4 Palaeobotanical evidence for Tertiary

D. R. GREENWOOD

Climate is the primary framework within which 1985;Upchurch & Wolfe, 1987),the termsmega- plant populations grow, reproduce and, in the thermal, and are used longer time frame, evolve. However, reconstruc- here to describe the temperature characteristics tions of palaeoclimateare often based on the of the vegetationof the main climatic zones.The marine record (e.g. Quilty, 1984 and Chapter 3, definitions of Nix (1982; Kershaw & Nix, 1988) this volume) or on computer modelling (Sloan & are used here," although Wolfe's (1979, 1985) Barron, 1990).The latter type, in particular,suf- vegetation classification uses slightly different fer from their dependenceon simplifiedscenarios definitions. and must be tested against the terrestrial This discussion is restricted to the Tertiary, palaeontologicalrecord. Individual species and since pre-Tertiary floras are dominated by plant whole plant communities are morphologicallyand groups with no or few modern analoguesand so physiologicallyadapted to their physical environ- palaeobotanical indices of are (as yet) ment, most stronglyto climate,and so plant mac- unreliable. Numerous accountsof the Tertiary cli- rofossils are a proxy record of past climates mate history of Australia basedon palaeobotanical (Wolfe, 1971, 1985; Upchurch & Wolfe, 1987). evidence have been published (Kemp, 1978, Severalstudies have indicated that plant macro- 1981;Christophel, 1981, 1988; Nix, 1982;Trus- fossils provide a potentially accurate record of well & Harris, 1982; Christophel & Greenwood, Tertiary terrestrial palaeotemperatures @olfe, 1989), in some casesas part of global accounts 1979,1990; Read et al., 1990). Interpretationof (Axelrod, 1984; Wolfe, 1985). Many of these the climatic signal preservedin plant fossil assem- interpretationshave been basedsolely on nearest- blages is dependent on understanding (1) how living-relative analogy(NLR) and provide qualita- plants (and vegetation)interact with climate, (2) tive reconstructionsof climate. how plant fossil assemblageswere formed, and Many North American studies have used a (3) how these assemblagesrelate to the original strong correlation berween the proportion of vegetation. woody plant specieswith entire leaf margins and The use of terms such as tropical, subtropical mean annual temperature (MAT) in modern and temperate in palaeoclimaticdiscussions is mesic forests (leaf margin analysis,Wolfe, 1979, potentially conflusing,as these terms are rarely 1990), and other aspectsof foliar morphologY,to (lati- defined climatically and have geographical 'C * Modi{ied liom Nix (1982): mcgathermal, > 24 (MAT); "C; "C tudinal) connotations that are inapplicable for meso-mcgathcrmal intcrzone, 20-24 mcsothcrmal, l4-20 much of the Tertiary. Following Wolfe (1979, (14-24 "C if the interzoncis includcd); microthcrmal,< l4'C.

l44l Palaeobotanicalevidence for Tertiarv climates 45 reconstructboth the regional distribution ofphysi- Francis, 1990) and epiphyllous ftttgt found on ognomically defined vegetationtypes and palaeo- leaf cuticle (Lange, 1978, 1981, 1982) are other climates for the Late Cretaceous and Tertiary sources of climatic information from plant fossil (e.g. Wolfe, 1985; Upchurch & Wolfe, 1987; assemblages.These approachesprovide quantitat- Parrish & Spicer, 1988). Wolfe (1971, 1979, ive estimates for critical components of climate 1985) has argued that foliar physiognomic that can be tested against estimates from the approachesare preferred to NLR methodologies, marine record and climatic models. primarily due to inaccurate systematictreatments for many Tertiary floras. However, foliar physiog- THE NATURE OF THE PLANT nomic analysisis used here accordingto the meth- MACROFOSSIL RECORD odolory proposed by Greenwood (l99la; Christophel & Greenwood, 1988, 1989),in com- Leaf accumulationsin streamsor lakes constitute bination with recent quantitative analysesof key a biassed,but neverthelessdetailed, record of the NLRs, of Tertiary taxa (Nix, 1982; Kershaw & parent vegetation. This bias results from a Nix, 1988;Read & Hope, 1990;Read et a1.,1990) number of processesacting on the plant parts (pri- to interpret early Tertiary macrofloras from Aus- marily leaves)from the time of abscissionor trau- tralia (Figure 4.1). Wood anatomy (Frakes & matic loss to their eventual entombment in sediment and lithification (Ferguson, 1985; 135o r5o" Spicer, 1989; Greenwood, l99la): 1. Deciduous and evergreentrees may have dif- ferent representationin an assemblageif leaf- fall is asynchronouswith periods of maximum sedimentation. 2. Different plant speciesmay haveleaves that are dehiscent or nondehiscent(e.9. many palms, herbs and forbs). Nondehiscent leavesare less olffio likely to be fossilisedthan are dehiscentleaves. 3. Plant organs may have varying capacities for transport from the plant body. For example, leaveswith a low weight per unit area tend to travel further in air than do denser leaves. 4. Leaves of individual species(e.9. sun versus shade leaves) and different species decay at different rates. 5. The varying productivity of the plants is sig- nificant, with canopy trees producing copious amounts of leaves,swamping the litter-fall. Fossil leaf accumulationstherefore reflect only a subset of the original surrounding vegetation, 135o 1soo and in most situations represent only the local Figure 4.1 Location map of AustralianTertiary macrofloras plant communities (Greenwood, 199| a). Further- discussedin the text, and modern thermal regimes(adapted more, smaller plant macrofossil accumulations from Nix, 1982;Christophel & Greenwood,1989): l, Stuart may represent geologicallyinstantaneous records Creek(silcrete floras); 2, Nelly Creek;3, GoldenGrove; 4, of past vegetation. Individual sedimentary facies Maslin Bay; 5, Nerriga;6, BacchusMarsh; 7, Anglesea;8, LatrobeValley (Yallourn & Monvell);9, Pioneer;10, Cethana; often reflect different components of the local ll, Monpeelyata. vegetational mosaic (see e.g. Christophel et al., 46 D. R. GREENWOOD

Table 4.1. Foliar physiognoniccharacteistia of leaf-litterfrom modem Australion rainforest A. Mean "/" leaaes

Leaf size "h Forest type" Micro- Noto- Mesophylls entire leaf margins CMVF 6-25 50-70 I l-39 90.0-98.3 CNVF 16-55 35-77 2-22 34.3-57.0 SI\IVF 59-85 14-36 0-ll 87.O-es.3(Q!d) 7.0-37.0(I,{s!v) MV-FF' 74-80 19-24 l-2 56.7-96.4 MFF' 90-96 4-10 0 1.0-13.0 NMF. 96-99 r-4 0 not available

B. "/" taxa

Leaf size Foresttype" Micro- Noto- Mesophylls LSI 7o entire leaf margins CMVF 1l-18 58-78 1l-30 47-s9 82.4-95.0 Cl\ryF 30-46 14-63 0-36 33-43 50.0-81.8 SNVF 40-75 25-50 0-20 13-38 84.2-e0.0(Qld) r43-2s.0 (NSIV) MFF r00 0 33.3-40.0

"CMVF, complexmesophyll vine forest;CNVF, complexnotophyll vine forest; SNVF, simplenotophyll vine forest;MV-FF, microphyllvine-fern forest;MFF' microphyllfern forest;LSI,leaf sizeindex (from Webb, 1959; Tracey, l98Z). ,Basedon a singlesite (four samples). "Basedon Macphailet al.,l99l, Table 3 (LakeDobson and Precipitous Bluff). Datafrom Greenwood,l99la (Table l, in part) and1992.

