Fossil Eucalyptus Remains from the Middle Miocene Chalk Mountain Formation, Warrumbungle Mountains, New South Wales W. B. K. HOLMES, F. M. HOLMES and H. A. MARTIN Holmes, W. B. K., Holmes, F. M., & Martin, H. A. Fossil Eucalyptus remains from the Middle Miocene Chalk Mountain Formation, Warrumbungle Mountains, New South Wales. Proc. Linn. Soc. N.S. W. 106 (4), (1982) 1983: 299-310. Eucalyptus fruits and leaves are preserved in the diatomite of the Middle Miocene Chalk Mountain Formation (new name), Warrumbungle Mountains, New South Wales. The fossil eucalypts show some features of advanced states in fruit and venation and are compared with some extant species. Based on present knowledge of Eucalyptus sensu lato, the fossils provide evidence of two separate phylogenetic lines being present in Middle Miocene time. The microflora recovered from the diatomite and associated lignite includes pollen grains of Myrtaceae (Myrtaceidites spp) of which a few are referable to Eucalyptus. W. B. K. Holmes, Hon. Research Fellow, Department of Geology, University of New England, Armidale, Australia 2351; F. M. Holmes, 'Noonee Nyrang', Gulgong Road, Wellington, Australia 2820 and H. A. Martin, School of Botany, University of New South Wales, Ken- sington, Australia 2033; manuscript received 8 June 1982, accepted for publication 18 August 1982. Introduction Despite the present day prominence of Eucalyptus and the great diversity of its species, little is known of its evolutionary history from the fossil record (Martin, 1978; Briggs and Johnson, 1979). For phytogeographical reasons alone, there can be little doubt that the genus is of ancient Australian origin (Barlow, 1981); however, the 'eucalypt-type' pollen, Myrtaceidites eucalyptoides Cookson and Pike, which is probably bloodwood-^4 ngophora first appears in the Oligocene (Martin, 1981). Fossil leaves which resemble eucalypts in gross form but which were inadequately described and figured, have been recorded from many poorly-documented Tertiary localities in New South Wales, Victoria, Tasmania, South Australia and Central Australia. References to these occurrences are given in Duigan (1951). Lange (1978) reported a diverse assemblage of Lepidospermae fruit, including a number of forms closely similar to Eucalyptus , preserved as impressions on silicified slabs of unknown age from the arid zone of South Australia. Ambrose et al. (1979) have also collected Eucalyptus fruit similarly preserved to those of Lange. The age of these fossils from Stuart Creek, northwest of Lake Torrens in South Australia is thought to be either Eocene-Oligocene or Miocene. Deane (1902) described several forms of eucalypt-like leaves from the Berwick Quarry in Victoria. A re-investigation of this flora by Nelson and Webb (in prep.) has shown that the cuticle of one of the leaves is similar to the cuticle of some extant Eucalyptus. Impressions of a eucalypt-like fruit and of rainforest leaves are also present but no Eucalyptus type pollen has been observed in the microflora which is dominated by Nothofagus pollen (Nelson and Webb, pers. comm.). Eucalyptus fruit and leaves and the microflora from the Middle Miocene diatomite deposit at Bugaldie near Coonabarabran in northwestern New South Wales are described below. Associated with the eucalypts are remains of leaves and flowers of plants (Holmes, in prep.) which have affinities with extant plants growing in situations with a much higher rainfall than the present Bugaldie average annual rainfall of 650 mm. Also in the diatomite are fossils offish (Hills, 1946), a bird described as an owlet- Proc. Linn. Soc. N.S.W. 106 (4), (1982) 1983 300 MIDDLE MIOCENE FOSSIL EUCAL YPTUS nightjar (Rich and McEvey, 1977), diatoms (Thomas and Gould, 1981 a and b), insects and freshwater bivalves. Geology and Age of the Diatomite Chalk Mountain, which lies 6 km south of Bugaldie village, is a remnant of an extensive dissected plateau of late Permian, Triassic and Jurassic freshwater sediments capped by basalt associated with the