Revision and biostratigraphic implications of Thore Halle’s Permian plant fossils from the Falkland (Malvinas) Islands Stephen McLoughlin1* , Rose Prevec2,3‡ , Bárbara Cariglino4 & Marc Philippe5 1Department of Palaeobiology, Swedish Museum of Natural History, Box 50007, S-104 05 Stockholm, Sweden 2Department of Earth Sciences, Albany Museum, 40 Somerset Street, Makhanda, Eastern Cape, 6139 South Africa 3Department of Botany, Rhodes University, P.O. Box 94, Makhanda, Eastern Cape, 6140 South Africa 4CONICET – Area de Paleobotánica y Palinología, Museo Argentino de Ciencias Naturales “B. Rivadavia”, Av. Angel Gallardo 470, C1405DJR, Buenos Aires, Argentina 5Univ Lyon, Université Claude Bernard Lyon 1, CNRS, ENTPE, UMR 5023 - LEHNA, F-69622 Villeurbanne Cedex, France Received 19 December 2023, Accepted 19 June 2024 INTRODUCTION Thore Gustaf Halle’s (1911) original paper on the Permian flora of the Falkland Islands, recognized by Argentina as the ‘Islas Malvinas’, was the fourth in the series of geologi- cal publications on the results of the Swedish Expedition to Patagonia and Tierra del Fuego, 1907–1909, led by botanist and explorer Dr Carl Skottsberg. Halle (at age 23) participated in Skottsberg’s expedition, collected the fossil material, carried out additional extensive geological investigations on the Falkland (Malvinas) Islands, and published the results for his doctoral dissertation at Uppsala University under the supervision of Professor Arvid-Gustaf Högbom. The geology of the Falkland (Malvinas) Islands had received little attention between Charles Darwin’s visits of 1833 and 1834 (Armstrong 1992) and the Swedish South Polar Expedition of 1901–1902. Andersson (1907) and Halle (1911) provided the first serious attempts to clarify the geology of the islands and provided age constraints on the stratigraphic succession using marine and continental fossils. Halle spent a little under four months on the Islands (26 October 1907 to 12 February 1908). He described elements of the Devonian, Permian and Quaternary fossil floras, of which only the Permian material is revised here. Nathorst (1904) published possibly the earliest report of pre-Quaternary fossil plants from the Falklands, based on material collected by Johan Gunnar Andersson from the southern end of East Falkland (Soledad) Island during the 1901–1902 Swedish South Polar Expedition. His identifica- tion of Asterocalamites was later revised to Phyllotheca (Nathorst 1906), suggesting a Permian age for these strata. Halle (1908, 1911) confirmed a Permian age for these plant-bearing beds and documented two species of sphenopsids, five Glossopteris leaf types, a glossopterid scale leaf, spine-leafed conifer branches, strap-shaped leaves attributed to Desmiophyllum, and two species of permineralized wood. Several subsequent studies have mentioned, described or illustrated Permian plant remains from the Falkland (Malvinas) Islands (Table 1). Key studies include those of Seward & Walton (1923) who described four morphotypes of Glossopteris and a new species of permineralized wood (Dadoxylon bakeri) from George Island, Speedwell Island and North Arm, near the southern tip of East Falkland (Isla Soledad). Baker (1924) recorded additional silicified ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 53 *,‡Authors for correspondence. E-mail: *steve.mcloughlin@nrm.se / ‡r.prevec@am.org.za Palaeontologia africana 58: 53–93 — ISSN 2410-4418 [Palaeontol. afr.] Online only Permanently archived on the 5th of December 2024 at the University of the Witwatersrand, Johannesburg, South Africa The article is permanently archived at: https://hdl.handle.net/10539/43014 The Permian fossil plant assemblages from the Lafonia Group on the Falkland (Malvinas) Islands collected by Thore Gustav Halle on the 1907–1909 Swedish Expedition to Patagonia and Tierra del Fuego are re-described and their systematic placement revised. Two species of sphenophytes based on foliage and one on axial remains are recognized. Eight morphotypes of Glossopteris are differentiated using more rigorously defined criteria than Halle’s original character sets. A single species each of cordaitaleans and conifers are recognized. The absence of ferns and lycophytes may indicate significant taphonomic filters on the composition of the plant assemblages. Re-assessment of the characters of the fossil woods and their nomenclatural and taxonomic problems suggests that only a single species is recognizable in the assemblage. Several of the wood and leaf species bear evidence of fungal degradation along with a broad array of arthropod herbivory and oviposition damage that add to the diversity of biotic interactions documented in the middle–high southern latitude Glossopterid Biome of the late Paleozoic. The ages of the various fossiliferous units on the Falkland (Malvinas) Islands remain equivocal, but similarities with chronostratigraphically constrained leaf assemblages from the Karoo Basin, South Africa, suggest that the Bay of Harbours Formation (uppermost unit of the Lafonia Group) is referable to the upper Guadalupian to lowermost Lopingian. Keywords: Glossopteris, sphenopsids, conifers, Lopingian, Lafonia Group, Karoo Basin. Palaeontologia africana 2024. ©2024 Stephen McLoughlin, Rose Prevec, Bárbara Cariglino & Marc Philippe. This is an open-access article published under the Creative Commons Attribution 4.0 Unported License (CC BY4.0). To view a copy of the license, please visit http://creativecommons.org/licenses/by/4.0/. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. The article is permanently archived at: https://hdl.handle.net/10539/43014 https://orcid.org/0000-0001-6723-239X https://orcid.org/0000-0001-6674-1547 https://orcid.org/0000-0002-4346-3502 https://orcid.org/0000-0002-4658-617X http://creativecommons.org/licenses/by/4.0/ mailto:r.prevec@am.org.za https://hdl.handle.net/10539/43014 mailto:steve.mcloughlin@nrm.se https://hdl.handle.net/10539/43014 54 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Table 1. Principal published records of Permian plant fossils (original published identifications) from the Falkland Islands with their location and stratigraphic assignment. Author/year Taxon Location Stratigraphic unit Nathorst (1906) Phyllotheca sp. Seal Cove, Bull Cove, Speedwell Island Egg Harbour Mbr, Bay of Harbours Fm Halle (1911) (1) Phyllotheca australis Brongniart (1) Speedwell Island + most See Materials and Methods (leaves and associated ribbed axes) Permian plant localities except Darwin Harbour and Dos Lomas (2) Phyllotheca deliquescens (Goeppert) (2) Same localities as (1) + Dos Lomas. (3) Glossopteris browniana Brongniart (3) South of Dos Lomas on Falkland Sound, south of Goose Green, and Speedwell Island (4) Glossopteris indica Schimper (4) Darwin Harbour, Goose Green, Bodie Creek Head, Dos Lomas, Low Bay, Speedwell Island (5) Glossopteris angustifolia Brongniart (5) Southernmost point of Speedwell Island and at Bodie Creek Head (6) Glossopteris damudica Feistmantel (6) Dos Lomas (7) Gangamopteris cyclopteroides Feistmantel (7) South of Dos Lomas var. major (Feistmantel) (8) Scale-frond of Glossopteris sp. (?) (8) Dos Lomas (9) Coniferous branches (9) Speedwell Island, north of the Halfway Cove settlement (10) Desmiophyllum sp. (10) South coast of East Falkland opposite Speedwell Island (11) Dadoxylon lafoniense (11, 12) See Table 2 (12) Dadoxylon cf. angustum Felix Seward & Walton (1923) (1) Equisetaceous stems (possibly (1) Unspecified Phyllotheca sp.) (2) Equisetaceous stems cf. Phyllotheca (2) Speedwell Island, George Island, australis Brongniart and P. deliquescens North Arm (Bay of Harbours), Dos Lomas (Goeppert) (3) Equisetaceous stems cf. Neocalamites (3) Cygnet Harbour, Egg Harbour carrerei (Zeiller) (4) Glossopteris indica Schimper (4) Speedwell Island, George Island, North Arm (Bay of Harbours), Dos Lomas, Goose Green (5) Glossopteris indica Schimper, cf. var. (5) North Arm (Bay of Harbours) wilsoni Seward (6) Glossopteris indica Schimper, cf. var. (6) George Island and Bay of Harbours decipiens Feistmantel (7) Glossopteris browniana Brongniart (7) George Island, North Arm (Bay of Harbours), Goose Green (8) Dadoxylon bakeri (8) Walker Creek and Fanny Cove, on the southern side of Choiseul Sound, East Falkland Trewin et al. (2002) (1) Plant debris (1) Unspecified (1) Cantera and Saladero members, Brenton Loch Formation (2) Glossopteris sp. (2) Unspecified (2) Praltos Member, Bay of Harbours Formation Stone (2010) Silicified wood Bleaker Island, SE of Lafonia, Praltos Member, Bay of East Falkland Harbours Formation Simões et al. (2012) Glossopteris sp. cf. G. communis Feistmantel Rorys Creek, East Falkland Saladero Member, Brenton Loch Formation Stone & Rushton Calamites-type plant stems; Glossopteris spp. Lafonia (possibly George and Probably Egg Harbour Mbr, (reviewing collections Speedwell islands) Bay of Harbours Fm of Baker 1924) Speedwell Island, George Island, Cygnet Harbour and Dos Lomas = Egg Harbour Mbr, Bay of Harbours Fm. North Arm = Praltos Mbr, Bay of Harbours Fm. Walker Creek and Fanny Cove = Saladero Mbr, Brenton Loch Fm. Goose Green = Possibly Saladero Mbr or Praltos Mbr wood at Walker Creek. Aldiss & Edwards (1999) summa- rized the occurrences of fossil plants on the Falkland (Malvinas) Islands and noted anecdotal reports of permineralized wood up to 40 cm in diameter and 4 m long in the southern part of East Falkland (Isla Soledad) and its adjacent islands. Trewin et al. (2002) noted that Glossopteris leaf fossils were abundant in certain beds of the Bay of Harbours Formation (Lopingian), exposed primarily in the central and southern portions of East Falkland (Isla Soledad) and its offshore islands, but no specimens were illustrated. Stone (2010) illustrated an unidentified permineralized wood with prominent growth rings from the same unit. Simões et al. (2012) illustrated two leaf specimens attributed to Glossopteris sp. cf. G. communis from the Saladero Member (Capitanian) of the Brenton Loch Formation at Rorys Creek, East Falkland (Isla Soledad). Stone (2012) mentioned that the Brenton Loch and Bay of Harbours formations host abundant plant fossils including Glossopteris, but no additional material was described or illustrated. The only other non-aquatic body fossil recovered from the Glossopteris-bearing beds is a single insect wing illus- trated by Halle (1911) and tentatively assigned to the Palaeodictyoptera, but later interpreted to derive from a damselfly by Tillyard (1928) or a basal group of damselfly- like Odonatoptera by Nel et al. (2012). The three fossil insect wings reported to have been recovered from the Permian of the islands by Schlüter (2003) apparently all represent the single specimen collected by Halle (1911). A sparse fauna of non-marine bivalves was documented by Simões et al. (2012) from the Saladero Member, Brenton Loch Formation. Various non-marine invertebrate grazing traces, trackways, and burrows, together with fish trails (Undichna) have also been reported from this unit (Trewin 2000; Trewin et al. 2002; Stone 2016). Seward & Walton (1923) wrote, in evaluating Dr Herbert Arthur Baker’s collection of Permian fossil plants from the Falkland (Malvinas) Islands, ‘A thoroughly satisfactory determination and specific separation of the numerous and, nearly always, incomplete leaves obtained from the different Permo-Carboniferous localities in the Falkland Islands is, we feel, a hopeless task’. We are less pessimistic about the demarcation of Permian leaf forms than those authors and we apply rigorous morphological and anatomical criteria to distinguish multiple fossil species and morphotypes from that region. Since Halle’s identifications of Permian plant impres- sions were supported mainly by sketches with limited or inaccurate morphological details, we aim to illustrate the principal components of the flora using colour photogra- phy. In doing so, we present additional observations on the fossils and confirm or revise various identifications presented in Halle’s (1911) report. We provide compari- sons primarily with fossils from South Africa and Argen- tina – the regions that were geographically closest to the Falklands (Malvinas) Block during the Permian (Trewin et al. 2002). Included in this analysis are the first Permian plant–insect interactions to be recognized on leaf and stem impressions from the Islands. We