1987). However, stratigraphically correlated The present analysis uses a dataset derived macroflorasmay be separatedin time (as a record from modern Australian rainforest leaf-litter and of the original vegetation, and hence climate) by fluvially deposited beds (Christophel & Green- tens of thousands of vears. wood, 1987, 1988, 1989; Greenwood, 1992; Table 4.1). A multivariate foliar physiognomic analysisfor the Australian Middle Eocene macro- evidence of climate Palaeobotanical floras is in preparation (D.R. Greenwood,unpub- Terrestrial plant fossil assemblagesrecord local lished data), however, here temperature climate usually over small time intervals. Each characteristics are derived primarily from uni- plant assemblagerepresents a single datum of cli- variate comparisons(e.g. Figure 4.2). Other Aus- mate, with separate assemblagesfrom a single tralian Tertiary macrofloras are discussed flora providing information about local variation. qualitatively using leaf size and margin data only, However, continuous deposition of terrigenous as full data are not available.These analysesare sedimentswithin a single basin (e.9. the Latrobe supplemented by quantitative and qualitative esti- Valley)preserve climate records spanningmillions mates based on NLR analysesfrom Read et al- of years. (1990; Read, 1990) and Nix (19821,Kershaw & Palaeobotanical evidence for Tertiarv climates +7

25 U o

qJ tr E20 0)

0)

(g EI)

C' 0,,

10 20 30 40 50 60

Leaf size index

Figure 4.2 Relationship (least-squares regression) between leaf size index of forest-floor leaf-liner from Australian rainforests and mean annual temperature (MAT), r : 0.94. (Adapted from Greenwood, 1992.)

Nix, 1988).Forest nomenclature and typologyin (LSI : (% microphylls+ 2/" notophylls+ 3% this discussionfollows those of Webb (1959, mesophylls 100)/2; Wolfe, 1978; Burnham, 1968), Tracey (1982; Table 4.1) and Wolfe 1989), and MAT (Greenwood,l99la, 1992; (re7e). Figure4.2).Mean leaf length(specimens) is also strongly correlated with MAT (Greenwood, r99Z\. Foliar physiognomy and climate The LSI of the Tertiary macrofloras (where Greenwood(1992; Greenwood& Christophel, applicable)has been used to calculatethe prevail- 1990, unpublished data) suggests that the ing MAT using a linear regressionequation from relationshipbenreen leaf margin type and MAT modern litter (Figure 4.2). These MAT values for Australianforests is different from, and more must be consideredto be minimum estimates;the "C complexthan, that derived by Wolfe (1979) for actual values are likely to be l-3 higher, as eastAsian forests.Leaf sizeis stronglycorrelated the taphonomic size bias is alwaystowards smaller with MAT in Australian rainforest canopies leaves than are found in the parent vegetation. (Webb, 1968), and, since the leaf size bias The spreadof MAT valuesin Figure 4.2 indicates betweenthe canopyand forest floor litter is con- the uncertainty of this magnitude. New data from sistent (Greenwood, l99la), modern leaf-litter the Australian rainforest litter are also used here from humid vegetationshows a strongrelationship to calculateMAT using leaf margin analysis.The betweenleaf size,calculated as the leaf sizeindex use of separate foliar physiognomic indices to D. R. GREF,NWOOD calculate MAT provides a check on the fidelity A common element in Nix's (1982), Axelrod's of the estimates. (1984) and Wolfe's (1985) interpretationsis for uniformly low seasonality(equable) climates in the Eocene. More recently, Sloan & Barron (1990) AUSTRALIAN TERTIARY have suggested that the Eocene mean annual MACROFLORAS AND range of temperature (MART) in continental PALAEOCLIMATES interiors should have been high. The new syn- This discussion is restricted to a small group of thesis presented here differs from Axelrod's macrofloras and is not a general review of the (1984)and Wolfe's (1985)interpretations in some Australian Tertiary climate. The macrofossil sites important details, and provides a quantitative (Figure 4.1) representa few key points in time: comparison with the estimations of Nix (1982) Middle-Late Eocene (Anglesea, Maslin B"y, and of Sloan & Barron (1990). southern Lake Eyre Basin silcrete and Nelly The Middle Eocene Maslin Bay (North Maslin Creek, Nerriga), Oligocene (Cethana), Late Sands; McGowran et al., 1970), Anglesea(East- Oligocene-Early Miocene (Pioneer, Monpeely- ern View Formation; Abeles et al., 1976) and ata) and the Early-Middle Miocene (Bacchus Golden Grove (Nonh Maslin Sands:Alley, 1987) Marsh, Latrobe Valley),and can be consideredto macrofloras(Figure 4.1) occur in mudstonelenses be individual 'snapshots',where climate can be within fluvial sands (Christophel & Blackburn, inferred for local areas.Late Tertiary macrofloras 1978; Christophel & Greenwood, 1987, 1989; are scarce in Australia, and, apart from Bacchus Christophel et al., 1987). Each clay lens is inter- Marsh and the Latrobe Vallev. are not considered preted as an infilled abandoned channel (oxbow) here. in a meandering stream sequence. A diverse Middle Eocene macroflora also occurs in well- laminated lake sediments (Titringo Siltstone) at Middle Eocene Nerriga (Hill, 1982). These macroflorasprovide Axelrod (1984) concluded that thermal regimes a record ofvegetation at near sealevel (Maslin Bay in southern Australia in the Late Paleocene- and Anglesea)and at moderateelevations (Golden Eocene were comparable to the present regime. Grove and Nerriga) for the continental margin Nix (1982)depicted a latitudinalzonation of ther- from which climate can be inferred (see Table mal regimes across Australia in the Middle 4.4, below). Eocene macrofloras from the Lake Eocene similar to that of the present, although Eyre Basin provide information on vegetationand with MAT values in southern Australia'2"4 "C climate in the continental interior (Lange, 1982; warmer than at present'.Wolfe (1985), however, Greenwoodet a1.,1990;Christophel et al.,1992). suggesteda more southerly latitudinal extension The continental margin Middle Eocenemacro- of tropical rainforest (megathermal), with para- floras have many taxa in common at the generic tropical rainforest (mesothermal-megathermal level (Christophel & Greenwood, 1989) and are interzonesensu Nix (1982)), in the southern half of typified by abundant leaf remains of Lauraceae. Australia, and notophyllous forest (mesothermal) A distinctive feature of the Nerriga macroflora on the adjacent part of Antarctica, in contrast to compared to other Australian Eocene macrofloras Nix's (1982) suggestionof microthermal climates is the common presence of hastatevine leavesof in southernmostAustralia and Antarctica. Early- Menispermaceae(Hill, 1989). This may reflect Late Eocene macrofloras from the Antarctic disturbance in the local vegetation as vines are Peninsula suggest that either cool temperate commonly enriched in modern tropical rainforest I{othofagu.sand podocarp rainforest (nanophyll canopiesafter canopydisruption (Tracey, 1982), mossy forest, NMF) or deciduous forests were or it may reflect vine-rich lake-edge vegetation. dominant over much of Antarctica at that time Many of the identified common plants from these (Tokarski et al., 1987; Case, 1988). Eocene macrofloras have NLRs with Palaeobotanicalevidence for Tertiary climates 49

Table 4.2. Thefoliar physiognomiccharacteistics of the macrofloras(specimen-based obsertsations)

