Climate change affects the duration of the reproductive season in birds

Transcription

1 Journal of Animal Ecology 2010, 79, doi: /j x Climate change affects the duration of the reproductive season in birds A. P. Møller 1,2 *, E. Flensted-Jensen 3, K. Klarborg 4, W. Mardal 5 and J. T. Nielsen 6 1 Laboratoire d Ecologie, Syste matique et Evolution, CNRS UMR 8079, Universite Paris-Sud, Baˆtiment 362, F Orsay Cedex, France; 2 Center for Advanced Study, Drammensveien 78, NO-0271 Oslo, Norway; 3 Cypresvej 1, DK-9700 Brønderslev, Denmark; 4 Skovvej 28, DK-9490 Pandrup, Denmark; 5 Legindvej 102, Sønderha, DK-7752 Snedsted, Denmark; and 6 Espedal 4, Tolne, DK-9870 Sindal, Denmark Summary 1. The duration of the reproductive season may depend on the duration of the growing season, with recent amelioration in spring temperatures allowing earlier start of reproduction. Earlier start of reproduction may allow a longer breeding season because of more broods a longer interval between broods for multi-brooded species. 2. We analysed extensive long-term data sets on timing of breeding in 20 species of birds from Denmark, based on records of over individual offspring, showing considerable heterogeneity among species in temporal change in duration of the breeding season. 3. Multi-brooded species increased the duration of their breeding season by 0Æ43 days year )1 while single-brooded species decreased the duration of their breeding season by 0Æ44 days year )1.This implies that recent climate change has allowed more broods or better temporal spacing of broods in multi-brooded species, while the time window for reproduction has become narrower in single-brooded species. 4. The single-most important predictor of change in duration of the breeding season was change in the date breeding started; there was no change in the date of end of breeding. Species advancing their breeding date the most also expanded the duration of the breeding season. In contrast, longdistance migration and generation time did not predict change in duration of the breeding season. Key-words: breeding season, comparative analyses, migration, multi-brooded species, variance in timing of breeding Introduction Climate change has occurred at an unprecedented rate during the last century, with mean temperature increases exceeding over two degrees in certain areas of the temperate and arctic climate zones (Houghton et al. 2001). Furthermore, extreme weather phenomena such as extreme temperatures and rates of precipitation are occurring at an increasing rate. However, there is a geographically uneven distribution of climate change even within these zones. Many different organisms have responded to these changes in environmental conditions either through phenotypic plasticity and or evolutionary responses (e.g. Walther et al. 2002; Parmesan & Yohe 2003; Root et al. 2003; Møller, Fiedler & Berthold 2004; Dunn 2004). For example, Crick & Sparks (1999) showed that median egg laying dates had advanced in many species, although it remains unclear whether this was due to an advance in laying date or a decrease in the duration of the breeding season. *Correspondence author. anders.moller@u-psud.fr Przybylo, Sheldon & Merilä (2000) showed no consistent change in laying date of collared flycatchers Ficedula albicollis Temm. during 25 years, although laying was earlier after warm winters. Sanz et al. (2003) showed that laying dates of pied flycatchers Ficedula hypoleuca L. did not change with increasing temperatures. Dunn & Winkler (1999) showed heterogeneous patterns of advanced laying dates among populations of tree swallows Tachycineta bicolor Vieill. in North America. Therefore, birds have advanced breeding, with substantial heterogeneity among species and areas (review in Dunn 2004). Likewise, many studies have shown significant changes in clutch size or reproductive success linked to climate change (review in Dunn 2004). Furthermore, several studies have shown dramatic changes in body size in birds during a period of only 50 years, mainly in resident species (Yom-Tov 2001; Yom-Tov, Benjamini & Kark 2002). These changes have been suggested to reflect the impact of abiotic conditions on the intensity of natural selection on body size in the same way as Bergman s rule of latitudinal trends in body size can be interpreted as an outcome of consistent Ó 2010 The Authors. Journal compilation Ó 2010 British Ecological Society

2 778 A. P. Møller et al. latitudinal patterns of natural selection (Yom-Tov 2001; Yom-Tov et al. 2002). Recently, Tryjanowski et al. (2004) have shown a rapid increase in egg size in red-backed shrikes Lanius collurio L. and linked this change to increasing temperatures during the breeding season. The changes in body size reported by Yom-Tov (2001), Yom-Tov et al should likewise have consequences for egg size, because large females tend to lay large eggs. Almost all published information on changes in phenology relates to changes in mean breeding dates, with large amounts of variation among species (e.g. Dunn 2004; Dunn & Winkler 2010). In contrast, there is little information on changes in variance in reproductive variables, including breeding date. The variance in breeding date is directly related to the duration of the breeding season because a larger variance implies a longer breeding season. Perrins (1970, 1991) and Drent (2006) have provided extensive evidence consistent with the hypothesis that many individuals lay too late for the seasonal food peak, and that this may be due