also provide a new anatomical analysis of the permineralized wood using modern transmitted-light microscopy techniques. We use these results to assess the taxonomic placement of the woods and to document the first record of permineralized fungus- and oribatid mite-interactions with probable glossopterid axes from the region. GEOLOGICAL SETTING Following the work of Adie (1952), the Falkland (Malvinas) Islands (and adjacent continental crust of the Falklands/Malvinas Plateau and Maurice Ewing Bank) were generally interpreted to have been contiguous with the southeastern margin of the Karoo Basin before Gondwanan breakup and rotation of the Falklands block in the Mesozoic (Mitchell et al. 1986; Marshall 1994; Aldiss & Edwards 1999; Stone 2016; and references therein). More recent work by Ramos et al. (2017, 2019), Lovecchio et al. (2019, 2020) and Bonifacio et al. (2023) has favoured a more southerly (south of the Cape Fold Belt) position of the Falklands (Malvinas) (and a fixed relationship to Patagonia) prior to continental breakup. Regardless of which reconstruction is correct, the Falkland (Malvinas) Islands would have been located at about 60–65°S during the Lopingian, with a cool- to cold-temperate climate (Torsvik & Cocks 2012; Prevec et al. 2022). Faulting and sediment accumulation in the Falklands (Malvinas) region may even have been linked to the western extent of a broad band of pan-Gondwanan rifting that formed in the Permian and provided structural weaknesses that were later reactivated during Gondwanan breakup (Harrowfield et al. 2005). Several studies have documented the geology of the Falkland (Malvinas) Islands in recent decades (Aldiss & Edwards 1999; Trewin et al. 2002; Simões et al. 2012; Stone 2016). Carboniferous–Permian strata are confined primar- ily to the southern part of East Falkland (Isla Soledad) and adjacent islands (Fig. 1), together with a few isolated expo- sures near Byron Sound on the west coast and along Falkland Sound on the east coast of West Falkland (Isla Gran Malvina) (Greenway 1972). Permian strata also probably occur in the subsurface of the South Falkland (Malvinas) Basin and in the Malvinas Basin to the south and west of the Falkland (Malvinas) Islands, respectively (Stone 2016; Lovecchio et al. 2019). Carboniferous and Permian strata on the Falkland (Malvinas) Islands are assigned to the Lafonia Group (Aldiss & Edwards 1999), which is synonymous with the Grupo Isla Soledad (East Falkland Group) of Borrello (1963, 1972). Distinct lithic, feldspathic and volcanigenic components of sedimentary strata in the Lafonia Group indicate that an orogenic belt (possibly the Cape Fold Belt) supplied sediments to the area (Aldiss & Edwards 1999). The Lafonia Group is up to 7 km thick and is generally divided into five formations (Fig. 2): The Bluff Cove For- mation, Fitzroy Tillite, Port Sussex Formation, Brenton Loch Formation and Bay of Harbours Formation, in ascending stratigraphic order. Several of these have been further subdivided into members (Aldiss & Edwards 1999), which Trewin et al. (2002) compared in detail to the lithostratigraphic succession in the Karoo Basin. The Bluff Cove Formation consists primarily of fine- ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 55 grained sandstones with interbedded siltstones and shales that lack identifiable body fossils but contain small indeterminate trace fossils and some disseminated unidentifiable plant debris (Aldiss & Edwards 1999). Sedimentary facies relationships suggest that the unit was deposited in shallow marine or deltaic settings prior to the Cisuralian (early Permian) glaciation of the region (Aldiss & Edwards 1999; Trewin et al. 2002). The Fitzroy Tillite consists predominantly of clast-poor diamictite with a sandy mudstone matrix. Deposition occurred in subaqueous to continental glacial settings (Trewin et al. 2002). Examples of the trace fossil Umfolozia Savage provide the only evidence of biological activity in the unit, and the formation is tentatively assigned a Stephanian–Asselian or Sakmarian age (Isbell et al. 2003), based primarily on correlation with the Dwyka Group in South Africa where it is relatively common in Dwyka and lowermost Ecca Group deposits (Anderson 1981; Aldiss & Edwards 1999). The Port Sussex Formation (Frakes & Crowell 1967) is divided into three subunits: the Hells Kitchen, Black Rock and Shepherds Brook members in ascending strati- graphic order (Aldiss & Edwards 1999). The Port Sussex Formation contains post-glacial, shallow-water, lami- nated, fine- to medium-grained sandstones, dark grey and purple, siliceous, locally bentonitic, marine mud- stones, and heterolithic facies containing sparse horizon- tal burrows (Planolites). Frakes & Crowell (1967) noted the presence of plant fossils in a claystone near the top of the Shepherds Brook Member at Port Sussex, and Aldiss & Edwards (1999) also reported plant debris from the lower half of this unit in the same area but no remains were identifiable. Age constraints on this unit are poor but lithological correlations with the Karoo Basin suggest an Artinskian–Roadian age (Aldiss & Edwards 1999; Trewin et al. 2002). The Brenton Loch Formation is also divided into three subunits: the Terra Motas, Cantera and Saladero members (Aldiss & Edwards 1999). Collectively, this formation is composed of variable associations of sandstones, mud- dominated rhythmites, and massive mudstones and siltstones. A rich ichnofauna dominated by arthropod trackways is preserved in the Cantera Member and, together with sedimentological features ranging from wave and current-rippled tops of sandstone beds to turbiditic facies, suggest deposition in shallow bay to marine slope settings possibly representing a prograding delta front (Trewin et al. 2002). Poorly preserved fossils of 56 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Figure 1. Geological map of the Falkland (Malvinas) Islands highlighting the distribution of Permian units and localities from which Halle’s plant fossils were recovered (adapted from Halle 1911; Aldiss & Edwards 1999; Trewin 2002; Simões et al. 2012). sphenopsid stems and glossopterid leaves occur in sand- stones of the Brenton Loch Formation on the east coast of West Falkland (Isla Gran Malvina) (Scasso & Mendia 1985), and Simões et al. (2012) illustrated Glossopteris leaves from the Saladero Member at Rorys Creek on East Falk- land (Isla Soledad). Based on its stratigraphic position, the unit has been inferred to be predominantly Guadalupian (middle Permian) in age (Simões et al. 2012). The Bay of Harbours Formation has a gradational lower contact with the Brenton Loch Formation and consists primarily of coarse- to fine-grained sandstones, interca- lated with siltstones and minor mudstones. The formation ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 57 Figure 2. Simplified composite stratigraphic log of the Permian sedimentary succession on the Falkland (Malvinas) Islands (after Trewin et al., 2002). C. = Carboniferous; G.–As. = Gzhelian–Asselian; A.–R. = Artinskian–Roadian; BC = Bluff Cove Formation; FT = Fitzroy Tillite; PS = Port Sussex Formation. is generally divided into the mudstone-dominated but coarsening-upward Praltos Member and an overlying succession of channel sandstones, intercalated massive sandstone–mudstone couplets, and sparse conglomerate beds constituting the Egg Harbour Member. The forma- tion is interpreted to have been deposited in a mix of delta slope to lower delta plain environments from subaqueous to exposed floodplain settings with weak palaeosol devel- opment (Aldiss & Edwards 1999; Trewin et al. 2002). This formation represents the uppermost unit of the Permian succession in the onshore part of Falkland (Malvinas) Islands and is overlain only by minor Cenozoic deposits. Bioturbation is common in siltstones of the Bay of Harbours Formation, but discrete identifiable traces are sparse (Aldiss & Edwards 1999). Plant fossils are relatively common in this unit and the vast majority of Halle’s Perm- ian fossil plant specimens probably derive from this formation (see Materials and Methods). The age of the Bay of Harbours Formation is poorly resolved but has been considered Guadalupian–Lopingian based on its strati- graphic position and tentative correlations with the Karoo Basin (Trewin et al. 2002; this study). MATERIALS AND METHODS Over 300 specimens were examined in this study; all are registered in the collections of the Department of Palaeo- biology at the Swedish Museum of Natural History (Naturhistoriska riksmuseet: prefixed NRM), Stockholm, under the numbers: NRM S047751–S047955 (all impres- sion fossils); and S04005, S004024, S004034, S004038, S004040, S004042, S004047, S004048, and S081825 (thin sections of permineralized wood). When more than one Glossopteris leaf was present on a slab, a numerical epithet was assigned, e.g. S0047795–01. Halle (1911) collected plant impressions and permineralized woods from 13 principal localities and received, from local residents, some donations of fossils from unspecified sources. Placed in the context of current stratigraphic resolution of the Lafonia Group (Trewin et al. 2002; Stone 2016), we assign the lithostratigraphic sources of his sampled beds as follows: 1. Localities on east coast of Speedwell Island near Half- way Cove: Egg Harbour Member, Bay of Harbours Formation. 2. Southwest coast of Lafonia near Black Point (opposite Speedwell Island): Egg Harbour Member, Bay of Harbours Formation. 3. North Arm, Bay of Harbours: Praltos Member, Bay of Harbours Formation. 4. Dos Lomas, Kelp Harbour, west coast of East Falkland (Isla Soledad): Egg Harbour Member, Bay of Harbours Formation. 5. Orqueta Park area, c. 10 km west of Goose Green: Praltos Member, Bay of Harbours Formation. 6. Bodie Creek House area, c. 7 km southwest of Goose Green: Praltos Member, Bay of Harbours Formation. 7. Bodie Creek Inlet (both banks), c. 2 km south of Goose Green: either Saladero Member, Brenton Loch Forma- tion or Praltos Member, Bay of Harbours Formation. 8. Darwin settlement area: either Saladero Member, Brenton Loch Formation or Praltos Member, Bay of Harbours Formation. 9. East of Tranquilidad (ruins): probably Praltos Member, Bay of Harbours Formation, less likely Saladero Mem- ber, Brenton Loch Formation. 10. The Back Beach, Low Bay: Praltos Member, Bay of Harbours Formation. 11. Arrow Harbour (not indicated on Halle’s map, but mentioned in his text): Saladero Member, Brenton Loch Formation. 12. Between Port Sussex and Cantera (not indicated on Halle’s map, but mentioned in his text): Saladero Member, Brenton Loch Formation. 