OCEANUS • 29

-1700

-160

0

-150

0

-1500

-150

0

-1500

-1400

-1400

-1400

-140

0

-1400

-1400

-1400

-1300

-1300

-1300

-130

0

-130

0

-130

0

-1300

-13

-1300

-1300

-130

0

-130

0

-130

0

1300

-130

0

-1200

-1200

-1200

-120

0

-120

0

-120

0

-120

0

-120

0-1

200

-1200

-1200

-1200

-1200

-1200

-120

0

-1200

-120

0

-1200

-1200

-1200

-1100

-1100

-1100

-1100

-110

0

-110

0

-110

0

-1100-1100

-110

0

-1100

-110

0

-1100

-1100

-1100-1100

-1100

-1100

-110

0-1

100

-110

0

-1100

-1100

-1100

-1100

-1000

-1000

-1000

-1000

-1000

-100

0

-1000

-1000

-1000

-1000

-100

0

-100

0

-100

0

-1000

-1000 -1000

-1000

-1000

-1000

-100

0

-1000

-1000

-100

0

-100

0

-100

0

-100

0

-900

-900

-900

-900

-900

-900

-900

-900

-900 -900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-900

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-800

-700

-700

-700

-700

-700

-700

-700

-700

-700

-700

-700

735B

1105A

Bathymetry of Atlantis BankAtlantis II Fracture Zone: South West Indian Ridge

I never imagined I would spend six weeks of mylife “wandering around” the seafloor exploringan 11 million year old beach, and it never oc-

curred to me to look for a fossil island. But that’swhat I did, and that’s what we found on two re-search voyages separated by more than a decade.

But, first, a bit of history: During World War II,Princeton mineralogist Harry Hess convinced USofficials to keep Navy ships’ echo sounders runningas they crisscrossed the Pacific. Hess, a captain andlater admiral in the Navy, even managed to have hismen bring back rock samples when they returned tothe ship after Pacific Island landings. We not onlywon the war, but also identified the mid-Pacificmountain range now known as the East PacificRise—a discovery that proved fundamental to thethen developing theory of plate tectonics.

Thus began modern marine geology and geo-physics. Prompted by Hess’s discoveries, whenpeace returned, American scientists initiated asystematic study of this little known region. One ofthe first results was the discovery that the oceancrust appeared to present a consistent, layeredseismic structure composed of sediments (layer 1 atthe seafloor) followed by two layers with rapidlyincreasing sound velocity, and then a reflector,dubbed the Moho after the its discoverer, a Czecho-slovakian seismologist named Andrija Mohorovicic.What surprised scientists was the Moho was foundat a very shallow, nearly constant depth of 6 to 7kilometers, and that seismic layers 2 and 3 abovealso appeared uniform in the oceans, whereas thisboundary ranges from 35 to 60 kilometers deepbeneath continents.

This led to a great debate about the nature of theMoho beneath the oceans and the composition ofthe layers above it. Hess proposed in his now famous1962 paper The History of the Ocean Basins * that theMoho was an alteration front created by hydrother-mal solutions converting the earth’s mantle to alighter rock known as serpentine. Serpentine formsby hydration of the primary mantle rock peridotite,which is named for the gem stone peridote (the mostcommon mineral, known to scientists as olivine). Theuniform depth of the Moho, then, represented the

Indian Ocean’s Atlantis BankYields Deep-Earth Insight

Henry J.B. DickSenior Scientist, Geology and Geophysics Department

With the Ocean Drilling Program Leg 176 and RRS James Clark Ross Cruise 31 Scientific Parties

Five-meter bathymetric map of Atlantis Bank created by WHOI scientist Maurice Tiveyfrom the James Clark Ross echo sounding tracks in collaboration with the other scientistson the cruise. The extraordinary high-resolution map shows the very top of Atlantis Bankwith the wave cut platform and the precipitous sea cliffs to either side. One can see oldheadlands and seastacks in the image, as well as the great arcuate fossil lagoon on the westside where oolitic limestones were dredged.

*Hess, H.H., 1962, The History of the Ocean Basins, in PetrologicStudies: A volume in honor of A.F. Buddington, A.E.J. Engel,H.L. James, and B.F. Leonard, eds., Geological Society ofAmerica, pp. 599-620.

upper temperature and pressure limit of peridotiteand olivine as stable elements beneath the oceanridges. Hess thought Layer 2 was the volcanic cara-pace formed as lavas erupted above the hydrother-mally altered mantle rock.

