21. deep-water continental-margin sedimentation, … · 2007. 5. 24. · david j.w. piper and...

21. DEEP-WATER CONTINENTAL-MARGIN SEDIMENTATION, DSDP LEG 28, ANTARCTICA

David J.W. Piper and Creighton D. Brisco, Departments of Geology and OceanographyDalhousie University, Halifax, Nova Scotia, Canada

ABSTRACTSites 268, 269, and 274 lie in deep water on the continental margin

of Antarctica to the south of Australia. Site 268, on the continentalrise of Wilkes Land, consists of a 150-meter-thick sequencedominated by turbidites, overlying a 300-meter-thick mudstone se-quence, with many silt laminae believed formed by bottom ("con-tour") currents. The turbidity currents were generated by slumpingof unstable masses of ice-rafted sediment, a resedimented pebblymud was recovered, and several otherwise well-sorted and well-graded medium sand beds have lutum contents of 25%-50%. Thelaminae formed by bottom currents can be distinguished from thoseof turbidity-current origin by their sedimentary structures and theirhigh content of microfossils, similar to those in interbedded muds.Site 269, on the Wilkes Abyssal Plain, is dominantly a silt-clay tur-bidite sequence. Sedimentary structures are exceptionally wellpreserved in the silts and show much evidence of synsedimentarydeformation. Although on the continental rise off Victoria Land,Site 274 was protected during most of its history from directsedimentation from the continent by an upslope graben.

The abundance and grain size of ice-rafted debris at these sites,and Sites 267 and 266, to the north of 268, in an area of pelagicsediments, have been studied. Ice-rafting extends back to at least thelower Miocene in both the Wilkes Land and Ross Sea sectors of Ant-arctica. Tentative interpretations of the type of ice cover in Ant-arctica since that time can be made.

At Site 274, ice-rafted sediment appears to derive mainly fromMarie ßyrd Land, but fine sediment and turbidites appeardominated by a local Victoria Land source. Substantial exposure ofthe Beacon Group from beneath the Victoria Land basalts probablyfirst occurred in the late Miocene. The petrology of detrital materialsat Site 269 is similar and somewhat intermediate to that at Site 268.Crystalline igneous and metamorphic rocks dominate the sourcearea for Site 268, but in post-Miocene sediments there is evidence ofsignificant amounts of Beacon-type detritus.

Dark micronodules of fine-sand size are common at severalhorizons. Most of these are similar in composition to manganesemacronodules and pebble coatings found at Sites 274 and 267. A feware dominantly of iron oxide and clay and contain virtually nomanganese. No close correlation appears between micronodules andbottom current activity.

Porcellanous cherts are found in Miocene and Oligocenesediments at Sites 268, 269, 272, and 274. They are formed by thediagenetic growth of disordered cristobalite in mudstones containingbiogenic silica. The variety of colors and bedding structures in themudstones are also found in corresponding cherts. Growth of cristo-balite appears to precede pronounced lithification. The source ofsilica is believed to be finely divided diatom debris but there is no un-equivocal evidence for this. Larger diatom and Radiolarian skeletonsare found in the cherts and appear to be uncorroded.

727

D. J. W. PIPER, C. D. BRISCO

INTRODUCTION

Leg 28 of the Deep Sea Drilling Project drilledthrough thick continental margin sediments in theAustralian-New Zealand sector of Antarctica at threesites (Figure 1). Sites 268 and 274 are located on the con-tinental rise; Site 269, on the Wilkes Abyssal Plain.Lower Tertiary sediments were penetrated at all sites.Sites 265, 266, and 267 are in pelagic sediments north ofSite 268; Sites 270-273 penetrated shelf sediments in theRoss Sea south of Site 274.

There is already much data available from Eltaninpiston cores in this region (Payne and Conolly, 1972;Fillon, 1972). These cores have been dated by magneticreversals coupled with paleontologic dating. Our inter-mittent near-surface recovery and core deformationpreclude the use of paleomagnetic stratigraphy; paleon-tologic dating in our cores is poor.

The sediment sequences are the result of many in-teracting sedimentation processes. Ice-rafting is of ma-jor importance, especially since later Miocene time. Tur-bidites are found at all three sites, and "hemipelagic"movement of fine sediments down continental slopesprobably also occurs. Bottom currents ("contourcurrents") are important at the continental rise sites.Siliceous microfossils (mostly diatoms) have ac-cumulated at Site 274 since the Eocene, but at Sites 268and 269, pre-Pliocene sediments also include calcareousmicrofossils.

CURRENT-DEPOSITEDDEEP-WATER SEDIMENTS

Site 268 (Unit 1)

Three sediment types are found in the upper 160meters at Site 268:

1) diatom ooze, diatom-bearing silty clay, and in-termediate lithologies;

2) clay, silty clay, and clayey silt, with laminae ofmedium silt; and

3) beds of fine sand.

GRAIN SIZE Φ

Figure 1. Location map for Sites 266, 267, 268, 269, 272,and 274. Shows principal Antarctic ice shelves, bathy-metry in meters, and subice drainage divides.

Figure 2. Grain size analyses of two turbidite sand bedsfrom upper part of Site 268. (a) bed with base at 268-2-2, 130 cm. (b) bed with base at 268-5-2, 59 cm.

The fine-sand beds are typically 2-20 cm thick, aresharp based, and grade upwards into clayey silt. Twotypical beds have been analyzed in detail for grain size(Figure 2). The bed with a base at Sample 268-2-2, 130cm consists of well-sorted sand, with a mode around 20,and very little sand in the 30-40 range. Yet the bed con-tains 20%-23% lutum (finer than 40), probably mostly ofclay size. Such high lutum contents in graded mediumsand beds are most unusual (Kuenen, 1966). A bed witha base at Sample 268-5-2,59 cm has a mode around 3.50;the sand fraction is poorly sorted, but graded, and thelutum content of the lower part of the bed is around60%, most of which is of clay size. Spot checks on twoother sand beds show high lutum contents.

In the clays and muds, silt laminae are common, bothsingly and in groups. Diatom contents in thesesediments are very low; ice-rafted granules and coarsesand are absent. The laminae are 0.5-2 mm thick, ofwell-sorted angular quartz and feldspar in the medium-and coarse-silt range. They usually have distinct basesand tops, and are not lenticular. Their appearance isquite different from the laminae in the lower part of thehole, described below. The absence of pelagic biogenicand ice-rafted material in the associated sediment mightsuggest a high rate of deposition, and hence a turbiditeorigin; however, there is no direct evidence for this inter-pretation.

Near the base of Core 1, Section 6, there is a 50-cm-thick bed of pebbly mud, overlain by a 4-cm-thick well-sorted sand bed. The pebbly mud consists of pebblesand granules of crystalline rocks and varicolored clastsof mud of pebble size in a matrix of sandy mud (Figure

728

DEEP-WATER CONTINENTAL-MARGIN SEDIMENTATION

C1ayey s i l t

Very f i n e quartz sani

pebbles absent

in upper 10 cm

Muddy

sand

with

pebbles

and

mud

clasts

pebbles more

abundant in

lower part

of core

©Θ

Clay

GEOCHEMICAL

SAMPLE

CORE

CATCHER

pebbles >l cm

major mud clasts

Figure 3. Sketch section of pebbly mud bed at base of 268-1-6, showing distribution of larger pebbles, and grain sizeanalyses.

3). Pebbles are much more abundant near the base of thebed.

In the sand bed, well-sorted foraminifera andforaminiferal fragments predominate in the lower partof the bed, while quartz is dominant in the upper part.The foraminifera are entirely Globigerina pachyderma.They are also found in other parts of Core 1, and are es-pecially abundant in the pebbly-mud bed. The sitegenerally appears to have been beneath the carbonatecompensation depth, and the accumulation offoraminifera presumably represents rapid deposition.

The high G. pachyderma content of the sand and theunderlying pebbly mud suggests a genetic relationship.Winnowing of the pebbly mud by bottom currents isprecluded by the relationship of quartz to foram sandand the lack of coarse sand and granules in the sand bed.The sedimentary sequence is very similar to that de-scribed by Kelts and Briegel (1971) from Lake Zurich,produced by a mudflow and a turbidity current ap-parently generated from it. Resedimentation ofshallower water sediments from above the carbonatecompensation depth would also explain the highforaminifera content.

Site 268 (Units 2 and 3)

Beds of obvious turbidite origin are absent in thelower two units at Site 268. No sand beds were

recovered. The dominant bedding feature is abundantsilt laminae (Figures 4, 5). These laminae are often len-ticular and frequently do not appear to have sharpbases. Where they occur in diatom- and radiolarian-richmuds, the laminae have concentrations of size-sorteddiatoms and Radiolaria.

These laminae appear quite different from thosefound in turbidite-dominated sequences, for example,Site 269, described below, or DSDP Sites 178-181 in theGulf of Alaska (Piper, 1973). The site is located on acontinental rise. There are thus a priori grounds for con-sidering the laminae to be produced by normal bottomcurrents or contour currents.

There are no generally accepted criteria in theliterature for distinguishing laminae deposited by tur-bidity and contour currents. Several authors (Field andPilkey, 1971; Bouma and Hollister, 1973) have erectedempirical criteria for distinguishing the two, but theyhave not satisfactorily demonstrated that their "con-tourites" are not distal turbidites. Furthermore, some ofHollister's generalizations about turbidites, while theymay be true for the area which he investigated, are notapplicable to all turbidites.

Turbidity currents differ from contour currentsprimarily in their much greater suspended sediment loadthat is necessary to keep them in motion. From this, wemay derive the following criteria for distinguishingturbidity-current and contour-current deposits:

1) Deposits with evidence of rapid deposition areprobably turbidites. Such evidence would include clim-bing ripples with no stoss side erosion, animal escapeburrows, and dewatering structures.

2) Deposits with evidence of sediment starvation areprobably contourites. Such evidence includes isolated("starved") ripples and perhaps shallow scours not filledwith unusually coarse sediments.

3) Evidence of highly turbulent erosion is probablyrestricted to turbidity currents.

4) The petrographic composition of contouritesshould follow that of the coarser components of the in-terbedded sediment; this may also occur in turbidites,but does not appear common.

Presently, no reliable empirical generalizations on tex-tural characteristics appear which allow a distinction ofturbidite and contourite laminae, although it may bepossible to distinguish thicker beds in this way.Demonstrable turbidite laminae show wide variations insorting, grading, and grain orientation (Piper, 1972b,1973). Geographic orientation of a pronounced fabricwould probably be a good criterion, but DSDP cores arenot oriented.

With so few criteria available, any interpretation ofthe laminae as contourites must be tentative. Thefollowing generalizations can be made about thelaminae:

1) They are never thick. Silt lenses are never morethan 5 mm thick.

2) Many laminae are lenticular, and thinner laminaeare discontinuous.

3) Some laminae have sharp bases and/or tops; inothers, the margins are gradual.

4) Where the laminae interbed with muds containingbiogenic material, this silt-size material is concentrated

729


grra/M•>iw^•w.•MViii * g ^

- - _••^-^-^^BBaifffca«irSb•«i —tfr^»--"~- - –"*~MM•r - '

cm

•0

% •

i - •.^ - ~~~~

• " ' ' • • • • — . ' " ' • " 1

• n

. . . . .

• •

. • β.-

Figure 4. Probable "contourite" silt laminae, 268-17-3. Note lenticularity of lami-nae, erosional scours not filled with coarser silt, and slight bioturbation.

CM

0

Figure 5. Probable "contourite" silt laminae in thin sections, Site 268, Cores 10 to 12.(a) 268-10-3, 108-110 cm. (b) 268-10-2, 103-105 cm.

in the laminae. Laminae may contain up to 25% diatomsand radiolarians in the coarse-silt size range, where theinterbedded sediment is a diatom-bearing mud. Wherethere is little or no biogenic material in the surroundingsediment, the biogenic component of the laminae is cor-respondingly reduced or absent.

5) Calcium carbonate fossils have not been found inthe laminae although occasional coccoliths andforaminifera are found in turbidites from Sites 269 and274.

