Abstract
Many aspects of an individual's biology derive from its interaction with symbiotic microbes, which further define many aspects of the ecology and evolution of the host species. The centrality of microbes in the function of individual organisms has given rise to the concept of the holobiont-that an individual's biology is best understood as a composite of the 'host organism' and symbionts within. This concept has been further elaborated to posit the holobiont as a unit of selection. In this review, I critically examine whether it is useful to consider holobionts as a unit of selection. I argue that microbial heredity-the direct passage of microbes from parent to offspring-is a key factor determining the degree to which the holobiont can usefully be considered a level of selection. Where direct vertical transmission (VT) is common, microbes form part of extended genomes whose dynamics can be modelled with simple population genetics, but that nevertheless have subtle quantitative distinctions from the classic mutation/selection model for nuclear genes. Without direct VT, the correlation between microbial fitness and host individual fitness erodes, and microbe fitness becomes associated with host survival only (rather than reproduction). Furthermore, turnover of microbes within a host may lessen associations between microbial fitness with host survival, and in polymicrobial communities, microbial fitness may derive largely from the ability to outcompete other microbes, to avoid host immune clearance and to minimize mortality through phage infection. These competing selection pressures make holobiont fitness a very minor consideration in determining symbiont evolution. Nevertheless, the importance of non-heritable microbes in organismal function is undoubted-and as such the evolutionary and ecological processes giving rise to variation and evolution of the microbes within and between host individuals represent a key research area in biology.