Many eukaryotes host mutualistic, maternally transmitted prokaryotic symbionts. Two kinds of evolution within symbiont genomes threaten to erode the benefits of these associations. First, because symbionts reproduce asexually, are sequestered within hosts, and undergo bottlenecks at infection, they are subject to the long-term accumulation of deleterious mutations through Muller's ratchet. Second, "selfish" mutations, benefiting symbionts at host expense, could cause the ultimate decline of both host and symbionts. We performed simulations to assess how the fate of each mutation type is affected by host population size, numbers of symbionts transmitted to progeny, selection within and between hosts, and mutation rate. Fixation rate always increases with decreasing host population size. However, fixation rates for uniformly deleterious and selfish mutations are oppositely affected by varying transmission numbers, with increased numbers slowing accumulation of strictly deleterious mutations, especially for effects concentrated within hosts, but speeding fixation of "selfish" mutations. In aphid symbionts, most genes underlie basic cellular processes and are probably selected at both levels, but a substantial minority of genes contribute only to host fitness. No inoculum size is optimal for minimizing deleterious evolution for both categories of gene.

Cellular slime molds (CSMs) possess a remarkable life cycle that encompasses an extreme act of altruism. CSM cells live as individual amoebae until starved, then aggregate and ultimately transform themselves into a multicellular fruiting body. This fruiting body consists of stalk cells (altruists that eventually die) and spores (the beneficiaries of this sacrifice). Altruistic systems such as this are vulnerable to cheaters, which are individuals unrelated to the altruists that obtain the benefits provided by them without reciprocating. Here, we investigate two forces that can maintain CSM altruism despite cheating: kin selection and anticheater adaptations. First, we present new kinship-based models based on CSM developmental biology to evaluate the efficacy of kin selection. These models show that stalk-making genotypes can still be maintained when aggregations are initiated by multiple "founder" spores, provided that spores of stalkless fruiting bodies have low rates of dispersal and dispersal success is a concave function of stalk height. Second, we review proposals that several features of CSM development, such as the chemical suppression of the redifferentiation of prestalk cells into prespores, act as anticheater adaptations.