REPRINT

Drilling and Dating New JerseyOligocene-Miocene Sequences: Ice Volume,

Global Sea Level, and Exxon RecordsKenneth G. Miller,* Gregory S. Mountain, the Leg 150 Shipboard

Party, and Members of the New Jersey CoastalPlain Drilling Project

Oligocene to middle Miocene sequence boundaries on the New Jersey coastal plain(Ocean Drilling Project Leg 150X) and continental slope (Ocean Drilling Project Leg 150)were dated by integrating strontium isotopic stratigraphy, magnetostratigraphy, andbiostratigraphy (planktonic foraminifera, nannofossils, dinocysts, and diatoms). The agesof coastal plain unconformities and slope seismic reflectors (unconformities or stratalbreaks with no discernible hiatuses) match the ages of global δ 1 8 θ increases (inferredglacioeustatic lowerings) measured in deep-sea sites. These correlations confirm a causallink between coastal plain and slope sequence boundaries: both formed during globalsea-level lowerings. The ages of New Jersey sequence boundaries and global δ 1 8 θincreases also correlate well with the Exxon Production Research sea-level records of Haqet al. and Vail et a/., validating and refining their compilations.

E>ustatic (global sea level) changes exert oneof the primary controls on the stratigraphicrecord (J, 2), although controversy surroundsthe age, magnitude, and mechanism of thesechanges (3). Vail et d. (4) and Haq et d. (5)reconstructed eustatic history by applying se-quence stratigraphy to a global array of pro-prietary Exxon Production Research (EPR)data comprising seismic profiles, wells, andoutcrops. Previously released EPR seismic datademonstrated that Oligocene to Recent se-quences are well defined beneath the NewJersey shelf, although the age control on thesesequences was poor ( ± 1 million years orworse) (6). To improve understanding of sea-level change, we collected additional mul-tichannel seismic data (cruise Ew9009) andtraced seismic sequences from the New Jerseyshelf to the slope (7). These sequences weredated at four slope sites drilled during OceanDrilling Project (ODP) Leg 150 (8) (Fig. 1).Drilling onshore at Island Beach, AtlanticCity, and Cape May, New Jersey (ODP Leg150X; Fig. 1), provided additional ages andfacies of these same sequences in much shal-lower paleodepths (9). This report synthesizesLeg 150 and Leg 150X chronologic studies ofOligocene to middle Miocene sequences thatare preserved onshore and have the clearlyvisible seismic reflection terminations off-shore. We compare the stratigraphic record ofthe New Jersey sequence with published δ 1 8 θrecords (Figs. 1 and 2) and with the inferredeustatic record of Haq et αl. (5).

K. G. Miller, Department of Geological Sciences, RutgersUniversity, Piscataway, NJ 08855, USA, and Lamont-Doherty Earth Observatory of Columbia University, Pali-sades, NY 10964, USA.G. S. Mountain, Lamont-Doherty Earth Observatory ofColumbia University, Palisades, NY 10964, USA.The members of the Leg 150 shipboard party and theNew Jersey Coastal Plain Drilling Project are listed (33).

*To whom correspondence should be addressed.

1092

Deep-sea δ 1 8 θ records provide a proxy forice volume and sea-level (glacioeustatic)changes during the Oligocene to Recent"Icehouse World" (10, II). Glaciomarinesediments near Antarctica and deep-sea ox-ygen isotopic records (10, 11) indicate thatlarge ice sheets have existed in Antarcticasince the earliest Oligocene [—35 million

