SEPM Society for Sedimentary Geology Paleontological Society Discovery of Fish Mortality Horizon at the K-T Boundary on Seymour Island: Re-Evaluation of Events at the End of the Cretaceous Author(s): William J. Zinsmeister Source: Journal of Paleontology, Vol. 72, No. 3 (May, 1998), pp. 556-571 Published by: Paleontological Society Stable URL: http://www.jstor.org/stable/1306654 Accessed: 21-10-2015 01:47 UTC REFERENCES Linked references are available on JSTOR for this article: http://www.jstor.org/stable/1306654?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp 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 support@jstor.org. SEPM Society for Sedimentary Geology and Paleontological Society are collaborating with JSTOR to digitize, preserve and extend access to Journal of Paleontology. http://www.jstor.org This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions JOURNAL OF PALEONTOLOGY,V. 72, NO. 3, 1998 556 G. 1945. The principlesof classificationand a classificationof mammals.AmericanMuseumof NaturalHistoryBulletin, SIMPSON, G. 85:1-350. STIRTON, R. A., AND V. L. VANDERHOOF. 1933. Osteoborus, a new genus of dogs, and its relations to Borophagus Cope. University of CaliforniaPublications,Bulletinof the Departmentof GeologicalSciences, 23:175-182. TEDFORD, R. H., M. E SKINNER, R. W. FIELDS, J. M. RENSBERGER,D. T. GALUSHA,B. E. TAYLOR,J. R. MACDONALD,AND P. WHISTLER, TORRES-ROLDAN, V., AND I. FERRUSQUfA-VILLAFRANCA. 1981. Cerdocyonsp. nov. a. (Mammalia,Carnivora)en Mexico y su significaci6nevolutiva y zoogeograficaen relaci6n a los Canidos sudamericanos.II CongressoLatino-Americande PaleontologiaAnais, 2:709-719. VANDERHOOF, V. L. 1936. Notes on the type of Borophagusdiversidens Cope. Journalof Mammalogy,17:415-416. , ANDJ. T. GREGORY. 1940. A review of the genus Aelurodon. Universityof CaliforniaPublications,Bulletin of the Departmentof Geological Sciences, 25:143-164. S. D. WEBB. 1987. Faunal succession and biochronologyof the ArikareeanthroughHemphillianinterval(late Oligocenethroughear- WEBB, S. D., AND S. C. PERRIGO. 1984. Late Cenozoic vertebrates liest Plioceneepochs)in NorthAmerica,p. 153-210. In M. O. Woodfrom Hondurasand El Salvador.Journalof VertebratePaleontology, burne,(ed.), Cenozoic Mammalsof North America:Geochronology 4:237-254. of California and Biostratigraphy. Press, Berkeley. University - , ANDZ. QIU. 1996. A new canid genus from the Pliocene of 4 NOVEMBER ACCEPTED 1997 Yushe, Shanxi Province.VertebratePalAsiatica34:27-40. J. Paleont., 72(3), 1998, pp. 556-571 Copyright ? 1998, The Paleontological Society 0022-3360/98/0072-0556$03.00 DISCOVERYOF FISH MORTALITYHORIZONAT THE K-T BOUNDARY ON SEYMOURISLAND:RE-EVALUATIONOF EVENTS AT THE END OF THE CRETACEOUS WILLIAM J. ZINSMEISTER Departmentof Earthand AtmosphericSciences,PurdueUniversity,WestLafayette,Indiana47907 ABSTRACT-The discoveryof a fishbone layerimmediatelyoverlyingthe K-Tiridiumanomalyon SeymourIsland,AntarcticPeninsula, whichmay representthe firstdocumentedmasskill associatedwith the impactevent,togetherwith new faunaldataacrosstheboundary has providednew insight into events at the end of the Cretaceous.The utilizationof a geographicalapproachand a new graphical of rangedata has revealedthat events at the end of the Cretaceouswere not instantaneous,but occurredover a finite representation periodof time. Althoughthe fish bone layer may containvictimsof the impactevent, the absenceof ammonitesin eitherthe iridiumbearinglayer or the overlyingfish layer suggeststhatthe extinctionevent at the end of the Cretaceouswas the culminationof several processesbeginningin the late Campanian.The impactwas the proverbial"strawthatbrokethe camel'sback,"leadingto the extinction of many othersforms of life that mighthave survivedthe periodof global biotic stress duringthe waning stages of the Mesozoic if therehad not been an impact.The absenceof mass extinctionfollowingcatastrophicgeologic events in a biotic robustworld,such as the Middle OrdovicianMillbrig-BigBentonitevolcanic event suggests that the biosphereis remarkablyresilientto majorgeologic catastropheswith mass extinctioneventsoccurringonly whenthereis a conjunctionof geologic eventsnone of whichmightbe capable of producinga global mass extinctionby itself. INTRODUCTION MORTALITY produced by mass kill events have been recognized in the fossil record throughout the Phanerozoic. Although the causes of mass kills vary, each resulted in the formation of a extirpation layer containing the victims of the event. A classic example is the "camposaurs bone bed" containing remains of thousands of individuals of the dinosaur, Maiasaurus, at Egg Mountain, Montana which was produced as a consequence of a vast herd of maiasaurs being overwhelmed by gases and ash from a nearby volcanic eruption (Homer and Gorman, 1988). One of the puzzling questions associated with the K-T extinction event is the absence of any type of extirpation layer at any of the well-studied boundary sections (Zinsmeister, 1993). The principal biotic data used to support the K-T extinction event is "species disappearance," which constitutes negative evidence (Zinsmeister, 1993). Although the absence of an extirpation horizon has not been addressed by proponents of the impact extinction hypothesis until recently (Cutler and Behrensmeyer, 1996), there are several possible explanations for the absence of a mortality layer at the boundary. Since most of the K-T boundary sections are finite localities with little or no significant geographic component, the chances of finding such HORIZONS evidence of a mass kill at any given boundary would be remote. Cutler and Behrensmeyer (1996, p. 375) have suggested that "... much of the terrestrial fossil record is time-averaged, any peak in bone abundance resulting from a mortality pulse is likely to be obscured by background attritional mortality.... No special explanation is needed for the absence of bones beds at the K/T boundary." To dismiss the absence of a mortality horizon at the K-T boundary to "time-averaging" ignores the existence of numerous examples of mass mortality horizons throughout the Phanerozoic which clearly demonstrates that "time-averaging" does not preclude the preservation of catastrophic kill events. Alternatively, the absence of a mortality horizon at the K-T boundary may reflect that events leading to the mass extinction at the end of the Cretaceous were not instantaneous, but occurred over a longer period of time, and, as a consequence, no single global extirpation horizon would have formed. The discovery of a layer of fish debris in close association with the iridium anomaly horizon on Seymour Island, northeast tip of the Antarctic Peninsula (Fig. 1), during the austral summer of 1994-95, may represent the first documented, direct evidence of a kill event associated with the bolide impact 65 m.y. ago (Zinsmeister, 1995). This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION FIGURE --Index mapand geologic mapof SeymourIsland.Rectangular box outlines location of the exposureof the "K-T glauconite"(see Figure 2 for detailedmap of outcropof the K-T glauconite).Major unitson SeymourIsland--, quaternary stratigraphic deposits;2, upper Eocene,La MesetaFormation;3, Paleocene,SobralFormation;4, uppermostMaastrichtianto lower Paleocene,Lopez de BertodanoFormation;and 5, upperNeogene,basalticdikes. The purpose of this paper is 1) to present a detailed description of the microstratigraphy five meter interval crossing the KT boundary and detail geologic map of the boundary sequence using the recently completed USGS topographic map (1:10,000 with 5 m contour interval) of Seymour Island, 2) to discuss the conditions leading to the formation of the fish bone layers immediately above the iridium anomaly, and lastly to examine biotic changes across the K-T boundary using spatial biostratigraphy, a new methodology for the analysis and graphical rep- resentationof biostratigraphic data.The sequentialnatureof the disappearanceof the ammonites,iridiumanomaly,and fish kill horizon on SeymourIsland provides new insight into the temporal sequence of events at the end of the Cretaceousand the effect of majorcatastropheson the Earth'sbiosphere. K-T BOUNDARY SECTION ON SEYMOUR ISLAND The Upper Cretaceous/LowerPaleocene sequence on Sey- mour Island consists of 1600 m of gently eastward, homoclinal dipping, mid-shelf clastic to inner shelf concretionary siltstones and silty sandstones of the Lopez de Bertodano Formation. Ma- cellari (1984) placed the K-T boundaryat a 50 cm layer of glauconitic sand approximately 1140 m above the base of the sequence that coincided with what he interpreted as an abrupt change in marine faunas. This glauconitic horizon was also used by Macellari as the boundary between his uppermost lithologic units (9 and 10) of the Lopez de Bertodano Formation. Because of the subtle lithologic changes within the glauconitic sand along strike and the occurrence of a second glauconitic sandy layer several meters higher in the section, locating the K-T boundary between sections with precision greater than four to five meters interval was not possible. As a consequence, the glauconitic interval between units 9 and 10 has been informally referred to by workers on Seymour Island as the "K-T glauconite" (Huber, 1988). Immediatelyunderlyingthe "K-T glauconite"is the medium to darkconcretionarysandysiltstoneof Macellari'sUnit 9. Con- cretions within the uppermost part of Unit 9 occur either as isolated spherical masses up to one meter in diameter, or as discontinuoushorizons which may extend as much as 200 m 557 along strike. The benthic fauna of Unit 9 is characterizedby a diverse high-latitudeassemblageof 77 species of molluscs and a varietyof otherinvertebratesand marinereptiles(Zinsmeister et al., 1989). The richnessof the faunadeclinesin the uppermost partof Unit 9 to 36 species with the abundanceand diversityat any given locality in the last five meters very low. The only exceptionsto this generaldeclinein abundancearethe continued occurrenceof large numbersof the serpulidRotulariaand echinoid spines. No topographicmaps of the island existed when Macellari began his work on the Cretaceousof SeymourIsland in 1981. The only map of the island availableduringhis firstseason was a crude drainagemap based on air photos. During his second field season on SeymourIsland in 1984, a draftof the first topographicmap (1:20,000 with a 10 m contourinterval)of the island was available (Brecher,1984). Although the absence of adequategroundcontrolpoints caused local problemswith accuracy,the Brechermap has been used for all subsequentfield work on SeymourIsland until 1994, when the U.S. Geological Survey completeda new GPS-basedtopographicmap (1:10,000 with 5 m contour interval). The new U.S. Geological Survey map provided an opportunityduringthe 1994-95 field season for detailedsmall-scalegeologic mappingof the island thatwas not previouslypossible. The objective of the 1994-95 field season was to constructa detailedgeologic map of the "K-T glauconite"acrossthe island (Fig. 2) and to documentany facies changesalong strike.It was quickly recognizedthat the "K-T glauconite"consists of three distinct facies (Fig. 3). Although each of the facies varies independentlylaterally,it is possible to recognize the three-fold relationshipat most places. Lower glauconite.