RESULTS AND DISCUSSION

One hundred and fifteen morphologies of cuticle parataxa were distinguished in the 120 samples (Appendix 3). Fifty eight taxa have been placed into extant families, and eight into extant genera. The remainder have not been identified, or have less certain, and only suggested affinities. The species-level biodiversity represented here makes a substantial contribution to that predicted from the diversity of one family in the fossil record, the Lauraceae. Based on the current relationship between the number of Lauraceae species and total tree diversity for rainforest sites in Australia, the 25 Lauraceae taxa known from the Manuherikia and East Southland Groups suggest at least 100 tree species grew with them, and probably at least 120 (Pole 2007a).

Some of the families described as fossils in this study occur in the extant flora of New Zealand, but others represent the first known occurrence of the family in New Zealand, and as such, add to another dimension of biodiversity. Among the unidentified taxa there are also likely to be totally extinct genera represented, but these will need more information, for instance evaluation of the corresponding whole leaf architecture.

A brief discussion of the significance of the identified taxa (Summarised in Table 1) follows and the detailed descriptions are relegated to Appendix 3. A fuller synthesis will await publication of some further taxa and will be integrated with more detailed stratigraphy and sedimentology.

Evidence for the presence of Gnetales in New Zealand is not new – palynological works have reported the pollen Ephedripites notensis Cookson (1957) for many years (e.g. Couper 1960). This pollen type has been said to have affinities with extant Ephedra, an arid-land plant of North America. Its presence in an obviously wet-environment is something of an anomaly, and the likelihood is that the family identification is correct, but that it represents some other, extinct genus. The presence of cuticle with similarities (but distinct from) Gnetum reinforces the view. The cuticle morphology is also present in the Late Eocene of the Waikato Coal Measures (Pole unpublished) and a similar parataxon is in the Early Eocene of Tasmania (Pole 2007d). It may represent a plant lineage which continued through the putative Oligocene drowning of New Zealand (Cooper and Cooper 1995), or it may represent long-distance dispersal. In this context it is of interest that DNA evidence has suggested that Gnetum has dispersed globally within the Neogene (Won and Renner 2006).

The Winteraceae are represented in New Zealand today by three species in the genus Pseudowintera (Allan 1982). The cuticle record suggests that other genera of the family were present in the Miocene (there is no Miocene cuticle record of Pseudowintera). The inference of higher generic diversity at this time supports the conclusions of the palynological record (Pocknall and Mildenhall 1984; Raine et al. 2006) where Harrisipollenites annulatus Mildenhall and Crosbie (1979) is regarded as representing Zygogynum, a New Caledonian genus, and Pseudowinterapollis couperi Krutzsch (1970) is consistent with Drimys, Pseudowintera and Tasmannia and P. cranwelliae (Stover and Partridge 1973) represents Tasmannia. Today the family is a prominent component of rainforests in the cool southern Hemisphere as well as New Caledonia.

The Atherospermataceae and Monimiaceae are also important families in cooler Southern Hemisphere rainforests and both are in the current New Zealand flora (Hedycarya in the Monimiaceae and Laurelia in the Atherospermataceae). Little is known of their prehistory; the long-ranging pollen type Liliacidites variegatus Couper (1953) may represent Atherospermataceae, though it may also be produced by the Lilliaceae (Raine et al. 2006), but fossil wood has been reported from Antarctica, but cannot be related to any single extant genus (Poole and Francis 1999).

The Proteaceae has two genera in the extant flora of New Zealand, Knightia, and Toronia, both members of the warmer rainforests. The cuticle evidence supports the palynological evidence that the Proteaceae were much more diverse in New Zealand. The identification of Placospermum in this study adds another genus to the New Zealand record and those known from the Miocene (Pole 1998a). Today a diverse component of Proteaceae is a feature of the wet, mesothermal rainforests of northern Australia (Webb and Tracey 1981).

The Menispermaceae represent a new addition to the New Zealand record. They are a family dominated by climbers or creepers, mostly tropical, reaching as far south as Victoria in Australia. Added to the Rhipogonum already known (Pole 1993c) this identification allows the inference of a second climber in the Manuherikia Group based on macrofossil evidence.

The genus Notothixos is a new addition to the New Zealand record and is the first fossil record of the genus, and perhaps of the Viscaceae. Notothixos is a mistletoe and occurs today from Ceylon to Malesia and down the east coast of Australia to Victoria (Barlow 1983).

Strasburgeria (Strasburgeriaceae) is currently endemic to New Caledonia, but has been identified with the pollen Bluffopollis scabratus (Couper) Pocknall and Mildenhall (1984), which is present throughout the Tertiary of New Zealand (Jarzen and Pocknall 1993) and is a common and widespread component in many of the samples studied here (pers. obs.). This is the first fossil record of the cuticle.

Griselinia is the only genus in the Griseliniaceae, which is found today in both New Zealand and South America. This is the first fossil record of the genus.

The Meliaceae and Sapindaceae both have a single rainforest species in New Zealand today; Dysoxylon and Alectryon respectively. In the Sapindaceae there is also Dodonaea, a shrub of more open vegetation. The cuticle evidence confirms the palynological evidence for these families in the Neogene, but like the pollen, is not helpful to generic identification. Although Couper (1960) identified pollen of Dysoxylum in the Cenozoic, neither the genus nor the Meliaceae is included in Raine et al. (2006). The fossil pollen may not be distinguishable from that of many other genera of Meliaceae (e.g. Mildenhall 1980).

The Cunoniaceae-Elaeocarpaceae are prominent components of the extant New Zealand flora as the genera Elaeocarpus and Weinamnnia. Miocene Elaeocarpaceae leaf impressions were described by Pole 1993c (which on the basis of their basic shape, domatia, and spinose attachments on the teeth, I am now certain are Elaeocarpus). The Myrsinaceae is present in New Zealand today as the widespread genus Myrsine and also the highly restricted Elingamita (Allan 1982). A compression fossil of the Mysinaceae was described by Pole (1996) from the Miocene Foulden Hills Diatomite (and based on the vein architecture, I am now certain that this is an Ardisia).

The Manuherikia Group would have lain at about 50ºS in the Early Miocene (Lawver and Gahagan 2003), about 5º further south than its present position. The latitude of the East Southland Group, like today, would have been only a degree or two higher than the Manuherikia Group. Both of these areas lie southwards of the southern limit of several plant families represented by the cuticle (Table 1), and therefore indicate a more southerly range extension than found today. This likely refects warmer conditions, consistent with other fossil (Hornibrook 1992) and isotopic evidence (Shackleton and Kennett 1975). Judged on the present distribution of these taxa, temperatures were in the microthermal to mesothermal range (sensu Nix 1982) and the vegetation was rainforest (sensu Bowman 2000).