Permain- Life Crisis on Land Author(s): G. J. Retallack Source: Science, New Series, Vol. 267, No. 5194 (Jan. 6, 1995), pp. 77-80 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/2886044 Accessed: 21/01/2010 15:50

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aaas.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org REPORTS

moved from the oocyte by digestion with collage- 19. The Ca2+ response to BKwas observed to decrease throughoutthe Bowen-Gunnedah- nase (2 mg/ml, type IA,Sigma). Two-electrode volt- substantially when multipleCE runs were performed age-clamp measurements were performed with a in a short period. basins of easternAustralia. Ironically, this TEV-200 amplifier (Dagan) in the virtual-current 20. A. B. Bush, L. A. Borden, L. A. Greene, F. R. Max- was the traditionallyrecognized - mode. IntracellularAg-AgCl electrodes were con- field, J. Neurochem. 57, 562 (1991). Triassic boundaryin Australiauntil 1970 structed with an initialinput impedance of -3 to 4 21. This result may be caused by sample dilution in the (6). At that time, correlationof palyno- megohms through 3 M KCIsingle-pull pipettes. The capillary-SCB transfer region. oocyte bath solution for measurements contained 22. T. B. L. Kist, C. Termignoni, H. P. H. Grieneisen, morph assemblagesfrom Australia with 140 mM NaCI, 2 mM KCI,2 mM CaCI2,and 10 mM Braz. J. Med. Biol. Res. 27,11 (1994); R. D. Smith, J. those of Pakistanencouraged the view that Hepes (pH 7.2). H. Wahl, D. R. Goodlett, S. A. Hofstadler, Anal. the boundarywas significantly higher in the 12. A narrow crack connecting the capillarychannel to a Chem. 65, 574A (1993). sequence:at the top of the palynozonechar- grounded electrolyte reservoir was made -10 cm 23. At gap distances less than 20 pum,damage was upstream from the capillaryoutlet with the technique sometimes apparent. Electricaland mechanical dis- acterized by Protohaploxypinusmicrocorpus described in M. C. Linharesand P. T. Kissinger [Anal. tress both may contribute to cell damage when very (7). Recent chemostratigraphicstudies of Chem. 63, 2076 (1991)]. small gap distances are used. boreholesin the CanningBasin of Western 13. T. Pozzan, F. DiVirgilio,L. M. Vicentini,J. Meldolesi, 24. "Grounding"through a crack in the capillarycreates B. M. E. Biochem. J. 234, 547 (1986); S. Sands and a voltage splitter, in which the relative current Australiahave demonstratedthat 13C/12C Fasolato, P. Barish, Brain Res. 560, 38 (1991); C. through the crack and to the end of the capillary is ratios in kerogen of marine became T. J. Biol. Chem. 265, 20351 (1990). Piz7o, Poz7an, determined by the relative resistances of the two interacted abruptly lower at the Permian-Triassic 14. Divalent cations in the separation buffer paths. Although most of the current passes through surface, causing a slow decrease in boundary(8). This dramatic isotopic ex- with the capillary the crack, some residual current flows through the was the electroosmotic flow rate until equilibrium capillaryoutlet and interferes with the voltage-clamp cursion is characteristicof numerousma- reached [see K. Yamamoto, S. Suzuki, M. Ueda, K. measurement. throughthe Permian-Triassic Kakehi, J. Chromatogr. 588, 327 (1991)]. This de- rine sections 25. M. Moriet aL, J. Biol. Chem. 263, 14574 (1988). crease reduced the velocity of a slow species by a boundary(9) and has been recognizedalso greater relativeamount than the velocity of a fast spe- 26. T. Jackson, Pharmacol. Ther. 50, 425 (1991). within nonmarine sequences of the Coo- 27. C. A. Briggs et al., Br. J. Pharmacol. 104, 1038 cies (assuming migrationvelocities remain positive). per, Bowen, and Sydney basins at the base 15. A 50-ml culture flask of PC-12 cells,differentiatedfor (1991). 