13. Seal Cove: Praltos Member, Bay of Harbours Forma- tion. Plant macrofossils were photographed using either a Canon EOS 40D digital camera or an Olympus SZX10 stereomicroscope equipped with a Sony Exmoor E3CMOS digital camera. In some cases, multiple photo- graphs were taken at different focal planes and a compos- ite image was obtained using the auto-stacking and alignment features available in Adobe Photoshop. ImageJ was used to measure Glossopteris leaves. Only leaves that showed minimal distortion, with a portion of medial lamina with midrib, leaf margin and a clear venation pattern visible, were included in the analysis. All leaves are preserved as impressions, and preservation is gener- ally poor, with very few complete leaves, many leaves torn and distorted prior to and during preservation, and vena- tion of highly variable quality. Descriptions focused on the medial portion of the leaves, as defined by Prevec (2011). This represents the largest section of the leaf (and is, there- fore, most commonly preserved in fragmented leaves), and venation characteristics are most consistent in this region. In the base and apex, venation varies markedly over a short distance, making standardization of measure- ments, such as vein angle and mesh characteristics, more challenging to define. As per the schema used by Prevec (2011), three zones were delimited within each half of the medial portion of a leaf, to either side of the midrib: (1) proximal, for the region close to the midrib; (2) mid- laminar, for the region halfway between midrib and margin and (3) marginal, for the zone close to the leaf margin. Changes in leaf characters between these regions are key to the identification of morphospecies, and here the degree of change between vein angle and mesh width across the zones was quantified using two new size- independant ratios Change in Vein Angle (CVA) and Change in Mesh Width (CMW). CVA values define vein course. CVA1 represents the ratio of the angle between the midrib and the secondary veins at the midpoint of the lamina-half versus the equivalent angle at secondary vein emergence from the midrib. In other words, this ratio is a measure of the change in secondary vein angle from adjacent to the midrib to the mid-laminar region, as defined by Prevec (2011). CVA2 represents the ratio of the angle between the midrib and secondary veins at the leaf margin versus the equivalent angle at the midpoint of the lamina half. This quantifies the degree of curvature of the veins from the mid-laminar portion of the leaf to the 58 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 margin. These ratios, therefore, reflect the strength or degree of arching of the veins across the three leaf sectors of the lamina-half; a value of 1 reflects an absolutely straight vein path, and the higher the value, the sharper the curve. CMW values allow for rapid assessment of the change in mesh width across the lamina. CMW1 reflects the change in mesh width from those adjacent to the midrib to the mid-laminar part of the lamina-half, in other words the ratio of proximal mesh width to mid-laminar mesh width as defined by Prevec (2011). A value greater than 1 reflects a decrease in mesh width from those adja- cent to the midrib to the mid-laminar region. Similarly, CMW2 reflects the change in mesh width from those in the mid-laminar region to those at the margin of the leaf. The most readily preserved and consistently observable features were used in the comparative analysis, viz. leaf width, leaf shape, form and width of midrib, mesh shape and size across the lamina, venation course across the lamina, tangential angle of the venation adjacent to the midrib, in the mid-lamina, and at the margin, and marginal vein density. Where possible, size class of the leaf morphotypes has been allocated, as per the schemes of Raunkiaer (1934) and Webb (1959). The designation ‘n ’ refers to the number of leaves assessed for any particular character, even though multiple measurements may have been taken per leaf; averages are of all measurements taken for a particular character, across multiple leaves. Photo-micrographs of thin sections of fossil wood were taken using an Olympus BX51 microscope with an Olym- pus DP71 digital camera. Nine sets of thin-sections of fossil wood were identified in Halle’s collection at NRM (Table 2). Among these samples, S004040 was identified as the specimen described by Halle (1911) as Dadoxylon cf. angustum; specimen S081825 was identified as the first sample described by Halle (1911) as Dadoxylon lafoniense, with well-preserved pith and poorly preserved second- ary xylem, and S004042 as the second specimen described by Halle (1911) as Dadoxylon lafoniense, lacking pith but with well-preserved secondary xylem. The six other samples are only alluded to in Halle’s contribution (1911). Museum labels provide locality information that is partly missing from Halle’s (1911) paper (see Table 2). A key goal of this study is to clarify the definitions of the wood taxa described by Halle as Dadoxylon lafoniense and D. cf. angustum, hence, we focused primarily on anatomi- cal information relating to the secondary xylem. Wood descriptions follow the template proposed by Boura et al. (2021). Biogenic traces on leaves are attributed to the Damage Type (DT) categories of Labandeira et al. (2007) where possible. Since the higher-level taxonomy of plants is in a state of flux, we organize the identified taxa only below order- level in the systematic section. Synonymy lists encompass only fossils previously illustrated from the Falkland (Malvinas) Islands. Some fossils were unidentifiable or could only be assigned to a genus; a full catalogue of the registered fossils and their geographic and stratigraphic sources is provided in Supplementary Table 1. SYSTEMATIC PALAEOBOTANY Order Equisetales Genus Paracalamites Zalessky, 1932 Paracalamites sp., Fig. 3A–D, K, L 1906 Phyllotheca sp.; Nathorst, pp. 74–76, pl. 7, figs 1–4. 1906 Phyllotheca or Schizoneura sp.; Nathorst, p. 75, pl. 7, fig. 7. 1911 Phyllotheca australis Brongniart; Halle, pp. 51–54, pl. 1, figs 12–16 (non figs 17–19). 1911 Phyllotheca cf. deliquescens Goeppert; Halle, pp. 54–55, pl. 1, fig. 26 (non figs 21–25). 1923 Equisetaceous stems cf. Phyllotheca australis Brongniart; Seward & Walton, pp. 318–319, text-fig. 2, pl. 19, figs 3, 4, 6, pl. 21, fig. 16. 1923 Equisetaceous stems cf. Neocalamites carrerei (Zeiller); Seward & Walton, pp. 319–321, text-fig. 3, pl. 20, figs 8, 10, 12. Material . NRM S047751 (Fig. 3A), S047752a-01, S047752a-02 (Fig. 3K), S047752b, S047753 (Fig. 3B), S047754-03, S047756, S047757, S047758, S047760a, S047760b, S047761a, S047761b, S047767 (Fig. 3C, D), S047768, S047769, S047770, S047792a, S047792b, S047798, S047823, S047825, S047837, S047858, S047885, S047914, S047915, S047925, S047926, S047927, S047952-01 (Fig. 3L). Description. Impressions of jointed axes, the largest exceeding 150 mm long and up to 20 mm wide (e.g. S047753, S047823), with longitudinal ribbing varying from pronounced (Fig. 3A) to indistinct (Fig. 3K), in almost every case crossing the nodes oppositely. Nodes thick, in some cases wider than internode width. Internodes generally longer than wide. Branching, where present, occurring from the node, where circular scars also indicate sporadic branch detachment. Remarks. Paracalamites accommodates isolated, large, articulated, leafless sphenopsid stems and pith casts, differing from the similar northern hemisphere Calamites in that the ribs remain mostly opposite and do not alter- nate as they cross the node. As noted by Rigby (1966), multiple sphenopsid genera, such as Phyllotheca, Schizo- neura and Raniganjia, among others, produce stems of indistinguishable morphology, with generic attribution dependent on the morphology and arrangement of attached leaves. He suggested using the fossil-genus Paracalamites for leafless sphenopsid stems found widely ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 59 Table 2. Locality details for Permian fossil wood specimens (and associ- ated thin sections) from the Falkland Islands studied by Halle (1911). Specimens Slides Locality on label Name on label Material: figs in pl. 4 of Halle (1911) S004005 3 Bodie Creek House Dadoxylon S004024 3 Tranquilidad Dadoxylon S004034 3 Tranquilidad Dadoxylon S004038 3 Port Sussex None S004040 3 Bodie Creek House D. cf. angustum 8 and 9 S004042 3 Tranquilidad D. lafoniense 5, 6 and 7 S004047 3 None stated None S004048 3 None stated None S081825 4 Arrow Harbour D. lafoniense 1, 2, 3 and 4 60 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Figure 3. Permian sphenophytes from the Falkland (Malvinas) Islands. A–D, K, Paracalamites sp.; E, H, J, Phyllotheca deliquescens (Göppert) Schmalhausen, 1879; F, G, Phyllotheca australis Brongniart, 1828; I, L, possible strobilus attached to Paracalamites sp. All scale bars = 10 mm, except G and J (= 1 mm). A, S047751; B, S047753; C, D, S047767; E, S047776-01; F, S047754-01; G, S047771-02; H, S047932; I, L, S047952-01; J, S047776-02; K, S047752a-02. across Gondwana in upper Paleozoic strata. Accordingly, Rigby (1966) erected Paracalamites australis and assigned to this species various specimens from the Falklands (Malvinas) that had been assigned to Phyllotheca australis and P. cf. deliquescens by Halle (1911, pl. 1, figs 12–14 and 26, respectively). We agree with Rigby’s diagnosis of Paracalamites and, therefore, place all of the specimens represented by artic- ulated leafless stems with (usually) opposite ribbing at the nodes under this taxon (Fig. 3A–D, K, L). However, as Rigby’s diagnosis for Paracalamites australis indicated that branch tubercles were absent, we leave the specific assignation open, since several specimens from the Falklands (Malvinas) are poorly preserved and lack evidence of branching (e.g. S047792) or branch tubercles adjacent to the nodes (Fig. 3D). Seward & Walton (1923) described plant material recov- ered from the Falklands (Malvinas) collected by H.A. Baker in 1922, among which were equisetaceous stems that differed from the others in their narrower, more crowded ribbing, and which they assigned to cf. Neocala- mites carrerei (Zeiller) Halle, 1908. Given the variation in ribbing expression evident in axes within Halle’s (1911) collection, we argue that these forms with slightly more dense ribbing are also better attributable to Paracalamites. Possible strobilus attached to Paracalamites sp., Fig. 3I, L Material. NMR S047952-01 (Fig. 3I, L) Description. Articulated stems with ribs that are opposite across the node. One of the stems (Fig. 3L: marked with a yellow dot) incorporates three internodes 1.7 cm long and 0.8 cm wide; nodes wider than internode width, with circular marks indicating branch bases. A stalk 0.9 cm long and 0.2 cm wide bearing an oval structure, 1.1 cm long and 0.6 cm wide, emerges from one of the nodes, indicating the possible presence of a strobilus. Remarks. Compared to the northern hemisphere, there are few records of sphenophytes with attached strobili in upper Paleozoic strata of Gondwana. Apart from a few reports of sphenophyllalean cones (Cúneo et al. 1993; Prevec et al. 2010, 2012; Cariglino & Prevec 2018), only strobili attributed to Schizoneura gondwanensis from the uppermost Permian of the Sydney Basin, Australia (Etheridge 1903), and Neocalamites sp. from the Permian La Golondrina Formation, Santa Cruz Province, Argen- tina (Escapa & Cúneo 2006), have been described to date. Neocalamites is typically employed for Mesozoic leafless ribbed, rarely branched axes, with continuous ribbing through the well-defined nodes, and linear leaves with acute apices and free bases. Such Mesozoic axes are other- wise indistinguishable from Permian Paracalamites (Brown 1977). Argentinean strobili affiliated with Neocalamites are stalked, compact, bractless and consist of peltate sporangiophores. Although the Falkland (Malvinas) specimen is too poorly preserved to resolve fine details, such as the leaves and sporangiophores, the presence of a stalked, apparently bractless strobilus, reveals similarities to Neocalamites. However, owing to the absence of leaves, the fragmentary nature, and poor preservational quality of this specimen, we provisionally retain this material under Paracalamites, axes of which are abundant else- where in the same beds. Genus Phyllotheca Brongniart, 1828 Phyllotheca australis Brongniart, 1828, Fig. 3F, G 1906 Phyllotheca sp.; Nathorst, pp. 74–76, pl. 7, figs 5, 6. 