ABSTRACT
ABSTRACT
A new species of fossil fern with in situ spores, Palaeosorum waipiata (Polypodiaceae), is described and figured for a microsoroid fern frond from earliest Miocene sediments in Otago, New Zealand. The fertile frond from the Foulden Maar fossil Lagerstätte is pinnatifid with deep narrow lobes with entire margins and bears circular sori without indusia in rows on either side of the lobe midrib. Monolete spores associated with the frond are differentiated from similar, widespread dispersed spores of Polypodiisporites radiatus in possessing very small proximal verrucae/granulae around the laesurae and more rounded verrucae and the description of Palaeosorum is expanded to include information about sori and spores. This is the first confident Southern Hemisphere record for a microsoroid fern macrofossil, as well as the first with in situ spores. It is probable that this fern was epiphytic on trees or lithophytic on rocks adjacent to the Miocene maar lake.
Introduction
Ferns are a distinctive component of present-day New Zealand rainforests and exhibit diverse habits and life forms ranging from tree ferns to epiphytes (Brownsey & Smith-Dodsworth 2000; Brownsey & Perrie 2014). Ferns have a long history in New Zealand based mainly on abundant and widespread fossil spores (Mildenhall 1980; Cieraad & Lee 2006; Raine et al. 2011). Fern macrofossils are less common in the fossil record, but a number have been documented by Pole (1992, 2012); Cieraad & Lee (2006), Conran et al. (2010, 2017) and Homes et al. (2015). Many more fern macrofossils from New Zealand, ranging in age from Paleocene to Miocene, held in various collections such as the Geology Museum, University of Otago, Museum of New Zealand Te Papa Tongarewa, Wellington and GNS Science, Lower Hutt remain undescribed (Holden 1983; DEL & JGC pers. obs.). However, most of these macrofossils are sterile, so the discovery of a fertile frond of a polypodiaceous fern with well-preserved sori and spores is a significant advance in our knowledge of the antecedents of the modern New Zealand fern flora.
At Foulden Maar, a warm temperate to subtropical rainforest surrounded a small deep lake formed by a volcanic eruption in the earliest Miocene. Numerous leaf, flower and fruit macrofossils representing more than 30 families of angiosperms have been recovered from alternating light (summer) and dark (winter) varved lacustrine diatomite deposited in the paleolake (Lee et al. 2016). In contrast only four fern macrofossils have been found, including two fragmentary and sterile fronds and two fertile fronds with in situ spores. One of these was described by Conran et al. (2010) as Davallia walkerii Conran, U.Kaulf., Bannister, Mildenh. & D.E.Lee with in situ spores of Polypodiisporites radiatus Pocknall & Mildenh.
One reason for the scarcity of fern macrofossils at Foulden Maar is that fern fronds, unlike angiosperm leaves, commonly die and decay attached to the plant (Drake & Burrows 1980; Ferguson 1985) and are thus unlikely to be transported and preserved in lacustrine settings. However, both Davallia (L.) Sm. and Microsorum Link have fronds that detach at senescence and as both genera can grow as lithophytes or epiphytes (Brownsey & Perrie 2014), could have been blown into the lake from rocky outcrops at the maar edge, or from overhanging trees. In contrast, vast numbers of locally-derived, wind-blown fern spores in the diatomaceous sediment include at least 25 fern and lycophyte sporomorphs (Mildenhall et al. 2014). Of these, five species belong to the form genus Polypodiisporites R.Potonié which has affinities with Davalliaceae and Polypodiaceae, amongst other fern families (Raine et al. 2011).
The extant New Zealand fern flora includes four indigenous genera of Polypodiaceae sensu PPG1 (2016): Loxogramme (Blume) C.Presl, Microsorum, Notogrammitis Parris and Pyrrosia Mirb. (Brownsey & Perrie 2014; Schönberger et al. 2018). New Zealand Polypodiaceae are commonly characterised by ‘long-creeping rhizomes, simple to once-divided fronds, reticulate venation and sporangia that are either arranged in round to oval, bulging sori, or scattered over much of the lamina surface’ (Brownsey & Perrie 2014, p. 1).