Lago

on

Headland

Lago

on

Headland

Headland

S e

a s

t a

c k

s

30 • Vol. 41, No. 1 • 1998

being recovered was too brittle, too broken, and itseemed too difficult to go deeper.

In 1984 my colleague Jim Natland (Scripps) and Iproposed using tectonic windows in the ocean crustto get to the deeper rocks. Simply go someplacewhere Mother Earth had done a lot of work for us bytectonically stripping off the shallow layers, andstart deep. In 1986, we surveyed the seafloor southof the island of Mauritius at a great oceanic trans-form fault named Atlantis II Fracture Zone, lookingfor a place to drill. The survey produced a map of a200-kilometer long, mid-sea mountain range flank-ing the fault. In the middle of these mountains wasa remarkable flat-topped peak rising up some 5kilometers from the seafloor. On the walls ofAtlantis Bank, as we named the feature, we sampledgabbro and peridotite, proof that drilling hereshould be productive.

Then, in 1987, Jim and I returned to AtlantisBank aboard the drill ship JOIDES Resolution alongwith several WHOI colleagues including co-chiefscientist Dick Von Herzen, who had backed theproject from the beginning. Over 16 days of drillingwe recovered 437 meters of gabbro from a 500 meterhole, making Hole 735B the most successful hard

0 20 40 80 10060

I

II

III

IV

V

VI

VII

VIII

X

IX

XI

0

500

1500

1000

XII

Lithologic Variability

unit

Disseminated FeTi Oxide Gabbro

Gabbronorite

Gabbro

Olivine Gabbro

Iroctolitic Gabbro

FeTi Oxide Gabbro

FeTi Oxide Gabbronorite

Lithologic columncompiled by JOIDESResolution scientists

showing the varia-bility of lithologiesdownhole and the

gradation frompredominantly oxide-

rich gabbros at thetop to olivine-rich

gabbros at the bot-tom. The figure was

constructed by usinga running average of

the different litho-logic units drilled

(almost 1,000 in all).

0

500

1000

1500

20000 4 8 16 20 23

Drilling Time in Weeks

Leg 70

Leg 83

Leg 111

Leg 140

Leg 148

Leg118

12/4/8712/13/79

1/1/82

10/9/864/25/91

11/6/912/9/93

Leg 176

Hole735B

Drilling & Penetration History Holes 735B & 504B

11/27/97

Hole504B

Hole 504B Recovery10/19/79

Leg 69

Leg 137

Hole 735BBRecovery

Drilling results for the two most successful hard-rock holesin the ocean crust: Hole 735B penetrated the lower oceancrust formed beneath the Southwest Indian Ridge andHole 504B the upper ocean crust formed at the Costa RicaRift in the eastern Pacific. Note that while Hole 504B isdeeper, it took nearly four times as long to drill, and only18 percent of the rock drilled was recovered while some 87percent of the rock drilled was recovered from Hole 735B.

This theory was rejected, however,because laboratory experiments onserpentine failed to produce the exactacoustic velocity characteristics in-ferred for seismic layer 3. Instead amodel where the Moho was the crust-mantle boundary, with layer 3 com-posed of the rock known as gabbrooverlying the unaltered mantle peri-dotite has become the generally ac-cepted paradigm. Gabbro is a coarsegrained rock similar in composition tothe basaltic lava that erupts fromsubmarine volcanoes. This rock isfound exposed on land where deeplayers of the earth have been upliftedonto mountain ranges and exposed byfaulting and erosion. It is believed torepresent the frozen remains ofmagma chambers where lava risingout of the earth’s interior pooled deepin the crust. Gabbro is also found in“ophiolite” complexes, bits of fossilocean crust and mantle commonlyfound on land in tectonic plate colli-sion zones and in island arcs (see mapon page 24). There it underliesextrusive pillow lava and sheeteddikes and overlies mantle peridotite; itis also believed to represent the re-mains of magma chambers. Theophiolite sequence of pillow lavas–dikes–gabbros–periodite provides a

convincing match to the seismic layering.However, ophiolite crust is generally much thin-

