A tumbler experiment using the 0.0625- to 4-mm fraction of four granodioritic grus samples was performed to investigate the nature of fracturing in the production of siliciclastic sand. Petrographical modal analysis was performed on 240 subsamples representing four major lithological constituents (quartz, potassium feldspar, plagioclase, and polymineralic rock fragments), four grain-size fractions (0.50 to 0.70, 0.35 to 0.49, 0.25 to 0.34, and 0.177 to 0.24 mm), and four time periods (untumbled grus; 2, 4, and 8 days). Each tumbling day represents a maximum of 25 km of transport in water. Tests of sample means and analysis of variance suggest that all subsamples are statistically homogeneous with respect to the number of quartz, potassium feldspar, and plagioclase grains among these size fractions. There is a highly significant difference in the number of rock fragments in the 0.50- to 0.70-mm fraction as compared to finer grained fractions, but no such differences occur within the medium- and fine-grained sand fractions. These results are inconsistent with the concept that lithic composition of the detrital light-mineral fraction is strongly size dependent within the 0.177- to 0.50-mm range. Where such compositional differences do occur, they may reflect sedimentological processes other than the comminution of grains from a single protolith. Grains from three grus samples examined for size changes displayed relatively rapid rates of disintegration during the first few days of tumbling, and then continued to fracture at reduced rates from days 4 through 8. From 8.9 to 16.2% of the relatively coarser grained fractions experienced comminution during 8 days (≃200 km) of transport. Of this amount, the sand fraction increased from 1.7 to 3.4%, whereas the silt plus clay fraction increased from 6.4 to 14.5%. For each sample, the weight percent of the fine- and/or medium-sand fraction remained nearly constant. Shape-ratio values (length of short axis/length of long axis of the maximum projection grain outline) as well as amplitude-ratio values for harmonics 2 and 3, as measured by Fourier analysis, were used to examine the character of quartz-grain fracture as evidenced in the 0.25- to 0.50-mm size fraction. Although some across-grain fracture occurred during the first day of tumbling, an overall increase in the frequency of smaller shape-ratio values indicated the dominance of grain-parallel fracture in all three samples. Scanning electron microscopy indicated that abrasion was not obvious even after 8 days (≃200 km) of transport. Average shape-ratio values were computed for at least 200 quartz grains in each of 49 samples from four high-gradient streams in southern California. Overall, the average shape-ratio values are remarkably consistent within each of the four streams, and closely resemble the value characteristic of the associated grus sample. Although significant decreases in average shape-ratio values occur along some local, very steep stream gradients due to high rates of grain-parallel fracture and/or shape sorting, the values reequilibrate after only 7 to 10 km of transport along lower gradient stream segments. Reequilibration probably reflects the dilution of more elongate by less elongate grains along lower stream gradients. No significant changes in the average shape-ratio values occur where small tributaries draining the same protolith enter the trunk streams, from sidewall erosion of the same source rock or across a contact with younger granitic source rock. Simple shearing is known to produce self-similar fracturing. Such fracturing is suggested by the lithological analysis as well as highly elongate quartz grains as observed in this and other studies. Self-similar fracturing may be an important mechanism in explaining the compositional and textural aspects of nascent siliciclastic sand derived from plutons subjected to emplacement and/or postcrystallization tectonism.

Abstract Petrographic modal analysis of the medium sand fraction (0.35 to 0.50 mm in diameter) of 274 samples principally from the shore zone of Lake Tahoe indicates that source-rock composition and mechanical-energy level at the depositional site explain approximately 46 and 32 percent of the total sample variance, respectively. A strong mechanical-energy signal may persist within a restricted grain-size interval. Q-mode factor analysis and stepwise discriminant function analysis were used to define six petrofacies, which reflect source-rock composition and relative mechanical-energy levels. Mean values for total quartz, total feldspar, and total rock fragments for petrofacies 1, 2, 3, 4, and 6 fall in the lithic arkose field of McBride (1963), whereas the mean for petrofacies 5 falls in the feldspathic litharenite field. The areal distribution of these six petrofacies may be used to identify nine shore-zone divisions, one of which is due to a major beach-nourishment project completed near Tahoe City in the early 1900s. Given the 1-km sampling interval employed, the areal positions of seven of the nine divisional boundaries showed no change from September 1978 to October 1988. This compositional stability reflects the consistency of the upland drainage systems and that of the wind-derived current systems operative in Lake Tahoe. The importance of source-rock composition, relative mechanical-energy level, and the apparent temporal consistency of the areal distribution of the shore-zone petrofacies should prove useful in constraining interpretations of older siliciclastic shoreline deposits associated with large lakes.

