- Email: [email protected]
Behavioral variation and its consequences during incubation for American kestrels exposed to polychlorinated biphenyls
ARTICLE IN PRESS
Ecotoxicology and Environmental Safety 63 (2006) 226–235 www.elsevier.com/locate/ecoenv
Behavioral variation and its consequences during incubation for American kestrels exposed to polychlorinated biphenyls Sheri A. Fishera, Gary R. Bortolottia,, Kimberley J. Fernieb, David M. Birdc, Judit E. Smitsd a
Department of Biology, University of Saskatchewan, 112 Science Pl., Saskatoon, Sask., Canada S7N 5E2 b Environment Canada, 867 Lakeshore Road, Burlington, Ont., Canada L7R 4A6 c Avian Science and Conservation Centre, McGill University, Ste. Anne de Bellevue, Quebec, Canada H9X 3V9 d Department of Veterinary Pathology, University of Saskatchewan, Saskatoon, Sask., Canada S7N 5B4 Received 12 May 2005; received in revised form 11 July 2005; accepted 25 July 2005 Available online 5 October 2005
Abstract We investigated whether polychlorinated biphenyl (PCB) exposure in American kestrels (Falco sparverius) influenced incubation behavior and whether altered behavior could lead to poor reproductive success. Captive kestrels were fed a mixture of PCBs (Aroclors 1248:1254:1260) at an approximate daily dose of 7 mg/kg body weight, 1 month prior to pairing and throughout incubation. Behaviors of 23 control and 23 PCB-exposed pairs were monitored throughout incubation using an electronic balance in the nest box. PCB exposure resulted in longer incubation periods and in altered incubation behaviors. Seven of 14 behavioral variables showed some association with treatment, with sex-specific effects largely biased toward disrupted male behavior. For most behaviors, the treatment effect was explained by the delayed clutch initiation induced by PCBs rather than by a direct physiological impact of the contaminants. PCB-exposed pairs with greater attendance to their eggs and better coordination of incubation duties had improved hatching success. r 2005 Elsevier Inc. All rights reserved. Keywords: American kestrel; Falco sparverius; Raptor; Incubation behavior; Polychlorinated biphenyls; PCBs; Organochlorine contaminants; Hatching success
1. Introduction Polychlorinated biphenyls (PCBs) are persistent, widespread environmental contaminants. As PCBs bioaccumulate, birds at higher trophic levels of the food chain are particularly vulnerable to PCB exposure (Elliott et al., 1991; Barron et al., 1995; Hoffman et al., 1996a). Historical population declines of raptors have been attributed to contaminant-induced reproductive failure (e.g., Kozie and Anderson, 1991; Clark et al., 1998; Valkama and Korpimaki, 1999). The breeding cycle and associated behaviors that are under hormonal control may be vulnerable to endocrine-modulating substances such as PCBs (Barron et al., 1995; Vos et al., 2000), especially since slight hormonal differences can elicit significant behavioral change (Silver and Ball, 1989; Jacobs and Wingfield, 2000). Behavioral Corresponding author. Fax: +1 306 966 4461.
E-mail address: [email protected] (G.R. Bortolotti). 0147-6513/$ - see front matter r 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2005.07.021
observations rather than intrusive physiological examinations are increasingly being used and encouraged in toxicological research to identify adverse effects of contaminants (Doving, 1991; Evangelista de Duffard and Duffard, 1996; Clotfelter et al., 2004; Zala and Penn, 2004). Altered behavior during the breeding season has been observed in birds exposed to organochlorine contaminants, including PCBs. Reduced nest defence behavior occurred in merlins (Falco columbarius) (Fyfe et al., 1976). Peregrine falcons (F. peregrinus) exposed to dichlorodiphenyldichloroethane displayed clumsy behavior in the nesting scrape, including stepping on the eggs (Nelson, 1976). Eggdestroying behavior occurred in contaminated gray herons (Ardea cinerea) (Milstein et al., 1970) and captive mallards (Anas platyrhynchos) (Risebrough and Anderson, 1975). McCarty and Secord (1999a) found that tree swallows (Tachycineta bicolor) in PCB-contaminated areas built lowquality nests, which were associated with reduced breeding success. Atypical parental behavior has also been suggested
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
to contribute to decreased reproductive success in cormorants (Phalacrocorax carbo) contaminated with PCBs, chlorinated dioxins, and dibenzofurans (Van Den Berg et al., 1995). Incubation behavior can be disrupted by sublethal exposure to environmental pollutants (Fry, 1995), including a lack of behavioral synchrony between the sexes which normally ensures that eggs are not left unattended (Koenig, 1982). Unattended nests are exposed to greater risk from predation, and embryos could die if exposed to a suboptimal thermal environment (Webb, 1987). Because the cooling rate of an unattended egg exceeds the heating rate when parents return, absences greater than about 1 min are also energetically wasteful to the parent (Drent, 1973). Exposure of parent birds to PCBs has been linked to lengthened and more frequent recesses, prolonged incubation, and increased incidence of nest abandonment (Peakall and Peakall, 1973; Fox et al., 1978; McArthur et al., 1983; Kubiak et al., 1989; McCarty and Secord, 1999b; Bustnes et al., 2001). Our study subject, the American kestrel (Falco sparverius), is a biparental incubator (Balgooyen, 1976; Bortolotti and Wiebe, 1993). We were able to monitor breeding pairs for continuous periods of time throughout incubation, whereas many previous studies of contaminated birds have used infrequent nest checks or monitoring for only short segments (Peakall and Peakall, 1973; Kubiak et al., 1989; Harris et al., 1993; Custer et al., 1998; Bustnes et al., 2001). In addition, and unlike some studies, we could easily differentiate males from females, enabling us to reliably investigate sex-specific effects. Complete data on reproductive performance also allowed us to investigate the consequences of variation in incubation. We predict that pairs with greater attendance to their eggs, and better coordination of their incubation duties, should have greater hatching success. 2. Materials and methods 2.1. Study animals This study took place at the Avian Science and Conservation Centre of McGill University, Canada, using American kestrels of known pedigree and age. Birds were randomly designated control (CTL) (n ¼ 25 pairs) or PCBexposed (PCB) (n ¼ 25 pairs) and were placed in flight pens (6 6 2.5 m) segregated by sex and treatment. The kestrels were fed ad libitum on their typical diet of dayold cockerels. Based on the PCB congeners found in the eggs and tissues of wild birds from the Great Lakes region (Braune and Norstrom, 1989, Clark et al., 1998) and PCB residue levels found in wild prey species of kestrels (Environment Canada, unpublished data), a dosing regime that would generate environmentally relevant levels of PCBs was calculated (Fernie et al., 2000). Chronic dietary exposure began 1 month before pairing on 18 March 1998 and
227
continued until the end of the incubation period of each pair (range 79–117 days of exposure; mean 95 days). A mixture of Aroclors 1248:1254:1260 (1:1:1 by weight; Monsanto, St. Louis, MO, USA) was dissolved in safflower oil at a concentration of 4.85 mg/g total PCB. As kestrels show a preference for consuming the heads of cockerels (I. Ritchie, pers. comm.), 100-ml aliquots of either PCB dosing mixture or plain safflower oil were injected into the heads of frozen–thawed day-old cockerels to be fed to the PCBexposed or control groups, respectively. The birds consumed approximately 7 mg/kg body weight/day (Fernie et al., 2000; Drouillard et al., 2001). The total PCB residue levels from the kestrel eggs averaged 34.1 mg/g (geometric mean) on a whole-egg wet-weight basis (Fernie et al., 2000). These levels are comparable to those found in the eggs of wild raptors exhibiting decreased reproductive success (Hoffman et al., 1996b; Clark et al., 1998; Valkama and Korpimaki, 1999). Genetically unrelated birds, all with previous breeding experience, were paired on 21 April 1998. The pairs were placed into outdoor breeding pens (2.3 0.9 3.6 m), which contained rope and wooden perches, a one-way glass window for observation, and a nesting box. 2.2. Monitoring of incubation behavior The incubation period was defined as the time from the completion of the clutch to the hatching of the first egg or until 28 days after clutch completion for pairs that failed to hatch any egg. A custom-built electronic balance system was used to quantify behavior. The monitoring system consisted of a wooden incubation box (20.3 17.8 10.2 cm) mounted onto a balance. Electrical impulses proportional to the mass of the incubation box were sent from the balance to a microchip once every minute, 24 h per day. Twenty electronic balances were randomly assigned to PCB-exposed or control pens. Every 5 days during the incubation period balances were switched between pens of the same treatment group to increase sample size (PCB n ¼ 23; CTL n ¼ 23). One PCB-exposed pair did not complete a clutch and was not monitored, while another PCB-exposed pair abandoned their clutch. Two control pens were not monitored to spread the effort equally between treatments. When pairs were not being electronically monitored, a wooden incubation box with the same dimensions was placed in the nest box. The bottom of the incubation box was filled with 2–4 cm of wood shavings. A door on the back of the nest box facilitated nest checks and removal of balances. Nest boxes were checked daily, and eggs were numbered to identify laying sequence. Ten days after clutch completion, eggs were candled to determine fertility and one egg was removed from each pair for PCB residue analysis (see Fernie et al., 2000). Another fresh but nonviable kestrel egg was put into the box to maintain the original clutch size. Female kestrels were heavier than their mates, allowing us to use mass to determine the sex of the bird in the nest
ARTICLE IN PRESS 228
S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
