Leaf fossils of the ancient Tasmanian relict Microcachrys (Podocarpaceae) from New Zealand^†

Abstract

• Premise of the study: Microcachrys tetragona (Podocarpaceae), endemic to the mountains of Tasmania, represents the only remaining taxon of one of the world's most ancient and widely distributed conifer lineages. Remarkably, however, despite its ∼150 Myr heritage, our understanding of the fossil history of this lineage is based almost entirely on the pollen record. Fossils of Microcachrys are especially important in light of recent molecular phylogenetic and dating evidence. This evidence dates the Microcachrys lineage to the Mesozoic and does not support the traditional placement of Microcachrys as sister to the southeastern Australian genus Pherosphaera.

• Methods: We undertook comparative studies of the foliage architecture, cuticle, and paleoecology of newly discovered fossils from the Oligo-Miocene of New Zealand and M. tetragona and discussed the importance of Microcachrys in the context of Podocarpaceae phylogeny.

• Key results: The fossils represent the oldest and first extra-Australian macrofossils of Microcachrys and are described as the new foliage species M. novae-zelandiae. These fossils confirm that the distinctive opposite decussate phyllotaxy of the genus is at least as old as the Oligo-Miocene and contribute to evidence that Microcachrys plants were sometimes important components of oligotrophic swampy habitats.

• Conclusions: Leaf fossils of Microcachrys closely comparable with the only extant species confirm that this lineage had a much wider past distribution. The fossil record and recent molecular phylogenetic studies, including that of Microcachrys, also serve to emphasize the important status of Tasmania as a refugium for seed plant lineages.

One of the most spectacular cases of a relictual plant is the Tasmanian endemic conifer genus Microcachrys (Podocarpaceae). A relictual organism may be defined as one that belongs to a lineage that is known to have been widespread in the past, but is now very restricted in distribution. Famous botanical examples include several other gymnosperms, notably Ginkgo, the dawn redwood (Metasequoia glyptostroboides Hu & W.C.Cheng) and the recently discovered Wollemi pine (Wollemia nobilis W.G.Jones, K.D.Hill & J.M.Allen) that is found only in isolated sandstone gorges west of Sydney, Australia (23). Fossil Wollemia-type pollen is known from sediments in Australia, New Zealand, and Antarctica and dates from ca. 90 million years ago (Ma) (36). However, Microcachrys has an even greater temporal and spatial heritage than Wollemia.

The single extant species of Microcachrys, M. tetragona (Hook.) Hook.f. produces highly distinctive, small, trisaccate pollen grains, whereas almost all other Podocarpaceae have bisaccate pollen. Similar trisaccate grains are first recorded from the Upper Jurassic (∼150 Ma) (13; 46). They include a range of extinct forms, but grains of at least two described fossil pollen species, Microcachryidites antarcticus Cookson and Podosporites parvus (Couper) Mildenh., conform very closely to the pollen of extant M. tetragona and are inferred to belong to the same lineage (e.g., 11; 12; 35; 42). Cenozoic records of these species occur widely across the southern hemisphere including Antarctica (46), Australia (e.g., 35), the Ninetyeast Ridge of the Indian Ocean (28), New Zealand (e.g., 42; 44), southern Africa (9) and South America (e.g., 1). However, despite the evidence from pollen for an extensive spatial and temporal range, pre-Quaternary macrofossil records of Microcachrys are essentially nonexistent. This absence of records is all the more surprising because M. tetragona has opposite and decussately arranged scale leaves that give the stem a square appearance in cross section, a highly distinctive state that is unique in Podocarpaceae, and superficially like some Cupressaceae. Microcachrys tetragona is now restricted to Tasmanian mountain thickets and boulder-fields, where it has a low shrub or prostrate habit.

