PCB Congener Profiles in Nestling Tree Swallows and Their Insect Prey

Transcription

1 Arch. Environ. Contam. Toxicol. 52, (2007) DOI: /s z PCB Congener Profiles in Nestling Tree Swallows and Their Insect Prey Zsuzsanna Papp, 1 Gary R. Bortolotti, 2 Mary Sebastian, 3 Judit E. G. Smits 1 1 Department of Veterinary Pathology, University of Saskatchewan, 52 Campus Drive Saskatoon, Saskatchewan S7N 5B4 2 Department of Biology, University of Saskatchewan, Saskatoon, 112 Science Place Saskatchewan S7N 5E2 3 Department of Biological Science, University of Windsor, 401 Suneet Avenue, Windsor Ontario, Canada N9B 3P4 Received: 13 March 2006 /Accepted: 1 August 2006 Abstract. Tree swallows (Tachycineta bicolor) are widely used as indicators of local polychlorinated biphenyl (PCB) contamination in North America. Although determining total PCB residues in tissues is useful in environmental monitoring, analysis of PCB congener profiles may reveal sources of contamination and thus prove to be a more refined tool to track contaminants through the food web. To show how differences in PCB congener patterns in birds can be linked to the PCB patterns in their prey, we evaluated PCB congeners in tissues of tree swallow nestlings and their insect prey using principal component analysis and Euclidean similarities. The PC1 scores for PCB residues in nestlings fell between those of the Hexagenia (Ephemeroptera, mayflies) and Chironomidae (Diptera, midges), the two major prey groups of the nestlings. The congener pattern was not related to the location of nest boxes within the study area. However, Hexagenia insects and the nestlings that consumed them were richer in less chlorinated congeners and had higher PC1 scores than Chironomidae insect. In concordance, congener pattern of nestlings that hatched earlier and consumed more mayflies was more similar than that of other nestling to the pattern of nestings texagenia as calculated by Euclidean similarities. We point to the importance of understanding the seasonal availability of specific types of insect prey and their PCB congener pattern before these data are applied in models of trophic transfer of individual PCB congeners within a food web. Tree swallows (Tachycineta bicolor) are migratory aerialfeeding insectivorous birds widely used as a sentinel species of contaminant exposure across North America (Bishop et al. 1995; Custer et al. 2003; Secored et al. 1999; Harris and Elliot 2000; Smits et al. 2005; McCarty 2001). For local biomonitoring purposes, nestling tree swallows are generally preferred to adults or eggs because their body burdens, which are derived from insects captured locally by their parents, are thought to better represent contaminant conditions in sediments in the vicinity of the nest box. Insects from the order Correspondence to: Zsuzsanna Papp; zsuzsanna.papp@usask. ca Diptera are usually the predominant prey item for tree swallows (McCarty 2001; Echols et al. 2004; Maul et al. 2006). However, mayflies (Ephemeroptera), which are seasonally emergent aquatic insects that can travel several kilometres from their emergence location in depositional sediments, can also significantly contribute to their diet (Wayland et al. 1998; Smits et al. 2005). Tree swallows have previously been used to evaluate trophic transfer of polychlorinated biphenyls (PCBs) (Froese et al. 1998; Custer et al. 2003; Echols et al. 2004; Nichols et al. 2004; Kay et al. 2005; Maul et al. 2006; Custer and Read. 2006). PCBs are a complex mix of compounds whose chemical and physical properties depend on the number and pattern of chlorine substitutions. The degree of chlorination, chlorination pattern, congener hydrophobicity and volatility of the congeners affect resistance to environmental weathering and biotransformation (Mackay et al. 1983; Drouillard et al. 2001). Lake Erie of the Great Lakes has been variably contaminated with PCBs and other persistent organic pollutants from various sources (Kelly et al. 1991; Gewurtz and Diamond 2003; Marvin et al. 2004; Heidtke et al. 2006). PCBs have been detected in sediments, water and biota, including tree swallows and their potential insect prey (Corkum et al. 1997; Bishop et al. 1995). In particular, PCBs were detected and their biologic effects evaluated in tree swallows nesting in the Point Pelee National Park (PPNP), located on a narrow peninsula in western Lake Erie, Canada (Smits et al. 2005; Papp et al. 2005). Insect prey collected from the crops or the stomachs of tree