6) Demonstrable turbidites from Sites 269 and 274,with a grain size similar to the laminae, contain very fewdiatom and radiolarian fragments, even though, at leastat Site 274, there is an abundance of these microfossils inassociated sediments.

7) The laminae appear well sorted in smear slides.The coarsest grains seen reach about 70µ.

8) There is no pronounced grain orientation visible inthin sections cut normal to the bedding plane.

9) No heavy-mineral placers have been seen.10) Both diatoms and radiolarians often have a dark

mineral coating. The nature of this coating has not beendetermined. It is also found in a few skeletons in the in-terbedded muds, but the proportion of mineralizedfragments in the laminae is much higher. Quartz andother mineral grains do not have the dark coating.

All of the first six points appear more consistent witha contour-current than a turbidity-current origin.

If the interpretation of these laminae as contourites iscorrect, then the criteria proposed by Bouma andHollister (1973) for the recognition of contourites mustbe modified, at least inasmuch as they refer to laminaerather than thicker beds. Specifically, heavy mineralplacers are absent, there is no pronounced grain orienta-

730


tion, matrix content is not a basis of distinction fromturbidites, and microfossils are commoner than in tur-bidites.

Site 269

The division of Site 269 into units is rather arbitrarybecause of the uniformity of the site and the discon-tinuous coring. It is most simply divided into two partson the basis of lithification: an upper unlithified partand a lower semilithified part (from which it is possibleto make polished slabs and thin sections). Cores in themiddle part of the site had very poor recovery, makingdetailed sedimentology difficult.

Upper Part of 269 (Units 1 and 2)Diatom oozes and diatomaceous muds dominate the

upper part of the section and interbed with sand and siltbeds (Figure 6). These beds include (a) Well-sorted veryfine sand beds, e.g., Sample 269-1-5, 124-130 cm (Figure7), 5-30 cm thick with a low lutum content around 10%(cf. Site 268); slight grading visible; sharp base, andsometimes a sharp top. (b) Well-sorted coarse silt beds,e.g., Sample 269-3-3, 14-20 cm (Figure 7), with a sharpbase and grading; usually 1-5 cm thick. These appearvery similar to those described by Piper (1973) from theGulf of Alaska. Some beds pass up into alternatinglaminae of silt and clay, (c) Alternating silt and mudlaminae, with the silt laminae occurring in groups inwhich there is an upward decrease in thickness and grainsize of the silt laminae. This is the "graded laminatedbed" of Piper (1972b).

Sands are commonest in Core 1, silts are commoner inlower cores. It is seldom possible to make anystatements concerning whether mud beds are associatedwith the silt and sand turbidites. Some such beds cer-tainly grade up into silty mud with few diatoms.

Lower Part of Site 269 (Unit 5)

A detailed study has been made of the lower part ofHole 269A (Sections 6-13), where a sequence of earlyMiocene or Oligocene claystones and siltstones were in-termittently cored. The rocks are semilithified, withwater contents in claystones as low as 14%. Recovery,especially in the lower cores, was excellent, with longsections of unbroken core being obtained.

Site 269 is located on the Wilkes Abyssal Plain, about200 km north of the lower edge of the continental rise. Itis probable, from seismic reflection profiles, that thelower part of Hole 269A was located on an abyssal plainin the mid-Tertiary. This interpretation is supported bythe similarity of the upper sedimentary sequence of un-doubted abyssal plain deposition with the lower part ofHole 269A. The main difference is in the lack of diatomsin the lower part of the hole. Sands are infrequentthroughout this hole (Figure 6) and where presentappear to be turbidites. A similar paucity of sands hasbeen found on other abyssal plains drilled by DSDP(e.g., Piper, 1973). The lower part of Hole 269A consistsof undeformed, fine-grained abyssal plain sediment withsedimentary structures well displayed.

I IQ- I— Cd •—iUJU O ZQ Σ CJ ZD

%SAND and SILT BEDS

0 10 20 30

200-

400H

600-

800-

1000J

Figure 6. Distribution of sand and silt beds at Site 269.

731


o.i

4 6

GRAIN SIZE Φ

Figure 7. Grain size analyses ofturbidite sand and silt bedsfrom upper part of Site 269, determined by pipette anal-ysis and sieving.

Cores were split with the band saw and washed clean.Lithologies were logged in detail (centimeter by cen-timeter) onboard ship. Thin slabs were taken for X-radiography.

Size analysis (Figures 8 and 9) was by conventionalsieve and pipette methods. An ultrasonic bath treatmentof 1 Vi hr was necessary to disaggregate samples. Resultsobtained are consistent with smear slide identifications,but are probably not entirely comparable with samplesof unconsolidated sediment. Sandstones with calciumcarbonate cement were successfully disaggregated withacetic acid.

The shipboard logging of cores was aimed at objectivedescription. As logging was completed, descriptionswere immediately rationalized into about 17 lithologictypes. A few core intervals could not be fitted into any ofthese types. Some core lengths were re-examined tocheck assignment to lithologic types, but there was nosystematic relogging.

From the 17 lithologic types, 10 major facies wererecognized (Table 1; Figures 8, 9). These are divided intoclay facies and silt facies as follows:

A) Olive-black claystone: Sometimes silty claystone.Mottling generally absent. (Facies A2 has very slight

6 7 8 9

GRAIN SIZE Φ

Figure 8. Grain size analyses of clay facies A, B, and D forthe lower part of Hole 269A, determined by pipette anal-ysis.

732

0.1 -,

0.014 6

GRAIN SIZE Φ

Figure 9. Grain size analyses of silts from the lower part ofHole 269A, determined by pipette analysis and sieving.

mottling: in this and other facies, mottling generallyappears to result from bioturbation.) Sometimes a veryindistinct horizontal lamination appears to be present.Organic carbon content around 0.4%.

B) Dark brownish-gray silty claystone: Without visi-ble mottling. Interbedded thin silt laminae of Facies Jare common, suggesting that bioturbation has not oc-curred. Organic carbon content around 0.4%.

C) Dark brown mottled silty claystone: Mottlingusually moderate to intense. Organic carbon contentaround 0.1%.


D) Dark greenish-gray mottled silty claystone: Mottl-ing usually moderate to intense. (Facies D2 is similar incolor, but appears unmottled.) Organic carbon contentaround 0.1%.

£) Intensely mottled coarse silt and fine sand: This oc-curs in beds usually several centimeters thick. Most bedshave a sharp top and are overlain by a variety oflithologies.

The beds appear to be bioturbated equivalents offacies F. They show a similar grain-size distribution oftheir sand component. At least one bed of facies Epasses down into an unbioturbated bed of facies F. Abed in Core 13, Section 1 is distinctly graded.

F) Silt and very fine sand beds: Commonly cementedwith calcium carbonate. Several different types havebeen recognized, but recognition may in part be depen-dent on preservation. (Fl) Well-sorted fine sand andcoarse silt beds, without visible structure. (F2) Beds ofvery fine sand and coarse silt, with cross-laminationand/or convolute lamination. The Bouma (1962) bedsequence has been recognized in several beds (Figure10). (F3) Beds of very fine sand and coarse silt, withlamination only. (F4) Beds of medium and coarse silt,with lamination, but not prominently graded. (F5)Medium and coarse silt beds, around 1 cm thick,without visible structure. (F6) Beds of medium andcoarse silt, with lamination and prominent grading.Almost all of these beds have sharp bases, but a fewappear to grade up from facies G and H.

G) Laminated silts: Comprising medium and fine siltlaminae alternating with laminae of clayey siltstone. In-dividual laminae are 0.5-2 mm thick, with beds of thisfacies usually several centimeters thick. Prominenthorizontal lamination is the commonest primarydepositional structure, occasionally with slight scourand fill. Cross-lamination is very rare.

Synsedimentary deformational structures are ubi-quitous (Figure 11). The commonest is the developmentof small load casts and ball and pillow structures, on thescale of about 1 mm, in individual silt laminae. Lesscommonly, load deformation affects composite units upto 1 cm thick. Synsedimentary faulting and extensionnecking or boudinage have been observed.

TABLE 1Summary of Facies in Lower Part of Hole 269A

ThicknessFacies Brief Description Color

A Olive-black claystone, no mottlingA2 As A, but with slight mottlingB Dark brownish-gray silty claystone, no mottlingC Dark brown mottled silty claystoneD Dark greenish-gray mottled silty claystoneD2 As D, but with no mottlesE Intensely mottled coarse silt and fine sandF Beds of silt and very fine sand, no mottlingG Finely laminated silt, often with deformational structuresH Horizontally laminated clayey siltstoneH2 As H, but with slight mottlingI Groups of silty laminae forming graded laminated bedsJ Single or randomly grouped silt laminae

5Y3/15Y3/11OR4/2, 5YR4/1, 5YR3/15YR5/15G4/1,5GY4/1,N4,N3(asD)

28.82.4

29.54.37.10.99.25.63.53.10.71.3a

3.5a

aThickness of sediment with abundant laminae, not total thickness of laminae.

733


269A-TI-3-75/100 269A-11-1-25/50

A

G

A

G

B

LJJ

269A-10-3-125/150

it

6

W

L••

B

1

F2

B

h

B

1

F

269A-12-5-50/75 269A-12-5-25/50 269A-11-4-50/75

Figure 10. Representative sequences ofsilty lithologies from the lower part of Hole 269A.

734


lcm

1cm

1cm

Figure 11. Laminated deformed lithified silts of fades G, lower part of Hole 269A. Scale bar = 1 cm; B = boudinage; C = con-volute lamination; G = fault; L = load deformation, (a) 269A-13-2, 120-130 cm. Shows lamination; load deformation in-volving both several laminae, and individual silt laminae; synsedimentary normal faulting, (b) 269A-13-3, 101-109 cm.Shows lamination; boudinage; and load deformation as in (a), (c) 269A-7-4, 47-69 cm. Shows massive graded silt bed offades F; load deformation; boudinage and associated faulting, (d) 269A-7-3, 50-58 cm. Shows load casting and convolutelamination. Lower part of slab is mostly of fades H. (e) enlargement of part of (a), showing faulting and load casting onseveral scales.

735


In places, the primary depositional structure appearsa rather irregular lenticular discontinuous lamination.This appears a common feature of silts: it is also foundin facies F and has been previously figured by Piper(1972a, fig. 4A, 4C, lower parts).

A few beds grade finer upwards, but in most the rangeof grain size is too small to visibly recognize grading.

H) Clayey siltstone: Has an indistinct horizontallamination. Generally found in beds 1-10 cm thick,which are sometimes size graded. Facies H2 has slightmottling, but generally bioturbation mottling is absent.There appears to be a complete gradation between thisfacies and facies G. Load structures are not seen, andthe lamination has a much less-pronounced alternationof grain size compared with facies G.

I) Groups of silt laminae: 0.2-2 mm thick, alternatingwith silty claystone, such that the abundance, thickness,and grain size of silt laminae decrease upwards. Thislithology is the graded laminated bed of Piper (1972b).

This facies is similar to facies G except that load struc-tures are rare, grading is developed, and interlaminatedsilty-claystone tends to be commoner.

J) Single or randomly grouped silt laminae: Less than1 mm thick, in silty claystone or claystone. Usually noprimary structures are visible, but load structures are oc-casionally found.

PetrologyX-ray diffraction analysis of six clay and mud samples

(three from facies A, one from B, two from D) shows nosystematic differences in petrology between the clayfacies, although more data would be desirable.

Silts comprise mostly quartz, with subordinateamounts of feldspar, and 0.2%-2% heavy minerals,notably hornblende, garnet, chlorite, biotite, zircon,tourmaline, and opaque minerals.

BioturbationBurrows are visible in facies A2, C, D, E, and H2.

Few can be assigned to any distinctive trace fossil taxon.Chondrites occurs occasionally in facies A2 and H2(Figure 12a, c, d). Zoophycos is common in facies D, andexceptionally in other facies (Figure 12c).