The mainly tropical genus Microsorum is represented by three species in New Zealand: the endemic M. novae-zealandiae (Baker) Copel., M. pustulatum (G.Forst.) Copel. and M. scandens (G.Forst.) Tindale, the latter two shared with Australia (Brownsey & Perrie 2014). However, molecular studies have found that Microsorum s. lat. is polyphyletic, with the Australasian members of the genus forming a well-supported basal grade that also includes Colysis ampla (F. Muell. ex Benth.) Copel. below the ant-fern genus Lecanopteris Reinw. The nomenclature of the microsoroid complex is still unresolved (Schneider et al. 2006; Testo & Sundue 2014).
Pole (2012, p. 473), in a checklist of plant macrofossils from New Zealand, listed three specimens that had been assigned to Polypodiaceae. The first is ‘Polypodium’ hochstetteri Unger, which was regarded by Unger (1864) as latest Jurassic or earliest Cretaceous in age. This taxon is now considered to be of Jurassic age and was placed into synonymy with Cladophlebis Brongn. by Arber (1917) and C. cf. indica (Oldham & Morris) Sahni & Rao in Johnston et al. (1987), but without affinity. The second taxon, ‘Psaronius’ huttonianus Crié (1889), was described from petrified tree fern stem sections of Jurassic age collected at Mataura and ‘Toïtoï’ (=Toetoes Bay) in Southland, but Arber (1917) related this fossil to Osmundites (Pettko) Unger (now Osmundacaulis (C.N.Miller) Tidwell) in the Osmundaceae. A third taxon, Platycerium morgani W.R.B.Oliv., was described from two fragmentary leaf impressions from a diatomite deposit at Ormond, near Gisborne, based on unusual ‘anastomosing veins of two orders’ (Oliver 1928, p. 291, Figure 3). The specimens are now known to be of mid-Pleistocene (∼620 ka) age (McDowall et al. 2006), rather than ‘Later Pliocene’ as given in Oliver (1928), but the specimens are too incomplete to allow for placement with certainty into a family.
The present paper records a new fossil species of Palaeosorum (Polypodiaceae) with in situ spores from the early Miocene Foulden Maar Fossil-Lagerstätte. This is the first confirmed record from New Zealand of a fertile fossil fern in the family Polypodiaceae. Microsorum-like ferns have not been recorded previously in New Zealand from before the Pliocene, apart from a possible sterile frond macrofossil referred to Phymatosorus diversifolius (Willd.) Pic.Serm. (=Microsorum pustulatum) from the Miocene Longford Formation (Holden 1983); however, that sample was not described and the figured material does not include features necessary to identify it as Polypodiaceae. Earlier reports of Microsorum listed in Cieraad & Lee (2006) were for spores thought to belong to the genus. To our knowledge, no other macrofossils attributable to microsoroid ferns are reported from Australasia.
Material and methods
Locality and age
The fossil was collected at Foulden Maar, near Middlemarch in Otago, southern New Zealand (45.5269°S, 170.2191°E). The biogenic, finely laminated and fossiliferous diatomites at the site were deposited on the floor of a small (∼1000 m diameter) and deep (>200 m) maar lake, in a crater formed by phreatomagmatic eruptions c. 23 Ma, with the surrounding crater rim covered subsequently by rainforest (Lindqvist & Lee 2009; Lee et al. 2016; Kaulfuss 2017). The fossil macroflora includes numerous leaf compressions, usually with cuticle, flowers, fruits, seeds and occasional bark and twig fragments, from taxa that are indicative of a warm-temperate to subtropical paleoclimate (Pole 1996; Bannister et al. 2012; Conran et al. 2014; Mildenhall et al. 2014; Lee et al. 2016). Pollen biostratigraphy supports an earliest Miocene age for the fossil deposit (Mildenhall et al. 2014).