swallow nestlings or from the parents as they arrive at the nest appear to possess a PCB pattern similar to that of the nestlings and the sediments from which the insects emerge, based on principle component analysis (PCA) and visual examination of congener-specific residue data (Novak et al. 1990; Echols et al. 2004). However, PCB congener profiles of potential insect prey collected with traps rather than from the birds appeared to be somewhat or substantially different from that of nestlings (Froese et al. 1997; Echols et al. 2004; Secord et al., 1999), likely because these prey items did not constitute the majority of food consumed by all the nestlings. An additional source of uncertainty is that the above analyses did not give an answer to how the congener patterns of species of prey are

2 258 Z. Papp et al. different statistically or how this difference influenced the congener pattern of the nestlings that consumed them. We believe that exploring such details would help in building better trophic transfer models because the major source of contaminant load in nestlings is their diet. We think there is a link between tissue PCB patterns in nestlings of insectivorous birds and their prey, i.e., the nestlingsõ congener pattern will resemble that of the insect consumed. To study this concept, it is important to know what the nestlings actually ate (insects are to be collected either from parents arriving at the nest or from the crop or stomach of nestlings) rather than assuming which insect may be important in their diet. It is also important to find a statistical tool that can provide information on the difference in congener pattern between nestlings rather than having to rely on visual and graphic estimations. Calculating Euclidean distances (or similarities) can be one such tool (Custer and Read 2006). Because the importance of fine-scale dietary differences to PCB exposure in tree swallows has not been adequately investigated, the purpose of this study was to examine the effects of dietary differences on such exposure. We investigated whether differences in congener profiles between the two major groups of insects eaten by the birds was reflected in congener profiles in the tree swallow nestlings in the PPNP. Methods Study Area The offspring of 20 breeding pairs of tree swallows were studied during 2002 in and around PPNP. All nest boxes were within 100 m of small bodies of water and within 400 m of Lake Erie. Boxes were placed in open areas with scattered shrubs and trees. Tissue Sample Collection The laying date for a clutch was defined as the Julian day the first egg was laid (for details and dates, see Smits et al. 2005). The eggs hatched 17 to 18 days after laying. Nestlings among nest boxes hatched during a 3-week period. At 13 days of age (approximately five days before fledging), 2 nestlings from each of the 20 nests were randomly selected, anaesthetized with inhalation anaesthetic (Halothane; Halocarbon Laboratories, River Edge, NJ), and killed by exsanguination. The liver, spleen, and bursa of Fabricius were excised and used for other studies. Additional organs and remaining tissues from each bird (including brain, kidney, a 2-cm piece of large intestine, plus any stomach contents) were stored in xylene rinsed foil at )20 C for chemical-residue analysis. Insect Collection Nighttime light trapping and day-time sweeping was used to collect insects for contaminant analyses in 2002 and For light trapping, a modified version of a standard Pennsylvania-type trap, developed by Kovats and Ciborowski (1989), was used. We collected insects using pairs of traps situated at each of two locations in the park from the vicinity of swallow boxes (DeLaurier-site A and Marsh Boardwalk-site B)( to W, to N). Nighttime light trapping (June 28, 2002) corresponded with the peak emergence period of mayflies (Smits et al. 2005). Light trapping collected large quantities of mayflies (Hexagenia, Ephemeroptera) and midges (Chironomidae, Diptera) in 2002 but relatively little biomass of terrestrial insects. In 2003, light trapping (July 8, 2003) provided sufficient mass of Hexagenia but not Chironomidae. Because organochlorine contaminant analysis requires up to 2 to 5 g insect mass (Ciborowski and Corkum 1988; Kovats and Ciborowski 1989), daytime sweep netting (aerial and bush) was also used both years in July to collect sufficient insect mass. Insects collected were frozen en mass immediately in hexane rinsed glass jars. In the laboratory, they