Figure legends
Fig. 1 Population densities across live-trapping sessions in 2012 (A) and 2015 (C) and the numbers of founder individuals in 2012 (B) and 2015 (D). The population density was estimated as the minimum number known to be alive (MNKA) in each enclosure. Data from the four enclosures in each of the two density groups were expressed as mean ± standard error (SE).
Fig. 2 Mean (± SE) recruitment in 2012 (A) and 2015 (B), mean the proportion of reproductive females in 2012 (C) and 2015 (D), and of reproductive males in 2012 (E) and 2015 (F), in the experimental periods. Data from the four enclosures in each of the two density groups were expressed as mean ± standard error (SE).
Fig. 3 The regression analysis of number of female founders with recruitment (A), the proportion of reproductive females (B) in 2012 and 2015, and the proportion of reproductive males (C) in 2015. Mean number of founders per enclosure was used as the independent variable and the corresponding recruitment and the proportion of reproductive condition per enclosure were the dependent variables. The female-founder number was negatively correlated with mean recruitment (F 1, 6 = 16.51, P = 0.007 for 2012;F 1, 6 = 15.93, P = 0.007 for 2015) and the proportion of reproductive females (F 1, 6 = 90.42, P < 0.0001 for 2012; F 1, 6 = 6.34, P = 0.045 for 2015). The male-founder number was negatively correlated with the proportion of reproductive males in 2015 (F 1, 6 = 24.28, P = 0.003), but not in 2012 (F 1, 6 = 2.93, P = 0.138).
Fig. 4 Recursive models in structural equation model, testing the direct and indirect effects of founder number on and FCM level, recruitment and the proportion of reproductive condition. The structural equation model considered all plausible pathways in which trapping sessions and number and FCM levels influence reproduction. Arrows represent the flow of causality. Solid arrows represent a significant positive pathway; double-headed solid arrows, a significant negative pathway; dotted arrows, a non-significant pathway. Path coefficients associated with each arrow show standardised effect sizes. Numbers indicate the standard path coefficients. CT, cumulative time of trapping session; N, number of founders; P, proportion of reproductive condition; R, recruitment. *P < 0.05, **P < 0.001, ***P< 0.0001. For the proportion of reproductive females,χ 2 = 4.00, P = 0.26, RMSEA = 0.07 in 2012 and χ 2 = 2.00, P = 0.91, RMSEA =0.00 in 2015. For the proportion of males, χ 2= 0.92, P = 0.34, RMSEA = 0.00 in 2012 andχ 2 = 3.34, P = 0.19, RMSEA = 0.07 in 2015. For recruitment, χ 2 = 3.27, P = 0.35, RMSEA =0.04 in 2012 and χ 2 = 2.17,P = 0.34, RMSEA =0.04 in 2015. RMSEA, root mean square error of approximation.