Recent molecular phylogenetic analyses (Fig. 1) place Microcachrys as sister to a clade that contains the small Australian genus, Pherosphaera (until recently known as Microstrobos: 5), and a large and geographically more widespread clade of other taxa, including Podocarpus, Dacrycarpus, and Dacrydium (2). However, Microcachrys has traditionally been regarded as the sister of Pherosphaera, a relationship strongly supported in morphological and combined morphological and molecular (18S rDNA) cladistic studies by 26, 27). Features shared by these genera include pollen with a finely verrucate body and three sacci with few alveoli, unordered embryo cap cells and cleavage polyembryony, and stomata in bands (see 26). In addition, among other similarities such as the presence of leaf marginal frills, 47 found that Microcachrys and Pherosphaera are the only fully epistomatic genera of Podocarpaceae, and the only genera that have irregularly shaped polar and lateral subsidiary cells, either in contact or shared between or within rows.

Figure 1

Simplified dated molecular phylogeny of Podocarpaceae adapted from 2, showing median estimated node dates only. Proposed evolution of trisaccate and bisaccate pollen types in the Microcachrys + relatives clade is shown, as well as age ranges for important fossil Microcachrys-like pollen species in New Zealand. Note however that these ranges are not proof of continuous presence in the region.

This conflict between molecular and morphological evidence implies that morphological characters in Microcachrys and Pherosphaera need to be reassessed. At least in terms of foliar characters, the condition of stomata in bands should be regarded as plesiomorphic in Microcachrys and Pherosphaera because this feature is shared throughout the large clade to which both genera belong, as well as in Saxegothea. Leaf marginal frills are also not phylogenetically informative, because they are evidently correlated with close leaf imbricacy throughout the family, and may be a convergent response to severe climatic conditions at higher altitudes (19). Other shared features between the genera may be more difficult to interpret and should be investigated in detail elsewhere. In particular, there is a need for a systematic review of the pollen of extant species with respect to similar trisaccate fossil types (42). As noted by 10, problems in interpreting historical character evolution in Podocarpaceae are to be expected given the ancient age of the family and that most of the extant genera are the last remnants of once more diverse lineages.

Molecular dating derived from analyses of chloroplast (matk, trnL-F intron, and spacer) DNA sequences using the Bayesian relaxed clock implementation BEAST program suggests a Lower to mid-Cretaceous divergence between Microcachrys and its large sister clade (Fig. 1) and also between Pherosphaera and the rest of this clade (2). Reconciling both the phylogenetic evidence and the fossil pollen record implies an earlier (Jurassic) evolution of the trisaccate form in the common ancestor of Microcachrys and its large sister clade, evolution of a differentiated form of this pollen in Pherosphaera, then reversion to bisaccate pollen in the rest of the clade, with a further, independent evolution of a distinct form of trisaccate pollen in Dacrycarpus (Fig. 1).

Therefore, the Microcachrys-like fossil pollen may be plesiomorphic, and at least some of the ancient and widespread pollen types may have been derived from stem lineages. Similarly, the paucity of recognizable Microcachrys macrofossils could be attributed to relatively recent evolution of the modern foliage form, although this could have occurred at any time since the mid-Cretaceous or earlier.

In this paper, we present the first detailed descriptions of fossil Microcachrys foliage, from the Oligo-Miocene of New Zealand. These fossils confirm that Microcachrys had a much wider distribution and that the same type of imbricate, opposite, and decussate foliage as that of M. tetragona was present in an extinct species by this time. We conclude that the relictual status of Microcachrys in Tasmania is shared by a number of other seed plant lineages and that the region may have had a long history as a biotic refugium for wet-habitat taxa.