were thawed, identified, and sorted into taxonomic groups, weighed, and stored refrozen in hexane-rinsed foil until analyzed. Samples from 2002 and 2003 were pooled before contaminant analysis. Contaminant Analysis and Analytical Procedures Contaminant analyses of insects and bird tissues were performed by the Great Lakes Institute of Environmental Research, Analytical Laboratory, University of Windsor, Windsor, Ontario according to standard analytical procedures using a Hewlett Packard (Orangeville, Ontario, Canada) 5890 gas chromatograph equipped with an HP-5972 mass selective detector and HP-7673 autosampler 25 (Drouillard et al. 2001). The tissues from both nestlings collected from each nest were pooled for contaminant analyses and were analysed for 85 PCB congeners. Of the 66 congeners detected, 59 were present in all nestling samples (Papp et al. 2005). A total of 53 and 34 congeners were detected in Hexagenia and Chironomidae, respectively. Thirty four congeners were common to all nestling and insect samples. Dietary Studies Food boluses were collected from 7-10-day-old nestlings from 19 nests as previously described by Smits et al.. (2005). Briefly, an elastic band was secured around the base of the neck of nestlings tightly enough that they could not swallow but loosely enough that their normal gaping behaviour was not affected. Thirty to 45 minutes later or after the parents had made four to six trips to feed the young, chicks were taken out of the nest box, and any food boluses were removed from their crops using a curved haemostat. Nests were sampled for dietary studies on 1 day only for both humane and logistic reasons. All food items collected from the chicks in one nest box were pooled and preserved in a jar with 95% ethanol. In the laboratory, insects were identified and biomass determined. The representation of the 17 different insect taxa in the nestlingsõ diet are described in detail in Smits et al Statistical Analysis Statistical analyses were carried out using SPSS Pearson correlation coefficients (r) were calculated on normally distributed variables with linear associations. Spearman correlation coefficient (r s ) was calculated when the association between variables was not linear. Two-tailed tests of significance were used for all analyses with significance limits of p < An independent samples t test was used to compare the means of Euclidean similarities to Hexagenia congener pattern between two groups of nestlings (nestlings that consumed > 20% or <20% mayflies).

3 PCB Congener Profiles in Nestlings and Insect Prey 259 For statistical analyses, PCB congener concentrations were PCB total- normalized by dividing each congener concentration (ng/g wet weight) with the total PCB concentration (ng/g wet weight) in the particular sample. This was followed by log transformation to achieve or improve normal distribution. PCA was performed on PCB total- normalized, log-transformed concentrations of 24 individual congeners. Although 34 common congeners were detected between nestlings and Hexagenia and Chironomidae samples, we limited the number of congener variables included in the PCA to 24 so as not to exceed the number of cases (i.e., n =24, comprised of 20 nestling and 4 insect samples) in our analysis for the purpose of test accuracy. We chose the 24 most abundant congeners for analysis. These congeners all comprised 1% the total PCB concentration in either or both the nestlings or the insects. In addition, we ran several PCA analyses, including ones with keeping every available congener in the analysis (n = 34) and found that these PCA all lead us to conclusions similar to those presented here. To assign a numeric value to the similarity in congener patterns between nestling and Hexagenia samples, Euclidean similarities were obtained by calculating the Pearson correlation coefficients between the congener patterns as variables. First, the congener pattern variable for the Hexagenia mayflies was generated by calculating the mean PCB total -normalized, log-transformed concentrations of each PCB congener of the two available individual Hexagenia samples (n = 53 congeners). These concentration values, as cases, constituted the variable Hexagenia pattern. The 20 nestling pattern variables were generated (one variable for each nest) similarly. Then we generated 20 Pearson correlation coefficients (r), one for each association between the 20 nest variables and the Hexagenia pattern variable (n = 53 for each association). The