-The lowermostfacies consists of green, medium-grained,unconsolidatedto occasionallyresistantledgeforming glauconitic sandstone.Although the maximumthickness of the Lower Glauconitereachesone meter,it occasionally disappearsfor short distances.The contactof the Lower Glauconite with the underlyingconcretionarysandy siltstonesis for the most partgradational.The outcropwhere the iridiumanomaly was documented(Elliot et al., 1994) clearlyshows a gradual transitionbetween the underlyingsandy siltstonesof the uppermost partof Unit 9 to the glauconiticrich basal partof the "KT glauconite."The bioturbatednatureof the glauconiteand the absenceof any primarystructuresindicatethatthe depositionof the LowerGlauconitewas below wave-base.The occasionaldisappearanceof the Lower Glauconite and occurrenceof fragmental weatheringmasses of glauconite, which may be glauconitic clasts, suggest that local scouring may have occurred duringsevere storm events. Irregularlyroundedconcretionsup to 50 cm in diameteroccur sporadicallywithinthe Lower Glauconite. Fossils in this unit are rare and generallylimited to the large thick-shelledbivalve, Lahillia larseni. The iridiumanomaly reportedby Elliot et al. (1994) occurs withinthis glauconitic intervaland is consideredthe markthe K-T bounedary. Fish bone layer.-Overlying the Lower Glauconiteis a 1 to 3 m layerof yellowish to tan weatheringsiltstoneto fine grained, slightly glauconitic,silty sandstonereferredto as the Fish Bone Layer because of the occurrenceof fish debris and sharkteeth. Fish materialoccursas: 1) disarticulatedbone massesin isolated, irregulardark gray to black, resistantmedium-grainedblocky sandy masses up to 10 cm in length;2) as compressedpartially articulatedindividualsin isolated sphericalconcretionswith diametersvarying between 20 and 30 cm; or 3) as isolated teeth and vertebrae.The bone debris in the blocky bone masses and concretionsconsist of black to darkreddishbrownopercula,jaw and other cranial bone elements, ray spines, teeth, and other postcranialbones. Bone materialabove the basal bone masses This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 558 JOURNAL OF PALEONTOLOGY,V. 72, NO. 3, 1998 A Scale: 1:10 000 0 100 200 300 400 50 Meters map of the "K-T glauconite"from Cross Valley to the southerncoast of SeymourIsland.The resistantnatureof the upper glauconitelayerand the low dip (typicallyless than 10?)of the UpperCretaceousandLowerTertiarystratatogetherwith the badlandstopography of the island causes the outcropof the "K-T glauconite"to vary dramaticallyin width. The arrowsmarkthe locationswhere fish debriswas observed.PointA is broadbeddingplane surfaceillustratedin Figure11; here, the last occurrenceof ammoniteswas observedapproximately 50 cm below the lower glauconiteof the "K-T glauconite." FIGURE 2-Geologic occurs as isolated elements, principally as bluish black-stained vertebrae and shark teeth. At localities where bone material is most abundant, it occurs as irregularly cemented masses of black bone debris. The absence of any primary sedimentary structures o .e =3 Tan to gray siltstone 5mUpper resistant Glauconite 4m- as c c 0 0) 0 0 0) 0o o 0o 0 c 0 Fish bone layerm 0 2m- a. Ct -. : lm- m Lower Glauconite as 'o of N Om C. o .S (0 Medium gray siltstone FIGURE3-Stratigraphic column of the "K-T glauconite" showing a three-foldlithologicdivisions. within the Fish Bone Layer indicates that deposition was below wave base with the disarticulation of the fish skeletons due to decomposition and bioturbation. The irregular masses of cemented fish debris and concretionary occurrences are restricted to the base of the Fish Bone Layer. The spherical bone-bearing concretions are occasionally found within the uppermost part of the Lower Glauconite layer. Although the three lithologic units that form the "K-T glauconite" vary along strike, the bone layer is surprisingly persistent laterally (Figure 3). Whereas bone debris is very abundant at some sites, at others it is rare or even absent for short distances along strike. At the section where the iridium anomaly was identified, bone debris is rare. Although bone material was not reported by Elliot et al. (1994) in the section where the iridium anomaly was identified, it was observed during the 1995 season in a concretionary sandstone layer 50 cm above the base of the Lower Glauconite. In addition to the fish fossils, bivalves, gastropods, solitary corals, and wood debris occur sporadically throughout the Fish Bone Layer. Only one species of bivalve within the fauna recorded in the 20 m spanning the uppermost part of Unit 9 has its last occurrence in the Fish Bone Layer. No species either makes its first appearance or is restricted to the Fish Bone Layer. The principal difference between the Fish Bone Layer and the glauconite layers above and below is decrease in the amount of glauconite within this horizon. Upper glauconite.-The uppermost facies in the "K-T glauconite" consists of a resistant, reddish-brown, irregularly blocky weathering, glauconitic sand that ranges from 0 to 1 m in This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION thickness.The lithifiednatureof the upperglauconitefrequently results in the developmentof broad resistantdip slopes which accounts for the irregularwidth of the outcroppatternof the "K-T glauconite" (Fig. 2). Although fossils are generallyrare in the UpperGlauconite,occasionally,they are locally abundant. At those sites where fossils are abundant,the molluscs are exceptionallywell preservedwith the fauna being dominatedby large numbersof pairedvalves of the large bivalve Lahillia. Overlyingthe "K-T glauconite"are the sandy concretionary siltstonesof Unit 10 of the Lopez de BertodanoFormation.Fossils in the basal part of Unit 10 are locally abundantand are characterizedby the bivalves, Lahillia larseni and Cucullaea ellioti and the aporrhaidgastropod,Struthiochenopusnordenskjoldi. The rotulariaand echinoid spines that were abundant throughoutthe UpperCretaceouson the island are absentabove the LowerGlauconite.Althoughmegafossilsarenot as abundant in the lower part of Unit 10, the fauna does not display any changein compositionbetweenthe base of the unitand the more fossiliferous upperpart of the unit or in the basal part of the overlying Sobral Formation.Although Huber (1988) has suggested the presenceof a dissolutionintervalin unit 10 to explain the absence of calcareousmicofossils, the presence of locally abundant,well-preservedmegafossils suggests that there were no unusualsea waterchemistryconditionsin the SeymourIsland region duringthe earliestDanian. 559 aboutevery six years.Thermalkills also occur frequentlyalong the boundaryregions of water masses with significantdifferences in thermal characteristics(Hansen, 1939 and Jensen, 1939). Numerousmass mortalitieshave been documentedalong southerncoast of Africa duringthe summermonths when the cold Agulhas Currentshifted shoreward(Dietrich, 1935). Mass kills resulting from salinity changes for the most part are restrictedto nearshoreregions eithernearriverssubjectto periods of high runoff (Beaven, 1946; Engle, 1946) or bay and coastal lagoon in aridregions(Gunter,1945;Hedgpeth,1947). Plankton blooms are by far the most common process producingmass fish mortalityevents (Brongersma,1947). Mass mortalitiesproduced by planktonblooms are the result of suffocationdue to low oxygen levels or poisoningby H2Sfromthe organismsproducing the bloom, typicallydinoflagellates.Otherprocessesthat have been documentedcausingsuddenfish kill events are severe storms(Engle, 1946), stranding(Bower andFassett, 1914), after effects of volcanism (Gislen, 1931), and earthquakes(Escher, 1948; Sieberg, 1923). While thereare a numberof mechanisms that are responsible for mass fish kills, Murray and Renard (1891) Eastman(1903) and Boggild (1916) remarkedaboutthe rarityof bones in modem marinesediments.Brongersma-Sanders (1949) in a review of the literaturereportingthe occurrences of fish remainsin Recent deposits also noted the strikingnearabsence of fish remainsin Recent sediments.Schafer(1962) in a discussionof taphomonicprocessesleadingto the preservation ORIGIN OF FISH BONE LAYER of fish also remarkedaboutthe rarityof fish remainsin Recent The occurrenceof the Fish Bone Layer in associationwith sediments. Both Brongersma-Sandersand Schafter noted that the iridiumanomaly at the K/T boundaryraises several ques- frequentreportsof abundantRecent fish remainsin the moder tions: Was the fish bed producedby changes in local environ- environmentsare in facies where the chance of burialand presmentalconditionsor by the same event thatproducedthe iridium ervationare remote,such in the swash zone. anomaly,the impact event? To determineif there is a relationA common misconception about the preservationof large ship between the Fish Bone Layer and iridiumlayer,it is nec- concentrationof fish remainsis the requirementof slow depoessary to examinethe processesthatproducedthe mass fish kills sitional rates which are believed to allow for the accumulation and the taphonomicprocesses leading to the preservationof a of abundantfish fossils without a catastrophickilling fish mortalitylayer.Thereare a numberof misconceptionscon- event. This belief is based on theinvoking accumulationof largenumbers cerningthe killing mechanismof mass kill events, frequencyof of sharkteeth and whale ear bones in deep marinebasins (Murmortalitylayers in the fossil record, and the depositionalproand Renard, 1891; Eastman, 1903, 1906). 