4 days in medium containing P-NGF (25 ng/ml) was 28. E. M. DeLorme and R. McGee Jr., J. Neurochem. of the P. microcorpuspalynozone (10). pelleted, and cells were lysed in -70 ,ul of 70% 50,1248 (1988). The transitionfrom Glossopteristo Di- methanol containing -0.1 mM eserine to inhibitcho- 29. L. A. Greene, M. M. Sobeih, K. K. Teng, in Culturing croidiumfloras is abruptand profound.Only linesterase activity. Nerve Cells, G. Banker and K. Goslin, Eds. (MIT 16. With the use of the Hagen-Poiseuille equation and Press, Cambridge, MA, 1991), pp. 207-226. four genera and one species of megafossil estimations of diffusion and spontaneous fluid dis- 30. We thank L. Stryer and R. Dadoo for stimulating plants are known to have survived the placement [see H. A. Fishman et a/, Anal. Chem. 66, discussions, and Z.-J. Xu, S.-C. Hsu, C. Sims, and boundaryin the SydneyBasin (1 1), an ex- 2318 (1994)], the injectionvolume was calculated to R. Schneevis for technical assistance. We gratefully be -500 pl. This volume corresponds to -10-5 of acknowledge the laboratories of E. Shooter and D. tinction of 97% of Late Permianfossil leaf the entire lysate volume. Cells were differentiated at Julius for the donation of PC-12 cells and 5HT1c species. Leaves of Glossopterishave been -10% confluence (that is, -5 x 106 cells were in- complementary DNA, respectively. J.B.S. is a found in claystone partingsof the upper- corporated into the lysate), so that the sample vol- Howard Hughes Predoctoral Fellow, and H.A.F. is a ume contained the equivalent of -50 cells. W. R. Grace Fellow. Supported by grants from the most (or Bulli) coal only 19 cm below 17. L. A. Greene and G. Rein, Nature 268, 349 (1977). National Instituteof Mental Health (MH45423-03 and shalesbearing Dicroidium callipteroides (12). 18. M. D. Greaney, D. L. Marshall, B. A. Bailey, I. N. MH45324-05) and Beckman Instruments. I restrictedmy analysisto fossil leaves (Fig. Acworth, J. Chromatogr. Biomed. Appl. 622, 125 to avoid of namesfor fructi- (1993). 2 September 1994; accepted 1 November 1994 1) duplication fications (13), but the Permian-Triassiccri- sis also curtailedthe Late Permianevolu- tionary adaptive radiation of glossopterid on Land fructificationsin Gondwana(14). Permian-Triassic Life Crisis Fossil plants replacing the Glossopteris G. J. Retallack flora were low in diversity (Fig. 1). Al- thoughthe zonalindicator is the distinctive seed fernD. callipteroides,many assemblages Recent advances in radiometric dating and isotopic stratigraphy have resulted in a are dominated by the conifer Voltziopsis different placement of the Permian-Triassic boundary within the sedimentary sequence (15) or the lycopod Cylomeia(16). These of the of southeastern Australia. This boundary at 251 million years ago was voltzialean conifers and small Isoetes-like a time of abrupt decline in both diversity and provincialism of floras in southeastern lycopodsare most closely allied to Eurasi- Australia and extinction-of the Glossopteris flora. Early Triassic vegetation was low in atic EarlyTriassic genera such as Annalepis, diversity and dominated by lycopods and voltzialean conifers. The seed fern Dicroidium Tomiostrobus,and Voltzia(17). Both Voltzi- appeared in the wake of Permian-Triassic boundary floral reorganization, but floras opsisand Cylomeiapersist into diverselater dominated by Dicroidium did not attain Permian levels of diversity and provinciality until floras dominated by the Gondwanan en- the Middle Triassic (244 million years ago). demic seed fern, Dicroidiumzuberi. Diversi- fication of Gondwananseed ferns contin- ued with the appearancein the Middle Triassicof Dicroidiumodontopteroides, a bio- The Permian-Triassicboundary has long for reassessingPermian-Triassic boundary stratigraphicevent that has been dated in been known as a majordiscontinuity in the events on land. High-precision 206Pb/238U New Zealandat about 244 Ma (18). The historyof life in the sea (1 ), but comparably radiometricdating of zirconsfrom a tuff at seed fern D. odontopteroidesis a prominent severeextinctions have not been apparent the Permian-Triassicboundary in marine element of diverse fossil floras that show fromrecent assessmentsof the fossil record sequences of China at 251 ? 