Supporting Information
Total Page:16
File Type:pdf, Size:1020Kb
Supporting Information Alfaro et al. 10.1073/pnas.0811087106 SI Text Age of the Root Between Osteichthyes and Outgroups. The sister Appendix 1: Comparison of Divergence Times from Earlier Studies. group of the Osteichthyes (bony fishes) is currently unknown. It Our inferred dates for the splits among the major gnathostome may be represented by the Chondrichthyes (cartilaginous fishes) lineages (Table S5) are largely consistent with those from or by the Acanthodii. The fossil record of the Chondrichthyes recently published studies (Table S6) (1–9). Among the sarcop- dates to at least the early-middle Caradoc of the Harding terygians, the mean age of the split between amniotes and Sandstone, Colorado (Ordovician, 456–454 Myr) (13). The amphibians is 350 million years (Myr) compared with 354 Myr acanthodian Onchus clintonii dates to the Sheinwoodian and inferred by Hugall (4). Within the amphibians, we date the split Homerian (Silurian, 436–428 Myr) (14). Because of the uncer- between the caecilians and the batrachia ϩ urodela clade at 320 tainty in the sister group relationships among Osteichthyes, Myr and between frogs and salamanders at 233 Myr compared Acanthodii, and Chondrichthyes, we assigned to the root a with Hugall, et al. (4), who estimated means of 292 to 322 Myr minimum age of 428 Myr. Placing an upper limit on the root is and 266 to 274 Myr. problematic because of uncertain relationships of many worm- We date the split between mammals and the remaining like Cambrian fossils that have been tentatively referred to as amniotes (Fig. S2, node 31) at 320 Myr, which falls within the chordates. Pikaia, from the Burgess Shale fossil deposits of the 95% credible interval obtained by Hugall et al. (4). Similarly, our late Cambrian of British Columbia, Canada (505 Myr), is a estimated age for the splits between monotremes and therians prevertebrate that has been identified as probably being related and between marsupials and eutherians (201 and 164 Myr, to cephalochordates (15). Our prior uses the age of these respectively) is similar to ages estimated by Hugall et al. (207 or deposits to set the 95% confidence interval. 227 Myr and 182 or 195 Myr, depending on the choice of Elasmobranchii (Lamnidae vs. Scyliorhinidae) (Fig. S2, node 1). Al- constraints). Our mean estimates are older than dates inferred though Elasmobranchii (sharks) have a fossil record that spans by Bininda-Emonds (2), but our 95% credible interval contains back to the Ordovician, crown sharks are much more recent, the ages they estimated (166 Myr for eutherians/marsupials and having appeared in the fossil record during the Mesozoic. The 148 Myr for montremes/therians). earliest relevant fossil is an unnamed and undescribed taxon Our mean age for the split between archosaurs ϩ turtles and from deposits of the Bathonian (Jurassic, 167.7–164.7 Myr), lepidosaurs (270 Myr) is slightly younger than the 285–289 Myr which was assigned to the family Scyliorhinidae (16). Our prior inferred by Hugall et al. but in closer agreement with the fossil assumed a minimum age of 165 Myr and placed 95% of the record. Within turtles, we estimate a mean age of the split weight on divergence times within the time span marked by the between cryptodires and pleurodires, node 34, at 215 Myr, which first Elasmobranchii (415 Myr) (16). is very close to the dates inferred by 2 recent studies (4, 7). Within the archosaurs, the mean age of the split between crocodilians Most Recent Common Ancestor (MRCA) of Osteichthyes (Sarcopterygii and birds, node 33, is dated at 249 Myr, slightly older than the vs. Actinopterygii) (Fig. S2, node 2). The oldest known fossil from 245 Myr of the fossil calibration by Muller and Reisz (10). The this clade is the stem actinopteygian Andreolepis hedei from the split between crocodiles and alligators is 87 Myr, which is Lludlowian of Gotland, Sweden (Silurian, 422–418 Myr) (17). significantly older than the 42 Myr inferred by Hugall et al. but Our prior assumed a minimum age of 418 and used the age of in much better agreement with both the crocodilian fossil record the Burgess Shale deposits (505 Myr) for the 95% confidence (10) and other molecular timescale studies based on mitog- interval. enomes (11). Within birds, the mean age of the split between ratites and the neognathans is dated at 118 Myr, in very good MRCA of Sarcopterygii (Dipnotetrapodomorpha vs. Coelacanthimor- agreement with Hugall et al. and Slack et al. (12), but younger pha) (Fig. S2, node 3). The oldest sarcopterygian is Psarolepis than the 139 Myr inferred by Pereira and Baker (8). romeri from the Lludlow-Pridoli of China (Silurian, 420–418 Among the Actinopterygii, the crown ray-finned fishes (Fig. Myr), but this is likely a stem taxon. The oldest taxon that marks S2, node 4) have a mean age of 298 Myr. This is Ϸ100 Myr older the split between