1889; ray, Murray cesses responsiblefor preservingthe victims of mass kill events Because of the low rate of sedimentationin deep marinebasins, 1957). The most commonmisconceptions (Brongersma-Sanders, and errorsin the literature,summarizedby Brongersma-Sanders large numbersof shark teeth and whale bones accumulateon the sea floor.Schaffer(1962) in a classic study to determinethe (1957), were that most mortalityevents were associated with volcanic activity and that the preservationof fish remainsre- length of time carcassesof differentspecies disintegratefound that destructionof fish carcasses was very rapid. Schafferob(1957) served that the quires slow rates of deposition(see Brongersma-Sanders disintegrationof fish carcassesfollowed one of and David (1957) publishedextensive annotatedbibliographies of marine mass mortalityevents and examples of extirpation two courseseven if externalconditionwere the same. Eitherthe carcasseswere refloatedby the trappingof gasses as the internal layers in the geologic record). The two most importantfactors in the formationof a fish organsdecomposedor the carcassesremainedon the bottom.If mortalitylayer are the killing mechanismand the depositional the carcasswas refloated,it never returnedto the bottomintact. setting. Althougha numberof mass fish kills have been attrib- Upon refloating,elements of the carcass begin to loosen and uted to volcanic activity (Weigelt, 1927), Brongersma-Sanders become separatedfrom the floating carcass. When the carcass (1957) noted that most of the recordsof fish kills reportedto be finally settles to the bottomit generallyconsists of disarticulated associatedwith volcanism were not based on observationsnot mass of bones held togetherby soft tissue. Those fish carcasses by oceanographers,but on anecdotalaccounts.Examinationof that remainon the bottom generallyremainintact, but may be these early accountssuggeststhatmost of these mass mortalities disarticulatedand scatterover a limited areaby scavengers.For were not the result of volcanic activity,but as the consequence a fish skeleton to remain intact, it has to be preservedin an of otheroceanographicfactors.The threeoceanographicfactors oxygen depletedenvironmentwhich inhibitsdecompositionand that account for most mass fish mortalitiesevents are thermal prevents destructionby scavengers. Schaffer showed that alshock, salinity changes,or planktonblooms. Suddenchangesin thoughrateof the disintegrationof fish carcassesvariedbetween temperaturefrequentlykills large numberof fish with sudden species, most of the soft tissue was completelydestroyedwithin drops in temperatureseem to be the most common (Gunter, three weeks. An aspect, not commonly recognized, is that the 1941, 1947). Johansen (1929), Blegvad (1929) documenteda phosphatein bone is an essential nutrientand many benthic numberof fish kills along the coast of Denmarkassociatedwith scavengersare known to actively feed on bone material(Blacksevere winter freezes. Storey and Gudger(1936) observed that welder, 1916;Martill,1991; and Haszpunar,1988) and any bone mass fish mortalitieswere a common featureof the coastal re- debris exposed on the sea floor is destroyedif not quickly burgion of south Texas duringthe winter,occurringon the average ied. Within marine siliciclastic facies characterizedby a well This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 560 JOURNALOF PALEONTOLOGY, V. 72, NO. 3, 1998 oxygenatedconditionswith an abundantbenthos,the preservation potentialof fish skeletonsis very low. The low preservation potentialof fish skeletonsin siliciclasticfacies is highlightedby David (1947) attemptto documentenvironmentalconditionsthat promotedthe accumulationof fish fossils. David studieda large numberof bottomsamplescollectedby ManleyL. Natlandusing small bottom grabs and wire drags from the CatalinaChannel, off the coast of southernCalifornia.The waterdepthrangefrom 10 to 758 m with bottom sedimentvaryingfrom mediumnearshore fine grained sands to fine grained deep water silts and muds. While most of the Natland'ssamplesdid not containany fish remains, those samples that containedfish debris, the remains were limited to scales, vertebraeand a few teeth. David noted that referencesto abundantfish remains in the Tertiary formationsalong the Pacific Coast, except for the spectacular fish beds in the siliceous facies of the UpperMiocene, are generallybased on occurrenceof fish scales andnot on the presence of partialor completedskeletalremains. The scarcityof vertebrateremainsin Recentsedimentsattests to the rarityand the special conditionsrequiredfor the preservation of mortalityhorizons. When evaluatingmass mortality events in the geologic record,it is importantto note the following: 1) Is it associatedwith other mortalityhorizons?2) Does the assemblageconsists of skeletonsof whole fishes, masses of skeletaldebrisor only as isolatedparts,scales, bone fragments, vertebrae,or otoliths and teeth? and lastly, the nature of the sedimentsentombingthe fish remains.The principalcauses of mortality events are a sudden drop in temperature,salinity changes,and the most commonand wide spreadbeing plankton blooms. Each of these processes are typically recurringevents. Brongersma-Sanders(1957) noted that the interval between these events is very short, frequentlyless than 10 years. The recurringnatureof mass mortalityhorizonsin certainfacies indicates that for the preservationof mortalityhorizonsrequires special conditions. Mortalityhorizons in the marine strataare commonlyassociatedwith hypertropic,upwellingregionsof the seas. Brongersma-Sanders (1957) noted that three factorsfavor the preservationof mortalityhorizons in hypertropicregions: abundanceof life, frequencyof planktonblooms, and the limiting of benthic scavengersby low oxygen and high H2Slevels below hypertropicregions. These factors increases the opportunity for the preservationof intact skeletons being preserved. Moder examplesof moder mortalityhorizonsassociatedwith upwelling regions have been reportedby Bonde (1928), Stub(1957), and Falke (1939 and bings (1939), Brongersma-Sanders 1950). Jordan(1920) reportedthe extraordinary Xynegrex mass kill horizonin a diatomitemine in the Miocene MontereyFormation near Lompoc, California.He also noted that the Xyne horizon was not a unique event, but one of a numberof other mortalityhorizonsin the Lompocsequence.David (1943) listed a numberof occurrencesof multiplemass mortalityhorizonsin the siliceous facies of the MontereyFormationthroughoutCalifornia. The finely laminatednature of the diatomitesclearly indicatesan absenceof benthicactivityas the diatomaceoussedimentsof the MontereyFormationaccumulationduringthe Late Miocene. While a numberof killing agents may have been responsiblefor the numerousmortalityevents, the key factorleading to the preservationof these mortalityhorizons in the siliceous facies of the Montereyformationwas the absence of active benthic scavengers. An examinationof the depositionalsetting and the natureof the occurrenceof the Fish Bone Layer on SeymourIsland indicatesthatits originis not associatedwith a recurringprocesses characteristicof a hypertropicfacies. The absence of primary sedimentarystructures,presenceof well developedbioturbation features,and moderatelydiverse benthicfaunaclearlyindicates a normalmarinedepositionalsetting below fair weatherwave base (Macellari,1988), an environmentalsettingwith the lowest potentialfor mass fish kills and preservationof mortalityevents. The uniquenatureof the Fish Bone Layeris highlightedby the fact that it is the only fish mortalitylayer within the 1600 m of uppermostCampanian,Maastrichtian,and Paleocene strataon SeymourIsland. The occurrenceof glauconite in close association with the Fish Bone layer led Norman MacLeod (per comm., 1996) to suggest that the fish horizon representeda condensedinterval resultingfrom decreasedrate of deposition associatedwith an increasedformationof glauconitewhich characterizesthe "K-T glauconitehorizon." Althoughthe presence of glauconitesuggests a decline in the rateof sedimentation,the absenceof manganese or phosphoritenodulesindicatesthatthe Fish Bed Layer did not accumulatein a sediment-starveddeep sea environment. It should be emphasizedthat none of the other similarglauconitic horizons in either the Lopez de Bertodanoor the Sobral formationscontain any extirpationhorizons. If other mortality layers were presentin the sequence or associatedwith any of the other glauconiticenrichedintervals,it might be possible to dismiss the close associationof the bone layer with the iridium anomalyto coincidenceresultingfrom local marineconditions, but the uniquenessof the fish horizon and its occurrenceimmediately above the iridiumanomaly,leads to the inescapable conclusionthat thereis some relationshipbetweenthe two. The uniquenessof the Fish Bone Layerin the UpperCretaceousand Lower Tertiarystrataof the James Ross Island Basin and the absenceof any evidence of recurrentmass mortalityevents suggests thatthe formationof the extirpationhorizonwas the result of unusualevent. Elliot (1988) and Elliot et al. (1994) notedthat the AntarcticPeninsulaareawas characterizedby increasedvolcanism duringlatest Cretaceousand earliestPaleocene.The absence of a volcanic ash layer with the fish horizon at the K-T boundarywould suggest that the fish kill was not relatedto a volcanic event, but a nutrientpulse from local volcanic activity causing a lethal planktonbloom cannot be ruled out. A nonvolcanic cause of the Fish Bone Layer is supportedby the absence of any extirpationhorizons associated with any of the bentonitesin either the Lopez de Bertodanoor Sobral formations. Althoughthere is no direct evidence to indicatedthe nature of the kill event that producedthe Fish Bone Horizon,its close associationwith the iridiumhorizonsuggests thatthereis a high likelihood it was producedby a planktonbloom or sea chemistrychanges following the impactevent. Recent evidence for a post-impactbloom was reportedby Heymannet al. (1997) from the Brazos River K-T boundarysection. They suggested thatthe occurrenceof nativesulfurwith an isotopic composition of 833S = -12.87%o, 834S = -24.89%o, and 836S = -46.4%o was producedby a sulfur-reducingbacteriaduringa local, transient bacterialbloom following the Chicexulubimpactevent. THE FISH BONE LAYER AND K-T FAUNAL TRANSITION Zinsmeisteret al. (1989) presentedmegafossil rangedatafor the uppermostLopez de Bertodanoand the lower half of the Sobral formations.The faunal data showed that there was no level with a largenumberof last occurrences, single stratigraphic suggesting that the faunal change across the boundaryin the high southernlatitudeswas gradationalin nature,characteristic of the acceleratedmass extinctionpattern(Zinsmeister,1994). The term "gradational"is preferredratherthan "gradual"when characterizinga decline in biotic diversityover a relativelyshort interval of time. The term gradationalextinction patternwas introduced by Kauffman (1988, p. 59) for "... acceleration of backgroundextinctionrates correlativewith increasedbiologic stress from rapidlydeterioratingglobal environmentsrelatedto, This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION 561 e.g., eustatic fall, greenhouse,glaciationand cooling, etc." Althoughthe changein diversityacrossthe K-T boundaryon Seymour Islandappearedto be gradational,the extent that artificial range truncationsof the Signor and Lipps Effect played in producingthe gradedextinctionpatternwas not addressed.The role of artificialrange truncationsraises questions about the nature of gradedextinctionpatternsacrossmass extinctionboundaries. Although Signor and Lipps (1982) stated that a gradedpattern may not be the result of artificialrange truncations,their only suggestionfor falsifying the Signor/LippsEffect was additional detailedsampling,but they did not proposeany new approaches or techniquesfor sampling.The following techniquehas been developed to test the potentialeffect of artificialrange truncations on extinctionpatterns. SPATIAL BIOSTRATIGRAPHY Since William Smith (1817) proposedthe principleof faunal succession, the science of biostratigraphyhas focused on establishing zonal schemesfor age datingand correlation.Traditional biostratigraphiczonations are based on Oppel's (1856-1858) techniquesof first occurrencesand over-lappingranges for establishingzonal boundariesand recognitionform the foundation of biostratigraphy. Althoughnew techniquesusing Oppel'sprinciples have been developed,such as graphiccorrelationmethods (Shaw, 1964), representationof biostratigraphicdata has also remainedvirtuallyunchangedsince the last century.Traditionally, biostratigraphicdata are