Burrows, often subvertical, with diameters of severalmillimeters, are common in facies E and associated bedsof facies C and D (Figure 12b, d). They are filled withsandy muds, with sand contents of up to 20%. The grainsizes are consistent with concentration of grains fromfacies E. Sometimes, there remains only a thin biotur-bated remanent of facies E, with none of the originalbedding (Figure 13).

Facies RelationshipsPreservation of structures was sufficiently good in

Cores 8 through 13 of Hole 269A that the great majorityof lithologic contacts were preserved unbroken. Threehundred sixty unbroken contacts (with an additional 40broken contacts) between different lithologies werelogged from well-preserved sections and used to con-struct a transition matrix. A simplified version of thematrix, with minor lithologies omitted (to give 332 con-tacts), is shown in Table 2. The data quality does not

allow sophisticated mathematical analysis. A qualitativeinterpretation of the transition matrix is possible basedon Selley's (1970) method of identifying those tran-sitions between facies which occur much more frequent-ly or much less frequently than would be expected in arandom sequence.

Salient features of this transition matrix include: (1)Facies A and B are commonly associated with facies Iand J. (2) Unbioturbated clays (facies A and B) are rare-ly overlain by bioturbated facies C, D, and E . With theexception of E overlain by A, the converse is also true.(3) The distribution of facies overlying facies E is almostrandom, but facies D is unexpectedly common. (4)Facies E generally overlies facies C or D.

Certain gradational sequences of facies observeddirectly from the cores provide further information onfacies associations. Figure 10 shows selected short corelogs showing relationships between sandy and siltyfacies.

Interpretation of FaciesThe lack of microfossils makes interpretation of the

abyssal plain sediments in this lower part of Hole 269much more difficult than in most such sequences.

Most abyssal plains have thick beds of turbidite clay,characterized by high organic carbon contents and fewmicrofossils. Facies A and B claystones are thought tobe such turbidite clays. They are both found resting witha sharp base on highly bioturbated facies D and E, in-dicating an abrupt termination of bioturbational con-ditions. Coccoliths have been found (below the car-bonate compensation depth) in three beds, suggestingresedimentation from shallower depths. The colordifference between the two facies may reflect the degreeof oxidation of the source material. The maximumthickness of these clay beds cannot be determined.

At least some of the thin sand and silt beds in facies Fare also turbidites. Graded bedding is common; the beddivisions of the Bouma sequence and convolute lamina-tion are found; cross-lamination with no stoss side ero-sion is present. Foraminifera in one bed suggestresedimentation from shallower depths. The beds resem-ble those described by Piper (1973) from the Gulf ofAlaska: sorting and clay contents are similar. In contrastto the upper part of the hole, the only sand bed analyzedcontains about 10% clay.

There is, however, no evidence that many of the bedsare not winnowed deposits of contour currents. Raregradational bases of beds overlying facies G and Hcould be interpreted in this way.

The bioturbated sandy muds of facies E have a rangeof grain sizes similar to those of facies F, and they are in-terpreted as bioturbated equivalents. In a few cases,facies E grades down into facies F, and some beds of Egrade finer upwards. The sharp upper contact of E withother facies suggests exposure on the sea floor for aperiod of time and that the unbioturbated original bedwas not overlain by a thick mud. Although turbiditesands without an overlying mud are common, com-parable silts have not been described. Thus a bottom(contour) current origin for at least some of the parentunbioturbated sands and silts is favored. However,

736


km

Figure 12. Bioturbation in lower part of Hole 269A. (a) 269A-5-1, 23-38 cm. Fades A and fades F, with occasionalChondrites. Note ?flute mark filled with cross-laminated silt at 32 cm. (b) 269A-10-3, 40-55 cm. Long vertical burrowsfilled with sandy mud extending down from fades F. (c) 269A-6-3, 26-32 cm. Fades G and J, with Zoophycos andChondrites. (d) 269A-6-3, 24-26 cm. Fades D and G, with large burrow filled with sandy mud, and Chondrites. Noteerosional base to fades G silt bed, apparently filling small burrows, and the progressive overstep of laminae in the silt bed.

fades D (bioturbated mud) frequently overlies fades Eunexpectedly: in these cases there may have been biotur-bation of an original sand and silt to mud couplet. Thepresence of foraminifera in one bed, and coccoliths inthree, suggests resedimentation from shallower depths.Any sands or silts exposed on the sea floor for anylength of time would be much more susceptible to

bioturbation than muds, since larger benthos are com-moner in granular sediments.

Facies G shows many examples of synsedimentary in-stability, which is otherwise uncommon in the sequence.This instability is most simply interpreted as the result ofrapid deposition from suspension, with much pore watertrapped in poorly packed silts and clays. It is in-

737


Figure 13. Sketch from an X-radiograph of bioturbatedremnant of fades E sand, overlain by unbioturbatedclaystone of fades B. Scale bar is 1 cm. 269A-12-2, 124-131 cm.

conceivable that such abundant and varied defor-mational structures would be produced by thepredominantly fractional motion of contour currents. Aturbidity current could readily produce the requiredrapid deposition from suspension.

Facies H shows a completely gradational range ofappearance with facies G, but deformational structureshave not been seen. A turbidite origin is preferred.

Facies I is interpreted as turbidite solely by com-parison with the examples discussed by Piper (1972b). Intwo cases (10-2, 96-103 cm and 11-4, 58.5 cm) individuallaminae show load deformation, supporting a turbiditeorigin, by analogy with facies G.

Isolated or random sequences of silt laminae (FaciesJ) are also common in other abyssal plain sequences.They are usually sharp-based and continuous, andsometimes show load deformation. Comparing themwith the lower part of Site 268, they are probably notformed by contour currents. There are insufficient datafor a firm interpretation.

No interpretation is offered for the bioturbated claysand muds of facies C and D. They may be pelagic or"hemipelagic" clays, or they may be bioturbated tur-bidites.

Site 274Site 274 is situated on the continental rise off Victoria

Land and has an unexpectedly thin sedimentary se-quence above oceanic basement. This is probablybecause a major graben upslope has protected it fromdirect sedimentation from the continental shelf for mostof its history.

Only in Unit 3 are beds of sorted sand and silt found.The base of the unit is marked by a disconformity: theunit is of lower and middle Miocene age. A few of these

sand and coarse silt beds (e.g., 18-3, 47-49 cm) are dis-tinctly graded. The thicker beds, including a 3.3-meter-thick sand bed at the top of Core 18, have sharp bases,but do not appear graded. Grain-size analysis (Figure14; Table 3) shows an almost constant median grain sizethrough the beds, but higher percentages of fine silt andclay in the lower parts of the beds, where large Radio-laria are also concentrated. In other words, the meangrain size remains constant, but sorting improves, pass-ing up through a bed.

Much thinner silty laminae are also found in Unit 3.Thin sections of these show the median grain size is 5-60.The silt laminae occur in groups which closely resemblethe graded laminated beds described by Piper (1972a,1972b). The silt laminae alternate with mud laminae inbeds 0.3-2 cm thick. Each bed has a sharp base and anupward decrease in the thickness, abundance, and grainsize of the silty laminae. The tops of some of the siltylaminae are very sharp. In one sample (274-18-4, 48-50cm) there is a cross-laminated silt at the base of such abed.

ICE-RAFTING OF SEDIMENT

IntroductionIce-rafting of sediment is manifestly an important

depositional process around Antarctica. Large numbersof dispersed pebbles were found in Leg 28 holes, some indeep-water pelagic sedimentary sequences. A represen-tative sample of a pebble population is rarely obtainedin DSDP cores. Relatively small samples, however, aresuitable for studying the sand fraction of the sediments.Sand has a less obvious ice-rafted origin than pebbles.Abundant sand in pelagic sediment sequences andcoarse sand in "hemipelagic" sequences, in the absenceof distinct current-concentrated sand beds, are inter-preted as being ice-rafted in origin. Specific cores wherethis interpretation is critical are discussed in detailbelow.

MethodsThe methods used are based on those described by

von Huene et al. (1973). Channel samples were taken bycutting out a narrow groove of sediment with a spatulafrom a 50-cm length of core. Disturbed lengths of corewere avoided. Sample size was around 10 g of dry sedi-ment.

Sediments were disaggregated with hydrogen perox-ide and Calgon and ultrasonic treatment wherenecessary. They were then wet sieved through a 20(250µ) and 40 (63µ) screen to determine the amounts of"coarse" and "fine" sand. The silt and clay fraction wasdetermined by conventional pipette analysis. Anypebbles (>4 mm) were removed by hand.

The sand fractions were dried and weighed and theproportion of detrital grains estimated. Many samplesin addition contained larger diatoms, Radiolaria, ferro-manganese nodules, and sometimes undispersed clayaggregates. Where very few detrital grains were present,they were individually counted while still on the sieve,using a binocular microscope. Sieves were cleaned in an

738


TABLE 2Transition Matrix Between the Principal Fades

in the Lower Part of Hole 269

Lower

Bed

A

Observed Transitions

Upper Bed

C D E F G

0

1

24

7

109215

50

3

2

04

10

12

9

82674

21

01

2

02

2

1

11

2

2

2

4

5

31

0322

1

3

7

9

01

0

0

021

8

16

1

34

03

02

37

11

10

01

0

2

0

02

26

7

10

0

0

1

4

0

0224

3

3

0

01

6

01

014

17

27

1

1

0

6

01

0

53

54

72

14

16

19

42

2626

12

51

332

Number of Transitions Predicted if Random

Upper Bed

ABCD

Lower E

Bed F

G

HIJ

A

10.8

2.1

2.4

2.9

6.3

3.9

3.9

1.8

7.7

B

12.0

3.1

3.6

4.2

9.4

5.8

5.8

2.711.4

C

1.82.4

0.5

0.61.4

0.9

0.9

0.4

1.7

D

3.6

4.8

0.9

1.3

2.8

1.8

1.8

0.8

3.3

E

3.4

4.6

0.9

1.0

2.7

1.6

1.6

0.8

3.2

F

6.0

8.0

1.6

1.8

2.1

2.9

2.9

1.3

5.7

G

4.2

5.6

1.1

1.3

1.5

3.3

2.0

0.9

4.0

H

3.9

5.2

1.0

1.21.4

3.0

1.9

0.7

3.7

I

2.3

3.0

0.6

0.7

0.8

1.8

1.1

1.1

2.7

J

8.6

11.5

2.2

2.6

3.0

6.7

4.2

4.2

1.9

ultrasonic bath and checked under a binocular micro-scope after every sample when analyzing those sampleswith low detrital grain contents. (In some sand-rich se-quences and for some preliminary analyses, cleaningwas less rigorous.) Some carbonate ooze samples weretreated with acetic acid, sieved, and the sand grainscounted. It was found to be very difficult to recognizesmall quantities of sand in a diatom-rich sediment.

Evidence of Earliest Ice RaftingSite 274

This site has an unusually complete stratigraphicsequence. It is situated on the continental rise, but atmost times was protected from downslope sedimenta-tion from the continent by a structural trough. Most ofthe sediment has a high biogenous component.

In the upper 113 meters (Quaternary to upperMiocene) coarse sand and pebbles are common. Belowthis, to 180 meters (middle and lower Miocene), most ofthe sediment is highly biogenous, but contains between1% and 0.01% sand, including a few grains of coarse

sand. Except at a few restricted horizons (e.g., Core 18,Section 2), there is no evidence of tractional emplace-ment of the sand by currents. The thick diamictites ofthe Ross Sea, only 500 km to the south, are timeequivalents of this sequence.

From 180 to 225 meters (middle Oligocene to upperlower Oligocene) extremely rare sand grains are found inthe sediment. None have been found below this. Thegrains are too small to show diagnostic surface texturesunder the scanning electron microscope. The Ross Seaholes do not indicate glaciation at the time this sequenceaccumulated. Such very small amounts of sand couldwell have been emplaced by processes other than ice-rafting.

Site 269

Granules and pebbles are exceptionally found in theuppermost two cores (Plio-Pleistocene) of this site.Because of the abundance of trubidites at this site, sanddistribution cannot be used as evidence of ice-rafting.

