Printer Friendly

Molluscan shell condition after eight years on the sea floor--taphonomy in the Gulf of Mexico and Bahamas.

ABSTRACT In 1993 and 1994, the shelf and slope experimental taphonomy initiative (SSETI) deployed shells of a suite of molluscan species in a range of environments of deposition (EODs) representing a range of depths, sediment types, and environmental conditions with the goal of measuring taphonomic rates over an extended period of time. In 1999 and 2001, SSETI retrieved skeletal remains from 41 locations in the Bahamas and on the Gulf of Mexico continental shelf and upper slope that had been on the seafloor for eight years. Here, we compare taphonomic processes in two different ocean basins, across 24 environments of preservation (EOP) to evaluate the influence of species, sedimentary environment, degree of burial, and water depth on the preservational process. Taphonomic signature after eight years was almost exclusively a function of location of deployment and, frequently, taphonomically-distinctive locations of deployment were subsumed within distinctive EODs. EOD-level characteristics were insufficiently discriminative to delineate environments of preservation. EOPs and EODs are not synonymous concepts. Across all sites and species, the dominant taphonomic process was discoloration. Dissolution was of penultimate importance; nevertheless the cumulative impact over eight years was insufficient to produce a significant loss in shell weight in any EOP. Maximum dissolution intensity was normally observed on the outer shell surface; the inner and outer shell surfaces are inherently different in their time course of shell deterioration. Principal components analysis (PCA) demonstrated limited co-occurrence of discrete taphonomic processes among the 24 EOPs. Breakage and edge rounding fell on the same PCA axis, but these two processes were independent of all others. PCA divided dissolution into three independent components that discriminated the inner and outer shell surface of bivalves (and spire and body whorl of gastropods) and pitting from the development of a chalky surface. Discoloration was dissembled into five distinctive discoloring processes: fading without subsequent discoloration, the development of a brown-to-red coloration, orange/orange mottled discoloration, development of a green/green mottled color, and gray-to-black discoloration. The only concordance of ostensibly distinctive taphonomic processes was the association of small pits on the shell surface with orange discoloration on the shell. Depth did not exert a single significant effect on any of the eight primary taphonomic factors resolved by PCA, likely because of burial processes. The trends in taphonomic signature cannot be explained by any simple combination of sediment type and degree of exposure. A comparison between two-year and eight-year deployments suggests that important revelations can be gleaned from short-term experimental deployments, yet the same comparison discloses the spuriousness of other inferences. Thus, long-term experiments are essential to understand the time course of preservation. The taphonomic process is, in general, slow, and nonlinearity in rates over time constrains the subset of inferences that can be deduced accurately from short deployment periods.

KEY WORDS: taphonomy, preservation, environment of preservation, dissolution, shell carbonate, burial, discoloration, mollusc, long-term experiment

INTRODUCTION

Taphonomic processes affecting the preservation of molluscan death assemblages often take place within the first few decades after an animal's death. Mounting evidence indicates that these processes can exert a significant impact on the composition of the subsequent fossil assemblage. Common taphonomic processes include decomposition, burial, fragmentation, dissolution, abrasion, precipitation, and infestation by epi- and endobionts (Powell et al. 1989, Smyth 1989, Brett 1990, Canfield & Raiswell 1991, Parsons & Brett 1991, Flessa et al. 1993, Walker et al. 1998). That any one of these can exert a significant influence within months of death, if not simultaneously, is well documented (e.g., Simon et al. 1994, Kiene et al. 1995, Christmas et al. 1997, Cadee 1999, Walker 2001, Zuschin et al. 2003). However, taphonomic rates, particularly as modulated by burial and encrustation (Kidwell 1986b, Powell 1992, Zuschin & Pervesler 1996, Zuschin et al. 1999, Walker & Goldstein 1999, Walker 2001, Powell et al. 2006), may extend the taphonomic process significantly in time. Taphonomic models have been developed to describe the process of preservation (Kidwell 1986a, Powell 1992, Kowalewski & Misniakiewicz 1993, Olszewski 2004, Powell & Klinck 2007). Such models suggest that taphonomic rates are likely themselves to be time-dependent. To date, however, we have only limited empirical information concerning the processes determining the fate of skeletal material after the first few months to years of taphonomic time (Peterson 1976, Allison 1990, Parsons & Brett 1991, Callender et al. 1994, Behrensmeyer et al. 2000, Powell et al. 2006).

In 1993, the shelf and slope experimental taphonomy initiative (SSETI) was established with the principal goal of measuring taphonomic rates of skeletal material over an extended period of time (Parsons et al. 1997, Parsons-Hubbard et al. 1999, Parsons-Hubbard et al. 2001). SSETI deployed experiments in the Bahamas off Lee Stocking Island (LSI, Fig. 1) and in the Gulf of Mexico (Fig. 2) in 18 distinctive environments of deposition (EODs). Experiments were retrieved from each of these EODs two years later. Parsons-Hubbard et al. (2001), Callender et al. (2002), Staff et al. (2002), and Powell et al. (2002) reported the results obtained from these two-year retrievals. White (2002) described the taphonomy of the encrusting foraminiferal fauna in detail and Walker et al. (1998, 2002) analyzed the results of tethered shell experiments run simultaneously with the primary SSETI deployments. In this contribution, we consider the fate of bivalves and gastropods left on the seafloor for a 6-to-8-y period, more than three times the period covered by these previous analyses, and longer than the heretofore longest running taphonomy experiment (Callender et al. 1994).

[FIGURE 1 OMITTED]

The central hypothesis of most taphonomic studies is that taphonomic characteristics co-occur predictably, defining "taphofacies," and that these can be used to characterize major environments of deposition (Brett & Baird 1986). Furthermore, the spatial distribution of these taphofacies should be associated with environmental gradients, such as depth and sediment type, that permit assemblage taphonomic signature to be interpreted within the framework of preservation potential and environment (Brett & Baird 1986, Kidwell et al. 1986, Meldahl & Flessa 1990, Callender et al. 1992, Callender & Powell 2000, Zuschin et al. 2003). The task, of course, is to understand how environmental gradients relate to changes in assemblage taphonomic signature. Most experimental taphonomic research has been restricted to nearshore regions generally shallower than 20 m (e.g., LaBarbera 1981, Plotnick 1986, Walker 1988, Walker & Carlton 1995), although exceptions exist: e.g., measurements of shell dissolution at hydrothermal vents (Killingley et al. 1980, Lutz et al. 1988, Lutz et al. 1994), estimation of the rates of taphonomy at petroleum seeps (Callender et al. 1994), and evaluation of the biodegradation of crushed skeletal material on shelves (Simon et al. 1994). SSETI was specifically designed to encompass a sufficiently wide range of continental shelf and slope environments to address the relationship between the taphonomic process, taphofacies, and the environment of deposition. In this contribution, we compare taphonomic processes, among 24 environments of preservation (EOP) and 10 molluscan species and evaluate the influence of species, sedimentary environment, degree of burial, and water depth on the preservational process.

[FIGURE 2 OMITTED]

METHODS

Site Description

Bahamas

Experiments were deployed in 1993 and 1994 by SCUBA and the submersible Nekton Gamma along transects AA and BA established by the Caribbean Marine Research Center/National Undersea Research Program (Fig. 1). Transect AA is located directly off an inlet that separates Lee Stocking Island from Adderly Cay to the NNW. The shallow sites on the AA transect are subject to fairly high wave and current energy and sediment transport because of the close proximity of this inlet. Transect BA is southeast of AA about half way down Lee Stocking Island and, as a consequence, is impacted by less current energy and sediment transport.

The experimental sites are located in seven distinctive EODs along each transect (Fig. 1): in sand channels on the platform top (15 m) and the platform edge (30 m), on ledges down the wall (70-88 m), on the upper (183 m--transect BA only) and lower (210-226 m) talus slope below the wall, and on the crest (256-264 m) and in the trough (259-267 m) of large sand dunes. In addition, a site was established in a shallow hypersaline coastal lagoon, Norman's Pond, on Adderly Cay. Parsons et al. (1997) and Parsons-Hubbard et al. (1999) provide descriptions of these experimental sites, which are reviewed here and summarized in Callender et al. (2002).

The platform on the eastern edge of Lee Stocking Island is characterized by a reef terrace containing carbonate sand flats, hardgrounds, and patch reefs that gradually deepen from shore to the platform edge at about 33 m. Experiments were placed in sand channels about midway across the platform top and near its edge. The 15-m arrays had moved significantly each year that they were observed and large scale sand movement was apparent from year to year. The arrays at 30 m were buried within the first year after deployment on both transects and not subsequently re-exposed (as far as observations permit). Parsons-Hubbard et al. (1999) provide details of burial at these sites.

