Across many taxa, clinal variation in body size has been observed to follow Bergmann's rule, which predicts larger body size in colder climates. For more than a century, this pattern has typically been ascribed to selection for large body size in cold winter climates. Here, in spatially distributed observational data from 30 populations of House Sparrow (Passer domesticus) introduced into Australia and New Zealand, we show that this relationship appears to be explained by a negative relationship with high temperatures during the breeding season. Our results suggest that higher temperatures during the breeding season could reduce body size through developmental plasticity, which should be considered in combination with or as an alternative to selection. Our findings would predict that a hotter climate during breeding could drive significant changes in morphology among populations (and potentially within populations as well, if climate varies temporally across a breeding season). This idea, and our support for it, could account for much of the variation in body size that drives the well-observed patterns first described by Bergmann, and that are still largely attributed to selection on adult body size during cold winters. Understanding the mechanisms behind any climate-dependent developmental plasticity could prove useful for understanding how endotherms may be affected by climate change in the future.
For over 100 years since the publication of Bergmann's rule in 1847 (Bergmann 1847), a clinal pattern of animals having larger body sizes in colder climates has been observed in a majority of the hundreds of species that have been examined (Mayr 1956, James 1970, Ashton et al. 2000, Ashton 2002, Meiri and Dayan 2003, Millien et al. 2006, Clauss et al. 2013, Teplitsky and Millien 2014). To date, most studies still cite Bergmann's original explanation that larger body size is favored by natural selection in colder climates because of the thermoregulatory benefits of a smaller volume to surface area ratio (Briscoe et al. 2015, Cardilini et al. 2016, Salewski and Watt 2017). A classic example of clinal variation in avian body size has previously been demonstrated in North American populations of the introduced House Sparrow (Passer domesticus; Johnston and Selander 1964, 1973, Murphy 1985). If winter temperatures are the selective force responsible for this clinal variation, as predicted by Bergmann's rule, then variation in body size between populations should be best explained by winter minimum temperatures. However, in hotter climates, smaller body size can also be advantageous to an individual's ability to thermoregulate by dissipating heat (Partridge and Coyne 1997), even though the benefits of minor changes in body size within species have been questioned for more than 40 yr (Scholander 1955, McNab 1971).
Understanding the mechanisms that create the morphological differentiation described by Bergmann's rule has gained fresh impetus as part of the study of the effects of a changing climate on animal populations (Gardner et al. 2011). Indeed, declining body size in a number of avian species has been linked to increasing temperatures consistent with climate change (Gardner et al. 2009, Van Buskirk et al. 2010), and it has been suggested that higher temperatures during development may act as an influence on plasticity in growth (Merilä and Hendry 2014). The idea that clines in body size are a result of phenotypic plasticity in morphology that is mediated by the effects of high temperatures on growth is now gaining traction (Teplitsky et al. 2008, Van Buskirk et al. 2010, Yom-Tov and Geffen 2011). In hot climates, nests can potentially act to buffer ambient conditions, but they can still get very hot. Recent work found that Zebra Finch (Taeniopygia guttata) nests in the Australian desert were typically several degrees warmer than ambient conditions and that internal nest temperatures occasionally exceeded 50°C (Griffith et al. 2016). Nest microclimates may therefore be a significant determinant of variation in developmental plasticity and may have the capacity to affect development and growth. Indeed, it has recently been found observationally in a wild population and experimentally in a captive population of the Zebra Finch that higher temperatures during development lead to reduced fledgling and adult body size (Andrew et al. 2017).
If temperature during development is indeed important, then, at the population level, summer maximum temperatures will be a better predictor of mean body size across locations than winter minimum temperatures. As with the House Sparrows studied in North America (Johnston and Selander 1964, 1971, Murphy 1985), the species was deliberately introduced into Australia and New Zealand in the mid-19th century from founders taken from northwestern Europe (Andrew and Griffith 2016). Over the next century, House Sparrows expanded their range to occupy most of the urban areas across both the North and South islands of New Zealand and the eastern half of Australia (Andrew and Griffith 2016), and today are found in a range of climates that are far more variable and extreme than those in the area from which they were sourced. The House Sparrow populations in Australia and New Zealand therefore provide an opportunity to assess the extent to which a species may respond to a changing climate in a relatively short period of time (∼160 yr at most, and <50 yr for the populations at the extreme edge of their range in Australia; Andrew and Griffith 2016). Here, we use these populations of a sedentary avian species (Anderson 2006) to test the extent to which clinal variation in body size is related to both winter minimum and summer maximum temperatures. This will provide new insight into the extent to which body size is a response to the climate experienced during development rather than a response to selection over the winter.
