Received: 26 October 2017 | Revised: 13 December 2017 | Accepted: 3 March 2018 DOI: 10.1111/eth.12747 RESE ARCH PAPER Patterns of extra-­pair paternity in mountain chickadees Erica S. Bonderud1 | Ken A. Otter1 | Theresa M. Burg2 | Kristen L. D. Marini3 | Matthew W. Reudink3 1 Natural Resources and Environmental Studies, University of Northern British Columbia, Prince George, BC, Canada 2 Abstract Extra-­pair paternity (EPP) is common in chickadees and often attributed to the good Department of Biological Sciences, University of Lethbridge, Lethbridge, AB, Canada genes hypothesis. Females generally seek dominant males, who are typically larger, 3 Department of Biological Sciences, Thompson Rivers University, Kamloops, BC, Canada songbird species, habitat quality and urbanization have been found to influence EPP. Correspondence Erica S. Bonderud, Natural Resources and Environmental Studies, University of Northern British Columbia, Prince George, BC, Canada. Email: bonderud@unbc.ca songbirds. Here, we ask how individual condition and urbanization influence rates of Funding information Natural Sciences and Engineering Research Council of Canada, Grant/Award Number: -, 418082 and RGPIN-2014-04300; University of Northern British Columbia, Grant/Award Number: - and 26830 Editor: E. Hebets older and sing at higher rates than subordinate males, as extra-­pair sires. In other Mountain chickadees commonly inhabit suburban habitat, and previous research on our population has shown urbanization may provide benefits to these adaptable EPP in mountain chickadees. Over three breeding seasons, we monitored mountain chickadee nests in urban and rural habitat, and determined parentage by genotyping adults and nestlings at six microsatellite loci. Extra-­pair paternity is common in mountain chickadees, with extra-­pair offspring (EPO) in 43.2% of nests and accounting for 17.9% of offspring. We found tenuous support for the good genes hypothesis with females tending to engage in EPCs with older males. However, we did not find an influence of male or female condition on the proportion of EPO in a nest. In addition, we did not find a significant effect of habitat on EPP rates, suggesting the impacts of urbanization on mountain chickadee reproduction may not extend to altering extra-­ pair behaviour. KEYWORDS condition, good genes hypothesis, habitat quality, mating strategies, Poecile gambeli, urbanization 1 | I NTRO D U C TI O N EPCs with males of higher quality than their social mates to obtain favourable genes for their offspring. Females may assess male quality In birds, social monogamy is widespread, but is often coupled with using phenotypic signals (e.g., plumage ornamentation, song or be- a mixed-­ mating strategy that includes extra-­ pair paternity (EPP). haviour) that convey information about physical condition (e.g., nu- There are several hypotheses as to why a female may choose to seek tritional state) and/or genetic quality. For example, female blue tits extra-­pair copulations (EPCs), including to insure fertilization of her (Cyanistes caeruleus) seek EPCs with older, larger males (Kempenaers, eggs, increase the genetic diversity of her offspring, or to receive Verheyen, & Dhondt, 1997) and males with brighter ultraviolet-­ direct benefits, like defence or resources, from the extra-­pair (EP) blue plumage (Kempenaers et al., 1992), while female black-­capped male (Griffith, Owens, & Thuman, 2002). Extra-­ pair paternity is chickadees (Poecile atricapillus) seek EPCs with more dominant males common in the Paridae family (chickadees and titmice), with extra-­ (Mennill, Ramsay, Boag, & Ratcliffe, 2004; Otter, Ratcliffe, & Boag, pair offspring (EPO) present in 30%–75% of nests, and accounting 1994; Otter, Ratcliffe, Michaud, & Boag, 1998; Smith, 1988). for 7%–25% of offspring (reviewed in Griffith et al., 2002). In Parids, Chickadee social structure revolves around dominance hier- EPP is often explained by the good genes hypothesis: females seek archies (Ratcliffe, Mennill, & Schubert, 2007). Although much of Ethology. 2018;1–9. wileyonlinelibrary.com/journal/eth © 2018 Blackwell Verlag GmbH | 1 2 | BONDERUD et al. our understanding of this system comes from black-­capped chick- (e.g., access to bird feeders, earlier leaf-­out and insect emergence) to adees, mountain chickadees (P. gambeli) are known to form linear conifer-­natives, as the mountain chickadees in our study population dominance hierarchies within winter flocks (McCallum, Grundel, initiate breeding earlier in urban habitat (Marini, Otter, LaZerte, & & Dahlsten, 1999), with males typically dominant to females, and Reudink, 2017). In addition, nestlings from urban nests have faster adults typically dominant to juveniles (Grava et al., 2012). This is par- feather growth than their rural counterparts (Marini et al., 2017), allel to the social rank structure of black-­capped chickadees; thus, it which could indicate nestlings in urbanized areas are being better is likely that other predictors of dominance in mountain chickadees provisioned, as has been shown in song sparrows (Melospiza melodia; are similar to those known for black-­capped chickadees, such as Searcy, Peters, & Nowicki, 2004). body condition (Schubert et al., 2007), and male song output (Otter, Chruszcz, & Ratcliffe, 1997). Extra-­pair paternity has been related to habitat quality in other songbird species. In house sparrows (Passer domesticus), experi- In black-­capped chickadees, dominant individuals gain increased mentally increased food availability resulted in pairs spending more access to resources (Ratcliffe et al., 2007) and are sought by fe- time together at the nest, which, in turn, led to a fivefold reduction males as both social mates (Otter & Ratcliffe, 1996; Ramsay, Otter, in EPP rates (Václav, Hoi, & Blomqvist, 2003). In superb starlings Mennill, Ratcliffe, & Boag, 2000) and EP partners (Otter et al., 1994, (Lamprotornis superbus), Rubenstein (2007) found EPP to be less 1998). Males signal their status through condition-­dependent traits, prevalent in higher quality territories (greater vegetation cover and which may provide females with mechanisms to assess male quality prey