Warrumbungle Volcano Complex to the south. Lacustrine sediments which are interbedded between basalt flows on the Jurassic sediments and a 25 metre thick overlying basalt flow, outcrop around the escarpment of Chalk Mountain. Diatomite in these sediments was commercially exploited from 1919 until 1980. Griffin (1961) gave a detailed description of the geological and commercial aspects of the deposit but he did not name the formation. Reports by Herbert (1968a and b) deal with the diatomite. Chalk Mountain Formation (Holmes and Holmes, new name). The Chalk Mountain Formation is a lake deposit over 15 metres in thickness and is composed of almost horizontal layers of mudstone, clay, diatomite, three interbedded horizons of tuf- faceous clay and a thin band of lignite. The type locality is Quarry A on the western face of Chalk Mountain immediately below the basalt flow capping the mountain (grid reference 199148, Gilgandra 1: 250 000 Geological Series Sheet SH 55-16). The base of the formation is obscured at this locality by the basaltic talus slope and by quarry waste. On the roadway up the northern slope of the mountain the basal beds of mudstone and clay rest on basalt. The section exposed in Quarry A is shown in Table 1. The area of lake sediments that have disappeared due to erosion around Chalk Mountain is unknown, but it is likely that the original lake was very much larger than the 38 hectares of the existing deposit. Plant macro-fossils have been collected only from the diatomite. Plant micro-fossils occur in both the diatomite and the thin band of lignite near the top of the formation. For a horizontal distance of about thirty metres in from the outcropping face, the diatomite has weathered to a pure white lightweight material much of which has been removed by open-cut mining and some tunnelling. Beyond the weathered zone the diatomite is of a grey to black colour due to the presence of organic matter and is known as 'black earth'. The 'black earth' was only sparingly exploited due to the necessity to process it to remove the organic matter. In the weathered zone of the diatomite the plant macro-fossils are preserved as colourless or limonite-stained impressions and in the unweathered zone as colourless impressions Table 1 Type Section of Chalk Mountain Formation exposedin Quarry A on Chalk Mountain, Bugaldie (after Griffin, 1961) Top Basalt 4.26 m Clay with lignite band 0.91m Tuff, clay and grit 4.05 m Clay and diatomite 0.13m Tuff, clay, diatomite and gritty sandstone 3.04m Diatomite high grade 0.08 m Clay and tuff 0.81m Diatomite high grade 3.0 m Mudstone, tuff and clay Bottom Basalt, trachyte and andesite overlying Jurassic Pilliga Sandstone Proc. Linn.Soc.N.S.W. 106(4), (1982) 1983 W. B. K. HOLMES, F. M. HOLMES AND H. A. MARTIN 301 or carbonaceous compressions. The diatomite exhibits seasonal varves each 0.33 mm to 0.5 mm thick (Thomas and Gould, 1981b). A deposition time of between 14 000 and 21 000 years would have been needed to accumulate the total thickness of 7 metres of diatomite, assuming the varves to be annual. The bands of tuff within the deposit indicate contemporaneous eruptive volcanic activity during the period of lacustrine sedimentation. K/Ar dating of igneous rocks from localities around the Warrumbungle complex (Dulhunty and McDougall, 1966; McDougall and Wilkinson, 1967; Wellman and McDougall, 1974) show that the volcanic activity took place during the period 17 million years to 14 million years before the present. The nearest sites to Chalk Mountain with K/Ar dates are Looking Glass Mountain (16.6 ± 0.6 m.y.), 15 km to the west-northwest, and Woorut Mountain (15.5 ± 0.4 m.y.), 20 km to the south (Wellman and McDougall, 1974). As no study has been made of the geology of the northern portion of the Warrumbungle Volcano, the stratigraphic relationships of these localities to Chalk Mountain are not known. Therefore, with present information, we can only date the diatomite as being within the age limits of the active Warrumbungle Volcano — 17 million