The Taphonomy of Plant Macrofossils.

Chapter 7 THE TAPHONOMY OF PLANT MACROFOSSILS David R. Greenwood INTRODUCTION Taphonomyis definedas the study of the transitionof organicremainsfrom the living organismto fossil assemblages (Efremov, 1940).As such, plant taphonomyincorporatesthe processes of plant parts, of the initial abscission their transport (by air and/or water) to a place of eventual deposition, entrapmentand eventualburial, and subsequentlithification. Within these processesa number of factors can be identified which influenceboth the characterof the eventualassemblage and its taxonomiccomposition.These factorsproducea taxonomicmix within the assemblage which is a subsetof the taxonomiccompositionof the originalsourceplant communityor commu- nities. The organographiccharacterof the assemblage may also be biased (Collinson,1983).Spicer(1989,p. 99) defineda plant fossilassemblage as'an accumulationof plant parts, derivedfrom one or severalindividuals,that is entombedwithin a volume of sedimentthat is laid down in essentiallythe sameconditions'.His definitionwill be usedhere. Plant macrofossilassemblages are fundamentallydifferent from most ani- mal assemblages in that they are almostentirely composedof disarticulated parts.This resultsin part from the continualproductionthroughouta plant's life cycle of generally temporary modular plant organs-leaves, stems, flowers and fruits-which are shed from the plant when their usefulnessis complete,throughtraumaticloss,or for the dispersalof propagules.Eachof thesepartsis given a separatenameuntil connectionwith other partscan be demonstrated(brgan taxa; for example,Stigmaria and Lepidophylloidesare the root systemand leaves,respectively,of the sameCarboniferouslycopod: Thomas and Spicer, 1987;Thomas, 1990). A further factor is that plant 142 David R. Greenwood communities,in contrast to most animal communities,are composedof organismsthat remain in one place throughout their whole life cycle. In common with animal remains, the behaviour of the disarticulatedplant organsas sedimentaryparticlesvaries,as doestheir preservationpotential. Considerationof the above factors in the interpretationof plant fossil assemblages hasonly occurredcomparativelyrecently.However,early work by Chaney(L924;Chaneyand Sanborn,1933)pioneeredthe idea of using modernleaf assemblages to interpretTertiary localities.Over the last decade a number of seminalworks have appearedon plant taphonomy (Hickey, 1980;Spicer, 1980;1981;Ferguson,1985;Scheihingand Pfefferkorn, 1984; Gastaldo,1986; 1988;Burnham, 1989).Spicer (1930; 1989) and Hickey (1980)stressedthe need to considerplant macrofossilassemblages within their stratigraphicand sedimentological context. In a seriesof experiments Spicer(1981)and Ferguson(1985)have investigatedthe behaviourof plant parts,mainly leaves,as sedimentaryparticlesto determinethe factorswhich control their dispersal.Other plant taphonomistshave concentratedon the influenceof the-characterof the standingvegetationand sedimentaryen- vironments on the resulting plant macrofossilassemblages(Drake and Burrows, 1980;Scheihingand Pfefferkorn, 1984;Gastaldo,1986;1988;Hill and Gibson,1986;Taggert,1988;Burnham,1989). Thesestudieshave contributedto a changein how plant assemblages are sampledand interpreted.This chapteris not intendedas a review of plant taphonomy,as thorough reviewshave appearedrecently (Gastaldo,1988; Spicer,1989).Rather, this accountemphasizes the influenceof vegetationon plant taphonomy,althoughnecessarily somemajor points are reviewed. The natureof theplant macrofossilrecord Most plant macrofossilassemblages are in fact fossilizedlitter (Figures7.1, 7.2).Plant macrofossilsconstituteany plant organor organpart whichis large enough to be recognizablewithout microscopy. Thus, plant macrofossils includeleavesand other foliar organs(suchas stipulesand raches),flowers, fruit and other reproductiveorgans (for example,cones,sporangia,etc.), stemsand stem fragments(includingwoody axes),and root systems.These organsare producedin considerablydifferentproportions,with, for example, most treesproducingonly a singletrunk throughouttheir life, but thousands of leavesin any given year. Although strictlynot macrofossils, the cuticleof fossilleaves(asdispersedcuticle)and its associated fungalmicrofloraare also studied by palaeobotanists,providing both biostratigraphicand palaeo- climatic information (Kovach and Dilcher, L984;Lange, 1976;L978;Wolfe and Upchurch,1987). Somepalaeobotanists have stressedthe varying preservationpotential of plant organs(Gastaldo, 1988;Spicer, 1989).Heavily lignified plant tissue, suchaswoodystemsand somefruits, decaysfar slowerthan, for example,the delicateperianth and stamensof most flowers,and so has a higher preser- vation potentialthan the latter. The chemicallyresistentcuticleof leavesalso Taphonomyof plant macrofossils 1tli| hasa very high preservationpotentialand may persistextremelywell in many different sedimenttypes.However, in contrastto many animals,most plant organshavea similarpreservationpotential.In many animalshardpartssuch as shellsor bones have high fossilizationpotential, whereasthe softparts (visceralmass)havea very low preservationpotentialand are only preserved in fossil-Lagerstiitten(Seilacher,L990).However, the relative preservation potential of relatedplant organsvariesenormouslybetweendifferent taxa, institutingan importantbiasin the plant fossilrecord. In general,mostplant organshavea poor preservationpotentialcompared to vertebratebonesor shellyfossils,and so havea very limited potentialto be reworked from older sedimentsinto much youngersediments.Thus, time- averaging-the emplacementof fossilsfrom older stratigraphicpositionto a youngerpart of the sequence-isfar lesscommonin plant macrofossilassem- blagesthan animalassemblages. Fossilizedwood, either as unalteredlogsor as petrifications,and (more so) pollen and spores,are highly resistantto decay,and may experiencesignificantreworking (see,for example,Kemp, L972).More rarely,blocksof sedimentmay be reworked,carryingportionsof an older macrofloraintact into youngersediment(see,for example,Hill and Macphail,1983). Plant fossilsare preservedprimarily within clastic sediments(including volcanoclastics): as compressions with much of the organicmatter preserved (but flattened);asimpressions with the organicmatter mostlylost, leavingan imprint or stainon the rock; as fusainthroughthe conversionof the cell wall into charcoal by burning; and as permineralizationsand petrifications (Schopf,1975;Scott, 1990).Compressionfossilsvary from essentiallymum- mified remainspreservingconsiderableanatomicaldetail to preservingonly the cuticularenvelopeof leaves.The sedimentsmay becomelithified, result- ing in the chemicaltransformationof the plant remainsinto petrificationswith little or none of the original organic material retained. This often occurs through the secondarydepositionof silica, carbonateor pyrite, preserving internal or externalmoulds of cells, tissuesor whole organs(Schopf,1975; Scott,1990).Examplesrangefrom silicifiedtransportedleaf litter (Petersand Christophel, t978; Greenwood et al., 1990) to in situ plant communities (Knoll, 1985).Silicifiedplant assemblages often preservea high degreeof internal anatomicaldetail (see,for example,Petersand Christophel,1978). The cell wall of plantsis primarily celluloseand generallyonly preservesin acidicanoxicconditions.Theseconditionsare most commonlvfound at the bottom of lakes, (generallyabandoned)streamchannels,and in swampsor river deltas. The plant macrofossilrecord is therefore random and strongly biased towards plant communitiesrepresentativeof high rainfall environ- ments, or environmentsassociatedwith water coursesand other water bodies.Important plant macrofossilassemblages are also found in volcano- clasticdeposits(see,for example,Burnhamand Spicer,1986).Theseinclude many importantNorth AmericanPalaeogene floras(Wing, 1987).Individual plant fossil beds consist primarily of disarticulatedintermixed individual organsof many different taxa from the local vegetation.Only rarely are whole plantsor whole plant communitiespreservedin a form which approxi- matesthe originalplant or vegetation.  David R. Greenwood Fi g u re7. 1 T he f or es t-fl o o rl i tte ro f C o m p l e xN o to p hyl lV i ne Forestfrom north- e a stOueens land( equ i v a l e nto t P a ra tro p i c aRl a i n fo r estof W ol fe, 1979). Taphonomy of plant macrofossils 145 Figure 7.2 The oxidized leaf-mat from Golden Grove (Middle Eocene,South Au stra l i a ) . 