7o entire Leaf size ("/") No. of leaf Mean length Site' leaves' margins Micro- Noto- Mesophylls (mm) No. of taxa. Middle Eocene Anglesea (l) 8l 73.0 32.0 46.0 22.0 4l (2\ 1548 79.7 82.9 16.l 0.9 49 ca35 MaslinBay (l) 1046 84.0 25.0 54.0 2t.0 57 (2) 90 76.7 43.2 4l.l t5.7 88 (3) 2t0 70.0 ?.4.0 53.0 23.0 ?r00 80 GoldenGrove (6) 338 47.4 52.6 40.9 6.5 77 ca30 Nerriga (t) t29 66.0 24.0 55.0 2r.0 38 (4) ttz 44.1 22.7 59.2 l8.l ?102 z5 576 39.6 (including fragmented leaves) Deane'sMarsh 100.0 0.0 0.0 Nelly Creek(5) 160 88.0 80.0 20.0 0.0 ; Oligocene Cethana 267 13.0 95.0 5.0 0.0 34 Pioneer 308 9.7 100.0 0.0 0.0 25 5 Early Miocene BacchusMarsh 100.0 0.0 0.0 6 Monpeelyata 630 57.5 100.0 0.0 0.0 7 >t

oCethana,Pioneer and Monpeelyatadata; R. S. Hill & R.J. Carpenter,personal communication. (l), Christophel,1981; Christophel & Blackburn,1978. (2),Christophel & Greenwood,1987, 1989. (3),L.J. Scriven,personal communication. (4),Hill, 1982,and personal communication. (5),Christophel el al., t992. (6),D. Barrett,personal communication; Greenwood, l99la. All'otherdata: Christophel & Greenwood,1987, 1988, 1989 and unpublished results. 'Number of wholeleaves used to calculatethe values. "Numberof leaftaxa encountered in sampleused; whole macroflora may be higher. mesothermal-megathermal modern ranges here. The Angleseaanalysis was based on a small (Kershaw & Nix, 1988; Christophel & Green- sample (81 leaves)and Christophel's interpret- wood, 19g8, 1989; Hill, 1990a). ations of all of the macrofloraswere based on Christophel (1981; Christophel & Blackburn, direct comparisons with canopy values (from 1978)originally considered the Anglesea,Nerriga Webb, 1959). Burnham (1989; Burnham et al., and Maslin Bay macroflorasanalogous to modern 1992) and Greenwood (l99la, I99Z) suggested complex notophyll vine foresr (CNVF) or simple that sample sizesin excessof 200-350 leavesare mesophyllvine forest (SMVF). The initial analysis required to sample leaf assemblagesadequately. of Maslin Bay was based on a large sample (1046 Subsequent analysisof the Anglesea and Golden leaves); however, only the tallies presented by Grove macrofloras used much larger samples Christophel & Blackburn (1978) were available (267 -1548 leaves;Tables +.2, 4.3) and a more 50 D. R. GRF,ENWOOD

Table 4.3. Thefoliar pfuisiognomiccharacteistics of the macrofl.oras(ta"xon-based obsentations)

7o entire Leaf size (/"') No. of No. of leaf Site" leaves taxa margins Micro- Noto- Mesophylls LSI' Middle Eocene Anglesea (l) 793 30 70.0 60.0 40.0 0 20.0 (2) 8l 4l 78.0 23.0 50.0 27.0 49.0 Maslin Bay (2) 1046 57 ,n: 25.0 47.0 28.0 51.5 GoldenGrove (5) 338 24 53.3 46.7 0 25.0 Nerriga (l) t29 38 66.0 25.0 45.0 30.0 52.5 (4) ll6 25 80.0 56.5 39.1 4.3 23.8 Nelly Creek(3) 160 l6 75.0 60.0 40.0 0 20.0 Oligocene Cethana 267 ?100 0 Late Oligocene-Early Miocene Pioneer 308 5 60.0 10000 0 Monpeelyata 630 >7 >60.0 10000 0 Early Miocene Bacchustr{arsh 66.7 90.0 10.0 10.0

"Cethana,Pioneer and Monpeelyatadata; R. S. Hill & R.J. Carpenter,personal communication. (l), Christophel& Greenwood,1987, 1988, 1989. (2),Christophel, 1981. (3),Christophel et al., t992. (4),Hill, 1982,and personalcommunication. (5),D. Barrett,personal communication; Greenwood, l99la. All otherdata: Christophel & Greenwood,1987, 1988, 1989 and unpublished data. 'LSI. leafsize index. detailed comparison with modern leaf-litter mesophyll vine forest (CMVF) from Queensland (Christophel & Greenwood, 1987, 1988, 1989). are typically wider and there are fewer leaf speci- In theselater analyses,the AngleseaSite II macro- mens and leaf specieswith nonentire margins pre- flora (Christophel et al., 1987) matched simple sent than the leaves from the fossil macrofloras notophyll vine forest (SNVF) litter, and the Ner- (Christophel & Greenwood, 1988; Tables riga, Golden Grove and Maslin Bay macrofloras +.1-+.3). Leavesfrom New South Wales SNVF compared most closely to CNVF litter; however, and CNVF are as narrow as the Anglesea and in this instancethe Maslin Bay analysiswas based Golden Grove leaves,but the proportion of leaves on a sampleof only 90 leaves(Tables 4.3, +.4). and species with nonentire margins is much The Middle Eocene continental margin macro- higher in the New South Wales SI{VF than in floras are dominated by simple, narrowly elliptic the Angleseamacroflora. A significant number of (length: width > 2.0), microphyllous to notophyll- hastate-baseand broad leaves(mostly Menisper- ous leaves (Anglesea) or notophylls (Christophel maceae) grves the Nerriga macroflora a unique & Greenwood, 1988), with entire margins more foliar physiognomic signature compared to the common than nonentire (Tables 4.2, 4.3). The other Eocene macrofloras; however, narrowly leavesin litter from SNVF, CNVF and complex elliptic leavesalso dominate this flora. The tend- Palaeobotanicalevidence for Tertiarv climates 5r

Table 4.4. Summaies of the sedimentologicaland geographicsetting of eachmacroflora and t he inferc d p alaeo era i ronm en t

Sedimentary Modern 7o entire MAT Vegetation Site environment elevn (m) leaf margins LSI fC) type" Middle Eocene Anglesea Oxbowpond 50 70 20.0 15-18 SNVF MaslinBay Oxbowpond 65 79 51.5 23-26 CMVF GoldenGrove Oxbowpond 120 25.0 17-20 CNVF Nerriga Lake 535-570 66 23.8 16-21 CNVF Nelly Creek River channel 8 75 20.0 >20 ScF/MVF Oligocene Cethana Lake 300 ?O IO-I2 MVFF Late Oligocene-Early Miocene Pioneer Alluvial fan Sea-level 60 O IO_I2 MFF/SNVF Monpeelyata Lake 920 >60 O 4-8 NMF/T Early Miocene BacchusMarsh Oxbowpond 400 67 l0 t2-t4 MFF

