Under natural and sexual selection traits often evolve that secure paternity or maternity through self-sacrifice to predators, rivals, offspring, or partners. Emasculation—males removing their genitals—is an unusual example of such behaviours. Known only in insects and spiders, the phenomenon's adaptiveness is difficult to explain, yet its repeated origins and association with sexual size dimorphism (SSD) and sexual cannibalism suggest an adaptive significance. In spiders, emasculation of paired male sperm-transferring organs — secondary genitals — (hereafter, palps), results in ‘eunuchs’. This behaviour has been hypothesized to be adaptive because (i) males plug female genitals with their severed palps (plugging hypothesis), (ii) males remove their palps to become better fighters in male–male contests (better-fighter hypothesis), perhaps reaching higher agility due to reduced total body mass (gloves-off hypothesis), and (iii) males achieve prolonged sperm transfer through severed genitals (remote-copulation hypothesis). Prior research has provided evidence in support of these hypotheses in some orb-weaving spiders but these explanations are far from general. Seeking broad macroevolutionary patterns of spider emasculation, we review the known occurrences, weigh the evidence in support of the hypotheses in each known case, and redefine more precisely the particular cases of emasculation depending on its timing in relation to maturation and mating: ‘pre-maturation’, ‘mating’, and ‘post-mating’. We use a genus-level spider phylogeny to explore emasculation evolution and to investigate potential evolutionary linkage between emasculation, SSD, lesser genital damage (embolic breakage), and sexual cannibalism (females consuming their mates). We find a complex pattern of spider emasculation evolution, all cases confined to Araneoidea: emasculation evolved at least five and up to 11 times, was lost at least four times, and became further modified at least once. We also find emasculation, as well as lesser genital damage and sexual cannibalism, to be significantly associated with SSD. These behavioural and morphological traits thus likely co-evolve in spiders. Emasculation can be seen as an extreme form of genital mutilation, or even a terminal investment strategy linked to the evolution of monogyny. However, as different emasculation cases in araneoid spiders are neither homologous nor biologically identical, and may or may not serve as paternity protection, the direct link to monogyny is not clear cut. Understanding better the phylogenetic patterns of emasculation and its constituent morphologies and behaviours, a clearer picture of the intricate interplay of natural and sexual selection may arise. With the here improved evolutionary resolution of spider eunuch behaviour, we can more specifically tie the evidence from adaptive hypotheses to independent cases, and propose promising avenues for further research of spider eunuchs, and of the evolution of monogyny.