fume 34, Number 4^Winter 1991/92 *^N -\ V r . . Mid-Qcean Ridges \ Index to Volume 34 (1991) Number 1, Spring Ocean Engineering & Technology Introduction: Ocean Engineering, Albert J. Williams 3rd High Resolution Optical and Acoustic Remote Sensing for Underwater Exploration, W. Kenneth Stewart SOFAR Floats Give a New View of Ocean Eddies, Philip L. Richardson Robotic Undersea Technology/, Dana R. Yoerger A Telescope at the Bottom of the Sea, George Wilkens, John Learned, and Dan O'Connor Ocean Data Telemetry: New Methods for Real-Time Ocean Observation, Daniel E. Frye, W. Brechner Owens, James R. Valdes The Role of the Microcontroller in Ocean Research Instruments, Albert M. Bradley The History of Salinometers and CTD Sensor Systems, Neil Brown Underwater Technology in the USSR, Deam Given Toward a Global Ocean Observing System, D. James Baker Modernizing NOAA's Ocean Service, Virginia K. Tippie and John H. Cawley Artificial Reefs: Emerging Science and Technology, Iver W. Duedall and Michael A. Champ Douglas Chester Webb: A Profile, Henry M. Stommel Number 2, Summer An Open Door: Soviet-American Cooperation From the Editor: An Open Door: Soviet-American Cooperation in Marine Science, Vicky Cullen Diving the Soviet Mir Submersibles, Cindy Lee Van Dover The Oceans and Environmental Security, James M. Broadus and Raphael V. Vartanov The History of Soviet Oceanology, Leonid M. Brekhovskikh and Victor G. Neiman Living Marine Resources, Viatcheslav K. Zilanov The USSR and the International Law of the Sea, Yuri G. Barsegov Soviet Polar Research, Arthur Chilingarov Exploring Pacific Seafloor Ashore: Magadan Province, USSR, Wilfred B. Bryan Developing a New Soviet Ocean Policy, Raphael V. Vartanov Dynamics of Ocean Ecosystems: A National Program in Soviet Biooceanology, Mikhail E. Vinogradov Satellite Oceanography, Vladimir V. Aksenov and Alex B. Karasev Good Morning, Comrades, Hugh D. Livingston and Stella J. Livingston Physical Oceanography: A Review of Recent Soviet Research, Yuri A. Ivanov A History of USSR-US Cooperation in Ocean Research, N.A. Ostenso, A. P. Metalnikov, and B.I. Imerekov Number 3, Fall Reproductive Adaptations in Marine Organisms An Introduction to Reproductive Adaptations in Marine Organisms, Lisa Clark Caribbean Reef Corals, Alina M. Szmant and Nancy J. Gassman Mating Strategies of Coastal Marine Fishes, Phillip S. Lobel Sex (and Asex) in the Jellies, Katherine A.C. Madin and Laurence P. Madin Larval Forms with Zoological Verses, Walter Garstang, illustrated by Rudolph Scheltema The Story of the Coelacanth, Keith S. Thomson Elasmobranch Fish: Oviparous, Viviparous, and Ovoviviparous, Carl A. Luer and Perry W. Gilbert Challenging the Challenger, Craig M. Young Hydrothermal Vent Plumes: Larval Highways?, Lauren S. Mullineaux, Peter H. Wiebe, and Edward T. Baker Photoessay: A World of Art Beneath the Waves, Kathy Sharp Frisbee Number 4, Winter Mid-Ocean Ridges Introduction — Mid-Ocean Ridges: The Quest for Order, Ken C. Macdonald The Segmented Mid-Atlantic Ridge, Jian Lin Modeling Ridge Segmentation... A Possible Mechanism , Hans Schouten and Jack Whitehead RIDGE: Cooperative Studies of Mid-Ocean Ridges (plus Box on InterRIDGE), Donna Blackman Map of Mid-Ocean Ridges and Research Locations Ridges and Rises: A Global View, Peter Lonsdale and Chris Small Onions and Leaks: Magma at Mid-Ocean Ridges, A Very Personal View, Joe Cann From Pillow Lava to Sheet Flow, Evolution of Deep-Sea Volcanology, Wilfred B. Bryan Tectonics of Slow-Spreading Ridges, Jeffrey A. Karson Mid-Ocean Ridge Seismicity, Eric A. Bergman Hydrothermal Vent Systems, Margaret K. Tivey The Biology of Deep-Sea Vents and Seeps Alvin's Magical Mystery Tour, Richard A. Lutz Megaplumes, Edward T. Baker Tomographic Imaging of Spreading Centers, Douglas R. Toomey Bruce C. Heezen, A Profile, Paul J. Fox I The World's Most Accurate CTD We go to sea with our instruments, get involved in the science, and continuously apply what we learn toward refinement. As a result, the Sea-Bird 9/11 CTD offers unmatched precision, superior static and dynamic accuracy, and exceptional reliability. We're serious about CTD performance. Call us. SBE SEA-BIRD ELECTRONICS, INC. Telephone (206) 643-9866 1808 -136th Place Northeast Bellevue, Washington, 98005 USA Telex: 292915 SBEI UR Fax (206) 643-9954 MID-OCEAN RIDGES 9 Introduction Mid-Ocean Ridges: The Quest for Order Ken C. Macdonald The last decade has brought significant advances in understanding of the seafloor and its spreading processes. nThe Segmented Mid-Atlantic Ridge Jian Lin A recent Mid- Atlantic Ridge expedition and other studies contribute to expanding knowledge of ridge segmentation. page 11 Ridge Segmentation: A Possible Mechanism Hans Schouten and Jack Wlritehead A laboratory experiment with glycerine and water provides a model for ridge segmentation resulting from the rise of hot mantle material. RIDGE and InterRidge Donna Blackman and Trileigh Stroh RIDGE is a cooperative effort to study the mid- ocean ridges and InterRidge is its international counterpart. Ridges and Rises: A Global View Peter Lonsdale and Chris Small An overview of current knowledge of the patterns, mechanisms, and the relief of mid-ocean ridges. page 68 Onions and Leaks: Magma at Mid-Ocean Ridges "% Joe Cann \^S\J A 35-year review of a dynamic period of Earth science and ridge models succeeding ridge models. M ^% From Pillow Lava to Sheet Flow J Evolution of Deep-Sea Volcanology "^— Wilfred B. Bryan An historic, current, and future look at knowledge of the rocks that make up mid-ocean ridges. page 84 Copyright © 1991 by the Woods Hole Oceanographic Institution. Oceanus (ISSN 0029-8182) is published in March, June, September, and December by the Woods Hole Oceanographic Institution, 9 Maury Lane, Woods Hole, Massachusetts 02543. Second-class postage paid at Falmouth, Massachusetts and additional mailing points. POSTMASTER: Send address change to Oceanus Subscriber Service Center, P.O. Box 6419, Syracuse, NY 13217-6419. Oceanus Headings and Readings 51 Tectonics of Slow-Spreading Ridges Jeffrey A. Karson Slow-spreading and fast-spreading ridges build structure that are quite different from one another. Mid-Ocean Rise Seismicity Eric A. Bergman Seismic waves signal earthquake locations and expand knowledge of ridge structures. page 75 "\ Hydrothermal Vent Systems Margaret K. Tivey In the 15 years since the first "black smoker" was sighted, much has been learned of hydrothermal vents. The Biology of Deep-Sea Vents and Seeps Richard A. Lutz \te>' Extensive submersible work in the past two years has brought new knowledge of deep-sea vent and seep communities. Megaplumes fc / 1 ... Edward T. Baker The megaplume dectectives are on the case studying a recently discovered vent phenomenon. page 100 Editor's Note I Glossary ', Map: Ridges & Rises 24 Books & Videos 108 Creature Feature 112 The Oceanus Annual Index may be found on the inside front cover. "\ ^^ Tomographic Imaging of Spreading Centers J Douglas R. Toomey ^X sLum A new tool yields three-dimensional images of Earth's dynamic processes working deep within mid-ocean ridge spreading centers. Profile Bruce C. Heezen Paul J. Fox A man of extraordinary vision and enormous research capacity changed thinking about the seafloor. ON THE COVER: An artist's concept of the mighty Mid-Atlantic Ridge and a glimpse of the Pacific Ridge are highlighted in the colors scientists use to indicate elevation. (Watercolor by E. Paul Oberlander, WHOI Graphics) Winter 1991/92 Vicky Cullen Editor Lisa Clark Assistant Editor Kathy Sharp Frisbee Editorial Assistant Robert W. Bragdon Advertising & Business Coordinator Lisa Poole Publishing Intern The views expressed in Oceanus are those of the authors and do not necessarily reflect those of Oceainif magazine or its publisher, the Woods Hole Oceanographic Institution. Editorial responses are welcome. Please write: Oceanus Magazine Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543, or telephone: (508) 457-2000, extension 2386. Members within the US, please write: Oceanus Subscriber Service Center, P.O. Box 6419, Syracuse, NY 13217-6419. Individual membership rate, $25 per year; Students, $20; Libraries and Institutions, $50. Current issue price, $6.25; 25% discount on current issue orders for five or more. Please make checks payable to the Woods Hole Oceanographic Institution. Subscribers outside the US, please write: Occninis, Cambridge University Press, The Edinburgh Building, Shaftesbury Road, Cambridge CB2 2RU, England. Individual subscription rate: £24 per year; Students, £15; Libraries and Institutions, £43. Single- copy price, £7. Please make checks payable to Cambridge University Press. Canadian subscriptions available through Faxon at P.O. Box 2382, London, Ontario N6A 5A7, Canada. Add $5 to basic rates. When sending change of address, please include mailing label. Claims for missing numbers from the US will be honored within three months of publication; overseas, five months. Oceanus and its logo are ® Registered Trademarks of the Woods Hole Oceano- graphic Institution. All Rights Reserved. Oceanus International Perspectives on Our Ocean Environment Volume 34, Number 4, Winter 1991 /92 ISSN 0029-8182 193O Published Quarterly by the Woods Hole Oceanographic Institution Guy W. Nichols, Chairman of the Board of Trustees James M. Clark, President of the Corporation Charles A. Dana, III, President of the Associates Craig E. Dorman, Director of the Institution Sallie K. Riggs, Director of Communications Editorial Advisory Board Robert D. Ballard, Director of the Center for Marine Exploration, WHOI James M. Broadus, Director of the Marine Policy Center, WHOI Henry Charnock, Professor of Physical Oceanography, University of Southampton, England Gotthilf Hempel, Director of the Alfred Wegener Institute for Polar Research, Germany John Imbrie, Henry L. Doherty Professor of Oceanography, Brown University John A. Knauss, US Undersecretary for the Oceans and Atmosphere, NOAA Arthur E. Maxwell, Director of the Institute for Geophysics, University of Texas Timothy R. Parsons, Professor, Institute of Oceanography, University of British Columbia, Canada Allan R. Robinson, Gordon McKay Professor of Geophysical Fluid Dynamics, Harvard University David A. Ross, Senior Scientist and Sea Grant Coordinator, WHOI Permission to photocopy for internal or personal use or the internal or personal use of specific clients is granted by Oceanus magazine to libraries and other users registered with the Copyright Clearance Center (CCC), provided that the base fee of $2 per copy of the article is paid directly to CCC, 21 Congress Street, Salem, MA 01970. Special requests should be addressed to Oceanus magazine. ISSN 0029-8182/83 $2.00 + .05. GST R 101 749 364. Oceanus From the Editor On Mid-Ocean Ridges he theory of plate tectonics, the idea that the surface of the earth is made up of eight large and several small rigid plates that are in constant motion (at least in geologic time), was born in the late 1960s, a synthesis of the concepts of continental drift and seafloor spreading. Observations on the apparent fit of the bulge of eastern South America into the indentation of Africa date back at least 300 years. The first detailed theory of continental drift was proposed in 1912 by Alfred Wegener, a German meteorologist. He suggested that a single supercon- tinent he called Pangaea existed through most of geological time and that it began to break up about 180 million years ago. In 1937, Alexander DuToit, a South African geologist, suggested that rather than one pri- mordial continent perhaps there were two, Gondwanaland in the south and Laurasia in the north. Neither proposed a mechanism for continental motion. New work in the 1950s brought mounting evidence for continental drift. During the early 1960s the theory of seafloor spreading was advanced by US geophysicist Harry Hess, who suggested that new crust is con- tinually being generated by volcanic activity at the crests of mid-ocean ridges. In a 1970 paper, US scientists Robert Dietz and John Holden reconstructed Pangaea and de- scribed a plausible sequence of continental dispersion, depicted overleaf, over the past 200 million years. "Continental drift,"' they wrote, "is a necessary consequence of plate tectonics in that the continents would be passively rafted on the backs of the conveyor-belt-like crustal plates. The drift of the continents may be conveniently thought of as a summation of sea-floor spreading." Plate motion and the tectonic forces that create and destroy the plates are still being defined and understood, and we share with you in this issue of Oceanus some of the continuing excitement among Earth scientists about these forces and the phenomena they create. It is along the mid-ocean ridges, as Harry Hess surmised, that the crustal plates are created as molten material, called magma, from deep beneath the surface rises to fill the gaps or rifts created between plates that are moving or "spreading" apart. Our authors tell us that different spreading rates result in different surface expressions: The slowly spreading Mid- Atlantic Ridge, where about 30 millimeters of crust are created annually, - rises steeply from surrounding seafloor and has a characteristic depres- Boundaries of the large and some of the small crustal plates are diagrammed here. Winter 1991/92 PERMIAN -225 million years ago TRIASSIC -200 million years ago JURASSIC -135 million yearsago Acknowledgment: The Oceanus staff thanks G. Michael Purdy, Woods Hole Oceanographic Institution Depart- ment of Geology and Geophysics Chairman, for his advice on this issue. sion at its crest that is called a rift valley, while the fast-spreading East Pacific Rise that expands some 60 to 170 millimeters per year presents a more gently rolling topography and no rift valley. Seawater circulates down through the porous new volcanic crust, heating as it moves and accumulating elements absorbed from the rock. Eventually the heated, altered water rises again to erupt through vents in the seafloor and support colonies of unusual animals. The new crust being created at ridges is balanced by destruction of crust on the opposite sides of the plates where deep-sea trenches mark subduction zones, areas where the heavier of two colliding plates is shoved back down toward the center of the earth. Characteristics of theses zones include curving island chains or island arcs, such as the Aleutian Islands and the islands of Japan, and volcanoes born of the melting edges of subducting plates. Alternatively, when there is no subduction, mountains result, such as the Himalayas that mark the colliding boundaries of the Indian and Eurasian plates. Readers new to these concepts may find it helpful to begin with a review of the glossary that begins opposite and the profile of Bruce Heezen (page 100) in which Heezen's student, Jeff Fox, describes the meticulous assembly of mid-ocean ridge topography by Heezen and Marie Tharp. Their successive "physiographic maps" revealed the sharp relief of a seafloor previously thought to be largely flat and uninterest- ing. A world map on pages 24 and 25 shows ridges and research areas to help readers follow this issue's wide-ranging discussions of ridge research. Author Joe Cann (Onions and Leaks. . .on page 36) notes that early in his career it was still possible to become well acquainted with every scientist working in marine geology while now that would be quite impossible. "Thirty-five years ago," he writes, "most geologists were secure in the knowledge that continents did not move." As we mark just over two decades since the theory of plate tectonics gained wide accep- tance, we invite you to join us for a review of the exciting, and still new, realm of a planet in motion. Oceanus CRETACEOUS -65 million years ago CENOZOIC Present Glossary accreting plate boundary— the border between two separating crustal plates where new oceanic lithosphere or crust is being created asthenosphere-the layer of the earth that extends roughly from 100 to 300 kilometers below the surface where temperature and pressure cause the rock to flow plastically. axial — relating to a line that bisects a mid-ocean ridge. For example, an axial rift valley runs down the center of the Mid-Atlantic Ridge. basalt — medium gray to black igneous rock that constitutes the upper- most 2 to 3 kilometers of oceanic crust and is the chief component of isolated oceanic islands, rich in iron, magnesium, and calcium bathymetry — measurement and charting of ocean depths using echo soundings plotted on a chart to show seafloor contours crust — the outermost layer of the earth, 6 to 8 kilometers thick beneath the ocean and 30 to 35 kilometers thick beneath the continents diapir — a vertical columnar plug of less dense rock or magma that has risen through more dense rock dike — a tabular body of igneous rock that intrudes pre-existing struc- tures (see photo page 38) echo sounding — generation of sound in water and recording the time lapse of the return or echo of the sound from a reflecting surface as a measure of depth fault — rock fracturing that displaces the sides of the fracture relative to one another. Fault movement may be continuous creep or a series of abrupt jumps (earthquakes). fault scarp — steep cliff formed by movement along one side of a fault fissure — an extensive crack in a rock formation fracture zone — area surrounding a large fault that crosses and displaces a mid-ocean ridge, often the site of intense seismic activity gabbro — a group of granular, dark-colored igneous rocks composed largely of plagioclase and clinopyroxene graben — a crustal block that is depressed relative to neighboring blocks, which are called horsts horst — a crustal block that is raised relative to neighboring blocks, which are called grabens hot spot — heat source from deep within the earth's mantle, surface manifestation of a rising plume of hot mantle material hydrothermal — relating to heated water and its actions or products igneous rock — rock of several types formed of molten material (magma) that upwells from the deeper part of Earth's crust and comprises most of the oceanic crust World maps similar to these were published in 1970 by Robert Dietz and John Holden depicting the breakup of the primordial super- continent Pangaea and subsequent dispersion of continents over the past 200 million years. The action began when the southwest Indian Ocean rifted, splitting West Gondwana (South America and Africa) from East Gondwana, and India lifted off Antarctica. Then Laurasia (Nortli America and Eurasia) separated from South America and the bulge of Africa. Later South America and Africa split. Spain rotated to form the Bay of Biscay, and Madagascar split from Africa. India continued its north- ward trek, and Austra- lia separated from Antarctica. As Antarc- tica rotated westward, Australia made a remarkable northward flight and New Zealand dropped off its east coast. The North and South Atlantic oceans continued to open, Greenland parted company with Europe, Africa moved slightly northward and rotated, and India collided with Asia, raising the Himalayan Mountains. Winter 1991/92 hydro thermal- relating to heated water and its actions or products hydrothermal vent isostasy — state of equilibrium with the earth's crust buoyantly sup- ported by the plastic material in the mantle lithosphere — Earth's outer shell including the crust and uppermost rigid layer of mantle. In plate tectonics, the lithospheric plates move over the plastic asthenosphere below. magma — molten, mobile rock that is the product of melting deep within the earth's crust or upper mantle, the source of igneous rocks. Lava is magma that reaches the earth's surface. Magma that solidifies below the surface is called intrusive or plutonic and that emerging and solidifying above the surface is called extrusive or volcanic rock. mantle — zone of the earth extending from below the crust to the core. The upper mantle extends to 400 kilometers depth followed by a transition zone from 400 to 1,000 kilometers and the lower mantle from 1,000 to 2,900 kilometers. Moho — abbreviation of Mohorovicic discontinuity, the boundary between the crust and the mantle offset — horizontal displacement of a topographic trend, commonly along a fault; also a spur or branch from a mountain range ophiolite — masses of igneous rocks of oceanic crustal origin that have been pushed up onto continents by plate collisions peridotite — coarse-grained igneous rock thought to be the primary component of the upper mantle, often associated with ophiolites rift valley — the deep central cleft with a mountainous floor in the crest of a mid-ocean ridge. The valley results from plate separation; at fast- spreading ridges upwelling magma fills the rift and smooths the topography while at slow-spreading ridges the upwelling magma does not fill the rift but adheres to the trailing edge of the separating plates. rift — a narrow opening in a rock caused by cracking or splitting Ring of Fire — chain of volcanoes occurring in a rough circle around the perimeter of the Pacific Ocean scarp — sequence of cliffs resulting from faulting seep — place of contact between deep-sea sediments and limestone walls where hypersaline waters seep onto the seafloor and feed sulfide- dependent biological communities seismic waves — the form (like sound waves) of the energy released by fracturing or abrupt slipping of rock along fault planes during an earthquake. Seismic waves provide valuable information about the regions they travel through; most importantly they map reflecting discontinuities, and measurement of the velocity at which the waves travel through different layers of rocks allows inferences to be made concerning the extent of various rock types within the earth. sheeting — ruptures in massive rocks characterized by tabular surfaces that are slightly curved and parallel to the topographic surface shield volcano — gently sloping volcano built by flows of very fluid basaltic lava erupted from a large number of closely spaced vents and fissures strike — the direction taken by a structural surface, directional trend strike-slip — movement parallel with the strike of a fault Glossary continues on page 112 8 Oceanus Introduction Mid-Ocean Ridges: The Quest for Order Ken C. Macdonald very year, the chain of active volcanoes that comprises the mid- ocean ridge erupts, on average, ten times as much lava as the dramatic and disastrous Mt. St. Helens eruption in 1980. This is enough lava to pave the entire US Interstate freeway system with a layer of rock 10 feet thick. The largest and most volcanically active chain of mountains in the solar system, the mid-ocean ridge wraps around the globe for over 70,000 kilometers, much like the seam of a baseball. Along the ridge, brittle plates that comprise Earth's surface separate at rates of 10 to 170 millimeters per year. As the plates move apart, rock melts, separates from the solid residuum, and wells up from tens of kilometers deep. Some of the molten rock ascends all the way to the seafloor, pro- ducing extensive volcanic eruptions and building volcanoes, while the rest adheres to the edges of the parting plates. The late Bruce Heezen (see profile on page 100) aptly called this world- encircling system of ridges "the wound that never heals." Over the last decade, an ex- traordinary confluence of diverse observations at mid-ocean ridges has led to a series of advances in our understanding of the seafloor and its spreading processes. Swath-mapping tools have been developed that can image large ar- eas of the deep seafloor accurately. Structural maps based on these charts, combined with geochemical studies of rock samples, seismic and gravitational studies of veloc- ity and density variations beneath the ridge, seafloor magnetization studies, and near-bottom imaging of hydrothermal-vent distribution, have revealed a fundamental parti- tioning of the ridge into segments bounded by discontinuities. These segments behave like giant cracks in the seafloor that can lengthen or shorten, and have cycles of in- creased volcanic, hydrothermal, and tectonic activity. Most observations support the concept of a hierarchy in the Merging Sea Beam and SeaMARC II swath bathymetry produced this shaded-relief image of the Office of Naval Research East Pacific Rise Natural Laboratory. In the fore- ground is the Siqueiros transform fault (a first-order disconti- nuity) and the 8°20'N seamount chain; the fast-spreading East Pacific Rise and the 9°N overlapping spreading centers are in the middle; and the Clipperton Transform Fault is in the background. Notice the numerous seamount chains. The actual image is approximately 300 by 300 kilometers, viewed toward the northeast. (This image is based on data from expeditions funded by the US Office of Naval Research.) Winter 1991/92 Is the architecture of the global mid-ocean ridge system really so orderly, or is this concept of a "segmentation hierarchy" merely a human construct? segmentation of mid-ocean ridges. First-order segments are generally hundreds of kilometers long, persist for millions to tens of millions of years, and are bounded by relatively permanent, rigid, plate-transform faults (first-order discontinuities). A first-order segment is usually divided into several second-order segments. These segments are shorter, survive for less than several million years, and are bounded by nonrigid, second-order discontinuities that can migrate along the length of the ridge. Thus second-order segments lengthen, shorten, and even disap- pear. There are third- and fourth-order segments (and discontinuities bounding them) that are increasingly short, short lived, and peripatetic. For example, fourth-order segments, approximately 10 kilometers long, may survive as distinct structures for only 100 to 10,000 years. The longevity of individual segments and associated cycles of volcanic/ hydrothermal/ tectonic activity must influence the distribution and survival of exotic faunal communities that flourish at mid-ocean ridge hot springs (see The Biology of Deep Sea Vents and Seeps, page 75). For example, a violent eruption on the East Pacific Rise near 9°50'N in March and April 1991 wiped out a large community of tube worms, mussels, and other benthic fauna (and might have done the same to divers in the submersible DSV Alvin who arrived only hours to days later!). Thus, amidst frequent volcanic eruptions and seafloor temblors, there seems to be an orderly spatial and temporal pattern to magmatic, volcanic, hydrothermal, and tectonic processes associated with the birth of new ocean floor. Is the architecture of the global mid-ocean ridge system really so orderly, or is this concept of a "segmentation hierarchy" merely a human construct? To be sure, this model may be vastly modi- fied or even abandoned, as new information and new minds contribute to the ongoing debate. In his superb book, The Principles of Physical Geology, Arthur Holmes recalled the words of Alfred North Whitehead, which are still appropriate to our exploration of the Mid-Ocean Ridge today: "There can be no living science unless there is a widespread instinctive conviction in the existence of an Order of Things and, in particular, of an Order of Nature." Ken C. Macdonald is Professor of Marine Geophysics at the University of California, Santa Barbara, and a member of the Woods Hole Oceanographic Institution Corporation. Over the last 20 years he has focused on the tectonics on mid-ocean ridges and has been fortunate enough to participate in some of the first explorations of the ridge, using swath-mapping systems, remotely controlled vehicles, and submersibles. 10 Oceanus The Segmented Mid- Atlantic Ridge Jian Lin early three decades ago, in 1964, an Oceanus article by Richard M. Pratt described an exciting R/V Chain expedi- tion to the Mid-Atlantic Ridge. On echo-sounding profiles across the ridge crest some 2,500 meters beneath the ocean surface, Pratt and his colleagues saw familiar mountains and valleys on the ocean floor. But a peculiar feature caught his eye: The rift valley in one area had shifted laterally for tens of kilometers. Pratt, a Woods Hole Oceanographic Institution (WHOI) scientist, speculated in his article (Volume 11, Number 2, December 1964) that the shift of the ridge may have been caused by a "transverse" feature of unknown origin. In the early 1960s, H. William Menard and Bruce Heezen discovered similar features on other parts of the Mid-Atlantic Ridge and on the East Pacific Rise. In 1965, J. Tuzo Wilson identified these transverse features as "transform faults:" boundaries formed perpendicular to the length of the Ridge. The anomalous transverse feature noted by Pratt is now known as the Atlantis Transform Fault. It is just one of many that offset the 60,000-kilometer-long global mid-ocean ridge system. The discovery and recognition of transform faults played an essential role in the develop- ment of the plate-tectonic theory in the late 1960s and early 1970s. In plate-tectonic theory, Earth's outer 100 to 250 kilometers, called the lithosphere, breaks up into a set of rigid plates that move with respect to each other. The lithospheric plates, such as those of North America and Africa, drift over underlying, less rigid asthenosphere, According to the theory of plate tectonics, Earth's lithosphere is broken into plates that move with respect to each other. The plates originate at mid-ocean ridges (A), subduct into the underlying astheno- sphere at trenches (B), and slide by each other at transform faults (C). ASTHENOSPHERE > MESOSPHERE Winter 1991/92 11 A computer-generated relief image of the rift valley of the Mid- Atlantic Ridge near 29°50'N (view towards the south). The rift valley is 20 to 30 kilometers wide and a few kilometers deep. The inner rift valley is covered by elongated volcanic hills and circular volcanoes. Large steplike normal faults run along the ridge axis, here shown at three-times vertical exaggeration. 28°N Rift Valley Rift Valley 42°W 43°W 1000 Speading Center Axis PR. Shaw/WHOI much like icebergs floating in the ocean. Plates are created at mid-ocean ridges, are consumed at subduction trenches, and slide by each other along transform faults. When plates interact at their boundaries, earthquakes strike, volcanoes erupt, and mountains grow. The plate-tectonic theory provided a fundamen- tal link between global tectonics, from ridges to trenches and from continents to ocean floors. It unified geology in the same way that the principle of evolution unified biology. The development of the rigid-plate concept and early, sparse observations of ocean ridges led to a simple idea of how the ridges worked. They were thought to be linear spreading segments, periodically offset by trans- form faults. Each spreading segment was hundreds of kilometers long and had the same two-dimensional In this computer-generated relief image of the Mid-Atlantic Ridge between 28°N and 30°45'N, thin red lines along the rift valley approxi- mate spreading axes based on magnetic data. Note the prominent rift valley, deep Atlantis Transform Fault trough (just below 43° W), and small non-transform offsets (regions where red lines do not meet). 