ner than the seismic crust in the oceans, and thelavas there often have quite different chemicalcompositions than those on the seafloor. Doophiolites represent broader ocean crust character-istics? Hess and geophysicist Walter Munk (ScrippsInstitution of Oceanography, University of Califor-nia, San Diego) cooked up the idea of drilling a holeinto the earth’s mantle and sold it to a group of theircolleagues, known as the “American MiscellaneousSociety,” who in turn sold it to Washington: Drill ahole into the mantle all the way through the oceancrust. Dubbed “Project Mohole,” the idea eventuallyfoundered on the rocks of the congressional budgetoffice in cost overruns and ballooning budgets. Itwas later reborn, however, as the successful DeepSea Drilling Project and its successor, the OceanDrilling Program. (See Oceanus, Vol. 36, No. 4 for“An Abridged History of Deep Ocean Drilling.”)These programs have incrementally sought to drilldeeper, first through the sediments, then throughlayer 2, which proved to be, as thought, pillowbasalts and sheeted dikes, the remains of the lavathat erupted onto the seafloor and its feeder chan-nels. But there the drilling stopped. The material

Author Henry Dick standing next to the Canadian ROVROPOS on site in the Indian Ocean.

OCEANUS • 31

rock hole ever drilled in the oceans. Although weexpected to find gabbro in the lower ocean crust,these rocks still produced a lot of surprises. Scien-tists had long debated whether the lower crust wascrystallized from one great steady-state magmachamber many kilometers across lying beneath theocean ridge. There was no proof for that here—rather there were many small intrusions, hundredsof meters in dimension, not kilometers. Crystalliza-tion here was ephemeral and short-lived. And, agreat surprise—the magmas deep in the earth weremoved around as the rocks deformed when theocean crust stretched and ripped apart duringseafloor spreading. The crust simply looked a lotdifferent than we expected.

The Indian Ocean is a tough place to get to. Ittook 10 years to get back. Finally, Jim (now at theUniversity of Miami) and I returned last fall aboardJOIDES Resolution as co-chief scientists andstarted drilling again to deepen Hole 735B. Wedrilled another kilometer, and surprise—every-thing changed again.

After careful study, the results showed that thedeep ocean crust is compositionally zoned. Its up-per layers are enriched in iron and titanium. In thelefthand figure on page 30, this isreflected in a change in the amountsof different rock types. Many largegabbro intrusions on land are alsocompositionally layered, but not thisway. In this piece of ocean crust, thelayers are created as tectonic forcesdrive the last liquid, with iron andtitanium in solution, out of numeroussmall magma intrusions when theysolidify—pushing it from the lower-most layers upward to “freeze” at thetop of the section. Our discovery ofthis “synekinematic igneous differen-tiation” reveals a whole new processnever before dreamed of—unique toocean ridges.

In the 10 years we had been absent,British colleagues led by Tim Minshull(Cambridge University) had been byto conduct a seismic experiment,which showed that the Moho beneathAtlantis Bank was some 5 to 6 kilome-ters down. This was much deeperthan expected or seemed reasonablein an area where the ocean crust wasbelieved considerably thinner thanthe usual 6 or 7 kilometers, and whereshallow seismic layer 2 (1.5 to 2 kilo-meters) was already gone.

Well, where was the mantle hereanyway? If it were on the walls ofAtlantis Bank, then the Moho be-neath wasn’t the crust-mantle

boundary. That’s whatwe set out to find inspring 1998. To dothis, Paul Robinson(Dalhousie Univer-sity), Chris MacLeod(Cardiff University),Simon Allerton (Uni-versity of Edinburgh),and I organized aninternational expedi-tion on the BritishAntarctic Survey shipJames Clark Ross.Using British GeologicSurvey rock drills andthe Canadian remotely operated vehicle ROPOS(Remotely Operated Platform for Ocean Science),we went back to see if we could find the crust-mantle boundary on the wall of the transform, andto map out the great gabbro massif through whichwe were attempting to drill. What we found wasmore surprises. We did find the crust-mantleboundary—outcropping on the side of AtlantisBank—only a few hundred meters deeper thanwhere we stopped drilling. So it appears the Moho

~50 Kilometers

A (11.7 million years ago)

B (11.5 million years ago)

C (10.5 million years ago)

D (9.5 million years ago)

Rift Valley

Rift Valley

Rift Valley

Rift Valley

Block FaultingRift Mountains

Rift Mountains

Basalt Lavas& Dikes

Gabbros

Gabbros

Gabbros

Gabbros

AlteredPeridotite

Basalt

Mantle Flow

Detachment Fault

Mantle Upwelling

Atlantis Bank

Water

Africa Antarctica

Hypothetical modelfor the unroofing anduplift of Atlantis Bankto sea level frombeneath the South-west Indian Ridgefrom the ScientificResults Volume of theOcean Drilling Pro-gram (Leg 118). Onceuplifted, the islandthen submerged asthe ocean crust onwhich it sat cooledand sank over the last10 million years.