values of these coefficients are the values of the Euclidean similarity between the nestlings and the Hexagenia. To visualize the difference in congener pattern between the two insect groups, the Hexagenia minus Chironomidae variable was created. The case values of this variable (data points on the Y axis in Fig. 4.) were calculated by subtracting the mean PCB total -normalized log-transformed concentration of a particular congener in Chironomidae from that in Hexagenia. Results Total concentration and lipid-normalized total concentration of PCBs and relative concentration of selected congeners for Hexagenia, Chironomidae, Formicidae (Hymenoptera, ants), and tree swallow nestlings are presented in Table 1. Because the Hexagenia and the Chironomidae insects were by far the two most important food groups (Smits et al. 2005), statistical analysis was focused on these two insects and the nestlings. PCA was performed on the PCB total -normalized, log transformed concentrations of PCB congeners in the nestlings and the Hexagenia and Chironomidae samples. The 24 congener variables used represented 74% to 77 % of total PCBs in the nestlings and 81% to 84 % of total PCBs in the mayflies and midges. The PCA yielded a major factor (PC1) that explained 50.1% of the variation in the original data and can be interpreted as a continuum with Hexagenia at one end and Chironomidae at the other (Fig. 1). PC2 explained 27.4% of variance in the data. The PC1 scores for nestlings fell between the range for that of the insects. There was no visually detectable relationship between location of nest boxes (i.e., distance from lake shore, marsh, latitude or longitude) and PC1 values of nestlings from neighbouring nest boxes were distributed along the range of PC1 scores. However, PC1 Table 1. Total PCB concentration and the relative concentration of selected congeners in tree swallow nestlings and insects from PPNP % of Total b 52 70/ / / /182 Lipid (ng/g) Wet weight (ng/g) Tree swallows n ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Nestlings a 20( ) Hexagenia A Hexagenia B Chironomidae A Chironomidae B Formicidae A ND ND ND ND ND ND ND Formicidae B ND ND ND ND ND ND ND a GM of total PCB concentration (ng/g wet weight), GM of lipid normalized total PCB concentration (ng/g lipid), and GM of relative concentration of selected congeners (percent of total) 95% confidence interval. b Relative concentration of selected congeners (percent of total). A= Site A (DeLaurier, PPNP). B= Site B (Marsh Boardwalk, PPNP). GM= Geometric mean. ND= Not defected.

4 260 Z. Papp et al PC1 scores Fig. 1. Dot-plot of the first PCA scores of PCB total -normalized, logtransformed concentrations of PCB congeners in tree swallow nestlings (s), Hexagenia (n), and Chironomidae (.) 2 3 Percent mayflies in the diet Euclidean similarities in congener patern between nestlings and mayflies high scores positively correlated with the percent of mayflies in the nestlingsõ diet (r s = 0.79, p = 0.001, n = 13) and negatively correlated with the laying date (r= )0.84, p < 0.001, n = 20). Almost identical results were obtained when Euclidean similarities in congener patterns between mayflies and the nestlings were correlated with the percent of mayflies in the nestlingsõ diet and the laying date (r s = 0.82, p = 0.001, n = 20, respectively) (r s = 0.82, P = 0.001, n = 13 and r = 0.83, P < 0.001, n = 20, respectively). By an alternative statistical approach, the congener pattern of nestlings whose diet consisted of > 20% Hexagenia was more similar to that of Hexagenia than the congener pattern of nestlings that consumed < 20% (regardless of whether the pattern was expressed as Euclidean similarities or PC1) (t test p < 0.001, Fig. 2). In other words, nestlings that hatched earlier and ate more mayflies had PC1 scores that were more similar to those of mayflies and exhibited higher Euclidean similarities to the congener pattern of mayflies than nestlings that consumed fewer mayflies. The percent of mayflies in the nestlingsõ diet very highly correlated with the number of days passed from the peak mayfly emergence (June 28, 2002 according to J.J.H. Giborowski, personal communication, 2006) at the date of low Percent mayflies in the diet Fig. 2. Association between percent Hexagenia mayflies in the nestlingsõ diet and Euclidean similarities in congener pattern ( SE). Euclidean similarities were calculated between the congener pattern of Hexagenia mayflies (mean PCB total -normalized, log-transformed