plottedto show the total rangeof a taxon from its firstoccurrenceto its last with a solid rangebar ratherthan the actual occurrencesof fossils. The gaps in the observed range of a taxon were assumed to be the result of collecting deficienciesor facies changes and were unimportant. Anothercommonprocedurewas to extendthe rangeof a lineage to stage boundarieseven though data providesno tangibleevidence for the extension.The K-T controversydramaticallyhighlighted the inherentshortcomingsto the traditionalapproachto biostratigraphicproceduresand graphicalrepresentationof biostratigraphicdata. An example that clearly illustratesone of these shortcomingswas the practice of extending of species ranges to stage boundarieswhich distortedthe true natureof biotic events and suggested instantaneousextinctiongroups of organismswhen such extinctionevents did not occur (Newell, 1982). Review of publisheddatatogetherwith new field studies, has shown that many of the groups portrayedas supportinga suddencatastrophicextinctionat the end of the Cretaceoushad declined dramaticallyin diversityand abundanceor became extinct well before the K-T boundary(Fig. 4). The Maastrichtian history of the rudistidsclearlyillustrateshow the traditionalapproachto biostratigraphycould distortbiotic events. The apparent sudden extinctions of rudistidsand inoceramidbivalves at the K-T boundarywere used by proponentsof the bolide hypothesis in the early stages of the K-T debate as prima facie evidence to supportthe catastrophicnatureof the impactevent. Re-evaluationof earlierrange data along with new distribution data obtainedduringthe last decade have shown that nearlyall rudistidsand inoceramidsbecame extinct several million years before the end of the Cretaceousand only a very few species actuallysurviveduntilthe very end of the Mesozoic (Kauffman, 1994). The use of continuous range bars representingtaxon rangeshas commonlybeen replacedby noting each occurrence of the taxon. Although collecting methodologieshave greatly improved with the emphasis on detailed, closely-spaced sampling, the natureof extinctionpatternsacross the K-T boundary continuesto generateconsiderablediscussion. The need for a new approachto evaluaterange terminations was clearlydemonstratedby SignorandLipps (1982) when they describedthe effect of artificialrange truncationson diversity of classical representationof rudistiddiversity and apparentinstantaneousextinctionpatternproducedby extending rangesto stage boundaries.1, Kauffman(1984) plottedrudistiddiversity data to illustratethe effect of extendingrange data (Coogan,in Moore,1969)to stageboundary.The plottingof rangesto stageboundaries producesan the apparentmass extinctionof the rudistidsat the K-T boundary.This type of paleontologicdatahas been cited as evidence of supportfor the catastrophicimpactinduce extinctionat the end of the Cretaceous;2, rudistiddiversitychangesduringthe last four million years of the Maastrichtian(Kauffman,1988; Johnson and Kauffman,1996). The apparentmass extinctionof rudistidsat the end of the Cretaceousvanisheswhen the datais accuratelyrepresented. FIGURE 4-Comparison patterns.They stated that gaps in the stratigraphicrecordhave the potentialof increasingthe numberof species with artificially truncatedranges, producingan apparentgradualdecline in diversity whether or not the real change was gradualor abrupt (Fig. 5). Fastovskyand Weishampel(1996), in a discussion of the artificialrange truncations,noted that even in a continuous stratigraphicsequence, as an arbitraryboundaryis approached (Fig. 6), the probabilityof findingfossils declines as the search interval decreases. As a consequences, the effect of artificial range truncationsproducedby collection omissions, preservation, and facies changescould resultin a catastrophicextinction patternappearingto be graded.AlthoughSignor and Lipps did not rule out the occurrenceof a gradationalextinctionacrossthe K-T boundary,they emphasizedthe need to evaluategradational patternsin respect to completenessof the stratigraphicrecord and artificialrange truncationphenomenon.The artificialrange truncationphenomenonclearly highlightsthe importanceof developing a methodology for determiningthe role of artificial range truncationsin diversity patternsduring mass extinction events. Although a statisticalapproachto test the extent of the Signer/LippsEffect has potentialin evaluatingthe role of range truncationphenomenon,it is limited by the data sets used. Several statisticalprobabilityanalysesof Macellari's1986 ammonite data (fig. 5, p. 7) have been undertaken(Strauss and Sadler, 1989, Springer,1989, and Marshall,1995) to determinethe natureof the extinctionon SeymourIsland,but none of these analyses was able to conclusively show that the disappearanceof the ammonites was abrupt. Unfortunately,no attempts were made by these authorsto determineif the range data had been modifiedsince its initial publication.It shouldbe noted thatthe location of the K-T boundaryin three of Macellari'smeasured sections (A, E, and F) were located below the actualboundary (Macellari, 1988). As a consequence,the ranges of several of the ammonitetaxa are not correctas originallydepicted. To determinethe role of artificialrange truncationsin the extinction patternsacross the K-T boundary,a new empirical This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 562 JOURNALOF PALEONTOLOGY, V. 72, NO. 3, 1998 Effect Signor/Lipps CatastrophicMass Extinction 1 ArbitraryBoundary GradedMass Extinction 2 A . .. Time Time A ... .. ..... Taxa . EaW ;9op ao .... ..... .X ..... ..... ... . ... 0 ..\ - Artificial RangeTruncationsof Signor/LippsEffect O O 3 r1 L, 0 ,1 0 -_. .... .... 11 II I Taxa FIGURE5-Artificial Range Truncationsof the Signor/LippsEffect. 1, instantaneousextinctionpatternproducedby a catastrophicmass extinction;2, gradedextinctionpatternproducedby gradualdiversity declineduringan AcceleratedMass Extinction;3, apparentgradedextinctionpatternis producedby artificialrangetruncations(solidcircles indicateobservedoccurrencewhile open circles are fossils that were not recoveredbecause of the of vagariesof preservationand collection). approach to data collection and graphically representing the data is needed. Although the phenomenon of artificial range truncations can never be completely eliminated, the following approach (referred to as spatial biostratigraphy) provides a technique for approximating lineage terminations and evaluating the robustness of diversity changes and extinction patterns. One of the features of artificial range truncation phenomenon is the decreasing probability of finding fossils, either as a consequence of gaps in the record or as the given search interval within a section decreases. Although there is no way to increase the search interval in a single section or core, increasing the number of sections studied within a geographic area, in effect, increases the search interval and the probability of obtaining a more complete record of fossil occurrences as the boundary is approached. The collection of range data from a numberof localities has been standardpracticesince the 19th century,but the traditional approachof combininggeographicdatainto single-dimensional composite range compilations effectively eliminates the geographic component within the data set. The spatial biostratigraphicapproachto analyzingrangedata is based on the retention of the geographiccomponentin the analysis by utilizing a new graphicalrepresentationof last occurrencedata. With this new graphical presentation, it is possible to preserve and integrate the spatial geographic component within a single dimensional temporal range data representation. Although a number of terms have been used to define bio- stratigraphicranges, the following terms used in this paperare definedto avoid confusionswith similartermscurrentlyused in the literature.Following Shaw (1964), each species exists for a lesser < > greater Search interval FIGURE 6-Effect of reducedsearchintervalon fossil occurrences.The figuredepictsthe occurrenceof fossils as randomlyanduniformlydistributedthrougha sectionof rock.Althoughthis assumptionsextreme, it illustratesthat the probabilityof finding fossils decreasesas the searchintervaldeceasesas an arbitraryboundaryis approached. (Modified fromFastovskyand Weishempal,1996) unique interval of time, biozone, consisting of the time between the orientation and termination of the species. The total observed range of a species is the interval of time between the first and last observed occurrences of a species. In contrast to traditional biostratigraphy, in which there is a single first or last observed occurrence, in spatial biostratigraphy there are as many first or last observed occurrences as there are sections sampled. Each section has its own first and last occurrences, and because of the vagaries of sample collection and preservation there is no basis to assume that these observed occurrences are synchronous with any other section. Because it is impossible to know either the very first or last individual, there will always be a gap between ends of the biozone and the total observed range. The goal of any range determination is to reduce the interval of time represented by the gap at either limit the biozone to a minimum. Although the following discussion will focus on the gap between the last observed occurrence and the termination of the biozone, the procedure could just as easily be applied to orientation of the biozone and first observed occurrence. The key to spatial biostratigraphy is the graphical representation of the data in a manner that retains the geographic component of the data (Zinsmeister, 1996) which may be accomplished by either of two approaches. The first approach is to plot only the last observed occurrence from each sampled section in a single composite section, disregarding all other non-last occurrence data. The second is to plot all the occurrence data, but using an easily recognizable symbol to distinguish last observed occurrences from non-last occurrence data. This technique is illustrated later in the paper when discussing faunal changes across the K-T boundary on Seymour Island. Traditionally, all occurrence data are plotted in an undifferentiated form and, as a consequence, the last occurrence from each section is unrecognizable. When only the last observed occurrences are plotted, a cluster develops which is termed a termination cluster (Fig. 7). With a limited set of geographic This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION 1 4 563 5 Composite section 2 Plot of Last Occurrences Termination cluster I II occurrenceof a species from FIGURE 7-Range terminationcluster.1, Compositeplot of rangedata.The compositesection shows the stratigraphic seven measuredsections. Althoughthere is a last occurrencein each of the measuredsection, the plottingof observedoccurrencesin a single dimensionalcompositesectionfiltersout any informationaboutthe last occurrencein any of the individualsections;2, Plot of only last occurrences. Ratherthanusing the traditionalsymbolsfor observedoccurrences,a single right-sidedash is used to indicatehighestgeographiclast