years to 14 million years, Middle Miocene. The Fossil Eucalypts „ Family MYRTACEAE Genus Eucalyptus L'Herit. Eucalyptus bugaldiensis Holmes and Holmes sp. nov. Fig. 1 A, Fig. 2A-D Diagnosis: Fossil infructescence of a group of three umbellasters of seven or fewer small, Fig. 1. A Eucalyptus bugaldiensis sp. nov. AMF61713 holotype x 2. B Eucalyptus sp. Leaf form B AMF61721 x Z.C Eucalyptus sy. Leaf form A AMF61724 x 2. Proc. Linn. Soc. N.S. W. 106 (4), (1982) 1983 302 MIDDLE MIOCENE FOSSIL EUCAL YPTUS hemispherical, ribbed, woody fruit attached by tapering pedicels to peduncles. Rim of fruit broad; exerted valves broad and low. Description: Infructescence of three umbellasters on peduncles 5 mm - 10 mm in length, and 0.8 mm in width attached at a common point to a stem. Umbellasters composed of seven or fewer fruit attached to the peduncle apex by pedicels (anthopodia) 5 mm - 10 mm in length widening upwards to merge into the base of the fruit. Fruit hemispherical, 4 mm - 5 mm in diameter. External surface ornamented by 7 - 10 longitudinal ribs which occasionally fork near the rim. Rim flat ca 1 mm in width. Exerted valves forming a broad low triangular projection 0.5 mm above the rim. Type material: Holotype AMF61713 and counterpart AMF61714. Paratypes AMF61715-20. The Australian Museum, Sydney, N.S.W., Australia. Type locality: Diatomite beds in Quarry A on Chalk Mountain, 6 km south of Bugaldie village near Coonabarabran, New South Wales. Grid reference 199148, Gilgandra 1: 250 000 Geological Series Sheet SH 55-16. Chalk Mountain Formation, Middle Miocene. Discussion: The fossil fruits of Eucalyptus bugaldiensis are preserved as impressions in the soft diatomite. The hollow moulds contain no organic material. The surface of the diatomite was hardened with Bedacryl 122X and latex casts were then taken of the holotype and its counterpart. When whitened with ammonium chloride, the latex casts revealed some of the finer details not readily seen in the original specimens. The distal ends of the peduncles show scars where additional buds or fruits have been aborted or lost during development or preservation. The umbellaster on AMF61715 has five fruit attached by pedicels to the peduncle. In addition to the still-attached fruit, the distal end of the peduncle shows the abscission scars of another two pedicels. This suggests that the total flower number per umbellaster was seven, a number which may be ex- pected in dichasial inflorescences in a family with primarily opposite and decussate phyllotaxy. Extant species in subgenus Symphyomyrtus very often have seven flowered umbellasters. The arrangement on the holotype of three umbellasters about a common point, is found in the eucalypt conflorescence sub-types of Tj and S 3 of Johnson (1972, fig. 3). The Tj sub-type characterizes the small tropical subgenus Telocalyptus (section Equatoria) and the S 3 sub-type occurs mainly in the subgenus Eudesmia, but the fossil fruits are quite different from any in Eudesmia. However, three umbellasters can be so attached in some variants of sub-type Se also (L. A. S.Johnson, pers. comm.) when the phyllotaxy in the inflorescence region is opposite rather than disjunct-opposite (Briggs and Johnson, 1979), as for instance in some inflorescences of Eucalyptus microtheca of subgenus Symphyomyrtus section Adnataris. The infructescence and the individual fruits of E. bugaldiensis are very close in form to some extant species of Eucalyptus . However, due to the type of preservation and to the lack of supporting organs, this new species cannot at present be placed with certainty in any particular subgenus. The fruits of E. microtheca F. Muell. (Blake, 1953, pi. 30, figs 20-36) are closely similar in size and form to those of E. bugaldiensis. E. microtheca is very widely distributed throughout the continent (Blake, 1953, map 24) on ground that is subject to seasonal flooding. E. raveretiana F. Muell. from areas near the coast in central and northern Queensland has similar but rather smaller fruit. None of