The slope steepens dramatically at the platform edge and forms a wall that drops steeply (>60[degrees]) into Exuma Sound. The wall is punctuated by narrow ledges upon which the experimental arrays were positioned. These ledges have a thin veneer of coral-Halimeda-molluscan sand over a lithified surface and are subject to occasional periods of high current flow. Arrays had been repositioned to some extent over the six-to-eight-year period on transect AA.

The photic zone extends to approximately the base of the wall. Below this point, the forereef slope is characterized by shingled carbonate debris, large talus blocks, and small lithified carbonate outcrops. The intervening areas are often covered by a thin veneer (2-3 cm) of unconsolidated sediment. Stalked crinoids are frequently observed on the carbonate outcrops. Below the talus slope is a field of large dunes, partially cemented, up to 10 m tall. The dune crests roughly parallel the slope of the reef wall, with each dune crest being deeper than the preceding one.

Gulf of Mexico

Experiments were deployed in 1993 by the submersible Johnson-Sea-Link on a transect off Galveston, TX and at selected EODs along the shelf edge and slope to the northeast (Fig. 2). The sites located off Galveston were typical of continental shelf and slope terrigenous sediments in the northwestern Gulf of Mexico. The outer continental shelf site is a relatively shell-rich terrigenous mud representative of the upper Texas outer continental shelf (Curray 1960, Parker 1960, Gallaway 1988, Balsam & Beeson 2003). Sulfate reduction begins within millimeters of the sediment surface (Lin & Morse 1991). The assemblage is composed of a diverse set of small bivalves and a few gastropod and urchin remains. Few shells are articulated, most show signs of dissolution. Breakage is common. Encrustation is rare. The continental slope site is representative of the upper continental slope throughout most of the western Gulf of Mexico. Sediments contain pelagic carbonates and a very sparse benthic fauna. Sulfate reduction begins well below the sediment-water interface (Lin & Morse 1991). Few shells are encountered; most are articulated.

In 1995, additional sites were established off Corpus Christi Bay on the outer Texas shelf (Fig. 2). This area is the purported depocenter for suspended sediments originating from the Mississippi River funneled by alongshore flow down the Texas coast (Cochrane & Kelly 1986, Siringan & Anderson 1994) and is characterized by a well-developed nepheloid layer (Shideler 1981). Halper et al. (1988) and Sahl et al. (1997) further discuss the offshore circulation. Sediment shell content is reduced in this area in comparison with Galveston. Skeletal material is dominated by burrowing echinoid remains (Hill et al. 1982; our unpubl, data).

Six EODs were sited at the East Flower Garden (EFG) Bank. The EFG is a deep reef atop a salt diapir (Bright & Powell 1983, Gardner et al. 1998, Lugo-Fernandez 1998) (Fig. 2). Experiments were deployed on the deep reefal coralline-algal dominated hardground for comparison with the Caribbean sites, in an anoxic brine pool (200 [per thousand]) where lagerstatten are preserved (Rezak & Bright 1981, Powell et al. 1986), in a brine filled canyon downslope from the pool (Gittings et al. 1984), at the canyon mouth where dilution returns salinity to near normal (Powell et al. 1986), on the carbonate sand of the canyon fan downslope of the canyon mouth, and in the gravel-to-sandy carbonate sediment downslope of the hard bank (Powell et al. 1983).

Experiments were deployed at two very different petroleum seeps with lush chemoautotrophic-based communities on the continental slope. GB425 is a massive clam (lucinid and thyasirid) community (Fig. 2). GC234 is dominated by mussel beds and tubeworm thickets (Callender & Powell 1999, Callender & Powell 2000). GC234 and GB425 are in the geologically and bathymetrically complex Garden Banks and Green Canyon regions characterized by active salt diapirism, associated faulting (Bouma et al. 1980, Bouma et al. 1981, Behrens 1988), widespread active seepage of liquid and gaseous hydrocarbons (Kennicutt et al. 1988a, Kennicutt et al. 1988b), and locally high sedimentation rates (Roberts & Carney 1997, MacDonald et al. 2000). Authigenic carbonate is common at and below the sediment-water interface (Brooks et al. 1986, Behrens 1988, Roberts et al. 1987). The carbonate forms from microbial degradation of hydrocarbons that produces excess CO2 in the sediment porewater. More details can be found in Carney (1994), Callender and Powell (1997), MacAvoy et al. (2002), and Bergquist et al. (2003).

Parker Bank is one of many continental shelf-edge banks in the northwest Gulf of Mexico situated on top of salt diapirs (Rezak et al. 1990) (Fig. 2). The center of this bank has collapsed leaving a central basin filled with large carbonate blocks and fine ooze. Current flow is low to absent in the basin, however the basin is oxic. Experiments were deployed on the floor of this collapse basin. The sides and top of the collapsed dome support a thriving deep-water carbonate hard-bottom community substantially different from that present on the deep-water reef of the East Flower Garden. Experiments were deployed at two locations 10 m apart in depth on this carbonate rim.

Experimental Design

The SSETI experimental design is described in detail in Parsons et al. (1997) and only summarized briefly here. Each experimental array consisted of a series of l-cm-mesh bags attached to a 1.5-m PVC pole. To each pole was added a 12-kg weight to counter a 25-cm square float made of 6-ram-thick sheet polypropylene. The float rose about 0.5 m above the array and served to mark the location of the experiment even when buried. The PVC pole was unattached to the bottom, anchored only by the 12-kg weight, and so was free to move, given sufficient current and wave action. Arrays were moved at deployment depths of up to 80 m, but the polypropylene float permitted relocation in every case. In only one case were arrays moved more than a few meters, however. Arrays at the 15-m location of Bahamian transect BA were moved tens of meters downcoast between 1996 and 2001 as a result of hurricanes.

Two of the mesh bags on each pole contained molluscan shells, typically five individuals of five different species, each species compartmentalized from the others by plastic cable ties. Species were chosen based on microstructure, composition, and availability in quantity. Molluscan species deployed included the ocean quahog Arctica islandica, the mussel Mytilus edulis, the lucinid Codakia orbieularis, the venerid Mercenaria mercenaria, the glycymerid Glycymeris undata, the scallop Argopeeten irradians, the cerithiids Rhinoclavus vertagus and Telescopium telescopium, the conch Strombus luhuanus, and the turritellid, Turritella terebra. Because of shell availability for different deployment efforts, not all species were deployed at each site. Arrays were recovered on LSI transect AA in 1999 by scuba and the submersible Delta. Arrays were recovered on LSI transect BA in 2001 by scuba and the submersible Clelia. Arrays were recovered at Gulf of Mexico sites in 2001 using the submersible Johnson-Sea-Link. Parsons-Hubbard et al. (1999, 2001) analyzed the results of experiments recovered in 1995 and 1996 two years after deployment and characterized each shell deployment as exposed, dusted (by a fine layer of sediment), or buried at the time of collection. These observations are updated herein (Table 1). The number of individuals of each species recovered and analyzed for the 6-to-8-y recoveries is provided in Table 2. For simplicity, these will be combined under the appellation "8-y recoveries" hereafter. Details of site locations are tabulated in Callender et al. (2002) and Powell et al. (2002). Callender et al. (2002), Staff et al. (2002), and Callender et al. (2002) described the taphonomic signatures of deployed shells after one and two years. Walker et al. (1998) compared the taphonomic condition of gastropods deployed in the more-cloistered bag arrays with the more-exposed tethered shells.

Laboratory Analyses

Laboratory analyses of shell specimens commenced soon after recovery and were completed within 48 h. During this time, shells were maintained in chilled seawater, except when under active analysis. Each shell was photographed, measured, assessed for taphonomic alteration, air-dried, and weighed. As the shells were not oven-dried, we refer to the measured variable as a wet weight. The types of taphonomic alteration assessed included breakage, edge alteration, periostracum condition, color alteration, and evidence and severity of dissolution, abrasion, and authigenic precipitation. We recognize that our study does not discriminate chemical dissolution from maceration (Poulicek et al. 1981, Glover & Kidwell 1993, Behrensmeyer et al. 2000) and we use the term dissolution to refer to both processes. Each of these taphonomic processes was evaluated using a semiquantitative scale described by Davies et al. (1990) on each of eight standard shell areas for bivalves on the inner and outer surfaces (see Davies et al. 1990) and on five standard shell areas for gastropods somewhat modified from Davies et al. (1990), namely the spire aperture up, spire aperture down, body whorl aperture up, body whorl aperture down, and the inside aperture/columella complex. Kidwell et al. (2001) provide a critical review of this methodology and Rothfus (2004) considered investigator bias in data collection of this type.