box. In pairs where the masses of both sexes were similar (two PCB and one control pair), data on only the presence or absence of a bird were used. While the scales detected the presence of the parent, the adult may not have been incubating the entire time that it was in the nest box. To confirm that the balances could accurately detect the presence/absence and sex of a parent, visual observations were conducted at 08:00, 10:00, 13:00, 15:00, 17:00, and 20:00 h each day by looking through the one-way glass window. In total, 3792 visual scans were performed, and 3743 of them concurred with the balance (99% scale accuracy). The few discrepancies were not necessarily errors as the birds may have flushed or incubation switches could have occurred after the observer completed the scan and before the scale reading was taken. 2.3. Behavioral data analysis To determine the percentage of the day that the nest was occupied (PCB n ¼ 23; CTL n ¼ 23), the number of minutes of nest occupation was divided by day length. We refer to day length as the time between the end of the overnight incubation bout and the start of the overnight incubation bout. Females were responsible for overnight incubation (see also Bortolotti and Wiebe, 1993). The variables chosen describe the role of each sex during incubation plus the combined effort of the parents (Table 1). Perhaps to ensure the continuity of incubation and thus not leave eggs exposed, kestrels performed rapid Nest Switches. Nest Switches occurred either when the bird in the box exited and was replaced within the same minute by its mate or when the mate entered the nest and the incubating bird left. The number of Nest Switches and Time of Last Recess indicate the coordination of the sexes (see below). The incubation period of kestrels is typically 28–29 days but can range from 26 to 31 days (Bird, 1988). The incubation period was divided into early, middle, and late (1–8, 9–16, 17+ days since clutch completion) observation
periods because behavior and hormones have seasonal patterns of variation (Rehder et al., 1986; Silver and Ball, 1989). Behavioral data for each pen were averaged within each of the three periods. Sample sizes for each period vary slightly as not all pens were observed each time. The average numbers of observation days for controls were 4.7, 4.6, and 5.8 days, and those for PCB-exposed birds were 6.1, 5.1, and 7.6 days for early to late periods, respectively. For repeated-measures analyses (across observation periods), the total sample sizes were: PCB n ¼ 21 and CTL n ¼ 19. Variables were transformed when they did not meet the assumptions for statistical tests (Table 1). PCB-exposed kestrels had later clutch completion dates (CCD) compared to controls (PCB: mean Julian date ¼ 141, SE ¼ 1.76, n ¼ 24; CTL: mean Julian date 134.8, SE 1.26, n ¼ 25; t ¼ 2.93, df ¼ 47, P ¼ 0.005). If some behaviors varied seasonally, then PCB and control groups could differ in behavior by virtue of when eggs were laid (i.e., CCD). In such a case, an analysis of variance (ANOVA) would reveal a significant treatment effect, but the result would be incorrectly interpreted as an effect of PCBs on the behavior per se. We therefore used analysis of covariance (ANCOVA) with CCD as a covariate. In the case where treatment is no longer significant after controlling for CCD, we refer to PCBs as having an ‘‘indirect’’ effect on behavior. While it is still important to recognize that the treatments differed, these differences in behavior would be explained by how early in the breeding season that the birds bred. We considered PCBs to have had a ‘‘direct’’ effect on behavior when treatment remained significant in the ANCOVA. Therefore, we performed two statistical tests on the same data set: repeated-measures ANOVA and ANCOVA. Few studies have made a distinction between such indirect and direct effects. 2.4. Hatching success To determine the consequences of behavioral variation due to contamination, variables that were directly and
Table 1 Definitions of behavioral variables and type of transformation used for statistical analysis Variable
Definition
Transformation
Percent In Percent Female In Percent Male In No. Female Bouts No. Male Bouts No. Recesses No. Switches Female Bout Length Male Bout Length Recess Length First Switch Last Switch First Recess Last Recess
No. min nest occupied /day length 100 No. min female occupied nest/day length 100 No. min male occupied nest/day length 100 No. of times female on nest No. of times male on nest No. of times nest was unoccupied No. of times one parent relieved the incubating bird No. min female occupied nest/No. Female Bouts No. min male occupied nest/No. Male Bouts No. min nest unoccupied/No. Recesses Time of day of the first switch Time of day of the last switch Time of day of the first recess Time of day of the last recess
Arcsine Arcsine Arcsine Square root Square root Square root Square root Square root Square root Log 10 None None None None
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
indirectly associated with PCB exposure were examined in relation to hatching success. The typical hatching rate of captive and wild populations of kestrels varies from 66% to almost 100% (see review by Bird, 1988 and G.R. Bortolotti, unpublished data). We categorized success as zero (0%), moderate (1–74%), and high (75–100%). In other words, allowing for the small variation in clutch size (four or five eggs), we compared pairs that failed to hatch all, two or three, or at most one egg. Because embryos vary in their sensitivity to the thermal environment depending on their stage of development (Webb, 1987), analyses were performed for each observation period. 2.5. Length of incubation period Inefficient incubation behavior may have negative consequences in addition to mortality to developing embryos. Eggs given inadequate heating may take longer to hatch; therefore, we examined the number of days for each viable egg to hatch from the date that the clutch was completed (PCB n ¼ 29, CTL n ¼ 44). Because CCD was defined as the start of incubation, that date was used to determine the length of the incubation period. Kestrels partially incubate eggs before the clutch is complete, resulting in hatching asynchrony, so the position of an egg in the laying sequence influences its incubation period (Bortolotti and Wiebe, 1993). Therefore, egg sequence and treatment were both considered factors in an ANOVA. For analysis of consequences of behavior we began with full ANOVA or ANCOVA models that included all effects and interactions and then repeated the analyses after iteratively removing nonsignificant interactions or factors. All statistical analyses were performed using SPSS software (Noru˘sis, 1991). 3. Results 3.1. Incubation behavior Exposure to PCBs produced detectable changes in frequency, length, and timing of behaviors. In addition to statistical results (Table 2) we present behavioral data (Table 3) to allow for comparison to other studies on wild or captive birds. Seven of 14 behavioral variables showed some association with treatment. In most cases, PCBs had an indirect effect as the ANCOVA showed that the treatment difference detected by ANOVA could be accounted for date of clutch completion. For the Number of Recesses and Number of Switches there was an interaction between treatment and observation period. There were no significant differences between the Number of Recesses taken by control and PCB pairs during periods 1 and 2 (P40:61); however, PCB pairs took more recesses during period 3 (Fig. 1a; t ¼ 2:18, df ¼ 38, P ¼ 0.036). Similarly, the Number of Switches did not differ between groups in periods 1 or 2 (P40.26), but PCB-exposed kestrels had fewer switches in period 3 (Fig. 1b; t ¼ 2.5,
229
Table 2 Results of repeated-measures ANOVA and repeated-measures ANCOVA for all incubation variables across early, middle, and late observation periods of PCB-exposed or control American kestrels Variable
Percent In Observation periods Treatment CCD Observation periods Observation periods Percent Female In Observation periods Treatment CCD Observation periods Observation periods Percent Male In Observation periods Treatment CCD Observation periods Observation periods Female Bout Length Observation periods Treatment CCD Observation periods Observation periods Male Bout Length Observation periods Treatment CCD Observation periods Observation periods Recess Length Observation periods Treatment CCD Observation periods Observation periods No. Female Bouts Observation periods Treatment CCD Observation periods Observation periods No. Male Bouts Observation periods Treatment CCD Observation periods Observation periods No. Recesses Observation periods Treatment CCD Observation periods Observation periods
ANOVA
ANCOVA
P
P
Treatment CCD
0.000 0.012 — 0.220 —
0.225 0.123 0.036 0.646 0.264
Treatment CCD
0.005 0.332 — 0.225 —
0.625 0.544 0.412 0.320 0.578
Treatment CCD
0.075 0.034 — 0.690 —
0.283 0.158 0.068 0.821 0.306
Treatment CCD
0.121 0.732 — 0.809 —
0.525 0.559 0.461 0.604 0.506
treatment CCD
0.497 0.129 — 0.390 —
0.890 0.479 0.031 0.365 0.882
treatment CCD
0.000 0.009 — 0.022 —
0.202 0.078 0.086 0.028 0.295
Treatment CCD
0.705 0.912 — 0.985 —
0.972 0.624 0.288 0.995 0.981
Treatment CCD
0.204 0.027 — 0.268 —
0.505 0.156 0.029 0.339 0.499
Treatment CCD
0.000 0.345 — 0.020 —
0.056 0.570 0.499 0.220 0.087
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
230 Table 2 (continued ) Variable
ANOVA
ANCOVA
P
P
Treatment CCD
0.018 0.085 — 0.082 —
0.017 0.219 0.273 0.381 0.021
First Switch Observation Periods Treatment CCD Observation periods Treatment Observation Periods CCD
0.012 0.214 — 0.350 —
0.351 0.730 0.016 0.693 0.332
Last Switch Observation periods Treatment CCD Observation periods Observation periods
Treatment CCD
0.044 0.195 — 0.988 —
0.173 0.496 0.100 0.750 0.209
treatment CCD
0.003 0.104 — 0.595 —
0.503 0.467 0.057 0.644 0.552
treatment CCD
0.009 0.002 — 0.019 —
0.212 0.018 0.139 0.155 0.258
No. Switches Observation periods Treatment CCD Observation periods Observation periods
First Recess Observation periods Treatment CCD Observation periods Observation periods Last Recess Observation periods Treatment CCD Observation periods Observation periods
Length, Number of Recesses) and better coordination of their incubation duties (Time Last Recess, Number of Nest Switches) should have better hatching success. 3.3. Length of incubation period
The date of clutch completion (CCD) was used as a covariate in ANCOVA models.