the diverse forms of fossil Eucalyptus fruits from South Australia (Lange, 1978; Ambrose et al. , 1979) appear to be similar to E. bugaldiensis but their different aspect of preservation makes comparisons difficult. The cluster of fruits from Berwick, Victoria (Nelson and Webb, in prep.) is superficially similar to E. bugaldiensis but lacks evidence of valves. PROC. Linn. Soc.N.S.W. 106(4), (1982) 1983 W. B. K. HOLMES. F. M. HOLMES AND H. A. MARTIN 303 A / H* B •i, > • ~ « i v % \* 1 fpn QifU B ^ /^j^ D c Fig. 2. A-D Eucalyptus bugaldiensis Holmes & Holmes sp. nov. Holotype. A, B AMF61713, A x 1 ; B x 4. C, DAMF61714 counterpart of AMF61713,C x 1;D x 4. Proc. Linn. Soc. N.S.W. 106(4), (1982) 1983 304 MIDDLE MIOCENE FOSSIL EUCAL YPTUS Fig. 3. A, B Eucalyptus sp. Leaf form A AMF61724 A x 1, B x 2. C, D Eucalyptus sp. Leaf form B MMF15284 x 1.DAMF61721 x 2. Proc.Linn.Soc.N.S.W. 106(4), (1982) 1983 W. B. K. HOLMES, F. M. HOLMES AND H. A. MARTIN 305 Eucalyptus sp. Leaf- form A Fig. lC.Fig. 3A-B Description: Leaf lanceolate to narrow lanceolate, falcate, asymmetrical about lamina base; margin scarious, entire, sometimes insect-damaged. Length 8 cm- 14 cm; width 1.5 cm-2 cm, widest at one quarter of length from the base and tapering gradually to an obtuse apex. Petiole curved, to 2 cm in length. Midvein prominent, tapering from the petiole to the apex. Secondary veins parallel, ca 1 mm apart, sometimes with a minor vein in between; attached to the midvein at 50°-60° and running straight to the intramarginal vein which runs close and parallel to the leaf margin. Tertiary veins form an irregular reticulum of about four rows of areoles between the secondary veins. Material: AMF61721-3. Discussion: The venation pattern of leaf-form A is of the Transversae class (Cambage, 1913; Maiden, 1922) (not to be equated with the section Transversaria of Pryor and Johnson, 1971) and is present in the subgenera Corymbia and Blakella of Eucalyptus and in Angophora which Pryor and Johnson (1971) have suggested is also of subgeneric rank in Eucalyptus s.l. Extant species with venation similar to leaf- form A are the widespread Angophora floribunda (Sm.) Sweet and Eucalyptus trachyphloia F. Muell., a bloodwood tree, which grows on the drier sandy soils of the Upper Hunter River and Pilliga areas of New South Wales and into Queensland. Eucalyptus sp. Leaf- form B Fig. IB, Fig. 3C-D Description: Leaf lanceolate, slightly falcate; margin entire; apex obtuse; lamina base asymmetrical. Width 16 mm-23 mm; length 8 cm- 12 cm. Secondary veins sub- parallel, spaced irregularly, 5-6 per cm, running a slightly undulate course to the intramarginal vein at an angle of 55°-60° to the midvein. Secondary veins occasionally fork and anastomose to form a very elongate reticulum. Petiole curved, to 2 cm in length. Material: AMF61 724-5. MMF15284. Discussion: MMF15284 (Fig. 3 C) shows portions of five leaves, one of which is at- tached to a slender stem which has sub-opposite projections from where the other leaves may have become detached. Leaf- form B differs from leaf- form A by the wider spaced and irregular course of the secondary veins. The venation pattern is in- termediate between the Transversae and Obiiquae classes (Cambage 1913; Maiden 1922). This general pattern of leaf venation is present in extant species within the subgenus Symphyomyrtus section Adnataria series Oliganthinae or series Pruinosae and in the subgenus Telocalyptus series Degluptae. Leaves of Eucalyptus raveretiana F. Muell. (a member of Telocalyptus) bear a close resemblance to this fossil form. Palynology of the Chalk Mountain Formation (H.A.M.) Table 2 lists the taxa found in the 'black earth' diatomite and in the lignite from near the top of the formation. Preservation is rather poor which limits identifications, and this may be the result of initial poor conditions at the time of deposition or of subsequent weathering on exposure. Nevertheless, the assemblages are workable and invaluable because they can be dated by independent evidence. The