Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 79, 1951
This text is also available in PDF
(1 MB) Opens in new window
– 340 –

The Spore and Pollen Flora of the Cocos-bearing Beds,
Mangonui, North Auckland

[Read before the Wellington Branch, August 10, 1950; received by the Editor, September 6, 1950.]

Introduction and Acknowledgments

For more than fifty years, fossil coconuts from Coopers Beach, near Mangonui, North Auckland, have excited interest both in New Zealand and elsewhere. The coconuts are very well preserved and are commonly washed up on the beach (Fig. 1), together with recent nuts of a naturalised Australian plant, Hakea acicularis R. Br., which is quite abundant in the area. In view of the numerous and well-preserved leaf impressions in the Cocos-beds of the Mangonui area noted by Hector (1872) and McKay (1892, 1894) it is rather surprising that so little botanical work has been done on this interesting flora.

In March, 1950, samples for spore and pollen study were collected from carbonaceous beds at Coopers Beach by Dr. J. Marwick and Mr. R. F. Hay, of the New Zealand Geological Survey, during a brief visit to the area. Identification of this fossil material was greatly facilitated by reference to a collection of spores and pollens of the New Zealand flora placed at the disposal of the writer by Mr. W. F. Harris, of Botany Division. Mr. Harris also made available a draft copy of a paper on the spores of the New Zealand Pteridophyta, which enabled specific identifications to be made in many cases.

Previous Botanical and Geological Work

The coconuts from Coopers Beach were described and figured by Berry (1926) as Cocos zeylandica n. sp. He considered a second type of fruit from this locality to be indistinguishable from Phymatocaryon bivalve Mueller (1878), a fruit of uncertain botanical position from auriferous drifts of probable Pliocene age in Australia. On the evidence of Phymatocaryon bivalve, Berry suggested a Pliocene age for the Mangonui beds.

Hector (1872, p. 155) detailed the following succession of beds at Coppers Beach:

White indurated clay 30
Sandstone 20
Conglomerate 20
Sandy clay 40
Blue clay with ironstone bands 30
Sandy clay with carbonaceous markings and lignite layers 20
Lignite 4–6
Grey laminated clay 8
Conglomerate 20
Slates (Palaeozoic)
– 341 –

The lower beds according to Hector contained fossil leaves “similar to those now existing and fruits like those of the mangrove.” In a later publication, Hector (1886) figured a palm nut (p. 60, fig. 5), giving Coopers Beach and Livingstone Tunnel, Oamaru, as localities. Park (1910) reproduced Hector's figure and apparently included the Livingstone Tunnel fossil nut in his Kaikorai beds, which he considered to be of Miocene age. (The Livingstone Tunnel beds are in fact early Tertiary, and the authenticity of the record of a fossil palm nut from this locality is open to question.)

McKay (1892, 1894), in an account of brief geological reconnaissance in the Mangonui area, detailed the succession of beds at Coopers Beach and Mangonui Township. He did not find Cocos zeylandica in place, but thought that the specimens washed up on the beach were derived from blue and white clay, interbedded in volcanic breccia, about 50 ft. above the contact with “Palaeozoic” rocks, near Mangonui Township. The following succession of beds at Coopers Beach was described by McKay (1892). No thicknesses were given:

  • White pipeclay

  • Lignite bands (up to 4 ft. thick)*

  • White pipeclay

  • Sand and blue clay

  • Conglomerate, sub-angular gravel and interbedded sand containing wood changed to lignite, also leaf impressions Unconformity

  • “Palaeozoic beds”

The lignites and associated clay, sand, silt and breccia were correlated by McKay with the Manukau or Trachyte Breccias (1894, map facing p. 80) and given a post-Miocene age.

Location and Nature of Material

Details of the localities of the carbonaceous material collected by Dr. Marwick and Mr. Hay are as follows:

Sample. No.
Sample No.
Grid Ref. Additional Details of Locality
N.7/507 L.63 N.7/023853 6 chs. S. of stream at N. end of Coopers Beach
N.7/508 L.62 N.7/022850 10 chs. W. of house at 15 ft. bluff of sandstone S. end of Coopers Beach
N.7/508A L.61 N.7/023853 Pieces of lignite washed up on Coopers Beach

L.63: Light-coloured mudstone from irregularly bedded, carbonaceous sand and silt overlying conglomerate lenses in sandy silt. Two specimens of Cocos zeylandica were found in place in this light-coloured mudstone, which yielded no spore or pollen flora.

L.62: Carbonaceous mudstone bands in sand and silt.