1911 Phyllotheca australis Brongniart; Halle, pp. 51–54, pl. 1, figs 17–19 (non figs 12–16). Material. NMR S047754-01 (Fig. 3F), S047766, S047771-01 (Fig. 3G), S047771-02, S047772, S047773. Description. Impressions of leaf-bearing stems or isolated leaf whorls. Largest axis (S047773) c. 70 mm long and 8 mm wide, incorporating five internodes of equal length (1.2 cm long); leaves c. 5 mm long emerging from the nodes. Isolated whorls incorporate 3–11 leaves (Fig. 3F, G), basally fused, with indistinct median vein and acute to slightly rounded apices (e.g. S047754-01, S047772). Remarks. Phyllotheca was established by Brongniart (1828) for upper Paleozoic plants resembling extant Equisetum. Brongniart noted that the articulated stems could be unbranched or monopodially branched, and grooves and ridges did not alternate through the nodes; leaves were basally fused into sheaths to varying degrees, but remained free for most of their extent, tapering distally. Much later, Townrow (1955) emended Brongniart’s diagnosis to include cuticle-bearing and fertile material, and described P. australis, the type species, in detail to better delimit and define the genus. Townrow (1955) defined Phyllotheca as plants with stems up to 20 mm wide, bearing branches and leafy shoots. The leaves had a single, commonly indistinct midrib, and were basally fused in a sheath. The cuticle revealed epidermal cells and stomata similar to those of other equisetitaleans. The asso- ciated (but not attached) cones consisted of alternating whorls of sterile leaf-like bracts and sporangiophores, abaxially bearing from few to numerous sporangia. When referring specifically to P. australis, he stated there were 18–23 leaves per whorl, usually 20 (Townrow 1955), although numbers ranging from 8–33 have been mentioned by other authors for this species (Maheshwari 1967; Rayner 1992). A very similar species, P. indica Bunbury 1861, has been an issue of discussion since many authors have argued that this species should be synonymized with P. australis (e.g. Seward, 1898; Du Toit 1932; Maheshwari 1967). Townrow (1955) also stated that the characters used to separate these species were probably inadequate; however, he regarded the slightly longer leaves in P. indica and the fact that they taper only very close to the apex as sufficient to distinguish them. We do not agree that these characters are adequately diagnostic or consistent and, therefore, also consider P. indica to be a junior synonym of P. australis. Although no cuticle can be recovered from the Falkland (Malvinas) specimens, and there is no evidence of fertile structures closely associated with the leaf-bearing speci- mens, those sphenophytes from the Falklands (Malvinas) that have linear leaves with an indistinct median vein, ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 61 either borne on an axis or preserved in detached leaf whorls, are consistent in morphology with the wide- spread P. australis. Phyllotheca deliquescens (Göppert) Schmalhausen, 1879 Fig. 3E, H, J 1911 Phyllotheca cf. deliquescens Göppert; Halle, pp. 54–55, pl. 1, figs 21–25 (non fig. 26). Material. NMR S047776-01 (Fig. 3E), S047776-02 (Fig. 3J), S047776-03, S047776-04, S047932 (Fig. 3H). Description. Small, fragmentary articulated stems bear- ing c. 12–14 leaves per whorl (estimated from S047776-02). Leaves 8–11 mm long, 0.5 mm wide, fused at the base to form a funnel-shaped, 1-mm-long sheath; free segments longer than internode. Median vein visible (S047776-01). Remarks. Phyllotheca deliquescens is generally character- ized by robust articulated stems, with well-defined ribs and whorls of branches, typically arising from the nodes. Leaf bases are fused to form a well-developed, funnel- shaped sheath, splitting into linear free segments with a median vein. Although it was first described from Siberia, Townrow (1955) considered P. deliquescens to be part of the ‘australis group’, in which he grouped the southern hemisphere Permian species of Phyllotheca commonly associated with Glossopteris. The main difference from Phyllotheca australis is that the free, slender leaf segments radiating from the sheath of P. deliquescens are traversed by a more promi- nent median vein, and tend to be upwardly curved near their apex (Arber 1905; Halle 1911). We agree and consider these characters sufficiently robust to separate P. deliquescens from P. australis. The Falkland (Malvinas) specimens of P. deliquescens (Fig. 3E, H, J) also tend to have longer free parts of their leaves compared to P. australis. Phyllotheca lawleyensis Anderson & Anderson also has strong similarities in its slender, univeined leaves united into a small basal cup, but the individual leaves on that species reach in excess of 35 mm long, and they are possibly associated with Giridia-like strobili (Anderson & Anderson 1985, pl. 41). Order Glossopteridales (= Dictyopteridiales) Genus Glossopteris Brongniart 1828 ex Brongniart 1831 General remarks on the classification of Glossopteris. Halle (1911) recognized four species of Glossopteris and one of Gangamopteris in the small collections he recovered from multiple fossil localities in the Falkland (Malvinas) Islands. We found that it is not possible to unequivocally assign the Falkland (Malvinas) glossopterid leaves to established species (particularly so where cuticular details are required for confident identification), hence we have described the various forms under eight informal morphotypes (MF), some of which are distinctive and easily differentiated, whereas others encompass a signifi- cant degree of variation. There is also some overlap in individual character ranges between morphotypes. The overlapping ranges of quantitative variables and subtle qualitative features used to distinguish morphotypes forms a data matrix that is difficult to represent graphi- cally with any clarity. Additionally, we recognize that, given the small number of leaves considered here, it is impossible to capture the full range of morphological variation within a taxon. With more comprehensive sampling, it is possible that some of these morphotypes will constitute end-members of more broadly circum- scribed species. Sampling bias is also heightened in small collections, where collectors may specifically target specimens that have an unusual appearance, in an attempt to represent the greatest range of leaf morphotypes. Ironically, in the case of Glossopteris, this approach does not work towards accurate representation of diversity, as there is great diffi- culty in demarcating fossil-species within the the mor- phological continua for which this group of plants is renowned. For example, it is possible that MF01 and MF02, distinguished primarily by the angle of the secondary venation to the midrib, lie on a morphological spectrum of a single species. But other morphotypes, such as MF05, are clearly distinct, even where based on a single leaf. Large collections can assist greatly in recognizing morphological continua, contributing to a far more robust taxonomic framework. In the Falkland (Malvinas) collec- tions, morphospecies MF03, MF05 and MF08 were repre- sented by only single specimens. This, in combination with the overall poor quality of preservation of the leaves and the absence of attached or associated fertile organs, has limited the precision of our identifications and, conse- quently, our biostratigraphic interpretations. Further exploration of the islands, and collection of larger numbers of (better preserved) plants, within a more detailed strati- graphic and sedimentological context, is critical to better understand the palaeoenvironmental, biostratigraphic and biogeographic implications of these floras, and partic- ularly how the Falklands (Malvinas) Block related to Argentina and southern Africa during the Permian. Quantifying Glossopteris leaves at the level of detail presented here is vexatious and time-consuming, but a necessary step if this field of palaeobotany is to progress. Even if consensus is never reached on the names used for different fossil-species, detailed descriptions at least facilitate accurate comparisons of these plants across Gondwana. We compare the glossopterid leaf morpho- types described below, with fossil leaves documented primarily from southern Africa and Argentina – the two regions likely juxtaposed with the Falklands (Malvinas) Block during the Permian. Glossopteris morphotype MF01, Fig. 4A–D; Suppl. Fig. 1D 1911 Glossopteris indica Schimper; Halle, pl. 2, fig. 4, 4a. 1923 Glossopteris browniana Brongniart; Seward & Walton, p. 325, pl. 20, fig. 11. 1923 Glossopteris indica Schimper; Seward & Walton, pp. 321–322, pl. 21, figs 14, 15, 17. Material. S047780 (identified by Halle as Glossopteris indica in NRM database: Fig. 4 A, B); S047867, S047840 (Fig. 4C, D; Suppl. Fig. 1D) (identified by Halle as Glosso- pteris sp.); S047953-1; S047953-2 (G. sp. cf. G. browniana). Localities. Bodie Creek Head Site 1; Low Bay; Speedwell Island (southern point). 62 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 63 Figure 4. A, B, Glossopteris morphotype MF01; C–H, Glossopteris morphotype MF04. A, central portion of leaf; B, details of venation; S047780; C, proximal portion of leaf tentatively assigned to this taxon; D, details of venation S047840; E, central portion of leaf; F, details of venation showing relatively straight course of veins in outer lamina; S047775; G, central portion of relatively broad-meshed leaf; H, details of venation; S047835-01. Scale bars = 10 mm for A, C, E, G; 1 mm for B, D, F, H. Description. Oblong to elongate-elliptical or lanceolate microphyll-sized leaves (Fig. 4A); estimated leaf area 810–1198 mm2 (n = 3), with acute apex; 52 mm long (n = 1), 14–(15.8)–16.2 mm wide (n = 4). Midrib indistinct, weak parallel veins evident in some specimens, unknown if persistent to apex; medial width 0.38–(0.58)–0.88 mm (n = 4). In medial portion of leaf, lateral veins arise sub-parallel to midrib, arch across lamina at a narrow angle of 13.8°–(22.3°)–31.4° (n = 4), flexing slightly from proximal one third of lamina to 30°–(47.7°)–62.1° (n = 4), intercepting the margin at 48°–(57.4°)–66° (n = 4) to midrib (Fig. 4B; Suppl. Fig. 1D). In medial portion of leaf, veins anasto- mose 2–5 times and bifurcate 3–5 times from midrib to margin; meshes proximal to midrib are elongate-elliptical to elongate-trullate; becoming linear, elongate-polygonal to elongate-elliptical in mid-laminar region and towards the margin. Proximal mesh width is 0.28–(0.35)–0.42 mm (n = 3), becoming slightly broader in the mid-laminar region 0.39–(0.43)–0.56 (n = 4), narrowing again at margin 0.29–(0.35)–0.46 mm (n = 4). Marginal meshes are similar in width to proximal meshes, with marginal vein density of 12–(15.6)–18 veins per 5 mm (n = 4). Remarks. Four leaves were identified and analysed. These are long, narrow microphyll-sized, with steeply angled, gently curving venation. The curvature from midrib to mid-lamina is sharper, with a CVA1 of 1.7–(2.3)–2.5 (n = 4), becoming gentler from mid-lamina to margin, with a CVA2 of 1–(1.3)–1.7 (n = 4). Meshes are narrow and elongate, of fairly consistent width across lamina, but broadening slightly in the mid-laminar region as reflected by a CMW1 of 0.73–(0.84)–0.91 (n = 3), and CMW2 of 1.04–(1.29)–1.63 (n = 4). The midrib is poorly defined. Comparisons with South African and Argentinean leaves. Long, narrow Glossopteris leaves with subparallel mar- gins, a midrib reaching the apex and gently arching secondary venation with fairly sparse anastomoses and elongate meshes are common in the middle to upper Permian of South Africa (Anderson & Anderson 1985; Bordy & Prevec 2008; Prevec et al. 2009, 2010; Botha et al. 2020) and Argentina (e.g. Menéndez 1966), and are strongly associated with various species of the ovuliferous fructification Lidgettonia. Those leaves most similar to MF01 include upper Permian examples from the Karoo Basin within the concept of ‘Lidgettonia lidgettonioides’ proposed by Anderson & Anderson (1985, pl. 127, figs 1, 4, 6, p. 302, Mooi River; pl. 128, figs 13–14, p. 303, Bulwer), Glossopteris loskopensis (Anderson & Anderson 1985, pl. 132, p. 307, Loskop quarry); Clouston Farm morphotypes C2b, c (Prevec et al. 2009, pl. IV, figs 6–8; pl. VII, C2b, c); Wapadsberg Pass morphotypes WP1 and WP2 (Prevec et al. 2010; fig. 9, E–L, p. 399) and leaves figured from Kwa Yaya (Bordy & Prevec 2008; pl. II, figs D, G, H). MF01 also has similarities with several leaves from older, probably middle Permian fossil assemblages, from Somkhele Mine (Bordy & Prevec 2008, pl. 3, fig. A, p. 450); Lawley (‘Lidgettonia lawleyensis’ leaves, Anderson & Anderson 1985, pl. 124, p. 299) and Hammanskraal (‘Otto- karia hammankraalensis’ leaves of Anderson & Anderson 1985, pl. 60, fig. 9, p. 235; Glossopteris leptoneura of Kovács- Endrödy 1991, pl. 5.3, figs 1–4, pp. 24–26). Specimens from the Bonete Formation that were assigned to Glossopteris angustifolia by Menéndez (1966, pl. 2, figs. 5–7; and previously by Harrington 1934) were described as long, narrow leaves with subparallel margins, a midrib reaching the apex and smoothly arched secondary vena- tion, rarely anastomosing and forming elongated meshes. This description corresponds well with MF01 in its slightly curved secondary venation, as opposed to MF02, which has markedly straight venation. MF01 also resembles some leaves attributed to Glosso- pteris browniana Brongniart from the Bonete Formation (i.e. FCEN 6501, 6510) and the Carapacha Formation (i.e. GHUNLPam 3134) (Menéndez 1966; Melchor & Césari 1997). These leaves have secondary veins that initiate almost parallel to the midrib then curve gently at c. 20°, to intercept the margin at c. 55°, with few anastomoses and forming fine, elongated meshes. Of particular interest are some leaves formerly assigned to G. indica from the Bajo de Véliz, Arroyo Totoral, Tasa Cuna and the Rio Genoa formations, which were later reclassified as Glossopteris wilsonii (Seward) Archangelsky et al., 1981 (Archangelsky et al. 1981; Césari 2014). Glosso- pteris wilsonii includes long, narrow leaves, with sub- parallel secondary venation, and very few (if any) anastomoses. In their revision, Archangelsky et al. (1981) proposed that Glossopteris indica leaves from South Africa, Antarctica and the Falklands (Malvinas) be transferred to G. wilsonii, including Halle’s (1911, pl. 2, figs 4, 4a) specimen from the Bodie Creek Head site (S047780). We consider that a more rigorous morphometric analysis of the Argentinian and South African material is needed before the Falkland (Malvinas) specimens can be confi- dently attributed to this species. Glossopteris morphotype MF02, Fig. 5A–J; Suppl. Fig. 1A, C 1911 Glossopteris indica Schimper; Halle, pp. 56–58, pl. 2, fig. 3. 