availability). As the author suggests, greater prey availability in (Otter et al., 1997). Dominant males are typically larger, but leaner, high-­quality territories may limit the distance females need to travel and have greater song output than subordinate males (Dixon, 1965; to forage, and consequently, decrease her probability of encounter- Grava, Grava, & Otter, 2009; Otter et al., 1997; Ratcliffe et al., 2007; ing an EP male. In contrast, serin (Serinus serinus) nests in territories Schubert et al., 2007). In black-­capped chickadees, male song output with greater food availability are more likely to contain EPO than is a condition-­dependent trait, with males in good condition (usually nests in poor-­quality habitat (Hoi-­Leitner, Hoi, Romero-­Pujante, & dominant males) singing for longer periods and at higher frequen- Valera, 1999). The authors postulate females on high-­quality terri- cies than males in poor condition (usually subordinate males) (Grava tories may be in better condition and more able to resist male mate et al., 2009; Otter et al., 1997). In a supplemental feeding experi- guarding efforts, and thus, may have greater opportunity seek EPCs ment, Grava et al. (2009) found male black-­capped chickadees that (Hoi-­Leitner et al., 1999). received additional food had greater song output than their unfed For spotted towhees (Pipilo maculatus) breeding in urban parks, counterparts. This trend was observed in both dominant and subor- park edges are sites of high food abundance due to anthropogenic dinate males, and in both high-­and low-­quality habitats, suggesting food sources (e.g., bird feeders), while interior habitat is compara- individual condition is the main component contributing to variation tively lower quality (Smith, McKay, Murphy, & Duffield, 2016). As in song output (Grava et al., 2009). Thus, the difference in song out- such, EPP rates may be expected to be greater at the habitat interior put between dominant and subordinate males appears to be a by-­ than the edge. Smith et al. (2016), however, found the relationship product of differential resource accessibility. between EPP rates and nest distance from habitat edge to vary non-­ Habitat urbanization can affect food availability and interspecific linearly: the probability a nest contained EPO was the greatest at the interactions, and consequently, may impact a female’s likelihood to habitat edge and interior, and lowest at intermediate distances. The both seek and encounter EPCs. While some species successfully authors suggest anthropogenic food sources may have drawn indi- colonize and thrive in urbanized habitat, how a specific species viduals from the interior habitat to the edge, resulting in increased fares is dependent on multiple factors. Mountain chickadees pres- contact with potential EP sires in edge territories and greater occur- ent a unique opportunity to investigate the effects of urbanization rences of EPP than predicted (Smith et al., 2016). In our study pop- on avian reproduction, as unlike other focal species that have been ulation, mountain chickadees in urban habitat are more dispersed investigated (Bailly et al., 2016), mountain chickadees are native to than those in rural habitat (E. Bonderud, personal observation), sug- coniferous forests (McCallum et al., 1999), rather than deciduous gesting any potential increase in food availability in urban habitat forests. Compared to deciduous forest, urban habitat has a lower does not attract individuals from neighbouring rural habitat. As such, density of deciduous vegetation, and consequently, food availability urban habitat may limit EPC opportunities for mountain chickadees. for deciduous-­specialists. Thus, urban habitat is often cited as lower Here, we investigate the ecological and social factors that influ- quality habitat for such species (Blewett & Marzluff, 2005; Marzluff, ence patterns of EPP in mountain chickadees. Specifically, we ask 1997). whether female condition influences her propensity to seek EPCs, For species native to coniferous forests, the opposite may be true. and whether male condition predicts his likelihood of losing pater- Although lower total canopy cover, there is greater deciduous vege- nity. Following the good genes hypothesis, we predict females in tation at our urban nest sites, as compared to the conifer-­dominated good condition (presumably paired to a male in good condition) will rural nest sites. Because deciduous trees typically bear greater in- be less likely to seek EPCs, and males in good condition will be less sect abundance and diversity (Southwood, 1961), it is possible urban likely to lose paternity. In addition, we ask how ecological charac- habitat is associated with greater prey availability, as compared to teristics of nesting habitat (urbanization, vegetation composition) coniferous forest. Indeed, urban habitat appears to provide benefits affect rates of EPP. We predict EPP to be less prevalent in urban | 3 BONDERUD et al. habitat (presumably high-­quality habitat) than rural habitat, although samples from nestlings by piercing the ulnar vein and drawing as shown by Smith et al. (2016), urban habitat may promote unex- 10–20 μl blood into a micro-­capillary tube. We stored blood samples pected breeding strategies. dried on filter paper at –20°C. 2 | M E TH O DS 2.3 | Dawn vocalization recordings 2.1 | Study site We recorded dawn vocalizations from males breeding in the urban (n = 9 unique males) and rural (n = 9 unique males) study sites be- We monitored mountain chickadees breeding in nest boxes in urban tween 4 May and 16 May during the 2014–2016 breeding seasons. and rural areas of Kamloops, BC, Canada during the 2014–2016 We recorded dawn vocalizations using a Sennheiser ME67/K6 mi- breeding seasons (May-­Jul). Nest boxes were distributed through- crophone and either an Olympus LS-­14 or a Marantz PMD670 digital out south Kamloops on a gradient of rural to urban habitat. Rural recorder on settings of at least 44 kHz sampling frequency and 16-­ nest boxes were located in Kenna Cartwright Park, an approximately bit digitization, or higher. We arrived on-­site approximately 30 min 2 8 km wilderness area