SEARCH

SEARCH BY CITATION

Keywords:

  • artificial selection;
  • corticosterone;
  • heritability;
  • life history traits;
  • stress hormone;
  • testosterone;
  • zebra finch

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Vertebrates respond to environmental stressors through the neuro-endocrine stress response, which involves the production of glucocorticoids. We have selected independent, duplicate divergent lines of zebra finches for high, low and control corticosterone responses to a mild stressor. This experiment has shown that over the first four generations, the high lines have demonstrated a significant realized heritability of about 20%. However, the low lines have apparently not changed significantly from controls. This asymmetry in response is potentially because of the fact that all birds appear to be showing increased adaptation to the environment in which they are housed, with significant declines in corticosterone response in control lines as well as low lines. Despite the existence of two- to threefold difference in mean corticosterone titre between high and low lines, there were no observed differences in testosterone titre in adult male birds from the different groups. In addition, there were no consistent, significant differences between the lines in any of the life history variables measured – number of eggs laid per clutch, number of clutches or broods produced per pair, number of fledglings produced per breeding attempt, nor in any of egg, nestling and fledgling mortality. These results highlight the fact that the mechanisms that underlie variation in the avian physiological system can be modified to respond to differences between environments through selection. This adds an additional level of flexibility to the avian physiological system, which will allow it to respond to environmental circumstances.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Organisms experience environmental variability and to accommodate these environmental changes, individuals must modify their morphology, physiology and behaviour. Large stochastic changes in the environment may require that organisms mount emergency responses to potentially catastrophic, unpredictable events [e.g. storms (Wingfield, 1994b)]. Vertebrates respond to such events through the production of glucocorticoids (Nelson, 2000), high levels of which result in what is known as an emergency life history stage (Wingfield, 2002). Corticosterone (the main avian glucocorticoid) is involved in mediating physiological and behavioural responses to unpredictable environmental events (Wingfield & Ramenofsky, 1997, 1999). Free-living animals respond to stressful events by up-regulating their corticosterone production (Silverin, 1998a; Romero et al., 2000), which modulates metabolic processes including the mobilization of energy stores, the shutdown of digestive processes and an increase in the peripheral blood supply (Silverin, 1998b; Buchanan, 2000). Chronically elevated corticosterone tends to suppress immune function (Wingfield et al., 1997; Råberg et al., 1998), reduces reproductive activity, but increases dispersal and foraging behaviours (Wingfield et al., 1997; Buchanan, 2000). For a contrary example in which raised corticosterone is associated with enhanced immune function and survival see (Svensson et al., 2002). In summary, the conventional view is that corticosterone acts to suppress activities that are concerned with long-term survival and reproduction, in favour of those concerned with short-term survival thus maximizing fitness by ensuring immediate survival of a potentially catastrophic event (Wingfield, 2002).

As physiological mechanisms have been selected to produce optimal life history strategies, it is not necessarily correct to interpret an individual with high levels of corticosterone as less fit. It is equally possible that this is an adaptive response produced through selection on individual physiological trade-offs. The key point is whether the amount of corticosterone is appropriate to the context in which it occurs. That selection can act on patterns of hormone secretion is shown by the fact that individuals in populations that are frequently exposed to environmental stresses tend to have greater tolerance of stress (Wingfield et al., 1992, 1995). The most likely explanation for this observation, is that selection has acted to modulate the physiological response to stress in a way that is appropriate for the environment.

During a natural stressful event individuals that were capable of mounting a high stress response are more likely to respond appropriately and survive. However, in a benign environment an individual with high levels of corticosterone should experience reduced fecundity, compared with an individual with low levels of corticosterone. Therefore the observed relation between the magnitude of the stress response and life history variables will depend on the frequency and severity of stressful events in the environment.

A number of artificial selection experiments on the stress response have been conducted; as far as we can determine these have all been on birds with the focus generally on welfare issues in farmed animals. The longest running corticosterone selection experiment is on Japanese quail (Coturnix japonica). The heritability of peak corticosterone response has been estimated in these lines and has been found to be 0.20–0.30 after nine generations of selection (Satterlee & Johnson, 1988) and 0.14–0.30 after 27 generations (although for some generations selection was relaxed) (Odeh et al., 2003). The principle aim of our work was to provide an estimate of heritability and determine the extent of additive genetic variation in the stress response in a passerine widely used in laboratory work.

