First Evidence for Wollemi Pine-type Pollen (Dilwynites: Araucariaceae) in South America

We report the first fossil pollen from South America of the lineage that includes the recently discovered, extremely rare Australian Wollemi Pine, Wollemia nobilis (Araucariaceae). The grains are from the late Paleocene to early middle Eocene Ligorio Márquez Formation of Santa Cruz, Patagonia, Argentina, and are assigned to Dilwynites, the fossil pollen type that closely resembles the pollen of modern Wollemia and some species of its Australasian sister genus, Agathis. Dilwynites was formerly known only from Australia, New Zealand, and East Antarctica. The Patagonian Dilwynites occurs with several taxa of Podocarpaceae and a diverse range of cryptogams and angiosperms, but not Nothofagus. The fossils greatly extend the known geographic range of Dilwynites and provide important new evidence for the Antarctic region as an early Paleogene portal for biotic interchange between Australasia and South America.


Introduction
The Southern Hemisphere monkey-puzzle tree family, Araucariaceae, was long believed to comprise two living genera: Araucaria Juss., with about 19 species endemic to the southwest Pacific and South America, and Agathis Salisb., with about 20 species distributed from Sumatra to New Zealand but absent in South America. Remarkably, a third araucarian genus was discovered in 1994 in New South Wales, Australia, whose sole species is Wollemia nobilis W.G. Jones, K.D. Hill & J.M. Allen, common name Wollemi Pine [1].
With fewer than 40 adult specimens known to survive in the wild, W. nobilis is one of the world's rarest trees. Adding to the spectacular nature of the discovery was the location of the stands, in a remote gorge within 150 km of Sydney, Australia's largest city; the large stature of the trees (up to 40 m tall); and the apparent similarity of the foliage to that of the Jurassic species ''Agathis'' jurassica M.E. White [2] and the Cretaceous to early Cenozoic genus Araucarioides [3][4][5][6]. So far, none of the similar macrofossils has been convincingly demonstrated to belong to Wollemia [6,7], and indeed ''A.'' jurassica differs from foliage of Wollemia in details of venation, leaf arrangement and leaf shape [8]. However, the presumed close relationship quickly led to W. nobilis being given the status of a ''living fossil from the age of dinosaurs'' in the popular press (e.g. [9]).
In contrast, once pollen was made available, it was quickly recognized that the Wollemia clade had a well-established fossil history provided by the morphogenus Dilwynites W.K. Harris [10], comprising D. granulatus W.K. Harris and D. tuberculatus W.K. Harris [6,11,12]. For example, in southern Australia, D. granulatus, the morphospecies that most closely resembles modern Wollemia pollen, can be traced back as far as the Turonian Age (89.8 to 93.9 Ma) of the Late Cretaceous [11][12][13]. So far, Dilwynites has been identified in Paleogene to Neogene deposits of western, central, and northern Australia (references in [14]), in Cretaceous to Neogene deposits of New Zealand [15], and in late Eocene deposits of East Antarctica [16]. However, apart from a possible record from the Paleogene of Seymour Island [17], no Dilwynites pollen has previously been recognized from West Gondwana.
Since 2000, the first author has also recognized that at least one species of Agathis produces pollen that is morphologically consistent with Dilwynites, and thus the nearest extant relatives of the plants that produced Dilwynites pollen are best regarded as both Wollemia and Agathis (e.g., [14]). This observation is consistent with the results of many recent molecular studies indicating that Wollemia and Agathis are sister taxa [18][19][20][21][22][23], along with several characteristics of the seed cones of the two genera that may be synapomorphic, especially the condition of the seeds being winged and nearly free from the fused bract and scale [24]. By contrast, in Araucaria, the seeds are embedded in the bract/scale complex.
We here present microfossil evidence that araucarians producing Dilwynites pollen were growing in southern Patagonia during the late Paleocene to early middle Eocene. This discovery greatly augments evidence for the past range of Wollemia and/or Agathis conifers that produce this pollen type and adds to the growing paleobotanical evidence for extensive trans-Antarctic interchange between Patagonia and Australia during the globally warm early Paleogene.