Site 563

years ago (Ma) (12)]. Because ice preferen-tially sequesters light oxygen isotopes, fluc-tuations in ice volume cause changes inglobal seawater δ 1 8 θ (δ w ) . These global δ w

changes are recorded by benthic and plank-tonic foraminifera along with variations inseawater temperature and local isotopiccomposition. Comparisons of benthic andlow-latitude (nonupwelling) planktonic fo-raminiferal δ 1 8 θ records can be used to iso-late ice volume effects from local isotopicand temperature changes (13). Using thisstrategy, Miller et d. (10) and Wright andMiller (14) identified 12 Oligocene to Mio-cene benthic foraminiferal δ 1 8 θ increases(all >0.5 per mil); these increases culminat-ed in δ 1 8 θ maxima that were used to definezones Oil to Oi2b and Mil to Mi7 (Figs. 1and 2 and Table 1). Six of the δ 1 8 θ increasesare also recorded by tropical or subtropicalplanktonic foraminifera; the other six lacksuitable low-latitude isotopic records. Milleret d. (10) interpreted coeval increases inbenthic and planktonic δ 1 8 θ records as theconsequence of glacioeustatic lowerings of~30 to 80 m. On the basis of the ODP Site747 δ 1 8 θ record (Fig. 1), we suggest that theMi3 increase (13.4 to 14 Ma; Table 1) canbe split into two increases (Mi3a and Mi3b).We assume that all 13 Oligocene to early-to-late Miocene δ 1 8 θ increases (Figs. 1 and2) reflect million-year scale increases in ice

Site 747

δ 1 8 o

10-

11

12 -

13 -

14 -

1 5 -

16 -

17

18 -

ene

Mio

c

late

die

mid

early

benthic

Site 608

Milb

oxygen isotopeincreases

Fig. 1 . Comparison of the timing of middle Miocene reflectors on the New Jersey slope with three benthicforaminiferal δ 1 8 θ records (units are per mil). Zones Mi1b to Mi6 are oxygen isotopic zones associatedwith the δ 1 8 θ increases. Reflectors m5.2 to ml are dated on the New Jersey slope. Two independentlydated sets of stippled lines are shown: (i) lines are drawn through inflections in the δ 1 8 θ records; (ii) agesof the reflectors are shown as best estimates (lines) and error bars (boxes) (Table 1). Oxygen isotope datafor ODP sites 563 (western North Atlantic), 608 (eastern North Atlantic), and 747 (Indian sector, SouthernOcean) are generated on Cibicidoides spp. after Wright and Miller {14). Inset map shows locations of theonshore and offshore drilling sites.

SCIENCE VOL. 271 23 FEBRUARY 1996

(Reprinted with permission from Science, Vol. 271 (23 Feb. 1996), 1092-1095. Copyright 1996 American Association for the Advancement of Science.)

REPRINT

volume, although additional low-latitudeplanktonic foraminiferal δ 1 8 θ data are need-ed to confirm this (15).

Oligocene to Recent seismic reflectionsbeneath the New Jersey shelf exhibit ero-sional truncation, onlap, downlap, and top-lap and are thus objectively identified assequence boundaries (4, 6, 8). We tracedthese sequence boundaries from the shelf tothe slope, using both EPR and Ew9009 mul-tichannel seismic data including Red, Tus-can, Yellow-2, Pink-2, and Green (6) plusOchre, Sand, True Blue, Pink-3, andGreen-2 (8). To simplify the nomenclatureand incorporate reflections restricted to theslope, we use a unified alpha-numericscheme (ol, m6; Figs. 1 and 2 and Table 1)based on the results of ODP Leg 150 (8).

We derived time-depth relations for cor-relating seismic profiles to the boreholesfrom three sources: the velocity log from theContinental Offshore Stratigraphic Test(COST) B-3 well, semblance velocitiesfrom analysis of Ew9009 Common DepthPoint (CDP) stacks on the adjacent shelf,and sonobuoy data from the continentalrise (8). Synthetic seismograms derivedfrom log (8) and core physical propertiesdata (16) were used to evaluate these cor-relations. The sedimentary expression of se-

quence boundaries on the slope is mutedbecause of relatively uniform Oligocene toMiocene lithologies (silty clays) (8), andseveral reflectors are associated with a cor-relative conformity (17). Still, many se-quence boundaries are associated with hia-tuses or increased sand content immediatelyabove the boundary, both of which yieldimpedance contrasts (8) and consequentlyseismic reflections.