L.61: Pieces of dark brown to blackish oil shale, bored by the Recent marine mollusc Anchomasa similis (Gray), which are washed up in abundance on Coopers Beach, apparently from an outcrop below low-tide level.

[Footnote] * A proximate analysis of the uppermost of these bands is given in Table I.

– 342 –

L.62 and L.61 yielded satisfactory spore and pollen floras, L.61 being considerably richer.

Proximate analyses of L.61 by Mr. W. Doherty, Mines Department, Greymouth, and of a sample of a 4 ft. lignite given by McKay (1892, p. 75) are shown in Table I.

Table 1
L.61 McKay (1892)
Moisture 13.8% 14.61%
Volatiles 34.5% 33.18%
Fixed Carbon 15.7% 11.17%
Ash 36.0% 41.04%

It is clear from the Volatile/Fixed Carbon ratio that both samples are oil shales rather than coals. The analyst's report on McKay's sample also emphasized this feature. From the similarity of the analyses, L.61 is probably from McKay's upper lignites.

The stratigraphical succession in the Coopers Beach area is rather obscure, as can be seen from the rather conflicting succession of beds given by McKay and Hector, but L.61, L.62 and L.63 probably all come from within a thickness of 100 ft. of beds, so that they are all of approximately the same age.

Preparation and Identification of Material

L.63 and L.62 were heavily mineralised. They were boiled in hydrofluoric acid for 15 minutes, then in 10 per cent. KOH for 15 minutes, washed, stained with basic fuchsin, and mounted in glycerine jelly. L.61 was simply boiled in 10 per cent. KOH for 15 minutes, washed, stained, and mounted, the same treatment as for a Recent peat sample.

The close similarity of spore and pollen forms within certain genera and families restricts some of the present identifications to genera or as in the case of the Cyperaceae (sedges) and Gramineae (grasses), for example, to families. New species indicated in the following floral list (Table 2) are, however, definitely distinct from any spores and pollens of living New Zealand species. No description of these new species is given in this paper as the problem of fossil spore and pollen nomenclature is to be discussed at the International Botanical Congress in 1950. Descriptions of these fossil forms will be published as soon as possible.

In both L.61 and L.62 there are a number of large, monosulcate pollens, 40μ–63μ, long, which conform very closely to the description and illustrations of the pollen of Cocos nucifera Linn. given by Selling (1947). The pollen of Rhopalostylis sapida Wendl. and Drude (Nikau Palm) also fits this description very closely. If, however, a size-frequency curve is constructed from the monosulcate pollens in L.61 and L.62, a distinct bimodal curve is obtained, with a peak on 49μ and another between 57 and 60μ. A size frequency curve constructed from reference slides of Rhopalostylis sapida gave a peak between 49 and 54μ. Thus, the pollen of R. sapida appears to be present, together with another member of the Palmae. It can be inferred that this pollen is probably that of Cocos zeylandica.

Use of the scientific names of living plants for pollens and spores which are identical with those of living species does not preclude

– 343 –

the possibility that some plants may have differed in leaf, flower or fruit to an extent which would justify specific distinction.

Table 2
L.61 L.62
Lycopodium laterale R. Br. X
Schizaea sp. X X
Hymenophyllum sp. X
Cyathea cf. smithii Domin. X
Blechnum sp. X X
Microsorium diversifolium Copeland X X
Polypodiaceae n. sp. A X X
Polypodiaceae n. sp. B X
Podocarpus sp. X
Laurelia novae-zealandiae A. Cunn. X X
Macropiper excelsum Miq. X ?
Muehlenbeckia sp. X X
Dysoxylum spectabile Hook. X X
Proteaceae n. sp. A X
Proteaceae n. sp. B X
Metrosideros sp. X X
Leptospermum sp. X
?Coriaria sp. ? X
Aristotelia sp. X
Weinmannia sp. X
Nothofagus n. sp. A X X
Griselinia sp. X
Alectryon cf. excelsum Gaertn. X
Nothopanax cf. arboreum Seem. X
? Avicennia officinalis Linn. X
Potamogeton cf. cheesemanii Benn. X
Rhipogonum scandens Forst. X
Typha angustifolia Linn. X
Cordyline sp. X X
? Rhopalostylis sapida Wendl. and Drude X X
? Cocos zeylandica Berry X X
Juncaceae X
Cyperaceae X X
Gramineae X X


In Table 3 the relative abundance of forest and non-forest spores and pollens is shown as percentages. Specimens too distorted for identification are listed under the heading “others.” In examining the slides, such unidentifiable spores and pollens are all recorded and then divided proportionally amongst the forest and non-forest elements.