occurrence horizonin the compositesection,the numberof fromeach sectionmeasured.If thereis moreone thata single occurrenceat a given stratigraphic occurrencesis recordedafterthe dash.CompositesectionI illustratesthe potentialscatterof last occurrenceswhenonly a few sectionsareplotted. CompositesectionII illustratesthe clusteringof last occurrencesas the numberof sectionsincreases.This clusteringof last occurrencesis defined as the rangeterminationcluster,and it is interpretedas approximatethe lineagetermination. data, scatter within the termination cluster will provide little insight into the length of the gap between the last occurrence and the true termination of the species. With each new set of data, the last occurrences will begin to cluster at a certain stratigraphic level approximating termination of the species. A point to emphasize is that as long as the last occurrences continue to fall within this termination cluster, we can be reasonably confident the species termination has been approximated. However, if new data sets include occurrences stratigraphically higher than the termination cluster, the cluster is then assumed to be due to some local facies problem or collecting phenomenon and must be rejected. This approach can be used just as easily for an entire assemblage as for a single taxon. Instead of developing a single termination cluster, the extinction clusters of all the elements within an assemblage will produce what is termed a termination band approximating the extinction pattern (Fig. 8). If a graded extinction pattern continues to be repeated, it must be assumed that the pattern represents a gradual decline in species diversity through a given interval of time. As in the case of a termination cluster, a gradational extinction pattern can be falsified and assumed to be the result of the Signor/Lipps Effect, if additional data demonstrate the occurrences of significant number of species within the assemblage significantly above the termination band (Zinsmeister, 1996). Central to the falsification of termination clusters or bands is the question: How much data must be collected before the cluster or band is considered to be valid? At the present time, there is no completely satisfactory answer to this question. As long as additional data continue to plot within the termination cluster or band it must be considered valid, with the caveat that the pattern would be falsified if new range data are obtained which plots significantly above the termination band. If the study is restricted to a single basin, the problem of emigration may play a significant role in the observed appearance or disappearance patterns. Support of an observed pattern would be greatly strengthened if the observed extinction pattern is duplicated in other geographic regions. SPATIAL BIOSTRATIGRAPHIC ANALYSIS OF THE K-T BOUNDARY SECTION ON SEYMOUR ISLAND The extensive exposure of the K-T boundary on Seymour Island provides an opportunity to examine events at the end of the Cretaceous, not only in a temporal, but also a spatial context. This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 564 JOURNAL OF PALEONTOLOGY,V. 72, NO. 3, 1998 E C A rl G Dx ,X E terminationband.As an arbitraryboundaryis approachedin a single section,the searchintervalfor fossils is reduced,increasing the probabilityof a distorteddiversitypatternby the artificialrangetruncationphenomenon.The searchintervalapproachingan arbitraryboundary can be increasedby incorporatingdata from additionalsections. If a gradedextinctionpatternis real and is the result of a gradualdecline in diversity,the terminationclusterof each taxon will producea band,termeda assemblageterminationband. FIGURE 8-Range The "K-T glauconite" is exposed along strike for approximately 6 km. The badlands topography, together with the low dip of the strata (less than 10 degrees), causes the outcrop of the "KT glauconite" to snake across the island in a belt about 2 km in width (see Fig. 3). Consequently, the sequence on Seymour Island provides a unique opportunity to examine events across the boundary from a geographic perspective over an area of Late Cretaceous sea floor of approximately 12 km2. None of the KT localities from the northern continents have aerial exposures that remotely approach those on Seymour Island. The faunal data (Fig. 9) are based on 24 measured sections (Fig. 3) across the "K-T glauconite." Depending on the nature of the exposure, the base of the sections varied from 5 to 20 m below the Lower Glauconite and extended to above the Upper Glauconite. While the stratigraphic "in place" occurrence of each taxon was noted, its spatial location on the outcrop was also noted. This was accomplished by selecting sections which had well exposed outcrop. The outcrop surface surveyed at each section is referred to as a panel. Each panel was carefully searched with the spatial location of each fossil within the panel in relation to the K-T boundary noted (Fig. 10). This procedure allowed for, not only the stratigraphic location of each species through the section, but also the spatial association and abundance of each species at each section. While molluscs at any single section are rare, combining the data from the 24 sections with other occurrences reveals a moderately diverse fauna of 36 species within this interval. Examination of the range date reveals that 20 species (55 percent) disappeared below the K-T boundary, two species (6 percent) make their last occurrence in the "K-T glauconite" with 14 species (39 percent) surviving the boundary event (Fig. 9). The mapping program also revealed that the reported "Tertiary ammonites" (Zinsmeister et al., 1989) were the result of inliers of Unit 9 in the broad dip slope produced by the resistant nature of the Upper Glauconite and displacement of the boundary by several minor faults (Zinsmeister, 1995). The "last" stratigraphic occurrence of ammonites on Seymour Island is from a 300 square meter bedding surface (Fig. 11), 50 cm below the lower glauconite bed (location A, Fig. 3). No ammonites were found in either the Lower Glauconitic iridium anomaly bed or in the overlying fish layer horizon. Although future discovery of ammonites within the Lower Glauconite, or above it, cannot be ruled out, the failure to find any ammonites during the focused collecting program through this interval during the 1995 season strongly suggests that the event that led to the disappearance of ammonites and associated Cretaceous marine life was independent and prior to, the events leading to the deposition of the iridium anomaly and subsequent fish kill. The use of the spatial biostratigraphic approach for the K-T boundary faunas on Seymour Islands indicates the faunal change across the This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 565 ZINSMEISTER-CRETACEOUS EXTINCTION I I I I I( I I . 4 2 Om .................................... Fish Layer Lower Giaiiconite i r, x , X' ,-~~~ 1 - -r- t r.r ~X . ' . ..... --...- ..- . - . ,X X '--i . -.X'.-........ r ' ' iconite Upper GIaL 3 r- r- rr. * r" ~ r-r-Ir- 2 r- x -~~~ I- ' *- X. r-~ X~~~~~~~~~~ :? - r r- X 3 4 5 ~~~~:: 6 7 x X *~~~~~ x* - 12 ,, rr-x- ~~~~~~~~~~~~X ,c x , b 13 14 ' ' E _ _ 15 X XX 16 I ?I ,E 'a'" - 0 LO:DXL< , E,~--0EL.~ Eo .*> <* Io X 8i^'S'Sl^tIc? 5?S Q CW -Dc w :~ 3 V m7 K , , - ? '): 0 O 0U-9a S_?i LC iSnoSe'SSgfflSCD _ ~' t5-0 a'.8' C * c w' sS-o 0 in ,o 0a - 11111 .Q ( 2 UOco^NE Q O WN &1- Eo D =il 0 8 <2 l lC0||ll|Qi0 u CbL 0 0 0 tl0 00 ~~~G E - o 3 n~~~~~~~c CI)OJ O<~~~~~~~O~~-OJO .3<0 r .0 I.C FIGURE9-Composite rangechartbased on the 24 measuredsections across last 20 metersof the unit 9 and throughthe "K-T glauconite."All occurrencedatarepresent"in-place"occurrences.Right-handdashesalongoccurrenceplots represent"lastobservedoccurrences"in the measured sections.X's representnon-lastobservedoccurrences.Five species withoutany observedoccurrenceshave documentedoccurrencesabove and below sampledinterval.Letterfollowing each species-A, arthropod;B, bivalve; C, cephalopod;E, echinoderm;G, gastropod;R, surpulid;S, scaphopod;V, vertebrate. boundary is more abrupt than originally believed (Zinsmeister et al., 1989). However, the sequential nature of the disappearance of the ammonites, iridium anomaly formation, and fish kill raise important questions about the nature of the events at the end of the Cretaceous. DISCUSSION A similar fish mortality layer at the K-T boundary at Stevns Klint has received considerable attention (Alvarez et al., 1984). The disturbed nature of the sequence at Stevns Klint together with fundamental differences in depositional environments between the Fish Bone Layer on Seymour Island and the Stevns Klint Fish Clay (siliciclastic verses carbonate) makes comparison difficult and tenuous at best. Although the Stevns Klint Fish Clay has been presented as evidence for the impact event (Kastner et al., 1984), the prevailing view is that the Fish Clay is Danian in age, a condensed section, and the result of a combination in the rate of sedimentation and carbonate dissolution and This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 566 JOURNAL OF PALEONTOLOGY,V. 72, NO. 3, 1998 FIGURE 1O-Spatial OccurrencePanel.Areaoutlinebelow the "K-T glauconite"representsthe spatialoccurrencepanel.Narrowline with numbers indicatesstratigraphic intervalin meterswith 0 being the base of the measuredsectionat lower marginof panel.Letterscorrespondto the location of each fossil occurrence(AO,unidentifiableammonitefragment;Al, Maoritesdensicostatus;A3, Grossourvitesgemmatus;E, echinoidspine;R, rotularia;L, Lahillialarseni;Cl, solitarycoral;T, Oistotrigoniapygoscelium;W, wood debris). thereforenot related to the extinction event at the end of the Cretaceous (Ekdale and Bromley, 1984; Officer and Ekdale, 1986; Elliott et al., 1989). The Danian age together with the sedimentologicdata suggests the Fish Clay is not the result of an impact,but reflectscomplex shallow-waterfacies thatexisted at the time of deposition.The absenceof similarFish Clay facies either at other Danish or any EuropeanK-T boundarysections providesadditionalsupportindicatingthatthe Stevns KlintFish Clay is a productof local environmentalconditions,not an impact inducedcatastrophe. The disappearanceof ammonitesand otherCretaceousmarine life prior to the depositionof the iridiumanomalyraises questions about the sequence of events across the K-T boundaryin the high southernlatitudes.Recent work on SeymourIslandand the nearby islands (Zinsmeisterand Feldmann,1995) together with the re-evaluationof fossil data from the high to low-latitudes by a numberof workers (see Kauffman,1994; Johnson and Kauffman,1996, mid- and low-latitudemarinedata;Archibald, 1996, terrestrialvertebraterecord; Askin and Jacobson, 1996; Sweet et al., 1993, high latitudefloralrecord)indicatethat the Earth'sbiospherewas alreadyundersevere stress,leadingto a decline in global diversity well before the end of the Cretaceous. This decline in biodiversityhas been notedby proponents of the impacthypothesis,but has been dismissedas eitherbeing unrelatedto events at the end of the Cretaceousor due to our incompleteknowledgeof the fossil recordduringthe