Supporting Appendix

File in this Data Supplement:

Supporting Appendix

Ichnological evidence has been ignored because it is not possible to know the identity of the trackmakers. This is precisely why a parataxonomy has been erected for such fossils. The problem is that in many cases, the tracks may have been left by an animal whose skeletal remains are known; thus, if the track were counted separately, actual biodiversity would be over-evaluated. This database includes all species of stegocephalians sensu Laurin (1998), namely, all taxa that are more closely related to extant tetrapods than to Panderichthys rhombolepis.

For various taxa, most of the data comes from the following references, with additional data from more recent papers: various stegocephalians traditionally considered reptiliomorphs (but most of which are probably stem-tetrapods), Carroll et al. (1972 [Carroll, 1972 #2398]), Kuhn (1969a [Kuhn, 1969 #2397]); Embolomeri, Panchen (1970 [Panchen, 1970 #1086]); Seymouriamorpha, Kuhn (1972 [Kuhn, 1972 #2581]), Tatarinov (1972 [Tatarinov, 1972 #2582]), Laurin (2000 [Laurin, 2000 #3379]); Amphibia, Bossy (1976 [Bossy, 1976 #2811]), Carroll et al. (1998 [Carroll, 1998 #4375]); Diadectidae, Olson (1947 [Olson, 1947 #2014]); various amniotes, Kuhn (1969b [Kuhn, 1969 #1731]); Synapsida, Reisz (1986 [Reisz, 1986 #15]).

Below, references are given only when the information does not come from the references mentioned above.

For lysorophians, I have ignored Molgophis, that may not be valid, judging by the material on which this was erected.

Vertebrates from the Permian part of the Argana region (Morocco) have been omitted because their age is too uncertain.

Frasnian: 2

Note: Livoniana multidentata may or may not be a stegocephalian (Ahlberg et al., 2000 [Ahlberg, 2000 #4879]). It has been excluded.

Famennian: 7

Note: Säve-Söderbergh (1932 [Säve-Söderbergh, 1932 #1905]) recognized four species of Ichthyostega and a species of Ichthyostegopsis, but most subsequent workers have ignored these and referred only to Ichthyostega sp. Thus, only one species of these (the one with page priority) has been retained.

Tournaisian: 3

Two unnamed stegocephalians from Horton Bluff (Clack and Carroll, 2000 [Clack, 2000 #5058])

Viséan: 21

Serpukhovian:

Note: Spathicephalus pereger is from the early Namurian (Baird, 1962 [Baird, 1962 #944]); I equate this with Serpukhovian.

Bashkirian:

Note: I assume, after Harland et al. (1990 [Harland, 1990 #10271]: 44) that the Bashkirian includes the Westphalian A and the Langsettian; this includes the localities of Jarrow and Joggins. Arizonerpeton wellsi has been put here, even though its age is equivalent to Langstettian/Duckmantian boundary and thus, it could also have been put into the Moscovian.

Moscovian:

Note: I assume, after Harland et al. (1990 [Harland, 1990 #10271]: 44) that the Moscovian includes the Westphalian B through D, and most of middle and upper Coal measures. Nyrany, Florence, Mazon Creek and Linton fit here. The Freeport coal and the whole Allegheny series is late Westphalian, according to Hook and Baird (1993 [Hook, 1993 #2933]).

I have place Danville, in the McLeansboro Formation (Clepsydrops collettii) here because the geological chart from the ISGS (downloaded from: http://www.isgs.uiuc.edu/coalsec/coal/pennstratigraphy.htm) indicates a Desmoinesian age for the Danville Coal. Harland et al. (1990) indicate that the Desmoinesian is Moscovian.

Kasimovian

Note: This includes the lower Stephanian. Boii crassidens is only dated Stephanian. I have inserted it here because the Kasimovian seems to have very few fossiliferous localities, but this may be a taphonomic artefact. According to Panchen (1970 [Panchen, 1970 #1086]), this includes the Bloomingdale and Reed Mills localities, and M. Conemaugh series.

The Mattoon Formation (Milosaurus mccordi) ranges from the Missiourian to the Virgilian age; according to Jacobson (2002 [Jacobson, 2002 #13209]). Since there are few taxa in the Kasimovian, I have put Milosaurus here, but it could as well be Gzhelian.