Some selection experiments have also shown life history trade-offs – experiments in domestic fowl have shown that birds selected for high stress responses were more susceptible to various diseases than low stress birds (Gross & Calmano, 1971; Thompson et al., 1980), while low corticosterone turkeys (Meleagris gallopavo) had higher growth rates and greater disease resistance than high line birds (Brown & Nestor, 1974). In the quail selection experiment low line birds have higher growth rates and faster onset of sexual maturity than high line birds (Satterlee & Johnson, 1985; Satterlee et al., 2002). From these studies, it would seem that there are trade-offs between the stress response and various other aspects of physiology, especially growth rates and immune defence. In our experiment we aimed explicitly to examine the effect of corticosterone selection on life history traits, predicting that if the stress response traded-off against investment in other costly traits then high corticosterone birds would have reduced fecundity and viability. In addition, we aimed to test the hypothesis that selection on corticosterone titres would produce correlated effects on testosterone titres, predicting that corticosterone and testosterone titres would covary.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We sought to establish duplicate independent lines of zebra finches selected for increased, decreased and random levels of corticosterone production (six lines in total). A pilot survey of corticosterone production in a small number of zebra finch (n = 4–6 at each time point) from blood samples obtained using standardized blood sampling and assay methods (see below) confirmed that peak production occurred after 20 min holding in a cloth bag [a standardized capture/restraint technique (Wingfield, 1994a; Fig. 1)]. These data also showed that peak corticosterone production showed considerable individual variation (range 3.06–183.29 nmol L−1n = 144). In July 1999 the founding generation (G0) was set up with 15 pairs of finches allocated to each of duplicate high, low and control lines (each line = 30 breeding individuals). To establish G0, control birds were selected randomly from the total available pool of birds, the remaining birds were divided into high and low lines by reference to their corticosterone titres (highest 50% into high lines, lowest 50% into low lines) and then randomly divided between the two duplicates (Table 1). Birds were sourced from a range of UK pet shops in order to reduce any initial relatedness. Birds in one line were housed together in a large internal aviary giving approximately 1 m3 per breeding pair on a 16 : 8 light : dark cycle and maintained at an ambient temperature of 20–24 °C (Jones & Slater, 1999). The humidity was maintained between 50% and 70% and the rooms sprayed with water two or three times per day. The birds were provided with ad libitum seeds (foreign finch mixture; Haith's Ltd., Cleethorpes, Lincolnshire, UK), Chinese millet sprays, mineralized grit, water and cuttlefish bone. The finches were provided with approximately 10 g of a 3 : 1 mixture of ‘nectarblend’ (Haith's Ltd.) and egg biscuit food (Haith's Ltd.) and either lettuce or cucumber daily with a cod liver oil supplement in the seed weekly. An excess of nesting baskets and boxes were provided. In each generation eggs were laid in all lines within 1 month of allocation. The resulting offspring were maintained in their natal nests for rearing. Cross-fostering chicks between nests would have been desirable to reduce the effects of common environmental effects confounding inherited genetic effects, but this was impractical.

image

Figure 1. Corticosterone titre changes with time restrained in bag in a classic stress response. Peak corticosterone titre is achieved around 20 min restraint. Symbols show mean ± SD corticosterone titre (nmol L−1) of four birds, except at 20 min when six birds were sampled.

Download figure to PowerPoint

Table 1.  Population measures of the corticosterone titres (nmol L−1) for the selected and control lines. Total number of individuals tested (n), population mean corticosterone titre (±SE), cumulative selection differential (Scum), heritability estimate (±SE) calculated as [response to selection (relative to control line)/cumulative selection differential] (h2), phenotypic variance (Vp) and additive genetic variance (VA).
 Generation nPopulation mean Scum h2 Vp VA
High line 1G15139.6 ± 3.813.80.26 ± 0.1687.022.6
G26434.4 ± 3.630.10.25 ± 0.10133.233.3
G35236.6 ± 3.045.70.21 ± 0.0556.811.9
G45127.5 ± 3.059.80.19 ± 0.0459.211.2
G55728.1 ± 2.579.10.20 ± 0.0341.88.4
High line 2G13933.0 ± 3.523.8−0.41 ± 0.0569.2−28.4
G27036.2 ± 3.042.00.08 ± 0.06142.211.4
G36137.7 ± 3.056.90.14 ± 0.0465.99.2
G44832.2 ± 2.670.50.20 ± 0.0342.18.42
G55430.7 ± 3.284.30.06 ± 0.0266.74.0
Control line 1G14733.4 ± 3.80.04 82.4 
G26824.0 ± 2.9−0.4 75.1 
G34824.6 ± 2.4−3.3 31.9 
G44714.2 ± 1.6−6.1 11.1 
G55510.8 ± 1.7−3.8 18.5 
Control line 2G15742.7 ± 3.8−3.7 98.2 
G24832.9 ± 3.8−7.0 154.2 
G35829.8 ± 2.4−4.5 40.7 
G44617.6 ± 2.0−4.7 27.4 
G55725.3 ± 2.7−1.5 50.4 
Low line 1G14226.7 ± 3.2−28.10.24 ± 0.0752.912.7
G24921.9 ± 2.6−35.90.06 ± 0.0565.13.9
G35820.3 ± 1.6−46.90.09 ± 0.0217.01.5
G45811.0 ± 1.1−56.40.06 ± 0.0110.00.6
G5547.40 ± 1.30−61.30.06 ± 0.0111.10.7
Low line 2G13330.6 ± 3.8−23.30.52 ± 0.1062.132.3
G27918.6 ± 1.8−37.90.38 ± 0.0674.828.4
G35521.9 ± 1.6−49.60.16 ± 0.0217.42.8
G45619.0 ± 2.0−58.9−0.03 ± 0.0232.81.0
G55116.0 ± 1.9−70.70.13 ± 0.0222.02.9