Maastrichtian (Alvarez et al., 1984). Changesin the Earth'senvironment may have led to biotically stressed biosphereduringthe latest Cretaceousresultingin a prolongeddecline in diversityand extinctionspriorto the impactevent. The most likely cause leading to the biotically stressedworld at the end of the Cretaceouswas widespreadvolcanic activity during the waning stages of the Cretaceous.Although most discussion of volcanic activity has focused on the Deccan Trapsof India(McLean,1985;Courtillot, 1990), volcanismwas widespreadglobally.Orogenicactivityresultingfrom subductionalong the westernmarginsof Northand South Americaproducedwidespreadigneous activitythoroughout the wanningstages of the Cretaceous(Dickinsonand Seely, 1979; James, 1971). Extensivevolcanismalso characterizedthe Pacific marginof West Antarcticathis period (Elliot, 1988) Because high southernlatitudesregions are particularlysensitive to climatic perturbations,biotic changes in the high latitudes may be viewed as a harbingerof futureevents. It is clear from the paleontologicalrecord on Seymour Island and the nearby islands that the biodiversityin the marine communitiesof the high latitudeswas in decline during the latest Cretaceous,beginning in the Late Campanian,with the diversityof a number of cosmopolitaninvertebrategroups either markedlydeclining or becoming extinct duringthis period (Zinsmeisterand Feldmann, 1996). In a recentcomprehensivereview of high latitude Late Cretaceousand earliest Tertiarymolluscan faunas of the South Island of New Zealand, Stilwell (1995) noted a similar patternof diversity decline and disappearanceof cosmopolitan molluscan taxa and no single instantaneousextinction event. This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions EXTINCTION ZINSMEISTER-CRETACEOUS 567 FIGURE11-Fossil occurrences on bedding plane surface 50 cm below the K-T boundary located at point A on figure 4. Continuous solid line is located at the base of the lower glauconite layer. Area of bedding plane below the glauconitic layer with the fossil is approximately 500 m 2. Letters indicate location of fossils observed at this location (F, fish debris; W, wood debris; N, nautiloid; AO, unidentified ammonite fragment; Al, Maorites densicostatus; A2, Diplomoceras maximum; A4, Kitchinites laurae; A6, Zealandites varuna; A7, Pseudophyllites loryi; Ce, Cucullaea ellioti; Pc, Panopea clausa; Co, Surobula nucleus; G, Goniomya hyriiformis; Cl, solitary coral). Thomas (1990) in an analysis of deep-sea benthic foraminifera from the Maud Rise located at approximatelythe same latitude (ODP sites 689, 64?31.009'S and 690, 65?9.629'S) as Seymour Island in the WeddellSea, found no evidence of a suddenmass extinctionat the K-T boundary. Changesin ammonitediversityare particularlyillustrativeof the polar climatic change during the Maastrichtian(Table 1). Ammonitespecies richnessdroppedfrom 45 species duringthe early and middleCampanian(Olivero, 1992) to 14 by the Maastrichtian(Macellari, 1986). Although 14 species have been reportedfrom the Maastrichtianon SeymourIsland,the maximum numberof ammonitesat any horizon is seven. There is some confusion concerningthe numberof ammonitetaxa at the K-T boundaryon SeymourIsland. This confusion stems from a review paper (Ward,1990b) which comparedMacellari'sammonite data from Seymour Island with diversity trends from Zumaya, Spain.AlthoughWardcorrectlystatedMacellari'sdatain the text thatonly four species surviveduntilthe boundaryevent, he provideda figure showing that there was an increase in diversity duringthe latter part of the Maastrichtianwith the ten ammonitespecies becomingextinctat the boundary.Zinsmeister and Feldmann(1996) based on faunaldatacollected subsequent TABLEI-Stratigraphic distributionof SantonianthroughMaastrichtianammonitefamilies on James Ross, Vega, Snow Hill, and Seymour Islands. (Modified from Zinsmeisterand Feldmann, 1996) Early Campanianto Late Campanian Late Santonian Phylloceratidae Tetragonitidae Guadryceratidae Scaphitidae Baculitidae Nostoceratidae Diplomoceratidae Pachydiscidae Latest Campanian/EarliestMaastrichtian Phylloceratidae Tetragonitidae Guadryceratidae Diplomoceratidae Desmoceratidae Kossmaticeratidae Pachydiscidae Diplomoceratidae Desmoceratidae Kossmaticeratidae Pachydiscidae Placenticeratidae Phylloceratidae Tetragonitidae Guadryceratidae Scaphitidae Baculitidae Nostoceratidae Latest Maastrichtian Kossmaticeratidae Desmoceratidae Diplomoceratidae This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions Tetragonitidae Guadryceratidae V. 72, NO. 3, 1998 JOURNALOF PALEONTOLOGY, 568 TABLE2-Number of ammonitespecies and families from the late Cretaceous of the James Ross Islandregion. (Modifiedfrom Zinsmeisterand Feldmann, 1996) Age Latest Maastrichtian Early to Late Maastrichtian Latest Campanian Early to Late Campanian Santonian No. species No. families 5 14 21 14 14 5 6 7 11 8 to Macellari'swork and discountingthe "Tertiaryammonites" showed that only four species survived until the end of the Maastrichtian. Duringthe 1995 field seasonan additionalspecies of ammonite was discovered within 50 cm of the boundary bringingthe total numberof ammoniteson SeymourIsland at the end of the Maastrichtianto five. Of greatersignificanceto understandingevents in the high southernlatitudesduring the waning stages of the Cretaceous are the changes in ammonitediversityat the family level (Table 2). Familial diversityin the James Ross Island region declined from 11 families in the middle Campanian,to six by the Maastrichtian.The five families of ammonites(Phylloceratidae,Placenticeratidae,Scaphitidae,Baculitidaeand Nostoceratidae)that disappearedaroundAntarcticacontinuedto prosperin lower latitudes until the end of the Cretaceous(Zinsmeisterand Feldmann, 1996). Duringthe 1994 field season, a sixth family,Pachydiscidae, a common element throughoutmost of the Maastrichian, was documentedas disappearingpriorto the K-T boundary. Traditionally,ammoniteworkershave believed that there was a global isochroneityin the disappearanceof ammonites duringthe waningstagesof the Cretaceous.The absenceof these importantMaastrichtianammoniteson SeymourIslandled several workersto suggest that eitherthe uppermostMaastrichtian is missing (Weidmann,1986) or thatK-T boundaryon the island had been place to low in the section Ward(1990a). This view has continuedto be expressedby Ward(1995) when he stated in referenceto Zinsmeister(1995) that the decline in ammonite diversityin the high latitudesduringthe Late Cretaceousis comparableto those in the mid- and low-latitudes.It is clear that absence of these ammoniteshas nothingto do with local stratigraphic problems,but was the result of these families disappearing from the high southernlatitudes several million years before they became extinct in the mid- and low-latitudesat the end Maastrichtian.The extinctionof the ammonitesalong with that of other Cretaceousextinction species on SeymourIsland occurs approximately50 cm below the iridium anomaly and their absence in the overlying fish layer, raises the fascinating questionof whetherthe impact was the sole factor of the TerminalCretaceousExtinctionEvent,or was the proverbial"straw that broke the camel's back" leading to the extinctionof many otherforms of life that might have survivedhad the impactnot occurred. It is interestingto speculatewhat might have happenedif the Earth'sbiospherehad been "robust"instead of stressed when the impact occurred.In terms of this discussion, a biotically robustworld is definedas state in which the global biodiversity is increasingor relativelystable with the numberof extinctions approximatingthe numberof appearanceof new species. In a robustbiosphere,catastrophicevents such as an impact appear to have minimalor no effect on global biodiversity.The apocalyptic mid-Ordovicianvolcanic eruption during the Taconic Orogeny,the greatestplinean eruptionduringthe Phanerozoic, covered several million square kilometers of eastern North Americaand Baltoscandiawith enough ash to form a 1 to 2 m thick Millbrig/Big Bentonite (Hay et al., 1992). Although the Millbrig/BigBentonitevolcanicevent left an easily recognizable signaturein the rock recordover a significantgeographicarea of the Earth'ssurface, it was not followed by any significant decrease in global diversity indicating that in a robust biotic world, the biosphereis resistantto majorphysicalcatastrophes. The effect of comparablecatastrophicevents on the biosphere understress such as existed at the end of the Cretaceouswould be greatlymagnifiedand could lead to a majormass extinction event. Although geochemical data and microscopic shocked quartzand spherulesindicate that an impact event occurredat the end of the Cretaceous,the absenceof any readilyobservable widespreadsedimentologicsignature,similarto the Millbrig/Big Bentonite, suggests that it did not approachthe magnitudeof otherdocumentedgeologic catastrophesduringthe Phanerozoic, but was large enough to precipitatethe final act of the mass extinctionat the end of the Cretaceousin a biospherein crisis. It is interestingto note that a multicausalscenario has been discussed by a numberof investigatorssince the initial impact hypothesiswas proposed(Clemens et al., 1980; Birkelundand Hakansson,1982; Wiedmann,1986 and 1988), but these multicausalscenarioshave been ignoredor dismissedby proponents of the impacthypothesis.Data from the high latitudesprovides new and importantsupportfor multicausalcause of the mass extinctionat the end of the Cretaceous.It is time to accept the fact that mass extinctionat the end of the Cretaceouswas not the consequenceof a single instantaneouscatastrophicevent,but was a multicausalevent involvingboth long termterrestrialprocesses and an extraterrestrial event. CONCLUSIONS The occurrenceof a fish bone layer immediatelyabove the K-T iridiumanomalyon SeymourIslandmay representthe first documentationof a mass kill producedby an impact event at the end of the Cretaceous.The absenceof any otherextirpation horizonsin the UpperCretaceousand Lower Tertiarysequence strongly suggests that the fish kill is related to the event that producedthe iridiumanomaly.The disappearanceof ammonites and otherCretaceousspecies priorto the formationof eitherthe iridiumanomalyor the fish bone horizon suggests that the Terminal CretaceousExtinctionEvent was not the resultof a single factor,but to the conjunctionof several events with the impact precipitatingthe final extinction.If the Earth'sbiospherehadnot been undera period of long term stress beginningin the Campanian,the impactevent at the end of the Cretaceousmay have had little or no effect on the Earth'sbiosphere.The absence of mass extinctionevents following otherdocumentedcatastrophic events, such as the mid-OrdovicianMillbrig/BigBentonitevolcanic event, suggests that the Earth'sbiospheretends to be resilient to catastrophicevents and mass extinctions occur only when thereis a conjunctionof processes,each incapableof producing a mass extinctionby itself. ACKNOWLEDGMENTS I wish to thankR. M. FeldmannandD. H. Elliot for providing importantdiscussion and criticisms during the early stages of this paper.This paperwould not have been possible withoutthe generouslogistical supportof the C.E. Rinaldiand R. del Valle, E. B. Olivero,and the staff of the InstitutoAntarticoArgentino, not only for their supportduringthe 1994 field season, but for their untiringsupportover the last 20 years. Special thanksis extendedto J. Mullins and his staff at the MappingDivision of the United States Geological Survey for preparingthe remarkable topographicmaps that we used during the field season. Withoutthese maps, it would have been impossible to obtain the detail geologic data that allowed me to recognize the true natureof the "K-T glauconite." Special thanksis extendedto This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION 569 andOtherCatastrophesin EarthHistory.GeologicalSociety of America Special Paper307. DAVID, L. E. 1943. Miocene fishes of southernCalifornia.In Geological Society of AmericaSpecial Paper43, 193 p. . 