- Top of page
- Materials and methods
Variation in demographic and physiological attributes of life history is thought to fall on one single axis, a phenomenon termed the Pace-of-Life. A slow Pace-of-Life is characterized by low annual reproduction, long life span and low metabolic rate, a fast Pace-of-Life by the opposite characteristics. The existence of a single axis has been attributed to constraints among physiological mechanisms that are thought to restrict evolutionary potential. In that case, physiological traits should covary in the same fashion at the levels of individual organisms and species. We examined covariation at the levels of individual and subspecies in three physiological systems (metabolic, endocrine and immune) using four stonechat subspecies with distinct life-history strategies in a common-garden set-up. We measured basal metabolic rate, corticosterone as endocrine measure and six measures of constitutive immunity. Metabolic rate covaried with two indices of immunity at the individual level, and with corticosterone concentrations and one index of immunity at the subspecies level, but not with other measures. The different patterns of covariation among individuals and among subspecies demonstrate that links among physiological traits are loose and suggest that these traits can evolve independent of each other.
- Top of page
- Materials and methods
Variation in demographic and physiological attributes of life-history strategies is thought to fall on one dominant slow–fast Pace-of-Life axis, with low reproductive rates, long life spans and low metabolic rates at one end, and the opposite traits at the other end (Promislow & Harvey, 1990; Ricklefs & Wikelski, 2002). This single dominant axis has been reported at different levels of organization, ranging from studies that include a wide array of different species (Promislow & Harvey, 1990; Ricklefs, 2000) to those that investigate among-individual variation within a single species (Ricklefs, 2000; Johnson, 2001; Tieleman et al., 2005). Traditionally, most studies connecting life-history and physiology measure a standardized level of metabolic rate (e.g. Pearl, 1928; Drent & Daan, 1980; Ricklefs & Wikelski, 2002; Wikelski et al., 2003; Speakman, 2005; Tieleman et al., 2006), but more recently, the endocrine system and the immune system have been included as well (Sheldon & Verhulst, 1996; Ricklefs & Wikelski, 2002; Tieleman et al., 2005; Lee, 2006). The idea behind this view is that the different physiological systems (e.g. immune, endocrine and metabolic) are intrinsically linked (e.g. Dhabhar & McEwen, 1997; Lochmiller & Deerenberg, 2000; Speakman, 2005; Landys et al., 2006), in addition to each playing its own role in the balance between reproductive effort and adult self-maintenance. Because physiology mediates the relationship between organism and environment, Ricklefs & Wikelski (2002) suggest that the organization of physiological mechanisms may constrain individual responses to the environment, thereby limiting life-history variation. Constraints at the individual level would result in limited evolutionary potential (Lande, 1979; Schluter, 1996; Duckworth, 2010; Ketterson & Nolan, 1999). Yet, the existence of these constraints remains to be revealed.
When evaluating whether and how physiological constraints play a role in shaping life-history variation, patterns of covariation of traits at different levels of organization can provide complementary insights: (i) if the connections among physiological mechanisms have restricted the independent evolutionary potential of physiological traits, one might expect that these traits covary in the same way within and among species. Moreover, if physiological mechanisms restrict the evolutionary potential of life-history traits, one would expect that these physiological traits, including immune and endocrine variables, covary with the single dominant axis described for demographic and metabolic traits. (ii) If patterns of covariation of physiological traits within species differ from those among species, links between physiological traits are unlikely to constrain the evolutionary potential of these traits, or life-history traits. In fact, the absence of correlations among two or more traits at the individual level would strongly indicate that these traits are not forced to covary through mechanistic connections. The presence, in contrast, of a strong correlation among traits at the individual level does not necessarily confirm the existence of physiological constraints, which could lead to forced coevolution, but could also reflect individual condition or quality. (iii) Finally, patterns of covariation among individuals can also stem from environmental conditions selectively favouring certain combinations of traits, potentially resulting in different patterns of covariation within and among species (Lande, 1986; Ketterson & Nolan, 1999; Duckworth, 2010). Therefore, in studies of the physiology/life-history nexus, the role of the environment must be explicitly considered as well.