Adult House Sparrows were sampled at 26 locations across Australia (Figure 1, Appendix Table 4), with ∼40 birds measured in each location (n males = 636, n females = 512). Measurements were taken from birds in Australia from April to September, 2014, and in March, 2015. Measurements were taken from birds in New Zealand in 4 locations (Appendix Table 4; n males = 511, n females = 242) between June and August, 2005, as part of earlier work. New Zealand House Sparrow measurements were collected by a separate team from Otago University (Dunedin, New Zealand). Birds were captured using mist nets and placed in bird bags until they were measured. Birds were not held for more than 30 min and were released as soon as possible after they had been measured and banded in accordance with the local bird banding authority.
We determined the age and sex of captured birds by plumage and bill color. All juvenile birds were removed from the morphological analyses. We recorded the tarsus length and body mass of all individuals sampled in Australia and New Zealand. Tarsus length was measured for the right leg, from the bottom of the tarsus with the toes bent forward to the ankle joint. Body mass was measured to the nearest 0.1 g using a Pesola spring scale (Pesola, Schindellegi, Switzerland).
All House Sparrows sampled in Australia were measured by 1 of 2 measurers (S. C. Griffith and M. Awasthy), and some birds were measured by both investigators to test the consistency of tarsus measurements (the regression between tarsus length measurements was significant: R2 = 0.89, t = 20.20, P < 0.001). In New Zealand, all measurements were taken by a single investigator (K. Ludwig, a research associate of S. Nakagawa). We used tarsus length and mass as surrogate measures of body size primarily because tarsus length is the most widely applied measure of skeletal size in passerine birds and body mass relates to overall size as well as being indicative of condition in passerine birds. Therefore, the 2 metrics provided 2 different surrogate measures of body size.
Geographic and Climatic Data
Latitude, longitude, and date of collection were recorded for each sampling site. We used the geographic coordinates to extract the average daily Maximum Temperature of Warmest Month (BIO5), the average daily Minimum Temperature of Coldest Month (BIO6), and average Temperature Seasonality (BIO4 = standard deviation × 100) from WorldClim Global Climate Data (Hijmans et al. 2005), which uses climatic data averaged over 30 yr from 1970 to 2000. The average daily Maximum Temperature of the Warmest Month (hereafter, ‘summer maximum') was our measure of high temperatures during the breeding season to look for a relationship between the average maximum temperature during the breeding season and body size. Average daily Minimum Temperature of Coldest Month (hereafter, ‘winter minimum') was used as a measure of winter extremes to test for a relationship due to the selection pressure of cold conditions.
To test for a relationship between variation in temperature during the breeding season and variation in body size within populations, we looked at the change in mean maximum temperature across the breeding season. In our study locations, the House Sparrow breeding season typically occurs from September to December (Duursma et al. 2017). Using the weather stations closest to our 26 Australian sampling sites, we took the difference in mean maximum temperatures between December and September (Australian Bureau of Meteorology; http://www.bom.gov.au). Data for New Zealand was sourced from the National Institute of Water and Atmospheric Research (NIWA; https://www.niwa.co.nz/education-and-training/schools/resources/climate). The close proximities of the relevant government weather stations to the sampling sites (mean distance = 10.7 km, range = 0.2–34.1 km; see Supplemental Material Data S3) meant that WorldClim and weather station data were highly correlated (e.g., using data for the hottest month from both sources: Pearson's correlation coefficient = 0.98, n = 30, P < 0.001; using data for the coldest month: Pearson's correlation coefficient = 0.96, n = 30, P < 0.001). A second measure of climatic variability in Australia used daily maximum temperatures from the 3 breeding seasons (September to December) prior to sampling (2011–2013). Daily maximum temperatures for these 360 days were downloaded from the Australian Water Availability Project (Jones et al. 2009, http://www.bom.gov.au/jsp/awap/). Temperature variability for this period was calculated, using the same method as for body size, by adding the log of sample standard deviation to sample variance (Nakagawa et al. 2015). Temperature variability was found to be highly correlated with breeding season range in Australia (estimate = 0.11, t24 = 6.17, P < 0.001, R2 = 0.61).