MATERIALS AND METHODS

The fossils were recovered from a thin leaf litter bed 5 m below the top of the 17 m thick Seam W6 of the middle Gore Lignite Measures (GLM), Newvale Mine, Waimumu, Southland (32; 14) that is most likely to be of latest Oligocene to earliest Miocene age (Waitakian on the New Zealand scale: 25.2–21.7 Ma) based on palynostratigraphy. Other leaf fossils that occur in the Newvale lignites include Araucariaceae (21; 32), epacrids (Ericaceae) (25), Gymnostoma (Casuarinaceae) (14), several Podocarpaceae (14) and diverse Proteaceae, including Persoonieae and Banksia (6, 8). As discussed in previous studies the leaf litter bed at Newvale is comprised of densely packed plant fragments. The specimens of Microcachrys in the present study were detected on these surfaces and catalogued using OU prefixes for the Department of Geology collection, University of Otago, Dunedin, New Zealand. Microscope slides and SEM stubs with material derived from the specimens are presently housed at the Universities of Adelaide and Tasmania. Cuticle preparations were made by soaking leaf fragments in household bleach (42 g sodium hypochlorite /L). The cuticles were then rinsed, cleaned with a fine paintbrush if necessary, and mounted on glass slides in glycerine jelly after staining with 1% aqueous safranin O for viewing with transmitted light microscopy (LM). Cuticles were photographed using an Olympus DP11 (Tokyo, Japan) digital camera mounted on a Zeiss Axioskop (Jena, Germany) microscope at the University of Adelaide. Other pieces were placed on double-sided adhesive tape on aluminum stubs and either sputter-coated with platinum and examined with an FEI Quanta MLA 600 (Hillsboro, Oregon, USA) environmental scanning electron microscope (SEM) operated at 15 kV at the University of Tasmania or carbon/gold coated and examined using a Philips XL 30 FEGSEM (Eindhoven, Netherlands) operated at 10 kV at Adelaide University.

The fossil leaves were compared with the large collection of conifer material held at the School of Earth and Environmental Sciences, University of Adelaide and the Department of Plant Science, University of Tasmania. These collections include cuticle slides of almost all species of scale-leaved Podocarpaceae and Cupressaceae that are comparable with the fossils.

RESULTS

Systematics

• Family: Podocarpaceae

• Genus: Microcachrys Hook.f.

• Species: M. novae-zelandiae R.J.Carp., G.J.Jord., Mildenh. & D.E.Lee (Figs. 2–9)

Diagnosis

Microcachrys foliage; each scale leaf with a marginal frill of finger-like cilia and ctenoidally arranged cells, often transitioning from cilia in basal half of the leaf to smooth frill in the apical half and apex. Florin rings prominent, segmented by grooves. Polar extensions present on cuticle associated with guard cells. Anticlinal cell wall cuticle not distinctly beaded or pitted.

Holotype

OU32896 (Figs. 2, 3, 8, 9), stored in the Department of Geology, University of Otago, Dunedin, New Zealand. Paratypes: OU32897 (Fig. 4), OU32898 (Figs. 5–7). Includes material mounted on slides and SEM stubs.

Figure 2–9

Specimens of fossil Microcachrys novae-zelandiae. 2. Holotype (specimen OU32896) in lignite matrix showing opposite and decussate phyllotaxy. 3. Scanning electron microscope (SEM) image of holotype (specimen OU32896). 4. SEM image of a broken shoot with the removal of a facing leaf, showing the basal junction of two lateral leaves. Cilia can be seen along the margins of the adjoining leaves (example arrowed), and there are no stomata near the leaf bases (specimen OU32897). 5. Light microscope image showing two partially cleared scale leaves (specimen OU32898). Note the marginal frill on the leaf at left, which transitions from cilia in the basal half of the leaf to more even ctenoidally arranged cells in the apical half, including around the leaf apex. 6. SEM image of inner cuticle from the adaxial leaf surface, showing files of stomata. Note that the polar and/or lateral cells of the stomata may be shared. Note also the polar extensions of cuticle associated with the guard cells, but little evidence of lateral extensions and that the epidermal cell anticlinal wall cuticle is only weakly beaded (specimen OU32898). 7. SEM image showing two stomatal complexes. Note the groove surrounding the cuticle associated with the guard cell of each stoma that corresponds with a Florin ring on the outer surface (specimen OU32898). 8. SEM image of the outer adaxial cuticle surface showing raised and disjointed or segmented Florin rings (holotype specimen OU32896). 9. SEM image of inner abaxial cuticle showing elongated epidermal cells and little evidence of anticlinal wall cuticle beading (holotype specimen OU32896). Scale bars = 2 mm for Fig. 2; 500 µm for Fig. 3; 300 µm for Fig. 4; 100 µm for Fig. 5; 50 µm for Fig. 9; 20 µm for Figs. 6–8.

Etymology

In recognition of a New Zealand species of Microcachrys.

