COMPARISON OF PLANT MICRO- AND MACROFOSSILS, KARIOITAHI, AWHITU PENINSULA

by Rewi Newnham1 and Chris Lusk2 1 Department of Geology, University of Auckland, Private Bag, Auckland 2 Department of Botany, University of Auckland, Private Bag, Auckland SUMMARY

Mid-Quaternary plant micro- and macrofossils from Awhitu Peninsula in• dicate coastal mixed broadleaf forest with warm-termperate conifer-broadleaf forest further inland. Differences between the regional pollen from taxa ab• sent or uncommon at the site. Beilschmiedia tawa and B. tarairi leaves were present, but no Beilschmiedia pollen was recorded.

INTRODUCTION

The Awhitu Peninsula, consisting of Quaternary sands and carbonaceous deposits, forms a near-rectangular block of high relief, bounded to the north• east by the Manukau Harbour, and to the south by the Waikato River. Quater• nary sediments are well-exposed at numerous localities on the west coast of the peninsula, where steep cliffs, capped by dunesands, are dissected by small streams. One such locality, about 1.5 km north of Karioitahi beach (NZMS 260 R12, grid ref. 564 357) contains a bed of carbonaceous sand and mud, rich in fossil wood and leaf matter. This plant bed, one of several known in the area, is found in association with sands of the Cochrane Formation, of upper Putikian age (Barter 1976). Fission track analysis of glass shards from within the sands has provided an 'absolute' date of 380,000±40,000 years B.P.(op.cit.). The occurrence of well-preserved plant leaves within carbonaceous sediments provides an opportunity to compare macro- and microfossil (pollen and spores) assemblages obtained from the same deposit, with a view to drawing some in• ferences abut the paleobotanical environment. The first and second authors respectively worked independently on identification of micro- and macrofossils.

POLLEN ANALYSIS

METHODS The pollen sample was prepared as follows. Approximately 0.5cc of the car• bonaceous sediment was crushed, broiled for 10 minutes in dilute KOH, pass• ed through a 150 micron wire mesh sieve, boiled in 40% HF for 45 minutes, then acetolysed as described by Erdtman (1969). The residue was stained with Basic Fuchsin, then mounted in glycerine jelly. Counting was undertaken on a Leitz Ortholux microscope at magnification x400, with critical identifica-

171 Fig. 1. Location. tions at xlOOO. A sample slide was sent to D.C. Mildenhall, palynolygist with Geological Survey, for his assistance with the identification of several 'unknowns'. Pollen recovery was low and most of the grains, although iden• tifiable, were poorly-preserved. Mr Mildenhall indicated that a better pollen recovery would be needed to determine the age of the sample more accurately, and that the possibility of reworking should not be discounted in a sample such as this, where preservation is poor. A total of 240 pollens and spores was counted.

RESULTS AND INTERPRETATION Results are shown in Table 1. The pollen assemblage suggests that mature warm-temperate lowland forest, indicative of a warm interglacial climate, was established in the area. Arboreal pollen and treefern spores dominate the assemblage, with the major taxa being lianous Metrosideros species and tall podocarps which typically reach emergent status. Agathis australis (kauri) and

172 Libocedrus (probably L. plumosa) were also present; these two large coniferous species are both found only in the warmer northern regions in present day New Zealand. The record of Ascarina lucida is of interest as the plant is rare in northern North Island today. However, it was common in western North Island localities during the early post-glacial (McGlone & Moar 1977). A. lucida is frost-sensitive (op. cit., Sakai & Wardle 1978) and has become an important indicator of warm, relatively drought-free conditions in Quaternary pollen assemblages. Although the assemblage indicates a forest origin, it lacks the abundance and diversity which would be expected if the site itself had been very close to or surrounded by forest. Most of the pollen grains show some deterioration, especially degradation and crumpling (cf. Cushing 1967), whereas the fern spores (generally with thicker walls), tend to be in a better state of preserva• tion. The presence of the colonial alga Botryococcus in this assemblage (D.C. Mildenhall, pers comm) indicates the proximity of freshwater, probably streams (Mildenhall 1979). It seems likely therefore that many of the pollens and spores were transported fluvially prior to their deposition in these sediments.

FOSSIL LEAVES

METHODS The deposits split readily along the bedding planes, revealing abundant leaf fossils and impressions. Several blocks totalling about 7,000 cc were carried back to Auckland University for further splitting and examination. Leaf iden• tifications were checked by comparing the fossils with herbarium specimens and fresh material. The leaf characters used in confirming the identifications are described in the appendix. The fossils were labelled, photographed, and stored in the A.U. Geology Department (Fossil Record No. R12/f041).

