Supporting Information
Dingemanse et al.
Methods
Some of the methods used in the present work (crossing designs and extraction / interpretation of principal component axes) have been described previously, so we present here only those details relevant to our study.
Crossing designs. We applied two types of standard crossing schemes(Lynch & Walsh 1998). Partial North Carolina II (PNC II) breeding designs were used whenever ≥3gravid females from the same population were available for breeding on the same day, and full crosses in all other cases. For PNC II crosses, each female was crossed with 2 of m males, such that the 1st of n females was crossed with the 1st and 2nd of m males, the 2nd of n females was crossed with the 2nd and 3rd of them males and so on, but the last female (n) was crossed both with the last (m) and the 1st male. This design was applied using either n = m = 3 or n = m = 4, resulting in respectively 6 or 8 unique full-sib families per cross.For full crosses, each of n females was crossed with each of m males, resulting in n × m unique full-sib families per cross. An overview of the breeding designs is given in Table S3.
Extraction of behavioural axes.PCA analyses summarized behaviours quantified in each of the five personality essays into a few principal components that reflected the traits of interest (Table S4), echoing the results from an earlier study on these populations (Dingemanse et al. 2007). The resulting PCA scores were given functional labels (presented in italics throughout the paper) and used for further analyses.
Principal components A1, B1, and C1 described activity levels in novel (A1) and familiar environments (B1, C1), with high positive loadings for all measures of activity. The component A1 appeared to reflect a fish’s exploration of the novel environment, whereas B1 and C1 measured, respectively, general activity 2 h or 4 h after release. Following Dingemanse et al. (2002, 2007), we interpreted these data in this way because the number of unique squares visited was 1.7 times higher at the time of release into the novel tank compared to two hours later (21.91 ± 0.396 (mean ± SE) vs. 13.09 ± 0.363; paired t-test: t1026 = 19.163, P <0.0001) and did not decrease further between 2 and 4 h after release (13.16 ± 0.366 vs. 12.65±0.368; t1011 = 1.364, P = 0.173).
The sociability test measured two distinct principal behaviour axes. D1 measured lack of sociability, with high positive loadings for both the latency to first reach the conspecific and the mean distance between the focal and stimulus fish during the entire test: social fish quickly approached and then stayed in close proximity to the novel conspecific, whereas solitary fish instead remained near the refuge.Notably, aggressive interactions were not observed. D2 appeared to measure the explorationof novel conspecifics,with high positive loadings for variables describing movement (distance travelled) and number of orientations: exploratory fish tended to orient towards and inspect the conspecific shortly following exposure compared to non-exploratory fish.
E1 described lack of boldness towards predators, with high negative loadings for variables describing movement (distance travelled) and the number of inspections, and high positive loadings for latency to first move and minimum approach distance: inspecting fish moved towards the stimulus (shortly following exposure), then turned around, returned to cover and initiated new inspections, whereas non-inspecting fish remained still and did not approach the predator.
Figure Legends
Fig. S1.Diagram of the experimental set-up.
(a) Set-up of the novel environment tank, showing the area in which the subject was confined directly after capture from its growth tank, and the arrangement of stones during the subsequent test. (b) Arrangement of the test tank used for the sociability and predator tests. The tank contained a refuge (large stone and two pebbles); two dotted lines indicate the position of transparent barriers, solid vertical lines indicate rows (facilitating movement recording), the asterisk (*) the position where the conspecific (sociability test) or live perch (predator test) were introduced, and ‘s’ the position of the subject.
Table S1.Genetic parameters for six behavioural and three morphological traits.
We give here trait means, additive genetic variances (VA), residual variances (VR), heritabilities (h2), and significance of VA for two populations (one predator-sympatric, one predator-naïve) and two treatments (control vs. exposed to predators during ontogeny) for various personality traits (principal component axes given functional labels; numbers between brackets refer to axes numbers in Table S4) and morphological traits.