3.4 (2u) regional differentiationthroughout south- of land plants (2) or animals (3). Geo- million years ago (Ma) (4), supportedby ern Pangaea(19). In Eurasiasimilarly, ex- chemicalapproaches to the vexing problem dating of tuffs in coal measures of the tinction of ruflorianand voynovskyancor- of correlationbetween marineand nonma- Gunnedahand Sydneybasins of New South daites was followed by an interregnumof rine biostratigraphicschemes give grounds Wales (5), now indicatesthat the Permian- conifersand lycopods,which weresupplant- Triassic boundaryis near the contact be- ed by diverse Middle Triassic floras domi- Department of Geological Sciences, University of Ore- tween coal measureswith Glossopterisand nated by the seed fern Scytophyllum(17). gon, Eugene, OR 97403-1272, USA. overlying fluvial deposits with Dicroidium The megafossil plant record of Permian-

SCIENCE * VOL. 267 * 6 JANUARY 1995 77 Triassic plant extinctions in the Sydney oldest preservedTriassic rocks in the basin, mian-Triassicboundary sequences world- Basinis supportedby basinwidepalynologi- as determined by sequence-stratigraphic wide (24, 25). cal data for each palynozone (20, 21). A correlationof onlap relations(23). Palyno- In addition, fungal, fern, and lycopod patternof declining then recoveringdiver- logicalcounts show dramatic floral overturn sporesshow transienthigh abundanceim- sity is once againapparent, with a clustering and a sharp peak of last appearancesbe- mediately above the Permian-Triassic of first and last appearancesnear the Per- tween the uppermostcoal and the first few boundaryin the LakeMunmorah core (26). mian-Triassicboundary (Fig. 1). In addi- meters of overlying sandstone and shale. Abundant fungal spores may represent a tion, EarlyTriassic palynozones are found The loss of about 30% of pollen and spore periodof plant destructionand decay,with throughoutAustralia, but by MiddleTrias- speciesreflects extinction of the peat-form- lycopodand fern sporesindicating regener- sic time there wasprovincial differentiation ing flora at the Permian-Triassicboundary. ation of herbaceousplants after catastrophe, of eastern and western Australianpalyno- Striatebisaccate pollen known to belong to as inferredfor the Permian-Triassicbound- floras(22). There are indicationsof greater a varietyof Permianglossopterids and Per- ary in other parts of the world (25, 27). continuityof species throughthe Late Per- mian-Triassicconifers (24) are displacedby Similar interpretationshave been made of mian to Middle Triassic interval from pa- an increasedabundance of nonstriatebisac- the better known "fernspike" at the Cre- lynologicaldata than from megafossildata. cate pollen. Of these, the most common is taceous-Tertiaryboundary (28). This may be due in part to recycling of Falcisporites,which was produced by Di- It could be arguedthat the abruptde- palynomorphs,which lessensthe severityof croidium,among other plants (24). This dra- cline then protractedrecovery in diversity extinctions to only 19% of Permianpaly- matic reorganizationof the dominantpol- of fossilfloras above the coal measuresis an nomorphspecies not found also in earliest len producers matches the megafossil artifact of an erosional disconformitybe- Triassicrocks. record. tween Permiancoal measuresand