concentrations of congeners in Hexagenia) and those in nestlings. High and low categories of percentage of mayflies in the diet indicates higher (n = 8) or lower (n = 5) than 20% Hexagenia in diet bolus collection (Fig. 3, r = )0.87, p < 0.00l, n = 19) which supports the accuracy and relevance of our dietary studies. In general, PC1 loaded highly positively on low-chlorinated congeners and negatively on high-chlorinated congeners (Table 2), statistically demonstrated by the negative correlation of PC1 component loadings with the logged octanol water partition coefficients (logk ow ; Hawker and Connell 1988) (r = ) 0.839, p < 0.001). Indeed, Chironomidae were richer in higher chlorinated congeners than Hexagenia, demonstrated by the negative correlation between the variable Hexagenia minus Chironomidae and logk ow (r = )0.654, p < 0.001) (Fig. 4) Chironomidae were richer in high chlorinated congeners compared with Hexagenia regardless of site collection of collection (i.e., when individual samples rather than means were compared), including comparisons with five Hexagenia samples from previous years from nearby lake Erie sites (data not shown; based on Corkum et al. 1997; Ciborowski and Corkum 1988). Nestling samples were tightly clustered around PC2, whereas insects were separated from nestlings in their PC2 scores (results not shown). PC2 was only associated with a few congeners (Table 2). Because Formicidae contained 1 order of magnitude lower total PCBs than Chironomidae and Hexagenia (Table 1) with only 17 congeners detected, and because they were a less important food group compared with the other two insects (Smits et al. 2005), statistical analyses were not focused on ants. We noted though, that Formicidae were similar to Chironomidae in PC1 scores and they were richer in high-chlorinated congeners compared with Hexagenia (data not shown). Discussion 0 5 Number of days passed from peak mayfly emergence Fig. 3. Correlation of percent Hexagenia mayflies in the nestlingsõ diet with the number of days passed from the peak mayfly emergence on the bolus collection dates (June 25 to July l3, 2002) Our study provides an example of how differences in PCB congener profiles between two major insect orders is reflected in the congener profiles in the tree swallow nestlings that consumed them. Although we showed that nestlings cluster close to their major insect food in terms of their PCB profiles as shown by PC1, as expected, the samplesõ PCB patterns are 10 15

5 PCB Congener Profiles in Nestlings and Insect Prey 261 Table 2. Loadings of PC1 and PC2 on individual congeners from PCA analysis of PCB total -normalized, log-transtormed concentrations of PCB congeners in n = 24 a PCB PCI PC ) ) / ) )0.42 ) ) ) / ) ) ) /190 ) ) ) ) )0.93 )0.15 a Twenty tree swallow nestlings, 2 Hexagenia, and 2 Chironomidae. "Hexagenia minus Chironomidae" /41/ / /48 56/ /31 153/ / / / logkow Fig. 4. Correlation of logk ow with Hexagenia minus Chironomidae (calculated by subtracting the log-transformed relative concentration of each particular PCB congener in mean Chironomidae from those in mean Hexagenia). Congener numbers are indicated different enough to reveal subtle differences and associations with their diet. The PCB profile of mayflies was characterised by higher relative concentrations of low-chlorinated congeners compared to the nestlings and especially compared to Chironomidae. As a result, nestlings that hatched earlier and were feeding during the time of emergence of mayflies had a PCB pattern more similar (higher Euclidean similarities) to that of mayflies than nestlings raised later in the nesting period, regardless of the location of the nest box within the area. Such small differences in PCB congener patterns, which are likely present in many other studies, can become important and interesting when used for building models of trophic transfer of PCBs. Previous studies have demonstrated that nestling tree swallows have PCB congener patterns similar to that of their major insect food. Based on the PCA approach, the congener pattern of Diptera insects appeared to be similar to that of tree swallow nestlings (Froese et al. 1997; Echols et al. 2004), whereas adults clustered with the pattern shown by eggs (Echols et al. 2004). However, the PCA approach can not test statistical differences in PCB patterns between samples. Instead of using PCA, other researchers have compared Euclidean distances in PCB congener pattern between pooled diet samples