df ¼ 36, P ¼ 0.016). Both of these behaviors varied in a strong seasonal pattern in control birds but not at all in exposed birds (Fig. 1). Three other behaviors were indirectly affected by PCBs regardless of observation period; Percent In, Percent Male In, and Number of Male Bouts were all lower in PCB-exposed kestrels. For variables directly affected by PCBs, the timing of Last Recess was later for PCB-exposed kestrels, and the effect on Recess Length was contingent on the period; it was longer in contaminated birds during periods 1 and 3 (PD1: t ¼ 3.8, df ¼ 38, Po0.001; PD3: t ¼ 1.98, df ¼ 38, P ¼ 0.055) but not period 2 (t ¼ 1.1, df ¼ 38, P ¼ 0.29). 3.2. Hatching success For simplicity, Table 4 presents only significant (P40.05) associations between behavior and hatching success categories per observation period. The results, all of which are for exposed birds, are all consistent with our prediction that pairs with greater attendance to their eggs (Percent In, Percent Male In, Number Male Bouts, Recess
Egg number in the laying sequence affected time from clutch completion to hatching (P ¼ 0.002) as expected, given hatching asynchrony. PCB exposure also increased the hatching time of each egg (P ¼ 0.026), with the mean incubation period for PCB-exposed birds being 26.9 days (SE ¼ 0.24, n ¼ 29) vs. 26.1 days for CTL pairs (SE ¼ 0.25, n ¼ 44). 4. Discussion Unlike most behavioral studies which provide snapshots of a few specific incubation behaviors, we were able to monitor multiple behaviors for half the actual time spent incubating across the entire incubation period. CTL kestrels spent more time in the nest box than PCB-exposed birds, with the treatment difference being primarily the result of reduced male contribution (Tables 2 and 3). Within studies of both captive and wild kestrels, the extent to which males incubate varies from 0% to 60% (Porter and Wiemeyer, 1972; Balgooyen, 1976; Wilmers et al., 1985). Control males contributed between 1.2% and 40.0% of daily incubation (16.8% overall), and PCB-exposed males contributed between 0.1% and 30.2%, with an overall average of 10.7%. It appeared that PCB females may have compensated slightly for the reduction in incubation by their mates (Table 3) but not enough to attain the same overall nest attendance (i.e., Percent In). The only behavior that may suggest an effect of PCBs on females was Time of Last Recess, as females were responsible for all overnight incubation. The time that the final recess began was later in PCB-exposed pairs; however, the problem may lie with the male for not performing a nest switch. The Number of Nest Switches occurred less often in PCB-exposed pairs. Fewer nest switches indicate poor parental coordination, which is vital to ensure that eggs are not unattended. Highly male-biased effects of PCBs have also been found in our study population for several physiological parameters (Smits and Bortolotti, 2001; Smits et al., 2002) and reproduction (Fernie et al., 2001 a, b). Furthermore, PCB-exposed males, but not females, exhibited altered sexual and flight behaviors during courtship (Fisher et al., 2001). Altered courtship may even explain the delay in clutch initiation that had such an overwhelming subsequent effect on behavior in this study. While the delay in breeding may be directly induced by PCBs, it is also possible that exposed males did not adequately stimulate their mates’ endocrine systems. The latter is plausible given that in this colony it has been found that young males breeding for the first time do not court as well as experienced males and as a result
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
231
Table 3 Means and 95% lower (LCL) and upper (UCL) confidence limits for all incubation behavioral variables for each treatment and observation period (PD) Variable
PD
Treatment Control X¯
PCB LCL
UCL
X¯
LCL
UCL
Percent In
1 2 3
89.6 93.2 95.2
87.5 91.3 93.7
91.5 94.7 96.6
85.4 91.7 91.8
81.9 89.7 89.6
88.6 93.5 93.7
Percent Female In
1 2 3
74.5 76.7 76.7
70.1 71.4 72.1
79.4 81.6 80.9
75.3 80.8 79.5
70.2 77.7 76.1
79.9 83.8 82.6
Percent Male In
1 2 3
14.2 16.2 17.3
9.7 10.1 12.8
18.7 22.3 21.8
9.0 10.2 10.4
5.3 6.3 6.9
12.8 14.2 13.9
Female Bout Length
1 2 3
30.3 33.0 33.9
25.4 26.9 27.6
35.6 39.7 40.9
31.3 35.3 33.9
26.3 29.1 29.8
36.8 42.2 38.2
Male Bout Length
1 2 3
9.7 10.6 9.7
7.5 7.3 7.6
12.1 14.6 12.1
7.0 7.5 8.2
5.2 5.2 5.8
9.2 10.2 11.0
Recess Length
1 2 3
3.9 3.4 3.0
3.4 2.9 2.6
4.4 3.8 3.5
5.8 3.8 3.8
4.9 3.2 3.2
7.1 4.5 4.6
No. Female Bouts
1 2 3
22.1 21.1 21.9
18.7 17.0 18.0
25.8 25.7 26.1
21.7 21.1 21.7
19.4 18.1 19.0
24.1 24.3 24.6
No. Male Bouts
1 2 3
11.1 12.2 13.7
8.4 9.2 10.6
14.2 15.6 17.2
7.9 8.3 8.1
5.3 6.1 5.4
11.1 10.9 11.3
No. Recesses
1 2 3
21.8 17.3 13.2
18.1 13.2 9.5
25.9 21.8 17.5
20.6 18.6 18.6
17.7 15.2 15.5
23.8 22.5 21.1
No. Switches
1 2 3
14.2 18.8 26.8
9.8 11.8 16.7
19.4 27.5 39.2
11.1 13.9 12.7
7.2 7.4 6.9
16.0 22.5 20.1
Female Switch
1 2 3
493.2 466.8 423.6
434.1 384.9 360.3
552.3 548.7 486.8
553.9 487.0 502.4
466.7 430.9 425.7
641.1 543.1 579.1
Last Switch
1 2 3
1097.7 1108.3 1142.2
1046.3 1045.4 1087.0
1149.0 1171.3 1197.5
1062.4 1076.3 1112.9
1013.0 1033.1 1078.8
1111.8 1119.6 1147.1
First Recess
1 2 3
374.1 403.4 407.4
349.7 374.8 372.8
398.5 432.0 442.1
357.1 371.9 376.7
338.2 346.9 346.9
376.0 396.9 406.5
Last Recess
1 2 3
1186.4 1162.8 1126.5
1176.6 1134.0 1080.9
1196.3 1191.6 1172.1
1206.6 1189.7 1201.6
1194.8 1170.3 1179.8
1218.4 1209.0 1223.5
When CCD was significant, estimated marginal means were used. Values are based on back-transformed variables where necessary.
their females may lay up to 2 weeks later (I. Ritchie, pers. comm.). Whatever the proximate cause for a delay in clutch initiation, the timing of reproduction (i.e., CCD) was a major factor explaining treatment effects of PCBs.
Most behaviors thus appeared to be affected indirectly by PCBs. Two behaviors, Time of Last Recess and Recesses Length, were directly affected by PCBs (i.e., treatment
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
232
26
Number of recesses
24 22 20 18 16 14 12
N = 19
23 1
(a)
21
23 2 Observation period
21
21 3
Number of nesst switches
40
30
20
10
0
N=
18
22 1
(b)
20
20 2 Observation period
20
20 3
Fig. 1. (a) Mean (71 SE) number of recesses and (b) number of nest switches that occurred in captive American kestrel pairs in early (1), middle (2), and late (3) observation periods for PCB-exposed (PCB) and control groups (CTL) (solid circles, CTL; open circles, PCB).
effect remained after controlling for CCD). Increase in the length or frequency of recesses have been reported for other birds exposed to PCBs (Larus argentatus: Fox et al., 1978; L. hyperboreus: Bustnes et al., 2001; Streptopelia risoria: Peakall and Peakall, 1973; McArthur et al., 1983). Regardless of whether the effects of PCBs were direct or indirect, there were marked differences between treatment groups. All of the seven variables associated with treatment also were shown to have some consequence on hatching success. Each behavior varied in a direction consistent with our initial prediction that pairs with greater attendance to their eggs and better coordination of their incubation duties should have better hatching success (Table 4). For example, exposed males with complete hatching failure were in the nest less than half the time of contaminated males with high hatching success (Table 4). Frequent or long recesses have been considered causal factors of low hatchability in other toxicology studies (Peakall and
Peakall, 1973; Fox et al., 1978; McArthur et al., 1983). Similarly, our results for nest switches suggest a need for continuous care of eggs. Pairs that were well coordinated in having more nest switches experienced the best hatching success (Table 4). Koenig (1982) also found that behavioral synchrony between the sexes was essential for successful hatching. Factors intrinsic to the egg may have acted synergistically with parental behavior to result in the lower hatching rate of PCB-exposed kestrels (Kubiak et al., 1989; Fernie et al., 2001a). Kestrels that completed clutches later in the breeding season laid eggs with higher PCB burdens. In many species of wild birds, pairs that complete clutches later in the season generally have smaller clutches composed of eggs with higher levels of PCBs (Becker et al., 1993; Custer et al., 1998). PCB exposure lengthened incubation period by almost 1 full day. Extended incubation was also found in PCBexposed doves (Streptopelia risoria; McArthur et al., 1983) and Forster’s terns (Sterna forsteri; Kubiak et al., 1989). A longer incubation may have various fitness costs to wild birds as there is more time for eggs to be vulnerable to predation, and there may be less time for fledglings to accumulate skills prior to migration (Webb, 1987). Detrimental developmental effects on the embryo have resulted from a lengthened incubation period caused by longer or more frequent recesses (Haftorn, 1988; Ancel et al., 1995; Lindstro¨m, 1999; Jacobs and Wingfield, 2000). Inconsistent incubation may also affect embryonic metabolism, immunocompetence, and even sexual attractiveness in adulthood (Lindstro¨m, 1999). Events during avian embryonic development have future consequences on adult fecundity (Lindstro¨m, 1999). PCB exposure of these kestrels also affected size at hatching and postnatal growth (Fernie et al., 2003). Our results are significant because they demonstrate an association between behavioral differences that result from toxicological contamination and decreased reproductive success. To conclusively determine whether factors extrinsic to the egg alone affect hatching success, nest-exchange experiments should be performed. Such studies of PCBexposed birds have illustrated that increased hatchability occurred when PCB-exposed eggs were removed from contaminated adults and incubated artificially or by uncontaminated pairs (doves: Peakall and Peakall, 1973; Forster’s terns: Kubiak et al., 1989). It is likely that our results are conservative as captivity is a relatively benign environment. Birds in the wild have further constraints imposed on them, including the need to forage for food which could induce more behavioral variation and negative consequences. Our identification of specific behavioral anomalies paves the way for more targeted studies in the future to elucidate related physiological mechanisms, especially endocrine modulation. Since prolactin regulates kestrel incubation behavior (Sockman et al., 2000), this hormone should be of particular interest.
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
233
Table 4 Means and 95% lower (LCL) and upper (UCL) confidence limits for incubation behaviors that varied with hatching success in observation periods (PD) 1, 2, or 3 of PCB-exposed American kestrels X¯
Variable
PD
Hatching success
LCL
UCL
Percent In
2
0 1 2
89.9 93.8 92.9
87.4 92.1 90.4
92.1 95.4 95.0
Percent In
3
0 1 2
91.3 94.7 94.0
88.6 92.8 91.3
93.7 96.4 96.2
Percent Male In
3
0 1 2
5.7 13.0 13.2
3.0 10.8 12.0
13.2 18.9 22.5
Recess Length
2
0 1 2
4.0 3.0 4.4
3.1 2.3 3.1
5.3 4.2 6.4
No. Male Bouts
1
0 1 2
5.3 10.9 11.2
3.0 8.2 7.6
8.2 14.1 15.5
No. Recesses
3
0 1 2
22.9 18.8 11.7
18.8 14.6 8.0
27.4 23.4 16.2
No. Nest Switches
1
0 1 2
6.8 12.3 16.4
2.7 6.2 8.1
12.7 20.5 27.6
No. Nest Switches
3
0 1 2
5.6 16.4 21.0
1.2 7.3 8.9
13.2 29.2 38.2
Last Recess
1
0 1 2
1209.5 1216.2 1188.0
1192.2 1196.8 1164.7
1226.9 1235.8 1211.2
Last Recess
3
0 1 2
1224.5 1200.1 1162.8
1194.2 1165.7 1122.1
1254.8 1234.5 1203.4
Hatching Success: 0 (zero ¼ 0%), 1 (moderate ¼ 1–74%), 2 (high ¼ 75–100%). Values are based on back-transformed variables where necessary.
Acknowledgments We thank J. Huff for developing the electronic balance, I. Ritchie for management of the kestrel facility, E. Woodsworth for aid with computer programming, J. Willson for assistance, and K. Wiebe for comments on the manuscript. S.A.F. thanks the University of Saskatchewan for financial assistance through a graduate teaching fellowship. This study was funded by grants from NSERC (G.R.B., J.E.S.) and the Canadian Network of Toxicology Centres (G.R.B., J.E.S.).
References Ancel, A., Liess, S., Girard, H., 1995. Embryonic development of the domestic Guinea fowl (Numida meleagris). J. Zool. London 235, 621–634.