Fern macrofossil
The compressed fern is preserved as part and counterpart on a dark (winter) bedding plane in weathered diatomite (Figure 1). Photographs of the fossil, including close-up images at high resolution were taken with a Canon EOS 1100D camera and photomicrographs with a Canon T3 camera (Canon Inc., Tokyo, Japan) attached to a Nikon SMZ1000 stereomicroscope (Nikon Corporation, Tokyo, Japan). Close up images of the fossil were also taken using different colour filters, as well as side-lighting and the images where then subjected to contrast manipulation in Photoshop CS5 Extended® version 12.1 × 64 (Adobe Systems Inc., San Jose, California) to try to enhance the venation.
Published online:
10 January 2019Figure 1. Palaeosorum waipiata sp. nov.; holotype OU35184. A, part (OU35184a); B, counterpart (OU35184b); C, interpretational drawing based on part and counterpart; D, close-up of a pinna showing sori and venation; E, interpretational drawing of D. Scale bars: A–C = 20 mm, D, E = 2 mm.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2019/tnzb20.v057.i01/0028825x.2018.1560336/20190227/images/medium/tnzb_a_1560336_f0001_oc.jpg"})
Figure 1. Palaeosorum waipiata sp. nov.; holotype OU35184. A, part (OU35184a); B, counterpart (OU35184b); C, interpretational drawing based on part and counterpart; D, close-up of a pinna showing sori and venation; E, interpretational drawing of D. Scale bars: A–C = 20 mm, D, E = 2 mm.
Kvaček (2001) pointed out the features preserved in fossil ferns that can be used to distinguish families and this has been expanded upon variously for polypodioid ferns by Jacques et al. (2013), Sundue & Poinar (2016) and Xie et al. (2016). These characteristics include sorus shape and position along the frond, presence or absence of an indusium and the type of spore morphology, amongst others. The fossil was therefore compared with living and fossil taxa for diagnostic features in order to justify its possible identification and affinities. Descriptions for morphologically similar extant taxa were derived from Kramer & Green (1990), but classified using family placements in PPG1 (2016).
Spores were removed from the sori using a needle and fine paint brush, placed in a small dish in warmed, dilute hydrogen peroxide to break up any clumps of spores and clear any residual material. The spores were then rinsed in water before mounting on SEM stubs and in thymol glycerine jelly on slides. Light-micrographs were taken with a Leica DM 1000 microscope and LAS software. Spores were examined and measured, using a 100× oil immersion lens, 10× objective, optivar 1.25, on Zeiss Axioplan 2 imaging light microscope number FA 1609 at GNS Science, Lower Hutt. The grains examined under the SEM (JEOL FE-SEM6700) are much smaller (by about 10 µm on average) and the measured specimens on the mounted slides are over-expanded due to processing during extraction from the matrix. Therefore, the measurements represent maximum possible sizes, even though the specimens are at different stages of development. Only the grains assumed to be fully mature and those with small proximal verrucae/granulae were counted.
Fern nomenclature follows Jacques et al. (2013). The fern fossil and spore samples are held in the collections of the Department of Geology, University of Otago (OU, see descriptions for reference numbers). Vouchers for modern taxa used for comparison are located at the Otago Regional Herbarium (OTA). All photographs are by the authors unless indicated otherwise.
Results
The Foulden Maar fossil was compared with genera bearing similar frond and spore characteristics. The pinnatifid frond morphology, circular sorus shape and laminar position, lack of an indusium and monolete, verrucate spores () all support a placement in Polypodiaceae (Brownsey & Perrie 2014). Overall frond, sorus and spore morphology tend to support the placement of the fossil within subfamily Microsoroideae, in the Australasian lecanopteriod clade of Schneider et al. (2006), Kreier et al. (2008) and Testo & Sundue (2014), although there are possibly other taxa from outside the region that may also show these characteristics (Jacques et al. 2013).