1rt6 David R. Greenwood The traditional approachin palaeobotanyhas been to view plant macro- 'flora' and primarily as depositoriesof specimens. fossillocalitiesas a single Thefossilswereviewed asexamples of extinctplant speciesand so, therefore, as evidenceof morphologicalevolution (phylogeneticpalaeobotany).The representation of particulartaxaor lineagesof taxawasthen usedasevidence foi the presenceof particular analogousmodern plant commy{ties and/or climates.This approachignoresthe reality that most individuallocalitiesare composed of a numberof separatedepositionalstructures,eachrepresenting discieteevents, separatedin time and/or space. Even within apparently uniform assemblages such as seen in low-rank coals (lignites), individual lithotypes often contain separate macrofloras, reflecting differences in edaphicconditionsand ecologicalprocesses suchas succession. In a review of North American Eocenemacrofloras,Wing (1987' p. 75I) 'actuallyconsistof the summedfloral listsof a number statedthat manyfloras of separatequarry sites', and cautioned that such floras are not strictly compirable with floras derived from a single excavation. This typg of approachis not always avoidable, however, as the diversity9t a single eiposure may often be very low, and outcrop of a single fossiliferousunit exlensiveand continuousover a wide area.It is commonin thesecasesto see regionalaccountsof the resultingfloral summation.It must be remembered, however,that suchsummedfloraspotentiallycontainseveralplant communi- ties. Separationof sedimentaryfaciesis essential.Burnham (1989,,p. T) found that the sizeof individual North American floras varied dramatically becauseof similar factors,rangingfrom an averageof as few as 64.3 speci- mensto 1667specimens. Ferguson(1985)and Gastaldo(1988)presentedflowchartswhich summar- ized t[e formation of a plant macrofossildepositand emphasizedthe role of taphonomicbiasesoccurringduring transportand deposition.According to their model, the selectivenature of the leaf-rain is further emphasizedby thesepost-abscission biases.Analysesof leaf accumulationswithin modern sedimentaryenvironmentshavedemonstratedalsothat the type of sedimen- tary faciesand the characterof the localvegetationhavea stronginfluenceon the compositionof the fossilmacroflora(Taggert,1988;Burnham' 1989).A brief reviewof the influencesof eachof thesefactorsis presentedbelow. TRANSPORTAND DEPOSITIONOF PLANT PARTS Transportof plant parts A potentiallystrong sourceof taphonomicbias is the transportof the dis- articulatedorgans(particularly leaves)from the sourceplant(s) to the site of deposition.Transportis generallyinitially by air, but ultimatelysomewater transportis involvedin mostcases.The shape,sizeand structureof leaves,as well as the relativedensityof leaf tissue(weightper unit area), are intrinsic factorswhich may influencethe passageof leavesthrougheither air or water (Spicer,1981;1989;Ferguson,1985).However,the characterand behaviour Taphonomyof plant macrofossils 141 of the sourcevegetationhave an initial controllingeffect by determiningthe compositionof the litter-rain. In modern Australian humid tropical forestsleavesfrom the canopytree speciesdominatethe litter-fall (Braseller al., 1980;Spain, 1984;Stockerer al., in press).Leavesfrom other forest synusiae-shrubs,terrestrialherbs, vines,parasites(for example,Loranthaceae,Viscaceae)and giant monocots (Zingiberaceae,Musaceae,etc.)-are rare componentsof the leaf-rain. However, where vines constitutea major componentof the canopy their leavesmay be a significantfractionof the leaf-fall.The volumeand taxonomic compositionof the leaf-rainis alsovariable,with a peak in absolutevolumes just prior to the wet season,althoughleaf-fallin somespeciespeaksat other timesof the year (Spain,1984;Rogersand Barnes,1986;Stockeret al., in press).The presenceof flowersand fruits in the litter-rain is highly variable, and generallyrepresentsa minor fractionof the resultantdebris. The initial compositionof the litter-rainis thereforebiased,with the leaves of the canopytreesswampingthe leavesof other synusiae,and non-foliage organs are a relatively minor component.Potentially,the taxonomiccom- position also variesthrough the year. Leaf-litter from the forest floor reflects this variation and the dominanceby canopytrees. In temperatedeciduous forestsFerguson(1985)found that the taxonomiccompositionof leaf litter was strongly influencedby the nearest trees. Greenwood (1987a)found similar resultsin Australianhumid evergreentropical forestsand evergreen temperaterainforests.Most treesin deciduousforestslose their leavesin a short spaceof time; however,treesin evergreenforestslosetheir leavesover the whole year with peak periods of leaf loss for the forest and particular species (Brasell et al,, 1980; Spain, 1984; Rogers and Barnei, 1986). Residencetime on the forest floor of litter in tropical forests is variable (Andersonand Swift, 1983);however,a significantamount of litter remains from previousmonths, and much longer during dry weather (Brasellet al., 1980;Spain,1984).The litter volumesin the Australianhumid tropicalforests can therefore be expected to contain a time-averagedsample of the litter-rain. The forest-floorlitter (see, for example, Figure 7.1) in the Australian forestsis dominatedby leavesfrom the principalcanopytree species(Green- wood, 1987a),reflectingthe bias seenin the litter-rain (Brasellet al., 1980; Spain, t984; Stockeret al., in press).In the low-diversityforests(temperate rainforestfrom New South Wales)little changewas seenbetweensamples, althoughthe relativeproportionof leavesfrom the non-canopysynusiaewas variable, reflectingspatial variability in the distribution and abundanceof theseplants(Greenwood,L987a).Variation betweensamplesfrom the high- diversity forests (humid tropical lowland and upland forests from northeast Queensland)was high, reflectinghigh spatial variation in abundanceand distributionof speciesin all woody synusiae.This evidenceimplies that the litter wasderivedfrom only the immediatelyadjacenttrees(see,for example, Ferguson,1985). The distancethat the componentsof the litter-rainmay travel is controlled by the behaviourof the individualleaves,flowersand fruits as sedimentary particles(Spicer, 1981; 1989). In a seriesof experimentsusing both leaf 18 David R. Greenwood models and actual leaves, Spicer (1980; 1981; 1989; and references therein) and Ferguson (1985) have examined leaves as sedimentary particles to deter- mine the relative significanceof taphonomic biases created during transport. The distance travelled by leaves in air was found to be largely controlled by the weight per unit area of the leaf (Spicer, 1981; 1989); however, leaf morphology and overall weight are contributing factors (Spicer, 1981; Ferguson, 1985). The general consequence of this is that lighter, usually smaller leaves, travel further than heavier (and denser), usually larger leaves. However, as Spicer (1981; 1989, p. 108) pointed out, the coriaceous'sun leaves' of the upper canopy have a higher weight per unit area than the comparatively larger but membranous 'shade leaves' of the understory, and so can be expected to travel less far than the 'shade leaves'. 45 40 35 30 number of 25 leaves in 10 x 10 cm 20 15 10 2 4 6 I 1 0 L 2 L 4 1 6 1 8 2 0 2 2 2 4 d i s ta n c e fro m s o urce tree (m) Figure 7.3 Number of leaves found in forest floor-litter at distancefrom the source tree (Prumnopitys amara - Podocarpaceae)in 10 x 10 cm quadrats a l o n g t wo t r ans ec t s -u p s l o p e a n d d o w n s l o p e -i n C ompl ex N otophyl l V i ne Forest(sensuWebb, 1959). In unrestricted fall, fewer leaves are found with increasing distance from a source tree in a negative exponential relationship (Figure 7.3; Spicer, 1981; Ferguson, 1985; Greenwood, 1987a).In a forest situation, the structure of the trunk space and the presence of any screening foliage may also significantly influence the passageof leaves. Ferguson (1985) concluded that, in general, leaves are unlikelv to travel further in a forest interior laterallv than the Taphonomy of plant macrofossils 149 height at which they abscissedfrom the tree. Measurement of the dispersal of leaves around their source trees in tropical forests (mesothermal, Notophyllous Forest sensu Wolfe, L979) in northeastern Queensland demon- strated that the complex trunk-space of these forests substantially reduced the effective distance travelled by leaves (Figure 7.3; Greenwood, 1987a). The dominance of the leaf-fall by canopy trees and the local nature of the taxonomic composition of the leaf-fall controls the physiognomic character of the resulting leaf assemblage.Correlations between the foliar physiognomy of modern vegetation (leaf size and margin type) and climate (mainly mean annual temperature (MAT): Wolfe, 1979; Upchurch, 1989; Spicer, 1990) have been used to predict palaeoclimate from leaf assemblages(see, for example, Wolfe, 1985; 1990; Wolfe and Schorn, 1989). The main criticism of this approach has been the extent to which a decodable climatic signal is preserved in the foliar physiognomy of leaf assemblages(Dolph and Dilcher, L979; Roth and Dilcher, 1978;Burnham, 1989;Christophel and Greenwood, 1988; 1989). A key assumption,however, has been that the leavesof canopy trees will dominate fossil leaf assemblages.Based on the evidence of leaf-rain and forest-floor litter in Australian forests, this assumption would appear well founded. Greenwood (1987a; Christophel and Greenwood, 1988; 1989) and Burnham (1989) have examined whether a decodable physiognomic signal can be detected in modern leaf assemblages. The forest-floor litter examined by Greenwood (1987a) contained leaves much smaller than expected based on canopy values for the same forests. Using taxon-based observations (leaf size index (LSI)t), the litter was found also to have fewer taxa with larger leaf-sizeclassesthan expected from canopy observations (that is, smaller LSI values). In general, the bias was in the order of. 3:2 smaller LSI values for litter compared to canopy (Figure 7.4), with departures between litter and canopy decreasingwith concomitant decreases in mean leaf size in the canopy. However, the proportion of species with entire leaf margins in the litter was generally different than that recorded for the canopy alone (Table 7.I), reflecting the influence of non-canopy synusiae. Similar biases in actual fossil assemblageswould bias estimations of MAT by similar ratios. Burnham (1989) sampled litter from a number of sedimentary subenviron- ments in a riverine environment within Paratropical Rainforest (sensrzWolfe, 1979) in Mexico. In general, Burnham found that the foliar physiognomy (LSI and leaf-margin analysis) of most of the subenvironments reflected the regional climatic signal. Burnham determined the leaf-sizeclassof each taxon found in the litter from canopy-collected herbarium samples, and so gave no measure of any changes in LSI due to biases from the actual leaf-rain. However. in the studies of both Greenwood and Burnham it was demon- strated that a decodable climatic signal (Burnham, 1989), or 'foliar physio- gnomic signature'for particular forest types (Greenwood, 1987a;Christophel and Greenwood, 1988; 1989) was preserved in leaf assemblages,which, when properly constrained, can be used to reconstruct palaeoclimates. An import- ant constraint is the influence of additional transport by water prior to deposition. Transport of plant parts in water is controlled by the rate at which they - Taphonomy of plant macrofossils 151 becomewaterlogged,their hydrodynamicpropertiesand the turbulenceof water flow (Spicer,1.989).The rate of decayalsocontributesto the transport and preservationpotentialof the part, with different ratesof decaydemon- stratedfor the leavesof differentspecies(Ferguson,l97l; 1985;Greenwood, I987a), and also for leavesof differing physiognomy(Heath and Arnold, L966).Ferguson(I97I;1985) found that whereastrees of.Ilex europeawere commonin local forest, leavesof this specieswere reducedto almostunrecog- nizablebagsof cuticle after only short periods of immersion. Similarly, litter collected in a stream surrounded by forest with a canopy dominated by Doryphora sassafras (Monimiaceae) and Ceratopetalumapetalum(Cunonia- ceae) was found to be dominated by leavesof C. apetalum,with few or no leavesof.D. sassafras.Litter collectedfrom the forest floor nearby contained leavesof both speciesin nearly equal proportions (Greenwood, L987a). Preferentialdecayin water of somespeciesmay, therefore,bias fluvial and lacustrineleaf assemblages. However, Hill and Gibson (1986) found that leavesfrom a number of speciesfrom Tasmaniansubalpineevergreenvegetationwere essentially intact after six monthsof immersion,in constrastto the high level of decay detected for leaves from deciduous temperate speciesafter two to four months (Spicer, 1981; Ferguson, 1985). This difference probably reflects differencesin the chemicaland anatomicalnature of the two sets of leaves (Ferguson,1985;Spicer,1989),with the evergreenleavestypicallycoriaceous and rich in both lignin and tannins(and other phenoliccompounds),and the deciduousleavestypically papery and containing fewer phenolics.Spicer (1989)discussed how leavesmay undergolimited transportwithin the water column,and so the periodleavesremainrelativelyintact free of sedimentis a controllingfactor. However,the initial behaviourof leaves(and other parts) as they enter a water body largelycontrolswater transportdistances. The leavesof most temperatedeciduoustreesare sheddry (Spicer,1981), whereasmany evergreenspeciesshedessentiallyunalteredleaves.A dry leaf will remain on the surface of still water for more than several weeks. Experimentswith aquariahave shown that thin papery leavesfloat for much shorter periods of time than thick coriaceousleaves (Spicer, 1981; 1989; Ferguson,1985).Duration of floatingappearsto be controlledby the rate at which the leaf tissuebecomeswaterlogged.Cuticle thickness,stomatalfre- quencyand size (for example,hydathodes),damageto the leaf lamina and petiole, and the conditionsof the water (temperatureand chemistry)would appearto be the main factorscontrollingwater uptakeby the leaves,and thus floatingtimes (Spicer,1989). Hill and Gibson (1986) found that the majority of leavesof Eucalyptus coccifera and Orites acicularissank within two days. The leavesof both of thesespeciesare sclerophyllous (markedlycoriaceous).The majority of other Tasmaniansubalpinetaxaexaminedby Hill and Gibson(1986)had significant numbersof leavesfloating after much longer periodsof time. Spicer(1981; 1989)also found variablefloating times, with the thin papery leavesof the deciduous speciesAlnus glutinosa sinking within hours, while significant numbersof evergreen(coriaceous)leaves of.Rhododendronremained float- ing after severaldays. Hill and Gibson (1986)concludedthat sinking rates 152 David R. Greenwood may be a significantfactor determining the distribution and abundanceof leavesin lake sediments. Christopheland Greenwood( 1988;1989;Greenwood,1987a)found that leaves collected in allochthonous deposits downstream were significantly smaller than leaves from an autochthonous(essentiallyforest-floor) leaf assemblage,approximately100m upstreamin SimpleNotophyll Vine Forest (northeastQueensland).Little differencein the relativewidth wasobseryed, althoughthe leavesin both assemblages tendedtowardsnarrow elliptic. This suggeststhat there was a trend towardsstenophyllyin the leavescontributing to the stream litter load, perhapsreflecting a bias towards input from river- margin vegetation(Greenwood, 1987a;Christopheland Greenwood,L988; 1989). Both assemblages were more diverse than equivalent forest-floor samples,with the allochthonoussamplethe most diversewith 36 leaf taxa from 249 specimens.Transport within streams,therefore, ray significantly alter the physiognomiccharacterof leaf assemblages, and resultsin enhanced samplingof the local vegetation(seeBurnham, 1989). The variableform and sizeof fruits, seedsand flowerscontributesto a wide variationin observedfloatingtimesby theseorgans(Collinson,L983;Spicer, 1989).Collinson(1983)concludedthat depositscomposedprimarily of seeds and fruits (that is, thick-walled, durable plant material) probably occur through the selectivebiodegradationof the lessdurableplant material(that is, leavesand flowers).Plant organswhich remainfloatingare more likely to be transportedfurther by streamsthan quickly sinkingorgans.Spicer(1989, p. 119)hashighlightedthe problemsattendantwith the highly transportable natureof tree logs(see,for example,Frakesand Francis,1990)and suggested that transportedlogs of unknown provenancepose problemswhen used as palaeoclimaticindicators.Significantaccumulationsof plant detritus occur at the mouthsof rivers,includinga rich assortmentof more durableplant parts, suchas logs,twigs,seedsand fruits. Much of this materialhasprobablybeen transportedconsiderabledistancesfrom upstream. Modern analoguesof plant fossil assemblages A primary distinctioncan be made betweenplant fossildepositsformed from the gradual accumulationof plant material in situ (autochthonousassem- blages) and assemblagesformed by the accumulationof transported plant materials(allochthonousassemblages). This dichotomyhasimportanttapho- nomic consequences as the autochthonousassemblages reflectprimarily the plants growing within the depositionalsite, including the potential preser- vation of in sitz whole or nearly completeplants. These assemblages are thereforelikely to representonly the immediatevegetation.The allochtho- nous assemblages contain plant material which may have been transported from a number of separateplant communitieswithin the local depositional basinand so could representseveralplant communities.Taphonomicbiases caused by transportation effects are also likely to be more profound in allochthonousassemblages. I I I I Taphonomy of plant macrofossils 153 Autochthonous assemblages:peat bogs and swamps (mires) The abundant Mesozoic and Cainozoic coal sequencesrepresent fossilized peat. Peatresultsfrom a long-termaccumulationof plant matter and usually occursthrough the suppressionof decayprocesses in the subsoiland humic layer of soils. The conditionsnecessaryfor the accumulationof peat are usuallyfound where soil water levelsare high, maintaininganoxicsoil con- ditions.As such,they can be found from the wet tropicsto the arctictundra (Moore, 1989).The presenceof coal doesnot imply high grossproductivity by theseplant communities,but rather high net accumulation(Moore, 1989). Petrological differences between coals reflect a combination of differing source sedimentaryenvironmentsand biological communitiesas well as subsurfaceprocesses subsequentto burial (Cameronet al., 1989).In particu- lar, differencesbetweenlithotypeswithin a singleseam,particularlyin low- rank coals,canbe attributedto the presenceof a varietyof plant communities in the original peat-formingvegetation(Luly et al., 1980;Cameron et al., 1e8e). Peat-formingplant communitiesrangefrom essentiallytreelesspeat-bogs dominatedby bryophytes(typicallySphagnum)and pteridophytesor swamps of herbaceousmonocots(often including Sphagnum)to swampscontaininga significantcoverof woody plants,includingtrees(for example,the Kerangas of Borneo; Bri.inig, 1983). Tree-dominatedswamps, or bog forests, are usually dominated by conifers at mid- to high latitudes (for example, Taxodium, Picea and Larix in the northern hemisphere;Dacrydium sensu lato and Dacrycarpusin the southern hemisphere),but contain significant angiospermcomponentsin tropical areas.A mosaicof different peat-forming communitiesmay occur within a singlebasin.Many of the plant speciesmay be restricted to these communities,or occur only rarely in other plant communities. Peat-formingcommunitiescanbe collectivelyreferredto as'mires'and can be classifiedaccording to the primary source of water to the community (Moore, 1989).Mires where all of the water is sourcedfrom rain are termed 'ombrogenous'(ombrotrophic)(raisedor domed mires), whereasmireswith additionalwater suppliedfrom ground water or inflowing streamsare termed 'rheogenous'(rheotrophic)(Cameron et al., 1989;Moore, 1989)or 'topo- genous'(Macphail and Hope, 1985).The rheogenousmires producepeats with much higher fractions of inorganic matter (mostly clastics)to organic matter than ombrogenousmires due to transportedmaterialsbrought in by inflowingstreams.Additional plant material,particularlypollen and leaves, may also be transported into a rheogenousmire, thus altering the floristic characterof the macrofossilsuite preserved. Peat-formingenvironmentsoccur within a number of sedimentarysettings (Flores,1981;Gastaldoet a|.,1987;Gastaldo,1988;Moore, 1989).Peatsthat form in alluvial floodplains, lake margins, deltaic wetlands and mangroves generallyaccumulatemacrodetritussolely from the in situ vegetation(Gas- taldo, 1988).Peatsmay also form from solely allochthonousaccumulations due to reworking of macrodetritusin channelsand coastalsettings(Gastaldo et al., L987),and are generallycomposedof highly fragmentedplant material (Gastaldo,1988). 