Triassic The floralcrisis is also blurredby coarse A strikingfeature of these high-resolu- deposits.However, the transienthigh abun- temporal resolution of the palynozones. tion palynologicaldata (Fig. 2) is the abun- dances of fern and lycopod spores and of Short-termchanges in vegetationacross the dance of acritarchsin samplesimmediately acritarchsis characteristicof the Permian- Permian-Triassicboundary are best seen above the Permian-Triassic boundary. Triassicboundary in high-resolutionmarine from counts of palynomorphabundance in These presumedalgal cysts represent a brief sequences(24-27, 29). The uppermostcoal samples of a single drill core from Lake episode of marine influence in a predomi- seam in the Sydney Basin is little eroded Munmorah,New South Wales (Fig.2). The nantly nonmarine sequence. Acantho- awayfrom local tectonic structuressuch as core includes the uppermost Permian morphacritarch swarms and abruptmarine the LochinvarDome (23) and can be cor- (Vales Point) coal and is overlain by the transgressionare found also in other Per- related over distancesof 200 km. The up-

First Last Diest and affinities Appeamnces ~~~~~~~~~~~~~~~~Diversity Age Rockunk Palynomorphzones (leaf species) (pollenand spore spp.) Megafossilplants 40 80 0 20 3 ______~~~~~~~~1020 10 20 BrinellySha Dicroidium t odontopteroides ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Other\.plants * E~Sed 244 MinchinburyBringellyShale Sandst. D ~ ~ ~~ ~~~~~~~~ .~~ J1 4fen zon Ma 4 ~~~~~~~~~Aratrisporites cm parvispinosusfic AshfieldShale Dicroidium C zbr Hawkesbury -;, I zober 4 Sandstone I ,Z. ~~~~elongatum ~~ ~~~~~~Taeniopteris~ ~ ~ . Newor io is

Blentriculifor JqGarieBald Hillsprie Formation ~.~Perml Bald HillClaystone l~~~~~~~~~~~~onegicau

4bConifers *44 StnwlSaBorulgh PakC t . groshapoyiLes Voltziopsis oifr

WombarraShale ~~~~~~~~~~~~~~~Dicroidium Lycopods Sandstone microcorPus callipteroideIszone 21-CoalcliffMa ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~BulliCoal Dulhuntyispora 9Gosotsens'rtir EckersleyFormation parvithola cmossopt4ri

Fig. 1. Changes in fossil floras across the Permian-Triassic boundary within first and last appearances, whereas only standing diversity and taxonomic the Sydney Basin, Australia, showing abrupt extinction of the peat-forming composition have been computed from megafossil data (13). Events recently Glossopteris flora and replacement by a low-diversity conifer-lycopod flora redated radiometricallyinclude the Permian-Triassic boundary (4) and first that was later supplanted by adaptive radiation of the Dicroidium flora. Pa- appearance of Dicroidium odontopteroides (18). Bar scales for pollen and lynological data (20, 21) have been used to calculate standing diversity and spores are 20 p.m and for plant fossils are 1 cm.

78 SCIENCE * VOL. 267 * 6 JANUARY 1995 REPORTS permostcoal is onlappedfrom the north by robustbut compromisedby taphonomicbi- saurusand Cynognathusfaunas of South Af- six fluvialparasequences, which can be tak- ases. Lacustrineshales with are un- rica and Antarctica(40). en as evidence of about 6 X 105 years of common in this dominantly fluvial se- In summary,terrestrial ecosystems suf- earliestTriassic nondeposition in the south- quence, but include the Late PermianBel- fered markedlosses in diversityat the Per- ern coalfield that is representedby rock in mont fauna with 145 species (32), a single mian-Triassicboundary, after which a cos- the northern coalfield (23). Around Lake wing from earliestTriassic of mopolitan flora of lycopods and conifers Munmorahin the north, unusuallythick the BirthdayShaft near Sydney (33), the prevailed as new groupsof endemic seed fluvial conglomeratesof Permianand Tri- MiddleTriassic () Brookvalefauna ferns