Quaternary Science Reviews.Vol.4, pp. Zf5-278,1985. 0277-379'tt85$0.00 + .50 Printed in Great Britaln. All rishts reserved. Copyright@ 1986Pergamon Journals Ltd.
AMINOSTRATIGRAPHY OF EUROPEANMARINE INTERGLACIAL DEPOSITS
Gifford H. Miller* and Jan Mangerudf * INSTAAR and Dept. of GeologicalSciences, Campus Box 450, Universityof Colorado, Boulder, CO 80309,U.S.A. I Departmentof Geology,Sec. B, All6gt. 41, Universityof Bergen,N-5000 Bergen, Norway
Received28 January1986
With contributionsb-v:
K.S. Petersen DanmarksGeologiske Undersogelse 31, Thoravej ,2400 Kobenhavn NV, Denmark H.P. Sejrup GeologiskInstitutt, Avd. B Universiteteti Bergen, 5000 Bergen, Norway W. Zagwijn Rijks GeologischeDienst. Spaarne 17,2000 AD, Haarlem.Netherlands B. Menke GeologischesLandesamt Schleswig-Holstein, Postfach 5049, 2300 Kiel 21, F.R.G. V. Gudina Institute of Geology and Geophysics,Siberian Branch Academy of Sciencesof the U.S.S.R..Novosibirsk-90. U.S.S.R. R. Paepe Quaternaryand MarineGeology, Belgische Geologische Dienst, Jenner- straat,13, 1(X0Brussels, Belgium V. Yevzerov Instituteof Geology,Kola Branchof the Academyof Sciencesof the U.S.S.R.,Apatity, U.S.S.R. J.M. Punning Institute of Geology,Academy of Sciencesof the E.S.S.R.,Tallinn 200001.Estonia 7. U.S.S.R. A. Makowska InstytutGeologiczny, ul Radowiecka4,00-975 Warszawa, Poland T. Pisse GeologicalSurvey of Sweden,Kungsgaten 4, 4lI 19 Gciteberg,Sweden K.-L. Knudsen Laboratorietfor Palaeontologiog Stratigrafi.Geologisk Institut, Aarhus Universitet,8000 Aarhus C, Denmark A.-L. Lykke-Andersen Labyrinten 17, 8220Brabrand, Denmark T. Meijer Rijks GeologischeDienst. Spaarne 17..2000 AD, Haarlem,Netherlands J. Somm6 Laboratoirede G6omorphologieet d/Etudedu Quaternaire,Universitd des Scienceset Techniquesde Lille, 59655Villeneuve d'Ascq C6dex, France G. Linke GeologischesLandesamt Freie und Hansestadt Hamburg, Oberstrasse 88, 2000Hamburg 13, F.R.G. C. Baeteman Quaternaryand MarineGeology, Belgische Geologische Dienst, Jenner- straat,13, 1040Brussels, Belgium W. Hinsch GeologischesLandesamt Schleswig-Holstein. Postfach 5049, 2300 Kiel 21, F.R.G. D. Goter INSTAAR, CB 450.University of Colorado.Boulder, CO 80309,U.S.A. J. Brigham-Grette Departmentof Geology,University of Alberta,Edmonton, Alberta. T6G 2E3. Canada J. Hollin INSTAAR, CB 450,University of Colorado.Boulder, CO 80309,U.S.A.
2t5 216 CiffordH. Millerand Jan Mangerud
Molluscanfossils collected from shallowwater marine sediment across NW Europeand havebeen analysed for the extentof isoleucineepimerization (D/L nearbyArctic regions 'classical' ratio)in indigenousprotein residues. The D/L ratiosconfirm that essentially all EemiansiteJ from NW Europeare of the sameage. and are correlativewith the type localitynear Amersfoortin the Netherlands;shells from interglacialmarine sediment beneaihthe typeWeichselian till in Polandalso correlate with the typeFemian sile. D/L ratiosin Holiteinianmarine shells (0.29) are substantiallyhigher than in their Eemian counterparts(0.17);'Late Cromerian'shells yield even higherratios (0.46). D/L ratiosin lateglaiial shells (0.06) and Middle Weichselian shells (0.09) permit differentiation from modirn (0.01)and last interglacialmaterial. Based on the positionof the Matuyama- Brunhesboundarv and the differencesin D/L ratios,the Eemianmust correlatewith isotopesubstage 5e, whereasthe Holsteinianis mostlikely substage 7c, possiblystage 9 but iertainly youngerthan stage 11. Intra-Saalianwarm periods may be terrestrial equivalentsof the youngersubstages of stage7. Extensive.pre-Eemianmarine sediments albngthe SW coastof Denmark previouslycorrelated with the Holsteinianare shownto be oi'Late Cromerian'age.The underlyingtill thereis thefirst widespread evidence of a pre-Elsteriantill in NW EuroPe. D/L ratios in moiluscsfrom last interglacialsites along the Arctic coastof the USSR. the Arctic Islandsand easternGreenland are substantially lower than in theirEuropean counterpartsdue to their low thermalhistories. The combinedmid- and high-latitude data are usedto developa predictivemodel for the expectedDiL ratio in any of several moderateepimerization-rate taxa for last interglacialsites with mean temperatures between-20 and +15'C. Not all sites could be unambiguouslyassigned to an establishedinterglacial. The Fjosanger(Norway) and Margareteberg (Sweden) sites previously_thought to be Eemian. yietOOll ratios higher than in securenearby Eemian material. It is yet unresolved whetherthese are aberrantsites or if they predatethe last interglacial.1n situ shoreline depositsencountered in boringsin SW Belgiumand in expo_sureson the Belgiumcoastal plain contain molluscsthat yield D/L ratios intermediatebetween secure Eemian and Late Weichselianratios, raising the possibilitythat'a late stage5 high-sea-levelevent attained near-modern levels in the southern North Sea basin. Resolution of these uncertaintiesis the focusof future work.
INTRODUCTION
Despitenearly a century of investigation,the Quaternarystratigraphy of Northwest Europehas yet to be fully resolved.The classicalland-based sequence in NorthwestEurope (Table1) is basedon end-morainesystems and correlative till beds.and the palynologyand faunal assemblagesin associatedterrestrial and shallow-watermarine sequences.Three mainglacial stages are widely recognized: the Elsterian,Saalian and Weichselianseparated by the Holsteinianand Eemianinterglacials. Over the last decadeseveral pre-Elsterian climatic oscillationshave also been described(e.g. Menke, 1970; Zagwijn, I9151. Brunnackeret at..1982 Urban, 1982;Nilsson, 1983), but their occurrenceis relativelyrare andregional correlation remains somewhat ambiguous. The Matuyama-Brunhesboundary 'Cromerian has been identified in the lower portion of the complex'(Zagwijn. 1975; Brunnackerand Boenik, 1976 Zagwijn and Doppert, 1978; Brunnacker1979). The number of glacial-interglacialcycles in Europe betweenMatuyama-Brunhes and the Elsterianis uncertain,but seldomare more than three proposed, giving altogether six after the Matuyama-Brunhesboundary. In contrast,the deep-seaoxygen-isotype stratigraphy (e.g.Shackleron and Opdyke,1973) implies that there have been eight glaciations since the Aminostratigraphy of European Marine Interglacial Deposits 2t7
TABLE 1. Classical subdivision of the middle and late Quaternary of NW Europe
Holocene (lnterglacial)
Late Pleistocene Weichselian (Glacial) Stage Eemian (Interglacial)
Middle Pleistocene Saalian (Glacial) Stage Holsteininian (Interglacial) Elsterian (Glacial) Stage Probably Four Temperate and Three Cold StagesCollectively Called the 'Cromerian Complex'
Earlv Pleistocene Six named stages
Matuyama-Brunhesboundary. The apparentdiscrepancy between the deep-seaand land- baseclstratigraphy has stimulated additional investigations into the continentalrecord, and the correlationand dating of Quaternarydeposits in northwesternEurope has been vigorouslydebated in meetingsand papersin recentyears. At presenta widelyheld view is that the classicalpost-Elsterian interglacial sequence is probably correct (e.g. Ehlers et al., 1984),and that knowledgeof the pre-Elsterian 'missing' sequ"nceis so fragmentarythat it may accommodatethe interglacials.In this interpretation,the post-Eemianwarm intervals(e.g. St. GermainI and II of Woillard, 1978) are consideredto be interstadialswithin the Weichselian,and the Wacken/ Dcimnitzianand Holsteinian(seruu stricto) are consideredto representa singleinterglacial complex,the Holsteinian(sensu lato). The inescapableconclusion from this perspectiveis thatisotope stage 1 is the Holocene,2-5d the Weichselian,5e the Eemian(e.g. Shackleton, 1967;Mangerud er a\.,1,979),stage 6 the Saalian,stage 7 the Holsteinianand stage8 the Elsterian. Opposingthis interpretationare severalrecent proposals that more than three post- Elsterianinterglacials occur. Kukla (1977),followed by Bowen (1978,1979) argued that sitesclassified as Eemian may representmore than one interglacialthat havebeen grouped as a singleevent. The type Eemiansite in the Netherlandswas proposed as the oldestof theseinterglacials (Kukl a, t977). In their scheme, the Holsteinian(sensu stricto) is ascribed to isotopestage 11 or 13rather than stage 7 asin a classicalinterpretation. Earlier Frenzel (1973)had arguedthat the pollenstratigraphy of interglacialsites in Germanycould not be accountedfor by the classicalinterglacial sequence and additionalinterglacials must be present;he proposedtwo intra-Saalian(sensu lato) interglacials.In East Germany,Cepek and Erd (e.g. Cepekand Erd, 1982;Cepek , 1967)have long claimedthat at leasttwo and possiblythree interglacialsoccurred between the Holsteinian(sensa stricto) and the Eemian:the Domnitzian,an un-namedSaalian I/II event,and the Rtigenan.Similarly' Stremme(1982) and Sarntheimand Stremme(1984) proposed correlation of the Eemian with isotopesubstage 5e, the Holsteinianwith isotopestage 11, with two intra-Saalian interglacialsas the terrestrialcorrelatives of isotopestages 7 and 9, a view supportedby Bowin (1985).The lack of reliableindependent dating methodologies that canbe applied to terrestrialsites has hampered resolution of this conflict,although new effortsin U/Th. electron-spinresonance and thermoluminescencedating are in progress. 2rB Gifford H. Miller and Jan Mangerud an In 1979,partially in responseto the challengeof possiblemiscorrelation. but alsoin interglacial attempt to obtain an integratedchronostratigraphy of the Europeanmarine of isoleucine ,.qu"nr". we inrtiatedu p.ogtu* to determinesystematically the extent along epimerizationin marinetnollut.t from interglacialsites across Northwest Europe and 1000 the Arctic coastof the Soviet Union. This paper presentsthe resultsof nearly isoleucineepimerization determinations in marine molluscshells f;om interglacialand the interstadialsites over this region, and also gives a review and appraisalof geochronologicalinterpretation of thesedata' Theprimaryquestionswewishtoaddresswiththesedataare: which (1) Are all sitesconsidered Eemian clearlyof a singleage? If they are not, representthe last interglacial? (2) Can we identify marineinterglacial beds intermediate in agebetween Holsteinian and Eemian? (3) Can the amino acid data be usedto test regionalcorrelation of Middle Pleistocene interglacialsin NW Europe and can it be usedto placeconstraints on the possible agesof theseevents? 5c (4) Are thereany indications that the highsea-level events of isotopesubstages 5a and are preservedaround the North Sea' in two The analyticalresults and an appraisalof the questionsposed above are presented controlling parts. Part 1 containsun on"rui"* of the methodologyemployed and the of parametersthat governhow the datacan be interpreted,and it includesthe evaluation site-specific ihe analyticaldita from a regional perspective.In Part 2, more detailed the basisfor informationts presented, including the primarystratigraphic relationships and all taxa from the proposedage assignmentfor itre sitis investigated.Amino acid ratiosfor primarily local each site are tabulaLd on the accompanyingtables. Conclusionsof significanceare givenwith the site descriptions' Aminostratigraphy of European Marine Interglacial Deposits 219
of part I - Amino Acid Geochronology, Regional Chronostratigraphy and Absolute Age EuropeanMiddle.andLate.PleistoceneMarinelnterglacials
AMINO ACID GEOCHRONOLOGY
The use of certain reactionsinvolved in the degradationof indigenousproteins and their constituent amino acids preservedin fossilsto estimate the age of the enclosingsediment (Amino Acid Geochronology) has gained increasing acceptancesince the pioneering to itudi"r of Hare, Abelson and colleaguesin the late 1960s.In particular, the application work marine carbonate fossils(see Wehmiller, 1983,for review) has been widespread,but hasalso been carried out on bone (e.g. Bada, 1985)and wood (Pillans, 1983;Rutter, 1984). Most studies rely on the fact that protein produced by organisms consists almost exclusivelyof amino acids in the L-isomer configuration. As the protein is degraded after death of the organism, the L-amino acids invert to their respectiveD-configurations; the inversion rate is dependentprimarily on temperature, and to a lesserdegree on taxonomy. Measured DiL ratios can be directly compared ('Aminostratigraphy', Miller and Hare' 1980) only for monospecific samples that have experienced similar (+1'g; thermal histories. ln this study, we have relied on the epimerization of the protein amino acid L-isoleucine to its non-protein diasteriomer,D-alloisoleucine. The ratio of thesetwo isomersis generalh' abbreviated as aIIe/Ile or D/L ratio; throughout the remainder of the text we use the D/L abbreviation.The ratio of D- to L-isomers increasesfrom near zero in a modern shell to an equilibrium ratio of 1.30 + 0.05, at which time the number of L-amino acids inverting to their D-configuration is balancedby the opposite reaction. The time required to reach an equilibrium siate is related to the effective diagenetictemperature (EDT) experiencedbY the fossil; at mid-latitude sites (e.g. 10"C) it takes ca. 2 Ma to attain equilibrium. whereas Arctic sites (<-10"C) will require 20 Ma or more to reach the same state.
ANALYTICAL PROCEDURE
Indigenousproteinaceous residues and associateddegradation products are releasedon decalcificationof the shell in acid; peptideremnants are further decomposedinto therr constituentamino acidsby laboratoryhydrolysis in 6N HCI f.or22hr at 110"C.All amino aciddeterminations reported herein follow the standardanalytical procedure in effectat the Coloradolaboratory since the springof 1982.From mostcollections three to five individual shellsof a singletaxon were prepared for analysisfollowing the procedure outlined in Miller (1985).Sepaiation and detecrionof individual amino acidswas accomplishedon an automated-high-pressure liquid chromatographutilizing cation-exchangeresin. post- columnderivatization with O-Phthaldehydeand fluorescencedetection. The ratio of D- alloisoleucineto L-isoleucine(D/L) in the total aminoacid population (free plus peptide- bound amino acids)is determinedby a recordingintegrator based on the ratio of peak heights.Each analysis requires less than 2 mg of shellmaterial.For thisreason, individual sheilscan be analyzedseparately for nearly all taxa; the only exceptionbeing the small gastropodBittium reticulatum,for which up to four individualswere combinedin some instances. 220 GiffordH. \liller andJan Mangerud
To ensurethat determinations made throughout the durationof thisproject were directlv comparablethree techniqueswere employedto monitor analyticaluncertainties: (1) a calibrationstandard; (2) an interlaboratoryshell standard; and (3) multiplepreparations of selectedsamples. Two nearlyidentical amino acid analyzers are in routineuse at the Coloradolaboratory and a similar instrumentis used in the Bergen laboratory.An amino acid control is processedon each analyserevery day. BetweenJune 1982 and December1983, the measuredD/L ratio in the controlaveraged 0.122 + 0.011(n : 123)on oneinstrument and 0.123+ 0.012(n : 234)on the other at the Coloradolaboratory. The Bergenlaboratory reportssimilar precision. Additionally, some samples have been processedat both the Coloradoand Bergenlaboratories with similarresults. In 1982J. Wehmiller circulatedsized shell powders of monospecificmollusc samples from depositsof three very differentages to all laboratoriesinvolved with amino acid geochronology(Wehmiller, 1984).The resultsof four seriesof analysesat the Colorado Lborutorl'and a singleseries from the Bergenlaboratory (Table 2) indicatethat therehas beenno statisticalshift in ratiosdetermined over the durationof the projectand that the two facilitiesproduce comparable results. As a finaltest of analvticalreliability, rve analyzed several samples more than once during the program.Similar results for samplesprepared near the startof the analyticalprogram and recentlvconfirm that there are no significantanalytical problems.
FACTORS AFFECTING THE MEASURED D/L RATIO IN MOLLUSCAN SAMPLES
Time The degreeto which a chemicalreaction has proceeded toward equilibrium is relatedto the time elapsedsince the onsetof the reaction.In the caseof the isoleucineepimerization reactionin a molluscanshell, the reactioneffectively starts with deathof the organismand proceedsro racemicequilibrium (D/L : 1.30). Measuringthe D/L ratio in a mollusc providesan indexto the elapsedtime sinceits death. The generalform of the relationshipbetween D/L ratio and time for a specific temperalureis distinctlynon-linear. As shownin Fig. 1, the conversionof L- to D-amino acidsproceeds most rapidly during the earlystages of diagenesis,decreasing in the later stageswhenthe forwardreaction is counterbalancedby the reversereaction. Consequently, the resolvingpower of the D/L ratioin shellsfrom aroundthe North Seawill be greatestfor youngsamples (modern through Holsteinian) and diminishfor older shells.
Temperature The rare of most chemicalreactions is stronglydependent on the temperatureof the reactionmedium. In this studyit is the temperaturein the sedimentenclosing the molluscs (calledthe effectivediagenetic temperature [EDTI) that governsthe epimerizationrate. Note that the temperatureat whichthe shelllived is inconsequentialin thisrespect. A shell Aminostratigraphy of European Marine Interglacial Deposits 221
calibration TABLE 2. Comparison of D/L ratios determined on the interlaboratory standard (Wehmiller, 1984)
(Total fraction) Sample ID AnalYsis date Lab ID D/L (1) 81 ILC A July i982 AAL-2694 0.1s1 October1983 AAL-3513 0.158r 0.00s (3) Ma1'1984 AAL-3832 0.1s8+ 0.005(2) Sept 1985 AAL-4510 0.16s+ 0.008 (3) July 1984 BAL. 334 0.167+ 0.003(3)
81 ILC B July 1982 AAL-2695 0.s4 (1) October1983 AAL-3514 0.s4 + 0.0s (3) May 1984 AAL-3833 0.49 + 0.02 (2) Sept 1985 AAL-4511 0.50 + 0.03 (3) July 1984 BAL- 335 0.54 + 0.02 (3)
81 ILC C July 1982 AAL-2696 1.08 (l) October 1983 AAL-3515 1.16 + o.o4 (3) May 1984 AAL-3834 r.06 + 0.02 (2) Sept 1985 AAL-4512 1.08 + 0.01 (3) Jul.v1984 BAL. 336 1.20 + 0.05 (3)
NW Europe FIG. l. Simplifieddiagram showing the increase in D/L ratioin a molluscanfossil from whereasthe over the lasi t3Oka. T-hefastest epiinerization rate occurs during the warm interglacials, reduced. ;pi;;;ir;ii." rate during the coolerinrervening stadials and interstadialswas substantially 222 Gifford H. Miller and Jan Mangerud rapidepimerization during that died earlyduring the lastinterglacial will haveexperienced during the coldestphases of the the rest of the warm interglacial,much reducedrates and a rapidrate againduring Weichselian.rntermediat"."ut", during the mild interstadials theHolocene(Fig.2).Thetemperaturedependencyofisolzucineepimerizationisan of the rate for each 4'c exponentialrelationsirip, with an approximatedoubling temperaturerise. Clearly'quantificationofthethermalhistoryforeachsiteisessentialforthe dataare generallyinsufficient interpre;tion of the measuredD/L ratios,yet paleoclimatic of the EDT for eachsite to assessadequately the thermalfactor. As a first approximation At mid-latitudes'mean annual we haverelied on the currentmean annual air temperature' (e.g..Gieger,1965). Implicit with this air and groundtemperatures are in closeagreement occurreduniformly across upp.ouJhis the assumptionthat Pleistocenetemperature changes NW Europe. complicationsin the This simplifiedapproach to sitetemperature may obscure important have-experienced significantly thermal history of specificsites. Even nearbysites may shellswere collected from ice- differentthermal histories. For example,some of the Danish havebeen below sea-level for thrustedfloes deposited in Late weichseliantime, and may for such collectionscould be most of their history. The integratedthermal history shortlyafter their deposition significantlydifferentihan for neaibySites that wereuplifted sites that were coveredby the and have remainedabove sea level ever since.other nearbysites outside the ice continentalice sheetmust have been closeto 0'C, whereas acrossthe Netherlands marsin mav havebeen considerably colder. Periglacial conditions
Species'A'
"B' ----- Species
= v.iv G t ) - 0.50
200 400 600 1000
Trme (ka)
two molluscantaxa at a constanttemperature FIG. 2. Theoreticalpattern of increasein D/L ratiofor of ca.l0"coveramillionyearinterval.Theratioincreasesnearlylinearlyoverthefirstfewhundred returningto a linearrate about %oth that of thousandvears after whicr, irr. ,.i" a".r"ases.eventually (A) taxa aie includedto illustratehow the initial period. Hyporn.ifu"f'l^i-iet and slow the time intervals' differencern D/L ratio betweentaxa can vary over long Aminostratigraphyof European Marine InterglacialDeposits L.L-)
and portions of West Germany at the glacial maximum resultedin lower temperaturesthan at ice-coveredsites in Scandinavia,thus reversingthe present-daythermal gradient. Finally. the thermal gradient acrossNW Europe during interstadialsor full glacial conditions mav have differed from the current gradient. However, it is important to keep in mind that most 'warm' of the epimerization occursin the periods; full glacialconditions are not as critical to the average reaction rate as are interstadial and interglacialregimes.
Taxonomic Effects As has been noted previously, the epimerization rate can vary between taxa. The reason for the different rates has not yet been demonstrated,but is probably in part related to the position of isoleucinein the peptide chains. The difference in epimerization rate between taxa may also change over time. The theoretical relationshipsare illustrated in Fig. 1. showing an initial difference in D/L ratio of a factor of 2 for two hypothetical taxa. decreasing over time until the epimerization rates are the same in the later stages of diagenesis. Such changes in relative epimerization rates are unlikely to alter the interpretation of samplesof Holsteinian and younger ages,but may become important in evaluating ratios from older deposits.