Ethics Statement
All necessary permits were obtained for the described study, which complied with all relevant regulations. Permits were issued by the Secretaría de Estado de Cultura de la Provincia de Santa Cruz, Argentina.
The Dilwynites specimens reported here came from an isolated, newly recognized, streamcut outcrop of the Ligorio Márquez Formation in Santa Cruz Province, Patagonia, Argentina, located along the Río Zeballos and 36 km south-southwest of the town of Los Antiguos (Fig. 1). Precise locality data are available on request from AI, PW, or Museo Padre Jesús Molina, Río Gallegos, Santa Cruz, Argentina (MPM), where material is stored. The Ligorio Márquez Formation, previously studied only on the Chilean side of the nearby border [25][26][27], comprises a sequence of coastal floodplain, fluvial and mire facies deposited in a foreland basin (the Ligorio Márquez Basin) that subsequently was uplifted by compressional Andean tectonic activity during and since the early Miocene [28,29]. The material studied here was derived from a ,0.5 m thick carbonaceous shale bed containing abundant fossil leaves (under separate study), from a probable coastal swamp.
The local exposure of the Ligorio Márquez Formation lies unconformably above the Cretaceous Río Tarde Formation and is itself overlain unconformably by marine rocks of the Centinela Formation [30][31][32]. Stratigraphic correlation of the fossil locality is extremely difficult due to limited outcrop area and local cover. All radioisotopic ages listed below are as originally reported and would need recalibration with new constants and reanalyses with updated methods for any detailed analysis. 40 K/ 39 Ar dates derived from ash beds at the top of the Río Tarde Formation at Lago Posadas (100 km south of our study area) were 97.163.8 and 99.165.6 Ma [33]. Whole-rock 40 Ar/ 39 Ar analyses of an altered ash bed from the Centinela  Formation south of Calafate (420 km to the south) yielded a range of ages with large scatter, from which the authors suggested a best estimate of 4662 Ma [34]. This suggests a middle Eocene minimum age for the Centinela transgression in western Santa Cruz and thus of the fossil flora studied here. However, it is not certain that the Centinela exposures in our study area are correlative with those dated [34]. Our attempts to date basaltic intrusions superposed above the Río Zeballos locality did not yield informative analytical results (B. Jicha, pers. comm. 2012), although some basic intrusions in the area are associated with the Los Antiguos Teschenite and the Posadas Formation. The Los Antiguos Teschenite intrudes the Río Tarde Formation and yielded early-middle Eocene 40 K/ 39 Ar ages of 4663 and 4864 Ma [35][36][37][38]. The basalts from the Posadas Formation were dated on the Argentinean side to 43.567 Ma [33]. At the Chilean type locality [26], the Ligorio Márquez coal mine (Fig. 1), the Ligorio Márquez Formation comprises a c. 55 m thick succession of alternating subhorizontal beds of mudstones, quartz-rich sandstones and thin coals, unconformably underlain by Lower Cretaceous tuffs, the Flamencos Tuffs, and overlain by basalts with a 40 K/ 39 Ar age on plagioclase of 47.660.78 Ma above the mine [27] but which elsewhere range in age from c. 57 Ma to c. 41 Ma [26,27,38,39].
In summary, all the geochronologic evidence, while greatly in need of revision, is most consistent with an early middle Eocene (Lutetian) minimum age for the fossil flora at Río Zeballos. This inference is best supported by the recent 40 Fig. 3E). In general, these taxa are indicative of regional microtherm to mesotherm rainforest vegetation [40]. Angiosperm pollen (Figs. 3J-L, 4) includes mesotherm to possible megatherm taxa such as Proxapertites sp. (Araceae/Arecaceae; Fig. 3L) and Bombacacidites (bombacoid Malvaceae; Fig. 4B). Pollen of the microtherm to mesotherm rainforest genus Nothofagus was not recorded. This absence of a taxon that is usually very abundant when present, from an otherwise diverse palynoflora, also implies that our samples are most likely to be early Eocene in age, an interval that in Patagonia has long been noted to lack Nothofagus, and certainly no younger than middle Eocene (e.g., [41][42][43][44][45][46]).
