RHIZODONTIDAE Traquair, 1881

Johanson, Zerina, Turner, Susan & Warren, Anne, 2000, First East Gondwanan record of Strepsodus (Sarcopterygii, Rhizodontida) from the Lower Carboniferous Ducabrook Formation, central Queensland, Australia, Geodiversitas 22 (2), pp. 161-169 : 161-169

publication ID	https://doi.org/ 10.5281/zenodo.4664614
persistent identifier	https://treatment.plazi.org/id/6B7F6971-FFCF-3702-A2BE-A666CB4E0AB7
treatment provided by	Felipe
scientific name	RHIZODONTIDAE Traquair, 1881
status

Treatment

Family RHIZODONTIDAE Traquair, 1881

Shoulder girdle material has been described or figured for a large number of Late Devonian and Carboniferous Rhizodontida including Strepsodus sauroides (Binney, 1841) , Screbodinus ornatus (Andrews & Westoll 1970; Andrews 1985), Rhizodus hibberti Owen, 1840 (Andrews & Westoll 1970),? Strepsodus anculonamensis Andrews, 1985 , Strepsodus sp. (Janvier et al. 1984) (these taxa are assigned to the Rhizodontidae , a derived family within the Rhizodontida ), Sauripteris taylori Hall, 1843 (Andrews & Westoll 1970), Barameda decipiens (Long 1989) , Gooloogongia loomesi (Johanson & Ahlberg 1998) and “ Notorhizodon ” macelveyi Young et al., 1992 (this taxon may be restricted to assigned shoulder girdle material [Johanson & Ahlberg, in press]). Undescribed rhizodontid shoulder girdle material (cleithra) from Horton Bluff (Tournaisian, Nova Scotia, Canada) has also been examined (by Johanson & Warren) and will be included in comparisons below.

One variable feature among these taxa is the degree of development of the unornamented and depressed flange along the posterior margin of the cleithrum. Of these taxa, Screbodinus and Barameda are said to have relatively smaller flanges ( Fig. 2E, H View FIG ). This is also the morphology seen in the? Strepsodus anculonamensis and “larger Foulden form”, associated with (and closely related to)? S. anculonamensis ( Fig. 2A, B View FIG ) (Andrews 1985: 85, figs 11b-e; 13a-c). The widest portion of these smaller flanges (visible in lateral view) appears to mirror the dimensions of the posterior scapulocoracoid attachment on the internal surfa- ce of the cleithrum (see below); dorsal to this, the flange narrows sharply. This widest point occurs at the narrowest part of the main body of the cleithrum (the pectoral incision [Andrews 1985]). It should be noted that the posterior flange figured for Barameda (Long 1989: fig. 9B) may be nearer to the scapulocoracoid and the pectoral incision than indicated (based on NMV P181699). This flange in Barameda (Long 1989: fig. 7E) is probably smaller than that of? Strepsodus anculonamensis and the “larger Foulden form” (Andrews 1985: fig. 11). The posterior flange on the Horton Bluff cleithra is most similar to? Strepsodus anculonamensis and the “larger Foulden form”. That of Strepsodus sp. from the latest Devonian of Turkey (Janvier et al. 1984: pl. 2.12) also appears to be smaller, especially when compared with Strepsodus sauroides (see below). These species can be separated from Screbodinus ornatus, which has a small posterior flange but lacks a pectoral incision (Andrews 1985) ( Fig. 2E View FIG ). The smaller flange in Barameda may be associated with the absence of a raised or distinct posterior scapulocoracoid attachment (Long 1989: fig. 7C; see below).

Gooloogongia loomesi lacks a posterior flange entirely ( Fig. 2G View FIG ) (Johanson & Ahlberg 1998), while “ Notorhizodon ” macelveyi appears to have a thin posterior unornamented area that is not depressed relative to the dorsal lamina of the cleithrum ( Fig. 2F View FIG ) (Young et al. 1992: fig. 40A). This may represent an incipient posterior flange. The posterior flange of Sauripteris taylori is not readily evaluated, as the specimen described is visible only in internal view (e.g., Andrews & Westoll 1970: pl. 8C). A new specimen of Sauripteris figured by Daeschler & Shubin (1997) may have more information in this regard following preparation.

