Corystospermaceae (or Umkomasiaceae of Petriella, 1981; Meyen, 1987; Anderson and Anderson, 2003; Holmes and Anderson, 2005; Pattemore et al., 2015) was established by Thomas (1933) to include reproductive and vegetative structures preserved in the Karoo Basin, South Africa. The family name proposed referred to the cupulate nature of the uniovulate organ (from the Greek corysto, a helmet wearer). For Petriella (1981), the Umkomasiaceae is a preferable familiar name, taking into account the “Triassic” cupulate organ Umkomasia (Thomas) Klavins, Taylor and Taylor (Klavins et al., 2002; Taylor et al., 2009), recently also found in upper Permian levels of India and in the Lower Cretaceous of Mongolia (Chandra et al., 2008; Shi et al., 2016). Subsequent works substantiated the use of Corystospermaceae for those Mesozoic pteridosperms (Townrow, 1957; Jain and Delevoryas, 1967; Petriella, 1979; Baldoni, 1980; Retallack, 1981; Holmes, 1982; Artabe, 1990; Gnaedinger and Herbst, 1998; Axsmith et al., 2000; Taylor et al., 2006; Artabe et al., 2007a, b; Bomfleur and Kerp, 2010).

Leaf morphologies representing the distinct genera of Corystospermaceae are diverse and sometimes exhibit mixed/transitional features or lead to the proposal of a hybridization model of speciation (Anderson and Anderson, 1983). Their distinct morphotypes have been mostly included within the genera Dicroidium (Gothan) Townrow or Zuberia (Frenguelli) Artabe, Johnstonia Walkom, and Xylopteris (Frenguelli) Stipanicic and Bonetti, associated by the general presence of basal bifurcate fronds and some cuticle features (Bomfleur and Kerp, 2010). There is no consensus regarding the separation from the other proposed genera nor about which characteristics prove useful for such a separation.

Corystospermaceae assemblages have been identified along palaeolatitudes 30° S-80° S, in areas marked by highly diversified environmental conditions (e.g., Anderson and Anderson, 1983; Cantrill et al., 1995; Troncoso and Herbst, 2007; Escapa et al., 2011). These were capable of growing from near coastal to highland areas and in the lowland interior areas of continents along the margins of temperate forests and/or lakes or even under somewhat arid conditions (Bomfleur and Kerp, 2010).

The oldest record of corystosperm leaves dates from the Anisian?-Ladinian of Australia, Argentina and South Africa (Beaufort Group) and is presumed to extend to the Rhaetian (Retallack, 1977; Anderson and Anderson 1983).

The placement of the genus Xylopteris among Corystospermaceae, initially proposed by Frenguelli (1943), was based on the basally bifurcated pinnate fronds (such as those of Dicroidium) and the mainly linear and one-veined pinnules. Frenguelli (1943) included in the new genus some previously described forms such as Sphenopteris Brongniart, Trichomanides Tenison-Woods and Stenopteris Saporta. Townrow (1957) adopted a comparable procedure when proposing the emended diagnosis of Dicroidium, headlining the similar cuticle features that characterise Dicroidium and Xylopteris. In spite of these similarities, the author maintained Xylopteris as a separate genus based on its univeined segments.

Petriella (1979), studying the Argentine forms, took an analogous posture of resorting to the univeined attribute to distinguish Xylopteris from Dicroidium and Johnstonia.

However, the occurrence of a single vein in the pinnules and the exclusive presence of pinnate fronds were later reanalysed after the finding of Xylopteris argentina (Kurtz) Frenguelli, in which bi- to tripinnate fronds are exhibited. Such features lead to the proposition of an emended diagnosis of Xylopteris drawing attention to the lobate nature of the pinnules and a “tendency to a sphenopteroid venation pattern” with free-ending veins (Stipanicic and Bonetti, in Stipanicic et al., 1996).

To better distinguish leaf morphologies, Retallack (1977) and Stipanicic et al. (1996) still proposed the use of varieties (or forms) respectively for the Australian and Argentine remains. Retallack (1977), although stressing the cuticular features shared among Xylopteris, Dicroidium and Johnstonia, did maintain the segregation of Xylopteris based on the diagnostic features expressed by Walkom (1925), Frenguelli (1943), Jacob and Jacob (1950) and Townrow (1957). Additionally, Retallack proved that thicker and straighter cell outlines, sunken stomata and cutine flanges overhanging the poles of the guard cells characterise Xylopteris cuticles. Taking such description into account, Baldoni (1980) provided a review of the anatomical features of X. elongata and X. rigida from Argentina, Australia and South Africa.

Subsequent researchers considered that the established cuticular similarities proved to be good enough criteria to maintain Xylopteris, together with Hoegia, Zuberia, Diplasiophyllum and Dicroidiopsis, in the genus Dicroidium because of its mixed and transitional morphologies (Thomas, 1933; Archangelsky, 1968; Holmes, 1982; Holmes and Anderson, 2005; Bomfleur and Kerp, 2010; McLoughlin, 2013). The segregation of Xylopteris was upheld by Meyen (1984, 1987), who additionally proposed that Dicroidium and allied forms present many epidermal features shared with the Peltaspermales. Many other researchers, especially those dealing with South American and Australian assemblages, also adopted this posture (Menéndez, 1951; Petriella, 1979; Baldoni, 1980; Gnaedinger and Herbst, 1998; Ottone, 2006; Morel et al., 2010; Ottone et al., 2011; Pattemore and Rigby, 2005).

The detailed cuticle analysis performed by Bomfleur and Kerp (2010) with Dicroidium specimens from East Antarctica (with forms hereby included in Xylopteris) contributed some critical and diagnostic characteristics. For them, Dicroidium forms exhibit a distinct and basic epidermal anatomy although also presenting variations in the overall thickness of the cuticle, cell dimensions, trichome bases, the cutinized subsidiary cells (in D. odontopteroides and D. crassinervis), the disposition of the amphistomatic stomata and anticlinal wall microstructures. However, a recent study conducted by Pattemore et al. (2015) suggests that even when reproductive structures and cuticle features are present, there is no support for including Johnstonia and Xylopteris in Dicroidium.