elevn, elevation;LSI, leaf size index; MAT, mean annual temperature. "CMVF, complex mesophyll vine forest (mesothermal-megathermalrainforest) MVF, (?semideciduous)mesophyll vine forest (monsoon forest) ScF, sclerophyllousforest (?markeddry season) CI{VF, complex notophyll vine forest (mesothermal rainforest) SI{\trF, simple notophyll vine forest (mesothermal rainforest) MVFF, microphyll vine- fern fore st (microthermal rainfore st) MFF, microphyll fern forest (microthermal Nothofagusrainforest) NMF/T, nanophyll mossy forest (subalpinerainforest or wet thicket). ency towards narrower leaves in the Eocene on Anglesea leaf cuticles are of type V, and type macrofloras versus analogous modern Queens- 6 manginuloid hyphae are also present, which land mesic vegetation(SNVF, CNVF and CMVF) indicate a humid climate of the type restricted to may reflect latitudinally controlled climatic differ- modern montanetropical rainforest (Lange, 1981, ences,as indicated by the narrower leavesfound 1982). The floristic character of the Anglesea in New South Wales CNVF and SNVF; however, macroflora is consistent with this interpretation the exact relationship is unclear. (Christophel & Greenwood, 1989; Christophel, The leaf size index for Anglesea (Table 4.3; Chapter ll, this volume). Leaf cuticles from Figure 4.2) suggestsa MAT of 15-18'C. Using Golden Grove and Maslin Bay have high grade leaf margin analysis(Wolfe, 1979,1985; data from fungal epiphyllous germlings (4. I. Rowen, p€r- Greenwood, 1992) the MAT is estimated to sonal communication) indicative of warm humid "C. be 17 The LSI values for Maslin Bay and climates (Lange, 1978, l98Z). Golden Grove suggestMAT values of 23-26"C The leaf size index (Table 4.3) of the Nerriga and 17-20 oC respectively (Figure 4.2). Mod- macroflorausing Hill's (1982, and personalcom- "C ern SNVF has a range of MAT of l2-18'C munication) data suggestsa MAT of 16-21 oC (mesothermal),CNVF MAT of 17-22 (meso- (mesothermal).If Christophel's (1981) data are "C thermal) and CMVF MAT of 20-27 used, the LSI is much higher, gtuittg a MAT of (mesothermal-megathermal interzone to mega- 25-28'C. Christophel's (1981) original figures thermal). Fungal epiphyllous germlings recorded were preliminary and probably overestimate 52 D. R. GREENWOOD species number (Hill, 1982; 38 versus ca 20 estimates relative to Maslin Bay (Table 4.4) can species) and LSI is sensitive to both species be explained. number and the proportion of mesophylls(Wolfe The MAT values and elevation of the Maslin & Upchurch, 1987; Greenwood, 1992). Leaf Bay and Golden Grove macrofloras suggeststhat margin analysisof the Nerriga macroflora based either nvo distinct and climatically controlled veg- "C. on Hill's data indicatesa MAT of 2l The etation types occurred within only a short dis- Nerriga macroflora sediments outcrop today at tance, or that taphonomic bias is causing a moderate elevation(Hill, 1982; Table 4.4). If this significant reduction in mean leaf size, resulting modern elevationis indicative of the Eocene elev- in an underestimate of MAT for Golden Grove ation the Nerriga macroflora can be expected to (an overestimatefor Maslin Bay is highly unlikely; indicate a lower MAT than do lowland Middle Christophel & Greenwood,1988). Christophel & Eocenesites at similar palaeolatitude(e.9. Maslin Greenwood (1987) have suggestedthat there may 'C) Bry). The MAT estimatefor Nerriga (16-21 be a significant collection bias towards smaller compared to Golden Grove and Maslin Bay curated leavesfor the Golden Grove macroflora. (Table 4.4) is slightly higher than would be The Maslin Bay macroflora is based solely on expectedon the basis of their present elevations. compressionsand impressionsin situ on the mud- Lake sediments may be enriched with the stone matrix, whereas all of the other macrofloras smaller sun-leaves of the upper canopy (Roth & are based on mummified leaves that have been Dilcher, 1978; Spicer, 1981)and so the MAT for isolated from the matrix. It is possible that this Nerriga basedon LSI may be an underestimate. preservational difference may enhance leaf size However, the significant number of entire- taphonomic biases. Casual inspection of margined vine leavesmay also bias the leaf margin compression-impression facies of the Golden analysis,producing a MAT estimate that is too Grove macroflora (seeChristophel & Greenwood, high. Greenwood (1992; Greenwood & 1987) was equivocal on this point. These factors Christophel, 1990) has demonstratedthat tapho- act to underscorethe degreeto which taphonomic nomic and ecological factors introduce a signifi- biases influence resolution of quantitative cant uncertainty in leaf margin analysisestimates variables. of MAT. The actual MAT value for Nerriga The Lake Eyre Basin contains extensivefluvial probably lies benveen the estimatesderived from and fluvio-lacustrine sequencesof Tertiary sedi- the LSI (Hill, 1982) and from leaf margins. In ments (Wopfner et al., 1974; Ambrose et al., either case, a mesothermal humid climate is 1979). An outcrop of the Paleocene-Eocene indicated. Eyre Formation (Sluiter, 1991) within the There are severalexplanations for the relatively southern margin of present Lake Eyre (e.9. Nelly high estimated MAT for Nerriga, and the appar- Creek site, Figure 4.1) contains leaf and fruit ent discrepancybetween Maslin Bay and Golden macrofloras(Wopfner et al.,1974). These macro- Grove values. The Nerriga macroflora is older floras are usually silicified and abundant capsular than the other Middle Eocene macrofloras,so cli- myrtaceous fruits, including putative Eucalyptus matic change over this interval may be the cause. (Lange, 1978, 1982), have been recorded from However, there is debate over the timing and correlated sedimentsin the sameregion (Ambrose degree of uplift of the Great Dividing Range et al., 1979; Greenwood et al., 1990). Middle (Ollier, 1986), including the upland area of the Eocene sedimentsof the Eyre Formation at Nelly Nerriga macroflora.It is possiblethat the Nerriga Creek contain mummified leaves and fruits macroflorawas depositedat a lower elevationthan (Christophel et al., 1992). at present, as significant uplift in the area may A significant number of taxa in the Lake Eyre have occurred during and since the Oligocene silcreteshave leaveswith broad lamina, large size (Ollier, 1986).If this is so, the apparentsimilarity and entire margins, suggestingthat they grew in between the Nerriga and Golden Grove MAT mesic environments;other taxa havevery narrow, Palaeobotanicalevidence for Tertiarv climates 53