1911 Glossopteris angustifolia Brongniart; Halle, pp. 58–60, pl. 3, figs 1–4. 1923 Glossopteris indica Schimper; Seward & Walton, pp. 321– 322, pl. 19, figs 5, 7, pl. 20, fig. 9. Material. S047779-01 (Fig. 5C) S047877, S047894 (Fig. 5G, H; Suppl. Fig. 1A), S047950a, b (all identified by Halle as Glossopteris indica in NRM database); S047782-01, S047783-01, S047783-02, S047795 (Fig. 5I, J; Suppl. Fig. 1C), S047801 (Fig. 5A, D) (all identified by Halle as G. angustifolia); S047779-02, S047782-02, S047797-02, S047832, S047861-01 (Fig. 5B, F), S047870, S047902, S047907 (all identified by Halle as Glossopteris sp.); S047800 (Fig. 5E) (identified by Halle as cf. Gangamopteris sp.); S047862 (identified by Halle as Glossopteris browniana). Localities. Bodie Creek/Bodie Creek Head Site 1; Speed- well Island Southern Point; between Darwin and Goose Green; Brenton Loch (Site 1); Speedwell Island (south of settlement); Low Bay; Darwin (south of Goose Green). Description. Leaves oblong, elongate-elliptical, or elon- gate-lanceolate, of large microphyll size with estimated leaf area of 823–(1528)–2000 mm2 (n = 3); 52 mm long 64 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 65 Figure 5. A–J, Glossopteris morphotype MF02. A, complete leaf; D, enlargement of venation; S047801. B, distal portion of leaf, F, details of venation; S047861-01; C, distal portion of leaf; S047779-01. E, distal portion of leaf with notched apex; S047800. G, details of venation; H, whole leaf; S047894; I, atypical narrow leaf; J, details of venation; S047795. Scale bars = 10 mm for A–E, H, I; 1 mm for F, G, J. (n = 1), 14–(15.8)–16.2 mm wide (n = 4) (Fig. 5A, H). Apex acute with rounded or pointed tip (Fig. 5B, C, H), base narrowly cuneate. Midrib well-defined but narrow, flat, becoming indistinct and very narrow apically (Fig. 5E); individual veins are visible in some specimens; width ranges from 1.2 mm at base (n = 1), 0.57–(0.89)–1.4 mm in medial part of leaf (n = 9), to 0.17–(0.4)–0.8 mm at apex (n = 6). In medial portion of leaf, lateral veins arise from midrib and traverse the lamina at a fairly constant angle between 27.6° and 60.4° (av. 41.3°) (n = 10), following a notably straight path to intersect the margin at 34.4°– (49.5°)–67.2° to the midrib (n = 9) (Fig. 5D, F, G; Suppl. Fig. 1A, C). In the medial portion of leaf, veins typically anastomose 3–4 times and bifurcate three times from midrib to margin; most meshes adjacent to midrib are strongly linear, but some are elongate-elliptical to narrowly trullate, becoming consistently and strongly linear in mid-laminar region and towards the margin. Proximal mesh width is 0.23–(0.37)–0.57 mm (n = 10), becoming progressively narrower across the mid-laminar region 0.19–(0.33)–0.47 mm (n = 9) to the margin 0.19–(0.26)–0.35 mm (n = 10). Marginal vein density variable, 14–(16.9)–21 mm per 5 mm of margin (n = 10). Remarks. The ten measured leaves are distinguished by their long and narrow lamina with a narrow midrib that is well-defined in the basal and medial portions of the leaf, becoming less distinct apically. The venation is distin- guished from other forms by curving slightly from midrib to mid-lamina with a CVA1 of 1.3–(1.9)–2 (n = 9), but is markedly straight from mid-lamina to margin, with a CVA2 of 1.1–(1.2)–1.7 (n = 10). This produces meshes that are narrow and of fairly consistent width across the lamina, with a CMW1 of 0.94–(1.18)–1.79 (n = 10), and a CMW2 of 1.01–(1.2)–1.48 (n = 10). A few leaves are especially slender (e.g. S047862) but they are here deemed to be variants (possibly early-season leaves) of MF02 based on their similar venation course. Leaf S047795 (Fig. 5I, J) is atypical, with a particularly broad, flat midrib, and disproportion- ately narrow lamina. MF01 and MF02 are very similar, and the latter may represent a morphological variant of the former. The main distinctions between the two, are the greater vein angles (on average) and more intense curva- ture of the veins from midrib to mid-lamina in MF01, and the more prominent midrib in MF02. Comparison with South African and Argentinean leaves. Leaves from several South African upper Permian floras figured by Anderson & Anderson (1985), resemble MF02, including some placed within ‘Lidgettonia africana’ (their pl. 114, p. 289; Lidgetton), ‘L. mooiriverensis’ (pl. 117, figs 8–11, p. 292, pl. 118, p. 293, Mooi River), ‘L. lidget- tonioides’ (pl. 126, figs 10, 11–13, p. 301, Mooi River; pl. 130, figs 1, 4, p. 305, Bergville), although these leaves are larger than MF02 and are elongate-obovate. Leaves associated with ‘L. inhluzanensis’ of Anderson & Anderson (1985; pl. 120, p. 295, Inhluzani; pl. 121, figs 18–31, p. 296, Bulwer; pl. 122, figs 15–17, p. 297, Loskop quarry) are also similar, but the venation in MF02 is denser, and the leaves larger. The anomalous leaf in Fig. 5I, J, is somewhat reminiscent of the glossopterid ovuliferous fructification Dictyo- pteridium flabellatum known from the upper Permian of South Africa (Anderson & Anderson 1985). This species of seed-bearing organ has a very high length to width ratio, and is particularly leaf-like in appearance, with promi- nent and steeply inclined venation visible across both the seed-bearing and opposing surfaces. The specimen figured here is, unfortunately, too poorly preserved to make a conclusive identification, although some of the tubercle-like features present on the surface of the organ are reminiscent of seed scars. Prevec et al. (2022, p. 7, fig. 4g,) figured undescribed leaves from the middle Permian flora at the Onder Karoo locality that bear a close resemblance to MF02. These narrowly oblanceolate leaves have a well-defined midrib and linear venation following a steep, straight path across the lamina, and may be conspecific with the Falklands (Malvinas) specimens. Although much larger than MF02, some long, slender leaves from the (probably late middle to early late Perm- ian) Lawley flora have a similar venation pattern (e.g. ‘Arberia allweyensis’ leaf figured by Anderson & Anderson 1985, pl. 110, fig. 3, p. 285). Similar leaves from the lower Permian, include those placed within the concept of ‘Hirsutum dutoitides’ proposed by Anderson & Anderson (1985) (a fructification now assigned to Gladiopomum dutoitides; Adendorff et al. 2002), from Vereeniging. These leaves are generally larger than those of MF02, and have a more pronounced midrib, particularly in the proximal parts of the leaf, but they share the very narrowly lanceo- late shape and the fine venation that follows a very steep, straight path from midrib to margin (e.g. Anderson & Anderson 1985, pl. 75, figs 1–6, 14, p. 250, pl. 76, p. 251; Adendorff et al. 2002, p. 6, figs 3–5). Almost straight, linear secondary venation is evident in a fragmentary leaf attributed to Glossopteris decipiens Feistmantel from the Bonete Formation (i.e. BA-PB 3504); however, the angle at which the secondary veins depart from the midrib is steeper (c. 10–15°), and the midrib is not well-defined (Menéndez 1966). Straight secondary vena- tion is also evident in a Glossopteris argentina Archangelsky leaf, figured from the Carapacha Formation (GHUNLPam 3130; Melchor & Césari 1997), but this is the only specimen with secondary venation similar to MF02. Other speci- mens assigned to G. argentina (Melchor & Césari 1997, fig. 9.1–7) have more curved secondary venation, similar to the type specimens from the La Golondrina Formation (Archangelsky 1958). Glossopteris morphotype MF03, Fig. 6F, G; Suppl. Fig. 1F Material. S047799 (identified by Halle as cf. Gangamopteris sp. in the NRM database. Locality. Bodie Creek Head Site 1. Description. Leaf incomplete, but probably oblong to elongate-elliptical, of large microphyll size; at least 23 mm wide (Fig. 6F). Midrib narrow, poorly preserved, details not distinct, medial width 1.26 mm. The meshes in the first row adjacent to the midrib are oriented at a steep angle to midrib 12.5°–(15.4°)–18.9° (n = 1); veins curve slightly to mid-lamina 23.3°–(27.2°)–29.9° (n = 1), and follow a straight course to the margin at 26.8°–(27.4°)–27.7° to the midrib (n = 1) (Fig. 6G; Suppl. Fig. 1F). In the median part of the leaf, veins anastomose and bifurcate multiple times 66 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 67 Figure 6. A–D, Glossopteris morphotype MF07; E, Glossopteris morphotype MF08; F, G, Glossopteris morphotype MF03; A, half lamina; B, details of venation; S047835-02. C, proximal lamina; D, details of venation and simple scalloped margin-feeding damage (arrowed); S047777-01. E, proximal portion of lamina with prominent midrib and elliptical oviposition scars (arrowed); S047785. F, central portion of lamina; G, details of venation; S047799. H, glossopterid scale leaf (cf. Eretmonia sp.) showing details of petiole and triangular lamina; S047786. All scale bars = 10 mm. across the lamina, forming an elongate-trullate row of meshes proximally, becoming linear, elongate-polygonal to elongate-elliptical in the mid-laminar region, and parallel towards the margin. Proximal mesh widths are 0.25–(0.26)–0.27 mm (n = 1), becoming slightly narrower in mid-laminar region 0.15–(0.2)–0.27 mm (n = 1) and near the margins 0.13–(0.14)–0.15 (n = 1), with a marginal vein density of 10 veins per 5 mm (n = 1), reflecting the strongly oblique angle of vein intersection with the margin. Remarks. One medial to sub-medial portion of a leaf corresponding to this morphotype was identified in the collection. This large microphyll-classed leaf is character- ized by venation that is steeply inclined with veins inter- secting the margin at a narrowly oblique angle. Veins are distinctive in curving slightly after leaving the midrib, with a CVA1 of 1.8, but then following a straight path to the margin (CVA2 of 1). The vein angle also decreases towards the apex, which is unusual, since vein angle is generally consistent in the medial portion of glossopterid leaves, with a decrease in angle only occurring close to the apex, converging on the leaf tip. Comparison with South African and Argentinean leaves. This leaf fragment is reminiscent of Palaeovittaria with its very steeply angled veins that follow an unusually straight path across the lamina, e.g. Palaeovittaria kurtzii from Vereeniging (Anderson & Anderson 1985, p. 229, pl. 54, figs 4, 6, 7, 9,). Further, the venation appears to radiate slightly from the base of the specimen, a feature typical of this taxon. However, upon close inspection, the veins in MF03 anastomose regularly, the meshes are elongate- elliptical to linear, and the midrib appears to have been disrupted/distorted, which might account for the peculiar decrease in vein angle away from the base. A poorly preserved leaf on the underside of slab BA-PB 7708 from the Bonete Formation (Menéndez 1966, unfig- ured; B.C. pers. obs.) has notably low-angled secondary venation emerging from a poorly defined midrib, and follows a straight path to the margin, in a similar manner to that of MF03. This is the only example of a leaf similar to morphotype MF03 recognized in the Argentinean fossil floras. Glossopteris morphotype MF04, Fig. 4E–H; Suppl. Fig. 1G, J 1911 Glossopteris browniana Brongniart; Halle, pp. 55–56, pl. 1, fig. 28, 28a, pl. 1, fig. 27. (non pl. 1, fig. 29) 1923 Glossopteris indica Schimper; Seward & Walton, pp. 321–322, pl. 21, fig. 18. 