Abstract From Shoreline to Abyss: Contributions in Marine Geology in Honor of Francis Parker Shepard - Francis P. Shepard left a rich scientific legacy including more than 230 published papers and books primarily addressed to the study of submarine canyons and turbidity currents, continental shelves and associated sediments, coastal processes and sediments and marine physiography and tectonics. He is best remembered for his work on submarine canyons; however, his broad range of scientific interests and his remarkable ability to break new ground in each of these disciplines have served as a model for at least four generations of ?Shepard? students. This new work from these Shepard students addresses problems in marine geology from the global scale to the local outcrop scale. Relationships among tectonics, eustacy and both siliciclastic and carbonate sedimentation create a unifying theme. Special topics include coastal processes, shelf and slope evolution, and submarine canyon and fan systems.

Abstract Francis Parker Shepard was born in Brookline, Massachusetts, on the 10th of May, 1897, to Thomas Hill and Edna (Parker) Shepard of the Shepard Shipping and Lumber Family. This was an auspicious year for geology because it was punctuated by the publication of Sir Archibald Geike's classic book on "the founders of geology" (Adams, 1954). Shepard was later to become one of the founders of marine geology for he would acquire the noble title of the Father of Marine Geology (Dietz and Emery, 1971). Thomas Hill Shepard was of strong New England mariner's stock and provided well for his family, which included a daughter Katherine and son Francis. Both youngsters learned to love the sea from their parents, and spent many happy summer hours aboard personal pleasure yachts sailing the coastal waterways of the New England coast. The winters were devoted to classroom activities in both public and private schools. After completion of his secondary-school education, F. P. Shepard entered Harvard University in 1916 with geology as a major. The next several years were filled with studies and summer field work in the Rocky Mountains with Professori. B. Woodworth. Professor Wallace A. Atwood (Ph.D., University of Chicago, 1903) directed his thesis work and Shepard graduated with a B.A. at the end of the summer in 1919. Professor Atwood joined Harvard University as Professor of Physiography in 1913, succeeding W. M. Davis. He went on to become President of Clark University in 1920. At the fiftieth Anniversary Celebration of the University

Abstract An attempt is made to understand some of the ways that marine geology developed during the past 50 years, essentially the working lifespan of an active but venerable scientist. This interpretation is aided by comparing marine geology with the development of land geology during a longer period, and by attempting to understand the relative roles of science and technology in the field of marine geology. Excursions from simple straight-line advance for all geology (and also for other fields) are provided by the unexpected appearances of broad generalizations, or paradigms, that commonly are developed by a few scientists and opposed by many, at least for a time. These sudden advances await the accumulation of critical masses of knowledge that, in turn, depend upon exceptional opportunities, partly in the form of adequate funding and partly by transfer of technology. These unusual circumstances make accurate prediction of future advances in mating geology (and in other scientific and technical fields) unreliable but still worthy of thought.

Abstract The areal distributions of active-beach widths, eolian-dune fields, coastal-terrace heights and beach grain size are examined along a 1,000-km open coastline in the convergent Cascadia margin of the northeast Pacific. Beaches range from 10 to over 500 m in width in a total of 42 continuous beach segments, which are considered proxies for littoral cells at least 2 km long. Tectonic upwarping has exposed coastal-terrace deposits (1–30 m thick) and/or underlying bedrock to elevations as much as 120 m above mean sea level. The regional distributions of short-beach segments (average of 8–12 km long) in northern Washington and northernmost California—southern Oregon correspond to areas of maximum variation in terrace height, reflecting relatively rapid rates of tectonic vertical deformation. The three largest beach segments, 65 to 165 km long, are associated with major dune fields, and correspond to three of the largest river sources, namely the Eel, Columbia and Umpqua Rivers. Tectonic downwarping of inner-shelf sediment sinks might account for narrow, short-beach segments associated with the Klamath and Rogue Rivers. Six beach segments in central Oregon are isolated from either major river or high-terrace sand sources, and are assumed to be derived from trasgressive shelf sands. However, the observed longshore variability in these beach deposits indicates that pre-existing distributions of offshore sands were not uniform along the coast. Finally, longshore trends in increasing beach width, increasing dune development and/or decreasing beach grain size to the north in about one-quarter of the analyzed beach segments suggest a small net northward transport. However, inconsistent trends of all three variables in the remaining segments confirm the occurrence of short-term conditions responsible for reversing longshore transport in littoral segments effectively bounded by prominant headlands.