Young birds were ringed while in the nest with a coloured plastic split ring (A.C. Hughes, Hampton Hill, Middlesex, UK) to identify birds hatched in the same nest from birds hatched in other nests. The offspring were blood sampled 6 weeks post-fledging, between 09:00 and 12:00 hours, to avoid any confounding problems from variation in corticosterone production with age or time of day. In order to minimize disturbance effects, we ensured that the room in which the birds were housed was not visited for at least 16 h prior to blood sampling. At this point all birds were ringed with commercially available leg rings with unique identification numbers (A.C. Hughes) and removed from their natal aviary and housed together until a month prior to selection after which they were housed in single sex cages until selection. Once 60 surviving offspring from a line had been produced, progeny were allocated to breed or be discarded, by ranking the birds by corticosterone titre in each line and drawing the top 15 individuals of each sex (for the high lines) or the bottom 15 individuals of each sex for the low lines (i.e. approximately 50% truncation selection). Selection of breeding birds in the control lines was achieved by randomly selecting 15 individuals of each sex in each of the two control lines, this ensured that the control lines were exposed to the same selection procedure and so would run the same risks of inbreeding as the selected lines. As far as was achievable, all the lines were treated similarly and so birds within each of the lines only differed in the nature of the selection experienced – i.e. selection for high, low or control corticosterone response. Data on the strengths of selection exerted in each line at each generation are presented in Table 1. Each generation was started synchronously across the lines.

Corticosterone was sampled at 6 weeks post-fledging when the birds were independent of their parents. They had also completed a partial moult into adult plumage by this time and so could be sexed on plumage characters. Testosterone was assayed in males at 6 months. This time point was chosen as males at this age were sexually mature and would be breeding, if given access to females. Blood samples for the two hormones were not taken at the same time in order to minimize sampled blood volume. Blood samples were taken from the brachial vein after 20 min restraint in a small cloth bag (20 × 30 cm). They were placed in individual cloth bags immediately after capture. This is the standard capture-restraint protocol used in studies of the corticosterone response (Wingfield, 1994a). Blood sampling any one bird took approximately 30 s and 100 μL of blood was taken in heparinized glass capillaries. The blood was centrifuged at 11 000 g for 15 min and the plasma frozen at −20 °C. Plasma samples from the birds were assayed within approximately 3 months of drawing, using standard radio-immunoassay techniques. Corticosterone concentrations were measured after extraction of 20 μL aliquots of plasma in diethyl ether, by radioimmunoassay (Wingfield et al., 1992; Maddocks et al., 2001). Anti-corticosterone antisera (code B21-42 and B3-163; Esoterix Inc, Endocrinology, CA, USA) and [1,2,6,7-3H]-corticosterone label (Amersham, Little Chalfont, UK) were used throughout. The interassay coefficient of variation was 15.7%, and the intra-assay coefficient of variation was 3.1%. The mean extraction efficiency was 72%. The assay was run with 50% binding at 134 pg per tube, and the detection limit (for 7.3 μL aliquots of extracted plasma) was 1.76 nmol L−1.

As we were interested in the covariance between corticosterone and testosterone production, around 6 months post-fledging a random sample [an average of ten males per line per generation (range 5–17)] of sexually mature males were blood sampled. These males were held physically separate from females but in visual and auditory contact, in order to standardize exposure to females. This sample was processed and assayed for testosterone titre. Testosterone concentrations were measured in plasma samples by direct radioimmunoassay using anti-testosterone antiserum (code 8680-6004; Biogenesis, Poole, UK) and [125I]-testosterone label (code 07-189126; ICN Biomedicals, Basingstoke, UK) (Parkinson & Follett, 1995). The assay was run with 50% binding at 11.0 pg per tube and a detection limit of 0.068 nmol L−1 for the 20 μL plasma volumes that were run in the assay. The interassay coefficient of variation was 15.5% and the intra-assay coefficient of variation was 2.2%.

During breeding the cages were visited three times a week and all nests checked, records were maintained on number of eggs laid per clutch, number of eggs hatching per clutch, nestling and fledgling mortality (the latter from fledging to 6 weeks post-fledging), from the records we calculate the number of clutches and fledglings produced per breeding pair.

The selection experiment has been running for 5 years and we report here results from the first six generations (G0–G5).

Data analysis

We initially conducted a mixed model anova (GLMM) with sex and line (high, low or control) as fixed factors and replicate (one or two) and family as random factors (family nested within replicate and line). Generation was used as a covariate to test for changes in peak corticosterone titre in the six lines across the generations. We tested for significant line by generation interaction terms, which would indicate divergence of the lines through the selection process. Mixed model anova were run using the ReML procedure on Genstat. Subsequently, we estimated (i) realized heritability by estimating the cumulative selection differential (deviation of the selected individuals used as parents from the population mean in that generation) and (ii) the cumulative response to selection (response compared with starting population expressed relative to controls) at each generation in each line. The realized heritability (h2) is given by the linear regression coefficient of the cumulative response to selection (R) on the cumulative selection differential (S) (Falconer & Mackay, 1996; Lynch & Walsh, 2005). The variance of the heritability estimate was calculated according to Falconer & Mackay (1996) (p. 210). Because of the possibility of exerting different amounts of selection on the two sexes, if different numbers of males and females had been produced, we calculated changes in corticosterone titre and heritability estimates for all individuals combined and for the two sexes separately. We tested for a sex effect on corticosterone selection by including sex as a factor in the anova described above. Changes in life history traits were examined using anova with line and replicate as factors and generation as a covariate. Additionally, we ran a parallel analysis that was similar to the previous one but replaced the two factors line and replicate with a single factor with six groups. We did this in order to check for the possibility that a significant effect of line could be produced by a large effect in one replicate but not in the other, in which case the effect was likely to be because of idiosyncratic factors in one population rather than to corticosterone selection. Analyses were conducted in Splus 2000 (Mathsoft, 1999) and Genstat Release 7 (VSN, 2004). The distribution of residuals was checked for normality and homoscedasticity for each anova model and appropriate transformations made when required.