Balgooyen, T.G., 1976. Behavior and ecology of the American kestrel, Falco sparverius, in the Sierra Nevada of California. University of California Publications Zoology, vol. 103. University of California Press, California. Barron, M.G., Galbraith, H., Beltman, D., 1995. Comparative reproductive and developmental toxicology of PCBs in birds. Comp. Biochem. Physiol. 112C, 1–14. Becker, P.H., Schuhmann, S., Koepff, C., 1993. Hatching failure in common terns (Sterna hirundo) in relation to environmental chemicals. Environ. Pollut. 79, 207–213. Bird, D.M., 1988. American kestrel. In: Palmer, R.S. (Ed.), Handbook of North American Birds, vol. 5, Part 2. Yale University Press, New Haven, pp. 253–290. Bortolotti, G.R., Wiebe, K.L., 1993. Incubation behaviour and hatching patterns in the American kestrel Falco sparverius. Ornis Scand. 24, 41–47. Braune, B.M., Norstrom, R.J., 1989. Dynamics of organochlorine compounds in herring gulls: III. Tissue distribution and bioaccumulation in Lake Ontario gulls. Environ. Toxicol. Chem. 8, 957–968. Bustnes, J.O., Bakken, V., Erikstad, K.E., Mehlum, F., Skaare, J.U., 2001. Patterns of incubation and nest-site attentiveness in relation to
ARTICLE IN PRESS 234
S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
organochlorine (PCB) contamination in glaucous gulls. J. Appl. Ecol. 38, 791–801. Clark, K.E., Niles, L.J., Stansley, W., 1998. Environmental contaminants associated with reproductive failure in bald eagle (Haliaeetus leucocephalus) eggs in New Jersey. Bull. Environ. Contam. Toxicol. 62, 247–254. Clotfelter, E.D., Bell, A.M., Levering, K.R., 2004. The role of animal behaviour in the study of endocrine-disrupting chemicals. Anim. Behav. 68, 665–676. Custer, C.M., Custer, T.W., Allen, P.D., Stromborg, K.L., Melancon, M.J., 1998. Reproduction and environmental contamination in tree swallows nesting in the Fox River drainage and Green Bay, Wisconsin, USA. Environ. Toxicol. Chem. 17, 1786–1798. Doving, K.B., 1991. Assessment of animal behaviour as a method to indicate environmental toxicity. Comp. Biochem. Physiol. 100, 247–252. Drent, R., 1973. The natural history of incubation. In: Farner, D.S. (Ed.), Breeding Biology of Birds. National Academy of Sciences, Washington, pp. 262–332. Drouillard, K.G., Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., Norstrom, R.J., 2001. Bioaccumulation and toxicokinetics of 42 PCB congeners in American kestrels (Falco sparverius). Environ. Toxicol. Chem. 20, 2514–2522. Elliott, J.E., Kennedy, S.W., Jeffrey, D., Shutt, L., 1991. Polychlorinated biphenyl (PCB) effects on hepatic mixed function oxidates and porphyria in birds—II. American kestrel. Comp. Biochem. Physiol. 99C, 141–145. Evangelista de Duffard, A.M., Duffard, R., 1996. Behavioural toxicology, risk assessment, and chlorinated hydrocarbons. Environ. Health Perspect. 104 (Suppl. 2), 353–360. Fernie, K.J., Bortolotti, G.R., Smits, J.E., Willson, J., Drouillard, K.G., Bird, D.M., 2000. Changes in egg composition of American kestrels exposed to dietary polychlorinated biphenyls. J. Toxicol. Environ. Health 60, 101–113. Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., 2001a. Reproductive success of American kestrels exposed to dietary polychlorinated biphenyls. Environ. Toxicol. Chem. 20, 776–781. Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., 2001b. In ovo exposure to polychlorinated biphenyls: reproductive effects on secondgeneration American kestrels. Arch. Environ. Contam. Toxicol. 40, 544–550. Fernie, K.J., Bortolotti, G.R., Smits, J.E., 2003. Reproductive abnormalities, teratogenicity, and developmental problems in American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. J. Toxicol. Environ. Health A 66, 2089–2103. Fisher, S.A., Bortolotti, G.R., Fernie, K.J., Smits, J.E., Marchant, T.A., Drouillard, K.G., Bird, D.M., 2001. Courtship behaviour of captive American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. Arch. Environ. Contam. Toxicol. 41, 215–220. Fox, G.A., Gilman, A.P., Peakall, D.B., Anderka, F.W., 1978. Behavioral abnormalities of nesting Lake Ontario herring gulls. J. Wildlife Manage. 42, 477–483. Fry, D.M., 1995. Reproductive effects in birds exposed to pesticides and industrial chemicals. Environ. Health Perspect. 103 (Suppl. 7), 165–171. Fyfe, R.W., Risebrough, R.W., Walker, IIW., 1976. Pollutant effects on the reproduction of the prairie falcons and merlins of the Canadian prairies. Can. Field Nat. 90, 346–355. Haftorn, S., 1988. Incubating female passerine do not let the egg temperature fall below the ‘physiological zero temperature’ during their absences from the nest. Ornis Scand. 19, 97–110. Harris, H.J., Erdman, T.C., Ankley, G.T., Lodge, K.B., 1993. Measures of reproductive success and polychlorinated biphenyl residues in eggs and chicks of Forster’s terns on Green Bay, Lake Michigan, Wisconsin—1988. Arch. Environ. Contam. Toxicol. 25, 304–314. Hoffman, D.J., Rice, C.P., Kubiac, T.J., 1996a. PCBs and dioxins in birds. In: Beyer, W.N., Heinz, G.H., Redmon, A.W. (Eds.), Environmental
Contaminants in Wildlife: Interpreting Tissue Concentrations. SETAC Publication, Lewis Publishers, pp. 165–207. Hoffman, D.J., Melancon, M.J., Klein, P.N., Rice, C.P., Eisemann, J.D., Hines, R.K., Spann, J.W., Pendleton, G.W., 1996b. Developmental toxicity of PCB 126 (3,30 ,4,40 ,5-pentachlorobiphenyl) in nestling American kestrels (Falco sparverius). Fundam. Appl. Toxicol. 34, 188–200. Jacobs, J.D., Wingfield, J.C., 2000. Endocrine control of life-cycle stages: a constraint on response to the environment. Condor 102, 35–51. Koenig, W.D., 1982. Ecological and social factors affecting hatchability of eggs. Auk 99, 526–536. Kozie, K.D., Anderson, R.K., 1991. Productivity, diet and environmental contaminants in bald eagles nesting near the Wisconsin shoreline of Lake Superior. Arch. Contam. Toxicol. Chem. 20, 41–48. Kubiak, T.J., Harris, H.J., Smith, L.M., Schwartz, T.R., Stalling, D.L., Trick, J.A., Sileo, L., Docherty, D.E., Erdman, T.C., 1989. Microcontaminants and reproductive impairment of the Forster’s tern on Green Bay, Lake Michigan—1983. Arch. Environ. Contam. Toxicol. 18, 706–727. Lindstro¨m, J., 1999. Early development and fitness in birds and mammals. Trends Ecol. Evol. 14, 343–348. McArthur, M.L.B., Fox, G.A., Peakall, D.B., Philoge`ne, B.J.R., 1983. Ecological significance of behavioural and hormonal abnormalities in breeding ring doves fed an organochlorine chemical mixture. Arch. Environ. Contam. Toxicol. 12, 343–353. McCarty, J.P., Secord, A.L., 1999a. Nest-building behavior in PCBcontaminated tree swallows. Auk 116, 55–63. McCarty, J.P., Secord, A.L., 1999b. Reproductive ecology of tree swallows (Tachycineta bicolor) with high levels of polychlorinated biphenyl contamination. Environ. Toxicol. Chem. 18, 1433–1439. Milstein, P.le S., Prestt, I., Bell, A.A., 1970. The breeding cycle of the grey heron. Ardea 58, 171–257. Nelson, R.W., 1976. Behavioural aspects of egg breakage in peregrine falcons. Can. Field Nat. 90, 320–329. Noru˘sis, M., 1991. The SPSS Guide to Data Analyses for SPSS/PC+. Prentice-Hall, Englewood Cliffs, NJ. Peakall, D.B., Peakall, M.L., 1973. Effect of a polychlorinated biphenyl on the reproduction of artificially and naturally incubated dove eggs. J. Appl. Ecol. 10, 863–868. Porter, R.D., Wiemeyer, S.N., 1972. Reproductive patterns in captive American kestrels (sparrow hawk). Condor 74, 46–53. Rehder, N.B., Bird, D.M., Lague, P.C., 1986. Variations in plasma corticosterone, esterone, estradiol-17b, and progesterone concentrations with forced renesting, molt, and body weight of captive female American kestrels. Gen. Comp. Endocrinol. 62, 386–393. Risebrough, R.W., Anderson, D.W., 1975. Some effects of DDE and PCB on mallards and their eggs. J. Wildlife Manage. 39, 508–513. Silver, R., Ball, G.F., 1989. Brain, hormone and behavior interactions in avian reproduction: status and prospectus. Condor 91, 966–978. Smits, J.E.G., Bortolotti, G.R., 2001. Antibody-mediated immunotoxicity in American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. J. Toxicol. Environ. Health A 62, 217–226. Smits, J.E.G., Fernie, K.J., Bortolotti, G.R., Marchant, T.A., 2002. Thyroid hormone suppression and cell mediated immunomodulation in American kestrels (Falco sparverius) exposed to PCBs. Arch. Environ. Contam. Toxicol. 43, 338–344. Sockman, K.W., Schwabl, H., Sharp, P.J., 2000. The role of prolactin in the regulation of clutch size and onset of incubation behaviour in the American kestrel. Horm. Behav. 38, 168–176. Valkama, J., Korpimaki, E., 1999. Nest box characteristics, habitat quality and reproductive success of Eurasian kestrels. Bird Study 46, 81–88. Van Den Berg, M., Craane, L.H.J., Van Mourik, S., Brouwer, A., 1995. The (possible) impact of chlorinated dioxin (PCDDs), dibenzofurans (PCDFs), and biphenyls (PCBs) on the reproduction of the cormorant Phalacrocorac carbo—an ecotoxicological approach. Ardea 83, 299–313.
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235 Vos, J.G., Dybing, E., Greim, H.A., Ladefoged, O., Lambre´, C., Tarazona, J.V., Brandt, I., Vethaak, A.D., 2000. Health effects of endocrine-disrupting chemicals on wildlife, with special reference to the European situation. Crit. Rev. Toxicol. 30, 71–133. Webb, D.R., 1987. Thermal tolerance of avian embryos: a review. Condor 89, 874–898.
235
Wilmers, T.J., Bowman, R., Samuel, D.E., 1985. Notes on Incubation by male kestrels in western Virginia, Pennsylvania, and southern Quebec. N. Am. Bird Bander 10, 6–8. Zala, S.M., Penn, D.J., 2004. Abnormal behaviours induced by chemical pollution: a review of the evidence and new challenges. Anim. Behav. 68, 649–664.