to a shortage of food, or a trade-off with other components of fitness. This is unfortunate because patterns of selection will depend both on means and variances and temporal patterns in means and variances in breeding date. Two different hypothetical patterns are possible. First, the timing of reproduction may change as a consequence of changing climatic conditions, with breeding now taking place earlier than before, without a change in the duration of the breeding season (the period during which clutches are started). Second, the breeding season may have advanced, while the duration also increased. In other words, mean timing or variance in timing may change in response to climate change. Whether one or the other of these possibilities reflects current conditions remain unknown. Winkler, Dunn & McCulloch (2002) suggested that changes in laying date during global warming may affect both the variance and the mean date of reproduction, and they showed that the variance in laying date of tree swallows was smaller in the warmest and hence the earliest years. In another study, Møller (2007a) showed for the barn swallow Hirundo rustica L., that commonly has two broods per year, that duration of the interval between first and second broods increased by more than a week during a period of 35 years, when climate during spring ameliorated. This change in inter-clutch interval was associated with an increase in annual fecundity, a change in quality of offspring, and a change in survival prospects of adults (Møller 2007a). Thus, there was an overall increase in the duration of the entire breeding season because the interval between broods increased. Husby, Kruuk & Visser (2009) showed for a population of great tits Parus major L. that the frequency of second broods has decreased in recent years. There have been few other attempts to investigate anything but first dates or mean dates, so little is known about changes in variances and higher modes of distributions during periods of climate change when the intensity of selection is likely to be strong. The objectives of this study were to investigate how variation in breeding date has changed in common birds. We investigated patterns of timing of breeding during a period of 38 years in response to recently increasing spring temperatures, relying on ringing records of nestlings and direct observations of nests as a source of information on timing of breeding. More specifically, we assessed (i) how variance in breeding date has changed during this period, (ii) how changes in variance in breeding date were associated with changing climatic conditions, and (3) how rate of change in variance in breeding date could be predicted by change in mean breeding date, migration distance and number of broods per season. Previous studies have suggested that longdistance migration may constrain phenological response of breeding to climate change (Both & Visser 2001), that number of clutches and interval between these may depend on start of the growing season (Møller 2007a), and that mean life span (and hence generation time) may affect the ability of different species to respond to climate change. Materials and methods BREEDING DATE We recorded breeding dates for 20 species of birds from specific study populations or national surveys (the latter for sandwich tern Sterna sandvicensis Lath., Arctic tern Sterna paradisaea Pont., common tern Sterna hirundo L., little tern Sterna albifrons L. and starling Sturnus vulgaris L.). KK studied hole nesting species in nest boxes around Rødhus (57 12 N, 9 40 E), Denmark during Nest boxes were mainly located in coniferous forests, and they were checked weekly for presence of eggs and nestlings. Breeding date was recorded as the date when nestlings were banded at an approximate age of 7 days, assuming that one egg was laid per day, and that the incubation period was 15 days (coal tit Parus ater L.), 14Æ2 days (blue tit Parus caeruleus L.), 13Æ9 days (great tit), 14 days (marsh tit Parus palustris L.), and 13 days (redstart Phoenicurus phoenicurus L.) (Cramp & Perrins ). APM studied wood pigeon Columba palumbus L., corvids, barn swallow, house martin Delichon urbica L., blackbird Turdus merula L. and sparrows around Kraghede (57 12 N, E), Denmark, The same nest sites were checked annually and timing of breeding was determined using back-dating based on the size of nestlings and the number of eggs present during nest checks, assuming incubation and nestling periods had durations as reported by Cramp & Perrins ( ). JTN studied sparrowhawk Accipiter nisus L. ( ) and goshawk Accipiter gentilis L in Vendsyssel, Denmark, checking all potential nest sites within the study areas for active nests. Breeding dates were determined from regular nest visits, back-dating from the size of nestlings, assuming incubation and nestling periods had durations as reported by Cramp & Perrins ( ). See Nielsen & Møller (2006) and Møller & Nielsen (2007) for further details. For the 15 species that were followed in population studies we searched extensively for nests within the study sites. For the five species based on national surveys we extracted information on banding dates for young or nestlings from Denmark for the period , using all published information and information provided by bird banders. We also recorded all ringing information from local ornithological publications (Danske Fugle, Dansk Ornithologisk Forenings Tidsskrift, Videnskabelige Meddelelser fra dansk Naturhistorisk Forening, Meddelelser Til Ringmærkerne, Preuss &