underwentan evolutionaryadaptive assic age are evidence of rapid sedimenta- with 15 species (34), and the later Middle radiationto dominanceof regionallydiffer- tion (23) and thus a high probabilityof Triassic(Ladinian) Glenlee fauna with 12 entiated floras by Middle Triassic time. stratigraphiccompleteness (30). A paleosol species(35). The compositionof these local This culminatedin the Dicroidiumflora of developed on the uppermostcoal at Wy- faunas is compatible with recent global southernPangea (13) and the Scytophyllum bung Head on the coast near Lake Mun- compilations(36) indicatingthat the Per- floraof northernPangea (17). The florasof morah includesrelict beddingexpressed to mian-Triassic boundary was a time of the EarlyTriassic interregnum in the Syd- a degree compatible with a hiatus of no markedextinctions and modernizationof ney Basin included needle-leavedconifers more than 104 years by comparison with insect faunas. Conchostracansare repre- and herbaceousaquatic lycopods in oligo- surfacesoils of known exposureage (31). sented by 25 species in the Belmont insect trophic forest and marsh assemblagesvery Resolutionof the Permian-Triassicbound- beds, but only four in Early and Middle different from the deciduous seed fern ary in this northeasternpart of the Sydney Triassicrocks (37). This low diversityis not swampsof the LatePermian and wet forests Basin is probablyof the orderof 105 years. due to scarcity,because conchostracans are of the Middle Triassic(13). EarlyTriassic It could also be arguedthat the apparent locally abundant in Early Triassic rocks. oligotrophiclycopod-conifer vegetation can catastrophicextinction of the Glossopteris These fossil arthropodlocalities are not be recognizedalso in Europe,Russia, and florain the SydneyBasin is compromisedby spacedsufficiently close stratigraphicallyto China (17). The Lystrosaurusfauna is sim- poor preservationof plant fossilsin Triassic reveal the abruptnessof faunal overturn ilarlywidespread and low in diversitycom- red beds, recycling of Permian palyno- from Permianto Triassic. pared with Permian therapsidfaunas and morphsinto Triassicdeposits, and preferen- Fossil land vertebratesare most abun- Middle to LateTriassic faunas tial paleobotanicalattention to fossil floras dantlypreserved in calcareousred paleosols (40). The Permian-Triassiclife crisis thus of large coastal exposuresof Permiancoal of the kindforming in dryclimates (30), but usheredin an EarlyTriassic terrestrial eco- measuresnear Newcastle and TriassicNew- paleosolsof the Sydney Basin are noncal- systemthat was lower in diversityand glo- port Formation of the northern Sydney careousand deeplyweathered, indicative of bally more homogeneousthan ecosystems coastal cliffs. These argumentslose force a humid climate (38) in which bones are of the Late Permianand MiddleTriassic. given the consistencyof the palynological dissolvedby soil acids.This mayaccount for and megafossilrecord of floral change, de- the poverty of vertebrate remains other REFERENCESAND NOTES spite the differentconstraints on preserva- than and aquatic labyrinthodontsin 1. D. H. Erwin, The Great Paleozoic Crisis (Columbia tion of leaves versuspollen (30). There are the SydneyBasin. There arefossil footprints Univ. Press, New York, 1993). carbonaceousshales with abundant fossil of a variety of terrestrialreptiles (39), but 2. A. H. Knoll, in Extinctions, M. Nitecki, Ed. (Univ. of plants at many levels within earliestTrias- identifiableskeletal remains have yet to be Chicago Press, Chicago, IL, 1986), pp. 23-68. sic rocks exposed in coastal exposuresthat reported.Fragmentary of 3. W. D. Maxwell, Palaeontology 35, 571 (1992). remains a procolo- 4. J. C. Claou6-Long, Z.-C. Zhang, G.-G. Ma, S.