We investigated whether parasites or hosts benefit from reduced reproduction in infected hosts. When parasites castrate their hosts, the regain of host reproduction is necessary for castration to be a host adaptation. When infecting Daphnia magna with Pasteuria ramosa, in a lake water based medium, 49 2% of the castrated females regained reproduction. We investigated the relationship between castration level, and parasite and host fitness proxies to determine the adaptive value of host castration. Hosts which regained reproduction contained less spores and had a higher lifetime reproduction than permanently castrated hosts. We also found a negative correlation between parasite and host lifetime reproduction. For hosts which regained reproduction we found no optimal level of castration associated with lifetime reproduction. These results support the view that host castration only is adaptive to the parasite in this system. In addition, we suggest that permanent castration might not be the norm under natural conditions in this system. Finally, we argue that a reduction in host reproduction is more likely to evolve as a property favouring parasites rather than hosts. To our knowledge this is the only experimental study to investigate the adaptive value of reduced host reproduction when castrated hosts can regain reproduction.