All statistical analyses were conducted using R 3.3.1 (R Core Team 2017). All R code and data used are provided as Supplemental Material ( Supplemental Material Data, Supplemental Material R Code). The individual body size measurements for mass and tarsus length showed a normal distribution. We calculated the mean body size (mass and tarsus length) and variability for each sample population. Variability was calculated by adding the log of the sample standard deviation to the sample variance; this method was chosen because it produces variability that is linearly related to the mean (Nakagawa et al. 2015). If temperature during development affects mass or tarsus length, then we would expect higher variability in body size in locations where the climate shows a higher degree of variation in temperature during the breeding season. For the mean and variability data, linear models were fitted using the standard lm function in R. Males and females were analyzed separately as well as combined. Summer maximum and winter minimum were used as fixed effects in models for mean size. Summer maximum and temperature range across the breeding season were used as fixed effects in size variability models. To account for differences in sample sizes, all linear models included weights (sample weight = 2(n − 1)). All linear models using variability included the log of the mean size as a predictor to account for any relationship between mean body size and variability.
Linear mixed models (LMM) were used for fitting individual measurements from all 1,901 birds. For these models, sample population was used as a random factor. LMMs were fitted using the R package lme4 (Bates et al. 2015). For the LMMs, P-values and degrees of freedom were calculated with the R package lmerTest (Kuznetsova et al. 2016). Summer maximum, winter minimum, and sex were included as fixed effects in the LMMs. Interclass correlation coefficients (ICC) were also calculated for the random effect of sample population to describe how much variation was partitioned between populations (Nakagawa and Schielzeth 2010). The interclass correlation coefficient (ICC), R2, and narrow sense heritability (h2) all estimate the proportion of the variance in the response variable that is explained by factors in the model (Nakagawa and Schielzeth 2013). Because ICC is a proportion, it can be compared between similar models (such as our LMMs) that share the same fixed and random effects. To calculate the proportion of variance explained by random factors, the residual variance and the variance explained by fixed effects (known as marginal R2) need to be included. As a result, the total variance explained by the model, that is the conditional R2 (Nakagawa and Schielzeth 2013), can also be calculated. We report the marginal (fixed effects) and conditional (total model) R2 for both mass and tarsus length models.
To compare the predictive power of the individual fixed effects, we used semipartial correlations for all of our main models (Schielzeth 2010). Semipartial correlations (hereafter, ‘semipartial r') scale the response and predictor variables so the mean is 0 and the standard deviation is 1. This scaling results in estimates that are able to be related to estimates of other response variables within and between models. However, all P-values and t-values remain unchanged due to scaling. Scaling also allows for binary variables, such as sex, to be coded as −1 and 1, which allows these binary factors to be directly compared with continuous variables (Schielzeth 2010). We chose summer maximum and winter minimum as our 2 bioclimatic variables because they were the most relevant to our hypotheses and because summer maximum was highly correlated with other bioclimatic variables such as latitude and seasonality, but not winter minimum (Appendix Table 5).
Using the mean mass and tarsus length measurements of sampled House Sparrow populations, we found that summer maximum was a better predictor of body size than winter minimum (Tables 1 and 2). The relationship between mean body size and summer maximum temperature was strongly negative for both males and females, as was the relationship between female mean tarsus length and summer maximum temperature (Figure 2). All relationships of mean body mass and mean tarsus length with winter minimum temperature were nonsignificant (Appendix Figure 4). Likewise, at the individual level, where both summer maximum and winter minimum were used as fixed effects in linear mixed models (LMMs), summer maximum was a substantially better predictor of body mass than winter minimum temperature (semipartial r = −0.34 vs. semipartial r = −0.01; Table 3). For tarsus length, the fixed effect of summer maximum temperature was a stronger predictor than winter minimum temperature, but the effects were similar (semipartial r = −0.13 vs. semipartial r = −0.10; Table 3). The random factor of location explained a similar amount of variance in both body mass and tarsus length (11% and 9%, respectively; Table 3).