154 David R. Greenwood There is some debate over the types of peat community and sedimentary setting which give rise to very large coal seams. On the island of Borneo, and elsewhere, modern and Holocene peat-forming (ombrogenous) raised mires have been studied as possible analogues for large-scale coal formation (Cameron et al., 1989). Floristically, the peat-forming communities in Borneo resemble the Miocene plant communities which produced extensive economic reservesof lignite in southern Australia (Christophel and Greenwood, 1989). Smaller peat (and ultimately coal) structures that form within fluvial settings in floodplain swamps and in deltas (Flores, 1981) are likely to contain some, and may be primarily composed of, transported material (Gastaldo et al., 1987;Gastaldo, 1988). Peat macrofossil assemblagesare generally dominated by seeds and fruits of the mire vegetation, their roots systems (especially the rhizomes of reeds, sedgesand other monocots) and their stems or woody axes (GreatRex, 1983; Raymond, L987; Gastaldo, 1988). Leaf. material tends to be rare. Material from all of the plants present (herbaceous and woody) is incorporated and herbaceous plants may contribute as much as 25 per cent of the accumulated material (Raymond, 1987). GreatRex (1983) found that seedsand fruits were generally transported no further than 1,m from their source plant, but that seeds and fruits which were adapted to wind or water transport may be transported further. Coal, however, generally shows a high level of alteration and often very little plant material is recognizable. In low-rank coals in the Latrobe valley, southeastern Australia (see, for example, Luly et al., 1.980), peat-surface horizons may be seen with in situ tree bases and rarely also with leaf litter beds. The relative abundance of root fossils in autochthonous peat communities is correlated with the abundance of the source taxon in the original plant community (Raymond, 1987). Allochthonous assemblages:fluvial and lacustrine deposits Plant macrofossil assemblagesare commonly formed in fluvial, deltaic and lacustrine settings. Fluvial plant fossil assemblagesare typically small in size and may occur within fine-grained sediments included within coarser fluvial sediments. Several depositional facies exist within fluvial environments: chan- nel, levee and crevasse splay, floodplain, and infilled abandoned channels (oxbow ponds). In addition, swamps are associatedwith some river systems, representing areas of impeded drainage beyond the levee banks which may be seasonallyflooded by the river. Where streams flow into larger water bodies, such as lakes or the sea, deltas form. Significant accumulations of plant macrodetritus may occur in deltaic deposits. These will be only briefly con- sidered here. Channel deposits of meandering river systemscan be divided into channel lag and point bar deposits, representing, respectively, deposition in the active channel and on the inside loops of meandering river channels (Collinson, 1986). In general, plant macrofossil assemblagesare rare in channel lag deposits, although larger plant parts such as tree logs may accumulate (Scheihing and Pfefferkorn, 1984; Wing, 1988). However, the upper portion of point bar structures typically preserves leaves, flowers and other delicate plant structures, whereas the lower part of these structures is typically barren Taphonomy of plant macrofossils 155 of plant fossils, or only contains more durable plant remains (Wing, 1988). Levees form through the preferential deposition adjacent to the existing channel course of suspended sediment by floodwaters. Crevasse splay de- posits occur when the levee is breached, depositing a fan of sediment onto the adjoining floodplain. Typically both of these deposits are characterized by extensive rooting by vegetation and other characteristics of soil formation (Collinson, 1986) and so are poor sites for plant macrofossil preservation (Scheihing and Pfefferkorn, 1984; Wing, 1988; Spicer, 1989), although occasionallyforest-floor litter, tree basesand whole herbaceousplants may be preserved in crevasse splay deposits, producing very rich assemblages (Scheihingand Pfefferkorn, 1984; Gastaldo, 1986; Gastaldo et al., 1987). Floodplain environments contain a number of subenvironments, including swamps, inactive channels (oxbow ponds) and alluvial dryland plant commu- nities. The interfluvial areas may be seasonally inundated, receiving a thin, extensive layer of generally fine-grained sediment. These sediments are gen- erally strongly rooted by the in situ vegetation and may also show signs of bioturbation. However, where the floodplain remains waterlogged for a substantial part of the year, plant debris buried by the sediment may produce assemblagesof finely preserved, often quite delicate structures (Wing, 1984; 1988; Spicer, 1989). Where plant growth is luxuriant, thick sequencesof peat may develop in this environment (Flores, 1981; Collinson, 1986). Infills of abandoned channelsare generally of little lateral extent and have a distinctly lenticular shape when viewed in section (see, for example, Potter 'lenses' are usually dominated by flat-bedded, and Dilcher, 1980). Individual fine-grained sediments, mainly clays, with coarser sediments at their base representing the old channel bedload. The clay lenses are a rich source of well-preserved plant macrofossils in Eocene sediments in both North America and southern Australia (Potter and Dilcher, 1980; Wing, t987; Christophel, 1981; Christophel and Greenwood, 1989). Fossilization is usually as compressions or impressions (for example, Golden Grove, Christophel and Greenwood, 1987: see Figure 7.2 herein), although sub- sequent lithification may produce other modes of preservation such as silici- fication (Ambrose et al., L979; Greenwood et al., 1990). Generally, numerous clay lenses are found at individual quarry sites (see, for example, Christophel et a|.,1987), or in closely associatedsites (Christo- phel and Greenwood, 1987; Potter and Dilcher, 1980), and in many cases were coeval or closely coeval, representing the meander track or braided channel zone of single river systems. The Australian Eocene clay lenses are generally finely laminated, with individual layers often defined by rich accumulations of mummified leaves and other plant macrodetritus (Christo- phel and Greenwood, 1987; Barrett and Christophel, 1990). The presence of very delicate plant structures, such as staminate conifer cones with in situ pollen (see, for example, Greenwood, 1987b), suggeststhat little transport of plant parts occurred in many of these assemblages. Sequencesformed on lake beds are generally more uniform in thickness and laterally continuous than fluvially deposited sediments. These lake sedi- ments are consequently tabular in section and may be finely laminated (lamina from less than 1 mm to 1 cm; Wing, 1988). Associated sedimentary 156 David R. Greenwood environments,suchasdeltafronts and swampfacies,may interruptlacustrine sequences. The foresetand toesetbedsof lacustrinedeltaspreserveabundant plant remains,mainly leaves(Spicer, 1980;1981;Spicerand Wolfe, 1987). The occurrence of plant remains in the finely laminated lake sediments generallyshowsa strongrelationshipbetweendiversityand abundance,and distancefrom shore;near-shoresamplesare generallymuchmore diverseand with more specimensfor the sameamount of sedimentthan samplesfurther from shore (Wilson, 1980;Spicer, 1981;Hill and Gibson, 1986;Spicerand Wolfe, 1987). A numberof studieshaveexaminedthe extentto which leaf assemblages in the lakes reflectedthe local plant communities.Very few have considered input of plant material into fluvial environments.Birks and Birks (1980) reviewedmany studieson the input of plant materialsinto lakes, but the majority of these concernedseedsand fruits. Severalother studieshave concentratedon leaves entering lakes from nearby woody vegetation (McQueen, 1969;Drake and Burrows, 1980;Spicer,1981;Hill and Gibson, 1986;Spicerand Wolfe, 1987).The generalconsensus is that lake sediments are dominatedby leavesfrom the localwatersideflora and nearbyvegetation. Additional transport of plant macrodetritusfrom upstreamsources(in the caseof lakes with stream infeeds) or by wind transport (storm effects and chance dispersal) may introduce elements from extralocal vegetation (McQueen,,1969;Drake and Burrows, 1980;Hill and Gibson, 1986;Spicer and Wolfe,1987). Carpenter and Horowitz (1988) and Burnham (1989) have examined modern fluvially depositedleaf assemblages to determinepotential bias in their taxonomicmembershiprelativeto the riparianvegetation.In Carpenter and Horowitz's study, Tasmanianhigh-rainfalltall evergreenforest domin- ated by Eucalyptusobliqua with an understoryof cool temperaterainforest containingNothofaguscunninghamiiand associatedspecieswas sampled. The compositionand relative abundanceof taxa in stream-bedand drift samplesof litter were comparedto the surroundingforest. In general,the litter matchedthe riparian forest; however,very few leavesof E. obliqua were found in any samples,possiblydue to rapid sinkingratesfor leavesof this species(for example,Hill and Gibson,1986).Leavesfrom the interfluvial forest were also encounteredin significantnumbers, indicating that the streambedlitter representedmore than just the riparianvegetation. Burnham (1989) found that most fluvial subenvironmentsreflectedthe local ParatropicalRainforest flora well, but that the adjacenttrees over- whelminglydominatedthe leafbeds.Overproductionof leafletsby legumi- nous treeswith compoundleaveswas a significantfactor (Burnham, 1989). The channel depositscontained a biased sample of the local vegetation, possiblyreflectingthe smaller samplesizesfrom this subenvironment(Burn- ham, 1989).Litter from SimpleNotophyll Vine Forestin New South Wales and northeastQueenslandcollectedby Greenwood(1987a;Christopheland Greenwood, 1988; 1989) in stream beds indicated a small but significant influenceby a specializedriparian flora, althoughthe canopytreesof the local forest overwhelminglydominatedthe deposits.Notably, in the New South Walesexample,leavesof one of the canopydominants,Doryphorasassafras, Taphonomy of plant macrofossils 157 were scarceor rare in streamlitter, but representeda significantcomponent of nearbyforest-floorlitter (seediscussionabove). It can be seen from the discussionabove that observationsof modern depositionalanaloguesdemonstratethat a consequence of the localisedleaf- rain and hydrodynamicinfluencesis that fluvial and lacustrineleaf beds generallyonly reflect the immediatelocal vegetationand thus local plant communities.In contrast,palynoflorasgenerallyreflectthe regionalflora and resolutionof individualplant communitiesis often dependenton comparison with modernfloristicassociations. Biofacies Plant communitiesare not evenlydistributedin spaceor time over modern landscapes and it is reasonableto assumethat this hasbeenthe casefor much of the Mesozoicand Cainozoic.Similarly, sedimentaryenvironmentsvary spatiallyand temporally,and this heterogeneityis expressedby the presence of discretefacies (Reading, 1986). Within each sedimentaryfacies quite characteristic plant assemblages can often be recognized.Theseassemblages are often termed'biofacies',but this term is usedin two differentways.In an ecologicalsense (as above) biofaciesrefer to a fossil assemblagewhich characterizes a regionor body of rock; in a stratigraphicsense,biofaciesrefer to a 'body of rock which is characterisedby its fossil content which dis- tinguishesit from adjacentbodiesof rock' (Moore, L949;Brenchley,1990, p. 395). The ecologicalsenseof the term emphasizesdifferencesbetween environments,whereasthe stratigraphicsenseemphasizeshorizontal(?bio- geographic)and verticalchangesin the biota (temporalchangesin communi- 'biofacies'is usedin the ecologicalsense. ties). In this discussion, Separatebiofacies may be causedby a numberof interactingfactorsand it is somewhatartificialto separatethem. For easeof discussion,they are here separatedinto influencesarisingfrom the sedimentarybehaviourof various plant organs,and thosecausedby the heterogeneityof vegetation.In fluvial- paludal sedimentarysettingsseveralsedimentarysubenvironmentscan be recognized:within-channel,leveebank, point bar, floodplainswamp,aban- donedchannels,deltasand crevasse splays.Hydrodynamicsortingwithin and betweentheseenvironments,and the individualcharacterof the localvegeta- tion surroundingthem, wilt all contribute to forming separatebiofaciesin each of these environments(see, for example,Scheihingand Pfefferkorn, 1984;Burnham, 1989). Thefate of leaves,fruit, wood and otherplant parts The unequal dispersalcharacteristicsand preservationpotential of different organs(leaves,fruits, flowers, wood fragments)and of different forest synu- siae (trees, vines, herbs) contribute to taphonomic biasesacting on the organographiccharacterof plant fossil beds (Collinson,1983;Spicer,1989) and the representationof structuralcomponentsof the standingvegetation (Scheihingand Pfefferkorn,L984;Gastaldo,1988).The differentialpotential 158 David R. Greenwood for incorporation and fossilization of some taxa also ensures that most plant fossil beds preserve a taxonomically biased subset of the local vegetation. Vegetational heterogeneity Differential sorting and ecological separation of the standing vegetation between individual sedimentary environments causes sedimentary facies within a single sedimentary unit to contain different taxonomic subsetsof the local vegetation, representing the subenvironments in the local vegetational mosaic (Scheihing and Pfefferkorn, 1984; Taggert, 1988; Burnham, 1989). At finer sedimentary scales,temporal variability in the character of the leaf-rain (phenological seasonality) may result in cyclic changes in the taxonomic and organographic character of individual horizons within a single sedimentary sequence.Similarly, lateral floristic heterogeneity in the local vegetation may be expressed in the lateral taxonomic heterogeneity of leaf beds (Taggert, 1 9 8 8 ). Several authors have attempted to portray the influence of different sedi- mentary environments on the composition of plant macrofossil assemblages (for example, Wing, 1988; Spicer, 1989). Ferguson (1985) and Gastaldo (1988) summarized the factors which influence dispersal, incorporation (bio- stratinomy; Gastaldo, 1988) and diagenesis.Their summaries emphasizedthe role of these processesand the segregation that occurs between sedimentary facies. Only limited attention has been paid to the initial influence of the vegetation itself (see, for example, Taggert, 1988; Spicer, 1989). In most instances,plant macrodetritus deposition reflects strongly the local flora, with 'dryland' vegetation the specialized riparian or water's-edge vegetation and represented to varying degrees. Figure 7.5 summarizes the interaction be- tween vegetational heterogeneity and the presence of separate sedimentary environments in a hypothetical landscape. In this hypothetical example, a landscape of generally moderate relief is crossed by a meandering river which has numerous cutoff meander lakes (oxbow ponds) in its lower reaches. Lenticular clay bodies are shown in section through the vertical fluvial sequence. Infeed from highlands is dammed forming a small lake and, in the interfluvial floodplain area, impeded drainage has formed a small internal basin within which lacustrine conditions fluctuate with swamp (peat-forming) conditions. Intertonguing lacustrine fine clastic and lignitic sediments are shown in sequencebeneath the present lake and marginal swamp. The vegetation is simplistically divided into a'wetland' (mostly swamp) flora associatedwith water's-edge and mire communities, a 'dry-lowland' flora subdivided into riparian elements and interfluvial forest, and a highland forest flora. Each of these units is indicated by different grades of shading (see legend, Figure 7.5). The relative contribution of each of the floristic elements to leaf assem- blages formed in particular sedimentary facies in this hypothetical landscape is indicated by a series of pie charts. The relative percentagesgiven in these pie charts is partly based on actual examples, but should be viewed as approximate and used as a rough guide only. It can be seen in this example that each sedimentary facies contains a different subset of the regional flora. The overall representation of speciesfrom each floristic element is respect- Taphonomy of plant macrofossils 159 :qE c F o POE E *x eEX : E oF o € rHi cr) ry $ Y - o C , E b EIE ( , g -'- 'SEs c YF (o.g (L c :E.e .9 € I6 gr oE 6 6CD EE: 8 . E o.o o o .=:(', b t.e o) 9 9 a 4.3 o .E C ) o o E 3 tr> o . 6 Ei: I - o O (ootr +t ' o. = ot r! O G l c c oF(,) 3 € E o +. t tr=r 6 sE I O)o r s 6 6 y >(oo) .o E -A o o, c: D x o O)C r- \),t CD .g 3'= 'a = d i ; 9 o 3 9; - e - .o Po ta9 v E e -oo) P - tt ts'oJg o _ - d c P f o :l e ( U . = E - L q g). .! so- )- ;= o- O 6 G ' . c , 6 = P L L - v c O O = . 