Some grammitid species bear pinnatifid fronds with similar sori (e.g. the New World Grammitis aspleniifolia (L.) Proctor), although most have entire fronds and the spores are globose-tetrahedral and trilete (Parris & Given 1976; Parris 1990). Similarly, Polypodium L. also has some species with similar overall frond, sorus and spore morphology to the fossil (Kramer & Green 1990), but the genus has different frond venation (Mitsuta 1983, 1984a, 1984b). Jacques et al. (2013) defined Palaeosorum F.M.B.Jacques & Z.K.Zhou for fossil microsoroid fern macrofossils on the presence of characteristic anastomosing reticulate venation (Nooteboom 1997, 1999); however, Drynarioid ferns can also display very similar mature venation to Microsorum, though there are differences in juvenile frond venation (Mitsuta 1984).
Unfortunately, the Foulden Maar fossil lacks sufficient venation preservation to allow definitive placement into a living genus. As a result, it is placed here into Palaeosorum, as this genus was erected for fossil microsoroid ferns of uncertain generic affinity and the generic description is amended to incorporate new information about frond and reproductive morphology.
Systematic paleontology
Class. Polypodiopsida (=Filicopsida) Cronquist, Takht. & Zimmerm. ex. Reveal, 1995
Order. Polypodiales Link, 1833
Family. Polypodiaceae J.Presl & C.Presl, 1822
Genus. Palaeosorum F.M.B.Jacques & Z.K.Zhou, 2013
Type species. Palaeosorum ellipticum F.M.B.Jacques & Z.K.Zhou, 2013
Emended description. Leaf simple, long elliptic to deeply 1-pinnatifid, petiolate; base acute; apex rounded to acute; frond or lobe margins entire. Venation reticulate with a strong costa; connecting veins forming a row of equally sized, polygonal areoles between two adjacent veins straight to slightly zigzag, with no prominent veinlets parallel to the veins, generally with 7–8 parallel connecting veins between two adjacent veins. Veinlets several per areole, free to once forked, ending in a club-shaped hydathode. Sori prominent, exindusiate, round to slightly elongated, superficial or impressed, in one or more rows throughout the lamina on either side of the costa, not confluent with age. Spores monolete, bilaterally symmetrical, granulate with rounded verrucae.
Species. Palaeosorum waipiata U.Kaulf., Conran, Bannister, Mildenh. & D.E.Lee, sp. nov. (Figure 1)
Holotype (Figure 1). Partial frond, 2017 (OU35184a (part) OU35184b (counterpart); Geology Museum, University of Otago, Dunedin, New Zealand.
Paratypes. A. OU35187a; co-ordinates 16.4 × 98.2, England Finder Reading J35/2; slide reference point (lower left-hand corner of coverslip as seen under the microscope) 02.2 × 93.5. B. OU35187b; co-ordinates 11.3 × 105.9; England Finder Reading O34; C. OU35187c; co-ordinates 20.0 × 105.9, England Finder Reading E43.
Type locality and stratigraphic horizon. Mining pit in early Miocene-aged Foulden Hills Diatomite at Foulden Maar, near Middlemarch, Otago. The specimen was collected in the uppermost 10 m near the centre of the crater-filling sedimentary sequence. The locality is registered as I43/f8503 in the New Zealand Fossil Record File administered by the Geological Society of New Zealand and GNS Science.
Collector. U Kaulfuss Foulden leaf 85, May 2017.
Diagnosis. Fronds deeply pinnatifid, sori circular, laminal in single rows, lacking indusia. Spores monolete, morphologically similar to Polypodiisporites R.Potonié, but characterised by very small proximal verrucae/granulae around the laesurae and more rounded verrucae.
Description. The partial fern frond is pinnatifid with deep, narrow lobes with entire margins (Figure 1A–C); rhizome and stipe not discernible. Length of (incomplete) blade 102.2 mm; pinnae 6.5–10 mm wide, the only completely preserved pinna 41.1 mm long. Venation very poorly preserved, midrib prominent, some lateral veins visible (Figure 1D,E); higher venation and areoles not preserved. Sori superficial, solitary, circular, 1.4–1.9 mm in diameter, without indusia, arranged in rows on either side of the lobe midrib, either midway between the midrib and the margin or slightly nearer the margin (Figure 1B). Circular or elliptical clusters of 22+ sporangia within sori (Figures 1D and 2A).
