Extended Data Fig. 7: Relationships between temporal activity patterns.

a-c, The phylogenetically corrected pairwise relationships between PC1 loadings, PC2 loadings, and total rest, along with the statistics from the fitting of a linear regression using phylogenetically corrected generalised least squares (pGLS). Dots represent each species and are coloured by their tribe. d-e, Ancestral reconstructions of PC1 and PC2 loadings. Highlighted species codes correspond to highlights in panel g. f, Relationship between PC1 and PC2 loadings for both extant (blue dots) and ancestral nodes (red dots). Nodes are connected by their phylogenetic relationship, and edges are coloured by the age of the ancestral node. g, Subset of the data in panel f highlighting recent branching events with differing PC2 loading values reconstructed at their ancestral nodes. Nodes are connected by their phylogenetic relationship, and edges are coloured by the age of the ancestral node. h, Comparison of discrete model fits (AIC scores and delta AIC) for the evolution of temporal activity patterns. Both the ER and bridge-only-SYM models were significantly favoured over other models, but neither was favoured over the other. i, Transition rates between diurnal, nocturnal, crepuscular, and cathemeral states based on the best fitting bridge-only-SYM model (symmetric rates without direct transitions between diurnal and nocturnal states). The thickness of each line is representative, but not directly scaled to the rates between states.