FigS1 v2

Figure S1. Schematic of the ‘short fuse’ Cambrian explosion simulation analyses. Using all branch lengths < 520 Ma identical to our Observed Tree (panel A) for a subset of taxa necessary to include all of our calibration points, branch lengths of basal lineages > 520 Ma were modified to conform to a ‘short fuse’ Cambrian explosion event at 540 to 530 Ma (panel B). Branch lengths of this resulting tree were then modified at different levels of autocorrelated rate variation (panel C). For no rate variation among branches, the coefficient of variation (CoV) among branch rates is 0 (and there is only 1 tree). For each of the higher levels of rate variation, 10 different realizations of autocorrelated branch length manipulation were generated (with some examples shown in panel C). Each of these trees was then used as a template for generating simulated DNA datasets and then analyzed using BEAST and the same parameters as our observed data. Histograms of the distribution of posterior estimates for the age of the Arthropoda are show, for the three different levels of CoV shown in panel C, with posterior values pooled among replicates.