Abstract Beach compartments or littoral cells form the framework for our understanding of the sources, transport, sinks, and storage of sand in the nearshore zone. in general, along the California coast, beach sand is derived from rivers or bluff erosion, moves alongshore under the influence of prevailing wave conditions, and ultimately is lost either to a submarine canyon or dune field. The Santa Cruz Littoral Cell appears to extend as far north as San Francisco Bay and terminates downcoast at Monterey Submarine Canyon. Northwesterly waves drive littoral drift at a rate of about 200,000 to 250,000 m 3 /yr at the Santa Cruz Harbor. The major sources of sand within the Santa Cruz cell are coastal streams draining the Santa Cruz Mountains and 130 km of coastal bluffs. On the basis of the grain-size distribution of beach and nearshore sediments, a littoral cutoff diameter (0.18 mm) was established in calculating a sediment budget. Fluvial sediment transport measurements combined with flow-duration curves and grain-size distributions were employed to calculate input from coastal streams to the cell. Cliff height and length, sand content, and average long-term (50 to 75 years of aerial-photograph coverage) erosion rates were utilized to determine the littoral contribution from seacliff and bluff retreat. Coastal streams supply about 75 percent of the total littoral-sand input to the cell, bluff erosion contributes about 20 percent and the remaining 5 percent is from gully erosion and sand-dune deflation. Sediment input to the cell is highly episodic in response to large and infrequent erosional events. The processes that deliver sand to the cell (principally bluff erosion and high stream flow) may operate at different frequencies than those that move sand through the cell (longshore transport). An additional complexity arises from potential changes in sand storage, either on the beach or along the inner shelf, which are capable of producing significant volumes of sand due to the large areas involved.

Abstract The identification of sand sources and the volumetric contribution from each during a given time interval is essential to our understanding of coastal processes and the computation of meaningful sediment budgets for the littoral zone. Fourier grain-shape analysis (FGSA) of detrital quartz from the medium-sand fraction of 83 fluvial, seacliff, beach and inner continental-shelf samples was performed to identify and quantify the contributions from local sand sources within the Oceanside Littoral Cell, which occupies the coastal reach between Dana Point and Point La Jolla in southern Orange and San Diego Counties. Differences in shoreline orientation, physiography and the occurrence of Oceanside Harbor may be used to divide the Oceanside Littoral Cell informally into five segments. These are, from northwest to southeast, the Capistrano subcell (Dana Point to San Mateo Point), the San Onofre subcell (San Mateo Point to the north breakwater at Oceanside Harbor), the Oceanside-Carlsbad subcell (south jetty at Oceanside Harbor to the onshore projection of Carlsbad Submarine Canyon), the Encinitas subcell (Carlsbad Submarine Canyon to La Jolla Submarine Canyon), and the La Jolla Cove subcell (La Jolla Submarine Canyon to Point La Jolla). FGSA demonstrates that each of these subcells is compositionally distinct with regard to grain-shape populations, which, in part, reflects a characteristic set of local sources supplying each subcell. Although shape-compositional differences occur between and among the five subcells, the shape composition within each cell is rather uniform with the possible exception of the Santa Margarita River sand in the San Onofre subcell. This homogenization may reflect sediment mixing as a result of seasonally related, bidirectional, longshore and onshore transport during a relatively recent but undefined time interval. With the probable exception between the San Onofre and Oceanside-Carlsbad subcells since 1942, an appreciable amount of sand has been exchanged naturally between each pair of adjacent subcells. Although the reason(s) is (are) not clear, the northernmost two subcells are dominated by downcoast and southern shelf sources associated with each subcell, the central subcell appears somewhat transitional with regard to upcoast and downcoast sources, and the southernmost two subcells are dominated by upcoast and northern shelf sources. Continued monitoring is required to estimate the volumes involved and the rates of exchange, which will determine the effectiveness of the subcell boundaries as barriers to littoral-sand transport. An average of approximately 36 percent of the quartz in the medium-sand fraction of the foreshore samples was derived from the inner continental shelf. Although it seems likely that storm activity, perhaps coupled with shoreward transport through ridge and runnel systems, may be involved, fundamental questions concerning the mechanism, volumetric aspects, rates of transport and transport pathways remain unanswered. Weighted local source estimates supplied by FGSA, grain-size and mineralogic analyses must be integrated with mass-balance computations to improve sediment budget analysis and to constrain sand-transport pathways to and within the littoral zone.