We developed the Oligocene to middleMiocene chronology on the slope by inte-grating Sr isotopic stratigraphy (17) andmagnetostratigraphy (18) with planktonicforaminiferal (19), nannofossil (20), dino-cyst (21), and diatom (22) biostratigraphy(Table 1). We do not discuss late Mioceneto Recent history here because (i) the chro-nology of the upper Miocene slope sectionsis still uncertain, (ii) Pliocene strata arepoorly represented in the slope boreholes,and (iii) the recovered Quaternary sectionswere restricted to the middle Pleistocene(stages 15 to 5.5) and Recent (23).

Onshore boreholes recovered fossilifer-ous Oligocene to middle Miocene strata;younger strata were mostly unfossiliferousand undateable (9, 24). We identified un-conformities (sequence boundaries) in theonshore boreholes using physical stratigra-

δ 1 8 obenthic foraminifera

Onshoresequences

"Eustatic" CurveHaq etal. (1987)

Fig. 2. Compari-son of the timingof Oligocene tomiddle Miocenereflectors on theNew Jersey slopewith a benthic fo-raminiferal δ 1 8 θrecord, a sum-mary of onshoresequences, andthe inferred eu-static record ofHaq etal. (5). Theδ 1 8 θ record is astacked compos-ite of Cibicidoidesspp. from severalsites that hasbeen smoothedto remove all pe-riods longer than~1 million years{32); Oil to Mi6are δ 1 8 θ maxima;dashed lines indi-cate inflections inthe δ 1 8 θ recordsimmediately be-fore the maxima.Reflectors o1 tom l are dated onthe New Jersey

slope and are shown with best age estimates indicated with thin lines and error bars indicated with boxes(Table 1). Onshore sequences are indicated by dark boxes; the white areas in between are hiatuses.Sequences 01 to 06 are Oligocene, and KwO to Kw-Cohansey (Coh) are Miocene onshore New Jerseysequences; cross-hatched areas indicate uncertain ages. Sequences TA4.4 to TB3.1 are from Haq etal.(5), and arrows are drawn at the inflection points in their inferred eustatic record.

SCIENCE VOL. 271 23 FEBRUARY 1996

phy, including erosional contacts, rework-ing, bioturbation, major facies changes,gamma-ray peaks, and paraconformities in-ferred from biostratigraphic and Sr isotopicage breaks. Onshore sequences consist ofbasal transgressive deposits (TransgressiveSystems Tracts; glauconitic in the Oligo-cene; occasionally shelly in the Miocene)that progressively shallow upsection to me-dial silts and upper sands (High-Stand Sys-tems Tracts); low-stand deposits are notfound on the coastal plain but are restrictedto beneath the shelf and slope (6). Mioceneonshore sequences were named KwO to Kw-Cohansey (9, 25), whereas Oligocene se-quences were termed 0 1 to O6 (26). Agecontrol for the Miocene onshore sectionsrelies primarily on Sr isotopic stratigraphywith an age resolution of ±0.4 million yearsfor the early Miocene and ±0.9 millionyears for the middle Miocene (27). Diatomand planktonic foraminiferal biostratigra-phy supplements Miocene onshore control(9, 25). We derived age control for Oligo-cene onshore sections by integrating mag-netostratigraphy, biostratigraphy (plank-tonic foraminifera and nannofossil), and Srisotopic stratigraphy, with a resulting strati-graphic resolution of better than ±0.5 mil-lion years in most cases.

There is excellent correlation betweenthe timing of the major Oligocene to mid-dle Miocene slope reflectors dated at theLeg 150 slope sites and glacioeustatic low-erings inferred from the δ 1 8 θ record (Figs. 1and 2 and Table 1). Reflectors ol , m6,m.5.6, m5.2, m5, m4, m3, m2, and ml cor-relate with the Oi l , Mil, Mila, Milb, Mi2,Mi3a, Mi3b, Mi4, and Mi5 δ 1 8 θ increases,respectively (Figs. 1 and 2 and Table 1).This similarity confirms a link between se-quence boundaries traced from the shelfand glacioeustatic changes. Of the reflec-tors, only m5.4 does not appear to have acorresponding δ 1 8 θ increase. Of the δ 1 8 θincreases, only Oi2b and Oi2 fail to haveequivalent reflectors because Oligoceneseismic resolution is limited by the thinsection and concatenated reflections on theslope (Fig. 2).