Ethical note

We have taken every care to minimize the welfare implications of this project. Corticosterone is known to be immunosuppressive and so it is possible that high corticosterone line birds could be relatively immunosuppressed when compared with low and control line birds. However, the results of the selection (see below) suggest that corticosterone titres in the high lines were maintained while they declined in the other lines. There were no mortality differences between the lines (as shown below) and so we are as confident as we can be that there were no deleterious consequences of this work on the birds. Birds were maintained on high quality diets through out and in large communal aviaries that meet current recommendations. All work was carried out under licence by the Home Office and after local ethical review.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Selection for corticosterone

In total we have bred and assayed 1714 birds (866 males and 848 females) from 10 to 20 families per line per generation (Table 1). Figure 2 shows how the peak corticosterone titre changed over the duration of the selection experiment. Overall there were significant effects of selection pressure (line) (Wald = 148.75, d.f. = 2, P < 0.001), sex (Wald = 7.37, d.f. = 1, P = 0.007) and generation (Wald = 36.35, d.f. = 1, P < 0.001). The effect of family was highly significant (change in deviance because of family effect = 187.78, d.f. = 139, P < 0.001), as was replicate (change in deviance because of replicate effect = 35.93, d.f. = 1, P < 0.001). There was also a significant line by generation interaction, showing that the lines diverged over time (Wald = 17.12, d.f. = 2, P < 0.001; Fig. 2). There has been a significant difference between the mean corticosterone titre of the high and control (and high and low lines), since generation 2, as shown by significant contrasts between corticosterone selections regimes. The figure also reveals that there has been a significant reduction in corticosterone titre in the two control lines (control 1 b = −6.5, r2 = 0.94, n = 6 P < 0.01; control 2 b = −5.1, r2 = 0.83, n = 6, P < 0.05).

image

Figure 2. Response to artificial selection per generation for four generations of selection, for positive selection (triangle symbols) and negative selection (square symbols), control lines are shown by circular symbols, in two replicate lines in each direction as shown by solid lines with filled symbols and dashed lines with open symbols. Points are population mean corticosterone titre (±SE). The box shows whether there was a significant post hoc contrast (Tukey's test) between corticosterone selection regimes for each generation.

Download figure to PowerPoint

The significant effect of sex in previous analyses suggests that it is justified to separate the two sexes. It is noticeable that females have significantly higher corticosterone titres than males, as shown by a significant sex effect in the analysis shown above. However, when the analyses are separated by sex they show that there have been qualitatively similar changes in each of the two sexes. There were also significant effects of line (males Wald = 115.04, d.f. = 2, P < 0.001; females Wald = 44.21, d.f. = 2, P < 0.001), and generation (males Wald = 37.49, d.f. = 1, P < 0.001; females Wald = 27.56, d.f. = 1, P < 0.001). A significant line by generation interaction term existed for both sexes showing there had been significant divergence in corticosterone titre between the lines during the process of selection (males Wald = 7.13, d.f. = 2, P < 0.05; females Wald = 10.32, d.f. = 2, P < 0.01). Both sexes also showed significant effects of replicate (males change in deviance because of replicate effect = 24.75, d.f. = 1, P < 0.001; females change in deviance because of replicate effect = 9.52, d.f. = 1, P < 0.01) but the family effect was nonsignificant for both sexes (males change in deviance because of family effect = 87.00, d.f. = 74, 0.01 > P > 0.05; females change in deviance because of family effect = 102.18, d.f. = 66, n.s.) In males considered separately, there have also been significant differences in mean corticosterone titre between the high and control lines (and high and low lines), since generation 2. This is similar to those seen in the combined dataset, but the males in the high line were significantly different from males in the low line by generation 1. In females however, the effect of selection is less marked, with significant differences between the high and low lines only emerging by generation 3 and between high and control lines by generation 4. There have been similar changes in corticosterone titre in the two control lines with time in both sexes, although these have been more marked in females than in males (males: control 1 b =−6.5, r2 = 0.98, n = 6, P < 0.001; control 2 b = −4.0, r2 = 0.67, n = 6, n.s.; females: control 1 b = −7.1, r2 =0.87, n = 6, P < 0.05; control 2 b = −6.7, r2 = 0.87, n =6, P < 0.05).

The fact that there have been significant declines in corticosterone titre in the control lines suggests that there have been consistent changes in the animals, independent of the selection regime. To remove this unwanted generation effect, the response to selection in the selected lines has been expressed as the difference from their respective control lines. Figure 3 shows the response to selection in the combined dataset against the cumulative selection differential. There were effects observed in the high lines with a significant regression coefficients in response to selection for high levels of peak corticosterone in one line and a near significant effect in the other (high 1 r2 = 0.87, n = 6, P < 0.05; high 2 r2 = 0.35, n =6, n.s.). However, there was no significant effect observed in either of the low lines which both changed in parallel with the controls, with nonsignificant regression coefficients in both these lines (low 1 r2 = 0.13, n =6, n.s.; low 2 r2 = 0.28, n = 6, n.s.). The data from male birds show similar results, with a significant response to upward selection and no significant effect of downward selection (high 1 r2 = 0.79, n = 6, P < 0.05; high 2 r2 =0.12, n = 6, n.s.; low 1 r2 = 0.04, n = 6, n.s.; low 2 r2 =0.21, n = 6, n.s.). The results obtained from female birds were also qualitatively similar to the combined dataset (high 1 r2 = 0.49, n = 6, n.s.; high 2 r2 = 0.36, n = 6, n.s.; low 1 r2 = 0.13, n = 6, n.s.; low 2 r2 = 0.23, n = 6, n.s.). Greater downward selection was exerted on females (cumulative selection differentials males low line 1 = −60.3, line 2 = −58.8; females low line 1 =−71.9, line 2 = −79.3), while greater upward selection was exerted on males (cumulative selection differentials males high line 1 = 83.6, line 2 = 102.5; females high line 1 = 75.7, line 2 = 74.9).