12 Oceanus Atlantis Transform Fault I cross-sectional view along its strike. Today, two decades after the birth of plate-tectonic theory, this view is rapidly changing. This article begins with a report on a recent expedi- tion to the Mid-Atlantic Ridge, where Pratt visited 30 years ago. Using this and other recent studies as background, I will review current ideas on seafloor spreading at mid-ocean ridges and explain how the earlier two-dimensional ridge model must be expanded to allow for variations along strikes and with time. The Mid-Atlantic Ridge The huge Mid-Atlantic Ridge (MAR) mountain range runs down the middle of the Atlantic Ocean from Iceland in the north to near Antarctica in the south. Since it was first studied in 1873 by the British survey ship HMS Challenger, the ridge has been the focus of intense scientific curios- ity. Perhaps the most detailed survey was Project FAMOUS (French- American Mid-Ocean Ridge Undersea Study) of the early 1970s, in which oceanographers investigated a 100-kilometer-long stretch of the rift valley near 37°N. Early exploration of the Mid-Atlantic Ridge only identified trans- form faults of very large offsets. From observations of Earth's magnetic field, however, oceanographers in the early 1980s proposed that the MAR was composed of a string of about 50-kilometer-long spreading segments separated by small "zero-offset" transform faults. To look closely at these "zero-offset" features, in 1988 and 1989, scientists from WHOI and the University of Washington examined a 900- kilometer-long stretch of the ridge. Our survey, which started near the Atlantis Transform Fault, included detailed ocean floor mapping and precise measurements of Earth's gravitational and magnetic fields. Our sonar sounding system, called Sea Beam, was much more capable than that used by Pratt in the early 1960s: It can map a 2-kilometer- wide This relief image of the Atlantis Transform Fault is viewed toward the east. The deepest parts of the trajisform valley are more than six kilometers below sea level. The vertical relief from the top to the bottom of the valley is more than four kilometers, and is shown at three-times vertical exaggeration. The red vertical bars on the top-right corner are an imaging artifact. Winter 1991/92 13 Seismic Velocity (m/s) 4.5-5.5 6.5-7.0 7.8-8.4 1 Below a ridge, the mantle of the astheno- sphere (orange) rises to fill the gap between two separating lithospheric plates (blue). As they rise, some rocks melt to form magmas. The buoyant magmas or melts then surge into a magma chamber (red). Material in the magma chamber further segregates into various layers of the oceanic crust (dark green). The crust is less dense than the mantle. In the mantle, density decreases with increas- ing temperature and with depth. swath of the seafloor in a single pulse. Never before had oceanographers seen such a long stretch of a slow-spreading ridge with such high- resolution sonar. With Sea Beam sending back numerous ocean-floor images, we soon recognized many familiar features. The central rift valley, which is 20 to 30 kilometers wide and a few kilometers deep, runs nicely along the ridge crest. Volcanic hills and circular volcanoes — each tens to a few hundreds meters tall — blanket the rift valley's inner floor. The crust cracks along steep steps of normal faults, some of which run along the ridge for tens of kilometers. We were intrigued by the im- mense size of the trough inside the Atlantis Transform Fault. This fault offsets the ridge axis by almost 70 kilometers, and the deepest parts of its floor are more than six kilome- ters below sea level. Its vertical relief from top to bottom is more than four kilometers, or twice the depth of the Grand Canyon. Further down the ridge axis, however, we were totally sur- prised: The "zero-offset" features that we were searching for were not transform faults after all. The ridge breaks into many 20- to 80-kilometer-long spreading segments, but these segments often overlap one another at "nontransform" offsets. Unlike a transform fault, a nontransform offset does not contain a trough perpendicular to the ridge axis. By themselves, these nontransform offsets constitute a new type of unstable, transitory plate boundary. Why do the MAR volcanic chains break into short spreading seg- ments? Is this phenomenon related to deep structures of the earth beneath the spreading axis? To answer these questions, we must first examine how oceanic crust is generated. The Origin of Oceanic Crust The mantle beneath lithospheric plates, known as the asthenosphere, creeps plastically because its temperature stays near its melting point. Below a ridge, however, the asthenosphere rises to fill the gap between two separating plates. While ascending, some mantle rocks fuse to form basaltic magmas, or melts, and the buoyant magmas float to the top of the mantle to form oceanic crust. Meanwhile, the unmelted mantle residual accretes to the oceanic lithosphere bottom. From studying the chemical composition of rocks dredged from the ocean floor, oceanogra- phers have determined that melts are produced at depths of 20 to 80 kilometers and at temperatures of 1,150° to 1,400°C. Theoretical models further suggest that the rising asthenosphere reaches its maximum velocity in a partial melting zone, inside which the mantle has its mini- mum density and viscosity. Buoyant melts from all depths surge into a magma chamber at the 14 Oceanus Transform Zones \ CD .C CL (f> O Asthenosphere base of the crust. Inside the chamber, melts further separate into layers according to their densities. The least-dense lava erupts to form volca- noes on the ocean floor. The most-dense, called gabbro, accumulates at the chamber floor to form the lower crust. Between these two layers lies a layer of vertical dikes, narrow slabs of cooled melt that have risen to fill fissures in the crust. The end product of the melt segregation process, therefore, is a stable layering of light crust (its density is expressed as 2.5 to 2.8 grams per cubic centimeter) overlying heavy mantle (3.1 to 3.3 grams per cubic centimeter). On the other hand, the partially molten astheno- sphere under the lithospheric plates is not stable. This is be- cause its density and viscos- ity are less than that of the overlying lithosphere, a situ- ation analogous to a layer of high-density fluid overlying low density fluid. Laboratory experiments show that if the density and viscosity con- trasts between two fluids are great enough, the less-dense fluid will rise and protrude into the upper layer, in the form of regularly spaced dia- pirs (see Box, page 19). In the mid-1980s, oceanographers proposed that diapirs may occur below mid- ocean ridges. They reasoned that the partial melting zone of the astheno- sphere can develop gravitational instability, inducing regularly spaced diapirs of melts. The melt diapirs then percolate toward the surface to form discrete magma chambers that feed individual spreading segments. Such diapir-induced segmentation models predict that ocean-floor topography should be shallowest at segment centers and deepest at seg- ment boundaries. The models also predict that crustal thickness, which is a measure of melt production, should be greatest at segment centers and decrease toward segment edges. The first prediction was confirmed readily by detailed ridge-crest topography, including that obtained in our survey. To confirm the second prediction, however, techniques for prob- ing into the earth are required. Two commonly used techniques are the studies of Earth's gravitational field and of seismic waves. Gravity, Seismic, and Faulting Evidence Geophysicists often use sensitive gravity meters to probe unseen mate- rial below ground. To examine the crust and mantle below spreading segments, we must first employ modeling to remove the gravitational effects of seawater and a model crust. The leftover signal, called the mantle Bouguer anomaly, should then reveal information about the mantle. Using this technique, researchers have detected an unusual "bull's-eye" shaped gravity low over a spreading segment in the South Atlantic. During our 1988 to 1989 survey, we found a string of such Crust Melt Depleted Mantle Mantle In this model of magma diapirs beneath a ridge, the partially molten asthenosphere (red) is not stable under the cold lithos- phere (green). The gravitational instabil- ity of this partial melting zone will induce regularly spaced diapirs of magmas. The magma diapirs then percolate toward the surface to form discrete spreading segments. Winter 1991 J92 15 44°W 30°N • III (MGAL) Gravity data reveals gravity lows beneath the spreading segments of the North Mid- Atlantic Ridge. This pattern may be caused by thicker crust or less dense mantle beneath the midpoints of the spreading segments. bull's-eyes in the North Atlantic. In both the North and South Atlantic, most of the circular regions of gravity lows are centered near the shallow middle points of the spread- ing segments. In contrast, large positive values are located over the Atlantis Transform Fault and the nontransform offsets. There are two possible explanations for this gravity pattern. The first is that the crust, which is less dense than the mantle, decreases in thickness from segment centers to segment edges. The second is that the mantle beneath the 42°W segment mid-points is of unusu- ally low density, most likely due to a combined effect of high temperature, the presence of limited melts, and density changes in the melted mantle. Both possi- bilities are consistent with the concept of diapir-induced seg- mentation. Similar gravity pat- terns have now been observed in other sections of the Mid-Atlantic Ridge, although substantial local variations exist. The study of seismic waves provides another powerful tool for probing the deep structure of ridges. In the 1980s, seismologists studied a half-dozen large-offset transform faults and a few nontransform offsets at the Mid- Atlantic Ridge. Their results generally confirmed that the oceanic crust is abnormally thin beneath segment boundaries, especially under large-offset transform faults. The gravity and seismic data together, therefore, confirmed that the punctuation of the ridge topography by offset features is indeed indicative of deep-seated, along-axis periodicity in the melt supply. In addition to the gravity and seismic data, imprints of segmentation were found in the pattern of tectonic faulting and earthquakes on the ocean floor. Based on deep-sea observations from submersibles, re- searchers in the 1970s and 1980s determined that lithospheric plates do not spread steadily; instead, they move in a "stop-and-go" fashion, with long periods of tectonic stretching interrupted by short periods of volcanic construction. During prolonged periods of stretching, tectonic faults developed at the Mid-Atlantic Ridge. Recently we observed that tectonic faults are quite linear within spreading segments, indicating that the crust of each segment breaks in parallel fault arrays. In contrast, oblique faults are common near segment offsets, suggesting that the tectonic-volcanic cycles of neighboring segments are not synchronous with each other. The major faults of each spreading segment are f ^DISCONT DISCONT. 28°N to o W o en o 16 Oceauus seismically active, as indicated by large numbers of moderate-sized modern earthquakes (See Mid-Ocean Ridge Seismicity, page 60). The Spreading-Rate Factor The overall magma supply, as well as the style of segmentation, vary dramatically from slow- to fast-spreading ridges. From early, sparse observations, oceanographers noted that the gross axial topography of a mid-ocean ridge depends on the plate-spreading rate. The crest of the slow-spreading Mid-Atlantic Ridge, with full spreading rates of about 30 millimeters per year, is rugged, with faulted crust and a median valley. In contrast, the axis of the fast-spreading East Pacific Rise, with full rates of 60 to 170 millimeters per year, is much smoother, and is defined by an elevated crust a few hundred meters high. Theoretical models suggest that the rift valley of a slow-spreading ridge may result from thinning of the lithospheric plate similar to the "necking"of a plastic beam under tension. For fast-spreading ridges, topography is caused mainly by the upward push of the buoyant magma chamber. There are other major differences between slow- and fast-spreading ridges. Geophysicists have imaged the top of magma chambers at the fast-spreading East Pacific Rise and the intermediately fast-spreading Juan de Fuca Ridge and the Valu Fa Ridge of the Lau Basin, but no comparable structure was found at the slow-spreading Mid-Atlantic Ridge. The lithospheric plate at the MAR has a 3- to 10-kilometer-thick brittle lid in which moderate to large earthquakes can nucleate; at the fast-spreading ridges, the brittle lid is thinner than 2 kilometers, and moderate and large earthquakes are essentially absent. Other differences include gravity and bathymetry, which vary substantially along the slow-spreading ridges, but only slightly along intermediate- and fast- spreading ridges. There is mounting evidence, then, that overall magma supply is greater at fast-spreading ridges than slow ones. Despite major differences in magma supply, both fast- and slow- spreading ridges break into spreading segments. Various types of ridge- crest offsets have been found at the East Pacific Rise, including transform faults, overlapping spreading centers, and deviations from axial lineality. Chemical-composition studies of seafloor rocks indicate that even small ridge offsets mark boundaries between two distinctive magma-supply units. Global variability in ridge magma supply is, in some ways, analo- gous to global variability in climate. The mean air temperature (or overall climate) of Earth's polar regions is substantially lower than that of the equatorial oceans. In both the cold polar and warm equatorial regions, however, the temperature varies from one local area to another. Similarly, the mean magma supply at the slow-spreading Mid-Atlantic Ridge is lower than that of the fast-spreading East Pacific Rise. But in both ridges the magma supply varies locally from one segment to another, and from one part of the segment to another part. The geophysical evidence discussed above and the present segmen- tation theories illustrate only the gross structure of ridge segmentation on length scales of a few to a few hundred kilometers. Smaller, shorter- lived segmentation features, which are beyond the detectabilitly of current instruments, are certainly possible. To further understand and eventually Tlie overall magma supply, as well as the style of segmentation, vary dramatically from slow- to fast-spreading ridges. Winter 1991/92 17 Why do slow- spreading ridges differ dramatically from fast ones? How do spreading segments evolve in time over tens of millions of years? predict the global pattern of magma supply and ridge segmentation, we must develop better instruments, expand the data base, and formulate new theories. The Future of Ridge-Magma Dynamics In the past decade, theories of ridge-magma dynamics have advanced rapidly in conjunction with new oceanographic instruments and discoveries of exciting ridge features. As a result, we have a better understanding of how two-thirds of Earth's solid surface — the oceanic crust — was created. The 1990s promise even greater understanding of ridge dynamics, a system once described by the late Bruce Heezen (see the Profile on page 100) as "the wound that never heals." Oceanographers have designed and are implementing an international decade-long program called RIDGE (Ridge Inter-Disciplinary Global Experiments; see the article on page 21) to study the interactions among complex ridge processes from magma dynamics to earthquakes, water column chemistry, and biology. In particular, the RIDGE program has designed specific oceanographic experiments to continuously explore the origin of mid-ocean ridge segments, posing such questions as, Why do slow-spreading ridges differ dramatically from fast ones, but both break into segments? How do spreading segments evolve in time over tens of millions of years? Why are transform faults longer and stable while nontransform offsets are shorter and transitory? Does segmentation in the volcanic ridge correlate with changes in ridge-crest hydrothermal vents, or even with the biological population at the ridge crest? These and many other questions await exploration in a new era of oceanographic studies. Acknowledgments: The research reported in this article was supported by the National Science Foundation and the Office of Naval Research. G. Michael Purdy and Hans Schouten of WHOI, and J.-C. Sempere of the University of Washington were the leaders of the 1988-1989 Mid- Atlantic Ridge expedition. Thanks to Brian Tucholke and Tom Reed for permission to use the previously unpublished images at the top of page 12 and on page 13 and to P.R. Shaw for permission to use the previously unpublished image at bottom of page 12. The figure on page 11 is after Isacks, Oliver, and Sykes, 1968; that on page 14 is modified after RIDGE Steering Committee, 1989; on page 15 after Whitehead, Dick, and Schouten, 1985, and Dick, 1989; and on page 16 after Lin, Purdy, Schouten, Sempere, and Zervas, 1990. Jian Lin is an Assistant Scientist at the Department of Geology and Geophysics, Woods Hole Oceanographic Institution. He was born when H. William Menard and Bruce Heezen discovered the first transform faults in the world's ocean basins. Today he enjoys the opportunity to further develop and challenge aspects of the great theory of plate tectonics. His research activity ranges from oceano- graphic measurements at sea to building quantitative ridge models on super computers. When not exploring undersea volcanoes, he studies earthquake faults in southern California and their threats to metropolitan areas. He is an Associate Editor of the Journal of Geophysical Research. 18 Oceamis MODELING RIDGE SEGMENTATION. ALL MID-OCEAN RIDGES seem to be segmented. In many places the ridges consist of a series of relatively straight segments divided by fracture zones. In other places they are divided by overlap- ping spreading centers. The spreading plates display a pattern of fairly orderly cellular structure, with spacing between the cells of approximately 30 to 80 kilo- meters. The spacing varies with the speed of spreading of the ridge. Those study- ing the ridge had often suggested that ridge segmentation was due to thermal contraction of the cooling plates as they spread apart. In 1984 and 1985 we, along with Henry Dick, hypothesized something quite different — that the segmentation results from forces produced by hot mantle material rising under spreading centers and liberating melt. We knew that a layer of material with either en- hanced melt or a higher temperature tends to develop a lower density, and there was reason to believe that it would also have a lower viscosity than the sur- rounding regions. We also knew that such a region is prone to develop fluid- dynamic instabilities. One example is called Rayleigh-Taylor instability. This happens when a layer of lower-density fluid underlies a layer of higher-density fluid. The interface between the two flu- ids develops undulations so that the lower density fluid can float upward through the denser fluid. To demonstrate this we conducted some simple experiments in which a water-glycerine mixture was quickly in- jected into glycerine with a hypodermic syringe along a horizontal line. Although this line gradually rises because the wa- ter-glycerine mixture is less dense than the glycerine, an instability also devel- ops (see the photos overleaf) and leads to the formation of semi-spherical pockets. It is reasonable to expect that a linear region of partially molten mantle in the earth will behave in a similar manner and will lead to fairly regularly spaced protrusions from which the melt will ascend to form magma chambers. We suggested this example as a possible model of what might be happening un- der oceanic ridges. To be specific, the idea was that segmentation was pro- duced by buoyancy-induced instability (which ultimately leads to volcanism) rather than by thermal contraction of the cold plates. Clearly, the model was too crude to apply to ridges in detail. However, at that time numerical models of spreading ridges were always taken to be two di- mensional for simplicity and ignored seg- mentation. Unfortunately, if the segmen- tation is a process that enhances up- welling the two dimensional models would be incomplete. Recently, observa- tions indicate that segmentation is in- deed not just a surface feature of cooling plates but extends "deep" (tens of kilo- meters) under ridges. In addition, recent three-dimensional numerical models have been developed with flows that do break up into segments. Thus the crude idea that segmentation has deep origins seems to be borne out even though the detailed mechanics of the break up may be different in detail than our Rayleigh- Taylor models. "^N (continued on next page) HANS SCHOUTEN AND JACK WHITEHEAD WOODS HOLE OCEANOGRAPHIC INSTITUTION Winter 1991/92 19 .A Possible Mechanism These two photographs were taken 30 seconds apart. In the upper photo- graph, the injecting needle was dragged from right to left. The gravitational instability of a horizontal line of water I glycerine mixture in a bath of pure glycerine is shown below. Consider a Gift Membership! A gift membership to a friend, relative or colleague, is a great way to share your concerns for our ocean environment. The next volume (four issues) of Oceanusvrill cover the four primary fields in marine science, constituting a course in itself. Just complete one of the Membership Order Forms, or call our office at (508) 457- 2000, extension 2386, and we will take your order. Oceanus International Perspectives on Our Ocean Environment 20 Oceanus RIDGE: Cooperative Studies of Mid-Ocean Ridges Donna Blackman and Trileigh Stroh he Ridge Inter-Disciplinary Global Experiments (RIDGE) Initiative is a cooperative effort to study the mid-ocean ridges as a dynamic global system of focused energy flow from Earth's interior outward. The National Science Founda- tion supports the RIDGE Initiative, part of the US Global Change Research Program, through both its Global Change and Ocean Sciences divisions. The program's key goals include: • characterizing the global ridge structure, • understanding crustal accretion and upper-mantle dynamics, • charting the variability over time of volcanic and hydrothermal systems, • mapping biological colonization and evolution at ridge crests, • determining the properties of multiphase materials at ridge crests, and • developing technology for ridge-crest experimentation. By characterizing the ridge structure, researchers intend to provide a global perspective for the mechanics of plate separation, variable lava types, circulation of hot seawater, and biological characteristic of indi- vidual ridge sections. Swath bathymetry, sidescan sonar imagery, and widely spaced geologic and hydrologic samples will be used to develop large-scale maps of the ridge system. This will provide a basis for estimating the total flux of materials through ridge crests (hydrothermal input to the oceans, for example), as well as for making site-location decisions for more detailed study. Crustal accretion results from convective upwelling of the mantle beneath a spreading center (see Onions and Leaks. ../page 36). Basaltic melt segregates from the rising, decompressing mantle, and is delivered to a magma chamber at the ridge axis where it solidifies to form oceanic crust. Understanding this process requires a variety of information: the mantle's flow geometry, temperature, and composition beneath mid- ocean ridges; the nature of the subaxial magma chamber and the mode of volcanic extrusion on the seafloor; and the role of hydrothermal circulation in cooling the crust. Constraints on the upper-mantle struc- ture are obtained from computer modeling and seismic- and electromag- netic-imaging studies that use large arrays of seafloor instruments. A magma chamber's size and shape are revealed by geophysical measure- ments including seismic refraction /reflection and gravity data; geo- chemical studies of rock samples from the ridge axes help determine the RIDGE Winter 1991/92 21 -- Megaplume ^p ^ "- - ;'J*t*-?- ?«» SHS^J Mantle Peridotite Basaltic melt arriving from the mantle either reestablishes or replenishes a crustal magma chamber that solidifies to produce gabbro, diabase dikes, or basaltic lava flows. The geometry, longev- ity, and circulation of a subaxial magma chamber are topics of active inquiry. history of the basaltic melt as it separates from the mantle and cools within the crust. Mapping the hydrothermal vent fields at ridge axes and the associated faulting of the seafloor illustrates the seawa- ter-circulation pattern in the upper crust. Although seafloor spreading is continuous on a geologic time scale, individual earthquakes, eruptions, and venting episodes affect only a short length of the ridge for a short time. Neither the spatial nor the temporal scales of specific ridge-axis events are currently known in detail, but both play critical roles in shaping seafloor morphology, local seawa- ter properties, and biological diversity. Two important means for studying temporal variability in ridge processes are event detection and response and long-term deployment of instruments in a seafloor observatory. When a ridge-crest volcanic event, such as an earthquake swarm, is detected, airborne and shipboard instruments can be deployed to chart the activity pattern and map any new eruptions. Long-term monitoring can reveal linkages among complex, interrelated physical, geological, and biological processes at ridges. Diverse coordinated measurements made at permanent ridge-crest observatories will be essential in studying these relationships and developing improved theoretical models for ridge processes. Mapping biological communities along mid-ocean ridge crests is fundamental for understanding the thermal and chemical requirements of these unique ecosystems. Sampling and laboratory studies will reveal the physiological and genetic requirements for living without sunlight at water depths exceeding 2,000 meters. Integrating biological and chemical studies will elucidate the dependence of different organisms on the tem- peratures and the chemical characteristics of hydrothermal vents. Determin- ing a biological community's response time to changes in volcanic and hydrothermal activities will be an important aspect of this research. Multiphase materials are present in virtually every part of the mid- ocean ridge system, from the upwelling mantle that contains basaltic melt, through the magma chamber where molten rock is crystallized, to the hydrothermal systems in which both liquid and gas aqueous solu- tions are likely to exist. Laboratory experiments on silicate aggregates under various temperature and pressure conditions are needed to define the behavior of the ascending mantle (viscosity, melt content, and composition). The chemical properties and crystallization sequences of mid-ocean ridge basalts must be determined at pressures appropriate for a crustal magma chamber. Modeling the effects of combined vapor and fluid phases in hydrothermal circulation should aid in understanding 22 Oceanus interactions between the cooling crust and seawater. Developing extended-deployment seafloor instrumentation that can accurately measure the changing conditions at the ridge crest is an integral part of many of the above research topics. Examples of new technological advances include chemical sensors that detect minute changes in trace elements and compounds (such as hydrogen sulfide, methane, iron, manganese, and oxygen), geodetic instruments to mea- sure uplift and tilt of volcano flanks, broadband ocean-bottom seismom- eters, and deep-water temperature and chemical profiling systems. Systems that can deploy and manipulate these sensitive instruments will also be required and may take the form of remotely operated seafloor vehicles or manned submersibles. Donna Blackman is a post-doctoral research associate at the University of Washington studying mid-ocean ridge processes. At the time of writing, she is assisting with several projects at the RIDGE office. Trileigh Stroh has been the RIDGE Coordinator and Editor of RIDGE Events newsletter since 1988. In 1992 she will instead serve as Executive Administrator for the InterRidge office. InterRidge INTERRIDGE IS an international effort to coordinate and expand ridge-crest re- search. Representatives from scientific communities in Australia, Canada, France, Germany, Iceland, Japan, Nor- way, Portugal, UK, USA, and USSR have been meeting since 1989 to establish means for effective communication, pro- gram coordination, and data exchange among various national programs for mid-ocean ridge research. Ratification of an InterRidge program plan and estab- lishment of an InterRidge office are ex- pected in 1992. The Program Plan will propose three primary program elements: global studies, observatory development, and regional dynamics studies. A recent response to events south- west of Iceland showed that international cooperation can produce insights into ridge-crest processes and their transient oceanographic signals. In November 1990, Icelandic scientists reported a se- ries of earthquakes on the Reykjanes Ridge. Their report was followed by de- ployment of sonobuoys and expendable bathythermographs from a US Navy P3 aircraft. British, Icelandic, and American scientists used an Icelandic research ship to conduct several days of on-site map- ping and sampling. Using the combined assets of several countries enhances the ability to quickly assemble a team of investigators at an eruption site, provid- ing valuable opportunities to document ongoing ridge-crest activity. Winter 1991/92 Earth's RiftS, Ridges & Rises: Areas of interest referred to within this issue 1-Albatross