Environmental conditions affect life history and physiology at different evolutionary timescales, and as a result, phenotypic variation in life history and physiology reflects a mixture of different evolutionary phenomena. These phenomena include phenotypic plasticity, genetic differences based on adaptation and genetic differences resulting from historical evolutionary pathways. Their contributions can be largely disentangled when raising organisms from different environments in a common environment (Falconer & Mackay, 1996). Such a common-garden experiment minimizes environmental components of variation and highlights the genetic component of the adaptive response to different environments. When the experiment is restricted to closely related taxa, the effects of historical pathways are minimized.
Stonechats (Saxicola torquata) are an ideal study system to address evolutionary questions related to variation in life history, physiology and environment, because of their widespread geographic range covering a variety of habitats, their well-documented variation in life-history and physiological traits (Table 1) and their ease of handling and acclimation to captivity (Gwinner et al., 1995; Raess, 2005). When raised and kept in a common garden, different subspecies of stonechats maintain differences in clutch size, metabolic rate, moult, development of the young and behaviour (Gwinner et al., 1995; Klaassen, 1995; König & Gwinner, 1995; Starck et al., 1995; Helm & Gwinner, 1999, 2001; Wikelski et al., 2003; Tieleman, 2007; Helm, 2009; Tieleman et al., 2009; Helm et al., 2009). Many of these traits are correlated with factors characterizing a subspecies' original environment and migratory strategy. The existing knowledge of stonechats combined with the common-garden set-up allows a powerful integrative and broad study of physiology and life history.
Table 1. Life-history traits of the studied stonechat subspecies (after Helm, 2003; Baldwin et al., 2010)
|Migratory status||Sedentary||Partially short distance||Short distance||Long distance|
|Migration distance (km)||0||0; 1200||1700||2600|
|Present on the breeding ground||Year-round||Year-round; migrants: February–September||March to mid-October||Early May to early September|
|Number of clutches||1–2||3–4||2–3||1|
|Number of eggs per clutch||3||5||5||6|
|Time of hatching||Related to rainy season||Early April–August||Mid-April–August||Late May/June|
In this study, we evaluate the hypothesis that variation in innate immunity, baseline and stress-induced corticosterone concentrations and metabolic rate falls along a single axis within and among four subspecies of stonechats kept in a common garden. We chose to measure six indices of innate and acquired immunity to capture the complexity of the immune system (Lee, 2006; Buehler et al., 2011). We measured corticosterone concentration because of its functional relationships with immunity and energy mobilization (McEwen et al., 1997; Sapolsky et al., 2000; Landys et al., 2004). We explore whether physiological constraints among the traits at the level of individual birds (the evolutionary potential) can explain variation in physiology and life histories at the level of subspecies (the evolutionary product). Specifically, we test how immune function, corticosterone response and metabolic rate are associated within and among four subspecies of stonechats. Metabolic rate correlates well with some life-history traits, such as migratory disposition and clutch size (Table 1; Wikelski et al., 2003; Tieleman et al., 2009). Therefore, traits that correlate with metabolic rate are likely to correlate with these life-history traits. In this study, we investigate the following scenarios: (i) if we find that physiological traits show the same patterns of covariation within and among subspecies, constraints among these traits may underlie limited evolution. (ii) If, however, the patterns of covariation of physiological traits differ within and among subspecies, it is unlikely that physiological constraints at the individual level have limited the evolutionary potential and have led to the variation in physiological and life-history traits we find at the subspecies level.