two assemblages are essentially the same, although there is a quantitative difference between the diatomite and the lignite. This difference could result from the different conditions of deposition. Diatomite is formed in lakes and so would ac- cumulate more wind-blown pollen. Because lignite forms in swamps most of the pollen Proc. Linn. Soc. N.S.W. 106(4), (1982) 1983 306 MIDDLE MIOCENE FOSSIL EUCAL YPTUS Table 2 The Pollen Frequencies Fossil Name Botanical Affinity Diato- Lig- mite nite ALGAE Botryococcus braunii Kutzing FERN SPORES Cyathea paleospora Martin Deltoidospora inconspicua Martin Laevigata sporites ovatus Wilson & Webster Polypodiidiles sp. Reticuloidosporites minisporis Martin Unknown spore GYMNOSPERMS Araucariacites australis Cookson Cupressaceae sp. indet Dacrydiumflorinii (Cookson & Pike) Cookson Phyllocladus palaeogenicus Cookson & Pike Podocarpus australiensis (Cookson & Pike) Martin Podocarpus elliptica (Cookson) Martin ANGIOSPERMS Cupanieidites orthoteichus Cookson & Pike Ericipites sp. Graminidites media Cookson Haloragacidites harrisii (Couper) Harris Loranthaceae Malvacipollis diversis Harris Milfordia sp. Monosulcites cf Palmae Myrtaceidites eucalyploides Cookson & Pike Myrlaceidiles cf A/, eucalyploides Cookson & Pike M. parvus Cookson & Pike Myrtaceae not identified further Nothofagus aspera Cookson N. emarcida Cookson Polyporina chenopodiaceoides Martin Proteacidites sp . Quintinia psilatispora Martin Tricolporites scabratus Harris Tricolporiles cf Cunoniaceae Tricolporites cf Goodeniaceae Tricolporopollenites ivanhoensis Martin Triorites minisculus Mclntyre T. orbiculatus Mclntyre Trio riles c f U lm ac e ae Unknown tricolpate/tricolporates Botryococcus braunii Cyathea spp. Total Fern Spores Agathis and A raucaria Cupressaceae Dacrydium Sec. B Phyllocladus spp. Podocarpus Sect. Dacrycarpus Podocarpus , most other sections Total Gymnosperms Tribe Cupaneae (Sapindaceae) Epacridaceae Gramineae Casuarma Loranthaceae Austrobuxus — Dissiliaria (Euphorbiaceae) Restionaceae Like Palmae Bloodwood — Angophora, possibly other genera Possibly other eucalypts Possibly Tristania — Backhousia — Baeckia and other genera Myrtaceae Total Myrtaceae Nothofagus, the menziesii type (e.g. N. moorei, N. cunninghamii) Nothofagus, the brassii type Total Nothofagus Chenopodiaceae and some Amaranthaceae Proteaceae Quintinia like Cunoniaceae like Goodeniaceae probably Rutaceae like Apananthe — Celtis 1.5 2.9 1.5 2.4 1.5 5.6 0.5 0.2 5.4 4.6 17.2 10.7 18.1 0.8 2.2 4.6 1.0 0.5 5.3 5.4 7.6 10.5 29.0 34.8 + 1.2 0.2 13.0 24.0 0.2 + + 0.7 0.5 0.8 6.8 0.2 17.5 5.2 25.2 3.2 0.8 0.5 1.5 2.3 0.5 0.8 3.0 Number of kinds of unknowns 22.1 No. 7 2.0 0.7 0.2 0.2 0.2 1.0 0.2 13.2 No. 6 Notes: (1) Subtotals may not add up exactly because individual percentages have been rounded off. (2) + indicates a presence but not counted. (3) The use of 'like' indicates a tentative identification. Proc. Linn. Soc. N.S.W. 106(4), (1982) 1983 W. B. K. HOLMES, F. M. HOLMES AND H. A. MARTIN 307 would come from the plants growing in and around the edges of the swamps. Fern spores are more abundant and diverse in the lignite. The swamp environment would be more suitable for fern growth. Gymnosperms are well represented in both assemblages but are slightly more abundant in the lignites and of this group, Araucariaceae are the most common. Casuarina is almost twice as abundant in the lignites. Some species of Casuarina grow on the swamp edges today. The Myrtaceae are more abundant in the diatomite than in the lignite. Un- fortunately, the identification of pollen to genera within this family is not particularly reliable. The blood wood-Angophora type is only found in the lignite in a very low concentration. Judging from surface samples, Rose (1981) found that this pollen type does not disperse aerially far from the source of production. However, water transport cannot be ruled out; but swamp vegetation would act as a filter and prevent the pollen from travelling far (Birks and Birks, 1980). The other 'possibly