Results from multiple linear regression models using House Sparrow mean body mass in relation to temperature variables. Mean body mass was calculated for 30 sampling locations across Australia and New Zealand (Appendix Table 4) and for males and females separately. Significant effects are in bold font. Summer maximum temperature (Summer max) had a significant negative relationship in all 3 models. Winter minimum temperature (Winter min) was not a significant predictor in any of the 3 models.
Results from multiple linear regression models using House Sparrow mean tarsus length in relation to temperature variables. Mean tarsus length was calculated for 30 sampling locations across Australia and New Zealand (Appendix Table 4) and for males and females separately. Significant effects are in bold font. Summer maximum temperature (Summer max) was significantly negatively related to tarsus length for females and all birds combined; however, the same negative trend was not significant for males. Winter minimum temperature (Winter min) was not a significant predictor in any of the 3 models.
Results from linear mixed models (LMMs) examining individual House Sparrow body mass and tarsus length measurements in relation to temperature. These LMMs used the measurements from 30 House Sparrow populations across Australia and New Zealand (n = 1,901 individuals) and included sampling location as a random factor. Significant effects are in bold font. Body mass had a significant negative relationship with summer maximum temperature (Summer max) but not winter minimum temperature (Winter min). Males were heavier than females, but the difference between the sexes was only marginally significant. Tarsus length decreased with increasing summer maximum temperature. Males had longer tarsi than females, but the semipartial correlation value was low, indicating a small size difference between the sexes.
We would expect populations breeding across a relatively long breeding period (September–December in our study region) to encounter a wide range of ambient temperatures. In locations with a broader range of temperatures, we would expect individuals to experience a wider range of temperatures during development, resulting in greater variation in body size in these populations that experience the greatest temperature ranges across the breeding season. We did not find any significant relationships between temperature range and variability in mass or tarsus length (Appendix Tables 6 and 7). However, the variability of body mass in our sample populations was positively related to summer maximum temperature; this positive relationship was also significant for females but not for males when they were analyzed separately (Figure 3, Appendix Table 6). For tarsus length, there was a nonsignificant positive relationship between summer maximum and tarsus variability for males, females, and both sexes combined (Figure 3, Appendix Table 7). There was no strong linear relationship between the temperature range across the breeding season and summer maximum temperature, indicating that these 2 variables were not conflated in this case (estimate = 0.18, t28 = 2.63, P = 0.01, R2 = 0.20).
Our observational work on the House Sparrow populations introduced into Australia and New Zealand essentially replicates the earlier work done in North America (Johnston and Selander 1964, 1973) and Europe (Murphy 1985) that revealed latitudinal clines in body size in this species. As with most other similar studies across animal taxa, in these earlier studies the clinal variation in body size was attributed to the selective effects of cold weather during the winter (Johnston and Fleischer 1981, Fleischer and Johnston 1984). However, there have been suggestions that a similar pattern may also be driven by constraints affected by the climate experienced during development (Van Buskirk et al. 2010, Gardner et al. 2011, Cunningham et al. 2013). We found support for this idea through our observation that summer maximum temperatures better predicted body size variation than winter minimum temperatures. As all variables were scaled, we were able to use the semipartial r values from our models to identify summer maximum as a stronger predictor than winter minimum temperature. This observational finding, from House Sparrow populations introduced into the range of climates found in Australia and New Zealand ∼150 yr ago (Andrew and Griffith 2016), supports the hypothesis that excessive environmental heat during development may affect growth (Van Buskirk et al. 2010, Gardner et al. 2011, Burness et al. 2013, Andrew et al. 2017). We also explored this hypothesis by looking at the relationship between variability in body size within populations and climatic variability. We did not find any significant relationships between temperature range across the breeding season and variation in mass or tarsus length (Appendix Tables 6 and 7). However, we found the expected positive relationship between summer maximum temperature and body size traits, although it was not always significant (Appendix Tables 6 and 7). The nonsignificant results could have been due to low statistical power (only 30 populations) or a weaker effect on skeletal measurements (tarsus length) than mass. The relationship between summer maximum temperature and variability in body size could have been due to warmer climates being more likely to exceed possible ‘threshold temperatures' that significantly affect development. The analysis of body size variability promises to be a useful avenue for future studies to explore, especially those with large numbers of sample populations, to test whether this result can be replicated. Only a small portion of the variation in body size observed in this study was explained by temperature; in addition to this, a meaningful portion (∼10%) of the variation in tarsus length and mass was partitioned between locations by the random factor of location. These differences between populations could have been linked to genetic differentiation (due to selection or drift) or to other environmental factors not included in the model, such as the time of year that birds were measured. It is also important to note that, in many contexts, temperature may not directly affect body size because of adaptations for mitigating the effects of temperature, such as behavioral adaptations that reduce the exposure of developing offspring to high temperatures.