1947. Significanceof fish remainsin Recentdepositsoff coast of SouthernCalifornia.Bulletin of the AmericanAssociationof PetroleumGeologists, 31(2):367-370. 1957. Fish (other than Agnatha),p. 999-1010. In H. Ladd (ed.), Treatiseon MarineEcology and Paleoecology, 2. Geological Society of AmericanMemoir67. W. R. AND D. R. SEELY. 1979. Structureand stratigraDICKINSON, phy of kore-arcbasins. In Bulletin of the AmericanAssociationof PetroleumGeologists.63: DIETRICH,G. 1935. Aufbau und dynamik des sudlichen Agulhas strombebietes.Veroffentlichungendes Institutfur Meeresforschung, 27:1-79. EASTMAN, C. R. 1903. Sharks'teeth and cetaceanbones. Albatross 1899-1900. Museumof ComparativeZoology, Memior,50:177-192. REFERENCES . 1906. Sharks' teeth and cetacean bones. Albatross 18991900. Museumof ComparativeZoology, Memior,26(40):75-98. F. SURLYK,L. W. ALVAREZ,F ASARO, ALVAREZ,W., E. G. KAUFFMAN, ANDH. V. MICHEL.1984. Impacttheory of mass extinctionsand EKDALE,A. A. AND R. G. BROMLEY. 1984. Sedimentology and ichnology of the Cretaceous-Tertiary boundaryin Denmark:implication the invertebratefossil record.Science, 223:1135-1140. for the causes of the TerminalCretaceousExtinction.Journalof SedARCHIBAD,J. D. 1996. DinosaurExtinctionand the End of an Era: imentaryPetrology,54:681-703. Whatthe Fossils Say. ColumbiaUniversityPress, New York,237 p. D. H. 1988. Tectonic setting and evolution of the James ELLIOT, ASKIN, R. A. AND S. R. JACOBSON. 1996. Palynological change Ross Basin, northernAntarcticPeninsula,p. 541-555. In R. M. Feldacross the Cretaceous-Tertiary boundaryon SeymourIsland,Antarcmann and M. 0. Woodburne(eds.), Geology and Paleontologyof tica: environmentaland depositionalfactors,p. 7-26. In N. MacLeod SeymourIsland,AntarcticPeninsula.Geological Society of America Mass Extinction:Biotic and and G. Keller(eds.), Cretaceous-Tertiary Memoir 169. EnvironmentalChanges.W. W. Nortonand Company,New York. --, R. A. ASKIN, F T. KYTE, AND W. J. ZINSMEISTER.1994. BEAVAN, G. 1946. Effect of Susquehannariver stream flow on Iridiumand dinocysts at the Cretaceous-Tertiary boundaryon Seyof on and mortalities salinities history past oyster ChesapeakeBay for the K-T mour Antarctica: event. Island, Implication Geology, 22: upper bay bars. Contributionto ChesapeakeBiological Laboratory, 675-678. 68:1-9. A. Oleinek,PurdueUniversityfor his help duringthe 1994 field season and his computergraphics skills. I would also like to acknowledgethe extensive and frankelectronicdiscussionswith N. MacLeod,BritishMuseumof NaturalHistoryconcerningthe origin of the fish bed. Althoughwe do not agree on a number of points, I found his comments very stimulatingand in come cases useful. I would like to thank S. Borg, Office of Polar Programs,for his efforts and supportof the field programand J. H. Lipps and A. J. Boucot for their reviews. Special thanks is extendedto J. Gardner,technicaleditor,Departmentof Earth and AtmosphericSciences, PurdueUniversityfor his many editorial suggestions.Lastly, I would like to acknowledgeM. 0. Woodburne.Withouthis effortpriorto the field season, it would not have been possible to accomplishmy field work program. This work was supportby OPP grantno. 94-17776. T. AND E. HAKANSSON. 1982. BIRKELAND, The terminal Cretaceous extinctionin Boreal shelf seas-A multicausalevent, p. 373-384. In L. T. Silver and P. H. Schultz (eds.), Geological Implicationof Impactsof LargeAsteroidsandCometson the Earth.GeologicalSociety of AmericaSpecial Paper 190. BLACKWELDER, E. 1916. The geologic role of phosphorus.American Journalof Science, Series 4, 42(250):285-298. H. 1929. Mortalityamong animals of the littoralregion BLEGVAD, in ice winters.Reportof the Danish Biological Station,35:49-62. der Siboga-Expedition. 0. B. 1916. Meeresgrundproben BOGGILD, Sibooga-Expeditie,79(65):1-50. BONDE,C. VON 1928. Reportno. 5 for the years 1925-1927. Fish and MarineBiological Survey,Union of South Africa, 10-85. H. H. 1984. 1984, Landsat3 RBV imageryas scale conBRECHER, trol for a topographicmap of SeymourIsland, Antarcticafrom nonmetricaerialphotographs.Surveyingand Mapping,44:253-258. BOWER,W. T. ANDH. C. FASSETT. 1914. Fishery industries. Alaska Fisheries and fur industriesin 1913. In Reportof the United States Commissionof Fisheriesfor 1913, 37-139. BRONGERSMA-SANDERS, M. 1947. On the desirabilityof research into certainphenomenain the region of upwelling water along the coast of South West Africa. Proceedingsof the KoninklijkeNederlandse Akademievon Wetenschappen,50(6):659-665. 1949. On the occurrenceof fish remainsin fossil and Recent marinedeposits. BijdragenDierkunde,Leiden, 28:65-76. 1957. Mass mortalityin the sea, p. 941-1010. In J. W. Hedgepath (ed.), Treatiseon marineecology and paleoecology, 1. Geological Society of AmericanMemoir,67. CLEMENS,W. A., J. D. ARCHIBALD,AND L. J. HICKEY. 1981. Evolution of terrestrialfaunas duringthe Cretaceous-tertiary transition. Memoirde la Societe Geologiquede France,New Series, 139: 67-74. COOGAN, A. H. 1969. Evolutionray trendsin rudist hard parts,p. N766-776. In R. C. Moore (ed.), Treatiseof InvertebratePaleontology. Geological Society of America,Boulder,Colorado. COURTILLOT, V. E. 1990. What caused the mass extinction:a volcanic eruption?ScientificAmerican,263(4):85-92. CUTLER,A. H. AND A. K. BEHRENSMEYER.1996. Models of verte- bratemortalityevents at the K/T boundary,p. 375-379. In G. Ryder, D. Fastovsky,and S. Gartner,(eds.), The Cretaceous-Tertiary Event ELLIOTT,C. C., J. L. ARONSON,H. T. MILLARD,AND E. GIERLOWSKI- KORDESCH.1989. The origin of the clay minerals at the Cretaceous/Tertiaryboundaryin Denmark.GeologicalSociety of America, 101:702-710. ENGLE,J. B. 1946. Commercialaspects of the upper Chesapeake Bay oysterbarsin the light of recentoystermortalities.In Convention of the NationalShellfishAssociation,New York,42-46. ESCHER,B. G. 1948. Grondslagender algemene geologie. 7th edi442 p. tion, Amsterdam-Antwerp, FLAKE,H. 1939. Uber recente sedimentbildungin der bucht von Concepcion(Mittel Chile). Petroleum,35(34):640-644. 1950. Das fischterbenin der bucht von Concepcion (Mittel Chile). Senchenbergiana,32:57-77. FASTVSKY, D. E. ANDD. B. WEISHAMPEL.1996. The Evolutionand Extinctionof the Dinosaurs.CambridgeUniversityPress, 460p. GISLEN,T. 1931. A survey of the marineassociationsin the Misaki districtwith notes concerningtheirenvironmentalconditions.Journal of the Faculty of Science, ImperialUniversityof Tokyo, Section 4, Zoology, 2:389-444. GUNTER, G. 1941. Death of fishes due to cold on the Texas coast, January1940. Ecology, 22(2):203-208. 1947. Catastrophismin the sea and its paleontologicalsignificance, with special referenceto the Gulf of Mexico. AmericanJournal of Science, 245:669-676. HANSEN,P. M. 1939. The age compositionof the stock of cod in West Greenlandwaters in the years 1924 to 1938. RapportProces Verbauxdu Conseil, 109:60-65. G. 1988. Anatomyand relationshipsof the bone-feedHASZPUNAR, ing limpets,Cocculinellaminutissima(Smith)and Osteopeltamirabilis Marshall(Archaeogastropoda). Journalof MolluscanStudies,54: 1-20. ANDD. R. KOLATA. 1992. Gigantic HAY, W. D., S. M. BERGSTROM, Ordovicianvolcanic ash fall in North Americaand Europe:biological, tectonomagmatic,and event-stratigraphic significance.Geology, 20:875-878. HEDGPETH, J. W. 1947. Whathappensin the LagunaMadre?Texas Game and Fish, 5:14-15, and 30. HEYMANN,D., T. E. YANCEY,ANDM. THEIMENS. 1997. Native sulfur in sedimentsfrom KT boundarysites of the BrazosRiver,Texas. Lunarand PlanetarySciences, XXVIII, 2:565-566. This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions 570 JOURNAL OF PALEONTOLOGY,V. 72, NO. 3, 1998 J. R. ANDJ. GORMAN.1988. Digging Dinosaurs.Workman HORNER, Publications,New York,210 p. B. 1988. UpperCampanian-Paleocene foraminiferafromthe HUBER, JamesRoss Islandregion,AntarcticPeninsula,p. 163-252. In R. M. Feldmannand M. 0. Woodburne(eds.), Geology and Paleontology of SeymourIsland,AntarcticPeninsula.GeologicalSociety of America Memoir 169. D. E. 1971. Plate tectonics model for the evolution of the JAMES, Andes. Bulletin of the Geological Society of America,82: A. S. 1922. Researcheson the distribution,biology, and JENSEN, systematicsof the Greenlandfishes. VidenskabekugeMeddellelser Dansk NaturhistoriskforeningKobenhavn,74:89-109. JOHANSEN, A. C. 1929. Mortality among porpoises, fish and the largercrustaceansin the waters aroundDenmarkin severe winters. ReportDanish Biological Station,35:63-97. extincJOHNSON,C. C. AND E. G. KAUFFMAN. 1996. Maastrichtian tion patternsof CaribbeanProvince rudistids, p. 231-274. In N. MacLeodand G. Keller (eds.), Cretaceous-Tertiary Mass Extinction: Biotic and EnvironmentalChanges. W. W. Norton and Company, New York. JORDAN, D. S. 1920. A Miocene catastrophe.NaturalHistory, 20: 18-22. W. ALVAREZ, ANDL. W. ALKASTNER, M., E ASARO,H. V. MICHEL, VAREZ. 1984. The precursorof the Cretaceous-Tertiary clays at Stevns Klint, Denmark,and DSDP Hole 465A. Science, 226:137143. E. G. 1984. The fabric of Cretaceousextinctions, p. KAUFFMAN, 151-245. In W.A. Berggrenand J. A. van Couvering(eds.), Catastrophesand EarthHistory.PrincetonUniversityPress, Princeton. 1988. The dynamics of marine stepwise mass extinction,p. 57-71. In M. A. Lamolda,E. G. Kauffman,and0. H. Walliser(eds.), Palaeontologyand Evolution:ExtinctionEvents, III Jornadasde Paleontologia,2nd InternationalConferenceGlobal Bioevents. Revista Espanolade Paleontologia,NumeroExtraordinario. . 1994. Commonpatternsof mass extinction,survival,and recovery in marineenvironments:Whatdo they tell us aboutthe future? p. 437-466. In G. D. Rosenberg(ed.), PaleontologicalSocietySpecial Publication7. C. E. MACELLARI, 1984. Late Cretaceous stratigraphy, sedimentol- de la Isla James Ross, InstitutoAntarticoArgentino,Buenos Aires, Argentina,p. 47-76. A. 1856-1858. Die Juraformation und OPPEL, Englands,Frankreichs des sudwestlichenDeutschland.Wurttemberg Naturwissenschaftlichen Verein Jahreshefte,v. xii-xiv (p. 1-438, 1856; 439-694, 1857; 695-857, 1858) Stuttgart. W. 1962. Aktuo-palaontologie,nach studienin der NordSCHAFER, see. VerlagW. Kramer,Frankfurt,ChicagoPress 1972 edition,translated by IrmgardOertel,567 p. SHAW,A. B. 1964. Time in stratigraphy.McGraw-HillBook company, New York,365 p. A. VON. 1923. Geologische,physikalischeundangewandte SIEBERG, Erdbebenkunde, Jena, 572 p. SIGNOR,P. W. ANDJ. H. LIPPS. 