eucalypt' type is smaller than the blood wood-Angophora type, and Rose found that it is more widely distributed. This second type has been found only in the diatomite. Myrtaceous pollen which is not like that of eucalypts is also present (see Table 2). The majority of the pollen cannot be identified beyond the family level because of poor preservation or crumpling which obscures diagnostic features. Nothofagus is present but in very low concentration, indicating transport from a distant source. The small amount of the brassii type may have come from plants growing a long way from the catchment of this lake. Little can be said of the taxa present in low frequencies, for these are usually low pollen producers. The in- determinate tricolpate/tricolporate group is quite abundant. These pollen assemblages must represent forests, judging from the low content of herbaceous elements such as Gramineae and the absence of Compositae. The high content of Myrtaceae in the' diatomite together with the low content of Nothofagus in both samples fits the Myrtaceae phase of the Pliocene (Martin, 1973). Malvacipollis diversus and Triporopollenites ivanhoensis are usually found only in the lower part of the older phase (Martin, unpubl.). The Pliocene age was based only on the general geology of the Riverina region (Martin, 1973), so it may not be reliable. None of the diagnostic species of the mid-late Miocene Triporopollenites bellus Zone, described from the Gippsland Basin of south eastern Victoria (Stover and Partridge, 1973), has been found at Chalk Mountain and the general description of this Zone is a poor fit for the present assemblages. The base of this Zone contains abundant Nothofagus which decreases upwards through it. Its upper limit possibly extends into the Pliocene but the relationship to the overlying Myrtaceae phase (older) has not been documented. The problem is one of geographic variation during the Tertiary. In time-equivalent assemblages, Nothofagus is more abundant in southeastern Victoria than in New South Wales. Conversely, Myrtaceae are more abundant in New South Wales, particularly in the most westerly sites (Martin, 1983). Since the change from an assemblage with high Nothofagus to one with high Myrtaceae depends on the climate becoming drier, this change should have occurred earlier in the drier, inland regions than on the wetter coastal southeastern part of Australia. The Chalk Mountain microflora! assemblages show that the change had occurred in this region by the mid Miocene, some 14-17 million years ago. The Middle Miocene Chalk Mountain Palaeoenvironment The Tertiary pollen record is largely that of rainforest assemblages (Martin, 1978, p. 200) and most deposits of plant macro-fossils show a preponderance of mesophyllous leaves which constituted the 'Cinnamomum' flora (Sussmilch, 1937). The preservation Proc. Linn. Soc. N.S.VV. 106(4), (1982) 1983 308 MIDDLE MIOCENE FOSSIL EUCAL YPTUS of pollen and other plant remains as fossils usually depends on the presence of lakes, swamps or bogs. Therefore, a strong bias exists against the fossilization of plants adapted to growing in drier and harsher environments. Beadle (1981) has suggested that scleromorphic plants began to differentiate at the margins of rainforests and developed along declining soil-fertility gradients. Johnson and Briggs (1981) discussed the possibility of differentiation of scleromorphic taxa during the Early Tertiary in nutrient-deficient forest sites. The large tracts of low-fertility soils derived from the underlying Jurassic Pilliga Sandstone surrounding the Warrumbungles would have provided such an environment. At the present time eucalypts characteristically occur in open forest and woodland associations and only a few species have been successful in rainforest margins (Barlow, 1981). In addition to the eucalypt remains, the Chalk Mountain fossil flora includes fern fragments, podocarp leaves, Ceratopetalum sp. flowers and myrtaceous, lauraceous and other leaves that show similarities in