In contrast, both Strepsodus sauroides and Rhizodus have more strongly developed posterior flanges of the cleithrum ( Fig. 2C, D View FIG ). That of Rhizodus does not narrow, but extends around the dorsal margin of the cleithrum (Andrews & Westoll 1970: pl. 6). This is the best developed posterior flange in the Rhizodontida . In Strepsodus sauroides (Andrews & Westoll 1970: pl. 11; Andrews 1972: pl. 1C), the flange also extends dorsally but does not reach the dorsal margin of the cleithrum. This is the widest flange among species of Strepsodus (compare the “larger Foulden form” with S. sauroides [Andrews 1985: fig. 13a-c, h]).

Of the taxa just described, the better preserved Middle Paddock rhizodontid cleithra (QMF34611, QMF36762 and QMF34602) show closest similarity to Strepsodus , with a welldeveloped pectoral incision and posterior flange. Of these, QMF36762 ( Fig. 3 View FIG A-C) is the best preserved. The width of this flange decreases more rapidly above the scapulocoracoid attachment than does that of S. sauroides , and may be more similar to? Strepsodus anculonamensis , the “larger Foulden form” and the Horton Bluff rhizodontid, where the main width of the flange occurs at the pectoral incision in external view ( Fig. 3A View FIG ). On the other hand, QMF34611 ( Fig. 3D View FIG ) and QMF34602, although more damaged, appear to show a better developed flange, and differences in the morphology of the dorsal cleithrum and scapulocoracoid attachment (see below) that may indicate the presence of more than one species at Middle Paddock. Comparisons between these two cleithra and S. sauroides are difficult due to damage in the former; still, the flanges of these appear better developed than QMF36762 and other species of Strepsodus .

Other differences exist among known rhizodont cleithra. For example, the dorsal cleithrum is more triangular in the best preserved specimen of Strepsodus sauroides (Andrews & Westoll 1970: pl. 11A), rounder in? S. anculonamensis and the “larger Foulden form” (Andrews 1985: fig. 11d) but squarer in the Middle Paddock ( Fig. 3 View FIG ) and Horton Bluff specimens.

The morphology of the scapulocoracoid attachment also differs among these, although it is always located on the dorsal lamina of the cleithrum at the pectoral incision (Andrews 1985). This position is consistent among the Rhizodontida . The morphology of the scapulocoracoid attachment on the Middle Paddock cleithra is somewhat variable. In certain specimens (QMF34611, Fig. 3E View FIG , arrow), ridges form a V- or U-shaped structure posteriorly (attachment for the glenoid buttress). A low anterior ridge forms the attachment for the supraglenoid and infraglenoid buttresses of the scapulocoracoid. In another specimen (QMF36762, Fig. 3C View FIG ), the posterior attachment area is flatter and less clearly composed of ridges. The area between the anterior and posterior attachment areas is slightly concave in QMF36762 but less well-developed in other Middle Paddock cleithra.