- Top of page
- Materials and methods
When we compared each physiological trait separately among the four stonechat subspecies, two general patterns of variation emerged: one that did not correlate with variation in mass-specific BMR and another one that did. The first general pattern included all immune indices except haptoglobin. Kenyan or Kazakh stonechats, the subspecies that differed most in mass-specific BMR and life-history traits, showed intermediate, and Irish or European showed the highest or lowest values compared with the other subspecies (Fig. 1a–e). Although the pattern was qualitatively present in all immune indices except haptoglobin, differences between subspecies were statistically significant only for haemolysis and microbicidal capacity against E. coli (Table 2; Fig. 1a,e). The second pattern, which did correspond to mass-specific BMR (Fig. 2e), consisted of a gradual change from Kenyan to Irish and European to Kazakh stonechats in haptoglobin (decrease, Fig. 1f), stress-induced corticosterone levels (increase, Fig. 2b), and body mass (decrease, Fig. 2c). In other words, for these indices, the Kenyan and Kazakh subspecies differed most from each other, and Irish and European subspecies were intermediate. Again, the pattern was qualitatively present in all indices, but differences between subspecies were statistically significant only for body mass and mass-specific BMR (Table 2).
Figure 1. Residuals of immune indices of four subspecies of stonechats. Microbicidal capacity of blood against (a) Escherichia coli, (b) Candida albicans and (c) Staphylococcus aureus. (d) haemagglutination, (e) haemolysis and (f) haptoglobin concentration in plasma. The residuals are calculated from linear models with covariates and all significant fixed effects except subspecies. Bars and whiskers refer to mean values ± standard error. Letters refer to significant differences among subspecies. Numbers refer to sample sizes (number of individuals).
Download figure to PowerPoint
Figure 2. Residuals of (a) baseline and (b) stress-induced corticosterone, (c) body mass, (d) basal metabolic rate (BMR) and (e) mass-specific BMR of four subspecies of stonechats. Residuals are calculated from linear models with covariates and all significant fixed effects except subspecies. Bars and whiskers refer to mean values ± standard error. Letters refer to significant differences among subspecies. Numbers refer to sample sizes (number of individuals).
Download figure to PowerPoint
Table 2. Results of the linear models for the fixed effects subspecies, sex and age, and the interaction subspecies × sex on measures of body mass, metabolic rate, immune function and corticosterone concentration of four subspecies of stonechats from different environments. Results are from mixed effects or linear regression models after backward elimination of nonsignificant terms (P > 0.05)
| ||Subspecies||Sex||Age||Subspecies × sex|
| P || F ||d.f.|| P || F ||d.f.|| P || F ||d.f.|| P || F ||d.f.|
|Mass||< 0.001||27.13||3,84||< 0.001||14.09||1,84||0.30||1.09||1,34||0.06||2.50||3,81|
|Basal metabolic rate (BMR)||0.32||1.18||3,84||0.02||5.50||1,84||0.58||0.31||1,34||0.37||3.16||3,81|
|BMR with mass as covariate||0.049||2.73||3,85||0.31||1.04||1,84||0.71||0.14||1,33||0.13||1.92||3,81|
|Mass-specific BMR||< 0.001||7.08||3,84||0.17||1.92||1,84||0.77||0.09||1,34||0.01||3.87||3,81|
|Baseline corticosteronea ||0.58||0.56||2,22||0.08||3.43||1,20||0.79||0.07||1,19||0.40||0.96||2,17|
|Stress-induced corticosteronea ||0.13||2.21||2,26||0.60||0.29||1,25||0.03||5.42||1,26||0.63||0.48||2,23|
|Haemolysisa ||0.03||3.21||3,42||0.13||2.31||1,56||< 0.001||14.61||1,57||0.47||0.84||3,39|
|Microbicidal ability against|
| Escherichia coli b ||< 0.001||6.48||3,50||0.61||0.26||1,50||0.18||1.85||1,49||0.008||4.35||3,50|
| Candida albicans b ||0.68||0.51||3,50||0.03||5.16||1,50||0.87||0.03||1,49||0.99||0.04||3,46|
| Staphylococcus aureus b ||0.51||0.78||3,51||0.03||4.73||1,51||0.72||0.13||1,50||0.71||0.47||3,47|
To explore the patterns of covariation of physiological traits at the level of subspecies, we correlated corticosterone and immune measures with mass-specific BMR (Fig. 3). We correlated subspecies' means of mass-specific BMR with residual values of baseline and stress-induced corticosterone concentrations. Mass-specific BMR was not significantly correlated with residual baseline levels of corticosterone (t 1 = 1.12, P = 0.47), nor, despite the positive trend, with residual stress-induced levels of corticosterone (t 1 = 5.50, P = 0.11) (Fig. 3a). Discriminant analysis of the immune measures resulted in one significant discriminant function and two nonsignificant discriminant functions (Table 3). The significant first discriminant function showed that Kenyan and Kazakh subspecies were most different and that Irish and European stonechats had intermediate values, a result consistent with the Pace-of-Life hypothesis (Fig. 3b). The correlation coefficient between this first function and mass-specific BMR was marginally nonsignificant (F 1,2 = 3.50, P = 0.07). Because haemagglutination was the immune variable that contributed mostly to the first function, and because this function was significant, we interpret this result as a significant difference in haemagglutination among subspecies. The nonsignificant second and third discriminant functions describe the variation in the other immune indices, including haemolysis, haptoglobin and microbicidal capacity against E. coli, C. albicans and S. aureus (Table 3; Fig. 3c,d). The second and third functions did not significantly correlate with mass-specific BMR, and in contrast to the predictions regarding the Pace-of-Life, scores for the Kenyan and Kazakh subspecies were similar for the second and third functions.