3 Climate and reproductive season 779 Andersen-Harild (1981), Pedersen & Rahbek (1999), and Zoological Museum ( )), and we supplemented this information with all ringing data from bird banders in Northern Jutland. As young or nestlings can be banded during a limited period from an age of a couple of days to some days before fledging, ringing date provides a reliable indication of breeding date. We assumed an incubation and a nestling period of 25 and 29 days for sandwich tern, 22Æ2 and 22Æ5 days for Arctic tern, 21Æ5 and 25 days for common tern, 20 and 19Æ5 days for little tern and 12Æ2 and 21 days for starling (Cramp & Perrins ). We assumed that young were banded half way through the nestling period, that each clutch consisted of two eggs, with the exception of starlings, where we assumed five eggs, and that the interval between eggs was 2 days, with the exception of the starling, where we assumed an interval of 1 day (Cramp & Perrins ). This allowed us to calculate the start of laying for each brood. For each year and each species we estimated the 10th and the 90th percentile of the breeding dates (with 1 = May 1), and the interval between these two dates was defined as the duration of the breeding season. For the barn swallow where we had information on first and second broods separately, because the population was colourbanded, we made these calculations separately for first and second broods, and for the two broods combined. We defined the duration of the breeding season from the 10th and the 90th percentile rather than the first and the last laying date of a year, because such extreme dates may be particularly subject to sampling effort. However, the estimates based on the percentile and the estimates based on the duration of the breeding season including all observations were strongly positively correlated (Pearson product-moment correlation weighted by sample size, r = 0Æ84, t = 36Æ84, d.f. = 553, P < 0Æ0001). Hence, the two estimates of duration of the breeding season provided similar information. A total of 9825 broods and chicks formed the basis for the present study (Table 1). Table 1. Models of the relationship between duration of the breeding season (days) and year for first clutch, second clutch and the entire breeding season of barn swallows Variable Sum of squares d.f. F P Slope (SE) Start of breeding First clutch 206Æ82 1 9Æ03 0Æ0049 )0Æ221 (0Æ074) Error 801Æ62 35 Second clutch 707Æ91 1 9Æ09 0Æ0048 )0Æ410 (0Æ136) Error 2727Æ11 35 Both 142Æ03 1 6Æ21 0Æ018 )0Æ183 (0Æ074) Error 799Æ86 35 End of breeding First clutch 0Æ26 1 0Æ01 0Æ94 Error 1503Æ63 35 Second clutch 5Æ70 1 0Æ26 0Æ61 Error 759Æ11 35 Both 5Æ70 1 0Æ26 0Æ61 Error 759Æ11 35 Duration of breeding season First clutch 221Æ69 1 4Æ87 0Æ034 0Æ229 (0Æ104) Error 1594Æ04 35 Second clutch 90Æ84 1 2Æ75 0Æ11 0Æ148 (0Æ089) Error 1156Æ89 35 Both 143Æ87 1 4Æ84 0Æ034 0Æ185 (0Æ084) Error 1039Æ70 35 A summary of the data is provided in Appendix S1, Supporting information. LIFE HISTORY Complete nesting failure We recorded complete nesting failure as the complete loss of a clutch or a brood between two nest checks for 15 of the 20 species. Incubation and nestling periods APM checked most barn swallow nests daily around the presumed period of hatching and fledging. Incubation period was estimated as the number of days between the first day when all eggs have hatched and the day when all eggs have been laid. Nestling period was estimated as the number of days between the first day when all nestlings have left the nest (even if sitting just outside the nest) and the first day when all eggs have hatched. Number of clutches We recorded the number of clutches per year as the maximum recorded in Denmark, supplemented with information in Cramp & Perrins ( ). Adult survival rate We recorded annual adult survival rate using Cramp & Perrins ( ) as a source. Generation time was estimated as T = A + P (1 ) P), where A is age at first reproduction (in this case 1 3 years) and P is adult annual survival rate. Migration distance Migration distance was estimated as the mean of the northernmost and southernmost latitude of the breeding range minus the mean of the northernmost and southernmost latitude of the winter range, using Cramp & Perrins ( ) as a source, providing an estimate of migration distance in degrees latitude. Body mass We used adult body mass from the breeding season from information in Cramp & Perrins ( ). Summary data are reported in Appendix S1, Supporting information. CLIMATE VARIABLES We used mean monthly temperature for March July in Denmark based on data recorded by the Danish Meteorological Institute in Copenhagen for predicting rate of change in timing of breeding and duration of the breeding season. STATISTICAL ANALYSES Migration distance, generation time, number of clutches per season, and body mass was log 10 -transformed before analyses. We made all analyses with JMP (SAS 2000). We developed general linear models to quantify change in mean breeding date with year and temperature, respectively. Climatic conditions were