consisting primarily of Great Basin grassland before sunrise to determine dawn vocalization start time and obtain habitat (e.g., sagebrush, Artemisia tridentata; saskatoon, Amelanchier full recordings. We considered male dawn vocalizations to be fin- alnifolia; Poaceae spp.), but with mature ponderosa pine (Pinus pon- ished following a five-­minute period of silence following the last vo- derosa) and Douglas fir (Pseudotsuga mensiesii) forests occupying calization. Dawn vocalization recordings began between 04:34 and approximately 20% of the park; it is these forests occupied by moun- 05:21, and lasted an average of 40 ± 13 min (mean ± SD, n = 18). To tain chickadees. Urban nest boxes were spaced in several clustered ensure the male being recorded was the male associated with a given patches distributed over approximately 37 km2 of various urban and nest, we began recording at the nest box itself and only recorded suburban areas of south Kamloops, including the Thompson Rivers within a 75 m radius of the nest box. Often, the vocalizing male was University campus, neighbourhood parks and backyards of partici- observed copulating with the female at the nest box following ces- pating citizens. The interspacing of nest boxes within these clustered sation of dawn vocalizations, providing further confirmation the re- patches was similar to the spacing with which they were deployed corded male was the male associated with that nest. We analysed in our rural site. The vegetation at these sites consisted primarily of recordings using Avisoft–SAS Lab Pro (Avisoft Bioacoustics, 2017) immature Douglas fir trees and various species of native and non-­ and calculated total vocalization (songs and calls) rate (vocalizations/ native deciduous trees and shrubs (e.g., maple spp., Acer spp.; moun- min) as a measure of male condition. tain ash, Sorbus spp.; various fruit trees). 2.2 | Nest monitoring and sampling We monitored nest boxes every 1–3 days to identify the breeding 2.4 | Habitat classification 2.4.1 | Habitat index pairs occupying boxes and measure breeding success. We caught Because our study sites varied along a gradient from natural habi- adults at the nest, either while brooding or feeding, and banded tat to urban neighbourhoods, we used a habitat index developed them with a Canadian Wildlife Service (CWS) aluminium leg band by LaZerte, Otter, and Slabbekoorn (2017; scripts available from with a numerical identifier, and a unique combination of three col- https://github.com/steffilazerte/urbanization-index) oured plastic leg bands. We classified age as either second-­year (SY) ground cover and measure the degree of habitat urbanization at to assess or after-­second-­year (ASY) by examining the shape and wear of the each nest location. We used R v3.3.2 (R Core Team, 2016) to plot a outer retrices (Pyle, 1997). We determined adult sex in the field by 75 m radius around each nest box location (approximately the size the presence (females) or absence (males) of a brood patch and by of the average territory) in Google Earth (Google Inc., 2015). We behaviour at the nest (e.g., only females incubate and only males imported these aerial images into the image manipulation software sing), and later confirmed sex genetically (Bonderud et al., 2017). For GIMP v2.8.16 (The GIMP Team, 2015), where we manually classified genetic analysis, we collected two tail feathers from each adult and the buildings, pavement, natural and non-­native grasses, deciduous stored the samples at −20°C. We measured adult fat score, weight, trees and coniferous trees around each point location. We grouped tail length, tarsus length and flattened wing chord to evaluate indi- buildings and pavement into a single variable (“urban features”) and vidual body size and condition. To estimate overall body condition, conducted a principal components analysis (PCA) in R to collapse we calculated the residuals from a mass x tarsus linear regression. the five variables into a single, continuous index of habitat. The first Using the same metric in black-­capped chickadees, Schubert et al. principal component (PC1) accounted for 68% of the total varia- (2007) found leaner males with larger skeletal frames had higher tion in habitat ground cover. Higher PC1 values corresponded to a dominance ranks. Thus, negative residual values may suggest an in- greater number of coniferous trees and natural grass cover, and less dividual is in better condition. cover of deciduous trees, non-­native grasses and urban features (i.e., Six days after hatching, we banded nestlings with a single CWS greater natural vegetation cover, increasing “rural-­ness,” decreas- aluminium band. Twelve days after hatching, we collected blood ing “urban-­ ness”) (PC1 loadings: coniferous trees = 0.35, natural 4 | BONDERUD et al. grasses = 0.40, deciduous trees = −0.46, non-­native grasses = −0.50, urban features = −0.51). 2.4.2 | Vegetation index Following methods similar to those used in the habitat index, we assessed only the vegetation at each nest location to determine vegetation type and cover, and proxy food availability. We obtained aerial images of each nest location from Google Earth (Google Inc., 2015), and using GIMP (The GIMP Team, 2015), manually classified the deciduous tree cover, coniferous tree cover and other ground cover (e.g., grass, pavement) within a 75 m radius of the nest. We TA B L E 1 Allelic variation at the six microsatellite markers. Size ranges for microsatellite alleles are given in base pairs, along with the total number of unique alleles observed, and observed (Ho) and expected (He) heterozygosities Locus Size (bp) # Alleles Ho He Pat14 132–176 20 0.93 0.86 Pat43 158–232 16 0.88 0.85 Titgata02 222–270 12 0.82 0.80 Titgata39 210–238 11 0.89 0.87 Escμ4 170–184 8 0.70 0.76 Escμ6 128–152 13 0.91 0.85 conducted a PCA in R (R Core Team, 2016) to collapse the three variables into a single value to describe vegetation cover. PC1 accounted a 50–350 bp size standard on each load/channel to ensure alleles for 81% of the total variation in vegetation cover. Higher PC1 values were sized consistently across gels. corresponded to greater coniferous tree cover and lower