small leaves,with coarse serrations (to spinose), macrofloras at continental margin sites or narrow-lobed leaves, which suggest dry or (Christophel et al., 1992). Nelly Creek leavesare seasonally dry climates (Givnish, 1984; Wolfe, significantlysmaller overall (Tables 4.2, 4.3),with 1985). The silcrcte floras suggest a mosaic of a different assortment of leaf types. Important sclerophyllous vegetation and broad-leaved gal- floristic differencesare discussedby Christophel lery forests(Greenwood et al.,1990; Greenwood, (Chapter I l, this volume). Lobed leaves and l99lb; Christophel et al., l99Z). Beadle (1967) specieswith toothed leaf margins are common in suggestedthat adaption to low soil fertility during both disturbed environments (early succession), the Tertiary may have produced sclerophyllous and in strongly seasonalenvironments (Givnish, plants, pre-adapting many lineages of Australian 1984; Upchurch & Wolfe, 1987). The foliar plants to seasonallydry and arid climates. It is physiognomy of the Nelly Creek and silcrete plausible therefore, that the sclerophyllous macrofloras compared to the other Eocene elements in the Lake Eyre Basin floras do not macrofloras suggeststhe presencein central Aus- represent dry climate vegetation,but low fertility tralia of seasonally dry (?monsoonal)palaeocli- soils. mates where mesic broad-leaved vegetation The foliar physiognomy of the Nelly Creek occurred in riparian corridors (Greenwood et al., macroflora partly matches the silcrete macro- 1990; Greenwood, l99lb; Christophel et al., floras; however, the presence of sunken stomata 1992). This interpretation is supported by sedi- and thick cuticles (Christophel et al.,1992) on the mentological evidence that suggests periods of leaves of this mummified macroflora emphasise aridity in central Australia in the early Tertiary the sclerophyllous and,/or xerophyllous character. (Q.rilty, 1984). The floristic composition of both the silcrete floras and the Nelly Creek macroflora alsosupport Oligocene this interpretation, with some taxa commonly associatedwith seasonallydry environments (e.g. The best-documented Early Oligocene macro- Brachychiton)and others associatedwith mesic- flora occurs at Cethana in Tasmania (Hill & Car- humid environments (Christophel, Chapter li, penter, l99l; Figure 4.1) in ?slumped lake this volume). Only scarce,low grade (types I and sedimentsat a present altitude of 300 m (R. J. II), epiphyllous fungal germlings occur on Nelly Carpenter, personal communication). The Late Creek leaf cuticles(Christophel et al.,l99Z), indi- Oligocene-Early Miocene Pioneer macroflora cating that rainfall was low (Lange, 1978, 1982) occurs in lenses of organic-rich clays in coarse or markedly seasonal.The common presenceof braided fan (alluvial) depositsclose to sea level in both Gymnostoma and Brachychiton in some north-easternTasmania (Hill & Macphail, 1983). exposuresof Eyre Formation (Greenwood et al., This macroflora representsthe lowland vegetation 1990; Greenwood, l99lb) indicate low seasonal of the area at that time (Table 4.4). The Late and diurnal temperature variation in central Aus- Oligocene-Early Miocene Monpeelyata macro- tralia in the Eocene as these genera today are flora occurs at 920 m above sea level and it is restricted to frost-free mesotherm-megatherm probable that the original palaeovegetation areas (data from Webb & Tracey, 1982). On the occurred at this or a higher elevation (Hill & basis of MAT estimates for more southerly Gibson, 1986;Macphail et al., l99l). Middle Eocene macrofloras (e.g. Maslin Bry, The Cethana macroflora is species rich and Golden Grove) and the common presence of contains a diverseassociation of microthermal lin- mesotherm-megatherm taxa (e.g. Gymnostoma), eages including three of the four modern the Lake Eyre Basin macrofloras probably reflect subgeneraof Nothofogu.r(Hill, 1984, 1990a,I 99 I "C. ). a MAT of > 20 Many of the taxa found in the Cethanaassemblage The foliar physiognomy of the Nelly Creek are today restricted to the montane tropical rain- macroflora is distinct from the Middle Eocene forests of New Guinea and New Caledonia (HilI, 54 D. R. GREENWOOD

1990a). Physiognomically, these modern forests and the predominance of microphylls in both "C. are analogousto Australian (microthermal) micro- macrofloras,suggest a MAT of 10-12 On phyll vine-fern forest (MV-FF) of upland areas the basis of analogousmodern forests (MFF and (>1000 m) in the northeast Queensland Humid MV-FF) and particularly key taxa such as Notho- Tropical Region, and New Guinea montane fr7 (Read, 1990), the MART of the Oligocene forests. palaeovegetation (particularly at Cethana) was The Pioneer macroflora is dominated by leaves probably low, of Nothofagus,with abundant conifer material and matically sensitive taxa (e.9. Nothofagussubgenus several species of angiosperm. The single site Brassospora)in the Cethana macroflora suggests used for this analysis is less diverse than the that freezing temperatures were unlikely (Hill, Cethana macroflora. 1990a; Read et al., 1990). Initial analysis of the foliar physiognomy of the The Monpeelyata macroflora is characterised Cethana and Pioneer macrofloras suggested a by entire-margined leaves(Table 4.2) and,imbri- match with Australian microphyll fern forest cate foliage conifers, including Araucaria (Hill & (MFF) (Hill & Macphail, 1983; Christophel & Gibson, 1986; Hill, 1990&).Significant numbers Greenwood, 1989). MFF in south easternAus- of small-sized toothed Nothofagusleaves (N. micro- tralia and Tasmania is speciespoor relative to the phylla) related to N. cunningharnii are present tropical microthermal rainforests (MV-FF) and (Hill, 1991), in addition to leavesof the modern appears physiognomically distinct (Webb, 1959, deciduous species,N. gunnii. Overall leaf size is 1968; Tracey, l98Z). Hill & Macphail (1983) very small, with nanophylls dominant (< 25 mm recorded grade V germlings (Lange, 1976) from length). This macroflora may represent alpine Pioneer and thus suggested that rainfall was in shrub vegetation or low alpine woodland (Hill & excess of 1500 mm annually. The relative fre- Gibson, 1986). Macphail et al. (1991) have sug- quency of the three main leaf size categories gested that the macroflora represents local cool (Tables +.1-4.3) in both macrofloras matches temperate rainforest (nanophyll mossy forest, MFF. These Tasmanian macrofloras have a NMF) or rainforest scrub (nanophyll mossy higher frequency of species with entire leaf thicket, NMT), with a lakeside coniferous margins than do modern Australian MFF, and microphyllous-angiospenn shrubby vegetation. SNVF from New South Wales, and reflect more Significantly also,Macphail et al. (1991) suggested the situation seen in modern MV-FF. Important that the Monpeelyata flora was deposited below floristic differences contribute to these discrep- the alpine treeline. On the basis of the small size ancieswhich in part reflect extinction events(Hill of the N. micropfuillaleavesand the modern toler- & Read, 1987; Carpenter et al., Chapter 12, this ancesof N. gunnii (Read, 1990), a MAT of 4-8 "C 'C. volume), and in part may reflect climatic differ- is indicated, and a MART of ca 15 The ences (such as a change in MART). presence of subalpine vegetation at Monpeelyata Leaf sizes are predominantly microphyllous in reflects the elevation at the time (Hill & Gibson, both macrofloras,although leptophyllous conifers 1986; Macphail et al., 1991). are also prominent. The LSIs of the Cethana and Pioneer macrofloras are not meaningful as both Miocene are zero. On the basisof modern leaf-litter studies (Greenwood, L992),mesic vegetationabove lZ'C An Early Miocene macroflora occurs at Bacchus MAT will have a LSI value ) 0, and a LSI of Marsh in Victoria (Figure 4.1) in a single clay lens zero below 10-12 'C MAT (Figure 4.2). Modern in fluvial sands (Werribee Formation; Abele et al., MFF in Australia ranges from a MAT of 4-15 1976).It is likely this site was at an elevationsimi- oC, with Nothofagus moorei occurring over the lar to the present situation (a00 m) in the Miocene range8.9-16.1 "C (Read, 1990).The presence (Christophel, 1985). The macroflora is poorly of leaves similar to those of modern N. moorei, known but is dominated by leaves from only a few Palaeobotanicalevidence for Tertiary climates 55 species(Christophel, 1985).Leaf sizesrange from although perhaps less seasonal.SNVF dominated microphyllous to nanophyllous, with toothed by Syzygium and MFF' with l/. cunninghamii are leaves (Nothof&gus) dominating, although the found in the nearby Strzelecki Ranges.The rela- second most common leaf type (Myrtaceae) has tively higher diversity of the Yallourn Clayscuticle an entire margin. Some nanophyllousconifers are and macroflora compared to Victorian MFF and present, including Araucaria and Dacrycarpus SNVF, and the presenceof.A/. mooreitype leaves, (Christophel, 1985, and personal communi- however, suggests a better parallel with New cation). Today, Araucaria may be found in associ- South Wales MFF and SNVF, and indicates a "C ation with N. mooreiin the SNVF-MFF ecotone MAT of 9-16 (seee.g. Read, 1990). at Mt Lamington in southeastQueensland under (Webb & Tracey, 1982), microthermal regimes CONCLUSIONS and in Chile and Argentina with a winter decidu- ous speciesof Nothofogu.lunder winter cli- From the review and new analysespresented here, mates (Veblen & Schlegel, 1982). Comparison of it is apparent that climates were not uniform the macroflorawith modern leaf assemblagesindi- across Australia during the Middle Eocene, and cates a match with N. moorei - MFF (Tables that local topography in the Oligocene and the 4.2-4.3). This vegetation type (although with N. Early Miocene influenced the climate and plant cunninghamii, not N. moorei)is found today near communities in southeastern Australia (Table to the Bacchus Marsh site, implying a similar cli- 4.4). Nix's (1982) conclusion that increasing mate in the Miocene. The BacchusMarsh macro- seasonality in both temperature extremes and 'C. flora LSI value indicatesa MAT of 12-14 rainfall and, less so, marked changesin the ther- In the Latrobe Valley (Figure 4.1), a more-or- mal regime, was the primary driving force of biotic less continuous sequenceof coal formation from change within Australia during the early to mid- the Eocene to the Pliocene has preserveda record Tertiary is supported by the palaeobotanicalevi- of swamp and terrestrial vegetation (Blackburn & dence of climate and vegetation patterns (e.g. Sluiter, Chapter 14, this volume). In contrast to Christophel & Greenwood, 1989; Read et al., the Latrobe Valley coal floras, most speciesfrom 1ee0). the Middle Miocene Yallourn Clay macroflora The Middle Eocene macroflorasat Maslin Bay, show no sclerophyllousfeatures; cuticles are thin Golden Grove, Angleseaand Nerriga (Figures 4.1 with few or no trichomes and possesssurficial and 4.3) reveal predominantly mesothermal- stomata (D. R. Greenwood, unpublished data). megathermal humid climates ranging from MAT "C However, there are floristic elements such as values of about 16 (Anglesea,mesothermal) to Dacrycarpuslatrobensr (Hill & Carpenter, 1991) perhapsas high as 25'C (Maslin Bay, megather- and Agathri common to the Late Miocene coals mal) in the coastallowlands and at moderate elev- and the clays. The Clay flora typically contains ations. Climates in the Middle Eocene were mainly microphyllous entire-margined leaves. In markedly wanner in southern Australia than at some samplesmicrophyllous toothed leavesof N. present. Rainforests analogousto SNVF, CNVF rnooreitype (aff. N. tasmanica)were common. The and CMVF grew under these climates cuticle flora is diverse (20-30 cuticle morpho- (Christophel & Greenwood, 1989). There is evi- types)and is dominated by severalspecies in each dence that a transitional zone from mesothermal of the Lauraceae, Myrtaceae (aff. Syzygium), to microthermal climates to the south, and meso- Cunoniaceaeand Proteaceae. thermal to megathermal climates to the north, The predominantly microphyllous nature of the existed in an irregular band across southeastern macroflora of the Yallourn Clays, together with Australia (e.g. Nix, 1982; Christophel & Green- the occasionalpresence of .A/.moorei type leaves wood, 1989); however, this interpretation and the suggeststhat the Middle Miocene palaeoclimate latitudinal position of such a zone remains specu- was similar to the present climate of the area, lative until more macrofloras are described in 56 D. R. GREENWOOI)