2012 Glossopteris sp. cf. G. communis Feistmantel; Simões et al., p. 544, fig. 4A, B. Material. S047775 (Fig. 4E, F; Suppl. Fig. 1G), S047781-01, S047835-01 (Fig. 4G, H; Suppl. Fig. 1J), S047836-01 (all identified by Halle as Glossopteris browniana in NRM data- base), S047838, S047884-01 (identified by Halle as Glosso- pteris sp.). Localities. Darwin Harbour Site 1; Dos Lomos Site 1; North Arm Site 1. Description. Large microphyll-class leaf (Fig. 4C, E, G), 12.4–(18.25)–24 mm wide (n = 4), oblanceolate, oblong or obovate, with cuneate base, apex acute. Midrib well defined, flat, with individual veins visible in some cases, persistent to apex; basal width 1.6 mm (n = 1), medial width 0.6–(0.81)–1.24 mm (n = 5), narrowing to 0.36 mm (n = 1). First row of meshes adjacent to midrib broader and at a steep (narrow–moderate) angle to midrib 20.3°–(36.1°)–58.5° (n = 4), veins arching strongly in second row of meshes, with medial meshes at a wider angle 61.6°–(64.4°)–69.4° (n = 4), curving gently to follow an almost straight path at 63.1°–(71.3°)–80.1° (n = 3) to margin (Fig. 4D, F, H; Suppl. Figs 1G, J). In the medial portion of the leaf, veins anastomose and bifurcate 3–4 times (n = 4) across lamina; first row of meshes rhombic, trullate, polygonal, broadly falcate, with a width of 0.5–(0.61)–0.74 mm (n = 3) becoming falcate, trullate to linear in second row, with medial meshes becoming ellip- tical, to linear, 0.37–(0.55)–0.8 (n = 6), and becoming progressively narrower towards margin 0.25–(0.4)–0.72 (n = 5). Marginal meshes similar in shape to mid-laminar meshes, with a marginal vein density of 9–(15)–19 veins per 5 mm (n = 5). Remarks. Six leaves were identified and analysed. This large microphyll-class leaf type is distinguished by a well- defined midrib from base to apex, with veins arising from the midrib at a narrow angle, then arching (CVA1 of 1.5–(1.9)–2.4, n = 4) and following an almost straight path to the margin (CVA2 of 1–(1.1)–1.2, n = 3) at a moderate angle. Mesh widths decrease consistently across the lamina from midrib to margin at (CMW1 of 1.09–(1.33)–1.71, n = 3; CMW2 of 1.12–(1.33)–1.74, n = 5), but the first few rows of meshes adjacent to the midrib are shorter and more angular in shape, becoming more elongate-elliptical to elongate-polygonal across the midlamina and towards the margin. There is a large range in leaf size, and in mesh width (and consequently in marginal vein density) in this morpho- type, but these do not appear to be correlated – the two largest leaves fall at opposing ends of the spectra for both vein density and mesh size. The meshes of MF04 are less linear in appearance than in MF01 and MF02, the leaf is shorter and broader. These leaves are smaller than those of MF07, and the venation is more consistent (with broader areolae) across the lamina. Comparison with South African and Argentinean leaves. Halle (1911) placed these leaves in G. browniana. Of all glosso- pterid leaf species, this is possibly the most misrepre- sented in the literature. The type material of G. browniana derives from the upper Permian Newcastle Coal Mea- sures in the Sydney Basin (Rigby et al. 1980), and the species is widely represented in the Sydney-Gunnedah- Bowen foreland basin complex of eastern Australia (McLoughlin 1994a), but beyond this province, its recog- nition is equivocal. This species is commonly used as a catchall for leaves with no particularly outstanding diag- nostic features, and could be called the ‘wastebasket’ of glossopterid taxonomy, a sentiment shared by Kovács- Endrödy (1991, p. 32). The general concept for this species is: a leaf that is oblanceolate, of small to medium size, with venation arising at a narrow angle and curving gently across the lamina to intersect the margin at a moderate angle; meshes are generally slightly broader near the midrib, polygonal to elongate-trullate, becoming more 68 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 elongate-polygonal to elongate-elliptical across the lamina, retaining a linear pattern and there is no dramatic decrease in mesh width towards the margin. In this broad sense of ‘G. browniana’, MF04 accords well. However, Rigby et al. (1980), in a study of topotype material, noted that this species is defined by very few (<3) anastomoses across the lamina in contrast to >3 in MF04. Moreover, Australian upper Permian examples (including the neotype) of G. browniana have vein marginal angles of 40–50° (Rigby et al. 1980; McLoughlin 1994a), i.e. much less that the >60° evident in the Falkland (Malvinas) specimens. The fragmentary nature of the studied specimens means that we do not have a full appreciation of the size and shape of these leaves. Although they would fall within the upper size ranges, they nevertheless bear a strong resemblance to some of the South African upper Permian leaves linked to Lidgettonia species by Anderson & Anderson (1985). For example, among the leaves grouped under ‘L. lidgettonioides’, are forms in which the venation pattern is very similar to that evident in MF04 (e.g. their p. 301, pl. 126, figs 4–6, from Mooi River). Kovács-Endrödy (1991, pp. 33–34, pl. 5.6, figs 3–6) figured several diminutive leaves from Hammanskraal, Karoo Basin outlier, that she attributed to G. browniana. These have very similar venation to MF04 and might be conspecific pending a reappraisal of the taxonomy of that leaf assemblage. Leaves associated with, or attached to, Scutum leslii fructifications from Vereeniging, South Africa (Anderson & Anderson 1985, p. 242, pl. 67, figs 22–23; p. 243, pl. 68; Prevec 2011, p. 11, figs 24–26, p. 14, figs 44–47) are also similar, although the meshes in these forms are generally narrower than in MF04, and the leaves are much larger. Despite the wide variation in shape and size that charac- terizes Glossopteris browniana leaves, the well-defined polygonal meshes separate it from other species. Notably, when compared to the Falklands (Malvinas) fossils, some Bonete Formation (Argentina) fossil leaves show strong similarities with MF04 (S047884-01, not figured by Halle 1911), and specimens attributed to G. browniana by Halle (1911, pl. 1, figs 27–28), particularly in the angle of second- ary vein departure from the midrib, ranging from acute to almost perpendicular in some cases, distinguishing them from similar leaves with elongate-polygonal meshes across the lamina. The Falklands (Malvinas) specimens, however, differ slightly in that the relatively broader elongate- polygonal to elongate-elliptical meshes continue to the margins in some cases (e.g. S047835-01 [Fig. 4G, H], S047781-01), whereas in leaves attributed to ‘G. browniana’ from the Bonete Formation, the meshes tend to reduce in width and become more elongate. A similar pattern is evident in leaves assigned to ‘G. browniana’ from the La Golondrina Formation (Archangelsky 1958; Cariglino 2011). In specimen GHUNLPam 3243 from the Carapacha Formation (Argentina), assigned to Glossopteris tortuosa Zeiller, the meshes in the row adjacent to the midrib are polygonal to trapezoid, and arise from a thick, well- defined midrib (Melchor & Césari 1997, fig. 6.5). This species, however, differs from MF04 in that the secondary veins emerge almost perpendicular to the midrib with little curvature towards the margin, and the meshes retain constant width across the lamina. The length-width ratio of the meshes is also considerably smaller than in MF04. Glossopteris morphotype MF05, Fig. 7A–C; Suppl. Fig. 1I 1911 Glossopteris damudica Feistmantel; Halle, pp. 60–61, pl. 3, figs 5–7. Material. S047784. Localities. Dos Lomas Site 1. Description. Large microphyll to small mesophyll-class leaf, probably elongate-elliptical (Fig. 7A), with a width of 34.8 (n = 1); base and apex unknown. Midrib narrow but well defined, individual veins not resolved; observed width of 0.84 mm (n = 1). Secondary veins arise from midrib at very wide angle 51.4°–(71.1°)–81.3° (n = 1), and abruptly curve to run almost perpendicular to midrib 80.6°–(83.7°)–86.1° (n = 1) following an almost straight course to intersect the margin at 82.3°–(85.6°)–89.1° (n = 1) (Fig. 7B, C; Suppl. Fig. 1I). In medial portion of leaf, meshes adjacent to midrib are broadly triangular, 0.46–(0.54)– 0.6 mm wide (n = 1), becoming strongly linear, narrowly elongate-elliptical, 0.26–(0.29)–0.36 mm wide (n = 1) in mid-lamina, and progressively narrowing towards margin 0.12–(0.18)–0.22 mm (n = 1), with a marginal vein density of 15–(26)–27 veins per 5 mm (n = 1). Remarks. This morphotype is represented by only a single leaf yet it is possibly the most distinctive form in the collections. It is relatively large (large microphyll/small mesophyll-class), with secondary venation following a straight course to the margin (CVA1: 1.2, n = 1; CVA2: 1, n = 1) at a very large angle (virtually perpendicular) to the midrib. Meshes are also distinctive in being polygonal and broad in the row adjacent to the midrib, but becoming abruptly more linear to narrowly elliptical, and steadily decreasing in width across the lamina (CMW1: 1.84, n = 1; CMW2: 1.67, n = 1). Among other leaf types in the collec- tion, it is most similar to MF07, but the vein angle in the latter is at a lesser angle, and the meshes are broader. Comparison with South African and Argentinean leaves. Among South African fossil leaves, MF05 is strongly remi- niscent of forms from Hammanskraal described by Kovács-Endrödy (1991, pp. 48–49, pl. 5.11, figs 7, 9) as Glossopteris divergens and (p. 65, pl. 5.15, fig. 7) Glossopteris cf. damudica. The Falklands (Malvinas) specimen (34.8 cm wide) is slightly smaller than the Hammanskraal leaves (c. 5 cm wide), but is much smaller than the very large G. cf. damudica leaves (c. 14 cm wide) from Lawley figured by Anderson & Anderson (1985, pl. 135, figs 1–3, pl. 136, figs 1–2). The latter leaves have veins that form several polygonal meshes at a low angle to the midrib, and then recurve at a very broad angle across the lamina to the margin. This recurved venation pattern was not observed in MF05, but it is otherwise similar. Other leaves figured by Anderson & Anderson (1985, pl. 100, figs 5–7), from the Cedara locality, are larger than M05 but have similar vena- tion. Lacey et al. (1975, p. 371, NM 1775b) figured one small section of leaf from Mooi River (Normandien Formation), which they compared to G. damudica Feistmantel, with ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 69 70 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Figure 7. A–C, Glossopteris morphotype MF05; S047784; A, large leaf with secondary venation almost perpendicular to the midrib; B, C, details of venation. D–H, Glossopteris morphotype MF06. D, elliptical leaf; E, details of venation; S047778a-01; F, central portion of leaf showing dense venation; S047872-01-02. G, obovate leaf; H, details of dense venation; S047778b-02. All scale bars = 10 mm. venation perpendicular to the midrib, including the first row of meshes adjacent to the midrib. Meshes in this leaf then appear to become narrower and more elongate- elliptical away from the midrib. This fragment is within the size range, and closely similar in morphology to M05 and might be conspecific. Leaves similar to MF05 with a narrow but well-defined midrib and secondary veins arising almost perpendicular to the midrib, following a straight, (sub)parallel course to the margin, and forming narrow, linear meshes have been described from the Carapacha Formation of Argentina and were assigned to Glossopteris damudica (Melchor & Césari 1997). The latter species has also been confirmed from the La Golondrina Formation (Archangelsky 1958; Cariglino 2011). Both of these leaves are very similar to MF05 and might indicate a Glossopteris species shared between South Africa, the Falklands (Malvinas) and Argentina. Glossopteris morphotype MF06, Fig. 7D–H; Suppl. Fig. 1B Material. S047778a-01 (Fig. 7D, E; Suppl. Fig. 1B), S047778a-02, S047778b-02 (Fig. 7G, H) (all identified by Halle as Glossopteris indica in NRM database), S047872-01- 02 (Fig. 7F: identified by Halle as Glossopteris browniana in NRM database). Localities. Low Bay. Description. Leaves lanceolate to oblanceolate, micro- phyll-class, 100 mm long (n = 1), 22.89–25.39 mm wide (n = 2), with estimated leaf area of 1781.2 mm2 (n = 1), apex acute and base cuneate (Fig. 7D, F, G). Midrib narrow, well-defined in basal and medial portion of leaf, persistent to apex; basal width up to 1.43 mm (n = 2), medial width 0.78–(0.99)–1.19 mm (n = 3) tapering to 0.65–0.31 mm (n = 2) near apex; individual veins weakly defined, depending on preservational quality. Secondary veins arise from midrib at steep (low) to moderate angle of 