Ecotoxicology and Environmental Safety 63 (2006) 226–235 www.elsevier.com/locate/ecoenv
Behavioral variation and its consequences during incubation for American kestrels exposed to polychlorinated biphenyls Sheri A. Fishera, Gary R. Bortolottia,, Kimberley J. Fernieb, David M. Birdc, Judit E. Smitsd a
Department of Biology, University of Saskatchewan, 112 Science Pl., Saskatoon, Sask., Canada S7N 5E2 b Environment Canada, 867 Lakeshore Road, Burlington, Ont., Canada L7R 4A6 c Avian Science and Conservation Centre, McGill University, Ste. Anne de Bellevue, Quebec, Canada H9X 3V9 d Department of Veterinary Pathology, University of Saskatchewan, Saskatoon, Sask., Canada S7N 5B4 Received 12 May 2005; received in revised form 11 July 2005; accepted 25 July 2005 Available online 5 October 2005
Abstract We investigated whether polychlorinated biphenyl (PCB) exposure in American kestrels (Falco sparverius) influenced incubation behavior and whether altered behavior could lead to poor reproductive success. Captive kestrels were fed a mixture of PCBs (Aroclors 1248:1254:1260) at an approximate daily dose of 7 mg/kg body weight, 1 month prior to pairing and throughout incubation. Behaviors of 23 control and 23 PCB-exposed pairs were monitored throughout incubation using an electronic balance in the nest box. PCB exposure resulted in longer incubation periods and in altered incubation behaviors. Seven of 14 behavioral variables showed some association with treatment, with sex-specific effects largely biased toward disrupted male behavior. For most behaviors, the treatment effect was explained by the delayed clutch initiation induced by PCBs rather than by a direct physiological impact of the contaminants. PCB-exposed pairs with greater attendance to their eggs and better coordination of incubation duties had improved hatching success. r 2005 Elsevier Inc. All rights reserved. Keywords: American kestrel; Falco sparverius; Raptor; Incubation behavior; Polychlorinated biphenyls; PCBs; Organochlorine contaminants; Hatching success
1. Introduction Polychlorinated biphenyls (PCBs) are persistent, widespread environmental contaminants. As PCBs bioaccumulate, birds at higher trophic levels of the food chain are particularly vulnerable to PCB exposure (Elliott et al., 1991; Barron et al., 1995; Hoffman et al., 1996a). Historical population declines of raptors have been attributed to contaminant-induced reproductive failure (e.g., Kozie and Anderson, 1991; Clark et al., 1998; Valkama and Korpimaki, 1999). The breeding cycle and associated behaviors that are under hormonal control may be vulnerable to endocrine-modulating substances such as PCBs (Barron et al., 1995; Vos et al., 2000), especially since slight hormonal differences can elicit significant behavioral change (Silver and Ball, 1989; Jacobs and Wingfield, 2000). Behavioral Corresponding author. Fax: +1 306 966 4461.
E-mail address: [email protected] (G.R. Bortolotti). 0147-6513/$ - see front matter r 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2005.07.021
observations rather than intrusive physiological examinations are increasingly being used and encouraged in toxicological research to identify adverse effects of contaminants (Doving, 1991; Evangelista de Duffard and Duffard, 1996; Clotfelter et al., 2004; Zala and Penn, 2004). Altered behavior during the breeding season has been observed in birds exposed to organochlorine contaminants, including PCBs. Reduced nest defence behavior occurred in merlins (Falco columbarius) (Fyfe et al., 1976). Peregrine falcons (F. peregrinus) exposed to dichlorodiphenyldichloroethane displayed clumsy behavior in the nesting scrape, including stepping on the eggs (Nelson, 1976). Eggdestroying behavior occurred in contaminated gray herons (Ardea cinerea) (Milstein et al., 1970) and captive mallards (Anas platyrhynchos) (Risebrough and Anderson, 1975). McCarty and Secord (1999a) found that tree swallows (Tachycineta bicolor) in PCB-contaminated areas built lowquality nests, which were associated with reduced breeding success. Atypical parental behavior has also been suggested
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
to contribute to decreased reproductive success in cormorants (Phalacrocorax carbo) contaminated with PCBs, chlorinated dioxins, and dibenzofurans (Van Den Berg et al., 1995). Incubation behavior can be disrupted by sublethal exposure to environmental pollutants (Fry, 1995), including a lack of behavioral synchrony between the sexes which normally ensures that eggs are not left unattended (Koenig, 1982). Unattended nests are exposed to greater risk from predation, and embryos could die if exposed to a suboptimal thermal environment (Webb, 1987). Because the cooling rate of an unattended egg exceeds the heating rate when parents return, absences greater than about 1 min are also energetically wasteful to the parent (Drent, 1973). Exposure of parent birds to PCBs has been linked to lengthened and more frequent recesses, prolonged incubation, and increased incidence of nest abandonment (Peakall and Peakall, 1973; Fox et al., 1978; McArthur et al., 1983; Kubiak et al., 1989; McCarty and Secord, 1999b; Bustnes et al., 2001). Our study subject, the American kestrel (Falco sparverius), is a biparental incubator (Balgooyen, 1976; Bortolotti and Wiebe, 1993). We were able to monitor breeding pairs for continuous periods of time throughout incubation, whereas many previous studies of contaminated birds have used infrequent nest checks or monitoring for only short segments (Peakall and Peakall, 1973; Kubiak et al., 1989; Harris et al., 1993; Custer et al., 1998; Bustnes et al., 2001). In addition, and unlike some studies, we could easily differentiate males from females, enabling us to reliably investigate sex-specific effects. Complete data on reproductive performance also allowed us to investigate the consequences of variation in incubation. We predict that pairs with greater attendance to their eggs, and better coordination of their incubation duties, should have greater hatching success. 2. Materials and methods 2.1. Study animals This study took place at the Avian Science and Conservation Centre of McGill University, Canada, using American kestrels of known pedigree and age. Birds were randomly designated control (CTL) (n ¼ 25 pairs) or PCBexposed (PCB) (n ¼ 25 pairs) and were placed in flight pens (6 6 2.5 m) segregated by sex and treatment. The kestrels were fed ad libitum on their typical diet of dayold cockerels. Based on the PCB congeners found in the eggs and tissues of wild birds from the Great Lakes region (Braune and Norstrom, 1989, Clark et al., 1998) and PCB residue levels found in wild prey species of kestrels (Environment Canada, unpublished data), a dosing regime that would generate environmentally relevant levels of PCBs was calculated (Fernie et al., 2000). Chronic dietary exposure began 1 month before pairing on 18 March 1998 and
227
continued until the end of the incubation period of each pair (range 79–117 days of exposure; mean 95 days). A mixture of Aroclors 1248:1254:1260 (1:1:1 by weight; Monsanto, St. Louis, MO, USA) was dissolved in safflower oil at a concentration of 4.85 mg/g total PCB. As kestrels show a preference for consuming the heads of cockerels (I. Ritchie, pers. comm.), 100-ml aliquots of either PCB dosing mixture or plain safflower oil were injected into the heads of frozen–thawed day-old cockerels to be fed to the PCBexposed or control groups, respectively. The birds consumed approximately 7 mg/kg body weight/day (Fernie et al., 2000; Drouillard et al., 2001). The total PCB residue levels from the kestrel eggs averaged 34.1 mg/g (geometric mean) on a whole-egg wet-weight basis (Fernie et al., 2000). These levels are comparable to those found in the eggs of wild raptors exhibiting decreased reproductive success (Hoffman et al., 1996b; Clark et al., 1998; Valkama and Korpimaki, 1999). Genetically unrelated birds, all with previous breeding experience, were paired on 21 April 1998. The pairs were placed into outdoor breeding pens (2.3 0.9 3.6 m), which contained rope and wooden perches, a one-way glass window for observation, and a nesting box. 2.2. Monitoring of incubation behavior The incubation period was defined as the time from the completion of the clutch to the hatching of the first egg or until 28 days after clutch completion for pairs that failed to hatch any egg. A custom-built electronic balance system was used to quantify behavior. The monitoring system consisted of a wooden incubation box (20.3 17.8 10.2 cm) mounted onto a balance. Electrical impulses proportional to the mass of the incubation box were sent from the balance to a microchip once every minute, 24 h per day. Twenty electronic balances were randomly assigned to PCB-exposed or control pens. Every 5 days during the incubation period balances were switched between pens of the same treatment group to increase sample size (PCB n ¼ 23; CTL n ¼ 23). One PCB-exposed pair did not complete a clutch and was not monitored, while another PCB-exposed pair abandoned their clutch. Two control pens were not monitored to spread the effort equally between treatments. When pairs were not being electronically monitored, a wooden incubation box with the same dimensions was placed in the nest box. The bottom of the incubation box was filled with 2–4 cm of wood shavings. A door on the back of the nest box facilitated nest checks and removal of balances. Nest boxes were checked daily, and eggs were numbered to identify laying sequence. Ten days after clutch completion, eggs were candled to determine fertility and one egg was removed from each pair for PCB residue analysis (see Fernie et al., 2000). Another fresh but nonviable kestrel egg was put into the box to maintain the original clutch size. Female kestrels were heavier than their mates, allowing us to use mass to determine the sex of the bird in the nest
ARTICLE IN PRESS 228
S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