1990a). Microthermal lineages(sensu Nix, 1982), particularly Nothofogu.r,were of increasingimpor- tance, compared to the situation in the Middle Eocene(Hill, 1990a;Read et a1.,1990). The foliar physiognomic and floristic character of these macrofloras suggests humid microthermal cli- mates, although modern Australasian tropical montane communities are better analogiesfor the Cethana Oligocene vegetation than are modern southern Australian humid microthermal forests such as the Nothofagus-dominatedMFF (Hill, 1990a;Read et a1.,1990).In contrast,the Pioneer and Miocene BacchusMarsh macroflorassuggest microthermal to modern Nothofogu.r-dominated MFF, particularly the l/. moorei-dominated forests of montane New South Wales. The absence from the Early Miocene macrofloras of some k.y Oligocene taxa suggests increasing seasonality with perhaps greater temperature extremes during the Early Miocene than in the r50" Oligocene(Hill & Read, 1987;Hill, 1990a).Sub- Figure 4.3 Reconstruction of Australia and region for the alpine woodlands occurred at this time at moder- Middle Eocene showing palaeocoastline superimposed on ate altitudes in Tasmania (Hill & Gibson, 1986; modern coastline. Based in part on Christophel & Greenwood Macphail et al., 1991). (1989).NMT, nanophyllmossy thicket or subalpinewoodland The Middle Miocene Yallourn Clay macroflora (microthermal); BDF, broad-leaved deciduous forest (micro- thermal); MFF, microphyll fern forest (microthermal); SNVF/ indicates mainly microthermal climates in south- CNVF, simple/complex notophyll vine forest (mesothermal); eastern Australia. Mesothermal humid vegetation CMVF, complex mesophyll vine forest (mesothermal- (SNVF) probably persistedin parts of the Latrobe megathermal); ScF, sclerophyllous forest (?megathermal); Valley during the Middle Miocene, as it does MVF, ?semideciduousmesophyll vine forest (monsoonal today. There is anecdotalevidence to suggestthat megathermal). Middle Miocene climates in the Latrobe Valley area were more equable than today, but that detail from a wider geographicalarea. The Lake seasonaldryness became establishedby the Late Eyre Basin macroflorasindicate a developmentby Miocene, and perhaps much earlier in the conti- the Middle Eocene of seasonallywet/dry ?mega- nental interior (Lange, 1982; Q.rilty, 1984; thermal climates in the continental interior, and Greenwood et al., 1990). that vegetation physiognomically pre-adapted to The poor knowledge of macroflorasof suitable aridity was in place in the interior at that time age and preservation over much of the continent (Greenwoodet a1.,1990; Christophel et al.,1992). prevents a continental synthesis at this time. Wolfe (1985) speculatedon the nature of Aus- There are further areasof researchon palaeobot- tralian Early Miocene vegetation and climate. anical climatic evidence, such as wood anatomy However,the limited number ofwell-documented and dispersed cuticle (e.9. Upchurch & Wolfe, Oligocene and Miocene macrofloras from south- 1987), which remain largely untapped. Important easternAustralia precludesany extraregionalsyn- information on Australian Tertiary climates can thesis.What is clear is that this region experienced be gained from analyses of macrofloras. Work cooler temperaturesin the Oligocene than in the continues on the well-presened early to mid- Eocene (see e.g. Hill & Gibson, 1986; Hill, Tertiary macrofloras of southeastern Australia Palaeobotanicalevidence for Tertiary climates 57