Taking into account the presence of pinnate fronds with one-veined segments and bipinnate fronds with lobate pinnules, features corresponding to the emended diagnosis proposed by Stipanicic and Bonetti (in Stipanicic et al., 1996), we herein adopted the inclusion of part of the leaf impressions previously identified from the Santa Maria Formation assemblage in the genus Xylopteris.

STRATIGRAPHIC AND GEOLOGICAL SETTING

Previous studies on a set of rare and partially preserved forms of Corystospermaceae from the Triassic of Brazil confirm its restriction to the southernmost areas of the country. In such reports, the Xylopteris-type morphologies were included as a subgenus of Dicroidium with two species, D. (Xylopteris) elongatum and D. (Xylopteris) argentinum (Bortoluzzi et al., 1985; Guerra-Sommer et al., 1999; Guerra-Sommer and Cazzulo-Klepzig, 2000). The plant-bearing levels were assigned to the type section of the Passo das Tropas Member (PTM), from the Santa Maria Formation (Bortoluzzi, 1974), and considered to be Anisian-Ladinian in age based on Dicroidium representativeness (Guerra-Sommer et al., 1999; Guerra-Sommer and Cazzulo-Klepzig, 2007).

The leaf material herein studied was collected south of the Santa Maria municipality, Rio Grande do Sul (Fig. 1), from a new exposure 500 m north from the original studied site. The corystosperm remains dominate the macroflora and are accompanied by ginkgophytes (the second element in abundance), cycadophytes, isolate Taeniopteris leaves, rare sphenophytes and conifers. Insect wings, fish scales and conchostracans complete the assemblage (Barboni and Dutra, 2015).

The PTM was considered to be the lowest member of the Santa Maria Formation by Bortoluzzi and Barberena (1967), Bortoluzzi (1974) and Andreis et al. (1980). Under a Sequence Stratigraphy approach conducted by Faccini (2000) and Zerfass et al. (2003), PTM deposits were considered to be the basal part of a second order sequence (Santa Maria 2 Sequence - SM2, Fig. 2). Later, Zerfass et al. (2004; 2005) proposed that tectonic control affecting all the Santa Maria Supersequence induces extensional efforts that generate a small and independent basin (Santa Maria Basin) included within the broader Paraná Basin.

Together with the Waterberg Basin from Namibia, the Santa Maria Basin is part of a set of Western Gondwana intracontinental basins that form depressed and aligned areas generated by the Waterberg-Omaruru fault system. Those geological constraints could explain the regional and exclusive nature of the deposition and its fossil assemblages which prove absent from other parts of Brazil but comparable with those preserved in South Africa and Argentina (Zerfass et al., 2005; Golonka, 2007).

The floristic assemblage (Gordon and Brown, 1952; Pinto, 1956; Bortoluzzi et al., 1985; Guerra-Sommer et al., 1999; Guerra-Sommer and Cazzulo-Klepzig, 2000, 2007; Barboni and Dutra, 2013) and tetrapod fauna (Schultz et al., 2000; Langer et al., 2007) of SM2 seem to be coherent with this geological approach. However, the correlation and age of those two distinct proxies (flora and tetrapods) are still under discussion. Based on the new allostratigraphic approach and the studied plant fossils, a preliminary ordination of the floral assemblages and their relation with the faunal content are herein proposed (Fig. 2). It is also supported by the absence of Dicroidium-related forms in the younger floras of the Caturrita Formation (Santa Maria 3 Sequence- SM3), the only known exception being the ex-situ fossil wood Rhexoxylon brasilensis Herbst and Lutz (Herbst and Lutz, 1988), of uncertain source area. We also considered the identification of a very distinct flora hereby informally referred to as “conifer and bennetittes flora” (Wilberger and Dutra, 2012; Barboni and Dutra, 2013) and exclusively composed of evolved Bennettitales (Williamsonia potyporanae Barboni and Dutra) and shoots of Pagiophyllum Heer in the upper SM3 levels. A probable Jurassic Spinicaudata fauna was described for the same beds (Rohn et al., 2014).

Figure 1.

1, Location of the Paraná Basin and the distribution of Paleozoic and Mesozoic strata in the central part of the State of Rio Grande do Sul, southern Brazil (modified from Faccini, 2000); 2, Location of the Passo das Tropas outcrop, south of Santa Maria County.

Those levels give place to an uppermost floral assemblage with still dubious field relations and containing only big fossil wood logs which were firstly assigned to the “Araucarioxylon Flora” Cenozone (Guerra-Sommer et al., 1999; Guerra-Sommer and Cazzulo-Klepzig, 2000) and identified in the informal unit “Mata Sandstones”, considered Norian in age (Faccini, 2007). Based on descriptions crafted with those materials in more recent years (Pires and Guerra-Sommer, 2004; Bardola et al., 2009, Crisafulli et al., 2012), which attest the presence of other groups of gymnosperms (e.g., Taxaceae, Ginkgoaceae and Araucariaceae), we proposed an assignment to “Gimnosperm xyloflora” (Fig. 2).

The all-red bed Triassic succession exposed in southern Brazil attests to a cyclic recurrence of aeolian, fluvial and lacustrine deposits (Faccini, 2000; Zerfass et al., 2003) whose influence over the vegetation was firstly observed in the Argentine basins by Spalletti et al. (2005). Due to the intracontinental nature of those areas of Western Gondwana and probably also because of the new reliefs emerging after tectonic activity, moonsonal climate conditions were established at the end of the Triassic over those areas (Parrish et al., 1982; Ziegler et al., 1993; Wilson et al., 1994; Mutti and Weissert, 1995). The extensive aeolian sands that cover the fluvial deposits with flora (Guará and Botucatú formations, Fig. 2) indicate progressively drier conditions (Bortoluzzi, 1974). According to Faccini (2000), the PTM facies attests to the dominance of fluvial systems containing lenticular limnic beds (Fig. 3) and represents the filling of abandoned river channels or depressed lacustrine areas protected from active erosion.

Figure 2.

Lithostratigraphy and the proposed Sequences to the Late Permian-Mesozoic strata of Paraná Basin in the State of Rio Grande do Sul (Faccini, 2000; Zerfass et al., 2004) and the associate tetrapod Assemblage Zones-AZ (Langer et al., 2007; Abdala and Ribeiro, 2010; Soares et al., 2011). The new informal zones of the taphofloras are based in more recent studies (Pires and Guerra-Sommer, 2004; Bardola et al., 2009, Crisafulli et al., 2012; Barboni and Dutra, 2013), and in their ordination in relation to time, faunal AZ and the depositional sequences. Preservational styles and lithofacies based on Faccini (2000).