Published online:
10 January 2019Figure 2. Morphology of monolete spores of Palaeosorum waipiata (A–G) and spores of extant M. pustulatum subsp. pustulatum for comparison (H–I). A, P. waipiata sorus (OU35184a); B–D, P. waipiata TLM photomicrographs of paratype (OU35187a); E–G, P. waipiata SEM images showing spore surface detail (OU35187b); H, I, M. pustulatum subsp. pustulatum SEM images showing spore morphology and surface detail (P. Johnson s.n. OTA071151). Scale bars: A = 1 mm, B–E = 20 μm, F, I = 5 μm, G = 1 μm.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2019/tnzb20.v057.i01/0028825x.2018.1560336/20190227/images/medium/tnzb_a_1560336_f0002_oc.jpg"})
Figure 2. Morphology of monolete spores of Palaeosorum waipiata (A–G) and spores of extant M. pustulatum subsp. pustulatum for comparison (H–I). A, P. waipiata sorus (OU35184a); B–D, P. waipiata TLM photomicrographs of paratype (OU35187a); E–G, P. waipiata SEM images showing spore surface detail (OU35187b); H, I, M. pustulatum subsp. pustulatum SEM images showing spore morphology and surface detail (P. Johnson s.n. OTA071151). Scale bars: A = 1 mm, B–E = 20 μm, F, I = 5 μm, G = 1 μm.
Spores bilateral, anisopolar, usually reniform, plano-convex to concavo-convex in equatorial view, ends rounded and of equal sizes; elliptical in polar view; contact faces concave to straight; monolete (Figure 2B,E); laesurae usually from one-half (mostly) to two-thirds (occasionally) spore length, 28–39 μm long (Figure 2C), gaping when under pressure; thin, psilate membranous margins; perine not preserved; exine 2–4 μm thick including verrucae (Figure 2D), 1–1.5 μm between verrucae, distally verrucate to proximately granulate, verrucae 5–10 μm in size on distal face, separated by narrow channels ca 1–2 μm wide sometimes containing small granulae or scabrae at meeting points, verrucae occasionally fused to become elongate; granules 1–2 μm in size on proximal face lining laesurae; verrucae circular to sub-circular in shape, occasionally slightly angular, gently rounded in distal optical section, tending to be slightly elongate between poles (Figure 2B,C). Under SEM, grooves between verrucae are psilate, surfaces of verrucae are psilate to faintly flecked and fused verrucae are quite variable in shape; laesurae lined by ±10–12 granulae (Figures 2E–G). Dimensions: equatorial length 52–76 μm (mean 65 μm); polar depth 36–53 μm (mean 46 μm) (25 specimens measured); P:E ratio 5.2:8.
Etymology. The species is named for the Waipiata Volcanic Field in which Foulden Maar is located.
Age. Earliest Miocene, ∼23 Ma; Aquitanian, New Zealand local stage: Waitakian.
Material examined. Part and counterpart of a partially preserved fertile frond, OU35184 a, b.
Remarks. Distinctive features of microsoroid ferns generally include a single row of sori (sometimes two) between the costa and margin, a relatively low number of pinnae with entire margins, width of the pinnae and the size and shape of the sori, all features displayed by the fossil. Anastomosing veinlets and forked free veins ending in hydathode are also regarded as defining features of both microsoroid ferns and Palaeosorum (Jacques et al. 2013). Within Microsorum s. lat. most species of the lecanopteroid clade studied by Schneider et al. (2006), Kreier et al. (2008) and Testo & Sundue (2014) display a characteristic, complex, anastomosing venation pattern (Type 5b) at least in the fertile fronds (Hetterscheid & Hennipman 1984; Bosman 1991; Nooteboom 1997, 1999; Bosman et al. 1998), whereas members of the other clades containing taxa from this polyphyletic genus showed different patterns (Schneider et al. 2006). Unfortunately, despite several attempts to view higher order venation on the fossil using differential colour imaging, side-lighting and contrast manipulation of high resolution images, the preservation was insufficient to recover enough detail for detailed comparison.