Becauseof the wide range in paleoenvironmentalconditions that characterizethe various localities, there is not a single taxon that is common to all sites. In our search for appropriate taxa, amino acid determinations were undertaken on nearly 30 species of marine molluscs.Some of these taxa were not useful for geochronologicalstudies. in some casesdue to the restrictedoccurrence of the taxon either temporally or spatially, whereas for other taxa the measuredD/L ratios were not reproducible.
The main criteria for evaluatingthe reliability of a taxon are empirical. In most instances. at least three individualshave been analysedfrom eachlocality. If the measuredD/L ratios scatterwidely (coefficientof variation greater than 15 to 20'/. of the mean), that speciesis potentially of questionablereliability. If a taxon yields similarly scatteredresults at several localities it is considered to be of low reliability. Another factor is the concentration of protein within the shell matrix, which can vary by a factor of 10 between genera.Those taxa containing the lowest levelsof protein contain so little alloisoleucineand isoleucinethat the instrumental detectionlimit is approachedand the precisionof the measurementdecreases.
Proteins in different structural layers of a shell frequently differ in their amino acid composition and rate of epimerization. Taxa with multiple structural lavers must be sampledfrom the samelayer to ensurereproducible results.In thick-shelledspecies such as Arctica, this can usually be accomplished easily, but thin-shelled taxa cannot alwavs be confidentially separated.This may be a contributing factor to the scatter in data obtained from the often diminutive shells of Cardium, an ecologicallydiverse taxon that is often a major constituent of Eemian marine deposits.The failure of Cardium to yield reproducible D/L ratios was a major disappointment of this study. Shells of relativelv homogenous structure such as Mya or Hiatella, require less rigorous sampling (cf. Brigham. 1983). although we routinely sample from the center of the shell near its apex to minimize intra- shell variation. Shells with multiple structural layers are sampled from the inner layer. which is senerallv denser and lesssusceptible to leachinethan the outer laver. 114 Gifford H. Miller and Jan Manserud
Our evaluation of taxa analysed in this study is given in Table 3. There is a group of severalpelecypod taxa for which isoleucineepimerization proceeds at a similar, moderate rate. Manv members of this group are widely distributed and have moderate to high reliability. The core of the interpretations of this project are based on results from five representatives of this group: Arctica islandica, Macoma balthica and Macoma calcarea, Venerupis senescensand Mya truncata. Important taxa that yield reliable ratios but have a significantly different epimerization rate include Littorina (several speciesbut principally littorea) which yields D/L ratios nearly half those in the moderate-rate group and Divaricella divaricatawhich has an epimerizationrate about 1.5 x faster than in moderate- rate taxa. Additional speciesin the moderate-rategroup that provide important control for some sites include Hiatella arctica, Corbula gibba and Spisula subtruncata.
Other factors Although the only establishedvariables that control the measuredD/L ratio in a shellare time, temperature and taxonomy, there may be other less obvious environmental parametersthat can alter the apparent epimerizationrate. The carbonate matrix effectively buffers the systemfrom local pH changes,and moisture content is not a limiting factor even under high-arctic conditions. Pressure(e.g. beneath a thick ice sheet) does not seemto be significant as there is no change in volume associatedwith the epimerization reaction, although hydrolysis, involving a volumetric increase, may be impeded. A possible explanation for acceleratedepimerization could the incorporation of impurities, such as radioactive elements, within the shell matrix that catalyze the epimerization reaction, although none have been identified yet in sufficient concentrations.The leaching of low molecular weight free amino acids,most severein the outer portion of a shell, can lower the D/L ratio in the total amino acid population. The sampling methods employed in our laboratories minimize this effect.
EPIMERIZATION DURING THE EEMIAN
Becauseof the sensitivityof the reaction rate to temperature, it is important to evaluate the extent of epimerization that occursduring an interglacial,the warmest and, hencemost rapid epimerization rate in the post-depositionalhistory of a sample. The D/L ratios in 1l to 13 ka old shells of the moderate-rate taxa range from 0.05 (Norway) to 0.07 (Scotland). Macoma balthicacollected from marine sedimentsin the Netherlandsdated 4500BP have a DiL ratio of 0.035 t 0.002 (3). When extrapolatedto 10 ka BP this givesa D/L ratio of 0.07. Subtractingthe D/L ratio in a modern shell givesa net epimerization of between0.035 and 0.055 during the Holocene. The Eemian was of similar duration to the Holocene (Mriller, 1974a), and at its thermal maximum was at least two degreeswarrner. Thus at least the same amount of epimerization is to be expected for Eemian time as occurred durine the Holocene.
Sites that span an interglacial should demonstrate whether shells from late in the interglacialyield ratios significantlylower than from early in the interglacial.Two suchsites are the Schnittlohe boring (w. Germany) and the Fjosangerexcavation (Norway). Aminostratigraphy of European Marine InterglacialDeposits 225
r - tr ) I r€€C N Or € - r -N F-*N N NN N _iN F * _J= 'i- d
-,d
!e o= 99999 I g 99 99 999 -G G666€ 6 qJ €c d€ aZ !LL!!L 666 =-fo'*i{{i{'-{ =i >ii--ii *= =riii => E i€ =r=ji>===ji5ji5=== i.::i===
o '=-x3 99999 6 6 c € ^€ ^"t .,o = d =s .=.: a Cu, 9 af,Ei -. -. -E:< : iaa€z:<++= EE!- Y- -'-&<*Z**-j;+++-+Z+ -_2++++
9
r+++a SFio -! !oo0 ooo ooooo6o : I LO ue OL - € s G G 6 6 6 F G < - € 6 ^- - ^,!): a a.! L Z= ,^, o o O O O O O O A O O O q O ra a =p!! =1r!__!!":=!! 3]!-3'' ::.:9 gt ;;=;*ESStrtr=S:;;iS* t,iii ;;3t3'-i - > ., c !si' :-i9
L>C: Y9 .,7 0 e :a :::tr '-E - € E : )!€ z ="E.E '- Z .2 € c-- '--l E o ::;_ ;* ! =a;: a ?=}'i.-5 j \ cs ^, - s { iI E =3 | \ : !l !9=+ n a: \sc SS i !E!.!SS FF lJ.s.c -! ,!.! .!,! I v ZFFr>FF Vv !J i c.9O a-9= =z:"o ae o 9-:: o ?= Y == !E iE 3 a --,@ #EI - e , a? = ^ O=aq =- ZA,-R J a;^E E3 -):.Y o P i :lE r -ii;g3F":^5 € i- :i= !=E E O E.g F u.= i. x v) '3>-:5 i?:;siiiiEi:.iri!E Eii = i*i a.Y a =>6e :is $lFi sEFi€ -\- c , :siSSsssi:sss:s:s;SfiiS;s iE:issF!!ir;;i 226 Gifford H. Miller and Jan Mangerud
The Fjosanger site essentiallycontains a complete sequenceof marine depositsfrom the onset through to the termination of an interglacial episode. We have analvsedthree taxa from Fjosanger that span the interglacial(Mya, Hiatella andLittorina. Table 4). D/L ratios from samples collected respectivelv at the base and top of the interglacial beds give conflicting information on the epimerization that occurred during the inierglacial period. Mva and Littorina show a clear decreasein D/L ratios up-section.whereas ratios in iliatella are similar throughout. The Schnittlohe I boring penetrated more than 5 m of Eemian marine sediments. Samplesof V. senescensand Corbula gibba were obtained from levelswithin the upper part of the Piceallowerpart of the Abies pollen zone (13 m), the base of the Tiliaron"'1tS'n,,;, while the 16 m sample predates the Corylus zone. Mean D/L ratios (Table 4) do not shorv the expected down-core increase,and the deepest sample gives the lowest ratio for both taxa. However. at the 1o confidence level the ratios are almost undifferentiable. The lack of a consistent increasein ratios through both sitesis surprising. Although the precisionof the data is lessthan expected,the resultsdo indicatethat D/L ritios in -Jllur., from a single interglacialprobably wilt include a range of rarios.The typical range of values may be between 0.03 and 0.06, and is caused both by 'warm' the differences in the duration of interglacial conditions experienced by different samples, and possibly orher unidentified factors.
TABLE4 MeanD/Lratiosinselectedtaxafromtwo-sitesthatspanmostofaninterglacial.Seetextfordiscussion
Fjosanger(Norway) Schnittlohe(West Germanv)
Bed MYa Hiatella Littorina Depth in core Venerupis Corbula
0.l9 G 0.22 H 0.11 m U I 0 0.14 m 0 0 J 0.12 K 0.14 16m 0 l6 U t6 L 0.14 M 0.28
RESOLVING POWER OF THE AMINO ACID DATA Of fundamental interestto usersof aminoacid data is the levelto whichdeposits can be separated.Because of the non-linearityof the reactionrate, the isoleucineepimerization reaction has greater resolvingpower for young samples-ot than for older samples.For Holocene sitesat temperaturesaround 10'C, the il*itr separationare probabtyabout 3 ka. This capabilitvdecreases rapidly for Weichseliansamples, largely bicause ih" lorr,., temperatures during the Weichseliansubstantially diminished epimerization rates. In Danishsamples, for instance,the measuredDiL ratioincreases from 0.013(modern) to ca. 0'055 by l2 ka in the moderate-rategroup, but increasesonly to ca. 0.085in shellsdated >'10ka. More epimerizationoccurred during the last 12ka than in the preceding30 ka or Aminostratigraphyof EuropeanMarinc Interglacial Deposits 22'7 more. Consequently, resolution within the Weichselian is probably no better than 20 to 30 ka.
For older sites, a useful rule of thumb is that two sites of similar thermal history are of significantly different age if their mean D/L ratios do not overlap at two standard deviations. In most initances there should be no ambiquity between Eemian and Holsteinian sites,but if there is an intra-saalianinterglacial, it might be difficult to separate it clearly from either the Eemian or Holsteinian. Pre-Holsteinian interglacialsprobably require at least a 100 ka separation,and by early Pleistocenetime, the resolvingpower of in the data is undoubtedly even further reduced. One of the primary contributing factors improving the resolvingpower of the data is the number of individualsmeasured. Based on thii study we suggesta minimum of five different individuals are required to characterize adequatelyu tp".ifi. stratigraphiclevel. The resolvingpower essentiallyincreases with the ,quu.. root of iire number of samplesanalysed (c/. Miller et a\.,1979). An increaseto 20 in the number of individualsprocessed would double the resolvingpower. Such an effort may be justified in certaininstances.
PRESENTATION OF THE DATA
In the following discussions,amino acid data are restrictedto wideli, distributed taxa of Divaricella and the the moderate-rate group. and the relatively fast-epimerizingbivalve to three significant siow-epimerizinggastropod, Littorina. The D/L ratios are measured figures and are io-pr.t"nt.d in the site tables and figures. However. inter-site variation ts ,ulh thut for regional correlations the ratios are cited to two significantfigures only. D/L ratios are presentedin the format 0.19 + 0.02 (5), where the first number is the mean ratio the of all individual shell preparations of a given taxon from the site(s) being discussed. second number is the standard deviation of the measurements and the number rn parenthesesrepresents the number of individual preparationsanalysed. ln most caseseach analysisis of adifferent individual. but in collectionsfor which the sample size is limited. this may include more than one analysisof some individuals. For a verl' feu' sitesonly one individual was available from which up to four separate preparations ma,v have been of analysed. Such data may be less representativethan ratios based on a similar number be analysesof different individuals, although frequently as much intra-shell variation can expected in the measured D/L ratio as inter-shell variation (e'g. Brigham. 1980)'
REGIONAL CORRELATION OF INTERGLACIAL AND INTERSTADIAL DEPOSITS ACROSS THE CORE AREA: THE NETHERLANDS. WEST GERMANY, DENMARK, SWEDEN AND POLAND
The Eemian marine depositswere selectedas the major focusof our studybecause thev havebeen well studiedfor nearlya century and are geographicallywidespread. We accept the boringsalong the Eem River nearAmersfoort, Netherlands (Zagu'ijn' 1961)as the ty'pe localitl,oithe Eimian and all other Eemian sitesare to be comparedto the ratiosobtained from the type localitY. 228 Gifford H. Miller and Jan Manserud
To test the controversywhether more than one interglacialis representedby deposits consideredEemian. we began.with the hypothesisthat all Eemiansites are of a singleage. within the restrictedarea of the Netherlinas, NW Germanyand Denmark,the dJnsiryof sites is sufficientlyhigh that site-by-sitecorrelation is possibledespite the decreasing thermalgradient across this region. All sitesproposed to be Eemianihould yield similar ratios with a modesttrend of decreasinguuiu.r in Germanyand Denmark. Siteswith anomalouslyhigh (or tolv) ra119sare potentiallymis-correlaied. outside of this limited geographicregion, possibledifferences in the integratedthermal history of the sites complicatesthe amino-acid basedcorrelations. However, we do expectto find a coherent pattern of changing D/L ratios that can be clearly correlatedwith a realistic gradient. thermal
q =-\F\q}. (\ )2 \^5 o \n\d ,_4v-,---'rV € l/edjtenanean\ sed
FIG. 3 Regional location.map ind,icatingsites studied in this survey outside the core area of Belgium. Netherlands,NW Germany and Denmark. Aminostratigraphy of European Marine InterglacialDeposits 229
FIG. 4. Detailedlocality map of NorthwestEurope showing sites from whichsamples have been analysed.Sample abbreviationsare as follows. Sweden:Margarereberg (Mb). Denmark: Skarumhede(SK). Hirshals (HI), NorreLyngby (NLlKis HoveO(kH). Hos*trup(Hp), Vognsboi (vo), Esbjerg(ES). Rogle (R@), Srettenshage(SE), Hormstrup(Ho), Strandeg""roisoJ, Stubberuphave(SB). _HjelmNakke (HN), RistingeKlint (RK). TrappeskovKlint (TKi. Mommarli (MO), Stensigmose(ST), Hojer (H@),Tonder (T@). West Germiny: Holnis(HS). Stidermarsch (SM)' Heide(HH), Rodemis,Offenbiittel and Schnittlohe(SL), Schirhorn(SH), Wacken(WA). Hummelsbirttel(HB), Billbrook (BB), Halstenbek(HT). Netherlands:NoorO eergum'(NBj, Bergen(BN). castricum(CA), ?underdoorp(ZU), Amerfoorr(AF). Belgium,z..tr-ugg. izgj. Koolkerke(KK), Sr Anroine(SA). France:Herzeele (AZ).
Eemian Type Sile and Adjacent Correlatives in The Netherlands
The type Eemian sediments were deposited at the head of an estuary near the maximum inland penetration of the Eem Sea into the Netherlands (Fig. 4) and the euryhaline fauna is relatively impoverished in the open-ocean normal salinity taxa that are found at manv other sites. The only representatives of the moderate-rate group are V. senescensfor which ten preparationsgave a mean D/L ratio of 0.20 + 0.03 (Fig. 5) and a few diminutive shellsof ff. + arctica that averase0.19 0.02 (4). The nearby Bergen boring, correlared by its similar pollen stratigraph\', penetrates more open-marine sediments that contain additional moderate-rate taxa. The D/L ratio in four Veneruplssamples substantiates the correlatron to the type Eemian locality. D/L ratios in moderate-rate taxa from nearby borings at Castricumand Zunderdorp are similar to the samegenera at the Bergen Boring (Fig.5). The amino acid data thus confirm the previously establishedcorrelation of theie borings with the type locality at Amersfoort. Collectiveli,.the resultsfrom thesefour sitesform the basisfor correlation to more distant localitres.
Eemian Sitesin l\'V' Germany, Denmark and Poland
Correlation of sitesin NW Germany and Denmark with the type Eemian locality and adjacentcorrelatives in the Netherlandsis given bv mean D/L ratios in four taxa (Fie. 6). 230 Gifford H. Miller and Jan Mangerud
0.30' n t;] E E fl E E E ll
T T 0.20 T T- '-'?-----ilT 'F 'ito +--+ (! I 1 ll IY -l- I l T n 1n+
0.00
FIG. 5. Mean and standard deviations of D/L ratios measuredin representativesof 5 taxa of the moderate-rategroup found in 4 borings in The Nerherlands.Boring abbreviationsare Amersfoort (A). Bereen(Bl, Cistricum (C) and Zunderdorp (Z).The number of individual analysesfrom which the mean is derived are shown in square boxes at the top of the figure. The mean D/L ratio in all moderate-rareraxa from the four borings is 0.i82 (dashedline); the standard deviation is +0.017 (dotted line).
The mean ratios for Mya, Venerupis and Macoma from Denmark and NW Germany are eenerally slightlv lower than in the same taxa from the Netherlands, in keeping with the thermal gradient. The mean ratios in specifictaxa at four sitesin Denmark fall outside the one standard deviation confidence interval. These sites (Strandegaard,Arctica. Corbula', Mommark. Arctica'.and Ristinge, Myo) are discussedin Part 2.
The marine embavment of the Eem Sea in the Vistula Valley of Poland (Fig. 3) supported an impoverished and diminutive fauna; from the available borings the onlv material of the moderate-rate group obtained common to other sites were a fe'*' small individuals of. Venerupis senescens, Corbula gibba and Spisula subtruncata: Divaricella divaricata (fast epimerizer) also occurs at both the Amersfoort (Netherlands) and Nowiny (Poland) boringsas well as in one of the Danish collections.The mean D/L ratios in three of the four taxa decreasein the direction of the current thermal gradient, with the highest ratios in the Netherlands,lower ratios in the sitesfrom NW Germany and Denmark, and the lowest ratios from Poland (Fig. 7). The only taxon that counters this trend is Venerupis, for which the large standard deviation on the Polish collection allows the possibilitv that a more precise determination would support the trend of the other three taxa. The parallelism of thermal gradient and trend in measuredD/L ratios confirms the correlation of the Tychnowy Sea depositsin the Vistula Valley of Poland with the type Eemian in the Netherlands. Aminostratigraphyof European Marine InterglacialDeposits .!-J I
Additional support for the correlation of Eemian sitesacross the region comes from the amino acid ratios in the slow-epimerizingspecies Littorina littorea (Table 5). This taxon rs present at two of the Dutch sites (Castricum and Zunderdorp). both of which have representativesof the moderate-ratetaxa that substantiatethe correlation to other Eemian sitesin the region. L. littorea from Rodemis in NW Germany and from Stensigmoseand
0.30
T +l .9 0.20 { .1. .. .q 1 T (! 1 I Tl ------T------* - ?rT---f------J I---r- i r---6 r dT_-L 0 T; r9l '{ rl -t -Ii9 o 'l
0.10
'%Y*r',\+;:*t%"?;:*;:;;:4:$;
FIG. 6. Mean D/L ratios in four moderate-rate taxa from Eemian sites in NW Germanl and Denmark (open circles); for comparison.mean ratios from The Netherlandsare indicated br filled circles. Sites are located on Fig.4. The mean DiL ratio for the four taxa from core area is 0.171 (dashed line) with a standard deviation of +0.023 (dotted line).