Type locality

Newvale Mine, Waimumu Coalfield, Southland, New Zealand. Site registered as F45/f0394 in the New Zealand Fossil Record File administered by the Geoscience Society of New Zealand. The NZ Map Grid reference on Infomap Series NZMS 260 is F45/817434 (46.1427°S, 168.7518°E).

Age

Late Oligocene to early Miocene (Waitakian on the New Zealand time scale: 25.2–21.7 Ma)

Description

Shoots ca. 1 to 1.5 mm wide with opposite and decussately arranged scale leaves each ca. 1 to 1.5 mm long (Figs. 2–5). Leaves keeled, with marginal frill up to ca. 50 µm long composed of finger-like cilia and ctenoidally arranged cells (Figs. 3–5). Cilia typically confined to basal half of leaf beneath overlapping scale leaf (Figs. 4, 5), and ctenoidal cells to exposed upper half (Fig. 5). Stomata (Figs. 6–8) confined to adaxial surface (i.e., epistomatic), mostly in bands on each side of midvein, each band comprising at least seven irregular rows of parallel-aligned stomata. Stomata with a variable number of surrounding cells, these of variable size and shape, and so that the lateral “subsidiary” cell of one stoma may be the polar “subsidiary” cell of a stoma in an adjacent row (Fig. 6). Polar and lateral cells often shared between adjacent stomata in the same row (Figs. 6, 7). Polar extensions present. Guard cell cuticle pairs (including polar extensions) 14–20 μm long, 5–10 μm wide. Florin rings present, typically appearing collapsed into lobed segments on outer surface (Fig. 8). External dimensions 17–30 µm long, 11–22 µm wide. Florin ring evident on inner cuticle surface as a furrow surrounding the cutin complex associated with the guard cells (Figs. 6, 7). Normal epidermal cells of extremely variable size and shape, squat rectangular to linear (Fig. 6). Anticlinal wall cuticle of these and stomatal surrounding cells not or only slightly beaded or pitted (Figs. 6, 7). Abaxial epidermal cells longitudinally aligned, of variable size and shape, rectangular to trapezoidal, mostly 35–65 µm long, 15–20 µm wide (Fig. 9). Abaxial anticlinal wall cuticle not or only slightly beaded, sometimes ragged-edged, periclinal walls smooth (Fig. 9).

Justification of assignment to a new species of Microcachrys

Within Podocarpaceae, opposite and decussate phyllotaxy is unique to Microcachrys (Figs. 10, 11) and must be regarded as a derived and diagnostic feature for the genus. The shared presence of this feature in the fossils enables their secure placement in Microcachrys and a number of other cuticular features are also shared with M. tetragona. Within Podocarpaceae, Microcachrys is one of only two genera that are fully epistomatic and that have irregularly shaped polar and lateral subsidiary cells, either in contact or shared between or within rows (47). The other genus is Pherosphaera. The fossils show all of these features and also exhibit the same type of leaf marginal frill as in M. tetragona, including the presence of cilia (Fig. 11; see also fig. 36 in 47). However, in the leaves of the fossil specimens examined, these cilia appear to be concentrated in the basal half of the leaf only (Figs. 4, 5), whereas in M. tetragona they are found around the whole margin (Fig. 11). Marginal frills are also found in Halocarpus, Lagarostrobos, Parasitaxus, and certain scale-leaved species of Dacrydium and Pherosphaera, but only the latter genus has a frill similar to that of Microcachrys (47). The prominent, raised Florin rings of the fossils (Fig. 8) are also shared with M. tetragona, although their disjointed nature is less clear in this extant species (Fig. 12; see also fig. 39 in 47). Within other imbricate-leaved Podocarpaceae, similarly raised Florin rings are present only in Pherosphaera (47). The fossils are also very similar to M. tetragona in details of the inner cuticle associated with the guard cells. In particular, although there appears to be greater development of polar cutin extensions in the fossils, both fossil and extant species lack lateral cutin extensions (compare Figs. 6, 7 with Fig. 13; see also fig. 38 and discussion in 47). A small difference between the fossils and M. tetragona is that the cuticle of the anticlinal wall of the epidermal cells of this species has more evidence of slight pitting and beading (Figs. 13, 14; see also fig. 38 and discussion in 47) than in the fossils (Figs. 6, 7, 9). Thus, in summary, M. novae-zelandiae shares many cuticular features with M. tetragona and Pherosphaera, but clearly belongs in Microcachrys on the basis of its phyllotaxy. A new species is proposed in recognition of the large geographic and temporal gap between M. tetragona and the fossils and minor differences in leaf details. Thus M. novae-zelandiae shows more pronounced development of polar cutin extensions associated with the guard cells, more clearly segmented Florin rings, evidence of a clear transition from marginal cilia to a uniform frill in the apical half of the leaf and less evidence of anticlinal wall pitting.