Figure 3. Subspecies level correlations between (a) residual baseline (black circles) and stress-induced (grey circles) corticosterone and (b–d) the three discriminant functions of constitutive immunity and the indicator for Pace-of-Life, mass-specific basal metabolic rate. Residual corticosterone is calculated from linear models with covariates and all significant fixed effects except subspecies. Dots and whiskers refer to mean values ± standard error. P-values of the correlation are shown in the graphs. If the correlation showed a trend (P < 0.1) a regression line was drawn.
Download figure to PowerPoint
Table 3. Results of the discriminant analysis of six measures of constitutive immune function among four subspecies of stonechats. The highest loading for each physiological trait shows to which function that trait contributes the largest explanatory value (compare values within rows, highest value is bolded). Significance of each discriminant function indicates whether that function can be significantly discriminated among subspecies. High loadings on the same discriminant function indicate correlation between physiological traits
| ||Function 1||Function 2||Function 3|
|Haemagglutination|| 0.757 ||−0.067||0.223|
|Haemolysis||−0.097||−0.374|| 0.388 |
|Haptoglobin||0.101|| 0.290 ||0.131|
|Microbial ability against|
| Escherichia coli ||0.125||−0.714 ||0.474|
| Candida albicans ||−0.071||0.249|| 0.905 |
| Staphylococcus aureus ||−0.180||0.106|| 0.188 |
|Prop. variance explained (%)||56.94||35.93||7.14|
|Cumulative prop. variance explained (%)||56.94||92.86||100.00|
The principal component analysis at the level of individual revealed grouping of physiological traits that was dissimilar from the result of the discriminant analysis at the level of subspecies (Tables 4 vs. 3). The first principle component was highly correlated with haemagglutination and haemolysis. This principal component showed a significant negative correlation with mass-specific BMR (F 1,44 = −2.45, P = 0.02; Fig. 4a). The second principle component explained most of the variation in microbicidal capacity against E. coli and C. albicans. The third principal component explained most of the variation in microbicidal capacity against S. aureus and haptoglobin (Table 4). These two principal components did not significantly correlate with mass-specific BMR (F 1,44 > −1.50, P > 0.14; Fig. 4b–c).
Figure 4. Individual-level correlations between the indicator for Pace-of-Life, mass-specific basal metabolic rate (BMR) and (a) PC1, (b) PC2 and (c) PC3 resulting from the principal component analysis on indices of immunity. The y-axes are labelled with the indices that had the highest loadings on the PCs. Mass-specific BMR was statistically corrected for subspecies. Correlation coefficients are calculated with pooled subspecies, but subspecies are represented by different symbols. Significance of the correlation is shown in the graphs, and if the correlation was significant (P < 0.05) a regression line was drawn.