deciduous tree and other ground cover (i.e., greater canopy cover, greater coniferous content) (PC1 loadings: coniferous trees = 0.63, deciduous 2.6 | Parentage assignment trees = −0.48, other ground cover = −0.61). Because deciduous trees We assigned parentage (both paternity and maternity) first by hand, typically bear greater insect abundance and diversity (Southwood, and then again using CERVUS 3.0 (Kalinowski, Taper, & Marshall, 1961), higher PC1 values may correspond to lower prey availability 2007). In some cases, we were not able to genotype individuals at all or quality. six loci due to insufficient quantities of DNA, which resulted in amplification failure. Previous parentage studies in black-­capped chick- 2.5 | Molecular methods adees employed only three microsatellite markers but still excluded sires with a high degree of confidence (Mennill et al., 2004; Otter We extracted total genomic DNA from feather samples using the et al., 1998). Thus, we only included nestlings with three or more loci standard protocol for the QIAamp DNA Micro Kit (Qiagen, Hilden, successfully typed in parentage analysis (only one offspring was ex- Germany), and from blood samples using the standard protocol for cluded for not meeting those criteria). We classified offspring as EP the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) with a if they had two or more mismatches with the putative mother or fa- modified lysis step. ther (Mennill et al., 2004). We then conducted parentage analysis in We assessed nestling parentage by genotyping all adults and CERVUS and combined these results with our manual assignments. nestlings at six avian microsatellite loci (Table 1): Pat14 (Otter In CERVUS, we set 99% strict and 95% relaxed confidence limits. We et al., 1998), Pat43 (Otter et al., 1998), Titgata02 (Wang, Hsu, estimated 75% of the male population had been genotyped based on Yao, & Li, 2005), Titgata39 (Wang et al., 2005), Escμ4 (Hanotte surveys of the proportion of banded vs. unbanded birds observed in et al., 1994) and Escμ6 (Hanotte et al., 1994). We amplified DNA our rural population over the breeding season, and based on similar in 10 μl reactions containing 1 × buffer, 1.5–2.5 mM MgCl2 estimates within the clustered study areas within the urban popula- (1.5 mM: Pat43, Escμ6; 2.0 mM: Pat14, Titgata39; 2.5 mM: tion. CERVUS did not identify any additional EPO; however, eight- Titgata02, Escμ4), 0.25 U Taq DNA polymerase, 0.5 μM forward een offspring we classified as EP in our manual assignments were primer, 1.0 μM reverse primer and 0.05 μM fluorescently labelled identified as within-­ pair by CERVUS. In these cases, if CERVUS M13 primer. All forward primers were synthesized with a M13 identified the social male as the first or second most-­likely father sequence on the 5′ end to allow for incorporation of the fluores- when all males in the population were considered, we accepted the cently labelled M13 primer. We added 1% formamide to reactions CERVUS assignment (n = 8); if not, we retained the manual assign- involving Pat14 and Escμ4. ment (n = 10). In some instances, CERVUS was unable to assign par- We amplified all loci using a two-­step annealing protocol: one cycle of 94°C for 2 min, 50°C for 45 s and 72°C for 1 min, followed entage due to an unknown putative male or female. In these cases, we used the manual assignment. by seven cycles of 94°C for 1 min, 50°C for 30 s and 72°C for 45 s, To identify EP sires, we used CERVUS to compare the genotypes followed by 25 cycles of 94°C for 30 s, 52°C for 30 s and 72°C for of EPO to all males in the population. Using the males CERVUS 45 s, followed by a final extension step of 72°C for 5 min. For two identified as the most-­likely fathers (≥95% confidence) and breed- loci (Escμ4 and Escμ6), the third step was increased from 25 to 31 ing information from our study population, we created an index of cycles. For one locus (Escμ4), we decreased annealing temperatures confidence in the realistic validity of the EP sire assignments (i.e., from 50°C and 52°C to 45°C and 48°C, respectively. PCR products we asked whether it was realistic for the identified male to have en- were run on a 6% acrylamide gel on a Licor 4300 (Licor Inc.). We countered the female and sired EPO in her nest). For each EPO, the included individuals of known allele sizes, a negative control, and identified father was only considered a realistic EP sire if he had held | 5 BONDERUD et al. territory within 500 m (measured from box-­to-­box) of the female 2015, or 2014 and 2016; however, we found a greater proportion at some point during the study period. Otter et al. (1998) found EP of EPO in 2016 as compared with 2015 (2014 vs. 2015: χ 2 = −0.74, males typically hold the adjacent territory in black-­capped chicka- p = .46; 2014 vs. 2016: χ2 = 1.08, p = .28; 2015 vs. 2016: χ2 = 2.12, dees, thus, we chose 500 m as a conservative cut-­off distance. As p = .03). the average interterritory spacing between territories in our rural To ask how male condition predicted the proportion of EPO study site is approximately 250–300 m, this distance would include in his nest, we constructed GLMMs with logit link functions and males up to two territories away. We did not restrict our criteria to binomial error distributions. We used the number of EPO in the only males having bred in the same year as the female because not nest as the response variable, the total number of offspring in the every male was recaptured in subsequent years. If the male was not nest as the binomial denominator, age (ASY/SY) and body condi- recaptured, we assumed he was still alive in subsequent breeding tion (mass x tarsus regression residual) as predictor variables and seasons, and that he bred in the vicinity of his original nest, as the included study year and male identity as crossed random effects mountain chickadees in our study sites have high site fidelity (E. to account for multiple observations of the same breeding adult Bonderud, personal observation). across study years, and interannual variation in EPP rates. We included “age x body condition” as an interaction term and dropped 2.7 | Statistical analyses it if non-­significant (p > .10) to derive the final model. We ran a second model using male total vocalization (songs and calls) rate Across all three breeding seasons, we monitored 46 nests and col- (vocalizations/min) as predictor variables as dawn vocalization data lected data on 260 nestlings and 59 adults. Of the 46 nests moni- were only available for a subset of males. To ask how female condi- tored, 31 had both the attending male and female identified, five tion predicted the proportion of EPO in her nest, we repeated the had an unidentified female, nine had an unidentified male, and one above analysis using measures of female condition (mass x tarsus had neither adult identified. In total, seven males and seven females regression residual, age) as the predictor variables, and study year were recaptured in more than one breeding season. In three cases, and female identity as crossed random effects. Finally, to ask how the same male and female paired in more than one breeding season. habitat influenced the proportion of EPO in a nest, we constructed In one case, a pair produced two successful broods within a single a similar GLMM using our habitat and vegetation indices as the breeding season. Only first broods within a year were included in predictor variables, and study year and female identity as crossed our analyses. random effects. Hatching success in our population was 87% (266/306 eggs hatched). Of the 266 nestlings, we were able to obtain genetic samples from 260 (98%) nestlings from 46 broods. Genetic samples were 2.8 | Ethical note not obtained from six nestlings from six broods because mortality All work was approved by the University of Northern British occurred before collection on day 12. We genotyped all 260 nest- Columbia Animal Care and Use Committee and was conducted lings and were able to assign maternity and paternity to all but two under a Canadian Federal Master Banding Permit and Scientific nestlings, one for which the paternal genotype was not known and Collection Permit no. 22806. the nestling was the only offspring in the brood (thus, we could not assign paternal alleles as coming from a WP or EP source), and the other for which only two loci amplified. In total, 252 nestlings (rural: n = 155, urban: n = 97) from 44 broods (rural: n = 28, urban: n = 16) were included in our analyses. 3 | R E S U LT S 3.1 | General patterns of parentage We conducted statistical analyses in STATA 14 (StataCorp, In high-­density populations, only a small proportion of the total 2015). To compare the condition of a male who lost paternity to population is typically sampled. However, in both of our study areas, the male who gained paternity in his nest (i.e., social male vs. EP mountain chickadees are found in low densities (E. Bonderud, per- sire), we conducted either a paired t test (male body condition) or sonal observation). Based on the proportion of banded adult males Wilcoxon signed-­r ank test (male age). To ask whether rates of EPP detected during breeding, we estimated 75% of the males had been differed between the study years, we constructed generalized lin- sampled in our study. As is typical of many studies, we were only ear mixed models (GLMMs) with logit link functions and binomial able to sample social males from a subset of the nests (n = 36) in error distributions. We included the number of EPO in the nest as which we genotyped offspring (n = 46). When conducting manual the response variable, the total number of offspring in the nest as parentage assignment for the 10 broods with no social male genetic the binomial denominator, and study year as the predictor vari- information, we used a conservative approach and assumed that if able. Because eight females produced more than one brood across all nestlings shared a single set of paternal alleles that these were ob- the study period, and we assumed EPCs to be sought by females tained from the social male rather than an extra-­pair male. EPP rates (Otter & Ratcliffe, 1996; Otter et al., 1999; Ramsay et al., 2000; were similar when all nests were included (17.9% EPO) and when Smith, 1988), we included female identity as a random effect. We nests with no sample from the social male were excluded (18.2% did not find the proportion of EPO to differ between 2014 and EPO). | BONDERUD et al. Nests with EPO Nests with WPO only Percentage of nests (a) Nests 100% 80% 60% 40% 20% 0% 15 2014 17 12 2015 2016 Extra-pair offspring Within-pair offspring (b) Offspring Percentage of offspring 6 100% 80% 60% 40% 20% 0% Year 81 2014 Overall, 17.9% (45/252) of offspring were sired by an EP male, 94 77 2015 2016 Year F I G U R E 1 (a) Percentage of nests containing at least one extra-­pair offspring (EPO) and all within-­pair offspring (WPO); (b) Percentage of offspring that were extra-­pair and within-­ pair. Numbers within bars represent total number of (a) nests and (b) offspring sampled in each study year 4 | D I S CU S S I O N and 43.2% (19/44) of nests contained EPO (Figure 1). The percentage of EPO within nests with mixed parentage ranged from 12.5% Extra-­pair paternity has been well-­studied in Paridae species, but, (1/8 offspring EP) to 100% (4/4 offspring EP) in a single nest; the ma- until now, intraspecific EPP has not been investigated in mountain jority of mixed parentage nests (13/19) contained under 50% EPO, chickadees. Here, we present evidence that mountain chickadees with an average 39.8% EPO per nest. Exclusionary power based on also frequently engage in EPCs: EPO were found in almost 50% of the six microsatellite loci was 0.987 for the first parent, 0.999 for the nests and represented almost 20% of offspring. These were some- second parent, and 0.999 combined. We were able to confidently what surprisingly high frequencies, as rates of EPP in the closely identify seven EP males from six (31.6%) of the mixed-­paternity related black-­capped chickadee tend around 30% of nests and 10%– nests, with two EP males siring EPO within a single brood in one 15% of offspring (Mennill et al., 2004; Otter et al., 1998; Ramsay case. Both the social male and EP male were known within four et al., 2000). However, similarly high, and even higher, rates of EPP nests, one of which was the nest with two EP males identified, re- to what we observed have been documented in the related blue tit sulting in a total of five social male/EP male pairs for comparison (see (60% of nests; Delhey, Peters, Johnsen, & Kempenaers, 2007) and below). All offspring were determined to be the genetic offspring of great tit (Parus major, 40% of nests; Lubjuhn, Gerken, Brün, & Epplen, their putative mother. 