TABLE 5. Comparison of D/L ratios measuredin Littorina from sitesin n-oru'a1. Denmark. West Germanv and The Netherlands
Country Locality Age Mean
7.5 Norway Fjosanger(Beds L-I) Fjosangerian 0.1.1 0.01 lo 7.5 Norwav Bo (BedD) Avaldsnes 0.097 0 017 E.0 Denmark Stensigmose(Tapes sand ) Eemian 0.086 0.008 3 6.0 Denmark Tonder(Boring 4, 11.3m) Eemian 0.11 0.01 -5 8.5 W. Germanv Rodemis Eemian 0.098 0 004 3 8.5 W. Germany Wacken Holsteinian 0.22 0.03 6 10.-s Netherlands Castricum(19C/6,18) Eemian 0.11 0.01 5 10.-5 Netherlands Zunderdorp(25E1341) Eemian 0.12 0.01 5
'Current mean annual temDeraturein the vicinitv of the collection site. 232 Gifford H. Miller and Jan Mangerud
0,40
'= (o cc J 0.20 o
0.00 L /, 7. 2a.o + 2o -a 'bo e( o
from.last-interglaciallocalit-i-9s FIG.7. Mean D/L ratios in three moderate- and one high-ratetaxa (N) and adjacentcorrelatives)' NW summarizedb-"- region for sitesin the Netherlands (Amersfoort Valley borings (Poland) (P)' The general c.r,nunl 1\\'f ani Denmark (o) and from the vistula thermal gradient and implies a trend of decreasrngratios acrossihis region conforms to the current over the area over last 100ka broadl-vsrmrlar gradient must have dom-inatedthe epimerization-rate
These data support Tonder in Denmark gave slightly lower ratios than in the Netherlands. associatedeposits are of a the conclusionsderived from the moderate-rategroup that the present thermal gradient' single age. with a trend toward decreasingD/L ratios along the differencein within the resolvingpower of the amino acid method, there is no significant Eemian correlatives age between the type"demian site in the NetherlandsanrJ the classical ide'tified in NW Germany, aialysed from othei localitiesin the Netherlands,or the sitesso east of the Netherlands Denmark and Poland. The decreasein D/L ratios to the north and presum.ablyreflects an integrated correspondswell with the present thermal gradient, and of the present' paleo-ihermalgradient that at leastin generalform, mirrored that
Pre-Eemian Sites depositsin NW Germany Molluscs have been obtained from classicalHolsteinian marine Heide)' The mean (wacken, Scharhornborings 55 and 57. Hummelsbi-ittel'Billbrook and separatethe interglacials D/L ratios in taxa that are io*rnon with the Eemian sitesclearly slow-rate taxa' the two (Table 6). Using the group means for both moderate- and interglacialsare separatedat the 95% confidencelevel' Aminostratigraphy of European Marine InterglacialDeposits 233
Schleswig- TABLE 6. D/L ratios in Eemian and Holsteinian molluscs from the Holstein reqion of West GermanY
Taxon Eemran sltes Holsteinian sites
0.16 + 0.01 (12) 0.30+ 0.02 (9) Arctica islandica + balthica 0.28 0.01(13)r Macoma + (9) Mya truncata 0.16 + 0.01 (8) 0.25 0.02 gibba 0.18 + 0.02 (9) 0 26+ 0.03(5) Corbula + (7) Spisula subtruncata 0.11 + 0.02 (2) 0.30 0.02 + Mean moderate-rate taxa 0.17 + 0.01 (29). 0.28 0.02(43) + (6) Liltorina liltorea 0.098+ 0.004 (3) 0.22 0.03
'spisula excluded from the averaging; ratios considered aberrantly low relative toratiosinthesametaxafromotherEemiansitesinthecorearea. I Macoma from Hummelsbiittel excluded from averaging'
Although the mean Holsteinian ratio is easilydifferentiated from the mean Eemian ratio. placed in an imporiant additional considerationis with what reliability a specificsite can be its appropriate interglacialperiod. There are no potential miscorrelationswithin the suites can be of sampies analysed from West Germany. Similarly. most of the Danish sites Table 6, unambiguouslyissigned to their appropriate interglacialusing the mean ratios in group means' althougf, a few siteJyield specifictaxa with ratios intermediate between the (sensustricto) The lowest Holsteinian ratios are in Mya and Corbula from the Holsteinian the beds at Wacken (Mya 0.25; Corbula 0.26). Within the Eemian sites in Denmark in highestratios for Mya are0.22ina singleindividualfrom Ristinge.and for Corbula.0.23 t*o indiuiduals from Strandegaard.Both examplesoverlap their Holsteinian counterparts in at the 1o level. In contrast, eight preparations of.Mya from the Siidermarschboring Germany average0.16 + 0.01, and 9 preparationsof.Corbula from the Schnittloheboring averageb.tq * 0.02, well below the ratios in the sametaxa from Wacken. Shells have also been analysed from a number of sites in Denmark that had been previously correlated with the Holsteinian sitesof NW Germany. Three Mva tuncata from age kaas Hoved gave a mean D/L ratio of 0.31 + 0.05. suggestinga possibleHolsteininan same but with considerablescatter in the data. In contrast, threeArctica islandicafrom the valuesin collection yielded a mean ratio of 0.37 + 0.01, well above the typical Holsteinian the site is Germany. Without additional analyses,the conclusion can only be made that certainly at least of Holsteinian age, but may be from an older interglacial' in The other Danish sitesexpected to be Holsteinian (Rogle Klint and the numerous sites West the Esbjerg area) gave ratios (Table 11) that are substantiallyhigher than in the 0.04 (19)' German Holsteinian sires. From the Esbjerg region Mya averaged 0.45 t + been Macoma calcarea0.46 + 0.04 (5), and Hiatella 0.47 0.03 (8). Only Hiatella has + and analvsedfrom Rogle, but the ratios [0.54 0.08 (3)], although limited in number We that rather scattered,are nonethelesswell above a possibleHolsteinian level. conclude all of these sites are from a pre-Holsteinian interglacial' results The age of the pre-Holsteinian interglacialin Denmark is best constrainedby the four valves of from Noord Bergum. a pre-Elsterian site in the Netherlands. D/L ratios in _-)+ Gifford H. Miller and Jan Mangerud
Macoma balthica from this site averaged 0.46 + 0.06, similar to the Esbjerg ratios and suggestingthat the sitesmay be from the same interglacial.Based on this correlation.the ratios obtained from the Esbjerg region and from Rogle Klint are interpreted to indicate that the associatedinterglacial beds are of pre-Elsterian rather than Holsteinian age as had been previously proposed. The till that underlies the marine bed at both sites is, bv implication, from a pre-Elsterian glaciation. Tills of this age have not been previously identified from NW Europe except at Harreskov, Denmark where glacial-fluvialsediments and till underlie limnic sediments of Harreskovian ('Cromerian') age (Sjorring, 1983). Glacial-fluvial sediment of pre-Cromerian age are reported from the Netherlands (de Jong and Maarleveld. 1983),but the oldesttill there is Elsterian. The Slettenshage(Denmark) site had been previously correlated with the Eemian based on the broadly similar mollusc-faunas,although the classical indicator species Bittium reticulatum and Venerupissenescens were absent. Arctica from the site gave ratios of 0.32 + 0.01 (3), well above the Eemian ratio for this species.The amino acid data suggesta correlation to the Holsteinian of NW Germany where the same speciesaverages 0.30 + 0.0i (9). We have analysed two taxa from the recently discovered marine interglacial beds at Margareteberg, SW Sweden(Fig. a). Ten preparationsof Hiatella gavea mean ratio of 0.25 and 6 preparations of.Macoma averaged0.26 (Table 10), well above the mean values in moderate-ratetaxa from Eemian sitesin nearby Denmark (ca. 0.16 to 0.17), but not as high as in classicalHolsteinian sitesin adjacent NW Germany (ca. 0.29). The pollen stratigraphv is similar to that at Fjosanger (Norway) and dissimilarto the Holsteinian. Primarily on the basis of the pollen stratigraphy an Eemian age has been proposed for the site (T. Pisse. written comm., 1985). The amino acid data suggestthat the marine beds pre-date the Eemian, but are certainly no older than Holsteinian. We favor a correlation of Margareteberg with the FjOsangerianof western Norway, for which we offer two age interpretations (seefollowing section);the two sitesmay be Eemian or they may be from an interglacial between the Eemian and Holsteinian (sensustricto).
Weichselian Sites
Shells of the moderate-rate group analysed from deposits known to be of Middle Weichselian age in Denmark (Hirtshals. Norre Lyngby, Holmstrup and the upper 100m of the Skaerumhedeborings) consistentlyyield D/L ratios between 0.08 and 0.10. inter- mediate betweenthose in the sametaxa from Eemian and Late Weichseliandeposits.
Lateglacial Sites Shells from radiocarbon-datedLateglacial and Holocene sites have been analvsedfrom localitiesin Svalbard,Norway, Denmark, Scotlandand the Netherlands(c/. Miller. 1985). DiL ratios are consistentwith present site temperatures.
INTERGLACIAL SITESIN BELGIUM AND NORTHERN FRANCE
To the south of the Netherlandsat leastsome of the depositsin Belgium consideredto be Eemian correlate on D/L ratios with the Eemian sites in the Netherlands. Venerupisfrom Aminostratigraphy of European Marine InterglacialDeposits 235
the OostendeFormation in the FlemishValley of Belgium has a mean D/L ratio of 0.17, andfromtheNetherlands0.2l;meanratiosinMacomafromBelgiumare0.19.thesameas in the Dutch sites.
We have analysedMacoma balthica and Cardium lamarcki from the brickyard at the t)'pe focality of the Herzeele Formation in northern France. M. balthica yielded mean rarios of 0.31 consistentwith a correlation to the Holsteinian marine depositsof NW Germany, and C. lamarcki, although not one of the primary taxa. did vield ratios appropriate for a Holsteinianage (0.28).
Borings in the Belgian coastal plain close to the French border and outcrops near Zeebruge have yielded moderate-ratetaxa with D/L ratios that are lower than expectedfor Eemian deposits (ca.0.11 to 0.13 in Mya and Hiatella). Becausethey are associatedwith high sea-levelevents well beyond the zone of isostaticdepression from the Fennoscandi- navian Ice Sheet, the deposits are tentatively associatedwith an Early Weichselian (late isotope stage 5) transgression,although this interpretation must await a larger data set before its validity can be tested.
INTERGLACIAL SITES IN WESTERN NORWAY
Two in situ marine interglacialsites are known from western Norway; the Fjosanger site near Bergen and the Bo site on the islandof Karmoy. 100 km to the south (Fig. 3). Details of the stratigraphy within the interglacial beds at Bo remain incompletelv understood (Andersen et al.,\981). DiL ratios in presumablv interglacial molluscs (Table 10) are slightly lower than ratios in the same taxa from Eemian sites in NW Germanv and Denmark. supporting an Eemian age for the interglacialat Bo as proposed by Andersen er c/. (1981).
The Fjosangersite is one of the most thoroughly studied interglacialsites in NW Europe. containing a complete depositional sequencethrough an interglacial cvcle and into the succeedingglaciation. Mangerud et al. (1979,1981a)have suggestedfrom the floral and faunal data that the interglacialbeds correlate with the Eemian of the continent. D/L ratios in molluscsand foraminifera from the Fjosangersite have been assembledfrom analvsesat both the Bergen and Colorado laboratories (Mangerud et al.. 1981a; Miller et at.. 1983: H.P. Sejrup, unpublisheddata). Note. however, that the ratios presentedin previous publications were prepared under a different system than used in this project, and are consistently higher than reported herein. Re-analysis of the earlier-prepared samples confirms the same apparent relationshipsas proposed previouslv. but the absolute values are ca. 70'/. of those reported earlier.
The D/L ratio in 40 preparations of 4 speciesof the moderate-rategroup at Fjosanger (Table 10) is 0.21 + 0.04, whereas11 specimensof two speciesof the samegroup ar Bo average0.14 + 0.01.The two sitesare separatedat 1o, but not at the 2o confidencelevel. As with the bivalve mollusc ratios, D/L ratios in Littorina from the Bo sitesare the sameor slightly fower than in Eemian sites in NW Germany and Denmark, whereas the Littorina ratios from Fjosanger are slightly higher, although still well belou' ratios in Liuorina from the Holsteinian beds at Wacken. Countering the argument that Fjosangerpredates Bo are 236 Gifford H. Miller and Jan Mangerud
f -i: €-: - .in\1 vfl: = - N
N--=s v,- r E .:ai -::
\l!,r tl rr rl {;
N :;!:J .: \ !s: i fl ,< s: fl E ,<-c' ! .l cl € '- a (r) '..; I 3 rth NN -Pc i;-a .ui! +l +t +l +l . h: nr€ ! -ro- 9N : J \= <: +r +t +l +l tl +l +l d cl :: ^': : r gJ = rr I s €-, t€ N F U .: : : ;: +l +l +l +l +t +i rq N 6 C.€ -N : :; :: -aBv a^nelau .E er { - F 4 €*, YE:f so >9: zn & .-i'at.Yz/ a _ a s; i Qi * 7iEee= fis" 2iYEZE?riE_i= f :E*E-t*$E a"z !-d{ ,i tsiglg r-; E=t;stEEE=E-#figg Interglacial Deposits 237 Aminostratigraphy of European Marine - 6-a I - ocrr &= -n.Y- = € -*c{ occ = =88 3 3 - E =-- i c=c c +l +l +l +l +l +l +l +l +l 9 a 5 o6€ 6 €59 d clclc'i J --{ - (, ccc i 5io - € -^ N a t = j= = cc +l+ rl +l - r RFl : N - ; : d -i 5 u.l =--; ? ?aa 5 s 8=8 o a c-c ; +l +l +l +l +l g €-r = ^i ':.1 - - - o-- .9 G 3 qJ cl C +1 +1 lLJ N ; ; € 7 ,,,,.r'.r++F L!sl HHH*F- IllIl]t!Ul t!o U € zzE = Vy :'22 dE '-ii 3;a ts. il = i5 Fi,:. ,C==ZE2=E! Z=z. 2? 6V=*A-EYEf s#iE 34+ THE EEMIAN IN THE ARCTIC A number of sampleshave been analysedfrom marine depositsalong the Arctic coast of the Soviet Union and some of the Arctic Islands.Due to the large differencein temperature between these sites and the secure Eemian sites in Europe, and lack of an independent absolute date on any of the Arctic samples,the interpretation of the measuredamino acid ratios must necessarilybe tentative. As a first approximation, we have assumedthat the agesprovided by the collectors are correct, and that the first deposit below the Holocene with a fauna beyond its present range is the most likely candidate for the last interglacial. Arctica islandica,which has a current northern limit in the White Sea.occurs in deposits along the Arctic coastof the U.S.S.R. from Kola Peninsulato Taimyr Peninsula.Following Kellogg et al. (1982), the most likely time that North Atlantic Drift water entered the Norwegian Sea and extended farther along the arctic coastof the Soviet Union than during the Holocene would have been during isotope sub-stage 5e, the last interglaciation. Certainly the depositscould not be younger. The measured D/L ratio in Arctica (Fig. 10) decreasesbetween Kola Peninsulaand the Yenisey Bay region. Mya truncatafrom Yenisey Bay have a mean D/L ratio about the same as in Arctica, but the spread in measured values is somewhat higher. Probable last -15"C) interglacial depositson Taimyr Peninsula(MAT -13 to contain Arctica, but onlv Mva tntncata andHiatella arcticawere availablefor analysis.The mean ratios (Fig. 10, J.K. Brigham-Grette, unpublisheddata), are lower than from Yenisey Bay, as are the ratios in M. truncata from a purported last interglacial deposit on Novaya Zemlva. The overall trend of these data are as would be expected for last interglacialdeposits under low thermal conditions, although it must be remembered that at such low temperatures the epimerization rate is so reduced that the potential resolvingpower of the Aminostratigraphyof European Marine InterglacialDeposits 239 :;! ! = :: -: tr =;rJa - ! = -e ; { J= - tr.*= = 7 s!:E"i -' .^t,i - i;sSS! F {:-! t/=_2Y .: = +j o _ !_ L O g ; z= ^.t q = -/ \c==--_: - = = c a2 t d a 3 aJir. ?'r!,2e !i;::: y a a -=- q.=r--- ;c it ! i .=?. t. ?::E:!' '- t ! \' u :-= c ? - . a= .= o ; : - '- 2 -= = - : = EJ -= i .: !: q :: =F== t3\ ; i:o-a-::; \zi t i:t_7Ei ). Sr .\ >\ z ii==-iz \ti -ci .=t==ia: -=_=1:= , a;::-:::*: J= E = / = ' q =j-=.aa= * - = I -v. = != = - = -=tai=-?:_ -r 2 ?- ;tazz:! 'qb.. J =.:- i r<11 ]*i- xr!g-: l# .+, -: Li-._ z=l=4 -!s;a :Et:s 210 Gifford H. Miller and Jan Manserud -LE =lrr X:"+! : ;.3ni r ! -* :;.c{ad L - -L :^^i- - t == > :r:!= = e a-= ui;-=i >5zl i.sl z 3 cY -n ! =: L I 6--- o ". - o 9 L v -i= =) t ; ;:r'E !-;qF ; xLo= go6.r o3Fr a-,; Ei 'e'lE5 g --: d6E ! :- - .: OC i.! E o $=\! ?- 3= OF:- -!Ey 9- {936 g-: i a fv u.=: S .:EH; i F*d.= a i.: -. > i i ]:- E P tr ?= ! ==II E:_ F - 6 -=-: I 6.gJ = z- dd;L F I-*i3 G =o"€ :l EE!; !qoF l-#r o it i +&L. F--a>l \\\\l ao. N N*\ l=-:.H *r,4 \'.. ..: ro --.{:-++i.rr-r>.], o x: s - \ a2 a a o+r,z : 'od '=:.== N ,=-.o{ to^ >itrL "J- u=>! 7s "'4, NSN c:= = "+ I\\N.J tr-tr ? a-iIrffll ;!e.* gL ! = =!:d NNsN- E;:c 6r;! N :F;! )59- >o';t !a:; Co€! zag=3 !=.d o i:*-a ,^ :a .! tr VFJ- - 7= Aminostratigraphy of European Marine Interglacial Deposits 241 -20 r?0 SITE TEMPEFATURE ( C) FIG. 10. Mean D/L ratios in moderate-ratetaxa all last-interglacialsites included in this studv. The hatured envelopedefines an exponentialrelationship between D/L ratio and site temperature.This relationshipcan be used as a predictive model to provide an estimateof the D/L ratio for siteswith temperaturesdifferent from ones already studied. methodis lessthan at mid latitudes.The measuredD/L ratiosgenerally decrease in an easterlydirection, with the lowestratios found on TaimyrPeninsula and the Arctic Islands. On Svalbard(Fig.3), the last interglacialis currentlyconsidered to be representedby depositsof aminozoneC (Miller, 1982)based on unpublishedU-series and TL datesand the associatedmicrofauna (G.H. Miller, H.P. Sejrup,S.J. Lehmanand S.L. Forman, unpublisheddata). DIL ratios in Mya andHiatella from depositsof aminozoneC average 0.044+ 0.005. Interglacialdeposits thought to be correlativewith the Eemian have recentlybeen reportedfrom Greenland(Petersen, 1982; Funder, 1984). Mya truncatafrom threelevels within the interglacialbeds produced mean D/L ratiosof 0.047+ 0.002,in keepingwith a last interglacialage. PREDICTIVE MODEL FOR D/L RATIOS IN LAST INTERGLACIAL SITES The wide thermalrange of lastinterglacial localities analysed (-15 to +11'C) allowsthe constructionof an empirical model describingthe relationshipof D/L ratio to site temperature.The mean D/L ratios in all moderate-ratetaxa from all last interglacial localitiesanalysed are plotted against current site temperatures on Fig. 10.The curvefitted -lA") Gifford H. Miller and Jan Mangerud generalform defined to thesedata describesan exponentialrelationship in keepingwith the model that allows the by epimerization kinetics. ttris relationshipalso provides a predictive taxa eitimation of the D/L ratio for last interglacial representativesof the moderate-rate a model will be from anv site within the temperature range of the current data set' Such than NW particulirly helpful for future studies in sites that are slightly colder or warmer Europe. ABSOLUTE AGE ESTIMATESFOR HIGH SEA-LEVEL EVENTS AROUND THE NORTH SEA Althoughthe conversionof D/L ratiosto absoluteage is hampered by uncertaintiesin the thermalhistories of the sites,the ratioscan be usedto placeconstraints on the possible ages.This appraisalis dependenton the correctinterpretation that the EemianInterglacial o-ccurredabbut 120ka ago.Based on that assumptionand the relativedifferences in D/L ratiosfor the other high sea-levelevents, limiting absoluteages can be assigned. We assumethat all Europeaninterglacials are correlatedwith odd-numberedstages tn the deep-seaisotope record, but that not all suchstages are necessarilyrepresented in the terrestrialrecord. Furthermore, with our presentunderstanding of the isotopicsignal. it is not possibleto saywhich of the isotopicpeaks represent terrestrial interglacials and rvhich are interstadials.In the introductionwe arguedthat if the land-basedinterglacial sequence is complete.then the Eemianmust representisotope substage 5e, the Holsteinianstage 7 and the youngestof the pre-Elsterian('Cromerian') warm stages,isotope stage 9' Alternatively,itt" Holtteinianinterglacial might representisotope stage 9, 11 or 13. prior to using the measuredD/L ratios to constrainthe absoluteages of European interglacials.it is essentialto note whether the samplesare from early or late in the interglacial.For example,most Holsteinian molluscs that we haveanalysed come from the late Elsterian-earlyHolsteinian transition or from early in the Holsteinianand have consequenrly.*p".i"n..d nearlythe full duration(ca. 15ka) of interglacialwarmth (rapid epimeiization)ai well asrhat of the subsequentWacken Interglacial. Our Eemiansamples. on the other hand.cover a wider rangeof the interglacial.We estimatethat they average from the middleEemian and so experiencedonly 5 ka of interglacialwarmth. Because the Holsteinianwas longerthan, and aboutas warm as the Eemian(e.g. Meyer, 197'l:H' Mtiller, 1974b),we expectthe Holsteinianshells to haveexperienced a higherproportion of interglacialwarmrh ihan did the Eemian samples.Consequently, the DiL ratios tn Holsteinianshells may be disproportionatelyhigher than expected for their agerelative to the Eemian rarios.*ith th"r. guidelinesand the assumptionthat the midpoint of the Eemiandates from about 120ka ago,the D/L ratioscan be usedto evaluatewhether the Holsteinianis isotopestage 7, 9, 11or 13' The mostconcentrated data set from a similartemperature is from the sitesin Denmark' been plotted NW Germany and the Netherlands.The mean D/L ratios (Table 8) have et againsttheir known or possibleages (based on the recentSPECMAP timescale; lmbrie ot., tSS+;in Fig. 11.The threescinarios depicted are for the Holsteinianto be correlated with stages7,-9 or 11. The slopeof the straight-linesegments connecting each point is duringthat time interval.Assuming that directlyfroportional to the integratedtemperature Depostts 243 Aminostratigraphy of European Marine Interglacial sea-levelevents in NW Europe (Denmark, NW. Germanv. TABLE g. Mean DiL ratios in shells from high oxygen-isotopestratigraphy and limiting absolute age Netherlands)with postulatedcorrelations to the-deep-sea estimates.These data are used to construct Fig 11 Mean D/L ratio Suggestedcorrelation Limiting absolute age Relative age Moderate-rate Slow-rate + Modern 0 Modern 0.013+ 0.002 0.013 0.002 1/2boundarY 10to 12ka Late 0.060+ 0.010 about0.0.{ Weichselian 3 > ,{0 ka Middle 0.090+ 0.010 Weichselian 5alc 70 to 100ka Early 0.12+ 0.01 Weichselian 5e 125ka Eemian 0.18 + 0.02 0.11a 0.01 + 0.02 0.22 + 0.03 7 (or 9) 2.10(or 330ka) Holsteinian 0.29 700ka 0.46 t 0.04 >11 J00 to Pre-Elsterian <19 0.20 ; (! _-a_ oPTIoN 1 tr -{- oPTION 2 J __{- oPTION 3 0.30 so AGE in NW Europe See text for FIG. 11. Possibleabsolute age relationshipsfor interglacialdeposits discussion of the most probable option' 244 Gifford H. Miller and Jan Mangerud the temperatures characterizingeach of the glacial-interglacialcycles should be about the same requires that the slopesof the various line segmentsbe similar. However, the situation is complicated by the general decreasein apparent epimerization rate over time as the proportion of isoleucine in the slower-epimerizing free form increases.Consequently a decrease in slope during successivelyolder glacial cycles may be expected, but not an increasein slope. If we assumethat the Holsteinian is early stage7 (option 1, Fig. 11) then the projection to 'late the Cromerian' ratios in shellsfrom the Netherlandspredicts a minimum ageof ca. 500 ka (stage 13) in keeping with their stratigraphic position above the Matuyama-Brunhes (M-B) boundary. Assuming that the Holsteinian shellsare from early in stage9 (330 ka; oprion 2), then the extrapolationto the'late Cromerian'bedsyields an ageof 670ka. This date is already close to the M-B boundary and leaves little time for additional climatic 'late oscillations in the normally magnetized Cromerian' sequence.Although we favor a stage7 age for the Holsteinian, we cannot exclude a stage9 correlation. Option 3, in which the Holsteinian is correlatedwith stage 11 can be excludedbecause the extrapolatedage for 'late the Cromerian' shellslies within the Matuyama reversedpolarity zone. These considerationssuggest that the Holsteinian (sensulato) is most likely correlative with isotope stage7. The stage7 isotope peak is complex, and has been subdividedinto two 'light'substages (7a and 7c) separatedby substage7b (Ninkovitch and Shackleton. I9'75). Substage7a, in turn, consistsof two distinctpeaks (e.g. Imbrie et a|.,1984, p. 68).These complexities are mirrored in the deep-sea faunal record; in the North Atlantic two interglacial assemblagesare separated by a brief return to a glacial fauna within isotope stage 7 (Ruddiman and Mclntyre, 1982). We hypothesize that the Holsteinian (sensu stricto) is the first light isotope peak in stage 7 (7c:245-230 ka); most of our molluscan samplescome from early within this interval. Substage7b is representedby the niveo-eolian sands separating the Holsteinian (sensu stricto) from the overlying peat at Wacken. NW Germany. The latter peat, the Wacken Interglacial. can be correlated with substage7a, or with the first peak of it. leaving room for an additional warm interval within the Saalian.If 'late these correlations are correct, then the Cromerian' beds can be no younger than stage 13 and no older than stage17. Consequentlya shallow-marineequivalent of at leaststage 9 and 11 is missingin NW Europe. Alternatively, if the Holsteinian (sensustricto) is stage9, the Wacken/Domnitzianmay be correlativewith stage7 and the'late Cromerian'could be stage15, 17, or 19. In this case, shallow-marine equivalentsof at least isotope stages7, 11 and 13 are missing (assuming Fjosangerian is not stage7). The D/L ratios can be used similarly to support strongly a last-interglacialage for the Eemian. If the Eemian is older than stage5e (e.g. stage7), then the minimum age for the Holsteinianis stage11 and the'late Cromerian'would have to be considerablyolder than 700 ka. again in contradiction to the establishedpaleomagnetic stratigraphv. Several recent campaigns have been directed at providing independent dates on terrestrial interglacialsequences across NW Europe. A number of U-seriesdates have been reported recently (e.g. Hennig, 1983;Brunnacker et a|..1983) primarily on travertinesand speleothems.The agesof thesedeposits show a strong grouping that correlateswith isotope stage5. In addition, a number of pre-Eemian sitesyield dates that cluster between300 and Aminostratigraphy of European Marine Interglacial Deposits 215 350 ka. However. correlation of the localitiesdated with the classicalinterglacial sequence is problematic due to the scarcity of pollen in the carbonate deposits. We conclude that these correlations cannot yet be securelyestablished independently of such dates. There have alsobeen attempts to date molluscsdirectly in well studied interglacialsites in the region. Stremme (1983) reports U-series dates on shell from the Herzeele Formation (Holstiinian) in northern France that cluster around 330 ka. More recently Linke et zzl- (USS; reporr 17 ESR dates on marine molluscsfrom the same Holsteinian sites in NW + d.r1nuny that were included in our amino acid survey.The ESR datesaverage220 20 ka, supporting our interpretation that the Holsteinian correlateswith isotope stage7. CONCLUSIONS We have identified a suite of several pelecypod genera from interglacial sites in NW Europe in which the isoleucineepimerization reaction proceeds at a similar, moderate rate. D/L ratios can be used more or less interchangeablybetween these taxa in developing a chronostratigraphyfor the region. It is necessaryto use several taxa in such a regional survey because no single taxon is present at all sites due to strong differencesin local environmentsof deposition. In addition to the moderate-rategroup that forms the basisof the interpretations,supplemental information has been obtained through a feu' additional taxa that produce reliable results, but have different epimerization rates. The relationship between the D/L ratios obtained from thesetaxa to those from the moderate-rategroup is not constantover time, complicating conversionof all measurementsto a singlenumber of universalapplicability. Based on our analysisof more than 1,000individual shells for this project we offer the following guidelinesfor future studies of this nature: (1) reliable amino acid data require the analysisof at leastthree, and preferably five individualsof an establishedtaxon at each site: (2) improved confidence can be obtained by similarly analysing additional primary species;(3) sitesfor which the mean D/L ratio differs from an establishedinterelacial group mean by more than one standard deviation require supplemental analysesof either the 'new' same species or another well constrained taxon before recommending that a interglacialbe introduced; (4) occasionallyaberrant individualsand possiblveven sitesmay occur. The following conclusionscan be drawn from the interpretation of the measured D/L ratios: (1) A first order verification of the validity of the D/L data is the correct ordering of shells from modern, Lateglacial,Middle Weichselianand interglacialsites as shown for Denmark and western Norway (Figs 8 and 9). (2) The amino acid survey essentiallyverifies that sitesin NW Europe previouslylabelled ,Eemian' are all of a single age and that they correlate with the interglacial beds in the Amersfoort boring in the Netherlands. We argue that the type locality for the Eemian lnterglacial, the Amersfoort boring, is demonstrablythe last interglacialin NW Europe and can b. correlated with the marine beds of the Tychnowy Sea that underlies the type Weichseliantill in Poland. 246 Gifford H. Miller and Jan N'lanserud (3) All sitesanalysed that containthe typicalEemian pollen stratigraphy(Amersfoort. Castricum,Bergen. Zunderdoorp,Rodemis, Schnittlohe, Ristinge Klint) have i'ielded secureEemian ratios. Sites that contain the classicalEemian index molluscsBittrum reticulatumand/or Venerupissenescens also producedEemian ratios (with the possible exceptionof FjOsanger,Norway). (4) Marine beds in the vicinity of Esbjerg and Rogle, Denmark. that have been correlatedpreviously with the Holsteinianof NW Germany,yielded D/L ratiosin molluscs that are consistentlyhigher than ratiosin the sametaxa from secureHolsteinian deposits. The amino acid data suggesta correlation with a pre-Elsterianinterglacial in the Netherlands.This correlationimplies that the till belowthe Esbjergmarine beds is of pre- Elsterian(sensu stricto) age; the first reportedwidespread occurrence of sucha till in NW Europe. (5) We acceptthe previouscorrelations of the Eemranto isotopesubstage 5e andan age of 120ka. Usingthis asa control,the aminoacid data imply that the Holsteinianprobably correlateswith isotopestage 7, lessprobably with stage9 but cannotbe asold asstage 11. If the Holsteinian(sensu lato) doescorrelate with isotopestage 7 thenthe Holsteinian(sensu stricto)must representsubstage 7c and the Wackenand other intra-Saalianwarm periods representlater peaksof stage7. (6) D/L ratiosin shellsfrom somelocalities on the Belgiancoastal plain are significantly lowerthan valuespredicted for an Eemianage, suggesting the possibilityof a post-Eemian (Early Weichselian)high sea-levelevent. A similar age high sea-levelevent has been identifiedin an amino acid surveyof Mediterraneanshoreline deposits (Hearty et al., m press). (7) One of the few remainingcontroversies is the aminoacid correlationof interglacial sites in westernNorway. The amino acid data remainopen to interpretationwhether Fjosangeris the sameage or is older than the interglacialat Bo, and by inference,the Eemianof NW Europe. (8) Interglacialsites in the U.S.S.R.,Svalbard and Greenlandcannot be securely correlatedwith the Eemiansites in NW Europeby D/L ratiosalone because of temperature differences.But by assumingthat the postulatedcorrelations are correctthis providesa framework for evaluatingthe temperature-sensitivityof the isoleucineepimerization reaction.Last interglaciallocalities that have beensampled span a temperaturerange of more than 20'C (-12 to +10'C). The D/L ratiosmeasured in shellsfrom thesesites demonstratean exponentialincrease in reactionrate with increasingtemperature. a A1 InterglacialDeposits Aminostratigraphy of European Marine and Measured D/L Ratios Part 2- Description of the Sites SitesfromwhichmolluscansampleswereobtainedarebrieflydescribedandtheD/L ratiosobtainedforallspeciesanalysedateachsite.arep.Tll,-l1;Referencesfordetaileddescribed basesfor previousige ascription'The sites' sitedescriptions are glven, as are the in Figs3 and 4' in o.0". fiom north to south,are shown for which resultsare presentedalphabetically The tablesfollow a standardformat, in eachrelativeagegroupfromeachcoully.Theanalysesare,listedbytheirlaboratorymean of all the D/L ratio is giu"n ut the arithmetic identificationnumber ind species; (o) and the standarddeiiation of the measurements analysesfor a givenr-J ro^(Mean), meanwas based(n)' Somecollections of sp".ifJp."purution, on which the the number of individualshells' Such ou, resultsinclude multiple analyses were limited in size unl analvses(tt) in parenthesesfollowing the numberof collectionsare designated by a number individualsutilized' Multiple laboratoryidentifi- that designatesthe nu*u"i of different are indicatedas footnoteswhere appropriate' cationnumbers for the samecollictions SOVIET UNION WehaveconcentratedouranalysesofinterglacialmolluscsfromtheU.S.S.R.onseveral regionsthatrepresentalargerangein.u,,.ntthermalregime'Mostanalyseswereonwater ispresently to areasinfluenced by Arctica islandicaanl'Mya trircoto."Arctica 1i*i,".9 its no,ih"u""tn limit Iiesin the White Sea)and of the North Atlantic Drift (the present coastof the SovietUnion is an indication presencein Pleistocenedeposits along the Arctic (Fig. 3) are discussedfrom westto east;measured tf interelaciut.onoltionr. 1't" localities D/L ratiosare givenin Table 9' Kola Peninsula describetwo seriesof sub-tillmarine sediments Yevzerovand Koschechkin(1977, 1981) Kola Peninsula'The ageof these in the Iowercourses'of river valleys on eastern that occur (Faustova,i984) that the J"uut"d, bur ir.now seemsto be agreed depositshas been tong assumed ponoy beds,were deposited during the-Mikulino lnterglacial. lower of these,the fauna The Ponoy beds contain a boreal mollusc to be correlativewith ihe Eemian. astoday. Several finite radiocarbondates between suggestingmarine ,.rnp"iu,.rr"s as. high in the ponoybeds, whereas U/Th dateson molluscs 34 and47 ka wereoutained from shelis and 1i4 ka (Yevzerovand Koschechkin'1977; from the samebeds range between86 bedsare separated from-th-e-Ponoy et al., Dtl).-it? upp". marineStrel'ninskiye Arslanov and Koschechkin,1977). The ,u,fu"",, un-oin one case,a till (Yevzerov bedsby erosional in the Pono,vbeds. bedsSuggest .ota". marineconditions than faunain the Strel,ninskiye Periipohjola Jirel'niskiyebeds a-recorrelated with the Through polten ,,ruiigruptrv_the interstadialofFinland(e.g.Hirvasetat.'tvztl.Allofthesampleswehaveanalyzedfrom Kola Peninsulacame from the Ponoybeds' ThePonoytypelocality,situatedo.nthesEcoastofKolaPeninsula,containsinterelacial to25 m abovethe sea.Samplesof Aruica islandica marinesediments ti,ui ur..urrently 20 u,eresubmittedbyV.Y"u,",ovinlgS0andlg8lfromthePonoybedsexposedinsecttons and Jan Mangerud 248 Gifford H. Miller for Tables 9 Union. An explanation of the format ratiosin molluscsfrom sitesin the Soviet TABLE 9. D/L to Part 2 of the text throught9 is givenin theintroduction Locality/samPlenumber Mean Lab ID Species Kola Peninsula 0.002 3 PonoyRiver, S of Village" Arctica blandica 0.035 m* AAL-2596 0.004 6(3) MalayaKachkovka 5.8 Arctica islandica 0.034 m* AAL-2786 0.001 3 MalavaKachkovka 4.2 Arctica islandica 0.037 Kachkovka4.2 m AAL-2'785 0.040 0.000 2(t) Malaya AAL-3753 Arctica islandica 2 MalayaKachkovka 4.2 m islandica 0.042 0.004 AAL-3'752 Arctica 0.004 3 R. Chapoma* Arctica islandica 0.012 AAL-2'781 0.050 0.006 2(l) R. Chapoma AAL-3754 Arctica islandica *Submittedby Yevzerov,others submitted by Punning Zaton, Archangeslk District Mezen 0.046 0.005 3(1) R. AAL-3917 Arctica islandica 2(1) R. Mezen islandica 0.0s3 0.008 AAL-3751 Arctica 0.006 2(1) R. Mezen Hiatella arctica 0.054 AAL-392i 0.044 0.000 2(1) R. Mezen AAL-3750 Mya tuncata 2 R. Mezen truncata 0.089 0.006 AAL-3920 Mya 0.005 2(1) R. Divina. Arctica islandica 0.094 AAL-3192 0.082 0.005 2(l) R. Divina" AAL-3192 Mva lruncala 'submitted by Gudina,others submitted by Punning Kanin Peninsula 0.036 I 3018-3 AAL-3190 Arctica islandica YeniseYBaY 0.022 0.000 2(1) A-262 AAL-2586 Arctica islandica 2(1) B-192-4 islandica 0.025 0.000 AAL-2584 Arctica I B-2'71 Arctica islandica 0.026 AAL-2587 0.02'7 0.001 2(1) A-252t1 AAL-2585 Arctica islandica 0.02'1 0.000 2(1) T-989 AAL-2590 Arctica islandica 0.035 0.004 2(1) A-323 AAL-2589 Arctica islandica 1 G-260 lruncata 0.015 AAL-2595 Mya 0.000 2(1) G-2'7011 Mya truncata 0.019 AAL-2588 0.020 0.003 3(1) G-122 AAL-2591 Mya truncata 2(1) D-21 truncata 0.031 0.001 AAL-2593 Mya 0.001 2(r) G-203tr Mva lruncata 0.035 AAL-2592 0.02'1 0.004 6(l) 12|5 AAL-2594 Cardium edule Arctic Islands 1 NovavaZemlaYa 182'3 truncala 0.021 AAL-319i Mva 0.000 2(1) Novaya ZemlaYa: K-3 Mya truncata 0.022 Islands: 1083 AAL-3189 0.028 1 New Siberian AAL-3187 Mya truncata alongthePonoyRiver,fromcorrelativebedsontheChapomaRiverl50kmtotheSW,and fromtheMalayaKatchkovkaRiver,50kmtothenorth.AttheMKatchkovkasite,T.8m 140m asl' In 1983Punning submitted additional thick marinebeds occur below till at ca. MalayaKatchkova River and chapomaRiver' Arcticaislandicarro* tr,. sectionson the Peninsulais of.Arctica islandicain7 collectionsfrom Kola The meanD/L ratio in 21 shells we localitiesrepresent a-single interglacial e\'ent. 0.039t 0.005(Table 9), suggestingthe with the EemianiMikulinointerglacial' and that assumethat this lnr"rgiu.iuii, .orr""lutiue Aminostratigraphy of European Marine Interglacial Deposits 249 the high elevation of some of the deposits is a result of uplift (Yevzerov, written comm. 198s). Zaton, ArchangelskDistrict In 1983,Punning sent well preservedvalves of Mya, Hiatellaand Arctica from interglacial marinebeds exposed along the River Mezen,near V. Zaton closeto the White Seacoast. The shellswere excavatedfrom marinesands 4.4 to 6.0 m belowthe groundsurface and wereoverlain by additionalsandy beds. The faunasuggests conditions as warm aspresent andthe depositis consideredto be Eemianby Punning(written comm., 1983).Recent TL datingof sandsbeneath the earlyWeichselian (Valdai) till in the Archangelskdistrict gave an ageof 93 ka (Hiitt etal., 1985).Arcticaand Mya from an interglacialsite on the River Divina were suppliedby Gudinain 1980. All but one of the valves submittedby Punning yielded similar D/L ratios. The anomalousvalve (AAL-3920;.M. truncata)gave a ratio twice as high as the other samples, but similarto the ratiosin both samplesfrom the River Divina.These results suggest that bedsfrom two discreteinterglacials occur in the Archangelskdistrict. The meanD/L ratio of the otherfour samplesis 0.051+ 0.006(Table 9), slightlyhigher than on Kola Peninsula. The currentMAT is aboutthe sameas on Kola Peninsula(0 to - 1'C). The Zatonratios are slightlyhigher than expected,but we suggestthe shellsare probablyEemian. Kanin Peninsula We have analyzeda single valve of Arctica islandicafrom Kanin Peninsula.at the junctureof the Barentsand White seas.The shellwas supplied by V. Gudina.Novosibirsk. The D/L ratio is 0.036 (Table 9), compatiblewith the correlationto the Kola and Archangelsksamples. YeniseyBay In 1980,Gudina submitteda suiteof samplesfrom depositsaround the mouth of the YeniseyRiver (MAT -10'C). Most of the sampleswere collectedwithin a 100km N-S strip in the inner reachesof YeniseyBay, betweenGolchikka and LadygrigYar. except sampleB-L92, which was collected i50 km westof the river mouth,sample G-122 collected midwaybetween 8-192 and the river mouthand sample A-323 from the vicinityof lake Oz Khaseynto,60 to 70 km westof the innerpart of the bay.It is our understandingthat all of the samplesare consideredto be of last (Kazantsevo)interglacial age. Certainlythe occurrenceof Arcticaislandica in thisarea suggests much warmer inshore marine conditions than existtoday. The interglacialbeds are locallyoverlain by till of an Earll' Weichselian glaciationand Late Weichselianglacial-lacustrine sediment, separated by N{iddleWeich- seliannon-glacial deposits (Isayeva, 1984). The D/L ratiosin Arctica(Table 9) clearlyindicate that all collectionsare of a similarage. The spreadin ratiosis greaterf.or Mya, possiblysuggesting some age differences. However. all samplesconsist of a singleshell from whichone to three preparations$'ere made for analvsis.At presentwe assumeall samplesare of approximatelythe sameage. The mean D/L ratio from elevenpreparations of .4. islandicais 0.027+ 0.004and in ten preparations M. truncatais 0.024I 0.008. 250 Gifford H. Miller and Jan Mangerud Taimvr Peninsula ln 1919D.M. Hopkins obtainedmolluscs from depositson the Taimyr Peninsula(MAT -15"C). These samples were processedby J. Brigham-Grette and are included here for comparison with the samplesfrom the more western localitiesalong the Arctic coastof the Soviet Union. The mean D/L ratio in 8 shellsof M. truncata and 11 shellsof H. arcticathat we believe are last interglacial in age is 0.023 + 0.008. Arctic Islands We have processeda few shells from Novaya Zemlaya and the New Siberian Islands supplied in 1980 by Gudina. Both areas are high arctic sites with MATs averagingabout -12"C. Only M. truncatahas been analyzed,but theseare from siteswhere the microfauna indicates an interglacial age (Gudina. written comm. 1980). The mean D/L ratio in 4 preparations is 0.023 + 0.004 (Table 9). GREENLAND An in situmarine unit a few meters above sea-levelon the west coast of JamesonLand. East Greenland contains a thermophilous fauna similar to that of the mid-Holocene marine optimum but includestaxa that did not live in the area during the Holocene as well as some that have sincemigrated out of the region (Petersen,1982; Funder, 1984).The sandvshell- bearing beds lie on sediment of glacial origin. Based on the faunal composition, infinite radiocarbon dates and relatively low amino acid ratios, the beds. which define the Langelandselv Interglacial, are correlated with the Eemian interglacial and isotope substage5e (e.g. Funder. 1984). D/L ratios have been measuredin Mya truncatafrom three levels in the interglacialbeds. The mean ratios for each level are 0.047,0.049 and 0.045. The mean D/L ratio for the sequenceis 0.047 + 0.02 (5). SVALBARD The Svalbardarchipelago preserves a largevariety of raisedmarine deposits, many of which containa rich and diversefauna. Many of the depositsfound alongthe north and west coastsof Spitsbergenare beyond the range of radiocarbonand may be of last interglacialage or older.Lacking a diagnosticpollen stratigraphy that canbe usedas a basis 'extra- for correlatingwith the standardEuropean sequence, most workers have relied on limital' molluscor foraminifersto differentiatethe lastinterglacial from interstadialevents. Amino acidratios supplement the biostratigraphy. Miller (1982)identified two marinedepositional episodes that might be correlatedwith the Eemian; episodeB or C. Recent micropaleontologicalanalysis of these siteshas demonstratedthat althoughboth unitsare indicativeof relativelywarm nearshoremarine Aminostratigraphv of European Marine InterglacialDeposits 251 waters, only the older (episode C) contains faunal elements suggestiveof temperatures warmer than present (H.P. Sejrup and S.J. Lehman. unpublisheddata). Consequently.we argue that episodeC is the most reasonablecorrelative of the Eemian depositsof Europe. The mean D/L ratio in Mya truncata and Hiatella arctica,s 0.044 + 0.005. NORWAY Fjosanger A complete glacial-interglacial-glacialmarine sequencewas uncoveredin excavationsat Fjosangernear Bergen (Mangerudet a\..1979,1981a).At the baseof the secrionis a till, overlain by silt with a cold-water mollusc and foraminifera fauna. In the overlying sandsand gravelsthe fauna gradesinto an assemblagerequiring sea-surfacetemperatures higher than at present. At the top of the interglacial beds the fauna changes back to a high arctic assemblageenclosed in glacial-marine sediment. The interglacial (the Fjosangeiian) has been correlated with the Eemian because of the strong similarity in the pollen stratigraphies. An Eemian age is supported by the occurrence of the index gastiopod Bittium reticulatum and a seriesof thermoluminescencedates (Hiitt et al.,I9p3f . Fjosanger was the first European mainland interglacialsite to be studied by the amino acid method and a large number of analyseshave already been published (Mangerud et al., 1981a; Miller et al., 1983). The assumption that the Fjosangerian is the lasi (Eemian) interglacial has been the basisfor estimating the agesof younger depositionalevents from their D/L ratios (N{angeruder a/., 1981b; Miller et a\.,1983). The averaee D/L rarios rn Fjosangerian taxa are given in Table 10. BO An excavation at Bo on the island of Karmoy, revealed interglacial nearshore marine sediments in the lower 3 m of the 21 m section. The interglacial beds (Avaldsnes Interglacial:Andersen et a\.,I98I,1983) are overlainby two marine interstadialdeposits separatedby a till, and the sectionis capped by a Late Weichseliantill. The interglaciaibeds have been correlated with the Fjosangerian (see above) and the Eemian of NW Europe basedon the interglacialcharacter of the floral and faunal elementsand amino acid ratios rn foraminifera (Andersen et a\..1983 Miller et a\.,1983). The marine gasrropodBiuittm reticulatum,common in the interglacial beds atBo, is generallvregarded as an index taxon for the Eemian. The upper interstadial marine unit, the Bo marine sands. is of Middle Weichselian age. The interglacial beds are overlain by Early Weichselian marine sands, althoush the contact, which is either erosional or representsa prolonged period of slou, sedimeniation. was not obvious in the field, and molluscsare not availablefrom this bed. A new excavarron at the site has revealed a more complete stratigraphyof the interglacialbeds (B.G. Andersen, oral comm., June 1985) and additional analyses are underway at our laboratories. 252 Gifford H. Miller andJan Maneerud TABLE 10.D/L ratiosin shellsfrom marinesires in WesternNorwav Lab ID Specres Mean D/L Bed Sample no Fjosanger Site AAL-2181 MVa Iruncata 0.180 0.000 2 F 235 AAL-286.{ Mva truncata 0.193 0.021 2(1) UI I l,s0 AAL-3681 Hiatella urctica 0.222 0.008 4(2) vz 847 AAL-456.{. Hiatella arctica 0.238 0.008 3(2) G2 847 AAL-2865 Mya truncata 0.219 0.002 2 1411 AAL-3678 Littorina littorea 0.108 I H 13,s2 AAL-3680 Turritella communis 0.295 0.02s 4(1) H I 118 BAL-I8I Arctica islandica 0.213 0.006 2 366i AAL-3100 Bittium reticulatum 0.063 0.035 1 295 AAL-3398 Bittium reticuiatum 0.130 0.059 4 JJJ trJ AAL-3683 Littorina littorea 0.138 0.006 3(l) 37? AAL-3846 Lucinoma borealis 0.26r 0.025 3(:2) 35'7 AAL-3817 Lucinoma borealb 0.311 I 328 AAL-2866 Mya truncata 0.222 0.008 2(t) 381 AAL-2867 Mya truncata 0.223 0.005 2(1) r293 AAL-3682 Turritella communis 0.i40 0.004 3(1) I lu9 AAL-3399 Turritella communis 0.157 0.004 3(l) 1199i AAL-3401 Turriella communis 0.21r 0.013 2 1462i AAL-2829 Turritella communis 0.219 I 1271 AAL-2828 Arctica islandica 0.177 1 1210 AAL-3631r Arctica islandica 0.197 0.018 2(1) 1210 AAL-.1563r Arctica islandica 0.191 0.003 4(1) 1210 BAL.182 Arctica islandica 0.208 0.002 2 1311 AAL-3675 Bittium reticulatum 0.154 1 1299 AAL-3679 Littorina littorea 0.123 0.007 4(1) 1208 AAL-3848 Lucinoma borealis 0.231 0.021 2(1) 663 AAL-3676 Turritella communis 0.285 0.026 4(1) 66-5 AAL-3228 Venerupis rhomboides 0.135 0.007 2 1350i AAL-1562. Venerupis rhomboides 0.162 0.027 s(2) 1350i AAL-3227 Venerupis rhomboides 0.154 0.007 3(1) 1222i AAL-.1561. Venerupis rhomboides 0.194 0.002 2(1) 1212i AAL-3.102 Littorina Iittorea 0.130 0.006 3 K 146-5 AAL-384.{ Littorina littorea 0.144 0.011 3 K 1290 AAL-38.15 Littorina littorea 0.150 0.007 3 K 12,17 AAL-38.19 Lucinoma borealis 0.265 0.00e 3(1) K 135I AAL-3850 Lucinoma borealis 0.255 I K 8.19 AAL-3681 Littorina littorea 0.125 0.007 3(1) K/L 137ti AAL-3397 Littorina littorea 0.135 0.017 4(2) WL l.{73 AAL-3677 Littorina littorea 0.153 0.008 s(2) L 1.16I AAL-3851 Lucinoma borealis 0.207 I L 6-s9 AAL-3685 Hiatella arctica 0.212 0.024 2(1) M 8,{8 AAL-rs65- Hiatella arctica 0.226 0.004 2(1) M 8-18 AAL-2868 M y-atrUnCAtA 0.264 M l1l5 AAL-2869 Mva truncata 0.286 M 606 AAL-2870 Mya truncata 0.274 M l 302 AAL-3685 Mya truncata 0.282 M 818 AAL-.r565' Mva truncata 0.283 0.010 3(1) M 84E Aminostratigraphy of European Marine InterglacialDeposits 253 TABLE 10(continued) Lab ID Species Mean D/L Bed Sample no. Bs Site AAL-2825 Arctica islandica 0.156 0.009 3D AAL-3403 Arctica blandica 0.140 0.010 4D AAL-3404 Arctica islandica 0.144 0.02'7 3D BAL-I83 Arctica islandica 0.130 0.026 2D AAL-3278 Bittium reticulatum 0.064 0.022 4D AAL-327'7 Corbula gibba 0.139 0.023 4D BAL-291 Littorina littorea 0.097 0.017 3D AAL-3853 Lucinoma borealis 0.232 0.022 3D AAL-2824 Turritella communis 0.150 1D AAL-3276 Turritella communis 0.193 0.057 3D Middle Weichselian Sites AAL-3630 Arctica blandica 0.083 0.005 3 Rogne.Alesun d 19'/7-16 AAL.3#1 Mya truncata 0.090 0.007 3 Rogne,Alesund 19'77 -I22 AAL-3001 Mya truncata 0.090 0.006 Bo. unitB AAL-4528 Mya truncata 0.tt2 0.011 3 Vigra.Alesund AAL-4528 Hiatella arctica 0.126 1 Vigra,Alesund AAL-3755 Mya truncata 0.171 0.021 3 Eidsvik197E-52 * Repreparationof sameshell(s) as in precedingsample. tRepreparationsof sameshell as AAL-2828. M iddle W eichselian Sites Shellytills commonlyoccur in the Alesundarea of westernNorway (Fig. 3). The shells are assumedto havelived duringthe Middle WeichselianAlesund Interstadial (Mangerud et al.,l98lb). For comparison,Arctica islandicaand Mya truncatawere re-preparedfrom the Rognesite. as wereMya from Eidsvik,a sitethat had previouslyyielded higher ratios than other sitesin the Alesund area, and new setsof Mya andHiatella were analyzedfrom Vigra (Table 10). The new analysesfrom Rogneand Eidsvik are ca. 70"h of the values reportedpreviously (see earlier discussion). Mya andHiatella from Vigraare slightly higher than at Rogne.The Eidsvikresults suggest that someof the sitesin the Alesundarea ma1' datefrom Eemianor Early Weichseliantime ratherthan from the Middle Weichselian. SWEDEN Marineinterglacial deposits were recently discovered near Margareteberg along the west coastof southSweden (Pisse et a|.,1984).Floral and faunal studies on thisdeposit are still in progress,but PisseQters. commun.,1985) argues that the pollenstratigraphy suggests the site is probably Eemian. The palynology,however, has similaritieswith the FjOsangerian,the ageof whichremains unresolved. Pisse supplied us with severalvalves of Hiatella arcticafrom the Margaretebergbeds in 1984and Macoma balthicain 1985.Ten individualsof Hiatellayielded a meanD/L ratioof 0.25+ 0.01and six Macoma gave a mean ratio of 0.26+ 0.02(Table 11). The ratiosin bothspecies are well abovetheir counterparts at Eemiansites in Denmark, althoughonly slightlyhigher than extremeratios in other epimerizingEemian moderate-rate taxa. 1< I and Jan Mangerud -JA Cifford H. Miller Margaretebergsite' sw TABLE 11. D/L ratlos rn marine molluscs from the Sweden Lab ID Specres Mean 5 AAL-126,{ Hiatella arctica 0.251 0.012 5 AAL-3911 Hiatella arctica 0.252 0.0i5 6 AAL--+552 Macoma balthica 0.263 0.022 than in Danish Analytically, it is certain that the ratios in Margareteberg are higher particularly with Eemian sites. Whether this showsthe site to be pre-Eemian is unresolved, Norway' Although we the lack of resolution on the age of the FjOsangerianbeds in western 'aberrant' high or low ratios have suggestedthat an occasional shell that yields anomalously 'aberrant' ratios. If the can be encountered it is not yet known if an entire site can yield additional factors FjOsangerianis Eemian. then so might be Margareteberg, and any D/L ratio would besidestime and temperature which might be influencing the measured reouire evaluation. PRE-EEMIAN SITES IN DENMARK Esbjerg Near the city of Esbjergin SW Denmark, shell-bearingmarine sedimentshave been frequentlyobserved in itaypits (Hansen,1965). The typicalsuccession is till at the base' generallyassumed to be oi Elsterianage (Sjorring, 1983), overlain by glacial-fluvialsand the Ina gta.iat-rnarinesilt with a cold-waterfauna (e.g. Portlandiaarctica). Higher up in siltrclaybeds is a more boreal fauna, but warm-waterspecies are absent,presumably becauserelative sea-level fell beforethe interglacialclimatic optimum. The sequencesare the cappedby a till or other glacialsediments. showing that the marinebeds are older than Eemianbecause this arealies outside the Weichselianice limit. The marinesediments have beenassumed to be of Holsteinianage (e.g. Hansen,1965; Feyling-Hanssen and Knudsen' 1980;Sjorring. 1983). ,\Iva truncaLafragments from the Tellina (Macoma) sandsand Hiatella arctica from Maade clay pit. boih at Esbjergwere submittedby K.S. Petersenin 1983;additional the materialfrom the Esbjergarea was supplied in 1984:all shellsare from collectionsof All GeologicalSurvev of 6enmark, manyfrom collectionsobtained late in the lastcentury. thanHolsteinian. They are correlated shellsfielded D/L ratiosindicating an agemuch older u,iththe 'late Cromerian'of the Netherlands' Vognsbol The Vognsbolboring, 2 km north of the town of Esbjerg,contains a Quaternarvrecord a similar to that at the Esbjerg localities.The shellscome from the VognsbOlSand' in fossiliferousmarine unit containinga fauna that gradesfrom boreal-arcticto boreal glacial- character.The marinebeds in the Vognsbolboring, which are overlainby 19m of to fluvialsediments and at least22m of till, are consideredto be a differentand later facies the marineunits at Esbjergand Rogle (Nordmann,I9l3, L922)' Aminostratigraphy of European Marine InterglacialDeposits 255 The D/L ratios in 5 valves of.Mya truncata from Vognsbol are indistinguishablefrom ratios in the same species from the Yoldia clay at the Esbjerg localities, confirming Nordmann's local correlation. Arctica ratios from Vognsbol give lower mean values than Mva.but are in statisticallyindistinguishable (Table 12), whereassix valves of.Littorina are not unexpectedlysubstantially lower. The ratios in both Itlya andArctica are well above the ratios in the same taxa from Holsteinian deposits in the Schleswig-Holsteinarea of NW Germany, with which they had been previously correlated (e.g. Nordmann, 1913). Rqgle Klint At the baseof the Quaternary sectionat ROgle(Sjorring, 1983)on the westerntip of Fyn, is a till and associatedglacial-fluvial sediment, overlain by marine clay of supposed late Elsterian age (Feyling-Hanssen and Knudsen, 1980). Two till units overly the marine sediment: the first, a thick sequence of glacial-fluvial sediment and till is generally consideredto be of Saalian age, although a Weichselian age has also been proposed; a younger overlying till is of Weichselian age. Hiatella arctica from the marine clay gave D/L ratios of 0.54 + 0.08 (3), considerablyabove the expectedHolsteinian ratios for this taxon. Although higher than the ratios in Hiatella and other moderate-ratetaxa from the Esbjerg area, the rather large standard deviationsoverlap and the ratios from the two regions are not statisticallydifferent. We argue that the beds at Esbjerg, VognsbOland Rogle predate the Holsteinian interglacial by a considerabletime interval basedon the large differences between D/L ratios measured in these shellsand those in the same taxa from the classical 'Cromerian' Holsteinian sites in NW Germany. Their correlation with the of the Netherlandsis discussedin Part 1. Sleuenshage Petersen(1973) describeda unit of marine clay within the next-to-youngesttill (till A) in a ciay pit at Slettenshageon the southern side of the RosnaesPeninsula, Sjaelland. Twenty TABLE 12.DIL ratiosin shellsfrom pre-Eemiandeposits in Denmark Lab ID Species Locality AAL-3507 Hiatella arctica 0.438 0.025 1(2) Esbjerg. Maade Claypit AAL-4003 Hiatella arctica 0.502 0.054 4(2) Esbjerg. Yoldia Clay AAL-4026 Macoma calcarea 0.415 0.064 Esbjerg, Yoldia Clav AAL-4023 Macoma calcarea 0.49'l 0.025 Esbjerg, Yoldia clay AAL-4002 Mya truncata 0.504 0.007 Esbjerg. Yoldia Clay AAL-4006 Mva truncata 0.412 0.044 Esbjerg. Maade Clavpit AAL-3506 Mva truncala 0.433 0.036 4 Esbjerg. Tellina Sand AAL-.{009 Arctica islandica 0.367 0.002 3(l) Kaas Hoved AAL-4010 Mya truncata 0.308 0.052 3 Kaas Ho'v'ed AAL-4004 Hiatella arctica 0.535 0.078 3 Rogle Klint AAL-2606 Arctica blandica 0.320 0.010 3 Slettenshage AAL-260r Turritella communis 0.308 0.008 5 Slettenshage AAL-2602 Turritella communis 0.292 0.008 5 Slettenshage AAL-4001 Arctica blandica 0.375 0.041 4 Vognsbol,13.8-1.1.8 m AAL-3981 Littorina liuorea 0.247 0.038 6 Vognsbol.13.8-11.8 m AAL-4000 Mya truncata 0.435 0.023 5 Vognsbol.15.3-15.6 m H' Miller andJan Mangerud 256 Gifford statedthat all of the from the clayand Petersen(1913: p' 42) mollusctaxa were identified tentatively ur"-t no*r, from Eem Seadeposits. He therefore generaand most"f th;;;;;", of the a Weichselianage for the youngertills' All suggestedan EemianugJto, ttrgcrgY and and submittedby him in 1980' shellsanalyzed were cJllectedby Petersen (Table 12) in Arctica islandicaand Turritella communi's The D/L ratios determined and but youngerthan the marinebeds at Esbjerg clearlyindicate un ug"-oiJ",than Eemian, the raliosin Arcticaare the sameas in Arctica Rogle.The shellsmay be of Holsteinianage; from the Holsteinianin NW Germany' KaasHoved ThelithologyandmarinemicropaleontologyofthesiteatKaasHovedinthewestern andKnudsen (198a)' A marineunit part of Limfjord nasrecentty been iescribed bl Jensen iectonized.glacial-fluvialsediments (Unit 4) telow two tills (Units 2 and 3) and glacialiy interglacialharineconditions' Based on the containsboreal toru*inir".u indicativeof ruit the-marinebed is assignedan Holsteinian faunalcomposition ""J-rii",igtaphic position, at Gyldenal(Jensen and Knudsen' 1984) age.Marine unitsoisimlta. aleiocaieo n"urby faunasof shallowarctic marine affinity andHostrup (Knudsen, 1977) lontain foraminiieral but marginallypredating the Kaas Hoved that are consideredto t" of iate Elsteriaoug., beds. Specimensof.ArcticaislandicaandMyaftuncatafromtheLedaClayatKaasHovedwere suppliedbyPeterseninlg84fromcollectionsattheGeologicalSurveyofDenmark.Three+ Arcticaratios g^roup tightly at 0'37 0'01' individualsof each;;;i;*"." analyzed;the ratiosat .u' b.zg and a third at 0'37' We butMyagave ratherconflicting "atresults *iih t*o agebut are more probablyolder' concludethat rhe r";;i;; ;.; leustof Holsteinian EEMIAN SITES IN DENMARK have been previouslysub-divided AssumedEemian marine sequencesin Denmark in theirmollusc faunas (odum' 1933: into two paleontologicalgroups, based on differences rwo faunalgroupings is geographicallydistinct' Hansen,ig65). ThJ dlstiibutionof these by the extinct venerupissenescens Sites in southwesternD"n-u.k are characterized (Nordmann,1908)andSeveralspeciesthattodayonlyliveSouthofDenmark(e.g.a Gastiana L) ' Bittium reticulatum' Divaricellaairoriroto, iyndrr^yo ivata Phil., fragilis Specieswhichis.on.ia"'"atobelimitedtotheEemianandtheHoloceneinNWEurope qSSt is alsofrequent' This thermophilous (Heide,1957; DecheiO-, f Mangerudet al.,I98t), andthe Netherlands,and the Danish is similarto the iemian fauna-sin N* Germany fauna the seaextending from the consideredto havebeen deposited in-an:1t".of sequencesare TrappeskovKlint' sitesincludea in ttti. study, RistingeKlint' EnglishChannel. oittre stricto) and HOjer belong to this Eemian (sensu Stensigmose,f,no*ui[ Faergegaard group. portion],Strandegaards Dyrehave' The easternDanish sites (Skaerumhede [interglacial (e'g' char;cterizJdby a boreal-arcticfauna StubberupHave and Hjelm Nakke).are most of the Lusitanian(southerly) elements' Macomacalcarea,-Nuruio tennus) and lack Aminostratigraphy of European Marine Interglacial Depostts 257 However,the index gastropodBittium reticulatum does occur in someof thesesites. The marinesequences in iasternDenmark are thought to havebeen deposited in a seawith an openconnection to the oceantoward the north, betweenDenmark and Norway.This sea hasbeen called the SkaerumhedeSea (Hansen, 1965), and the associateddeposits have beensuggested to representa youngerevent during the lastinterglacial antedating the Eem Sea(e.g. Hansen, 1965). This interpretationhas been questionedby Petersen(in Bahnsonet al', 1974)who considersthe two faunasto representdifferent faciesof the sameage a view supportedby the geographicseparation and lackof superpositionof the two faunalgroupings' In contrastto the paleontologicalargument, maps of the distributionof Eemianmarine sedimentsshow that the Eem Seadid not extendacross the Jutland-SchleswigPeninsula into the BalticSea (Krogh, 1979; Behre et a\.,1979;Oele and Schuttenhelm, 1979; color chart).On this basisall of the sitesinvestigated in Denmark,the siteat Holnis, Germany and the sitesin Polandare associatedwith a paleo-BalticSea that was connectedto the North Atlantic betweenDenmark and Norway,much asit is today. To compareas many sitesas possible we concentratedthe analyticalprogram on two of the most common raxa in Danish interglacial sites, Ardica islandica and Turritella communis.Most of the specimenswere suppliedby Petersenin 1980and 1982from collectionshoused at the GeologicalSurvey of Denmark. A. islandiccis a reliablespecies for amino acid geochronology,but D/L ratios in Turuitellashow considerablescatter in many sites,and we suggestthat the mean D/L ratios in all Turritellaanalyzed in this study +20% stroutObe consideredto havea confidencelevel no smallerthan (1o). The analyticalresults for Eemiansites are summarizedin Table 13; site locationsare givenin Fig.4. Mean D/L ratiosin Arcticarange between 0.14 and 0.21,with most sites separationof the two faunalgroups: the lrouping between0.15 and 0.17. There is no clear Eernsei sitestend to be lower,although the singleArctica from RistingeKlint gavesome of the highestratios (0.20). The SkaerumhedeSea group includes high (Strandegaard), and intermediate(Stubberup Have) ratios.The mean D/L ratio in the moderate-rategroup (Arctica,Macoma, Venerupis, Corbula, Mya) is 0.17 + 0'02 (80)' The implicationsare that if thesedeposits do not belongto the sameinterglacial, the time betweenthe two extremevalues is not greatand doesnot appearto follow the established faunal/geographicgrouping. At presentthe geological,biostratigraphic and amino acid evidenceappears to supporta strongcase for a similarage for thesedeposits. Nevertheless, u,e acknowledgethe fact that the evidencefrom somesites is debatableand acceptthe possibilitythat more thanone age may be represented.None of the sitesgave ratios as high is in Holsteiniandeposits or aslow asin Middle Weichseliandeposits. RistingeKlint At RistingeKlint more than 30 glaciotectonicallystacked floes occur, with a nearly identicallithostratigraphical sequence within each floe (Sjorring,1983). From the base the stratigraphyis: (1) glaciolacustrineclay; (2) i0 cm sand with freshwater upwards, 'Cyprina-clay', molluscs,assumed to b- Early Eemian;(3) Eem with a brackish-waterfauna ar the base,and a more open-marinefauna higher up (Madsenet al., 1908).This clay 258 Gifford H. Miller and Jan Mangerud containsa rich, thermophilousmarine fauna, including the speciesconsidered guide-fossils in the Eemian; (4) sand; (5) till; (6) glaciofluvialsand; (7) till. Unconformablyoverlying theseunits is a further till bed. Together with the shellsfor amino acid analysis,we collectedpollen samplesthat were countedby J.I. Svendsen(University of Bergen).The pollenstratigraphy (Fig. 12) is similar to the Eemianpollen stratigraphyin southernDenmark-northern Germanv,and supports an Eemian agefor the sequence.Picea rises to more than 8"h at the top of the diagra., but the diagnosticCarpinus-Picea phase of the upper part of the Eemian is missing,which shows either erosion, and/or a regressionwell before the end of the interglacial.The significantrise of Tilia (to 6%) in the highestsample may be the start of thehajor Tilia phaseknown from nearby localities(S.T. Andersen, L975).The marinetransgresiion (the boundary betweenlacustrine and marine beds)took place before the Corylui rise, which accordingto Miiller Q97aa)occurred only 750years after the start of the Eemian.This is in accordancewith the resultsof Jessenand Milthers (1928,pp. 178-179)from Stensigmose and suggeststhat the Danish sitesmay include more of the interglacialthan sorneof th" Dutch localities. Sampleswere collected by Mangerudin 1983,under the guidanceof S. Sjdrringfrom floe 15 (Sjorring, 1983).A single,whole specimenof. Arctica islandicawas recoveredfrom the upper portion of the Cyprina (Arctica) clay and several individuals of. Venerupis were collectedfrom lower levelsin the samebed. Nine separatepreparations of the singleArctica specimengave a mean D/L ratio of 0.20 + 0.03. Although the mean ratio is higher than most other nearby Eemian sites(Table 13) the large standarddeviation indicatei thar the +:Pollen present< l7o FIG. 12. Pollen diagram for the interglacialbeds in floe 15 at RistingeKlint. Denmark(J.I. Svendsen,Univ. of Bergen,unpubl. data). Aminostratigraphy of European Marine InterglacialDeposits 259 TABLE 13. D/L ratios in shellsfrom Eemian localitiesin Denmark Lab ID Species Mean Localitv AAL-2604 Turritella communis 0.200 0.000 2 Hjelm Nakke AAL-2600 Turritella communis 0.162 0.001 5 Hojer AAL-2599 Turritella communis 0.230 0.007 5 Hojer AAL-3698 Arctica islandica 0.131 0.003 3 Mommark AAL-26r2 Arctica islandica 0.144 0.030 5 Mommark AAL-3615 Arctica islandica 0.208 0.02t) 6(1) Ristinge AAL-3916" Arctica islandica 0.182 0.033 3(1) Ristinge AAL-3616 Mya truncata 0.223 0.002 3(1) Ristinge AAL-3913 Nassarius reticulatus 0.099 0.008 3(1) Ristinge AAL-3612 Venerupis senescens 0.150 0.025 3 Ristinge AAL-3614 Venerupis senescens 0.160 0.02,1 3 Ristinge AAL-3613 Venerupb senescens 0.r72 0.006 2 Ristinge AAL-3616 Venerupis senescens 0.204 0.023 3 Ristinge AAL-2610 Arctica islandica 0.r47 0.026 3 Stensigmose AAL-3697 Arctica islandica 0.i56 0.006 3 Stensigmose AAL-3509 Bittium reticulatum 0.140 0.026 t Stensigmose AAL-2611 Cardium edule 0.174 n.0.16 4 Stensigmose AAL-4005 D ivaricella divaricata 0.261 0.030 4 Stensigmose AAL-3982 Littorina littorea 0.086 0.008 3(1) Stensigmose AAL-4025 Macoma balthica 0.r59 0.012 4 Stensigmose AAL-4007 Macoma balthica 0.17? 0.013 1 Stensigmose AAL-4034 Mactra subtruncata 0.139 0.015 3 Stensigmose AAL-3510 Spbula subtruncata 0.145 0.033 4 Stensigmose AAL-3508 Venerupis senescens 0.149 t).006 4 Stensigmose AAL-4035 Venus gallina 0.143 0.020 3 Stensigmose AAL-3699 Arctica islandica 0.205 0.012 3 Strandegaard AAL-zffi'l Arctica blandica 0.209 0.010 3 Strandegaard AAL-4058 Corbula gibba 0.229 0.006 2 Strandegaard AAL-2603 Turritella communis 0.190 0.019 5 Strandegaard AAL-z795 Arctica islandica 0.156 0.019 8(1) Stubberup AAL-2605 Turritella communis 0.222 0.032 4 Stubberup AAL-3915 Arctica blandica 0.171 0.01,5 5 Trappeskot AAL-2608 Arctica islandica 0.174 0.012 3 Trappeskov AAL-3983 Littorina littorea 0.1l4 0.014 5 Tonder AAL-4008 Venerupb senescens 0.180 0.016 5 Tonder *Repreparationof same shell as utilized in AAL-3615. sample cannot be shown to be significantlydifferent thanArctica from other Eemian sites. A single specimen of.Mya also produced a mean ratio higher than expectedfor an Eemlan sire. To check the validity of the high ratios obtained from the single individuals of.Arcrtca andMya,12 individuals of Veneruprscollected from four levelsin the marine clay bed u'ere also analyzed, and produced a mean D/L ratio of 0.16 + 0.02. The similarity in Ristinge Klint Vener,Lpisratios with other Eemian correlatives suggeststhat the Mya and Arcrica resultsrepresent aberrant individuals, reinforcing our conclusionsthat reliable D/L values require the analysisof several individuals. Trappeskov Klint Trappeskov Klint is situated only 50 km west of Ristinge Klint. The entire sequence. describedby Madsen et al. (1908), has been glaciotectonicallyupthrusted and moderately disturbed. Madsen et at. (1908) stressthat the lithostratigraphyand biostratigraphyof the 260 Gifford H. Miller and Jan Mangerud Eemian beds (freshwatersand and'Cyprina clay') are nearly identicalto the beds of RistingeKlint. At the baseis a freshwatersand, overlain by brackishwater sediments and 1.5m of 'Cyprinaclay'. The sequenceis cappedby two (?) tills and glaciofluvialsediments. Samplesfrom the collectionat the GeologicalSurvey of Denmark were submittedby Petersenin 1980.Eight preparationsof. Arctica islandica gave a meanD/L ratio of 0.7'7+ 0.01.which confirman Eemianage for the site' Mommark This sitewas described by Johnstrupin 1882and referred to by Madsenet al. (1908),who correlateit with RistingeKlint, TrappeskovKlint andother Eemian sites on the basisof its mollusc fauna. Only Arctica islandicafrom the Cyprina (Arctica) clay was supplied by Petersenfrom the DGU collection.The first setof analyses,confirmed by a secondseries (Table 13), gaveratios lower than in Arcticafrom other Eemiansites. Citing Johnstrup (1982),Madsen et al. (1908,p. 128)indicate that neitherof the classicalEemian index speciesBittium reticulatum nor Venerupissenescens were found at Mommark. For the momenr,we assumethat the site is probablyEemian and that the ratiosare simplyat the lowestend of the Eemianrange. but additionalanalyses would be importantto testwhether the site might be from a post-Eemianhigh sea-levelevent. Stensigmose The sequenceat Stensigmoseis glaciotectonicallydisturbed by thrustfaults. According to Madsenet at. (1908)the lithostratigraphyis, however,clear: at the baseare lacustrine sediments,overlain by marine clay('Cyprina clay'),marine sand ('Iapes sand')and till. Most of the sequenceis similarto RistingeKlint, with respectto both lithostratigraphyand molluscanbiostratigraphy. The main differenceis the presenceof sandat the top of the interglacialsequence at Stensigmosewhich indicates a marineregression. This sand also has an extremelyrich molluscfauna. Madsenet al. (1908)consider the Stensigmosesite to representa more completesequence of the Eemianthan at Ristingeor Trappeskov;both transgressiveand regressivefacies are confirmedby foraminifera-stratigraphy(Konradi' 1976).This interpretationalso agrees with recentreconstructions of sea-levelchange during the Eemian(Mangerud et al., 1981,Zagwijn,1983). Although the foraminiferafauna from Stensigmose(Konradi, 1976)includes more cold-waterspecies than the molluscfauna it generallysupports the proposedinterpretations. Samplesfrom the collectionsat the GeologicalSurvey of Denmarkwere submittedby 'Tapes Petersenin 1980and 1982.Ten generafrom the Sand'have been analyzed for their D/L ratios(Table 13); mostof them stronglysupport an Eemianage. Tonder Boringsin the Tonderregion of southwestDenmark were initially undertaken late in the lasrcenrury (e.g. Madsen et a|.,1908). The arealies outside the Weichselianice-limits and the interglacialdeposits have not beenglacially disturbed. The top of the marinebeds are encountered7 to 10 m below sea-levelwhere they overlieSaalian till, and glacial-fluvial sediments,and are overlain solely by Weichselianmeltwater sediments and Holocene depositssu.r,-u.in];';J:.' ;.Hffi:ffi;" o'ast,",..",;:l age based both on stratigraphic position and presenceof Lusitanian molluscan elemenrs such as venerupis senescens,Divaricella divaricata and others (Hansen 1965). Samplesof Venerupissenescens and Littorina littoreafrom 11.2to 12.7 m depth in Tonder boring 4 from the Geological Survey of Denmark collectionswere supplied by petersen rn 1984.The D/L ratios (Table 72) in Venerrzplsconfirm the correlarion of Tonder ro orher Eemian sitesin Denmark and NW Germany and the Littorina valuesprovide an important Eemian calibration point for this slow-epimerizingspecies. H6jer Marine depositsof supposedEemian age have been encounteredin boreholesalons the coast west of Hojer where they rest on till, glaciofluvial sandsor Eemian limnic depJsits. The marine beds are clayey at the base, coarseningupward into sand at the toD und o.. overlain by Weichselian glaciofluvial sands. The region lies beyond the limits of Weichselian ice in Denmark, and the depositsare consideredto be in situ.They were first describedby Nordman (1928) while the microfauna was describedby Sorensen(tOaOy. fn. foraminiferal faunas in the interglacial beds include boreal-lusitanian shallow-water assemblagesof a composition similar to those found in Eemian sitesin NW Germanl, and Denmark. In 1981K.-L. Knudsen provided us with samplesof Turritella communis from 19 and 24 m depth in the interglacial beds of a boring near Hojer. D/L ratios (Table 13) are within the range of Eemian values. Strandegaards D1,rehave A clay rich in molluscs of interglacial aspect occurs ar rhis site (Odum. 1933). The interglacial clay is less than 1 m thick and is strongly folded and thrusted. According ro Petersen and Konradi (I974), the clay is overlain by three till-beds, allsepararedbv sand or gravel. They assumethe clay is of Eemian age, from the characteristicsof the moilusc and foraminifera faunas. Samples were collected by Petersenand submitted by him in 19g0. Arctica, Corbula andTurritella have been analvzedfrom the site; all vield D7L ratios at the very highest end of the range for these taxa in nearby Eemian sites. The mean rario in Arctica' in particular, lies above the other secureEemian Arctica ratios from this resron. Tentatively we considerthe high ratios to indicate a slightly warmer thermal historv th-anat other sitesin the region, rather than indicatingan older age. Stubberup Have A marine clay, rich in molluscs,is found beneatha till bed (ca.5 m) ar this site (Odum. 1933).Well-sorted sand and clav without fossilslie below the clav and below these beds there is anothertill. Thrust-planesare frequentin the interglacialbeds. The marine clay is correlated with the interglacial beds at Strandegaardon the basis of the mollusc fauna. Arctica and Turritella were submitted by Petersenin 1980;Arctica yietded typical Eemian D/L ratios. 262 Gifford H. Miller and Jan Mangerud Hjelm Nakke Below a complex sequenceof tills and glaciotectonicallydisturbed sedimentsoccur beds of marine sand (10 cmj and clay (20-50 cm) which are also glaciotectonicallyup-thrusted. Berthelsen et al. (\977) correlated the sand and clay with StrandegaardsDyrehave and Stubberup Have by their similar mollusc faunas. Turritella collected by Petersen were submittedby him in 1980. WEICHSELIAN SITESIN DENMARK SkaerumhedeBorings The Skaerumhedeborings from northernJutland (Fig. a) weretaken from an elevation about24 m asl.The first boring(Jessen et al.,I9L0',Knudsen and Lykke-Andersen'1982) penetrated200 m of Quaternarysediment before reaching bedrock. Close to 20m of till and glacial-fluvialsediments overlie bedrock at the baseof the coreand arein turn overlainby IZf * of marine sedimentrhat are cappedby ca. 57 m of glacial-fluvialsediment. The SkaerumhedeII boringpenetrated the upper 122m of glacial-fluvialand marinesediment (Bahnsonet at..1974)-Thefauna in the marinebeds follows a generallycold-warm-cold sequence.Samples from the GeologicalSurvey of Denmark from both boringswere submittedby Petersenin 1980,1983 and 1984. A thoroughstudy of isoleucineepimerization through the two boringsis currentlyon- goingand *itt U" reportedin due course.Here we reportonly the analyticalresults of the upp.r f ZOm of the two borings.Four preparationsof a singlefragment of. Arctica islandica from a depth of 98 m (0.082+ 0.002)and 5 preparationsof Macomacalcarea fragments from 66 m below the ground surface(0.093 + 0.007),both in the more recentboring' confirmthat the levelsabove 100 m in the boringare probablyof Middle Weichselianage and younger.Bahnson et al. (1974)suggested that the Eemian-Weichselianboundarv migirttie it t+Om in the originalboring. Samples of Turrilellafrom 95 to 120m depthin boih boringswere analyzed.The similar ratios in both borings(Table 14), suggestthat correlationlspossible, but the ratiosare lower than in Turritellafrom othersecure Eemian localitiessuggesting that the upper 120m of the boringsare indeedpost-Eemian. Hirtshals A sequenceof glacio-tectonicallythrusted and folded silt andsand beds occur at Hirtshals (Lvkke-And".r"n. 1981,1982). Both the foraminiferaand molluscassemblages indicate that mirine conditionscolder than today persisted throughout the depositionof the silt andsand units.Lykke-Andersen subdivided the sequenceinto interstadialsand stadials of supposed MiddleWeichselian age. Radiocarbondates gave ages ranging from 14 to 47 ka. Lykke-Andersen(1982) concludedthat the two datesfrom her zone C between14 and 15 ka BP were severely contaminated,and that all shellswere older than 34 ka' We analyzedshells from zoneC partlyto obtainMiddle Weichselian D/L ratios,but also in part to testwhether the youngradiocarbon dates could be due to mixedpopulations' in Aminostratigraphy of European Marine InterglacialDeposits 263 TABLE 14. DIL ratios in Middle Weichselianshells from Denmark Localrty Lab ID Species Mean u AAL-3704 Hiatella arctica 0.099 0.007 4 Hirtshals.zone C AAL-3114 Hiatella arctica 0.085 0.008 3 Hirtshals. lower zone C AAL-3'726 Macoma calcarea 0.098 0.009 3 Hirtshals. zone E AAL-3301 Zirphaea crispa 0.131 0.030 3 Hirtshals. zone E AAL-3115 Zirphaea crispa 0.119 0.021 3 Hirtshals. zone E AAL-1050 Macoma calcarea 0.110 1 Holmstrup AAL-4057 Macoma calcarea 0.318 0.083 2 Holmstrup AAL-{036 Macoma calcarea 0.26i) 0.071 5 Holmstrup AAL-405i) Macoma calcarea 0.193 0.059 2 Holmstrup AAL-4051 Macoma calcarea 0.076 1 Holmstrup AAL-3113 Hiatella arctica 0.074 0 002 3 Norre Lyngby AAL-3112 Hiatella arctica 0.097 0.017 I Norre Lvngby AAL-2968 Turritella communis 0.125 0.006 2 SkaerumhedeI. 115m AAL-3306 Turritella communis 0.148 0.001 ) SkaerumhedeI. 12-5m AAL-3511 Arctica islandica 0.082 0.002 1(l) SkaerumhedeII. 98 m AAL-1022 Macoma calcarea 0.093 0.007 5 SkaerumhedeII. 66 m AAL-2598 Turritella erosa 0.1l6 0.030 2(l) SkaerumhedeII. 96 m AAL-1597 Turritella communis 0.129 0.002 l(1) SkaerumhedeII, 11tim which casethe glacial-tectonicswould post-date14 ka BP. Hou'ever.we did not haveshells from the same collections that were radiocarbon dated. Hiatella arcticafrom zone C (193 m in Fig. 1 of Lykke-Andersen. 1982)gave a mean D/L ratio of 0.085 (Table 14). Macoma balthica from zone E (ca. 185 m). underlving zone C. yieldeda mean ratio of 0.098.Both sampleswere collectedb1' A.-L. Lykke-Andersenand iubmitted by her in 1982. Samplesfrom the same site of Zirphaea were analyzedbut are difficult to interpret becauseof the lack of comparative specimens.Additional H. arctica collected by Mangerud in 1983 under guidanceof Lykke-Andersen from a site in zone C have a mean ratio of 0.099 similar to the earlier analyses.None of the shellsanalyzed from the Hirtshals sequence yielded a ratio within the range of Lateglacial samples from Denmark (0.052 + 0.007 f.or Hiatella, Miller, 1985). Our results support a Middle Weichselianrather than a 1.1ka age for the Hirtshals collections. Nprre Lvngbl, ln 1982.Knudsen sent two samplesof. Hiatellaarctica from interstadialdeposits at Norre Lyngby. N Jutland. The siteshave been radiocarbondated and in both casesthe resultsare slighllv older than l4 ka (Knudsen, 1978).The mean D/L ratio in one of the collections baied on four different individuals is 0.097whereas in the other collection.three individuals gavea ratio of 0.074.The higherratio stronglysuggests a Middle Weichselianage, whereas ih. lo*.r ratio is significantly higher than ratios obtained in the same taxon dated I to 2 ka rrC vounger. There remains some question whether the age is correct, or if the depositsare of UiAOt. Weichselianage as favored by the amino acid results(see also discussionof the Hirstshalssamples). Gifford H. Miller and Jan Mangerud Hohnstrup Dislocated marine sediments at Holmstrup on the island of Sjaelland have been described by Petersen and Buch (1971). They overlie a till correlated with the Earl.v Weichselian Old Baltic till that overlies Eemian sediments at Ristinge Klint and Strandegaard. The foraminiferal assemblage in the Holmstrup marine sediment is correlated with the Middle Weichselianzone C assemblageat Skaerumhede. Amino acid analyses have been completed on eleven individuals of.Macoma calcarea submitted from the Holmstrup locality by Petersenin 1984(Table 14). The measured D/L ratios scatter widely, ranging from ca. 0.08 to ca. 0.4, suggestingthat the fauna has been derived from deposits of more than one age. Three of the shells gave Weichselian ratios, averagin_e0.11 + 0.03. The remaining individuals all contained ratios similar to or greater than in Eemian representativesof this taxon. These samplesdemonstrate the importance of analyzing individual shells in dislocatedsediments. The three shellswith a mean D/L ratio of ca. 0.11 support the contention of Petersenand Buch (1974) rhat the Holmstrup deposit was emplaced during the Weichselian.However, the apparent mixed ages of the Macomc population analyzed in this study, raises the quesrion of whether the microfauna in the same sediment is similarly of a mixed origin. If rhe stratigraphicinterpretations of Petersenand Buch (I974) and Petersen{1973,1984) are correct, then glaciersmust have reachedDenmark before the Late Weichselian,contrary to the common view of the last decade that the area was glaciated only during the Late Weichselian. POLAND In the lower Vistula valleytwo seriesof marinesediments of assumedEemian age are strarigraphicallyseparated by continentalsediments(Makowska, 1980. 1982). According to pollensrrarigraphy, the lowermost(Sztum Sea) is from early Eemian,whereas the upper (Table15) iTychno*y Sea)is from aroundthe climaticoptimum. All the analyzedsamples are from the Tychnowy Sea beds, and were obtained from A. Makowska in 1982; insufficientmollusc remains were presentin the SztumSea layers for this project. The mollusc fauna are diminutive and relatively rare, suggestingstressed environmental conditions. TABLE 15. D/L rarios in shells from Tvchnowv Sea Sediment in cores from the Vistula Valley, Poland Lab ID Species Mean o n Core. Depth AAL-3r75 Nassarius reticulatus 0.121 0.020 KwidzynXXI. 30 m AAL-3666 Nassarius reticulatus 0.1'79 0.004 2 KwidzynXXI. 30 m AAL-3719 Bittium reticulatum 0.l0i 0.017 J KwidzynXIX. 