Statistical Analyses

For statistical analysis, taphonomic attributes were treated in one of three ways. (1) For dissolution, each shell area was assigned a numerical value according to the degree of dissolution; in some cases this assignment also included amalgams of originally-discrete levels of dissolution defined by Davies et al. (1990) (Table 3). The average condition for a shell was taken as the weighted average of the values assigned to each shell area. Weights were proportional to the fraction of total shell surface area contributed by that shell area. For bivalves, a weighted average was calculated separately for the inner and outer shell surfaces. For gastropods, a weighted average was calculated separately for the spire and body whorl. The spire, like the outer surface of a bivalve, is exposed for a long period while the animal is alive. The body whorl, like the inner surface of a bivalve, is a relatively new surface for taphonomy, although not as new as the bivalve inner surface. (2) For the remaining taphonomic attributes, and as an alternate way to evaluate dissolution, the most extreme condition observed on any of the discrete shell areas was taken as the value for that shell (Table 3). This method of evaluation follows the approach used by Staff and Powell (1990), Callender and Powell (1992), and Callender et al. (1994) and was recommended by Kidwell et al. (2001). (3) The semiquantitative scales used for the previous two groups of taphonomic attributes permit statistical evaluation because they represent a hierarchical ranking of increasing taphonomic impact (Table 3). Some taphonomic attributes, including various types of shell discoloration and various styles of dissolution, cannot be apportioned hierarchically and so were analyzed as presence/absence attributes (Table 4).

Periostracum state was evaluated on a subset of four species with easily observed periostracum in fresh shells: Arctica islandica, Mytilus edulis, Telescopium telescopium, and Glycymeris undata. Shell wet weight, per se, is not a taphonomic attribute. The attribute of interest is the change in weight over time. Accordingly, prior to statistical analysis, a regression was established between length and wet weight using all specimens, control and experimental, for each species, following the convention of Callender et al. (1994), and the standardized residuals from the observed wet weights were calculated:

observed wet weight--expected wet weight/observed wet weight

These standardized residuals were used in subsequent analyses. The analysis was limited to bivalves, as many of the gastropods were partially filled with sediment after eight years on the seafloor and this infill could not be consistently removed.

The dataset was analyzed in two ways. The first focused on whether the individuals recovered in each array at each site were significantly different from control shells. We conducted Tukey Studentized Range tests comparing the 41 sites for each of nine taphonomic attributes and evaluated whether the number of significant results could be attributed to chance using binomial tests. We caution the reader that the mix of sites and taphonomic attributes tested to evaluate the difference between recovered shells and controls in this study diverges from the groups analyzed by Staff et al. (2002) and Powell et al. (2002). Consequently, no time course inferences concerning taphonomic rates should be made from comparisons between this analysis and analogous analyses from two-year recoveries, as summarized by Powell et al. (2002). We examined the distribution of significant comparisons by grouping sites by species, ocean (Caribbean or Gulf of Mexico), sediment type, degree of exposure, and depth. In each of the latter four, we summed over all species rather than testing species separately for each taphonomic trait at each site. The taphonomic traits included were: the maximum degree of dissolution, the average degree of dissolution on the outer shell surface/spire, the average degree of dissolution on the inner shell surface/body whorl, the maximum degree of abrasion, the maximum degree of edge rounding, the maximum degree of discoloration, the degree of breakage, the degree of periostracum loss, and the standardized residual of wet weight. For sediment type, we allocated sites to a series of sediment types, namely carbonate sand, terrigenous sand, hardground, veneered hardground (hardgrounds covered with a thin veneer of sand), brine, carbonate mud, and terrigenous mud. We converted depth into a polytomous variable, grouping sites <50 m, 50 to <100 m; 100 to <200 m, 200 to <300 m, and [greater than or equal to] 300 m. Degree of burial for each recovered array was assigned from video documentation to four exposure categories: exposed, dusted, partly buried, and buried following the convention of Parsons-Hubbard et al. (1999, 2001). An array with both shell bags completely buried was assigned to the "buried" category. The category "dusted" applied to arrays with shell bags covered by a fine layer of sediment through which the bags could still be discerned. In addition, for some arrays, bags were partially buried: either one bag was buried and the other was not, or a portion of a bag was buried and the remainder exposed. Arrays of this type were assigned to the "'partially buried" category.

The second approach to analysis focused on the comparison of sites, sometimes grouped according to EOD, depth, sediment type, or state of burial. Routine statistical analysis becomes problematic, however, when taphonomic attributes are correlated directly or inversely. An inverse correlation between dissolution and abrasion was observed by Callender et al. (2002) and Powell et al. (2002), for example. Consequently, principal components analysis (PCA) was used to obtain unique taphonomic descriptors from combinations of taphonomic attributes. One PCA was conducted for all species using all taphonomic variables in Tables 3 and 4, except for periostracum status. This latter variable was not included because it could not be evaluated in species having hard-to-observe periostracum (e.g., Codakia orbicularis). A second PCA was conducted solely for bivalves. This PCA included the residual of wet weight. Each variable was standardized to a mean of zero and a standard deviation of 1 prior to the PCA. All species were included in the PCA.

PCA factor scores were tested statistically with two ANOVA models. The first used an ANCOVA analysis. Main effects were EOD and species. Depth was included as the covariate rather than a polytomous main effect because many potentially important variables such as light, food supply, and benthic secondary production are expected to vary more or less monotonically with depth (e.g., El-Sayed 1972, Rowe 1983, Rowe et al. 1990). The second ANOVA model addressed the influence of sediment type. Main effects were sediment type, degree of exposure, and species. All ANOVAs also included all pair-wise interaction terms. A posteriori Tukey Studentized Range Tests were used to identify sources of significance within the ANOVA, however we caution the reader that significant interaction terms where they occurred limit the interpretation of the results of Tukey Studentized Range tests.

Preliminary analyses for the primary taphonomic attributes identified in Table 3 designed to detect significant differences between bags when a species was deployed in both bags on the same array demonstrated significant differences infrequently (9.8%), but significantly more often than expected by chance (binomial test, p = 0.05, q = 0.95, n = 1476, P < 0.0001). Many species were present only in one of the two bags on a given array, however, obviating a nested analysis of variance that otherwise might be used to accommodate this variability (Hurlbert 1984). As a consequence, when a species was present in both bags on an array, these data were combined for subsequent analyses. The reader is cautioned that some variability in these cases originated from bag placement and this variability has not been treated independently in these analyses.

In some cases, the collected arrays were grouped into 18 environments of deposition (EODs) (listed in the succeeding as "EOD abbreviation--number of arrays"). The 10 Gulf of Mexico EODs providing the eight-year recovery data will be referred to as East Flower Garden (EFG) deep-reef hardground (EFGHD-2), EFG brine pool (EFGBR-2), EFG brine canyon (EFGCS-1), EFG canyon mouth (EFGCM-1), EFG canyon talus slope (EFGCF-1), EFG deep carbonate sand (EFGDS-1), petroleum seep GB425 (GB425-4), petroleum seep GC234 (GC234-3), south Texas outer continental shelf (STXSH-2), and Parker Bank carbonate rim (PARKR-2). The petroleum seep at GC234 is characterized by significant taphonomic variability between deployment sites (Cai et al. 2006). Nevertheless, this complex taphonomic environment has been retained as a single EOD. The nine Bahamian EODs will be referred to as Norman's Pond (NORMP-1), platform top (15M-3), platform edge (30M-4), wall (WALL-4), upper talus slope (183M-1), lower talus slope (TALUS-3), dune crest (CREST-3), and dune trough (TROFF-3). In some cases, the AA and BA transects at the Lee Stocking Island sites have been distinguished by adding the prefixes "AA" and "BA" to the respective EOD designation. Some EODs were represented by multiple array collections (e.g., GB425), whereas others were represented by a single array collection (e.g., EFGCS), thus obviating a nested analysis of variance that otherwise might be used to accommodate this variability. The reader is cautioned that some variability in EOD comparisons may have originated from array placement and this variability has not been treated independently in these analyses.

RESULTS

Average taphonomic condition for these EODs for each of the taphonomic attributes listed in Tables 3 and 4 and for each of the 10 deployed species are provided in Figures 3-17. Representative Gulf of Mexico and Caribbean shells are pictured in Plates 1 and 2.