739


991

gr;

0.1

0.01

GRAIN SIZE Φ

Figure 14. Grain size analyses of turbidite silt beds fromSite 274. (a) Near base, and from upper part of 2.7-meter thick sand bed in Core 18 (base at 274-18-2, 120).Compare with Table 3. (b) Samples from middle andlower part of a 40-cm-thick silt bed, grading up into 50cm alternating silt and clay (base at 274-16-2, 146).

TABLE 3Grain Size Analyses From a Single

Bed at Site 278

Sample(Interval in cm)

18-1, 120-12218-1, 145-14718-2, 20-2218-2, 50-5218-2, 80-8218-2, 110-112

Percent Finer Than20

0.030.020.030.010.270.14

40

64.161.762.865.067.164.0

60

94.693.493.193.891.791.3

80

96.395.595.996.295.396.8

Site 268

This is a normal continental rise site, so that bottomtransport of coarse sand is not unlikely. Induration andchertification hinders analysis. It is difficult to obtainuncontaminated soft sediment from short cores whichconsist mostly of short blocks of indurated sediment.

Ice-rafting is well established down to Core 10 (mid-dle lower Miocene), with high percentages of dispersed

sand. Below this, a little sand is still found in abun-dances comparable with those at Wite 274 between 180and 225 meters: its origin is uncertain. In Core 14 thereis a single granite pebble: this could be downhole con-tamination, as in Cores 22-34 at Site 274. A chert inCore 19 (pre-middle Oligocene) has an irregular layer ofgranules and coarse sand grains. These are not suf-ficiently dispersed to be considered as ice-rafted, norsufficiently bedded to be due to bottom or turbiditycurrents. Their origin is uncertain.

Site 267

Ice-rafted sand and pebbles are found in the Plioceneand Quaternary sequence at Site 267. The Miocenesediments are generally rich in siliceous organisms andno sand was detected. Acid residues of upper Oligocenecarbonate ooze from Core 4 contained a few sandgrains, which may indicate ice-rafting.

Site 266

This is a fully pelagic site, well away from the abyssalplain. Small amounts of detrital sand are found in manysamples down to 157 meters (upper Miocene), althoughsome samples are barren. Below this, acid residues of allbut one of the samples analyzed contained no sand. Thethree quartz sand grains in Core 20, Section 4 may bedue to contamination.

Summary

In the Ross Sea sector, the data from Site 274substantiate the evidence from the Ross Sea sites thatsignificant glaciation was initiated in the late Oligocene.The significance of trace amounts of sand in the earlyOligocene is unclear.

In the Wilkes Land sector, there is unequivocalevidence of early Miocene, and possible evidence of lateOligocene ice-rafting. Earlier minor ice-rafting cannotbe ruled out.

Grain-Size Distribution of Ice-Rafted Sediment

Grain-size distribution of ice-rafted sand and largerclasts is very variable, even at any one site. This varia-tion is best documented at Sites 274 and 270 (Barrett,Chapter 22). In the grain size analyses in Table 4, thevariation can be recognized from differences in the ratioof coarse to fine sand. (However, this statistic isprobably unreliable in siliceous oozes with low detritalgrain contents. It is much easier to see a coarse sandgrain in an otherwise empty 20 sieve than a fine sandgrain in a 40 sieve full of diatoms.)

Barrett (Textural Characteristics of CenozoicPreglacial and Glacial Sediments, this volume) hasshown that grain-size variations at Site 270 do notappear to be the result of variations in the petrology ofsource material; and that diamictites both with high andwith low proportions of coarse sand and pebbles areequally poorly sorted. The pebble-rich diamictites can-not be the result of sea-floor winnowing of granule-poordiamictites.

At Site 274, Cores 10-12 are unusually rich in coarsesand. Above Core 10, there is a spectrum of coarse tofine sand ratios, but it is generally lower than in Cores

740


TABLE 4Size Analyses of Channel Samples From Sites 266, 267, 268, 269, 272, and 274


Site 266

1-3,100-1502-2, 90-923-1, 100-1504-4, 50-1005-6,50-1006-5, 50-1007-5, 50-1008-6, 50-10094, 70-15010-6, 50-10011-7,50-10012-3, 0-4013-3, 50-10014-4, 0-5015-2, 0-5016-2,100-15017-5, 0-5020-4, 52-5422-2, 50-52Site 267

1-3, 90-1503-3, 50-1004-1, 50-1004-3, 50-1001A-1, 50-1001A-1, 100-1501A-2, 0-501A-2, 50-1001A-2, 100-1501A-3, 0-501A-3, 50-1001A-3, 100-1501A4, 0-501A-4, 50-1001A4,100-1501A-5, 0-501A-5, 50-1001A-5, 100-1501A-6, 10-501A-6, 50-1001A-6, 100-1503A-2, 100-1501B-1, 45-902B-1, 110-1503B-2,113-1504B-3, 59-1505B-6, 50-1001-2, 90-1001-3, 0-462-1,43-1002-4, 72-902-5, 50-1012-6, 60-1003-1, 60-1003-2, 65-1004-2, 50-1005-1,45-955-2, 50-1007-2, 50-1008-2, 55-968-3,60-1109-1,60-10010-2, 50-10012-1,100-15017-3, 95-15018-2,70-10018-3, 85-115

Depth(m)

4.0-4.531.9-31.9245.0^5.571.0-71.590.0-90.5

107.5-108.0128.0-128.5137.5-138.0147.2-148.0156.5-157.0166.0-166.5184.5-184.9199.5-200.0219.5-200.0235.5-236.0246.0-246.5260.5-261.0318.5363.0

3.5-4.093.5-94.0

128.5-129.0131.5-132.0

4.5-5.05.0-5.55.5-6.06.0-6.56.5-7.07.0-7.57.5-8.08.0-8.58.5-9.09.0-9.59.5-10.0

10.0-10.510.5-11.011.0-11.511.6-12.012.0-12.512.5-13.070.0-70.5

113.5-114.0142.5-143.0152.0-152.5171.0-171.5189.5-190.0

2.4-2.53.0-3.5

28.43-29.033.2-33.934.5-35.136.1-36.557.1-57.558.7-59.087.0-87.5

113.9-114.4115.5-116.0146.0-146.5172.5-172.9174.1-174.6199.6-200.0229.5-230.0285.5-286.0388.5-389.0410.2410.5

411.85-412.15

Clay(%)

56.3554.1248.7557.2279.8475.0918.4932.0724.1878.7687.2114.84

28.53

59.88

68.9563.95

53.56

50.64

50.7347.9162.0975.2779.2268.24

31.56

10.09

52.2267.0576.7062.9662.9657.0967.0064.5461.1069.0079.5880.4773.04

51.01

62.3470.37

98.7597.2098.86

96.81

97.9798.6698.85

98.98

99.4899.9697.6999.55

98.07

99.89

Süt(%)

41.3436.5549.4842.0819.0123.1180.0066.4774.7120.7612.5384.91

71.14

29.85

30.0834.28

44.49

42.73

47.7750.2234.9423.0018.4921.16

68.37

87.60

16.7031.2623.0019.7634.9732.6526.3125.1616.5125.8019.8018.3320.37

48.87

37.5029.50

Sand(%)

2.319.331.770.701.151.791.511.461.110.480.250.25

0.33

10.27

1.371.771.252.801.141.953.196.642.031.341.151.501.862.961.732.29

10.601.020.070.520.042.310.452.311.93

21.081.690.303.332.07

10.266.69

10.3022.395.200.621.206.600.110.11

0.150.12

Detrital Sand(%)>2Φ

tr0.0

4 grains3 grains6 grains

0.00.02

5 grains0.03tr?0.00.00.00.00.00.00.0

1 grain0.0

1.230.0

2 grains0.00.050.130.191.320.580.210.190.210.670.220.400.080.141.600.180.052.580.0

5 grains0.100.00.00.00.790.80

24.410.750.141.590.785.884.773.538.371.330.180.822.460.050.020.00.00.0

Detrital Sand(%) 2-40

0.040.00.0

1 grain0.020.020.00.00.00.020.00.00.00.00.00.00.0

2 grains0.0

0.200.00.0

7 grains0.040.720.380.200.160.300.181.120.210.100.120.120.020.020.210.040.080.0

1 grain0.070.00.00.00.350.953.000.370.141.211.284.281.140.63

12.161.680.080.042.730.030.020.00.01

tr

741



Site 269

1-3, 100-150Site 272

1-1,80-1101-2, 50-1001-3, 50-1001-4, 50-1001-5, 50-1002-2, 50-1002-3, 50-1002-4, 50-1002-5, 50-1003-1, 50-100Site 274

1-2, 50-1001-3, 50-1001-4, 50-1001-6, 50-1002-2, 50-1002-3, 50-1002-4, 50-1002-5, 50-1002-6, 50-1003-2, 50-1003-3, 50-1003-4, 50-1003-5, 50-1003-6, 50-1004-1, 50-1004-2, 50-1005-2, 50-1006-1, 50-1006-2, 50-1006-3, 50-1006-4, 50-1006-5, 50-1006-6, 50-1007-1, 50-1007-2, 50-1009-1,50-1009-2, 50-1009-3, 50-1009-4, 0-509-4, 50-1009-5, 50-1009-6, 0-509-6, 50-10010-2, 50-10010-2,100-15010-3, 50-10010-4, 50-10010-5, 0-5010-5, 50-10010-6, 50-10011-2,50-10011-3, 50-10011-4, 0-5011-4,50-10012-2, 50-10012-3, 50-10012-4, 50-10012-5, 50-10012-6, 50-10013-1,40-10013-2, 50-10013-3, 50-10013-4, 50-10013-5, 50-10013-6, 50-100

Depth(m)

4.0-4.5

4.8-5.16.0-6.57.5-8.09.0-9.5

10.5-11.017.0-17.518.5-19.020.0-20.521.5-22.026.5-27.0

2.0-2.53.5-4.05.0-5.58.0-8.5

11.5-12.013.0-13.514.5-15.016.0-16.517.5-18.021.0-21.522.5-23.024.0-24.525.5-26.027.0-27.535.5-36.037.0-37.544.545.048.0-48.549.5-50.051.0-51.552.5-53.054.0-54.555.5-56.063.5-64.065.0-65.576.5-77.078.0-78.579.5-80.080.5-81.081.0-81.582.5-83.083.5-84.084.0-84.587.5-88.088.0-88.589.0-89.590.5-91.091.5-92.092.0-92.593.5-94.0

100.0-100.5101.5-102.0102.5-103.0103.0-103.5106.5-107.0108.0-108.5109.5-110.0111.0-111.5112.5-113.0114.4-115.0116.0-116.5117.5-118.0119.0-119.5120.5-121.0122.0-122.5

Clay(%)

52.12

62.5723.7837.4740.2541.2331.8234.4432.7731.6439.74

56.75

63.5354.2666.0148.3261.1657.7560.4858.7962.8365.7863.4460.6170.2473.4961.5370.1360.8461.8559.6964.3867.7873.3764.7757.9071.8160.08

70.9172.64

65.4766.01

68.8371.80

71.1369.58

69.68

64.7554.7557.4455.7452.6028.86

62.6866.3664.5869.56

TABLE 4 -

99.07

95.70

93.95

90.03

91.10

81.80

93.75

98.44

99.92

Continued

Silt(%)

44.07

29.1328.5328.4927.3428.0025.1823.7025.5624.9532.32

39.68

34.1040.2332.1950.2735.4935.1238.7140.0536.7132.0834.5734.2927.9625.4232.9928.3438.1836.7239.4231.9030.7625.8728.6130.1224.6329.03

25.0022.94

23.2524.58

22.4720.82

20.0523.00

22.32

27.3530.6028.3331.1031.5853.58

33.4132.4933.3430.39

Sand(%)

3.80

8.2957.6934.0532.4130.7642.9941.8541.6743.4127.95

3.571.012.375.521.801.413.357.140.811.160.472.141.995.091.801.095.471.530.981.430.903.721.451.766.62

11.993.56

10.894.304.094.426.05

11.289.419.978.717.388.908.827.43

18.208.006.257.91

14.6514.2213.1715.8217.56

1.563.901.102.090.050.08

Detrital Sand(%) >2</>

0.21

5.3327.1017.6314.2715.1421.8422.5217.2326.07

8.52

0.340.190.602.660.490.270.910.220-100.610.051.270.732.490.730.412.720.630.200.220.201.720.870.702.654.611.734.622.101.452.102.685.854.304.733.673.533.792.641.997.272.282.472.535.306.735.496.596.05

1 grain2 grains3 grains3 grains3 grains

tr

Detrital Sand(%) 2-40

0.08(?)