38 m AAL-3179 Corbula gibba 0.189 0.010 5 Nowinvboring. 62 m AAL-31'77 D iv aric eI la div aricata 0.213 I Nowinyboring. 62 m AAL-3176 Spisula subtruncata 0.131 0.023 A Nowinyboring. 62 m AAL-3171 Venerupis senescens 0.177 0.034 3 Nowinvboring, 62 m AAL-3665- Venerupis senescens 0.219 0.037 3 Nowinvboring. 62 m ' Repreparationof AAL-317'1 Aminostratigraphyof European Marine InterglacialDeposits 265 Nov'iny Boring The lower marine Eemian beds(Sztum Sea)occur at ca. 70 m below the surface.overlarn bv ca. 5 m continental and lacustrinesediments and the upper (Tychnowy Sea)marine beds (lr{akowska, 1980). More than 50 m of Weichseliansediments overlie the younger marine beds. Ratios in five individuals of.Corbula gibba (0.19) and six analysesof three Venerupis (0.20)confirm a correlation to the type Eemian. A single specimanof Divaricella divsricata gave a ratio of 0.24, whereas four Spisulasubtruncata produced a mean ratio of 0.13, but u'ith considerablescatter. Kv'idzyn From the Tychnowy Sea beds (Makowska, 1980), Nassariusreticulatus and Bittiunt reticulatumwere analvzed (Table 15). PRE-EEMIAN SITESIN NORTHWESTGERMANY The oldestglacial sediments identified in NW Germany are tills of Elsterianage (Ehlers. 19E-l:Ehlers et a|.,1984); the Holsteinianbeds overlying the Elsteriantills are the oldest knorvnPleistocene marine deposits.The top of the Holsteinianmarine bedsare generallr 20 to 25 m below sea-level,except where upthrusted(Ehlers. 1983).Thel'are wide-spread and are consideredto be an importantmarker horizon (Duphorn et al..1973 Behre et al.. 1979). We have determinedD/L ratiosfrom molluscsat six sitesof Holsteinianage (Table 161. Unfortunately, except for Wacken, most sites are represented by only' a single genus. reducingthe power of the amino acid method for correlation.With the exception of Hummelsbrittel,the resultsdo supportthe assumptionof a similarage for all deposits.and that they are clearlv older than the Eemian. ll'acken The glaciotectonicallyupthrusted sequence at Wackenconsists of 3 m of marine clal' and 'flaser 'white sandoverlain by sand' and sand' followed by a peat. Pollen stratigraphvof the lou'er marine beds indicatesinitial accumulationearly in the Holsteinian(Menke, 196E. 1980).The overlying'white sand'is interpretedas niveofluvial-eolian(Diicker, 1969).The palynologyof the overlying peat demonstratesthe developmentfrom a Betula-Pinus phase into full interglacial vegetation. the Wacken Interglacial. The main stratigraphic significanceof this site is that it documentsa cold phase between the Holsteinian (sensustricto) and the overlying Wacken Inter-elacial.Manv workers (e.g. Duphorn et a\..7973: Mriller, 1974b;Menke, 1968,1980) consider the Holsteinian(sensrr stricto)and Wacken to be a single interglacialcomplex. containing a short-lived cool intervalbetween two warm intervals,all of which underliesthe entire Saalian(sensu lato). Others (e.g. Cepek and Erd, 1982)consider the Wacken Interglacial(correlated with the Domnitzian of East Germany) to be separatedfrom the Holsteinian (sensustricto) by the first glaciation of the Saaliancomplex. 266 Gifford H. Miller and Jan Mangerud TABLE 16. D/L ratios in pre-Eemian shells from West Germanv Lab ID Species Mean o Locality AAL-372.r Littorina sp 0.263 0.041 o(jr Halstenbek AAL-3851. Littorina sp 0.213 0.047 3 Halstenbek AAL-.1061 Spisula subtruncata 0.309 0.02r 5 Billbrook AAL-3182 Spisula subtruncata 0.280 0.015 2 Heide-Holtex,J7 m AAL-2613 Cardium edule 0.278 0.021 A Hummelsbi.ittel AAL-.1062 Macoma balthica 0.382 0.016 5 Hummelsbiittel AAL-3183 Portlandia arctica 0.315 0.007 L\ 6) Mtildsberg AAL-3105 Macoma balthica 0.289 0.004 3 Scharhorn55, 3tl m AAL-3.106 ,\,lacoma baLthica 0.289 0.002 3 Scharhorn51,34 m AAL-3909 Arctica islandica 0.305 0.0c8 \(1) Wacken AAL-3617 Arctica islandica 0.286 0.020 4 Wacken AAL-3911 Corbula gibba 0.263 0.027 5 Wacken AAL-3721 Littorina litorea 0.221 0.030 6( Wacken AAL-3912 ,Vacoma calcarea 0.293 0.013 2 Wacken AAL-3910 ,Vacoma sp 0.271 0.048 5 Wacken AAL-37251 Mva truncata 0.223 0.057 3 Wacken AAL-3919+ Mva truncata 0.215 0.052 3 Wacken AAL-2943t Mya truncata 0.253 0.015 3 Wacken AAL-3618 Nucula sp 0.31'7 0.045 3 Wacken *Repreparation of the same three shells utilized in AAL-3724. tAll three preparationsutilized the same three individuals. Molluscsfrom the marineHolsteinian at Wackenwere submitted by'B. Menke in 1982 and additionalspecimens were collected by J. Mangerudin 1983under guidance of Menke. The analyzedtaxa include Arctica, Corbula, Littorina. Macoma and My'a. The averageD/L ratio in moderate-ratetaxa is 0.28+ 0.02(43) compared to a ratioof 0.17+ 0.01(29) in nearbyEemian representatives of the sametaxa. Heide Hinsch submittedSpisula subtruncata from assumed Holsteinian beds at 37 to 38 m depth. The meanD/L ratio is 0.28 + 0.02 (Table 16). similar to other membersof the moderate-ratetaxa. Muldsberg The Holsteinian marine sediments at Muldsberg began to accumulate during late Elsterian as indicated by the cold-water fauna; the fauna becomes increasinglythermo- philous upward. Portlandia arcticasubmitted by Menke in 1982from the late Elsterian beds gave a mean ratio of 0.32 + 0.01 (2). Scharhorn Specimensof iVacoma balthica from borings on the island of Scharhorn in the German Bight just ourside the river Elbe, were provided by G. Linke in 1983.Pollen analysisof the cores confirms an early Holsteinian age for levels from which the shells were obtained (below the Abies and Carpinu.rincrease; Linke. 1970).Three individuals of.Macoma from Aminostratigraphyof European Marine Interglacial Deposits 267 33.10to 34.10m depth in Boring 55 and an additionalthree indrvidualsof the samespecies from 33.15to 34.0 m depth in Boring 57 gaveidentical results (0.29 + 0.01; Table 16). Hamburg Region From Hamburg, the type area for the Holsteinian, we have analyzed shells from two sites:Hummelsbiittel and Billbrook. The palynology of the Holsteinian beds at Hummels- btittel (Hallik, 1960) demonstratesthe difference in vegetation changes between Hol- steinianand Eemian interglacials.and the typical Holsteinianpollen stratigraphyis usedfor regional correlation. Whether the Holsteinian pollen stratigraphyis unique and was never repeatedin another interglacialis difficult to show. We assumethat the similar palynology of the'Holsteinian'sites we have analyzedshows them to be the same age. Huntmelsbilttel The Hummelsbiittel locality is situated in the city of Hamburg. It is in an old clay pit cut into glacial-lacustrineclay of late Elsterian age. The interglacialbeds comprise 0.7 m of freshwater gyttja overlain by 3 m of fossiliferousmarine sands,all of which is capped b1' Saaliantill (Grube. 1959, 1963).The entire complex has been upthrusted from its original position by subsequentglacial tectonics. Both the pollen (Hallik, 1960) and foraminiferal assemblages(Knudsen, 1979) indicate full interglacial conditions during deposition of the marine beds. The late Elsterian arctic-subarctic faunas found belou' other Holsteinian marine layers in NW Germanv are not present in the Hummelsbtittel sequence.presumablv becauseof the freshwater sedimentationat the start of the interglacial. Analysis of five individuals of Macoma balthica supplied from this site by Linke gave a mean D/L ratio of 0.38 + 0.02. well above the ratios measuredin the same taxon from Holsteinianbeds in the nearbyScharhorn borings (0.29) or at Wacken (0.29).Because the Hummelsbiittel site is Holsteinian by definition, we can only note the discrepanc)' at present. Additional amino acid analyses are required to verify whether the difference apparent on the few analvsesavailable is real or an artifact of analytical errors. Billbrook The Billbrook locality, also from the Hamburg area, is an object of ongoing research.It is considered to represent the same sedimentary sequencethat is exposed in the Hummels- biittel section except in an undisturbedposition. The late Elsterianclay (Lauenberg Clay) is overlain by freshwatergyttja and 25 m of Holsteinian marine silt and sand. The marine beds represent the lower portion of the temperate stage of the interglacial (Linke, written contm. 1985). Spisulasubtruncata from the marine beds were supplied by Linke from a depth of 42 to 44 m below the surface; none of the primary' moderate-rate taxa were present.The D/L ratio measuredin 5 individualsis 0.31 + 0.02.similar to that in the same taxon from the Holsteinian beds at Heide. Halstenbek Littorina were collected i5 m asl from the Halstenbek site. u'heremarine beds, presumed to be ice-thrustedupwards, occur between6 and12 m belou'ground level. The sequence 'Gifford 268 H. Miller and Jan Mangerud was thought to be of Holsteinian age, but recent U/Th dates suggestthev are Eemian (Stremme, personal comm. 1982). The specimenswere provided by Hinsch in 1982.The mean D/L ratio derived from nine preparationsof three individuals of.Littorina sp. (0.25 + 0.05) indicatea Holsteinian(or possiblyeven older) age;the depositis certainlvolder than Eemian. EEMIAN SITES IN NORTHWEST GERMANY The end morainesof the Weichselianglacial maximum run throughSchleswig-Holstein (NW Germany).Outside that boundary,the marineEemian beds are not disturbedby ice. However, the top of thesebeds lie approximately5 to 7 m below presentday sea-level (Duphornet al., 1973:Behre et a|.,1979)probably due to generalsubsidence of the region sincethe Eemianand samplesare thereforeonly obtainedby coring.A paleogeographical map indicatingthe Eemian shorelineis givenin Behre et al. (1979).The samplesfrom Rodemis,Schnittlohe, Offenbi.ittel and Siidermarschare all from coresin the lowlands outsidethe Weichselianglacial limits. For Rodemisand Schnittlohean Eemian age is shownby pollen stratigraphy.For the others,an Eemianage is assumedbv correlation. Holnis lies insidethe Weichselianboundary and providesa link betweenthe Danishsites and the other West Germanylocalities. The samplesfrom Rcidemiswere submittedby Menke in 1980;all otherswere providedto Miller by Grube, Menke and Hinschduring a visit to GeologischesLandesamt Schleswig-Holstein, Kiel, in 1982. Rddemis In the RodemisIV coring,marine silt from ll.2 to 74.7m depth is assignedan upper Eemianage from the pollenstratigraphy (Menke. 1985). Littorinalittorea from 12.3to12.6 m depth were suppliedby Menke in 1980.The measuredratios in this slow-epimerizing genus(0.098; Table 17) supportan Eemianage. Schnittlohe The SchnittloheI coring site is situated 1 m asl; Eemian marine sedimentswere encounteredbetween 12.3 and 17.7 m below ground surface.The guide fossilfor the Eemian. Bittium reticulatum,occurs through most of the core (Hinsch. 1985). and Venerupissenescens is alsocommon. The D/L ratiosdetermined in V. senescens,B. reticulatumand Corbttla gibba from 13, 15 and 16 m depthin the coreare listedin Table 17.Pollen stratigraphically (lvlenke. 1985), the 13m sampleis from the upperpart of the Picea- lowerpart of theAbies zone. the 15 m sampleis from the baseof the Tiliazone. and the 16m samplepredates the Corylus zone. The lack of a clearincrease in ratiosdown-core is discussedin Part l. Nevertheless.the meanratios in Venerupis[0. tZ + 0.02(9)] andCorbuls [0. tg + 0.02(9) j areas expected for Eemiansites. Aminostratigraphy of European Marine InterglacialDeposits 269 TABLE l7. DIL ratios in shellsfrom Eemian depositsin West Germany in core Lab ID Species Mean o n Localitl . Depth AAL-3180 Arctica islandica 0.153 0.010 4 Holnis AAL-3008 Arctica islandica 0.165 0.008 4 Holnis AAL-3723 Arctica islandica 0.161 0.014 Holnis AAL-3181 Spisula subtruncata 0.111 0.019 2 Offenbiittel AAL-2581 Cardium edule 0.137 1 Rodemis IV, 14.5 m AAL-2580 Cardium edule 0.106 0.010 4 RodemisIV, 12.5m A Rodemis IV, 14.5 m AAL-2582 H 1-drobia ulvae 0.105 0.033 AAL-2583 Hydrobia ulvae 0.103 0.012 6 Rddemis IV. 12.5 m AAL-3635 Littorina littorea 0.098 0.004 3 RodemisIV, 12.5m AAL-2580 Turritella communis 0.128 0.001 2 Rodemis IV, 11.5 m AAL-3184 Bittium reticulatum 0.099 0.007 4 SchnittloheI. 1,1.8m 0 016 2 SchnittloheI. 1.1.8m AAL-3632 Bittium reticulatum 0.104 'tl5\ AAL-3633 Bittium reticulatum 0.093 0.019 SchnittloheI. 16.2 m AAL-3003 Corbula gibba 0.185 0.007 2 SchnittloheI. 13.1m AAL-3185 Corbula gibba 0.191 0.004 2 SchnittloheI. l{.8 m AAL-300-s Corbula gibba 0.204 0.035 3 SchnittloheI. 1.1.8m AAL-3007 Corbula gibba 0.16-s 0.000 2 SchnittloheI. 16.1m AAL-3002 Vcnerupis senescens 0.155 0.008 3 SchnittloheI, 13.2 m AAL-3001 Venerupb senescens 0.184 0.019 3 SchnittloheI. ll.8 m AAL-3006 Venerupis senescens 0.157 0.006 3 SchnittloheI. 16.2m AAL-3'720- Mya truncata 0.164 0.024 4 Sridermarsch AAL-2941 M v-atruncala 0.157 0.008 4 Siidermarsch AAL-2942 Turritella communis 0.162 0.002 4 Siidermarsch 'Repreparation of same four individuals used in AAL-2911 Offenbiluel Thissite is locatedclose to Schnittlohe(Menke. 1985;Hinsch. 1985). Two individualsof + Spisutasubtruncata from a depth of 14.0to 14.3m gavea meanD/L ratio of 0.11 0.02. This ratio is lower than expectedfor an Eemian site; however,onlv two valveswere analyzedand the taxa has not been thoroughll'studied for its reliabilit.vin amino acid geochronology.At presentwe can safelyconclude that the site is no older than Eemian. Sridermarsch From the Siidermarsch29 boring the following samples were analvzed: four large fragmentsof Turritellacommunis from 20.5to 21.5m depthand four umbo fragmentsof M. truncata.The meanD/L ratios(Table 16)support an Eemian age. Holnis This site is assumed to be Eemian from its sedimentarl' (clayey) facies. It is situated within the Weichselian glacial boundary, close to the Danish sites Stensigmoseand Mommark. Our main interest in the site was the occurrenceof Arctica islandicaat a depth of 24 to 27 m. Nine individual fragments of this specieswere prepared at three different + tlmesover a rwo year period. The resultsare consistentin all preparationsaveraging 0.16 0.01 (9). similar to the ratio in Arctica from most Danish Eemian sites. 270 Gifford H. Miller and Jan Mangerud THE NETHERLANDS Amersfoort Interglacial deposits near Amersfoort were named Eemian after the nearby river Eem (Harting. 1874;referenced in Zagwijn.1961).A descriptionof the stratotypenearthe city of Amersfoort is given in Zagwijn, 1961.The lower part of the Eemian beds, includine the sedimentation euercus (oak) zone, consistsof peat and lacustrine gyttja. Brackish-water began during the Corytlus(hazel) increaseand marine sedimentationcontinued through the Cirpinw (hornbeam) zone ending with a return to brackishsedimentation during the Picea (spruce)zone. This sequenceis, by definition, the Eemian. For this project it was essentialto obtain reliable D/L ratios from a variety of taxa from the type locality. Existing shell collections had been heated during processingand were not suitable for the method (note the higher ratios for Venerupis and Divaricellain the heated material; Table 18). Under the direction TABLE 18.D/L ratiosin shellsfrom interglacialsites in The Netherlands Lab ID Species Mean Locality, depth AAL.279I Bittium reticulatum 0.r62 0.021 3(6) Amersfoort,17-18 m AAL-3637 Bittium reticulatum 0.r79 0.021 s(8) Amersfoort,17-18 m AAL-3638 Bittium reticulatum 0.165 0.022 s(8) Amersfoort.l8-19 m AAL-2792 Bittium reticulatum 0.17"1 0.012 3(e) Amersfoort,20 m AAL-2789 Cardium edule 0.1't5 0.029 3 Amersfoort,17-18 AAL-2790 Cardium edule 0.141 0.067 3 Amersfoort,20 m AAL-2794 D iv arice lla div aric ata 0.290 0.024 3 Amersfoort.20 m AAL-.{218 Hiatella arctica 0.196 0.019 .t Amersfoort,26 m AAL-2788 Nassariusreticulatus 0.1,10 0.017 3 Amersfoort,20 m AAL-2'793 Venerupis senescens 0.201 0.031 6 Amersfoort,17-18 m AAL-3394 Venerupis senescens 0.210 0.022 4 Amersfoort,17-18 m AAL-3221 D ivarice lla div aricata 0.355 0.037 Amersfoort(heated) AAL-3395 Div aricella divaricata 0.368 0.051 + Amersfoort (heated) AAL-3390 Spisula subtruncata 0.173 0.025 A Amersfoort(heated) AAL-3225 Spisula subtruncata 0.189 0.039 3 Amersfoort(heated) AAL-3393 Venerupis senescens 0.256 0.019 4 Amersfoort(heated) AAL-36{0 Bittium reticulatum 0.160 0.022 s(8) Bergen19.4'/263,66 m AAL-3391 Corbula gibba 0.229 0.011 4 Bergen194/263,71 m AAL-3387 Macoma balthica 0.172 0.012 Bergen194/263,59 m AAL-3386 Macoma balthica 0.179 0.0i7 A Bergenl9,4"/263, 66 m AAL-3223 Mva truncata 0.176 0.021 5 Bergen194/263. 68 m AAL-3226 Spisula subtruncata 0.181 0.041 3 Bergen194/263,66 m A AAL-3396 Spisula sublruncata 0.16.1 0.039 Bergen194/263,66 m AAL-3389 Spisula subtruncata 0.r54 0.022 Bergen194/263.77 m AAL-3392 Venerupis senescens 0.2r0 0.036 Bergen19A/263.6E m AAL-3222 Mya tuncata 0.200 0.014 3(1) Bergenboring AAL-4230 Macoma balthica 0.035 0.002 3 Castricum,21 m (4.,ika) AAL-.1221 Littorina Iittorea 0.122 0.0r3 5 Castricum.38-39 m AAL-.1223 Macoma balthica 0.162 0.016 4 Castricum,38-39 m AAL-122'7 Macoma balthica 0.165 0.020 4 Castricum.45-46 m AAL-1221 Macoma balthica 0.216 0.060 Castricum.38-39 m AAL-.{220 Mya truncata 0.r55 0.016 A Castricum38-39 m A AAL-4228 Mva truncala 0.180 0.019 Castricum,46-17 m ,16-.17 AAL-1215 Littorina littorea 0.il9 0.007 5 Zunderdorp, m AAL-,12i6 Macoma balthica 0.189 0.018 5 Zunderdorp.16-17 m AAL-3388 Macoma balthrca 0.458 0.062 4 NoordBergum 6D/3S. i9 m Aminostratigraphy of European Marine InterglacialDeposits 271 of W. Zagwijn, the Geological Survey of the Netherlands undertook a special boring for this project at Amersfoort. According to de Jong (written comm., 1981) the lithostrati- graphy of the Eemian beds in the new core is similar to that of the Amersfoort boring I of Zagwrjn (1961). The molluscs were identified by G. Spaink and the samples,all of which were from the Carpinu.rzone, were submitted by de Jong and Zagwijn in 1981. The Eemian marine beds at the type locality were deposited far inland from the coast near the head of an estuary during the highest stand of the Eem Sea. Consequently, the local marine environment was of reduced salinity and the fauna is restricted with several open-marine speciesconspicuous by their absence.The only representativesof the core taxa used in this survey are V. senescens(0.21), and a few diminutive shells of Hiatella (0.1e). Bergen Boring Becauseof the limited fauna in the Amersfoort boring at the type Eemian locality. we choseto analyzeadditional shellsfrom nearby borings that penetrateddeeper water, more open-marine conditions from the last interglacial and for which the shell material had not been heated during the extraction procedure. Zagwijn submitted molluscsof severaltaxa in 1983 from a boring near the city of Bergen (boring 19A1263.unpublished) in which fossiliferous marine beds were located between 59 and 80 m depth where they directlv overlie Saalian fluvial-glacial sediment. The boring is situated in an old tidal inlet that connectedIjsselmer with the Eem Sea; a similar inlet also existedduring the Holocene but was later abandoned. Pollen analysisof a nearby boring indicatesthat the marine beds are of late Eemian age (pollen zone E6b, Zagwijn, personal comm., 1984). The ratios in Venerupisand Spisulasubstantiate the biostratigraphiccorrelation of the two borings. Two other important species analysed from the Bergen boring. Macoma balthica and My-a truncata, gave D/L ratios similar to those in Venerupis(Table 18). A single whole valve of Mya truncatawas also submitted from a different boring in the same region (AAL-3222, Table 18). Although the precise location of the boring and depth from which the shell came are not known, the regional similarity in subsurfacesediment around Bergen allows us to correlate to the Eemian marine beds of boring 19A1263. Castricum The boring at Castricum (i9C/648), 15 km south of the Bergen boring, u'as selectedfor additional analysisbecause the Eemian marine bedsthat were encounteredbetween 38 and 46 m depth contained specimensof severaltaxa usedin other localitiesin this study. and the shellshad not been heated above 50'C during their extraction. Samplesof Littorina littorea and Mya truncata were given to Miller by Zagwijn and Meijer during a visit to Rijks GeologischeDienst, Haarlem in 1984. Zunderdorp The boring at Zunderdorp (25EBa ), north of Amsterdam. contained Eemian deposits correlative with pollen zone E4l5 and a molluscanfauna indicativeof an environment in the intertidal zone. Of particular interestwas the occurrenceof Littorina and Macoma. Samples H Miller andJan Mangerud 272 Gifford in 1984.The measuredD/L ratios(Table wereprovided by Zagwijnand Meijer' Haarlem' 18) supportthe Eemiancorrelation' Noord Bergum the olderinterglacial sequences, we analvzed ln an attemptto providelimiting values on Bergum(Boring 6D/38' unpublished)'In this Macomabalthicafrom a boringnear Noord fill wlich containsa Holsteinianpollen spectra' boringa Saaliantill overliesa riarine valley bedsof the Elsterianglaciation' These The valleyhad beeneroded into glacial-lacustrine are'cromerian'marinesediments; the shells bedsin turn fill " ""u;;,;i;n. uui" of which analyzedcamefromthesebedsatadepthof5gto60m.Theshellsmayhavebeenheated duringextraction.Thesitehasbeenknownsincethelg30s,butthemarinebedswere originallythoughttobeHolsteinian;onlyrecentlyhavethey'beenreinterpretedtobe .Cromerian,(e.g.TerWee'1983),morespecifically.CromerlV'(Zagwijn'writtencomm.+ 'late cromerian'' The measuredD/L ratio [0'46 1gg0).They will be referredto hereas 0.06(4)] supportsa pre-Holsteinianage' BELGIUM AND NORTHERN FRANCE in boringsin the Flemishvalley and Interglacialmarine deposits have been penetrated Formationincludes the tidal flat deposits alongthe coastalpruin ofn.tgium. The Oostende generallyconsidered to be a guidefossil for that contain venerupis senescens(Paepe,1965), an interglacialfreshwater bivalve (see e'g'' the last interglacialand Corbicuia ftiminatis, peatdeposits overlying the marine paepeet al.,lgSlfor a review).Pollen spectra within correlationwith the Eemian(pollen zones E4a sandsof the oostendeFormation indicate a of the molluscanassemblages has not to E6b) in the Nethe.tands,but a detailedtreatment molluscanassemblages are from middle yet been published.ihe implicatiorris that the at-,1983)recent survey of molluscsextracted Eemiantime. However,Spaink's (Denys et sedimentsnear De Panne,on the western from a boringtnrough sufiosed Eemianmarine the Frenchborder indicated a lack of diagnostic BelgianCoastal pta1-n tesi ir,an Z km from depositsin- the Netherlandsand the presence faunalelements that characterizethe Eemian demian beds' Spaink raisedthe question of numerousraxa noi f."r.n, in the Dutch age' whetherthese deposits were actuallyof Eemian sp' obtainedfrom a boring near we have analyzedMacoma balthica and venerupls BruggeintheFlemmishValley(Koolkerke3.Boring#368)'inwhichmarineshellswere in the boring(shells were obtainedfrom R' encounrereduetr"een iO.S und i+.g n.,depth Paepeinl981).ThemeasuredD/LratiosinthesetwotaxaSupportacorrelationtothe Eemiansites in the Netherlands(Table 19)' interglacialunit' underliethe oostende Depositsof the HerzeeleFormation, an older for amino acid analyseshave yet to be Formation, but adequatemolluscan material Maco,mabalthica from the Heem obtained trorn g"igi],;' Instead,we have analyzed n-ea-1the-town of Herzeele,northern Brickyardlocated u-fe* km wesrof the Belgianqoraer gavea consistentset of analyticalresults France.Cardium lamarckifrom the ,u-" J.porit to thosein Macoma(Table 19)' The (unlike C. edule),with measuredD/L ratiossimiLr guidedby R. Paepeand J' Somm6'The D/L Herzeeleshells were collectedby J. Hollin Aminostratigraphy of European Marine InterglacialDeposits zI-) TABLE 19. D/L ratios in shellsfrom Belsium and Northern France Lab ID Species Mean o Locality AAL-2976 Macoma balthica 0.163 0.028 4 Koolkerke 3 AAL-3664- Macoma balthica 0.201 0.004 4 Koolkerke 3 AAL-4090 Macoma balthica 0.1fi 0.025 A Koolkerke 3 AAL-3302 Macoma balthica 0.179 0.041 2 Koolkerke 3 AAL-3301 Venerupis senescens 0.157 0.063 2 Koolkerke 3 AAL-2970 Venerupk senescens 0.178 0.037 6(4) Koolkerke 3 AAL-3117 Venerupis senescens 0.136 0.014 4 Koolkerke 3 AAL-3296 Corbicula fluminalis 0.343 0.0,{4 3 Zeebrugge AAL-4094fl Glycymeris sp 0.201 0.012 2(1) Zeebrugge AAL-329'7 Glycymeris sp 0.20'7 I Zeebrugge AAL-3298 Macoma balthica 0.312 0.012 2 Zeebrugge AAL-4095 Macoma balthica 0.275 0.011 2(1) Zeebrugge AAL-2940 Mya truncala 0.116 l Zeebrugge AAL-3299r Mya truncata 0.r3"1 0.01i 2(1) Zeebrugge AAL-40911 Mva truncata 0.153 1 Zeebrugge AAL-3299 Hiatella arctica 0.132 I Zeebrugge AAL-3634+ Hiatella arctica 0.120 I Zeebrugge AAL-4103 Cardium lamarcki 0.280 0.033 5 Herzeele AAL-4102 Macoma balthica 0.250 0.085 2(l) Herzeele AAL-4098 Macoma balthica 0.302 0.087 5 Herzeele AAL-4104 Macoma balthica 0.306 0.018 3 Herzeele *Repreparationof samefour shellsas AAL-29'16. iRepreparationof sameshell as AAL-2940. gRepreparationof sameshell as AAL-3299. !1Repreparation of sameshell as AAL-329'1. ratios suggest a correlation with the Holsteinian. supporting a similar conclusion based on the palynology (Somm6 et al.,1978). However, U/Th dating of shells has given an age of 326ka (Stremme,1983). Additional shell materialhas been analyzedthat was collectedfrom an exposureat Zeebruggeon the Belgian Coastal Plain. The age of the deposit and stratigraphic associationof the shellsare obscure. The aminoacid results suggest the shellsare of several ages.including some that appearto be youngerthan the last interglacial.particularh, individualsof Hiatella arctica and possiblyMya truncatc.On the other hand, Corbicula flttminalisyielded D/L ratiossuggestive of a Holsteinianage. We suggestthat the depositis complexand may containshells of significantlydifferent ages. Regardless of stratigraphic complexities,the presenceof molluscswith D/L ratiossignificantly lower than in secure Eemiansites raises the possibilityof a major post-Eemianmarine transgression across the Belgianc_oast. ACKNOWLEDGEMENTS This project was conceived while Miller was a Visiting Professorat the University of Bergen in 1979. The active collaboration and assistanceof colleaguesin all of the countries included in this surveyhas been exceptional;the researchcould not have been undertaken without their cooperation. We wish to thank them all for unselfishlyproviding samplesfrom personal and museum collections and the necessarvdocumentation and details of the collection sites. 114 Gifford H. Miller and Jan Mangerud Amino Acid Geochronology Laborator,v' Amino acid separatlons were accomplished at the the operation of the Laboratort. and Universitv of colorado. we thank Daniei Goter for overseeing Ernst for preparing samplesfor analvsrs. John Rohde, Franco Gaudiano. Steve Schilling and Rick with the contributing authorsas well as Our interpretationshave benefitedfrom numerousdiscussions P.J' Hearty' J'T Andrews and B' additionalcolleagues, r.,;;it D.O. Bowen. N.J. Shackleton' Frenzel. ScienceFoundation Grants AEO- This researchhas been supportedby the United StatesNational from Nansenfondet (Oslo)' The Norwegian 8109013and AEO-83fcb32i';ith suiplemental funds (NAvF) and the Norwegian-Marshall Fund' ResearchCouncil for Scienceand the'Humanities thanks' To all of these individuals and institutions we extend our sincere REFERENCES depositsat Bo on Andersen,B.G., Sejrup.H.P. and Kirkhus,@. (1983)._Eemianandweichselian ru.*oy.Apreliminaryreport.NorgesGeologiske.Undersokelse.S8o'189-201. Jylland,and-the.Eemtan enO"rr"n, S.T.^(1975).The Eemianfreihwater dJposit at Egernsund.South Arbok 1974,49-70' landscapedevelopment in Denmark. Danmaris GeologiskeUndersdgelse, at Tornskovin south nnJ"ri.n,'S.T. (19b3).Pollen analysisof the Quaternaiymarine deposits Jutland. Danmarks'Geologiske Undersdgelse Raekke, 4' 26 pp' geochronologvand Arslanov,Kh.A., Lavrov,A.3. andNikifor6va, L.D. (1981).On thestratigraphy, of the RussianPlain' 1n: .firnut. changesof the middle and late Pleistocenein the northeast EuropeanPlain. Velichko. A.A. and Faustoua.M.A. (eds),Pleistocene Glaciations of the Eastern pp. 37-52,Nauka Press, Moscow. Annual Reviewof Earth and Bada, J.L. (1985).Amino acid racemizationdating of fossilbones. PlanetarySciences' 13. 214-268. (1974). of Qlaternary Bahnson,--O"poii,r H., Petersen,K.S., Konradi,P. and Knudsen,K.L. Stratigraphy in the SkaerumhedeII boring: Lithology, molluscs and foraminifera. Dannnrks Giologiske Undersdgelse,Arbog l!7?,27 6-?' developmentof the Behre.- K.E., Menke,'S. anO Str?ii, t{. (1979).The Quatemary.geological Symposia Universirutes G.i.nun part of the North Sea. Acta UniversitatesUppsaliensis, (lp ps alie nsisA nnum QuinqentesimumC-e-lebranns, 2, 86- 113' Uppsala' og struktureri Vestmons Berthelsen.R., KonradilP. undP.t..t.tt, K'S' (1977)'Kvartae,re lagfolger klinter. DanskGeologisk Forening, Arskrift for 1976'93-99' acidstratigraphy for late Boulton,-_Quut..nu'yeventsinSpitsbergenandtheArctic,Nature,29S'437-44|.G.S. andorher; (1982).A {lacio-isostaticfacies model and amino 221 pp' Bo*in, D. a. ( 1978 ). Qu it er n ari Ge o lo.gv.' Pergamon P:::t',9lfgld"' Bowen,D.Q. (1979).Quaternarv correlations' Nalure,277, ttL-Ltz' juarcrnarv Press'oxford' go*.n, o.O. ilsssi. Geology'(2nd edn), Pergamon andisoleucine epimerization giigh^;. l.K. irSS:). intrashellvariatiois in aminoacid concentrations iatios in f.ossilHiatella arctica.Geology., ll, 509-513' untersuchungenim Brunnacker,K. andno.nigL. w. (1976)."ijberden Stand der paliiomagnetischen und Gegen.wart,27l-17 ' Plioziin und Pleistozdn?.i gunA.trepublik Deutschland.Eiszeitalter ' of the ouaternary Brunnacker,- K., Loscher,M.. Tillmans,w. and Urban, B. (1982).Correlation of Central Europe' i"..u." systemin the lower Rhine Valley and northernAlpine Foothills Research,18, 152-173. Quaternary Matuyama/Brunhes-Grenze Brunnacker.K.. Boenigi. w.. tcoci,A. andTillmanns,w' (1976).Die Abhandlungen' 157' am Rhein und an ?", Donau. NeuesJahrbuch Geologie Paleontologie 1{r'1sR-?7R probleme im Nordtleil der DDR' C.;k,-;.G."irsozi.- -'Ber'ichtigungen Stand und der Quartiirstratigraphil der DiutsrhenGesellschaft Geologischen Wissenschaften A Geologie Paldontologie. 12,375-404. the Holsteinianand Saalian C.;;k, e.c- ano Erd, K. (1982).Classification and stratigraphyof D.J., Havlicek. comDlexln ttreeuaternary of tt. GermanDemocratic Republic. 1n: Easterbrook, Aminostratigraphy of European Marine InterglacialDeposits 275 P., Jdger,K.-D. and Shotton,F.W. (eds),IGCP ProjectT3-l-21,Quaternarv Glaciations in the Northern Hemisphere,Report 7, Prague,pp. 50-56. Dechend.W. (1958).Marines und brackischesEem in Raum der Eemsmundung.Geologische Jahrbuch,76,175-189. Deny,S.L.. Lebbe. L., Sliggers,B.C., Spaink.G., vanStlijdonck, M. andVerbruggen. C. (1983). Litho- and biostratigraphicalstudy of Quaternarydeep marine deposits of the uesternBelgium coastalplain. Bulletin van de BelgischeVereniging voor Geologie,92. 125-154. Duphorn,K., Grube,F., Meyer,D.-D., Streif, H. andVinken, R. (1973)Pleistocene and Holocene. in: Schonhals,E. and Huckreide,R. (eds),State of Researchon the Quaternarvof the Federal Republicof Germany.Eiszeitalter und Gegenwart,23124'222-250. Diicker,A. (1969).Der Ablauf der Holstein-Warmzeitin Westholstein.Eiszeitaker und Gegenwart, 20, 46-57. Ehlers,J. (1983).The glacialhistory of north-westGermany. /n: Ehlers,J. (ed.),Glacial Deposits of North-WestEurope, pp.229-238, A.A. Balkema.Rotterdam. Ehlers.J., Meyer,K.-D. and Stephan,H.-J. (1984).The pre-Weichselianglaciations of North-West Europe.Quaternary Science Reviews,3, 1-40. Feyling-Hanssen,R.W. andKnudsen, K.-L. (1980).Foraminiferer og deresbetydning i skandinavisk kvartaergeologi. Dansk Natur - Dansk Skole, Arsskrift 1979, 3-47- Frenzel,B. (1973).Some remarks on the Pleistocenevegetation. Eiszeitalter und Gegenv'art.23124, 281-292. Funder,S. (1984).Chronology of thelast interglacial/glacial cycle in Greenland:First approximation. 1n: Mahaney,W.C. (ed.) Correlationof QuaternaryChronologies. pp.261-279. Geo Books, Norwich.UK. Geiger.R. (1965).The Climate Near the Ground. Harvard University Press. Cambridge. MA. -598pp. Grube, F. (1959).Die Bedeutungdes Holstein-Interglazial-Aufschlussesvon Hamburg-Hummels- biittel Fiir die GeologieNord-deutschlands. Jahrbuch Alsterverelns, 38. 1-7. Grube, F. (1963). Geologie der Ziegelei-Tongrubenvon Hamburg-Hummelsbiittel.lahrbuch Alstervereins,42, I-8. Gudina,V. (1981).Correlation of marinePleistocene deposits of northUSSR, west Europe and north Alaska using foraminifera.In: Sibrava,V. and Shotton.F.W. (eds). IGCP Project73-1-21. QuaternaryGlaciations in the NorthernHemisphere, Report 7' pp 87-90. Gudina,V.I. andYevserov, V.Ya. (1973).Stratigraph;" and foraminifera of LatePieistocene of Kola Peninsula(in Russian).Transaction of the Instituteof Geology'and Geophysics.175. Novosibirsk. Hansen,S. (1965).The Quaternaryof Denmark.1n: Rankama, K. (ed.). TheGeologic Systems: The Quaternary,Vol. 1. pp. 1-90, IntersciencePublishers, Wiley and Sons. Hearty. P.J., Miller. G.H., Stearns,C.E. and Szabo,B.J. (in press).Aminostratigraphv of Quaternarvshorelines in the MediterraneanBasin. GeologicalSocierl- of America.Bulletin. Heide,S. van der (1957).Correlation of marinehorizons in the middleand upperPleistocene of the Netherlands.Geol. Mijnbouw (nieuweserie), 7. 272-276. Hinsch.W. (1985).Die Molluskenfaunades Eem-Interglazials von Offenbiittel-Schnittlohe(Nord- Ostee-Kanal,Westholstein). Geologische Jahrbuch A' 85' (in press). Hirvas,H., Korpela.K. andKujansuu, R. (1981).Weichselian in Finlandbefore 15.000 BP. Boreas. r0.423-43t. Hiitt. G.. Punning,J.M. and Mangerud.J. (1983).Thermoluminescence dating of the Eemian/Early Weichseliansequence at Fjosanger,western Norway. Boreas. 12,27'7-231. Hiitt er ai. (1985). Imbrie,J., Hays,J.D.. Martinson, D.G., Mclntyre.A.. Mix. A.C.. Morley,J.J.. Pisias. N.G.. Prell. W.L. and Shackleton,N.J. (1984).The orbital theoryof Pleistoceneclimate: Support from a revisedchronology of the marine6180 record. .ln: Berger,A.L. et a/. (eds),Milankovitch and Climate.Part 1.269-305,D. Reidel,Holland. Velichko, (ed.), Isayeva,-LareL.L. (1984).Late Pleistocene glaciation of north-centralSiberia. 1n: A.A. Quaternarl'Environments of the SovietUnion, pp.21-30, Universityof lvlinnesotaPress. Minneapolis. Jensen.J.B. andKnudsen, K.L. (i984).Kvartaerstratigrafiske undersogelser ved Glldendal og Kaas Hoved i det vestligeLimfjordsomraade. Dansk GeologiskForening, Arsskrift 1983,35-54. Jessen.A.. Milthers.V., Nordmann.V., Hartz,N. andHesselboe. A. (1910).En Boringgennem de KvartaereLag ved Skaerumhede.Danmarks Geologiske Undersdgelse Raekke 11.25.175 pp. 276 Gifford H. Miller and Jan Manserud Jessen,K. and Milthers,V. (1928).Stratigraphical and palaeontologicalstudies of interglacialfresh- waterdeposits in Jutlandand northwestGermany. Danmarks Ge-otogiske UndersdgelFe iii**, tt. 48, 379pp. Joh-nstrup' (1882). F. Nogle lagtagelserover Glacialphoenomenerneog Cyprina- Leret i Danmark. . Univ. Program.Kongens, Fodselsdag 1881, Kobenhavn,T4 pp. de Jong,J.D. and Maarleve-ld.G-C. (1983).The glacialhistory'of the Netherlands.1n: Ehlers,J. ,- (.0 ), GlacialDeposits of North-westEurope, pp.:s:-:so, A.A. B"lk..a, Rotterdam. Kellogg, T'8., Duplessy,J.C. and Shackleton,Ni (1973).Planktonic foraminiferal and oxygen isotope-stratigraphy -- and pale.oclimatologyof Norwegian Sea deep-sea cores. Boreas,7,6l-73. Knudsen,K.L. (1977). Foraminiferalfaunal of the eua-ternaryHostrup Clay from northernJuttanO, Denmark. Boreas6. 229-245. Knu-dsen, (1979) K L. Foraminiferalfaunas in marineHolsteinian Interglacial deposits of Hamburg- Hummelsbiittel.Mitt GeologischePaliiontologische Institutionen Unii. Hambirg, qc, tii-1,q. Knudsen, {,1- 1nd Lykke--Andersen,A.-L. (1982)Foraminifera in Late Saalian,f?1nlun, Eurfv unO Middle Weichselian of the SkaerumhedeI boring. Bulletin Geological Societyoi"iiri, lo. 97-109. Konradi, (1976)' P.B. Foraminiferain Eemiandeposits at Stensigmose,southern Jutland. Danmarks __ GeologiskeUnderscjkelse Raekke II,105,57 pp. Krog, H. (1979) Late Pleistoceneand Holocine shorelinesin westernDenmark. In: Oele. E. schrittenhelm,TT F and wigger, A.J., (eds), The euarernaryHistory of the North sea.Acta Universitates_Uppsaliensis, Symposia IJniversitatesUppsaliensis'Annum' elinquentesimu; Cek- brantis.2.75-83. 5t\tu'G.J. (1977).Pleistocene land-sea correlations I. Europe.Earth SciencesReview,13,307-374. Linke, G. (1970).Uber die _g-eologischenVerhdltnisse im GebietNeuwerk/Scharhorn. Hamburger Kiistenforschung,17, 17-58, Hamburg. Linke.G', Katzenberger. (t9j5). 9 3ndGrtin, n. Correlationof theHolsteinian Interglaciation with the deepse.a stratigraphy by meansof ESR-dating.Abstracts of the INeUA Subc"ommissionon _ EuropeanQuaternary Stratigraphy, ETH Zurich. Lykke-Andersen' raC^^datering A.-L. (1981).En. n1 fra Aeldre YoldiaLer i HirtshalsKyst-Klint. Dansk GeologiskForening, Arsskrift 19g0,1-5. Lykke-Andersen, raC A.-L. (1982)."Noglenye dateringerfra Aeldre Yold.iaLer i HirshalsKt,stklint DanskGeologisk Forening, Arsskrift'l9gl, 119-12i. Madsen,V., Nordmann, V. andHartz, N. (1908).Eem-Zonerne. Studier over Gyprinaleret og andre Eem-Aflejringer i Danmark, Nord Tyskland og Holland. Rdsume en francais. Dattnrarks - GeologbkeUndersokelse, Raekke II, 17, 302 pp. Mal interglaciallokal i SV Sverige. Meddelandenfrdn Stockholms Universitet Geologiska Institutionen, 255. t71. Ruddiman, W.R. and Mclntyre. A. (1982). Severity and speed of Northern Hemisphere glaciation pulses: The limiting case'l Geological Society of America Bulletin, 12. 1273-1279. Rutter, N. and Crarvford. R. (1984).Utilizing wood in amino acid dating. /n: W.C. Mahanney (Ed.). Quaternarv Dating .l4ethods,pp. 195-209.Elsevier, Amsterdam. Salvigsen, O. and Nydal, R. (1981). The Weichselian glaciation of Svalbard before 15,000 BP. Boreas. 10. 433--116. Sarntheim,M., Stremme,H.E. and Mangini.A. (1984).The HolsteinInterglacial: Timestratigraphic position at paratype localitiesand correlationto stable-isotopestratigraphv of deep-seasediments. 1n: Aarseth, I. and Sejrup, H.P. (eds), Quaternary Stratigraphy of the North.Sed. Abstract Volume, Symposium Univ. of Bergen. Dec. 1984, Bergen, p. 54. Shackleton.N.J. (1967). Oxvgen isotope analysesand Pleistocenetemperatures reassessed. Nature, 2r5. 15-t-7. Shackleton, N.J. and Updyke, N.D. (1973). Oxygen isotope and palaeomagneticstratigraphyof equatorial Pacificcore V28-238:oxygen isotope temperatures and ice volume on a l0'year and10n yearscale. Quaternary Research,3. 39-55. Sjorring,S. (1983).Ristinge Klint. .fn: Ehlers, J. (ed.),Glacial Deposits in North-West Europe, pp 219-228.A.A. Balkema.Rotterdam. Somm6.J., Paepe.R., Baeteman.C., Beyens,L., Cunat,N., Geeraerts,R.. Hardv,A.F., Hus,J., Juvignd.E.. Mathieu, L., Thorez,J. and Vanhoorne,R. (1978).La formationd'Herzerele: Un nouveaustratotvpe du Pleistocenemoven marin de la Mer du Nord. Bulletin de I'Associauon frangaisepour l'€tudedu Quaternaire,pp. 54-56, Paris. Stremme,H.E. (1983).Pedostratigraphy of glaciatedand non-glaciatedareas in westernEurope. 1n: Billard,A., Conchon,O. andShotton. F.E. (eds),IGCP Project73-l-24 Quaternary Glaciations in the Northern Hemisphere,Report 9, Paris.pp. 140-147. Stremme,H.E. (1985).Alterbestimmung und PalAobcidenin der Qudrtarstratigraphie.Bulletin de I'Associationfrangaise pour l'etudedu Quaternaire,22-33, 159-166. Urban, B. (1982).Quaternary paleobotany of the Lower Rhine Basin.Zeitschrift Geomorphologie N.F..42.201-213. ter Wee, M.W. (1983).The ElsterianGlaciation in the Netherlands.1n: Ehlers,J. (ed.), Glacial Depositsin Norrh-WestEurope, pp.413-,116, A.A. Balkema,Rotterdam. Wehmiller,J.F. (i983). A review of amino acid racemizationstudies in Quaternarymolluscs: Stratigraphicand chronologicapplications in coastaland interglacialsites. Pacific and Atlantic coasts,United States.United Kingdom.Baffin Islandand tropicalislands. Quaternary Science Rev,iews,l,83-120. Wehmiller,J.F. (198"1).Interlaboratory comparison of amino acid enantiomericratios in fossil Pleistocenemollusks. QuaternaryResearch, 22, 109-120. Woillard. G. (1975).Grand Pile peat bog: a continuouspollen recordfor the last 140,000vears. Quaternarl'Research. 9, 1-21. Yevzerov,V.Ya. and Koschechkin,B.I. (1977).Kola Peninsulatill stratigraphy.In: Prospecritrgtrr Areas of GlaciatedTerrain, pp. 30-33. London. Yevzerov.V.Ya. and Koschechkin,B.I. (1981).The Quaternarystratigraphv and chronology of the westernPart of the Kola Peninsula./n: Gorbunov.G.H. (ed.), GlacialDeposits and Glacial History in EasternFennoskandia. pp. 16-24. Apatity, Kola Branchof the Academvof Sciences. USSR. Zagwijn.W.H. (196i). Vegetation,climate and radiocarbondatings in the late Pleistoceneof the Netherlands.L Eemian and Earlv Weichselian.Medelingen van de GeologischeStichting, 14, Zagwijn,W.H. (1975).Variations in climateas shownby pollenanalysis,especially in the Lower Pleistoceneof Europe.1n: Wright, A.E. andMoseley, F. (eds),IceAges: Ancient and Modern, pp. 137-152.Seel House Press. Liverpool. Zagwijn.W.H. (1983).Sea-level changes in the Netherlands during the Eemian. Geologie en Miinbouw.62.137-150. Zagwijn. W.H. and Doppert. J.W.C. (1978).Upper Cenozoicof the sourhernNorth Sea basin: paleoclimaticand paleogeographicevolution. Geologie en MijnbouwN.s., s7, .177-58g.