Table S1continued
(a) Exploration of novel environment (A1)Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / -0.153±0.083 / 0.208±0.104 / 0.745±0.100 / 0.219±0.102 / 10.182 / 0.001
Naïve / Exposed / Ne / -0.135±0.091 / 0.300±0.125 / 0.612±0.101 / 0.329±0.121 / 14.97 / <0.001
Sympatric / Control / Sc / 0.132±0.103 / 0.189±0.124 / 0.905±0.125 / 0.172±0.108 / 5.602 / 0.018
Sympatric / Exposed / Se / 0.196±0.090 / 0.128±0.090 / 0.800±0.102 / 0.138±0.093 / 5.248 / 0.022
(b) Activity 2 h after release (B1)
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / -0.136±0.069 / 0.105±0.073 / 0.741±0.086 / 0.124±0.083 / 4.36 / 0.037
Naïve / Exposed / Ne / -0.025±0.066 / 0.027±0.082 / 1.063±0.118 / 0.025±0.075 / 0.096 / 0.757
Sympatric / Control / Sc / -0.051±0.114 / 0.285±0.143 / 0.645±0.113 / 0.307±0.137 / 15.356 / <0.001
Sympatric / Exposed / Se / 0.197±0.117 / 0.268±0.143 / 0.840±0.127 / 0.242±0.118 / 12.466 / <0.001
(c) Activity 4 h after release (C1)
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / -0.117±0.059 / 0.023±0.056 / 0.833±0.088 / 0.026±0.065 / 0.220 / 0.639
Naïve / Exposed / Ne / 0.008±0.077 / 0.137±0.104 / 0.878±0.113 / 0.135±0.099 / 2.300 / 0.129
Sympatric / Control / Sc / -0.001±0.122 / 0.322±0.169 / 0.745±0.133 / 0.301±0.141 / 12.104 / 0.001
Sympatric / Exposed / Se / 0.113±0.114 / 0.258±0.138 / 0.776±0.121 / 0.249±0.121 / 12.912 / <0.001
(d) Sociability (-D1)
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / 0.087±0.064 / 0.046±0.069 / 0.860±0.097 / 0.051±0.076 / 0.588 / 0.443
Naïve / Exposed / Ne / 0.002±0.087 / 0.235±0.116 / 0.761±0.108 / 0.236±0.109 / 9.154 / 0.002
Sympatric / Control / Sc / -0.176±0.113 / 0.234±0.147 / 0.962±0.141 / 0.196±0.115 / 6.532 / 0.011
Sympatric / Exposed / Se / 0.050±0.126 / 0.364±0.158 / 0.506±0.112 / 0.418±0.151 / 28.407 / <0.001
(e) Exploration of novel conspecific (D2)
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / -0.175±0.075 / 0.156±0.089 / 0.651±0.089 / 0.193±0.105 / 5.884 / 0.015
Naïve / Exposed / Ne / -0.186±0.060 / 0.029±0.066 / 0.837±0.094 / 0.033±0.076 / 0.198 / 0.656
Sympatric / Control / Sc / 0.178±0.133 / 0.391±0.190 / 0.811±0.148 / 0.326±0.139 / 17.396 / <0.001
Sympatric / Exposed / Se / 0.259±0.108 / 0.215±0.130 / 0.743±0.118 / 0.225±0.126 / 7.012 / 0.008
(f) Boldness towards predators (-E1)
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / 0.090±0.091 / 0.261±0.119 / 0.718±0.108 / 0.267±0.112 / 13.64 / <0.001
Naïve / Exposed / Ne / 0.011±0.087 / 0.197±0.126 / 0.898±0.128 / 0.180±0.110 / 3.984 / 0.046
Sympatric / Control / Sc / -0.004±0.121 / 0.316±0.166 / 0.579±0.124 / 0.353±0.161 / 8.841 / 0.003