[FIGURE 3 OMITTED]

Comparative Statistics--Eight-year Recoveries vs. Controls

The status of shells, irrespective of species, recovered from the 41 sites after eight years was compared with that of the original controls with respect to each of the primary taphonomic attributes. The number of significant differences did not vary much between the Gulf of Mexico (26% of site-by-taphonomic trait comparisons) and Caribbean (23%) sites (Fig. 18). The frequency of significant differences between eight-year recoveries and controls for sites in the Gulf of Mexico was not significantly different from that expected from the equivalent frequency among Caribbean sites.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Shells recovered from certain depth zones differed more frequently from controls than shells from other depth zones (Fig. 19). For depth zones, the number of significant differences tallied over all taphonomic traits exceeded the number anticipated by chance (binomial test, p = 0.05, q = 0.95, P < 0.05). Of particular interest were the relatively few differences between shells deployed at depths of 50-100 m and controls: these differences occurred much less frequently than expected from the frequency of occurrence of significant differences between shells deployed at depths of 100-200 m and controls (binomial test, P < 0.005), as well as at depths >300 m versus the controls (binomial test, P < 0.0001), the two highest frequencies observed. In fact, significant differences between shells deployed at 50 100 m and the controls occurred less frequently even from that expected from the intermediate frequencies of significance for shells deployed at depths <50 m (P < 0.05) or 200-300 m (P < 0.05) versus the controls. Moreover, the frequency of occurrence of significant differences between shells recovered after eight years and controls at depths of <50 m and 200-300 m was significantly lower than the highest frequency observed for shells deployed at >300 m (binomial test, P < 0.01).

[FIGURE 6 OMITTED]

Interestingly, the number of significant deviations from controls was fewest for shells deployed on hardgrounds, veneered hardgrounds, and carbonate sands. Shells deployed in brine-exposed sites, on terrigenous sands and muds, and on carbonate mud were most likely to differ significantly from controls after eight years (Fig. 20). The difference between the three highest frequencies, brine-exposed sites, terrigenous sand, and terrigenous mud, and the three lowest frequencies, hardgrounds, veneered hardgrounds, and carbonate sands, was significant (P < 0.0005). At an intermediate position was carbonate muds. Thus, two groups of sediment types exist as measured by the frequency at which shells deployed in them for eight years deviated from controls; one group in which deployed shells frequently deviated in taphonomic condition from controls, the brine-exposed sites and terrigenous sediments, and the second group, the carbonate habitats, in which deployed shells much less frequently deviated from controls.

[FIGURE 7 OMITTED]

Some species differed from controls more often than others (Fig. 21). Because all species were not deployed in all EODs and the comparisons were made between sites for each taphonomic trait by species, some portion of these differences may accrue from the suite of EODs at which these species were deployed. Nevertheless, Mytilus edulis (26% of site-by-taphonomic trait comparisons) and Telescopium telescopium (28%) deployed for eight years differed much more often from their controls than did the remaining species, most of which differed significantly in 11% to 15% of the sites (Fig. 21). This difference between M. edulis and T. telescopium and the remaining species was significant (binomial test, P < 0.01). Of particular note are the frequencies for M. edulis, Strombus luhuanus, and Codakia orbicularis. These species were deployed at nearly all EODs (Table 2) and therefore are fully comparable. The frequency of occurrence of significant differences between controls and eight-year deployments for M. edulis (26%) far exceeds that for the other two species, C. orbicularis (16%) and S. luhuanus (14.3%).

Shells deployed for eight years were significantly more discolored than the controls in 87.8% of the sites (Fig. 22). Shells differed relatively frequently from controls after eight years on the seafloor in dissolution intensity. The inner shell surface differed somewhat more frequently than the outer shell surface (31.7% versus 24.4%); however, the deviation was not significant (P > 0.05). On the other side of the spectrum, shells recovered after eight years differed significantly from controls at few sites in the amount of breakage (2.5%), the degree of edge rounding (2.4%), or the degree of abrasion present (12.2%). The residual of wet weight did not differ significantly from controls at any site after eight years (Fig. 22). Of the 12% of the sites with shells that differed significantly in the degree of abrasion from controls, nearly all differed by having an abrasion index lower than the controls. This bias is unexpected (P < 0.05).

Comparative Statistics--Within Species

Three species were deployed at all or nearly all of the SSETI EODs, Mytilus edulis. Codakia orbicularis, and Strombus luhuanus. Trends for these three species confirm that the general trends for all species were not influenced by the deployment of some species in a subset of the EODs. Discoloration was by far the most common taphonomic effect. Shells differed from controls in 79%, 90%, and 62.5% of the cases for M. edulis, C. orbicularis, and S. luhuanus, respectively. The penultimate taphonomic process was dissolution. However, in two of the three species, M. edulis and S. luhuanus, the longer-exposed portion of the shell, the outer surface of bivalves and spire of gastropods, differed more frequently from controls than did the inner shell surface or body whorl, in contrast to the trend for all species. In only one case was this difference significant however, for M. edulis (P < 0.005). In this case, the condition of the outer shell surface differed from controls significantly more frequently. The percentages are, for M. edulis, 20.8% and 54.2% for the inner and outer shell respectively. The equivalent percentages are 9.5% and 14.3% for C. orbicularis and 25% (body whorl) and 29.2% (spire) for S. luhuanus.

Deployed shells differed rarely from controls in the degree of breakage and edge rounding: 10.4% for Mytilus edulis, 0% for Codakia orbicularis, and 2.1% for Strombus luhuanus. The degree of breakage and edge rounding after 8 y diverged significantly from controls for C. orbicularis significantly less often than for M. edulis (P < 0.02). No significant differences were found for abrasion or the residual of wet weight.

[FIGURE 8 OMITTED]

Principal Components Analysis--Coherent Taphonomic Processes

PCA revealed eight significant factors. Factor 1 was defined by the maximum degree of dissolution (Fig. 9), the incidence of deep dissolution (Fig. 13, Plate 2F), and the average degree of dissolution on the outer surface of bivalves and spire of gastropods (Fig. 8). Factor 1 was, therefore, a general dissolution factor describing the surface of the shell exposed for the longest period of time in life. Factor 2 was defined by the incidence of brown discoloration (Fig. 16), the maximum degree of discoloration (Fig. 7) and, as the negative, the incidence of original coloration (Plate 2A, 2D). Factor 2 was, therefore, a general discoloration factor. Hence, dissolution and discoloration proceeded along decidedly different trajectories.

Factor 3 was defined by the incidence of a soft shell surface (Fig. 12, Plate 2E), the average degree of dissolution on the inner shell surface or body whorl (Fig. 8), the incidence of deep dissolution resulting in holes forming in the shell, and the incidence of chalkiness (Fig. 11, Plate 2B, 2C). Factor 3 was, therefore, a second dissolution factor that described a particular type of dissolution, the development of a chalky surface and its ultimate evolution into a soft shell surface, and this process was observed principally on the inner shell surface of bivalves and body whorl of gastropods. Factor 4 was defined by the incidence of faded coloration, and the incidence of brown discoloration (Fig. 16), as the negative. Factor 4, therefore, described a specific style of discoloration that produced a faded shell surface without the appearance of altered coloration.

Factor 5 was defined by the degree of edge rounding and breakage (Figs. 3 and 5). Factor 6 was defined by the incidence of pitting (Fig. 10) and, interestingly, the incidence of orange/ orange-mottled discoloration (Fig. 15, Plate 1 C, 1F). This is the only factor to combine a dissolution process with a discoloration process. Factor 7 was defined by the incidence of green discoloration, likely caused by boring green algae (Fig. 14, Plate ID, IE). Factor 8 was defined by the incidence of gray-to-black discoloration (Fig. 17, Plate 1A, 1B).

[FIGURE 9 OMITTED]

Abrasion (Fig. 4) was the only important taphonomic process not resolved by these eight factors. When just bivalves are considered, and the residual of wet weight added to the PCA, the first eight factors remain very similar and the residual of wet weight also was not resolved by them. The single exception is that orange discoloration fell onto two PCA axes rather than one. The first fell on an axis with the incidence of pitting as was observed for the all-molluscan PCA. The second fell on an axis with gray-to-black discoloration. The latter relationship is resolved in the smaller dataset because of the proportionally greater contribution of Mercenaria mercenaria, which often expressed a reddish coloration caused by the oxidation of pyrite present within the shell when deployed (Callender et al. 2002). As a consequence of the similarity between the two PCAs, further PCA analysis is concentrated on the full PCA including bivalves and gastropods and lacking the residual of wet weight. The residual of wet weight is considered as an additional variable in all subsequent analyses. Abrasion is not further analyzed.