1999). In addition, high rates of EPP have been observed in black-­ capped chickadees hybridizing with Carolina chickadees (Poecile 3.2 | Condition and extra-­pair paternity We did not find either female age (χ2 = −0.15, p = .88) or body condi2 carolinensis, 55.6% of nests; Reudink, Mech, & Curry, 2006) and mountain chickadees (62.5% of nests; Grava et al., 2012). The presence of EPP in Paridae species has often been at- tion (χ = −0.14, p = .89) to influence the proportion of EPO in her tributed to the good genes hypothesis: females engage in EPCs to nest (Table 2). Likewise, we did not find any measures of male condi- obtain more favourable genes for their offspring than their social tion (age, body condition or vocalization rate) to influence the pro- mate can provide to increase their own fitness. In black-­capped portion of EPO in the nest (all p > .13; Table 2). chickadees, females engage in EPCs with males of higher dominance rank than their social male (Mennill et al., 2004; Otter et al., 1998). 3.3 | Social male vs. extra-­pair male comparisons In blue tits, several measures of condition appear to influence the decisions of females, with larger males (Kempenaers et al., 1997), When we compared the condition of the male that lost paternity to older males (Kempenaers et al., 1997), males with greater song out- the EP male, we found no difference in body condition (t4 = 1.38, put (Kempenaers et al., 1997) and males who begin singing earlier p = .24) or age (W = −1.73, p = .08) between the two, but the small (Poesel, Kunc, Foerster, Johnsen, & Kempenaers, 2006) being sought sample size (n = 5) for these comparisons resulted in low power. In as EP sires. Here, we asked how male condition, as measured by age, all five social male/EP male pairs, the EP male was of equivalent age weight relative to body size and dawn vocalization rate influence his (ASY vs. ASY, n = 2) or older than the social male (ASY vs. SY, n = 3). likelihood to lose paternity. Contrary to these studies, we did not In the two comparisons where the males were of equivalent age, find a significant relationship between male condition and the pro- two males holding neighbouring territories on the Thompson Rivers portion of EPO in his nest. Although we had few cases where both University campus sired EPO in each other’s nests (i.e., both gained the social male and the EP male were known, among those where paternity, but also both lost paternity to one another). this information was available, the EP male was either of equivalent age to the social male or older. In the cases where the males were of equivalent age, both males were ASY, although their age in years 3.4 | Habitat effects was not known. Together, these anecdotes suggest female mountain 2 We found neither the habitat index (χ = −1.52, p = .13) nor vegeta- chickadees engage in EPCs with adult males (5/5 cases), and males tion composition (χ2 = 1.37, p = .17) of nesting habitat to have an ef- older than their social male (3/5 confirmed cases)—findings consis- fect on the proportion of EPO in a nest (Table 2, Figure 2). tent with the good genes hypothesis. | 7 BONDERUD et al. TA B L E 2 Results of generalized linear mixed models asking how female condition (age, mass x tarsus regression residual), male condition (age, mass x tarsus regression residual, vocalization rate) and habitat characteristics (habitat index, vegetation index) influence the proportion of EPO in nests Estimate SE χ2 p n (broods) Female age −0.16 1.06 −0.15 .88 37 Female body condition score −0.12 0.86 −0.14 .89 37 Male age −1.49 0.97 −1.53 .13 35 Male body condition score −1.00 1.03 −0.97 .33 35 Male vocalization rate (vocalizations/min) 0.06 0.18 0.35 .73 24 Habitat index −0.84 0.55 −1.52 .13 44 Vegetation index 0.83 0.60 1.37 .17 44 Variable Female condition models Male condition models Habitat models (b) 1.00 0.75 Count 1 0.50 2 15 0.25 Proportion EPO in nest Proportion EPO in nest (a) 1.00 0.75 Count 1 0.50 2 0.25 0.00 0.00 4 Urban 3 2 1 0 1 Rural Habitat index 3 2 Deciduous 1 0 1 2 Coniferous Vegetation index F I G U R E 2 The proportion of extra-­pair offspring (EPO) in a nest was not influenced by either (a) the habitat index or (b) vegetation index of the nesting habitat. These indices were derived to describe (a) the degree of habitat urbanization and overall quality, and (b) vegetation composition, canopy cover and food availability In contrast to other studies, we failed to find a significant effect of habitat on EPP in mountain chickadees. Several stud- habitat urbanization and quality, and vegetation composition and food availability, respectively. ies investigating habitat and EPP have considered differences in Other studies with sites bordering urban and suburban neigh- food availability in otherwise similar habitat as the determinant of bourhoods have found effects of urbanization on EPP. In blue tits, habitat quality (Hoi-­L eitner et al., 1999; Rubenstein, 2007; Václav Kempenaers, Borgström, Loës, Schlicht, and Valcu (2010) found ar- et al., 2003). Here, we asked how EPP varied along a gradient tificial night lighting (i.e., street lights) in suburban habitat influences from rural to urban habitat. Unlike other studies, we had no direct EPP. Compared to males with territories in interior forest, males in measure of habitat quality (e.g., food availability). However, urban edge habitat bordering lighted suburban neighbourhoods were not habitat appears to be of slightly better quality than rural habitat in any better condition, but were more successful at gaining pater- to the mountain chickadees in our study population (Marini et al., nity in other nests. In addition, males in lighted territories began 2017), possibly because of greater food availability due to the singing