Island, and Tasmania; however, there is an urgent need CASF.,,J. A. (1988). Paleogenefloras from Seymour Antarctic Peninsula. GeologicalSociety of Ameica Memoir, to find and analyse macrofloras from outside of 169,523-30. to apply a wider range of these areas and CHRISTOPHEL, D. C. (1981).Tertiary megafossilfloras approaches. as indicators of floristic associationsand palaeoclimate.In EcologicalBiogeography of Australia, ed. A. Keast, pp. This researchwas completed partly at the Botany 379-90. The Hague: W. Junk. Department, University of Adelaide, South Australia, CHRISTOPHEL, D. C. (1985).First recordof well pre- and partly at the Biolory Department, University of senred megafossilsof Nothofagu.sfrom mainland Australia. the South Pacific, Fiii Islands. The manuscript was Proceedingof the RaltalSociety of Victoia,97,175-8. completed during a NSERC International Fellowship CHRISTOPHEL, D. C. (1988).Evolution of the Austra- at the University of Saskatchewan,. The lian flora through the Tertiary. Plant Systematia and Anglesea data were collected with D. C. Christophel Eaolution, 162, 63-78. by Earthwatch volunteers in 1986. Data from Golden cHRISTOPHE,L, D. C. & BLACKBURN, D. T. (1978). Grove (in part) were supplied by D. J. Barrett from The Tertiary megafossil flora of Maslin Bay, South Aus- research funded by the South Australian Department tralia: a preliminary report. Alcheinga, 2, 3ll-19- of Mines & Enerry, and Monier Pty Ltd. Additional CHRISTOPHF,I,,D. C. & GREENWOOD,D. R. funding was provided by the Australia Council (to (1987). A megafossil flora from the Eocene of Golden Christophel). I thank D. C. Christophel, A. I. Rowett, Grove, South Australia. Transaclionsof the Royal Society L. J. Scriven, R. S. Hill and R. J. Carpenter for of SouthAustralia,lll, 155-62. supplying unpublished data, and J. F. Basinger, CHRISTOPHEI-. D. C. & GRF,ENWOOD,D. R. B. Pratt, M. E. Collinson and B. A. LePage for (1988). A comparison of Australian tropical rainforest and critically reading the manuscript. Tertiary fossil leaf-beds. In The Ecolog ofAustralia's Wet Tropia, ed. R. L. Kitching. Proceedingsof the Ecological Societyof Australia,15, 139-48. CHRISTOPHE,I,,D. C. & GREENWOOD, D. R. REFERENCES (1989). Changes in climate and vegetationin Australia dur- ing the Tertiary. Rniep of Palaeobotanyand Pafunolog, ABEI-F..C., GLOE, C. S., HOCKING, J. 8., HOLD- 58,95-109. GATE, G., KENI,EY,P. R., I,AWRENCE,C. R., D. C., HARRIS,W. K. & RIPPER, D. & THRELFAL, W. F. (1976).Tertiary. CHRISTOPHEI,, SYBER, A. K. (1987). The Eocene flora of the Anglesea ln Geolog of Victoia, ed. J. G. Douglas & J. A. Ferguson' locality, Victoria. Alcheinga, ll, 303-23. GeologicalSociety of Australia Special Publication, S, 177-274. cHRISTOPHEI., D. C., SCRIVEN,1..J. & GREEN- WOOD, D. R. (1992). An Eocene megaflorafrom Nelly ALLEY, N. F. (1987). Middle Eocene age of the megafossil South Australia. Transactionsof the Royal Societyof flora at Golden Grove, South Australia: preliminary Creek, Australia, 116, 65 -7 6. report, and comparison with the Maslin Bay flora. Trans- South FERGUSON, D. K. (1985).The origin of leaf- artions of the Royal Societyof South Australia, lll, - ?tt-12. assemblages new light on an old problem. Rniew of Palaeobotanyand Palynolog, 46, ll 7-88' AMBROSE,G.J., CAI-LEN,R. A., FLINT, R. B. & FRAKES, L. A. & FRANCIS, E. (1990).cretaceous LANGE, R. T. (1979). Eucalyptusfruits in stratigraphic J. palaeoclimates. In CretaceousRaources. Euenls and Rhythms, context in Australia. Nature,280, 387-9. ed. R. N. Ginsberg & B. Beaudoin,pp. 273-87. Dor- AXEI-ROD, D. L (1984). An interpretationof Cretaceous drecht: Kluwer Academic publishers. and Tertiary biota in polar regions. Palaeogeography, (1984). Leaf and canopyadaptations in P alaeoclimatologt,P ala.eoecologt,45, 105- 47 . cIVNISH, T. J. Ecolog of Plants of the lilet BEADI-E, N. C. W. (1967). Soil phosphateand its role in tropical forests. ln Ph.ysiological & C. Vazquez-Yanes' Tasks molding segmentsof the Australian flora and vegetation, Tropia, ed. E. Medina for -84. with special reference to xeromorphy and sclerophylly. VegetationS cience,12, 5 (l99la). of plant Ecologt, 47, 992-1007 . cRE,ENWOOD, D. R. The taphonomy macrofossils. ln The Processesof Fossilization, ed. BURNHAM, R. J. (1989). Relationshipsbetween standing BelhavenPress. vegetation and leaf litter in a paratropical forest: implica- S. K. Donovan, pp. l4l-69. : tions for paleobotany. Rniep of Palaeobotanyand Pa$n- cREENWOOD, D. R. (l99lr). Middle Eocenemega- ologt,58,5-32. floras from central Australia: earliest evidence for Aus- BURNHAM, R.J., WING, S. L. & PARKER,G. G. tralian sclerophyllousvegetation. American Journal of (1992). The reflection of deciduous forest communities in BotanySupplement, T8, I l4- 15. leaf litter: implications for autochthonous litter assem- cREENWOOI), D. R. (1992). Taphonomic constraintson blages from the fossil record. Paleobiolog,18, 30-49. foliar physiognomic interpretations of Late Cretaceous 58 D. R. GREF,NWOOD