Gzhelian

Note: I assume that the Gzhelian includes the second half of the Stephanian (2/3 of Stephanian B and all of Stephanian C), after Harland et al. (1990 [Harland, 1990 #10271]: 44). I have placed Diplocaulus salamandroides here, but this is tentative. It is from the Upper Pennsylvanian. The age of Diploceraspis burkei is also somewhat uncertain, but I have followed Bossy (1976: fig. 96). The Badger Creek (where Desmatodon hesperi was found) and Pitcairn localities (Pitcairn is in the Red Knob formation, originally Pittsburgh Red Shale) should fit here because they are late Pennsylvanian (Berman & Sumida, 1995), rather than in the Lower Permian.

I have put Stereorhachis dominans here because there is little information on its provenance, other that it was found near Autun. Reisz suggests and Upper Carboniferous age, citing Romer & Price (1940). Now, most authors put the Autunian in the Lower Permian, but the base of this seems to be poorly constrained. Some of it could conceivably be Carboniferous.

The Monongohela Group is Gzhelian and Asselian according to Kozur (1984 [Kozur, 1984 #3255]).

The Hamilton quarry fits into the Gzhelian because Schultze et al. (1994: 444 [Schultze, 1994 #2826]) indicate that it fits in the upper part of the Shawnee group, and Kokur (1984: 578) places this within the Gzhelian.

Asselian

Note: I follow Hoffmann et al. (1989 [Hoffmann, 1989 #189]: fig. 1) in considering that the Autunian is equivalent to the Asselian. However, Ross (1970 [Ross, 1970 #1465]: 26) suggested that the Autunian could also include into all of the Sakmarian and Harland et al. (1990: 46) indicate that some authors divided the Permian into three parts, the lower of which was Autunian (even more extensive). Manning (1993 [Menning, 1993 #3853]: 83) seems to equate the Autunian with the Lower Rotliegend, which includes Asselian, Sakmarian and Artinskian (top of page), but below in the same page, he equates the Autunian with lowermost Rotliegend, which is consistent with an Asselian age of the Autunian. Gand et al. (1996: 383, fig. 4) equate the Autunian with Asselian and Sakmarian, while Châteauneuf (1989 [Châteauneuf, 1989 #5130]: 179) equates Autunian with Asselian, Sakmarian and Artinskian. Saxonerpeton geinitzi is only dated Lower Rotliegend, so it could also fit into the Sakmarian.

I have placed ?Euryodus bonneri in the Asselian because the Speiser Shale is dated Wolfcampian (Asselian and Sakmarian), but the fossil comes from the lower part of these shales. Furthermore, Hembree et al. (2004 [Hembree, 2004 #9616]: 128) indicate that the Speiser Shale is in the top of the Council Grove Group, that seems to be the lower half of the Wolfcampian.

The Cutler/Abo and Sangre de Cristo formations are Asselian. Personal communication from D. Berman (25/1/2006): The Sangre de Cristo is a prograding deposit whose origin is mts in what is now southern Colorado. So, depending on what level you are in the deposit or distance from the source, the age is variable from Late Pennsylvanian to Early Permian. In Colorado at Peter's locality, the Howard site, it is Late Pennsylvanian but in northcentral New Mexico it is probably Early Permian.

The Cedar Mesa Sandstone and Organ Rock Shale are considered to fit here (Tseajaia campi comes from there).

I tentatively put much of the Dunkard group (where Ambedus pusillus comes from, among others) into the Asselian, but this is poorly constrained; the only thing clear is that this fits in the Wolfcampian. Wellstead (1991: 7 [Wellstead, 1991 #2785]) shows a good geological chart with stratigraphic correlations between various localities, formations, and the standard geological time scale. The Washington formation of the Dunkard group is Asselian, and the Greene formation (from which Lysorophus dunkardensis originates) of the Dunkard group is Sakmarian.

The Pueblo formation of the Cisco group fits at the base of the Wolfcampian (Asselian), but the Cisco group starts in the Pennsylvanian and may extend well into the Sakmarian (Hentz, 1988: 9).

I consider that the Moran formation is Asselian.