-H. Du, are the equal of those of both the coal phonid and a kannemeyeriddicynodont Earth Planet. Sci. Lett. 105, 182 (1991). measuresand NewportFormation. from the Bowen Basin of Queenslandpro- 5. J. J. Veevers, P. J. Conaghan, S. E. Shaw, Geol. In contrast,the fossil recordof nonma- vide evidence that EarlyTriassic terrestrial Soc. Am. Spec. Pap. 288,187 (1994). 6. T. W. E. David, The Geology of the Commonwealth rine animalsin the Sydney Basin is not so faunasof Australiawere similarto Lystro- of Australia (Arnold,Sydney, 1950). 7. B. E. Balme, in StratigraphicBoundary Problems, B. Kummel and C. Teichert, Eds. (Univ. of Kansas Press, Lawrence, 1970), pp. 305-453. Last First Standing 8. R. J. Morante, in New England Orogen, Eastern Aus- Sequence Appearances diversty tralia, P. G. Flood and J. C. Aitchison, Eds. (Univ.of Age Rock units (pollen and spore spp.) Palynomorph percent New England Press, Armidale, New South Wales, m ?og 10 20 10 20 20 60 20 40 60 80 1993), pp. 293-298. 70 9. M. Magaritz, R. V. Krishnamurthy,W. T. Holser, Am. J. Sci. 292, 727 (1992); K. Wang, H. H. J. Geld- Munmorah stri~~vJ~ cate setzer, H. R. Krouse, Geology 22, 580 (1994). Conglomerate 8 Sandstone Coal b 10. R. J. Morante, A. S. Andrew, J. J. Veevers, P. J. Hamilton, J. Aust. Pet. Explor. Assoc. 34, 330 (1994). ig Dooralong t 44 11. Surviving genera have been taken to include Glos- Vales Point 90 W sopteris as a Lazarus taxon [W. B. K. Holmes, Geo- phytology 22, 19 (1992)], although there is meritalso Coal 4 NO4444Stlt in referringdistinctive Middle Triassic leaves of this ad soKarignan(000 tt Spores kindto Gontriglossa. Other survivingform genera are Conglomerate 100 Fv~ pole Cladophlebis, Sphenopteris, and Phyllotheca. The single survivingspecies is Schizoneura australis. Wallarah 12. The seed fern "Thinnfeldia"callipteroides has been Coal 10 referred to D. callipteroides by M. E. White [Austra- lia's Prehistoric Plants (Reed, Balgowlah, New South Lithoogicl~e :o~Conglomerate []Sandstone Coal Wales, 1984)]. 13. Data from G. J. Retallack [N.S. W. Geol. Surv. Bull. Fig. 2. Changing taxonomic COmpositionof palynoflorasacross the Permian-Triassicboundary 26, 383 (1980) and references therein],together with measured by counting palynomorphsin preparationsfrom Newvale diamonddrill hole number28, published additions by S. R. Ash [Alcheringa 3, 73 LakeMunmorah, New South Wales All are shown in (1979)] and unpublished observations of fossil plants (21). palynomorphs percentcomposition (to the in the Australian Museum from the right),but presumedmarine algal acritarchs were not includedin diversityand appearances of pollen near Glenlee and the basal Narrabeen Group at Coxs and spores (to the left). Gap railwaytunnel, New South Wales.

SCIENCE * VOL. 267 * 6 JANUARY 1995 79 14. S. McLoughlin,in Gondwana Eight, R. H. Findlay,R. trated and referred to "? cf. Cymatiosphaera" by tion in boreholesusing small dual-axis, leaf- Unrug, M. R. Banks, J. J. Veevers, Eds. (Balkema, Hilda Grebe (21) are here interpreted as fungal Rotterdam, 1993), pp. 253-264. spores. springtiltmeters (7). We also measuredthe 15. J. A. Townrow, Pap. Proc. R. Soc. Tasman. 101, 27. K. J. Hsu and J. A. McKenzie, Geol. Soc. Am. Spec. shearstrength of the till with a roughened 173 (1967). Pap. 247, 61 (1990). horizontalcylinder with conical ends that 16. I use Cylomeia only for herbaceous /soetes-like 28. D. J. Nicols and R. F. Fleming, ibid., p. 445. forms, based on specimen AMF5891 figured by M. 29. Indicator species of the latest Permian Playfordia- was draggedthrough the till. The force on E. White [Rec. Aust. Mus. 33, 724 (1981)]. spora crenulata zone of the Bowen Basin [C. B. the cylinderwas measured, and geotechnical 17. S. V. Meyen, Fundamentals of Palaeobotany (Chap- Foster, Rev. Palaeobot. Palynol. 