While group formation provides antipredatory defences, increases foraging efficiency and mating opportunities, it can be counterintuitive by promoting disease transmission amongst social hosts. Upon introduction of a pathogen, uninfected individuals often modify their social preferences to reduce infection risk. Infected hosts also exhibit behavioural changes, for example, removing themselves from a group to prevent an epidemic. Conversely, here we show how Trinidadian guppies infected with a directly transmitted ectoparasite, Gyrodactylus turnbulli, significantly increase their contact rates with uninfected conspecifics. As uninfected fish never perform this behaviour, this is suggestive of a parasite-mediated behavioural response of infected hosts, presumably to offload their parasites. In the early stages of infection, however, such behavioural modifications are ineffective in alleviating parasite burdens. Additionally, we show that fish exposed to G. turnbulli infections for a second time, spent less time associating than those exposed to parasites for the first time. We speculate that individuals build and retain an infection cue repertoire, enabling them to rapidly recognize and avoid infectious conspecifics. This study highlights the importance of considering host behavioural modifications when investigating disease transmission dynamics.

Parasites are detrimental to host fitness and therefore should strongly select for host defence mechanisms. Yet, hosts vary considerably in their observed parasite loads. One notable source of inter-individual variation in parasitism is host sex. Such variation could be caused by the immunomodulatory effects of gonadal steroids. Here we assess the influence of gonadal steroids on the ability of guppies (Poecilia reticulata) to defend themselves against a common and deleterious parasite (Gyrodactylus turnbulli). Adult male guppies underwent 31 days of artificial demasculinization with the androgen receptor-antagonist flutamide, or feminization with a combination of flutamide and the synthetic oestrogen 17 β-estradiol, and their parasite loads were compared over time to untreated males and females. Both demasculinized and feminized male guppies had lower G. turnbulli loads than the untreated males and females, but this effect appeared to be mainly the result of demasculinization, with feminization having no additional measurable effect. Furthermore, demasculinized males, feminized males and untreated females all suffered lower Gyrodactylus-induced mortality than untreated males. Together, these results suggest that androgens reduce the ability of guppies to control parasite loads, and modulate resistance to and survival from infection. We discuss the relevance of these findings for understanding constraints on the evolution of resistance in guppies and other vertebrates.

Parasites typically have low reproductive fitness on paratenic hosts. Such hosts offer other significant inclusive fitness benefits to parasites, however, such as increased mobility and migration potential. The parasite fauna of the guppy (Poecilia reticulata) is dominated by the directly transmitted ectoparasites Gyrodactylus bullatarudis and Gyrodactylus turnbulli. In the wild, close predatory and competitive interactions occur between the guppy and the killifish Rivulus hartii. Previous observations suggest that these fish can share gyrodactylids, so we tested experimentally whether these parasites can use R. hartii as an alternative host. In aquaria, G. bullatarudis was the only species able to transmit from prey to predator. Both parasite species transferred equally well to prey when the predator was experimentally infected. However, in semi-natural conditions, G. bullatarudis transmitted more successfully to the prey fish. Importantly, G. bullatarudis also survived significantly longer on R. hartii out of water. As R. hartii can migrate overland between isolated guppy populations, G. bullatarudis may have an enhanced ability to disperse and colonize new host populations, consistent with its wider distribution in the wild. To our knowledge, this is the first empirical study demonstrating a predator acting as a paratenic host for the parasites of its prey.

The enemy release hypothesis proposes that in parasite depleted habitats, populations will experience relaxed selection and become more susceptible (or less tolerant) to pathogenic infections. Here, we focus on a population of guppies (Poecilia reticulata) that are found in an extreme environment (the Pitch Lake, Trinidad) and examine whether this habitat represents a refuge from parasites. We investigated the efficacy of pitch in preventing microbial infections in Pitch Lake guppies, by exposing them to dechlorinated water, and reducing gyrodactylid infections on non-Pitch Lake guppies by transferring them to Pitch Lake water. We show that (i) natural prevalence of ectoparasites in the Pitch Lake is low compared to reference populations, (ii) Pitch Lake guppies transferred into aquarium water develop microbial infections, and (iii) experimentally infected guppies are cured of their gyrodactylid infections both by natural Pitch Lake water and by dechlorinated water containing solid pitch. These results indicate a role for Pitch Lake water in the defence of guppies from their parasites and suggest that Pitch Lake guppies might have undergone enemy release in this extreme environment. The Pitch Lake provides an ideal ecosystem for studies on immune gene evolution in the absence of parasites and long-term evolutionary implications of hydrocarbon pollution for vertebrates.

Mating was observed in populations of Gyrodactylus turnbulli from guppies. Copulation took place between older (post-1st birth) flukes and resulted in the insemination of spermatozoa into the seminal receptacle. New-born flukes, and flukes which had only recently given birth for the first time (in which the penis was absent or developing), were never inseminated. Flukes were promiscuous, copulating with several partners in succession. The occurrence of mating was related to the size, age and crowding of the G. turnbulli population, taking place most frequently when the parasites were most densely crowded upon the peduncle of the fish on days 5 and 6 of the infection. Maximum population size was positively correlated with the number of flukes initiating an infection. As a result, mating began earlier, and continued for longer in populations founded by several parasites rather than those begun by single flukes. These observations are discussed with respect to the population biology of G. turnbulli: it is likely that sexual and asexual reproduction alternate, with sex being important only during phases of epidemic population growth. The significance of sex in gyrodactylids is discussed, and the potential importance of continuous sexual reproduction in G. turnbulli infections on cultivated guppies is noted.