In (a) a schematic shows the hypothesized change in the sex difference in lifespan (upper) if castration has the greatest effects on lifespan in populations where males have a shorter lifespan than females. The bottom schematic shows the hypothetical change that would occur to the absolute variance in lifespan between sexes (i.e. the absolute gap in survival time between males and females, irrespective of whether males or females live longer) with castration if the presence of male gonads contributes to sex differences in lifespan. Each dot represents a hypothetical sex difference in survival for a given species (e.g. intact male mean survival/intact female mean survival), and how this sex difference would change when males are castrated. If males have a shorter lifespan than females in some species because male-specific gonadal hormones cause them to have poorer survival in these instances, then males of these species should show the biggest increase in survival with castration and the sex difference in survival would become closer to zero. Similarly, the absolute sex gap in survival should also be reduced because castration will improve male survival where males live shorter than females but will have less impact where males and females normally have a similar lifespan (i.e. where intact male survival is similar to intact female survival). If male castration increases survival irrespective of the sex difference then the variance in lifespan will not change. In (b), the log response ratio (lnRR, upper) and absolute lnRR (i.e. the absolute gap in survival, lower) is shown for male/female comparisons of survival from the zoo dataset where comparable data was available for surgically sterilized male life expectancy, and non-contracepted male and female life expectancy. The centre dot shows the meta-analytic effect size, with bars showing 95% confidence intervals (thick lines) and 95% prediction intervals (thin lines) for each comparison. This demonstrates that there is substantial variation in the sex difference in lifespan across species, but that there is no change in the overall sex difference in lifespan across species when males are surgically sterilized (upper). Similarly, there is no reduction in the overall sex gap in life expectancy (i.e. the total variance in lifespan between males and females, irrespective of the direction) when males are surgically sterilized. Each dot represents a different population, with the size of each dot corresponding to the precision of the estimate; k is the number of effect sizes within each comparison. In (c), the log response ratio (lnRR, upper) and absolute lnRR (i.e. the absolute gap in survival, lower) is shown for male/female comparisons of survival from previously published data used in the meta-analysis for castrated male survival, male non-castrated survival and female non-contracepted survival for a particular population. The centre dot shows the meta-analytic effect size, with bars showing the 95% confidence intervals (thick lines) and 95% prediction intervals (thin lines) for each comparison. Comparing non-contracepted males and females, there is substantial variation in lifespan between the sexes but no overall sex bias in survival. However, castration of males leads to a sex bias in survival, with castrated males having higher survival on average compared to females. This is because castration increases male survival irrespective of the underlying sex difference in survival between non-contracepted individuals (i.e. survival of males is increased even where there is no sex difference in survival) and there is no overall reduction in the absolute sex gap in survival (lower) when males are castrated. Each dot represents a different population, with the size of each dot corresponding to the precision of the estimate; k is the number of effect sizes within each comparison.
Source data