The macroflora is currently the subject of a separate study, but so far, Dacrycarpus (presumably corresponding to the Dacrycarpites australiensis pollen) has been identified as well as a possible cycad and many angiosperms, including several species of Lauraceae, a family that is also well represented in Chilean samples of the Ligorio Márquez Formation [47]. In accord with the palynological data, Nothofagus macrofossils are absent. Dacrycarpus macrofossils are otherwise known in Patagonia only from early and middle Eocene strata [48], although Dacrycarpites australiensis pollen is present in the region until the Miocene [49].
Although more study is clearly needed, the combined geological and paleobotanical evidence suggests an early Eocene age for the Río Zeballos material studied here, and most conservatively, its age lies within the late Paleocene to early middle Eocene interval.
Illustrations. Figs Age. Late Paleocene to early middle Eocene, and most probably early Eocene.
Distribution. So far known only from the Río Zeballos locality, Santa Cruz, Patagonia, Argentina.
Remarks. The specimens from the Ligorio Márquez Formation differ from Dilwynites granulatus (Figs. 2G, H) and D. tuberculatus (Figs. 2I-K), described from the Danian to Selandian (early to middle Paleocene) Pebble Point Formation, South Australia [10], in that the coarse ornamentation of the fossils studied here consists of clavae (which may appear as gemmae in poorly preserved specimens) rather than granula and verrucae-tuberculae, respectively. The exines of the new fossils are also thinner. Pseudolaesurae created by folding superficially resemble the trilete apertures on baculate-rugulate spores assigned to Baculatisporites, e.g. B. turbioensis Archangelsky in Argentina (see Fig. 5A in [50]). Lauraceae pollen is similar to that of Araucariaceae in being spheroidal and inaperturate, but it typically differs from Dilwynites in having ornamentation that consists of regularly-spaced, sharply pointed echinae, spinulae or foveolae (see e.g., Plate 33, images 391-393 in [51]). Moreover, it is well known that Lauraceae pollen is generally absent from fossil assemblages, mostly because it has only very thin exine with little sporopollenin [52], and probably very low production [53].
Here, we infer that the Dilwynites specimens from the Ligorio Márquez Formation potentially indicate the past presence of Wollemia in Patagonia because the specimens closely resemble pollen of W. nobilis [6,11,12]. Alternatively, the fossil pollen could be attributed to Agathis because it has recently become apparent that several extant species of Agathis produce grains ornamented with relatively coarse granules (M.K. Macphail, unpublished data) and thus would be accommodated within Dilwynites if found as fossils. Examples of Agathis species producing this type of pollen include the New Caledonian A. ovata (Vieill.) Warb. (Fig. 2L) and A. moorei (Lindl.) Mast. Other extant Araucariaceae (and in sculptural elements that are similar to those of the Río Zeballos specimens (A-F), and which are more pronounced and more widely spaced than in D. granulatus (G, H). I, Cethana, Tasmania (early Oligocene). J, Ti-tree Basin, Northern Territory (early Eocene). K, Lowana Rd, Tasmania (early Eocene). L, Agathis ovata recent specimen from Mts. des Koghis (Queensland Herbarium specimen AQ 391532: W.G. Ziarnik 34), New Caledonia. Note strong similarity to Dilwynites spp. Scale bars: 10 mm. doi:10.1371/journal.pone.0069281.g002 particular Araucaria) pollen most obviously differ in having much less prominent surface ornamentation [12]. Further refinement of relationships between Dilwynites and extant taxa may be possible following more detailed comparisons.

Discussion
The Ligorio Márquez Formation specimens of Dilwynites are the first known record of Wollemia-type pollen in South America. It is uncertain whether the newly recognized Patagonian clavate morphotype of Dilwynites represents a new species, given the wide range of variation observed in the granulate sculptural elements characterizing D. granulatus and the baculate to tuberculate sculptural elements characterizing D. tuberculatus populations in Australia (see Figs. 2G-K). The same is true from preliminary observations of other gymnosperm pollen taxa in the Ligorio Márquez Formation sample, which differ from the ranges of morphologies observed in Australian populations and those recorded from the Falkland (Malvinas) Islands (compare Figs. 3D-I with, e.g., Fig. 21 in [54]). A not unreasonable conclusion is that degrees of geographic differentiation occurred over the very long distances of these plants' ancient ranges.
