Comparison of the afrotherian divergence times (my) and associated 95% credibility intervals obtained in Springer et al. (2003) (based on their full data set - see their supporting data 1 on the PNAS web site) and in our two analyses.

Analysis 1: 12S, 16S, tRNA-Val, NADH2, CO2 (3366 bp)

Analysis 2: 12S, 16S, tRNA-Val, CREM (2049 bp)

Node / Springer et al. (2003) / Analysis 1 / Analyse 2
prior / posterior / prior / posterior
Afrotheria / 79.9 (73.9-85.8) / 91.2 (62.1-121) / 85.52 (75.6-96.8) / 90.4 (61.6-118.3) / 84.2 (77.4-91.6)
Afroinsectiphillia / 76.5 (70.3-82.2) / 73.1 (33.5-110) / 76.62 (64.6-89.6) / 72.3 (33.7-108) / 76.5 (68.8-85.1)
Macroscelidea + Afrosoricida / 74 (68.2-80.1) / 54.9 (16.8- 96.7) / 72.53 (59-86.5) / 54.5 (16.3-94.5) / 74.9 (66.7-83.5)
*Macroscelidea / 18.9 (14.6-23.6) / 27.2 (1-73.6) / 34 (22.1-48) / 27 (1-70.9) / 36.9(29.3-45.5)
Afrosoricida / 66.4 (59.5-72.4) / 36.8 (5.9- 80.3) / 62.5 (47.9-77.7) / 36.4 (5.5-78.5) / 67.9 (58.5-77.7)
**Paenungulata / 62.5 (56.9-64.9) / 59.6 (54.3- 64.7) / 62.1 (55.9-64.9) / 59.6 (54.2-64.7) / 63(58.7-64.9)
Procavia + Dugong / 59 (53.1-62.7) / 29.6 (1.3- 59.1) / 58 (50.3-63.6) / 29.8 (1.5-58.9) / 57.6(51.4-63)
Chrysochloridae / --- / 18.2 (0.5-56.6) / 28.5 (21.5-36.5) / 18.1 (0.5-56.9) / 26.9 (14.7-41.8)

*Afrotherian divergence times obtained during this study are very similar to those obtained by Springer et al. (2003) except for the Macroscelidea. The divergence estimates for the Macroscelidea contained in the Springer et al. (2003) study (18.9 my, 95% Cred. Int. = 14.6-23.6) were based on their full data set (mtDNA and nuclear DNA sequences) and, in fact, are similar to Douady et al. (2003)’s values (21.3 +/- 3.3 my) that are derived from different data (i.e. 12S, 16S, vWF and IRBP). This gives added support to the Springer et al. (2003) dating. However, Springer et al. (2003) also report divergence estimates in their supplementary data which are based only on 12S and 16S sequences (mtDNA sequences). Within Afrotheria, these estimates are all very close to those from the full data set with a single exception, the Macroscelidea. In this case the estimate (37.28 my, 95% Cred. Int. = 25.8-51.3) is twice that based on their full data set (see above). This difference is not discussed in the Springer et al. (2003) paper, but most likely means that the estimate for the macroscelids is biased by using only mitochondrial DNA. This assumption is further strengthened by the fact that the other divergence estimates within Afrotheria based on 12S and 16S are very close to those suggested by the full data set thus underscoring the anomalous situation for Macroscelidea. Our results for the Paenungulata, Procavia + Dugong, Afroinsectiphillia, Afrosoricida and the Afrotheria nodes are all consistent with Springer et al. (2003)’s analysis of their full data set, and give credence to our estimate for golden moles reported herein

**The node Paenungulata was calibrated between 54 and 65 my ago following Springer et al. (2003).

Reference

Douady CJ, Catzeflis F, Raman J, Springer MS, Stanhope MJ 2003 The Sahara as a vicariant agent, and the role of Miocene climatic events, in the diversification of the mammalian order Macroscelidea (elephant shrews). PNAS 100: 8325-8330.