The analysed outcrop is located at the margins of BR 392 highway, 3.5 km away from Santa Maria city (29° 44′ 40″ S; 53° 47′ 36″ W; Figs. 1, 3). The fossiliferous section exposes 2.20 m of reddish-brown to pink-yellow coloured laminated mudstones. The only disturbing event in the monotonous mudstone interval is the cyclic deposition of a great concentration of fossil leaves together with occasional lenses of iron oxyhydroxides, which sometimes form thin coats over the fossil impressions, thus highlighting their anatomical details. The basal psamitic interval is characterised by an alternation of massive to low angle trough cross-bedded sandstones and conglomerates (with gravels and boulders of mud). The exposed facies fit well with those described by Bortoluzzi (1974) for the PTM type section, today covered by urban expansion.

Figure 3.

Stratigraphic section, sedimentary features and the occurrence of Xylopteris remains in the outcrop of the Passo das Tropas Member (PTM), Santa Maria Formation. Numbers in the shale interval indicate the levels submitted to the detailed taphonomic work.

MATERIAL AND METHODS

The good preservation of the leaves in the PTM outcrop attests to the dominance of Corystospermaceae, Ginkgoaceae and Taeniopteris and allows for the evaluation of their distinct leaf physiognomy. In Corystospermaceae related forms, the basal bifurcation, rare in previously known materials, is generally preserved. Vegetative structures are abundant and prove accompanied by reproductive ones. They appear spread along the whole section yet concentrated in some specific levels and with superimposed and multidirectionally oriented fragments.

A detailed analysis was performed in the mudstone interval, with quadrats of 20 cm² for every 10 cm of height. This favoured the exploration and identification of twenty levels (N1 to N20, bottom to top) of which N5 and N7 exhibit greater richness and contain 15–20 faunal and floral organs in each quadrat (20 cm²). Additionally, from N4 to N17, an average of 13–14 distinct organisms per level were found. Most of these organisms remain under taxonomic study. Xylopteris-related frond impressions and isolate reproductive structures represent 14.30% of the total assemblage and appear accompanied by Dicroidium, Zuberia and Sphenobaiera leaf impressions (Corrêa, 2014). Despite such proportionally low representation Xylopteris-related forms occur along the whole profile and even in the uppermost levels, where other components become rarer (Fig. 3).

After collection, samples were wrapped in PVC plastic film for two weeks or until they dried completely, therefore preventing cracks from forming on air exposure of the mudstone materials (Barboni et al., 2008).

The 64 Xylopteris-related leaf impressions or fragments are stored in the palaeontological collection (LaViGeae) of the Museu de História Geológica do Rio Grande do Sul (MHGEO), at the Graduate Program in Geology of UNISINOS (Universidade do Vale do Rio dos Sinos), under the label ULVG.

For the analysis and graphic record of the distinct morphotypes we used an Olympus SZH Stereomicroscope with camera lucida. Photographs were taken with a Fujifilm Finepix hs50exr digital camera. Cross-polarized light and close-up filters were used to enhance contrast, and fine details were recorded under low-angle light. Corel Draw Graphic SuitesTM X6 was used to control or enhance contrast and brightness and to record particular morphological features.

Among the leaf impressions, those which prove more coriaceous and present thicker venation or rachis also exhibit a fine dark-brown iron oxyhydroxides coat which covers the leaf and results in rarer and exceptional preservation which includes cuticles and mesophyllous cell molds over the Dicroidium and Ginkgo-related forms (Barboni and Dutra, 2015). Such phenomenon suggests the occurrence of a process related to post-mortem bacteria-mediated stabilization activities marked by, in the first phase of the taphonomic process, leaves and other plant remains favouring the adsorption of metallic ions and generating ferrihydrites which enhanced the preservation during burial (Dunn et al., 1997; Gupta and Pancost, 2004).

The taxonomic treatment of type-morphologies takes into account previously studied assemblages from Gondwana. Higher taxonomic categories follow Stewart and Rothwell (1993). As previously discussed, a preferable use of Corystospermaceae instead of Umkomasiaceae was adopted for family ranking.

SYSTEMATIC PALAEONTOLOGY

Pinnate Fronds

Xylopteris rotundipinnulia nov. comb. gen., nov. sp. (Figures 4.1–4)

1979 Xylopteris spinifolia (Tenison-Wood) Freng. Petriella (p. 97 and 100, lam. 2, fig. 12) .

1980 Xylopteris spinifolia (Tenison-Wood) Freng. Baldoni (p. 143, lam. 1, figs. 3? and 6).

1983 Dicroidium elongatum (Carruthers) Archangelsky forma rotundipinnulium Anderson and Anderson (p. 17, lam. 38, figs. 27–28, basionym).

Holotype. ULVG 9155A, from the LaViGæa Collection, UNISINOS.

Material. ULVG 11131.

Etymology. Refers to the expanded, rounded and multiveined lobes on the margins of the pinnae, as proposed by Anderson and Anderson (1983) in the diagnosis of D. elongatum forma rotundipinnulium.

Type locality and horizon. Levels N4 (base) and N17 (top) of the rhythmic mudstone interval (PTM).

Diagnosis. Basal bifurcate and pinnatifid frond with laterally expanded rachis. Linear pinnules characterised by two-three bilaterally symmetrical and basal rounded lobes that dilute into a unique and longer one at the apices. Pinnules opposite to alternate and regularly spaced, inserted at acute angles and with veins variable from parallel to the margins to basally ramified, each branch freely ending in the marginal lobes.

Description. Incomplete fronds of probable delicate texture (3.5 cm long in the preserved portion), with expanded rachis (1.5 to 2 mm wide) and bifurcated at an angle of 13o. Pinnules diversified in form and size, ranging from opposite to alternate (angles of 45°-50°), varying from linear (3–4 mm wide, 1.8–2.2 cm long) and regularly spaced (5–8 mm wide) to pinnatifid, with a maximum of three symmetrical and opposite basal pairs of lobules. The main vein runs along the pinnules, giving rise to ramified free branches.