box. In pairs where the masses of both sexes were similar (two PCB and one control pair), data on only the presence or absence of a bird were used. While the scales detected the presence of the parent, the adult may not have been incubating the entire time that it was in the nest box. To confirm that the balances could accurately detect the presence/absence and sex of a parent, visual observations were conducted at 08:00, 10:00, 13:00, 15:00, 17:00, and 20:00 h each day by looking through the one-way glass window. In total, 3792 visual scans were performed, and 3743 of them concurred with the balance (99% scale accuracy). The few discrepancies were not necessarily errors as the birds may have flushed or incubation switches could have occurred after the observer completed the scan and before the scale reading was taken. 2.3. Behavioral data analysis To determine the percentage of the day that the nest was occupied (PCB n ¼ 23; CTL n ¼ 23), the number of minutes of nest occupation was divided by day length. We refer to day length as the time between the end of the overnight incubation bout and the start of the overnight incubation bout. Females were responsible for overnight incubation (see also Bortolotti and Wiebe, 1993). The variables chosen describe the role of each sex during incubation plus the combined effort of the parents (Table 1). Perhaps to ensure the continuity of incubation and thus not leave eggs exposed, kestrels performed rapid Nest Switches. Nest Switches occurred either when the bird in the box exited and was replaced within the same minute by its mate or when the mate entered the nest and the incubating bird left. The number of Nest Switches and Time of Last Recess indicate the coordination of the sexes (see below). The incubation period of kestrels is typically 28–29 days but can range from 26 to 31 days (Bird, 1988). The incubation period was divided into early, middle, and late (1–8, 9–16, 17+ days since clutch completion) observation
periods because behavior and hormones have seasonal patterns of variation (Rehder et al., 1986; Silver and Ball, 1989). Behavioral data for each pen were averaged within each of the three periods. Sample sizes for each period vary slightly as not all pens were observed each time. The average numbers of observation days for controls were 4.7, 4.6, and 5.8 days, and those for PCB-exposed birds were 6.1, 5.1, and 7.6 days for early to late periods, respectively. For repeated-measures analyses (across observation periods), the total sample sizes were: PCB n ¼ 21 and CTL n ¼ 19. Variables were transformed when they did not meet the assumptions for statistical tests (Table 1). PCB-exposed kestrels had later clutch completion dates (CCD) compared to controls (PCB: mean Julian date ¼ 141, SE ¼ 1.76, n ¼ 24; CTL: mean Julian date 134.8, SE 1.26, n ¼ 25; t ¼ 2.93, df ¼ 47, P ¼ 0.005). If some behaviors varied seasonally, then PCB and control groups could differ in behavior by virtue of when eggs were laid (i.e., CCD). In such a case, an analysis of variance (ANOVA) would reveal a significant treatment effect, but the result would be incorrectly interpreted as an effect of PCBs on the behavior per se. We therefore used analysis of covariance (ANCOVA) with CCD as a covariate. In the case where treatment is no longer significant after controlling for CCD, we refer to PCBs as having an ‘‘indirect’’ effect on behavior. While it is still important to recognize that the treatments differed, these differences in behavior would be explained by how early in the breeding season that the birds bred. We considered PCBs to have had a ‘‘direct’’ effect on behavior when treatment remained significant in the ANCOVA. Therefore, we performed two statistical tests on the same data set: repeated-measures ANOVA and ANCOVA. Few studies have made a distinction between such indirect and direct effects. 2.4. Hatching success To determine the consequences of behavioral variation due to contamination, variables that were directly and
Table 1 Definitions of behavioral variables and type of transformation used for statistical analysis Variable
Definition
Transformation
Percent In Percent Female In Percent Male In No. Female Bouts No. Male Bouts No. Recesses No. Switches Female Bout Length Male Bout Length Recess Length First Switch Last Switch First Recess Last Recess
No. min nest occupied /day length 100 No. min female occupied nest/day length 100 No. min male occupied nest/day length 100 No. of times female on nest No. of times male on nest No. of times nest was unoccupied No. of times one parent relieved the incubating bird No. min female occupied nest/No. Female Bouts No. min male occupied nest/No. Male Bouts No. min nest unoccupied/No. Recesses Time of day of the first switch Time of day of the last switch Time of day of the first recess Time of day of the last recess
Arcsine Arcsine Arcsine Square root Square root Square root Square root Square root Square root Log 10 None None None None
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
indirectly associated with PCB exposure were examined in relation to hatching success. The typical hatching rate of captive and wild populations of kestrels varies from 66% to almost 100% (see review by Bird, 1988 and G.R. Bortolotti, unpublished data). We categorized success as zero (0%), moderate (1–74%), and high (75–100%). In other words, allowing for the small variation in clutch size (four or five eggs), we compared pairs that failed to hatch all, two or three, or at most one egg. Because embryos vary in their sensitivity to the thermal environment depending on their stage of development (Webb, 1987), analyses were performed for each observation period. 2.5. Length of incubation period Inefficient incubation behavior may have negative consequences in addition to mortality to developing embryos. Eggs given inadequate heating may take longer to hatch; therefore, we examined the number of days for each viable egg to hatch from the date that the clutch was completed (PCB n ¼ 29, CTL n ¼ 44). Because CCD was defined as the start of incubation, that date was used to determine the length of the incubation period. Kestrels partially incubate eggs before the clutch is complete, resulting in hatching asynchrony, so the position of an egg in the laying sequence influences its incubation period (Bortolotti and Wiebe, 1993). Therefore, egg sequence and treatment were both considered factors in an ANOVA. For analysis of consequences of behavior we began with full ANOVA or ANCOVA models that included all effects and interactions and then repeated the analyses after iteratively removing nonsignificant interactions or factors. All statistical analyses were performed using SPSS software (Noru˘sis, 1991). 3. Results 3.1. Incubation behavior Exposure to PCBs produced detectable changes in frequency, length, and timing of behaviors. In addition to statistical results (Table 2) we present behavioral data (Table 3) to allow for comparison to other studies on wild or captive birds. Seven of 14 behavioral variables showed some association with treatment. In most cases, PCBs had an indirect effect as the ANCOVA showed that the treatment difference detected by ANOVA could be accounted for date of clutch completion. For the Number of Recesses and Number of Switches there was an interaction between treatment and observation period. There were no significant differences between the Number of Recesses taken by control and PCB pairs during periods 1 and 2 (P40:61); however, PCB pairs took more recesses during period 3 (Fig. 1a; t ¼ 2:18, df ¼ 38, P ¼ 0.036). Similarly, the Number of Switches did not differ between groups in periods 1 or 2 (P40.26), but PCB-exposed kestrels had fewer switches in period 3 (Fig. 1b; t ¼ 2.5,
229
Table 2 Results of repeated-measures ANOVA and repeated-measures ANCOVA for all incubation variables across early, middle, and late observation periods of PCB-exposed or control American kestrels Variable
Percent In Observation periods Treatment CCD Observation periods Observation periods Percent Female In Observation periods Treatment CCD Observation periods Observation periods Percent Male In Observation periods Treatment CCD Observation periods Observation periods Female Bout Length Observation periods Treatment CCD Observation periods Observation periods Male Bout Length Observation periods Treatment CCD Observation periods Observation periods Recess Length Observation periods Treatment CCD Observation periods Observation periods No. Female Bouts Observation periods Treatment CCD Observation periods Observation periods No. Male Bouts Observation periods Treatment CCD Observation periods Observation periods No. Recesses Observation periods Treatment CCD Observation periods Observation periods
ANOVA
ANCOVA
P
P
Treatment CCD
0.000 0.012 — 0.220 —
0.225 0.123 0.036 0.646 0.264
Treatment CCD
0.005 0.332 — 0.225 —
0.625 0.544 0.412 0.320 0.578
Treatment CCD
0.075 0.034 — 0.690 —
0.283 0.158 0.068 0.821 0.306
Treatment CCD
0.121 0.732 — 0.809 —
0.525 0.559 0.461 0.604 0.506
treatment CCD
0.497 0.129 — 0.390 —
0.890 0.479 0.031 0.365 0.882
treatment CCD
0.000 0.009 — 0.022 —
0.202 0.078 0.086 0.028 0.295
Treatment CCD
0.705 0.912 — 0.985 —
0.972 0.624 0.288 0.995 0.981
Treatment CCD
0.204 0.027 — 0.268 —
0.505 0.156 0.029 0.339 0.499
Treatment CCD
0.000 0.345 — 0.020 —
0.056 0.570 0.499 0.220 0.087
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
230 Table 2 (continued ) Variable
ANOVA
ANCOVA
P
P
Treatment CCD
0.018 0.085 — 0.082 —
0.017 0.219 0.273 0.381 0.021
First Switch Observation Periods Treatment CCD Observation periods Treatment Observation Periods CCD
0.012 0.214 — 0.350 —
0.351 0.730 0.016 0.693 0.332
Last Switch Observation periods Treatment CCD Observation periods Observation periods
Treatment CCD
0.044 0.195 — 0.988 —
0.173 0.496 0.100 0.750 0.209
treatment CCD
0.003 0.104 — 0.595 —
0.503 0.467 0.057 0.644 0.552
treatment CCD
0.009 0.002 — 0.019 —
0.212 0.018 0.139 0.155 0.258
No. Switches Observation periods Treatment CCD Observation periods Observation periods
First Recess Observation periods Treatment CCD Observation periods Observation periods Last Recess Observation periods Treatment CCD Observation periods Observation periods
Length, Number of Recesses) and better coordination of their incubation duties (Time Last Recess, Number of Nest Switches) should have better hatching success. 3.3. Length of incubation period
The date of clutch completion (CCD) was used as a covariate in ANCOVA models.