4 780 A. P. Møller et al. quantified as mean temperature in the month when breeding started. We analysed separately first and second clutches and both combined for the barn swallow in an attempt to understand the impact of first and second broods on changes in phenology and duration of the breeding season. Closely related species are more likely to have similar phenotypes than species that are more distantly related. Therefore, species cannot be treated as statistically independent observations in comparative analyses because apparent phenotypic correlations among species may result from species sharing a common ancestor rather than convergent evolution. We controlled for similarity in phenotype among species due to common phylogenetic descent by calculating standardized independent linear contrasts (Felsenstein 1985), using the software CAIC (Purvis & Rambaut 1995). All branches of the phylogenies were assigned uneven branch lengths, assuming a gradual evolution model as implemented in the software, although a second set of analyses based on similar branch length produced qualitatively similar results to those reported here. We tested the statistical and evolutionary assumptions of the comparative analyses (Garland, Harvey & Ives 1992) by regressing absolute standardized contrasts against their standard deviations. To test for effects of problems of heterogeneity in variance: (i) we excluded outliers (contrasts with Studentized residuals >3) in a second series of analyses (Jones & Purvis 1997), and (ii) analyses were repeated with the independent variable expressed in ranks. These analyses are conservative tests of the null hypothesis, explicitly investigating the robustness of the conclusions. In neither case did these new analyses change any of the conclusions, and they are therefore not reported here. The composite phylogeny used in the comparative analyses was based on Sibley & Ahlquist (1990), combined with information from Jønsson & Fjeldså (2006) and Bridge, Jones & Baker (2005) (Appendix S2, Supporting information). Because information for the composite phylogeny originated from different studies using different molecular and phylogenetic methods, consistent estimates of branch lengths were unavailable. Therefore, branch lengths were transformed assuming a gradual model of evolution with branch lengths being proportional to the number of species contained within a clade. The results from the phylogenetic analyses were also qualitatively similar to those found when making the calculations using the taxonomy of Sibley & Monroe (1990). Values reported are means (SE). Results CHANGE IN MEAN DATE OF REPRODUCTION AND DURATION OF THE BREEDING SEASON Mean temperature in April increased dramatically during , while that was not the case for mean temperature in May [April: F =24Æ75, d.f. = 1, 36, r 2 =0Æ41, P < 0Æ0001, slope (SE) = 0Æ071 (0Æ014); May: F =1Æ23, d.f. = 1, 36, r 2 =0Æ03, P =0Æ27, slope (SE) = 0Æ017 (0Æ016)]. This amounted to an increase in temperature by 2Æ7 C in April and 0Æ6 CinMay. Nine of the 20 species showed a significant advance in mean laying date and one a significant delay. For all 20 species taken together, the change in mean breeding date differed significantly among species (species year interaction: F =4Æ68, d.f. = 19, 515, P <0Æ0001). Change in start of the breeding season varied among species (species year interaction: F =5Æ85, d.f. = 19, 515, P <0Æ0001). Finally, there was a significant change in the 90th percentile of the breeding season during the study period among species (year species interaction: F = 8Æ93, d.f. = 19, 515, P <0Æ0001). The duration of the breeding season varied significantly during the study among species (species year interaction: F =10Æ37, d.f. = 19, 515, P <0Æ0001). The duration of the breeding season expanded by on average 0Æ18 days year )1, or in total 6Æ7 days (SE = 2Æ3 days) during , ranging from a shortening of 0Æ94 days year )1 (or 35Æ5 days in total) (SE = 12Æ6 days, sandwich tern) to an elongation of 0Æ95 days year )1 [or in total 36Æ2 days (SE = 3Æ9 days, wood pigeon)] (Appendix S1, Supporting information). Examples of changes in duration of the breeding season for three species with a shortening pattern, no change, and an expanding pattern are shown in Fig. 1. Species that bred late in the year did not significantly expand the duration of their breeding season more than early breeders [F =3Æ23, d.f. = 1, 18, r 2 =0Æ15, P =0Æ09, slope (SE) = 0Æ0075 (0Æ0042)]. Furthermore, species that advanced the date of breeding the most did not expand their breeding season the most (F =0Æ72, d.f. = 1, 18, r 2 =0Æ04, P =0Æ41). CHANGE IN PHENOLOGY, DURATION OF THE BREEDING SEASON, CLUTCH SIZE, REPRODUCTIVE FAILURE AND DEVELOPMENTAL PERIODS IN THE BARN SWALLOW The barn swallow was the only multi-brooded species in which we were able to analyse the importance of second clutches in determining the duration of the breeding season. The 10th percentile of breeding date advanced significantly for both first and second clutches of the barn swallow, while there was no significant trend for the 90th percentile of breeding date for either clutch (Table 1). Therefore, there was an increase in the duration of the breeding season for first clutches and both clutches combined, but not for second clutches (Table 1). Therefore, both clutches combined produced a rate of expansion in duration of the breeding season that was qualitatively similar to the rate of change for first and second clutches separately (Table 1). Total breeding failure in the barn swallow did not change significantly during (the proportion of failure square-root arcsine-transformed: first clutch: F = 1Æ83, d.f. = 1, 36, r 2 =0Æ05, P =0Æ19; second clutch: F =0Æ16, d.f. = 1, 36, r 2 <0Æ01, P =0Æ69). Hence, there were no clear trends in breeding failure that could have affected the duration of the breeding season. The duration of the incubation period did not change significantly over the years (5138 incubation periods ; F <0Æ01,