Abstract Pleistocene nearshore deposits (part of the Merced Formation) exposed in sea cliffs south of San Francisco, California, contain elongate, gravel-filled casts at the bases of some gravel or sand beds. The size and shape of these linear features resemble those of gutter casts of sand protruding into shale in other deposits. The gravel-filled gutter casts of the Merced Formation show a strongly preferred east to northeast orientation, approximately normal to the shoreline trend inferred on the basis of the modern shoreline trend and the regional structural grain. Truncation and absence of deformation in the lamina of the sand into which the gutter casts extend indicate that the features are erosional rather than constructional (due to loading), and their orientation implies that they were cut by oscillatory currents beneath shoaling waves. As such, they provide a useful indication of the shoreline trends in other ancient gravelly nearshore deposits. in the Merced Formation, high-angle cross-bedding is abundant and dips primarily in a longshore or obliquely offshore direction. The variability of the cross-bedding makes it a less reliable indicator of shoreline trends. Steep, in some cases vertical or overhanging, walls to the gutter casts and the presence of clasts of sandstone in the casts that are indistinguishable from the underlying material indicate that the structures are formed by rapid cutting and filling of nearshore sand that is unexpectedly cohesive.

Abstract Sedimentation in regions of high (100–1,000 Bubnoff units [B]) and superhigh accumulation rates (more than 1,000 B) differs drastically from that characterized by normal rates (less than 100 B), in terms of distribution, mass, composition and structure. One Bubnoff unit is 1 mm/1 ka. High-accumulation rates result in high-water content, high-sediment mobility, and high organic-carbon content. The ability of sediments to slide and flow into depressions increases, and the main transport processes are gravitational flows. Sedimentation associated with avalanche rates is the principal sedimentation process on Earth. The thickness of normal deposits in the pelagic realm does not exceed 1 km, but in the avalanche-sedimentation regions, thickness reaches as much as 10 to 15 km. Normal sedimentation is associated with dispersion of sedimentary material, whereas avalanche sedimentation is associated with the concentration and formation of thick, autonomously developing sedimentary bodies. Avalanche sedimentation occurs where suspension concentration exceeds 5 mg/l, where sediment accumulation rates of more than 100 B occur, and absolute masses of more than 5 mg/cm 2 /l ka accumulate. According to the hypographic curve of the Earth, three global levels are revealed: 1) river mouths (river-sea border), 2) continental rises, and 3) deep-sea troughs. They are each separated by 3 to 6 km of topographic relief, but are related closely by sedimentary source material. Sediment translocation occurs with hiatuses and is a function of sea-level changes. With an increase in sea level, deposition occurs on the first level of concentration, and/or hiatuses form on the second and 3rd levels; with a decrease in sea level, hiatuses form on the first level and concentration occurs on the second and third levels. These three vertical levels of sediment deposition, related to each other in space and time, are unified by sediment composition (including organic carbon). These sedimentary systems (each system includes collection, transportation and accumulation of sediment) include subaerial and subaqueous aspects. Concentration of organic-rich material with isostatic crustal downwarping takes place in these zones, and in some cases results in oil and gas formation. This is the principal zone of oil and gas formation. Capture of the main supply, more than 90 percent of the river-suspension load on the first level, results in a deficiency of pelagic material. Precipitation of the dissolved load forms sediment by bioprocesses (CaCO 3 , SiO 2 amorphous, C organic, etc.). However, river-suspension residues (7–8 percent), absorption, eolian, ice and endogenous-material input are most important. The concept of avalanche sedimentation requires new approaches to the questions of sedimentation, geochemistry and the formation of mineral deposits.