2.9613.8111.5914.5715.3213.7515.3917.0216.5219.42

0.240.550.751.720.440.140.240.090.140.140.060.571.082.600.860.671.930.810.390.200.131.990.570.192.634.401.635.612.02.112.272.655.161.001.041.990.973.961.811.783.240.420.382.313.934.424.595.536.830.040.040.01.630.0010.02

742


TABLE 4 - Continued


Depth(m)

Site 274 - Continued

14-1, 50-10014-2, 50-10014-3, 50-10014-4, 50-10014-5, 50-10014-6, 50-10015-2, 50-10015-3, 100-15015-4, 50-10015-5, 50-10015-6, 50-10016-3, 50-10017-2, 50-10017-3, 50-10018-5, 50-10018-5,100-15019-2, 50-10019-6, 50-10020-5, 50-10021-6, 50-10021-6, 100-15022-5, 0-5022-5, 100-15023-5, 50-10023-6, 50-10024-5, 0-5024-5, 50-10025-6, 50-10026-6, 50-100

26-6,100-15027-4, 50-10028-6, 50-10029-3, 50-10030-6, 0-5030-6,50-10030-6, 100-15031-6, 50-10032-6, 50-10032-6,100-15033-6, 50-10034-6, 50-100

124.0-124.5125.5-126.0127.0-127.5128.5-129.0130.0-130.5131.5-132.0135.0-135.5137.0-137.5138.0-138.5139.5-140.0141.0-141.5151.0-151.5158.5-159.0160.0-160.5169.5-170.0170.0-170.5173.0-173.5179.0-179.5187.0-187.5198.0-198.5198.5-199.0207.0-207.5208.0-208.5215.5-216.0217.0-217.5224.5-225.0225.0-225.5236.0-236.5245.5-246.0

246.0-246.5255.5-256.0264.5-265.0274.0-274.5283.0-283.5283.5-284.0284.0-284.5293.0-293.5302.5-303.0303.0-303.5312.0-312.5321.5-322.0

Clay

(%)

66.2671.85

72.9970.7461.20

52.14

73.9155.10

74.27

12.829.25

19.738.257.29

15.928.33

6.208.308.88

11.446.81

8.1114.2112.865.716.37

98.99

99.07

99.99

99.9899.5299.7899.8499.9199.93

99.95

99.7999.37

99.89

99.34

Silt

(%)

33.6928.14

26.9628.9838.69

47.73

26.0342.51

25.52

86.9590.25

89.3791.5292.5282.8991.11

93.4191.3590.7088.0692.77

91.6785.3686.8093.7393.13

Sand(%)

1.010.040.020.030.050.280.110.010.13

0.062.400.020.480.220.160.090.070.210.050.230.510.210.630.900.230.191.190.56

0.110.390.170.430.500.420.660.220.430.330.560.50

Detrital Sand(%) >2Φ

1 grain0.00.00.00.0

1 graintr

0.00.0050.00.0

1 grain0.00.0

1 grain2 grains1 grain

0.00.00.00.00.00.00.00.00.00.00.0

2 grainsvolcanicglass

0.00.00.00.00.0

3 grainsa

0.00.0

2 grainsa

0.00.00.0

Detrital Sand(%)2-40

0.010.010.010.0030.010.030.04

tr0.01

0.020.710.020.450.150.130.080.06

trtr

4 grains3 grains

0.0tr

0.03 grains

tr0.0

1 grainvolcanicglass

0.00.00.00.0

tr0.00.00.00.00.00.00.0

Note: Determined by conventional wet-sieve and pipette techniques. Percent sand includes gravel-sizematerial and sand-sized, detrital, biogenic, and authigenic components; the proportion of detritalmaterial in the sand fraction (finer than 4 mm) has been estimated using a binocular microscope,tr = trace.

aEarly analysis in which contamination was less carefully avoided - believed due to contamination.

10-12. As at Site 270, sediments with both a high and alow coarse to fine sand ratio are equally poorly sorted.Grain-size analyses (Figure 15) show no evidence ofwinnowing. Neither is there evidence to suggestpetrologic control of grain-size distribution.

There are several possible explanations for this type ofvariation. One is a control of available grain size bypetrology and weathering of source material that is sosubtle that it cannot be recognized from an examinationof the petrology of the sedimentary sequence. Variationsbetween sites may largely result from differences insource material.

Winnowing at or near the depositional site seems im-probable from the grain-size data. However, the sedi-

ment could result from mixing of two different types ofpoorly sorted glacial debris in different proportions, onewith a high content of sand and pebbles, the other richin silt and clay. So long as both parent sediments werepoorly sorted, it would be very difficult to distinguishdistinct modes resulting from mixing. Two such parentsediments might be englacial material, little modified bythe action of melt water, and till, which has had somefine material winnowed out by running water duringablation (probably while being transported in a supra-glacial position).

The proportion of clay-rich till to sand-rich till willtend to vary from berg to berg. Mixing on the sea floorwill result from successive dumping of material by

743


TABLE 5Core Sections Containing Micionodules at

Sites 266, 267, 268, 272, and 274

Figure 15. Gram s/ze analyses of sediments in upper partof Site 274. Sediment mostly emplaced by ice rafting,but in part biogenic. Percentage ofbiogenic components(estimated from smear slides) is indicated.

icebergs carrying varying amounts of these two parentsediments.

Rates of Ice-Rafted SedimentationBecause of discontinuous coring and poor paleon-

tologic control, it is not possible to make any more thangeneral estimates of the rate of ice-rafted sedimentation.It is difficult to distinguish ice-rafted silt and clay, andthe very small amounts of coarse sand and pebbles inmany cores may not be representative samples. InFigure 16, we have attempted to show variationsthrough time in the rate of sedimentation of ice-rafted20-40 sand. Note that this grain size does not form aconstant proportion of the total ice-rafted sediment.Furthermore, it could be argued that the biostrati-graphic control is inadequate for this type of compila-tion. We have found that the several versions of thisfigure that were produced, as biostratigraphic correla-tion was modified and refined, have varied substantiallyin detail, but have still showed the same general trends.An equally serious problem is the lack of recovered corein some critical intervals. So the figure must be regardedas highly speculative.

In both the Wilkes Land and Ross Sea sectors, thehighest rate is found in Gauss times. Rather lower ratesare found in younger sediments. Our dating is not suf-ficiently refined to detect the type of variations seen byFillon (1972). Gilbert (early Pliocene) rates at Sites 266and 267 appear low. Dating is poor at Site 268; theabundance of ice-rafted sand decreases downwards inthe Gauss-Gilbert interval but there is no informationon overall rates of sedimentation. At Site 274, the Gaussrate may have been underestimated, or Gauss sedimentsmay not have been recovered. Note that the core catcher


Site 266

3-14 45-68,CC9-310-611-6

Abundance Type Age

T s

TF** 1FF >

, Pleistocenex Plio-Pleistocene)

} Dark > Pliocene

Ji ' Upper MioceneMiddle Miocene

(No routine examination below Core 12)

Site 267A

1-1,50-1001-1, 100-1501-2, 0-501-2, 50-1001-2, 100-1501-3,0-501-3, 50-1001-3, 100-15014, 0-5014, 50-1001-4, 100-1501-5,0-501-5, 50-1001-5, 100-1501-6, 0-101-6, 10-501-6, 50-1001-6, 100-150

Site 268

1-21-32-63-13-2

TTFFFFMFFMMMMM

> Dark Plio-Pleistocene

Manganese-oxide-coated pebble bedF 1T 'M J

FMFFT

" / Pliocene

Light / Pleistocene

(No routine examination below Core 12)

> Light i Plio-Pleistocene

Site 272

1-5 F2-2 F2-5 F

(No routine examination below Core 3 Section 1)

Site 274

10-2 through Pliocene to12-3 F Dark middle Miocene

Macronodules from 10-5 through 12-6

Note: T = trace, F = few (1-5% of sand fraction); M = many(>5% of sand fraction); * in smear slides, but not inadjacent sand fractions. Based on examination of sandfractions of all grain-size analyses reported in thischapter.

of Core 8 (the only material recovered), of probableGauss age, contained unusually pebbly sediment. Thesudden appearance in Site 274, Core 12 of abundant ice-rafted sediment above middle Miocene strata with littleice-rafting is exaggerated by the very low overall rate ofsedimentation in the late Miocene. The early Miocenerate of ice-rafted sedimentation is much higher than thatin the middle Miocene. The middle and late Miocene at

744

0-1 1AGE m.y.depth in ' L

0 i meters


RATE OF ICE RAFTED SEDIMENTATION OF 2 to 4<j> SAND2 6 6 267 268 272 274

cm/m.y.10100 0-1 100 0-1 100 Q71 < t 100 0-1 100

-oo

l

3-

4-

5-note scale

change

10-

15-

20-

78-

148.5 J

4-

11.5

CD

CO

I %100-

^160

5.5

<34O

23.51- ? — ? J 85.5-

Figure 16. Rates of accumulation of ice rafted 2 to 40 sand, in cm/m.y. Based on core thicknesses, abundance of sand(Table 4), and available biostratigraphic data. The biostratigraphic summary shown in Figure 3 of Chapter 1 (Introduction)has been used, with the following exceptions: Site 266: the assignment of the base of the Gauss is speculative. Site 267:dating of Core 1A (4-13 m) as Gauss disconformable on upper Gilbert is speculative, partly based on Radiolaria, andsilicoflagellates (Ciesielski, personal communication). Hole 272: recognition of Brunhes part of sequence based on litho-logic correlation with the cores of Fillon (1972) (see Barrett, Chapter 22). Note the Brunhes thickness is very uncertain.Hole 274: Position of top of Gauss uncertain; it may be lower than shown, in which case the Gauss might not be repre-sented by recovered core. Detailed dating of Gilbert based mainly on silicoflagellates (Ciesielski, personal communica-tion). Base of the upper Miocene may have been picked too low, so that the higher upper Miocene sedimentation rate mayalso apply to the topmost middle Miocene.

745


TABLE 6Electron Microprobe Analyses of Micronodules


274-10-6and

274-11-2

266-11-6

272-2-2268-2-1

Mn

29.626.429.230.928.428.259.462.90.150.04

Percent Weights of Elements Expressed as OxidesFe

1.91.41.51.92.32.0

0.50.3

0.923.9

Ti

0.71.11.01.21.01.2

0.40.6

1.10.1

Ca

1.90.60.70.90.60.9

0.91.0

0.40.1

Si

29.022.118.410.319.026.9

3.31.0

70.549.5

Mg

3.22.42.02.11.11.4

3.03.1

0.83.6

Ni

n.d.a

n.d.n.d.n.d.n.d.n.d.0.10.2

n.d.0.00

Cu

n.d.n.d.n.d.n.d.n.d.n.d.0.40.6

n.d.0.03

Al

n.d.n.d.n.d.n.d.n.d.n.d.n.d.n.d.n.d.n.d.

TotalAccounted

for

66.253.952.747.252.460.668.1

85.577.3

n.d. = element not determined after initial scan suggested very low concentration.

Site 268 were not recovered and this part of the sectionmust be quite thin at this site. So little sand was found inthe Miocene at Sites 267 and 266 that rate calculationsare meaningless.

Fillon (1972) found abundant ice-rafted sand inGauss and early Matuyama sediments in the Ross Seaand the area to the north around Site 274. Abundanceswere much lower in the late Matuyama, and in-termediate amounts were found in Brunhes sediments.A correlation of Site 274 with Fillon's data suggests thatthe middle of Core 3 is mid-Matuyama. Note that theice-rafted debris in the lower part of Core 2 and the up-per part of core 3 (i.e., upper Matuyama by this cor-relation) has a high coarse- to fine-sand ratio.