Sympatric / Exposed / Se / -0.180±0.120 / 0.288±0.151 / 0.733±0.127 / 0.282±0.133 / 12.014 / 0.001
Table S1continued
(g) Body lengthPopulation / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / 17.82±0.226 / 2.334±0.714 / 1.069±0.435 / 0.686±0.149 / 51.404 / <0.001
Naïve / Exposed / Ne / 17.62±0.193 / 1.665±0.498 / 0.884±0.309 / 0.653±0.142 / 61.818 / <0.001
Sympatric / Control / Sc / 16.78±0.303 / 2.589±0.871 / 0.182±0.469 / 0.934±0.178 / 69.494 / <0.001
Sympatric / Exposed / Se / 17.06±0.275 / 2.087±0.709 / 0.126±0.382 / 0.943±0.180 / 71.940 / <0.001
(h) Body shape
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / 3.797±0.027 / 0.035±0.003 / 0.000±0.000 / 1.000±0.000 / 104.738 / <0.001
Naïve / Exposed / Ne / 3.785±0.025 / 0.030±0.008 / 0.005±0.005 / 0.853±0.151 / 61.630 / <0.001
Sympatric / Control / Sc / 3.875±0.036 / 0.036±0.012 / 0.001±0.006 / 0.963±0.175 / 92.322 / <0.001
Sympatric / Exposed / Se / 3.876±0.030 / 0.023±0.009 / 0.010±0.005 / 0.693±0.181 / 44.418 / <0.001
(i) Relative spine length
Population / Treatment / Label / Mean±SE / VA±SE / VR±SE / h2±SE / / P
Naïve / Control / Nc / 1.243±0.013 / 0.006±0.003 / 0.015±0.002 / 0.286±0.114 / 13.792 / <0.001
Naïve / Exposed / Ne / 1.237±0.016 / 0.010±0.004 / 0.014±0.003 / 0.413±0.139 / 14.362 / <0.001
Sympatric / Control / Sc / 1.343±0.021 / 0.010±0.004 / 0.013±0.003 / 0.432±0.159 / 21.756 / <0.001
Sympatric / Exposed / Se / 1.358±0.022 / 0.010±0.005 / 0.019±0.004 / 0.347±0.149 / 15.012 / <0.001
Table S2.Fit of four alternative models for variation in genetic parameters.
Variation in additive genetic (VA), and residual variances (VR) among four population-treatment groups, for various personality traits (principal component axes given functional labels; numbers between brackets refer to axes numbers in Table S4) and morphological traits. Four models were considered four each trait (separate analyses for VA and VR): (i) equal variances across all groups (“Constant”), (ii) population differences (“Population”), (iii) treatment effects (“Treatment”), or (iv) unequal variances across all groups (additive or interactive effects of population and treatment; “Both”). We give AIC (Akaike Information Criterion), AIC (the difference between AIC of the given model and that with the lowest AIC), AIC weight (AICw: measure of support), and Deviance (-2*Loglikelihood) (Akaike 1973; Burnham & Anderson 2002). Models with AIC<2 fit the data relatively well. The model best supported by the data (i.e. that with AIC=0) is underlined, separately for each variance component.
Table S2 (continued)
(a) Exploration of novel environment (A1)Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 1017.4 / 0.0 / 0.494 / 1007.4 / 1019.6 / 0.3 / 0.336 / 1009.6
Population / 1018.6 / 1.2 / 0.275 / 1006.6 / 1019.4 / 0.0 / 0.382 / 1007.4
Treatment / 1019.4 / 2.0 / 0.184 / 1007.4 / 1020.9 / 1.5 / 0.183 / 1008.9
Both / 1022.1 / 4.7 / 0.048 / 1006.1 / 1022.1 / 2.7 / 0.099 / 1006.1
(b) Activity 2 h after release (B1)
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 1037.5 / 1.6 / 0.255 / 1027.5 / 1040.8 / 3.2 / 0.116 / 1030.8
Population / 1035.9 / 0.0 / 0.557 / 1023.9 / 1041.5 / 3.9 / 0.081 / 1029.5
Treatment / 1039.5 / 3.6 / 0.094 / 1027.5 / 1037.6 / 0.0 / 0.576 / 1025.6
Both / 1039.5 / 3.6 / 0.094 / 1023.5 / 1039.5 / 1.9 / 0.227 / 1023.5
(c) Activity 4 h after release (C1)
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 1028.9 / 2.5 / 0.171 / 1018.9 / 1024.0 / 0.0 / 0.517 / 1014.0
Population / 1026.3 / 0.0 / 0.605 / 1014.3 / 1025.4 / 1.4 / 0.254 / 1013.4
Treatment / 1030.3 / 3.9 / 0.084 / 1018.3 / 1025.9 / 2.0 / 0.193 / 1013.9
Both / 1029.3 / 2.9 / 0.140 / 1013.3 / 1029.3 / 5.3 / 0.036 / 1013.3
(d) Sociability (-D1)
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 987.6 / 0.6 / 0.241 / 977.6 / 991.4 / 3.6 / 0.082 / 981.4
Population / 987.2 / 0.3 / 0.287 / 975.2 / 991.2 / 3.3 / 0.093 / 979.2
Treatment / 986.9 / 0.0 / 0.333 / 974.9 / 987.9 / 0.0 / 0.494 / 975.9
Both / 988.7 / 1.7 / 0.140 / 972.7 / 988.7 / 0.8 / 0.330 / 972.7
(e) Exploration of novel conspecific (D2)
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 946.7 / 1.2 / 0.236 / 936.7 / 943.8 / 0.0 / 0.467 / 933.8
Population / 945.5 / 0.0 / 0.430 / 933.5 / 945.8 / 2.0 / 0.176 / 933.8
Treatment / 947.2 / 1.7 / 0.184 / 935.2 / 944.8 / 1.0 / 0.285 / 932.8
Both / 947.6 / 2.1 / 0.150 / 931.6 / 947.6 / 3.8 / 0.071 / 931.6
(f) Boldness towards conspecific (-E1)
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 933.2 / 0.0 / 0.542 / 923.2 / 935.9 / 0.0 / 0.376 / 925.9
Population / 935.0 / 1.8 / 0.221 / 923.0 / 936.7 / 0.8 / 0.246 / 924.7
Treatment / 935.1 / 1.9 / 0.205 / 923.1 / 936.4 / 0.5 / 0.292 / 924.4
Both / 938.8 / 5.6 / 0.032 / 922.8 / 938.8 / 2.9 / 0.086 / 922.8
(g) Body length
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / 1770.1 / 0.0 / 0.483 / 1760.1 / 1773.5 / 2.3 / 0.201 / 1763.5
Population / 1771.8 / 1.7 / 0.207 / 1759.8 / 1771.2 / 0.0 / 0.633 / 1759.2
Treatment / 1771.3 / 1.2 / 0.267 / 1759.3 / 1775.4 / 4.3 / 0.075 / 1763.4
Both / 1775.0 / 4.9 / 0.042 / 1759.0 / 1775.0 / 3.9 / 0.091 / 1759.0
(h) Body shape
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / -2587.3 / 0.0 / 0.501 / -2597.3 / -2585.8 / 0.3 / 0.348 / -2595.8
Population / -2585.6 / 1.7 / 0.212 / -2597.6 / -2584.5 / 1.5 / 0.185 / -2596.5
Treatment / -2585.8 / 1.5 / 0.239 / -2597.8 / -2586.1 / 0.0 / 0.396 / -2598.1
Both / -2582.6 / 4.7 / 0.048 / -2598.6 / -2582.6 / 3.4 / 0.071 / -2598.6
(i) Relative spine length
Comparison of VA / Comparison of VR
Model / AIC / AIC / AICw / Dev. / AIC / AIC / AICw / Dev.