Summer temperatures explained variation in both tarsus length and mass, having a larger effect on mass. A comparison of the 3 locations that had the hottest summer maximum temperatures with the 3 locations that had the coolest summer maximums revealed that sparrows in the hottest locations were ∼6% lighter and had tarsi that were ∼2% smaller than sparrows in the coolest locations. The greater magnitude of the effect of summer temperatures on mass than the skeletal measure of tarsus length is consistent with earlier studies that showed greater plasticity and lower heritability of body mass (Alatalo et al. 1990, Jensen et al. 2003). Similarly, a study on North American migratory birds (249 species migrating during all 4 seasons) found that increases in summer temperature caused a larger percentage decline in mass (0.34% per degree Celsius) than wing chord length (0.09% per degree Celsius; Van Buskirk et al. 2010). By comparison, in the House Sparrow populations studied here, mass declined by 0.33% and tarsus length declined by 0.11% per degree Celsius. Although observational, the findings that we report here are similar to those from a recent study of the Zebra Finch, in which the temperature during development in an observational study in the field (fledglings were ∼8% lighter in hot vs. cold breeding attempts by the same pair), and in an experimental study in the laboratory (mass was 5% lower in the high temperature treatment), caused similar decreases in body size (Andrew et al. 2017).
Our findings regarding the House Sparrow populations in Australia and New Zealand are consistent with Bergmann's rule, but not the widely cited mechanism that cold temperatures select for large adults. Future studies should explore the mechanistic link between the climate experienced during development and the body size attained. Temperatures experienced during development could also be relevant to another ecogeographical rule, that of larger extremities relative to core size in warmer climates, as predicted by Allen's rule (Allen 1877, Symonds and Tattersall 2010). For example, a large proportion of the variation in the bill surface area (82–89%) among species of North American tidal salt marsh sparrows is explained by summer temperature (Greenberg et al. 2012). Possible mechanisms for determining plasticity in morphological development include physiological constraints (Gardner et al. 2009), constraints on parental provisioning (Cunningham et al. 2013), and parental effects (Mariette and Buchanan 2016). We believe that our study of the House Sparrow and recent experimental work on the Zebra Finch (Andrew et al. 2017) lead to the prediction that increasing summer temperatures at a given site will drive down the average body size of that population, consistent with the effect reported by Gardner et al. (2009). However, while our data could suggest that this decline in body size might simply be the result of developmental plasticity, our data also do not exclude the possibility that selection may contribute to a change in body size over time. For example, the effect that we describe here may be accounted for, wholly or partly, by higher reproductive success or differential survival of smaller adults in hotter locations. Both of these possible scenarios remain to be tested in the House Sparrow. However, in a recent study of the Zebra Finch, the same pairs produced offspring of different sizes in the laboratory when breeding in cool vs. hot experimental temperatures (Andrew et al. 2017), which certainly suggests that developmental plasticity can contribute to the size differences observed across climates. Yet, even in Zebra Finches, which are highly adapted to breeding in a very hot climate (Griffith et al. 2016), the question remains regarding the extent to which size differences can be attributed to selection on the genes underlying body size and/or developmental plasticity. Our study highlights the fact that, when addressing the possible selective response of body size to a changing climate (Gardner et al. 2011), we should be considering not only selection on adults during the cold of winter, but also selection on both adults and offspring during the breeding season and the hot extremes of the summer climate.