image

Figure 3. Response to selection expressed as deviation from relevant control in each generation, in relation to the cumulative selection differential. Symbols as in Fig. 2. Lines represent regression lines for each replicate, note that the regressions for the low selection lines are not significant but have been added for consistency. Some jitter has been added to the points at zero.

Download figure to PowerPoint

Realized heritability is calculated as the regression coefficients from the relations between cumulative responses to selection and cumulative selection differentials (Falconer & Mackay, 1996; Lynch & Walsh, 2005). Table 2 summarizes these results and shows that there was significant heritability in corticosterone production in the high lines, but not in the low lines and that heritability estimates were similar in the two sexes.

Table 2.  Realized heritability estimates (h2) ±SD for peak corticosterone response in high and low line zebra finches, calculated as the slope of the regression between cumulative selection differential and response to selection.
Direction of selectionReplicateAll birds combinedMales onlyFemales only
High10.24 ± 0.040.27 ± 0.040.20 ± 0.06
High20.10 ± 0.030.09 ± 0.040.11 ± 0.05
Low10.08 ± 0.020.04 ± 0.020.09 ± 0.02
Low20.14 ± 0.020.21 ± 0.030.10 ± 0.02

Correlated effects on testosterone

There was no significant difference in the plasma levels of testosterone in adult males in breeding condition in the different corticosterone selection lines although there were significant effects of generation (F1,206 = 21.54, P < 0.001), suggesting a decline in testosterone titre of 0.02 ± 0.001 nmol L−1 per generation that was similar in all lines. There was also a significant effect of replicate (F1,206 = 12.26, P < 0.001; replicate one being an average of 0.01 ± 0.001 nmol L−1 lower than replicate two) there are no significant effects of either line (F2,202 =0.35, P = 0.70) or line interacting with generation (F2,202 = 0.54, P = 0.58). When the line by generation interaction term is removed, the line term does not become significant (F2,204 = 0.20, P = 0.81). Therefore, the change in testosterone with time appears to have occurred in similar ways in the different lines (Fig. 4). If there has been a change in corticosterone titres over time and no change in testosterone titres over time, we should expect that there would be a significant corticosterone titre by generation interaction term to emerge in the data analysis of the birds for which we have data on both hormones. This is indeed the case (F1,191 = 17.10, P < 0.001).

image

Figure 4. Plasma testosterone titres in random samples of adult, breeding condition males from generations 2–4. Symbols as in Fig. 2. An average of 10 males were sampled per line per generation (range 5–17) representing about a third of the males in any generation.

Download figure to PowerPoint

Correlated effects on life-history traits

We have measurements of four life history traits – the number of eggs laid per clutch, the number of young fledged per breeding attempt, the number of broods per pair and the number of clutches produced per pair. In addition, we have assessed two aspects of viability – the proportion of nonhatching eggs in each clutch, nestling and fledgling mortality (to 6 weeks). The results obtained from the analysis of these traits are presented in Table 3. Table 3 reveals that the main changes in these traits has occurred with generation, with the number of eggs laid per clutch gradually increasing over the course of the selection in all lines. At the same time there have been slight increases in the both the number of eggs failing to hatch and nestling mortality. These factors have effectively cancelled out each other, as the number of young fledging per breeding attempt has remained at approximately 3–3.5 fledglings per breeding attempt throughout the experiment. No consistent, significant effects of either line or replicate could be determined in the data, although there were effects specific to particular populations – for example the low line replicate 2 demonstrated consistently low clutch sizes and relatively few clutches per pair. These effects are probably because of idiosyncratic differences in particular populations rather than to the selection regime.

Table 3.  Changes in life history and viability traits over the course of the selection experiment.
Source of variationTrait
Eggs laid perclutchFledglings per breeding attemptClutches per pairNumber of broods per pairEgg mortality (nonhatching)Nestling mortalityFledgling mortality
  1. ***P < 0.001; **0.001 < P < 0.01; *0.01 < P < 0.05.

  2. Statistics are presented when P-value is <0.1, statistics are from anova models containing generation as a covariate, corticosterone selection regime (three levels) and replicate (two levels) as a factors.

GenerationIncreases, F1,414 = 20.8***Declines, F1,413 = 10.3**Increases, F1,414 = 55.5*** Increases, F1,416 = 21.7***Increases, F1,364 = 6.4*Declines, F1,306 = 7.3**
Corticosterone selection lineLow lines fewer, F2,414 = 5.3** Low lines fewer, F2,414 = 6.8**    
ReplicateReplicate 2 low only group different from mean, F5,407 = 4.3** Replicate 2 low only group different from mean, F5,407 = 3.0*    

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The stress response shown by the zebra finches in this study (Fig. 1) is similar to that recently found in wild zebra finches, which have basal levels of corticosterone of around 15 nmol L−1 with a peak corticosterone titre of around 60 nmol L−1 after 30 min of restraint (N. Perfito, pers. comm.). This suggests that despite many generations in captivity stress responses are similar to those seen in wild populations.