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

In summary, of the eight factors, one defines the process of fragmentation and breakage (Factor 5), two define a suite of dissolution processes (Factors 1 and 3), four apply to various types of discoloration (Factors 2, 4, 7, and 8), and one confounds dissolution with discoloration (Factor 6). The dissolution factors resolve pitting from chalkiness and taphonomic attack on the inner surface/body whorl from that of the outer surface/ spire. The discoloration factors resolve a series of likely independent discoloring processes modifying shell appearance: fading, brown-to-red discoloration, green discoloration, gray-to-black discoloration, and orange/orange-mottled discoloration.

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

ANOVA Analysis: EOD and Depth

EODs were first defined as described in Powell et al. (2002) based on visually obvious similarities between sites of shell deployment; all wall sites, for example, on the two Bahamian transects were combined as a "Wall" EOD and all 15-m Bahamian sites were combined as a 15-m EOD. In both of these examples, the appearance of all sites subsumed into these EODs by visual inspection was similar. Wall sites were all ledges of exposed hardground at similar depths on what was otherwise a near vertical drop-off. All sites at 15 m were characterized by patch reefs surrounded by veneered hardgrounds, the sediment cover consisting of coarse carbonate sand. Example ANOVAs for Factors 1 4 are shown in Table 5. These statistical results are characterized by the common occurrence of significant interaction terms between EOD, species, and depth, the frequent occurrence of significant main effects for EOD and species, and the rarity of a significant main effect for depth.

[FIGURE 14 OMITTED]

[FIGURE 15 OMITTED]

However, Staff et al. (2002) identified significant differences in taphonomic signature in shells deployed at visually very similar EODs on the two Bahamian transects AA and BA after two years. Staff et al. (2002) recognized the visual similarity of these sites, but emphasized the differential trends in shell condition between the two transects that support the presence of a different preservational regimen. Staff et al. (2002) suggested that downslope transport may be more important on transect AA where the influence of tidal flow into and out of the nearby inlet, Adderly Cut, is significant. Stalked crinoids are more abundant downslope of the wall on transect AA, for instance, suggesting a greater flux of food in this region. As a consequence of the differences observed in shell condition after two years on the seafloor by Staff et al. (2002), these seemingly similar EODs on these two transects were separated and the ANOVA analyses rerun. The results for Factors 1M are displayed in Tables 6-9. Note that the separation of these seemingly similar EODs eliminates all significant interaction terms and species main effects; that is, differences in the distribution of species deployed between the AA and BA transects along the depth gradient and the inherent differences between the two transects were manifested statistically in the earlier analyses (Table 5) in part as spurious significances in species main effects and in part as species*EOD interactions. In accordance with the evidence that these two transects encompass taphonomically distinctive EODs, though not visually evident as such, we have conducted all subsequent statistical analyses on this larger set of discrete environments, termed environments of preservation (EOP) hereafter.

[FIGURE 16 OMITTED]

[FIGURE 17 OMITTED]

ANOVAs run using the set of 24 EOPs revealed that depth was not a significant covariate for any of the eight PCA factors or for the residual of wet weight (Tables 6-12). Taphonomic processes were not significantly influenced by depth. Species, as a main effect, was significant in only one case, Factor 8 (Table 11). Mercenaria mercenaria and Rhinoclavus vertagus were discolored gray-to-black more frequently than other deployed species. Only two interaction terms were significant, the interactions between depth and EOP for the residual of wet weight (ranked ANOVA) (Table 12). The frequency of significant interaction terms, 2 in 36 such comparisons, is no more than expected by chance (binomial test, p = 0.027, q = 0.973, P > 0.05). In contrast, EOP was significant as a main effect for four of eight PCA factors, Factors 2 (Table 7), 3 (Table 8), 7 (Table 10), and 8 (Table 11). Three of these four factors described different styles of discoloration. Thus discoloration style was strongly influenced by EOP.

The maximum degree of discoloration and the incidence of brown discoloration were significantly influenced by EOP (Factor 2--Figures 7, 16; Table 7). A few EOPs, specifically the brine pool and brine stream at the East Flower Garden and the 30-m site on Bahamian transect AA, were discolored less often than other EOPs (Fig. 7, Plate 1A, 1B). Chalkiness (Fig. 11) and its more extreme form, a soft shell surface (Fig. 12), were also significantly affected by EOP (Factor 8--Table 8). Three EOPs were most heavily impacted by processes leading to a chalky surface, the brine stream canyon at the East Flower Garden (Plate 2C, 2E), the petroleum seep at GC234 (Plate 2F), and a group of sites on Bahamian transect AA (Fig. 11). A number of other East Flower Garden sites downstream of the brine seep, the other petroleum seep site at GB425 (Plate 2B), and several other sites on the AA transect also had relatively high incidences of this type of taphonomic attack (Fig. 11).

Factor 7 was also significantly influenced by EOP (Table 10). Factor 7 is principally determined by the incidence of green discoloration (Fig. 14, Plate 1B, 1D). Usually, this coloration is produced by boring green algae and, not surprisingly, the sites with highest Factor-7 scores were nearly all sites in the photic zone characterized by hardground sediment types, where burial was limited (Fig. 14). Gray-to-black discoloration determined Factor 8 and this factor was also significantly influenced by EOP (Table 11). The most discolored sites were the deeper sites on the AA transect and a large fraction of all BA-transect EOPs (Fig. 17; Plate 1B, 1D).

Finally, the residual of wet weight was not significantly affected by EOP (Table 12), but two interaction terms with depth were significant. Shells relatively heavier than the controls were from dominantly deeper-water carbonate sites in the Bahamas, with both transects represented. Shells less heavy than the controls were from petroleum seeps, brine-exposed sites, and shallow-water Bahamian sites. The order of sites by the change in wet weight of the shells deployed there (Table 12) did not match well with the order of the sites by the degree of dissolution on the outer shell surface (Factor 1, Table 6) or with the order of the sites by the degree of dissolution on the inner shell surface (Factor 3, Table 8).

[ILLUSTRATION OMITTED]

ANOVA Analysis: EOP and Depth

Deployment sites were classified into a series of sediment types and the degree of burial was assessed visually at the time of recovery (Table 2). The taphonomic process is thought to be related in some way to the degree of exposure of shells on the bottom, as expressed by the concept of the TAZ [taphonomically-active zone, Davies et al. (1989)] (Walker & Goldstein 1999, Walker 2001, Barbieri 2001), and the sedimentary environment should influence sediment chemistry. Accordingly, taphonomic signature should be significantly influenced by sediment type and degree of exposure.

[ILLUSTRATION OMITTED]

Results of analyses of Factors 1-4 are shown in Table 13. The significance of species and interaction terms with species stems to some extent from the unequal distribution of species among sediment types and degrees of exposure (Table 1). When the ANOVA was restricted to Mytilus edulis and Strombus luhuanus, the two species present in every sediment type and degree of exposure, the results were moderately improved in that interaction terms were much less frequently significant and species main effects were not significant in every case, but qualitatively the trends were the same. The main effects of sediment type and degree of exposure are routinely highly significant as are many interaction terms between them regardless of whether all species or the latter two are used (Table 13). The remaining four factors did not diverge in any meaningful way from this pattern (Table 13). Thus, the analyses suggest that sediment type and degree of exposure are routinely highly significant determinants of factor scores for all eight factors and that interactions between these two main effects and between them and species, being frequently significant, record a complex interplay between these three main effects in determining taphonomic signature.

[FIGURE 18 OMITTED]

However, the complexity of these ANOVA results contrasts starkly with the simplicity of most ANOVA results in which location (EOP) was used as a main effect and generates suspicion that the main effects of sediment type and degree of exposure are poor descriptors of the taphonomic environment. That is, EOPs of disparate taphonomic signature are allocated under the same sediment-type and degree-of-exposure categories. The corollary suspicion is that environments of preservation (EOPs) are not well distinguished by visual distinctions in environments of deposition (EODs).

[FIGURE 19 OMITTED]

EOPs in Factor Space

Figure 23 positions EOPs with respect to the two primary dissolution factors, Factors 1 and 3, and the residual of wet weight. Factor 1 and the residual of wet weight are positioned on the y and x axes, respectively, and Factor 3 is depicted in "bubble" form. We caution the reader that, in these and subsequent figures, the extremes in x and y are as important as the extremes in bubble size. Readers should focus on the location of the sites in x-y space as diligently as the visually more distinctive third dimension expressed in bubble form.