RESULTS AND INTERPRETATION The leaves were largely those of dicotyledonous trees and shrubs, but the majority could not be identified with any certainty. Only ten taxa were iden• tified with confidence; eight dicotyledons, one conifer and one fern (see Table 1). Examples are shown in Figure 2. The species' label numbers that appear in the photos are listed in the Appendix. In many cases only part of the leaf outline was visible, and many characters are lost in fossilisation, with the result that impressions of unrelated taxa sometimes appear very similar. The pro• blems of identifying fossil leaves are discussed by Mildenhall (1973). Unfor• tunately there were no finds of fossil seeds or fruits, which are often readily identifiable (Watts 1978). Several monocotyledonous leaf fragments were found, but these were not identified. Although the leaf fossil assemblage is therefore clearly not fully representative of the local vegetation, it suggests the presence of warm-temperate mixed broadleaf forest, with few conifers.

173 Fig. 2 (a) and (b). Examples of fossil leaf material Scale is in cm.

COMPARISON AND DISCUSSION

Birks and Birks (1980, Table 5.1) discuss the relative merits of macrofossil (including leaf) and microfossil (pollen) analyses in the reconstruction of vegeta• tion history. Whereas pollens can be transported great distances and will therefore reflect vegetation at the regional scale, leaves are more likely to have come from plants growing locally; i.e. at or very near to the site. Differences in pollen production, dispersal and preservation lead to bias in pollen assemblages in favour of wind-pollinated taxa and those with more durable pollens and spores. As a result, many plants are under-represented in modern pollen assemblages, and this may explain the paucity (and in some cases absence) of certain taxa in the fossil record, e.g. Beilschmiedia tawa (McPhail 1980). Fossil macrofloras are often species-poor compared with pollen assemblages (Watts 1978). Poor preservation of delicate-leaved species, and also the identification difficulties mentioned earlier, are probably major factors con• tributing to this species-paucity. Thus both micro- and macrofossil analyses have their own inherent limitations, but combining and comparing data from the two sources can greatly enhance interpretation of a paleoenvironment. There are marked differences between the micro- and macrofossil lists of this study (see Table 1), both in terms of taxa present in one list but absent in the other, and in the relative abundances of those taxa common to both lists. We suggest that these differences largely result from the above-described tendency for microfossil assemblages to reflect vegetation on a larger scale than macrofossil data, because of differential dispersal. As conifers are often poorly- represented in New Zealand coastal forests (Poole & Adams 1980), the dicot-

174 Table 1. Comparative frequencies of vascular plant micro- and macrofossils Species Palynomorph frequency Leaf frequency

Large Trees Agathis australis 4 Beilschmiedia taraire 3 B. tawa 1 Dacrydium cupressinum 6 Elaeocarpus 3 5 Knightia excelsa 4 cf. Laurelia novae-zelandiae 1 6 Libocedrus 5 Metrosideros excelsa group* 13 Nestegis 4 Nothofagus fusca group 2 Phyllocladus 6 Podocarpaceae undiff.** 14 Podocarpus totara group 20 1 Prumnopitys ferruginea 1 cf. P. taxifolia Weinmannia 3 Small Trees, Shrubs Araliaceae undiff. 3 Alectryon excelsus 5 1 Ascarina lucida 3 Coprosma 4 Dodonaea viscosa 4 cf. Geniostoma 2 Griselinia 3 Lagarostrobus 1 Leptospermum 9 Macropiper excelsum 7 1 Myrsine 3 3 Neomyrtus type 1 Pittosporum 1 Pseudopanax 15 Lianes Metrosideros undiff. (liane spp) 34 Meuhlenbeckia 3 Rubus 1 Herbs Astelia 2 Cyperaceae 1 Gramineae 5 cf. Haloragis 1 Pteridophyta Cyathea 19 1 cf. Isoetes 6 monolete fern spores 16 Extinct Species Proteaciditus minimus 2

* = M. excelsa, M. robusta, M. umbellata ** = excluding Dacrydium, Halocarpus *** - M. salicina

175 dominated leaf assembly possibly indicates that the site was very close to the sea, as it is today. The considerable coniferous pollen component is probably largely 'regional', i.e. from further inland. The major differences between micro- and macrofossil representation of taxa in the results can be categorized as follows:

1. Abundant pollen, leaves rare or absent: Podocarpaccae undiff. Podocarpus totara group Metrosideros excelsa group Metrosideros (liane spp.) Pseudopanax Leptospermum 2. Leaves but no pollen: Beilschmiedia tarairi B. tawa

The first four taxa in group one are all tall forest elements; their typically emergent status favouring good pollen dispersal. The pollen from these was probably of regional origin. Pseudopanax and especially Leptospermum are mainly forest margin and serai species of lesser stature, and although no macrofossils were found, their pollen is likely to have originated closer to the site. Species of the genus Beilschmiedia are known to be low producers of pollen (Gardner 1974, McPhail 1980). This factor probably explains their absence from this and other Quaternary pollen assemblages.