Figure 10–15

Scanning electron microscope (SEM) images of (Figs 10–14) extant Microcachrys tetragona and (Fig. 15) Diselma archeri. 10. Foliage showing opposite and decussate phyllotaxy. 11. Broken shoot with the removal of a facing leaf. Note the basal junction of the lateral scale leaves where no stomata can be seen and the cilia along the margins of each leaf. Compare details with those of the fossil M. novae-zelandiae (especially Fig. 4) and contrast with Diselma archeri (Fig. 15). 12. Outer adaxial cuticle showing Florin rings. Note that these do not show clear evidence of segmentation as in the fossils (Fig. 8). 13. Inner adaxial cuticle showing four stomatal complexes. Note the absence of lateral cutin extensions associated with the guard cells as in the fossils (Figs. 6, 7), but that polar extensions are not typical. Note also the beaded nature of the anticlinal epidermal cell wall cuticle. 14. Inner abaxial cuticle surface showing beaded to slightly buttressed epidermal cell anticlinal walls. 15. Broken shoot with the removal of a facing leaf. Note the basal junction of the lateral scale leaves where clusters of stomata can be seen (examples arrowed at center) and that there are no cilia along the margins of each leaf. Rather, there is a smooth frill only. Scale bar = 1 mm for Fig. 10; 500 µm for Figs. 11, 15; 50 µm for Figs. 12–14.

Scale leaves of similar size and shape and arranged in an opposite and decussate manner are also found in some Cupressaceae. Most striking is Diselma archeri Hook.f. (Fig. 15), a montane Tasmanian endemic conifer that often co-occurs with Microcachrys. However, Diselma has the typical cuticular morphology of Cupressaceae (e.g., 49) that is very different to Podocarpaceae, as well as clear differences in stomatal distribution and marginal frill morphology. For instance, in Diselma there is a group of stomata present on the abaxial surface at the base of each of the lateral pair of leaves. These are visible (arrowed in Fig. 15) following removal of a facing leaf. Also, although a marginal frill is present in Diselma, it is composed of uniform ctenoidally arranged cells (Fig. 15) and does not show the finger-like cells (cilia) seen in Microcachrys.

DISCUSSION

History and biogeography of Microcachrys

Whatever the problems in interpreting the phylogeny of Microcachrys and Pherosphaera, they do not affect our conviction that the New Zealand fossils described here belong to Microcachrys. These fossils represent the first foliar evidence that Microcachrys formerly occurred remote from the Tasmanian region and indicate that opposite and decussate phyllotaxy in this lineage has been present since at least the Oligo-Miocene. The only previous macrofossil reports are from the early Pleistocene of Tasmania (24), where Microcachrys still occurs and from the Oligo-Miocene Morwell Seam of the Latrobe Valley coals of the nearby Australian mainland (as ‘aff. Microcachrys’: 3). The latter record has not been critically substantiated as Microcachrys, although 18, p. 398) considered that the leafy twig “has more than a superficial resemblance to that genus”. Regardless of the status of this fossil, previous records of Microcachrys are overwhelmingly pollen-based. These records indicate that the Microcachrys lineage was more diverse in the past and particularly so in the Cretaceous to early Miocene (41; 42).