The rate of ice-rafted sedimentation is closely relatedto distance offshore. The highest rates occur at Sites 274and 268 on the continental rise and fall off rapidlynorthwards through Sites 267 and 266.

Rates at Site 269 are difficult to determine because ofrapid turbidite sedimentation, but are comparable withthose at Site 266. This may indicate that the rate ofsupply is dependent on proximity to a major source ofbergs. Sites 266 and 268 lie a little way downcurrentfrom the Shackleton Ice Shelf; there is no major sourceof bergs immediately to the west of Site 269. Site 274would be fed by Ross Ice Shelf bergs. If this interpreta-tion is correct, it negates Warnke's (1970) suggestionthat bergs depositing sediment travel great distances inthe circumpolar current.

Rates will also depend on the preferred paths oficebergs, controlled by pack ice distribution and stormtracks. The effect of this on local variations in the rate ofice-rafted sedimentation is difficult to evaluate, but itseems inadequate to explain an overall very low rate ofice-rafting for tens of millions of years.

DETRITAL PETROLOGY

IntroductionThis summary of the detrital sand petrology at Sites

267, 268, 269, and 274 is based on the following data:1) Binocular microscope examination of about 120

sand fraction samples.2) Petrologic microscope examination of about 20

grain mounts of sand fraction samples.

3) Petrologic microscope examination of nine thin-sectioned light mineral grain mounts in the size range2.5-40 (Table 7).

4) Ten heavy mineral separations, size range 2.5-40,separated with tetrabromoethane (Table 8).

5) About 20 thin sections of impregnated or lithifiedsediments.

6) Incidental data from shipboard smear slidedescriptions by Ford (Site 267), Kemp (Site 268), andBarrett (Site 269).

Our data are generally consistent with those given byBarrett (Characteristics of Pebbles, this volume) for peb-ble lithology, and by Payne and Conolly (1972) for tur-bidite sands on the Wilkes Abyssal Plain.

Sand-size sediment at the sites investigated isprobably transported in three ways: (a) ice rafting, ac-tive at all sites, but least important at Site 269; (b)transport by turbidity currents, especially at Site 269; (c)transport by bottom currents, especially at Site 268.

Sand PetrologyTurbidites from Site 274 show a strong influence of

the Victoria Land basalts. The plagioclase to potassiumfeldspar ratio is higher than in the Wilkes Land sector(Table 7; Cook et al., X-Ray Mineralogy, this volume)and pyroxene is the dominant heavy mineral. In con-trast, the only ice-rafted sample studied in detail wasrich in hornblende, biotite, and opaques, and relativelypoor in pyroxenes and garnet. This lithology is consis-tent with the lithology of the ice-rafted pebbles found atthis site (Barrett, Characteristics of Pebbles, thisvolume). The quartz: feldspar ratio in the ice-raftedsands from Site 274 is very variable.

Site 269 has also received some sediment from abasaltic source. Pyroxene is the dominant heavymineral, but hornblende and garnet are (as at Site 268)also present in quantity. The X-ray diffraction data ofCook et al. (this volume) show plagioclase to potassiumfeldspar ratios similar to those of Site 274, although ourlimited data do not confirm this.

Garnet, hornblende, and opaque minerals are theprincipal heavy minerals at Site 268, consistent with theice-rafted lithologies. Cook et al. found subequalamounts of potassium feldspar and plagioclase.

746


TABLE 7Light Mineral Petrology of Selected Samples From Sites 268, 269, and 274 (in percent)


268-2-2, 109-126268-5-2, 53-59

269-1-6, 41-52269-3-3, 15-20269A-8-2, 84-117

274-12-5, 100-6-50274-16-2, 130-145274-18-1, 120-122174-18-2, 20-22

<D ^ 3

50.964.6

62.050.050.9

38.946.154.557.4

Quartz

a >.

_0.9

1.06.31.9

1.1_1.80.9

17.06.0

16.023.223.6

18.98.9

13.610.2

2

6.64.3

6.04.55.7

3.34.06.75.6

Feldspar•*^>-*•

8uo

"f17.98.6

1.07.11.9

4.59.85.5

10.2

•ö

<D

<D

3.86.0

7.02.73.8

15.66.95.57.4

αCD

RJ

1f§

3.86.0

3.02.71.9

12.211.8

6.73.7

1

_-_—-_1.0

0.9-

<D

"S

ö_-_

0.9-_

4.91.81.9

o

D

_3.4

4.02.71.0

5.66.93.62.8

c

cβ

<D "5

II<D i-.

OOPL,

TUTU

TUTUTU

IRTUTUTU

Note: TU = turbidity current; IR = ice-rafting. Light mineral 2-40 fraction impregnated in epoxy resinand thin sectioned. At least 100 grains counted.

TABLE 8Heavy Mineral Petrology of Selected Samples From Sites 268, 269, and 274 (in percent)


268-2-2,119-125268-5-2, 53-59

269-1-6,41-52269A-8-2, 84-87 & 113-117

274-12-5 & 6274-18-1, 120-122274-18-1, 145-147274-18-2,20-22274-18-2, 80-82274-18-2, 110-112

Gar

net

17.26.8

14.312.5

2.22.43.33.02.54.1

Oliv

ine

0.7-

0.4-

4.5_

2.90.42.50.8

Gre

enH

ornb

lend

e

10.423.5

24.212.5

21.06.38.39.89.3

11.0

Bro

wn

Hor

nble

nde

1.52.2

2.6-

1.9_

0.40.40.42.0

Cli

nop

yrox

ene

1.514.2

25.338.7

9.041.152.946.838.641.9

Hyp

erst

hene

_

0.6

0.41.2

0.4_

1.51.70.8

En

stat

ite

_

1.9

5.66.4

1.56.75.04.29.88.1

Zoi

site

1.10.6

1.12.5

3.42.84.11.52.1-

Epi

dote

1.92.6

0.40.4

0.41.20.40.40.80.4

Chl

orit

e

3.712.6

3.01.7

1.12.80.81.94.72.4

Bio

tite

3.08.4

1.50.4

11.21.61.73.41.71.6

Roc

k F

ragm

ents

-

-

1.90.40.41.5_

0.4

Alt

ered

Gra

ins

7.85.5

2.63.0

13.913.4

8.710.2

8.510.6

Opa

ques

49.315.8

11.311.2

20.67.95.48.36.88.1

Hea

vyM

iner

als

5.83.2

2.33.1

2.43.43.25.24.53.7

Sed

imen

tati

onP

roce

ss

TUTU

TUTU

IRTUTUTUTUTU

Note: TU = turbidity current; IR = ice-rafting. About 300 grains counted in 2V2 - AΦ size range.

Magnetite is very common in the upper part of the hole,where mafic plutonic rocks make up about 10% of thepebbles. Well-rounded quartz grains are common andcorrespond to those found in pebbles of medium- tofine-grain sandstone at this site (Barrett, Characteristicsof Pebbles, this volume). Similar quartz grains are foundat Site 267 (where the concentration of sand is insuffi-cient for heavy mineral analysis). Our heavy mineraldata in this sector differ from those of Payne and Conol-ly (1972) in that we find more garnet and less biotitethan they do.

Clay Mineralogy

Clay mineral abundances have been replotted inFigure 17, with mica normalized to a value of 1.0, fromthe <2µ X-ray diffraction data reported by Cook et al.(X-Ray Mineralogy, this volume).

Kaolinite is found in the Pliocene-Quaternary of allthree sites, but except for the upper Miocene at Site 274,is absent in older sediments. The kaolinite is presumablydetrital from kaolinite-bearing rocks on the Antarcticcontinent. Terrigenous sequences (such as the BeaconGroup) are frequently rich in kaolinite, although Barrett(1966) did not detect its presence in Beacon rocks that heexamined.

Montmorillonite contents are highest in the lowerpart (Oligocene to early Miocene) of Site 269 (20%-30%;0.7-3.0 normalized to mica = 1); in the upper part of thesection, they decrease to 9%-15% (0.3-0.6 normalized tomica). Montmorillonite is abundant at Site 274, es-pecially in the lower Pliocene and Miocene parts of thesection (0.9-1.3 normalized to mica). Above and belowthis the normalized content ranges from 0.3 to 0.7, ex-cept for a montmorillonite-rich sample near the base of

747


the hole. At Site 268, montmorillonite contents are low(0-0.3 normalized to mica) and show no general trends.There appears to be no relationship between the abun-dances of montmorillonite and volcanic ash seen insmear slides (Figure 17) (although operator variancemight be the reason for this); but there is a broad cor-relation between the abundance of pyroxene of basalticorigin and the montmorillonite content. The mont-morillonite is presumably detrital from the basalts anddolerites overlying the Beacon Group.

The first appearance of kaolinite in the latest Mioceneindicates the time that the basaltic cover was breachedand significant erosion of the Beacon Group occurred.The high montmorillonite content of the Miocene at Site274 suggests rapid erosion of the basalts at that time;whereas at Site 269, maximum erosion appears to havetaken place in late Oligocene times, and by the Miocenethere was substantial dilution by erosion of other rocks(not Beacon, since kaolinite is lacking). The mont-morillonite at Site 268 could perhaps have been suppliedfrom volcanic rocks within the Beacon Group, or from

ultramafic and mafic intrusions (cf. pebble data,Barrett, this volume); but the lack of Miocene or earlierkaolinite at both Sites 268 and 269 suggests that theBeacon was not substantially eroded before the Plio-cene.

These data can be interpreted in several ways. We donot know how drainage basin areas have been modifiedsince the middle Tertiary. Was there Tertiary uplift ofthe Transantarctic mountains? How much are theMiocene data for 268 modified by transport of clays bybottom currents? Much more detailed analysis is neededto resolve these problems.

FERROMANGANESE MICRONODULES

DistributionManganese nodules or micronodules were found in

several cores from Leg 28. Macronodules wererecognizable in split cores; micronodules were identifiedeither in smear slides or in sieved sand fractions ofsamples.

TOTAL FELDSPAR MONTMORILLONITE CHLORITE

j 0;5 1 METERS

% DETRITAL SAND 2-4*.01

Figure 17. Variations in petrology with depth at Sites 267, 268, 269,and 274. Note that vertical scale is different for Site 269. The per-cent detrital sand (Table 4) is believed to represent ice-rafted sedi-ment in most cases. Montmorillonite, kaolinite, and chlorite valuesfrom the less than 2 M X-ray diffraction data of Zemmels (Chapter00), normalized to mica = 1.0. Total feldspar, and percentage of feld-spar that is K feldspar, based on bulk X-ray diffraction analysis byZemmels, normalized to quartz = 1.0. Percent cristobalite based onbulk X-ray diffraction data. Note that adjacent data points have beenjoined purely to facilitate reading of the graph. No geologic predic-tions are intended. Presence of chert based on visual core description.Presence of volcanic glass (generally in trace amounts) based onsmear slide identifications. Note the probability of operator variancein the recognition of volcanic glass.

748


TOTAL FELDSPAR MONTMORILLONITE CHLORITE DETRITAL SAND 2-4-.01 .1 1 10

K-FELDSPAR0 100

KAOLINITE " CRISTOBALITE0 O. I 0 . 2 0 50 100

Figure 17. (Continued).

749


i DETRITAL SAND 2-4*

TOTAL FELDSPAR MONTMORILLONITE0 1 2 0 1 2 3

KAOLINITE % CRISTOBALITE!0.1 0.2 0 50 100 !

0 • PLEIST

-600


In the upper Miocene and lower Pliocene section ofSite 274, from Core 10, Section 5 through Core 12, Sec-tion 6, manganese nodules are common. They are up to2 cm in diameter and have a characteristic lumpy sur-face. Manganese-coated pebbles or clusters of pebblesare also common; coatings up to 2.7 mm thick wereseen. Core logs suggest a 1 cm or larger manganesenodule per meter of core.

The only other site with macronodules was 267, wherea cluster of pebbles with manganese oxide coatings200µ-600µ thick occurs in Sample 1A-6, 10-13 cm. Thissection is probably of lower Pliocene age.