Figure 23 identifies a group of EOPs with high residuals of wet weight and extremely negative Factor-1 scores. Most have positive Factor-3 scores, although not particularly high. These sites encompass essentially all of the Bahamian transect-AA EOPs. These shells are uniformly characterized by lower than average dissolution indices on the shell's outer surface (spire), and a gradation towards higher than anticipated wet weights, and moderately-high degrees of chalkiness on the inner shell surface/body whorl. At the upper center of Figure 23 are EOPs with unusually high dissolution indices on the outer shell surface (spire). Several of these also have high dissolution indices on the inner shell surface (body whorl) documented by high Factor-3 scores. These sites include the EFG brine stream and the petroleum seep at GC234. Both EOPs are characterized by high sulfide concentrations. Also shown on Figure 23 are a group of EOPs with low Factor-3 scores. The majority of these have moderate to high Factor-1 scores. These sites include the carbonate rim of Parker Bank and a number of shallow sites (30 m, 183 m, and wall) on Bahamian transect BA. These sites are characterized by relatively little dissolution on the inner surface/body whorl and moderate-to-high dissolution on the outer surface/spire. Finally, two sites are interesting in having relatively negative residuals of wet weight, but low dissolution scores on the outer shell surface: the 15-m sites on Bahamian transect BA and the brine pool at the East Flower Garden. Figure 23 clearly establishes the taphonomic distinctiveness between transect-AA and transect-BA EOPs.

[FIGURE 20 OMITTED]

Figure 24 positions EOPs with respect to the three primary discoloration factors. Two sites have been affected little by discoloration as shown by low Factor-2 scores and negative Factor-7 and Factor-8 scores. These are the 30-m site on Bahamian transect AA and the East Flower Garden brine pool. Arrays deployed at the former EOP were buried soon after deployment and arrays deployed at the latter are in 200 [per thousand] brine (Brooks et al. 1979). A series of EOPs are characterized by highly discolored shells as indicated by high Factor-2 scores. Many of these also have high Factor-8 scores, indicating that discoloration is gray-to-black as well as brown. One of the most discolored of these sites, the 15-m EOP on Bahamian transect BA, also ranks high in green discoloration. The remainder include the deepest transect-AA sites. These are discolored brown to gray-to-black, but are too deep for green algae to have added that distinctive coloration. These sites are relatively uninfluenced by dissolution and confirm a fundamental difference between discoloration and dissolution earlier identified by Staff et al. (2002) and Powell et al. (2002). The deeper BA-transect sites are also high in gray-to-black discoloration, suggesting that this is a deepwater phenomenon off LSI. Similar discoloration has occurred as well at the nearby saline lagoon Norman's Pond. A few East Flower Garden sites yielded shells that were also discolored brown, the canyon mouth and canyon fan sites.

[FIGURE 21 OMITTED]

[FIGURE 22 OMITTED]

All EOPs with shells having high Factor-2 and/or Factor-8 scores are carbonate sites, so that the most heavily discolored shells are essentially all deployed in carbonate EOPs. Green discoloration dominates at the shallow-water carbonate EOPs at LSI and at Parker Bank. These are photic sites, mostly hardgrounds, where boring green algae have invaded the shells. Again, however, the extremes in discoloration are nearly all carbonate EOPs.

Figure 25 positions EOPs with respect to Factors 4, 5, and 6: fading, edge rounding, and pitting/orange discoloration. Few EOPs attain extreme factor scores. The East Flower Garden brine canyon is unique in having an unusually low incidence of pitting and orange discoloration. These shells and those from the 15-m and 30-m EOP on transect BA, along with shells from the East Flower Garden brine lake, are faded more than most. Among these are EOPs that show the greatest and the least degrees of dissolution, confirming that colored discoloration and the development of a chalky surface are not incompatible, nor do they consistently co-occur. Edge rounding was relatively high at two East Flower Garden sites, the canyon talus fan and the deeper carbonate-sand apron, and at the two shallowest Bahamian sites, the 15-m sites on transects BA and AA. The talus-fan EOP was also unusual in having shells inordinately affected by pitting and orange discoloration. The latter was also an attribute of the East Flower Garden canyon mouth, suggesting that this unique shell discoloration and surface deterioration were associated with oxic, sulfidic water of near-normal salinity (Powell et al. 1983, Powell et al. 1986).

Sediment Types and Degrees of Exposure in Factor Space

Two sediment type-exposure pairs have unusually low discoloration indices, as evidenced by negative Factor-2 scores (Fig. 26). These are the East Flower Garden brine pool (BB in Figure 26) and the arrays buried in terrigenous sand (TB in Figure 26). These latter are the south Texas shelf sites. Two sediment type-exposure pairs are most discolored: arrays buried in carbonate mud (PB) at the Norman's Pond site and arrays exposed in terrigenous mud (ME) at petroleum seeps. Of these two, the former is characterized by gray-to-black discoloring, as shown by high Factor-8 scores (Fig. 26), but both sites tend to be relatively low in green discoloration, as shown by low Factor-7 scores. A number of other sediment type-exposure pairs are characterized by high Factor-8 and low Factor-7 scores. These include exposed arrays (CE) and partly buried arrays (CP) in carbonate sand. Shells discolored green include partially buried arrays deployed in veneered hardgrounds (SP), dusted arrays in carbonate sands (CD) and to a lesser extent, exposed arrays in veneered hardgrounds (SE). All of these are carbonate sites, probably because most terrigenous sites are not in the photic zone. More interesting is the tendency for gray-to-black discoloration to be dominated by carbonate sites and the tendency for most terrigenous sites to have low discoloration scores.

No sediment type-exposure pairs were characterized by unusually low or high dissolution indices on the outer surface/ spire (Fig. 27), with the exception of Norman's Pond (PD). The inner surface/body whorl was much more variable in its dissolution index. Exposed arrays from carbonate sand (CE) and exposed and buried arrays from terrigenous mud (MB, ME) were more dissolved than most (Fig. 27). Partly buried arrays from terrigenous mud (MP), though less so, were distinctly dissimilar from the remaining sediment type-exposure pairs. Some sediment type-exposure pairs were unusual in having low dissolution indices on the inner shell/body whorl surface. These include buried arrays in Norman's Pond (PB), partly buried arrays from veneered hardgrounds (SP), partly buried arrays from carbonate sands (CP), and exposed arrays from veneered hardgrounds (SE). These are all carbonate sediments, notably unusual in including no hardgrounds, but encompassing a range of burial states. Sediment type-exposure pairs uniquely identified by pitting include the exposed terrigenous mud arrays (ME) and buried carbonate mud arrays (PB) (Fig. 27). Pitting was unusually rare from partly buried arrays recovered from veneered hardgrounds (SP).

Four sediment type-exposure pairs were characterized by unusually large gains in wet weight (Fig. 28), shells buried in terrigenous sand, shells exposed and dusted on carbonate sands, and dusted shells from veneered hardgrounds. These sediment type-exposure pairs covered a range of discoloration and dissolution factor scores (Figs. 26-27). Faded shells were unusually uncommon on arrays from exposed terrigenous mud sites (ME), but this was because these shells were discolored in other ways (Fig. 26). Faded shells were unusually common at brine-exposed sites and from the buried shells in the carbonate mud of Norman's Pond (BB, PB). These latter shells were also discolored gray-to-black (Fig. 26), whereas shells from brine sites were otherwise unaffected by discoloration (Plate 2D). Breakage and fragmentation were generally unimportant as observed from the nearly uniformly low Factor-5 scores.

DISCUSSION

Dominant Taphonomic Processes

Shells of bivalves and gastropods were deployed and recovered six-to-eight years later at a total of 41 locations in the Bahamas off Lee Stocking Island and throughout the western Gulf of Mexico continental shelf and upper slope (Figs. 1 and 2). These sites were assigned to 24 EOPs. Ten different species were deployed, although not at every site. Across all sites and species, the dominant taphonomic process was discoloration, where over 80% of species-site combinations differed from controls after eight years on the seafloor. Dissolution was clearly the process of penultimate importance, but species-site pairs differed in only 20% to 30% of cases (Fig. 22). Thus, the process of dissolution, though important, did not have the ubiquity or cumulative intensity of discoloration after eight years. Other taphonomic processes were distinctly less important; these included abrasion, edge rounding, and breakage. A critical deduction is that, were it not for discoloration, the vast majority of species deployed at the vast majority of sites would still not have differed significantly in taphonomic signature from controls after eight years on the seafloor.