Table 3
Per cent. Abundance
L.61 L.62
Ferns 3 21
Podocarpus sp. 1
Laurelia 8 15
Proteaceae n. spp. 5
Metrosideros 16 9
? Coriaria 1 1
Aristotelia 4
Weinmannia 2
Nothofagus n. sp. 7 12
Griselinia 4 4
Alectryon 3
Nothopanax 2 6
Macropiper 2 2
– 344 –
Muehlenbeckia 18 10
Dysoxylum 3 3
Rhipogonum 2
Cocos 4 8
Grass 1 1
“Others” 14 8
Lycopodium 1
Potamogeton 15
Leptospermum 36 20
Cyperaceae 17 70
Typha 13
Juncaceae 3
Cordyline 4 8
“Others” 11 2

Forest Vegetation

In both L.61 and L.62 a forest of a coastal broadleaf nature is clearly indicated. Its chief constituents are trees allied to (if not identical with) Dysoxylum spectabile (Kohekohe), Laurelia novaezealandiae (Pukatea), Metrosideros (Rata or Pohutukawa), Nothopanax, Rhopalostylis sapida. (Nikau Palm), Weinmannia (Kamahi or Tawhero) and presumably Cocos zeylandica. Although Agathis australis (Kauri) pollens were particularly searched for, none was seen. It is possible that, because of the delicate nature of the exine, this pollen may not always be preserved, but Cookson (1947) reported Araucarian pollens from Tertiary lignites of the Kerguelen Archipelago, and the writer has noted possible Araucarian pollens in beds of Jurassic age in New Zealand.

The high percentage of Muehlenbeckia pollens in L.61 is rather unusual, because, although woody lianes like M. australis are particularly abundant in tropical or sub-tropical rain forests, such pollens usually form only a small percentage of the forest pollen in New Zealand Recent peats. It may simply be a case of local over-representation either from a forest species or from a coastal dominance of a form like M. complexa.

Non-forest Vegetation

The non-forest vegetation of L.61 and L.62, consisting of Cyperaceae (sedges), Typha (Raupo), Potamogeton (pond-weed), Juncaceae, Cordyline (Cabbage Tree) and Lycopodium (Club Moss), is typical of that found on poorly drained ground. The occurrence in L.61 of pollens which are similar to Avicennia officinalis (Mangrove) is noteworthy in view of the comment by Hector (1872, p. 155) of the presence of “fruits like that of the mangrove” in the Coopers Beach beds. Whether these fruits were actually Cocos zeylandica or Phymatocaryon bivalve is not made clear in any later publication, but as these species are quite distinct from the fruit of the mangrove, it is possible that fossil mangrove fruits may also be present.

Conditions of Accumulation

From the proximate analysis of L.61 given in Table 1 it is evident that this sample is more correctly termed an oil shale than a lignite. The floral evidence also indicates somewhat different conditions of

– 345 –

accumulation from that of a peat, for typical peat-forming swamp or bog plants such as Cyperaceae, Juncus, Typha, Phormium, Gleichenia and Hypolaena are either absent or make up only a small percentage of the total number of spores and pollen identified. The non-forest elements are only 25 per cent. of the total in this sample.

Oil shales, according to Twenhofel (1926, pp. 289–290) are for the most part “laid down in ponds, lakes, lagoons, inland seas and shallow gulfs too deep for vascular plant growth in places, too extensive to permit a considerable contribution of detritus of land origin and sufficiently protected against tidal scour and aerating currents to favour the necessary degree of conservation of humic products.” The high ash content (36 per cent.) is additional evidence in favour of the type of environmental conditions postulated by Twenhofel, but neither in a hand specimen, nor under the microscope, is mineral matter at all evident.

In L.62. Cyperaceae are of greater importance, both in the non-forest vegetation and in the flora as a whole. The sample is a carbonaceous silt in which occasional macroscopic remains of a sedge-like plant can be seen. This deposit probably represents either periods of flooding by a river, which buried and preserved the surrounding sedge and manuka vegetation, or a drift accumulation of plant material. Forest pollens are considerably fewer and less varied, suggesting that forest was further from the site of accumulation.