Detailed comparison of the ages of slopereflectors and their corresponding error es-timates with three of the middle Miocenebenthic foraminiferal δ 1 8 θ records used todefine the Mi zones (Fig. 1) shows remark-ably similar ages for the δ 1 8 θ inflectionsand reflectors. This comparison indicatesthat the sequence boundaries formed duringintervals of rapid glacioeustatic fall, as pre-dicted by various models (28).

This link between offshore New Jerseysequences and δ 1 8 θ records is furtherstrengthened if one compares the slope se-quences with their correlative onshorecounterparts (Fig. 2). Early to middle Mio-cene onshore sequence boundaries correlate

1093

REPRINT

well with major δ 1 8 θ increases (24, 25)(Fig. 2 and Table 1), indicating that theseunconformities were formed by global sea-level lowerings. Oligocene Oil, Oi2, andOi2b δ 1 8 θ increases also correlate with on-shore sequences Ol, O3, and O5, respec-tively (26). Sequence boundaries O2, O4,and O6 may correlate with minor δ 1 8 θincreases noted in recently publishedrecords (29).

The onshore and offshore sequencescompare well with each other and with theδ 1 8 θ record. The exceptions are as follows:(i) The Kwlc sequence boundary correlateswith the m5.4 slope reflector but with noδ I 8 O change within 1 million years. EitherKwlc or m5.4 sequences are the result of alocal lowering of base level, or they maycorrelate with a minor δ 1 8 θ increase atabout 21 Ma (14). (ii) The Kw2c sequenceboundary has no definite offshore counter-part. We are uncertain of the significance ofthis sequence boundary onshore because ithas been recovered at only one borehole(Cape May), (iii) The Oligocene onshoreboreholes record sequences not resolved onslope seismic profiles because of slope sedi-ment starvation.

Although the record of Haq et al. (5)has come under criticism as a reliable in-dicator of eustatic change (3), there isexcellent correlation between the recordof Haq et al. and the New Jersey records inboth the number and ages of Oligocene tomiddle Miocene sequences (Fig. 2 and Ta-ble 1). Comparison of the ages of the twoindependent sets of sequences shows thefollowing essentially identical ages: TB2.6and m2 sequences ( — 12.6 Ma); TB2.5 andm3 (-13.6 Ma); TB2.3 and m5 (16.5 to

16.9 Ma); and TB1.5 and m5.6 (22 Ma).The ages of the Oligocene TB1.4 and m6,TB1.3 and O6, TB1.2 and O5.TB1.2 andO3, TB4.5 and O2, and TB4-4 and Olsequences are similar when they are cor-rected for differences in the time scaleused in each study (30). The record of Haqet al. (5) also compares well with the δ 1 8 θincreases (Table 1). However, on the basisof our correlation to the New Jersey se-quences and δ 1 8 θ records, there are dif-ferences compared to the ages of severalother of the Miocene sequences of Haq etal. (Table 1). It appears that TB3.1 (10.5Ma), TB2.4 (15.5 Ma), and TB2.2 (17.5Ma) correlate with —11-Ma, 14.8-Ma, and18.5-Ma slope reflectors and with the-11.4-Ma, 14.4-Ma, and 18.5-Ma δ 1 8 θincreases, respectively (Table 1). The mi-nor differences in age (Table 1) among thesequences of Haq et al., New Jersey slopereflectors, and the δ 1 8 θ increases are gen-erally within the errors in dating the mar-gin sequences. For example, differences inage between the δ 1 8 θ inflections and theNew Jersey sequences are less than 0.6million years in all cases but one (Table1); differences with the record of Haq et al.are larger because the latter relied on wellcuttings [particularly on the New Jerseymargin (6)] and not on continuously coredboreholes.