9 . L9 o -c oE Ese 3T'.C, = o'= tffiilr " Eg * :- t f ; E 3 i L = - o.Y ;i 6 1 9 o l i'o, ir- E E 160 David R. Greenwood ively diluted or enhanced in different facies due to selective taphonomic processes. Essentially autochthonous deposition in the swamp-coal facies (1) produces a flora biased to the 'wetland' flora, including a significant component of herbaceous plants, whereas the lake deposits (2) may include significant elements from the lowland dryland flora due to fluvial transport and also hydrophytes (see, for example, Collinson, 1988). Similarly, the second lake (3) will probably contain significant macrofossil representation of the high- land flora, in addition to the lowland flora, due to infeed from upland fluvial sources. The abandoned channel infill deposits (4) in the main river-channel zone will essentially be dominated by the riparian and lowland flora. Successionalmosaics within this area may also be expressedby differences in the macrofloras of the clav lenses. THE PALAEOECOLOGY OF PLANT MACROFOSSIL ASSEMBLAGES Several plant macrofossil localities have been studied in some detail taphono- mically (by, for example, Gastaldo ,1986; Burnham, 1990). However, Eocene floras from southern Australia are the most familiar to the author and so these will be used to discusssome of the points made above. The main Eocene plant macrofossil localities from Australia are Middle Eocene and occur in the southern part of the present continent; the Maslin Bay and Golden Grove floras of the North Maslin Sands (St Vincents Basin: Lange, 1970; 1982; Christophel, 1981; 1988; Christophel and Greenwood, 1987; 1989) and the Anglesea flora of the Eastern View Formation (Christophel et al., 1987; Christophel and Greenwood, 1989; Figure 7 .2 herein). These Eocene macrofloras were all deposited within fluvial settings and in most instances represent either temporary lacustrine conditions within aban- doned channels, or other channel facies. The Golden Grove, Maslin Bay and Anglesea macrofloras are found in large lenticular clay bodies within cross- bedded coarse-grained sandstones (Christophel and Greenwood, L987; Lange, 1970; Blackburn, 1981; Christophel et al., 1987), indicating low- energy deposition within a larger high-energy fluvial setting, possibly braided streams, although the Anglesea deposits were interpreted by Christophel er al. (1987) as being deposited by a meandering river system. The level of preservation at the Anglesea and Golden Grove localities is generally high, preserving delicate flowers with attached intact stamens (Christophel, 1984; Basinger and Christophel, 1985) and leaf domatia with oribatid mites (O'Dowd et al., in press). The high standard of preservation and delicate nature of much of the fossil material suggestsvery little transport of the plant material prior to deposition, and thus it may be inferred that the leaf assemblagewill only reflect the local vegetation. The presence of arbor- eal, as opposed to soiUlitter, oribatid mites in the fossil leaf domatia suggests that leaves were abscisseddirectly into the Golden Grove clay lens from the canopy of the surrounding forest (O'Dowd et al., in press). Bulk dis- aggregation of the sediment containing the mummified plant matter generally Taphonomy of plant macrofossils 161 releasesmost of the original leaves, and thus avoids the problem of under- estimation of taxon abundance due to overlapping specimens seen in im- pression floras (Ferguson, 1,985;Spicer, 1988). Each of the Eocene macrofloras share common floristic characters, with a high incidence of leaves from the Lauraceae and Elaeocarpaceae, and low- diversity (but consistent) representation of leaves from Myrtaceae, Podocarpaceae and Proteaceae. Macrofossils of Nothofagrzs (Fagaceae) are rare or absent from (mainland) Australian Palaeogenemacrofloras (Christo- phel, 1988; Christophel and Greenwood, 1989). In contrast, the palynofloras associated with these macrofloras are dominated by Nothofagidites (palyno- morph attributed to Nothofagrzs) and a highly diverse assemblage of grains assignableto the Myrtaceae, Podocarpaceaeand Proteaceae. Pollen grains of the Lauraceae are absent. This discrepancy between the microfloras and macrofloras has led to differences in regional vegetational reconstructions for the Australian Palaeogene, with palynologists stressing the role of Nothofagus and microthermal Podocarpaceae in Palaeogene vegetation (Martin, 1978;1981; L982;Truswell and Harris, 1982), whereas the macrofos- sil workers have stressedthe presence of mesothermal rainforests of similar floristic and physiognomic character to the modern tropical rainforests of northeast Queensland (Christophel, 1988; Christophel and Greenwood, 1987; 1988; 1989). A single fossiliferous clay lens is known to occur at Golden Grove (Christo- phel and Greenwood, 1987; Barrett and Christophel, 1990). This structure is finely laminated with numerous laterally extensive, thin mats of leaves defin- ing the layers. Preservation varies from densely packed mummified leaves, fruits, staminate flowers (Christophel and Greenwood, 1987), sporangiate fern fronds (Lygodium sp.) and leaf domatia with oribatid mites (O'Dowd er a|.., in press) to an organic stain (occasionally preserving cuticle) revealing high detail of venation on a leached lighter matrix (Christophel and Greenwood, 1987;Figure 7.2 herein). Barrett and Christophel (1990) collected macrofossils from the Golden Grove macroflora laterally along a single bedding plane for three vertically separated horizons. The separation between successivebedding planes in the Golden Grove clay lens was variable, although typical separation was in the order of a few centimetres, with two lower layers separated by about 2 cm and separated by about 50 cm from the upper layer sampled. Barrett and Christophel (1990) found significant differences in taxonomic membership and frequency between the upper and lower layers at Golden Grove; how- ever, summing the leaves from the two lower layers and comparing this to the upper layer gave a different result (Figure 7.6).This procedure gave similar sample sizes (upper layer 433 leaves and lower layers 506 leaves). The same taxon was dominant (leaves of Elaeocarpaceae aff . Sloanea) laterally within and between each of the layers and the frequency and representation of other taxa changed only marginally between the lower layers and the upper layer. Some distinctive elements from the upper layer, such as aff,.Neorires (Protea- 'taxon 25' (affinities unknown), were rare or absent in the two ceae) and lower layers (Barrett and Christophel, 1990). The floras of the separatelayers were neverthelessessentiallv the same. The mean size of leaves varied, with 162 David R. Greenwood 14 L2 of leaves (r) 10 I 6 4 5s 75 85 95 105 115 L25 r35 145 155 165 17s leaf size classes (nm) upper layer - 433 leaves lower layer - 506 leaves 160 140 L20 100 number of leaves 80 60 1 5 1 6 4 1 0 5 1 4 8 leaf parataxa Figure7.6 Comparison of leafsizes(specimens)andtaxonomicmembership, lower leaf layersversusupper leaf layers,EoceneGoldenGrovemacroflora (adaptedfrom Barrett,unpublisheddata; Christopheland Greenwood,1987; BarrettandChristophel,1990). the upper layer yielding slightly larger leaves than the two lower layers (Figure7.6). The LSI, basedon the relativesizeof eachtaxon (Wolfe, 1979; Burnham, 1989),is neverthelesssimilar for the upper and lower layers(25 and22.5). It is likely that the separateorganic-richlayersat Golden Grove represent depositionof litter over perhapsonly one to a few generationsof treesin the original standingvegetation,a time period of hundredsor tens of yearsto 7 Taphonomyof plant macrofossils 163 perhapsonly betweenseasons(Barrett and Christophel,1990).The small observedvariationof taxonomicmembershipdominancebetweenthe three beddingplanesmay, therefore,reflecteitherseasonalchangesin the leaf-rain contributingto the leaf assemblage, or changesin the compositionof the sourcestandingvegetation(Barrett and Christophel,1990).Alternatively, the changesmay reflect differencesin the size and orientation of the source area sampledby the transportingmedium. If the differencesin the physi- ognomyof the leavesbetweenthe lower and upperlayersare significant,then this would imply a climaticchangebetweenthe times of depositionof these layers,thus supportingthe interpretationof ecologicallycontrolled change over longer time periods (hundredsof years). However, comparisonsbe- tween litter samples within a single stream flowing through Complex Notophyll Vine Forest in Queenslandhave produced similar variation (Greenwood,I987a; Christopheland Greenwood,1989)and so it is likely that the differencesobservedin foliar physiognomy(Figure 7.6) are due to taphonomic,not ecologic,factors.It can be concluded,however,that within this singlelens the sameflora recursin separatelamina. At the Anglesealocality severallensesare fossiliferous.Plant remainsin most of the lensesare typically mummifiedand can be extractedfrom the matrix intact by bulk disaggregation of the sediment.Each separateclay lens hasbeenfound to containa characteristic macrofloraand, to a lesserdegree, a characteristicmicroflora (Christophelet al., 1987). Overall macrofossil diversity is high with over 100 leaf and flower/fruit taxa recognizedfor the whole flora, althoughindividual lensestypically have much lower diversity with 15-20 taxa, with a single lens containing50 taxa. The analysisof the Golden Grove clay-lensmacroflora (Barrett and Christophel, 1990; and above)suggeststhat the macrofloraof eachindividualclay lens at Anglesea representsa discreteplant communitygrowingwithin very closeproximity to eachof the channelinfill deposits(Christophelet al., 1987). The floristic variation between the separate clay-lens macrofloras at Angleseahas been interpretedas reflectingthe original vegetationalmosaic of the floodplain(Christophelet a|.,1987;Christopheland Greenwood,1988; 1989).In modern mesicmesothermalenvironments(suchas subtropicaland tropicalrainforests),a mosaicof plant communitiesis often observed,reflect- ing edaphicallyand successionally controls on speciesmembership.Each separatelens representsa different biofacies(see, for example, Taggert, 1988;Burnham, 1989).If the macroflorasof eachclay lenshad beenlumped together(as a singleAnglesea'flora'), a quite different interpretationwould