Apart from Microcachryidites antarcticus and Podosporites parvus, a number of other fossil trisaccate pollen types that have been linked to Microcachrys occur in New Zealand, including Trichotomosulcites subgranulatus Couper, Trisaccites spp., and other species of Podosporites (42). Of the taxa that are considered to be morphologically closest to the pollen of extant M. tetragona, Microcachryidites antarcticus has the longest temporal span, ranging from around the Jurassic-Cretaceous boundary to the early Miocene, whereas the more strictly circumscribed Podosporites parvus ranges from the middle Eocene until the early Pleistocene (ca. 1.3 Ma) (Fig. 1) (42). Podosporites brevisaccatus (Couper) Mildenh. is another very similar Cenozoic trisaccate Microcachrys-like pollen type found only in New Zealand and Australia. All three of these pollen types are recorded from various horizons in the Newvale Mine in trace amounts, except in one sample (F45/f153) from Seam W6 where P. brevisaccatus accounts for up to 30% of total pollen (analyzed by D. T. Pocknall, GNS Science, New Zealand, unpublished records) and is thus indicative of abundant local source plants. It is therefore possible that the same plants that produced this pollen contributed the Microcachrys novae-zelandiae foliage to the lignite. Podosporites brevisaccatus became extinct in the early Pliocene in New Zealand (43) (Fig. 1).

Overall, the fossil record strongly suggests that the range of the lineage now represented only by M. tetragona in Tasmania formerly spanned the southern hemisphere. Several species occurred in New Zealand and plants persisted there until relatively recently.

Ecology

Even if the distinctive phyllotaxy and cuticular features of M. tetragona evolved relatively recently, the multimillion year absence of Microcachrys macrofossils from sediments that host Microcachrys-type pollen across the southern hemisphere is enigmatic (e.g., 18), particularly in the Mesozoic to early Paleogene, when this pollen was relatively abundant (42). A simple explanation for the absence of macrofossils is that potential source plants were growing at some distance from depositional sites and/or were similar to M. tetragona in being low or more or less prostrate shrubs that were unlikely to contribute litter to sediments. Such plants could have been common at high latitudes during the Mesozoic to early Paleogene because the vegetation must have been relatively open due to low sun angles. This niche became increasingly scarce for conifers during the Cenozoic, because of the ice-related extirpation of vegetation in the Antarctic and through canopy closure at lower latitudes as Australia drifted northwards (19).

Cenozoic assemblages in which Microcachrys-type pollen is relatively abundant and/or in which Microcachrys macrofossils have previously been reported have important features in common with that of the Newvale locality. In the early Pleistocene of New Zealand, the proportions of Podosporites parvus pollen suggested to 42 that the source plants occurred in moderate abundance in lowland Nothofagus and podocarp-dominated forest, in association with a diverse shrubland and acidic swamps. In the early Pleistocene of Tasmania, Microcachrys tetragona foliage is abundant and co-occurs with a diverse range of other conifers, including regionally extinct species (24). It has previously been noted that the Latrobe Valley coals have floristic similarities to the Newvale lignites (6). Both sites reflect deposition of plant material in highly oligotrophic, low altitude bog situations and both have diverse conifer assemblages and numerous taxa of scleromorphic angiosperms including epacrids (Ericaceae), Banksia, and other Proteaceae. The fossil assemblages of the Morwell Seam of the Latrobe Valley are interpreted as being formed under relatively stable conditions with little influence of climatic seasonality and fire (3; 4). Similarly, the complete lack of charcoal in the Newvale lignite shows that its source vegetation was not subjected to burning. There is thus an emerging picture that at least during the Neogene, Microcachrys grew under mild climates and formed part of conifer-dominated vegetation in oligotrophic, perhaps waterlogged sites that were not or only infrequently burnt. At least some of this vegetation was relatively open (6). Open, wet habitats dominated by conifers are now very rare across the southern hemisphere because of the widespread development of at least seasonally dry climates during the Neogene and the almost ubiquitous incidence of fire in modern situations. Thus, the present restriction of Microcachrys to the Tasmanian mountains could be primarily related to the persistence of relatively stable, wet, fire protected niches in these areas. Clearly, however, Microcachrys was unable to survive anywhere in New Zealand beyond the late Pleistocene (42), perhaps because there was a paucity of suitable wet and ice-free habitats during the sustained decline in mean annual temperature in the late Miocene–Pliocene that was probably a principal cause of wholesale extinctions of the New Zealand flora (33). Considering the past presence of Microcachrys in low altitude, presumably warmer environments, it would be interesting to know whether M. tetragona has the physiological capacity to grow successfully in sufficiently wet sites at lower altitudes in Tasmania or elsewhere.