We suggest that the ages of the δ 1 8 θincreases (inflections on Table 1) providethe best estimates on the timing of Oligo-cene to Miocene eustatic falls and that un-conformities (including seismic sequenceboundaries) are formed during falls in sealevel. Our records show that deposition re-sumed in the coastal plain by the time of the

lowest low stand (maximum δ 1 8 θ values;Fig. 2). Our margin chronologic resolution isinsufficient to evaluate small leads and lags(<l/4 of a cycle or a resolution of betterthan 0.25 to 0.5 million years) between eu-static falls and the timing of unconformitiesor hiatuses on the New Jersey margin. Reyn-olds et al. (31) used forward models to predictthat the unconformities begin to form onold, slowly subsiding margins such as NewJersey early in the fall of sea level (before theinflection and the maximum rate of fall).We cannot yet evaluate at what point in aeustatic fall the unconformities begin to formon this margin.

Although it is not possible to evaluatefully the age errors in the EPR records, ourscan be specified. Stratigraphic resolution iscoarse in some intervals (for example, re-flectors m5 and m5.4 have age uncertaintiesof at least ±0.9 and ±1.1 million years,respectively; Table 1), whereas others arewell dated by integrating Sr isotopic, mag-netostratigraphic, and biostratigraphic data.For example, the small uncertainty in theage of reflector m6 (23.8 ± 0.2 Ma) allowsa precise and unequivocal correlation withthe Mil oxygen isotopic increase (inflec-tion at 23.8 Ma; Fig. 2).

Given that some reflectors and sequenc-es have age errors of greater than 0.5 mil-lion years, one could argue that the corre-lations shown on Figs. 1 and 2 are at bestfortuitous and, at worse, are beyond theprecision of the geochronology that wehave used. Using this argument, Miall (3)claimed that stratigraphic resolution maynot be sufficient to document precise cor-relation and causal links between sequencesand the global synthesis of Haq et al. (5). In

Table 1. Comparison of Sr isotope-based age estimates of Oligocene-middle Miocene seismic reflectors, New Jersey continental slope with on-shore sequences (24, 26), oxygen isotopic increases (10, 14), and the se-quences of Haq et al. (5). The column labeled Best uses the older (1985) timescale, whereas the column labeled BKSA95 provides the ages of sequences

using the new (1996) time scale of Berggren etal. (12). We obtained correctedages of Haq et al. (5) by linearly interpolating ages between TB1.4 correctedfor time scale differences (24.2 versus 25.5 Ma), the revised age of the TB1.1sequence of 32.2 Ma (30), and the revised age of TA4.4 of 35.9 Ma (30).

Slopereflector

ml (Tuscan)m2(Yellow-2)m3 (Blue)m4 (Pink-2)m5 (Green)

m5.2 (Ochre)m5.4 (Sand)m5.6 (True blue)m6 (Pink-3)

o1 (Green)

Age estimate (Ma)

Best (error)

- 1 1 (10.5-11.3)12.5 (12.5-12.6)13.6 (12.8-13.6)14.8 (13.8-15.0)

-16.9 (16.3-18.0)

18,2 (18.0-18.4)

19-20 (18.4-20.6)- 2 2 (21.5-22.5)

23.8 (23.6-24.0)

35.8-36.7 (32-36.7)

BKSA95

-11.512.713.614.7

-16.617.7

18.8-19.821.823.8

Onshoresequence

Kw-Coh?Kw3Kw2cKw2bKw2a

KwiCKw1a,bKwOO6O5

04030201

Zone

Mi5Mi4Mi3bMi3aMi2Mi1b

?minor

Mil aM i l?minorOi2b?minorOi2?minorOil

δ 1 8 θ

Maximuminflection

11.3-11.412.6-12.813.5-13.614.1-14.4*16.1-16.318.1-18.5

720.6-21.1*21.8-22.423.5-23.826.0-26.2*28.0/28.2

?31.5-32.0

?35.8-36.0

Haq et al.

Sequence

TB3.1TB2.6TB2.5TB2.4TB2.3TB2.2

TB2.1TB1.5TB1.4TB1.3TB1.2

TB1.1TB4.5TA4.4

Age/corrected

age

10.512.513.815.516.517.5

21.022.0

25.5/24.226.5/26.328.4/29.4

30.0/32.233.0/34.436.0/35.9

*Not a formal isotopic zone.