The presence, in abundance, of an extinct species of Cocos is undoubtedly the best palaeoclimatic indicator in the Coopers Beach flora. Of living species, Cocos nucifera Linn. is the only one with a wide distribution, the others being confined to Tropical South America. Reliable data on the natural northern and southern distributional limits of C. nucifera are hard to obtain, owing to the difficulty in deciding in many cases whether or not the plant was introduced for food purposes. Regarding its northern limits, Selling (1947, p. 337) commented as follows: “The Hawaiian Islands situated on the northern limit of the Pacific range of the species are generally considered almost too cold for it; the most vigorous groves are situated on the warmer, leeward sides of the southernmost islands, Maui and Hawaii.” These islands are about Lat. 20°N. and the mean annual temperature at Honolulu is 23·3°C. (74°F.). In the Southern Hemisphere, the occurrence of C. nucifera on the Queensland coast at about Lat. 25°S. can perhaps be taken as an extreme southern limit. The mean annual temperature at this latitude is approximately 70°F. (Good, 1947, map facing p. 226). For comparison, the mean annual temperature at Waipapakauri, near Mangonui, at the present day is 58°F. Thus, if Cocos zeylandica grew within similar temperature limits to C. nucifera, the mean annual temperature at the time of the deposition of the Cocos-beds may have been at least 12°F. higher than at present.

The remainder of the forest elements in the flora (apart from Nothofagus and Podocarpus) would, by themselves, suggest a climate comparable to that of the present North Auckland region. Close relatives and species of such genera as Dysoxylum, Metrosideros, Wein-

– 346 –

mannia, Alectryon, Macropiper and Nothopanax are to-day important constituents of tropical and sub-tropical rain forests. There are other genera, viz., Laurelia, Griselinia and Aristotelia, which are more characteristic of warm temperate rain forests. Consequently an analogous climate would possibly be that of coastal Queensland and not truly tropical.

The presence of Nothofagus pollen in the same flora as Cocos may appear anomalous. It does not imply, however, that Cocos and Nothofagus grew side by side, as the pollen of the latter could have been blown to the site of accumulation from trees growing at a considerable altitude. In Australia the most northern occurrence of beech forest is at an altitude of 3,000 ft. in the MacPherson Range (Lat. 28°S.), near the Queensland-New South Wales border (Herbert, 1935, p. 363). The occurrence of beech forest within 3° of the Equator is described by Archbold, Rand and Brass (1942, p. 282) in their summary of the 1938–1939 Archbold Expedition to New Guinea. “In the mountains south-west of Bernhard Camp, where it (the beech forest) occupies the upper parts of the ridges and its lower contact is with the rain forest, an outlier body is encountered at the very low altitude of 850–900 metres (2,750 ft.–2,920 ft.).” This is the lowest altitude at which Nothofagus was encountered on the expedition, the greatest development of beech forest being from 4,900 ft. to 10,200 ft. By analogy with the present distribution of Nothofagus in Southern Queensland, the habitat of the Mangonui Nothofagus (and probably also Podocarpus) may be estimated as some 3,000 ft. higher than that of Cocos zeylandica, although the presence of Nothofagus in New Guinea at a little below 3,000 ft. suggests that this estimate may be too high.


There is no evidence to suggest that L.61 and L.62 are markedly different in age. To give an estimate of that age is rather difficult at present, because of the lack of dated material for comparison. Judging from the climatic facies of the flora and from the number of spores and pollens which have no living counterpart in the present New Zealand flora, the beds are certainly pre-Nukumaruan and probably pre-Wanganuian (pre-Pliocene). The beds do not appear to be older than Miocene, judged by comparison with available New Zealand spore and pollen floras of probable Oligocene age, although Nothofagus n. sp. A and Proteaceae n. sp. B are present in the Mataura lignites of probable Duntroonian (lower Oligocene) age (Willett, 1948, p. 232). The Miocene is considered to be a warm period (Finlay, 1947, p. 337) and a Miocene age is therefore suggested for the Mangonui Cocos-beds.

McKay (1894, map facing p. 80) correlated the Mangonui Cocos-beds on lithological grounds with the Manukau breccia, which is now included in the Waitemata group of Altonian age (lower Miocene, Finlay and Marwick, 1948, p. 24). Altonian molluscan and foraminiferal faunas are interpreted as indicating at least sub-tropical sea temperatures (Finlay, 1947, Finlay and Marwick, 1947), and the Ostracod faunule of this age from Pakaurangi Point (F.5730) also indicates similar sea temperatures. (Personal communication from

Picture icon

Fig. 1—Coopers Beach from south-east end. Concentrations of fossil wood, fruits of Cocos zeylandica, and recent nuts in foreground; low tide.
Photo. by Dr. J. Marwick.

Picture icon

Location of Mangonui

N. de Hornibrook, New Zealand Geological Survey.) Thus climatic evidence is consistent with an Altonian age for the Mangonui Cocos-beds.


In the beds underlying volcanic breccia and conglomerate from a locality given as Parengarenga North Head, McKay (1894, pp. 82-83) collected specimens of large palm fronds.