Abstract Shepard's classic paper on delta-front gullies on the Mississippi delta slope initiated a host of modern studies that revealed the complexity and importance of subaqueous mass-movement processes. Differential weighting and diapirism, rotational slumps, retrogressive slides and their gullies and depositional lobes are but a few of the processes that mold the morphology of the delta-front area. These instabilities form on extremely low-angle slopes (often less than 0.5°) and are responsible for the transport of large volumes of shallow-water sediment to deeper environments.

Abstract Regressive coastal deposits containing internal downlapping surfaces are common on continental shelves of the world. Through theoretical considerations and evaluation of examples from the literature and our own studies in California and Italy, we have examined the conditions that lead both to the formation and preservation of these deposits. Coastal downlapping deposits form by progradation of coastal and deltaic lithosomes during stable and falling sea level. in addition to the rate and direction of sea-level change, the major controls governing the development of downlapping deposits are sediment availability and shelf morphology (gradient, surface irregularity, and depth of shelf break). Continental margins receiving a large, continuous supply of sediment commonly have vertical stacks of thick, laterally extensive deposits; those fed by relatively small coastal streams may have well-developed shelf-margin deposits if accommodation space was available. in the absence of feeder streams, shelf-margin downlapping deposits can form from locally derived sediment. Regressive coastal deposits are preserved in a variety of settings, but they are least likely to be preserved on broad, low-gradient (coastal plain) shelves, where small drops in relative sea level were accompanied by large seaward shifts of the shoreline. Even on high-gradient tectonic margins, downlapping deposits commonly are not preserved on the midshelf. The process of shoreface erosion during a transgression is efficient at planing off deposits from previous depositional cycles. Where the deposits are thin, the deposit is partially or wholly reworked; where they are thick, the basal part may be preserved. Regressive deposits are most likely to be preserved along the shelf margin, where relatively thick sequences form at or below the position of the lowstand sea level. On many continental shelves, the coarse sandy texture of outer-shelf sediment largely reflects original deposition in a regressive coastal environment during a fall and lowstand of relative sea level.

Abstract The floor of the Yellow Sea is a geologically mundane surface: it is nearly horizontal, lacks relief, and, with few exceptions, is devoid of conspicuous geomorphologic features. However, it is the principal repository for the prodigious sediment load of the Huanghe (Yellow River); and, due to its inherent shallowness (average depth is 40 m), it is frequently stressed by waves generated by winter storms and typhoons. Analyses of mass physical properties of cores representing the upper few meters of sediment in the central and north-central Yellow Sea (near the Shandong Peninsula), in conjunction with analyses of slope stability, failure modes, and erodibility, permit an assessment of the likelihood and effect of dynamic, transient geologic events on the seabed. Vane shear-strength profiles along with consolidation test data indicate that the present surface of the seabed is in a depositional mode and is compacting normally. in addition, liquid-limit profiles imply that in the study area these neritic sediments have been accumulating in an environment that probably has not been modified significantly since sea level reached its current level. There is no geotechnical evidence in the nine cores recovered that slope failures have occurred, and clasts, sand lenses or other manifestations of mass movements, including flows, also are absent. These observations support previous interpretations of seismic records. Moreover, slope stability analysis for static conditions shows that the sea floor is quite stable. Regardless, shear-stress levels generated by cyclic loading during major storms may approach the sediment shear strengths, and, when coupled with concomitant excess pore pressures, could cause slope failure. Unless the failed beds collapsed or flowed, however, there probably would be little conspicuous evidence of such a failure. in fact, evaluation of the potential of these sediments for disintegrative behavior suggests that they are not prone to either collapse or flow. Storm waves also generate oscillatory bottom currents that may erode the seabed. Whether the sediment is considered as cohesionless or cohesive, typhoons could have the potential to erode at all water depths within the Yellow Sea (i.e., to 90 m), and winter storms to water depths of 60 m or more. However, in the case of cohesive behavior, it could be that the effect of winter storms and most typhoons is generally less extreme. If the sea floor is repeatedly scoured, it is likely limited to the top few centimeters. Despite the fact that storm waves may cause slope failure and are certainly responsible for frequent scouring, they probably leave only a subtle sedimentologic imprint on the seabed.