1094 SCIENCE VOL. 271 23 FEBRUARY 1996

427

REPRINT

contrast, we propose that it is unnecessaryto demonstrate that every event correlateswith a resolution of better than 0.5 mil-lion years. We have anchored key strati-graphic levels (such as reflectors ml to m4and m6) to a precise chronology and re-port a similar number of events in boththe margin and δ I 8Q records, indicatingthat unconformities (sequence bound-aries) correlate with glacioeustatic lower-ings. By firmly dating the sequences andproviding error estimates for these ages,we provide a template of Oligocene toMiocene sequences that will be comparedwith records from other margins.

REFERENCES AND NOTES

1. L L Sloss, Geol. Soc. Am. Bull. 74, 93 (1963).2. J. Imbrieefa/., Report on the Second Conference on

Scientific Ocean Drilling (European Science Founda-tion, Strasbourg, France, 1988).

3. A. D. Miall, J. Sediment. Petrol. 61, 497 (1991).4. P. R. Vail etal., Mem. Am. Assoc. Pet. Geol. 26, 49

(1977).5. B. U. Haq, J. Hardenbol, P. R. Vail, Science 235,

1156(1987).6. S. M. Greenlee, W. J. Devlin, K. G. Miller, G. S.

Mountain, P. B. Flemings, Geol. Soc. Am. Bull. 104,1403(1992).

7. K. G. Miller and G. S. Mountain, Proc. Ocean DrillingProgram Init. Rep. 150, 11 (1994).

8. G. S. Mountain etal., Eds., ibid., p. 1.9. K. G. Miller ef a/., ibid. 150X, 5 (1994); K. G. Miller ef

a/., Proc. Ocean Drilling Program Sci. Results, inpress.

10. K. G. Miller, J. D. Wright, R. G. Fairbanks, J. Geophys.Res. 96, 6829 (1991); K. G. Miller, R. G. Fairbanks, G.S. Mountain, Paleoceanography 2,1 (1987).

11. J. C. Zachos, L. D. Stott, K. C. Lohmann, Pale-oceanography 9, 353 (1994).

12. W. A. Berggren, D. V. Kent, J. J. Flynn, J. A. vanCouvering, Geol. Soc. Am. Bull. 96,1407 (1985). Weuse this time scale throughout, except as noted onTable 1. Although the revised time scale of W. A.Berggren, D. V. Kent, C. C. Swisher, and M. P. Aubry[in Geochronology, Time Scales, and Global Strati-graphic Correlation (SEPM Special Publ. 54, Societyfor Sedimentary Geology, Tulsa, OK, in press)] dra-matically revises the Oligocene ages (with minorchanges in the Miocene), we report the Leg 150studies using the older scale to maintain consistencyamong all leg results.

13. N. J. Shackleton and N. D. Opdyke, Quat. Res. 3,39(1973).

14. J. D. Wright and K. G. Miller, Proc. Ocean DrillingProgram Sci. Results 120, 855 (1992).

15. There are higher frequency (104 to 105 years, "Mi-lankovitch scale") 818O and sea-level variations em-bedded within the longer term (106 years) changes(Figs. 1 and 2) that we do not address.

16. J. M. Lorenzo and S. P. Hesselbo, Proc. OceanDrilling Program Sci. Results, in press.

17. K. G. Miller, C. Liu, M. Feigenson, ibid., in press.18. M. Van Fossen and M. Urbat, ibid., in press.19. S. W. Snyder, K. G. Miller, E. Saperson, ibid., in press.20. M.-P. Aubry, ibid., in press.21. L. de Verteuil, ibid., in press.22. L. H. Burckle, ibid., in press.23. B. A Christensen, B. Hoppie, R. Thunell, K. G. Miller,

L. Burckle, ibid., in press.24. K. G. Miller and P. J. Sugarman, Geology 23, 747

(1995).25. P. J. Sugarman, K. G. Miller, J. P. Owens, M. D.