Plateau 2-Atlantis Transform Fault 3-Aleutian Trench 4-Carlsberg Ridge 5-Chile Rise 6-Clipperton Fracture 7-Cox Transform Fault 8-East Pacific Rise 9-Easter Island 10-Galapagos Rift 11-Gorda Ridge 12-Guaymas Basin 13-Guaymas Transform Ridge 14-Gulf of Aden 15-Gulf of California 16-Hawaiian Islands 17-Juan de Fuca Ridge 18-Kula Rise 19-Labrador Sea 20-Lesser Antilles Arc 21-Louisiana Slope 24 Oceanns This computer-generated map is courtesy of Peter W Sloss, NOAA National Geophysical Data Center. Boulder. CO. 22-AMAR Area FAMOUS Area 23-TAG Area 24-MARK Area 25-Mariana Trench 26-Mathematician Rise 27-Mendoza Rise 28-Mid-Atlantic Ridge 29-Middle America Trench 30-Monterey Canyon 31-Pacific-Antarctic Ridge 32-Red Sea 33-Siqueiros Transform Fault 34-South Shetland Arc 35-Troodos Ophiolite 36-Valu Fa Ridge 37-West Florida Escarpment 38-Iceland 39- Azores 40-Tristan da Cunha 41-Vema Transform 42-1 5°20' Transform 43-Kane Transform Winter 1991/92 25 Slow spreading produces relatively steep- sided "ridges," while fast spreading produces more gently sloping "rises." Ridges and Rises: A Global View Peter Lonsdale and Chris Small eafloor spreading — the process that creates new material to fill in gaps between Earth's separating crustal plates — results in broad elevations with spreading centers along their crests. This is simply because crust formed by volcanic activity deepens as it moves away from the axes, cools, and contracts. The young- est, hottest crust stands highest, and the rate of deepening, which determines the regional slope gradients away from the spreading center, is proportional to the horizontal rate of crustal aging. Slow spreading produces relatively steep-sided "ridges," while fast spreading produces more gently sloping "rises." The regional side-slopes of spreading ridges and rises are concave curves, flattening out to imperceptible gradients where the crust is about 100 million years old. Even on the steepest, youngest part of slow-spreading ridge flanks, the regional gradients are actually so low that "ridge" may seem a misnomer; the popular concept of the Mid-Atlantic Ridge as a mighty chain of undersea mountains seems somewhat overblown given that the regional slope halfway down its flanks is no steeper than the eastward slope of the North American Great Plains. Of course, the small-scale topography of the Mid-Atlantic Ridge is rougher, but it took the truly global perspective provided by the very low resolution of exploratory bathymetric data (piano-wire sound- ings, hundreds of kilometers apart) for 19th-century oceanographers to recognize this "ridge" as a major feature of Earth's surface. Long before oceanic crust attains an age of 100 million years, the cooling-induced slope of its upper surface is altered by other seafloor processes; the outer margins of mid-ocean ridges are generally defined by the topographic boundary between the landward structural slope of cooling lithosphere and the seaward depositional slope of continent- derived sediment. In the Atlantic Ocean, this boundary usually occurs where the crust was created about 70 million years ago, so the Mid- Atlantic Ridge is a 1,500- to 2,000-kilometer- wide structure that includes all the crust created since then. It covers half of the seafloor. Narrower basins have proportionately narrower mid-ocean ridges. Young ex- amples are in the Gulf of Aden and the mouth of the Gulf of California, where the ridges are less than 100 kilometers wide. Ridges may even be disproportionately narrow where basins are exposed to the rapid influx of sediment from adjacent continents, and in extreme cases (such as within the Gulf of California and the Red Sea) smothering sediment completely inhibits ridge-building volcanic eruptions, and seafloor spreading proceeds without construction of a mid-ocean ridge. Rapid 26 Ocentins burial of a ridge and obliteration of its characteristic relief can also occur if spreading stops because of a change in continental drift patterns. A buried mid-ocean ridge underlies a sediment plain in the center of the Labrador Sea, where spreading between Greenland and Labrador stopped 45 million years ago. Mid-Ocean Ridges and Ocean Basins These examples can be arranged in sequence to illustrate a familiar model of the development and demise of mid-ocean ridges in growing "Atlantic-type" intercontinental ocean basins. Such basins have been abundant on Earth for the last 300 million years, as the supercontinent Pangaea broke up and its fragments drifted apart. There are other types of ocean basins, with other types of spreading ridges, especially in the whole hemisphere occupied by the Pacific Ocean. Speculative reconstructions of supercontinents that preceded Pangaea suggest that the Pacific Ocean may have originated by spread- ing at a mid-ocean ridge between North America and Antarctica, but that was in late Precambrian times, about 700 million years ago, and all the crust formed during this expansionary phase has long since been re- cycled into Earth's mantle at subduction zones. None of the Pacific Ocean's present floor is older than 250 million years, and all the time it was being created the ocean basin was getting smaller as adjacent conti- nents converged on it. During this prolonged contractional phase, two distinctive types of spreading ridges have been active: Pacific rises (typi- fied by the East Pacific Rise), and back-arc ridges (such as the Mariana Trough Ridge). Rises are generally considered variants of mid-ocean Rift Valley New Ocean Basin Mid-Ocean Ridge Fossil Ridge Ocean Water Sediment Continental Crust Oceanic Crust Mantle Rocks Jack Cook/WHOI Graphics Winter 1991/92 The development and demise of a mid-ocean ridge. A: Incipient separation of two continental blocks causes faulting and thinning of the continental crust, and development of a rift valley (e.g., East African Rift). B: Continued crustal separation produces a gap that is partly filled by sediment washed off the continents and partly by melting of the mantle to produce oceanic crust (e.g., Gulf of California). C:As the gap between the separating continents increases, oceanic crust formation by seafloor spreading at the crest of a rifted mid-ocean ridge becomes fully developed (e.g., North Atlantic). D: If continental separation stops, seafloor spreading ceases and the mid- ocean ridge subsides as it cools and gradually becomes covered with sediment. 27 These equal-area projections reveal the global distribution of spreading centers on mid-ocean ridges, Pacific rises, and back- arc ridges. The relative importance of the three types varies from region to region. ridges, mere components of the same global ridge system that connects divergent plate boundaries, though they neither originate as intercontinental rifts nor occupy mid-ocean posi- tions. Back-arc ridges are not even connected to a global ridge system, and instead of straddling boundaries between major diverging plates they adjoin subduction zones that mark sites of plate convergence and destruction of oceanic crust or lithosphere. Whereas authors with an "Atlantic fixation" have adopted Pacific rises as an eccentric, errant variety of the familiar mid-ocean ridge, to the amusement of researchers studying nearshore parts of the East Pa- cific Rise or the Juan de Fuca Ridge, back-arc ridges are often slighted as isolated second-order com- plications to the global scheme. The striking contrasts in the geographic distributions of the three types of spreading ridge encourages such parochial assessments of their relative importance, and the view from Tokyo or San Diego can be quite differ- ent than that from Boston or London. •— Mid- Ocean Ridge Back-Arc Ridge • • • • • Inactive Back- Arc Ridge Pacific Rise Pacific Rise Spreading and Destruction Pacific rises are concentrated in the eastern half of the ocean, where multiple, branching rise crests have developed at the boundaries of purely oceanic plates. They tend to spread faster than mid-ocean ridges, because oceanic plates move faster than partly continental ones. Despite this faster rate of crustal accretion, there is a net loss of Pacific crust each year, because the rate of recycling into the mantle, by the process of subduction at marginal trenches, is even greater. As oceanic plates descend into subduction zones, one or both flanks of the rises that grow on their trailing edges eventually enter the trenches and are destroyed. If one flank is completely consumed, the actively spreading rise crest may collide 28 Oceanus with the continental margin, as parts of the Chile Rise are now doing off southern Chile. More commonly, a rise crest drifting toward the margin of the ocean basin ceases spreading before it ever enters the trench, and it is an inactive or "fossil" rise crest that is consumed. The northernmost part of the Pacific Basin is now occupied by the south flank of a rise (Kula Rise) that was 2,000 kilometers long and 2,000 kilometers wide when it stopped spreading 42 million years ago; since that time the Aleutian Trench has consumed almost all of its north flank, and all but 75 kilome- ters of the fossil rise crest. In the same period a trench along the western margin of North and Central America has consumed most of the east flank of the northern East Pacific Rise. Along parts of this margin, off northern and southern California, there was a collision between the active rise crest and the continent, but off central California and Baja California, spreading ceased when the rise crest was within 50 to 100 kilometers of the trench. Fortunately, subduction (removal of oceanic crust) ceased at the same time, so the record of 20- to 10-million-y ear-old fossil-rise crests is preserved on the present ocean floor. Further north, off Oregon and Washington, part of the East Pacific Rise that had been approaching North America began to move away from it about 20 million years ago, thereby escaping subduction and surviving as the rise system now called the Juan de Fuca Ridge. The Juan de Fuca Ridge acquired its new identity, and a unique rate and pattern of spreading, when it became isolated from the main rise system by continental collision. Some altogether new rises got started with the fission of an oceanic plate that was being pulled in two different directions toward trenches along different parts of its margin. About 25 million years ago this was the fate of the largest eastern Pacific plate, which split into the "Cocos Plate" moving toward the Middle America Trench and the "Nazca Plate," which built a rise that extends east-west just north of the equator. A more common origin is when new spreading centers open up along the flanks of existing rises, generally in response to changes in plate-motion direction. After a few million years with both Crest of Mature Rise Incipient Spreading Center on Rise Flank Trench \ B Rift Valley 1 New Rise Trench \ Jack Cook/WHOI Graphics Winter 1991/92 The development of a new Pacific rise by replacement of an old one. A: A Pacific rise, producing crust that is reentering the mantle at a marginal trench, develops a new extensional plate boundary (with rifted oceanic crust) on its flank. B: The site of seafloor spreading shifts to the new plate boundary, where young hot crust forms a new rise (e.g., East Pacific Rise), right. This new rise may in turn become inactive, especially if it ap- proaches a trench too closely. The original rise crest becomes inactive and its axis forms a rift valley (e.g., Mathematician Rise), left. The color key is identical to the one for the figure on page 27. 29 The origin of a back-arc ridge. A: Crust overlying a downgoing slab of old oceanic crust is thickened by arc volcanism. B: The seaward movement of the trench and adjacent part of volcanic arc causes rifting of the arc and growth of a back- arc ridge in a new basin. Trench rise crests active, the older one is generally replaced by the new one and becomes extinct. This process commonly recurs at different times along various parts of the same rise, making these parts of different age, and thereby confus- ing the nomenclature. Some authors consider almost all the Pacific floor east of the Hawaiian Islands to be the partly subducted remnant of the East Pacific Rise, while others restrict the term to crust that has formed since the most recent new start and reorientation, which occurred 20 million years ago at latitude 20°S (replacing the now-fossil Mendoza Rise), 10 million years ago at 10°S (replacing the Galapagos Rise), and 5 million years ago at 18°N (replacing the Mathematician Rise). As further illustration of this rise's hybrid origin, the northernmost and youngest part of the present East Pacific Rise crest, in the mouth of the Gulf of California, is an exceptional 200-kilometer-long mid-ocean ridge, where intercontinental spreading between Baja California and the Mexican mainland has occurred for the past 3.5 million years. This local expan- sion of the Pacific Basin occurred when the tip of the spreading center, which had intersected the continental margin, propa- gated a short distance into the interior to link up with fault systems developing in the Gulf of California. Seaward Mantle melts to supply volcanic arc B Volcanism ceases on this fragment of the former arc Back- arc Ridge Arc Platelet Volcanic Arc -Trench Mantle supplies both volcanic arc and back-arc ridge Water Sediment Back-Arc Ridges: Episodic Rifting The Pacific Ocean's complex western margin is the locale for most "back-arc ridges," a phrase that describes their tectonic setting at the back or landward side of the rows or arcs of subduction-zone volcanoes that form the "Ring of Fire," where oceanic plates underthrust the Pacific Rim at marginal trenches. Around a contracting ocean basin, the subduction zones must migrate seaward. As they do, a narrow sliver of the rim, ^ including the landward side of the trench and the volcanic arc, migrates with them. Detachment of this "arc platelet" and its subsequent drift away from the parent land- ward plate causes back-arc spreading. The process is characteristically episodic. Rifting begins in the volcanically weakened arc crust, which is split lengthwise, and the landward half becomes inactive. Spreading between the volcanically active and inactive halves of the arc opens up a back-arc basin, with a spreading ridge whose crest migrates seaward at only about half the speed of the arc platelet, thereby becoming increas- Old Oceanic Crust New Oceanic Crust Mantle Jack Cook/WHOI Graphics 30 Oceanus ingly distant from the trench. The back-arc ridge generally becomes inactive after spreading and building the basin floor for several million years, but if subduction and arc volcanism continue, the process may repeat, with rifting renewed in the arc. In this manner, a series of succes- sively younger back-arc basins, floored by extinct or actively spreading ridges, has been added to the western margin of the Pacific. Exceptions to the rule that back-arc ridges are features of contracting trench-ringed ocean basins are two isolated examples associated with Atlantic island arcs: an inactive (and sediment-smothered) one behind the Lesser Antilles Arc in the eastern Caribbean, and an active back-arc ridge behind the remote South Shetland Arc. Characteristics of Spreading Ridges How do mid-ocean ridges, Pacific rises, and back-arc ridges differ in gross topography? Their diverse histories and tectonic settings result in a variety of sizes, sediment covers, and symmetries. Short-lived, back-arc ridges tend to be narrower, their ocean-margin location makes them more vulnerable to sediment smothering, and an asymmetric sediment supply (mainly from adjacent volcanic arcs) threatens that even if ridge development is not suppressed, the seaward flank may be buried by sediment fans. In many back-arc basins there is even evidence for encroachment of arc volcanism onto the seaward flank of the ridge. Asymmetric topography, in contrast to the striking bilateral symmetry of mid-ocean ridges, is also a feature of some nearshore Pacific rises that have unequal sediment loading on the two flanks; for example, most of the landward east flank of the Juan de Fuca Ridge lies beneath a thick lens of sediment brought by turbidity currents from the nearby continen- tal margin, whereas the west flank is sheltered from the effects of such currents by the relief of the rise crest itself. A more fundamental cause of asymmetry on many Pacific rises is varying amounts of flank removal at marginal subduction zones. Once a whole flank has been consumed and the crest collides with the continental margin, the "rise" is merely a steadily deepening ramp leading from the margin to the continental interior, as exemplified by the westward slope of the seafloor between California and Hawaii. The processes of sediment burial and plate consumption that cause some gross differences between the three genetic types of spreading ridges are secondary to the volcanic, tectonic, and hydrothermal pro- cesses that create oceanic crust and shape the medium- and small-scale topography at ridge and rise crests. These processes seem to work in remarkably similar ways at all three types. Variations in ridge-crest structure and relief are more clearly related to the rate of crustal accre- tion than to the origin and history of the spreading center. The current spreading rate (the width of the crustal strip added per unit of time) at the divergent boundaries of major plates, varies from 12 to 14 kilometers per million years (or, 12 to 14 millimeters per year) along mid-ocean ridges in the Arctic and between the slow-moving African and Antarctic plates. It is more than 10 times this speed along most of the East Pacific Rise. Active back-arc spreading centers cover a similar spectrum of spreading velocities, and geologic study of old The Pacific Ocean's complex western margin is the locale for most "back-arc ridges/' at the back or landward side of subduction- zone volcanoes Winter 1991/92 31 oceanic crust indicates that the same range of rates has prevailed throughout the past 200 million years, though the worldwide average rate has fluctuated significantly during this period. The spreading rates of the mid-ocean ridges and Pacific rises that are now active fall into four classes: slow, medium, fast, and ultra-fast. The longest ridges are in the slow-spreading class, which includes many Atlantic and Indian mid- ocean ridges; but the most productive in terms of total area of crust added each year is the ultra-fast class, which only includes the central part of the East Pacific Rise. Unrifted Rises and Rifted Ridges Although there are four speed classes, and speed of opening affects ridge crest structure, we recognize just two fundamental structural types: unrifted rises and rifted ridges. The former, characteristic of both the fast and the ultra-fast classes, have narrow 100- to 300-meter-high, 2- to 10- kilometer-wide "axial ridges" along their spreading axes. The elevation of the axial ridge is readily explained by its location over a body of hot, partly molten rock, including a thin, narrow magma chamber (see Tomographic Imaging of Spreading Centers, page 92). Despite the "unrifted" apellation, the axial ridge contains a volcanic rift zone much like the rift zones on Hawaiian volcanoes: an elongated zone of weak- ness into which vertical bladelike sheets of molten rock are injected from underlying magma sources. When the sheets freeze within the ridge, tabular intrusions known as "dikes" are formed, and gaping fissures open up in the overlying seabed. If the magma pressure is great enough, the sheets reach the seafloor and lava erupts from fissures along the axial ridge crest. The remarkably narrow zone of intense dike injection and fissure eruptions, typically less than 1 kilometer wide, is usually marked by a shallow "axial summit graben" only 10 to 100 meters deep and formed by collapse of the axial ridge crest between major eruption events. The fissured floor of this summit graben is a favored site for mineral-precipitating discharges of hydrothermal fluids that are heated Both rifted ridges and unrifted rises occur in this gravity anomaly plot. • Axis of Rifted Ridge • Axis of Unrifted Rise Gravity Anomaly (mgal) , Normal Fracture Zone t Oblique Fracture Zone -30 -20 -10 10 180 -ISO 32 Oceanus 270 E 180 E mGal <-30 by contact with the hot frac- tured seabed. Rifted ridges are those mid- ocean and back-arc ridges on which the plate boundary does not crop out along the crest of the highest volcanic ridge, but along the floor of a 100- to 3,000-meter- deep axial rift valley that is typically 10 to 40 kilometers wide, and bordered by uplifted fault blocks called rift mountains. Much has been written on the nature and origin of axial rift valleys since they were discov- ered by reconnaissance echo sounding more than 50 years ago (see Onions and Leaks. . ., page 36 and Tectonics of Slow Spreading Ridges, page 51 for some of the newer ideas and observations). The discussion here summarizes their global distribution, which has recently been clarified by satellite observations. Radar altimeters carried by some mapping satellites, notably Seasat, Geosat, and ERS-1, measure variations in the shape of the sea surface, which is affected by seafloor topography because water piles up over ridges that exert a gravitational attraction. After processing to remove the influence of waves, tides, and long-wavelength variations, altimeter profiles can be displayed as maps of sea-surface gravity anomalies, with positive anomalies over axial ridges and negative anomalies over the floors of axial rift valleys. Data from systematic global coverage by satellite altimeters combined with more patchy mapping by ships equipped with echo sounders shows that axial rift valleys are most impressive on slow-spreading ridges; at spreading rates of 15 to 60 millimeters per year, there is a crude negative correlation between spreading rate and rift-valley depth. However, a rift valley's presence and size are probably controlled by the rate at which molten rock is supplied from the mantle to the accreting plate boundary, and not necessarily the correlated spreading rate. Where the magma supply is voluminous and steady (a prerequisite for fast spreading and a feature of slow-spreading ridges that happen to be near unusual "hot spot" magma sources), a permanent reservoir of partly molten rock underlies the spreading axis and creates an axial ridge. At spreading centers with smaller, more episodic magma supplies, like most slow-spreading ridges that are far from unusual "hot spot" magma sources, plate separation and crustal extension leads to rift valley formation. In the medium- spreading speed class, many axes have well developed, albeit shallow, rift valleys, but axial ridges similar to those of fast-spreading rises also occur, sometimes on adjacent parts of the same spreading center. At 90 E >+30 T/7/s map depicts the gravity field of the Southern Ocean, as derived from Geosat altimetry. The south- ernmost portions of the Pacific-Antarctic and SoutJieast Indian ridges appear as a complex chain of positive gravity anomalies between 150° and 210° E. This map is available in both digital and poster form from the NOAA National Geophysical Data Center (Report MGG- 6, Marks and McAdoo, 1992). Winter 1991/92 33 Brazil, South America o I 1000 km Continent & Continental Slope Sediment Plain Mid -Atlantic Ridge \ Speeding Segment Transform Fault Offset Fracture Zone Spreading Direction Where the overall strike of the ridge is highly oblique to the spreading direction, as in the equatorial part of the Mid-Atlantic Ridge, the offsets between spreading segments are longer and more closely spaced. these intermediate spreading rates (60 to 80 millimeters per year) the presence or absence of a rift valley is probably sensitive to local and spatial and temporal changes in the magma supply rate. Ridge-Crest Segmentation Fracture zones, and the rise-crest offsets that create them, are also essential features. Ridge-crest segmentation is one of the most funda- mental features of spreading centers and one of the most active areas of current research. The crustal-accretion belt along a ridge crest is not continuous, but is broken by several types and sizes of ridge offsets into laterally displaced segments tens or hundreds of kilometers long. Individual spreading segments are oriented at right angles to the spread- ing direction, and are frequently arranged in a staircase with offsets systematically stepping left or right. In such cases the relative lengths of spreading segments and offsets is determined by how oblique the ridge is to the spreading direction; for instance, the equatorial part of the Mid- Atlantic Ridge strikes almost east-west, and has long left-stepping offsets linking short spreading segments, while further south, where the overall strike is more nearly north-south, lateral offsets are shorter and more widely spaced. There is also empirical evidence of an inverse correlation between segment length (offset spacing) and spreading rate. Only a small fraction of the global ridge system has been surveyed with the high-resolution tools needed to locate small offsets, however, so their mapped abundance partly reflects the relatively small survey effort. The spreading rate certainly influences the structure of the ridge offsets. On slow-spreading ridges, all but the shortest offsets contain transform faults, which are narrow zones of "strike-slip" (horizontally sliding) faulting at right angles to the spreading segments. Similar transform faults subdivide fast-spreading rise crests, but only where offsets are longer than 50 to 100 kilometers. Shorter steps in the plate boundary are much more abundant, and have broad zones of deformation that are said to be "nontransform" because they lack strike-slip faults. Both transform and nontransform offsets, known as "fracture zones," leave recognizable Oceanus trails on the rise flanks, belts of distinctive topography that interrupt the abyssal hill pattern because they have spread from disruptions of the spreading center. Fracture zones produced at long transform faults have high relief that was easy to discern even with early echo sounders. Conversely, locating fracture zones produced at nontransform offsets generally requires high-resolution mapping of abyssal-hill or crustal-age patterns, because their subtle relief is unpredictable: Nontransform offsets migrate along the rise crest at variable speeds and directions, rather than maintaining a stable location, as transform offsets do. The V-- shaped feature in the middle is a good example of a pair of oblique fracture zones produced by a migrating nontransform offset, in this case one that has migrated steadily southwest at a rate about twice that of the spreading rate. Much remains to be learned about the origin of ridge-crest segmen- tation, and the reasons for segmentation-pattern changes. Why, for instance, do some spreading segments grow in length at the expense of their neighbors, causing the offsets between them to migrate along the plate boundary? Different approaches being used to tackle this problem include making detailed studies of the rise-crest processes at a few convenient locations and preparing a global inventory of all such offsets, to see how their directions and rates of migration correlate with such factors as spreading rate, segment length, rise-crest depth, etc. For the latter task, a much more complete description of the global ridge system is needed, implying many more months of survey effort with the sophis- ticated multibeam mapping sonars now available on research ships. Satellite observations may have sufficient resolution to partly supple- ment the shipboard work, and provide an immediate global perspective. Unfortunately, the best satellite data now available has been classified as a military secret (except in the "nonstrategic" Antarctic region), and is therefore unavailable to most researchers, just as the results of multibeam sonars were a decade ago. This impediment to understanding the pattern and relief of spreading ridges will disappear as military satellites are replaced by civilian ones, like the ERS-1 that is now in orbit and busy collecting altimeter profiles across all the world's ridges. Peter Lonsdale is a Professor of Oceanography and Research Geologist with the Scripps Institution of Oceanography (SIO). He has spent two or three months in each of the past 20 years examining the ocean floor with echo sounders, cameras, and submersibles, about half of this effort being on Pacific rises and back-arc ridges. Chris Small is a graduate student at SIO, with a special interest in using satellite altimeters for structural studies of mid-ocean ridges. Ridge-crest segmentation is one of the most fundamental features of spreading centers and one of the most active areas of current research. Winter 1991 /92 35 Thirty -five years ago, most geologists were secure in the knowledge that continents did not move. Onions and Leaks: Magma at Mid-Ocean Ridges A Very Personal View 36 Joe Cann n 1992 we see mid-ocean ridges clearly, forming a complex, 50,000-kilometer-long web of seafloor mountain chains that encircle Earth. Along the mountain crests there is a narrow belt of activity, marked by shallow earthquakes, seafloor volcanic eruptions, and hot springs, where new ocean crust is constructed at the rate of a few centimeters every year (about as fast as fingernails grow). Recent intense study of this zone has sharpened our picture, refocused it here and there, brought sudden insights, and revealed errors of perception, until we have reached new levels of clarity. This year seems especially propitious for reviewing mid-ocean ridges. We are pleased that our new models are good, that our under- standing is secure. There are difficulties to be sorted out, but most are within our grasp. Now we should settle down to explain what we know. And in that spirit we write, and you read, this issue of Oceanus. Our certainty is not new. Thirty-five years ago, most geologists were secure in the knowledge that continents did not move, that the oceans were permanent, unchanging features of Earth's surface, containing sediments as old as ocean water itself and interleaved here and there with lava flows. Mid-ocean ridges might be fold-mountain belts like submarine Rockies, or rift mountains like submerged East African highlands, but were certainly explainable in sensible continental terms. Within a few years this comfortable picture was to be turned upside down by the very people who then possessed such certainty of belief. Thirty five years ago I was a geology undergraduate. Our first-year text was by Arthur Holmes who, in about 1930, focused attention on the mid-ocean ridges with his concept that Earth's deep interior might be slowly convecting as it was heated by the radioactive decay of potas- sium, uranium, and thorium. He thought that deep-Earth convection currents might move the continents apart, and that upwelling currents might rise in the centers of those oceans that had matching coastlines on either side, such as the Atlantic and Indian oceans. At first, Holmes thought that the Mid- Atlantic Ridge was a strip of continent left behind as Africa and America split apart, but in our 1944 textbook he replaced Oceanus the continent with oceanic crust. That diagram looks very much like the sketches we draw today. In student seminars we talked about Wegener and du Toit, pioneers of continental drift theories. We argued whether the oceans might be young, as they said, or ancient, and whether animals had crossed the oceans on land bridges or floating tree trunks, or maybe had wandered from place to place when the continents were joined to form Gondwanaland, Laurasia, or, earlier, Pangaea. Our professors cautioned us against believing Holmes too literally, indicating that ideas about drift were based on woolly speculation. Harold Jeffreys, the most eminent geophysicist of the time, had proved that drift was impossible. How could we disagree? The pioneers of rock magne- tism certainly did so. They showed that ancient rocks are magnetized very differ- ently from recent ones, suggesting that they originated at latitudes other than those they now occupied. Could Earth's spin axis have changed? Or had the continents moved? Jeffreys skeptically pointed out that iron could be remagnetized by striking it with a hammer, a tool traditional with geologists. Then marine geology and geophysics began to take a hand. Inspired by service at sea and trained in antisubmarine warfare, young marine scientists brought new talents and instruments to the oceans. It was a curious time. Marine scientists were few, and nearly everyone knew nearly everyone else. Even when I came into marine geology in 1962 as a young post doc, I was able to rapidly meet almost all of the players in the game. That would be quite impossible today. Many of the leaders in the field worried more about their next expedition than about publishing the results of their last, but all were willing to talk. Ideas developed by word of mouth, shortcutting publication, so that often it was — and still is— difficult to lay credit where it properly belongs. Harry Hess, one of the most charismatic scientists of his time, was a reluctant but remarkable speaker: Quiet, with a cigarette dangling from his fingers, he was seemingly casual, yet profoundly convincing. He first took a semi-fixist view, in which convection currents stirred the mantle, but continents did not move. He ascribed the existence of mid-ocean ridges to transformation of dense mantle peridotite to light serpentinite, by way of water seeping up from the rising convection currents. Soon he moved into the mobilist camp, and allowed that convection moved continents. In the first breakthrough since Holmes, Hess suggested that new ocean crust is continually created at mid-ocean ridges and spreads away as the ocean grows. This became the theory of ocean-floor spreading, and from it emerged the first model of mid-ocean-ridge processes. Hess thought that the rising limbs of deep-Earth convection currents not only split the ocean floor apart, but also contributed basalt magma and water to the growing crust. Magma would be erupted at the ocean floor to form the Drum Matthews (red shirt), Tony Laughton (blue hat), and Ron Oxburgh work amid basalt lavas at the Gulf ofTadjura, Djibouti, in Janitan/ 1967. The Gulf ofTadjura is where the Gulf of Aden spreading center comes ashore, and is splitting apart slowly. The lavas come from magma chambers below the seafloor. Winter 1991/92 37 :-, - ./ •--:_ -- ,,, :. , ... iA - , ?