Both Araucaria and Agathis have substantial macrofossil records in the Southern Hemisphere, but there is no strong macrofossil evidence for Wollemia (reviews [55][56][57][58][59]). Araucaria occurs extensively in Patagonia from the Early Jurassic to present, in West Antarctica from the Jurassic or Early Cretaceous to Eocene, and in Australia and New Zealand from the Early Cretaceous. The much more fragmentary Agathis record formerly came only from Cenozoic Australia and New Zealand, but abundant macrofossil Agathis specimens from the early and middle Eocene of northwestern Patagonia are now being described [60]; these include pollen cones, but pollen grains are not preserved within them. Reliable macrofossil evidence for Agathis is so far unknown from the Mesozoic [7,55,57].
At present, the macrofossils most likely to have close affinity to Wollemia are leaves of Araucarioides from Australia and New Zealand [6], and it is especially interesting that at the early Eocene Lowana Road site in Tasmania, these leaves co-occur with relatively abundant Dilwynites tuberculatus pollen [61]. As stated previously, recent molecular and reproductive data resolve Wollemia and Agathis as likely sister taxa [18][19][20][21][22][23][24]. This evidence, combined with the fact that Dilwynites first appears in the fossil record much later (Turonian: Late Cretaceous) than Araucaria suggests that at least some Mesozoic fossils that cannot be assigned to Araucaria can now be regarded as belonging to the stem lineage of the Agathis+Wollemia clade [24]. These fossils include winged seeds and cone scales with seed detachment scars from the Early to mid-Cretaceous in southeastern Australia [6,62], New Zealand [63] and Alexander Island, West Antarctica [64]. It should also be noted that at least some of the pollen included in the generalized, widespread form Araucariacites australis Cookson, which extends to the Triassic in the Southern Hemisphere, and which broadly accommodates pollen of modern Araucaria and many Agathis (e.g. [15]), could have been produced by extinct close relatives of Agathis and Wollemia.
The only extant Araucariaceae in South America are Araucaria angustifolia (Bertol.) Kuntze, native to southern Brazil and northeastern Argentina, where it is a dominant in temperate to subtropical rainforest, and A. araucana (Molina) K. Koch, native to Andean central and southern Chile and western Argentina between latitudes c. 37 to 40uS, where it associates with Nothofagus spp. to form mixed forests above c. 600-900 m elevation. The two species are apparently the survivors of the considerably more diverse Mesozoic araucarian flora of South America, represented by wood, foliage, cone, and pollen material (e.g., [58,[65][66][67][68]). This flora reached its maximum diversity and dominance in Patagonia, where araucarians were often co-dominant with cheirolepidiaceous conifers [69,70] during the Jurassic to Early Cretaceous. By the early and early middle Eocene, Araucaria and Agathis were abundant, but not diverse in Patagonia, occurring in association with crown group Podocarpaceae and Cupressaceae conifers with Australasian affinities and very diverse angiosperms [48,60,71,72]. However, so far as is known, none of the previously reported South American fossil species is comparable to Wollemia.

Materials and Methods
Blocks of wet sediment containing abundant mummified leaves (under separate study) were collected 3-4 May, 2011 at the Argentine Ligorio Márquez Formation outcrop (Fig. 1), wrapped in plastic to minimize water loss, and temporarily stored in a large refrigerator. Sediment samples selected for palynological processing were then oven-dried. Microfossils were extracted and replicate slides prepared by M. Rueda, Paleoflora Ltd, Bucaramanga, Colombia, using standard protocols. Microfossils were examined and photographed at ANU, Canberra, Australia using a Leica Axiophot transmitted light microscope fitted with AxioVision image capturing software. Residues were also examined and microfossils photographed at La Plata University, La Plata, Argentina using a JEOL JSM-6360LV scanning electron microscope operated at 10 kV. Adobe Photoshop Elements 6.0 software was used to optimise brightness and contrast of images, and to compose figures.