Remarks. Fossil leaf impressions referred to this morphotype prove rare in the PTM assemblage and are herein assigned to Xylopteris (Frenguelli) Stipanicic and Bonetti based on the expanded rachis, the linear leaves and the lobate margins in the pinnules, with a proximally bifurcated central vein (sphenopteroid). Stipanicic and Bonetti (in Stipanicic et al., 1996) suggested that the sphenopteroid venation serves as a criterion in their emended generic diagnosis. Yet, the authors did emphasise on how rare the presence of such venation is and on the exclusivity of its occurrence in fronds with lobate pinnules. Frenguelli (1943) and Baldoni (1980) also drew attention to this feature in some Xylopteris and suggested that transitions between linear to lobate margins in the pinnules could reflect distinctive ontogenetic phases. However, Jones and Jersey (1947: fig. 16) proposed that the different pinnules along the rachis of Stenopteris Thomas (~ Xylopteris) could represent evolutionary series of stratigraphic value.

The size of the fronds, the lobate margins and the bifurcate venation, each branch coinciding with the marginal lobes, characterise the new species herein proposed. Its description agrees with the form illustrated by Anderson and Anderson (1983) for Dicroidium elongatum “forma” rotundipinnulium from the Molteno beds, South Africa, in nearly all aspects. The only feature observed, not highlighted (but illustrated), is the presence of expanded rachides that give rise to the final segments. The “formas” and subspecies assigned to D. elongatum were later abandoned (Anderson and Anderson, 2003) but the marked morphological affinity and the presence of round lobate pinnules, similar to those of the Molteno fossils (Anderson and Anderson, 1983, 2003), reveal that the new specific epithet is useful to characterize the species herein proposed. Additionally, it guarantees the age inferences and the correlations with other Gondwanan areas.

Baldoni (1980) categorised leaf morphologies similar to that of D. elongatum forma rotundipinnulium into her Group B of Xylopteris. Even if present in both South Africa and in the Clarence Morton Basin of South Australia (Table 1), this morphotype is rarer and normally restricted to scarce levels in the successions.

Figure 4.

1–4, Xylopteris rotundipinnulia nov. comb. gen., nov. sp. 1–2, General aspect and the lobate pinnules (ULVG 9155-A); 3–4, Detail of the pinnae bifurcation and the “sphenopteroide” kind of venation in the pinnules (ULVG 9155-A). Scale bar= 5 mm.

Xylopteris elongata (Carruthers) Frenguelli, 1943
(Figures 5.1–2)

Materials. ULGV 9171, 9210; 9211; 9269; 9406a; 9747A; 9748B; 9748A; 11142; 11164; 11091A, 11196a-11196b.

Type locality and horizon. Distributed between the middleupper levels (N5 to N16) of the rhythmic mudstone interval.

Description. Well-preserved pinnate fronds (6.5 cm large) with rigid pinnae (2 mm wide) that bifurcate at acute angles (15°-20°). Rachis basally wide (2.7–3 mm in preserved samples) and carrying elongate and linear pinnules (0.8–1.5 mm wide) partially covered with fine iron deposits. Apices not preserved. Planar one-veined pinnules with parallel margins (ribbon-like) diverging from a lateral expansion of the rachis (winged?), with a regular spacing (4 to 6 mm) and suboppositely inserted at acute angles of 25°-35°. Some seem to present apical bifurcations, not well preserved in the samples.

Remarks. As previously noted by other researchers (e.g., Ganuza et al., 1995, and other references in Table 1), morphological features, such as the pinnate frond with linear regularly spaced pinnules and a faint main vein, fall into the range of characteristics associated with Xylopteris elongata Frenguelli (Frenguelli, 1943). Of the two varieties of X. elongata proposed by Retallack (1977), only X. elongata var. rigida (Dun) Stipanicic and Bonetti (Stipanicic et al., 1996) fits within the characteristic profile discussed here.

Although the cuticles described by Baldoni (1980, pl. 11, p. 146) differs in X. elongata from those of X. Rrigida, both species are characterised by linear and coriaceous pinnae. According to Baldoni (1980), in the former the anticlinal and periclinal walls are smooth while the stomata are circular to ovate and irregularly distributed in relation to the main vein. The presence of papillae on subsidiary cells is occasional. However, Bomfleur and Kerp (2010) allude to the similarity of the cuticular features in that kind of frond and propose its inclusion within the variation of D. elongatum. To them, the distinct anatomical features would be independent from the gross morphology and could only represent intraspecific variables. They also illustrated pinnae with widely spaced pinnules, such as those observed herein, sometimes presenting an apical cusp-like segment. Such combination of features opens the possibility of having some forms of D. elongatum that could be bipinnate.

Finally, in the list of synonyms of X. elongata var. elongata presented by Retallack (1977), the author emphasises the similar morphologies found in some X. densifolia morphotypes. For the second variety, X. elongata var. rigida, Retallack (1977) proposes a correlation with X. densifolia Frenguelli. However, the X. densifolia of Frenguelli (1943) exhibits characteristics absent in the morphotype herein discussed; namely, closely spaced pinnae and broader pinnules which are truly inserted oppositely.

Xylopteris elongata is abundant and well distributed in the PTM levels. Analogously, such conditions also characterise the species in other Gondwana assemblages, including those from Antarctica (Bomfleur and Kerp, 2010). The finding of Rhaetian forms in Argentina extends its time distribution, previously considered to be restricted to the middle and early Late Triassic (Tab. 1-2).

Figure 5.

1–2. Xylopteris elongata. 1, General aspect of the pinnae showing the ample angle of divergence of its linear, univeined and long pinnules (ULVG 9171); 2, Detail of the basal part of the frond, with pinnules diverging directly from the rachis (ULVG 11142); 3–4. Xylopteris rigida. 3, Basal part of a frond showing the slight and rigid pinnules, some upwardly curved, directly inserted on the main rachis (ULVG 11535Ba); 4, Curved pinnules (arrow) of X. rigida, associated with a pinna of a probable Zuberia (ULVG 9747-A). Scale bar= 5 mm.

