“…With the provisos discussed above, our data indicate that three complementary lines of evidence (mean differences, classification accuracy and morphological disparity) suggest that the size of horse premolars has changed to a variable degree over time and space in both archaic and modern horses (Figures 1c (box plots) and 2 (size phenogram) and Supplementary Information, available online), but changes to shape have been modest, with variation mostly overlapping among the archaic samples (Figures 3 and 4) until the development of modern breeds in recent centuries, when shape has become hugely distinctive (Figures 3 and 4). With an estimated time since the divergence of PRZ and the lineage leading to domestic horses of >100,000 years (Goto et al, 2011; Steiner et al, 2013; but see Der Sarkissian et al, 2015, for an alternative perspective), the data suggest that conservativeness may have characterized premolar shape for almost 99% of the history of Equus caballus , whereas in the last 1000–2000 years, under conditions of strong selective breeding, shape differences have more than doubled compared with those observed in archaic horses (Figure 3). Thus, we propose, for premolar shape, a hypothesis of a ‘long-fuse’ model of phenotypic change in domestication, whereby a long initial period of small variation was followed by an explosive acceleration in the magnitude of shape change.…”