RESULTS

The fossils found based on dispersed cuticle include a cycad (Bowenia) and a possible cycad (Pterostoma), a gnetalean, three families of conifers; Podocarpaceae (including Acmopyle and Dacrycarpus, Prumnopitys), Araucariaceae (including Araucaria, Araucarioides and Agathis), and Cupressaceae (including Libocedrus), and 55 angiosperm taxa. These include the Lauraceae, Proteaceae, Winteraceae, Aquifoliaceae, Rhizophoraceae, Rhipogonaceae, and other monocots. The mangrove palm, Nypa, was clearly an important element based on its pollen presence (Pole and Macphail 1996), but its delicate cuticle does not survive bulk preparation, but occasional scraps indicate other palms were present as well. Gymnostoma also rarely survives bulk cuticle preparation, but its woody articles often survive the initial sediment disaggregation and sieving and were common in some samples. The taxonomic affinities of these plants are consistent with a rainforest (sensu Bowman 2000; no charcoal is present in palynological preparations, pers. obs.) and mangrove vegetation.

An absence of Myrtaceae in this study is notable, given its prominence in Australia today, and its presence in most extant Australian rainforests, but is consistent with other evidence from the early Eocene on the Australian mainland (e.g., Martin 1994; Sluiter 1991; Carpenter et al. 2004). Myrtaceae cuticle can typically be identified by the presence of paired lid cells (Lange 1980; Christophel and Lys 1986). Clearly the timing and nature of the spread of Myrtaceae into Australian vegetation may be more complex than generally thought.

The highest diversity was found in sample R-75, which had 19 taxa in all, including 15 angiosperm taxa as well as Araucariaceae, Podocarpaceae, and Bowenia. Bowenia was widespread, occurring in 16 (33%) of the samples. Conifers occurred in 36 of the 49 samples (74%). Podocarpaceae were the most widespread group, occurring in 33 (67%) of the samples. Some samples have only conifer remains, and these are probably the result of weathering or taphonomic processes resulting in the destruction of all but the typically robust conifer cuticle. However, sample R-30 is visibly packed with conifer remains and was probably genuinely conifer-dominated vegetation.

Sample R-21 stands out as having the second highest diversity (12 taxa) yet is one of the few samples with no conifer remains at all. The most diverse angiosperm groups recognised are Lauraceae (11 taxa) and Proteaceae (eight taxa; see personal commun. in Jordan et al. 1998). The most widespread angiosperm was CUT-Z-JAE, occurring in 11 samples.

On Strahan Point, sample R-102 is from a layer of leaf fragments lying at the base of a small clay-filled channel, about 2 m deep, that is cut into a bed of clay penetrated by many small nodulated roots. These roots are probably remains of Podocarpaceae, which grew on waterlogged, gleyed soils above the high tide levels. However, the channel itself preserves an entirely angiosperm flora (without Gymnostoma). This suggests that there may have been segregation at high taxonomic levels across the broader environment. Conifer-dominated vegetation may have been more a feature of waterlogged habitats closest to the mangrove zone. A rise in relative sea-level then flooded these swamps, and the vegetation was replaced by mangroves, including Nypa and Rhizophoraceae. At the top of this section, a bed of metre-high foresets in sand with a lag of leaf and wood fragments at its base is interpreted to represent a fluvial channel that migrated into the tidal sand flat. Sample R-25 from this lag has both conifers and angiosperms.

The highly carbonaceous zones near the base of Regatta Point and Regatta Tavern are interpreted as freshwater swamps (Pole 1998a) where conifers were dominant and generally diverse. In both locations they are overlain within a metre of section by mangrove mud, so it is likely they grew immediately adjacent mangrove vegetation, just above high tide level. Because of their highly carbonaceous nature, the fossils are inferred to have accumulated in situ with little likelihood of contaminant material being washed in from elsewhere. Taxa that are found together in these facies are likely to have grown together. Sample R-30, from this facies, had all three conifer families (but no cycads), and had the highest conifer diversity of 10 species. Its angiosperm diversity was low, only three species, and not including any Lauraceae. The high diversity of conifers agrees well with previous work indicating extraordinary levels of conifer diversity in Tasmania's Paleogene (Pole 1992; Hill and Brodribb 1999).