Discussion
Polypodiaceae (including Grammitidaceae) are the second most diverse group of living ferns (PPG1 2016; Testo & Sundue 2016), based in part on the high number of mainly epiphytic species and their ‘vast range of morphological variation’ (Kreier et al. 2008, p. 1155). However, research on relationships and diversification within the family have been hindered by the poorly known fossil record for this group (Jacques et al. 2013). When the fossil is compared morphologically with descriptions of fronds of extant Polypodiaceae (Hennipman et al. 1990), it matches closely to Microsorum (Figure 3), particularly Australasian taxa such as M. pustulatum from Australia and New Zealand (Bostock & Stokes 1998; Brownsey & Perrie 2014) and M. vieillardii from New Caledonia (Brownlie 1969).
Published online:
10 January 2019Figure 3. Extant Polypodiaceae species for comparison with the fossil foliage. A, Microsorum novae-zelandiae (Nga Whanau A Ruapani Tarns, Te Urewera, M. Thorsen, NZPCN, used with permission); B–E, M. pustulatum subsp. pustulatum (B, D, Dunedin Botanic Gardens; C, E, P. Johnson s.n. OTA071151); F–I, M. scandens (F, Dorrigo National Park, NSW; G–I, P. Johnson s.n. OTA071152). Scale bars: A, B, F = 10 cm; C, G = 5 cm; D, E, H, I = 5 mm.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2019/tnzb20.v057.i01/0028825x.2018.1560336/20190227/images/medium/tnzb_a_1560336_f0003_oc.jpg"})
Figure 3. Extant Polypodiaceae species for comparison with the fossil foliage. A, Microsorum novae-zelandiae (Nga Whanau A Ruapani Tarns, Te Urewera, M. Thorsen, NZPCN, used with permission); B–E, M. pustulatum subsp. pustulatum (B, D, Dunedin Botanic Gardens; C, E, P. Johnson s.n. OTA071151); F–I, M. scandens (F, Dorrigo National Park, NSW; G–I, P. Johnson s.n. OTA071152). Scale bars: A, B, F = 10 cm; C, G = 5 cm; D, E, H, I = 5 mm.
The record of polypodiod fern macrofossils globally was summarised in Jacques et al. (2013) and Xie et al. (2016). However, very few, well-substantiated specimens have been found and all, up to now, have been from the Northern Hemisphere. Collinson (2001, p. 173) commented that ‘Cainozoic records of … Polypodiaceae are here considered as unconfirmed’, noting that the pyrite-permineralised fragment of rachis from the Eocene London Clay flora attributed to Polypodiaceae by Poole & Page (2000) could also belong to other taxa. In 2001, Kvaček described a new fossil species of Polypodium from the Oligocene of the Czech Republic, based on a frond with well-preserved in situ spores. Kvaček et al. (2004) re-examined records of Polypodium of Cenozoic age from North America and concluded that only a single fossil species (based on fertile material) could be accepted as Polypodium. Well-preserved, fertile grammitid fern macrofossils have also been described from Cenozoic Dominican amber as the extinct taxon Polymniopteris succinea (L.D.Gómez) Sundue & Poinar (Gómez 1982; Sundue & Poinar 2016). The first fossil microsoroid fern macrofossil to be described formally was Palaeosorum ellipticus F.M.B.Jacques & Z.K.Zhou from the middle Miocene of China (Jacques et al. 2013), with recent reports of Miocene Chinese species of Neolepisorus Ching (Xie et al. 2016) and Goniophlebium (Blume) C.Presl. (Xu et al. 2017).
Three fern macrofossils are now known from the earliest Miocene at Foulden Maar: a fertile frond of Davallia walkerii (Davalliaceae), a Davallia-like sterile frond (Conran et al. 2010) and now P. waipiata sp. nov. (Polypodiaceae) described herein. However, the spore assemblage from the site indicates a much more diverse fern flora, including taxa in Cyatheaceae, Dennstaedtiaceae, Dicksoniaceae, Gleicheniaceae, Grammitidaceae, Lycopodiaceae, possible Marattiaceae, Osmundaceae, Psilotaceae, Pteridaceae and Schizaeaceae (Mildenhall et al. 2014). This discrepancy suggests it is likely that further macrofossil fern taxa will be present.
