LATE WEICHSELIAN MARINE SEDIMENTS CONTAINING SHELLS,FORAMINIFERA, AND PoLLEN, AT ACOTNES, WESTERN

JAN MANGERUD

Mangerud, J.: Late Weichselian marine sediments containing shells, foramitli- fera, and pollen, at Agotnes, . Norsk Geologisk Tidsskrift, Vol. 57, pp.23-54. Oslo 1977.

Marine sediments, mainly clay' and silt, deposited at a water depth of 20-30 + m are describeil. The base of the succession is radiocarbon dated to 12.220 150 B.P., the upper part to 10,230 + 180. The shells indicate water tempera- tures similar to those of northern Norrvay today. These shells and shells from other localities suggest that warm Atlantic water entered the Norrvegian Sea prior to 12,600 B.P. The pollen diagram from the sediments has the same main trends as in cliagrams from limnic sediments covering the same period, except for high percentagesof Alnus. The age and correlation of several Late Weichselian events are discttsscd.

J. Mangerud, Geologisk instittttt, Avd.B, Olal Ryesvei 19' N-5014 Bergett- U niv ersitetet, N orw aY.

Large areasof were deglaciatedduring the A11er0dChronozone (Figs. I and l4) (Mangerud 1910,1912b, Aarseth & Mangerud 1974).LaIer' during the Younger Dryas Chronozone, the ice front re-advancedby at least 40 km. The maximum extent of the inland ice during this re-advance is marked by the Herdla moraines(Aarseth & Mangerud 1974). The Agotnes locality discussedin the present paper lies 2 3 km outside the Herdla moraines (Fig. 1), and includes sedimentsof both Allerpd and Younger Dryas age. The stratigraphicalterminology used here is in accordancewith the pro- posals of Mangerud et a]l. (1974). The boundaries of the chronozones (Younger Dryas, Allerod, etc.) are defined in conventional radiocarbon years.

The Agotneslocality The coast of Hordaland, including Sotra, is characterizedby bare bedrock. Loose, Quaternary sedimentsare mainly found in topographicaldepressions. Investigation of these sedimentsis usually possibleonly by coring. In the investigatedlocality, an excavatedsection was accessiblefor just one day in the summer of 1970.Field work was, however, limited by dangerousslides in the clayey sediments. The locality is situated on the landward side of the island Sotra (Figs. I and 2). close to the shore of Hjeltefjorden, however, only 6 km from the open North Sea. J. MANGERUD

--:NbBTH= :SEA

LOnd

Herdlo endmoroines {Younger Dryos)

lTTl ,ornn.,Dryos isoboses

non ot tce_itow

, 10km

Fig. 1. Location of the Agotnes locality in Hordaland. younger Dryas isobasesare ex- trapolated from Aarseth & Mangerud (1974, fig. ll).

It lies on a small peninsula (Fig. 2), with a maximum height of 29 m a.s.l.From the highestpoint the bedrock,with small patchesof till and lit- toral sand, slopesgently eastwardtowards the investigatedlocality (Fig. 3) at ca.12 m a.s.l. The bedrock is gneiss,forming smal escarpmentswith crestsparallel to the NNw strike, and dip slopeinclined eastwards.The locality lies at the foot of one of theseescarpments (Figs. 3 and 4). Details of the locality are shown at Figs. 4 and 5. The sedimentswere best exposedin the southernwall of the excavation,and the sequencein this wall is shown at Fie. 6.

Grain-size analysis Particlesabove 16 mm were excludedfrom the analysis,those coarser than 0.063 mm were sieved,with a sieveinterval of one phi unit. The finest (< 0.063mm) were analysedby the pipettemethod, with readingsfor each phi unit. In all samplesorganic material was removedwith HoO" before analvsis. NORWAY L) LATE WEICHSELIAN SEDIMENTS IN WESTERN

! locaritv [.--q,r.;....',o.T.runouoou" .o.u ",. [I.Tl Lando-3o ma s.l. = "uu rhe Asotnes

Fig.2.MapoftheareaaroundAgotnes.DuringtheLateWeichselianthesealevelwas and the map therefore gives an im- "p-pt"*mrLfv 30 m above that oithe present day' pressionof the contemporarypaleogeography'

Fig.3.PhotographyoftheAgotneslocalityafterconstructionworkhadremovedmost The photograph is of the sediments described. ihe bedrock wall is thus man-made. northern edge of the sediments iut"r, to*u.d, the summit of the peninsula (NW). The is marked with stippled line along the base' J. MANGERUD

Bedrock

l-l un"on.ol,,lut"o L*-J sedrments

7 r^"uuuuon

PI]OFiLE ]

the geologicalpro- Fig. 4. To the left sketchmap of the Agotneslocality. To the right 6' fills. The'describedsection' indicated on themap is shownin Fig'

(1968' fig' 1) In Norway a modified Wentworth scaleproposed by Doeglas is now used,and this scaleis applied here (Fig' 7)' McCammon(|962)andFolk(|966)discussedtheefficiencyofthedif- ferentparametersusedforgrain-sizedistribution.However,inroutine obtain the neces- analysisof fine-grained sediments,one does not always the present study, I sary percentilesfor the most efficient parameters.In (1957)for the mean grain have ttreretoreused the measureof Folk & Ward sizs: M2:(iDl6+o50+084)/3, and that of Inman (1952) for sorting (standard deviation) 5: (A84-Alol2. to be extrapolated' Even so, to obtain the value for iD16,some curves had Selmer-olsen(1954) analyseda large number of Norwegian Quaternary in a useful Md - sedimentsof differeni geneses,and presentedthe results Sodiagram.InordertocomparethesedimentsfromAgotneswithhisre- (Fig. assuming that sults, I have constructed a similar diagram 9), Mz: Md and recalculating the sorting coefficient Q75 so : log," Q25 LATE WEICHSELIAN SEDIMENTS IN WESTERN NORWAY 27

Fig. 5. Photograph of the excavation taken towards the sE. The main section shown in Fig. 6 is in the shadow, with the bedrock wall to the left.

of Selmer-olsen(1954) to be relatedto the standarddeviation expressed in phi units (S): S : 2.465o

Co.12mo.s I . co.5m LITHOST RAT I GRA PHY XX Chrono- Beds Inlormal stratrgraphy 0T Bed /\ names rock X Boreal XX Peat

x/ Sand upper Agotnes )- XL Preboreal Gravel ( 10cm) sano 6 '4,i* Agotnes Greenish YoungerDrVas Xkq.,."t grey srrr srlt \ a-rzo 6 ..::. ::x25. :1.21. ;.:':::: i)3 ["*"rr,,," Agotnes Allet@d ,22 cray clay 5 I L sano I ano Agolnes with Older Drvas lgravel ano boulderl sano X g ravel lsome

i. /2) . : lv_l Marlne sheils/ . .rn tenses Radiocarbon datesr sample 21t 10 230 ! 1go (T_1s74) X Sample 20: 12 22O ! 1SOfi_1oal Fig. 6. The described section. The position for all samplesdiscussed is given in numbers. J. MANGERUD

CLAY SILT SAND GRAVEL

- I

./ 95

l-_1-..- 1: /"1 I ,/i -t- I 2a 2a iI 20 14 IJ if ll ilT I I

109 876543210-t-2 -3 phi tAtz. t/ta .t/4 vtz l/e vz 1248mm t248163163 rooo x, micron Fig. 7. Grain-size distribution. Cumulative curves on probability-paper. For locations of samples, see Fig. 6. Sample 27: Sand and gravel in the bottom of the excavation. Sample 16: The lense of sand and gravel between Agotnes clay and Agotnes silt. Sample 10:

The sedimentsequence The sequenceconsist of 5 main units (Fig. 6). At the baseis a bed of sand and gravel at least 2 m thick, followed by a ca. I m thick bed of brown gyttja clay, with a gradual upward transition to a bed of greenishgrey silt. Marine shellsare frequent in both the clay and silt beds. Above the silt is a I m thick bed of well-sortedsand (sample 10, Fig. 7) with a thin underlying bed of gravel. This sand is a littoral sediment,depos- ited during the Holocene uplift. The sequenceis cappedby a I m thick bed of peat of Holocene age. NORWAY LATE WEICHSELIAN SEDIMENTS IN WESTERN

I

1_--

I

-3 phi 7654321 Vzz t/to \/a 1/a Vz 4 8mm micron 8 16 31 63 * gyttja clay (Agotnes cla,v'').Sample 13 Upper Agotnes sand. Sample 23 & 25: Brownish & 14: Greenish grey silt (Agotnes silt)'

Allthebedsarephysicallyrestrictedtothesmalldepression'andlwill thereforenotdefineanyformallithostratigraphicunits.Forthecorrelation to each bed with other sequences,the name Agotnes is, however, attached (Fig. 6) in an informal nomenclature(Hedberg 1970:16)'

gravel) The sand and Sravel (lower Agotnes sand ond sedimentscon- This bed was poorly exposed,partly becausethe overlying of the excavationwas stantly slumpeddown, and partly becausethe bottom gravelwere seenin direct filled with water. In the northern part the sandand 30 J. MANGERUD contact with the beclrock, and I assume this to be the case also in the the deeper parts of the pit, though possiblywith a veneer of till between bedrock and this bed. It consistsmainly of very poorly sorted sand and gravel (sample 27, Ftg. 7). However, some boulders were also found throughout the bed, and near the beclrockwall (to the east in Fig' 6) the a bouldersdominate. All the pebblesare angular or sub-angular,suggesting short clistanceof water transport. An encrustationof calcareousalgae was observedon some Pebbles. I assumethat this beclwas depositedmore or lessdirectly from the glacier front, followecl by some rolling and sliding on the sea floor' From other field observationsin Hordaland the water depth during the deposition can be estimatedto 20 -30m.

The brown gytt ja clot- (Agotnes clay) This was the most striking sediment of this sequenceas marine sediments from the cleglaciationperiod very seldomly have brownish colours. Shells occur frequently throughout the bed, giving clear evidence of its marine origin. The shellsoccurred partly in small lenses,and partly as isolatedin- dividuals,many of them in growth positions.The variable thicknessof the bed, from 30 cm up to 100 cm, and the variable strike and dip of some internal bedding planesare to be expectedin such a small sedimentaryba- sin. However, due to the slumping, it was not possibleto map the internal structures.

I rds I .' '1 ro ?o lo 1o 5c ba i.) "u to 8 6 4 2

T t0

I l2 ''l

ll

1t

2^

23

22

27

Fig. 8. Parametersof grain-sizedistributions and losses on ignition for some of samplesarranged in stratigraphicalorder (Fig. 6). The same samPles rvere used in polien analysis(Fig. 12). LATE WEICHSELIAN SEDIMENTS IN WESTERN NORWAY 31

The brownish colour is most probably causedby the high content of or- ganic material. The loss on ignition (Fig. 8) is generally higher in the brown Agotnes clay than in the greenish grey silt above. From the well- defined difference in colour, one should have expectedeven more consis- tent differencesin organic content betweensamples from the two beds. The sedimentsin the Agotnes clay are very poorly sorted (Figs. 7 and 8), being mixtures of gravel, sand, silt and clay. Even some cobblesand boul- ders were observed.Thus the term Agotnes clay is used for the bed as a whole and becauseof the high clay content. In the middle of the section(Fig. 6) a lens of sand and gravel (sample 16, Fig. 7) occurs betweenthe Agotnes clay and the Agotnes silt, indicating an unstable sedimentaryenvironment. Most probably this is a slump deposit from shallowerwater sediments.Below this lens the Agotnes clay is thinner, partly due to compaction,but probably mainly as a result of erosion.

The greenish grey silt (Agotnes silt) The boundary between the Agotnes clay and the Agotnes silt is a smooth transition, showing no hiatus, and indicating that the depositional en- vironment changedgradually. The Agotnes silt is 50-120 cm thick, and has a very sharp boundary with the overlying gravel. Marine shellsoccur also in this bed, but lessfrequently than in the Agot- nes clay. The sedimentof the Agotnes silt is poorly sorted (Figs. 7 and 8), although the sorting is better than in the clay below (Figs. 8 and 9). Some cobbles and boulderswere observed.

The genesisol the Agotnes clay and silt beds As stated above, there can be no doubt that both beds are marine. The sedimentsare very poorly sorted, compared with normal marine sediments (Fig. 9). Some of the sampleshad a considerablecontent of calcareousma- rine organisms,and to find the influence of theseon the grain-sizedistribu- tion, the CaCO, in 4 sampleswas removed with 10 % HCl. It can be as- sumed that this treatment did not influence the mineral particles signifi- cantly, as limestonesdo not occur in the bedrock. The CaCO, was mainly found in the clay fraction, the result of treatment being that the mean grain size increasedsomewhat, and the sorting became considerablybetter (Fig. 9). One must, however, realizethat some of the CaCO, removed was bio- genetic material depositedas sedimentaryparticles, and that the grain size distribution of the physically depositedsediment, probably lay somewhere between the curves of the HCl-treated and untreated samples.(Sylvi Hal- dorsen (pers. com.) has later found that severalclay minerals were also dis- solvedin the HCl-treatment, and my conclusionson the CaCOacontent are therefore erroneous.) -)L J, MANCERUD

Samples lfom lne brown gyltra.r i!

Samples lrom lhe greenrsh grey s!ll

The same samPle:; when CaCo3 s

\i\. \ 6 \- \-,ur .-\a' - .26 "crrli

Meangratn size (Mz) from the Agotnes clay and Fig. 9. Mean grain size verslls sorting (= standard deviation) (1954) for some sediment types are Aiot.,c' silt. ihe boundaries found by' Selmer-Olsen recalculated according to formulas indicated. His values for sorting and median are given on p. 26.

Thesedimentsstudiedweredepositedinadepressionwithadiameterof peninsula situated approxi- approximately 50 m. From the summit of the depression(Fig' 3)' mately 200 m away, there is a gentleslope down into the WithhigherSealevelprevailingatthetimeofdeposition,thisdepression rolling, or slumping must have acted as a sedimenttrap, retaining particles downtheseafloor,inadditiontothedepositionoffineparticlesfrom of these sub- suspension.This probably explains the grain-sizedistribution deposition of clay, littoral sediments:The water was calm enough for the but neverthelessthere was a supply of coarserparticles' TherelativeSealevelmayhavefluctuatedSomemetresduringthesedi. mentationperiod.Althoughthiscannotbesubstantiated,itisclearthatin 20-30 m, sedimentation this sublittoral environment, with water depths of vertical variation in was very sensitiveto variation in the sealevel, and the The infilling of the the grain sizesmay partly be due to sea-levelvariations. content of clay may depressionitself is also of importance. The decreasing I also presume simply be a result of the decreasingdepth of the depression' infilling. and later ,t-,ui it " top of the Agotnes silt is determined by this wave erosion. Theinfluenceoftheglacieronthesedimentaryenvironmentisverydif- curves of grain-size ficult to interpret. Nevertheless,the parallelismfor the (e'g' Fig' 12) is striking clistribution(Fig. 8 to the left) and pollen Betula' of mineral particles' anclindicates a climatic (glacial)influence on the supply LATE WEICHSE,LIAN SEDIMENTS IN WESTERN NORWAY

The coloursalso indirectly depentl on climate.as the brown Agotnesclay is from the Allerod Chronozone,while the greenishgrey Agotnes silt is from the Younger Dryas, with advancingglaciers tFig' I a)' As mentioned,the lower Agotnes sandand gravelwere probably deposited closeto the ice front. The sharp lithologicalboundary to the Agotnesclay indicatesthat a hiatus exists,and the clay was probably depositedsome tens of kilometres distant from the glacier front (see foraminifera p. 38). The colour reflects high organic production during the Allerld, and the upward change in colour the lower productivity during the climatic deterioration. During the Younger Dryas Chronozone,the ice advancedtowards the local- ity, and stoppedas a calving glacier in the fjord, only 2-3 km to the east' The sedimentshave not been disturbedby ice, and they have been used to prove that the ice advancedid not reach this island (Aarseth & Mangerud lg74: 16).The increasinggrain size upwards in the Agotnes silt may partly be due to the approaching ice, through sediment supply from icebergs' Possibly,however, the secliments

Radiocarbondates Two radiocarbon dates have been obtained, both from marine shells.The datingswere carried out at The University of Trondheim, RadiologicalDat- ing Laboratory. The calculation methocl useil includes a correction of 410 years for the apparent age of sea water (Mangerud 1912b: 146)' However' the apparent age of marine shells from the coast of Norway is 440 years (Mangerud & Gulliksen 1975),and the dates are therefore corrected for the aclclitional30 years.This additional correction could be ignored for the geologicalproblems, but is significant for dateshaving a standarddeviation of only 100-150years. Other shell datesreferred in this paper are corrected in a similar way, and should therefore be directly comparablewith dates of terrestrial material. Also whale-bone dates are corrected for an apparent age of 440 years.The Libby half life of 5570years is usedfor all dates. Both samplesof shellswere rinsed in destilledwater, and the outer l0- 20 /c removed in diluted HCl. The first sample(sample 20, Fig. 6) was from a shell lensesituated at the baseof the Agotnes clay, closeto the bedrock wall, containing Balanus sp. and Mya truncatL. Measurementsof the stablecarbon isotopesgave btsc: +1.5%0rel.PDB.Thedatingresultwas(T-1023):|2,220-+150years B.P. The secondsample (sample 21 ,Fig.6) was from a shell lens approximate- ly in the middle of the Agotnes silt. The dated shellswere mainly of Balanus balanus,with a few piecesof Mya truncata, Hiatella arctica, and Astarte : elliptica. The correspondingresults of this sample were d"rC 0'2/1,,,rel' PDB, the radiocarbonage being (T - 1514) 10,230t 180 years B'P' The radiocarbon dates provide the most important information on the 34 J. MANGERUD

(Fig' 6)' age of the secliments,and the resultsare indicated on the section (Fig' tie pollen diagram (Fig. l2), and the correlation chart 14)'

Marine shellsand palaeo-positionsof the Polar Front the main aids At the beginning of this century, marine shells were one of

20 and 2l (Fig' 6)' Table 1. Identified marine shells(molluscs and balanides)in samples broken, and the frag- Numbers are only approximate, as many of the bivalves were ments are counted as individuals.

Number of in the shells Recent distribution and references Species samples 20 2l

Gastropods (Fretter & Gra- Puncturella noachina Cold waters on both hemispheres (Linn6) ham 1962).Svalbard-Greenland to the British Isles(Sars 1878). to the British Lepeta caeca t4 N. America, Svalbard-Greenland (Miiller) Isles (Sars 1878,Fretter & Graham 1962)' Sea Gibbula cl. cineraria 21 Northern Norway to the Mediterranean (Linn6) (Sars1878 Nordsieck 1968). Norway (Sars Moelleria costulata Greenland-NE America, Northern (Nordsieck1968)' (Mciller) 1878),to the Bay of Biscay and the ex- Alvania miglrclsi N. America, Greenland,Svalbard, (Stimpson) treme NE NorwaY (War6n 1974). the extreme Alvania scrobiculata Greenland,Iceland, Svalbard,and (M,ijller) NE Norway (War6n 1974). MediterraneanSea OmalogYra atomus Northern Greenland to the (Philippi) (Fretter & Graham 1962). Biscay Trophon truncatus Svalbard-Greenlandto the Bay of (St16m) (Feyling-Hanssen1955, Nordsieck 1968). Troplrcn sp. 1 Lora sp. 2 Arctic waters(Sars 1878). Bivalves Bay of Modiolus modiolus Circumpolar. The White Sea to the (Linn6) Biscay (Tebble 1966,Nordsieck 1969)' A few occur- Chlamys islandica Greenland-Svalbardto Lofoten. (Miiller) rencesin western Norway (Wiborg 1963). British Astarte elliptica Greenland-White Sea to the northern (Brown) Isles and Massachusetts(Ockelmann 1958' Tebble 1966). (Jensen& Sptirch Thyasira sarsi Novaya Zemlya to Oslofjorden (Philippi) 1934,Feyling-Hanssen 1955)' Sea Macoma calcarea Arctic circumpolar. Greenland to the North (Chemitz) (Ockelmann1958). to Mya truncata t0 N. America, North Greenland and Svalbard (Linn6) the Bay of BiscaY(Strauch 1972). Hiatella arctica World-wide (Jensen& Spiirck 1934' (Linn6) Strauch1968).

Sum 68 36 Lusitanian region Balanides Innumerable Circumarctic, extending to the Balanus balanus fragments (Feyling-Hanssen1955). (Linn6) LATE WEICHSELIANSEDIMENTS IN WESTERNNORWAY 35 for Quaternary stratigraphic studies in Norway (e.g' Brggger 1900-1901' C. F. Kolderup 1908). After the introcluction of the radiocarbon method, interest in shellswas renewed,mainly from the point of view of dating. It is hoped that a comprehensivestudy of severalproperties (e.g. Strauch 1968, lg72b, Andrews 1972,1973,Mangerud 1912b)of the Late weichselian and Holocene shells of Norway can be undertaken. An improved knowledge of marine shallow-waterenvironments would be of great importance in cor- relation of the terrestrialand the deepsea records. Here only radiocarbon dates and a list of determined shellsare included (Table l). Most of the shellswere identified by cand.real. Per Wikander; the specimensof Alvania were identified by Dr. Anders War6n' Shellsoccurred throughout the Agotnes clay and the Agotnes silt' How- ever, large numbers of shells were only collected from the two lensesof shells (sample 20 and 21, Fig. 6). As sampleswere collected from lenses which were mainly mechanical accumulations,they contain shells derived from different habitats.For instance, Balanus balanus,dominating in both samples,was probably partly attached to the bedrock wall, whlle Mya trun' catalived burrowed in the bottom. The faunas of both samples(Table l) indicate colder water than in the area today. Severalof the speciesnow live in more northerly waters, and near their southernlimits they only occur at great depth. As the water depth at the presentlocality did not exceed30 m, the speciesshould be compared with their present-dayshallow-water occurrence. In open water, Chlamys islandica does not occur today south of Lofoten. Further south it is only found in fjor{s with relatively cold deepwater. The two Alvania species(in sample 2l) are found along the northeastcoast of Finnmark only (Fig. 10)' and in even colder waters (Table 1). In both samples,however, some more southerly species(Thyasira sarsi, Modiolus modiolus, and Gibbula cineraria) also occur; their northern limit lies at presentbetween Norway and Svalbard.Of these three speciesThya- siro found in sample2l seemstoday to extend further into the Barents Sea than Modiolus and Gibbula found in sample 20. I conclude that sample2l (Younger Dryas) suggeststhat the water tem- perature was approximately the same as today along the northern most coast of Finnmark (Fig. l0), or even slightly colder. Using the surface tem- peratures given by Saetre(1973, table l) for Korsfjorden and Vardg, this means that the mean yearly temperaturewas 3.8o C lower than today. The difference was greatestfor the summer (July-Sept. 6.5" C) and smallestfor the winter (Jan.-May 1.6'c). The assemblageof sample20 indicatesslight- ly warmer water than for sample 21, and can be comparedwith the fauna of larger parts of the northern Norwegian coast from Troms to Finmark (Fig. l0). Of specialinterest is the occurrencein sample 20 of Modiolus modiolus, a specieswith large shellsand therefore usually included in field studiesof Quaternary geology. The northern boundary of the present-day distribution 36 J. MANGERUD

Polar Front (Fig' of Modiolus modioluslies very closeto the oceanographic Atlantic water 10), that is the physical-oceanographicboundary between andPolar(Arctic)water(e.g.Dietrich1963,fig'221)'Modiolusmodiolus northern (:volsella modiolaLinn6 in Feyling-Hanssen1955: 133)occurs in (1964)reports Norway and the southwesternpart of the Barentssea. wiborg thatitshabitatisnearBjgrn1ya,butitdoesnotreachaSfarnorthasSval- climatic op- baril, where it is a guide fossil for deposits from the Holocene timum (Feyling-Hanssen1955: 133)' to Modiolus mo- Littorina littorea (Linn6) has a very similar distribution diolus(Fey|ing-Hanssen1955:l60,Nordsieck1968:40).A|soMytilusedulis to tolerate slightly colder Linn6 has a nearly similar distribution, but seems 63, Hjort & Funder water. as it occurs in Greenland (Ockelmann 1963:61

'guide-fossils' Atlantic water south of the Table 2. The occurrence of three for the in Horclaland' Polar Front in radiocarbon dated Late Weichselian sediments 0 - Shells of the sPecies are dated x - The species occur in the samPle

'-=.: Radiocarbon --t= Lithostrati- blUN Locallt) Chronozones dates S-: %^ ',*'on^ graphy o A :.o li.: = c - c :: i.5 i .5 >i >a nJ Not found Younger DrYas Os 11,250+110' T - 1021 Trengereid 11,230+200' T- 1161II Aller@d 11,530+150, T-fi628 Eikanger- 11,900 + 140. Older Dryas vag T-846 Agotnes 12,220+150, T - 1023

BOlling Ulv/y Till BlomvAg 12,400+ 90 T - 1882 + BlomvAg 12,540 150, 0 'f - Beds 1697 12,540+ 180, 0 T- 1696 i2,670+ 350, T-139 37 LATE WEICHSELIAN SEDIMENTS IN WESTERN NORWAY

(Feyling-Hans- lg74) andis more frequent in Holocene depositsat Svalbard (Feyling-Hanssen ,"r, ilSS; where it possiblyhas reappearedin recent time & Olsson1960:125)' Thesethree species,and other low-arcticand borealspecies, can therefore 'guide-fossils' re- be used as for Atlantic water south of the Polar Front, of membering that their limiting factors are not identical to the definition the Polar Front. Probably,the most important factor for thesethree species These is temperature,while the Polar Front is usually defined by salinity. two factors are, however,inter-related in the surfacewater'

30o 20. tu- v Position of the Fig. 10, Map of the North Atlantic ocean and the Norwegian sea. of the prlent-day iolar Front according to Dietrich (1963, fig. 227). Palaeo-positions -polar (1973, Front in the North Atlantic after Ruddiman & Mclntyre tie. 6). and Troms Along the Norwegian coast the Polar Front was situated between Hordaland B.P. the Polar il;G ;h" p"riod- 12,000_11,000B.p. During the period 11,000-10,000 the Norwegian coast' Froniwas pressedsouthwards, and it is uncertain if it reached J. MANGERUD

'guide-fossils' Table 2lists the known occurrencesof the three discussed in radiocarbon-datedLate Weichselian sedimentsin Hordaland. Even at 12,600B.P. all three speciesoccurred, and one must concludethat the Polar Front was situated north of Hordaland at that time. This makes it possible to extend the results of Ruddiman & Mclntyre (1973) from the Atlantic Ocean into the Norwegian Sea(Fig. l0). Soon after (12,200-12,000B.P.), the inland-ice reached the North Sea (Fig. la) at the coast of Hordaland, but this event probably did not influ- ence the position of the Polar Front significantly. From the Allerod Chronozone,Modiolus modiolus is found at severallo- calities (Table 2), and Littorina littorea at one, again indicating that the Polar Front was situatedfurther north. The faunas of the Younger Dryas sediments are clearly colder' and none of the three mentioned speciesare found. This may indicate that the At- lantic water did not reach the coast of Hordaland. It may, however, only be a local environmental responseto the major glacial re-advancewhich took place in Hordaland during the Younger Dryas Chronozone(Fig. l4). Along the coast of Norway north of Hordaland, marine shellsoccur fre- quently in Late Weichselian lEarly Holocene sediments.However, very few of them are preciselyenough dated to be used in a discussionon the loca- tion of the Polar Front. From Troms (Fig. 10), Andersen (1968:70-71) has given an extensivelist of speciesfrom radiocarbon-datedsediments, cover- ing the period from ca. 12,500years B.P. (the Bglling Chronozone)to ca. 9500 years B.P. (the Preboreal Chronozone).The faunas of Troms are of distinctly colder type than the contemporaneousfaunas in Hordaland, and 'guide-fossils' none of the three were found there. Several of the faunas {escribed from Troms lived, however, near ice fronts, and at that time a steep ecologicalgradient obviously existed from the calving glaciersin the fjords to the open ocean. My preliminary conclusionis that the Polar Front between the Atlantic water and the cold Arctic water was situated between Hordaland and Troms during the BBlling and Allerpd Chronozones. In both areas the Younger Dryas faunas were clearly colder than the Allergd faunas, indicating that the front moved southwardsduring the Younger Dryas (cf. Ruddiman & Mclntyre 1973),and possiblythat the warm Atlantic water did not reach the coastof Norway during the Younger Dryas.

Foraminifera The foraminifera were identified by cand.real.Ivar Miljeteig. The methods and the taxonomy follow the practice of Feyling-Hanssenet al. (1971)' The foraminifera were investigated for two reasons: firstly to follow the change of environment from the brown Agotnes clay to the greenishgrey Agotnes silt; secondly,as a contribution to the correlation of terrestrial and marine sediments. NORWAY 39 LATE WEICHSELIAN SEDIMENTS IN WESTERN

Young chronozones Arerod l;;;;: I -] al- | (oa) u:s!.9P I ea[Ea;=rl 1lesl| tsf;\Y^ E -] =, ; -1-.-i'-l ;;;,,", :l

t !lll'F;,r'= t a , [-Egroertandrcumuusnman-] Nf [:*" 7 '7 l i-'*""rT#r" r llf"n l.kt'''"^"r*'" TYJl=H*:T:*' in at least one sample with Fig. 11. Composition of the benthic foraminifera occurring a frequency of more than 5 Vo' 40 J. MANGERUD

TOTAL P Calc Dwarf shnrbs

r,ees sh,ub, ?#ilr, n.,b, Other Erical€

Fig. 12. Diagram of pollen, spores, and Hystrix. The black silhouette curves are on the same scale for all pollen and spore types. Five times exaggeration is indicated by sil- houettes with horizontal lines at the sample levels;

In the Agotnes clay foraminifera were abundant; between600 and 5,000 specimensper 100 g dried sedimentwere found in all samples.Even in the lens of sand and gravel (samples16 and 17) foraminifera were frequent. However, in the.A.gotnessilt (samples1l-14 and 26) foraminiferaswere rare, and therefore one of the aims of the investigation co.uldnot be ful- filled. In a small sample(no. 21) from the shell-rich rensin the Agotnes silt, foraminifera were abundant (Fig. 11). I therefore assumethat foraminifera were also originally present in the silt, and that their absencetoday is due to post-depositionaldissolution. In the shell lens the interstitial water was more alkaline and the foraminifera were preserved. In all samples(Fig. 11) the foraminifera assemblagesare typical for shal- low-water environments.The speciesidentified are those common in Late Pleistoceneand Holocene sedimentsidentified in Norway and Denmark (Feyling-Hanssen1964, Feyling-Hanssenet al. r97r, Aarseth et al. 1975). According to the records in Feyling-Hanssen(1964) and Feyring-Hanssen et al. (I97I), Elphidium umbilicatulum (: E,excavatumin Feyling-Hanssen 1964) and Elphidium albiumbilicatum indicate that the sea water was not extremely cold. Both speciesoccur frequently in the lower two units (Fig. LATE WEICHSELIAN SEDIMENTS IN WESTERN NORWAY +l

LEN DIAGRAM on P

Rosaceae caryo- P eo r 1.5i3 -o excl. Filip.iphyllacea9 ! #3 38 5 l0 20 5l0i l0l 5 5 5

on the 11). For the Agotnes clay, this accordswith the conclusionsdrawn that the basesof the molluscs (sample 20), and also with the assumption glaciershad retreatedfar into the fjords during the Allerld. concerning the lensofsandandgravelbetweentheclayandsilt(Fig.6and11),thefora- to the minifera assemblagesindicate that it belongsto the Allerld and not YoungerDryas. Thetwospeciesmentionedabovearenotfoundinsample2lfromthe is Agotnes silt. on the other hand, the assemblagefound in this sample n*tly identical to that describedby Nagy (1965: 113) frorn Treskelbukta, possible to Spitsbergen,collected 3-6 km from calving glaciers. It is not was too diaw definite conclusionsfrom this single sample,which in addition indicates small to be statistically reliable. Nevertheless,it is striking that it the sameenvironment as indicated by the sedimentand shells.

Pollen,spores, and HYstrix pollen analysiswas mainly performed to investigatethe possibilitiesof this the method for the correlation of terrestrial and marine sequencesalong 42 J. MANGERUD

pollen anal- coast of Norway. concerning the general problems involved in (1966)' ysisof marine sediments,reference is made to Groot & Groot & The preparation technique was slightly modified from that of zagwijn Veenstra (1966: 546) and Vorren (1972: 237)' Separation in bromoform' find that diluted with alcohol to a sp. gr.2.2, was repeated2 to 4 times. we the HF separationin heavy liquids gives much cleaner preparationsthan The treatment (Fregri & Iversen 1966: 69), generally used in scandinavia. pollen counts were carried out by cand.real.E. s@nstegaardand cand.real. K. Skdr.

Diagram and zones with small The resultsare shown in a total pollen diagram (Fig' l2), which in north- modifications is commonly used for Late Weichselian sediments in the western Europe. Long-distancetransported pollen are also included sum pollen sum (Mangerud 1970: 121).The sporesare not included in the (e'g' 1973: of pollen, even though this would perhapsbe more logical Birks so low 222). ln the present diagram, the number of spores are, however, spores that the pollen curves would not have significantly changed if the had been included. pollen I have subdividedthe sedimentsbiostratigraphically into informal assemblagezones(Mangerudetal.|914).Manyofthepollencurvesshow but a marked change between sample 25 and 26. Betula starts to decrease, placed between for this speciesthe boundary could just as well have been sample26and14.BecauseofthegenerallyhighNAPcontentandtheBe. tula maximum, the lowermost zone is called the Betula-NAP assemblage and I have zone. Also the succeedingzone has a very high NAP content' the lower selectedtwo of the important specieswith a marked increaseat boundary,andtermedittheArtemisia-Lycopodiumassemblagezone. content of sample l0 is different from all the others; as it has a Betula zone' 62.5 % I have simply named it the Betula assemblage

Sedimentation conditions I or pollen f,hepollencurves(Fig.12)arecontinuousandrelativelyeven.Itherefore has been assume that dynamic sorting of pollen during the sedimentation Berglund minimal in spite of the near-shore depositional environment. (1g73)found much more irregular curvesin similar sedimentsfrom Gothen- The im- burg. Sample 10, from the well-sorted sand' may be an exception' Betula poriunt feature of this sample, however, is that it contains 62.5 % both in and no Corylus. Since the pollen of BetulaandCorylus are similar' sample there- size and form, a dynamic separationseems unlikely, and the fore was almost certainly deposited before the immigration of Corylus.

Lon g-dis t ance tr onspor t ed pollen during the It is generally assumedthat pine (Pinus) did not grow in Norway Berglund (1966: Allergd and Younger Dryas chronozones. According to LATE WE,ICHSELIANSEDIMENTS IN WESTERNNORWAY 43

144) the northern boundary of pine during the Allerpd lay in northwestern Germany and southeasternSweden. In the present diagram Pinus consti- the per- tutes 4-l | 7o of the total pollen. These figures are very similar to centagesof Pinus in the Blomgy diagram (Mangerud 1970) and the Brond- myra diagram (chanda 1965),while Hafsten (1963) found slightly higher percentagesfrom Lista, which was, however, situated closer to the pine forests. of the mixed oak forest constituents only I tJlmus pollen and 2 Querqus pollen were found altogether. These numbers are also comparable with the numbers usually found in Late Weichselian limnic sediments.Very few other exotic pollen are identified.

Alnus Alnus pollen is also generally considered to be long-distance transported when found in Late Weichselian sediments,even in southwesternSweden (Berglund 1966:38). Consideringthe low percentagesgenerally found, this belief is probably correct. In the present diagram, however, Alnus has a maximum of ll Vo in the lowermost sample, and is found in all samples except 25. The mean for the Late weichselian samplesis 3.6Vc, compared sediments to 6.5 /6 for Pinus. This is remarkably higher than for limnic (Table 3) and I am therefore consideringfour possibleinterpretations: l. Alnus pollen has been redeposited. 2. Alnus pollen is enriched in marine sediments' 3. The Alnus pollenpresent is exotic, and brought in through atmospher- ic circulation. 4. Alnus grew in westernNorway during the Late Weichselian'

1. Redepositionof Alnus is unlikely, as pre-Weichseliansediments are ex- tremely rare in the area (Mangerud 1972a).lf Alnus were redeposited,also other redepositedpollen should have occurred'

in sor-rthern (Lista) Table 3. The content of Alnus pollen in Late Weichselian sediments and western Norway. '

Maximum -- No. of sam- Mean of ull Depositional in one pl.. in which rotal no. environment spectrum. -fam-Pres Alnus of samples 7o of tolal Vo of total is found

Lista Limnic z <1 15 (Hafsten 1963) Brpndmyra Limnic 0.5 ru0 3 35 (Chanda 1965) BIomPy Limnic 1A ru0 I t2 (Mangerud 1970) o Agotnes Marine 11 3.6 8 Four other localities Marine 5.6 1.8 8 8 in Hordaland J. MANGERUD

2. A marine overrepresentationwould explain the difference in content of Alnus pollen in marine and limnic sediments(Table 3). Especiallystrik- (Fig. 12)' ing is the difference between the marine sedimentsat Agotnes (Mangerud with abundant Alnus pollen, an{ the limnic sedimentsat Blomly north' 1970),deposited during the same period, located only 16 km to the and having nearly no Alnus Pollen. pre- However, marine overrepresentationof Alnus is not reported from vious analysis of marine sediments.Combined with the other arguments in below, I find this explanation unlikely, though it has to be considered future investigations. 3.Consideringlong-distancetransport,itisstatedabovethatthefre-

o_ t/! s (, q, JJ t n

5 Atnuso/o of >P o x Somptesfrom Agotnes o Somplesfrom other Lote Weichseliqnmqnne sedimentsin Hordolqnd calculated as in Fig. 13. Relative pollen frequences of Alnus vetsus Plnus. Percentages the pollen diagram (Fig. 12). 45 LATE WEICHSELIAN SEDIMENTS tN WESTERN NORWAY quenciesof Pinus anil QM are similar to those of the other Late Weichselian signif- ,liugru-r, and very few other exotic pollen occur' Alnus is the only and icant exception. However, when comparing the percentagesof Alnus in- Pinus in each spectrum(Fig. l3), a slight correlation appears'which may dicate a long-distanceorigin of Alnus. im- In present-daytundra environments Alnus is one of the relatively portant constituentsof the pollen rain (Birks 1973),especially in Greenland more fre- anil canaila (Fredskild 1913: 198 201) were Alnus occurs much quently in the northern forestsnear the tundra than in Europe' In Svalbard (Flyvdrinen 1968, 1970)Alnus constitutesa considerablepart of the exotic pollen during the Holocene. The percentagesare, however, always very low comparedwith Pinus and Betula. of A striking feature of the present diagram, favouring the hypothesis Allerld exotic origin of Alnus pollen is that the Alnus curve is lowest in production and highest in older Dryas and Younger Dryas, when the local was lowest. 4. The main inclication of the presenceof Alnus in Hordaland during In pol- the Late Weichselianis the high relative frequenciesof Alnus pollen' len diagramsfrom the boundary area of Alnus incana in northern Norway, of alder the percentagesof Alnus seldom exceed 2Vo, even when copses 1974' com- occur closeto the locality (K. D. Vorren, 1912 and pers' comm' pare also Kelly & Funder 7974 12-16, and Tallantire l9l4: 533)' The pollen diagram (Fig. 12) exhibits a normal vegetationaldevelopment re- for the Late Weichselian in western Norway, indicating that it really flects the existingvegetation, and thus that also Alnus was present' Reite(1968:21821g)reportsafindofAlnuswoodinaclaydepositat clay were sunnmore, ca. 250 km north of Hordaland; marine shellsin the -+ the wood radiocarbon dared to Aller@d (11,620 120). He believed that and as- found in these depositsdid not come from the A11er9dvegetation, However, sumed that it was either driftwood, or that it had been rebedded. Norway this find would be in accordancewith Alnus growing in western during the Late Weichselian. Two speciesof Alnus occur in the present vegetation of Scandinavia (L.) (Hu1t6n 1971: 153): Alnus glutinosa (L') Gaerth. and Alnus incana certainly Moench. of these A. glutinosa is a southerly species,and could thermic de- not have grown during the Late Weichselian' A' incana has have mands similar to Pinus silvestris(Tallantire 1974: 536), and should Allerod' Its been able to grow in the fringe areasof Scanclinaviaduring the probably it was actual distribution at that time is, however, not known; locaterl further to the east (Tallantire 1914: 539)' growing in However,the possibilitycannot be excludedthat a speciesnot identify Alnus Scandinaviatoday was present,though we have not tried to possible speciesis the North American Alnus pollen'crispa, to specieslevel. One which in Greenland today has a southern limit nearly identical 199)' As- with the northern limit of Betula pubescens(Fredskild 1973: 46 J. MANGERUD suming that this speciesdid grow in western Norway during the Late Weich- selian,the maxima and minima in the pollen diagram (Fig. 12) and the lack of its pollen in southern Scandinaviacould be explained. If it occurred in local copses,the difference in Alnus frequenciesin the diagrams from dif- ferent localitiesis also reasonable. At presentI must leave the question of the abnormally high percentages of Alnus open; however, my personal opinion is that a speciesof Alnus did grow in westernNorway during the Late Weichselian'

Lat e W eichselian ve ge t ation The conclusion regarding the number of exotic pollen indicates that the source area for the major part of the pollen and spores is relatively re- stricted. This is also the conclusionwhen consideringthe main composition of pollen, which is typical of the Late Weichselianof southwesternNorway (Fagri 1940,Hafsten 1963,Chanda 1965,Mangerud 1970)' Instead of attempting a closer identification of the area of provenance,I emphasizethat over short distancesthere must have been severaldifferent vegetational types. Within 20 km of the locality there were long shores against the open ocean, deep inlets and fjords, hills and mountains of all elevationsup to 600 m a.s.l.and slopesin all directions.As pointed out be- fore (Mangerud 1970: 128), the uneven and patchy distribution of soil is also a very important factor when interpreting pollen diagrams from the naked rock coast of Hordaland. We are forced to envisage a landscape with a mosaic of different plant communities. Pollen diagramsfrom small ponds,such as the one from Blomly (Mange- rud 1970), represent the vegetation of a small area, and therefore relatively few of these communities. The responseto climatic changesof this vege- tation will often be clear, and the interpretation of the pollen diagram rel- atively simple. Diagrams from larger lakes, and especially the present one from the sea, representin this area the entire mosaic of the vegetational types,and the resultsof climatic changesare much more complex' I have previously concluded (Mangerud 1970: 127) that arboreal birch grew in Hordalancl during the Aller@d Chronozone. This is also indicated by studiesof pollen morphology from other localities in Hordaland. In the present diagram the maximum of Betula is only 34/c, and I therefore as- sume that there were only copsesof birch on the most favourable habitats, with deep soil and good exposition. Betula decreasesmarkedly from the Betula-NAP zone (Aller4d) to the Artemisia-Lycopodium zone (Younger Dryas). In favourableplaces there were also juniper shrubsduring both Al- lerld and Younger DrYas. compared with the diagram from Blomoy, the percentagesof Sa/jx and Cyperaceae are much lower, whlle Artemisia is much higher. This must imply that the present diagram mainly reflects the vegetation of drier hab- itats that the Blomgy diagram, probably a grass-herb heath or tundra on hill slopes. It cannot be determined whether Artemisia was scattered LATE WE,ICHSELIANSEDIMENTS IN WESTERNNORWAY 47 throughout the vegetation,or restrictedto certain habitats (cf' discussionin Berglund 1966: 125-126). Nevertheless, it indicates large areas of other communities than the snow-bed communities (Hafsten 1963 334) during both Allerod and Younger Dryas, as concluded by Ovstedal & Aarseth (1915)on other evidence. Artemisia increased from the Betula-NAP zone (Allerpd) to the Arte- misia-Lycopodium zone (Younger Dryas), as it did in Lista (Hafsten 1963), while it decreasedin Brgndmyra (Chanda 1965).In the present area I as- sume that the light-demanding Artemisia took over many of the favourable habitats from retreating Betula and luniperus. Hyvdrinen (1975) has described an Artemisia zone of Younger Dryas age from northern Norway and correlatedit with similar zonesin easternFin- land and central Russia. An Artemisia-rich vegetation was also typical for southern Scandinavia cluring the Younger Dryas (Berglund 1966: 126), and the present diagram indicates that a similar vegetation even occurred in western Norway. If this vegetationindicates a continental climate, the lat- ter would explain some of the contradictory climatic conclusionsreached from Younger Dryas snowlinesand fossil icewedges(Mangerud & Skreden 1972: 94). Lycopodium (almost exclusively L. setago) has a remarkable increase from the Betuta-NAP zone to the Artemisia-Lycopodium zone' and during the latter zone it was an important constituent of the tundra vegetation.

Hystrix A group of microfossilsusually called Hystrix (or Hystrichosphaerids)by Scandinavianpalynologists, and used as indicators of marine environments, are now being identifieclas resting spores(cysts) of dinoflagellates,and sev- eral specieshave been identified (Nordli 1951, Wall & Dale 1967, Wall 1970).In this study Hystrix was only counted, without identification of the types present. Their relative occurrence (Fig. 12) is, however, rather re- markable. In the lowest 4 samplesthere are 3-16 Hystrix for each pollen, while in the other samplesthere are 0.04-0.02Hystrix for each pollen. The drastic change in the frequency of Hystrix occurs simultaneously with the change of the colour of the sediments,anil also with the onset of the changeof the terrestrial components.Certainly, the decreaseof Hystrix indicateschanges in the environment; whether the important factor was temperature, ice- drift cover, salinity, depth or others, is, however, impossible to determine on the basisof the presentmaterial.

Correlations The correlation of the sequenceat Agotnes with eventsin Hordaland during the Allergcl and Younger Dryas has been discussedearlier, and the main conclusionsare summarizedin Fig. 14. J. MANGERUD

:: E \\ S o \ : :S :\ \. EziN:l:\ \.::..,. :.:.:l : :-'-..:.:' \ \ :.1.'..' \ ! r o{ (t z o UO NO E{ F

o

F a(= -o z N N o a G != u^

6 N

gN F 9 F F *+ +F + 8 oo Oo o,l ; 12 | o o N d eg N i; I XR N o{ | o -"r! t ols lNrav!1slNNlA 169108t:0t60t ols lN avn r/ct I nQt;ocz nl czor.ros_r.ozz zr slN rotv NSAaSgig \\ - -AU -lr-va ndre o-s r Nr rLg I czer 09ti0t€6 08110160l l:::ltSNS -[\

I [^ EuN zu [\\ a

I

E N o.^ D o E co I ; TY D oo - Oo @

rlrrrrlrr'' AAa ooo ooo ?;s o ooo E

The age of the Agotnes clay is, however, of importance in the dating of older events,and radiocarbon dates recently obtained from other localities necessitatea re-evaluationof some earlier correlations. At Blomvig, on the island BlomOy (Fig. 1), fossil-bearing sediments underlying a till were excavatedin 1942143(Und6s 1942, Mangerud 1970: fig. 5). Since this sequenceis decisive in the interpretation of the Late Weichselian in western Norway, the beds are given formal lithostratigraph- ical names, the type locality being the graveyard at Blomvig. The till is named the Ulvgy Till, a name derived from an island just to the south of the type locality. The sedimentsbeneath the till become the Blomvdg Beds, which include all the sedimentsbetween the Ulv@yTill and the underlying till or bedrock at the type locality. The first radiocarbon dates from the Blomv6g Beds gave 12,670-r- 350 and 12,200-+ 350 (Nydal 1960: 88), but the exact position of the dated ma- terial within the Blomv6g Beds is not known. Further samplesfrom the collections of the Geological Museum, University of , were dated recently. Two shell samples,both mainly Modiolus modiolus, collectedfrom the baseof the Blomvig Beds gave 12,540-p 180 (T-1696) and 12,540-+ 150 (T-1697).Also from the top of the unit two sampleswere selected,one of shells (Mytilus edulis) which gave (T-1882) 12,400-+ 90, and one of whale bone, which was dated with two different pretreatments:Treatment with diluted HCI gave (T-1899/l) l2,ll0 -+ 100 and the EDTA method recom- mended by Olsson et al. (1914: 180) gave (T-189912)11,920-+ 80, when correctedfor an apparentage of 440 years.The correlation with other dates (Fig. la) indicatesthat the whale-bonedates may be slightly too young. Although the section at BlomvAg, and thus the Ulvgy Till, is not acces- sible today, both the site and the samplesfrom the excavation have been studied.After discussionswith the scientistswho visitedthe localitv in 1942- 43, there can be little doubt that the Ulv4y Till representsa basal till, and indicatesan ice advance. At Dale, just 600 m north of the Blomvig locality, a pollen diagram (Mangerud 1970)from limnic sedimentssituated stratigraphically above the Ulvgy Till has been obtained(Fig. l4). Dating of the lowermost limnic sedi- ments gave 12,070-+ 180 (T-612), slightly younger than most dates from the top of Blomvig Beds. The Agotnes locality is situated l5 km southeastof Blomvig (Fig. 1). The direction of glacial striae suggeststhat Agotnes was ice covered during the deposition of the Ulvgy Till; the entire sequenceat Agotnes must,

Fig. 14. Schematic profile of the Hordaland area, modified from Mangerud (1970, 1972a, 1973), with additional radio-carbon dates from Aarseth & Mangerud (1974) antl, unpublished material. The profile is mainly parallel to the direction of ice movement, the North Sea coast lying to the left. Dates between vertical stippled lines are from sediments undisturbed by ice. All the others are from sediments overrun by glaciers. All dates are corrected for apparent age (p. 33). 50 J. MANGERUD therefore, be younger than the Ulvgy Till (Fig. l4)' Nevertheless,the date obtained from the Agotnes clay (12,220 -+ 150, T-1023) is slightly older than those for the whale-bonefrom the Blomvig Beds. Clearly, the duration of the ice aclvance that produced the rJlvfrY Till was short (100-200 years?), and is within the limits of uncertainty for radiocarbondates from sediments depositedimmediately before and after the advance.The clear stratigraph- ical evidencefor the advanceat Blomgy is therefore very important. Based on the dates cited above,the age of the Ulvoy Till can be bracketed within the time interval 12,400-12,000years B.P., and the most probable age is 12,300-12,200. A section in Sandviken, Bergen (Mangerud 1910, 1912b), was correlated with the Blomvig Beds on the basisof 1aCdates. However, an error in the calculation of the agesof two samplesfrom this exposurehas been discov- ered; only the first date obtained (12,440-+ 150, T-750) now supports this correlation. At present 5 dates are availablefrom Sandviken(Fig. l4), the mean age being 12,000B.P. The conclusionmust be that the Sandvikensed- iments are, in all probability, younger that the Ulvgy Titt. In spite of uncertaintiesin the exact age of the Ulvly Till, it now seems quite clear that it is older than the Older Dryas Chronozone(Fig. l4). The vegetationalchanges during the Older Dryas Chronozonein southern Scan- dinavia (Mangerud et al.1974), and the British Isles(Pennington 1975)'and the ice advancewhich depositedthe Ulvgy Till, were therefore either the result of two separateclimatic events,or there existedsome time lag in the responseto the same climatic change.In the latter case,the change must have taken place before the Older Dryas Chronozoneas defined by Mange- rud et a]. (1974). The Skarpnes (cold) event in northern Norway (Andersen 1968:35) seemsto be of the sameage as the Ulvpy Till. No inclicationsof another re-advanceduring the Older Dryas Chronozone have so far been found in Hordaland. It is, however, very difficult to iden- tify

Conclusions At Agotnes there is a sequenceof marine sedimentsfrom the Older Dryas Chronozone, through the Allerpd and Younger Dryas, to the Preboreal Chronozone.The sedimentsare not disturbed by ice, and thus prove that the Younger Dryas re-advancedid not extend as far as this locality. The age of the baseof the sequence(12,220 + 150 years B.P.), combined with other radiocarbondates from Hordaland, indicatesthat the Ulvgy Till is older than the Older Dryas Chronozone, and thus that the re-advance which depositedthe till took place in the Bplling Chronozone. Warm Atlantic water enteredthe Norwegian Seaprior to 12,600B.P. The Polar Front. betweenthe Atlantic water and the Arctic water, was situated LATE WEICHSELIAN SEDIME,NTS IN WESTERN NORWAY 5I between Troms (northern Norway) and Hordaland during the Bglling and pos- Allergd chronozones. During the Younger Dryas the Atlantic water sibly did not reach the coastof Hordaland. The environment on the coast of Hordaland during the Allerod Chrono- zone seemsto have been very similar to that of the coast of Troms and Finnmark (northern Norway) today. The vegetation was open with copses of birch (Betula pubescens coll.) and shrubs of juniper (luniperus com- munis) and willow (Sc/ix sp.). The littoral and shallow-water shell faunas inclucled species(e.g. Modiotus modiolus, Thyasira sarsi, Littorina littorea, Gibbuta cineraria) whosepresent-day distribution has a northern limit in or immecliately to the north of Finnmark, and also species(e' g' Chlamys is- landica, Alvania scrobiculqta) which today are uncommon south of Troms and Finnmark. Due to glaciers, the ecological gradients from the inland towards the coast were certainly steeperin Hordaland during Allerod times, compared with Troms and Finnmark today. we do not know the exact position of the glacierswhich existed at that time, but they were probably calving in the inner parts of the fjords. gla- During the Younger Dryas the environment changed drastically: The ciers advancedmany tens of km; most of the trees disappeared,as did the most of the temperate molluscs; the foraminifera faunas changed; and the organic production in the sea,including Hystrix, decreased' The pollen diagram from the shallow-watelmarine sedimentsat Agotnes pol- can be correlatedwith diagramsfrom limnic sediments,indicating that re- len can be an important tool in the correlation of terrestrial and marine cords along the coast of NorwaY. In Hordaland, Late Weichselian marine and terrestrial sediments are closely correlated by means of radio-carbon datings. we have, however, also tried to investigateseveral aspects of the sedimentsand fossils,in order to provicle a basis for the correlation of sedimentswhich cannot be radio- carbon dated.

Ivar Acknowledgements.-I wish to expressmy gratitude to Mr. Johan Lund, cand.real. Wikander, Miljeteig, c"and.real.Kire skir, cand.real.Eivind S@nstegaard,cancl.real' Per fossils' The an<1Dr. Anders War6n for help in the laboratory and with identification of raC datingswere performed under the direction of Dr. Reidar Nydal and siv.ing Steinar Profes- Gulliksen. The manuscript was read critically by Professor Bjorn G. Andersen, Kjell R' sor Knut Fregri, ProfessorHans Holtedahl, cand.real.Inge Aarseth, cand.real. parts of it by cand. Bjorklund, cand.real.Karl-Dag Vorren, canil.real.Tore vorren, and the English real. Eirik Lande and Dr. Rolf Feyling-Hanssen.Dr. Brian Robins corrected pelsonsI proffer my language.The figures were drawn by Mr. Jan E. Lien. To all these Researchcoun- sincerethanks. The work was financially supportedby The Nor*'egian cil for Scienceand Humanities (NAVF). March 1976 52 J. MANGERUD

REFERENCES J' P'' Lorentzen-Styr' T' J ' Aarseth, I., Bjerkli, K. Bjgrklund, K. R., BOe, D', Holm, Myhre,L.A-.,Ugland,e.S.AThiede,J.1975:LateQuaternarysedimentsfromKors- fjorden, western Norway. Sarsia 58,43-66' between Hardanger- Aaiseth, I. & Mangerud, J. 1974: Younger Dryas end moraines fjorden ancl Sognefjorden, Western Norway' Boreas 3' 3-22' North Norway' Nor' Geol' Andersen, B. G. 1968: Glacial geology of western Troms' Unrlers. 256, 160 P. Anrlrews,J.T.1972:Recentandfossilgrowthratesofmarinebivalves,canadianArc- tic,andLate-QuaternaryArcticmarineenuironments'Palaeogeogr'Palaeoclim'Pa' Iaeoecol- 11,157-176. and carbon isotoptc compo- Andrews, J.T. 1973: Late Quaternary variations in oxygen Palaeoclim. Palaeoecol' 14, sitions in Canadian Arctii marine bivalves. Palaeogeogr. 187-192. Berglund,B.E.l966:Late-QuaternaryvegetationineasternBlekinge'south-castern Sweden. I. Late-Glacial time. Opera Botanica 12' 1' l-180' the Gothenburg core B 873 Berglund, B. E. 1973: A pollen analytical investigation of ol Lund. Dept. ol Quat' to establish the Pleistocene/Holocene boundary. University Geol. Report 1, 12 PP. Isle ol Skye. A Palaeoecological Birks, H. J. B. 1973: past and Present Vegetation ol the StudJ. Cambridge University Press. Cambridge' 415 p' pcstglaciale niviLforandringer i Kristi- Br,ijgger, w. c. 1900-1901: om de senglaciale og p' unlf.tt.t (Molluskfaunaen). Nor. Geol Unders' 3l ' 731 A Late-glacial and early' Chancla,- S. 1965: The history of vegetation of Br,ijndmyra. Bergen, Mat.-nat.set. losi_glacial cleposrt in Jaeien, South Norway. Arbok LJniv. 1965,1,5-17. Sons' N Y' 588 p' Dietrich, G. 1963: General Oceanography' John Wiley & and environmenl' Sediment. ]0, Doeglas, D. J. 1968: Grain-size inalc"s,'classification 83-100. Norwegen' II' Zur sptit- Fegri, K.: QuatdrgeologischeUntersuchungen im westlichen rekke 7,201' p' quartarenGeschichte Jerens. Bergens Mus.Arbok 193940, Naturvit' Copenhagen' Fegri, K. & Iversen J. 1966: Textiook ol pollen analysis.Munksgaard. 237 P. of Billefjor- Feylin!-Hanssen,R. W. 1955: Stratigraphy of the marine Late-Pleistocene den Vestspitsbergen'Nor. Polarinst' Skr' 107, 226 p' depositsfrom the Oslo- Feyling-Han.ss.,],R. w. 1964:Foraminifera in Late Quaternar-v fjord area. Nor. Geol. [Jnders.225' 383 p' Lykke Andersen' A'-L' Feyling-Hanssen,R. W., Anker Jcirgensen,J', Knudsen, K' L' & Denmark and Sandnes,Nor- lg1l: LaIe Quaternary Foraminilera from ven

Hyvdrinen, H. 1968: Late-Quaternarysediment cores from lakes on Bjornoya. Geogr. Ann. 50 A, 235-247.Stockholm' Hyvlrinen, H. 19'70:Flandrian pollen diagrams from Svalbard. Geogr. Ann. 52 A, zt )-LaL, Fenno- Hyvdrinen, H. 19':.5 Absolute and relative pollen diagrams from northernmost scandia. Fennia 112, l-23. Fauna. Jensen, A. S. & Spbrck, R. 1934: Bliiddlr ll. saltvandsmuslinger. Danmarks G. E. C. Gads forlag. Kpbenhavn. 208 p. sediments Kelly, M. & Funder, S. 1974: The pollen stratigraphy of late Quaternary lake of South-West Greenland. CrQnlands Geol' (Jnders' Rapport 64' 26 p' postglacial tid' Kolclerup, C. F. 1908: Bergensfeltet og tilstddende trakter i senglacial og Bergens Mus. Arbok. 1907, No. 14' 268 p Bergen Mangerutl, J. 1970: Late Weichselian vegetation and ice-front oscillations in the clistrict, western Norway. Nor. Geogr. Tidsskr. 21, 121-148' glaciations Mangerutl, J. l9':,2a: The Eemian Interglacial and the succession of during the last ice age (Weichselian) in southern Norway. Ambio Spec' Rep 2' 3944' ap- Mangerud, I.19':-26 Radiocarbon dating of marine shells, including a discussion of parent age of Recent shells from Norrval'. Boreas 1, 143-l'72' til Fusa. Mangerud, J. 19'1-3:Horclalands natur, under og like etter istiden' Frd Fjon Arh. Nord.- og Sogelag, 26"743' Mangerud, J., Andersen, S. T., Berglund, B. E. & Donner, I. J. l9'74 Quaternaly strati- graphy of Norclen, a proposal for terminology and classification. Boreas 3,109-127. shells uu"g"r"o, J. & Gulliksen, S. i975: Apparent radiocarbon age of Recent marine from Norway, Svalbard and Ellesmere lsland. Quat' Research 5' 263-273' in sediments Mangerud, J. & Skreden, S. A. 1972: Fossil ice wedges and ground v'edges below till at , Western Norway. Nor. Geol. Tidsskr' 55"73-96' mean McCammon. R. B. 1962: Efficiencies of percentile measures for describing the size antl sorting of sedimentary particles. J. Geol. 70' 453465' in Vestspits- Nagy, J. 1965: Foraminifera in some bottom samples from shallow waters bergen. Nor. Polarinst. Arb. 1963' 109-128. Nordii, E. 1951: Resting spores in Goniaulax polyedra stein. Nytt Mag. Naturvid.88, 207-212. Nor

Strauch,F.19.12a:Phylogenese,AdaptionundMigrationeinigernordischermariner Mya)' Ahh' Senckenberg'' Molluskengenera (Nepiunea,'Panomya' Cyrtidaria und naturlorsch.Ges. 531,l*211. und ihre geologisch-paldon- Strarcn, F. 1972b:Zur Klimabinclungmariner Organismen tologischeBedeutung. N. ib. Geol' Paliiont' Abh' 140' 2-127' (Alnus incana (L') Moench ) fufiunii.., p. A. 1974:The palaeohistoryof the grey alder andblacka|der(A'glutinuLsa(L.)Gaertn.)inFennoscandia'NewPhytol.73'529_546. (Natural History)' Lon- Tebble, N. 1966:British Bivalve Siashells.The British Museum don.213 P. Undis, I. 1942:Fossilfunnet i BlomvAg'Naturen'9'7-10'l' palsa bog in Northern Norway' Vorren, K.D. 19'.2:Straiigraphical investigationsof a Astarte5, 39-71. Vorren,T.o,|972:Interstadialsedimentswithrebeddedinterglacialpollenfrominner Sogn, west Norway. Nor. Geol' Tidsskr' 52' 229-240' Wall,D.1970:QuaternarydinoflagellatemicroPaleontology-19591o7969'Proc'N' Am. Paleont.Conv.' Sept.1969, Part G' 844-866' marine dinoflagellatesand their wall, D. & Dale, B. 1967;The resting cysts of modern put.o"tofogical significance'Rev' Palaeobot'Palvnol 2' 349-354' (Gastropoda,Prosobranchia)' w;;;;, A. DIT+:n"uirlo., or the Arctic-Atlantic Rissoidae Zoologica ScriPta 3' 121-135' (Modiola modiolus (L.\)' Fiskeridir' Wiborg, K. F. 1946: UnCIr.gt"tr". over oskjellet Skr. Ser.Havunders.8, 85 P' sca||opChlamys islandica (Mi|. Wiborg, K. F. 1963:some ob,"ruations on the Icelancl ler)inNorwegianwaters.Fiskeridir.Skr.Ser.Havunders.13,No.6'38_53. Zagwijn,w.H.&Veenstru,H'J.1966:Apollen-analy'ticalstudyofcoresfromthe Juter Silver Pit, North Sea'Marine Geol' 4' 539-551'