Download figure to PowerPoint
Table 4. Results of principal component analysis after varimax rotation for constitutive immune indices of individual stonechats pooled across the four subspecies. Loadings on the first three principal components (PC's) had an eigenvalue larger than 1. Loadings larger than 0.50 are bolded. Variables are correlated if they have a high loading on the same PC
|Haemagglutination|| 0.836 ||−0.092||0.042|
|Haemolysis|| 0.849 ||0.168||0.093|
|Microbicidal ability against|
| Escherichia coli ||0.268|| 0.720 ||−0.135|
| Candida albicans ||−0.213|| 0.822 ||0.130|
| Staphylococcus aureus ||0.297||0.200|| 0.789 |
|Prop. variance explained (%)||28||24||18|
|Cumulative prop. variance explained (%)||28||52||70|
- Top of page
- Materials and methods
Immune function, corticosterone response and BMR did not fall on a single dominant axis in stonechats, neither among subspecies nor among individuals (Figs 3 and 4). Our finding of dissimilar patterns of covariation of immune, endocrine and metabolic traits on the levels of subspecies and individual birds does not support the hypothesis that intrinsic connections between these physiological systems constrain the evolution of physiological attributes. Because we used different methods to describe the patterns of covariation among and within subspecies, a formal statistical comparison was not possible. Yet, the striking difference in patterns within and among subspecies make it unlikely that constraints among immune, endocrine and metabolic traits have limited the evolution of life-history variation (Ricklefs & Wikelski, 2002). If the strength of environmental selection factors is strong enough, it can uncouple originally constrained traits over evolutionary time (Ketterson & Nolan, 1999; Duckworth, 2010). This can lead to a lack of correlations at the level of subspecies between traits that are constrained within individuals. In our study, covariation among some physiological traits at the level of individuals is present and potentially points to such constraints. Recent related studies (discussed below), however, find that these same traits can also vary independently of each other. Synthesizing our results we conclude that the phenomenon of a single Pace-of-Life axis does not result from constraints among the physiological traits we measured. Combining our results with results of other studies shows that universal constraints are unlikely to exist among these traits. A possible alternative hypothesis is that environmental factors, such as food availability, select sets of covarying physiological traits and life-history characteristics, potentially leading to constrained evolution. Other environmental factors, like pathogen pressure, might select traits (e.g. immune and endocrine) which do not covary with life-history characteristics.
Physiological variation among subspecies
We found that only some physiological traits covaried with mass-specific BMR at the level of stonechat subspecies. Because previous studies demonstrate a link between life-history and metabolic traits in stonechats, we can use mass-specific BMR as an index for Pace-of-Life (Table 1; Klaassen, 1995; Wikelski et al., 2003; Tieleman et al., 2009). Concentrations of corticosterone after an acute stressor have been hypothesized to be high in birds with a slow Pace-of-Life, because of the hormone's role in favouring self-maintenance over current reproduction (Wingfield et al., 1995; Wikelski & Ricklefs, 2001). Therefore, we expected stress-induced corticosterone to be highest in Kenyan and lowest in Kazakh stonechats. Our results do not support this prediction: concentrations of stress-induced corticosterone were higher in fast-living Kazakh stonechats than in relatively slow-living Kenyan stonechats, and intermediate in European stonechats. Levels of baseline corticosterone concentration were also higher in Kazakh than in Kenyan stonechats. These results corroborate several other studies that find high levels of baseline (Martin et al., 2005; Bókony et al., 2009; Hau et al., 2010) or stress-induced corticosterone concentrations (Martin et al., 2005) in fast-living species, subspecies or populations. However, other studies, including one on free-living stonechats during the breeding season (Wingfield et al., 1995; Silverin et al., 1997; Breuner et al., 2003; Goymann et al., 2006; Bókony et al., 2009), offer support for the hypothesized link between higher corticosterone concentrations with a slower Pace-of-Life. A third set of studies finds no relationship between corticosterone concentrations and Pace-of-Life characteristics (Breuner et al., 2003; Lindström et al., 2005; Hau et al., 2010). Therefore, we conclude that the corticosterone response does not overlay the single axis of variation fitting with the Pace-of-Life syndrome.