Small shells are known to rapidly succumb to taphonomic attack in many habitats, leading to a bias in size frequency in most and to a bias in species preservation in some (Powell et al. 1984, Cummins et al. 1986, Glover & Kidwell 1993, Kidwell & Flessa 1995, Callender & Powell 2000, Kidwell, 2001). SSETI deployed adults of all species and these larger shells, by their very size, are more resistant to taphonomic degradation. For bivalves, we examined the residuals of wet weight to detect anticipated loss of shell weight over time (e.g., Lutz et al. 1985, Lutz et al. 1988, Callender et al. 1994). In many Bahamian sites, the residuals averaged above the controls. In a relatively few sites from shallow-water carbonates, petroleum seeps, and brine-exposed sites, the residuals averaged negative. However, never did the residuals diverge sufficiently to differ significantly from controls. The weight data conforms to the trends shown by dissolution indices that the loss of shell integrity was proceeding at a slow rate at the vast majority of these EOPs. Cutler (1995) found little weight loss in most habitats in deployed shells and Callender et al. (1994) observed weight loss only in deployed Mytilus edulis. Neither experiment encompassed the time period of SSETI deployments, but the inference from SSETI deployments that significant weight loss of large shells in most habitats, if it occurs, takes a very long time is consistent with most experimental deployments in which loss of shell integrity is not dominated by large common boring organisms such as Cliona sponges (e.g., Warburton 1958, Farrow & Fyfe 1988). Even M. edulis failed to diverge significantly from controls in any EOP, although this species was decidedly more altered at most sites than other species (see also Callender et al. 1994, Zuschin et al. 2003).

The process of discoloration can occur rapidly (Pilkey et al. 1969a, Pilkey et al. 1969b, Cutler 1995, Walker 2001, Parsons-Hubbard 2005). In our experimental deployments, in contrast to dissolution, shells discolored rapidly in nearly all EOPs. At the other end of the spectrum was edge rounding and breakage that proceeded at a very slow rate. The emplacement of shells in large mesh bags to permit deployment and recovery is probably responsible for some portion of this slowness. Breakage due to predation of hermitized gastropods is certainly higher at these SSETI sites in tethered, rather than enclosed, gastropods (Walker et al. 2002).

Kidwell et al. (2001) discussed the benefits of focusing on the inner shell surface of bivalves to assess taphonomic signature as this surface is unambiguously degraded after death (e.g., Cutler 1995), save perhaps for a light chalky surface characteristic of metabolic anoxia (Crenshaw & Neff 1969, Wilkes & Crenshaw 1979). By comparison, the outer shell surface can degrade significantly during life (Lescinsky 1993, Mao Che et al. 1996, Kaehler 1999), unless protected by a thick periostracal coat (Smyth 1989, Kaehler 1999). Shells deployed in 1993 1995 by SSETI were selected to have relatively little degradation of the outer shell surface, thereby minimizing the differential between these two surfaces. Nevertheless, the maximum dissolution intensity was normally observed on the outer shell surface, in conformance with the expectation from Kidwell et al. (2001) and Cutler (1995). Interestingly, the inner shell surface degraded more rapidly than the outer shell surface during the eight years of deployment (Fig. 22), as assessed by the average dissolution index. This difference was not significant however, when compared with the respective conditions of the controls. Although no definitive conclusions can be drawn from these eight-year recoveries, the data are not consistent with the suggestion that the longer exposure period that included life for the outer shell surface permitted an accelerated surface degradation after death. However, our data are in agreement with the inference by Kidwell et al. (2001) that the inner and outer shell surface should inherently diverge in their time course of shell deterioration and that the two surfaces should be considered separately in analyses of the preservational process. This was particularly true in the comparison between EOPs for which rarely did an EOP with a high dissolution index on the inner shell/body whorl also have a high index on the outer shell surface/spire. Tables 6 and 8 show a remarkable inconsistency in the order of the EOPs by these two criteria.

[FIGURE 23 OMITTED]

Staff et al. (2002) suggested that dissolution was overprinting the evidence of abrasive wear that was present on the shells when deployed. Evidence for this process was the inverse correlation between dissolution and abrasion that was revealed statistically in that study. Overprinting is likely a common process when shells are moved from one taphonomic environment to another (e.g., Cutler 1995). After eight years, an inverse correlation between abrasion and dissolution was no longer apparent. Overprinting was likely a function of both increased dissolution and increased discoloration and these two processes, being unrelated, forestalled any simple correlation with abrasion after eight years, in addition, the overprinting process was likely complete in many EOPs so that any remaining evidence of abrasion may have been produced in situ.

[FIGURE 24 OMITTED]

Species Differences

Three species were deployed at nearly every EOP: Mytilus edulis, Codakia orbicularis, and Strombus luhuanus. Mytilus edulis was clearly more taphonomically altered on the average than S. luhuanus or C. orbicularis. Discoloration was relatively similar, but M. edulis was much more heavily dissolved and degradation by breakage and edge rounding occurred more frequently. Codakia orbicularis and S. luhuanus were similarly and little influenced, in comparison. In fact, the differential between deployed Mytilus edulis and control M. edulis was greater than for all other species except Telescopium telescopium after eight years. After eight years, deterioration was remarkably consistent across most taxa. Much of the explanation for this similarity was the ubiquity and intensity of discoloration, which occurred relatively equivalently on all species and throughout most EOPs. However, dissolution indices increased as well across all species so that taphonomic attack in general was not strongly biased towards any particular taxon, exclusive of M. edulis.

[FIGURE 25 OMITTED]

Callender et al. (1994) remarked on the more rapid deterioration of mussel shells in comparison with other species after four years of deployment in a petroleum seep EOP. Zuschin et al. (2003) also considered the greater rate of taphonomic degradation of mussels relative to many other molluscs. This tendency for mussels to deteriorate more rapidly was already apparent in SSETI deployments after two years (Staff et al. 2002).

The Conundrum of Shell Weight

We expect shells to lose weight over time as taphonomic processes take their toll. Dissolution and maceration should slowly reduce total carbonate mass and the activities of boring bionts should also reduce carbonate mass. Both processes can produce substantive changes in shell weight over relatively short periods of time (e.g., McNichol et al. 1988, Powell et al. 1989, Parsons & Brett 1991, Kleeman 1996, Sartoretto 1998). Trends in weight could be analyzed only in bivalves. Surprisingly little evidence of weight loss was present in any bivalve species after eight years on the seafloor. The processes of shell deterioration, chemical and biological, were active in most EOPs and significant in some; nevertheless their cumulative impact was insufficient to produce a significant loss in weight in any EOP. Some portion of this failure was the counterweighing influence of biont attachment. Over eight years, in many EOPs, the accumulation of skeletal bionts has added carbonate mass to these shells. This process was already observable after one year (Walker et al. 1998). However, in the main, the added weight of bionts has not been great and, as a consequence, bivalve shell weight has remained relatively unchanged over the eight-yr period.

[FIGURE 26 OMITTED]

A potentially interesting trend exists to the generally limited change in shell weight, however. Shells of most bivalve species deployed in most deeper-water Bahamian EOPs averaged higher in weight than controls, though not significantly so. This tendency was also noted after two years (Callender et al. 2002, Staff et al. 2002). Authigenic carbonate precipitation and carbonate recrystallization in supersaturated sediments are well described (Morse et al. 1992, Tribble 1993, Callender et al. 1994; Stentoft 1994, Al-Agha et al. 1995). Moir (1990) describes a particularly enticing scenario from observations of fresh and "seawater-aged" Tridacna from tropical seas in which seawater aging rapidly increased shell hardness. Alexandersson (1972) comments on the rapidity at which carbonate infilling can occur on the Bahamian platform. Other authors have remarked that the formation of carbonate precipitates on shells and recrystallization can occur relatively rapidly (Cutler 1995, Kiene et al. 1995). Regardless, in these EOPs, eight years is insufficient time to unequivocally identify differences in weight or to assign a unique taphonomic signature to such trends.

[FIGURE 27 OMITTED]

Concordance in the Taphonomic Process

Principal components analysis demonstrated limited cooccurrence of discrete taphonomic processes among the 24 EOPs. Change in weight was unrelated to any other taphonomic process. Independent as well was abrasion. Breakage and edge rounding fell on the same PCA axis as might be anticipated, but these two processes were independent of all others. The two most important taphonomic processes, dissolution and discoloration, were even more complex. PCA divided the dissolution process into three independent components described by three separate PCA axes. Discoloration, likewise, was dissembled into four PCA axes describing four distinctive discoloring processes. One PCA axis offered the only concordance of ostensibly distinctive taphonomic processes in combining the shell degrading process evinced by the formation of small pits on the shell surface with a discoloring processes producing an orange discoloration on the shell. This concordance was produced by one of the two processes generating orange discoloration. The type associated with pitting appears to be caused by a biological process exemplified by Plate 1F. The second type, revealed distinctly by the bivalve-only PCA, was nearly exclusively associated with Mercenaria mercenaria (Fig. 15). Unlike other species, orange discoloration in this species was a byproduct of the incorporation of pyrite into the shell during life [see Clark & Lutz (1980) for a description]. In some EOPs, this pyrite was oxidized to iron oxide during exposure, producing the orange discoloration recorded upon collection (Plate 1B, 1C).