Xylopteris rigida (Dun) Jain and Delevoryas, 1967
(Figures 5.3–4)

1983 Dicroidium elongatum subsp. matatifolium Anderson and Anderson (p. 30, pl. 63, figs 1–18).

Materials. ULVG 9269; 9767aA-9767b; 11088B; 11190; 11194a-11194b; 11535a-11535bA; 111536.

Type locality and horizon. Concentrated at the boundary of middle to upper levels (N14-N15) of the mudstone interval.

Description. Partially preserved coriaceous fronds (6–7 cm large) with expanded rachis (1 mm wide) and a very acute bifurcation angle (15o). Linear and characteristically slender univeined pinnules (0.5–0.8 mm wide) opposite to subopposite and widely spaced (5–8 mm), inserted at angles of 50°-60°. Usually, pinnules are loosely arranged and straight on one side while decurrent and slightly curved upwards on the other. Apices are acute. Few and wider (2.4 mm) pinnules directly connected, immediately under the basal bifurcation, to the main rachis.

Remarks. The diagnostic characteristics of this morphotype are the pinnate fronds as well as the very narrow and wellspaced pinnules, of which some directly diverge from the main rachis. This type of morphology is very similar to that which can be recognised in Dicroidium elongatum subsp. matatifolium, in Anderson and Anderson (1983). X. rigida also resembles X. elongata (e.g., Jain and Delevoryas, 1967; Baldoni, 1980) in presenting the noticeable feature of linear and regularly spaced pinnules. This is probably the reason for the joint taxonomic treatment of both species in previous works.

In the original description of X. elongata by Townrow (1962, 1967), the presence of denser pinnules was considered to be a criterion for distinguishing the species from X. rigida. Townrow (1967) also proposed that the rarer “slightly sinuous pinnae” in some X. elongata could represent only varieties. Additionally, Baldoni (1980) differentiated X. rigida (lam. 11, p. 146) from X. elongata in terms of the former's cuticular features and included all the Townrow species (1962) in X. rigida. However, Bomfleur and Kerp (2010) observed that the greater distances between the pinnules in the pinnate/bipinnate fronds of D. elongatum distinguished the species from X. elongata.

The previously discussed morphological features were used herein for distinguishing these two species, grouping those presenting slightly curved to straight pinnules within X. elongata and those with lower bifurcation angles, asymmetrical and lax fronds and, sometimes, boasting highly curved pinnules within X. rigida.

Morphotypes comparable to the herein studied X. rigida occur in Argentina (e.g., Ottone, 2006; Table 1), Australia and South Africa. However, in Australia and South Africa, such forms were included in Dicroidium (Retallack, 1977; Anderson and Anderson, 1983).

Xylopteris densifolia (Du Toit) Frenguelli 1943
(Figures 6.1–4)

1927 Stenopteris densifolia Du Toit, text-fig. 13a (basionym).

1977 Xylopteris elongata var. elongata (Carruthers) Frenguelli, Retallack, Frame 19 (illustration).

1995 Xylopteris elongata var. elongata (Carruthers) Frenguelli, Ganuza et al., lam. 1g.

Material. ULVG 9743

Type locality and horizon. A rare form found in an ex situ sample.

Description. Impression of a vastly incomplete pinnate frond (4 cm large in the preserved sector) exhibiting longitudinal striations (3–4) over the wide and expanded (winged?) secondary rachis (3 mm wide). The straight and ribboned pinnules (1.3–2.0 mm) present a proximal and regular insertion (2.2 mm spaced) at acute angles (30°) and acute apices.

Remarks. The dense foliage together with the form and disposition of the pinnules attest to a relation with X. densifolia. In his first description of the species from the Molteno beds, Du Toit (1927) emphasised those features and the wide “winged” nature of the rachis and included the latter Stenopteris densifolia, which was afterwards transferred to Xylopteris by Frenguelli (1943). Baldoni (1980), while searching for cuticular features in the straight pinnules of X. densifolia, proposed that such characteristic could represent a distinct and initial phase of development related to the lobulated bi- to tripinnate fronds of her Group B foliage (with X. tripinnata and X. spinifolia).

In the materials studied herein, X. densifolia proved scarce and represented by fragmentary impressions, a condition also observed in samples from Argentina and South Africa (Table 1).

Table 1

- Previously known distribution of Xylopteris and its related forms in the Gondwana assemblages. In clear grey, the inferred wetter areas, in white the seasonally dryer ones. Obs.: (1), includes X. elongata var. rigida, X. elongata var. elongata (Retallack, 1977; Artabe et al., 1994; Ganuza et al., 1995; Morel et al., 1999) and some Dicroidium elongatum (Townrow, 1967; Archangelsky, 1970; Boucher et al., 1997); (2), includes Stenopteris spinifolia Jones and De Jersey (1947), Dicroidium elongatum forma remotipinnulium (Anderson and Anderson, 1983) and D. elongatum (Anderson and Anderson, 2003); (3), with the D. spinifolium Anderson and Anderson (Bomfleur and Kerp, 2010); (4), with Stenopteris tripinnata Jones and De Jersey (1947) and Dicroidium tripinnata Anderson and Anderson (1970). (*) Age data of 211+5 in interbedded basalts of the Nymboida area, Australia (Retallack et al., 1977), indicates a Norian age to the fossil floras (Wotzlaw et al., 2014; Gradstein et al., 2014).

Continuation

Bipinnate fronds

Xylopteris spinifolia (Tenison-Woods) Frenguelli, 1943
(Figures 7.1–4, 8.1–3)

1883 Trichomanides spinifolia Tenison-Woods (p. 197, pl. 3, fig. 7).

1898 Trichomanites elongata v. spinifolia Shirley (pl.V, fig.2, p. 19).

1983 Dicroidium elongatum forma spinifolium Anderson and Anderson (pl. 47, figs. 1, 4, 10).

1983 Dicroidium superbum forma bipinnatum Anderson and Anderson (pl. 50, figs. 19, 21).

Material. ULVG 8994; 9150; 9191; 9213; 9226; 9313Aa-9313b; 9345; 9531; 9552; 9747C; 9750; 9753; 9754; 11007; 11018; 11056; 11061; 11062; 11088; 11094; 11150; 11172; 11177; 11189; 11191; 11194; 11220; 11235; 11398; 111399A; 11414; 11431.