and nestlings and found them to be a valuable statistical tool for testing differences in congener patterns between samples (Custer and Read, 2006). We independently used a Euclidean similarities approach along PCA and found both methods to be equally informative regarding detailed differences between samples. Although Euclidean similarities of the nestlingsõ PCB patterns to the pattern of the mayfly appeared high at a first glance for all nestlings, they are variable enough to reveal a significant statistical association with other variables (laying date, percent of mayflies in the diet). The peak emergence of Hexagenia mayfliesõ in 2002 from Lake Erie was approximately June 28 (J. H. H. Ciborowski, 2002, personal observation). Mayflies emerging from the lake can drift several kilometres on the wind from the site of emergence. In contrast, Chironomidae likely developed in the marshes within the boundaries of PPNP (J. H. H. Ciborowski, 2005, personal communication). Nestlings that hatched later in the season missed the peak mayfly emergence and consumed a greater variety of insects including Chironomidae, Hymenoptera, Asilidae and Odonata during time span of their development in the nest (Smits et al. 2005), which is more characteristic of the usual diet of a tree swallow (Menglekoch et al. 2004). The earlier nestlings consumed a more homogeneous mayfly diet (Fig. 3; Smits et al. 2005). We demonstrated that although both insect groups were contaminated with PCBs, Hexagenia were richer in low-chlorinated congeners than Chironomidae regardless of the site of collection. The octanol water partition coefficient (logk ow ), which is strongly associated with congener size and degree of chlorination (Hawker and Connell 1988) correlated with the congenersõ relative concentration in mayflies compared to midges. Although we were limited to two Chironomidae samples, they were almost identical in their PC1 scores; therefore, we are confident that they are different from mayflies. Although our graphs show only two Hexagenia samples, we were not limited to these two in our study. The congener patterns of mayflies from previous studies from nearby areas of western Lake Erie (five samples from Corkum et al. [1997] and Ciborowski and Corkum [1988]) closely matched ours in their PC scores. i.e. their PC1 scores were the same or higher than our Hexagenia when ran together with our samples (Papp et al, 2005, unpublished observation). In addition, they all exhibited higher Euclidean similarities to each other than to Chironomidae.

6 262 Z. Papp et al. There are several possible explanations for the difference in PCB congener pattern between mayflies and midges. Perhaps Aroclors with different amount of high- versus low-chlorinated congeners were the source of contamination of the marshes where the Chironomidae likely developed compared with the Lake Erie sediments where the mayflies spent their larval stages. Alternatively, the source of PCBs in the marshes may have been from aerial deposition, whereas the western part of Lake Erie was mostly contaminated from industrial effluents through the Detroit River (Heidtke et al. 2006). It is also possible that PCBs in the marshes were exposed to more weathering or more biodegradation than those in Lake Erie, resulting in the enrichment of marsh sediments and Chironomidae in high chlorinated congeners. In general, high chlorinated congeners possess fewer open (i.e., unchlorinated) para-meta sites on the biphenyl structure and therefore are more resistant to chemical changes and metabolic breakdown than the less-chlorinated congeners. Finally, mayflies might have a lower capacity for biotransformation of PCBs than midges despite the fact that the congener patterns of both appear to differ from sediment patterns (Gobas et al. 1989; Drouillard et al. 1996; Froese et al. 1998; Maul et al. 2006). Our results and those of other studies (Froese et al. 1998; Echols et al. 2004; Maul et al. 2006) suggest that the difference between the congener patterns of nestlings (i.e., relatively high versus relatively low concentration of higher chlorinated congeners in their tissues) can be ascribed to different congener patterns in their insect food. Although we had no data available on eggs for this article we believe that the amount of insects consumed and the rate of growth are so substantial in tree swallow nestlings that contaminants from the egg would be diluted (egg weight is only 10% of body mass of a 13-dayold