*-.' - jg- ; • -. ^ '-+- :-• , V^;> ^|^ --' - ^:'3: -^'f. -.- - / "^^' ^X .;- ' £-''' Troodos ophiolite complex in Cyprus, a road cut shows the top of a seafloor magma chamber formed 1 to 2 kilometers below the ancient seafloor. The pale rocks are gabbros and trondhjemites produced by crystalli- zation of the top of the magma chamber. The gray stripes are dikes intruded from another chamber nearby. Hazel Prichard is the figure, once a student of the author and now at the Open University in the United Kingdom. seafloor lavas that were now being collected regularly, and water would circulate down through the porous lavas to alter the uppermost mantle to serpentinite, creating the lower part of the crust. The base of the crust would thus mark the lowest level that serpentinite could form at the ridge axis, representing the temperature at which serpentinite dehy- drates back to peridotite. The demonstration that Hess had been broadly right was a triumph of the 1960s. From data gathered by towing a magnetometer across the oceans, Fred Vine and Drum Matthews explained that magnetic anomalies were created as the result of ocean floor spreading, while Earth's magnetic field periodically reversed. Tuzo Wilson invented the concept of transform faults to account for the great oceanic fracture zones. A sequence of other important papers trans- formed ocean-floor spreading to plate tectonics and con- vinced all but the most recalcitrant oil-company geologist that the mobilist view was correct. But that is all part of a different story, and shed no further light on what is happen- ing at mid-ocean ridges. That revelation was already being achieved elsewhere, namely in the Cyprus Geological Survey. In Cyprus (and in other places such as Newfoundland, Oman, and Papua New Guinea, a combination that accounts for some curious stamps in my passport), there is a thick slab of rock, an ophiolite complex, made up of basalt, peridotite, and serpentinite, containing seafloor lavas and deep-sea sediments. Smaller fragments of similar rocks had long been known from mountain belts, and had been studied by, among others, Harry Hess. Now the Cyprus Survey, spurred by the discovery of iron, copper, and zinc sulphides, and chrome ore, decided to map the Troodos ophiolite, which was 100 kilometers long by 50 kilometers wide. The first map was started by R.A.M. Wilson. His work was a masterpiece of acute observation and justified interpretation that is a pleasure to read, even today. Within the ophiolite structure, which forms a gently warped and eroded sheet several kilometers thick, he found a unit composed entirely of dikes soon called the sheeted-dike complex. Dikes are thin, vertical sheets of magma, relics from when magma intruded into vertical cracks and became frozen there. They are common in the rock record, and show that the rock has been stretched when magma was around. In some places on the continents they make up perhaps five percent of the terrain, which up to that time had been considered a large amount. In Cyprus, Wilson showed that they make up 100 percent of one unit that is 1 kilometer thick and stretches for 70 kilometers across the mountains. 38 Ocean us Here was ocean-floor spreading frozen into geology, 70 kilometers of it, though Wilson did not make the connection at first. By the mid-1960s the link had been made, and the oceanic and ophiolitic strands of evidence became inextricably tangled. In the oceans it is possible to observe active mid-ocean ridges, especially using geo- physical methods, but very difficult to see what is happening below the seafloor, except by inference. In ophiolites the processes ceased long ago, but it is possible to wander over the countryside, passing deeper below the ancient seafloor at will, and reconstruct past events using the stan- dard tools of geology. The two approaches are complementary, but communication between them presented problems. Certainly there seemed to be a conflict between what Hess predicted for mid-ocean ridges and what was observed in ophiolites. When I first came into marine science, I trod warily, watching in admiration as my geophysical colleagues manipulated mathematics, patched instruments at sea, and set apparently arbitrary constraints on what was and was not possible. After all, I was an impeccably orthodoxly trained microscope man. Some moments were magical, as when Tuzo Wilson first propounded his transform-fault theory, grinning like a Cheshire cat that had swallowed the cream and snipping newspa- pers with a large pair of scissors to show how his theory worked. Other moments were more prosaic, and I gradually realized that geophysicists did not hold a monopoly on truth — or perhaps I just learned some geophysics. Was it possible to make a simple model of ocean-crust construction at mid-ocean ridges that drew on all of the evidence available, ophiolitic and oceanic, geophysical and geological? I had a false start: My first model was undone by a graduate student's simple question, "What determines the position of the Moho, the boundary between crust and mantle, in your model?" I said something in reply, floundering, hoping that he wouldn't notice. I expect he did. Then, a month or so later, digging the sandy soil of our garden in Norwich, I sud- denly saw what was wrong- perhaps also what was right. Suppose that magma rose up from the mantle as the plates moved apart. Suppose it rose high in the crust, not far below the seafloor, and collected there as a magma cham- ber, stretching along the axis of the mid-ocean ridge at a shallow depth. When the crust cracked above it, magma could rise along the crack to make a dike and then feed seafloor lava flows. The dike would intrude older dikes, and in turn cut yet older ones. When the magma chamber froze it would make a layer of gabbro in the lower crust. Crust produced Author Cann cooking porridge over a steam vent in northern Iceland in August 1991. The vent is part of the Theistareykir hot springfield, lying in the Theistareykir rift zone, and heated (almost certainly) by a magma chamber deep below the rift. Note that Iceland marks where the Mid- Atlantic Ridge comes ashore, though it is anomalous in many ways. Winter 1991/92 39 8 10 I B 8 10 Me\\ The " infinite onion" model (above) for magma chambers beneath fast-spreading ridge segments is compared to the "infinite leak" model (below) for magma storage beneath slow- spreading ridge segments. this way would have the same structure as the Cyprus (and now the Oman) ophiolites, and a shallow magma chamber might fit the geophysical observations, too. This outline evolved into the infinite-onion model, since in its ideal form it required a magma chamber that was onion-shaped in cross section and as long as that part of the ridge. Soon I was involved in stout defense of the onion. It proved very difficult to make the seismic observations that would test it properly, and inconclu- sive tests were regarded by skeptics as negative evidence. Soon it became clear that, in its simplest form, the model did not hold at slow-spreading ridges such as the Mid-Atlantic Ridge. Euan Nisbet and Mary Fowler devised an alternative, punning infinite-leak model to cope with that. Recently Debbie Smith and I have come up with observations that support infinite leaks in the Atlantic. But in the Pacific, where spreading rates are faster, there seemed every reason to expect the infinite onion. People looked for and found hot mantle, but no magma; they were looking for magma in the wrong place, it turned out. John Orcutt said he could see the magma. I liked his evidence, perhaps naturally, but many others stonewalled. Eventually, Bob Derrick (I simplify — John and Bob will have to stand for the teams they led) managed to image the top of the magma chamber for tens of kilometers along the ridge, using seismic reflection, just as the oil companies do to find oil — oil and magma look surprisingly similar by seismics. The chamber was much thinner than I had originally predicted, but the top was just at the right level (1 to 2 kilometers below the seafloor), as could have been predicted from ophiolites. And the onion? So far Bob and John stand out against the spike on the top of the chamber that would make the onion complete, but one day in the future...? Do I need to say that I still feel it is there? And what then? There is no space to tell of the other successful models of mid- ocean ridges, of George Constantinou in Cyprus showing that the ore deposits there were formed from hot springs on the ocean floor, thus leading the way toward black smokers; or the recent recognition that the third dimension, the variation of ridges along the axis, has as much of a Temporary High-level Chamber Gabbros and Cumulates Trapped Melt Relict Crystals 40 Oceanus story to tell as the across-ridge models we started with. But models are there to be overthrown: Perhaps the second-best experience as a scientist is to see a model elegantly destroyed. The best? To do it yourself by creating a new one, of course. In spring 1992, Debbie Smith and I will be leading an expedi- tion to the Mid- Atlantic Ridge, trying hard to do just that. Joe Cann is Professor of Earth Sciences at the University of Leeds in the UK and Adjunct Scientist at Woods Hole Oceanographic Institution. He took his Ph.D. in 1962 and is thus one of the old fogies of marine geology, but he is still trying hard to destroy his and other people's models of mid-ocean ridges. For the last few years he has been worrying more about black smokers than magmas, but recently he has come back into seafloor volcanoes. He works happily at sea or on land (in Greece or Cyprus) with a microscope, or a computer, or an X-ray set, especially on figuring out how the different aspects of mid-ocean ridges knit together. SL 1970 L A L 1974 lavas dikes isotropic gabbro . (Rising Asthenosphere layered gabbro M|i!!!!lharzburgite tectonite 1982 Lava Thickness TL Dike Thickness T. Cumulate Thickness T Magma chamber Blob of primitive melt deflated volume V0 volume V In the last 35 years, geological "certainty" has changed enor- mously. Models are created, proved, then sometimes disproved— with the end result, ultimately, of better understanding. Simple diagrams of models from the last 20 years illustrate (in a punctu- ated manner) this evolution. Winter 1991/92 41 The first volcanic rocks from a mid-ocean ridge were accidentally sampled during cable-laying operations in the North Atlantic in 1874. Sketches for this article are by the author. From Pillow Lava to Sheet Flow Evolution of Deep-Sea Volcanology Wilfred B. Bryan he black, fine-grained volcanic rock called basalt has long been associated with ocean basins, though sometimes for the wrong reasons. Today, basaltic lava is a familiar sight to millions of tourists who have visited Hawaii, and millions more have watched it flowing into the sea on television news programs. But 200 years ago in western Europe, basalt was known mostly by its association with sedimentary rocks containing marine fossils, and so was widely regarded as a chemical precipitate from seawater. A few practioners of the new science of geology at that time recognized the similarity of basalt to the lavas of nearby volcanoes such as Vesuvius. This led to one of the first major controversies in geology, between the so-called "neptunists" and the "plutonists," who believed that basalt was the product of volcanic eruptions. That issue was eventu- ally solved when one of the supposed basalt precipitates was traced back to its source at an obvious volcanic vent. But the nature and extent of volcanic rock on the deep seafloor would not be known for many more years. Prior to the mid-1960s, scientific papers on this subject still were largely constrained to rocks observed and collected on land; their association with sedimentary rocks typical of the deep seafloor contin- ued to be the principal evidence for their deep-sea origin. The first volcanic rocks from a mid-ocean ridge were accidentally sampled during cable-laying operations in the North Atlantic in 1874, about 200 nautical miles east of what we now know to be the Mid- Atlantic Ridge (MAR). The dark basalt was dismissed as having been dropped from a drifting iceberg, although the 27.5-ton tension required to recover the cable would seem to suggest this was not a loose fragment. In 1898 P. Termier described basaltic glass also recovered from the MAR at about 47°N, during cable repairs. He correctly deduced that this material indicated a volcanic origin for the seafloor at this location, but it would require another 60 years of study before the true extent and nature of the Mid-Atlantic Ridge would be known. Meanwhile, in the Pacific, widely scattered dredges recovered by the Challenger Expedition included samples of dark basaltic rock, also indicating a likely volcanic origin for the deep seafloor. Throughout the first half of the 20th century the seafloor was widely 42 Oceanus 75°- 30°- 60°- 120C 180C 120° 0 60C assumed to be basaltic, but evidence for this assumption was still sketchy and indirect. A "basaltic" and therefore "volcanic" seafloor was consis- tent with the arguments based on isostasy and bathymetry that remain valid today: The continents must stand high, because they are composed of relatively thick, light granitic rock that literally floats higher on the underlying mantle than does the thinner, heavier rock comprising the oceanic crust. Also, petrologists generally assumed that basalts of volcanic islands such as Hawaii or Iceland were representative of the rocks to be found on the deep seafloor. Although there are often striking differences between continental volcanic rocks and the deeper crustal rocks on which they have been erupted, the shaky logic of this analogy as applied to the seafloor does not ever appear to have been challenged. Finally, with the recognition of the reality of seafloor spreading and plate tectonics in the mid-1960s, mid-ocean ridge volcanism became a logical geometric necessity for creating new seafloor. The spreading model predicted that seafloor of similar basaltic composition but of regularly increasing age should extend to the margins of the ocean basins, a relation that was soon confirmed by basement samples recov- ered during legs 2 and 3 of the Deep Sea Drilling Program. Attention could now be redirected toward defining the nature of volcanic pro- cesses on mid-ocean ridges and the nature and extent of compositional variation in volcanic rocks erupted there. Locations are plotted from which oceanic basalts were dredged as early ridge petrologists defined compositional and structural boundaries between oceanic and continental crust. Marker's "Pacific" boundary and Hobbs 's "andesite line" bracket the circum-Pacific "Ring of Fire." Winter 1991/92 43 Chemical Variations Some of the most intriguing compositional features of ocean-ridge basalts are found in their trace-element and isotopic signatures. Because of their very fine-grained or glassy nature, volcanic rocks are most easily studied quantitatively by their chemical composition. Chemical analyses of volcanic rocks in and around the major ocean basins began to appear in the latter half of the 19th century. By the beginning of the 20th century there were already enough data to support speculation on the global distribution of volcanic rock types; in these schemes it was implicit that the volcanic rocks somehow reflected the nature of the ocean floor with which they were associated. One of the best-known global distributions was proposed by the British petrologist Alfred Marker in 1909. He recognized three main groups, which were named for the ocean basins in or adjacent to which they were first identified. The "Atlantic" type was characterized by the dominance of soda (sodium oxide), the "Mediterranean" type by potash (potassium oxide), and the "Pacific" type by lime (calcium oxide). Marker's "Pacific" type, however, was based entirely on data from volcanoes and volcanic islands from the "Ring of Fire" around the Pacific margin. Almost immediately, new data from various Pacific Islands proved similar to those from the Atlantic, and Marker's scheme was discredited. About 20 years later, W.H. Hobbs called attention to the composi- tional differences between volcanic rocks from islands within the Pacific Ocean basins and those of the volcanic-island arcs and continental volcanoes along the Pacific margins of Asia and North and South America. It is interesting to compare this boundary, which Hobbs called the "andesite line," with the boundary drawn by Marker between the "Pacific" and "Atlantic" rock groups. Following Hobbs, most geologists and volcanologists quickly accepted the andesite line as the structural and compositional boundary between oceanic and continental crust. It was not until detailed mapping and sampling of some of the circum- pacific volcanic-island arcs in the 1950s and 1960s that Harker's bound- ary was rediscovered. Chemical analyses of rocks specifically associated with mid-ocean ridges were not published until the 1930s in papers by C.W. Correns and J.D.H. Wiseman. Their samples came from the Mid-Atlantic Ridge and the Carlsberg Ridge in the Indian Ocean. Both authors recognized the unusually low potash contained in these rocks compared to both the island basalts and continental rocks, and correctly deduced some of the chemical effects of seawater alteration on basalt. Wiseman's paper contained the first carefully detailed drawing of crystal forms observed with a petrographic microscope; Correns recognized the similarity of his sample to those collected in the Pacific by the Challenger Expedition, and suggested that these might be typical of the seafloor as a whole. Some of the most intriguing compositional features of ocean-ridge basalts are found in their trace-element and isotopic signatures, but these data had to await the mid-1960s development of more sophisticated analytical technology. Analyses of basalts dredged both from the Mid- Atlantic Ridge and the East Pacific Rise showed that, compared to typical basalts of continents and oceanic islands, ocean-ridge basalts are highly depleted not only in potash but in many trace elements chemically similar in behavior to potash, such as lanthanum, rubidium, thorium, 44 Oceanus and uranium. Because these elements are concentrated in typical volca- nic rocks on Earth's surface and upper lithosphere, geochemists refer to them as "large-ion-lithophile elements." Based on these data, some geochemists emphasized the depleted and homogeneous nature of ocean-floor basalt. This view was quickly challenged when new analyses of basalts from the northern Mid-Atlantic Ridge that were enriched in these same chemical elements were pre- sented. However, the most extensive early collections of samples from a mid-ocean ridge were recovered from the Mid-Atlantic Ridge between 22° and 30°N, and the "depleted" chemical character of these basalts became established as the definitive signature of "normal MORB" (mid- ocean ridge basalt). At the University of Rhode Island, Jean-Guy Schilling published a series of pioneering papers that first conclusively showed the gradational nature of geochemical variability along ocean ridges. First demonstrated in the North Atlantic, these along-ridge variations are now known to continue through the equatorial region into the South Atlantic, and are also present along the Galapagos Rift and southern East Pacific Rise. Sections of ocean ridges enriched in potash, trace elements such as lanthanum, thorium, and uranium, and with high strontium-87/stron- tium-86 were shown to be associated with shallow bathymetry or island platforms such as Iceland and the Azores. This appeared consistent with the idea that these are "hot spots," characterized by extensive melting of a mantle source enriched in these elements and in radiogenic isotopes such as strontium-87. Hot Spots: How Normal is Normal? Although models remain sketchy and highly speculative, a popular view is that hot spots are the locis of upwelling "mantle plumes" that bring new, hot, and previously undepleted mantle from a deep, previously untapped source to a sufficiently shallow level, permitting partial melting and the escape of basaltic magma to the surface. On the other hand, depleted, supposedly "normal" mid-ocean ridge basalt is pre- sumed to be derived from relatively shallow mantle, perhaps the "low- velocity zone" defined by seismic surveys, which may have been de- pleted by partial extraction of magma in previous melting events. Mixing between melts derived from these two sources may account for much of the intermediate isotopic and trace-element variability. Major hot spots are now recognized along the Mid-Atlantic Ridge at Iceland, near 45°N; the Azores, near 15°N; and near Tristan da Cunha, at about 36°S in the South Atlantic. The chemical signature of the larger plumes extends hundreds of kilometers along-ridge, and it can be asked, at least in the Atlantic, if "plume" MORB isn't actually more normal than "normal" MORB! In the Pacific, the Galapagos Rift crosses the best-documented hot spot, but another must exist on the East Pacific Rise near Easter Island, and there are several small ones along the Gorda-Juan de Fuca Ridge systems. Ridges have been less systematically sampled in the Indian Ocean, but the available dredges and Deep-Sea-Drilling-Program basement samples indicate both "normal" and "plume" chemistry in Major hot spots are now recognized along the Mid- Atlantic Ridge at Iceland, the Azores, and near Tristan da Cunha. Winter 1991/92 45 Author Bryan (left) and T.H. Pearce (riglit) on an expedition in Quebec stand in front of classic Archean pillow lavas. basalts recovered there. There now is even a "cold spot" recognized along the Pacific-Antarctic Ridge, characterized by extreme depletion in large-ion-lithophile elements. Most recently, researchers at the Lamont- Doherty Geological Observatory have shown that systematic along-ridge variations can be demonstrated in major-element chemistry as well as in trace elements and isotopes, and also can be correlated with bathymetry and the location of hot spots. Petrography and Mineralogy Much of the early work on ocean-floor basalts was based on chemical analyses and ignored mineralogical details of the rocks. In 1972, 1 described in detail for the first time the sometimes-bizarre crystal morphology that results from rapid underwater quenching of magma. An unexpected result of this paper was the recognition of similar quench-crystal morphology in Archean pillow lavas that are up to 3.5 billion years old. Previously these morphologies were believed to have been caused by chemical changes over time, accompanied by recrys- tallization. Now it was obvious that these basalts had changed little since they originally erupted on ancient seafloor, and both their chemistry and morphology could be used to interpret volcanic processes in some of the oldest seafloor known, now uplifted and exposed on land. While many of these ancient basalts resemble their modern counterparts, others do not, including some unique varieties that are very enriched in magnesium, nickel, and chromium. Mineralogically, these rocks, known as komatiite, are unusually rich in olivine, the major mineral component of the upper mantle. One possible interpretation of these komatiites is that they were derived from an oceanic lithosphere much thinner and with a much steeper thermal gradient than that observed today, resulting in more complete melting of the mantle source. Morphology and Volcanic Processes The size, shape, and other morphologic details of ocean-ridge lava flows and associated volcanic structures provide important clues to the loca- tions of eruptive vents, rates of eruption, flow mechanisms, and lava distribution on the seafloor. As for deep-sea basalt compositions, the question whether deep-sea lavas had a unique morphologic character was heatedly debated for nearly a century. Exposures of these lavas in cliffs, road cuts, or on glacially eroded and smoothed outcrops on land were largely two-dimensional, and left much room for arguments about the lateral extent of individual flows, 46 Oceanus their three-dimensional forms, and the nature of the larger volcanic structures they built. In cross section many of these lava flows consist of elliptical to circular masses, 2 meters to over 1 .5 meters in diameter; these classic "pillow lavas" have been cited as proof of eruption underwa- ter at least since the first half of the 19th century. However, whether these pillows are spherical or tubular in three dimensions, and whether they uniquely indicate underwater eruption, was argued for many years. Central to these debates was a Tlie sizes and shapes of lava flows reveal information about the mechanics of eruptions. Sheet flows from the East Pacific Rise (above) and a layered lava tube in subglacial pillow lava in Iceland (left) are vastly different morphologically. A typical flow from the FAMOUS area (below, left) is further illustrated with a schematic cross section that reveals the draining of lava. Winter 1991/92 47 This "elephant seal" pillow formed at the end of a Hawaiian pahoehoe lava flow. question: Should the term "pillow lava" be reserved only for circular or spherical structures in lava erupted underwater (or at least in wet mud), or should it also be applied to morphologically similar lavas formed on land? The case for purely descriptive use of the term "pillow" was well argued as long ago as 1938 by J. T. Stark, who pointed out that in his even earlier 1914 review of the subject, J.V. Lewis had cited 98 descrip- tions of "pillow lava" dating back to 1834, of which more than half were probably formed on land. Nevertheless, questions continued to arise as to whether similar morphologies could be produced in different ways. For example, in 1968, J.G. Jones documented "pillow lava" composed of interconnected and elongated tubular lava fingers analogous to the tubular "pahoehoe" lava commonly observed in Hawaiian lava flows (as also advocated by Lewis in 1914!). This interpretation was challenged, and the issue would not be put to rest until the mid- 1970s, when scuba divers ob- served pillows forming on the submarine extension of an active lava flow in Hawaii, and the first direct observation of deep-sea lavas was made by diving scientists in the Project FAMOUS (French- American Mid-Ocean Undersea Study) on the Mid -Atlantic Ridge in 1974. These lavas were indeed composed of elongated tubes, which grow downslope by budding, as Lewis long ago deduced. Recent observations of new submarine flows in Hawaii also confirm that they are fed by master feeder channels