Type locality and horizon. Numerous and regularly distributed frond imprints in middle to upper levels (N7 to N16) of the mudstone interval.

Description. Basal-middle part of a bipinnate, apparently delicate frond with a well-developed and striated main rachis 6–7 cm long and 8 mm wide (preserved sector) bifurcating at an angle of 40°. Secondary pinnae oppositely inserted (2 mm wide), some directly on the distal end of the main rachis and below the main bifurcation. The penultimate univeined pinnae (1 mm wide) give rise to at least five pairs of triangular and short (6–9 mm long) pinnules which are also one-veined, regularly spaced and present acute to rounded apices (Fig. 7.3). Partial coverings of iron coats (Fig. 7.2) and dispersed red points probably related to glands (Fig. 7.4) characterise the preservation.

Remarks. The main distinctive characteristic of X. spinifolia is the bipinnate frond carrying successive triangular pinnules with acute apices, in regularly spaced ribbon-like and opposite segments (akin to a sequence of Vs along the penultimate pinnae).

Such leaf morphology became the target of different taxonomic interpretations depending on the areas from which they were collected and the proposed age of the deposits. Based on Australian materials, Tenison-Woods (1883) assigned them to Trichomanides spinifolium and highlighted the membranaceous nature of the expanded rachis (winged) and the great number of short and linear pinnules with acute apices, a morphology that fits well with the morphotype illustrated herein. Also working with Australian materials, Etheridge (1892) reassigned this kind of foliage to Stenopteris Saporta, ignoring the non-bifurcated nature of this genus and its cuticular features (Carruthers, 1872). Seward (1903) described similar morphologies from South Africa (Cape Colony) and assigned them to Trichomanites (Sphenopteris) elongata.

Following Walton (1940) and taking into account the affiliation of the austral forms to Corystospermaceae, Frenguelli (1943) considered that the weak and dubiously one-veined pinnules were different from those of Stenopteris. He also alluded to the multiple striations and coriaceous nature of the Stenopteris leaves, which probably cloak the venation. The presence of a distinct and prominent rachis, the cuticle features and the age (Jurassic) assigned to the aforementioned forms, support Frenguelli's (1943) configuration of the Xylopteris genus. However, the author questioned the real winged and bipinnate disposition of the fronds and the basal constriction in the pinnules and suggested that they probably represent “lobulations”.

Other researchers rejected the pinnate condition of this form and considered that the generic assignment to Stenopteris was invalid (Menéndez, 1951; Petriella, 1979; Baldoni, 1980; Stipanicic et al., 1996). Stipanicic et al. (1996) proposed, in the emended diagnosis of Xylopteris, a bipinnate to tripinnate condition of the leaves and the presence of other kinds of venation rather than the exclusively one-veined ones. However, X. spinifolia was absent from their materials.

Another problem in the discussion of the present morphotype involves the description of X. spinifolia provided by Frenguelli (1943: fig. 32) for material from the Cacheuta levels from Argentina, where the two superimposed pinnae differ from those of T. spinifolium from Australia (Tenison-Woods, 1883) in presenting round apices and an unclear bifurcate condition. The Argentine form seems to be more similar to “Dicroidium” elongatum forma remotipinnulium from South Africa (Anderson and Anderson, 1983). The same occurs with Baldoni's X. spinifolia (1980: lam. I, figs. 3, 5-6), which presents lobulated and rounded pinnules as well as closely spaced pinnules, a common morphology in X. densifolia.

To add a new component to such taxonomic controversy, Jones and De Jersey (1947) proposed a preferred affinity of the previously described Trichomanites spinifolia Etheridge (in Jack and Etheridge, 1892; p. 367, pl. 18, fig. 8) with the genus Stenopteris and of the species T. elongata var. spinifolia Shirley 1898 with S. spinifolia. The authors considered that the bipinnate fronds of S. spinifolia, with one to six one-veined linear pinnules inserted at acute angles and thin cuticles covering the subsidiary cells, proved different from those of Stenopteris elongata.

Bomfleur and Kerp (2010) later confirmed the cuticle features implied by Jones and De Jersey (1947) in their emended diagnosis of Dicroidium spinifolium (Tenison-Woods) Anderson and Anderson, which included X. spinifolia Frenguelli as a synonym. Furthermore, according to Bomfleur and Kerp (2010), the presence of small to mediumsized fronds with linear and narrow triangular pinnules presenting particular cuticle features distinguished D. spinifolia from D. elongatum. Anderson and Anderson (1970, 2003) also reported that the “X. spinifolia physiognomy” must be considered within the diversity of Dicroidium.

Conversely, other researchers supported the existence of X. spinifolia (e.g., Hill et al., 1965; Gould, 1967; Retallack, 1977; Petriella, 1983) even though the emended diagnosis of Xylopteris proposed by Stipanicic and Bonetti (in Stipanicic et al., 1996) was not formally published. More recently, Pattemore and Rigby (2005: fig. 5A) adopted the same procedure for the plant fossils from the Ipswich Basin. Data in Table 1 confirm that X. spinifolia, X. elongata and X. rigida prove to be the species with broader distribution in Gondwana.

Table 2

- Biostratigraphic distribution of Xylopteris and its allied forms in Gondwana. 1 refer to the species identified in southern Brazil and the asterisks to the new specific affinity herein proposed.

Figure 6.

1–4. Xylopteris densifolia. 1, General aspect (ULVG 9743); 2, Drawing of the specimen; 3, Detail of the pinnules and its round-acute apices (arrow); 4, Part of pinnae and the ribbon-like pinnules and striated rachis, covered by mud and iron coats. Scale bar= 5 mm.

Figure 7.

1–4. Xylopteris spinifolia. 1, General aspect of the straight bipinnate frond (ULVG 11007), showing the many sectors covered by dark brown iron coats; 2, Distinct parts of the frond (ULVG 11062) probably attesting different taphonomic phases: initial one with light colours produced by the mud cover (left) and the advanced ones with the remains covered by dark brown iron coats (lower right); 3, Closer view evidencing the V disposition of the ultimate pinnules (ULVG 11062); 4, Detail of pinnules (ULVG 9531) showing small red points (glands or papillae?). Scale bar= 5 mm.