earlier. In natural, forest habitat, female blue tits engage in presence of bird feeders and greater deciduous tree content. Still, EPCs with early-­singing males, suggesting the timing of dawn singing we did not find EPP to be related to the habitat index or vegetation may be an indicator of male quality (Kempenaers et al., 1992). Thus, composition of nesting sites, indices that were derived to describe females may have perceived early-­singing males in suburban habitat 8 | BONDERUD et al. as being high quality and, consequently, sought these males as EP partners. As in our study population (Marini et al., 2017), Kempenaers et al. (2010) found females nesting in suburban territories began laying eggs earlier. Females should time breeding so that peak nestling food demand aligns with peak food availability; however, as the authors suggest, earlier laying may have led to a mismatch between the two in suburban habitat (Kempenaers et al., 2010). Thus, rather than indicating better habitat quality, as we had speculated, earlier clutch initiation in urban habitat may be a maladaptive behaviour instigated by features of urban habitat (e.g., artificial lighting). Investigations of the relative timing of breeding in relation to insect abundance peaks in either habitat would have to be conducted to discern whether these mismatches occur. Together, these examples illustrate the complex dynamics of urban habitat and suggest the differences we have previously observed between mountain chickadees in urban and rural habitat (see Marini et al., 2017) may represent more than simply differences in habitat quality. AC K N OW L E D G E M E N T S We acknowledge this research was conducted on the traditional territory of the Tk’emlúps te Secwépemc and Skeetchestn First Nations. We thank the City of Kamloops for permission to conduct research in Kenna Cartwright Park, and members of the Kamloops Naturalist Club and other participating citizens for permission to erect and monitor backyard nest boxes. We also thank B.W. Murray for the use of laboratory space at the University of Northern British Columbia, M. Prill and other members of the Burg Lab at the University of Lethbridge for their assistance with laboratory work, and J.M. Bailey and members of the BEAC Lab at Thompson Rivers University for their assistance in the field. Funding was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC) through a Canada Graduate Scholarship to E.S.B, and through Discovery Grants to M.W.R. and K.A.O., and by the University of Northern British Columbia through a Graduate Entrance Scholarship, a Graduate Entrance Research Award, and a Research Project Award to E.S.B. CONFLIC T OF INTEREST We declare no conflicts of interest regarding this research. ORCID Erica S. Bonderud http://orcid.org/0000-0001-8307-3755 REFERENCES Avisoft Bioacoustics. (2017). Avisoft-SASLab Pro 5.2.12. Germany. Retrived from https://www.avisoft.com/. Bailly, J., Scheifler, R., Berthe, S., Clément-Demange, V.-A., Leblond, M., Pasteur, B., & Faivre, B. (2016). From eggs to fledging: Negative impact of urban habitat on reproduction in two tit species. Journal of Ornithology, 157, 377–392. https://doi.org/10.1007/ s10336-015-1293-3 Blewett, C. M., & Marzluff, J. M. (2005). Effects of urban sprawl on snags and the abundance and productivity of cavity-­ nesting birds. The Condor, 107, 678–693. https://doi.org/10.1650/0010-5422(2005)10 7[0678:EOUSOS]2.0.CO;2 Bonderud, E. S., Otter, K. A., Marini, K. L. D., Murray, B. W., Burg, T. M., & Reudink, M. W. (2017). Effects of parental condition and nesting habitat on sex allocation in the mountain chickadee. Behaviour, 154(11), 1101–1121. https://doi.org/10.1163/15685 39X-00003460 Delhey, K., Peters, A., Johnsen, A., & Kempenaers, B. (2007). Fertilization success and UV ornamentation in blue tits Cyanistes caeruleus: Correlational and experimental evidence. Behavioral Ecology, 18, 399–409. https://doi.org/10.1093/beheco/arl096 Dixon, K. L. (1965). Dominance-­ subordination relationships in mountain chickadees. The Condor, 67, 291–299. https://doi. org/10.2307/1365577 Google Inc. (2015). Google Earth. Retrived from https://www.google. com/earth/. Grava, A., Grava, T., Didier, R., Lait, L. A., Dosso, J., Koran, E., … Otter, K. A. (2012). Interspecific dominance relationships and hybridization between black-­c apped and mountain chickadees. Behavioral Ecology, 23, 566–572. https://doi.org/10.1093/beheco/arr229 Grava, T., Grava, A., & Otter, K. A. (2009). Supplemental feeding and dawn singing in black-­c apped chickadees. The Condor, 111, 560–564. https://doi.org/10.1525/cond.2009.080061 Griffith, S. C., Owens, I. P. F., & Thuman, K. A. (2002). Extra pair paternity in birds: A review of interspecific variation and adaptive function. Molecular Ecology, 11, 2195–2212. Hanotte, O., Zanon, C., Pugh, A., Greig, C., Dixon, A., & Burke, T. (1994). Isolation and characterization of microsatellite loci in a passerine bird: The reed bunting Emberiza schoeniclus. Molecular Ecology, 3, 529–530. https://doi.org/10.1111/j.1365-294X.1994.tb00133.x Hoi-Leitner, M., Hoi, H., Romero-Pujante, M., & Valera, F. (1999). Female extra-­pair behaviour and environmental quality in the serin (Serinus serinus): A test of the “constrained female hypothesis”. Proceedings of the Royal Society of London, Series B: Biological Sciences, 266, 1021– 1026. https://doi.org/10.1098/rspb.1999.0738 Kalinowski, S. T., Taper, M. L., & Marshall, T. C. (2007). Revising how the computer program cervus accommodates genotyping error increases success in paternity assignment. Molecular Ecology, 16, 1099–1106. https://doi.org/10.1111/j.1365-294X.2007.03089.x Kempenaers, B., Borgström, P., Loës, P., Schlicht, E., & Valcu, M. (2010). Artificial night lighting affects dawn song, extra-­pair siring success, and lay date in songbirds. Current Biology, 20, 1735–1739. https://doi. org/10.1016/j.cub.2010.08.028 Kempenaers, B., Verheyen, G. R., & Dhondt, A. A. (1997). Extrapair paternity in the blue tit (Parus caeruleus): Female choice, male characteristics, and offspring quality. Behavioral Ecology, 8, 481–492. https:// doi.org/10.1093/beheco/8.5.481 Kempenaers, B., Verheyen, G. R., Van