and Tertiary palaeoclimates. Rniap of Palambotany and palaeoclimatic estimatesfrom pollen data using bioclim- Palynolog, Tl, 142-96. atic profiles of extant taxa.Journal of Biogeography,15, GREENWOOD,D. R.,CAI,I,EN, R. A. & ALLEY, N. F. 589-602. (1990). The conelationand depositionalmaironmnt ofTmiary LANGE,R. T. (1976).Fossil epiphyllous'germlings', strata baed on manoforas in the Southen Lake Eyre Basin. their living equivalentsand their palaeohabitatvalue. South Ausnalian Department of Mines & Energy Report Nna Jahrbuch fir Geologieund Palihntologie,l5l, No. 90/15. 142-65. GREENWOOD,D. R. & CHRISTOPHE,L,D. C. I-ANGE, R. T. (1978). Southern AustralianTertiary epi- (1990). Leaf Margin Analysis of Cretaceousand Tmiary cli- phyllous fungi, modern equivalents in the Australasian mata: modtrn Southern Htmisphere rainforats as analogua. region, and habitat indicator value. CanadianJoumal of Abstracts. Geological Society of America Annual Meet- Botany, 56, 532-41. ing, Dallas, Texas, p. 350. I-ANGE, R. T. (1981). Arnaud's'enigmatic liftle marks': HILL, R. S. (1982).The Eocene megafossilflora of Ner- an extension (type 6) to the manginuloid hyphae series of riga, New South Wales, Australia. Palaeontographica epiphyllous fung1. Experimtia, 37, 720. Abt. B, 181,44-77. I.ANGE, R. T. (1982). AustralianTertiary vegetarion,evi- HII.L, R. S. (1984). Tertiary Nothofagusmacrofossils from dence and interpretation. ln A History of Australasian Cethana, Tasmania. Alcheringa,8, 8l-6. Vegetation,ed.J. M. B. Smith, pp.44-89. Sydney: HII-I., R. S. (1989). Early Tertiary leavesof Menisperma- McGraw-Hill. ceae from Nerriga, New South Wales. Alcheringa,13, MACPHAII,,M. K., HILL, R. S., FORSYTH,S. M. & 37-44. WF.LLS, P. M. (1991).A Late Oligocene-Early HILL, R. S. (1990a).Evolution of the modern high latitude Miocene cool climate flora in Tasmania. Alcheringa,15, flora: evidence from the Australian 87- 106. macrofossil record. In Proceedingsof the Third IOP Confer- McGOWRAN, B, HARRIS, W. K. & LINDSAY, J. M. etce, ed. J. G. Douglas & D. C. Christophel, pp. 3l-42. (1970). The Maslin Bay flora, South Australia, l. Evidence Melbourne: A-Z Printers. for an Early Middle Eocene age. Neua Jahrbuch fi)r Geolo- HII-1., R. S. (1990b).Araucaia (Araucariaceae)species gie und Paliiontologie,8, 481-5. from Australian Tertiary sediments - a micromorphological NIX, H. (1982). Environmental determinants of biogeogra- study. A ustrali an Sys tenatic B otany, 3, 203 - 20. phy and evolution in Terra Ausnalis. ln Eoolution of the HII-L, R. S. (1991).Tertiary Nothofagus(Fagaceae) macro- Flora and Fauna of Aid Australia, eds W. R. Barker & fossils from Tasmania and Antarctica and their bearing on P.J. M. Greenslade, pp. 47 -66. Frewville: PeacockPubli- the evolution of the genus. BotanicalJournal of the Linnean cations. Society,105, 73- I12. OI-LIER, C. D. (1986). The origin of alpine landforms in HII-L, R. S. & CARPENTE,R,R. J. (1991).Evolution of Australasia. ln Flora and Fauna of Alpine Australasia, ed. Aonopyle and Danycarpru (Podocarpaceae) foliage as B. A. Barlow, pp. 3-26. Melbourne: CSIRO. inferred from macrofossils in south-eastern Australia. PARRISH, J. T. & SPICER, R. A. (1988).Late Cre- Australian SystematicBotany, 4, 449 -79. taceous terrestrial vegetation: a near-polar temperature HILL, R. S. & clBSON, N. (1986).Macrofossil evidence curve. Geolog, 16, 22-5. for the evolution of the alpine and subalpine vegetation of QUII-TY, P. c. (1984). Mesozoic and Cenozoichistory of Tasmania. ln Flora and Fauna of Alpine Australasia: Aga Australia as it affects the Australian biota. ln Aid Aus- and Oigins, ed. B. A. Barlow, pp. 205-17. Melbourne: tralia, ed. H. G. Cogger & E. E. Cameron, pp.7 -56. CSIRO. Sydney: Australian Museum. HILL, R. S. & MACPHAIL, M. K. (1983).Reconstruc- READ, J. (1990). Some effects of acclimation temperature tion of the Oligocene vegetation at Pioneer, northeast on net photosynthesis in some nopical and extra- Tasmania.Alcheinga, 7, 281-99. tropical Australasian Nothofagusspecies. Journal of Ecolog, HILL, R. S. & READ, J. (1987).Endemism in Tasmanian 78,100-12. cool temperate rainforest: alternative hypotheses. Botan- READ, J. & HOPE, c. S. (1990).Foliar frost resistance icalJounal of the Linnean Society,95, ll3-24. of some evergreen tropical and extratropical Austra- KE,MP, E. M. (1978).Tertiary climatic evolution and lasian Nothofagusspecies. Australian Journal of Botany, 37, vegetation history in the southeast Indian Ocean region. 36r-73. P alaeogeography, P alaeodimatologt, Palacoecologt,24, RE,AD,J., HOPE, c. S. & HILL, R. S. (1990).Integrat- 169-208. ing historical and ecophysiological studies in Nothofogus KE,MP, E. M. (1981).Tertiary palaeogeographyand the to examine the factors shaping the development of cool evolution of Ausnalian climate. ln EcologicalBiogeogra- rainforest in southeastern Australia.ln Proceedingsof the phy of Australia, ed. A. Keast, pp. 3l-50. The Hague: Third IOP Confercnce,ed. J. G. Douglas & D. C. W. Junk. Christophel, pp. 97 - 106. Melbourne: A-Z Printers. KERSHAW, A. P. & NIX. H. A. (1988).Quantitative ROTH, J. I.. & DILCHER, D. L. (1978).Some consider- Palaeobotanicalevidence for Tertiary climates 59

ations in leaf size and margin analysis of fossil leaves. ecologica de los bosques del sur de Chile. Bosque,4, Couier Fonchung-Ins titut S enckenberg,30, 165-7 l. 73-l 15. 'Equable'cli- SI-OAN, t,. C. & BARRON, E. J. (1990). WEBB, L. J. (1959). A physiognomicclassification of Aus- mates during history? Geologr, 18, 489-92. nalian rainforests.Journal of Ecolog, 47, 551-70. SI-UITER, I. R. K. (1991). Early Tertiary vegetationand WEBB, I-. J. (1968). Environmentaldeterminants of the climates, Lake Eyre region, northeastern South Australia. structural types of Australian Rain Forest Vegetation. Ecol- ln The Cainozoic of Australia: A Re-appraisal of the Eoidence, ogt,49,296-311. ed. M. A. J. Williams, P. De Dekker & A. P. Kershaw. WE,BB, L.J. & TRACEY, J. G. (1982).Rainforats: Data Geological S ociety of Aus ralia Special Publication, 18, on Floistia and Site Characteristia. Canberra: Biogeo- 90- I 18. graphic Information System, Bureau of Flora and Fauna. SPICER, R. A. (1981). The sorting and depositionof WOLFE, J. A. (1971).Tertiary climatic fluctuationsand allochthonous plant material in a modern environment at methods of analysis of Tertiary floras. Palaeogeography, Silwood Lake, Silwood Park, Berkshire, England. United Palaeoclimatologt, Palaeoecologt,9, 27 - 57 . S tata Geological Sun:ey Profasional Paper, 1143. WOLFE, J. A. (1978). A paleobotanicalinterpretation of SPICER, R. A. (1989).The formation and interpretation Tertiary climates in the . American of plant fossil assemblages.Adaanca in Botanical Scientist, 66, 694 -7 03. Raearch,16, 96- l9l. WOLFE, l. A. (1979). Temperature parametersof humid TOKARSKI, A. K., DANOWSKI, W. & ZASTAWN- to mesic forests of eastern and relation to forests IAK, E. (1987). On the age of fossil flora from Barton of other regions of the Northern Hemisphere and Austra- Peninsula, King George Island, West Antarcnca. Bitish lasia. United Stat,a GeologicalSun:ey Profasional Paper,1106. Polar Raearch, 8, 293 - 302. WOLFE, J. A. (1985). The distribution of major vegeta- TRACEY, J. c. (1982). The Vegetationof the Humid Tropia tional types during the Tertiary. ln The Carbon Cycleand of North Qreensland. Melbourne: CSIRO. AtmosphericCOz. Natural Vaiation Archaen to Praent, ed. TRUSWELL, E. M. & HARRTS,W. K.(1982).The E. T. Sundquist & W. S. Broeker. American Geopfusical Cainozoic palaeobotanical record in arid Australia: fossil Union Monograph,32, 357-75. evidence for the origin ofan arid-adapted flora. In Eaol- WOLFE, J. A. (1990). Palaeobotanicalevidence for a ution of the Flora and Fauna of Aid Australia, ed. marked temperature increase following the Cretaceous/ W. R. Barker & P.J.M. Greenslade,pp. 67-76. Frew- Tertiary boundary. Nature, 343, 153-6. ville: Peacock Publications. WOLFE, J. A. & UPCHURCH, c. R. (1987).North UPCHURCH, c. R. & WOr,FE, J. A. (1987).Mid- American non-marine climates and vegetation during Cretaceous to Early Tertiary vegetation and climate: the Late Cretaceous. Palaeogeography, Palazodimatologt, evidence from fossil leaves and woods. ln The Oigin of Palamecologt,61, 33 -77 . Angiospenns and their Biological Consequenca, ed. WOPFNER, H., CALI-EN, R. A. & HARRIS, W. K. E. M. Friis, W. G. Chaloner & P. R. Crane, pp. 75-105. (1974). The Lower Tertiary Eyre Formation of the Cambridge: Cambridge University Press. southwestern Great Artesian basin. Joanal of the Geologi- VEBLEN, T. T. & SCHLEGEL, F. M. (1982).Resefra cal Societltof Auslralia,2l, 17-52.