5 Climate and reproductive season 781 (a) Duration of season (days) (b) Duration of breeding season (days) (c) Duration of season (days) Year Year Year Fig. 1. Duration of the breeding season (days) in (a) goshawk Accipiter gentilis during , (b) starling Sturnus vulgaris during and (c) wood pigeon Columba palumbus during The lines are the linear regression lines. d.f. = 1, 36, r 2 <0Æ01, P =0Æ96). The duration of the nestling period did not change significantly over the years (4830 nestling periods ; F =0Æ81, d.f. = 1, 36, r 2 =0Æ02, P =0Æ37). Hence, there were no clear temporal trends in incubation or nestling periods. The frequency of total breeding failure in all species except the terns did not change consistently during (the proportion of failure square-root arcsine-transformed: F <1Æ35, d.f. = 1, 36, r 2 <0Æ04, P >0Æ25). Change in duration of breeding season (days year 1 ) CHANGE IN DURATION OF THE BREEDING SEASON AND CLIMATE CHANGE The duration of the breeding season increased significantly with increasing mean temperature during the month when breeding started [mean slope (SE) = 1Æ73 days degree )1 (0Æ54), t = )3Æ19, d.f. = 19, P =0Æ006]. ECOLOGY, LIFE HISTORY AND DURATION OF THE BREEDING SEASON The duration of the breeding season expanded with the number of broods per season, with single-brooded species shortening their breeding seasons, while species with more broods expanded their breeding seasons (Fig. 2; Table 2). There was no significant quadratic effect of number of clutches in this analysis (F = 0Æ46, d.f. = 1, 16, P = 0Æ51). In addition, there was a weak, but non-significant effect of migration distance (Table 2). Discussion No. clutches per year Fig. 2. Rate of change in duration of the breeding season (days per year) in relation to maximum number of broods per year in different species. The main findings of this study were that the duration of the breeding season of 20 species of birds became significantly longer in eight species, significantly shorter in five species, and did not change significantly in seven species. Change in duration of the breeding season was directly linked to change in temperature during the breeding season. The number of broods produced per year, and perhaps weakly migration distance, but not other life-history variables accounted for these interspecific differences in change in duration of the breeding season. Mean temperature for March and April increased dramatically in our study areas since This change in start of the growing season in the temperate study area was associated with a general advance in phenology. We found extensive evidence of advancing start of the breeding season, but little or no evidence of change in termination of the breeding season. Therefore, the duration of the overall breeding season

6 782 A. P. Møller et al. Table 2. Best-fit models of the relationship between temporal change in duration of the breeding season and life history and ecological variables. The two-first models (A) that included all variables had the statistics F =6Æ52, d.f. = 4, 13, r 2 =0Æ67, P =0Æ0042 and F =4Æ17, d.f. = 4, 13, r 2 =0Æ24, P =0Æ022, while the two last models (B) that only included number of clutches and migration distance had the statistics F =17Æ54, d.f. = 2, 17, r 2 =0Æ67, P <0Æ0001 and F =9Æ46, d.f. = 2, 17, r 2 =0Æ36, P =0Æ0017 Variable Sum of squares d.f. F P Slope (SE) (A) Species No. clutches per year 1Æ Æ05 0Æ0074 0Æ46 (0Æ14) Migration distance 0Æ05 1 0Æ47 0Æ51 0Æ09 (0Æ13) Generation time 0Æ00 1 0Æ04 0Æ84 0Æ10 (0Æ48) Body mass 0Æ29 1 2Æ56 0Æ14 )0Æ30 (0Æ19) Error 1Æ49 13 Contrasts No. clutches per year 0Æ Æ41 0Æ0029 1Æ02 (0Æ28) Migration distance 0Æ00 1 0Æ14 0Æ71 0Æ06 (0Æ16) Generation time 0Æ02 1 0Æ82 0Æ38 0Æ56 (0Æ62) Body mass 0Æ06 1 1Æ98 0Æ18 )0Æ33 (0Æ24) Error 0Æ36 13 (B) Species No. clutches per year 2Æ Æ97 0Æ0002 0Æ39 (0Æ08) Migration distance 0Æ55 1 5Æ75 0Æ028 )0Æ27 (0Æ11) Error 1Æ61 17 Contrasts No. clutches per year 0Æ Æ57 0Æ0008 0Æ90 (0Æ22) Migration distance 0Æ07 1 3Æ09 0Æ10 )0Æ29 (0Æ17) Error 0Æ39 17 increased in recent decades, causing the breeding season by some species to expand by as much as 30 days during the study. The average expansion of the duration of the breeding season was 6Æ7 days during , with considerable variation among species, with the sandwich tern now having a breeding season lasting 36 days (SE = 13) less than in 1970 (a decrease by more than 70%), while the wood pigeon now has a breeding season lasting 36 days (SE = 4) longer than in 1970 (amounting to an increase by 48%). The changes in mean breeding date and change in duration of the breeding season were related to change in temperature in spring, consistent with the hypothesis that it is change in climatic conditions that is driving change in phenology. We found significant differences in change in duration of the breeding season among species, mainly caused by a change in the start of the breeding season. In contrast, there was less change in date of termination of the breeding season. Therefore, we investigated several potential predictors of such interspecific variation in response to climate change. Several hypotheses have been proposed to account for interspecific differences in response to climate change including long-distance migration, generation time and number of clutches produced per season (Both & Visser 2001; Dunn 2004; Møller 2007a). Crick, Gibbons & Magrath (1993) suggested that the number of clutches produced per year