that are direct extensions of the adjacent island's pahoehoe lavas. Diving scientists have provided abundant photographic records, direct observations, and descriptions of the great variety of morphologic details in lavas of the mid-ocean ridges. These observations make it clear that elongated, tubular lava units are common on steep-flow fronts, but pillows take many forms: On the upper flow surfaces some are hollow bubbles, but others are highly ornamented sculptures that resemble animal or human forms. Many similar forms are also found on land on pahoehoe lava flows The first submersible dives on the East Pacific Rise showed that many lavas are not pillowed at all, but are composed of slabby plates ornamented with swirls and wrinkles suggestive of drapery or a wrinkled tablecloth. These "sheet flows" form when lava is temporarily ponded. The lava sheets are produced by quenching against the overly- ing seawater. When lava pressure breaks the barrier and lava drains away, successive layers form as the level of the pond drops. Hollow columns of lava surrounded by "bathtub rings" form within these pits, where trapped water vapor has risen through the lava and quenched it. On land, analogous features are found in "shelly pahoehoe," where lava has temporarily ponded around trees, and in collapse pits formed in lava 48 Oceanus that erupted onto wet ground. It is now obvious that this diversity of form can be related to a variety of factors, including the steepness of the flow surface, the rate and volume of lava extrusion, and the influence of the underlying seafloor morphology, and that the morphologic differ- ences resulting from quenching in air or water are relatively minor. Small conical or moundlike volcanic structures form over eruptive vents on the seafloor as they do on land, but few have been described in detail. Some appear to be typical extrusive lava mounds similar to those that form over active lava tubes or along eruptive fissures in Iceland or Hawaii. Larger cratered cones and mounds, common on the Mid- Atlantic Ridge between 22° and 26°N, have been a special focus of study by geologists at Woods Hole Oceanographic Institution. One of these, named Serocki Volcano after one of the Ocean Drilling Program engi- neers, has been mapped in detail, observed at close range from a sub- mersible, and even penetrated by drilling. These studies indicate Serocki has a flattish, pancakelike form and is probably not a true volcano but rather a "rootless vent." Originally a thick lava delta, the north flank of Serocki broke open, allowing lava trapped within to escape to a lower level, where it again ponded temporarily to form another delta. This delta in turn also broke open, and was drained; collapse of the unsup- ported surface crust on both deltas created the central craters. Looking Toward the Future About 25 years ago, marine geologists and geophysicists first became aware of the vast extent and importance of volcanic activity along mid- ocean ridges. Following initial hopes that the resulting volcanic rocks Author Bryan sketched this cross section of the Serocki volcano region based on Sea Beam bathymetry, Sea MARC sidescan images, and observa- tions by diving scientists in DSV Alvin. The Serocki volcano is the large opening at left. Winter 1991/92 49 Long-term observatories will be required on selected parts of the Mid- Ocean Ridge system to document eruptive events. would prove to be unique and homogeneous, we are continually recog- nizing the great geochemical and mineralogical variability in ocean-ridge basalts. The morphological similarities of submarine lavas and other volcanic structures to lava flows on well-studied land volcanoes indi- cates that processes of magma generation and eruption are similar in both environments; thus, lessons learned in the study of the more accessible land-based volcanoes can be applied to volcanic processes on the deep seafloor. Already, some consistent correlations are beginning to appear between certain chemical parameters and first-order geophysical and morphological seafloor properties such as depth, gravity field, and spreading rate. The most profitable future work is likely to come from geophysical and petrologic studies carefully designed to integrate both compositional data and physical properties of the ocean crust into comprehensive models for melt generation, ascent, and the "plumbing system" beneath ocean ridges. Just as has been true of land volcanoes, long-term observatories will be required on selected parts of the Mid-Ocean Ridge system to docu- ment eruptive events, associated seismic activity, and subsequent hydrothermal processes. Although the long controversy about the significance and mode of pillow-lava formation has ended, much re- mains to be learned about the growth of submarine volcanoes and the mechanisms of lava distribution on the deep seafloor. Individual lava flows must be mapped and sampled in detail, and their morphologies carefully documented. Many morphologically diverse small volcanoes and seamounts associated with active ridges must be imaged, sampled, and restudied as they evolve with successive eruptions. Such long-term observations are being discussed and planned as part of the National Science Foundation-sponsored RIDGE initiative (See article, page 21), but the magnitude of the commitment required for definitive results is sobering. For example, observations carried on for over 50 years at Kilauea Volcano in Hawaii are only now beginning to yield a meaningful understanding of the volcano's eruption mechanics and deep plumbing system. Further, this length of time still has not been long enough for all styles of activity, as deduced from older lava and ash deposits, to have been repeated for recording and analysis using modern instrumentation. Emulating this work on our largest terrestrial basaltic volcano, the 60,000-kilometer-long Mid-Ocean Ridge system, remains a major challenge. ""% Wilfred B. Bryan is a Senior Scientist in the Department of Geology and Geo- physics at the Woods Hole Oceanographic Institution. He was Chief Diving Scientist in Project FAMOUS, and has participated in studies of volcanic activity on other parts of the Mid-Atlantic Ridge and East Pacific Rise. He was a principal investigator in lunar volcanic landform studies for the Apollo Program and has documented volcanic morphology and processes in Hawaii, the Southwest Pacific, Iceland, Italy, and the western US and Canada. 50 Oceanus Tectonics of Slow-Spreading Ridges Jeffrey A. Karson s oceanic plates diverge at mid-ocean ridge spreading centers, two major processes produce and modify oceanic lithosphere. The most familiar of these is magmatic con- struction in the form of volcanic extrusion onto the seafloor, probably accompanied by the intrusion of dikes and larger bodies of coarse-grained, igneous material beneath the seafloor. Just as important, however, are the effects of mechanical extension, faulting of brittle surface materials, and plastic flow of hotter material in the lower crust and upper mantle. At fast-spreading ridges, magmatism nearly keeps pace with plate separation, so each increment of separation is accompanied by sufficient igneous activity to fill any cracks and fissures in the seafloor and bury most of the minor fault scarps created since previous eruptions. In general, the wound inflicted along the ridge axis is regularly healed, resulting in the formation of what geologists call an "axial summit graben atop a very elongated shield volcano." At any instant in time, the plate boundary resembles a series of linked cracks in brittle material similar to cracks in a pane of glass. In contrast, slow-spreading ridges display a completely different interplay of mechanical extension and magmatism. Here the magma supply is insufficient to completely restore the faulted axial crust. Magmatism is discontinuous and episodic along the ridge axis despite the relentless separation of the plate edges. The result is that the axial crust at those edges is stretched and faulted in a manner similar to that of continental rifts. This article describes some new insights gained from submersible studies and continental rift analogs. Ridges and Rifts: A Morphologic Comparison In the late 1950s Bruce Heezen and colleagues at Lamont-Doherty Geological Observatory discovered a deep cleft in the crest of many parts of the mid-ocean ridge system. Based on similarities to profiles of the East African Rift, they considered this cleft to be a rift valley produced by extensional faulting of the oceanic crust. Much that has been learned in the past 30 years about the geologic architecture of land and the seafloor has stimulated a cross-pollination of ideas that derive from the different constraints and limitations of these environments. At any instant in time, the plate boundary resembles a series of linked cracks in brittle material similar to cracks in a pane of glass. Winter 1991/92 51 A B Accomodation Zones 20 40km The generalized geologic structure of the MARK Area on the Mid-Atlantic Ridge (A) and the Turkana Rift of northern Kenya (B) allow a comparison of slow-spreading oceanic and continental rifts. Both are composed of a series of discrete rift segments several tens of kilometers in length linked by accommodation zones (stippled). Some segments have neovolcanic ridges (A) or quaternary volcanic centers (B) (black). Hatched areas are rift-shoulder uplifts; dashes are exposures ofplutonic rocks; squiggles are serpentinites; bold lines with boxes are major normal faults; and lines with or without tick marks are normal faults and fissures. Bars labeled A, B, C indicate cross sections referred to in the text. Although continental crust is typically about 35 kilometers thick compared to only about 6 kilometers for oceanic crust, rifts in these two settings are very similar both in scale and form. This is a result of the dominating effect of temperature, which determines the lithosphere's strength and thickness. The lithosphere is hot, thin, and weak at rifts, but becomes thicker and stronger as it cools or moves away from a rift. As it cools, the mantle beneath the crust becomes strong and controls the rifting process. Oceanic lithosphere has a greater proportion of this strong mantle than does continental lithosphere of similar thermal structure or lithospheric thickness, limiting rift development in old oceanic lithosphere. 52 Oceajius If the oceans were drained, Earth's slow-spreading ridge systems would resemble the well-known continental rift valleys, for example, the 4,000-kilometer-long East African Rift. The Mid-Atlantic Ridge occupies nearly one-third of the seafloor beneath the Atlantic Ocean. It is broad and undulating, with crests at hot spots like the Azores and Iceland. This large-scale morphology is similar to the 100-kilometer- wide topographic domes of Kenya and Ethiopia, upon which the East African Rift is superimposed. On a finer scale, the continental and oceanic rift valleys are segmented, that is, they are made up of a series of discrete fault- bounded rift valleys. Each segment is several tens of kilometers long and is linked end-to-end with adjoining valleys to form a nearly continuous structure thousands of kilometers in length. Minor offsets, misalign- ments, and overlaps of the rift-valley segments are typical of both oceanic and continental rifts. This segmentation is also evident in the gravity, magnetics, and seismic characteristics of rifts in both settings. Viewed in profile, opposing rift valleys are commonly asymmetrical; one bounding wall is higher and steeper than its mate across the axis. Major faults with hundreds to thousands of meters of displacement occur on the steep sides, and smaller faults and smoothly bent layers occur on the lower sides. Thus, half-graben forms are more common than the symmetrical full grabens with equal-sized faults on both sides of the valley. The valley depths are comparable, generally around 2,000 meters. Lavas partially fill both types of rift valley, and sediments deposited in rivers, lakes, and deltas reach several-kilometer thicknesses in the continental rifts. Where the faulted rift-valley walls overlap, the roughly symmetrical fault- bounded troughs called grabens or uplifted blocks called horsts are created. Other areas, where no overlap occurs, may have no rift valley at all, just a rugged, faulted terrane that occupies the ridge axis. The Neovolcanic Zone The neovolcanic zone of the mid-ocean ridge system is the fresh bead of lava that welds the ridge axis together. Along slow-spreading ridges this most-recent volcanic material forms an imperfect seam, with many large globs and gaps. The globs are referred to as neovolcanic ridges; their lustrous lavas and very thin sediment dusting indicate they are only a few thousand years old. The gaps are filled with faulted and fissured lavas that erupted tens to hundreds of thousands of years ago, an earlier version of the neovolcanic zone. The discontinuous nature of the young lavas reveals that these areas are fed by a sputtering magma supply and that the temperature of the lithosphere along the ridge axis is highly variable. The discontinuity of the neovolcanic zone, as well as seismic 0 10km 71 Symmetrical Graben Asymmetrical Half-Graben Symmetrical rift segments produced by graben structures are common in many rifts, for example areas marked by bars labeled "A" opposite. Asym- metrical half-graben rift segments are also common, for example nrens marked by bars labeled "B" opposite. Note that half-grabens may overlap to create a symmetrical graben morphology. Winter 1991/92 53 Various types of normal faults occur in rifts, including steeply dipping planar faults (top), rotated planar faults creating a domino fault-block pattern (second), Hstric (curved) faults merging doumward into a horizontal detachment fault (third), and detachment faults cutting across the full thickness of the crust (bottom). Steep planar faults dominate the median valley walls of the Mid-Atlantic Ridge creating a stair-step shape and cutting any earlier faults of the median valley floor. Median Valley Wall Median Valley Floor studies of the rift valley, indicate that there is no continuous magma chamber beneath the axis of slow-spreading ridges. A similar scenario applies to continental rifts. Major differences exist in the eruptive styles of typical oceanic and continental rifts. In continental (subaerial) rifts, extensive basaltic outpourings may precede the formation of a rift valley, and explosive volcanism that produces cinder and ash cones is common. In contrast, submarine rifts are dominated by monotonous fields of pillow-lava mounds and ponded sheet flows. In both continental and oceanic rifts, it appears that young magmatic centers are associated with the rift-valley segments. Whereas small volcanoes appear to have reached the surface through conduits provided by various faults, the largest neovolcanic ridges appear to be centered in well-defined grabens or half-grabens; thus the spacing of volcanic centers appears to be similar to that of rift-valley segments. The distribu- tion of very young volcanic rocks along slow-spreading ridges is known in only a few areas, but in the eastern branch of the East African Rift, extending from the Afar Triangle to northern Tanzania, the young volcanic centers have a remarkably regular spacing of about 3 to 5 kilometers along the rift axis. Future seafloor mapping will determine if a similar volcanic chain exists along the mid-ocean ridge axis. Fault Structure Axial lithosphere stretching is accommodated by faulting in the upper crust. A number of factors determine the depth of faulting; the most important is the temperature of the lithosphere. In some places, where the lithosphere along the ridge axis is very cold, and where there has not been a recent (less than 1 million years) magmatic event, faulting marked by earthquakes affects the entire crust and even the upper mantle beneath the spreading center. Because the axial lithosphere is relatively thick and strong in such cases, faulting affects a wide area, including the rift-valley walls. In ridge segments where magmatic events have occurred recently (perhaps marked by neovolcanic ridges), the lithosphere is relatively thin and hot and faulting affects only the narrow axial region. It is likely that some ridge segments vacillate between these two situations as axial temperatures wax and wane. Almost certainly a similar variation occurs in continental rifts, but it is complicated by the breaking of thicker, less uniform continental crust with its many preexisting zones of weakness. Still, hot, weak areas are expected to develop a rift architecture distinct from that of cooler areas. Scientists participating in the first submersible studies of the mid-ocean ridge in the early 1970s de- scribed rift-valley walls with numerous closely spaced faults separating narrow blocks of crust and displacing them to form a stair-step structure. Subsequent studies have shown that this is just one of a family of ridge fault 10 km structures that includes simple planar faults, curved 54 Oceanus Inactive Faults Active Faults Dike injection Periodic Magma "teth^olphere Chamber Uplifted Rift Shoulder Detachment Minor (lateral?) Lavas & Dikes in Upper Crust Gabbroic Lower Crust Mantle Ductile Deformation B Hanging Wall Lithosphere Asthenosphere Old Shear Zone Major Shear Zone 10 km "listric" faults, and very large continuous fault zones. All of these have been studied in great detail in continental environments, but links between processes of plastic flow and magmatic intrusion in the lower crust and upper mantle are poorly understood at present. Although only a few examples of rift-valley faulting have been studied to date, they appear to follow a relatively simple pattern. Areas with relatively high magma supplies that do not display large amounts of extension have simple planar fault structures, and tend to form symmetrical rift valleys where only basaltic rocks are exposed. Good examples are known from the FAMOUS and AMAR rift valleys of the Mid-Atlantic Ridge. Areas with somewhat larger amounts of stretching and less magmatism become asymmetric as faulting on one side of the rift valley begins to dominate. Listric and low-angle detachment faults occur in some limited areas such as the TAG site. More stretching results in extreme extension along gently to moderately inclined fault zones, and may result in the exposure of materials once deeply buried. In the most extreme situation known, a chaotic faulted assemblage of mixed upper-crustal and mantle materials occurs across a ridge axis in the MARK area. Although some cross sections in this area lack a clearly defined rift valley, the ridge axis is nevertheless very highly extended. This spectrum of fault structures mimics that of continental rifts. In both settings, the amount of crustal stretching and displacement that indi- vidual faults have sustained can be read in the types of rocks exposed just beneath the fault surfaces. In the oceans, small amounts of extension result in exposure of only basaltic rocks of the upper crust, while large amounts can expose once deeply buried lower crustal rocks. In oceanic areas, where the crust is only 6 kilometers thick, even upper-mantle rocks can be exposed. Ductile stretching of the crust and mantle is punctuated by periodic injection of basaltic dikes beneath a symmetrical rift valley with a neovolcanic ridge (A). A deeper, broader, asymmetrical rift valley is created by concentrated slip on a major detachment fault in cooler lithosphere (B), for example section "C"on the left figure on page 52. Episodes of magmatic (A) and amagmatic (B) spread- ing may alternate over periods of a few hundred thousand years in the same ridge segment. Adjacent spreading segments may be as different as these two extreme examples or an intermediate stage. Winter 1991/92 55 In the past few years, seismic reflection studies have provided remarkable new images of the oceanic crust's internal structure. Highly Extended Terranes In both oceanic and continental rifts, prolonged periods of plate separa- tion with little or no magmatic activity result in extreme stretching of the crust and upper mantle as described. Such highly extended terranes are well known in continental areas such as the Basin and Range Province of the western US, where faulting has been localized along individual fault surfaces called "detachment faults." These gently inclined dislocation surfaces cut across rock units and smaller faults of the upper crust that are free to rotate in a fragmented upper plate. Beneath, more plastic flow occurs in a lower plate. There is continuing debate concerning the inclination of these faults when they were actually slipping. One school of thought argues that only steep faults are mechanically feasible, and that the detachments were formed by the rotation of steep fault segments that coalesced into a single longer segment. Others propose that low-angle detachments have maintained their near-horizontal attitude as fault blocks rotated above them, allowing the detachment fault and lower-plate rocks to come closer to Earth's surface. Still another hypothesis holds that the detach- ments are individual giant faults, along which lower-plate rocks have been pulled from deep beneath the overlying upper plate. In this case, the detachment surface would be warped by vertical movements driven by gravitational adjustments, maintaining a gently inclined attitude. Regions of significant stretching also appear to exist along mid- ocean-ridge spreading centers. Like continental detachment faults, they are marked by major fault surfaces that expose deep crustal or even upper-mantle rocks at the surface, or juxtapose them with shallow-level rocks. In these areas, faulting must have been localized along single-fault surfaces for long periods of time as plate separation continued. As a result, 2 to 5 kilometers of crustal rocks have been stripped of underlying deep-crustal and upper-mantle materials. Unfortunately, areas of such exceptional faulting are difficult to find. At present, there is no unambiguous link between the morphology of the rift valley and the type of fault structure and rocks exposed there. This is because numerous steep faults often cut and break up the large detach- ment surfaces, giving even highly extended rift valleys a form not unlike their less-stretched cousins. Studies of continental rifts show that earth- quakes detectable with conventional instruments occur only on steeply inclined faults. Slip on buried, low-angle detachments is known to occur in some areas from the study of surface structures, like upper-crustal faulting. Although required to link upper-crustal faulting to flow in the lower crust, slip on such surfaces appears to occur without major earth- quakes. In the oceans, such seismically quiet displacement could be taking place undetected, because we have no detailed seismic studies or maps of surface faults. In the past few years, seismic reflection studies have provided remarkable new images of the oceanic crust's internal structure. One surprising feature of these sonograms is that they reveal numerous low- angle reflectors cutting across the entire crust and sometimes intersecting the surface at probable fault scarps. These structures appear to be major detachment faults in the oceanic crust, spaced at about 30-kilometer 56 Oceanns intervals. This translates to a periodicity for major faulting and magmatic events of about 300,000 years, some 30 times longer than estimates based on the apparent ages of lavas in the median valley floor. At present, the interpretation of these features (as well as other intracrustal reflectors ) is still debated; however, considering the rift valley's presently known geology, detachment faulting is the most logical explanation. Detachment faulting is also the most likely means of exposing coarse-grained gabbroic and peridotitic rocks of the deep crust and upper mantle along oceanic rift walls. The conditions that result in extension with little or no magmatism could conceivably occur in any ridge segment. However, persistently cool spots along the spreading centers would be very likely places for these conditions. The intersec- tions between rift segments and oceanic transform faults where a ridge axis abuts a cold transform-fault wall are thought to be lithospheric cool spots that may have very limited magma supplies. They are, therefore, likely places for this type of extreme faulting. If cool, stretched crust were formed at a ridge-transform intersection, it would pass laterally along a transform fault and become the wall of an oceanic fracture zone. This may explain the common occurrence of deep-level rocks along fracture zones. Connecting Structures Individual rift segments are connected end-to-end by various types of linkages. Some are discrete crustal faults while others are more diffuse regions of bending or shattering. These features can be considered collectively as "transfer zones," a term first used to describe linkages in compressed and folded rocks of mountain belts, and useful to describe the geometry and kinematics of rift linkages as well. Transfer zones in continental rifts take many different forms, depending upon the charac- ter of the faults in the rift segments they join. The diversity of fault structures along slow-spreading ridges suggests that transfer zones are important components of rift valleys. For example, different types of transfer zones link rift-segment pairs that differ in amount of crustal stretching, amount and timing of magmatic events, rate of stretching, and style of faulting. In many cases asymmetrical rift-wall faults overlap to create a class of transfer structures referred to as "accommodation zones" that are typical of continental rifts. If small amounts of extension have occurred, simple ramps or folds may suffice to transfer the effects of faulting from one rift segment to the next. This geometry appears to be typical of many oceanic and continental rifts. However, major faults with large horizontal displacements may be required if large amounts of extension occur in even one segment. The well-known transform faults that occur along spreading centers can be regarded as just the largest of a family of these transfer structures. The asymmetry of the median valley of the Mid-Atlantic Ridge and the geometry of steep linear slopes that suggest major faults create a pattern very similar to that of continental rifts. At present, however, the details of fault geometry and slip directions for mid-ocean-ridge spread- ing-center segments and possible linking transfer zones are almost completely unknown. The morphologic similarity, however, suggests Jlie diversity of fault structures along slow- spreading ridges suggests that transfer zones are important components of rift valleys. Winter 1991/92 57 Fast-spread crust is likely to be uniform and continuous; slow-spread crust is apt to be much more heterogeneous. that the growing body of detailed information on continental-rift faulting can be applied at least in a general way to slow-spreading ridges. It is important to recognize that the morphology of oceanic and continental rifts is not necessarily a reliable indicator of their fault structures. In particular, low-angle faults and strike-slip faults that do not produce significant topographic or bathymetric relief are difficult to detect with remote mapping systems such as multi-beam echo sounders, including Sea Beam, or sidescan sonar systems like Sea MARC. Thus, the present perception of oceanic rift structure is strongly biased by steep faults with significant vertical offsets. These are certainly important components of oceanic rifts; however, in some cases they represent only a small amount of the total extension revealed by fault structures and lithologic associations documented by detailed near-bottom investiga- tions. It is clear that mapping seafloor morphology with remote-sensing systems will not be sufficient to evaluate the geometry and extent of faulting for segments of the mid-ocean ridge system. Much more de- tailed near-bottom sampling and mapping from submersibles such as Alvin or remotely operated vehicles like Argo-Jason will be required. Implications of Major Faulting on Slow-Spreading Ridges The recognition of major faulting in the median valley of slow-spreading ridges has some important implications for understanding seafloor- spreading processes and oceanic-crust production in these environments. First, the extreme faulting found in some places, like the MARK area and some transform-valley walls, suggests that long periods (perhaps as long as 1 million years) of plate separation occur with little or no magmatic construction. This implies that long magmatic "droughts" occur at least locally along the ridge axis, contrasting sharply with observations on fast-spreading ridges that suggest a persistent, robust magma supply. It follows that the end products of accretion of fast- and slow-spreading ridges may be very different geologically. Whereas fast-spread crust is likely to be characterized by a generally uniform and continuous geo- logic structure, slow-spread crust is apt to be much more heterogeneous. Swaths of slow-spread crust tens of kilometers across probably resemble that created at fast-spreading ridges. However, if the geology of the present-day median valley of the Mid-Atlantic Ridge is a valid guide to slow-spreading processes in general, there must also be patches of highly disrupted and broken crust that might be essentially a jumbled mass of faulted oceanic crustal and upper-mantle blocks that are locally welded together by intrusive dikes and lava flows. This inference must somehow be reconciled with the well-documented fact that the seismic structures of fast- and slow-spread crust are nearly identical. It is probable that the seismic structure is dictated by fractures and rock porosity rather than rock compositions, a possibility that would limit the usefulness of seismic studies in oceanic geology investigations. The extreme type of faulting described above also raises some questions regarding the origin of lineated marine magnetic anomalies. How can they persist if the basaltic layer of the crust, generally thought to be the source of magnetization, is highly faulted and even discontinu- ous? Typical lineations are found over several areas where basaltic rocks 58 Oceanus have been faulted away to expose deep-crustal or even upper-mantle rocks. These occurrences suggest that magnetic lineations can also be produced by the magnetization of metamorphosed