Feigenson, Geol. Soc. Am. Bull. 105, 423 (1993).26. S. F. Pekar, thesis, Rutgers University (1995);

and K. G. Miller, unpublished manuscript.27. J. S. Oslick, K. G. Miller, M. D. Feigenson, J. D.

Wright, Paleoceanography 9, 427 (1994).28. W. C. Pitman, Geol. Soc. Am. Bull. 89,1389 (1978).29. E. Barrera, J. Baldauf, K. C. Lohmann, Proc. Ocean

Drilling Program Sci. Results 130, 269 (1993).30. See also K. G. Miller, P. R. Thompson, and D. V. Kent

[Paleoceanography 8, 313 (1993)] for discussion ofthis in Alabama boreholes.

31. D. J. Reynolds, M. S. Steckler, B. J. Coakley, J.Geophys. Res. 96, 6931 (1991).

32. J. D. Wright, K. G. Miller, Antarct. Res. Ser. 60, 1(1994).

33. The Leg 150 shipboard party includes K. G. Millerand G. S. Mountain; P. Blum and S. Gartner, TexasA&M University, College Station, T× 77845, USA;P.-G. Aim, University of Lund, S-221 00 Lund,Sweden; M.-P. Aubry, Institut des Sciences de1'Evolution, Montpellier Cedex 5, France; L. H.Burckle, G. Guerin, M. E. Katz, Lamont-DohertyEarth Observatory; B. A. Christensen, University ofSouth Carolina, Columbia, SC 29208, USA; J.Compton, University of South Florida, St. Peters-burg, FL 33701, USA; J. E. Damuth, University ofTexas, Arlington, T× 76019, USA; J.-F. Deconinck,Universite de Lilie, 59655 Villeneuve D'Asq Cedex,France; L. de Verteuil, University of Toronto, Ontar-io M5S 3B1, Canada; C. S. Fulthorpe, University ofTexas, Austin, TX 78759, USA; S. P. Hesselbo,University of Oxford, Oxford OX1 3PR, UK; B. Hop-pie, University of California, Santa Cruz, CA 95064,USA; N. Kotake, Chiba University, Chiba 263, Ja-pan; J. M. Lorenzo, Louisiana State University, Ba-ton Rouge, LA 70803, USA; S. McCracken, Univer-sity of Western Australia, Nederlands, 6009 Aus-

tralia; C. M. McHugh, Queen College, Flushing, NY11367, USA; W. C. Quayle, University of New-castle, Newcastle on Tyne, NE1 7RU UK; Y. Saito,Geological Survey of Japan, Higashi 1-1-3,Tsukuba, Ibaraki 305, Japan; S. W. Snyder, EastCarolina State University, Greenville, NC 27858,USA; W. G. ten Kate, Free University, Amsterdam,The Netherlands; M. Urbat, Universitat zur Koln,500 Koln, Germany; M. C. Van Fossen, RutgersUniversity; and A. Vecsei, Geologisches Institut derUniversitat, 79104 Freiburg i. Br., Germany. TheNew Jersey Coastal Plain Drilling Project includesK. G. Miller; P. J. Sugarman and L. Mullikin, NewJersey Geological Survey, Trenton, NJ 08625,USA; S. Pekar, J. V. Browning, C. Liu, M. C. VanFossen, M. D. Feigenson, M. Goss, D. Gwynn,Rutgers University; D. V. Kent and L. H. Burckle,Lamont-Doherty Earth Observatory; M.-P. Aubry,Institut des Sciences de 1'Evolution; D. Queen, D.Powars, T. Heibel, U.S. Geological Survey, Reston,VA 22092, USA; and D. Bukry, U. S. GeologicalSurvey, Menlo Park, CA 94025, USA.

34. We thank three anonymous reviewers for com-ments. Supported by NSF grants OCE89-11810,OCE92-03282, and EAR92-18210 and by the JointOceanographic Institute of the U.S. Science Advis-ory Council. This is Lamont-Doherty Earth Observa-tory contribution 5456.

29 September 1995; accepted 21 December 1995

SCIENCE VOL. 271 23 FEBRUARY 1996 1095

428