nestling). Consequently, only 10 % of total PCBs in nestlings are calculated to originate from the egg in most studies where local bioaccumulation of PCBs from food is evident in the nestlings (Custer et al. 2003; Maul et al. 2006, McCarty 2001; and based on Bishop et al. 1995; Froese et al. 1998, Ankley et al. 1993). The effect of biotransformation of dietary PCBs by nestlings on the PCB pattern differences was not considered important in our study because there is no reason to believe that biotransformation capacity differed from nest to nest among these clinically normal, healthy nestlings. Even if we consider some biochemical differences among nestlings, the effect of metabolism on PCB levels and pattern in nestlings would be negligible considering the amount of PCBs consumed in a short period of time. In fact, CYP 450 activity was lower in the nestlings rich in high-chlorinated congeners because of lower total PCBs (Papp et al. 2005), the opposite of what would happen if enzyme activity significantly degraded the more easily biodegradable low-chlorinated congeners. However, metabolic differences between nestlings and insect prey could lead to different congener profiles between nestlings and prey (Froese et al. 1998). A uniform biotransformation of PCBs by nestlings could result in a shift in congener profiles from lower chlorinated to higher chlorinated compounds, therefore decreasing the difference in congener profile between high and low mayfly consumers. In conclusion, the present study demonstrates how a small difference in laying date among nests can influence the PCB congener pattern of nestlings within a study area because of availability of different insect foods with different contaminant patterns. It emphasizes the importance of studying actual diet in studies of contaminant transfer. If major classes of food items possess different PCB composition profiles, and at least one of these insect groups is available only seasonally, nestlings developing at slightly different time in the season are raised on food with a different PCB pattern compared with other nestlings. These fine differences in congener profiles between nestlings need to be recognized before data are used for congener-specific modeling of trophic transfer of PCBs. Acknowledgments. We thank K. G. Drouillard for GC analyses and preparations; J. J. H. Ciborowski for helpful comments on the manuscript; and A. Anaka, P. Garcia and M. Alvarez for research assistance. We would like to thank numerous people from Point Pelee National Park (Parks Canada) G. Mouland, D. Reive, T. Linke and the park wardens for their continuous support. We also thank the two anonymous reviewers of the manuscript for helpful comments. This research was funded through grants from NSERC, Suncor Energy, and Syncrude Canada Ltd (J. S., G. B., and J. J. H. Ciborowski). References Ankley GT, Niemi GJ, Lodge KB, Harris HJ, Beaver DL, Tillitt DE et al. (1993) Uptake of planar polychlorinated-biphenyls and 2,3.7,8-substituted polychlorinated dibenzofurans and dibenzo-p-dioxins by birds nesting in the lower fox river and green bay, Wisconsin, USA. Arch Environ Contam Toxicol 24: Bishop CA, Koster MD, Chek AA, Hussell DJT, Jock K (1995) Chlorinated hydrocarbons and mercury in sediments, red-winged blackbirds (Agelaius phoeniceus) and tree swallows (Tachycineta bicolor) from wetlands in the Great Lakes St. Lawrence river basin. Environ Toxicol Chem 14: Ciborowski JJH, Corkum LD (1988) Organic contaminants in adult aquatic insects of the St. Clair and Detroit rivers, Ontario, Canada. J Great Lakes Res 14: Corkum LD, Ciborowski JJH, Lazar R (1997) The distribution and contaminant burdens of adults of the burrowing mayfly, Hexagenia, in Lake Erie. J Great Lakes Res 23: Custer TW, Custer CM, Hines RK (2002) Dioxins and congenerspecific polychlorinated biphenyls in three avian species from the Wisconsin River, Wisconsin. Environ Pollut 119: Custer CM, Custer TW, Dummer PM, Munney KL (2003) Exposure and effects of chemical contaminants on tree swallows nesting along the Housatonic River, Berkshire Country, Massachusetts, ESA, Environ Toxicol Chem 22: Custer CM, Read LB (2006) Polychlorinated biphenyl congener patterns in tree swallows (Tachycineta bicolor) nesting in the Housatonic River watershed, western Massachusetts, USA, using a novel statistical approach. Environ Pollut 142: Drouillard KG, Ciborowski JJH, Lazar R, Haffner GD (1996) Estimation of the uptake of organochlorides by the Mayfly Hexagenia limbata (Ephemeroptera: Ephemeridae). J Great Lakes Res 22:26 35 Drouillard KG, Fernie KJ, Smits JE, Bortolotti GR, Birds DM, Norstrom RJ (2001) Bioaccumulation and toxicokinetics of 42 polychlorinated biphenyl congeners in American kestrels (Falco sparverius). Environ Toxicol Chem 20: Echols KR, Tillitt DE, Nichols JW, Secord AL, McCarty JP (2004) Accumulation of PCB congeners in nestling tree swallows (Tachycineta bicolor) on the Hudson River, New York. Environ Sci Technol 38:

7 PCB Congener Profiles in Nestlings and Insect Prey 263 Froese KL, Verbrugge DA, Ankley GT, Niemi GJ, Larsen CP, Giesy JP (1998) Bioaccumulation of polychlorinated biphenyls from sediments to aquatic insects and tree swallow eggs and nestlings in Saginaw Bay, Michigan, USA. Environ Toxicol Chem 17: Gewurtz SB, Diamond ML (2003) Distribution and burdens of bioaccumulative contaminants in the Lake Erie food web: A review. Environ Rev 11: Gobas FAPC, Bedard DC, Ciborowski JJH, Haffner GD (1989) Bioaccumulation of chlorinated hydrocarbons by the mayfly (Hexagenia limbata) in lake St. Clair. J Great Lakes Res 15: Harris ML, Elliott JE (2000) Reproductive success and chlorinated hydrocarbon contamination in tree swallows (Tachycineta bicolour) nesting along rivers receiving pulp and paper mill effluent discharges. Environ Pollution 110: Hawker DW, Connell DW (1988) Octanol-water partition coefficients of polychlorinated, biphenyl congeners. Environ Sci Technol 22: Heidtke T, Hartig JH, Zarull MA, Yu B (2006) PCB levels and trends within the Detroit River Western Lake Erie basin: A historical perspective of ecosystem monitoring. Environ Monit Assess 112:23 33 Kay DP, Blankenship AL, Coady KK, Neigh AM, Zwiernik MJ, Millsap SD, et al. (2005) Differential accumulation of polychlorinated biphenyl congeners in the aquatic food web at the Kalamazoo river superfund site, Michigan. Environ Sci Technol 39: Kelly TJ, Czuczwa JM, Sticksel PR, Sverdup GM, Koval PJ, Hodanbosi RF (1991) Atmospheric and tributary inputs of toxic substances to Lake Erie. J Great Lakes Res 17: Kovats ZE, Ciborowski JH (1989) Aquatic insect adults as indicators of organochlorine contamination. J Great Lakes Res 15: Mackay D, Shiu WY, Billington J, Huang GL (1983) Physical chemical properties of polychlorinated biphenyls. In: Mackay D, Pattersom S, Eisenreich SJ, Simmons MS (eds) Physical behavior of PCBs in the Great Lakes. Ann Arbor Science, Ann Arbor, MI, pp Marvin CH, Painter S, Charlton MN, Fox ME, Thiessen PAL (2004) Trends in spatial and temporal levels of persistent organic pollutants in Lake Erie sediments. Chemosphere 54:33 40 Maul JD, Belden JB, Schwab BA, Whiles MR, Spears B, Faris JL, et al. (2006) Bioaccumulation and trophic transfer of polychlorinated biphenyls by aquatic and terrestrial insects to tree swallows (Tachycineta bicolor). Environ Toxicol Chem 25: McCarty JP (2001) Use of tree swallows in studies of environmental stress. Rev Toxicol 4: Menglekoch JM, Niemi GJ, Regal RR (2004) Diet of the nestling tree swallow. Condor 106: Nichols JW, Echols KR, Tillitt DE, Secord AL, McCarty JP (2004) Bioenergetics-based modeling of individual PCB congeners in nestling tree swallows from two contaminated sites on the upper Hudson River, New York. Environ Sci Technol 38: Novak MA, Reilly AA, Bush B, Shane L (1990) In situ determination of PCB congener-specific first order absorption/desorption rate constants using Chironomus tentans larvae (Insecta: Diptera: Chironomidae). Water Res 24:32l 327 Papp Z, Bortolotti GR, Smits JE (2005) Organochlorine contamination and physiological responses in nestling tree swallows in Point Pelee National Park, Canada. Arch Environ Contam Toxicol 49: Secord AL, McCarty JP, Echols KR, Meadows JC, Gale RW, Tillitt DE (1999) Polychlorinated biphenyls and 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents in tree swallows from the upper Hudson River, New York State, USA. Environ Toxicol Chem 18: Smits JEG, Wayland ME, Miller MJ, Liber K, Trudeau S (2000) Reproductive, immune, and physiological end points in tree swallows on reclaimed oil sands mine sites. Environ Toxicol Chem 19: Smits JEG, Bortolotti GR, Sebastian M, Ciborowski JJ (2005) Spatial, temporal, and dietary determinants of organic contaminants in nestling tree swallows in Point Pelee National Park, Ontario, Canada. Environ Toxicol Chem 12: Wayland M, Trudeau S, Marchant T, Parker D, Hobson KA (1998) The effect of pulp and paper mill effluent on an insectivorous bird, the tree swallow. Ecotoxicology 7: