Supplementary material: Colony genetic diversity affects task performance in the red ant Myrmicarubra,Behavioural Ecology and Sociobiology
Slaa EJ, Chappell P and Hughes WOH, Institute of Integrative and Comparative Biology, University of Leeds, UK,
Table S1 Details of study colonies (number of workers and brood during experiments 1 and 2 and original queen number when colonies arrived at the lab during summer 2009), number of workers genotyped per colony (only the individuals for which information of the specific locus was obtained are listed) and several measures of colony genetic diversity. Most colonies came from Population 1 (near Berlin), except for colonies 5, 6, 7 and 8 (Population 2, near Boppard). Colony 19 had two queens up to 30 October 2009, but one of the queens died during hibernation. During Experiment 2, colony 15 had already migrated into the dark tube before destruction of the old nest, and the queen had disappeared from colony 18, so that these two colonies were excluded from further analyses. The SE around genetic relatedness (r) was calculated from jacknifing over loci, using the Relatedness 5.0.8 software (Queller and Goodnight 1989). Standardized allelic richness was calculated using the rarefaction method (using PAST software, rarefied to lowest number of alleles scored and averaged over loci). Nei’s gene diversity (i.e. expected heterozygosity averaged over loci) was calculated from the squared allele frequencies (1 - ∑ki=1pi2), where pi is the frequency of allele i). Shannon’s index of diversity was applied to allele frequencies and averaged over loci (H’ = -∑si=1pi log10pi) where pi is the frequency of allele i). Our measure of genetic relatedness was significantly correlated with Nei’s gene diversity (Pearson’s correlation r = -0.84, df = 15, P = 2.0e-05) and Shannon’s index of diversity (Pearson’s correlation r = -0.66,df = 15, P = 0.004), and tended to be negatively correlated with Allelic richness (Spearman rank correlation rho = -0.39, P =0.12).
Colony / Workers / Brood / Workers / Brood / Original # queens / Workers
genotyped / Relatedness (r) / SE / # Alleles / Standardised
allelic richness / Nei's gene
diversity / Shannon's index
of diversity
1 / 20 / 7 / 1 / 16 / 0.08 / 0.14 / 16 / 2.53 / 0.37 / 0.26
2 / 110 / 18 / 90 / 20 / 1 / 31 / 0.07 / 0.07 / 11 / 1.83 / 0.34 / 0.21
3 / 115 / 25 / 102 / 28 / 1 / 86 / 0.27 / 0.07 / 14 / 1.91 / 0.45 / 0.21
5 / 65 / 40 / 53 / 40 / 1 / 23 / -0.11 / 0.05 / 16 / 2.63 / 0.44 / 0.32
6 / 124 / 83 / 106 / 75 / 1 / 96 / 0.40 / 0.27 / 18 / 2.32 / 0.53 / 0.26
7 / 62 / 45 / 68 / 48 / 1 / 25 / -0.02 / 0.04 / 16 / 2.62 / 0.43 / 0.31
8 / 158 / 130 / 135 / 86 / 27 / 0.19 / 0.17 / 17 / 2.59 / 0.36 / 0.27
10 / 82 / 80 / 81 / 60 / 1 / 27 / 0.58 / 0.20 / 13 / 2.07 / 0.27 / 0.20
11 / 87 / 70 / 88 / 55 / 1 / 22 / 0.67 / 0.22 / 11 / 1.74 / 0.23 / 0.15
12 / 163 / 50 / 180 / 80 / 1 / 29 / 0.14 / 0.14 / 11 / 1.78 / 0.31 / 0.20
13 / 125 / 75 / 146 / 80 / 1 / 27 / -0.02 / 0.08 / 14 / 2.12 / 0.37 / 0.25
14 / 58 / 27 / 58 / 18 / 1 / 22 / 0.06 / 0.07 / 11 / 1.83 / 0.34 / 0.21
15 / 120 / 58 / 1 / 25 / -0.01 / 0.13 / 13 / 2.06 / 0.37 / 0.24
16 / 85 / 22 / 81 / 29 / 1 / 86 / 0.37 / 0.07 / 11 / 1.80 / 0.41 / 0.18
17 / 80 / 22 / 77 / 26 / 1 / 65 / 0.36 / 0.20 / 15 / 1.92 / 0.45 / 0.20
18 / 75 / 30 / 87 / 10 / 1 / 24 / 0.21 / 0.17 / 13 / 2.08 / 0.33 / 0.22
19 / 86 / 47 / 86 / 48 / 2 / 16 / 0.14 / 0.19 / 10 / 1.67 / 0.32 / 0.19
Table S2Loci used with number of individuals used, number of alleles observed, frequency of the most common allele, observed heterozygosity, expected heterozygosity, and amplification conditions. Thermal cycling conditions were: an initial denaturation phase for 5 min at 95 °C, then 40 or 45 cycles consisting of 40 s at 95 °C, 40 s at annealing temperature, and 40 s at 72 °C, with a final extension phase for 15 min at 72 °C. Observed and expected heterozygosity were calculated in Arlequin v3.1 and averaged over colonies.
MP-67 / Msca7 / Msca50 / MS26 / MS86 / MS3.62Number of individuals analysed / 600 / 669 / 588 / 581 / 587 / 537
Population 1 / 448 / 504 / 433 / 435 / 435 / 399
Population 2 / 152 / 165 / 155 / 146 / 152 / 138
Number of alleles observed / 6 / 3 / 2 / 2 / 5 / 7
Population 1 / 3 / 3 / 2 / 2 / 4 / 6
Population 2 / 5 / 1 / 2 / 2 / 4 / 5
Frequency most common allele
Population 1 / 0.70 / 0.52 / 0.76 / 1.00 / 0.82 / 0.46
Population 2 / 0.53 / 1.00 / 0.75 / 0.56 / 0.63 / 0.70
Observed heterozygosity
Population 1 / 0.45 / 0.47 / 0.50 / 0.001 / 0.35 / 0.54
Population 2 / 0.64 / 0.00 / 0.50 / 0.52 / 0.60 / 0.51
Expected heterozygosity
Population 1 / 0.40 / 0.43 / 0.38 / 0.001 / 0.27 / 0.51
Population 2 / 0.59 / 0.00 / 0.38 / 0.49 / 0.53 / 0.43
Amplification conditions
Total volume (µl) / 25 / 25 / 25 / 15 / 15 / 15
DNA template (µl) / 5 / 5 / 5 / 3 / 3 / 3
MgCl2 (mM) / 2.5 / 3.0 / 1.5 / 2.0 / 1.5 / 1.5
dNTP (µM) / 100 / 100 / 100 / 250 / 250 / 250
Forward primer (µM) / 0.5 / 0.2 / 0.5 / 0.2 / 0.2 / 0.2
Reverse primer (µM) / 0.5 / 0.2 / 0.5 / 0.2 / 0.2 / 0.2
Taq (U) / 0.9 / 0.5 / 0.9 / 0.8 / 0.8 / 0.8
Ta (°C) / 46 / 62.5 / 63 / 50 / 50 / 46
Number of cycles / 40 / 40 / 40 / 45 / 40 / 45
Table S3Summary of the Minimum Adequate Model testing the effects of genetic diversity (D), colony size (number of workers (W) or amount of brood (B)), and population (P) or day of experiment (Day) on nest site selection (tube (N = 11) or box (N = 4) Experiment 1, binomial regression model, link = log). We considered colonies that migrated partly into the box as having selected the box. N = 15 colonies (two colonies migrated under the tube and were excluded from this analysis). Data are non-orthogonal (i.e. observational), with a significant correlation between the number of workers and brood (Pearson’s correlation: r = 0.58, N = 15, P = 0.02). Therefore, the variables workers and brood were not included in the same model. When workers and genetic diversity were included in the model, neither variable was significant. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Nest site selection, Experiment 1
MAM (D+P/Day+B)
(Intercept) / 5.84 / 3.08 / 1.90 / 0.001**
Brood (B) / -0.08 / 0.05 / -1.75 / 0.006**
Non-significant variables removed, in reverse order of removal
Diversity (D) / 0.75
Population (P) / 0.71
Day / 0.54
D / 0.70
MAM (D+P/Day+W)
(Intercept) / 1.01 / 0.58 / 1.73 / 0.065
Non-significant variables removed, in reverse order of removal
Workers (W) / 0.11
Diversity / 0.11
Population / 0.20
Day / 0.44
Table S4Summary of the Minimum Adequate Model testing the effects of genetic diversity (1 - r), colony size (brood or workers), and population (P) or day of the experiment (Day) on the total migration time during Experiment 1 (negative binomial regression model, link=log). Only colonies that migrated into the tube and contained one queen and at least 20 workers were included in the analysis. N = 11 colonies. Data are non-orthogonal (i.e. observational). P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. The model including Population had a lower AIC score than the model including Day, and was thus preferred.
Estimate / Std Error / Z value / Pr(Chi)Total migration time, Experiment 1
MAM (D*P, AIC: 124.18)
(Intercept) / 12.91 / 4.44 / 2.91
Diversity(D) / -7.41 / 4.17 / -1.78 / 0.022
Population[1] / -8.79 / 4.46 / -1.97 / 0.034
D:Population[1] / 8.82 / 4.20 / 2.10 / 0.054
Non-significant variables removed
workers / 0.35
brood / 0.61
MAM (D*Day, AIC: 127.54)
(Intercept) / 7.38 / 1.48 / 4.98
Diversity(D) / -2.49 / 1.65 / -1.51 / 0.098
Day / -1.49 / 0.75 / -1.98 / 0.180
D:Day / 1.79 / 0.89 / 2.02 / 0.064
Non-significant variables removed
workers / 0.12
brood / 0.22
Table S5Summary of the Minimum Adequate Model testing the effects of genetic diversity (1 - r), number of workers and amount of brood (larvae and pupae) on the total migration time during Experiment 1 (negative binomial regression model, link=log). Only colonies from Population 1that migrated into the tube were included in the analysis. Data are non-orthogonal (i.e. observational), without significant correlations between the tested variables (Pearson’s correlations: P > 0.25 for all combinations). Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 9 colonies. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. Including “day of experiment” as a factor showed no significant effect of “day” on migration time (last rows of table).
Estimate / Std Error / Z value / Pr(Chi)Total migration time Experiment 1, Population 1
MAM (D+B*W)
(Intercept) / 4.12 / 0.44 / 9.42 / 3.17e-07***
Diversity(D) / 1.41 / 0.53 / 2.63 / 0.027*
Non-significant variables removed, in reverse order of removal
Workers (W) / 0.39
Brood (B) / 0.98
B:W / 0.15
(D*Day)
Day / 0.55
Day:D / 0.60
Estimate / Std Error / Z value / Pr(Chi)
Discovery time, Experiment 1
MAM (D+Day+W/B)
(Intercept) / 0.82 / 0.95 / 0.87 / 0.48
Diversity (D) / 2.84 / 0.90 / 3.16 / 0.020
Day / 0.64 / 0.23 / 2.72 / 0.030
Non-significant variables removed, in reverse order of removal
Workers / 0.96
Brood / 0.85
Model including ‘Population’ instead of ‘Day of Experiment’ ( D*Population+W/B)
Non-significant variables removed, in reverse order of removal
D / 0.16
Population / 0.13
Workers / 0.96
Brood / 0.94
D:Population / 0.94
1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111Table S6Summary of the Minimum Adequate Model testing the effects of genetic diversity (1 - r), population orday of experiment (Day), number of workers and amount of brood (larvae and pupae) on the time to discover a new nest site during Experiment 1 (negative binomial regression model, link = log). Data are non-orthogonal (i.e. observational), with a significant correlation between the number of workers and brood (Pearson’s correlation: r = 0.60, N = 17, P = 0.01). Therefore, workers and brood were not included in the same model, but both models resulted in the same MAM. Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 17 colonies. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. For the model that included population instead of day of experiment, none of the variables were significant (last rows of table).
Table S7Summary of the Minimum Adequate Model testing the effects of genetic diversity (1-r), number of workers and population on the number of ant minutes exploring (i.e. present around the new nest sites) before a new nest site was found during Experiment 1 (negative binomial regression model, link = log). Data are non-orthogonal (i.e. observational). Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 17 colonies. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. Day of experiment (Day) had no significant effect on total exploration time (last rows of table).
Estimate / Std Error / Z value / Pr(>|z|)Total exploration time (ant minutes), Experiment 1
MAM (D+W+P)
(Intercept) / 1.96 / 1.39 / 1.41
Diversity / 3.12 / 1.17 / 2.66 / 0.048*
Workers / 0.015 / 0.007 / 2.21 / 0.052
Population1 / 1.47 / 0.59 / 2.50 / 0.051
Model including day of experiment instead of population (D+W+Day)
Diversity / 0.08
Day / 0.12
Workers / 0.14
Table S8 Summary of the Minimum Adequate Model testing the effects of genetic diversity, number of workers, time to discover the new nest site, and population (P) or day of experiment (Day) on quorum threshold (Experiment 1, linear regression model). Only colonies which migrated into the tube, contained one queen and at least 20 workers were included in the analyses. N = 11 colonies. Data are non-orthogonal (i.e. observational), without significant correlations between the continuous variables(P > 0.17). The two populations differed in colony size (Welch two sample t-test t = 2.37, df = 8.195, P = 0.045) and genetic diversity (Welch two sample t-test t = 3.68, df = 6.07, P = 0.01). Variables did not differ significantly among the three days (AnovaworkersF1,9= 0.19, P = 0.67, diversityF1,9 = 2.04, P = 0.18, discovery timeF1,9 = 2.12, P = 0.18). P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. The MAM derived from the model including Population (upper rows) explained more of the variance in quorum threshold (r2 = 0.60) than the model including Day (lower rows, r2 = 0.30), and was thus preferred.
Estimate / Std Error / Z value / Pr(Chi)Quorum threshold, Experiment 1
MAM (P+T+W+D; r2 = 0.60, P = 0.01)
(Intercept) / -1.40 / 3.30 / -0.43 / -1.4e-14***
Population [1] / 8.59 / 3.24 / 2.66 / 0.029*
Discovery time (T) / 0.066 / 0.029 / 2.26 / 0.054
Non-significant variables removed, in reverse order of removal
Workers (W) / 0.25
Genetic diversity (D) / 0.88
MAM (Day+T+W+D; r2 = 0.30, P = 0.048)
(Intercept) / 2.24 / 4.23 / 0.53 / 0.61
Workers (W) / 0.10 / 0.04 / 2.29 / 0.048*
Non-significant variables removed, in reverse order of removal
Day / 0.12
Discovery time (T) / 0.33
Genetic diversity (D) / 0.65
Table S9Summary of the Minimum Adequate Model testing the effects of genetic diversity, number of workers, time to discover the new nest site, and population or day of experiment (Day) on time to achieve quorum after discovery (Experiment 1, negative binomial regression model, link = log). Only main effects were tested in this model. Only colonies that migrated into the tube and contained one queen and at least 20 workers were included in the analyses. N = 11 colonies. Data are non-orthogonal (i.e. observational), without significant correlation between the variables (Pearson’s correlations: P > 0.33 for all three pairs). P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. A different model, where the variable “workers” was replaced by the variable “quorum threshold” gave very similar results, and quorum threshold was not significant (with population: X21 = 0.10, P = 0.29; with day of experiment: X21 = 2.05, P = 0.15).
Estimate / Std Error / Z value / Pr(Chi)Time to quorum, Experiment 1
MAM
(Intercept) / 5.471 / 0.69 / 7.97 / 3.4e-9***
Discovery time / -0.018 / 0.007 / -2.49 / 0.037*
Non-significant variables removed, in reverse order of removal
Genetic diversity (1-r) / 0.21
Population / 0.12
Day / 0.51
Workers / 0.70
Table S10 Summary of the Minimum Adequate Model testing the effects of genetic diversity, number of workers and number of brood on migration time after achieving quorum (Experiment 1, negative binomial regression model, link = log). Only main effects were tested in this model. Only colonies that had migrated into the tube and which contained one queen and at least 20 workers were included in the analyses. N = 11 colonies. Data are non-orthogonal (i.e. observational), without significant correlations between the variables (P > 0.25 for all comparisons). When day of experiment (Day) was included as a factor instead of population, it had no significant effect on migration time after quorum. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Migration time after quorum, Experiment 1
MAM
(Intercept) / 0.61 / 0.63 / 0.971 / 0.36
Genetic diversity / 1.78 / 0.59 / 3.02 / 0.01**
Brood / 0.024 / 0.0055 / 4.3 / 0.0008***
Time to quorum / 0.006 / 0.002 / 3.29 / 0.007**
Non-significant variables removed, in reverse order of removal
Workers / 0.097
Population / 0.30
Day / 0.40
Table S11 Summary of the Minimum Adequate Model testing the effects of genetic diversity, whether or not the new nest site was already discovered (discovery), population or day of experiment, and colony size (number of workers or amount of brood) on total migration time (Experiment 2, negative binomial regression model, link=log). Only colonies which contained one queen and at least 20 workers, were included in the analyses. N = 14 colonies. Data are non-orthogonal (i.e. observational), with a highly significant correlation between the number of workers and brood (Pearson’s correlation: r = 0.73, N = 14, P = 0.003). Therefore, the variables workers and brood were not included in the same model. This Table gives the results for the model including workers. There is a highly significant interaction effect between diversity and discovery, and migration time increased with colony size. Neither the day of the experiment, nor population had an effect on total migration time. If brood was included instead of workers, neither brood nor population or day of experiment had a significant effect on total migration time, but the interaction between diversity and discovery remained significant (X21 = 5.52, P = 0.019). P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. Including “day of experiment” as a factor showed no significant effect of “day” on migration time (last rows of table).
Estimate / Std Error / Z value / Pr(Chi)Total migration time Experiment 2
MAM (D*W)
(Intercept) / 4.24 / 0.31 / 13.77
Diversity (D) / 0.50 / 0.31 / 1.61 / 0.011
Discovery (Disc)yes / 1.54 / 0.50 / 3.09 / 0.008
Workers (W) / 0.005 / 0.002 / 2.26 / 0.042
D*Disc / -2.26 / 0.64 / -3.56 / 0.003**
Non-significant variables removed, in reverse order of removal
(D*Disc + D*Day + Workers)
Day / 0.13
D:Day / 0.14
(D*Disc + D*Population + Workers)
Population / 0.66
D:Population / 0.63
Table S12Summary of the Minimum Adequate Model testing the effects of genetic diversity, population (P) or day of experiment (Day), and colony size (number of workers (W) or amount of brood (B)) on migration time after discovery of the new nest (Experiment 2, negative binomial regression model, link=log). Only colonies that had discovered the new nest site before destruction of the old nest, and which contained one queen and at least 20 workers, were included in the analyses. N = 9 colonies. Data are non-orthogonal (i.e. observational), with a highly significant correlation between the number of workers and brood (Pearson’s correlation: r = 0.82, N = 9, P = 0.007). Therefore, the variables workers and brood were not included in the same model. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Migration time after discovery, Experiment 2
MAM (D*W/B+P/Day)
(Intercept) / 6.07 / 0.57 / 10.62 / 4.3e-07
Diversity (D) / -1.49 / 0.69 / -2.16 / 0.054
Non-significant variables removed, in reverse order of removal
(D*P+W)
Workers (W) / 0.08
Population (P) / 0.45
D:P / 0.83
(D*P+B)
Brood (B) / 0.30
Population (P) / 0.22
D:P / 0.93
(D*Day+W)
W / 0.08
Day / 0.25
D:Day / 0.15
(D*Day+B)
Day / 0.25
B / 0.22
D:Day / 0.27
Table S13Summary of the Minimum Adequate Model testing the effects of genetic diversity, number of workers, time to discover the new nest site, and population or day of experiment (Day) on quorum threshold (Experiment 2, negative binomial regression model, link = log). Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 14 colonies. Data are non-orthogonal (i.e. observational), with a significant correlation between discovery time and workers (spearman rank correlation rho = -0.66, P = 0.01).None of the variables differed significantly between the two populations (t-test workersP = 0.72, diversityP = 0.65, wilcoxon test discovery timeP = 0.28). None of the variables differed significantly between the three days (AnovaworkersF1,12= 0.031, P = 0.86, diversityF1,12 = 1.23, P = 0.29, Kruskal-Wallis testdiscovery timeX2= 0.53,df= 2, P = 0.77). All other correlations between variables: P > 0.14. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Quorum threshold, Experiment 2
MAM
(Intercept) / 1.86 / 0.19 / 9.98 / 3.8e-10***
Discovery time / -0.015 / 0.007 / -2.14 / 0.036*
Non-significant variables removed, in reverse order of removal
Genetic diversity / 0.062
Workers / 0.61
Population / 0.83
Day / 0.28
Workers / 0.75
Table S14Summary of the Minimum Adequate Model testing the effects of quorum threshold, genetic diversity, number of workers, and time to discover the new nest site, on time to achieve quorum after discovery (Experiment 2, negative binomial regression model, link = log). Only main effects were tested in this model. Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 14 colonies. Data are non-orthogonal (i.e. observational), with a significant correlation between the variables quorum threshold and workers (spearman rank correlation rho = 0.53, P = 0.05) and between discovery time and quorum threshold (spearman rank correlation rho = -0.66, P = 0.01). None of the variables differed significantly between the two populations (t-test workersP = 0.72, diversityP = 0.65; wilcoxon test discovery timeP = 0.28, quorum thresholdP = 0.43). None of the variables differed significantly between the days (AnovaworkersF1,12 = 0.03, P = 0.86, diversityF1,12 = 1.2, P = 0.29; Kruskal-Wallis test discovery timeX2 = 0.53, df = 2, P = 0.77, quorum thresholdX2 = 0.02, df = 2, P = 0.99).For all other correlations between variables: P > 0.14. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Time to achieve quorum, Experiment 2
MAM
(Intercept) / 4.33 / 0.066 / 65.75 / 0***
Discovery time / -0.018 / 0.002 / -7.27 / 1.6e-06***
Non-significant variables removed, in reverse order of removal
Quorum threshold / 0.29
Genetic diversity / 0.56
Population / 0.80
Workers / 0.91
Genetic diversity / 0.37
Day / 0.40
Quorum threshold / 0.55
Workers / 0.86
Table S15 Summary of the Minimum Adequate Model testing the effects of genetic diversity (D), nest site discoveredbefore destruction old nest (yes or no, Disc), population, and number of workers on migration time after achieving quorum (Experiment 2, negative binomial regression model, link = log). Only colonies which contained one queen and at least 20 workers were included in the analyses. N = 14 colonies. Data are non-orthogonal (i.e. observational), with a significant correlation between workers and brood (Pearson’s correlation: r = 0.73, df = 12, P = 0.003). Therefore, the variables workers and brood were not included in the same model. When “brood” was included in the model instead of “workers”, none of the variables resulted significant. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05.
Estimate / Std Error / Z value / Pr(Chi)Migration time after quorum, Experiment 2
MAM (D*Disc+W+Day) AIC = 133.01
(Intercept) / 5.30 / 1.06 / 5.00
Genetic diversity / -0.088 / 0.59 / -0.15 / 0.04*
Nest site discovered[yes] / 1.50 / 0.91 / 1.64 / 0.03*
Workers / 0.01 / 0.004 / 3.61 / 0.005**
Day / -0.45 / 0.15 / -2.96 / 0.01*
D*Disc[yes] / -2.73 / 1.18 / -2.32 / 0.048*
MAM (D*Disc+W+P) AIC = 137.52
(Intercept) / 2.52 / 0.69 / 3.67
Genetic diversity (D) / 0.67 / 0.69 / 0.98 / 0.022*
Nest site discovered[yes] / 2.75 / 1.12 / 2.45 / 0.015*
Workers / 0.01 / 0.005 / 3.15 / 0.014*
D*Disc[yes] / -4.38 / 1.43 / -3.06 / 0.007**
Non-significant variables removed, in reverse order of removal
Population / 0.79
Table S16Summary of the Minimum Adequate Model testing the effects of genetic diversity, population, number of workers and number of brood on migration time after achieving quorum (Experiment 2, negative binomial regression model, link = log). Only main effects were tested in this model. Only colonies that had already discovered the dark tube before being forced to migrate, and which contained one queen and at least 20 workers were included in the analyses. N = 9 colonies. Data are non-orthogonal (i.e. observational), with a significant correlation between workers and brood (Pearson’s correlation: r = 0.82, P = 0.007). Therefore, the variables workers and brood were not included in the same model. In the model that contained “brood”, none of the variables resulted significant. However, in the model that contained “workers”, both “genetic diversity” and “workers” were significantly correlated with migration time after quorum. P - values give results of stepwise deletion tests (Likelihood ratio tests). ***P < 0.001; **P < 0.01; *P < 0.05. When day of experiment was included as a factor instead of population, it proved to have a significant effect on migration time after quorum, together with genetic diversity and the number of workers (see bottom rows).