36, 165 (1982)] are theory associated with cone-penetration man and Hall, London, 1987); K. B. Pigg, Ann. Mo. also present in the Sydney Basin, but this zone has testing was adaptedto estimate the shear Bot. Gard. 79, 589 (1992). not been widely recognized (20-22). 18. G. J. Retallack, P. Renne, D. Kimbrough, N. Mex. 30. G. J. Retallack, Paleobiology 10, 59 (1984). strengthfrom this force (8). We determined Mus. Nat. Hist. Sci. Bull. 3, 315 (1993). 31. __, Soils of the Past (Unwin-Hyman, London, surfacespeed by measuringthe down-glacier 19. J. M. Anderson and H. M. Anderson. Palaeoflora of 1 990). displacementof a stake at the site every 10 Southem Africa, Molteno Formation (Triassic), vol. 1, 32. E. F. Riek, Rec. Aust. Mus. 27, 303 (1968). Introduction, Dicroidium (Balkema, Rotterdam, 33. R. J. Tillyard,Proc. Linn.Soc. N.S.W. 43,260 (1918). min with an automateddistance meter (9). 1983). 34. K. C. McKeown, Rec. Aust. Mus. 20,31 (1937); E. F. Water pressurewas measuredcontinuously 20. R. J. Helby, thesis, University of Sydney, Sydney Riek, ibid. 22, 254 (1950); ibid. 23,161 (1954); J. W. in boreholes. (1969). Evans, J. Entomol. Soc. Queensland 2, 17 (1963); 21. H. Grebe, Rec. Geol. Surv. N.S.W. 12,125 (1970). and undescribed cockroach and dragonfly on dis- Duringthe periodof recordin 1992 (Fig. 22. R. J. Helby, R. Morgan, A. D. Partridge,Assoc. Aus- play in the AustralianMuseum. 1), a tiltmeterinserted in sedimentnear the tralas. Palaeontol. Mem. 4,1 (1987). 35. R. J. Tillyard,Proc. Linn.Soc. N.S.W. 42, 721 (191 7). ice-bed interfacerotated down-glacier a to- 23. C. Herbert in Proceedings of the 27th Annual Meet- 36. C. C. Labandeira and J. J. Sepkoski Jr., Science ing for Advances in the Study of the Sydney Basin- 261, 310 (1993). tal of 520, correspondingto a mean shear- Newcastle, New South Wales (University of New- 37. T. Kobayashi, J. Fac. Sci. Univ. Tokyo Sect. Il 9, 1 strainrate of 25 year-1.The till thicknessat castle, New South Wales, 1993), p. 61; in New En- (1954); J. A. Webb, Alcheringa 2, 261 (1978). the site, estimatedfrom penetrationtests, gland Orogen, Eastem Australia, P. G. Flood and J. 38. G. J. Retallack, J. Geol. Soc. Aust. 24, 19 (1977). m. If the thicknesswere shear- C. Aitchison, Eds. (Univ. of New England Press, Ar- 39. L. F. Harper,Mem. Geol Surv. N.S. W 7,153 (1915); was 0.33 full midale, New South Wales, 1993), pp. 127-135. T. Naing, Proc. Symp. Adv. Stud. Sydney Basin ing uniformly,the displacementof the sur- 24. B. E. Balme, Bur. Miner. Resour. Geol. Geophys. Newcastle 27, 77 (1993). face of the glacierwould have been at least Rec., Canberra 1989/37,1 (1989). 40. R. Molnar, in Vertebrate Palaeontology of Austral- twice what was observed.The observeddis- 25. H. Visscher and W. A. Brugman, Mem. Soc. Geol. asia, P. Vickers-Rich,J. M. Monaghan, R. F. Baird,T. Ital. 34, 115 (1988); Y. Eshet, in Permo-Triassic H. Rich, Eds. (Monash Univ. Press, Melbourne, placementthus indicatesthat shearingwas Events in the Eastem Tethys, W. C. Sweet, Z.-Y. 1992), pp. 605-702. not uniformand was instead focusednear Yang, J. M. Dickins, H.-F. Yin, Eds. (Cambridge Univ. 41. Supported in part by a grant from the Australian the glaciersole, as has been both measured Press, New York, 1992). Research Council to J. J. Veevers and NSF grant 26. Fern and lycopod spores conspicuously abundant OPP 9315228 to G.J.R. C. Herbert, R. M6rante, T. (2) and predictedfrom theory (2, 10, 11). immediatelyabove the last coal, but not elsewhere in Naing, and J. J. Veevers of Macquarie University Duringtwo peaksin velocity resultingfrom the section, areApiculatisporitesbulliensis and Lund- offered invaluable comments and discussion. rain storms on 15 and 22 July, negative bladispora fibulata, respectively (21). Some of the unicellularto multicellular,thick-walled spores illus- 15 August 1994; accepted 19 October 1994 shear-strainrates were recorded (Fig. 1). Such negativevalues, which have also been observedelsewhere (7), implythat the tilt- Flow Mechanism of Glaciers on Soft Beds meter recordingthe deformationwas rotat- ing up-glacier,the oppositeof that expected. Neal R. Iverson,* Brian Hanson, Roger LeB. Hooke, During accelerationof the glacier toward these peaks, the sediment shear strength Peter Jansson tendedto decrease;it then increasedrapidly duringdeceleration of the glacier(Fig. 1). Subhourly measurements of bed deformation, bed shear strength, subglacial water pres- The glacierand till exhibitedsimilar be- sure, and surface speed at Storglaciaren, a glacier in northern Sweden, showed that the havior in 1993 (Fig. 2). Diurnalpeaks in shear-strain rates of the bed decrease during periods of high water pressure and surface surfacevelocity were coeval with maximain speed. High water pressures appear to be accompanied by a reduction in the coupling water pressure,which never exceeded the of ice with the bed that is sufficient to reduce or eliminate shearing. The instability of large ice overburdenpressure (12). The velocity ice masses may result from similar decoupling rather than from pervasive bed deforma- peaks occurredduring or slightly after till tion, as has been commonly thought. strain-rateminima, which wereusually neg- ative. The total down flow rotationof the tiltmeterwas 90, which would account for only 5% of the glacierdisplacement, even if Many glaciersare underlain,at least in and glacier surface velocity on Storgla- the full 0.35 m of till were deforminguni- part, by unlithified sediment that may de- ciaren in northern Sweden. Basal water formly.The tiltmeterwas inserted0.10 m formreadily if basalwater pressures are near pressuresvary over a wide rangeof valuesat lowerin the till layerin 1993 than in 1-992, the ice overburdenpressure. Such deforma- Storglaciaren;thus, this glaciercan be used which may account for the lower mean tion has been invoked to explain the rapid to studythe criticalrelation between water strainrate. motion and instability of modern and pressure/andbed deformationrate. It appearsfrom these data that the cou- formerice masses (1-4). To furtherstudy Storglaciarenis a well-studied,wet-based pling betweenthe ice and the till is reduced this process,we obtainedsimultaneous con- valley glacier(5). Resistivitymeasurements duringperiods of high water pressureand tinuous measurementsof bed deformation (6), penetration testing, and sampling in most rapidflow. This decouplingis presum- boreholesdrilled to the bed indicate that ably caused by uplift of the b'asalice and N. R. Iverson and R. LeB. Hooke, Department of Geology and Geophysics, University of Minnesota, Minneapolis, much of the ablationarea is underlainby a reduction of the applied shear stress. A MN 55455, USA. layer of till that is decimetersthick. We natural conjecture is that the negative B. Hanson, Center for Climatic Research, Department of studied the deformationof this till in the strainrates arise because of squeezingof the Geography, Universityof Delaware, Newark, DE 19716, USA. lower part of the ablationarea on the up- till into the zone of uplift.This wouldthick- P. Jansson, Department of Physical Geography, Stock- streamside of a gentletransverse ridge in the en the till and cause the tiltmeters, which holm University,S-106 91 Stockholm, Sweden. bed. The averageglacier thickness in this were inclined in the down flow direction, to *To whom correspondence should be addressed. area is -95 m. We measuredbed deforma- rotate up-glacier. The up-glacier rotation

80 SCIENCE * VOL.267 * 6 JANUARY1995