gross form with present-day rainforest species in high rainfall areas of northeastern New South Wales and eastern Queensland. This mixed fossil flora should not be regarded as evidence that eucalypts in the Middle Miocene were adapted to rainforest conditions. Two present-day environmental situations provide a similar mixture of leaves. In moist sclerophyll forests on the North Coast of New South Wales, rainforest will develop, in the absence of fire, beneath tall Eucalyptus trees such as E. grandis W. Hill ex Maiden, E. saligna Sm. and occasionally E. microcorys F. Muell. Along the coastal strip of New South Wales and Queensland, rainforest species sometimes form a fringing forest along the banks of watercourses. Sclerophyll forest abuts and often overhangs this rainforest. Leaf-litter in the stream bed is a mixture of eucalypt and rainforest plant remains. At Chalk Mountain in Middle Miocene time, the rich soils derived from the basalt, the moist conditions adjacent to the lake, and the probable higher rainfall (Kemp, 1978) would have favoured rainforest growth. Sandstone hills around the lake and areas away from the soil-improving influence of the basalt would have supported a sclerophyll forest or woodland to provide the eucalypt, herb and grass remains that would have been carried by wind or water to mingle with the rainforest remains in the lake sediments. The high percentage of gymnosperm and Casuarina pollen probably indicates the proximity of these plants to the site of deposition. Conclusion On the basis of leaf venation, the Chalk Mountain fossils show that by the Middle Miocene the Eucalyptus genus s.l. had differentiated into at least two of the phylogenetic groups as proposed by Johnson (1972, fig. 1). The fossil eucalypts show some features of advanced states in fruit and venation (L. A. S. Johnson, pers. comm.). This suggests that a suitable environment for their development must have existed very much earlier in the Tertiary. The paucity of eucalypt remains in the fossil record probably reflects the lack of opportunity for scleromorphic plants to become fossilized. The lack of similarity between the Miocene pollen assemblages from the Gipp- sland Basin and Chalk Mountain is probably due to geographic and climatic factors rather than a great difference in age of deposition. K/Ar dating of the basalts above and below the Chalk Mountain Formation is desirable to determine the exact age of the locality. Acknowledgements The authors wish to thank Mr V. Mills and Mr and Mrs G. Hawker of Bugaldie for assistance and information on the Chalk Mountain area; Dr J. W. Pickett and Dr Proc. Linn. Soc. N.S.W. 106 (4), (1982) 1983 W. B.K. HOLMES, F.M.HOLMES AND H.A.MARTIN 309 A. Ritchie for the loan of specimens from the Mining and Geological Museum and the Australian Museum respectively; Mr R. S. Jones for preparing the latex casts; Mr P. J. Althofer for making available his private collection; Professor N. C. W. Beadle, Mr J. B. Williams, Dr A. G. Floyd, Dr L. A. S. Johnson and Mr D. F. Blaxell for helpful discussion on the relationship of the fossils to extant species; Professor J. F. G. Wilkinson for information on the Warrumbungle Volcano; Mr D. Nelson and Dr J. A. Webb for useful comments and for information on their unpublished investigations of the Berwick fossil eucalypts. W. B. K. Holmes gratefully acknowledges support from the Science and Industry Endowment Fund. References Ambrose, G. J., Callen, R. A., Flint, R. B., and Lange, R. T., 1979. — Eucalyptus fruits in stratigraphic context in Australia. Nature 280: 387-389. Barlow, B. A., 1981. — The Australian Flora: Its origin and evolution. In Flora of Australia, Vol. 1. Canberra: Australian Govt Printer. Beadle, N. C. W., 1981. — Origins of the Australian angiosperm flora. In Keast, A., Ecological Biogeography of Australia . The Hague: W. Junk. BlRKS, H.J. B., andBlRKS, H. H., 1980. — Quaternary Palaeoecology . London: Edward Arnold . BLAKE, S. T., 1953. — Botanical contributions of the Northern Australia regional survey. 