Associated Umkomasia fructification

Associated with frond remains of X. spinifolia in a unique sample (ULVG 1169B), a record of an Umkomasia megasporophyll was detected although no organic connections were preserved (Fig. 8.1–3). These remains consist of three rounded and curved pedicellate (or stalked) cupules of which the opposite lateral ones are inserted on the main axis. One of the seeds exhibits a pronounced bifid and outwardly curved micropyle (Fig. 8.3, arrow). Those features approximate this seed-bearing organ to that of U. macleanii Thomas, from the Molteno Formation (Anderson and Anderson, 2003).

Thomas (1933), when studying the fossil leaves of the Molteno beds, was the first one to propose a connection between Umkomasia and Xylopteris. According to Thomas (1933), the Umkomasia female organs could be related to both Dicroidium and Stenopteris densifolia; the latter transferred to Xylopteris densifolia by Frenguelli (1943).

Umkomasia is very common in Gondwana deposits despite its lower proportion in relation to “Dicroidium-like” fronds (Holmes, 1982, 1987; Klavins et al., 2002; Anderson and Anderson, 2003; Holmes and Anderson, 2005). The genus is also rare in Laurasia, where it has been associated with Thinnfeldia fronds (Kirchner and Müller, 1992; Zan et al., 2008). Due to the controversies herein observed about the segregation of Xylopteris with respect to Dicroidium, the great majority of researchers consider Umkomasia to be more closely related to Dicroidium (Lacey, 1976; Retallack, 1977, 1980; Pole and Raine, 1994; Klavins et al., 2002; Anderson and Anderson, 2003; Holmes and Anderson, 2005; Pattemore and Rigby, 2005). Petriella (1980), working with Argentine fossils, and Axsmith et al. (2000), in Antarctica, found a form -U. uniramia- that suggests to be in organic connection with shoots of D. odontopteroides. These occurrences are, so far, the only record of a more reliable association between Umkomasia and Dicroidium. Yet, it is questioned by some researchers (Anderson and Anderson, 2003; Artabe and Brea, 2003; Holmes and Anderson, 2005). The recent identification of Umkomasia in both Permian and Upper Cretaceous levels attests to a broader time span in its fossil record and introduces a new component regarding its exclusive relation with Dicroidium-Thinnfeldia allied forms, as discussed by Shi et al. (2016), and their usage in age inferences. According to those researchers, the fact that the seeds originate directly from the axis suggests a possible relation between corystosperms and Ginkgo.

Figure 8.

1, Xylopteris spinifolia frond (ULVG 1169A) and an associated terminal part of the megasporophyll from Umkomasia macleanii Thomas; 2, X. spinifolia (detail) and the one-veined pinnules (arrow); 3, Umkomasia macleanii (ULVG 1169B), with two opposite and curved pedicellate cupules and the pronounced micropyle in one of the seeds (arrow); 4, Xylopteris remotipinnulia (ULVG 11177), general aspect of the bipinnate frond showing pinnules and pinnae partially covered by iron coats and asymmetric weakly developed pinnules (arrow); 5, X. remotipinnulia (ULVG 11178), part of the frond highlighting the curved penultimate segments and short pinnules (arrow). Scale bar= 5 mm.

Considering the absence of organic connections and the common occurrence of Umkomasia remains together with other corystosperms along the Triassic outcrops from Brazil, a preferable relationship with Xylopteris or Dicroidium is difficult to establish. Anyhow, forms of Xylopteris are the element most commonly associated with Umkomasia in the PTM beds, followed by Dicroidium lancifolium and Zuberia feistmantelii.

Xylopteris remotipinnulia (Anderson and Anderson) Ottone, 2006 (see remarks)
(Figures 8. 4–5)

1927 Stenopteris densifolia Du Toit (pg. 365, only text. fig. 14a and 14b).

1947 Stenopteris spinifolia Jones and De Jersey (only text-figure 17, p. 28).

1951 Xylopteris elongata (Carruthers) Frenguelli, Menéndez, lam. XV, fig. 10

1983 Dicroidium elongatum forma remotipinnulium Anderson and Anderson (pl. 48, figs. 26–27, 30–31; pl. 63, fig. 29; pl. 84, figs. 8–9).

Material. ULVG 9210A; 9408; 9747B; 9747D; 1231; 11415; 1439.

Type locality and horizon. Rare form, nearly exclusively from N6, lowermost level from the middle part of the mudstone profile.

Description. Incomplete medium sized (5 cm large) bipinnate fronds. Main rachis longitudinally striated (1.8 mm wide) and bifurcated at low angles (15–20°), thus giving place to opposite/alternate elongate secondary pinnae also inserted in acute angles (25°-35°). Secondary pinnae (1.2 mm wide) regularly spaced (4–5 mm) and rarely curved. Few and short triangular pinnules (1 mm wide and 4–6 mm long) with round to acute apices, simple, rarely paired, disposed in the proximal part of the ultimate pinnae and forming an asymmetrical general pattern. A characteristic main vein is absent but fine ridges cover the pinnules. A great number of scattered red points cover the whole leaf, including the rachis.

Remarks. The bipinnate frond with elongate secondary pinnae herein assigned to X. remotipinnulia carries few paired or solitary pinnules. This form is rare in the assemblages but boasts a broad distribution throughout the Triassic of Gondwana. Regarding morphology (see synonym), X. remotipinnulia is comparable with fronds described in Australia (Jones and De Jersey, 1947), Argentina (Menéndez, 1951) and from the Upper Karroo beds in South Africa (Du Toit, 1927). The latter location presents remains sharing a great number of morphological features with the herein described form. Anderson and Anderson (1983) included them in a variety of D. elongatum (D. elongatum form remotipinnulium).

Ottone (2006), working with analogous forms from the Rincón Blanco Group floras from Argentina, observed the affinity of some South African materials with Xyloperis and suggested the new combination Xylopteris remotipinnulia, mainly based on the linear nature of the pinnae and its few and short pinnules. However, Ottone (2006) neither justified the changes proposed in relation to the original affinity with Dicroidium nor explained the reason for raising the “form” remotipinnulium to a specific rank. Nevertheless, the new generic and specific epithet together with the diagnosis presented by Ottone (2006) are consistent with the general architecture of the fronds (bipinnate and univeined pinnules). Both the leaf impressions described herein and those from Ottone (2006) expanded the paleogeographic occurrence of the species to Western Gondwana.

