Ancient sea water

NEWS AND VIEWS GEOCHEMISTRY------------------had modest objectives, proving the principles of operation with components that are well established and available. The computer uses a respectable 66 switches and 200 fibres. This number of components and the stored program mean that it is significantly more complex than previous collections of optical components. Access to the information circulating in the delay line is serial, as in a rotating disk file, and the 20microsecond memory time is reminiscent of the electronic memories of the 1960s. Can this be extended to larger, more powerful versions? All physical parameters of the computer will need to be changed by several orders of magnitude for it to make a competitive processor. Simply multiplying the number of components by some factor would not seem advantageous, as the dependence on time of flight would slow a machine in proportion to its physical dimension. And an optical-fibre/lithium-niobatebased system does not seem as susceptible to miniaturization as semiconductor devices. The use of larger numbers of gates would make it impractical to synchronize the elements through the adjustment of each separate connection length. The machine stores 64 16-bit words, 10,000 times fewer than the simplest personal computer. Large increases in the amount of data in memory would make random access most desirable. A more aggressive version of the system based on integrated optics suggests itself, but is far beyond the capabilities of known technology. Propagation velocities and path lengths could not be as well-controlled, nor would lengths be accessible for adjustment. These accomplishments and others require much ingenious invention and mastery of a great deal of technology to circumvent the physics that obstructs the introduction of optics into the essential logic of a computer. There is no doubt that optical communication will have a role in large computing systems; indeed, products that link major parts of a large system, for example, a disk file to a processing unit, by optical communication are already commercially available. The question is: how far will optical signalling penetrate into the parts that implement the logical action of the machine? The marketplace does not honour technology just because it is clever. Will optical communication between devices or chips, ever offer advantages, rather than merely be possible? And will the low-cost, low-power and high reliability that are needed to build a system of thousands to millions of devices be achieved? D Robert W. Keyes is at the IBM T. J. Watson Research Center, Yorktown Heights, New York 10598, USA. Ancient sea water Paul Knauth oceans retreated from the continents. Samples with excessive salinity were explained as sea water that had evolved chemically by reaction with the host rocks. The subsurface brines appeared to be a form of fossil sea water, albeit strongly modified, and their study offered a possible approach for inferring ancient ocean chemistry. In 1966, the hydrogen and oxygen isotope composition of water in sedimentary basins was convincingly interpreted as indicating that local rainfall had pervasively flushed out all traces of any ancient sea water that might have been deposited with the rocks 3 . At the same time, other isotope studies seemed to indicate that the original mineralogy and geochemistry of the host rocks themselves were obliterated and overprinted by this continuous flux of fresh ground waters that slowly evolved into concentrated Fluid inclusions of evaporation-concentrated sea water on by water/rock brines growth surfaces of 400-million year-old (Silurian) halite one-two interactions. This crystal from Michigan basin. Cubic inclusions are negative punch effectively wiped crystals of the cubic form of halite. Previous efforts to out hope of finding fossil deduce the chemical and isotopic history of the oceans sea water or of inferring have focused on these types of inclusions, altered by the effects of evaporation .. much or anything about the chemistry of the thermal alteration of rocks. Careful ex- ancient oceans from analyses of amination of the sequence of evaporite sedimentary rocks. minerals that formed in places where More recently, a rapidly growing numshallow seas once dried up suggests that ber of investigators has begun to reintermodern ocean chemistry was achieved at pret the chemical and isotope data for least several hundred million years ago both deep ground water and rocks. The and has remained fairly constant ever isotope data for many subsurface brines since 1 . However, uncertainties in this can now be reasonably interpreted in approach allow plenty of room for signi- terms of mixing of later fresh water ficant variations over time, and changes with ocean water that experienced variin seawater chemistry are often invoked ous degrees of evaporation on ancient to explain biological extinctions and salt flats 4 , and an elegant method has secular changes in the mineralogy of been established for back-tracking chemancient sedimentary rocks. ical evolution of brines to their original What is needed to evaluate the chem- seawater chemistry5 . Numerous workers ical history of sea water are actual are presenting strong arguments that examples of fossil water that can be marine minerals formed early and can be analysed chemically. Do they exist? found in very ancient sediments; the Johnson and Goldstein present candi- isotope and geochemical record of date samples on page 335 of this issue 2 . marine sedimentation has not been The idea that remnants of ancient obliterated by later alteration6 . In addiocean water are preserved is not new. tion, fluid inclusions trapped along the Ground water extracted from wells in oil growth faces of ancient halite crystals are fields becomes progressively saltier with now interpreted by some as evaporationdepth and eventually reaches the salinity concentrated sea water that has been of sea water. Until 1966, it was com- preserved for hundreds of millions of monly thought that this salt water was years (see micrograph above). Chemical ancient sea water that had seeped down analyses of these inclusions have been and was left stranded whenever the used to infer aspects of the isotope and ANYONE who has ever swallowed a mouthful of sea water has probably wondered why the ocean is salty. The question has never been fully answered, but the most tightly reasoned arguments suggest that chloride was outgassed from primitive rocks with water vapour shortly after the origin of the Earth and that sodium and other principal cations were released from the weathering and hydro- 290 NATURE · VOL362 · 25MARCH 1993 © 1993 Nature Publishing Group NEWS AND VIEWS chemical history of the hydrosphere 7 •8 • Amidst this ongoing controversy over the origin of subsurface fluids and the preservation of the ancient rock record, Johnson and Goldstein identify tiny mineral cements bearing strong textural, geochemical (...truncated)