1. Studies on Northern Australian species of Eucalyptus. Aust.J. Bot. 1 (2): 185-352. BRIGGS, B. G., and JOHNSON, L. A. S., 1979. — Evolution in the Myrtaceae — evidence from inflorescence structure. Proc. Linn. Soc. N.S. W. 102 (4): 157-256. CAMBAGE, R. H., 1913. — Development and distribution of the genus Eucalyptus . J . Proc. R. Soc. N.S. W. 47: 18-58. DEANE, H., 1902. — Preliminary report on the fossil flora of Pitfield, Mornington, Sentinel Rock (Otway Coast), Berwick and Wonwron. Rec. Geol. Sun. Vict. 1: 13-32. DUIGAN, S. L., 1951. — A catalogue of the Australian Tertiary flora. Proc. R. Soc. Vict. 63: 41-56. DULHUNTY, J. A., and McDouGALL, I., 1966. — Potassium-argon dating of basalts in the Coonabarabran- Gunnedah district, New South Wales. Aust.J. Set. 28: 393-394. GRIFFIN, R. J., 1961. — The Bugaldi (Chalk Mountain) diatomaceous earth deposit, N.S.W. Tech. Rept Dept Mines N.S.W. 7:1 9-36. HERBERT, C, 1968a. — The mineral industry of New South Wales. 14. Diatomite. Sydney: Dept of Mines, N.S.W. , 1968b. — Diatomite deposits in the Warrumbungles. Rept. Geol. Surv. N.S. W. , 1968/195 unpubl. HlLLS, E. S., 1946. — Fossil Murray Cod (Maccullochella macquariensis) from diatomaceous earths in New South Wales. Rec. Aust. Mus. 21 (7): 380-382. JOHNSON, L. A. S., 1972. — Evolution and classification in Eucalyptus. Proc. Linn. Soc. N.S. W. 97(1): 1 1-29. , and BRIGGS, B. G., 1981. — Three old southern families — Myrtaceae, Proteaceae and Restionaceae. In K.EAST, A., Ecological Biogeography of Australia. The Hague: W.Junk. Kemp, E. M., 1978. — Tertiary climatic evolution and vegetation history in the south-east Indian Ocean region. Palaeogeogr. Palaeoclimatol. Palaeoecol. 24: 169-208. Lange, R. T., 1978. — Carpological evidence for fossil Eucalyptus and other Leptospermeae (sub-family Leptospermoideae of Myrtaceae) from a Tertiary deposit in the South Australian arid zone. Aust. J. Bot. 26: 221-233. MAIDEN, J. H., 1922. — A critical revision of the genus Eucalyptus. Vol. 6. Sydney: Government Printer. MARTIN, H. A., 1973. — Upper Tertiary palynology in southern New South Wales. Geol. Soc. Aust. Spec. Publ. 4: 35-54. , 1978. — Evolution of the Australian flora and vegetation through the Tertiary: evidence from pollen. Alcheringa 2: 181-202. , 1981. — The Tertiary flora. In K.EAST, A., Ecological Biogeography of Australia . The Hague: W.Junk. , 1983. — Changing Cenozoic barriers and the Australian paleobotanical record. Annals Missouri Bot. Gdns 69: 625-667. McDouGALL, I., and WILKINSON, J. F. G., 1967. — Potassium-argon dates on some Cainozoic rocks from north-eastern New South Wales.,/, geol. Soc. Aust. 14: 225-234. NELSON, D., and WEBB, J. A., in prep. — A Tertiary flora from Berwick, Victoria. PRYOR, L. D., and JOHNSON, L. A. S., 1971. — A classification of the Eucalypts . Canberra: A.N.U. Press. Proc Linn. Soc. N.S.W. 106 (4), (1982) 1983 310 MIDDLE MIOCENE FOSSIL EUCALYPTUS RICH, P. V., and McEvEY, A., 1977. — A new owlet nightjar from the Early to Mid-Miocene of Eastern New South Wales. Mem. Nat. Mus. Vict. 38: 247-253. ROSE, S., 1981. — Palynology and history of the Holocene at Dry Lake, Thirlmere, N.S.W. Kensington: Univ. N.S.W. , B.Sc. (Hons) thesis, unpubl. Stover, L. E., and Partridge, A. D., 1973. — Tertiary and late Cretaceous spores and pollen from the Gippsland Basin, southeastern Australia. Proc. R. Soc. Vict. 85: 237-286. SUSSMILCH, C. A., 1937. — The geological history of the Cainozoic Era in New South Wales. Proc. Linn. Soc. N.S. W. 62: viii-xxxiii. THOMAS, D. P., and GOULD, R. E., 1981 a. — Tertiary non-marine diatoms from eastern Australia: descriptions of taxa. Proc. Linn. Soc. N.S. W. 105 (1): 23-52. , 1981 b. — Tertiary non-marine diatoms from eastern Australia: palaeoecological in- terpretation and biostratigraphy. Proc. Linn. Soc. N.S. W. 105(1): 53-63. WELLMAN, P., and McDouGALL, I., 1974. — Potassium-argon ages on the Cainozoic volcanic rocks of New South Wales./, geol. Soc. Aust. 21(3): 247-271. Proc. Linn. Soc. N.S.W. 106 (4), (1982) 1983