Birds with longer life spans – often indicated by lower metabolic rates – might benefit from more robust immune defences (Boots & Bowers, 2004; Lee et al., 2008; Horrocks et al., 2011). Stonechat subspecies, however, show no clear covariation among constitutive immunity and metabolic rate. When each index was analysed separately, haptoglobin was the only index that covaried with metabolic rate. When immune indices were collapsed in a discriminant analysis, only haemagglutination covaried with metabolic rate. Earlier studies that measured constitutive immunity in relation to metabolic rate or other Pace-of-Life indicators show ambiguous patterns. Among Neotropical birds, species with a low metabolic rate had higher microbicidal ability against E. coli than species with a high metabolic rate (Tieleman et al., 2005). In contrast, a population of Garter snakes (Thamnophis elegans) with a slow Pace-of-Life (i.e. with small litter size and a long life span) had lower microbicidal ability against E. coli than a fast-living population (Sparkman & Palacios, 2009). The slow-living population also had comparatively lower haemagglutination and haemolysis titres. In an inter-specific study of seven Peromyscus mice species, life-history strategy was not associated with microbicidal ability against E. coli or with other immune measures (Martin et al., 2007). Investigations of induced immune responses in a life-history context also show mixed results (Tella et al., 2002; Ardia, 2005; Palacios & Martin, 2006; Martin et al., 2007). The ambiguous patterns arising from previous studies suggest that variation in immune function exhibited by different taxonomic units does not clearly align with a single life-history/physiology (i.e. Pace-of-Life) axis. Our study shows that even within closely related subspecies physiological constraints do not lead to a single general axis among immunity, corticosterone and metabolic rate.
Physiological variation among individuals
At the among-individual (i.e. within subspecies) level, only haemagglutination and haemolysis were associated with mass-specific BMR, a proxy for Pace-of-Life. That is, only the first, and not the second or third, principal components covaried with mass-specific BMR (Fig. 4, Table 4). Studies investigating the correlation between indices of constitutive immunity and Pace-of-Life characteristics at the individual level are still scarce, and results are mixed. In contrast to our current findings, Tieleman et al. (2005) report that microbicidal ability against E. coli correlates with mass-corrected BMR in tropical house wrens (Troglodytes aedon). Rubenstein et al. (2008), however, find no relationship between investment in breeding, another proxy for Pace-of-Life, and microbicidal ability against E. coli in superb starlings (Lamprotornis superbus). Ots & Horak (1996) found that great tits (Parus major) with high breeding investment have higher heterophil/lymphocyte ratios as compared to great tits with low breeding investment, suggesting a decrease in health status. Investigations of induced immunity and Pace-of-Life characteristics at the individual level also lead to equivocal results (Nordling et al., 1998; Ilmonen et al., 2002; Apanius & Nisbet, 2006). The set-up that we used, with a common environment with one set of environmental conditions (day length, temperature), reflects only one point along the reaction norms of each subspecies (Nussey et al., 2007). If the reactions norms of the different subspecies differ, this may have influenced the relationships among traits that we found (Stearns, 1989; Nussey et al., 2007). However, the ambiguous picture painted by the diversity of associations between immune indices and Pace-of-Life characteristics at the individual level in our study as well as in the literature implies that at the individual level other factors than constraints must be important; immunity does not simply overlay the Pace-of-Life axis.