We have shown that, at least in the upward direction, it has been possible to select for peak corticosterone titre with a realized heritability of about 0.20. This is similar to that achieved in Japanese quail, in which heritability of peak corticosterone in a similar selection experiment was initially found to be 0.20–0.30 (Satterlee & Johnson, 1988) and 0.14–0.30 after 27 generations (although for some generations selection had been relaxed; Odeh et al., 2003). Heritability estimates obtained from selection of young turkeys for corticosterone responses to cold stress have been reported to be 0.14–0.25 (Brown & Nestor, 1973). A selection experiment on heat stress tolerance in Japanese quail showed that heat stress tolerance also showed heritability of 0.13–0.24 (Bowen & Washburn, 1984). While the levels of heritability found in these other studies are comparable to our estimates of heritability from the high lines they are obviously much greater than those found in the low lines of the present study. However, in all these studies the realized heritability estimates found when selecting for low corticosterone responses were lower than for upward selection (quail 0.19 vs. 0.33 (after nine generations and 0.14 vs. 0.30 after 27, turkeys 0.14 vs. 0.25). Our study differs from these previous ones, as it failed to find significant divergence in the low line from the control line, despite the selection pressure being larger in the present study (selection differentials per generation our study averaging 14–17 nmol L−1 per generation; quail 4–10 nmol L−1 per generation calculated from (Satterlee & Johnson, 1988).

One explanation for our failure to produce divergence between the control and low lines could be because the birds in the control lines have shown a steadily declining peak corticosterone response over the course of the experiment (about 7 nmol L−1 per generation). The reason for this decline is not known as it implies selection against peak corticosterone response over and above the selection pressure that we have been exerting. If the selection differentials on the control lines are calculated it can be seen that over the experiment to date there has been a cumulative selection differential of −0.42 nmol L−1 in control line 1 and −4.5 nmol L−1 in control line 2. These are <10% of the comparable figure for the selected lines. It seems likely that the observed decline in the control lines has occurred because of adaptation to the animal house environment (which typically has few disturbances) rather than the selection process per se. Indeed, the fact that the raw data from the high lines are flat with respect to generation (Fig. 2) suggest that the selection for decreased corticosterone response in the animal house environment was comparable in magnitude to the selection pressure exerted by our experiment (i.e. about 20 nmol L−1 per generation). Interestingly, in the quail selection experiments, changes in peak corticosterone responses were also observed in the control line, although these changes were in upward rather than downward direction but were also ascribed to environmental effects outwith the selection process. Whatever the explanation for the downward drift of peak corticosterone responses in the control birds, the fact remains that the low lines have changed in parallel with the control lines and we have observed little, if any, divergence between them. It is likely that whatever the source of the negative selection on peak corticosterone responses in control lines, also acts on the selected lines and has swamped any effects of our selection. It should be noted that this decline across time must be the result of selection and is not habituation as each bird was only exposed to the stressor on a single occasion during its lifetime and so habituation cannot have occurred. An interesting way to test for animal house adaptation would be to test for correlations between fitness parameters and corticosterone titres of individual birds within lines.

A difference between the nature of the selection regimes on the high and low lines is that in the high lines we have been able to exert consistent selection on the birds (ranging between 14 and 23 nmol L−1 per generation). In contrast, in the low lines the selection pressure has been uneven with large selection differentials in the first bout of selection (−24 and −28 nmol L−1) and much less since (ranging between −7 and −14 nmol L−1 per generation). The initial bout of selection produced a large response to selection, while the subsequent smaller selections have produced little response. In fact in some generations the response has been less than the change in the control lines. In contrast, there have been fairly consistent responses in the high lines (Figs 2 and 3). It may be that if we were able to exert more consistent selection on the low lines then a greater effect might be seen. A final explanation for the failure to produce an effect in the low lines may be the fact that the corticosterone titres are generally low and close in some cases to the lower detection limit of the endocrine assay. This limits our ability to select in a downward direction.

We have observed that the population mean corticosterone titre has changed as the selection experiment progressed. We have interpreted this as changes in the peak stress response, however it is possible that it could be because of changes in the number of birds responding to the stressor. We feel this is unlikely firstly because the stressor is a standardized one in all generations. In addition, if this were true we should see changes in the proportion of birds remaining at basal levels, which was not observed.

Table 1 also shows that the additive genetic variance in the selected lines appears to decline during the course of the selection experiment. This is expected as theory suggests that additive genetic variance should diminish after strong directional selection. This might suggest that extending the selection experiment further may be difficult, as there would be a diminishing amount of additive genetic variance to utilize. It would also appear that particularly after the first generation, the additive genetic variance in the low lines diminished more quickly than that in the high lines. This suggests that the potential for selection in the low lines may have been more limited than in the high lines and so may provide an additional explanation for the failure to see significant divergence between the low lines and controls.