Constant / -2706.5 / 0.0 / 0.496 / -2716.5 / -2706.0 / 0.0 / 0.524 / -2716.0
Population / -2704.9 / 1.5 / 0.230 / -2716.9 / -2704.0 / 2.0 / 0.197 / -2716.0
Treatment / -2704.9 / 1.5 / 0.230 / -2716.9 / -2704.2 / 1.7 / 0.220 / -2716.2
Both / -2701.6 / 4.8 / 0.044 / -2717.6 / -2701.6 / 4.4 / 0.059 / -2717.6
Table S2 (continued)
1
Table S3.Overview of breeding designs.
For each type of cross, we give the number of males (m) and the number of females (n) involved in the cross, the number of times the cross was applied (n replicates), and total numbers of males and females used (Full = full cross; PNC II = partial North Carolina II cross) for the (a) predator-sympatric population and (b) predator-naïve population.
(a) Design / m / n / n replicates / total m / total nPNC II / 3 / 3 / 2 / 6 / 6
Full / 2 / 2 / 2 / 4 / 4
Full / 3 / 2 / 1 / 3 / 2
Full / 3 / 1 / 3 / 9 / 3
Total / 8 / 22 / 15
(b) Design / m / n / n replicates / total m / total n
PNC II / 3 / 3 / 7 / 21 / 21
PNC II / 4 / 4 / 1 / 4 / 4
Full / 2 / 2 / 1 / 2 / 2
Full / 3 / 1 / 1 / 3 / 1
Total / 10 / 30 / 28
Table S4.Principal component analyses.
Results of principal component analyses (combined analyses for all the groups) on behavioural measures recorded (italics), for each of five behavioural tests. Loadings, eigenvalues and explained variance are given for the emerging axes.
(a) Novel environment test / Comp. A1No. of unique squares visited / 0.90
Distance travelled / 0.90
Eigenvalue / 1.63
% variance explained / 81.67
(b) Activity test (2 h after release) / Comp. B1
No. of unique squares visited / 0.91
Distance travelled / 0.91
Eigenvalue / 1.64
% variance explained / 82.08
(c) Activity test (4 h after release) / Comp. C1
No. of unique squares visited / 0.90
Distance travelled / 0.90
Eigenvalue / 1.61
% variance explained / 80.48
(d) Novel conspecific test / Comp. D1 / Comp. D2
Distance travelled / 0.05 / 0.93
Mean distance from conspecific / 0.92 / 0.21
No. of orientations / -0.11 / 0.85
Latency to reach conspecific / 0.91 / -0.22
Latency to first move / 0.56 / -0.50
Eigenvalue / 2.01 / 1.95
% variance explained / 40.14 / 38.94
(e) Predator-inspection test / Comp. E1
No. of inspections / -0.79
Distance travelled / -0.72
Minimum approach distance / 0.80
Latency to first move / 0.76
Eigenvalue / 2.35
% variance explained / 58.77
Reference List
Akaike, H. 1973 Information theory and an extension of the maximum likelihood principle. In International symposium on information theory (eds. B. N. Petran and F. Csáki), pp. 267-281. Budapest: Akadémiai Kiadi.
Barber, I. and Arnott, S. A. 2000 Split-clutch IVF: a technique to examine indirect fitness consequences of mate preferences in sticklebacks. Behaviour 137, 1129-1140.
Burnham, K. P. and Anderson, D. R. 2002 Model selection and multimodel inferences: a practical information-theoretic approach, 2nd Edition edn. New York: Springer-Verlag.
Dingemanse, N. J., Both, C., Drent, P. J., van Oers, K., and van Noordwijk, A. J. 2002 Repeatability and heritability of exploratory behaviour in great tits from the wild. Anim Behav 64, 929-937.
Dingemanse, N. J., Wright, J., Kazem, A. J. N., Thomas, D. K., Hickling, R., and Dawnay, N. 2007 Behavioural syndromes differ predictably between 12 populations of stickleback. J Anim Ecol 76, 1128-1138.
Lynch, M. and Walsh, B. 1998 Genetics and analysis of quantitative traits. Sunderland, MA: Sinauer Associates.
1