Gulf of AdenArabian

Basin

Bay ofBengal

Carlsberg Ridge

SomaliBasin

CentralIndianBasin West Australian BasinCh

agos

-Lac

cad

ive

Ridg

e

CentralIn

dia

nR

idge

MascareneBasin

MadagascarBasin Perth

Basin

SouthAustralia

BasinCrozet Basin

So

ut h

e a s tI n d i a n

R i d g e

Southwest

In

dian

Ridg

Mad

agas

car

Rid

ge

Mo

zam

biq

ue

Bas

in

K

Atlantis Bank

Highly productivedrilling on AtlantisBank in the IndianOcean gave scientistsa surprising newlook at the lowerocean crust.

32 • Vol. 41, No. 1 • 1998

An outcrop of layeredgabbros is exposed

on a wave cut benchon the top of

Atlantis Bank.

is an alteration front, as Harry Hess once sup-posed, at least in this one spot in the ocean. Andwe did, with our British and Canadian colleagues,map out an enormous block of gabbro over the topof the bank—proving that the crust here is rela-tively intact and not a great tectonic jumble. Butwe also found a fossil island.

And what about that fossil beach? About two-thirds of the bank is covered by limestone, withripple marks, just like those in the sand at a modernbeach. However, these were “frozen,” lithified aswhat was once an island sank beneath the wavesmillions of years ago. There are little pot holesground into gabbro rock, still partially filled withpebbles and sand, and headlands and fossilseastacks (isolated erosional remnants of the is-land). No beach umbrella, but an old rotting fishingnet—the place had a very large population of lob-sters, crabs, sharks, sea fans, siphonophores,sponges, and other critters. It seems that long ago,the whole massif popped up out of the earth some 6or 7 kilometers at the intersection of the SouthwestIndian Ridge and the Atlantis II Fracture Zone torise above sea level. This great tectonic island withan area of at least 25 square kilometers then slowlysubsided back beneath the waves to its presentposition some 700 meters below the sea surface. It isa remarkable place, both for its modern biologiccommunity and its unique fossil island characteris-tics. Flat topped seamounts called guyots are wellknown in the Pacific. They are submerged volcanicislands whose tops are worn away by wave action asthey sink beneath the sea surface. But this is thefirst tectonic guyot anyone has ever studied. Andwith it comes a unique paleontological record.Between the many seastacks, boulders, and ledgesof gabbros, we deployed our short, over-the-siderock drills—and came back with a lot of limestonein addition to the gabbro we sought. There werefossil clams, snails, a limestone made of sea urchinspines, and a kind of limestone known as oolitic,something that forms in lagoons. But they’re allnow 700 meters down. Jim, formerly a Moho chaserlike myself, found it so fascinating that he has con-verted himself into a carbonate sedimentologist toexplore the mysteries of that fossil island.

This research was funded by the US National ScienceFoundation with additional support for the author from WHOI’sVan Allen Clark, Jr., Chair. International support was providedby the Canadian Natural Sciences and Engineering ResearchCouncil and the British Natural Environment Research Councilfor remotely operated vehicle and rock drilling operations atAtlantis Bank aboard RSS James Clark Ross.

Henry Dick first came to Woods Hole Oceanographic Institutionfrom Yale University in 1975 as a postdoc with Bill Bryan of theGeology and Geophysics Department. He has been a SeniorScientist since 1990. He has a remarkably dedicated wife namedWinifred and three small children, Helene, Spencer, and Lydia,who think their daddy goes to sea too much. He’s promised themhe won’t even look at the ocean for at least a year, much less gethis feet wet—after his next cruise this fall!

A pothole on awave cut ledge of

Atlantis Bankcontains a pebble.

Fossil ripple marksindicate where

Atlantis Bank beachsands have hardenedinto limestone with a

light dusting ofmodern sediment.

Seafloor photographof Atlantis Bank

shows a wave-cutoutcrop of gabbro

smoothed and pol-ished by the sands

with complex layer-ing emerging from

the fossil beach some700 meters down on

the seafloor.