I consider the Upper Gladlauter beds of the Saale Basin Asselian because they are well under the Bacov horizon (Werneburg, 1989 [Werneburg, 1989 #2332]: 128).

Note: The Putnam formation (now Archer City Formation) is near the Asselian/Sakamrian boundary; here, I have considered it Sakmarian. The Admiral formation is Sakmarian. The Bacov horizon, zone 6 is here asssigned tentatively to Sakmarian. All Klembara says is that it is from the basal Saxonian sensu Werneburg 1989, and the latter puts that just above the Autunian, coeval with Bromacker or just slightly younger. Berman et al. (2000) indicate that the Tambach formation in which Bromacker is located is the basalmost unit of the Upper Rotliegend. Benek et al. (1996) indicate that the boundary between lower and upper Rotliegend could be anywhere between early Sakmarian and mid-Kazanian. The latter is unlikely. I consider that "Middle Rotliegendes" (Kadaliosaurus priscus) fits in the Sakmarian, as the Figure 3.7 of Harland et al. (1990) suggests. Sumida et al. (1996 [Sumida, 1996 #4347]) indicate that the Tambach formation is probably of Wolfcampian age, which would be Asselian to Sakmarian. Werneburgh (1989 [Werneburg, 1989 #2333]) indicates that the Tambach Formation straddles the Autunian/Saxonian boundary, and Kozur (1984) indicates that the Tambach Formation is mostly Sakmarian (it may extend into the Artinskian, but this is not too clear from the figure). I conclude that the Tambach Formation is Sakmarian. This is confirmed by Werneburg & Schneider (1996 [Werneburg, 1996 #3521]: table 1). This suggests that Bacov horizon, zone 6 (where Discosauriscus comes from) is Sakmarian. Discosauriscus pulcherrimus comes from zone 5 (in Werneburg's 1989 [Werneburg, 1989 #2333] definition), which I guess would be either lower Sakmarian or Asselian. I think that lower Sakmarian is slightly more likely because zones 3, 4 and a small part of zone 2 are all in the Lower Permian. Diadectes absitus and Orobates pabsti are also from the Tambach formation.

I assume that the "middle Rotliegend" (Kadaliosaurus priscus, Edaphosaurus? credneri) is Sakmarian.

I consider that the Petrolia formation is Sakmarian, as most of my colleagues have done.

I tentatively put the Fort Riley Limestone of the Chase Group in the Sakmarian (Ophiacodon hilli is from there), but all that is sure is that it is Wolfcampian, and it is probably fairly high in the Wolfcampian.

For the age of the Niderhäslich, I follow Werneburg and Schneider (2001 [Werneburg, 2001 #9644]), who indicate that it is high in the Unter-Rotliegend; this suggests a Sakmarian age, according to the chart of Harland et al. (1990: 48).

The Nahe-Gruppe (Nahe Group) is Sakmarian ( Dietze, 2000 [Dietze, 2000 #13336]).

Artinskian

Note: I assume that the Choza, Clyde, Arroyo and Vale formations and Fort Sill fit in the Artinskian. Pantylus cordatus is known from several formations (Moran, Putnam, Belle Plains, and Clyde, all in the Wichita Group) that range from the Asselian to the Artinskian in age. I have scored it as present in all these stages.

The Hennessey and Belle Plains formations fit into the Artinskian.

I put the Clear Fork group into the Artinskian.

Similarly, Thrausmosaurus serratidens was excluded because it is very fragmentary and not very diagnostic.

Kungurian

Note: The low biodiversity in the Kungurian is likely to be a stratigraphic artifact because the exact correlations of many "leonardian" strata, here systematically attributed to the Artinskina, may often have fit into the Kungurian. I suggest that Artinskian and Kungirian be lumped together for analyses.

Ufimian

Note: I assume that the Chickasha (equivalent to middle of Flowerpot formation) and San Angelo formations are Ufimian, as Reisz and Laurin (2002 [Reisz, 2002 #6423]) argued. This is the last stage that I will cover in this database.

I have not included some highly dubious forms described by Olson (1962 [Olson, 1962 #9]) whose affinities are probably wrong and that may not even all be valid taxa, such as Driveria ponderosa or Gorgodon minutus. See Sidor & Hopson (1995 [Sidor, 1995 #3150]).