In January 1951, the writer in company with Mr. R. F. Hay and Mr. N. de Hornibrook, visited Parengarenga Harbour and obtained a lignite sample from beds which appear interbedded with volcanic breccia and conglomerate 15 chains at 34° from Trig., Coal Point, Muriwhenua S.D. The microflora from this sample (N 2/506, L. 235) is similar to that of the Coopers Beach oil shale (L. 61) and includes a number of large palmaceous pollens.

At the same time was collected a sample of sandstone from 10 ft. below the volcanic breccia and conglomerate near Dyson’s old station (N 2/508, F. 7168, 150 chains at 260° from Trig., Coal Point, Muriwhenua S.D.), which yielded a fair microfauna. Mr. Hornibrook reports on this sample as follows: "The important species of Foraminifera are: Amphimorphina cf. acuminate (Hantken) (first appears in the Hutchinsonian); Siphogenerina rerensis Fin. (first appears in the Altonian); Uvigerina miozea Fin. (first appears in the Altonian); Anomalinoides vitrinoda (Fin.), sensu stricto (first appears in the Altonian); Plectofrondicularia parri Fin. (first appears in the Hutchinsonian). In the absence of Orbulina universa d’Orb. and other species indicating a middle or upper Southlandian age, this faunule appears to be Altonian and is undoubtedly no older."

In the light of this more recent work, the evidence outlined above for a Miocene and probable Altonian age for the Cocos-bearing beds of Coopers Beach appears considerably strengthened.

Literature Cited

Archbold, R., Rand, A. L., and Brass, L. J., 1942. Results of the Archbold Expeditions. No. 41. Bull. Amer. Mus. Nat. Hist., vol. 79, art. 3, pp. 197-288.

Berry, E. W., Cocos and Phymatacaryon in the Pliocene of New Zealand. Amer. Journ. Sci., Series 5, vol. 12, no. 69, pp. 181-184.

Cockayne, L. 1928. Vegetation of New Zealand. Die Vegetation der Erde, 14 (2nd Edition).

Cookson, Isabel C., 1947. Plant Microfossils from the Lignites of Kerguelen Archipelago. Br. Austr. N.Z. Ant. Res. Exped. (1929-1931), Reports Series A, vol. 2, part 8, pp. 127-142.

Finlay, H. J., 1947. The Foraminiferal Evidence for Tertiary Trans-Tasman Correlation. Trans. Roy. Soc. N.Z., vol. 76, pt. 3, pp. 327-352.

— and Marwick, J., 1947. New Divisions of the New Zealand Upper Cretaceous and Tertiary. N.Z. Journ. Sci. and Tech., vol. 28, no. 4, Sect. B, pp. 228-236.

— 1948. “Tertiary” in The Outline of the Geology of New Zealand. Dept. Sci. and Ind. Res., H. H. Tombs Ltd., Wellington.

Good, R. 1947. The Geography of the Flowering Plants. Ballantyne Press, London.

– 348 –

Hector, J., 1872. Report on the Coal Seams at Wangaroa and Mangonui, Auckland. Rep. Geol. Expl., 1871–1872, vol. 7, pp. 153–158.

—– 1886. Outline of New Zealand Geology, Wellington.

Herbert, D. A., 1935. The Climatic Sifting of Australian Vegetation. Report, 22nd Meeting A. and N.Z. Assoc. for Advancement of Science, vol. 22, pp. 349–370.

McKay, A., 1892. On the Lignites of Coopers Beach, Mangonui, Auckland. Rep. Geol. Expl. 1891–1892, vol. 21, pp. 72–76.

—– 1894. On the Geology of Hokianga and Mangonui Counties, Northern Auckland. Rep. Geol. Expl., 1893–1894, vol. 22, pp. 70–90.

Mueller, F., 1878. Ann. Rep. Dept. Mines, N.S.W., p. 170.

Park, J., 1910. Geology of New Zealand. Whitcombe and Tombs, Wellington.

Selling, O. H., 1947. Studies in Hawaiian Pollen Statistics. Part 2. The Pollens of the Hawaiian Phanerogams. Bernice P. Bishop Museum, Spec. Pub. 38, pp. 1–430.

Twenhofel, W. H., 1926. Treatise on Sedimentation. Waverley Press, Baltimore, U.S.A.

Willett, R. W., 1948. Preliminary Report on the Lignite Deposits of the Mataura Valley, Eastern Southland Coalfield. N.Z. Journ. Sci. and Tech., vol. 29, no. 5, Sect. B, pp. 228–246.