1982. Samplingbias, gradualextinctionpatterns,and catastrophesin the fossil record,p. 291-298. In L. T. Silver and P. H. Schultz (eds.), Geological Implicationof Impactsof LargeAsteroidsandCometson the Earth.GeologicalSociety of AmericaSpecial Paper 190. SMITH, W. 1817. Stratigraphical system of organizedfossils. E. Williams, Bookseller,London, 118 p. M. S. 1991. The effect on randomrange truncationson SPRINGER, patternsof evolution in the fossil record.Paleobiology, 16(4):512520. STILWELL, J. D. 1995. LatestCretaceousto earliestPaleogenemolluscan faunas of New Zealand:changes in compositionas a consequence of the breakupof Gondwanaand extinction. Unpublished Ph.D. Dissertation,Universityof Otago, New Zealand,1400p. STOREY, M. AND E. W. GUDGER. 1936. Mortalityof fishes due to cold at SanibelIsland,Florida,1886-1936. Ecology, 17:640-648. STRAUSS, D. ANDP. M. SADLER. 1989. Classicalconfidenceintervals and Bayesian probabilityestimates for ends of local taxon ranges. MathematicalGeology, 21:411-427. STUBBINGS, H. G. 1939. The marine deposits of the ArabianSea. An investigationinto theirdistributionandbiology. ScientificReports of the John MurrayExpedition,1933-34, London,3:31-158. SWEET, A., D. R. BRAMAN, AND J. F LERBEKMO. 1993. Northern and mid-continentalMaastrichtianand Paleocene palynofloristicextinctionevents.In GeologicalAssociationof CanadaProceedingsand Abstracts,p. A103. Eocene mass extinctionin THOMAS, E. 1990. Late Cretaceous-early the deep sea, p. 481-495, In V.L. Sharptonand PD. Ward(eds.), Global Catastrophesin the Earth History. Geological Society of AmericaSpecial Paper247. extinctionsin the maWARD, P. D. 1990a. The Cretaceous/Tertiary rine realm:a 1990 perspective,p. 425-432, In V. L. Sharptonand P. D. Ward(eds.), GlobalCatastrophesin the EarthHistory.Geological Society of AmericaSpecial Paper247. 1990b. A review of Maastrichtianammoniteranges,p. 519530. In V. L. Sharptonand P. D. Ward(eds.), GlobalCatastrophesin the EarthHistory.GeologicalSociety of AmericaSpecial Paper247. . 1995. The K/T Trial.Paleobiology,21:245-247. and the Cretaceous-TertiWEIDMANN, J. 1986. Macro-invertebrates ary boundary,p. 397-409. In 0. Walliser(ed.), Global Bioevents. LectureNotes in EarthSciences, Volume8, Berlin. Springer-Verlag ogy, and macropaleontologyof SeymourIsland,AntarcticPeninsula. UnpublishedPh.D. dissertation,The Ohio State University,Columbus, 599 p. -. 1986. Late Campanian-Maastrichtian ammonites from Seymour Island, AntarcticPeninsula.Journalof Paleontology,60(2 of 2):1-55. . 1988. Stratigraphy, sedimentology,andpaleontologyof Upper Cretaceous/Paleoceneshelf-deltaicsedimentsof SeymourIsland, p. 25-53. In R. M. Feldmannand M. 0. Woodburne(eds.), Geology and Paleontologyof SeymourIsland,AntarcticPeninsula.Geological Society of AmericaMemoir 169. C. R. 1995. Distinguishingbetweensuddenand gradual MARSHALL, extinctionsin the fossil record:predictingthe position of the Cretaceous-Tertiaryiridiumanomalyusing the ammonitefossil recordon SeymourIsland.Geology, 23:731-734. D. M. 1991. Bones as stones: the contributionof verteMARTILL, . 1988. Ammonoid extinction and the "Cretaceous-Tertiary bratesremainsto the lithologic record,pp. 270-292. In S. K. Donboundaryevent," p. 117-14. In J. WeidmannandK. Kullmann(eds.), ovan (ed.), The Processes of Fossilization. Columbia University Stuttgart. Cephalopods-Presentand Past. Springer-Verlag, Press, New York. J. 1927. Rezentewirbeltierleichenund ihrepalaobiologisD. 1985. Deccan traps mantle degassing in the terminal WEIGELT, MCLEAN, che bedeutung.Max Weg, Leipzig, 227 p. Cretaceousmarineextinction.CretaceousResearch,6:235-259. J. 1889. On marinedeposits in the Indian,Southernand ZINSMEISTER,W.J. 1993. It's time to look at all the evidence:a view MURRAY, of the K/T from Antarctica.Geological Associationof CanadaProAntarcticOceans. ScottishGeographicMagazine,5:405-436. ceedings and Abstracts,p. A135. -, AND RENARD,A. F. 1891. Report on the deep-sea deposits. 1994. What can the fossil record tell us about the terminal Reportof the Scientific Results of the Voyage of the HMS "ChalCretaceousExtinctionEvent and the disappearanceof the dinosaurs? lenger" 1872-1876, 525p. D. D. 1982. Mass extinction:illusionsor realities?p. 257p. 487-500. In G.D. Rosenberg,(ed.), PaleontologicalSociety Special NEWELL, Publication7. 263. In L. T. Silver and P. H. Schultz (eds.), Geological Implication . 1995. An extirpationhorizon at the K/T boundaryon Seyof Impactsof Large Asteroidsand Comets on the Earth.Geological mourIsland?GeologicalSociety of America,ProgramwithAbstracts, Society of AmericaSpecial Paper 190. C. B. ANDA. A. EKDALE. 1986. Letters.Science,234:26226(6):A143. OFFICER, . 1996. Spatialbiostratigraphy: 264. empericalapproachto approxE. B. 1992. Asociaciones de amonites de la formacion imatinglineage terminationsand evaluationextinctionpatterns.GeoOLIVERO, SantaMarta(CretacicoTardio),Isla JamesRoss, Antartida.Geologia logical Society of America,Programwith Abstracts,27(6):A249. This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions ZINSMEISTER-CRETACEOUS EXTINCTION . AND R. M. FELDMANN. 1995. Antarctica, the forgotten stepchild; a view of the K-T extinction from the high southern latitudes. Lunar and Planetary Institute, Contribution, 825:134-135. AND --. 1996. Late Cretaceousfaunal changes in the high southern latitudes: a harbinger of biotic global catastrophe? p. 356378. In N. MacLeod and, G. Keller, (eds.), Cretaceous-Tertiary Mass 571 Extinction: Biotic and Environmental Changes, W. W. Norton & Co., New York. M. 0. WOODBURNE, AND D. H. ELLIOT. 1989. Latest I, Cretaceous/earliest Tertiary transition on Seymour Island, Antartica. Journal of Paleontology, 63:731-738. ACCEPTED 15 SEPTEMBER 1997 J. Paleont., 72(3), 1998, pp. 571-576 Copyright ? 1998, The Paleontological Society 0022-3360/98/0072-0571$03.00 PRIMIlIVE POLLENCONE STRUCTUREIN UPPERPENNSYLVANIAN (STEPHANIAN)WALCHIANCONIFERS GENEMAPESAND GAR W. ROTHWELL Department of Environmental and Plant Biology, Ohio University, Athens 45701 large number of conifer pollen cones associated with primitive walchian conifers, including Emporia, occur within an exceptionally preserved fossilized biota near Hamilton, Kansas, USA. The fossils are derived from channel deposits within the Upper Pennsylvanian (Stephanian B/C) Topeka Limestone and show excellent external morphology, internal anatomy, and cuticular features. Pollen cones are cylindrical, up to 5 cm long and 1 cm wide and are simply organized. Each pollen cone consists of a nonwoody central axis from which sporophylls diverge in a helical arrangement. Vascular tissue comprises a ring of tiny cauline bundles that divide at intervals to produce a single trace to each sporophyll. Sporophylls consist of a terete stalk and a heeled distal lamina. Sporangia are attached as a cluster on the adaxial surface of the sporophyll stalk. Stomata occur in one or two bands on the inner surface of sporophyll laminae and consist of guard cells surrounded by a ring of subsidiary cells, each with one overarching papilla. The Hamilton specimens provide conclusive evidence that pollen cones wit-h a distinctive morphology were produced by some of the most ancient walchians from both North America and Europe. Contrary to traditional interpretations, these cones differ significantly from those of the Pinaceae. The broad geographic and stratigraphic distribution of this morphology reveals that conifer pollen cones similar to those at Hamilton are more widespread than previously suspected, and provides evidence for the potentially ancestral morphology of all conifer pollen cones. ABSTRACT-A INTRODUCTION A DRAMATIC increase in informationabout Paleozoic spe- cies from Europe, North America, and South America is significantly altering traditional concepts of ancient conifers of these data in understanding the origin and evolution of primitive conifers. OCCURRENCE, MATERIAL, AND METHODS (Schweitzer,1963;Miller andBrown, 1973; Clement-Westerhof, Specimens described in this study were derived from a rich 1984, 1987, 1988;Mapesand Rothwell, 1984, 1988, 1991;Win- biotic assemblage(Mapes and Mapes, 1988) preservedin and ston 1984; Archangelskyand Cuneo, 1987; Kerp et al. 1990; around abandoned limestone quarries (Mapes and Rothwell, Kerp and Clement-Westerhof,1991). These new data about 1984, 1988). The Hamiltonquarriesexpose rocks of the Topeka structure and diversity are allowing us to recognize ancestral Limestone and CalhounShale. These sedimentshave been incharacters more clearly and to evaluate more thoroughly phyterpretedto representtidal depositionin a paleovalley cut into logenetic relationships among extinct and living species (Roth- cyclothemicstrataof the VirgilianShawneeGroup(Stephanian well and Serbet, 1994). Until now this new information has had Cunninghamet al. 1993; Feldman et al. 1993). Conifer the greatest impact on our interpretationsof ovulate conifer B/C; fossils from these deposits range in preservationfrom imprescones and Roth1984, 1987; 1988; (Clement-Westerhof, Mapes well, 1984, 1991; Mapes, 1987; Mapes et al. 1989) and the anatomy of vegetative conifer shoot systems (Rothwell, 1982; Mapes and Rothwell, 1988). Ancient conifer pollen cones have been less completely known because of their rarity and generally incomplete preservation. Because conifer pollen cones are ephemeral and nonwoody, they are the conifer remains least sions that lack organic residues, to coalified compressions, to cellularpermineralizations with cuticle. The presentworkis basedprimarilyon datafrompollen cones that display internal preservedas calcareouspermineralizations anatomicaldetail as well as overall morphologyand cuticular features. Structuraldetails were determinedfrom morphology exposed on split surfaces of the rock, from internalanatomy likely to be preservedin the fossil record. revealed in peel preparations,and from cuticularmacerations. of the remarkable Late Pennsylvanian During investigation lagerstattefrom Hamilton, Kansas (Mapes and Mapes, 1988), Anatomicaland cuticularpreparationswere made by standard we have discovered an unexpectedly large number of well-pre- paleobotanicaltechniques. Serial cellulose acetate peels were served conifer pollen cones (Mapes and Rothwell, 1988). Over used for reconstructingoverall cone morphology.Featuresof 200 specimens that display a wide range of variation (Mapes vascular architecturewere determinedprimarilyfrom ground 1983; Mapes and Rothwell, 1988) represent the first reliable thin sections. Some specimenswere photographedas described evidence for the internal anatomy of Paleozoic conifer pollen in Mapes and Rothwell (1984). Additional specimen images cones. The purpose of this report is to clarify the overall struc- were capturedelectronicallyusing a Kodak DSM 100 digital ture and broad geographic distribution of currently known wal- camera,storedas TIFF files, processedusing Adobe Photoshop chian pollen cone morphology and to emphasize the importance 3.0, and printedon a Shinko CHC446i dyesublimationprinter. This content downloaded from 129.180.1.217 on Wed, 21 Oct 2015 01:47:29 UTC All use subject to JSTOR Terms and Conditions