21.9°–(28.9°)– 39.8° (n = 2), arch gently to mid-lamina where they follow an almost straight path at 54.6°–(62.7°)–70.6° (n = 2) across the lamina, intersecting the margin at 65.3°–(72.1°)–78.6° (n = 2) (Fig. 7E, H; Suppl. Fig. 1B). Veins anastomose 3–4 times and bifurcate four times from midrib to margin in medial part of leaf; meshes adjacent to midrib are elongate- trullate, polygonal to triangular, 0.27–(0.41)–0.72 mm wide (n = 3), becoming strongly linear and elongate-elliptical across mid-lamina 0.29–(0.33)–0.38 mm (n = 3), becoming narrower progressively towards margin 0.17–(0.22)– 0.27 mm (n = 3), with a marginal vein density of 18–(20)–22 veins per 5 mm (n = 3). Remarks. Leaves attributed to this morphotype are in the large microphyll size-class with an acute apex and a narrow, well-defined midrib. This morphotype is distinguished by the secondary veins emerging at a narrow to moderate angle from the midrib, initially arching strongly with a CVA1 of 2.2 (n = 2) into the mid-lamina, then following an almost straight path to the margin, with a CVA2 of 1.2 (n = 2). The meshes in the first few rows adjacent to the midrib are slightly broader, but then mesh size decreases gradually into the mid-laminar region with a CMW1 of 0.95–(1.23)–1.69 (n = 3), and then decreasing more intensely towards the margin with a CMW2 of 1.40–(1.53)–1.78 (n = 3). Marginal vein density is high. The relatively straight path of the venation in some specimens of this morphotype is reminiscent of MF07 but the vein angles are much steeper (lower), and the venation more gently curved and denser in the former. The prominent midrib and leaf shape of MF06, which appears to be oblanceolate with a sharply cuneate base, separates it from MF01 and MF02, which are narrower, more linear leaves. Comparison with South African and Argentinean leaves. This form of Glossopteris has generalized features and is poten- tially common in South African floras, depending on how the morphotypes are categorized. Some of the leaves attributed to various ‘Arberia’ associations by Anderson & Anderson (1985) have similar shape and venation, although they are all much larger, e.g. ‘A. madagascariensis’ from Hammanskraal (their pp. 280–281 pl. 105, figs 1, 2; pl. 106, figs 1–4); ‘A. leeukuilensis’ fromVereeniging (their p. 283, pl. 108); ‘A. allweyensis’ from Lawley (their p. 285, pl. 110, figs 1–3). There are also similarities in venation between MF06 and some leaf taxa attributed to the ‘Lidgettonia’ associations of Anderson & Anderson (1985), e.g. ‘Lidgettonia africana’ from Lidgetton (their p. 289, pl. 114, figs 3–10); ‘Lidgettonia mooiriverensis’ from Mooi River (their pp. 292–293, pls 117–118), also figured by Lacey et al. (1975) and attributed to G. indica (p. 366, NM 1774). Leaves figured by Anderson & Anderson (1985, pl. 142, figs 1–4) from Zwartskraal and included within their concept of Glossopteris symmetrifolia, are larger than MF06, but have very similar venation. Among the leaf forms represented in Argentina and the Falklands (Malvinas), Menéndez (1966) found close similarities between his Glossopteris indica and those described by Halle (1911, pl. 2) and here attributed to MF06, although he stated that in some, the secondary veins tended to follow a straighter course than those from the Bonete Formation. Menéndez (1966) also regarded specimens that Seward & Walton (1923, pl. XX, fig. 9, pl. XIX, fig. 7) attributed to G. indica from the islands also coincided with his view. Based on the morphotypes proposed herein, the Bonete Formation G. indica shares an oblanceolate to lanceolate shape, a well-defined midrib that is persistent to the apex, and secondary venation with few meshes, curving slightly across the lamina to reach the leaf margin at 50–70°. Glossopteris indica described from the La Golondrina For- mation could also be conspecific with material assigned to MF06, based on the presence of a well-defined persistent midrib, and secondary venation that moderately arches away from the midrib, taking a straight path to the margin at c. 60–70°, and forming long, narrow, parallel meshes (Archangelsky 1958). Glossopteris morphotype MF07, Fig. 6A–D; Suppl. Fig. 1E,H 1911 Glossopteris indica Schimper; Halle, pp. 56–58, pl. 2, fig. 1, pl. 2, fig. 6, 6a. 1923 cf. Glossopteris indica var. wilsoni Seward; Seward & Walton, p. 322, pl. 21, fig. 13. Material. S047777-01 (Fig. 6C, D; Suppl. Fig. 1E), ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 71 S047781-02, S047811, S047835-02 (Fig. 6A, B; Suppl. Fig. 1H) (all identified by Halle as Glossopteris indica in NRM database). Localities. Dos Lomos Site 1; Speedwell Island Site 2; Bodie Creek Head Site 1. Description. Leaf large, mesophyll-class, probably elongate- lanceolate with a width of 30–43 mm (Fig. 6A, C); base narrowly cuneate, truncated/sessile. Midrib wide (>3.2 mm) and prominent in leaf base, becoming narrow (0.67 mm) but well-defined in medial part of leaf; individual veins distinct. In medial portion of leaf, veins depart from midrib at a moderate angle 47.3°–(48.9°)–50.7° (n = 1), curve immediately to follow a straight path across lamina at 73.5°–(74.3°)–74.8° (n = 1) to intersect margin at 73.9°–(74.6°)–75.6° (n = 1) (Fig. 6B, D; Suppl. Fig. 1E, H); venation very linear in pattern, first two rows of meshes markedly broader, then decreasing gradually in width across lamina to margin. In medial portion of leaf, veins anastomose and bifurcate four times from midrib to margin, forming meshes that are falcate, triangular and elongate-polygonal, 0.37–(0.54)–0.75 (n = 2) mm wide, adjacent to midrib, gradually narrowing to 0.3–(0.4)–0.5 (n = 2) in mid-lamina, and becoming abruptly and strongly linear or narrowly elongate-elliptical, 0.14–(0.23)–0.3 mm wide (n = 1) towards margin. Marginal vein density of 21 veins per 5 mm. Remarks. Four incomplete leaves were identified and two of the larger fragments were analysed. This relatively large leaf (small mesophyll-size class) is also distinguished by its prominent midrib, and broader, triangular/polygo- nal meshes adjacent to the midrib, and the abrupt curva- ture of the venation from these proximal meshes (CVA1: 1.5, n = 1), forming strongly linear meshes that follow a straight path at a moderate angle (CVA2: 1, n = 1) to the margin. Meshes are longer near the midrib and in the mid-lamina, than evident in MF05, and the vein angle is slightly steeper (lesser). MF07 differs from MF04, in being much larger, and there is a greater decrease in mesh width from midrib to margin in the former. Comparison with South African and Argentinean leaves. There are similaries between MF07 and leaves from the South African Hammanskraal flora assigned by Anderson & Anderson (1985) to ‘Hirsutum leslii’ (pp. 257–258, pl. 82–83), some attached to Elatra leslii fructifications. Kovács-Endrödy (1991 p. 64, pl. 5.15, fig. 6) assigned these leaves to Glossopteris taeniopteroides. Both the Falkland (Malvinas) and South African forms are linear, very elongate-obovate leaves with a long, cuneate base, although some of the Hammanskraal leaves have an expanded base associated with attached E. leslii fructifica- tions. The midrib is very prominent, particularly at the base. Both MF07 and leaves assigned by Kovács-Endrödy (1991) to G. taeniopteroides have very linear venation with slightly broader, trullate meshes near the midrib, becom- ing finer towards the margin and crossing the lamina at a wide angle to the midrib. Lacey et al. (1975; NM1761, p. 371) figured a leaf with similar venation, but possibly a more broadly lanceolate lamina, that they compared with G. ampla. As in MF07, the venation is steeply angled at the midrib, curves fairly sharply and then crosses the lamina on a straight path to the margin, the linear meshes becoming progressively narrower. A characteristically strong, thick midrib has been observed in very few leaves from the Permian of Argen- tina. Specimen GHUNLPam 3224 from the Carapacha Formation (Melchor & Césari 1997, fig. 4.9) is character- ized by a 3-mm-wide midrib with secondary veins form- ing long, linear meshes throughout the lamina, reaching the margin at c. 50°. The leaf was assigned to Glossopteris browniana, but it is very similar to MF07. Another specimen with a thick midrib and also previously assigned to G. browniana, is BA-PB 13832 from the La Golondrina Formation (Cariglino & Gutiérrez 2011). This specimen in particular, closely resembles S047777-01 (Halle 1911, pl. 2, fig. 1) in the moderate arching of the secondary veins near the midrib and the linear, straight path to the margin, forming few, elongated meshes. Glossopteris morphotype MF08, Fig. 6E 1911 Gangamopteris cyclopteroides var major Feistmantel; Halle, pp. 61–62, pl. 3, figs 8, 9. Material. S047785 (Fig. 6E). Localities. Dos Lomos Site 1. Description. Relatively large mesophyll-class leaf, leaf base broadly cuneate, truncated at width of 8 mm; proxi- mal margins of lamina diverge at a moderate angle to form broad leaf (>40 mm wide). Midrib very broad and flat basally, 4.7 mm wide; individual veins anastomose and bifurcate (Fig. 6E). In basal portion of leaf, secondary veins run parallel to midrib, depart at a very narrow angle of 10.3–17° (n = 1), arch strongly and consistently across mid-lamina to 38.5°–(42.5°)–47.3° (n = 1), before curving more gently to intersect margin at 52°–(54.8°)–58° (n = 1) to the midrib. In medial portion of leaf, meshes adjacent to midrib are elliptical, falcate to linear and 0.3–(0.34)–0.38 mm wide (n = 1), becoming more linear and narrowly elongate- elliptical across the lamina, narrowing slightly towards the mid-lamina to 0.28–(0.31)–0.34 mm wide (n = 1), then remaining consistent with a width of 0.27–(0.3)–0.32 mm (n = 1) at margin. Marginal vein density is 17 veins per 5 mm (n = 1). Remarks. Only the lowermost medial portion of a single leaf base, with one half torn and distorted, is present in the collection. Experience has shown that leaf bases tend to have convergent or conservative morphology across glossopterid species that are otherwise quite distinct. Hence, the available features alone are not reliable for species recognition. However, this leaf base is clearly distinct from other morphotypes available in Halle’s collection. The shape of the base, extremely broad midrib incorporating clearly anastomosing and bifurcating veins, the venation that arches strongly from midrib to mid- lamina, with a CVA1 of 3.1, flattening slightly from the mid-lamina towards the margin with a CVA2 of 1.3, are all distinctive features of this morphotype. Glossopterid scale leaf: cf. Eretmonia sp., Fig. 6H 1911 Scale-frond of Glossopteris sp. (?) Halle, p. 62, pl. 3, fig. 10. Material. S047786. Locality. Dos Lomas Site 1. 