In our experiment, we have been obliged to select male and female finches independently of each other, as we needed to have equal numbers of males and females in each line. This has obviously raised the possibility that different selection differentials and possibly different responses to selection could be produced in the two sexes. Despite this, the heritability estimates for males and females considered separately were very similar. It is noticeable that females generally have higher peak corticosterone responses than males and that the control declines in peak corticosterone were steeper in females than in males. This may be an artefact of the experiment as in order to maximize genetic diversity between our birds we obtained males and females from different suppliers so these differences could be because of chance differences in the founding populations rather than being adaptive. Alternatively, it may be that females benefit from higher corticosterone titres than males in the same environment. This would be consistent with a study of rats, that has shown that estradiol in females rats raises their peak stress response (Viau & Meaney, 1991).

The fact that in this experiment, we have achieved significant divergence in peak corticosterone titre while during the same period there has been no divergence in testosterone titre in males suggests that we have changed the relationship between these two hormones. This is confirmed by the finding of a significant testosterone by corticosterone by generation interaction term, which shows that the relation between corticosterone titres and testosterone titres has changed over time. These results imply that the observed tendency for the two hormones to vary together is because of physiological effects, rather than any form of genetic linkage between alleles producing high testosterone titres and alleles producing high corticosterone responses.

We have shown that in the life history traits examined there have been some changes during the course of the experiment. The majority of these have been similar in all lines, which could indicate that they were because of inbreeding or to adaptation to their environment.

The only life history variables to differ significantly between lines were the number of eggs laid per clutch and the number of clutches produced per pair which were both lower in the low lines. Closer examination of the data suggests that this effect only really existed in low line, replicate 2 which consistently had small clutch sizes and fewer clutches per pair. It is unclear whether these effects were because of the selection regime or whether they were an idiosyncratic effect in one replicate only, although the latter seems more likely in this case.

The fact that there have been few changes in any of the life history traits, despite a large change in corticosterone titre, implies that any trade-offs that occur in response to changes in the stress response have occurred elsewhere. This is in contrast to the accumulating evidence that investment in immune defence appears to trade-off against investment in reproduction or survival (e.g. Moret & Schmidt-Hempel, 2000; Ardia et al., 2003), which has led to the argument that immune function should be regarded as a life history trait in its own right (Norris & Evans, 2000). A possible explanation for this difference is that while we are selecting for differences in the response to stress, the birds are living in a low-stress environment. It could be that having the ability to produce large amounts of corticosterone in response to stress is not particularly costly, although using it may be. Potentially, if the birds were moved into a more stressful environment where the differences in their ability to respond to stress might matter then differences would be observed more readily. Studies in poultry have shown that birds in high stress response lines are more susceptible to various diseases than low line birds (Gross & Calmano, 1971; Thompson et al., 1980); that low line birds had higher growth rates and greater disease resistance than high line birds (Brown & Nestor, 1974); low line birds have higher growth rates and faster onset of sexual maturity than high line birds (Satterlee & Johnson, 1985; Satterlee et al., 2002). From these studies, it would seem that there are trade-offs between the stress response and various other aspects of physiology, especially growth rates and immune defence.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

