Paleoecologists have speculated that post-glacial migration of tree species could have been facilitated by mycorrhizal symbionts surviving glaciation as resistant propagules belowground. The general premise of this idea, which we call the ‘paleosymbiosis hypothesis’, is that host plants can access and be colonised by fungal root symbionts that have been inactive for millennia. Here, we explore the plausibility of this hypothesis by synthesising relevant findings from a diverse literature. For example, the paleoecology literature provided evidence of modern roots penetrating paleosols containing ancient (>6000 years) fungal propagules, though these were of unknown condition. With respect to propagule longevity, the available evidence is of mixed quality, but includes convincing examples consistent with the paleosymbiosis hypothesis (i.e. >1000 years viable propagules). We describe symbiont traits and environmental conditions that should favour the development and preservation of an ancient propagule bank, and discuss the implications for our understanding of soil symbiont diversity and ecosystem functioning. We conclude that the paleosymbiosis hypothesis is plausible in locations where propagule deposition and preservation conditions are favourable (e.g. permafrost regions). We encourage future belowground research to consider and explore the potential temporal origins of root symbioses.