The in situ miospores of the fossil differ from those of previously described palynomorphs for the monolete dispersed spore genera Polypodiisporites and Verrucatosporites H.D.Pflug & P.W.Thomson (). They differ from P. radiatus in having very small (1–2 μm) proximal verrucae/granulae around the laesurae and more rounded verrucae in general. They are smaller than P. inangahuensis (Couper) R.Potonié emend. Pocknall & Mildenh., larger than P. minimus (Couper) A.M.Khan & A.R.H.Martin emend. Pocknall & Mildenh., have less pronounced verrucae than P. perverrucatus (Couper) A.M.Khan & A.R.H.Martin emend. Pocknall & Mildenh. and larger verrucae than V. speciosus T.M.Harris and V. rugufavus Krutzsch (Raine et al. 2011).
The fossil also displays quite different sculpturing from the scabrae and granules of P. variscabratus Mildenh. & Pocknall, which had also been described previously as P. irregularis Pocknall & Mildenh. (Pocknall and Mildenhall 1984; Mildenhall & Pocknall 1989) and attributed to Microsorum by Cieraad & Lee (2006), apparently in error. The nearest modern equivalent to the fossil appears to be M. pustulatum (G.Forst.) Copel., but whose spores have lower, more rounded verrucae (Figure 2H, I; see also Large & Braggins 1991).
Polypodiisporites spores are very common in Cenozoic sediments in New Zealand (Cieraad & Lee 2006; Raine et al. 2011) and Carpenter et al. (2004) attributed spores of P. perverrucatus from the Eocene of Tasmania to Microsorum. However, based on the range of morphotypes displayed within Polypodiisporites miospores, their botanical affinities could potentially not only include several modern genera within Polypodiales, especially Polypodiaceae, Davalliaceae and Dennstaedtiaceae, but also some Schizaeaceae (Schizaeales) and Psilotaceae (Psilotales) (Raine et al. 2011).
In general, the range of spore morphology seen within modern species in these fern families is consistent. However, differing percentages of spores within these species can still be more or less indistinguishable from those of other species. When polypodiaceous spores are mixed, as they are in the fossil record, a broader morphological range is generally accepted for each species of dispersed spores. It is only when abundant fossil spores from sporangia and fertile fronds are found that spore morphology can be defined more closely. This is the case with fertile fossil Davallia fronds (Conran et al. 2010) in which the spores were within the range of morphology accepted for the dispersed spore species P. radiatus. Similarly, the new miospore-bearing frond described here allows us to recognise dispersed spores with distinctly smaller verrucae around the laesurae as Palaeosorum waipiata. However, even within the sporangia of this fossil, 3%–5% of spores are indistinguishable from P. radiatus under a light microscope.
Biogeography and paleoecology
Fossil-calibrated divergence time estimates suggest that Polypodiaceae originated in the Early Eocene at ∼55.8 Ma (Schuettpelz & Pryer 2009). The earliest fossil referred to Polypodiaceae comes from the Eocene/Oligocene boundary in Russia and the family diversified and was widely distributed by the Oligocene and Miocene (Van Uffelen 1991; Sundue & Poinar 2016). Based on fossil records and molecular data it has been suggested that New Zealand pteridophytes include an ‘ancient’ pre-Gondwana break-up group and a ‘modern’ group, including all polypodiaceous ferns, that arrived post-Gondwana (Brownsey 2001; Perrie & Brownsey 2007). Some elements of the ‘modern’ group are shared with and may have arrived from Australia, although Schneider et al. (2006) suggested a close relationship of M. pustulatum and the New Zealand endemic M. novae-zealandiae with the New Caledonian endemic M. vieillardii. The sparse macrofossil record means that arrival times are difficult to ascertain. The finding of a fertile frond of Palaeosorum waipiata at Foulden Maar provides evidence for the existence of this genus in a mesothermal rainforest ecosystem in southern New Zealand in the earliest Miocene, shortly after major reduction of terrestrial habitats in the Oligocene (Lee et al. 2014).
