Conclusions and other considerations towards a potential model
-The deathbed behavior of spiders infected with Gibellula spp.likely represents an extended phenotype of parasite gene expression. Parasite induced changes in this host-pathogen system fulfil all four criteria set forth by Poulin (1995) to assess if the observed manipulation is ‘truly’ adaptive for the parasite. However, web building prior to death is not ubiquitous among spiders infected by Torrubielloid anamorphs and other species in this genus exhibit similar phenotypes but with characterizable differences in host choice and death location. For example, another Torrubielloid anamorph (Gibellula leiopus ) from the same area infecting the spider Trachelas tranquillus , invariably cause hosts to die on elevated tree branches and never on leaves. In this species, a fungal mat appears to form the only attachment of the host spider to the tree surface. Nonetheless, this behavior of dying in elevated (1-3 meters) locations on the underside of tree branches also provides fitness benefits for the parasite via efficient dispersal of infectious propagules (Araujo and Hughes 2016). The differences in host death locations between these two related pathogens likely reflects the differences in the specific ecologies of the host species, and like O. unilateralis infected ants, may be indicative of adaptive manipulation by the parasite.
Parasitized hosts dying on the underside of leaves is a common feature in the life cycle of many entomopathogenic fungi and represents a clear fitness advantage for the parasite (Araújo and Hughes 2016). The locations where infected hosts die can be phylogenetically informative for some highly specialized species and relates directly to the specific ecological niche exploited by the host (Evans et al. 2018, Araújo et al. 2018). The Torrubielloid anamorphs such as previously mentioned species of Gibellula appear to follow the same trend, though no studies have yet explored specialized fungal pathogens of spiders with similarly comprehensive detail. Despite compelling empirical evidence, the current literature is bereft of any experimental support for adaptive manipulation in this particular parasite-host system. These specialized fungal parasites of spiders may therefore provide a valuable experimental model for further exploring the dynamics of parasitic manipulation since rearing of large numbers of hosts in the laboratory is fairly simple and low cost. Such models will endeavor to better define the elusive distinction between adaptive and non-adaptive manipulation, as it will likely provide valuable insight into the evolution of specialized parasite-host systems.