df ¼ 36, P ¼ 0.016). Both of these behaviors varied in a strong seasonal pattern in control birds but not at all in exposed birds (Fig. 1). Three other behaviors were indirectly affected by PCBs regardless of observation period; Percent In, Percent Male In, and Number of Male Bouts were all lower in PCB-exposed kestrels. For variables directly affected by PCBs, the timing of Last Recess was later for PCB-exposed kestrels, and the effect on Recess Length was contingent on the period; it was longer in contaminated birds during periods 1 and 3 (PD1: t ¼ 3.8, df ¼ 38, Po0.001; PD3: t ¼ 1.98, df ¼ 38, P ¼ 0.055) but not period 2 (t ¼ 1.1, df ¼ 38, P ¼ 0.29). 3.2. Hatching success For simplicity, Table 4 presents only significant (P40.05) associations between behavior and hatching success categories per observation period. The results, all of which are for exposed birds, are all consistent with our prediction that pairs with greater attendance to their eggs (Percent In, Percent Male In, Number Male Bouts, Recess
Egg number in the laying sequence affected time from clutch completion to hatching (P ¼ 0.002) as expected, given hatching asynchrony. PCB exposure also increased the hatching time of each egg (P ¼ 0.026), with the mean incubation period for PCB-exposed birds being 26.9 days (SE ¼ 0.24, n ¼ 29) vs. 26.1 days for CTL pairs (SE ¼ 0.25, n ¼ 44). 4. Discussion Unlike most behavioral studies which provide snapshots of a few specific incubation behaviors, we were able to monitor multiple behaviors for half the actual time spent incubating across the entire incubation period. CTL kestrels spent more time in the nest box than PCB-exposed birds, with the treatment difference being primarily the result of reduced male contribution (Tables 2 and 3). Within studies of both captive and wild kestrels, the extent to which males incubate varies from 0% to 60% (Porter and Wiemeyer, 1972; Balgooyen, 1976; Wilmers et al., 1985). Control males contributed between 1.2% and 40.0% of daily incubation (16.8% overall), and PCB-exposed males contributed between 0.1% and 30.2%, with an overall average of 10.7%. It appeared that PCB females may have compensated slightly for the reduction in incubation by their mates (Table 3) but not enough to attain the same overall nest attendance (i.e., Percent In). The only behavior that may suggest an effect of PCBs on females was Time of Last Recess, as females were responsible for all overnight incubation. The time that the final recess began was later in PCB-exposed pairs; however, the problem may lie with the male for not performing a nest switch. The Number of Nest Switches occurred less often in PCB-exposed pairs. Fewer nest switches indicate poor parental coordination, which is vital to ensure that eggs are not unattended. Highly male-biased effects of PCBs have also been found in our study population for several physiological parameters (Smits and Bortolotti, 2001; Smits et al., 2002) and reproduction (Fernie et al., 2001 a, b). Furthermore, PCB-exposed males, but not females, exhibited altered sexual and flight behaviors during courtship (Fisher et al., 2001). Altered courtship may even explain the delay in clutch initiation that had such an overwhelming subsequent effect on behavior in this study. While the delay in breeding may be directly induced by PCBs, it is also possible that exposed males did not adequately stimulate their mates’ endocrine systems. The latter is plausible given that in this colony it has been found that young males breeding for the first time do not court as well as experienced males and as a result
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
231
Table 3 Means and 95% lower (LCL) and upper (UCL) confidence limits for all incubation behavioral variables for each treatment and observation period (PD) Variable
PD
Treatment Control X¯
PCB LCL
UCL
X¯
LCL
UCL
Percent In
1 2 3
89.6 93.2 95.2
87.5 91.3 93.7
91.5 94.7 96.6
85.4 91.7 91.8
81.9 89.7 89.6
88.6 93.5 93.7
Percent Female In
1 2 3
74.5 76.7 76.7
70.1 71.4 72.1
79.4 81.6 80.9
75.3 80.8 79.5
70.2 77.7 76.1
79.9 83.8 82.6
Percent Male In
1 2 3
14.2 16.2 17.3
9.7 10.1 12.8
18.7 22.3 21.8
9.0 10.2 10.4
5.3 6.3 6.9
12.8 14.2 13.9
Female Bout Length
1 2 3
30.3 33.0 33.9
25.4 26.9 27.6
35.6 39.7 40.9
31.3 35.3 33.9
26.3 29.1 29.8
36.8 42.2 38.2
Male Bout Length
1 2 3
9.7 10.6 9.7
7.5 7.3 7.6
12.1 14.6 12.1
7.0 7.5 8.2
5.2 5.2 5.8
9.2 10.2 11.0
Recess Length
1 2 3
3.9 3.4 3.0
3.4 2.9 2.6
4.4 3.8 3.5
5.8 3.8 3.8
4.9 3.2 3.2
7.1 4.5 4.6
No. Female Bouts
1 2 3
22.1 21.1 21.9
18.7 17.0 18.0
25.8 25.7 26.1
21.7 21.1 21.7
19.4 18.1 19.0
24.1 24.3 24.6
No. Male Bouts
1 2 3
11.1 12.2 13.7
8.4 9.2 10.6
14.2 15.6 17.2
7.9 8.3 8.1
5.3 6.1 5.4
11.1 10.9 11.3
No. Recesses
1 2 3
21.8 17.3 13.2
18.1 13.2 9.5
25.9 21.8 17.5
20.6 18.6 18.6
17.7 15.2 15.5
23.8 22.5 21.1
No. Switches
1 2 3
14.2 18.8 26.8
9.8 11.8 16.7
19.4 27.5 39.2
11.1 13.9 12.7
7.2 7.4 6.9
16.0 22.5 20.1
Female Switch
1 2 3
493.2 466.8 423.6
434.1 384.9 360.3
552.3 548.7 486.8
553.9 487.0 502.4
466.7 430.9 425.7
641.1 543.1 579.1
Last Switch
1 2 3
1097.7 1108.3 1142.2
1046.3 1045.4 1087.0
1149.0 1171.3 1197.5
1062.4 1076.3 1112.9
1013.0 1033.1 1078.8
1111.8 1119.6 1147.1
First Recess
1 2 3
374.1 403.4 407.4
349.7 374.8 372.8
398.5 432.0 442.1
357.1 371.9 376.7
338.2 346.9 346.9
376.0 396.9 406.5
Last Recess
1 2 3
1186.4 1162.8 1126.5
1176.6 1134.0 1080.9
1196.3 1191.6 1172.1
1206.6 1189.7 1201.6
1194.8 1170.3 1179.8
1218.4 1209.0 1223.5
When CCD was significant, estimated marginal means were used. Values are based on back-transformed variables where necessary.
their females may lay up to 2 weeks later (I. Ritchie, pers. comm.). Whatever the proximate cause for a delay in clutch initiation, the timing of reproduction (i.e., CCD) was a major factor explaining treatment effects of PCBs.
Most behaviors thus appeared to be affected indirectly by PCBs. Two behaviors, Time of Last Recess and Recesses Length, were directly affected by PCBs (i.e., treatment
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
232
26
Number of recesses
24 22 20 18 16 14 12
N = 19
23 1
(a)
21
23 2 Observation period
21
21 3
Number of nesst switches
40
30
20
10
0
N=
18
22 1
(b)
20
20 2 Observation period
20
20 3
Fig. 1. (a) Mean (71 SE) number of recesses and (b) number of nest switches that occurred in captive American kestrel pairs in early (1), middle (2), and late (3) observation periods for PCB-exposed (PCB) and control groups (CTL) (solid circles, CTL; open circles, PCB).
effect remained after controlling for CCD). Increase in the length or frequency of recesses have been reported for other birds exposed to PCBs (Larus argentatus: Fox et al., 1978; L. hyperboreus: Bustnes et al., 2001; Streptopelia risoria: Peakall and Peakall, 1973; McArthur et al., 1983). Regardless of whether the effects of PCBs were direct or indirect, there were marked differences between treatment groups. All of the seven variables associated with treatment also were shown to have some consequence on hatching success. Each behavior varied in a direction consistent with our initial prediction that pairs with greater attendance to their eggs and better coordination of their incubation duties should have better hatching success (Table 4). For example, exposed males with complete hatching failure were in the nest less than half the time of contaminated males with high hatching success (Table 4). Frequent or long recesses have been considered causal factors of low hatchability in other toxicology studies (Peakall and
Peakall, 1973; Fox et al., 1978; McArthur et al., 1983). Similarly, our results for nest switches suggest a need for continuous care of eggs. Pairs that were well coordinated in having more nest switches experienced the best hatching success (Table 4). Koenig (1982) also found that behavioral synchrony between the sexes was essential for successful hatching. Factors intrinsic to the egg may have acted synergistically with parental behavior to result in the lower hatching rate of PCB-exposed kestrels (Kubiak et al., 1989; Fernie et al., 2001a). Kestrels that completed clutches later in the breeding season laid eggs with higher PCB burdens. In many species of wild birds, pairs that complete clutches later in the season generally have smaller clutches composed of eggs with higher levels of PCBs (Becker et al., 1993; Custer et al., 1998). PCB exposure lengthened incubation period by almost 1 full day. Extended incubation was also found in PCBexposed doves (Streptopelia risoria; McArthur et al., 1983) and Forster’s terns (Sterna forsteri; Kubiak et al., 1989). A longer incubation may have various fitness costs to wild birds as there is more time for eggs to be vulnerable to predation, and there may be less time for fledglings to accumulate skills prior to migration (Webb, 1987). Detrimental developmental effects on the embryo have resulted from a lengthened incubation period caused by longer or more frequent recesses (Haftorn, 1988; Ancel et al., 1995; Lindstro¨m, 1999; Jacobs and Wingfield, 2000). Inconsistent incubation may also affect embryonic metabolism, immunocompetence, and even sexual attractiveness in adulthood (Lindstro¨m, 1999). Events during avian embryonic development have future consequences on adult fecundity (Lindstro¨m, 1999). PCB exposure of these kestrels also affected size at hatching and postnatal growth (Fernie et al., 2003). Our results are significant because they demonstrate an association between behavioral differences that result from toxicological contamination and decreased reproductive success. To conclusively determine whether factors extrinsic to the egg alone affect hatching success, nest-exchange experiments should be performed. Such studies of PCBexposed birds have illustrated that increased hatchability occurred when PCB-exposed eggs were removed from contaminated adults and incubated artificially or by uncontaminated pairs (doves: Peakall and Peakall, 1973; Forster’s terns: Kubiak et al., 1989). It is likely that our results are conservative as captivity is a relatively benign environment. Birds in the wild have further constraints imposed on them, including the need to forage for food which could induce more behavioral variation and negative consequences. Our identification of specific behavioral anomalies paves the way for more targeted studies in the future to elucidate related physiological mechanisms, especially endocrine modulation. Since prolactin regulates kestrel incubation behavior (Sockman et al., 2000), this hormone should be of particular interest.
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
233
Table 4 Means and 95% lower (LCL) and upper (UCL) confidence limits for incubation behaviors that varied with hatching success in observation periods (PD) 1, 2, or 3 of PCB-exposed American kestrels X¯
Variable
PD
Hatching success
LCL
UCL
Percent In
2
0 1 2
89.9 93.8 92.9
87.4 92.1 90.4
92.1 95.4 95.0
Percent In
3
0 1 2
91.3 94.7 94.0
88.6 92.8 91.3
93.7 96.4 96.2
Percent Male In
3
0 1 2
5.7 13.0 13.2
3.0 10.8 12.0
13.2 18.9 22.5
Recess Length
2
0 1 2
4.0 3.0 4.4
3.1 2.3 3.1
5.3 4.2 6.4
No. Male Bouts
1
0 1 2
5.3 10.9 11.2
3.0 8.2 7.6
8.2 14.1 15.5
No. Recesses
3
0 1 2
22.9 18.8 11.7
18.8 14.6 8.0
27.4 23.4 16.2
No. Nest Switches
1
0 1 2
6.8 12.3 16.4
2.7 6.2 8.1
12.7 20.5 27.6
No. Nest Switches
3
0 1 2
5.6 16.4 21.0
1.2 7.3 8.9
13.2 29.2 38.2
Last Recess
1
0 1 2
1209.5 1216.2 1188.0
1192.2 1196.8 1164.7
1226.9 1235.8 1211.2
Last Recess
3
0 1 2
1224.5 1200.1 1162.8
1194.2 1165.7 1122.1
1254.8 1234.5 1203.4
Hatching Success: 0 (zero ¼ 0%), 1 (moderate ¼ 1–74%), 2 (high ¼ 75–100%). Values are based on back-transformed variables where necessary.
Acknowledgments We thank J. Huff for developing the electronic balance, I. Ritchie for management of the kestrel facility, E. Woodsworth for aid with computer programming, J. Willson for assistance, and K. Wiebe for comments on the manuscript. S.A.F. thanks the University of Saskatchewan for financial assistance through a graduate teaching fellowship. This study was funded by grants from NSERC (G.R.B., J.E.S.) and the Canadian Network of Toxicology Centres (G.R.B., J.E.S.).
References Ancel, A., Liess, S., Girard, H., 1995. Embryonic development of the domestic Guinea fowl (Numida meleagris). J. Zool. London 235, 621–634.