Brown micronodules in the 63µ-250µ size range havebeen found in several cores (Table 5). They are of twotypes: (a) dark, rough nodules with a matte surface verysimilar in appearance to the macronodules; (b) lightercolored micronodules, with a smooth shiny surface.

The dark rough nodules are found in the upperMiocene-lower Pliocene sequence at Site 274 which con-tains macronodules and are also common in a section ofsimilar age in Cores 10 and 11 at Site 266. Dark, rough,Pliocene micronodules are found at Sites 266 and 267.

Smooth, light-colored micronodules are restricted toPlio-Pleistocene sediments at Sites 268 and 272. Fillon(1972) recognized apparently similar nodules in lowerPleistocene Ross Sea sediments in piston cores.

Chemical Composition of MicronodulesThe chemical compositions reported here have been

determined by electron microprobe. In some cases, due

KAOLINITE % CRISTOBALITE0 0.1 0.2 0 50 100

COARSE TON"INE SAND.-'RATIO <


to the high friability of the nodules, considerable dif-ficulty was encountered in preparing them for analysis.This problem, coupled with slight variations in elemen-tal composition over nodule surfaces, has resulted inanalyses which may only be regarded as semi-quantitative. Furthermore, most nodules probably con-tained substantial amounts of water which could not bedetected on the microprobe.

Analyses in Table 6 are for micronodules from Sites266, 268, 272, and 274. In some cases, it was necessary toprepare samples with micronodules hand-picked fromseveral adjacent lengths of core.

The dark rough micronodules contain high concen-trations of manganese. Silica contents are also high inthe nodules analyzed from Site 274; spot checks onalumina content suggested it was low and similar totitania. This silica is probably dispersed biogenic opalinesilica.

750


The two light, smooth nodules analyzed were low inmanganese, but relatively high in silica, alumina, andiron. They may be some type of iron oxide aggregate orconcentration associated with clays.

Discussion

The nodules and pebble coatings found in thePliocene at Site 267 and the upper Miocene-lowerPliocene at Site 274 are associated with low rates ofdeposition. The pebble concentration at Site 267 wasprobably winnowed out by bottom currents, but there isno evidence from grain-size analyses (Figure 15) of win-nowing at Site 274. Associated with the macronodules atthese sites are micronodules of very similar physicalappearance and limited analyses show close chemicalsimilarities. Sequences with micronodules only also havelow rates of sedimentation. There is no reason to sup-pose, however, that bottom currents are necessarilyresponsible.

Grains and aggregates of iron oxides are not uncom-mon in deep-sea sediments (see, for example, Cronan,1973). The reason for their high concentration at certainlevels at Sites 268 and 272 is not apparent.

CHERTS

Distribution

Porcellaneous cherts were found at Sites 268, 269,272, and 274, in strata of Miocene and Oligocene age.All appear to result from cherty lithification of normalmudstone lithologies in sequences containing siliceousmicrofossils.

At Site 268, cherts occur from Core 11 (256 m)downward. No trends with depth were detected. Chertsare recovered as short blocks, suggesting originalpebble-sized nodules or thin beds. They interbed withmudstones of varying colors and bedding structures.Most of the mudstones contain up to 5% diatoms andRadiolaria. The cherts appear to be lithified equivalentsof the mudstones: all variants of the mudstones, in-cluding those with silt laminae, are represented bycherts. All degrees of lithification, from mudstonethrough porcellaneous chert, are present. The age of thesequence at Site 268 is probably middle Oligocene toearly Miocene.

The occurrence of chert at Site 269 is almost identicalto that at Site 268. Cherts are found in Cores 10 through2A (359.5-435.5 m), of early Miocene age. Some chertyfine sandstones are found. Interbedded mudstonesusually contain siliceous biogenic material. The sectionbelow the chert sequence appears to contain calcareousrather than siliceous microfossils, and there are car-bonate cemented sandstone beds.

Chertified claystone occurs in blocks up to 20 cm longat Site 272, Cores 39-44 (365-422 m), of middle Mioceneage. As at other sites, the cherts appear to be of lithifiedclaystone or mudstone lithologies. Biogenic silica is veryrare in the sediments associated with cherts; overlyingcores contain a few diatoms and Radiolaria.

At Site 274, cherts occur from Core 35 to 39 (323-380m), but are absent below this down to basalt basementat 408.5 meters, with the exception of a minor chertified

mudstone in Core 43, Section 2. This cherty sequence isof Eocene age. Claystones interbedded with the chertscontain only trace quantities of biogenic silica, as dothose in the section below the cherty sequence. Abovethe chert sequence is a thick section of diatomite anddiatom ooze.

Observations in Hand SpecimenThe hardness of chert is revealed in two ways in hand

specimen. Conchoidal fractures are developed only inchert and not in associated mudstone. The diameter ofthe sediment core decreases with increasing hardness.There are no dense vitreous cherts composed ofrecrystallized quartz. The most indurated cherts de-scribed here have a hardness of only about 4 on Moh'sscale.

Observations in Thin SectionsCherts from Sites 268 and 269 are indistinguishable in

thin section from sections of impregnated mud from thesame cores. Distinct diagenetic silica is not recognizable.Under crossed polars there is usually a gross extinctionin the groundmass parallel to bedding. In muds this isdue to parallel orientation of clay minerals, but could beenhanced by silica in cherts. In one chert thin section(268-13-1, 100-150 cm), the specimen is divided intomany different fields of parallel extraction, with sharpirregular boundaries between the fields, suggestive ofdiagenetic mineral growth. Fragments of diatoms andRadiolaria are preserved in all the cherts examined,without showing any visible solution effects, and insome sections are very abundant. Radiolaria with acoating of an opaque mineral (?pyrite), similar to thosein unlithified sediments, are found in some cherts fromSite 268.

In cherts from Site 272, Radiolaria have not beenseen, but diatom fragments are very common, both inthe cherts and in immediately adjacent mudstones.

Examination with Scanning Electron Microscope

A few chert samples have been briefly examined underthe SEM. No clear evidence has been seen of solution ofbiogenic siliceous fragments. No growth of nodularcristobalite as reported by Heath and Moberley (1971)has been seen, but the coarseness of detrital quartzgrains in the cherts may obscure such features.

X-Ray DiffractionThe main silica phases prrsent have been identified by

X-ray diffraction. Samples were crushed with a stainlesssteel morta and pestle and then ground to powder by a20-min treatment in a tungsten carbide ball mill. Ran-dom mounts were made in an aluminum holder. Thesamples were run on a Phillips PW4620 X-ray diffrac-tometer. Instrument settings were: 40 kv; 20 mA; timeconstant 4 sec; 400 cps; chart speed 20 mm/min; scan-ning speed 1°20 per min; scanning range 15° to 28° 20.

At Site 268, quartz gives prominent peaks in allsamples. In addition, almost all cherts have a broadpeak between 21.5° and 22. 20 which is identified aspoorly crystallized cristobalite (cf. Heath and Moberly,

751


1971), otherwise known as lussatite (von Rad andRösch, 1972). This peak is sharpest, indicating thegreatest ordering of the cristobalite, in the hardest cherts(G and H, Figure 18). Only one chert (J, Figure 18) hasno detectable cristobalite peak: in thin section this chertis seen to consist of a nodule (?load cast) of well-sorted20µ quartz silt, apparently cemented with silica (the ce-ment is difficult to resolve). Quartz-dominated chertslike this are very rare in Leg 28 sites. Cristobalite isfound in mudstones within 1 cm of chert beds ornodules, but is not found in mudstones more than 1 cmfrom chert (e.g., C and I, Figure 18).

At Sites 272 and 274, the cristobalite peak is muchmore prominent, perhaps in part because of the smallerdetrital quartz content of the sediments. It occurs notonly in cherts, but also in indurated mudstones bothassociated with chert (e.g., Figure 18: N, 2 cm fromchert; P, 1 cm from chert) and also from lengths of corein which chert appears absent (e.g., T and Q, Figure 18).It is, however, not found in unlithified diatom clay oozefrom the lower part of Site 274 (Figure 18: U from 274-33-6).

DiscussionThe cherts from Leg 28 differ from most cherts

recovered by DSDP in being developed in terrigenoussediments. Detrital quartz and clay are a hinderance ininvestigating these cherts.

The diagenetic sequence inferred from these cherts issimilar to that found by Heath and Moberly (1971) andvon Rad and Rösch (1972). Disordered cristobalitedevelops in sediments inferred to originally containbiogenic opaline silica; and the cristobalite becomesmore ordered with time. The final stage of inversion ofcristobalite to quartz has not been observed in Leg 28cherts. There is no evidence at all of possible volcanicsource materials for the silica. But neither do thesiliceous skeletons preserved in the chert appear corrod-ed.

The clearest evidence of a biogenic origin for the silicain the cristobalite is the restriction of chert at Site 269 tothe deepest part of the section with common siliceousmicrofossils. Chert is absent from the underlying se-quence which contains coccoliths and foraminifera. Theabrupt reduction of siliceous microfossil abundances atabout the level where chert first appears in Hole 274may also be evidence of a biogenic source.

Smear slides of mudstones associated with chertsshow abundant fragments of diatoms, apparently rang-ing in size down to the limits of resolution (say, 5µ). Inthin sections, it is difficult to resolve fragments of lessthan 20µ. It is possible that the main source of silica isfinely comminuted diatom debris. Fragments in the few-micron size range would tend to dissolve readily, even iflarger fragments showed little sign of corrosion.

DISCUSSION AND SYNTHESIS

Ice-Rafting

SourceOur petrologic data and that of Barrett (Characteris-

tics of Pebbles, this volume) suggest that the source ofice-rafted debris for Site 274 was the Ross Sea and to a

lesser extent Victoria Land, while that at Sites 267 and268 came from elsewhere, perhaps the Shackleton IceShelf. The very low percentage of ice-rafted debris atSite 269 is partly an effect of the high sedimentationrate, but primarily reflects a lack of sediment-bearingicebergs.

Factors Controlling Rates of Ice-Rafted SedimentationFour main factors appear to control the rate of ice-

rafted sedimentation (Carey and Ahmad, 1961; Warnke,1970; Fillon, 1972; Anderson, 1972): (1) rate of con-tinental erosion; (2) thermal structure of glaciers or iceshelves, and hence both their erosional and meltingbehavior; (3) size of ice shelves; (4) sea temperatures.

How these factors affect sedimentation rate is dis-puted. A dry-base ice shelf may result in basal freezing,and less sediment will be deposited beneath or close tothe ice shelf. The larger a wet-base ice shelf is, the moresediment will be deposited beneath it, and the less willmove out to sea in bergs.

Cold bottom water generated by a large dry-base iceshelf may form a powerful deep-water bottom current.However, evidence of bottom currents in cores mayresult from other causes; for example, reduced ice-raftedand biogenic sedimentation or bottom currents resultingfrom bathymetric constrictions of deep-water move-ment. We have definite evidence of bottom currents inthe Miocene at Site 268 and in the early Pliocene(?) atSite 267. Manganese nodule accumulations may reflectbottom-current activity (Goodell et al., 1971), but thiscannot be demonstrated from our data.

Since glaciation was initiated at least as early as upperOligocene, by late Miocene time major variations in therate of continental erosion are unlikely to have resultedfrom variations in the resistance of source rocks, es-pecially from such a large source area. Barretfs pebbledata (Characteristics of Pebbles, this volume) is consis-tent with this interpretation. However, since the X-raydiffraction data (Cook et al., X-Ray Mineralogy, thisvolume) suggests little erosion of the Beacon Group un-til the early Pliocene, there may be some small long-termtrends in the rates of erosion.

We can say little about the detailed history of Ant-arctic ice cover because of intermittent core recoveryand poor biostratigraphy. Our dating is good to perhapsone million years; fluctuations in ice conditions are like-ly to be on the order of tens to hundreds of thousands ofyears.

Despite these limitations, we can roughly charac-terize the dominant ice-shelf conditions during each ofthe last four major magnetic epochs, on the basis ofvariations in the type and rate of ice-rafted sediment ac-cumulation.