CONCLUSION

The samples used in this study are too small to permit a thorough reconstruc• tion of the site vegetation, or for detailed work on the relative representation of micro- and macrofossils. However, there are sufficient data to suggest that differences between the two assemblages can be explained by regional (favour• ing pollen) versus local (favouring leaf) dispersal patterns. Although Beilschmiedia pollen was not found, the presence of leaves of both B. tarairi and B. tawa indicates that these species were growing locally. This supports the contention that the genus was an important component of North Island lowland forests during the Quaternary (as it is today), despite its absence from the pollen record. We believe this exploratory work suggests that the Quater• nary carbonaceous plant beds of the Awhitu peninsula have potential for more detailed paleobotanical investigations.

ACKNOWLEDGEMENTS

We are grateful to D.C. Mildenhall for his palynological assistance. We thank J.A. Grant-Mackie and J.E. Braggins for reviewing the manuscript, and J.E.B. again for Cyathea leaf identification.

176 REFERENCES

Barter T.P. 1976: The Kaihu Group (Plio-Quaternary) of the Awhitu Peninsula, Southeast Auckland. Unpubl. PhD thesis, University of Auckland. Birks H.J.B. & H.H. Birks 1980: 'Quaternary Palaeoecology.' Edward Arnold Ltd., London. Cushing E.J. 1967: Evidence for differential pollen preservation in Late Quaternary sediments in Minnesota. Review of Paleobotany and Palynology 4: 87-101. Erdtman G. 1969: 'Handbook of palynology'. Munksgaard, Copenhagen. Gardner R.O. 1974: Trinucleate pollen in Beilschmiedia Nees (Lauraceae). New Zealand Journal of Botany 12: 2A1-2AA. McGlone M.S. & NT. Moar 1977: The Ascarina decline and post-glacial climatic change in New Zealand. N.Z. Journal of Botany 15 (2): 485-489. Mildenhall D.C. 1973: A guide to morphological characteristics for use in the description of extant and fossil leaves. Tuatara 20 (2): 75-87. Mildenhall D.C. 1979: Holocene pollen diagrams from Pauatahanui Inlet, Porirua, New Zealand. N.Z. Journal of Geology and Geophysics 22 (5): 585-591. McPhail M.K. 1980: Fossil and modern Beilschmiedia pollen in New Zealand. N.Z. Journal of Botany 18: 453-457. Poole A.L. & Adams N.M. 1980: 'Trees and shrubs of New Zealand.' Government Printer, Wellington. Sakai A. & Wardle P. 1978: Freezing resistance of New Zealand trees and shrubs. N.Z. Journal of Ecology 1: 51-61. Watts W.A. 1978: Plant macrofossils and quaternary palaeoecology. pp. 53-67 in 'Biology and Quaternary Environments' (eds D. Walker & J.C. Guppy). Australian Academy of Science, Canberra.

APPENDIX. Characters used to identify fossil leaves.

Label no* Identification Characters used

(2) Laurelia novae-zelandiae Subacute to obtuse, 45-55mm long, bluntly, ser• rate, venation slightly distinct to indistinct.

(3) Alectryon excelsus Ovate-lanceolate. Secondary venation distinct, opposite to sub-opp.; emergent angle 60-70 degrees.

(4) Myrsine salicina Linear-lanceolate, obtuse, c.90x20mm margins entire. Secondary venation indistinct to slightly distinct, + parallel, emergent angle < 30 degrees.

(5) Beilschmiedia tarairi Oval to orbicular, obtuse to retuse. Secondary venation variable opp. to alternate.

(6) Macropiper excelsum Cordate leaf base.

(7) Elaeocarpus sp. Oblanceolate, 60-75 x c. 15mm, finely-serrate margins. Prominent primary domatia.

(9) Beilschmiedia tawa Lanceolate, subacute, 45? x 11 mm. Slightly distinct secondary venation emergent at 45 degrees. 'Grainy' appearance of fine reticulations - consistent with herbarium specimens of B. tawa.

177 (10) Cyathea sp. ID from inspection of sori by Dr J. Braggins.

(11) Griselinia sp. Very glossy, with oblique base.

Podocarpus totara group ID in field; not found in recovered samples.

* Labels on fossils stored in the A.U. Geology Dept. collection.

178