Although our data cast doubt on the generality of a relationship between immunity, corticosterone and metabolic rate, the correlations among immune indices which we identified at the individual level may still represent constraints within the immune system. We found correlations between haemagglutination and haemolysis, between microbicidal ability against E. coli and C. albicans and between microbicidal ability against S. aureus and haptoglobin. If these correlations represent universal physiological constraints (i.e. intrinsic mechanistic connections between indices), then we would expect to find the same correlations in other species. Red knots (Calidris canutus) and several species of waterfowl show the same correlations between haemagglutination and haemolysis (Matson et al., 2006b; Buehler et al., 2008b,2008c, 2011). Furthermore, these correlations are in line with expectations based on the underlying physiological mechanism: natural antibodies, a central player in haemagglutination, interact functionally with lytic enzymes, like complement, to cause haemolysis in vitro (Ochsenbein & Zinkernagel, 2000; Janeway et al., 2004). However, there are also two studies that report insignificant correlations between haemagglutination and haemolysis (Mendes et al., 2006; Parejo & Silva, 2009), suggesting that natural antibodies and complement can also operate, at least partly, independently from each other. Of the three studies of red knots, microbicidal ability against E. coli and C. albicans is correlated in only one (Buehler et al., 2008b,2008c, 2009). In the previous studies, the correlation between haptoglobin and microbicidal ability against S. aureus was consistently nonsignificant (Matson et al., 2006b; Buehler et al., 2009, 2011). This panoply of results begs for more studies but also leads us to tentatively conclude that the correlations that we identified among immune indices at the level of individual birds are not universal and, therefore, unlikely to result from physiological constraints. Therefore, environmental factors are likely to have influenced the evolution of correlated traits and should be taken into account. For example, recently, ecological immunologists have stressed the importance of measurement of pathogen pressure in studies of the immune system (Horrocks et al., 2011; Pedersen & Babayan, 2011), and measurement of such environmental factors will improve our understanding of the role of physiology in life-history evolution.
After exploring variation in indices from three major physiological systems to determine the potential presence and role of physiological constraints in explaining life-history traits, we validate that metabolic rate covaries with traditional demographic life-history traits, but conclude that corticosterone and immune system do not. Therefore, if physiological constraints limit the evolutionary potential of life-history traits (Sheldon & Verhulst, 1996; Ricklefs & Wikelski, 2002; Lee, 2006; Hau et al., 2010), then these constraints have to be at least partly located in the metabolic system. Metabolic rate summarizes the physico-chemical substrate of life, and life-history activities, such as breeding, are likely to be tightly linked to availability of metabolic energy (Brown et al., 2004). Moreover, metabolic constraints continue to be at the centre of studies of metabolic damage and other molecular mechanisms of ageing (Speakman, 2005; Dowling & Simmons, 2009). Alternative, more ecologically based hypotheses can also explain the covariation among traits that characterize the Pace-of-Life syndrome. Combinations of environmental factors (e.g. food availability, predation risk, pathogen pressure) play simultaneous selective roles and may result in particular combinations of traits (Lande, 1986; Duckworth, 2010). The enormous global diversity in combinations with environmental selective factors might help explain why the metabolic system, and not the endocrine and immune systems, covaries with a more general (slow–fast) life-history axis. On the one hand, environmental factors that impact energy metabolism (e.g. temperature, moisture, food availability (Tieleman et al., 2003, 2004) might covary with factors that impact demographic traits (e.g. effects of food availability, nest predation or seasonality on clutch size or mortality (Ashmole, 1963; Chalfoun & Martin, 2007; Biancucci & Martin, 2010; Griebeler et al., 2010)). In this case, metabolic and demographic traits might covary without the presence of constraints. On the other hand, environmental factors that impact immune system (e.g. pathogen pressure (Guernier et al., 2004; Buehler et al., 2008c; Horrocks et al., 2011)) or endocrine function (e.g. environmental predictability) might vary in a way that is unrelated or tangentially related to metabolism and demographic life-history traits. We therefore advocate including environmental factors in investigations of the links between life history and the immune and endocrine systems.
- Top of page
- Materials and methods
We thank B. Helm, L. Trost, S. Kuhn, M. Trappschuh, W. Jensen, E. Koch and W. Goymann, in addition to the late E. Gwinner for support. E. Croese, S. Engel, A. Lohrentz, C. Muck, J. Partecke, M. Raess, C. Schmidt-Wellenburg, L. Trost, A. Wittenzellner and E. Yohannes helped collect the blood samples, and J. Leenders, A. Foucher and M. Versteegh helped analyse the blood samples. We also thank M. Visser and J. Tinbergen for helpful discussions. D. Buehler, W. Goymann, K. Matson and R. Mauck and five anonymous referees provided valuable comments on the manuscript. This work was supported by the Netherlands Organization for Scientific Research (Veni 863.04.023, BIT) and an E-bird grant (MAV).