affected the optimal timing of reproduction, with species having more than a single brood starting reproduction earlier than the annual peak in food abundance, while single-brooded species matched their timing of reproduction with the annual peak. Møller (2007a) analysed time series of breeding date for the long-distance migratory barn swallow, showing that the interval between clutches increased with increasing spring temperature, causing a reduction in selection against two clutches per year. In other words, constraints on timing of reproduction due to the production of multiple clutches can be ameliorated by an increase in duration of the breeding season in a seasonal environment. Husby et al. (2009) have shown that the breeding season can decrease in duration because of a reduction in the frequency of second clutches. Here, we have shown that the number of clutches per season explains more than 60% of the variance in change in duration of the breeding season among species (Fig. 2; variance explained in model of contrasts = F (F + numerator d.f.) = 22Æ97 (22Æ ) = 0Æ64 (Rosenthal 1994)). This pattern is consistent with observations of timing of fall migration, showing that long-distance migrants that usually have a single clutch per year advanced their migration while short-distance migrants that usually have more than a single clutch per year delayed their migration (Jenni & Kéry 2003). It is also consistent with a negative relationship between phenological response to climate change and the number of broods in European birds (Møller, Rubolini & Lehikoinen 2008b). We suggest that the change in interval between clutches may account for these patterns, although clearly other mechanisms must be invoked for species with only a single clutch per year. The shorter duration of the breeding season for single-brooded species with increasing temperatures (Table 2) could potentially be predicted from higher temperatures shortening the duration of full gonadal maturation (Dawson 2005; Silverin et al. 2008). These results are consistent with the hypothesis that multiple clutches per year constitute a strong selective force expanding the duration of the breeding season. They also show that the duration of the breeding season is a labile life-history trait that can change rapidly in response to altered environmental conditions. Although this result is not surprising given latitudinal clines in duration of the breeding season among many species of birds (e.g. Cramp & Perrins ), this study constitutes the first to demonstrate such a correlation between climate change and duration of the breeding season. Many bird species are strongly affected by long-distance migration that has important implications for their annual cycle and the timing of different events during the annual cycle (Berthold 2001). Migrants will spend parts of the year in the breeding and the wintering grounds, with remaining parts being migration and stop-over, exposing such species to different weather systems than may not change in synchrony. Arrival date has been suggested to act as a constraint on timing of breeding, because breeding cannot advance further than arrival date to the breeding grounds (Both & Visser 2001). While this hypothesis is possible, several alternative interpretations also exist. These include the possibility that long-distance dispersal in migratory species (Paradis et al. 1998; Belliure et al. 2000) may prevent adaptation to local

7 Climate and reproductive season 783 changes in environmental conditions, and that sex differences in migration and selection may constrain timing of breeding (Møller 2004, 2007b). Saino & Ambrosini (2008) have shown that weather anomalies in Africa in fact can be predicted by weather anomalies in Europe, implying that migrants may have cues to climate conditions in Europe, even when in their African winter quarters. These factors do not seem to severely constrain evolution of the breeding season because migration distance did not significantly predict change in duration of the breeding season (Table 2). We assumed that the estimates of duration of the breeding season as derived from the time required for egg laying, incubation and care of nestlings according to the literature were unbiased (i.e. the duration of any of these parts of the breeding cycle does not change as a consequence of climate change). A previous study has shown that incubation and nestling periods may change in response to climate change (Cresswell & McCleery 2003). Here, we found no evidence of significant changes in duration of the incubation or the nestling period during in the barn swallow. Even if such changes may occur, there is generally little intraspecific variation in duration of developmental periods, with most individuals varying less than a couple of days (as in the barn swallow). Hence, such small differences cannot account for the differences in duration of the breeding season reported here. Likewise, complete nesting failure rates could change over time in response to climate change. If more nests failed as temperatures increased this could affect estimates of duration of the breeding season. However, our analyses of 15 species showed no significant evidence of such temporal change in failure rate. We found no clear evidence of body size, adult survival rate or generation time being related to change in duration of the breeding season (Table 2). Body size is related to many different life-history traits (e. g. Bennett & Owens 2002), and it was thus