deep-crustal and upper-mantle material. This seems feasible if the magnetization is acquired during faulting close to the median valley, the same place the basalts have their magnetization frozen-in during normal spreading. Faulting produces the major fracture porosity in all parts of Earth's crust. If the fault patterns of continental rifts can be used as a template for slow-spreading ridges, it should be possible to predict the fault and porosity patterns of the seafloor at least in a general way. Faulting and fracture porosity are likely to control the locus of magmatism and hydrothermal venting along spreading centers, just as they do in conti- nental settings. This relation could help explain patterns of volcanoes and black smoker vents that are just beginning to emerge from near- bottom studies, and might even prove important as a prospecting tool for spreading-center ore deposits. There is a growing awareness that fast- and slow-spreading ridges function in very different ways. The sputtering magma supply of slow- spreading ridges results in substantial periods of plate separation that involve stretching and faulting of relatively cool oceanic lithosphere with little or no magmatism. The fault patterns of the median valley appear to mimic those of continental rifts; however, at least locally, very highly stretched and thinned masses of crust and upper mantle occur. The median-valley geology and fault structure documented by near-bottom studies predict a very heterogeneous geological structure in slow-spread crust. This result is yet to be clearly defined or reconciled with the geophysical expression of the crust away from spreading centers. Future studies of the geometry and kinematics of faulting on slow-spreading ridges will determine the nature of faulting over much larger areas than have been studied to date, and will help contribute to the overall under- standing of how the lithosphere is pulled apart to form rifts in both the continents and the seafloor. Jeffrey A. Karson is an Associate Professor in the Department of Geology at Duke University. He is a field-oriented structural geologist who has studied the nature of faulting on the seafloor at both fast- and slow-spreading ridges and associated transform faults during seven diving programs using DSV Alvin and other submersibles. He maintains a parallel research program in the East African Rift and ophiolite terranes. There is a growing awareness that fast- and slow- spreading ridges function in very different ways. Winter 1991 192 59 Seismologists soon realized that a relatively narrow band of earthquakes could be traced through many of the world's ocean basins. Mid-Ocean Ridge Seismicity Eric A. Bergman he earliest observations of earthquakes in deep-ocean basins were reported by seamen whose ships were rocked by undersea disturbances, and by residents of oceanic islands such as the Azores and Iceland. Scholarly studies based on eyewitness accounts of so-called "seaquakes" began in the late 19th century, but systematic investigations of oceanic seismicity did not begin until a global network of earthquake observatories was estab- lished in the early decades of this century. By the 1920s the International Seismological Summary (ISS) was routinely compiling data from cooper- ating stations, and publishing earthquake locations. Seismologists soon realized that a relatively narrow band of earthquakes could be traced through many of the world's ocean basins. The earthquakes were associated with the mid-ocean mountains that had been revealed by early oceanographic surveys, but the global continuity of mid-ocean ridge seismicity was not demonstrated until the mid-1950s. For several decades after these initial discoveries, mid-oceanic earthquakes attracted relatively little research interest. Most seismic stations were in the northern hemisphere and most instruments recorded at low gains (that is, the signals were not magnified). As a result, the long-distance, or teleseismic, detection threshold varied dramatically in different regions, and the accuracy for locations in remote oceanic areas was poor. Magnitudes could be estimated only for the largest earth- quakes. Limited knowledge of the geology of the mid-ocean ridge system also inhibited seismological research. This humble status began to change in the late 1950s. Seismology was invigorated by the emphasis on global geophysical observations during the 1957-58 International Geophysical Year (IGY) project. Also, deployment of new seismic stations to monitor tests of nuclear explo- sions resulted in improved detection and location capabilities for earth- quakes as well. In addition, the introduction of computers for data processing and earthquake location made global monitoring of smaller earthquakes possible. The importance of mid-ocean ridge seismology soared in the late 1960s, when it provided compelling evidence for the plate- tectonic hypothesis. Growing catalogs of accurately located earthquakes brought clear delineation of plate boundaries on seismicity maps, and first- motion studies confirmed predictions regarding the geometry of faulting 60 Ocennus at different types of plate boundaries. This success can largely be attrib- uted to the establishment in 1963 of the World-Wide Standardized Seismograph Network (WWSSN), a global network of about 100 seismic stations equipped with well-calibrated and standardized seismometers. The improved global distribution of these stations and their relatively high magnification of seismic signals significantly lowered the detection threshold in oceanic regions. A notable advance was establishment of a central archive from which complete seismograms from all the WWSSN stations could be obtained quickly. For the first time it became relatively easy for a researcher to collect all the data needed for a detailed source study of an earthquake almost anywhere in the world. Analytical techniques for digitized wave-form seismic data (both body and surface waves) were developed in the 1970s. Except that they had to be hand-digitized, the long-period WWSSN data were well-suited for studies on the moderate-sized earthquakes typical of mid-ocean ridges. At the same time, the first digital seismometers were deployed. These stations eliminated the unpleasant task of digitizing analog records for analysis, but the early instruments were best suited to earthquakes that were larger than is typical for mid-ocean ridges. Also, there were simply too few stations deployed to use them exclusively for source studies in most areas. This situation is changing rapidly: Broad- banded digital stations are being deployed by institutions and countries, with the goal of obtaining global coverage at least as good as what WWSSN provided. Seismologists will soon have the luxury of working exclusively with digital data for global earthquakes studies. With the proliferation of digital seismic stations, digital analysis techniques for earthquake source studies have become quite sophisti- cated. It is now standard practice to perform a formal inversion for source parameters such as depth, focal mechanism, and seismic moment (a measure of earthquake size that has many advantages over magni- tude). Systematic studies of mid-ocean ridge earthquakes have produced many insights concerning the tectonics of accreting plate boundaries. Application of these techniques is still limited to the largest mid-ocean ridge earthquakes, however; further progress depends on improvements in our ability to study smaller earthquakes. Establishment in 1964 of the International Seismological Centre (ISC) stimulated global seismology significantly. The ISC collects phase readings from thousands of seismic stations, associates them with particular earthquakes, locates the earthquakes, and publishes the data and results in a widely circulated bulletin. A whole new class of seismo- logical research is based on the availability of the entire ISC data set since 1964 in machine-readable format (magnetic tape and CD-ROM). Techniques are now being applied to mid-ocean ridge earthquakes that improve epicenter location for smaller earthquakes and help to correlate epicenters with geological features. These techniques, which simultaneously locate many earthquakes in a region, provide useful information about the location of mid-ocean ridge earthquakes with magnitudes as low as 4.5 on the Richter scale (a logarithmic scale of earthquake magnitude from -3 to 9 — the highest recroding to date is 8.9). In the 1970s we began to study smaller oceanic earthquakes using temporary (up to a few weeks) deployments of sonobuoy arrays and Systematic studies of mid- ocean ridge earthquakes have produced many insights concerning the tectonics of accreting plate boundaries. Winter 1991/92 61 Tlie maximum magnitude of earthquakes occurring in close association with spreading segments is strongly correlated with spreading rate. ocean-bottom seismometer (OBS) and hydrophone arrays. Epicenters and focal depths can be determined with great accuracy for microearth- quakes occurring within the array, but the relatively low rate of occur- rence of earthquakes on mid-ocean ridges and the technical difficulty of deploying (and recovering) enough instruments to permit useful seismo- logical analysis has doomed many such studies to disappointment. Progress in battery and data-storage technology, however, is extending deployment times. Technology that permits rapid deployment of seis- mometers on targets of opportunity will play a significant role in the future of mid-ocean ridge seismology. A Note on the Importance of Thermal Structure: Before discussing the characteristics of mid-ocean ridge earthquakes, it is worth emphasizing that the depth distribution of earthquakes is closely linked to the thermal structure of the lithosphere. Earthquake faulting is commonly considered to be analogous, if not identical, to the phenomenon of brittle failure, which is the usual mode of deformation of small samples of crustal and upper-mantle rocks at low temperatures. At sufficiently high tempera- tures, ductile deformation relieves an applied stress before the brittle- failure limit is reached, and no earthquakes occur. In any given tectonic environment, therefore, the depth distribution of seismicity provides information about the thermal structure. Conversely, prior knowledge of the thermal structure can be used to interpret the seismic data more fully. Earthquakes on Spreading Ridges Given the difficulty (so far) in observing an actual spreading episode on any deep-ocean ridge segment, earthquakes are perhaps the most dramatic indicators of the tectonic processes involved in the creation of new oceanic crust. A major issue in mid-ocean ridge seismology has been the extent to which earthquakes signal active magmatism at shallow crustal levels. The observations summarized below are beginning to reveal the answer to this question, but much work remains to be done. The maximum magnitude of earthquakes occurring in close associa- tion with spreading segments of the mid-ocean ridges, so-called "ridge- axis" events even though few probably occur in zero-age (true ridge axis) crust, is strongly correlated with spreading rate. At slow-spreading ridges with well-developed rift mountains and a median valley, such as the Mid-Atlantic Ridge (MAR), earthquakes reach a maximum magni- tude of about 6.0, while at the fast-spreading ridges of the East Pacific Rise (EPR), very few earthquakes reach even the teleseismic detection threshold, about 4.7, and the ridge segments appear to be aseismic on global seismicity maps. Microearthquake surveys show that these ridge crests are seismically active at lower magnitudes, however. Ridge-axis seismicity frequently occurs as swarms of earthquakes that usually last a day or two, although a few sequences lasting weeks are known. As much as 50 percent of the ridge-axis seismicity is esti- mated to be associated with this sort of cluster, so understanding the tectonic significance of this seismicity is an important research goal. Because swarm seismicity is frequently observed in terrestrial volcanic centers, ridge-axis swarms are sometimes thought to reveal ongoing 62 Oceanus 5 °CVJ s •§ s .S .5d ^ i 1^-2 t*>s\ ^ c\\ §°-^ ^ K a. •u- ^u **** «^t "I ^ ^ ~ ~ Ji •2: ^: s: x. ^. -- ~~* *-« ^» ^ -~- ^) •*-> ^^ -*— '-r*- cr ^ 5 5 ° 6 "w j?~' O . x. •5 '* = EX. =: o £ -2 ^ >P '£ uj v> i^~ ^ J~^ ^ g O ^ ^ « ° K ^.v ^ e ~ > a t* rfi S WJ r- ~ J= ij s; t: .« **« p*. .2. 2 - § I 52 % 2 K w ^j J^ ID eg in (N m CM in ~ C S: 'ft 5: -C 5 ^ ~ s ^ _ rft -G •*• • «3 ? I1 *^v C* ;^ TO .— <-> g "Q i ;§l§-! -t*. ^"x *o »^! H *^- -i< i^r ^ 2 g ~ Q = 0| ^j O Vj • ~ S CD W CD ^ CO ±= 03 5 yj **> t^-* o ,2 5 'b 'i 8 '« ' ^ ^ JS s C> % ^ cs N >• o ^J ^ -It c f" t 0 "^ Si p _ 'X O T3 O O CD C C CD O O ±± .^ .^ 03 •5« •H f 1 I ^J w> O S V) VI VI x. ^ o O ^ K 2: r- -=: o O 0 ^j- CD 00 CD CD CD CO CO CO (sjd}duj) wi ~ O **• C .20 mixing of hot hydrothermal fluids and ambient seawater (before the temperature of this mixture falls below 125°C) subsequently dissolves within a matter of days in cold seawater. Third, and most exciting, the plume was the residue of a brief but very massive discharge event, quite unlike the familiar steady flow from small chimneys. The megaplume's history was gleaned from detailed mapping that revealed an almost perfect three- dimensional symmetry. In geo- metrical jargon, the megaplume formed an oblate spheroid — a Frisbee — with a diameter of more than 20 kilometers and a thickness of 700 meters at its center. "Nor- mal" plumes issuing steadily from a collection of chimneys in a vent field are imperfectly mixed and habitually asymmetric, growing erratically more dilute as deep currents sweep them away from their sources, like wood smoke from a lazy campfire. The megaplume's precise symmetry implied a bomblike event, lasting only several hours to, at most, a few days. Searching for a Motive: Two Possibilities With these clues in hand, the researchers working on the case next sought to construct a picture of the physical processes that created the megaplume: the suspect's modus operandi. Because the megaplume had known, symmetrical boundaries, we could confidently calculate its burden of hydrothermal heat and chemicals. Even though its average temperature increase above the surrounding seawater was only about 0.1 °C, its volume of over 130 cubic kilometers contained about 2 x 1016 calories of hydrothermal heat, or enough energy to electrify New York City for almost a year. By knowing the hydrothermal heat content, it was possible to calculate the original volume of hydrothermal fluids: 100 million cubic meters at a temperature of 350° C. The megaplume event unleashed a staggering volume of hydrother- mal fluid. By comparison, it would take 100 to 200 years for 100 million cubic meters to escape from a single familiar "black smoker" chimney. New ideas about fluid discharge from the seafloor were thus needed. A geologically attractive alternative proposed that the megaplume fluids erupted not from a forest of standard chimneys, nor from one yawning megachimney, but rather from a long but narrow fissure cleaving the vent field. Fissuring is common along the crest of the MOR, and pictures returned by a deep-towed camera sled revealed a prominent band of fissures cutting through the megaplume area. 86 Oceanus 45 N 50' 40'- 44 30' / i^//r \ l Megapiume fj 130 30' The fissures imaged by the camera sled were, of course, only the exit ways of the kilometers-deep and tortuous paths through which hydrothermal fluids must slowly percolate to reach the seafloor. The normal concentration of pores and microcracks deep in the MOR's vol- canic rocks is far too low to permit the fluid flow rate required by the megaplume. Ideas about what triggered the megaplume, therefore, had to have a common thread: the catastrophic rupturing of the crust beneath an ordinary vent field. To speculate how crustal permeability can catastrophically increase, consider the megaplume fluids from two points of view: active and passive. An active model views hydrothermal fluids, circulating deep in the crust, as soup simmering in a pressure cook- er afflicted with a stubborn safety valve. Crustal fluids fracture the imprisoning vol- canic rocks when fluid pressure increases faster than normal venting can release it. Pressure can be raised past the critical point by a pulse of magmatic heat that raises the temperature of the fluids, or by a sudden release of magmatic gas. The alternate view holds that a megaplume is the passive result of an impatient chef cracking open the pressure cooker while the soup is still boiling. An earthquake on the MOR fractures the crust, tem- porarily opening deep and spacious fluid pathways, re- leasing the trapped fluids as a gushing megaplume. This view evolves from a long his- tory of observations at terres- trial volcanic sites. Most inter- esting are those from Iceland, a segment of MOR obligingly lifted above the concealing ocean. Records kept almost since the time of the Vikings show that every 100 to 150 years a years-long episode of crustal rifting and volcanic outpouring assaults Iceland. Careful mea- surements collected during the 1975 to 1982 episode indicate that inter- mittent crustal stretching was accompanied by subcrustal movements of magma and subaerial eruptions of molten lava. Expanding the Investigation Several pieces of the puzzle were now in place, enough for tantalizing speculations, but too few to convince a sober jury of scientific peers. To find the crucial pieces needed for an airtight case, several investigators decided on a long-term stakeout of the megaplume area. After several years they hit pay dirt, in the form of two first-of-their-kind discoveries on the MOR. Vance / i y Segmenl / ', '•••! Cleft Segment' 134C 130 126 122- 130 OO'W DSVAIvm/Robert Embley. NOAA A plan view of the comparative extent of the 1986 megaplume mid the underlying "normal" plume along the axis of Cleft Segment is shown above. The size, symmetry, and hydrothermal tempera- ture anomaly of the megaplume are all in sharp contrast to the steadily emitted normal plume. At left is the sea floor fissure the megaplume may have erupted from. Taken from DSV Alvin (whose instruments are in tlie foreground), the fissure is several meters wide and perhaps 10 meters deep. Winter 1991/92 87 Within an hour of the beginning of two Kmfla eruptions, a photographer in an airplane captured the image at right and the one on page 2. The eruptions began on October 18 and November 18, 1990, and lasted five days each. Both were similar, occurring on an 8-kilometer-long fissure extending from the center of the Krafla caldera, northwrd along the rift zone, and were part of a series of rifting and magmatic events at the divergent plate boundary in nortlieast Iceland that lasted from 1975 to 1984. Sudden decreases in the elevation of the Krafla caldera often corresponded to instances ofcrustal widening and volcanic eruptions (below). Volcanologists believe magma flows out of the caldera of the Krafla volcano, filling and sometimes overflowing the newly widened rift zone as the crustal plates separate. (After Bjornsson, 1985, and Tryggvason, 1984.) The first solid evidence that the megaplume was a signal flare for magmatic activity appeared in annual analyses of water samples from the "normal" plume that always blankets the megaplume site. In August of 1986, immediately after the megaplume release, the ratio of helium-3 to hydrothermal heat in the normal plume exceeded by several times any value previously measured anywhere on the global MOR. The ratio then decreased every year until 1988, when it reached a level typical of other MOR vent sites, and of the megaplume itself. Such variability was no more antici- pated than the megaplume itself. Upwelling magma from the mantle is the sole source of both helium-3 and heat, and work elsewhere on the MOR had found no evidence of a changing ratio. These new observations had an exciting explanation: a sudden change in a pocket of fluid magma had injected a surge of helium-3 into the hydrothermal fluids feeding the vent field. Laboratory experiments have demonstrated that magma can be rapidly stripped of helium-3 and other volatile gases both by bubble formation as rising magma depressurizes and by crystal- lization as magma cools. The second discovery provided definitive evidence of the magmatic activity needed to produce a burst of helium-3. Magma can suddenly depressurize and cool when it intrudes into deep cracks opened in the upper crust by the retreating plates. Occasionally these dikes of intrud- ing magma overflow their cracks, spreading in shimmering lakes or piling up in blocky mounds of fresh basalt on the seafloor. We know this happens: The evidence is the MOR itself. But not until 1989 at the megaplume site had oceanographers been able to identify and sample a specific lava flow newly emplaced upon the seafloor. Careful bathymet- ric mapping in 1987 and 1989, repeating survey lines originally run in Cumulative Rift Widening (m) 88 Ocean us 1981 and 1983, revealed a series of new lava mounds stretching for 16 kilometers along the trail of fissures that runs through the center of the megaplume area. The volume of the mounds is roughly 0.5 cubic kilome- ters, somewhat less than half the volume extruded during the Icelandic eruption episode of 1975 to 1982. Revealing the Crime — and the Collaborators The discovery that the cataclysmic megaplume event of 1986 was closely associated with both an escape of magmatic gas and an eruption of fresh lava supplied enough pieces of evidence to make the case decipherable. A closing argument might sound like this: The separating tectonic plates on either side of the Juan de Fuca Ridge had been raising the tension along the axis of the Cleft segment for some time, perhaps decades. In mid-August of 1986, its crust failed and the width of the axial crest increased by a few meters. A huge mass of hydrothermal broth impris- oned in pores and crevices was released almost instantly. A several- kilometers-long line of hydrothermal heat and chemicals surged 1,000 meters above the seafloor, turbulently mixing seawater over a 300- square-kilometer area. The megaplume discharge ended abruptly as cold seawater filled the emptying fissures. On the heels of this release, a slab of molten magma pushed upward through the cracked crust, lumping up mounds of fresh lava wherever it breached the seafloor. Helium-3 and other volatile gases dissolved in the magma were liberated as the magma rose and cooled. Newly forming hydrothermal fluids absorbed these gases, and soon afterwards the flow from seafloor chimneys contained extraordinary concentrations of helium-3. The magmatic dike reaching up from the base of the crust was only a few meters wide and stretched perhaps 20 kilometers along the fissure line running through the north end of the Cleft segment. By 1988 it had completely solidified, and extraction rates of helium-3 and heat were once again similar. New communities of vent creatures colonized the lava mounds, attracted by warm, chemical-rich water leaking up through the lava from the buried fissure system. A new sliver of ocean crust had been added to the seafloor. This hypothesis of the events surrounding the megaplume release satisfied our puzzlement, while stimulating our curiosity with new questions. How common are megaplume events, and how important are they in the global budget of hydrothermal venting? Does volcano- tectonic activity on segments of the MOR follow the Icelandic pattern of several years of concentrated vigor separated by long quiescent periods? And, most basically, how can we test our hypothesis? The importance of episodic events relative to the more familiar steady venting can only be surmised until we know more about the frequency of megaplumes. Megaplume observations are still under- standably scarce. A second megaplume, somewhat smaller than the first, was found over the Juan de Fuca Ridge in 1987. The precise origin of that plume is unknown; when found, the plume was several weeks to months old and may have drifted far from its source. Researchers have also claimed evidence of a megaplume in the Fiji Basin just southwest of Samoa. More extensive evidence may actually exist in the geologic record, where the prevalence of hydrothermal breccias, deposits of A huge mass of hydrothermal broth imprisoned in pores and crevices was released almost instantly. Winter 1991/92 89 Cartoon of the hydro- thermal and geological observation program of the Vents Program at the megaplume site. The geometry of the magma chamber and magma dike are conjectural. fractured and fragmented debris, suggests that cataclysmic releases of hydrothermal fluids have not been uncommon along the MOR. A few simple calculations can yield a rough estimate of the relative importance of large hydrothermal events. In terms of the power released by hydrothermal discharge on various scales, the megaplume is an awesome force. During its short lifetime, it releases heat at a rate equal to perhaps 10 percent of steady venting along the entire global MOR axis. In terms of global budgets, however, we should more appropriately compare the supply of hydrothermal emissions over a longer time period. On an annual basis, the contribution of hydrothermal elements such as iron and silicon from a single megaplume is about equal to that of an entire vent field, or about 0.1 percent to 0.01 percent of the global total. As a first estimate, then, the global supply of hydrothermal fluid from megaplumes would equal that from steady venting if every vent field produced a megaplume each year. What little experience we have suggests that this schedule is too ambitious, so megaplume events are unlikely to dominate the global budget. They may, however, be a principal contributor to the hydrothermal supply from MOR segments undergoing spreading episodes. Megaplume Plume Monitoring Steady Venting Acoustic Extensometer Volcanic System Monitors Magma Dike Magma Chamber 90 Oceanus Vents to Remain Vigilant The progression from simple detection of megaplumes to a test of the complex hypothesis in which we have embedded them requires the establishment of a long-term observational program. The NOAA Vents Program is currently developing such a program at the site of the original megaplume. The seafloor observational system will initially include three components. Sensors on moorings along the ridge axis will monitor temperature and current velocity above the vent field, looking for perturbations that signal a sudden change in the distribution or intensity of steady venting, or the occurrence of hydrothermal events such as megaplumes. Volcanic system monitors resting on the seafloor will contain seismic recorders tuned to detect tremors indicative of magma rumbling through the crust, and sensitive inclinometers and pressure gauges to measure seafloor deformation caused by swelling or contraction of the magma chamber. Pilot deployments of these first two components began in 1991. The third component, an acoustic extensom- eter network, will be added in 1992. Acoustical beacons on either side of the rift zone will search for evidence of crustal spreading events by continuously monitoring the distance between themselves, much as a land-based laser-ranging system currently monitors movement along the San Andreas Fault system in southern California. In addition to these seafloor monitors, the Vents Program will tap into an existing array of military hydrophones throughout the Pacific to listen for the telltale underwater sounds, called T-phases, made by the cracking of earth- quakes and the rumbling of magma. In the world of pulp detective novels, the criminal always returns to the scene of the crime. Nature, however, may not so obligingly furnish another megaplume to the scientific Sam Spades now on a stakeout at the Cleft segment. But even without one, a long-term study of its birth- place will increase our understanding of the interrelationships of crustal rifting, magma movement, and hydrothermal activity, providing new puzzles and clues for the next detectives. The NOAA Vents Program is currently developing a long-term observational program at the site of the original megaplume. Edward T. Baker is a Research Oceanographer at the National Oceanic and Atmospheric Administration's Pacific Marine Environmental Laboratory, Seattle, and an Affiliate Associate Professor in the School of Oceanography, University of Washington. He presently serves on the RIDGE Program Steering Committee and has been mapping hydrothermal plumes since 1984. Winter 1991 /92 91 Seismic tomography will help us understand plate- boundary mechanics. Tomographic Imaging of Spreading Centers Douglas R. Toomey ince we cannot observe the vast regions beneath the seafloor directly, we must use remote-sensing methods such as sound waves to help us answer fundamental questions about the structures that make up mid-ocean ridges. We want to know the size, shape, and location of spreading-center magma- storage zones and the physical properties of the solid and molten rocks beneath the axis of accretion. We need to know the strength of ridge- building materials and the spatial and temporal relationships of various ridge components. Seismic tomography, a powerful method for mapping three dimensional physical properties within Earth's interior, promises to help us gain this knowledge, which in turn will help us understand plate-boundary mechanics. By analyzing the propagation of vibrational waves generated by earthquakes or man-made sources such as explo- sions, we can map seismic structure, that is, the travel speeds of different types of vibrational or seismic waves and their attenuation with distance due to frictional energy loss. One of the most frequently studied seismic waves, the P-wave, is similar to the acoustic waves we hear. A shock wave generated by a large explosion is a graphic analogue of a P-wave. By measuring the transit time of P-waves passing around and through an active spreading center, we can develop a three-dimensional image of the seismic structure beneath the plate boundary. Seismic tomography can be used to image three-dimensional physi- cal properties over a wide spectrum of length scales, including, for example, images at a scale of hundreds of meters of the upper-oceanic crust near regions of hydrothermal venting, reconstruction of kilometer- sized velocity anomalies characterizing axial magma chambers, and maps of physical properties within the zone of mantle upwelling and melt generation, located deep (10 to 100 kilometers) beneath seafloor- spreading centers. Several fundamental improvements in the descriptive and theoretical models of oceanic ridges await detailed, three-dimen- sional mapping of velocity structure. 