have been producedof the palaeovegetation of the Angleseaarea in the Eocene. SUMMARY Experiments with individual plant organs and modern vegetation have demonstratedthat the leaf-rainpotentiallycontributingto plant fossil beds reflectstreeswithin only short distancesof the area of deposition.Separate sedimentaryfaciesin fluvial, paludal and lacustrineenvironmentspreserve 164 David R. Greenwood plant macrofossilassemblageswhich reflect varying biasesin the level of transport (autochthonousto allochthonousdeposition)and hydrodynamic sorting(Figure7.5).Different vegetationtypeswithin any landscape will have a varied proportional representationin these sedimentaryfacies, reflecting proximity to depositionalsites,the modeof depositionof both plant partsand sediment,and the energy of transport. Each 'flora' presentwithin an ex- posureof particular facieswill representa subsampleof the total vegetational mosaic,in somecasesstronglybiasedtowardsindividualplant communities, in other casescontainingelementsfrom severalcommunities. In consequence of theseobservations,plant macrofossilstudiesof palaeo- vegetationmust (wherepossible)samplefrom within discretebeddingplanes and considersedimentaryfacieswhen attempting floristic reconstructionsof palaeovegetation. While the potentialsourcesof bias are great,observations of modern plant fossil sedimentaryanaloguesallows predictivemodelsto be constructedthat allow palaeovegetationreconstructionsto accountfor sedi- mentary facies,biofaciesand differential dispersal(and small-scalevariation through seasonaleffects?).Such applicationsof taphonomyare reliant on carefuland systematicstratigraphicsamplingand result in a finer resolutionof the palaeocommunity.Previous approachesof treating single plant fossil localitiesas a 'flora' must be abandonedin favour of suchan approach. ACKNOWLEDGEMENTS I thank D.C. Christophel,K.L. Johnsonand A.I. Rowett for constructive criticismson the text, and R.A. Spicer,D.K. Ferguson,C.R. Hill, D.J. Barrett and M.E. Collinsonfor many usefuldiscussions on plant taphonomy. This chapter was written with the generousprovision of facilities by the Botany Department,Universityof Adelaide. NOTE 1. Leaf size index is defined as LSI : [% microphyllspp. + 2(% notophyllspp.) + 3(% mesophyllspp.)- 1001x O.5. (seeWolfe,1979;Burnham,1989). REFERENCES Ambrose, G.J., Callen, R.A., Flint, R.B. and Lange, R.T., 1979,Eucalyptu.sfruits in stratigraphic context in Australia, Nature, 280 (572I): 387-9. Taphonomy of plant macrofossils 165 Anderson, J.M. and Swift, M.J., 1983, Decomposition in tropical forests. In S.L. Sutton, T.C. Whitmore, and A.C. Chadwick (eds), Tropical rainforest: ecology and management, British Ecological Society Special Publication, 2: 287409. Barrett, D.J. and Christophel, D.C., 1990, The spatial and temporal components of Australian Tertiary megafossil deposits. In J.G. Douglas and D.C. Christophel (eds), Proceedings of the 3rd Conference of the International Organisation of Palaeobotanists, Melbourne, 1988. Bassinger, J.F. and Christophel, D.C., 1985, Fossil flowers and leaves of the Ebenaceae from the Eocene of southern Australia, Canadian Journal of Botany, 63 (10): 1825-43. Birks, H.J.B. and Birks, H.H., 1980, Quaternary palaeoecology, Edward Arnold, London. Blackburn, D.T., 1981, Tertiary megafossil flora of Maslin Bay, South Australia: numerical taxonomic study of selected leaves, Alcheringa, 5 (1): 9-28. Brasell, H.M., LJnwin, G.L. and Stocker, G.C., 1980,The quantity, temporal distri- bution and mineral-element content of litterfall in two forest types at two sites in tropical Australia, Journal of Ecology, 68 (1): 123-39. Brenchley, P.J., 1990, Biofacies. In D.E.G. Briggs and P.R. Crowther (eds), Palaeobiology: a synthesis, Blackwell Scientific Publications, Oxford: 395-400. Briinig, E.F., 1983, Vegetation structure and growth. In F.B. Golley (ed.), Tropical rainforest ecosystems: structure and function, Ecosystems of the world, l4L, Elsevier, Amsterdam: 49-76. Burnham, R.J., 1989, Relationships between standing vegetation and leaf litter in a paratropical forest: implications for paleobotany, Review of Palaeobotany and Palynology, 58 (1): 5-32. Burnham, R. J., 1990, Some Late Eocene depositional environments of the coal- bearing Puget Group of western Washington, International Journal of Coal Geology,lS (1): 27-51. Burnham, R.J. and Spicer, R.A., 1986, Forest litter preserved by volcanic activity at El Chich6n, Mexico: a potentially accurate record of the pre-eruption vegetation, Palaios, | (2): 158-61. Cameron, C.C., Esterle, J.S. and Palmer, C.A., 1989, The geology, botany and chemistry of selected peat-forming environments from temperate and tropical latitudes. In P.C. Lyons and B. Alpern (eds), Peat and coal: origin, facies, and depositional models, International Journal of Coal Geology, 12: 89-103. Carpenter, R.J. and Horowitz,P., 1988, Leaf litter in two southern Tasmanian creeks and its relevance to palaeobotany, Papers and Proceedings of the Royal Society of Tasmania, 122 (2): 3945. Chaney, R.W., 1924, Quantitative studies of the Bridge Creek flora, American Journal of Science, 8 (2): 12744. Chaney, R.W. and Sanborn, E.I. , 1933, The Goshen flora of west-central Oregon, Contributions to Paleontology, Carnegie Institute of Washington, Publication 439: 1-103. Christophel, D.C., L98L, Tertiary megafossil floras as indicators of floristic associ- ations and palaeoclimate. In A. Keast (ed.), Ecological biogeography of Australia, Junk, The Hague: 379-90. Christophel, D.C., L984, Early Tertiary Proteaceae: the first floral evidence for the Musgraveinae, Austalian Journal of Botany,32 (2): 177-86. Christophel, D.C., 1988, Evolution of the Australian flora through the Tertiary, Plant Systematicsand Evolution, 162 (1): 63-78. 166 David R. Greenwood Christophel, D.C. and Greenwood, D.R. , 1987, A megafossil flora from the Eocene of Golden Grove, South Australia, Transactions of the Royal Society of South Australia, ll1 (3): 15542. Christophel, D.C. and Greenwood, D.R., 1988, A comparison of Australian tropical rainforest and Tertiary fossil leaf-beds. In R. Kitching (ed.), The ecology of Australia's wet tropics, Proceedings of the Ecological Society of Australia, lS: 139-48. Christophel, D.C. and Greenwood, D.R., L989, Changes in climate and vegetation in Australia during the Tertiary, Review of Palaeobotany and Palynology, SS (2): 95-109. Christophel, D.C., Harris, W.K. and Syber, A.K.,1987, The Eocene flora of the Anglesea locality, Victoria, Alcheringa, ll (2): 303-23. Collinson, J.D., 1986, Alluvial sediments. In H.G. Reading (ed.), Sedimentary en- vironments and facies (2nd edn), Blackwell Scientific Publications, Oxford:2042. Collinson, M.E., 1983, Accumulations of fruits and seeds in three small sedimentary environments in southern England and their palaeoecological significance, Annals of Botany,52 (4): 583-92. Collinson, M.E., 1988, Freshwater macrophytes in palaeolimnology, Review of Palaeobotany and Palynology, 52 (4): 31742. Dolph, G.E. and Dilcher, D.L., 1979, Foliar physiognomic analysis as an aid in determining paleoclim ate, P alaeonto grap hica, Bl7 0 : 151.:72. Drake, H. and Burrows, C.J., 1980, The influx of potential macrofossils into Lady Lake, north Westland, New Zealand, New Zealand Journal of Botany, 18 (2): 257-74. Efremov, LA., 1940, Taphonomy: a new branch of paleontology, Pan American Geologist, 7a Q): 81-93. Ferguson, D.K.,1977, The Miocene flora of Kreuzau, Western Germany: 1. The leaf remains, North Holland, Amsterdam. Ferguson, D.K., 1985, The origin of leaf-assemblages-new light on an old problem, Review of Palaeobotany and Palynology, a6 Ql2): 117-88. Flores, R.M., 1981, Coal deposition in fluvial palaeoenvironments of the Paleocene Tongue River Member of the Fort Union Formation, Powder River area, Wyoming and Montana. In F.G. Ethridge and R.M. Flores (eds), Modern and ancient nonmarine depositional environments, Special Publication of the Society of Economic Paleontologists and Mineraloglsrs, 31: 169-90. Frakes, L.A. and Francis, J.E., 1990, Cretaceous palaeoclimates. In R.N. Ginsburg and B. Beaudoin (eds), Cretaceous resources, events and rhythms, Kluwer Academic Publishers, The Netherlands: 273-87. Gastaldo, R.A., 1986, Implications on the paleoecology of autochthonous Carboniferous lycopods in clastic sedimentary environments, Palaeogeography, Palaeoclimatology, P alaeoecology, 53 Qa) : Dl-2I2. Gastaldo, R.A., 1988, Conspectusof phytotaphonomy. In W.A. DiMichelle and S.L. Wing (eds), Methods and applications of plant paleoecology, Paleontological Society Special Publication, 3: 14-28. Gastaldo, R.A., Douglass, D.P. and McCarroll, S.M.,1987, Origin, characteristics and provenance of plant macrodetritus in a Holocene crevassesplay, Mobile delta, Alabama, Palaios, 2 (3): 22940. GreatRex, P.A., 1983, Interpretation of macrofossil assemblagesfrom surface sam- pling of macroscopic plant remains in mire communities, Journal of Ecology, Tl (3\:773-91. Taphonomy of plant macrofossils 167 Greenwood, D.R., 1987a, The foliar physiognomic analysis and taphonomy of leaf- beds derived from modern Australian Rainforest, Ph.D. thesis, University of Adelaide. Greenwood, D.R., 1987b, Early Tertiary Podocarpaceae: megafossils from the Eocene Anglesea locality, Victoria, Australia, Australian Journal of Botany, 35 (2):rrL43. Greenwood, D.R., Callen, R.A. and Alley, N.F., 1990, Tertiary macrofloras and Tertiary stratigraphy of Poole Creek palaeochannel, Lake Eyre Basin, Abstracts, 10th Australian Geological Congress, Hobart. Heath, G.W. and Arnold, M.K., 1966, Studies in leaf-litter breakdown. II. 