The dispersal ecology of Microcachrys has not been studied in detail. However, M. tetragona is sometimes known as strawberry pine because of its attractive sweet, red, ripe female cones. The seeds are known to be dispersed widely by birds, enabling the recolonization of open ground (29). The source plants for Microcachryidites antarcticus pollen in New Zealand may have been continuously present there from the Cretaceous, but it is possible that the source plants for Podosporites parvus and P. brevisaccatus only reached the region around the Paleogene–Neogene boundary, possibly by long-distance dispersal from Australia.

Tasmania as a refugium for Microcachrys and other isolated plant lineages

The high concentration of monotypic or species-poor genera of conifers in equably moist, often montane regions within and surrounding the Pacific Ocean has long been known (e.g., 34), and in the southern hemisphere, New Caledonia, New Zealand, and Tasmania are exemplar regions of such diversity. Tasmania and New Zealand are both known to have had extremely high conifer diversity in the Cenozoic, and in the early Oligocene, Tasmania may have had the most diverse conifer flora on Earth (19). However, the traditional status of New Caledonia and New Zealand as ancient continental island centers of conifer evolution, perhaps extending to the Cretaceous, has increasingly been questioned. This is because although the geological history of the southwest Pacific is very complex and not fully understood (30), the New Caledonian land surface apparently did not exist until after the late Eocene, and a few geologists and biogeographers claim that New Zealand was entirely submerged in the late Oligocene (31). Moreover, there is increasing phylogenetic and molecular dating evidence that the ages of many taxa in both regions, including often-presumed Gondwanan vicariants such as taxa of Nothofagus, are too young to be explained by Gondwanan vicariance (see 17; 37 for reviews). Whatever their degree of validity, doubts about the antiquity of the terrestrial habitat and biotas of New Caledonia and New Zealand serve to highlight the significance of the Gondwanan links of Tasmania and its biota. As the southern tip of the Australian continent this region maintained an important terrestrial connection to Antarctica and beyond during most of the Paleogene. Also, it has been argued that there has been a very long history of physiographically diverse but relatively stable moist environments in the region that served as mesic refugia (7). This argument is exemplified by Tasmanian Paleogene records of Mesozoic taxa such as Ginkgo (20) and seed ferns (39) and the possibility that the mountainous interior of Tasmania provided the core Australian refuge for cool-adapted Nothofagus during the global “greenhouse” period of the early Eocene (38). A strong climatic gradient has probably existed across the island since the development of the Antarctic circumpolar current in the mid-Cenozoic (see 38), contributing to a wide diversity of habitats. Also, there is emerging evidence from phylogeographic studies that mesic refuges still existed in the Tasmanian mountains during periods of glaciation and glacial aridity beginning in the late Neogene (50).

Microcachrys tetragona is only one of numerous endemic and/or monotypic, and phylogenetically isolated plant taxa that typically occupy very wet and often montane habitats in Tasmania. Other examples include Pherosphaera hookeriana W.Archer bis, Athrotaxis (Cupressaceae; only genus of subfamily Athrotaxidoideae, sister to a large clade comprising almost all other Cupressaceae: 15), Bellendena montana R.Br. (Proteaceae; only species of subfamily Bellendenoideae: 48), Cenarrhenes nitida Labill. (Proteaceae; sister to two monotypic genera found only in Madagascar and New Caledonia: 45), Agastachys odorata R.Br. (Proteaceae), Isophysis tasmanica (Hook.) T.Moore (sister to the rest of Iridaceae: 16) and Campynema lineare Labill. (sister to the New Caledonian Campynemanthe, and this clade may be sister to the rest of Liliales: 22). Although the co-occurrence of these taxa in Tasmania cannot prove the continuous presence of their lineages in the region, it does enable such a hypothesis to be considered seriously.

The outstanding biological heritage of Tasmania is thus a product of its Gondwanan connections and its diverse landscapes and climates. Improved conservation and management measures will be required to meet the challenge of maintaining this heritage into the future (40).