a likely candidate variable explaining variation in response to climate change. Furthermore, body mass is negatively related to the amount of standing genetic variation in local populations of birds (Møller, Garamszegi & Spottiswoode 2008a), thus potentially affecting response to climate change. We found little support for those expectations. In fact, the weak effect of body mass in the species-specific analyses and the lack of effect of generation time may suggest that the responses reported here are more likely to be caused by phenotypic plasticity rather than micro-evolutionary change. The present study has a number of implications for future studies of the effects of climate change. First, this study is one of the first to pinpoint an important factor associated with response to climate change. Additional factors need to be identified in order to improve our understanding of the factors that affect interspecific differences in response to climate change. Second, the present study did not suggest that generation time and hence the number of generations available for adjustment to climate change was important. This result is surprising if response to climate change is mainly caused by micro-evolutionary response to change, but not if phenotypic plasticity is the most important factor. This raises the question to which extent response to climate change can be attributed to micro-evolutionary or phenotypic response. Finally, the present study suggests that the interval between clutches constitutes an important constraint on fitness because a larger number of clutches is traded against a reduction in survival due to reproductive costs (Møller 2007a). Further studies of this trade-off and how it varies with changing climatic conditions may shed light on the factors constraining response to climatic conditions. In conclusion, we have presented extensive analyses for 20 species of birds demonstrating advancing breeding dates, but less change in the date of termination of breeding, implying that the duration of the breeding season has increased considerably in some species, but not at all or even contracted considerably in others. These changes in phenology and duration of the breeding season are directly linked to changes in mean temperature during spring. Interspecific differences in changes in duration of the breeding season are predicted by the number of clutches per season, rather than migration distance or generation time that have been suggested to predict response to climate change. Acknowledgements We are grateful for the efforts of the numerous volunteers among the bird banders who contributed to this study. N.O. Preuss at the Ringing Centre at the Zoological Museum, Copenhagen, Denmark, kindly provided information on banding dates and recoveries. Two anonymous referees kindly improved the paper. References Belliure, J., Sorci, G., Møller, A.P. & Clobert, J. (2000) Dispersal distances predict subspecies richness in birds. Journal of Evolutionary Biology, 13, Bennett, P.M. & Owens, I.P.F. (2002) Evolutionary Ecology of Birds. Oxford University Press, Oxford, UK. Berthold, P. (2001) Bird Migration. Oxford University Press, Oxford, UK. Both, C. & Visser, M.E. (2001) Adjustment of climate change is constrained by arrival date in a long-distance migrant bird. Nature, 411, Bridge, E.S., Jones, A.W. & Baker, A.J. (2005) A phylogenetic framework for the terns (Sternini) inferred from mtdna sequences: implications for taxonomy and plumage evolution. Molecular Phylogenetics and Evolution, 35, Cramp, S. & Perrins, C.M. (eds) ( ) The Birds of the Western Palearctic. Vols Oxford University Press, Oxford, UK. Cresswell, W. & McCleery, R. (2003) How great tits maintain synchronization of their hatch date with food supply in response to long-term variability in temperature. Journal of Animal Ecology, 72, Crick, H.Q.P. & Sparks, T.H. (1999) Climate change related to egg-laying trends. Nature, 399, 423. Crick, H.Q.P., Gibbons, D.W. & Magrath, R.D. (1993) Seasonal changes in clutch size in British birds. Journal of Animal Ecology, 62, Dawson, A. (2005) The effect of temperature on photoperiodically regulated gonadal maturation, regression and moult in starlings: potential consequences of climate change. Functional Ecology, 19, Drent, R.H. (2006) The timing of birds breeding seasons: the Perrins hypothesis revisited especially for migrants. Ardea, 94, Dunn, P. (2004) Breeding dates and reproductive performance. Effects of Climatic Change on Birds (eds A.P. Møller, W. Fiedler & P. Berthold), pp Elsevier, Amsterdam, the Netherlands. Dunn, P.O. & Winkler, D.W. (1999) Climate change has affected the breeding date of tree swallows throughout North America. Proceedings of the Royal Society of London B, 266,

