asymmetry in wings, measured as the difference in length of the third primary on each wing (likelihood ratio test, SAS 1990). Data on Glaucous Gulls from Bear Island in the Barents Sea, 2000
In birds the optimal phenotype is perfect wing symmetry and there is a high aerodynamic cost of asymmetry (Thomas 1993). This study indicates that organochlorine pollutants are stressors in Arctic breeding Glaucous Gulls, and that FA may be used as a method to distinguish the effects of various contaminants. Furthermore, our results suggest that there is a negative feedback loop where birds with high concentrations of organochlorines also get the highest flight costs, potentially increasing their energy costs during reproduction. In our study population it has previously been observed that birds with high blood concentration of OCs spend more time on feeding trips than birds with low levels (Bustnes et al. 2001a). Increased flight costs may be an important factor in explaining this pattern.
There are several studies of passerines that have shown asymmetry in primaries related to nutrient stress, social rank and also pollution (Swaddle & Witter 1994; Witter & Lee 1995; Carrascal et al. 1998; Eeva et al. 2000). However, this is to our knowledge the first study to demonstrate an association between blood concentration of organochlorine contaminants and asymmetry between wings. While most studies have compared population means, the novelty of this study was that we were able to link FA to OCs in individual birds.
How organochlorines influence feather growth leading to asymmetries between wings is not known. The endocrinology behind moult and feather growth is poorly understood, but thyroid hormones seem central in regulating the feathers shedding and replacement. For example birds that have been thyroidectomized will grow highly abnormal feathers (reviewed by Hahn et al. 1992). HCB and various PCB congeners are known to reduce the circulating levels of thyroid hormones (Murk et al. 1994; Hoffman et al. 1996; Gould, Cooper & Scanes 1997; Alvarez et al. 2000), and this may be a link to the mechanism of distorted feather growth in birds with high circulating OC concentrations.
This study may suggest that FA is a potential way to assess the strength of the effects of various compounds. In field studies, relating concentrations of various persistent OCs to effects is often very difficult because the compounds are highly correlated (Jones & Voogt 1999; Bustnes et al. 2001b). In Glaucous Gulls it has proved difficult to disentangle the effects of specific contaminants when relating OC concentrations to fitness traits and behaviour (Bourne & Bogan 1972; Gabrielsen et al. 1995; Bustnes et al. 2001a). In this study, different OCs varied in the strength of the association with FA. HCB had by far the strongest effect on FA although it is found in relatively low concentrations in Glaucous Gulls, making up only about 3·5% of the measured OCs (Table 1). In comparison, the PCB congeners made up about 75% (Bustnes et al. 2001a). This study may indicate that HCB, a fungicide and by-product from industrial processes, which is widely distributed in the environment (Burton & Bennett 1987; McKone, Daniels & Goldman 1996; Alvarez et al. 2000), is the compound encountered by Arctic breeding Glaucous Gulls that yields the most stress. We cannot, however, exclude the possibility that other contaminants than those measured here, e.g. heavy metals, brominated flame retardants and others, could influence the probability of developing asymmetric wings. However, the levels of such compounds are very low in this area compared to organochlorines such as PCBs, DDE and HCB (Savinova, Gabrielsen & Falk-Petersen 1995; Burkow et al. 2001).
In conclusion FA measurements of wing feathers may be used as an indicator of both exposure and effects of such contaminants, and further testing will determine if it may function as a general tool for monitoring early effects of organochlorine pollution in bird populations.
We are grateful to Vidar Berg and Anuschka Polder for conducting the laboratory work, and two anonymous referees who provided comments that helped improving an earlier draft of this manuscript. The study was funded by the Norwegian Research Council (project no. 13402/720) and the Norwegian Environmental Ministry.