Balgooyen, T.G., 1976. Behavior and ecology of the American kestrel, Falco sparverius, in the Sierra Nevada of California. University of California Publications Zoology, vol. 103. University of California Press, California. Barron, M.G., Galbraith, H., Beltman, D., 1995. Comparative reproductive and developmental toxicology of PCBs in birds. Comp. Biochem. Physiol. 112C, 1–14. Becker, P.H., Schuhmann, S., Koepff, C., 1993. Hatching failure in common terns (Sterna hirundo) in relation to environmental chemicals. Environ. Pollut. 79, 207–213. Bird, D.M., 1988. American kestrel. In: Palmer, R.S. (Ed.), Handbook of North American Birds, vol. 5, Part 2. Yale University Press, New Haven, pp. 253–290. Bortolotti, G.R., Wiebe, K.L., 1993. Incubation behaviour and hatching patterns in the American kestrel Falco sparverius. Ornis Scand. 24, 41–47. Braune, B.M., Norstrom, R.J., 1989. Dynamics of organochlorine compounds in herring gulls: III. Tissue distribution and bioaccumulation in Lake Ontario gulls. Environ. Toxicol. Chem. 8, 957–968. Bustnes, J.O., Bakken, V., Erikstad, K.E., Mehlum, F., Skaare, J.U., 2001. Patterns of incubation and nest-site attentiveness in relation to
ARTICLE IN PRESS 234
S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235
organochlorine (PCB) contamination in glaucous gulls. J. Appl. Ecol. 38, 791–801. Clark, K.E., Niles, L.J., Stansley, W., 1998. Environmental contaminants associated with reproductive failure in bald eagle (Haliaeetus leucocephalus) eggs in New Jersey. Bull. Environ. Contam. Toxicol. 62, 247–254. Clotfelter, E.D., Bell, A.M., Levering, K.R., 2004. The role of animal behaviour in the study of endocrine-disrupting chemicals. Anim. Behav. 68, 665–676. Custer, C.M., Custer, T.W., Allen, P.D., Stromborg, K.L., Melancon, M.J., 1998. Reproduction and environmental contamination in tree swallows nesting in the Fox River drainage and Green Bay, Wisconsin, USA. Environ. Toxicol. Chem. 17, 1786–1798. Doving, K.B., 1991. Assessment of animal behaviour as a method to indicate environmental toxicity. Comp. Biochem. Physiol. 100, 247–252. Drent, R., 1973. The natural history of incubation. In: Farner, D.S. (Ed.), Breeding Biology of Birds. National Academy of Sciences, Washington, pp. 262–332. Drouillard, K.G., Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., Norstrom, R.J., 2001. Bioaccumulation and toxicokinetics of 42 PCB congeners in American kestrels (Falco sparverius). Environ. Toxicol. Chem. 20, 2514–2522. Elliott, J.E., Kennedy, S.W., Jeffrey, D., Shutt, L., 1991. Polychlorinated biphenyl (PCB) effects on hepatic mixed function oxidates and porphyria in birds—II. American kestrel. Comp. Biochem. Physiol. 99C, 141–145. Evangelista de Duffard, A.M., Duffard, R., 1996. Behavioural toxicology, risk assessment, and chlorinated hydrocarbons. Environ. Health Perspect. 104 (Suppl. 2), 353–360. Fernie, K.J., Bortolotti, G.R., Smits, J.E., Willson, J., Drouillard, K.G., Bird, D.M., 2000. Changes in egg composition of American kestrels exposed to dietary polychlorinated biphenyls. J. Toxicol. Environ. Health 60, 101–113. Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., 2001a. Reproductive success of American kestrels exposed to dietary polychlorinated biphenyls. Environ. Toxicol. Chem. 20, 776–781. Fernie, K.J., Smits, J.E., Bortolotti, G.R., Bird, D.M., 2001b. In ovo exposure to polychlorinated biphenyls: reproductive effects on secondgeneration American kestrels. Arch. Environ. Contam. Toxicol. 40, 544–550. Fernie, K.J., Bortolotti, G.R., Smits, J.E., 2003. Reproductive abnormalities, teratogenicity, and developmental problems in American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. J. Toxicol. Environ. Health A 66, 2089–2103. Fisher, S.A., Bortolotti, G.R., Fernie, K.J., Smits, J.E., Marchant, T.A., Drouillard, K.G., Bird, D.M., 2001. Courtship behaviour of captive American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. Arch. Environ. Contam. Toxicol. 41, 215–220. Fox, G.A., Gilman, A.P., Peakall, D.B., Anderka, F.W., 1978. Behavioral abnormalities of nesting Lake Ontario herring gulls. J. Wildlife Manage. 42, 477–483. Fry, D.M., 1995. Reproductive effects in birds exposed to pesticides and industrial chemicals. Environ. Health Perspect. 103 (Suppl. 7), 165–171. Fyfe, R.W., Risebrough, R.W., Walker, IIW., 1976. Pollutant effects on the reproduction of the prairie falcons and merlins of the Canadian prairies. Can. Field Nat. 90, 346–355. Haftorn, S., 1988. Incubating female passerine do not let the egg temperature fall below the ‘physiological zero temperature’ during their absences from the nest. Ornis Scand. 19, 97–110. Harris, H.J., Erdman, T.C., Ankley, G.T., Lodge, K.B., 1993. Measures of reproductive success and polychlorinated biphenyl residues in eggs and chicks of Forster’s terns on Green Bay, Lake Michigan, Wisconsin—1988. Arch. Environ. Contam. Toxicol. 25, 304–314. Hoffman, D.J., Rice, C.P., Kubiac, T.J., 1996a. PCBs and dioxins in birds. In: Beyer, W.N., Heinz, G.H., Redmon, A.W. (Eds.), Environmental
Contaminants in Wildlife: Interpreting Tissue Concentrations. SETAC Publication, Lewis Publishers, pp. 165–207. Hoffman, D.J., Melancon, M.J., Klein, P.N., Rice, C.P., Eisemann, J.D., Hines, R.K., Spann, J.W., Pendleton, G.W., 1996b. Developmental toxicity of PCB 126 (3,30 ,4,40 ,5-pentachlorobiphenyl) in nestling American kestrels (Falco sparverius). Fundam. Appl. Toxicol. 34, 188–200. Jacobs, J.D., Wingfield, J.C., 2000. Endocrine control of life-cycle stages: a constraint on response to the environment. Condor 102, 35–51. Koenig, W.D., 1982. Ecological and social factors affecting hatchability of eggs. Auk 99, 526–536. Kozie, K.D., Anderson, R.K., 1991. Productivity, diet and environmental contaminants in bald eagles nesting near the Wisconsin shoreline of Lake Superior. Arch. Contam. Toxicol. Chem. 20, 41–48. Kubiak, T.J., Harris, H.J., Smith, L.M., Schwartz, T.R., Stalling, D.L., Trick, J.A., Sileo, L., Docherty, D.E., Erdman, T.C., 1989. Microcontaminants and reproductive impairment of the Forster’s tern on Green Bay, Lake Michigan—1983. Arch. Environ. Contam. Toxicol. 18, 706–727. Lindstro¨m, J., 1999. Early development and fitness in birds and mammals. Trends Ecol. Evol. 14, 343–348. McArthur, M.L.B., Fox, G.A., Peakall, D.B., Philoge`ne, B.J.R., 1983. Ecological significance of behavioural and hormonal abnormalities in breeding ring doves fed an organochlorine chemical mixture. Arch. Environ. Contam. Toxicol. 12, 343–353. McCarty, J.P., Secord, A.L., 1999a. Nest-building behavior in PCBcontaminated tree swallows. Auk 116, 55–63. McCarty, J.P., Secord, A.L., 1999b. Reproductive ecology of tree swallows (Tachycineta bicolor) with high levels of polychlorinated biphenyl contamination. Environ. Toxicol. Chem. 18, 1433–1439. Milstein, P.le S., Prestt, I., Bell, A.A., 1970. The breeding cycle of the grey heron. Ardea 58, 171–257. Nelson, R.W., 1976. Behavioural aspects of egg breakage in peregrine falcons. Can. Field Nat. 90, 320–329. Noru˘sis, M., 1991. The SPSS Guide to Data Analyses for SPSS/PC+. Prentice-Hall, Englewood Cliffs, NJ. Peakall, D.B., Peakall, M.L., 1973. Effect of a polychlorinated biphenyl on the reproduction of artificially and naturally incubated dove eggs. J. Appl. Ecol. 10, 863–868. Porter, R.D., Wiemeyer, S.N., 1972. Reproductive patterns in captive American kestrels (sparrow hawk). Condor 74, 46–53. Rehder, N.B., Bird, D.M., Lague, P.C., 1986. Variations in plasma corticosterone, esterone, estradiol-17b, and progesterone concentrations with forced renesting, molt, and body weight of captive female American kestrels. Gen. Comp. Endocrinol. 62, 386–393. Risebrough, R.W., Anderson, D.W., 1975. Some effects of DDE and PCB on mallards and their eggs. J. Wildlife Manage. 39, 508–513. Silver, R., Ball, G.F., 1989. Brain, hormone and behavior interactions in avian reproduction: status and prospectus. Condor 91, 966–978. Smits, J.E.G., Bortolotti, G.R., 2001. Antibody-mediated immunotoxicity in American kestrels (Falco sparverius) exposed to polychlorinated biphenyls. J. Toxicol. Environ. Health A 62, 217–226. Smits, J.E.G., Fernie, K.J., Bortolotti, G.R., Marchant, T.A., 2002. Thyroid hormone suppression and cell mediated immunomodulation in American kestrels (Falco sparverius) exposed to PCBs. Arch. Environ. Contam. Toxicol. 43, 338–344. Sockman, K.W., Schwabl, H., Sharp, P.J., 2000. The role of prolactin in the regulation of clutch size and onset of incubation behaviour in the American kestrel. Horm. Behav. 38, 168–176. Valkama, J., Korpimaki, E., 1999. Nest box characteristics, habitat quality and reproductive success of Eurasian kestrels. Bird Study 46, 81–88. Van Den Berg, M., Craane, L.H.J., Van Mourik, S., Brouwer, A., 1995. The (possible) impact of chlorinated dioxin (PCDDs), dibenzofurans (PCDFs), and biphenyls (PCBs) on the reproduction of the cormorant Phalacrocorac carbo—an ecotoxicological approach. Ardea 83, 299–313.
ARTICLE IN PRESS S.A. Fisher et al. / Ecotoxicology and Environmental Safety 63 (2006) 226–235 Vos, J.G., Dybing, E., Greim, H.A., Ladefoged, O., Lambre´, C., Tarazona, J.V., Brandt, I., Vethaak, A.D., 2000. Health effects of endocrine-disrupting chemicals on wildlife, with special reference to the European situation. Crit. Rev. Toxicol. 30, 71–133. Webb, D.R., 1987. Thermal tolerance of avian embryos: a review. Condor 89, 874–898.
235
Wilmers, T.J., Bowman, R., Samuel, D.E., 1985. Notes on Incubation by male kestrels in western Virginia, Pennsylvania, and southern Quebec. N. Am. Bird Bander 10, 6–8. Zala, S.M., Penn, D.J., 2004. Abnormal behaviours induced by chemical pollution: a review of the evidence and new challenges. Anim. Behav. 68, 649–664.