72 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Description. Bract consisting of a 10-mm-long robust (6-mm-wide) petiole and 13-mm-long, 20-mm-wide rounded triangular lamina (Fig. 6H). Margin entire. Petiole containing multiple dense reticulate veins. Lamina lack- ing an obvious midrib, veins indistinct, reticulate, forming elongate-polygonal to elliptical meshes typically 2 mm long and 0.4 mm wide. Veins in mid-lamina departing from midline at 30°–40° then passing relatively straight to margin. Remarks and comparison. Bracts of this type are fairly common in Permian strata across Gondwana and are generally assigned to Eretmonia where a pair of sporangio- phores is inserted on the midline. No sporangiophores were detected on the Falkland (Malvinas) specimen, but these may have detached before burial. The specimen from the Falklands (Malvinas) is similar in size and shape to microsporophylls of Lidgettonia inluzanensis Anderson & Anderson 1985 (otherwise referable to Eretmonia du Toit 1932) from the Lopingian Normandien Formation of the northeastern Karoo Basin, although examples of the latter tend to have slightly more pointed apices. Order Cordaitales Genus Noeggerathiopsis Feistmantel, 1879 Noeggerathiopsis sp., Fig. 8A–K 1911 Desmiophyllum sp.; Halle, pp. 63–64, pl. 4, figs 1–7. Material. NRM S047789 (Fig. 8A), S047790 (Fig. 8B, G), S047910 (Fig. 8C, H), S047911 (Fig. 8D, I), S047914 (Fig. 8F, J), S047930-02, S047934-01 (Fig. 8E, K), S047935. Description. Oblanceolate to linear leaves (possibly falcate in one case), up to 107 mm long and 11 mm wide (possibly reaching 18 mm in one case); margin entire, apex rounded, base narrowly tapered (Fig. 8A–C); venation parallel to slightly divergent proximally, slightly conver- gent distally; c. 10 veins or grooves per 5 mm across the mid-region of leaf (Fig. 8D–F). Remarks. Halle (1911) assigned these leaves to Desmio- phyllum, a heterogeneous fossil-genus of parallel-veined, strap-shaped leaves, normally recognized from Mesozoic strata and possibly having czekanowskialean, ginkgoalean or pinalean affinities (van Konijnenburg-van Cittert et al. 2019). However, Desmiophyllum tends to be represented by leaves that are more slender (narrowly linear) with numerous interveinal striae (Miller & Hickey 2010; van Konijnenburg-van Cittert et al. 2019). We consider the Falklands (Malvinas) examples to be of cordaitalean affinity, given the wide distribution of morphologically similar parallel-veined oblanceolate leaves in Permian strata across Gondwana (Pant & Verma 1965; Rigby et al. 1980). The type specimen of Noeggerathiopsis hislopii (Bunbury) Feistmantel, 1879 (the type of the genus) lacks cuticle and is difficult to differente from species of the predominantly northern hemisphere Cordaites Unger, 1850 on macro- morphological grounds. On this basis, several researchers have proposed that all such cordaitalean leaves should be assigned to Cordaites (Meyen 1969; Maheshwari & Meyen 1975; Rigby et al. 1980). However, both cuticular and anatomical studies of Gondwanan cordaitalean leaves show them to differ significantly from northern hemi- sphere Cordaites (Pant & Verma 1965; McLoughlin & Drinnan 1996; Degani-Schmidt & Guerra-Sommer 2019). Consequently, Rigby (1984) erected Pantophyllum for Gondwanan cordaitalean leaves retaining cuticular details but this name has generally not been used by later workers and Noeggerathiopsis continues to be employed, collectively, for impressions, cuticle-bearing compres- sions, and petrifactions (Pant & Verma 1964; McLoughlin & Drinnan 1996; Srivastava & Tewari 2004). Noeggerathiopsis leaves vary in their micromorphology, hence Pant & Verma (1965) erected N. bunburyana, N. papillosa and N. fibrosa, and Lele & Maithy (1964) established N. indica, N. gondwanensis and N. zeilleri for leaves with a gross form equivalent to the type species, N. hislopii (Bunbury) Feistmantel, 1879, but having distinc- tive suites of cuticular characters. In some cases, the differ- ences are very subtle and several or all of these species may be synonyms (Srivastava & Tewari 2004). Further, Cordaites dumanii was erected by Chandra & Srivastava (1991) for Noeggerathiopsis-type leaves with major veins separated by apparent minor veins. However, the latter features probably represent stomatiferous grooves that, typically, are indistinct on impressions, but have been well- characterized in permineralized specimens (McLoughlin & Drinnan 1996). Leaf form is variable within individual populations of Noeggerathiopsis leaves. Even within a single cluster of attached leaves, lamina shape and dimensions can vary significantly as part of a seasonal growth series (Etheridge 1918). Leaves from the Falkland (Malvinas) Islands lack preserved cuticle, and the preservational quality is insuffi- cient to resolve whether the linear features on the leaf represent veins or interveinal stomatiferous grooves in most cases (Fig. 8D–K). Some leaves appear to have a wrinkled texture that might be the result of desiccation before burial (Fig. 8H, I). The linear features dichotomize in one case (Fig. 8I), but in other cases appear to initiate between adjacent lineations (Fig. 8J), hence these features probably represent a combination of both veins and stomatiferous grooves. Owing to their indistinct features, we can not definitively assign these leaves to any estab- lished species, although we note close similarities to the narrowly oblanceolate, densely veined leaves of Cordaites (=Noeggerathiopsis) dumanii from the Karharbari Forma- tion (upper Cisuralian) of India (Chandra & Srivastava 1991) and Noeggerathiopsis elongata from the Vryheid For- mation (Artinskian) of the northeast Karoo Basin, South Africa (Anderson & Anderson 1985). The two slightly stronger veins in the long, tapering base of S047790, are also strongly reminiscent of Sphenobaiera eccaensis Anderson & Anderson from Vereeniging, Karoo Basin (Anderson & Anderson 1985, p. 323). That South African species also possesses apparent stomatiferous grooves between veins, lacks lamina divisions typical of Spheno- baiera, and might better be placed in Noeggerathiopsis. The very gently tapered base of the Falkland (Malvinas) specimens and their slightly convergent venation in the distal portion of the lamina is also similar to that of ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 73 74 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Figure 8. Noeggerathiopsis sp. (Cordaitales) leaves. A–F, leaves showing gently tapered base with slightly divergent venation, rounded apex with slightly convergent venation, and entire margins. G–K, enlargements of leaf mid-regions showing the variable prominence of sparsely dichotomizing veins and interveinal ridges and wrinkles. A, S047789; B, G, S047790; C, H, S047910; D, I, S047911; E, K, S047934-01; F, J, S047914. All scale bars = 10 mm except for J (= 5 mm). Metreophyllum lerouxii Anderson & Anderson, 1985 from the Vryheid Formation of South Africa except that the vein density of the Falkland (Malvinas) specimens is greater than in N. elongata and M. lerouxii. Order Buriadiales Family uncertain Genus Brasiliocladus Yoshida (in Bernades de Oliveira & Yoshida 1982) Brasiliocladus sp. cf. B. acicularis Yoshida (in Bernades de Oliveira & Yoshida 1982), Fig. 9A–F 1911 Coniferous branches Halle, p. 63, pl. 3, figs 11–15. Material. S047754-02 (Fig. 9C), S047787 (Fig. 9A, B), S047921 (two specimens), S047924 (listed as Phyllotheca australis on museum labels by Halle 1911). Localities. Speedwell Island (Site 1); Speedwell Island (north of settlement); Black Point (Site 1). Description. Fragments of leaf-bearing axes, sparsely branched, up to 38 mm long, 0.5 mm wide (Fig. 9A). Leaves attached at variable orientations (10–70°) to axis, linear, straight, univeined, typically 3–5 mm but rarely up to 8 mm long, and 0.5 mm wide, with blunt to pointed apices and broadly clasping bases (Fig. 9B–F). Leaves inserted in apparent helical arrangement (Fig. 9C). Remarks. Several of the specimens illustrated by Halle (1911, pl. 3, figs 11–15) are not readily recognizable in the collections and it appears that the lithograph illustrations of the conifers prepared for his publication include some degree of artistic license. The short, spirally arranged, stiff, spine-like leaves of these fossils suggest affiliation with the voltzialean conifers. Fossils of this group are widely distributed but are relatively uncommon in Permian strata of Gondwana. Nevertheless, multiple genera have been established and some (e.g. Walkomiella) can be distinguished from the Falkland (Malvinas) specimens based on their triangular to rhombic foliage (White 1981). Others (e.g. Paranocladus, Ugartecladus and Ferugliocladus) generally have narrowly ovate to triangular leaves, in many cases appressed to the axis, and are typically differentiated on cuticular or reproductive characters (Fittipaldi & Rösler 1978; Bernardes de Oliveira & Yoshida 1982; Archangelsky & Cúneo 1987; Ricardi-Branco et al. 2013; Mune et al. 2016) that, unfortunately, are not available for the Falkland (Malvinas) specimens. Leaves of Buriadia differ in being polymorphic, incorporating forms with bifid and multifid apices (Pant & Nautiyal 1967; Ricardi- Branco et al. 2013; Serbet et al. 2010). Searsolia differs by both its bifid leaf tips and decussate leaf arrangement (Pant & Bhatnagar 1975). Krauselcladus has robust needle- like leaves, some of which have bifid apices (Fanton et al. 2006). Coricladus is characterized by well-spaced, high- angled, small, univeined leaves with acute apices (Jasper et al. 2005). The Falkland (Malvinas) specimens are most similar to Brasiliocladus acicularis Yoshida described from the Rio Bonito Formation (?Sakmarian–Roadian) of Brazil, based on the acutely orientated, univeined, needle-like, <12-mm-long, homomorphic leaves, lacking bifid apices (Yoshida 1968; Bernardes de Oliveira & Yoshida 1982), although the Brazilian specimens tend to have slightly denser leaf helices. The Falkland (Malvinas) specimens are also notably similar in their stout, straight, broad- based, spinose leaves to the leafy twigs assigned to Cyparissidium sp. A. from the Vryheid Formation (Artinskian) of the Karoo Basin by Anderson & Anderson (1985). Incertae sedis Dispersed seed, Fig. 9G Material. NRM S047884-02, possibly an additional speci- men on S047948. Localities. North Arm (Site 1). Description. A single seed (S047884-02: Fig. 9G) has a damaged transversely elliptical central body 4.5 mm long and 5.5 mm wide, flanked by marginal wings that are 2 mm in maximum width, extending slightly beyond the central body to have lengths of 5.5 mm, and which are contracted at the chalazal and micropylar ends. The seed is otherwise featureless. Remarks. Winged seeds are widespread in Permian assemblages across Gondwanan and are typically assigned to Samaropsis Goeppert, Cordaicarpus Geinitz or various other genera (Maithy 1965; Millan 1981; McLoughlin 1992; Singh 2002; Bernardes-de-Oliveira et al. 2007; Souza & Iannuzzi 2009) but relatively few can be assigned with consistency to particular plant clades. The specimen illustrated here (S047884-02: Fig. 9G) might fall within the broad circumscription of Samaropsis but has too few diagnostic characters to enable assignment to any one gymnosperm group represented in the leaf assemblages from the Falkland (Malvinas) Islands. A second, roughly circular 4.5-mm-diameter, object on sample S047948 from south of the settlement on Speedwell Island possibly represents a seed imprint but is too poorly preserved to resolve specific morphological features. Forked roots, Fig. 9H–K Material. NRM S047931, S047942, S047947, S047948. Localities. Speedwell Island (north and south of settle- ment). Description. Equi-dichotomous (Fig. 9I, K) to pseudo- monopodial-style (Fig. 9J) branched roots ranging from 0.3 to 5.0 mm wide (Fig. 9H, K). All forms with fine, discon- tinuous, longitudinal striae. No nodes, other modifica- tions or appendages apparent. Remarks. In the absence of internal chambers typical of Vertebraria, we suggest that the majority of such forked roots derive from non-glossopterid plants but they can not be ascribed to a particular clade. They typically occur in bleached palaeosols, either parallel to or transecting bedding. Order and family uncertain Genus Australoxylon Marguerier ex Blazer, 1975. Type. Australoxylon teixeirae Marguerier ex Blazer, U.S. Geol. Surv. Bull., 1396, p. 5. 1975. ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 75 76 ISSN 2410-4418 Palaeont. afr. (2024) 58: 53–93 Figure 9. Conifer, seed and root remains from Permian strata of the Falkland (Malvinas) Islands. A–F, Brasiliocladus sp. cf. B. acicularis Yoshida; G, winged seed; H–K, branched roots. A, B, slender axis bearing ill-defined short spinose leaves; S047787. C, branch apex bearing acutely disposed, spirally arranged, spinose leaves; S047754-02. D, E, branch tips with slightly curved spinose leaves; S047921. F, cluster of spinose univeined leaves; S047924. G, winged seed with degraded central body; S047884-02. H, probable root system penetrating bedding plane with radially disposed rootlets, S047788. I, small forked roots; S047947. J, pseudomonopodial rooting system; S047931. K, large forked roots; S047942. Scale bars for A, C–E, J, K = 10 mm; B, F, H = 5 mm; G, I = 1 mm. Nomenclatural note 1. Australoxylon was not validly published by Marguerier (1973a) because more than one species was described without designation of a type (ICBN, art. 38.5; Turland et al. 2018). This problem was not remedied in a subsequent publication by that author (Marguerier 1973b). The valid publication of Australoxylon was made effective (ICBN, art. 38.1, Turland et al. 2018) by Blazer (1975), whereby he selected A. teixeirae Marguerier as the type species. However, this species name was not validly published (ICBN, art. 38.5) by Marguerier (1973a), as the genus was not validly published in that paper. Nomenclatural note 2. In 2001, after having reviewed the type slides, Bamford & Philip