Kreier et al. (2008, p. 1166) noted that ‘extensive studies of the biogeographic history of these ferns are needed to determine the area of origin and migration routes of the microsoroid ferns’. Testo & Sundue (2016) found that the microsoroids and several other major polypodioid lineages diverged in the early to middle Eocene (55–40 ma) close to the time of the Paleocene–Eocene Thermal Maximum (PETM), but after the time when Zealandia had separated from Australia and Antarctica. Similarly, the basal members of the microsoroid clade are largely Asian taxa, with the former Zealandian members nested as crown lineages within a broader Australasian lecanopteroid clade (Testo & Sundue 2014). The discovery of a well-dated fertile microsoroid fern from New Zealand provides new data in this story, reinforcing the idea that it is unlikely that New Zealand was an area of origin for microsoroid ferns. Nevertheless, we can state confidently that by 23 Ma, an extinct species of fern with very close resemblance to Microsorum was living around a small volcanic crater lake in southern New Zealand, suggesting that by at least the end of the Oligocene, microsoroid ferns had dispersed thousands of kilometres from a possible centre of origin in Southeast Asia (Kreier et al. 2008; Testo & Sundue 2014).
Poole & Page (2000), Collinson (2002) and Dubuisson et al. (2009) noted that there is little fossil evidence for epiphytic ferns in the Cenozoic and suggested that this life habit may have evolved relatively recently in more advanced ferns, although they found the absence of evidence for epiphytic Mesozoic ferns to be puzzling. Based on molecular studies, Testo & Sundue (2016) suggested that most basal polypodioid lineages probably occupied largely xeric niches, diversifying as rainforest epiphytes with the spread of these ecosystems around the PETM and Kreier et al. (2008) and Testo & Sundue (2014) noted that the majority of modern microsoroid ferns are either epiphytes or grow mainly on rocks.
Extant New Zealand Microsorum species mostly grow from sea level to subalpine environments, either terrestrially or as epiphytes and lithophytes, although M. novae-zelandiae is restricted to North Island where it grows as an epiphyte in damp montane to subalpine forests and scrub (Lehmann et al. 2002; Brownsey & Perrie 2014). Testo & Sundue (2014) found that the basal microsoroid lineages were all epiphytic, with the climbing and hemiepiphytic conditions seen in several modern Australasian members of the leconopteroid clade being derived characteristics, with the Australian hemiepiphyte C. ampla (F.Muell. ex Benth.) Copel. placed as sister to M. scandens.
It is unlikely that ferns would have been transported into the Foulden Maar paleolake unless they were perched on branches overhanging the lake edge or living on rock outcrops at the lake margin and the presence of Davallia and Palaeosorum fronds, together with our earlier reports of Astelia (Maciunas et al. 2011) and fossil orchid leaves (Conran et al. 2009) suggests that epiphytes and lithophytes were common in the rainforest around Foulden Maar.
| Palynomorph | Affinity | Age range |
|---|---|---|
| Polypodiisporites spp. | Polypodiaceae, Davalliaceae | |
| P. inangahuensis | Polypodiaceae, Davalliaceae | Middle Eocene–late Miocene |
| P. perverrucatus* | Polypodiaceae, Davalliaceae | Early–late Miocene |
| P. radiatus | Polypodiaceae, Davalliaceae | Early Miocene–middle Pleistocene |
| P. variscabratus | Psilotaceae (Tmesipteris), Polypodiaceae, Schizaeaceae | Late Oligocene–late Miocene |
*Allied to Microsorum in Carpenter et al. (2004).
Acknowledgements
We thank the Gibson family at Foulden Hills and RSC Mining Limited, Dunedin, for allowing access to the study site, and Liz Girvan, Otago Centre for Electron Microscopy, for help with preparing and analysing SEM samples. We particularly thank Dr Peter Johnson for providing us with examples of the various New Zealand Microsorum species and Mike Thorsen for permission to use his image of M. novae-zealandiae. We would also like to thank two reviewers and Dr Leon Perrie for their helpful comments on the manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.