den Broeck, M., Burke, T., Van Broeckhoven, C., & Dhondt, A. (1992). Extra-­pair paternity results from female preference for high-­quality males in the blue tit. Nature, 357, 494–496. https://doi.org/10.1038/357494a0 LaZerte, S. E., Otter, K. A., & Slabbekoorn, H. (2017). Mountain chickadees adjust songs, calls and chorus composition with increasing ambient and experimental anthropogenic noise. Urban Ecosystems, 20, 989–1000. https://doi.org/10.1007/s11252-017-0652-7 Lubjuhn, T., Gerken, T., Brün, J., & Epplen, J. T. (1999). High frequency of extra-­pair paternity in the coal tit. Journal of Avian Biology, 30, 229– 233. https://doi.org/10.2307/3677134 Marini, K. L. D., Otter, K. A., LaZerte, S. E., & Reudink, M. W. (2017). Urban environments are associated with earlier clutches and | 9 BONDERUD et al. faster nestling feather growth compared to natural habitats. Urban Ecosystems, 20, 1291–1300. https://doi.org/10.1007/ s11252-017-0681-2 Marzluff, J. M. (1997). Effects of urbanization and recreation on songbirds. In W. M. Block, & D. M. Finch (Eds.), Songbird ecology in southwestern ponderosa pine forests: A literature review (pp. 86–102). Fort Collins, CO: United States Department of Agriculture. McCallum, D. A., Grundel, R., & Dahlsten, D. L. (1999). Mountain chickadee (Poecile gambeli). In P. G. Rodewald (Ed.), The birds of North America. Ithaca, NY: Cornell Lab of Ornithology. Mennill, D. J., Ramsay, S. M., Boag, P. T., & Ratcliffe, L. M. (2004). Patterns of extrapair mating in relation to male dominance status and female nest placement in black-­c apped chickadees. Behavioral Ecology, 15, 757–765. https://doi.org/10.1093/beheco/arh076 Otter, K., Chruszcz, B., & Ratcliffe, L. (1997). Honest advertisement and song output during the dawn chorus of black-­capped chickadees. Behavioral Ecology, 8, 167–173. https://doi.org/10.1093/ beheco/8.2.167 Otter, K., McGregor, P. K., Terry, A. M. R., Burford, F. R. L., Peake, T. M., & Dabelsteen, T. (1999). Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London B: Biological Sciences, 266, 1305–1309. https://doi.org/10.1098/ rspb.1999.0779 Otter, K., & Ratcliffe, L. (1996). Female initiated divorce in a monogamous songbird: Abandoning mates for males of higher quality. Proceedings of the Royal Society of London, Series B: Biological Sciences, 263, 351– 355. https://doi.org/10.1098/rspb.1996.0054 Otter, K., Ratcliffe, L., & Boag, P. T. (1994). Extra-­pair paternity in the black-­c apped chickadee. The Condor, 96, 218–222. https://doi. org/10.2307/1369083 Otter, K., Ratcliffe, L., Michaud, D., & Boag, P. T. (1998). Do female black-­ c apped chickadees prefer high-­ ranking males as extra-­ pair partners? Behavioral Ecology and Sociobiology, 43, 25–36. https://doi. org/10.1007/s002650050463 Poesel, A., Kunc, H. P., Foerster, K., Johnsen, A., & Kempenaers, B. (2006). Early birds are sexy: Male age, dawn song and extrapair paternity in blue tits, Cyanistes (formerly Parus) caeruleus. Animal Behavior, 72, 531–538. https://doi.org/10.1016/j.anbehav.2005.10.022 Pyle, P. (1997). Identification guide to North American birds. Point Reyes Station, CA: Slate Creek Press. R Core Team. (2016). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.https:// www.R-project.org/. Ramsay, S. M., Otter, K. A., Mennill, D. J., Ratcliffe, L. M., & Boag, P. T. (2000). Divorce and extrapair mating in female black-­ c apped chickadees (Parus atricapillus): Separate strategies with a common target. Behavioral Ecology and Sociobiology, 49, 18–23. https://doi. org/10.1007/s002650000270 Ratcliffe, L. M., Mennill, D. J., & Schubert, K. A. (2007). Social dominance and fitness in black-capped chickadees. In K. A. Otter (Ed.), The ecology and behaviour of chickadees and titmice: An integrated approach (pp. 131–146). Oxford, UK: Oxford University Press. https://doi. org/10.1093/acprof:oso/9780198569992.001.0001 Reudink, M. W., Mech, S. G., & Curry, R. L. (2006). Extrapair paternity and mate choice in a chickadee hybrid zone. Behavioral Ecology, 17, 56–62. https://doi.org/10.1093/beheco/ari098 Rubenstein, D. R. (2007). Territory quality drives intraspecific patterns of extrapair paternity. Behavioral Ecology, 18, 1058–1064. https://doi. org/10.1093/beheco/arm077 Schubert, K. A., Mennill, D. J., Ramsay, S. M., Otter, K. A., Boag, P. T., & Ratcliffe, L. M. (2007). Variation in social rank acquisition influences lifetime reproductive success in black-­ c apped chickadees. Biological Journal of the Linnean Society, 90, 85–95. https://doi. org/10.1111/j.1095-8312.2007.00713.x Searcy, W. A., Peters, S., & Nowicki, S. (2004). Effects of early nutrition on growth rate and adult size in song sparrows Melospiza melodia. Journal of Avian Biology, 35, 269–279. https://doi. org/10.1111/j.0908-8857.2004.03247.x Smith, S. M. (1988). Extra-­ pair copulations in black-­ c apped chickadees: The role of the female. Behaviour, 107, 15–23. https://doi. org/10.1163/156853988X00160 Smith, S. B., McKay, J. E., Murphy, M. T., & Duffield, D. A. (2016). Spatial patterns of extra-­pair paternity for spotted towhees Pipilo maculatus in urban parks. Journal of Avian Biology, 47, 815–823. https://doi. org/10.1111/jav.00931 Southwood, T. R. E. (1961). The number of species of insects associated with various trees. Journal of Animal Ecology, 30, 1–8. https://doi. org/10.2307/2109 StataCorp. (2015). Stata Statistical Software: Release 14. College Station, TX: StataCorp LP. The GIMP Team. (2015). GNU Image Manipulation Program 2.8.16. Retrived from https://www.gimp.org/. Václav, R., Hoi, H., & Blomqvist, D. (2003). Food supplementation affects extrapair paternity in house sparrows (Passer domesticus). Behavioral Ecology, 14, 730–735. https://doi.org/10.1093/beheco/arg059 Wang, M. T., Hsu, Y. C., Yao, C. T., & Li, S. H. (2005). Isolation and characterization of 12 tetranucleotide repeat microsatellite loci from the green-­backed tit (Parus monticolus). Molecular Ecology Notes, 5, 439– 442. https://doi.org/10.1111/j.1471-8286.2005.00955.x How to cite this article: Bonderud ES, Otter KA, Burg TM, Marini KLD, Reudink MW. Patterns of extra-pair paternity in mountain chickadees. Ethology. 2018;00:1–9. https://doi.org/10.1111/eth.12747