92 Oceanus The Importance of Seismic Imaging Within a spreading center, the geology and morphology of the sea floor and the thermal and mechanical structure of the newly formed oceanic plate are controlled by the complex interplay of magmatic injection, tectonic rifting, and hydrothermal cooling. These dynamic processes all exhibit pronounced spatial and temporal dependencies. Moreover, while an individual process may express itself on the seafloor as a volcano, an uplifted mountain range, or a hydrothermal vent field, the majority of the dynamic activity invariably occurs at some depth beneath the seafloor. Understanding the nature of oceanic spreading centers requires knowledge of the behavior of these dynamic processes as they evolve directly beneath the axis of accretion. In recent years, working models of oceanic spreading centers have evolved from two-dimensional, steady-state idealizations to more realistic three-dimensional, time-dependent systems (see the segmented Mid- Atlantic Ridge, page 11). The new dimension added to the working models is the pervasive along-strike variability of mid-ocean ridge processes, notably in the production of melt beneath the spreading center. Current hypotheses suggest that ascending melt within the mantle is focused into magmatic centers separated on the order of tens to a hundred kilometers. Each magmatic center supplies the greater portion of melt and heat to a single ridge segment. Within an individual ridge segment, processes such as faulting, hydrothermal circulation, and magmatic accretion vary systematically as a function of distance from the magmatic center. The hypothetical structural unit, consisting of a local maximum of magmatism bounded by along-axis minima, became known as a spreading-center segment or cell. This simple model of cellular segmentation provides an improved, but controversial, working hypothesis for mid-ocean ridge studies. What follows is a review of results of seismic tomography studies used to investigate the spatial variability of physical properties and processes deep within spreading centers. Divergent plate boundaries display several different structural forms, including the classic rift valley of the Mid-Atlantic Ridge, the pronounced en echelon or steplike structure of the Reykjanes Peninsula within southwest Iceland, or the more morphologically subdued East Pacific Rise. Each of these different spreading centers is the topic of a tomographic study. These investiga- tions are unified by a common purpose: furthering our knowledge of physical structure beneath ridge axes, and using these observational constraints to improve working hypotheses of the mechanics of diver- gent plate boundaries. Tomographic Images of Rifts and Rises Mid-Atlantic Ridge The cellular model of focused magmatic accretion is particularly apt for characterizing the rugged, slow-spreading Mid-Atlantic Ridge. Undula- tions in the ridge-parallel profile of axial seafloor depth are thought to result from variations in melt production. Magmatic centers presumably coincide with the shallower portions of the ridge axis, while the far ends of magmatic cells (regions of low melt production) are thought to The majority of the dynamic activity invariably occurs at some depth beneath the seafloor. Winter 1991 192 93 Tliis cross-section of the Mid-Atlantic Ridge median valley along-axis deep near 23°N, shows microearthquake hypocenters (circles) and contours of P-wave velocity (in kilometers per second) obtained from two-dimensional tomographic imaging. The seafloor bathymetry reveals the rift's relatively flat inner floor and rugged mountains to the east. A conjectural fault plane for recent large earthquakes is shown (see Mid-Ocean Ridge Seismicity, page 60). At the far end of th is 40-kilometer-long ridge segment, tomogra- phy data shows nearly normal oceanic crustal structure. correlate with the deeper parts of the ridge profile. Two recent studies conducted on the Mid-Atlantic Ridge by researchers from the Massachu- setts Institute of Technology (MIT) and the Woods Hole Oceanographic Institution (WHOI), one within an axial low near 23°N and the other astride an axial high near 26°N, demonstrate the utility of tomographic imaging for characterizing crustal seismic structure throughout spread- ing-center cells. The first application of tomographic methods to the study of mid- ocean ridge crustal structure occurred during an investigation of the seismicity and seismic structure of the Mid-Atlantic Ridge near 23°N, south of the Kane Fracture zone. The micro- earthquake study was located in an along-axis deep at the far end of a ridge segment approximately 40 kilometers long. During a two- week deployment within the rift valley, hundreds of microearth- quakes were recorded by ocean- bottom receivers. In addition to locating the earthquakes, a tomographic analysis of travel- time data was conducted; the P-wave data comprised transit times from earthquakes and several man-made explosions (seismic refraction data) to the ocean-bottom receivers. By analyzing variations in the transit times among many different paths, images of anomalous volumes of seismic velocity were obtained. The two-dimensional seismic structure across the rift-valley inner floor and transecting the axial deep was similar to normal off-axis oceanic crustal structure, excepting a small decrease in mid-crustal (1 to 4 kilometers beneath the seafloor) velocities at zero-age crust. These low velocities quickly evolved with age (or with off-axis distance) within the first few hundred thousand years of crustal formation. The similarity in structure between axial crust within an along-axis deep and normal off-axis oceanic crust that had undergone extensive cooling as a result of aging was remarkable. From these and other observations we hypothesized that at this far end of a ridge seg- ment, considerable time (about 10,000 years) had elapsed since an episode of significant magmatic accretion. Without the addition of new crustal material that results from magmatic injection, we also inferred that this rift-valley deep had undergone horizontal extension; in effect, the spreading of oceanic plates at the end of a ridge segment was accommo- dated by stretching and thinning of the axial crust. A second study near the Transatlantic Geophysical Profile (TAG) hydrothermal field at 26°N was located close to an along-axis high near a ridge-segment center also approximately 40 kilometers in length. In contrast to the magmatically quiescent, cool crust beneath the along-axis deep at 23°N, the axial high of the ridge segment at 26°N is characterized by high-temperature, black smoker chimneys. At 26°N the two-dimen- sional seismic structure along the rift valley, including the region of the axial high, was remarkably heterogeneous in comparison with the 94 Oceanus structure near 23°N. Anomalously high velocities in the upper crust were detected near the axial high, and a low velocity anomaly was detected beneath an axial volcano. Both of these crustal velocity anomalies were interpreted to be the result of recent magmatic intrusion. In contrast, toward the axial deep at the southern end of the TAG ridge segment, the seismic structure is comparable to normal oceanic crustal structure, similar to the axial deep near 23°N. The axis-parallel velocity patterns near 26°N, including complex structures near the ridge segment center and a transition to more homogeneous, almost normal crustal structure near the segment's far end, appear consistent with the hypothesis that crustal accretion is focused centrally beneath a slow-spreading ridge segment. We don't yet know whether or not the observed variations in axial seismic structure are fundamentally related to axial segmentation and focusing of magmatic accretion. Current models of mid-ocean ridge processes suggest that slow-spreading Mid-Atlantic Ridge segments of equal length, such as the 23°N and 26°N ridge segments, are magmatically fed by either upwelling plumes of similar size or volu- metrically similar amounts of melt. If the model predictions hold true, the tomographic images resulting from these two microearthquake experiments may begin to characterize the seismic structure near the axial high and axial low of a 40-kilometer-long ridge segment. Reykjanes Peninsula, Iceland A tomographic study of the Hengill-Grensdalur volcanic field in south- western Iceland provides further indication of the power of seismic methods for imaging the interior of active spreading centers. Working with Gillian Foulger of the UK, we resolved the anomalous three- dimensional crustal structure underlying the magmatic center of a slow- spreading Icelandic rift segment. Geologic maps of the area show that the magmatic center or central volcanic region incorporates the recently active Hengill Volcano, the inactive Grensdalur Volcano, and the high- temperature geothermal field associated with these features. To either side of the Hengill central volcano, and extending to the far ends of the rift segment, are a set of fissure swarms indicating the locus of past eruptions. Our scientific objective was to tomographically image the seismic structure of the center of this magmatic cell. These results also aided the Icelandic Energy Authority in their search for volumes of anomalously hot rock beneath the spreading center and to evaluate this volcano's geothermal energy potential. A significant advantage of land-based surveys is the ease of record- ing seismic data for a longer period of time than is typically possible for marine seismic experiments. A longer recording period provides a larger data set, and thus more extensive sampling of the study volume; as expected, higher resolution tomographic images are obtained when larger quantities of data are available. Using P-wave travel times re- corded during a four-month period by over 20 seismometers, we tomographically imaged seismic velocities within a 14-by-15-by-6-cubic- kilometer volume that underlies the high-temperature Hengill- Grensdalur geothermal field. A dense distribution of sources and receivers permits structural resolution to within approximately 1 and 2 kilometers in the vertical and horizontal directions, respectively. The A dense distribution of sources and receivers permits structural resolution to within approximately 1 and 2 kilometers. Winter 1991/92 95 Three-dimensional tomographic image of P-wave velocity beneath the Hengill- Grensdalur volcanic complex, Iceland. The color scale denotes percentage difference in velocity from the regional structure. For display purposes, the model is represented by constant-velocity cubic blocks of dimension 0.25 km. Two views of the tomographic image are shown; both views are from the northeast. Positions of the surface expressions of the Grensdalur and Hen- gill volcanoes (red circles) and the axis of crustal accretion (solid bar) are shown . final model of the area's structure is charac- terized by distinct bodies of anomalously high velocities: Two of these bodies are continuous from the surface to about 3 kilometers depth, and each is associated with a site of past volcanic eruption; the third body of high velocity lies beneath the center of the active geothermal field at a 3- to 4-kilometer depth. The volcanic features we directly ob- serve on the surface are clearly the expres- sion of igneous processes occurring at great depths. They include the crustal-level stor- age of molten magma and the cooling of such bodies to form magmatic intrusions or plutons. For crustal-level rocks, the P-wave velocity varies little at temperatures below 500°C and decreases rapidly at tempera- tures in excess of 500° to 800°C (basalt be- gins to melt at about 800°C). We thus infer that neither molten magma nor rock hotter than about 500°C exist presently in large volumes beneath the Hengill-Grensdalur volcanic and geothermal field. The presence of hot springs and fumaroles at the surface with water temperatures between 300°and 370°C, however, indicates the presence of intrusive rock at similar temperatures. We interpret the tomographic images of anomalously high velocity to be the result of recently solidified magmatic intrusions into the upper crust. Furthermore, those intrusions beneath active geothermal fields, while solid, are most likely hot enough (about 400°C) to provide a usable source of thermal energy. The tomographic image of the Hengill-Grensdalur volcanic field may provide an analog to the type of three-dimensional seismic structures possibly present beneath the TAG area of the Mid-Atlantic Ridge. Both sites are coincident with the center of a ridge segment, and both are characterized by profound structural heterogeneity suggestive of recent crustal-level intrusion of magma. East Pacific Rise Long-standing fundamental questions surround models of the size, shape, and physical properties of mid-ocean-ridge axial magma cham- bers. Mid-ocean ridge magmatism is a significant function of spreading rate, and at faster spreading rises, such as the East Pacific Rise, the volume of melt and the amount of heat delivered to the crust greatly exceeds that of the slower-spreading Mid-Atlantic Ridge. Consequently, thermal models for mid-ocean ridges predict that shallow crustal tem- peratures are generally higher, and axial magma chambers are generally larger and more long-lived along faster spreading rises. To test these models, a seismic tomography experiment was recently conducted on the East Pacific Rise (EPR) near 9°30'N by MIT and WHOI. It employed 96 Oceanus A mn}] and his charge: Beecher Wooding of WHO/ prepares to launch tin explosive charge (in the cardboard box) during the 1988 East Pacific Rise seismic tomography experiment aboard R/V Washington. 2 km South a, Q) -8 -6 9°28W deval -4 -202 West-East, km 1 km West 15 ocean-bottom receivers and over 450 shots to image for the first time the three-dimensional seismic structure of an axial magma chamber. The 15 receivers included ocean-bottom hydrophones and seismometers, designed and built by engineers and technicians at WHOI and MIT. Unlike the so-called passive tomog- raphy studies of the Mid- Atlantic and Icelandic rifts that used P-waves generated by local earthquakes, the EPR experiment was an active seismic-imaging experiment that used P-wave energy generated by explosives. We deployed the individual explosive shots in a dense grid to ensure good sampling of the crustal volume beneath an 18-by-16- square kilometer area centered on the EPR axis. Over 7,000 seismo- grams were recorded, each provid- ing some measure of the crustal seismic structure along a different path connecting a source to a receiver. A vertical section (at right) of the EPR tomographic reconstruction shows the anomalous P-wave structure across the rise axis and cutting through the axial magmatic system, which appears primarily as the anomalously low seismic velocities (orange and red areas) about 2 to 4 kilometers beneath the seafloor. From laboratory studies of 8 deval -8 -202 South— North, km Tliese vertical cross sections through the P-wave velocity structure were obtained by tomographic imaging of the East Pacific Rise. The top and bottom sections are transverse and parallel to the rise summit, respectively. The colors show departures of the three- dimensional model from an average one-dimensional, depth- dependent velocity structure: blues are faster than average and greens to reds are slower tlian average. Tlie contour interval is 0.2 kilometers per second. Two deviations in the along-axis trend of the rise summit (devals) are shown on the rise parallel section. Both images pass through the axial mag)natic system. Winter 1991/92 97 Juxtaposing the EPR seismic tomography results and the seafloor bathymetry permits a perspective view. Two map-view sections through the three- dimensional model are shown, one near the seafloor and the deeper one at a depth of 2 kilometers beneath the rise summit; the deeper section passes through the lowest seismic velocities of the axial magrnatic system. The colors show variations in seismic velocity structure; blues are faster than average, greens to reds are slower than average. The three-dimensional mesh depicts undula- tions of seafloor bathymetry; the axial summit is depicted by shallowing of the seafloor. The location of two devals are noted, as is the vertical projec- tion of these seafloor features down to the depth of the axial low- velocity volume. P-wave velocity with increasing temperature we infer that the subaxial crustal region comprising low seismic velocities is extremely hot, with temperatures well over 500°C We think the concentration of lower seismic velocities near a 2-kilometer depth results from the accumulation and storage of molten magma within a thin melt-filled sill; this magma lens is frequently observed by other types of seismic experiments and its maxi- mum cross-axis width and thickness are inferred to be 1 to 2 kilometers and less than a few hundred meters, respectively. The estimated volume of melt 9°35'N deval 9°28'N de 98 stored within this sill is comparable to that of a typical seafloor lava flow, suggesting that a volcanic eruption along a fast-spreading ridge draws melt from this region. The tomographic images also show a large region of low seismic velocities that presumably envelope the much smaller magma lens near a 2-kilometer depth. In general, the seismic velocities throughout the larger volume encompassing the melt lens are consistent with elevated temperatures, but not necessarily with molten rock. The size and shape of the seismic anomalies across the rise axis strongly constrain the size and shape of the axial magrnatic system. A vertical section parallel to the EPR axis shows a variation in seismic velocity suggestive of an along-axis segmentation of the crustal magma chamber and the axial thermal structure. Again, the elevated temperatures associated with crustal-level magmatism are effectively mapped as the regions of lower seismic velocity. Along this section of the EPR, the observed seismic structure is noticeably segmented on a scale of about 10 kilometers, with the lowest velocities observed immediately south of the experiment center; from this we infer that thermal structure is segmented in a similar manner. The along-axis segmentation of the axial magrnatic system gives rise to an observed segmentation of seafloor morphology. During the experi- Oceanus ment, we mapped seafloor bathymetry over a 3,600-square-kilometer area. Inspecting these maps we found that along axis the trend of the EPR axial summit was variable; within the aperture of our seismic experiment, the rise summit was easily divided into three adjacent linear segments. Our seismic tomography images included one complete 12- kilometer-long linear segment and parts of the bordering rise sections. At either end of this linear segment, the axial summit deviates from linear- ity, a seafloor morphologic feature referred to as a deval. The along-axis tomographic section shows that the axial devals coincide with a relative increase in along-axis seismic velocities near a depth of 2 kilometers. The interpretation is that at mid-crustal depths the temperature is highest in the center of the morphologically defined linear-rise segment, and lowest at the segment ends. The correlation of seafloor bathymetry with subseafloor thermal structure shows that magmatic processes occurring at great depths strongly affect surface geology. A perspective plot (opposite page) shows a different view of the seismic tomography results including anomalous seismic velocities near the seafloor and 2 kilometers beneath the seafloor; variations in seismic velocity are indicated with color. As in the other figures, anomalously low and high seismic velocities are indicated by warmer and cooler colors, respectively. Seafloor bathymetry undulations are represented by a three-dimensional mesh, clearly showing the shallowing of the seafloor that demarcates the axis of seafloor spreading. Seismic velocities near the axial summit seafloor are notably high (shown as blue colors). Two bathymetrically defined devals are shown as vertical lines penetrating the seafloor and continuing downward to the horizontal section at a 2- kilometer depth. The deeper section lies near the depth of the melt-filled sill and through the core of the axial magmatic system. Our interpreta- tion is that melt generated in the mantle tens of kilometers beneath the seafloor is injected into the shallow crust at intervals of about 10 kilome- ters along the rise axis, giving rise to magmatically defined rise segments of similar length. The segmentation of crustal-level axial magmatism and its relationship to segmentation of seafloor morphology is an important new observation made possible by seismic imaging. Tomographic studies of seismic velocity structure beneath local segments of the East Pacific Rise, the Mid-Atlantic Ridge, and the Icelandic rift represent a new and powerful approach to the seismologi- cal study of divergent plate boundaries. Future seismic tomography experiments will continue to provide images of the physical properties deep within spreading centers, and the study of these images, in con- junction with other geological and geophysical data, will greatly im- prove models of the tectonic, magmatic, and hydrothermal processes responsible for the formation of oceanic regions. "*\ Acknowledgements: Much of the research reported here was done in collaboration with G. Michael Purdy (Woods Hole Oceanographic Institution) and Sean C. Solomon (Massachusetts Institute of Technology). The results from the TAG area of the Mid-Atlantic Ridge are from the Ph.D. thesis of Laura Kong (MIT/WHOI Joint Program in Oceanography /Oceanographic Engineering). Douglas Ft. Toomey is an Assistant Professor in the Department of Geological Sciences at the University of Oregon, and a graduate of the MIT/WHOI Joint Program in Oceanography/Oceanopgraphic Engineering. Correlation of seafloor bathymetry with subseafloor thermal structure shows that magmatic processes occurring at great depths strongly affect surface geology. Winter 1991/92 99 Bruce Heezen, aboard Vema, /// the 1970s. Bruce C. Heezen A Profile Paul J. Fox nice Heezen died prematurely at the age of 54 in June of 1977, as he was preparing to dive aboard the Navy research subma- rine NR-1 . His intended destination was the Mid-Atlantic- Ridge axis, a limb of the world-encircling ridge system where oceanic crust is created. Bruce had been fascinated with the Mid -Atlantic Ridge since he first studied and explored it 30 years before, as an undergraduate research assistant for Maurice Ewing at Woods Hole Oceanographic Institution (WHOI). Bruce's passing was untimely, and marine geology and geophysics lost one of the great visionary minds of the science, but the way he died was in a sense heroic: He was at sea, poised to enter the abyss in his never-ceasing quest to better understand how Earth works. If asked, I cannot imagine that he would have scripted his death any differently. It is a daunting task to adequately profile the depth of character of a man who contributed more than 300 publications and two books to marine geological literature, who was the mentor and colleague of 13 100 Oceanus Ph.D. students, who spent more than eight years at sea pursuing knowl- edge about the seafloor, and who, through these achievements, changed the way we think about the processes that create and modify the sea- floor. I will try, however, to focus on a few highlights of his early years, his science, the particular gifts that allowed him to see further than most, and the generosity and wisdom that made him such a memorable teacher. As the son of a successful turkey farmer in Iowa, Bruce Heezen spent a great deal of his childhood outdoors, rambling about the countryside attending to agrarian chores and developing a keen interest in the natural sciences, a focus en- couraged by one of his grandfathers. He entered the University of Iowa as World War II drew to a close, with a desire to study science and an intention to never have anything to do with turkeys, whether it be their care or their consump- tion. I cannot help but believe that Bruce's intense dislike for these creatures contrib- uted to his desire to distance himself from the Iowa turkey pens and seek the sea's far shores and mysteries. His path to the sea, however, was indirect, and conditioned by serendipi- tous twists. His initial focus while an undergraduate geology major was paleontology, the study of fossil plants and animals. An outstanding under- graduate, he was selected to spend a summer in the western US helping a graduate student collect specimens for his Ph.D. thesis. Bruce later said that at this stage he had no doubt he would go on to graduate school seeking a Ph.D. in paleontology and spending his summers out west in search of fossils. By the 1940s the study of the fossil record was a mature science with a 100-year record of scholarship; advances came slowly and only after a great deal of careful work. This characteristic was clear to Bruce during the winter term of 1947, as he labored on a study of toothlike fossil elements called conodonts, sampled from 300-million-year-old rocks around Iowa. When Maurice Ewing, a geophysicist from Columbia University, arrived on campus as a visiting lecturer and spoke about the vast terra incognita that lay beneath the obscuring blanket of the world's oceans and the exciting science to be done at sea, Bruce was intrigued at the seemingly great opportunities for discovery in this young science. It was no coincidence that Ewing emphasized the romantic qualities of oceanographic research: He was looking for undergraduates to join him for a summer of work. During a tour of the geology department after the lecture, Bruce was introduced to Ewing over a tray of fossils. Out of this came a seductive invitation to join a National Geographic-sponsored Aboard Atlantis in the North Atlantic, Bruce Heezen arms a surplus World War II bomb for a seismic refraction experiment. This photo was taken in the late 1940s or early 1950s. Winter 1991/92 101 Again on Atlantis in the North Atlantic, Bruce Heezen (right, facing) with Maurice Ewing (left, facing) arms an explosive charge. cruise aboard WHOFs R/V Atlantis to explore a long linear swell, the Mid-Atlantic Ridge, that lay along the North Atlantic's center line. The opportunity to participate in an investigation of the first-order properties of an unknown mountain range beneath the sea offered a refreshing change in perspective and scale from the microscopic study of subtle changes in conodonts. Bruce accepted the invitation and arrived in Woods Hole in June to join Ewing's team preparing for the cruise. For the first several weeks, they worked feverishly to fabricate equipment for the voyage. Bruce constructed a photographic laboratory on Atlantis, helped to build several deep-sea cam- eras, and searched Harvard libraries for literature about the Mid-Atlantic Ridge. At that time, almost nothing was known about the Mid- Atlantic Ridge excepted that it existed. It came as a shock and a surprise to Bruce when Ewing told him, during a walk home late one evening in Woods Hole, that he was not to go on Atlantis to the Mid-Atlantic Ridge. Instead, he was to be chief scientist aboard a small Navy ship, Balamis, that had unexpectedly become available. He was to use one of the newly constructed bottom cameras to take photographs of the submerged conti- nental margin off the east coast, an environment that had never been photographed. Even in those expan- sive days of oceanographic science following World War II, it was most unusual to be chief scientist on one's first cruise. Ewing must have sensed that despite his inexperienced state, Bruce could be counted on to do the job. This change in plans proved providential: It set the stage for Bruce's uncompromising love for the seafloor and the processes that shape it. The stomach of a flatlander from Iowa was in no way prepared for the lively nature of a small ship at sea, and Bruce experienced terrible seasickness. He reflected later that had he gone to the ridge with Ewing aboard Atlantis, he would have been the youngest of a large number of students and, as just one of many under Ewing's tutelage, would have lacked the incentive to rally against relentless seasickness. Instead, he was put in charge of a ship and given the responsibility of carrying out a program. He persevered that summer, successfully taking 200 bottom photographs of the uncharted abyss. He took these photographs home with him that fall, and spent his senior year at the University of Iowa trying to understand and interpret them. Bruce found that he had more questions than he had answers, and the photographs became a catalyst for dedicating his professional career to the study of the seafloor. The 102 Oceanus Bruce's disdain for all things turkey remained inviolate for over 20 years until an oceanographic cruise in the late 1960s. The cook aboard the ship had his roots in southern cuisine and all animal and most vegetable products were fried in a cavernous deep-fat fryer that seemed never to be turned off. After six days of nothing but fried foods, one's interest in meals was low indeed. Finally, at lunch on the sixth day we arrived in the mess to find a roasted turkey on the menu; roasted, because the 20-pound bird was too big for the fryer. Bruce, who had not eaten turkey for over 20 years, hesitated, stared, poked, and then descended ravenously upon his portion. following summer he returned to the sea with Ewing, this time as a beginning graduate student, and spent two months aboard Atlantis continuing the Mid- Atlantic Ridge investigation. In the fall of 1948 he joined Ewing and his growing entourage of graduate students at Lamont Geological Observatory at Columbia University to begin his formal training. Marine geology was never to be the same. Like many talented people, Bruce was blessed with a very quick, multidimensional mind and an exceptional memory. He had an insatiable appetite for books on all aspects of ^^^^^^^^^^^^^^^^^^^^^ Earth science. These characteristics combined to form an ability to see and think imaginatively about link- ages both between and across dif- ferent, but complementary, investi- gative results. One of the great break- throughs in our understanding of the Mid-Ocean Ridge system pro- vides a good example. When Bruce started working with investigators at Lamont on the data from Atlantis