Incorporation of fossils into biogeographic studies can have a profound effect on the conclusions that result, particularly when fossil ranges are nonoverlapping with extant ranges. This is the case in archaeid spiders, where there are known fossils from the Northern Hemisphere, yet all living members are restricted to the Southern Hemisphere. To better understand the biogeographic patterns of archaeid spiders and their palpimanoid relatives, we estimate a dated phylogeny using a relaxed clock on a combined molecular and morphological data set. Dating information is compared with treating the archaeid fossil taxa as both node calibrations and as noncontemporaneous terminal tips, both with and without additional calibration points. Estimation of ancestral biogeographic ranges is then performed, using likelihood and Bayesian methods to take into account uncertainty in phylogeny and in dating. We find that treating the fossils as terminal tips within a Bayesian framework, as opposed to dating the phylogeny based only on molecular data with the dates coming from node calibrations, removes the subjectivity involved in assigning priors, which has not been possible with previous methods. Our analyses suggest that the diversification of the northern and southern archaeid lineages was congruent with the breakup of Pangaea into Laurasia and Gondwanaland. This analysis provides a rare example, and perhaps the most strongly supported, where a dated phylogeny confirms a biogeographical hypothesis based on vicariance due to the breakup of the ancient continental plates.
Supp. Fig. 7 analysis ii
Analysis (ii). Plots are created from a 20 million generation BEAST analysis performed both with data and with an empty alignment. Plots compare the posterior distribution from an empty alighment (red line) with the posterior distribution when data was included (green line), and for calibrated nodes, with the theoretical prior (blue line). A-D, plots non-calibrated nodes: compares the posterior distribution from an empty alignment with the posterior distribution with data, and shows that the data is influencing the results. E-K, plots calibrated nodes: compares the posterior distribution from an empty alignment, the posterior distribution with data, and the theoretical prior placed on the node. For plots E-J, the expected prior (blue line) matches the observed prior (red line). For plot K, the expected prior does not match the observed prior: this is likely because the youngest soft bound on the root node calibration, which was left intentionally broad in order to not bias the results and to contain the true age of Araneomorphae divergence, overlaps with the oldest soft bound on other node calibrations, which pushes the observed root prior back in time. However, we do not feel this is a problem because the observed prior is still broad, still contains the true age of Araneomorphae divergence, and is still reasonable.
Supp. Fig. 8 analysis v
Analysis (v). Plots are created from a 20 million generation BEAST analysis performed both with data and with an empty alignment. Plots compare the posterior distribution from an empty alighment (red line) with the posterior distribution when data was included (green line), and for calibrated nodes, with the theoretical prior (blue line). A-E, plots non-calibrated nodes: compares the posterior distribution from an empty alignment with the posterior distribution with data, and shows that the data is influencing the results. F-K, plots calibrated nodes: compares the posterior distribution from an empty alignment, the posterior distribution with data, and the theoretical prior placed on the node. For plots F-J, the expected prior (blue line) matches the observed prior (red line). For plot K, the expected prior does not match the observed prior: this is likely because the youngest soft bound on the root node calibration, which was left intentionally broad in order to not bias the results and to contain the true age of Araneomorphae divergence, overlaps with the oldest soft bound on other node calibrations, which pushes the observed root prior back in time. However, we do not feel this is a problem because the observed prior is still broad, still contains the true age of Araneomorphae divergence, and is still reasonable.
analysis_i_CORRECTED_link_tip_dates
BEAST file for analysis (i)
analysis_ii_
BEAST file for analysis (ii)
analysis_iii_
BEAST file for analysis (iii)
analysis_iv_
BEAST file for analysis (iv)
analysis_v_CORRECTED_link_tip_dates
BEAST file for analysis (v)
LAGRANGE 7 area
File for LAGRANGE 7 area range reconstruction
LAGRANGE 2 area
File for LAGRANGE 2 area range reconstruction
LAGRANGE_tree
Pruned tree used for input in LAGRANGE 7 and 2 area range reconstruction
Palpimanoidea Morphology + DNA_Bayes
Nexus file containing molecular and morphological characters used for MrBayes total evidence phylogenetic analysis, from Wood et al. 2012