We thank the large number of people who helped us with access to urban study sites for sampling the human-commensal House Sparrow. We are grateful to Elizabeth L. Sheldon, Anna Feit, and Peter Bird for their participation in fieldwork in Australia and Karin Ludwig for fieldwork in New Zealand. Daisy E. Duursma assisted us with climatic data. We extend our gratitude to Laura L. Hurley, Mylene M. Mariette, Kate Buchanan, Sasha Dall, Peter Dunn, Janet Gardner, Tim Parker, and Andy Russell for their comments on earlier drafts of the manuscript.
Funding statement: S.C.A. was supported by Macquarie University Research Excellence Scholarship (no. 2013077). S.C.G. was supported by an Australian Research Council Future Fellowship (FT130101253). S.N. was supported by a Rutherford Discovery Fellowship (New Zealand) and an Australian Research Council Future Fellowship (FT130100268). None of our funders had any input into the content of the manuscript, nor did they require the approval of the manuscript before submission or publication.
Ethics statement: Australian birds were sampled under an Animal Research Authority of the Animal Ethics Committee at Macquarie University (ARA 2014/248). New Zealand samples were collected through Otago University (Animal Ethics reference number 87/08).
Author contributions: S.C.A. and S.C.G. conceived the idea for the paper. S.C.A. and S.N. analyzed the data. All authors contributed to data collection, and wrote and/or substantially edited the paper. S.C.G. and S.N. secured funding for this work.
- Alatalo, R. V., L. Gustafsson, and A. Lundberg (1990). Phenotypic selection on heritable size traits: Environmental variance and genetic response. The American Naturalist 135:464–471. Google Scholar
- Allen, J. A. (1877). The influence of physical conditions in the genesis of species. Radical Review 1:108–140. Google Scholar
- Anderson, T. R. (2006). Biology of the Ubiquitous House Sparrow: From Genes to Populations. Oxford University Press, New York, NY, USA. Google Scholar
- Andrew, S. C., and S. C. Griffith (2016). Inaccuracies in the history of a well-known introduction: A case study of the Australian House Sparrow (Passer domesticus). Avian Research 7:art.9. Google Scholar
- Ashton, K. G. (2002). Patterns of within-species body size variation of birds: Strong evidence for Bergmann's rule. Global Ecology and Biogeography 11:505–523. Google Scholar
- Ashton, K. G., M. C. Tracy, and A. de Queiroz (2000). Is Bergmann's rule valid for mammals?The American Naturalist 156:390–415. Google Scholar
- Bates, D., M. Mächler, B. M. Bolker, and S. C. Walker (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67:1–48. Google Scholar
- Bergmann, C. (1847). Ueber die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Grösse. Göttinger Studien 3:595–708. Google Scholar
- Briscoe, N. J., A. Krockenberger, K. A. Handasyde, and M. R. Kearney (2015). Bergmann meets Scholander: Geographical variation in body size and insulation in the koala is related to climate. Journal of Biogeography 42:791–802. Google Scholar
- Burness, G., J. R. Huard, E. Malcolm, and G. J. Tattersall (2013). Post-hatch heat warms adult beaks: Irreversible physiological plasticity in Japanese Quail. Proceedings of the Royal Society B 280:20131436. Google Scholar
- Cardilini, A. P. A., K. L. Buchanan, C. D. H. Sherman, P. Cassey, and M. R. E. Symonds (2016). Tests of ecogeographical relationships in a non-native species: What rules avian morphology?Oecologia 181:783–793. Google Scholar
- Clauss, M., M. T. Dittmann, D. W. H. Müller, C. Meloro, and D. Codron (2013). Bergmann's rule in mammals: A cross-species interspecific pattern. Oikos 122:1465–1472. Google Scholar
- Cunningham, S. J., R. O. Martin, C. L. Hojem, and P. A. R. Hockey (2013). Temperatures in excess of critical thresholds threaten nestling growth and survival in a rapidly-warming arid savanna: A study of Common Fiscals. PLOS One 8:e74613. Google Scholar
- Duursma, D. E., R. V. Gallagher, and S. C. Griffith (2017). Characterizing opportunistic breeding at a continental scale using all available sources of phenological data: An assessment of 337 species across the Australian continent. The Auk: Ornithological Advances 134:509–519. Google Scholar