cruises to the Mid-Atlantic Ridge, the existence of a world- encircling ridge system was un- known. However, a ridge of some kind was known to extend the length of the Atlantic based largely on results from English and German oceanographic studies carried out before World War II. These investiga- tions showed the bottom water of the eastern and western basins to be different and, therefore, separated by a barrier with unknown properties. Also, investigators sailing under flags of a variety of countries before World War II had documented, with widely spaced soundings, the existence of the Albatross Plateau in the equatorial eastern Pacific (known today as the East Pacific Rise) and the Carlsberg Ridge in the northeastern Indian Ocean. Bruce's initial project at Lamont was to compile all the available sounding profiles across the ridge in the Atlantic, in an attempt to characterize its spatial properties. He was assisted in this task by Marie Tharp, a new research assistant, who had recently completed an M.S. in geology at the University of Michigan. Marie tackled the tedious and demanding task of creating coherent profiles from noisy sounding data. She compiled six widely spaced profiles across the ridge in the North Atlantic and made the interesting observation that on each profile, the crest of the ridge appeared to be notched by a several-thousand-meter-deep valley that was 40- to 60-kilometers wide. Bruce was skeptical about the existence of such a valley at a regional scale, but intrigued by the notion. Coincidentally, Bruce was working with Ewing on another project evaluating the linkage between underwater avalanches of sediment, called turbidity currents, and earthquakes. For this study, Bruce created a plot of earthquake epicenters on a North Atlantic map drawn to the same scale as Marie's sounding-profile map. This was no coincidence; Bruce believed that plotting different kinds of data at the same scale facili- Winter 1991/92 103 Marie Tharp (in about 1956) working with the first-edition physiographic map of the North Atlantic. The famous six profiles across the Mid- Atlantic Ridge are on her right, two sound- ing records on her left, and an early globe that she and Bruce created of ridges in North and South America is at center. In collaboration with National Geo- graphic, Marie and Bruce made a physi- ographic globe of the earth in the late 1960s (below). tated comparisons. Laying the epicenter map over the profiles revealed that the earthquakes defined a broad belt of activity down the center of the North Atlantic, and in the six profiles Marie plotted across the ridge, the earthquake locations fell within the bounds of the axial valley. The occurrence of earthquakes along the Mid-Atlantic Ridge had been recognized previously, but the association of the seismic events and the axial valley Marie proposed was startling. Bruce became convinced that Marie's insightful observation about the existence of a rift valley was correct. The coincidence of the valley and earthquake activity, which is an indicator of rupturing of the earth's brittle yet elastic outer shell in response to forces that exceed the shell's strength, indicated to Bruce and Marie that this feature was dynamic and shaped by currently active processes. The earthquake belt was continuous along the length of the North Atlantic, leading them to speculate that the earthquake-belt location could be used to predict the axial-valley location in the absence of sounding data. As they slowly accumulated sounding profiles across the North Atlantic, their plots began to reveal a deep axial valley coincident with the seismic belt along the crest of the Mid-Atlantic Ridge. Ewing and Heezen then embarked on a project to plot the locations of earthquakes in ocean basins throughout the world, and noted once again the coincidence between belts of earthquakes and the crests of known, but seemingly separate, ridge segments scattered about the ocean basins. In addition, 104 Oceanus the earthquake locations defined a diffuse but continuous belt that linked the known ridges in the Atlantic, Indian, Arctic, and Pacific oceans, leading them to suggest that the world was encircled by a mid-ocean system of ridges that by their very existence, scale, and continuity, were central to the history of the ocean basins. They also observed that when sounding profiles crossed a branch of the ridge system, earthquake epicenters fell within the boundaries of the axial valley that was inter- preted to be an active rift zone. In 1956, they published their idea for a continuous ridge system, and were met with some skepticism. Cruises were planned to test their predictions by surveying unexplored portions of the southern Pacific and Indian oceans to see if, indeed, a ridge existed where proposed — and a ridge was always found. During this global synthesis, Bruce noticed that a limb of the mid- ocean seismic belt could be traced across the northwestern Indian Ocean into the Gulf of Aden, where it linked with a north-south trending zone of continental seismicity in East Africa. This belt of continental seismicity was associated with the network of East African Rift valleys that con- tained the great lakes of Africa, such as Victoria and Rudolf, and where fieldwork by British and German geologists had documented that the earth's crust was being extended in an east-west sense to create a north- south trending rift system. Bruce and Marie constructed topographic profiles across the rift valleys of East Africa and compared them with profiles across the Mid-Atlantic Ridge. The similarity of the profiles was striking. This, along with the continuity of the seismic belt, indicated to Bruce that the axial terrain of the ridge and the rift valleys of East Africa are genetically related. He proposed that the crust along the axis of the Mid-Ocean Ridge system is stretched at right angles to the axis. By the late 1950s, evidence for large displacements of the continents (based on paleomagnetic studies of rocks by British investigators) was compelling. Bruce suggested that these displacements were accommodated by the creation of crust at the ridge axes, and that the history of continental displacements were recorded in the seafloor's structural fabric. With this, a major pillar in our understanding of how the earth works was in place. During this phase of exploration and insight, Bruce and Marie realized that it was difficult to create improved maps of the seafloor because of the vast scale of the underwater terrain and the slow rate of sounding data acquisition. Following the techniques developed by continental cartographers, they created physiographic diagrams of the seafloor. Unlike a contour map that links points of similar depth with lines, a physiographic diagram creates an interpretive three-dimensional view of the seafloor. Such a presentation also allowed Bruce and Marie to extrapolate the seafloor's textured variations between widely spaced sounding lines. They finished their first physiographic diagram of the North Atlantic in 1956, and it was followed over the next 20 years by a series of physiographic maps that, in one form or another, covered all the world's oceans. These maps are remarkable because Bruce and Marie had an ability to visualize seafloor morphology in three dimensions and intelligently extrapolate trends and relationships into areas of sparse data to create depictions that have since been shown to be remarkably accurate. This collection is probably the most widely distributed set of seafloor maps. As such, it has provided a pictorial gateway to the earth's Earthquake locations defined a diffuse but continuous belt that linked the known ridges in the Atlantic, Indian, Arctic, and Pacific oceans. Winter 1991 /92 105 last frontier and captured the imaginations of students and researchers around the world. Bruce was happiest when he was at sea learning something new. He was indefatigable in this environment, where he seemed to derive the strength he needed to work night and day by feeding off the realization that a major discovery was in the making if he could collect the right kinds of data in the right way. Given the great expanse of unexplored ^^_^_^_^^___^^^^_^^_ ocean, more was always better— and Bruce worked himself, his colleagues, the ship, and its crew to the limit. Data were not mindlessly accumulated; each new observation of interest was studied and as- sessed for telling clues about the Bruce became an avid and enthusiastic user of submersibles in the late 1960s when this technology became available and permitted manual presence in the deep sea. One day, during an explanation to a graduate student about what to expect when this student made his first dive, Bruce found his descriptions about the experi- seafloor. With each new insight ence to be lacking. In mid-sentence, he jumped up from his came hypothesis testing and cruise- desk, turned off the lights, grabbed a flashlight and crawled plan modification to create a more under his desk. From his confined quarters under the desk, effective investigative strategy. To Bruce held the flashlight above his head and pointed the be a student working with Bruce at narrow beam out across the floor slowly sweeping the shaft of light across one partially illuminated object after another. About this time his secretary opened the door to a darkened room to find Bruce squeezed under his desk, shining a light about the room with a silent but bemused graduate student standing off to one side. Her look was incredulous, but before she could say a word, Bruce announced that he was diving in a submersible and was not to be disturbed. sea under these conditions (which I had the fortune to be during several cruises) was at once exhila- rating and fearful: exhilarating in that I learned so much because the arrival of new information in the form of a sample, photograph, or profile always precipitated lively and intense discussions about all the data's aspects and implications, and fearful because Bruce asked penetrating questions and expected intelligent answers. I knew when I had run aground with an idea if Bruce likened me to one of his feathered friends from his early days in Iowa. Bruce was a marvelous advisor to his graduate students. He under- stood and respected the sanctity of research and the freedom to follow one's own ideas. He exposed incoming students to the broad venue of research possibilities that lay between the shorelines of the world's oceans, and gave them free reign to choose problems of interest. He recognized the importance of exploring unknown and untested avenues of research and the time-consuming nature of this process. He had as many as 11 Ph.D. students at one time, all working on a broad range of problems, and, given his many involvements, he only had time to measure a student's research progress every few months. When the call did come, however, his students knew they had better be prepared, because Bruce would expect to be challenged by new observations and ideas. The sessions were often lengthy, as Bruce explored every aspect of the student's work. I always left these encounters exhausted but enlight- ened, because his probing questions and great depth of knowledge had provided new insights about my results, opening up new avenues of exploration. No matter how thoroughly I had analyzed a problem, Bruce always seemed to see farther. 106 Oceamis The science of the seafloor was Bruce's life, and the boundary between his work and everything else that constituted his being was invisible. When your goal is to map the world-ocean floors and under- stand how this great expanse was formed, you need a great deal of space, more than Bruce found available at the Lament Observatory, so certain mapping and writing projects were carried out at his or Marie's homes. Over a period of years, both homes evolved into laboratories with drafting tables, a multi- tude of maps, and books piled everywhere. Because these en- vironments were quieter than the chaos of his offices and la- boratory space at Lament, where technicians and graduate students swirled about, he would often work at one house or the other, and this is where a student might go to work with him, especially when preparing a manuscript. These gatherings could go on for hours as Bruce probed every sentence for clarity and insight, oblivious of the time. I remember that when projects were under way, there was al- ways a welcome place at the dinner table. After dinner, typically a very rare meat and an excellent bottle of Bordeaux, the manuscript honing would continue long into the night. Bruce would reluctantly loosen his grip on our text when he observed that I had fallen asleep. During his career Bruce received awards from many scientific societies for his fundamental contributions to marine geology and our understanding of the earth. He was a man for his time, because his wide- ranging interests and probing intellect were free and unconstrained by the lack of disciplinary boundaries found only in new fields of science. Today, marine geology and geophysics is a much more mature science with rigorously defined investigative disciplines, and it is difficult, if not impossible, to work on and contribute to the range of problems that Bruce examined. We were lucky to have him when we did. Paul J. (Jeff) Fox is Professor of Oceanography at the University of Rhode Island Graduate School of Oceanography. He was one of Bruce's graduate students at Lamont-Doherty Geological Observatory of Columbia University and worked with him at sea and in the laboratory from 1964 until Bruce's death. Under Bruce's tutelage, he was introduced to the intriguing mysteries and romance of the Mid- Ocean Ridge system; it is a love affair that continues to this day. On an expedition to the Caribbean in the early 1970s, Bruce Heezen discusses a dredged limestone sample with students aboard R/V Eastward. Winter 1991/92 107 BOOK & VIDEO REVIEWS Fire Under The Sea By Joseph Cone, 1991. William Morrow and Company, Inc., New York, NY; 286 pp. - $25. The discovery of hot springs on the seafloor is one of the most dramatic findings in marine science in the last 15 years. From the for- mation of mineral deposits to the existence of previously unknown biological communities, studies of these hot springs have had profound effects on our understanding of ocean floor processes. In Fire Under The Sea, Joseph Cone traces their exploration in an action-packed story that not only conveys the chal- lenges and excite- ment of exploring the ocean bottom, but also provides a glimpse into the lives and work of sea-going scientists. Addressing the field in general, Cone concentrates on the exploration of hot springs on the northwest coast of the United States. The book opens aboard the research vessel Atlantis II off the Oregon coast on a typical morning as scientists prepare to dive in the submersible Alvin to the ocean floor for a day of observations and sampling. This is the first of many "dives" the reader makes during the book's course and, through the thoughts and comments of the sci- entists, the story captures the essence of being part of an oceanographic expedition. Cleverly interwoven into the story of hot springs on the Gorda and Juan de Fuca Ridges is an account of the development of modern ideas of seafloor spreading and continental drift. As in any good mystery story, a number of unconnected pieces — in this case, studies done independently by continental and marine geologists and geophysicists — have been fitted together to produce a model of the Earth's surface plates created at mid-ocean ridges. From the time when Alfred Wegener first noted the "fit" of the continents of South America and Africa early in this century, the reader is led through scientists' work as they developed new ideas, designed experiments, and debated their results. Although the story is full of information, Cone manages to keep the reader's interest with anecdotes that illustrate the personalities of those involved. By the 1960s, the plate tectonics paradigm had gained general acceptance. Against this background, scientists predicted the occurrence of hot springs on the ocean floor and took on the challenge of proving their existence. Apart from the highly visible aspects of scientific discovery, there are many other important facets — from the technological developments that frequently pave the way for exploration, to the policy decisions necessary when new, potential mineral resources are found. Whether describing the development of sophisticated echo-sounding techniques or Law of the Sea negotiations, Cone's writing is authoritative, easy to read, and without jargon. In writing fire Under The Sea, Cone inter- viewed an impressive list of scientists and consulted a large number of references, both of which are presented in the book. The inclusion of a chronology of events at the beginning is helpful, and eight pages of color photographs illustrate the strange chimney structures and unusual organisms associated with seafloor hot springs, fire Under The Sea is an entertaining and compelling account of discoveries in a young and exciting field, and should be of interest to anyone curious about the ocean. — Susan Humphris Dean, Sea Education Association Woods Hole, Massachusetts WE'D LIKE TO HEAR FROM YOU. Oceanus welcomes and occasionally publishes letters from readers regarding editorial content or other ocean- science issues. Please write to the address on page 4. 108 Oceanus BOOK & VIDEO REVIEWS 4000 Meters Under the Sea By Films for the Humanities and Sciences, Princeton, N.J. 1991. 28 minutes - purchase $149/rent $75. This 28-minute video, made by NHK (Japan Broadcasting Corporation) and distributed by Films for the Humanities and Sciences, con- cerns a 1987 joint US-Japan expedition to study the geology and biology of the Marianas Trough spreading center. The film depicts the findings of hot vents and associated fauna at the accreting plate margin in the trough using the deep submersible Alvin and R/ V Atlantis II. It is aimed toward the lay audience. The film nicely captures the essence of a multidisciplinary expedition and the use of Alvin as a research tool. A voice-over commen- tary identifies the principal investigators, Bob Hessler and Jim Hawkins of Scripps Institution of Oceanography, and their Japanese col- leagues, but, regrettably, with the exception of a closing sentence from Hessler, we do not hear first-hand from the scientists. The film quality is good with some excellent close-ups of the vents and the vent communities. The development of the film's theme — the finding of and questions regarding hot vents — is well done (with some reservations that I will discuss later), and draws the audience along nicely. It makes for an interesting 28 minutes. There are a few, minor irksome aspects: Principally I found it most annoying to be continually told in the introduction, and near the end, that the expedition was to the Marianas Trench. Presumably this assertion, accompanied by continual references to the deepest part of the world's oceans, is an attempt to add glamour and excitement— certainly not needed here. In actuality, some excellent graphics clearly show the Marianas Trough and its relative position to the Trench. These graphics were so good and visually striking, I regretted that they didn't spend a couple of minutes more explaining the spread- ing center and its relevance to the Trench, plate motions, etc. The other irksome aspects were in the commentary: Generally it was well done, but a little better quality control could have avoided small things like referring to the Alvin manipulators as "magic hands" or, more importantly, reference to the planktonic food falling to the seafloor and being converted into bacteria (rather than by). On the whole, the scientific aspects were well covered and well explained. A few errors slipped through: References to "magma in the venting solutions making them cloudy," rather than products of rock-water interaction; the chimneys or smokers being alluded to as "cooled molten lava," rather than precipitates of sulfides and sulfates; the observed minerals being "crystals of iron, copper, and zinc," rather than sulfides of these elements. I think the interaction of seawater with the rocks, subsequent extraction of metals, gases, etc., and how they reach the sea surface had the weakest coverage, but the main points did come through. While recognizing the educational and EGRET EAUISIONS urith (Burt Jones and !Maurine Shimiock^ 'Travel & Marine-Life Photography Secret Sea tours scheduled for 1992 include **• Truk Lagoon, from March 23 to April 5; **- Dominica from June 17 to 25; and **• Sipadan Island, Malaysia (Borneo), from July 25 to August 9. Join us! For further information or consultation about photography and travel with Secret Sea Visions, write or call Burt Jones and Maurine Shimlock, Secret Sea Visions, P.O. Box 162931, Austin, Texas, 78716; (512) 328-1201. Winter 1991/92 109 entertainment aspects needed for this kind of video, I was most disturbed by the implicit (and at times explicit) suggestion that this was the first discovery of vents, and the expression of great "surprise" to find life at this depth; similar vents and life had been discovered and investigated in many locales, and at similar depths, before this expedition. The questions being addressed (implicitly for the first time): Where does the hot water come from? Could the cloudy waters be the key to life? How does it sustain life? How do the animals feed? All these have been previously investigated at other vents and are known to some degree. The really important aspects of the Marianas Trough vents were not made clear — their setting in a back-arc basin as opposed to a mid- ocean ridge, and their biological community that is slightly different and dominated by new species (hairy gastropods) compared to other vents. This is never spelled out, and previous work is not properly referenced, except for one comment from Hessler about how exciting it was to find "new friends" (animals) as well as "old ones" (previously discovered vent fauna) at the Marianas vents. All in all, in spite of my minor misgivings, I found it an interesting video. It should excite the lay audience, and the shots of the vent communities are certainly worth a look by serious submarine-hot-spring researchers. —Geoffrey Thompson Senior Scientist, Chemistry Department Woods Hole Oceanographic Institution Special offers for Students and Teachers! A student Oceanus subscription is only $20 a year, a savings of $5 off the cover price. For teachers, we offer a 25 percent discount on bulk orders of five or more copies. A discount also applies to a one-year subscription for class adoption ($20 each). THE NOW GENERATION of CONTINENTAL SHELF INSTRUMENTS and HOUSINGS. (HOSTAI-'ORM C™ PLASTIC) BATTERIES ELECTRONICS ACOUSTIC RELEASE Programmable TRANSPONDER SELF-RECORDING TIDE GAUGE with or without an ACOUSTIC RELEASE 1 Other MORS oceanographic packages • O.E.M. housings & internal frame- work for your specific needs Low cost. No external metal parts. 600 m operational depth. No corrosion or biofouling. MORS MORS ENVIRONNEMENT- MASSY, France (1) 64 47 19 19 MORS ENVIORNNEMENT- BREST. France 98458510 OCEANO INSTRUMENTS UK-EDINBURGH 031- 552 6619 OCEANO INSTRUMENTS-SEATTLE, WA (206) 363 0500 Interested in an internship? Oceanus would like to hear from anyone interested in being an intern with us in 1992. Excellent English skills and an interest in magazine publishing are the requirements. For futher information, contact Lisa Clark at (508) 457-2000, ext. 2393. ORIGINAL ANTIQUE MAPS & SEA CHARTS US. & WORLDWIDE GRACE GALLERIES, INC Box 2488, RR5 Brunswick, ME 040 11 (207)729-1329 Call or Write for Listings ANTIQUE MAPS • PRINTS CARTOGRAPHIC BOOKS 110 Oceanus continued from page 8 strike-slip fault — a fault showing predomi- nantly horizontal movement parallel to the strike; vertical displacement is absent subduction zone — area of crustal plate colli- sion where one crustal block descends beneath another, marked by a deep ocean trench caused by the bend in the submerg- ing plate. The downward movement of the subducting plate results in earthquakes, volcanos, and intrusions on the far side of the trench. swath mapping tools — instruments installed on a research vessel that use sound re- flected from the seafloor to map the shape of the seafloor along a band or swath that extends as far as 1 to 5 kilometers on either side of the vessel's track. Common instru- ment names are GLORIA, Hydrosweep, Sea Beam, and Sea MARK. tectonics — the forces and movements that create Earth's larger features terrane — the area or surface over which a particular rock or group of rocks prevails transform fault — a strike-slip fault of a par- ticular type where displacement stops abruptly, especially associated with offsetting of mid-ocean ridges transverse feature — a geological feature whose strike is generally perpendicular to the general structural trend of the region vent — place where water heated and altered by circulation through porous volcanic rock erupts from the seafloor, precipitating minerals and supporting sulfide-depen- dent biological communities The Oceanus staff acknowledges the valuable aid in assembly of this glossary of The Facts on File Dictionary of Geology and Geophysics (© 1987 by Dorothy Farris Lapidus, Facts On File Publications, New York, New York, and Oxford, England). Be a Member of The Woods Hole Oceanographic ASSOCIATES Join the growing number of people who care about our ocean environment as it is today. ..and could be tomorrow. For almost forty years, WHOI Associates have helped make possible the Woods Hole Oceanographic Institution's cutting-edge research. You can share the excitement of our research through our magazine Oceanus, newsletters, tours, and special visits to the Institution. For more information, please contact: E. Dorsey Milot, Director of the Associates Woods Hole Oceanographic Institution Woods Hole, MA 02543 or call (508) 457-2000, ext. 2392 The Young Associates ^W ,;_, 1 Let Ocean Explorer introduce a child to the excitement and challenge of ocean science. Recommended age group: 11 to 13. For information about membership in the Young Associates Program, write to E. Dorsey Milot, Director of the Associates, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, or call (508) 457-2000, ext. 4895. Winter 1991/92 111 ••*A CREATURE FEATURE: These photographs represent the many interesting animals found in the colonies that thrive around mid-ocean ridge hydrothermal vents. (For a broad look at recent work on vent communi- ties, please see The Biology of Deep-Sea Vents and Seeps... by Rich Lutz on page 75.) The large photo is a field of tube worms, Riftia pachyptila, on the Galapagos Rift, and the upper right inset is a closeup of these animals. The lower left inset is an Alvin camera view of giant clams, Calyptogena magnified, and scavenging galatheid crabs, Munidopsis sp. At upper left, a line of crabs marches along the edge of a bed of mussels, Bathymodiolus thermophilus. At lower right, the Pompeii worm, Alvinella pompejana, was photographed on the surface with the tube it calls home. (Large photo by Kathleen Crane, small photos clockwise from lower left by Alvin exterior camera, Robert Hessler, Dudley Foster, and John Porteous.) MBL WHO! LIBRARY COM EXT 1992 in Oceanus: Ocean Sciences — Four Disciplines • The four issues of Oceanus for 1992 will make up a volume on four basic disciplines in oceanography. Each will offer a summary of the discipline and articles that expand on several topics. • The Spring Issue features Marine Chemistry, including an update on carbon dioxide and climate, how tracers aid the study of marine processes, a look at natural marine chemicals and their uses, and the chemistry of seafloor vents. • We will also bring you several regular departments- Focus on the Coast, Toolbox, Issues in Ocean Law & Policy, and Creature Features. • Our Summer Issue will revolve around Physical Oceanography, Biological Oceanography will follow in the fall, and Marine Geology & Geophysics will wrap up the series next winter. • Don't miss this interesting, educa- tional review of oceanography in the '90s! ORDER BACK ISSUES! What's Still Available? • Reproductive Adaptations Vol. 34/3, Fall 1991 • Soviet- American Cooperation Vol. 34/2, Summer 1991 • Naval Oceanography Vol. 33/4, Winter 1990/91 • The Mediterranean Vol. 33/1, Spring 1990 • Pacific Century, Dead Ahead! Vol. 32/4, Winter 1989/90 • The Bismarck Saga and Ports & Harbors Vol. 32/3, Fall 1989 • The Oceans and Global Warming Vol. 32/2, Summer 1989 • DSV Alvin: 25 Years of Discovery Vol. 31/4, Winter 1988/89 • Sea Grant Vol. 31/3, Fall 1988 • and many, many, more... To place your order, send a check or money order (payable to WHOI) to: Oceanus Back Issues WHOI Woods Hole, MA 02543 Please enclose $8.00 plus $1.00 shipping and han- dling for each magazine ordered, and include a street address and daytime telephone number with your order. Allow 3 to 4 weeks for delivery. All payments must be made in US dollars drawn on a US bank. For orders outside the US, please add an additional $1 .00 per item for shipping. Sorry, we cannot deliver to Canadian addresses. For information on other available back issues, con- sult the Oceanus editorial offices at the address listed above. Back issues of Oceanus are also available on microfilm through University Microfilm International, 300 N. Zeeb Road, Ann Arbor, MI 48106. Explore Educational Programs at the Woods Hole Oceanographic Institution With a commitment to the future, the Woods Hole Oceanographic Institution, one of the world's leading Oceanographic centers, offers unique experiences in graduate and post-graduate studies. WHOI sponsors numerous fellowship and traineeship programs, and offers graduate degrees through a Joint Program with the Massachusetts Institute of Technology, providing a first-rate education through both hands-on experience and classroom instruction by many of the world's leading oceanographers. WHOI also offers select educational outreach programs for the K-12 grades. Find out today what WHOI can do for you. For more information contact the Education Office of the Woods Hole Oceanographic Institution. Applied Ocean Sciences & Engineering, Biology, Chemistry, Qeology & Qeophysics, Physical Oceanography... Postdoctoral Awards, Summer Student Fellowships, Research Fellowships, Summer Study and more . . . / Woods Hole Oceanographic Institution 1930 Education Office 86 Water Street Woods Hole, Massachusetts 02543-9903 Telephone (508) 457-2000, ext. 2200 Fax (508) 457-2188