Abstract Profiles of slope-normal gradients of continental slopes of the U.S. Pacific margin can be related to the large-scale structural framework of the margin, and to the secondary effects of sedimentation. Tabulations of average gradient, steepest gradient, slope width and the relation of steepest segment to midslope depth define three major slope provinces. The first two regions form the area extending from the Agua Bianca Fault Zone and its offshore extensions off northern Baja California to the Mendocino Fracture Zone. The boundary between region 1 and region 2 has been taken at Point Conception. The boundary separates 1) the outer steep-slope province off the California Borderland, the Patton Escarpment, and 2) a northern province with gentler declivities, usually because of modification by deposition. Both provinces show possible effects of the intersection of oceanic fracture zones with the margin. The northern province (region 3), from Cape Mendocino to Juan de Fuca Strait, can be subdivided into two subregions approximately at Heceta Bank, Oregon. The major provinces correspond on the first-order level to the two main modes of margin interaction with the adjacent ocean plates, underthrusting and transform motion. Second-order factors are the location of fracture zones, and deposition. Within the California Continental Borderland, the shallower base-of-slope depths and narrower shelves produce slopes that have lower relief and are generally steeper than the main continental slopes. These are secondarily modified by depositional processes and mass failure. There is a general increase in depth of slope base and in steepness to the south, illustrating the regional depression to the south, as well as the decrease in sediment supply in that same direction. It is evident that regional characteristics of the morphology of slopes is a function of the structural setting and of the local sediment supply.

Abstract Lithified stromatolites, attaining 2-m heights and associated with cement-encrusted carbonate-mud beds, are found in many inter-island channels of the southern chain of Exuma islands and cays. They occur within migrating flood-tidal, ooid-sand bars and dunes. Reversing tidal and wind-driven currents, with velocities up to 3 kts (150 cm/s), flow across depositional sites for 3 hrs of each 6-h tidal period. Such a high-energy, bank-margin environment is not usually considered to be the site of stromatolite growth or of mud-size particle deposition; however, mud beds occur in water depths of 4 to 8 m. The normal thickness is usually about 10 cm, but one large bed has a total thickness of 1 m. If exposed to current flow for several weeks, the hard crusts become colonized by a diverse microbial community of algae, diatoms and marine plants similar to that growing and trapping ooids on nearby lithified stromatolites. The white lime muds forming the interior of the crusted-mud beds are cohesive, having the consistency of "tooth paste." At first glance, the interior muds appear to be homogeneous; however, they are pelleted. Uncemented ooid sand occurs above and below the laminated, crusted-mud beds, a relation that shows they occupy the same depositional environment. Their age of less than 4 ka and their depth of occurrence argue against the lenticular-bedded muds being relict deposits associated with the Holocene flooding of the Bahamian platform. Scanning electric microscope (SEM) images reveal that the muds contain pelletoids of aragonite needles. The ooid grains are held in place and cemented by organic remains of the microbial mat. It is clear from the SEM photomicrographs that the surface crusts on the mud beds are actively undergoing diagenetic changes, whereas the encased uncemented mud is not. Strong currents scour and undermine the crusted beds, forming "rip-up clasts" and "mud chips." Microbial mats colonize many of the larger chips, trapping ooid sand. The chips become part of larger stromatolites as they grow and coalesce. Cross sections of lithified stromatolites show that they contain clasts of mud chips incorporated in their laminated internal structure and are thus contemporaneous with stromatolite growth. The probable source of the lime-mud needles is precipitation from "whitings" and from resuspended biogenic muds. During tropical storms and hurricanes, fine-grained sediment that blankets the Pleistocene surface of the Bahama carbonate platform forms dense clouds of muddy bank water. The ebb tides and offbank, wind-driven currents carry the mud-charged water mass to the bank edge. There, every 6 hours, the muddy water mass comes in contact with flood-tidal oceanic water and intense mixing takes place. It is below the mixing zone of sediment-laden, slightly hypersaline bank and cooler open-oceanic waters that crusted, laminated-mud beds, interbedded with well-developed ooid-sand bodies, lithified stromatolites, hardgrounds and areas of grapestone, are found. The occurrence of Bahamian crusted muds within ooid-sand dunes, rip-up clasts and lithified stromatolites, all subjected to strong subtidal currents, represents a newly defined association. If seen in ancient deposits, such associations might be interpreted as supratidal or shoreline deposits. Finding these features in high-energy subtidal channels at the eastern margin of the Great Bahama Bank mandates caution when making such interpretations in the geologic record.