Throughout the Brunhes-Matuyama (biostra-tigraphy inadequate to subdivide), ice-rafted sedimentsare characterized by moderate rates of deposition at allsites, decreasing offshore, and with a low coarse to finesand ratio at Site 274. During the (probable) upperMatuyama, Site 274 experienced a low sedimentationrate of ice-rafted material, with a high coarse to finesand ratio. The (approximate) Gauss has universallyhigher rates of sedimentation. Because of limitedrecovery at Site 274, the coarse to fine sand ratio is un-

752


1A f\f |

B A*\/ J

1c ^λflU

E Λ V Λ / W

F yf*

i

quartz

1

11 A: 268-18-1-56/58 CHERT

Rf ß; 268-18-1-56/58 MUDSTONE 1 cm from cher t

VA

-3111 268-20-1-

\ 122/124

\

> %

F

E

D

C

1

quartz

disordered 1cristobalite/\

CHERT

S l i g h t l y che r t y MUDSTONE

MUDSTONE / I 1

Ass \\

/VAT V

I I I I I

DEGREES 2<t>DEGREES 20

M: 272-41-1-36/40 CHERT NODULEN: 272-41-1-36/40 MUDSTONE 2 cm from che r t nodule0 : 272-42-1-139/145 CHERT NODULEP: 272-42-1-139/145 MUDSTONE 1 cm from cher t noduleQ: 272-41-1-77 MUDSTONE

JlR rs Vv/

T Λ /

I

ft R:S:

i T:U:

i' V

274-38-2-4/6 HARD CHERT274-36-1-81/83 SOFT CHERT274-41-2-120/121 MUDSTONE274-33-6-100/150 SOFT DIATOM MUD

A

Λ M / V / I T /Λ \ .

^ Λ-~^ A /A V\.Λ« i ft ^y\rl m A /̂ V.

Λ / ^ ^ I\ V A' / > v \ _ w ^ ~/ / \ V

V '

1 1 1 1 1 1

DEGREES 2Φ DEGREES ZΦ

Figure 18. X-ray diffractograms of cherts and mudstones from Sites 268, 272, and 274, showing development of disorderedcristobalite in different lithologies.

753


certain. During the (approximate) Gilbert, there weremoderate to high rates of ice-rafted sedimentation atnearshore sites, but very low rates at offshore Sites 266and 267. At Site 274, the coarse to fine sand ratio is high.

The data for the lower Matuyama and the whole ofthe Brunhes suggest that mostly wet-base ice-shelf con-ditions dominated, with the upper Matuyama intervalmost simply explained by an intensification of wet-baseconditions. The universally higher rates in the (ap-proximate) Gauss indicate that less material was beingdeposited further inshore than our sites, possibly due tothe formation of a dry-base ice shelf. The Ross Sea un-conformity might have formed at this time. In the (ap-proximate) Gilbert, the rapid decrease in rates offshoresuggests a resumption of wet-base ice-shelf conditionsaccompanied by rapid melting of icebergs. This wouldbe in agreement with the warm water temperaturessuggested by Ciesielski and Weaver (1973).

We must emphasize, however, that these suggestionsare speculative.

Turbidity Currents

Initiation of Turbidity CurrentsThe shelf break around Antarctica is unusually deep

and would have remained in quite deep water even dur-ing eustatic lowerings of sea level at glacial maxima. Onmost continental margins of the world, submarinecanyons were eroded when sea level was lowered. Exceptoff major deltas, most present-day and late Pleistoceneturbidites appear to have been derived from submarinecanyons heading in very shallow water. Only thesecanyons that have maintained their heads in shallowwater during the Recent transgesssion actively funnelsand to deep water at the present time. Turbiditycurrents in such submarine canyons are probably in-itiated by rip currents during storms, and transport largeamounts of sand.

There are few documented turbidity currents deriv-ed from mudflows or slides, other than on major deltas,or in areas of glaciomarine sedimentation. The GrandBanks Slump and turbidity current of 1929 was initiatedin an area of major Pleistocene deposition from floatingice. Marlowe (1968) presents evidence that turbidites inBaffin Bay are generated by muddy slumps.

In turbidity currents generated by rip currents flowingover nearshore sandy sediments, even proximal sandswill be relatively well sorted and free of mud. In con-trast, a slump or mudflow will start off with very poorsorting and large amounts of mud. Lengthy travel willbe needed for a well-sorted sand to separate out at thebase of the turbidity current. Thus graded sands with ahigh content of mud will be expected from such slump-generated turbidity currents, in contrast to the well-sorted graded sands of rip current-generated turbiditycurrents.

Such muddy sands (and indeed, also a mudflow) arefound at Site 268; the much better sorting of sands andsilts at Site 269 probably results from the turbiditycurrents having traveled much further.

These ideas could be further evaluated by seeing if thecanyons and valleys of the Antarctic continental slope

resemble normal submarine canyons, or if they arecloser to the slump scars characteristic of the area of theGrand Banks Slump, where the continental shelf breakis also well below the depth of glacially lowered sealevel.

Proximal-Distal ChangesSite 269 confirms the observations of Payne and

Conolly (1972) on proximal to distal changes in tur-bidites in the south Indian Abyssal Plain. Sediments inthe distal part of the plain consist of 60% turbidite clayand mud, 20% turbidite silt, 2% turbidite sand, and 18%mud of uncertain origin. Turbidite sands and coarse siltsare increasingly common on the more proximal parts ofthe plain.

Continental Margin ProgradationLithologic changes at Sites 268 and 269 are evidence

of substantial progradation of the continental margin,which is also shown in seismic reflection profiles. Atboth sites there is a transition from more distal facies inthe lower parts of holes to more proximal facies nearsurface.

At Site 267 there is evidence of bottom current activi-ty throughout the late Cenozoic. In the Miocene strataof Site 268 laminated silts, probably indicating contourcurrent activity, predominate and middle and upperMiocene sediments are thin or absent; but in the Plio-Pleistocene there is no evidence of contour currents.Instead, there are thicker turbidite silts and in the top-most core, sands and pebbly muds.

Likewise at Site 269 turbidite silts and clays dominatethe lower part of the section and sands become in-creasingly important in the upper part of the hole.

These data could alternatively be interpreted asevidence of increased supply of sand and coarse silt,compared with fine silt and clay, from the Antarcticcontinent through the late Cenozoic. Such an interpreta-tion is not consistent with evidence in the Ross Sea sec-tor and is not a necessary hypothesis.

ACKNOWLEDGMENTSResponsibilities for authorship are: detrital petrology and

ferromanganese micronodules—C. D. Brisco and D. J. W.Piper; other sections—D. J. W. Piper.

Laboratory work was supported by Canadian NRC GrantNo. A8390. R. MacKay assisted with electron microprobeanalysis. X-ray diffractograms were produced at the Depart-ment of Geology, Memorial University, Newfoundland,through the courtesy of Roger Slatt. Romi von Borstel carriedout some of the size analyses. We thank Roland von Huene foruseful discussion and Peter Barrett for encouragement.

REFERENCESAnderson, J. B., 1972. Nearshore glacial-marine deposition

from modern sediments of the Weddell Sea: Nature Phys.Sci., v. 240, p. 189-192.

Barrett, P. J., 1966. Petrology of some Beacon rocks betweenthe Axel Heiberg and Shackleton Glaciers, Queen MaudRange, Antarctica: J. Sediment. Petrol., v. 36, p. 794-805.

Bouma, A. H., 1972. Sedimentology of flysch deposits:Amsterdam and New York (Elsevier).

754


Bouma, A. H. and Hollister, C. D., 1973. Deep ocean basinsedimentation. In Middleton, G. V. and Bouma, A. H.(Eds.), Turbidites and deep water sedimentation, Pacificsection SEPM: p. 79-118.

Carey, S. W. and Ahmad, N., 1961. Glacial marine sedimenta-tion. In Raasch, G. O. (Ed.), Arctic Geol.: v. 2, p. 865-894.

Ciesielski, P. F. and Weaver, F. M., 1973. Southern oceanPliocene paleotemperatures based on silicoflagellates fromdeep-sea cores: Antarctic J. U.S., v. 8, p. 295-297.

Cronan, D. S., 1973. Basal ferruginous sediments cored duringLeg 16, Deep Sea Drilling Project. In van Andel, T. H.,Heath, G. R., et al., Initial Reports of the Deep Sea Drill-ing Project, Volume 16: Washington (U.S. GovernmentPrinting Office), p. 601-604.

Field, M. E. and Pilkey, O. H., 1971. Deposition of deep seasands: comparison of two areas of the Carolina continentalrise: J. Sediment. Petrol., v. 41, p. 526-536.

Fillon, R. H., 1972. Late Cenozoic geology, paleo-oceanography and paleo-climatology of the Ross Sea, Ant-arctica: Ph.D. dissertation, University of Rhode Island.

Goodell, H. G., Meylan, M. A., and Grant, B., 1971. Ferro-manganese deposits of the South Pacific Ocean, DrakePassage and Scotia Sea: Am. Geophys. Union AntarcticRes. Ser., v. 15, p. 93-168.

Heath, G. R. and Moberly, R., 1971. Cherts from the westernPacific: Leg 7, Deep Sea Drilling Project. In Winterer, E.L., et al., Initial Reports of the Deep Sea Drilling Project,Volume 7: Washington (U.S. Government Printing Office),p. 987.

Kelts, K. and Briegel, V., 1971. Sediment slides and turbiditycurrent models from Swiss lakes: Internatl. sediment, con-gr. 8th, Heidelberg. (Abstracts)

Kennett, J. P. and Brunner, C. A., 1973. Antarctic lateCenozoic glaciation: evidence for initiation of ice raftingand inferred increased bottom water activity: Geol. Soc.Am. Bull., v. 84, p. 2043-2052.

Kuenen, P. H., 1966. Matrix of turbidites: experimental ap-proach: Sedimentology, v. 7, p. 267-297.

Marlowe, J. I., 1968. Unconsolidated marine sediments in Baf-fin Bay: J. Sediment. Petrol., v. 38, p. 1065-1078.

Normark, W. R. and Piper, D. J. W., 1972. Sediments andgrowth pattern of Navy deep sea fan, San Clemente Basin,California borderland: J. Geol., v. 80, p. 198-223.

Payne, R. R. and Conolly, J. R., 1972. Turbidite sedimenta-tion off the Antarctic Continent: Am. Geophys. Union An-tarctic Res. Ser., v. 19, p. 349-364.

Piper, D. J. W., 1972a. Sediments of the middle CambrianBurgess Shale, Canada: Lethaia, v. 5, p. 169-175.

, 1972b. Turbidite origin of some laminatedmudstones: Geol. Mag., v. 109, p. 115-126.

., 1973. The sedimentology of silt turbidites from theGulf of Alaska. In Kulm, L. D., von Huene, R., et al.,Initial Reports of the Deep Sea Drilling Project, Volume18: Washington (U.S. Government Printing Office), p. 847.

Selley, R. C, 1970. Studies of sequence in sediments using asimple mathematical device: Quart. J. Geol. Soc. London,v. 125, p. 557-581.

von Huene, R., Larson, E., and Crouch, J., 1973. Preliminarystudy of ice-rafted erratics as indicators of glacial advancesin the Gulf of Alaska. In Kulm, L. D., von Huene, R., etal., Initial Reports of the Deep Sea Drilling Project,Volume 18: Washington (U.S. Government Printing Of-fice), p. 835-842.

von Rad, U. and Rösch, H., 1972. Mineralogy and origin ofclay minerals, silica and authigenic silicates in Leg XIVsediments. In Hayes, D. E., Pimm, A. C, et al., InitialReports of the Deep Sea Drilling Project, Volume 14:Washington (U.S. Government printing Office), p. 727.

Warnke, D. A., 1970. Glacial erosion, ice rafting and glacialmarine sediments: Antarctica and the Southern Ocean:Am. J. Sci., v. 269, p. 279-294.

755

21. deep-water continental-margin sedimentation, … · 2007. 5. 24. · david j.w. piper and...

Documents