Introduction
There is increasing attention on the dilution effect of biodiversity on infectious diseases. The dilution effect hypothesis (DEH) suggests that high species diversity in a community including low competent and incompetent hosts can reduce risk of infectious disease (Ostfeld & Keesing 2000). Encounter (contact) reduction, transmission reduction (following encounters), and susceptible host regulation were proposed as underlying mechanisms of the DEH (Keesing et al. 2006; Ostfeld & Keesing 2012), although altered susceptibility due to stressors may also play a role (Buss & Hua 2023). However, the generality of the DEH remains controversial given that increasing host species also introduces more pathogens in complex systems (Randolph & Dobson 2012; Wood & Lafferty 2013; Wood et al. 2014). Meta-analysis suggested that the magnitude of the dilution effect on plant diseases varied between ecosystem type, pathogen type, study design (observational versus manipulative), parasite life history and latitude (Liu et al.2020). In contrast, the magnitude of the dilution effect on zoonotic diseases was independent of host density, study design, and type and specialization of parasites, but related to the relative abundance of the focal host (Civitello et al. 2015). Therefore, in addition to identifying patterns using correlative studies, mechanisms should be experimentally tested in different disease systems to determine when and where the DEH occurs (Halsey 2019).
A meta-analysis of published studies revealed that the dilution effect occurs commonly on plant, wildlife and human diseases (Civitelloet al. 2015). The DEH also has been observed in aquatic host-parasite systems involving anuran, teleost and molluscan hosts infected with fungal, monogenean and trematode parasites (Johnsonet al. 2008; Thieltges et al. 2008; Johnson & Hartson 2009; Johnson et al. 2009; Johnson & Hoverman 2012; Dargentet al. 2013; Venesky et al. 2014; Gendron & Marcogliese 2017). The occurrence of a dilution effect may rely on the specific composition of the host community (Johnson et al. 2013; Salkeldet al. 2013), as variation in host quality is a condition necessary for reduction in pathogen transmission (Keesing & Ostfeld 2021a).
Encounter reduction, considered one of the most common mechanisms behind the dilution effect (Johnson & Thieltges 2010), may occur via a number of means, including mortality of free-living parasite stages, mortality of parasitic stages on hosts, or alterations in host behavior and activity (Johnson & Thieltges 2010). Changes in host behavior may affect the likelihood of parasite encounter and subsequent risk of infection (Johnson & Thieltges 2010). For example, the presence of predators affects anticercarial behavior in tadpoles, contributing to the increase in exposure rates to trematode parasites (Rohr et al. 2015). Sex differences in the degree of shoaling behavior and inter-individual contact rates determines relative infection levels ofGyrodactylus turnbulli in the guppy (Poecilia reticulata ) males and females (Richards et al., 2010). Given that shoaling behavior in guppies affects transmission of gyrodactylids (Richards et al., 2010; Johnson et al., 2011), any changes in schooling behavior will likely affect the risk of infection. How one fish species with schooling behavior responds to the presence of other fish species may then result in the occurrence of a dilution effect. However the effects of host schooling on the occurrence of the dilution effect have never been tested in previous empirical studies and predictive models.
The monogenean Gyrodactylus kobayashii is the most common gyrodactylid on ornamental goldfish (Carassius auratus ) (Liet al. 2014), and the goldfish-G. kobayashii model has been established to study the transmission of the parasite under laboratory conditions (Zhou et al. 2017). While monogeneans tend to show high specificity to a particular host, gyrodactylids are known to infect other host species under experimental conditions, albeit with low infection success (King & Cable 2007; King et al. 2009). The viviparous gyrodactylids mainly transmit via contact between hosts, and have the capability of continuous transmission during their entire lifetime (Boeger et al. 2005).
Using this system, we first determined if host competence was a necessary condition for the dilution effect to occur in the goldfish-G. kobayashii model system. To test the influence of host competence on parasite transmission, we added other fish species of varying competence to the experimental system. This experiment also examined effects of host diversity and abundance on transmission.
Goldfish display schooling behavior, and mean distance among individuals and schooling pattern varies with the population size of goldfish (Leemet al. 2012; Jeon et al. 2013). Importantly, the frequency of social contact affects gyrodactylid transmission in guppies (Johnsonet al. 2011). We thus examined if schooling behavior in goldfish plays a role in parasite transmission and, potentially, the occurrence of the dilution effect in the presence of other host species of varying competence.