- Fleischer, R. C., and R. F. Johnston (1984). The relationships between winter climate and selection on body size of House Sparrows. Canadian Journal of Zoology 62:405–410. Google Scholar
- Gardner, J. L., R. Heinsohn, and L. Joseph (2009). Shifting latitudinal clines in avian body size correlate with global warming in Australian passerines. Proceedings of the Royal Society B 276:3845–3852. Google Scholar
- Gardner, J. L., A. Peters, M. R. Kearney, L. Joseph, and R. Heinsohn (2011). Declining body size: A third universal response to warming?Trends in Ecology & Evolution 26:285–291. Google Scholar
- Greenberg, R., R. Danner, B. Olsen, and D. Luther (2012). High summer temperature explains bill size variation in salt marsh sparrows. Ecography 35:146–152. Google Scholar
- Griffith, S. C., M. C. Mainwaring, E. Sorato, and C. Beckmann (2016). High atmospheric temperatures and ‘ambient incubation’ drive embryonic development and lead to earlier hatching in a passerine bird. Royal Society Open Science 3:150371. Google Scholar
- Hijmans, R. J., S. E. Cameron, J. L. Parra, P. G. Jones, and A. Jarvis (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965–1978. Google Scholar
- James, F. C. (1970). Geographic size variation in birds and its relationship to climate. Ecology 51:365–390. Google Scholar
- Jensen, H., B.-E. Sæther, T. H. Ringsby, J. Tufto, S. C. Griffith, and H. Ellegren (2003). Sexual variation in heritability and genetic correlations of morphological traits in House Sparrow (Passer domesticus). Journal of Evolutionary Biology 16:1296–1307. Google Scholar
- Johnston, R. F., and R. C. Fleischer (1981). Overwinter mortality and sexual size dimorphism in the House Sparrow. The Auk 98:503–511. Google Scholar
- Johnston, R. F., and R. K. Selander (1964). House Sparrows: Rapid evolution of races in North America. Science 144:548–550. Google Scholar
- Johnston, R. F., and R. K. Selander (1971). Evolution in the House Sparrow. II. Adaptive differentiation in North American populations. Evolution 25:1–28. Google Scholar
- Johnston, R. F., and R. K. Selander (1973). Evolution in the House Sparrow. III. Variation in size and sexual dimorphism in Europe and North and South America. The American Naturalist 107:373–390. Google Scholar
- Jones, D. A., W. Wang, and R. Fawcett (2009). High-quality spatial climate data-sets for Australia. Australian Meteorological and Oceanographic Journal 58:233–248. Google Scholar
- Mariette, M. M., and K. L. Buchanan (2016). Prenatal acoustic communication programs offspring for high posthatching temperature in a songbird. Science 353:812–814. Google Scholar
- Mayr, E. (1956). Geographical character gradients and climatic adaptation. Evolution 10:105–108. Google Scholar
- McNab, B. (1971). On the ecological significance of Bergmann's rule. Ecology 52:845–854. Google Scholar
- Meiri, S., and T. Dayan (2003). On the validity of Bergmann's rule. Journal of Biogeography 30:331–351. Google Scholar
- Merilä, J., and A. P. Hendry (2014). Climate change, adaptation, and phenotypic plasticity: The problem and the evidence. Evolutionary Applications 7:1–14. Google Scholar
- Millien, V., S. K. Lyons, L. Olson, F. A. Smith, A. B. Wilson, and Y. Yom-Tov (2006). Ecotypic variation in the context of global climate change: Revisiting the rules. Ecology Letters 9:853–869. Google Scholar
- Murphy, E. C. (1985). Bergmann's rule, seasonality, and geographic variation in body size of House Sparrows. Evolution 39:1327–1334. Google Scholar
- Nakagawa, S., and H. Schielzeth (2010). Repeatability for Gaussian and non-Gaussian data: A practical guide for biologists. Biological Reviews 85:935–956. Google Scholar
- Nakagawa, S., and H. Schielzeth (2013). A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133–142. Google Scholar
- Nakagawa, S., R. Poulin, K. Mengersen, K. Reinhold, L. Engqvist, M. Lagisz, and A. M. Senior (2015). Meta-analysis of variation: Ecological and evolutionary applications and beyond. Methods in Ecology and Evolution 6:143–152. Google Scholar