coelacanths, lungfishes, and tetrapods is Eo- than the oldest fossil but also Ϸ100 Myr younger than recent actinistia foreyi, the oldest coelacanthimorph from the Pragian mitogenomic studies (1, 6). Similarly, our estimated ages for (Devonian, 409–407 Myr) (18). The prior thus assumed 407 Myr splits within the teleost are younger than Azuma et al. (1) but in as the minimum age of the split and a 95% upper bound of 505 close agreement with dates provided by time-calibrated nuclear Myr (reflecting the upper age of the Burgess shale deposit). gene divergences (5). This discrepancy might be due to an overall higher rate of evolution in mitochondrial genomes as discussed MRCA of Actinopterygii (Actinistia vs. Acipenseriformes vs. Neote- by Hurley et al. (5). Our estimates for actinopterygian divergence leostei) (Fig. S2, node 4). The oldest fossil known to belong to this times incorporate the largest number of fish fossil calibrations in clade is the stem actinopteygian Andreolepis hedei from the a study to date and also cover a very broad phylogenetic scope. Lludlowian of Gotland, Sweden (Silurian, 422–418 Myr) (17). Thus our estimated ages should be more reliable than those with The oldest taxon that can be assigned to the crown is the fewer fossils or those based on a smaller number of represen- neopterygian Brachydegma caelatum, from the Artkinsian of tative fish lineages. Texas (early Permian, 284 Myr) (5). Our prior thus assumed 284 Myr as the minimum age and 418 Myr for the upper bound. Appendix 2: Description of Timetree Calibrations. The priors for the Tetrapoda were obtained from Hugall et al. (4), and readers are MRCA of Neopterygii (Amiiformes vs. Lepisosteiformes vs. Teleostei) referred to that paper or the literature therein cited, for a more (Fig. S2, node 5). The oldest crown taxon is the stem teleost detailed discussion. Here we list only minimum age and upper Pholidophoretes salvus (19). The oldest stem is Brachydegma boundaries. caelatum, from the Artkinsian of Texas (early Permian, 284 Myr) Alfaro et al. www.pnas.org/cgi/content/short/0811087106 1of20 (5). Our prior thus assumed 225 Myr as the minimum age and incertae sedis materials from the Adamantina Formation of 284 Myr for the upper bound. Brazil, Turonian-Santonian (Late Cretaceous, 93.6–83.5 MyrR) (28). Our prior thus assumed 73 Myr as the minimum age, and MRCA of Teleostei (Elopomorpha vs. Osteoglossomorpha vs. Clupeo- 83.5 MyrR as the upper boundary. cephala) (Fig. S2, node 6). The oldest taxon assigned to this clade is Anaethalion, from the late Kimmeridgian lithographic lime- MRCA of Bagridae (Fig. S2, node 14). The oldest bagrid fossils, stone of Nusblingen, Germany, and Cerin, France (Jurassic, 152 Eomacrones wilsoni, Nigerium gadense, Nigerium wurnoe¨nse, have Myr) (20). The oldest stem teleost is Pholidophoretes salvus all been assigned to deposits in Niger of Paleocene age (Ϸ59 (Pholidophoridae), from the early Carnian (Julian) of Polzberg MyrR) (27). Our prior assumed 59 Myr as the minimum age, and bei Lunz, Austria (Triassic, 228–225 Myr) (19). Our prior 73 Myr, age assigned to the crown siluriforms, as the upper assumed 152 Myr as the minimum age and 225 Myr for the upper boundary (27). bound. MRCA of Callichthyidae (Fig. S2, node 15). The oldest callichthyid, MRCA of Osteoglossomorpha (Fig. S2, node 7). The oldest crown Corydoras revelatus, dates from the Thanetian of Argentina osteoglossomorphs are several taxa from the Lycoptera assem- (Paleocene, Ϸ55 Myr) (27). Our prior assumed 55 Myr as the blage of the Barremian of China (Early Cretaceous, 130 Myr), minimum age, and 73 Myr, the age assigned to the crown such as the Hiodontidae Yanbiania wangqingica (21). To estab- siluriforms, as the upper boundary (27). lish an upper boundary, we used the elopomorph Anaethalion, from the late Kimmeridgian lithographic limestone (Jurassic, MRCA of Argentiniformes (Fig. S2, node 16). The argentiniform 152 Myr) (22). Our prior assumed 130 Myr as the minimum age Nybelinoides brevis from the Barremian-Aptian of Bernissart, and 152 Myr as the upper bound. Belgium (Early Cretaceous, 127–124 Myr) marks the earliest appearance of the crown argentiniforms (29). Leptolepides sprat- MRCA of Elopomorpha (Fig. S2, node 8). The oldest crown elopo- tiformis (Orthogonikleithridae) from the late Kimmeridgian morph is the Albulidae Albuloideorum ventralis, from the Early lithographic limestone of Cerin, France (Jurassic, 152 Myr) is the Hauterivian (Jurassic/Cretaceous border, 135 Myr) (23). The oldest euteleost (20). Our prior assumes 124 Myr as the mini- oldest stem elopomorph is Anaethalion, from the late Kimmer- mum age, and 152 Myr as the upper boundary. idgian lithographic limestone of Nusblingen, Germany, and Cerin, France (Jurassic, 152 Myr) (22). Our prior assumed 135 MRCA of Galaxiidae (Fig. S2, node 17). Stomporia rogersmithi from Myr as the minimum age and 152 Myr for the upper bound.