The presence of Rhizophoraceae cuticle (CUT-Z-JAG) with affinities to extant Bruguiera, Ceriops, and Rhizophora, clearly indicates a further mangrove taxon. It is present in sample R-74 in association with the Nypa described by Pole and Macphail (1996), and also within the mangrove mud on Regatta Point (R-12, 130, and overlying the freshwater swamp facies on Strahan Point (R-26). In this later sample it is associated with dinoflagellates (Pole 1998a) but no Nypa was found. This was interpreted as being the edge of a freshwater swamp, where saltwater incursions may have washed dinoflagellates in. Rhizophoraceae cuticle suggests that mangroves fringed the freshwater swamps here.

It is highly unlikely that some of the other plants found in mangrove facies in association with Nypa macrofossil, pollen and dinoflagellates, also grew as mangroves. For instance, Bowenia occurs with Nypa macrofossils in R-74 but today it grows as a small plant in the understory of rainforest. For instance, it grows within a rainforest only a metre from the upper limit of mangrove vegetation along the Mardja walk, in Cape Tribulation National Park (pers. obs.).

Gymnostoma is abundant in three samples (R-46, 47, 50) from mangrove mud on Regatta Point, where it is found with a number of other angiosperms but almost no conifers. Like Bowenia, Gymnostoma is unlikely to have been a mangrove, although today members of the Casuarinaceae can dominate regions adjacent to mangroves. The abundance of Gymnostoma in some samples suggests that it was, as members of the family often are today, typically gregarious. Members of the Casuarinaceae host nitrogen-fixing bacteria, and as such they are well-suited to be pioneering plants after a disturbance, perhaps on a fluvial point bars.

Other than CUT-Z-JAG (Rhizophoraceae) none of the cuticle morphologies appear similar to any known mangrove, and those that have been identified, like Lauraceae and Proteaceae, do not have mangrove taxa today. Whereas it can not be ruled out that some of these taxa may have had mangrove representatives in the past, some taphonomic mixing must surely have occurred. It is likely that many of these were plants growing along the edge of supra-tidal, freshwater reaches of rivers, which were washed into the mangrove environment. It is possible that some of the cuticle parataxa of unknown affinities represent extinct mangrove taxa, perhaps in families which today do not include mangroves. This would certainly be difficult to prove, but identifying taxa consistently restricted to mangrove facies would be a start.

Based on Figure 2 samples were allocated simply to either "freshwater' or "marine" (mangrove mud and sand flat) facies. Multivariate analysis based on the presence-absence data for all species (Figure 3.1) does not give significant results in terms of this environmental partition. However, most taxa are only present in one or two samples and are likely just introducing noise into the analysis. When the analysis was limited to taxa that are present in at least 20% of the samples, the results were significant (Figure 3.2), and suggest that the broad facies difference accounts for about 30% of the variability between samples. Samples from the "marine" facies probably include taxa from both mangrove and immediately adjacent vegetation.

Paleoclimate

As a genuine lowland and coastal Early Eocene site, Regatta Point could provide globally important paleoclimatic data. However, the method of foliar physiognomy (Wolfe 1979, 1995) is of limited use at Regatta Point as intact leaves are uncommon. Carpenter et al. (1994) included an average leaf length for Regatta Point fossils on a chart as about 70 mm, with a corresponding mean annual temperature of about 17ºC. However, none of the specimens that this figure was based on (number unknown) could be located, and in the four years of this study only a single partially complete angiosperm leaf impression (no cuticle could be isolated) was found, with an original length of about 100 mm.