8 784 A. P. Møller et al. Dunn, P.O. & Winkler, D.W. (2010) Effects of climate change on timing of breeding and reproductive success in birds. Effects of Climate Change on Birds (eds A.P. Møller, W. Fiedler & P. Berthold), in press. Oxford University Press, Oxford, UK. Felsenstein, J. (1985) Phylogenies and the comparative method. American Naturalist, 125, Garland, T. Jr, Harvey, P.H. & Ives, A.R. (1992) Procedures for the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology, 41, Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K. & Johnson, C.A. (eds) (2001) Climate Change 2001: The Scientific Basis, Cambridge University Press, Cambridge, UK. Husby, A., Kruuk, L.E.B. & Visser, M.E. (2009) Decline in the frequency and benefits of multiple brooding in great tits as a consequence of a changing environment. Proceedings of the Royal Society of London B, 276, Jenni, L. & Kéry, M. (2003) Timing of autumn bird migration under climate change: advances in long-distance migrants, delays in short-distance migrants. Proceedings of the Royal Society of London B, 270, Jones, K.E. & Purvis, A. (1997) An optimum body size for mammals? Comparative evidence from bats. Functional Ecology, 11, Jønsson, K.A. & Fjeldsa, J. (2006) A phylogenetic supertree of oscine passerine birds (Aves: Passeri). Zoologica Scripta, 35, Møller, A.P. (2004) Protandry, sexual selection and climate change. Global Change Biology, 10, Møller, A.P. (2007a) Interval between clutches, fitness and climate change. Behavioral Ecology, 18, Møller, A.P. (2007b) Tardy females, impatient males: protandry and divergent selection on arrival date in the two sexes of the barn swallow. Behavioral Ecology and Sociobiology, 61, Møller, A.P. & Nielsen, J.T. (2007) Malaria and risk of predation: A comparative study of birds. Ecology, 88, Møller, A.P., Fiedler, W. & Berthold, P. (eds) (2004) Effects of Climatic Change on Birds, Elsevier, Amsterdam, the Netherlands. Møller, A.P., Garamszegi, L.Z. & Spottiswoode, C. (2008a) Genetic similarity, distribution range and sexual selection. Journal of Evolutionary Biology, 21, Møller, A.P., Rubolini, D. & Lehikoinen, E. (2008b) Populations of migratory bird species that did not show a phenological response to climate change are declining. Proceedings of the National Academy of Science of the USA, 105, Nielsen, J.T. & Møller, A.P. (2006) Effects of food abundance, density and climate change on reproduction in the sparrowhawk Accipiter nisus. Oecologia, 149, Paradis, E., Baillie, S.R., Sutherland, W.J. & Gregory, R.D. (1998) Patterns of natal and breeding dispersal in birds. Journal of Animal Ecology, 67, Parmesan, C. & Yohe, G. (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, Pedersen, K.T. & Rahbek, C. (1999) Ringmærkning og Genfund af Ringmærkede Fugle Indsendt til Zoologisk Museum i 1996, Zoological Museum, Copenhagen, Denmark. Perrins, C.M. (1970) Timing of birds breeding seasons. Ibis, 112, 242. Perrins, C.M. (1991) Tits and their caterpillar food-supply. Ibis, 133, Preuss, N.O. & Andersen-Harild, P. (1981) Fugle Ringmærket i Danmark , Zoological Museum, Copenhagen, Denmark. Przybylo, R., Sheldon, B.C. & Merila, J. (2000) Climate effects on breeding and morphology: evidence for phenotypic plasticity. Journal of Animal Ecology, 69, Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts (CAIC). Computer and Applied Bioscience, 11, Root, T.L., Price, J.L., Hall, K.R., Schneider, S.H., Rosenzweig, C. & Pounds, A.J. (2003) Fingerprints of global warming on wild animals and plants. Nature, 421, Rosenthal, R. (1994) Parametric measures of effect size. The Handbook of Research Synthesis (eds H. Cooper & L.V. Hedges), pp Russel Sage Foundation, New York. Saino, N. & Ambrosini, R. (2008) Climatic connectivity between Africa and Europe may serve as a basis for phenotypic adjustment of migratory schedules of trans-saharan migratory birds. Global Change Biology, 14, Sanz, J.J., Potti, J., Moreno, J., Merino, S. & Frı as, O. (2003) Climate change and fitness components of a migratory bird breeding in the Mediterranean region. Global Change Biology, 9, SAS (2000) JMP. SAS Institute Inc., Cary, NC. Sibley, C.G. & Ahlquist, J.E. (1990) Phylogeny and Classification of Birds, a Study in Molecular Evolution. Yale University Press, New Haven, Conn and London, UK. Sibley, C.G. & Monroe, B.L. Jr (1990) Distribution and Taxonomy of Birds of the World. Yale University Press, New Haven, Conn and London, UK. Silverin, B., Wingfield, J., Stokkan, K.-A., Massa, R., Järvinen, A., Andersson, N.-Å., Lambrechts, M., Sorace, A. & Blomqvist, D. (2008) Ambient temperature effects on photo induced gonadal cycles and hormonal secretion patterns in Great Tits from three different breeding latitudes. Hormones and Behavior, 54, Tryjanowski, P., Sparks, T.H., Kuczynski, L. & Kuzniak, S. (2004) Should avian egg size increase as a result of global warming? A case study using the red-backed shrike (Lanius collurio). Journal of Ornithology, 145, Walther, G.-R., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T.J.C., Fromentin, J.-M., Høgh-Guldberg, O. & Bairlein, F. (2002) Ecological responses to recent climate change. Nature, 416, Winkler, D.W., Dunn, P.O. & McCulloch, C.E. (2002) Predicting the effects of climate change on avian life-history traits. Proceedings of the National Academy of Science USA, 99, Yom-Tov, Y. (2001) Global warming and body mass decline in Israeli passerine birds. Proceedings of the Royal Society of London B, 268, Yom-Tov, Y., Benjamini, Y. & Kark, S. (2002) Global warming, Bergmann s rule and body mass are they related? The chukar partridge (Alectoris chukar) case. Journal of Zoology, 257, Zoological Museum ( ) Ringmærkning og Genfund af Ringmærkede Fugle Indsendt til Zoologisk Museum. Zoological Museum, Copenhagen, Denmark. Received 16 September 2009; accepted 5 February 2010 Handling Editor: Simon Verhulst Supporting Information Additional Supporting Information may be found in the online version of this article. Appendix S1. Summary information on bird species analysed for changes in timing and duration of the breeding season. Appendix S2. Phylogenetic relationships among the 20 species of birds analysed for change in duration of the reproductive season. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.