Important considerations for adaptive vs. non-adaptive manipulation
Parasitic manipulation of host behavior is not always adaptive. Both direct and indirect mechanisms can lead to increased parasite transmission (Adamo 2002) and making the distinction between ‘advantageous by-products’ of infection and true adaptive manipulation requires an understanding of which mechanisms are involved (Poulin 1995, 2010; Thomas et al. 2005). In effort to deal with this, Poulin (1995) synthesized four criteria by which one can theoretically distinguish between ‘true’ adaptive traits and advantageous ‘side-effects’. The first criteria regard “complexity”, considering that complex traits are unlikely to arise by chance or accident and require an organizing principle like natural selection. Poulin (1995) acknowledges the difficulty in assessing these criteria since even simple changes in behavior can result from complex mechanisms. Nevertheless, the behavior presented in this paper whereby the host spider is manipulated to perform a task serving the post-mortem needs of the parasite, is indicative of a complex mechanism refined by selection. The second criteria is of “purposiveness of design”, referring to characters or traits seemingly “too well fitted” to fulfil a certain function to have arisen by chance. Such observations of apparent conformity between purpose, design, and function, used to be presented as evidence of divine creation, but are now seen as strong evidence of adaptation through selection (Bekoff and Allen 1995). The locations of host spiders killed by Gibellula spp. certainly appear to serve an obvious purpose by ensuring that the host is in an elevated and protected position for the dispersal of conidia. The third criterion is regarding the “convergence”, or independent origins of an adaptive trait in the phylogenetic lineage of a group of organisms and is seen as a “strong” indicator that the trait is not an accidental side-effect (Allen and Bekoff 1995). Poulin (1995) stipulates that traits do not have to be identical but only analogous, performing the same function (i.e. dispersal). As mentioned above, the closely related genusOphiocordyceps contains a huge assemblage of ant-pathogenic species wherein hosts secure themselves to the substrata via irreversible death-grip behavior immediately prior to death. This behavior, we argue, is analogous to the deathbed behavior of infected spiders. The final and most important of the four criteria according to Poulin (1995) are the “fitness effects”, though he admits that measuring fitness is logistically difficult and not on its own conclusive since even undetectable fitness advantages can be adaptive. As mentioned above, the specificity of these death locations ostensibly serves no other purpose than to maximize fungal spore dispersal and thus, parasite fitness.