Abstract Two alternative explanations still compete regarding the formation of atoll lagoons: the classic Darwinian theory and the karstic-saucer theory. Although the deep drillings at Mururoa Atoll, French Polynesia, certainly favor the first alternative, the discussion remains open elsewhere. Progress in knowledge may come from investigations of coral pinnacles or knolls, which dot lagoons in various quantities. Data on sets of pinnacles in the Tuamotu, Society, and Gambier Islands are discussed. The recent endo-upwelling theory has tried to explain the formation of these features by an ascent of deep oceanic water rich in nutrients through the pervious mass of atolls. Shallow-core borings were made in 1988 into one pinnacle and the adjoining rim of Tikehau Atoll, Tuamotu Islands. Studies of nutrients in the interstitial waters support the endo-upwelling theory. However, preliminary examination of the cores shows the presence of a highly dolomitized Tertiary reef underlying the Holocene coating without intervening Pleistocene rocks. Tikehau thus appears as a rather special structure and the same type of investigations should be continued in other atolls in the Tuamotus. Finally, the discussion is extended to the Gambier almost-atoll, a structure akin to Truk, Carolines, described in 1970 by Francis Shepard. Here the question of pinnacle origin also exists but is complicated by peculiarities associated with the shape and subsidence of the volcanic basement, differential tilting of the barrier, and absence of a deep passage into it. Thus, the discussion needs to be enlarged to consider diverse parameters of coral reef history. Results expected from multiple borings are a good approach to understanding atoll formation, still in question in spite of various proposals since the middle of the 19th century.

Abstract The Florida Middle Ground (FMG), the northernmost living coral reef in the Gulf of Mexico, was investigated by high-resolution seismic-reflection methodology, surface-sediment analyses, and direct observation by submersible to identify controls and processes responsible for recent geologic development, and to determine how it relates to development of the surrounding west Florida continental margin, a non-reef-rimmed carbonate platform. Modern reef growth is developed on a foundation consisting of a dead reef that probably was formed during one or more sea-level highstands of the Quaternary. Initial reef growth probably was controlled by bathymetry and regional circulation patterns that served to recruit Caribbean fauna to the FMG. Surface sediments are distributed in indistinct zones exhibiting a patchy distribution controlled by physical processes and the rugged bathymetry. Sediments throughout the FMG consist dominantly of molluscan shell hash derived from both the reef ridges and surrounding shelf environment. Barnacle fragments are the only major constituent found associated with reef ridges and not the surrounding shelf sediments. Barnacles probably are prevalent on the FMG because of the suitable sites for attachment provided by the dead-reef foundation. Thus, barnacles provide the only major sedimentological indicator of a transition from an open-shelf to reefal environment. The lack of a clear distinction between sediment types of the two environments may have implications for those investigating ancient carbonate environments. Recent geologic development of the FMG parallels that of the west Florida margin. Driven by high-frequency sea-level fluctuations, similar carbonate sediments are produced in both settings during periods when the shelf surface is flooded. Thick surface sediments on the FMG, relative to the surrounding shelf, indicate reef ridges act to trap sediments much like shelf-edge reefs trap sediments on a reef-rimmed carbonate platform surface. Reef growth and sediment accumulation cannot keep pace with sea-level rise, which results in the drowning of FMG reefs along with the remainder of the west Florida carbonate margin.