'sun' and 'shade' leaves, Pedobiologia,6 Breakdown rate of Ql$:23843. Hickey, L.J., 1980, Paleocenestratigraphy and flora of the Clark's Fork basin. In P.D. Gingerich (ed.), Early Cenozoic paleontology and statigraphy of the Bighorn Basin, Wyoming, University of Michigan Papers in Paleontology,24: 3349. Hill, R.S. and Gibsotr, N., 1986, Distribution of potential macrofossils in Lake Dobson, south central Tasmania, Australia, Journal of Ecology, T4 (2): 373-84. Hill, R.S. and Macphail, M.K., 1983, Reconstruction of the Oligocene vegetation at Pioneer, northeast Tasmania, Alcheringa, 7 (2): 28"1.-99. Kemp, E.M., L972, Reworked palynomorphs from the west Ice Shelf area, east Antarctica, and their possible geological and palaeoclimatological significance, Marine Geology, f3 (3): 145-57. Knoll, A., 1985, Exceptional preservation of photosynthetic organisms in silicified carbonates and silicified peats. In H.B. Whittington and S.C. Morris (eds), Extraordinary fossil biotas: their ecological and evolutionary significance, Philosophical Transactions of the Royal Society of London, 8311: lll-22. Kovach, W.L. and Dilcher, D.L. ,1984, Dispersed cuticles from the Eocene of North America, Botanical Journal of the Linnean Society, SS (1): 63-104. Lange, R.T., 1970, The Maslin Bay flora, South Australia, 2. The assemblage of fossils, Neues Jahrbuch ftir Geologie und Paliiontologie Monatshefte, 1970 (8): 48G90. 'germlings', their living equivalents and their Lange, R.T., 1976, Fossil epiphyllous palaeohabitat value, Neues Jahrbuch ftir Geologie und Paliiontologie Abhand- lungen,lSf (2): L4245. Lange, R.T., 1978, Southern Australian Tertiary epiphyllous fungi, modern equival- ents in the Australasian region, and habitat indicator value, Canadian Journal of Botany,56 (5): 5324I. Lange, R.T., 1982, Australian Tertiary vegetation, evidence and interpretation. In J.M.G. Smith (ed.), A history of Austalasian vegetation, McGraw-Hill, Sydney: 44-89. Luly, J., Sluiter, I.R. and Kershaw, A.P., 1980, Pollen studies of Tertiary brown coals: preliminary analyses of lithotypes within the Latrobe Valley, Victoria, Monash Publications in Geography,23: l-78. Macphail, M.K. and Hope, G.S., 1985, Late Holocene mire development in montane southeastern Australia: a sensitive climatic indicator, Search, 15 (11-12) : 344-8. Martin, H.A., t978, Evolution of the Australian flora and vegetation through the Tertiary: evidence from pollen, Alcheringa,2 (2): l8l-202. Martin, H.A., 1981, The Tertiary flora. In A. Keast (ed.), Ecological biogeography of Australia, Junk, The Hague: 391406. Martin, H.A. , 1982, Changing Cenozoic barriers and the Australian palaeobotanical record, Annals of the Missouri Botanical Gardens,69: 62547. 168 David R. Greenwood McQueen, D.R. ,1969, Macroscopic plant remains in Recent lake sediments, Tuatara, 17 (1): 13-19. Moore, P.D., 1989, The ecology of peat-forming processes:a review. In P.C. Lyons and B. Alpern (eds), Peat and coal: origin, facies, and depositional models, International Journal of Coal Geology, 12: 89-103. Moore, R.C., 1949, Meaning of facies, Geological Society of America Memoir,38: r-34. O'Dowd, D.J., Brew, C.R., Christophel, D.C. and Norton, R.A., (in press), Evidence for 40 million year-old mite-plant associations from the Eocene of southern Australia, Science. Peters, M.D. and Christophel, D.C., 1978,Austrosequoia wintonensis,a new taxodia- ceous cone from Queensland, Australia, Canadian Journal of Botany, 56 (24): 3tr948. Potter, F.W. and Dilcher, D.L., 1980, Biostratigraphic analyses of Eocene clay deposits in Henry County, Tennessee. In D.L. Dilcher and T.N. Taylor (eds), Biostratigraphy of fossil plants, successionaland paleoecological analyses, Dowden, Hutchinson and Ross; Pennsylvania: 2ll-25. Raymond, A., L987, Interpreting ancient swamp communities: can we see the forest in the peat?, Review of Palaeobotany and Palynology, 52 (213):217-31. Reading, H.G., 1.986,Facies. In H.G. Reading (ed.), Sedimentary environments and facies (2nd edn), Blackwell Scientific, Oxford: 4-19. Rogers, R.W. and Barnes, A., 1986, Leaf demography of the rainforest shrub Wilkiea macrophylla and its implications for the ecology of foliicolous lichens, Australian Journal of Ecology, 11 (4): 3414. Roth, J.L. and Dilcher, D.L., t978, Some considerations in leaf size and margin analysis of fossil leaves, Courier Forschungsinstitut Senckenberg,30: 165:71. Scheihing, M. and Pfefferkorn, H. , L984, The taphonomy of land plants in the Orinoco delta: a model for the incorporation of plant parts in clastic sediments of Late Carboniferous Age of Euramerica, Review of Palaeobotany and Palynology, ar Ql$:20s40. Schopf, J.M., L975, Modes of fossil preservation, Review of Palaeobotany and Palynology, 20 (1): 27-53. Scott, A.C., 1990, Anatomical preservation of plants. In D.E.G. Briggs and P.R. Crowther (eds), Palaeobiology: a synthesis, Blackwell Scientific, Oxford: 2634. Seilacher, A., 1990, Taphonomy of Fossil-Lagerstiitten. In D.E.G. Briggs and P.R. Crowther (eds), Palaeobiology: a synthesis,Blackwell Scientific, Oxford: 26G70. Spain, A.V. ,1984, Litterfall and the standing crop of litter in three tropical Australian Rainforest s, J ournal of Ecology, 72 (3) : 9474I. Spicer, R.A., 1980, The importance of depositional sorting to the biostratigraphy of plant megafossils. In D.L. Dilcher and T.N. Taylor (eds), Biostratigraphy of fossil plants, successional and paleoecological analyses, Dowden, Hutchinson and Ross, Pennsylvania: 171-83. Spicer, R.A., 1981, The sorting and deposition of allochthonous plant material in a modern environment at Silwood Lake, Silwood Park, Berkshire, England, US Geological Survey Professional Paper, 1143: I-77. Spicer, R.A., 1988, Quantitative sampling of plant megafossil assemblages.In W.A. DiMichelle and S.L. Wing (eds), Methods and applications of plant paleoecology, Paleontological So ciety Sp ecial Publication, 3: 29-5I . Spicer, R.A., 1989, The formation and interpretation of plant fossil assemblages, Advances in Botanical Research, 16: 96-191' - Taphonomy of plant macrofossils 169 Spicer, R.A., 1990, Fossils as environmental indicators: climate from plants. In D.E.G. Briggs and P.R. Crowther (eds), Palaeobiology: a synthesis, Blackwell Scientific, Oxford: 401-3. Spicer, R.A. and Wolfe, J.A., 1987, Plant taphonomy of late Holocene deposits in Trinity (Clair Engle) Lake, northern California, Paleobiology,13 (2):227a5. Stocker, G.C., Thompson, W.A., Irvine, A.K., Fitzsimon, J.D. and Thomas, P.R., (in press), Rainforest litterfall in tropical Australia. 1. Annual and site patterns, Journal of Ecology. Taggert, R.E., 1988, The effect of vegetation heterogeneity on short stratigraphic sequences.In W.A. DiMichelle and S.L. Wing (eds), Methods and applications of plant paleoecology, Paleontological Society Special Publication,3: I47-71. Thomas, B.A., 1990, Rules of nomenclature: disarticulated plant fossils. In D.E.G. Briggs and P.R. Crowther (eds), Palaeobiology: a synthesis, Blackwell Scientific, Oxford: 42I-3. Thomas, B.A. and Spicer, R.A. , 1987, The evolution and palaeobiology of land plants, Croom Helm. London. Truswell, E.M. and Harris, W.K. , L982,The Cainozoic palaeobotanical record in arid Australia: fossil evidence for the origin of an arid-adapted flora. In W.R. Barker and P.J.M. Greenslade (eds), Evolution of the flora and fauna of arid Australia. Peacock Publications. South Australia: 67-76. Upchurch, G.R., Jr., 1989, Terrestrial environmental changesand extinction patterns at the Cretaceous-Tertiary boundary, North America. In S.K. Donovan (ed.), Mass extinctions: processesand evidence, Belhaven Press, London: 195-216. Webb, L.J., 1959, A physiognomic classificationof Australian rainforests, Journal of Ecology, a7 Q):551-70. Wilson, M.V.H., 1980, Eocene lake environments: depth and distance from shore variation in fish, insect, and plant assemblages, Palaeogeography, Palaeoclimatolo gy, P alaeoecology, 32 (I) : 2I4a. Wing, S.L., 1984, Relation of paleovegetation to geometry and cyclicity of some fluvial carbonaceous deposits, Journal of Sedimentary Petology,54 (1): 5246. Wing, S.L., 1987, Eocene and Oligocene floras and vegetation of the Rocky Mountains , Annals of the Missouri Botanical Gardens, T4:748-84. Wing, S.L., 1988, Depositional environments of plant bearing sediments. In W.A. DiMichelle and S.L. Wing (eds), Methods and applications of plant paleoecology, Paleontological Society Special Publication, 3; l-ll. Wolfe, J.A., 1979,Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions of the Northern Hemisphere and Australasia, US Geological Survey Professional Paper,1106: I-37. Wolfe, J.A., 1.985,The distribution of major vegetational types during the Tertiary. In E.T. Sundquist and W.S. Broecker (eds), The carbon cycle and atmospheric CO2, natural variations Archaen to present, American Geophysical Union, Geophysical Monograph, 32: 357:75. Wolfe, J.A., 1990, Palaeobotanical evidence for a marked temperature increase following the Cretaceous/Tertiary boundary, Nature, y3 (6254): 153-6. Wolfe, J.A. and Schorn, H.E., 1989, Paleoecologic, paleoclimatic, and evolutionary significanceof the Oligocene Creede flora, Colorado, Paleobiology,lS (2): 180-98. Wolfe, J.A. and Upchurch, G.R. , Jr.,1987, North American nonmarine climates and vegetation during the Late Cretaceous, Palaeogeography, Palaeoclimatology, Palaeoecology, 6l (1): 33-77 .