Venomous snakes and flood-prone rivers predict serpent and flood myths across 958 societies

Whether the physical environment shapes the content of a culture’s myths has been argued mostly through single cases. Using a catalog of folklore motifs for 958 societies, we test at global scale whether exposure to six physical hazards predicts the prevalence of the matching myth, controlling for documentation, spatial diffusion, and shared ancestry. Two relationships stand out. Societies with more venomous snakes nearby tell proportionally more serpent myths (standardized b = 0.158, p = 0.012), and societies near rivers tell proportionally more flood myths (b = 0.108, p = 0.004). The snake result is the more robust: it replicates on expert range maps that carry no sampling bias, though whether it is specific to venom or to the tropical gradient venomous snakes track cannot be resolved, because the two are too collinear to separate. The flood result is weaker and measure-dependent, holding for river proximity but not for a flood-event catalog or a modeled hazard layer, so it tracks the river’s presence more than its hazard. Volcano, earthquake, lightning, and large-carnivore myths show raw associations that do not survive spatial controls, and two ecology-to-psychology relationships, including the parasite-stress account of distrust, fail to appear in myth. Under correction for six tests both sit just above a strict threshold (Benjamini-Hochberg q = 0.07 and 0.09), so we report them as leads rather than confirmed effects.