An experiment such as this requires the input of a large number of people and we are indebted to many individuals. The main ones are Joan Barnett and her team of animal facility staff who looked after the finches with great dedication. We have been helped in taking blood samples by several people but Thaís Martins, Kirsty Park and Sheila Donald have helped on more than one occasion. We would like to acknowledge the use of unpublished data generously provided by Nicole Perfito, University of Princeton. We would like to thank Allen Moore for constructive suggestions about the quantitative genetic sections of this paper. The project has been funded at various times by the Royal Society, Association for the Study of Animal Behaviour, University of Stirling and Natural Environment Research Council. All work reported here was conducted under Home Office licence number 60/2584. Finally, MRE would like to thank Mike Cherry for providing time and space to write this manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • Ardia, D.R., Schat, K.A. & Winkler, D.W. 2003. Reproductive effort reduces long-term immune function in breeding tree swallows (Tachycineta bicolor). Proc. R. Soc. Lond B Biol. Sci. 270: 16791683.
  • Bowen, S.J. & Washburn, K.W. 1984. Genetics of heat tolerance in Japanese quail. Poultry Sci. 63: 430435.
  • Brown, K.I. & Nestor, K.E. 1973. Some physiological responses of turkeys selected for high and low adrenal responses to cold stress. Poultry Sci 52: 19481954.
  • Brown, K.I. & Nestor, K.E. 1974. Interrelationships of cellular physiology and endocrinology with genetics. 2. Implications of selection for high and low adrenal response to stress. Poultry Sci. 53: 12971306.
  • Buchanan, K.L. 2000. Stress and the evolution of condition-dependent signals. Trends Ecol. Evol. 15: 156160.
  • Falconer, D.S. & Mackay, T.F.C. 1996. Introduction to Quantitative Genetics, 4th edn. Prentice Hall, Harlow, Essex, UK.
  • Gross, W.B. & Calmano, G. 1971. Effect of infectious agents on chickens selected for plasma corticosterone response to social stress. Poultry Sci 50: 12131217.
  • Jones, A.E. & Slater, P.J.B. 1999. The zebra finch. In: UFAW Handbook on the Care and Management of Laboratory Animals, Vol. 1 (T.Poole, ed.), pp. 722730. Blackwell, Oxford.
  • Lynch, M. & Walsh, B. 2005. Evolution and Selection of Quantitative Traits. http://nitro.biosci.arizona.edu/zbook/volume_2/vol2.html .
  • Maddocks, S.A., Cuthill, I.C., Goldsmith, A.R. & Sherwin, C.M. 2001. Behavioural and physiological effects of absence of ultraviolet wavelengths for domestic chicks. Anim. Behav. 62: 10131019.
  • Mathsoft, I. 1999. SPlus 2000 Professional Edition for Windows Release 1. Mathsoft, Cambridge, MA, USA.
  • Moret, Y. & Schmidt-Hempel, P. 2000. Survival for immunity: the price of immune system activation for bumblebee workers. Science 290: 11661168.
  • Nelson, R.J. 2000. An Introduction to Behavioral Endcrinology. Sinauer, Massachusetts.
  • Norris, K. & Evans, M.R. 2000. Ecological immunology: life history trade-offs and immune defense in birds. Behav. Ecol. 11: 1926.
  • Odeh, F.M., Cadd, G.G. & Satterlee, D.G. 2003. Genetic characterisation of stress responsiveness in Japanese quail. 2. Analyses of maternal effects, additive sex linkage effects, heterosis, and heritability by diallel crosses. Poultry Sci. 82: 3135.
  • Parkinson, T.J. & Follett, B.K. 1995. Thyroidectomy abolishes testicular cycles of Soay rams. Proc. R. Soc. Lond B Biol. Sci. 259: 16.
  • Råberg, L., Grahn, M., Hasselquist, D. & Svensson, E. 1998. On the adaptive significance of stress induced immunosuppression. Proc. R. Soc. Lond B Biol. Sci. 265: 16371641.
  • Romero, L.M., Reed, J.M. & Wingfield, J.C. 2000. Effects of weather on corticosterone responses in wild free-living passerine birds. Gen. Comp. Endocrinol. 118: 113122.
  • Satterlee, D.G. & Johnson, W.A. 1985. Metabolic traits in Japanese quail selected for high or low corticosterone response to stress. Poultry Sci. 64: 176.
  • Satterlee, D.G. & Johnson, W.A. 1988. Selection of Japanese quail for contrasting blood corticosterone response to immobilization. Poultry Sci. 67: 2532.
  • Satterlee, D.G., Marin, R.H. & Jones, R.B. 2002. Selection of Japanese quail for reduced adrenocortical responsiveness accelerates puberty in males. Poultry Sci. 81: 10711076.
  • Silverin, B. 1998a. Behavioral and hormonal responses of pied flycatchers to environmental stressors. Anim. Behav. 55: 14111420.
  • Silverin, B. 1998b. Stress responses in birds. Poultry Avian Biol. Revs. 9: 153168.
  • Svensson, E.I., Sinervo, B. & Comendant, T. 2002. Mechanistic and experimental analysis of condition and reproduction in a polymorphic lizard. J. Evol. Biol. 15: 10341047.
  • Thompson, D.L., Elgert, K.D., Gross, W.B. & Siegel, P.B. 1980. Cell-mediated immunity in Marek's disease virus-infected chickens genetically selected for high and low concentrations of plasma corticosterone. Am. J. Vet. Res. 41: 9196.
  • Viau, V. & Meaney, M.J. 1991. Variations in the hypothalamic-pituitary-adrenal response stress during the estrous-cycle in the rat. Endocrinology 129: 25032511.
  • VSN, I.L. 2004. Genstat Release 7.1, Genstat for Windows, 7th edn. Rothampstead Experimental Station, Harpenden, UK.
  • Wingfield, J.C. 1994a. Hormone-behavior interactions and mating systems in male and female birds. In: The Difference between the Sexes (R.V.Short & E.Balaban, eds), pp. 303330. Cambridge University Press, London.
  • Wingfield, J.C. 1994b. Modulation of the adrenocortical response to stress in birds. In: Perspectives in Comparative Endocrinology (K.G.Davey, R.E.Peter & S.S.Tobe, eds), pp. 520528. National research Council, Ottawa, Canada.
  • Wingfield, J.C. 2002. Ecophysiological studies of hormone-behavior relations in birds. In: Hormones, Brain and Behavior, Vol. 2 (D.W.Pfaff, A.P.Arnold, A.M.Etgen, S.E.Fairbach & R.T.Rubin, eds), pp. 587647. Academic Press, San Diego.
  • Wingfield, J.C. & Ramenofsky, M. 1997. Corticosterone and facultative dispersal in response to unpredictable events. Ardea 85: 155166.
  • Wingfield, J.C. & Ramenofsky, M. 1999. Hormones and the behavioral ecology of stress. In: Stress Physiology in Animals (P.H.M.Balm, ed.), pp. 151. Sheffield Academic Press, Sheffield.
  • Wingfield, J.C., Vleck, C.M. & Moore, M.C. 1992. Seasonal changes of the adrenocortical response to stress in birds of the Sonoran desert. J. Exp. Zool. 264: 419428.
  • Wingfield, J.C., O'Reilly, K.M. & Astheimer, L.B. 1995. Modulation of the adrenocortical responses to acute stress in arctic birds – a possible ecological basis. Am. Zool. 35: 285294.
  • Wingfield, J.C., Maney, D.L., Breuner, C.W., Jacobs, J.D., Lynn, S., Ramenofsky, M. & Richardson, R.D. 1997. Ecological bases of hormone-behaviour interactions: the ‘emergency life history stage’. Am. Zool. 38: 191206.