On a smaller specimen of Strepsodus sp. from Turkey (Janvier et al. 1984: pl. 2.10), the posterior scapulocoracoid attachment appears to be composed of two short, obliquely oriented and parallel ridges. The anterior ridge runs to about the level of these two ridges, forming a shallow V-shaped area between them. The comparable area on described specimens of Strepsodus sauroides and? S. anculonamensis is either covered by the scapulocoracoid (Andrews & Westoll 1970: pl. 11A) or not visible (Andrews & Westoll 1970: pl. 11H; Andrews 1985). The attachment on the “larger Foulden form” appears flat and not clearly composed of distinct ridges (Andrews 1985: fig. 11e), as in QMF36762 ( Fig. 3C View FIG , arrow). The nature of the anterior ridge is not clear. The posterior scapulocoracoid attachment in the Horton Bluff specimens is similar to the “larger Foulden form” and QMF36762. These areas are distinctly separate from the main body of the cleithrum. As noted above, this is not the case for Barameda (NVM P181699, Long 1989: fig. 7C). On the Horton Bluff specimens, the anterior ridge for the supraglenoid and infraglenoid buttresses is sharply developed and separated from the posterior by a deep and rounded groove. The scapulocoracoid is preserved on certain of these specimens and appears firmly fused to the cleithrum. That is, no fossae are present between the scapulocoracoid and the cleithrum. The presence or absence of these fossae in Strepsodus sauroides (Andrews & Westoll 1970: 442) is not well established. The scapulocoracoid is entirely absent from the Middle Paddock cleithra.

Young et al. (1992: fig. 47) restricted the British Carboniferous rhizodonts, including Screbodinus, Strepsodus and Rhizodus to the Rhizodontidae , with other rhizodonts (including Barameda ) forming more inclusive, basal branches within the Rhizodontida . Johanson & Ahlberg (1998) established that Gooloogongia was phylogenetically the most basal rhizodont, with Barameda forming a polytomy with the Rhizodontidae . Problematic in these phylogenetic analyses is Gooloogongia , which has little information on the nature of the pectoral fin supports, and Sauripteris , for which information is limited entirely to the latter. This results in uncertainty in character state codings and unresolved phylogenetic relationships.

One thing that can be established is that Rhizodus and Strepsodus represent the most derived rhizodonts, for example, in the development of the posterior flange of the cleithrum, and in the course of the postotic canal crossing onto the postparietal bone of the skull roof. This canal runs along the bones flanking the postparietal in Barameda and Gooloogongia (Young et al. 1992; Johanson & Ahlberg 1998). Screbodinus ornatus, although also possessing a reduced posterior flange (Andrews & Westoll 1970), shows a similar course of the postotic canal to Rhizodus and Strepsodus (Andrews 1985: fig. 7c).

BIOGEOGRAPHICAL REMARKS

Skull material from Middle Paddock is still under investigation; nevertheless, the similarity in shoulder girdle material suggests that the Middle Paddock rhizodontid is more closely related to the Rhizodontidae (although the number of species at Middle Paddock and their affinities within Strepsodus are difficult to determine at this time) than to other eastern Gondwanan taxa, which appear to be primitive in a variety of characters including those mentioned above.

The age of eastern Gondwanan rhizodonts ranges from Givetian-early Frasnian (“ Notorhizodon ” macelveyi) to late Frasnian ( Gooloogongia ) and into the Carboniferous ( Barameda ,?Tournaisian; Middle Paddock specimens, mid-Viséan). The referral of the Middle Paddock specimens to Strepsodus extends the range of this taxon, previously described from numerous northern hemisphere localities (Andrews 1985), including shoulder girdle material from the late Tournaisian Calciferous Sandstones at Foulden, Scotland (? S. anculonamensis and the closely related “larger Foulden form” [Andrews 1985]) and various Late Carboniferous localities in the British Coal Measures (Bashkirian-Westphalian; Andrews & Westoll [1970]; Andrews [1985]). Among the Rhizodontida , shoulder girdle material (cleithra) from Horton Bluff (Nova Scotia, early Tournaisian) also shows strongest similarity to Strepsodus . This distribution indicates interchange between Laurentian localities and East Gondwana in the Carboniferous. This is supported by the presence of Strepsodus sp. in Upper Devonian (latest Famennian) deposits of Turkey (Janvier et al. 1984) and Iran (Dashtban 1996; Janvier pers. comm.), along the northern margin of Gondwana.