- Partridge, L., and J. A. Coyne (1997). Bergmann's rule in ectotherms: Is it adaptative?Evolution 51:632–635. Google Scholar
- Salewski, V., and C. Watt (2017). Bergmann's rule: A biophysiological rule examined in birds. Oikos 126:161–172. Google Scholar
- Schielzeth, H. (2010). Simple means to improve the interpretability of regression coefficients. Methods in Ecology and Evolution 1:103–113. Google Scholar
- Scholander, P. F. (1955). Evolution of climatic adaptation in homeotherms. Evolution 9:15–26. Google Scholar
- Symonds, M. R. E., and G. J. Tattersall (2010). Geographical variation in bill size across bird species provides evidence for Allen's rule. The American Naturalist 176:188–197. Google Scholar
- Teplitsky, C., and V. Millien (2014). Climate warming and Bergmann's rule through time: Is there any evidence?Evolutionary Applications 7:156–168. Google Scholar
- Teplitsky, C., J. A. Mills, J. S. Alho, J. W. Yarrall, and J. Merilä (2008). Bergmann's rule and climate change revisited: Disentangling environmental and genetic responses in a wild bird population. Proceedings of the National Academy of Sciences USA 105:13492–13496. Google Scholar
- Van Buskirk, J., R. S. Mulvihill, and R. C. Leberman (2010). Declining body sizes in North American birds associated with climate change. Oikos 119:1047–1055. Google Scholar
- Yom-Tov, Y., and E. Geffen (2011). Recent spatial and temporal changes in body size of terrestrial vertebrates: Probable causes and pitfalls. Biological Reviews 86:531–541. Google Scholar
APPENDIX TABLE 4.
Summary of sampled House Sparrow populations in Australia (Aus) and New Zealand (NZ) used to test the extent to which clinal variation in body size was related to winter minimum and summer maximum temperatures. Shown are sampling locations, sample sizes, and climatic variables (Summer max = mean maximum temperature of the hottest month; Winter min = mean minimum temperature of the coldest month). Sample populations are ordered from north to south within country.
APPENDIX TABLE 5.
Relationship between latitude and summer maximum and winter minimum temperature at our sampling sites in Australia and New Zealand. Latitude had a significant negative relationship with both summer and winter temperature. The highest t-value and most significant relationship was for summer maximum temperature. However, there was no significant relationship between summer and winter temperature for our 30 sampling locations (estimate = 0.634, t28 = 1.951, P = 0.06, R2 = 0.120). The summer maximum temperature (of the hottest month) had a positive relationship with seasonality, which is a metric of climatic variability (estimate = 0.038, t28 = 6.505, P < 0.001, R2 = 0.602). There was a weaker relationship between breeding season temperature range and summer maximum temperature (estimate = 0.177, t28 = 2.631, P = 0.01, R2 = 0.198).
APPENDIX TABLE 6.
Results from multiple linear regression models using variability in House Sparrow body mass per population (see Appendix Table 4) in relation to temperature in Australia and New Zealand. We found a significant positive relationship between body mass variability and summer maximum temperature (Summer max) for female birds and all birds combined. The positive slope for summer maximum temperature for male birds was not significant. A positive slope shows that in warmer climates there is more variability in body mass at the population level. There was no significant effect of temperature range (Temp range) across the breeding season.
APPENDIX TABLE 7.
Results from multiple linear regression models using variability in House Sparrow tarsus length per population (see Appendix Table 4) in relation to temperature in Australia and New Zealand. There was no significant relationship in any of the 3 models between temperature and variability in tarsus length. There was a consistent, nonsignificant, positive slope for summer maximum temperature (a significant positive slope would have meant that in warmer climates there was more variability in tarsus length at a population level).