Rainfall levels are even harder to estimate, and annual totals are almost worthless without knowing how this was seasonally distributed. Like the Regatta Point Eocene, there are also high diversities of (mainly Podocarpaceae) in the microthermal rainforests of Tasmania and New Zealand today (e.g., Jarman et al. 1984; Reid et al. 1999), but without the diversity of Lauraceae and Proteaceae in the fossil assemblages. This is likely to be a function of very wet conditions and poor soils. The abundance of conifer remains at Regatta Point likewise may suggest very wet conditions throughout the year, but this needs to be balanced against the general lack of almost any coal. Several of the cuticle taxa at Regatta Point have surface papillae or pronounced ridges, which might have some climatic significance. However, their significance is ambiguous. Hill (1998) argued that the firmest evidence for xeromorphy included the presence of individually protected stomata (by being "surrounded by raised epidermal structures"). But he also argued that these could be evidence of wet conditions, and Carpenter et al. (2004) listed trichomes, papillae, and ridging as characters that obscured the stomata, some of which "would also be advantageous in generally wet conditions ..." Clearly some data on extant plants are needed to clarify this issue. This leaves floristics as a further climatic indicator. Nix (1982) classified "thermal regimes" as mean annual temperatures (MAT) of >24ºC = megatherm, >14<20ºC = mesotherm, and <12ºC = microtherm. In a broad sense, the prominence of conifers (especially Podocarpaceae), Lauraceae, and Proteaceae compares well with extant mesothermal rainforest vegetation, which is found in mid-montane altitudes of the tropics, and which extends down towards sea level at higher latitudes (Whitmore 1984; Richards 1996). Conifers are important in some tropical swamp situations, for example the low-nutrient raised peat-swamps of Borneo where they can compete with angiosperms (e.g., Brünig 1974). However, they do not dominate these communities, and there is no suggestion that the Regatta Point environment included raised peat swamps. The other 'dry-land' taxa that have been identified at Regatta Point are consistent with this interpretation. The mere presence of broad-leaved Lauraceae in Tasmania, where they do not occur today, suggests warmer conditions than the present. In Australasia the Lauraceae reach their southern limit near the southern margin of mainland Australia, and at similar latitudes in New Zealand (Pole 2007a). This is likely to be a temperature-related limit. At these limits today the Lauraceae are present at low diversities (for instance 1-2 species in a local flora). The 11 species occurring at Regatta Point suggest temperatures were well above the cold-limit for the family, and reflect a high overall tree diversity. A similar southern limit on the Australian mainland exists for mangroves and palms (Duke et al. 2002; Cameron 1987). Greenwood et al.'s (2003) characterisation of Ilex as a "megathermal" taxon is simply incorrect. As summarised by Martin (1977) it has wide limits, and in fact in a biomass sense, is probably more characteristic of microthermal conditions. However, superimposed on this essentially mesothermal combination is the presence of the mangrove palm, Nypa. N. fruticans lives today within about 15º of the equator, i.e., it is wholly tropical and would seem to provide an excellent indicator for tropical, or megathermal temperatures. In the fossil record the broad history of the genus confirms that it has always been a plant of relatively warm conditions (e.g., Gee 1990, 2001; Collinson 1993). But there is a clear danger of extrapolating too far from a single extant species. The Tasmanian Nypa australis is specifically different from the extant species (Pole and Macphail 1996) and is likely to have had different environmental tolerances. The fact that the fossil Nypa occurs with a variety and abundance of conifers should warn against assuming the environment as megathermal, although the presence of palms, cycads, and the mangrove life-style indicates a largely frost-free environment (e.g., Wing and Greenwood 1993).

Finally, Carpenter et al. (2004) noted that an unusual proportion of their Proteaceae taxa had average stomata; lengths less than 20 µm. Jordan et al. (1998) proposed this was a general phenomenon for Palaeogene Proteaceae from Tasmania and linked it to high carbon dioxide levels. This phenomenon is not apparent in the Proteaceae studied here, and only one taxon (CUT-P-EJD) has an average stomatal length less than 20 µm.

To summarise, temperatures at Regatta Point in the Early Eocene were warmer than today (12ºC), despite the locality being 20º of latitude or further south (perhaps as far south as the Polar Circle at 66 ºS). Mean annual temperatures somewhere in the middle of the mesothermal range (between 14 and 20ºC) were likely and rainfall was continuously high (probably similar to much of Tasmania today).