Thulborn et al. (1996) noted that the Middle Paddock fauna was similar to faunas included in Milner’s (1993) Mississippian Tetrapod Province, previously restricted to Laurentia (North America and western Europe). If correct, a significant extension of the Mississippian Tetrapod Province into East Gondwana is indicated during, or prior to, the Viséan. The Ducabrook rhizodontids show similarities to both Mississippian (Foulden, Horton Bluff) and Pennsylvanian/Late Carboniferous ( Strepsodus sauroides, Coal Measures of Britain) species of Strepsodus . This may indicate links or interchange beyond the Mississippian. Further study of the remainder of the Middle Paddock fauna, and description of other taxa from faunas of the Mississipian Tetrapod Province (e.g., from Iowa, West Virginia and Nova Scotia [Schultze & Bolt 1996; Clack & Carroll in press]) will help to clarify these suggestions.

Nevertheless, the presence of Strepsodus in the Middle Paddock fauna marks the appearance of the Rhizodontidae in East Gondwana. This might represent a distinct faunal change within East Gondwana from the Devonian to the Carboniferous, and perhaps more specifically, within the Carboniferous. The Middle Paddock fauna includes the most derived East Gondwana rhizodont known in association with at least one tetrapod.

By comparison, the rhizodont Barameda occurs in the Lower Carboniferous Mansfield Group of Victoria; tetrapods are absent. Although there is some uncertainty as to the phylogenetic position of Barameda (see above), a referral to the Rhizodontidae (with Strepsodus and Rhizodus ) is unlikely. This upper level of the Mansfield Group may be older than previously accepted (possibly Devonian), with the Carboniferous age based largely on the absence of placoderm material present in lower levels (Long & Campbell 1985). Undoubted Lower Carboniferous localities (Tournaisian-early Viséan, and so older than the Middle Paddock fauna) in the Upper Telemon and Raymond Formations of the Narrien Range, northern Drummond Basin, Queensland, contain an osteolepid, actinopterygians, acanthodians, xenacanths and other sharks and possibly rhizodonts (Turner 1993; Fox et al. 1995), but again, no tetrapod material.

With regards to other taxa, a ctenodont lungfish is present at Middle Paddock (Turner et al. 1999), while in the Mansfield Group, the lungfish Delatitia Long & Campbell 1985 appears less derived in characters such as the course of the supraoccipital commissure through the A rather than the B bone at the back of the skull (Schultze & Bolt 1996: character 3). The ctenodont lungfish Tranodis Thomson 1965 from the Goreville (Illinois) and Delta (Iowa) localities (Viséan- Namurian) is associated with taxa that are also found at Middle Paddock, including tetrapods, rhizodonts, palaeoniscids, xenacanths, hybodonts, other sharks and gyracanthid acanthodians (Schultze & Bolt 1996; S. T. pers. obs.). The former localities were included in Milner’s (1993) Mississipian Tetrapod Province and a comparison between the rhizodonts here and those from Middle Paddock will be important.

The differences just described suggest some type of major change in East Gondwana within the Early Carboniferous. This may not be entirely due to variation in the position of Laurentia and Gondwana, as strong evidence exists for an association between these landmasses in the late Frasnian and into the Famennian. This is based upon the presence of the lungfish Soederberghia Lehman, 1955 as well as tristichopterid sister taxa (Tetrapodomorpha) in New South Wales and various Laurentian sites (Ahlberg & Johanson 1997; Johanson & Ahlberg 1997; Ahlberg et al. in press) .

The presence of Strepsodus in the Carboniferous of Middle Paddock, eastern Canada and Britain, and in the Upper Devonian of Turkey and Iran corroborates suggested interchange between Laurentian and Gondwanan landmasses in the Late Devonian and Carboniferous. Additionally, Frasnian-aged Strepsodus -like material has been recovered from South America (in Colombia, Janvier & Villarroel 1998). The absence of Strepsodus or other more derived rhizodontid taxa from the Upper Devonian East Gondwana ( Australia + Antarctica, Young [1981]) may require further explanation, as Strepsodus was probably already established in western and northern Gondwana during this time.