RESULTS AND DISCUSSION

The detailed analysis performed along the Passo das Tropas Member outcrop revealed that the Xylopteris related forms, although a minor component in the plant assemblage, boast a cyclic recurrence that is maintained until the uppermost levels, where other components became rare. Such behaviour and the dominant presence of the species X. elongata and X. spinifolia (75%) in the PTM profile, like in other Gondwana areas, favour the confirmation of the constant arrival of leaf remains to the depositional areas and that distinct species of Xylopteris were part of the near-lake vegetation. Furthermore, the xerophytic nature observed in their dissected leaves suggest seasonal climatic dry conditions that seem to become more critical in the final stages of the lake deposition. The associated occurrence of narrow-leaved Sphenobaiera supports those inferences (Retallack, 1977; Barboni and Dutra, 2015).

During the Middle-Upper Triassic, the southern Brazilian areas were located under or at 30oS of palaeolatitude (Golonka, 2007) and subjected to an intense tectonism that generated small isolated basins similar to those in northwestern Argentina (Zerfass et al., 2004, 2005). The Pangea conformation at this time is also affected by a megamonsoonal atmospheric circulation (Parrish, 1993; Kutzbach and Gallimore, 1989; Preto et al., 2010) which, in the Northern Hemisphere, leaded to extensive evaporite deposits alternated with siliciclast deltaic successions (Wilson et al., 1994; Mutti and Weisser, 1995).

Comparable rich plant assemblages from the late Ladinian of Argentina were described and suggest a past location within the boundary of dry/wet climatic belts (Spalletti et al., 1999, 2003; Mancuso and Marsicano, 2008; Colombi and Parrish, 2008; Preto et al., 2010). This context appears modified during the Carnian which, in southern Brazil, coincides with the deposition of massive mudstones of the uppermost Santa Maria Formation (Alemoa Member), rich in tetrapod faunas. The signals of seasonal dryness become rarer or are only substantiated by discontinuous levels of calcium carbonate concretions. Such phenomena seem to reflect the global “Carnian Pluvial Event”, which is still under discussion in terms of effect pondering (Preto et al., 2010). On the other hand, the diversified and abundant leaf remains from the lenticular lacustrine deposits of the lower PTM (e.g. Dicroidium, Taeniopteris and Ginkgophytes) support a short stable condition of the water bodies also confirmed by the homogeneous mudstone succession containing the plant fossils.

In Australia, the hardly fertile sandy soils formed over the braided river deposits validate the relative stability of the base level after the emergence of the Lachman Fold belt, considered important in the appearance of the xerophytic woodlands (Retallack, 1977; Holmes and Anderson, 2005). Moreover, the megaflora from the shale lenses of the Hawkesbury fluvial sandstones (Middle-Late Triassic) indicate that the sandy floodplains and channel bars were colonised by a xerophyte woodland of complex structure (the “Dicroidietum zuberi and odontopteroideum xylopterosum” community of Retallack, 1977). In that assemblage, an analogous condition to that of the floristic composition of the PTM can be observed in how Xylopteris is associated with Dicroidium and with narrow-leaf Ginkgoaceae (Guerra-Sommer and Cazzulo-Klepzig, 2000; Corrêa, 2014; Barboni and Dutra, 2015).

In terms of age, the correlation of data performed on distinct communities from the Triassic of Argentina confirmed Xylopteris as a minor component in those shrubby or herbaceous communities. Its scarce appearance in the Ladinian of lower latitude basins is expanded in the Late Triassic to include the southern areas of Patagonia and South Africa (floral zones BNP to DLM from Spalletti et al., 1999; Morel et al., 2003; Artabe et al., 1998, 2007c). The data compiled and depicted in Table 1 reveals that, after such event, types with the Xylopteris morphology retract to lower latitudes and become extinct towards the end of the Rhaetian. Additionally, Table 2 demonstrates that the first record of the genus, proposed by Retallack (1977) for the Anisian, is not confirmed. Based on those data as well, the presence of X. remotipinnulia in the PTM proves useful for restricting the age of the herein studied assemblage to the Carnian.

CONCLUSIONS

The assignment to the genus Xylopteris proposed herein for some frond types detected in the Dicroidium Flora from southernmost Brazil is based on the presence of slender pinnate (X. elongata, X. rigida, X, densifolia and X. rotundipinnulia nov. comb. gen., nov. sp.) and bipinnate fronds (X. spinifolia and X. remotipinnulia). Such taxonomic proposition is also supported by the rigid nature of the pinnae, which sometimes present expanded or winged rachis, and the generally one-veined pinnules.

Our survey regarding the fossil record of Xylopteris confirms its restriction to the Middle-Late Triassic deposits (Ladianian-Rhaetian) and its value for establishing chrono-correlations in the continental Triassic. The fossil record additionally suggests that the western part of Gondwana, which contains both the oldest and the youngest records of the genus, was marked by maintaining favourable conditions for its life and survival. The assemblage studied in the present work also serves as corroboration for the proper usage of Xylopteris-like physiognomy as a tool, which taking into account taxon-independent morphological features can be resorted to establishing paleogeographic and palaeoclimatic inferences. Additionally, because of the elongate and coriaceous pinnae and pinnules of the species, more common in the younger levels of the Dicroidium dominated successions (Retallack, 1977; Bomfleur and Kerp, 2010: fig. 7), Xylopteris can also use for substantiating age inferences for Gondwana assemblages.

The shale deposits containing plant fossils, generated in abandoned and more stable depressions, seem to result from the soil fixing promoted by the permanent growth of riparian vegetation together with open woodlands, which developed over the drained soils of the exposed sandbars. The occurrence of well-preserved pinnae, associated with little carapaces of Spinicaudata and fish scales, confirmed the near-lake growth of the vegetation and the parautochtonous preservation of its remains.

Considering all such aspects, the presence of Xylopteris in southern Brazil extends the age of the Santa Maria Formation -previously considered Anisian-Ladinian- to the beginning of the Carnian, as well the paleogeographic distribution of the genus. The proposed age is supported by the joint occurrence of diversified Ginkgophyta and after establishing comparisons with the Molteno (South Africa), Ipswich (Australia) and northern Argentina taphofloras.