[FIGURE 28 OMITTED]

Discoloring, whether orange, brown, or gray-to-black, is largely incompatible with the development of a chalky surface. Figure 29 shows that EOPs with higher Factor-3 scores (chalkiness) cover a wide range of discoloration indices. Low chalkiness is much more often associated with high gray-toblack discoloration (e.g., the 15-m, 183-m, and crest sites on transect BA), but important exceptions exist (e.g., the 30-m and trough sites on transect BA and Parker Bank). In fact, discoloration is decoupled from dissolution overall, the single exception being the coincidence of pitting and orange discoloration. The consistent orthogonality between dissolution and discoloration is singularly noteworthy, also having been remarked upon earlier by Callender et al. (2002) and Powell et al. (2002). Possibly, however, dissolution, when active enough to significantly change the texture of the shell surface, simultaneously prevents the colonization and development of the microbial films necessary to drive the discoloring process. Those EOPs with shells both discolored and chalky mostly were discolored gray-to-black and came from sulfidic habitats, rather than being discolored green, brown-to-red, or orange. Whether the degradation of the shell occurs chemically or through maceration is not discriminated by the orthogonality between discoloration, much of which is suspected to be of microbial origin, and dissolution as the bacteria responsible for the breakdown of the organic matrix and, consequently, maceration, do not seem to impart significant coloration to the shell surface (e.g., Herrera-Duvauh & Roux 1986, Poulicek et al. 1981, Poulicek & Jeuniaux 1982, Simon et al. 1994).

[FIGURE 29 OMITTED]

Correlates of the Taphonomic Process

Environments of deposition are assumed to modulate the taphonomic process such that unique taphonomic signatures can be associated with empirically distinguished EODs defined by edaphic, physical, and macro-biological processes (e.g., Cutler 1995, Powell et al. 2002, Parsons-Hubbard 2005). As a consequence, certain basic descriptors of the environment should exert a significant effect on taphonomic signature. Initial analyses compared locations of deployment by combining sites into visually distinctive EODs based on sediment type, depth, and visually distinctive biotic and sedimentary regimes. This analysis produced a complicated array of statistical results in which interaction terms between species and EOD were routinely significant. The source of this complexity was immediately revealed when Bahamian transect-AA locations were separated from Bahamian transect-BA locations of the same EOD designation. Subsequent analyses using the larger set of site groups showed consistent location main effects, few species main effects, and no significant interaction terms. That is, taphonomic signature after eight years was nearly exclusively a function of location of deployment.

Distinctive locations of deployment were subsumed into what one would normally consider an EOD-level distinction. Sand wave crest and trough habitats existed on both transect AA and BA, for example, as did sites on the talus slope below the wall and at 15 m on the carbonate platform. Nevertheless, these within-EOD location pairs were significantly different in taphonomic signature. That is, functionally, these location pairs were separate EOPs. Obviously, EOD-level characteristics were poor descriptors of environments of preservation; hence our preferred use of the acronym EOP (environment of preservation). Different environments of preservation may not be distinguishable by visual inspection of biotic and sedimentary regimen.

To further evaluate the distinction between EOP and EOD, we examined depth, sediment type, and degree of exposure. Depth did not exert a single significant effect on any of the eight primary taphonomic factors resolved by the PCA regardless of whether sites were defined coarsely into EODs or distributed into EOPs. This trend was already noteworthy after two years (Callender et al. 2002; Powell et al. 2002) and simply persisted after eight years. Why does no trend exist with depth? One of the likely reasons is that burial removed some shallow-water arrays from the photic zone. The 30-m sites on Bahamian transects AA and BA are of this type. Both sites are in the photic zone, but arrays at both sites were rapidly and, at least as can be judged by aperiodic examination over the intervening eight years, permanently buried within the first year of deployment.

Sediment type and degree of exposure routinely generated significant main effects and often significant interaction terms. The analysis also produced a number of significant species main effects and significant interaction terms between species and degree of exposure or species and sediment type. Moreover, interaction terms between degree of exposure and sediment type were significant for seven of eight factors. The complexity of the statistical results are directly analogous to the results of analyses using the eighteen visual EODs. That is, combining locations into sediment types or sediment type-exposure pairs is as inappropriate or unenlightening as combining locations into visual EODs. The simple fact is that the trends in taphonomic signature cannot be explained from any simple descriptor of environment of deposition. What is particularly disturbing is the great similarity in habitat, by visual criteria, of some transect AA and BA locations, including the crest and trough sites in water depths below 260 m, and the substantial dissimilarity in taphonomic signature. This discrepancy between expectation and reality suggests that assuming equivalence in the taphonomic process using visual similarity in EOD should be treated with a healthy measure of skepticism. EOPs and EODs are not synonymous concepts.

What parameters need to be measured to develop a more objective set of criteria for determining similarity in EOP is unclear. Cai et al. (2006) examined sediment geochemical measures with some degree of success. Whereas SSETI data do not directly confirm the efficacy of this approach, the comparison of shells on the visually similar Bahamian EODs suggest that the carbonate chemical milieu varies considerably between transects AA and BA. Certainly, the conflation of taphofacies and environment of deposition, though understandably important, is still illusory.

Discoloration and Dissolution: Origins and Interactions

The three PCA factors describing the dissolution process discriminated the inner and outer shell surface (and spire and body whorl). The three factors also distinguished pitting from the development of a chalky surface. Behrensmeyer et al. (2000) and Glover and Kidwell (1993) have emphasized that the process normally referred to as dissolution includes a complex of distinctive shell surface-degrading processes driven differentially by chemical attack of calcium carbonate (Tribble 1993, Cutler 1995, Isaji 1995, Cai et al. 2006), biological attack of the supporting organic matrix (Fitzgerald et al. 1979, Alexandersson 1979, Frerotte et al. 1983, Simon et al. 1994), and microboring (Wilkinson & Burrows 1972, Hook et al. 1984, Cutler 1995, Mao Che et al. 1996). The relative importance of these various processes in any death assemblage and the sedimentological and geochemical factors influencing their importance are poorly understood. Analysis of SSETI deployments does not permit discrimination of the causative factors behind the development of a dissolution surface on the shell. Nevertheless, in SSETI deployments, pitting was a distinctive shell surface phenomenon distinguished from the development of a chalky surface and the more extreme forms of surface degradation and, presumably, developed in a different way. Thus, SSETI deployments confirm that significantly different processes exist between EOPs and implicate microbiological processes as being at least as important as physical chemistry.

One of the curiosities of this dataset is the lower dissolution indices for shells deployed on carbonate hardgrounds and veneered hardgrounds compared with the remaining sediment types such as carbonate muds, terrigenous sands, and brine-exposed sites (Fig. 27). The simplest explanation is that the latter three sediment types are characterized by higher acid production from sulfide oxidation than the former two. High sulfide production rates are known to be associated with some of the latter EOPs, including the petroleum seeps (Scott & Fisher 1995, Callender & Powell 1997) and the East Flower Garden brine seep (Powell et al. 1986). However, another possibility is the protective nature of encrusting bionts. That bionts can degrade shell is well known, but bionts as a protective coat have also received some consideration (Kidwell 1986b, Zuschin & Pervesler 1996, Zuschin et al. 1999).

Dissolution was not the only taphonomic process that could be dissembled into a set of relatively unrelated components. Discoloration came in at least five flavors: fading without subsequent discoloration, the development of a brown-to-red coloration, orange/orange mottled discoloration often associated with pitting of the shell surface, development of a green/ green mottled color, and gray-to-black discoloration. For a few of these, mechanisms can at least be guessed. Sites with shells discolored the least were locations where the exposure of shells to oxic overlying seawater was minimized. These sites included Bahamian sites where arrays were buried rapidly and permanently soon after deployment in 1993/1994 and arrays deployed in anoxic brine. Exposure to overlying seawater for significant periods resulted in discoloration that ranged from simple fading to brown, green, orange, and gray-to-black discoloration. Green discoloration was most common in shallow-water hardgrounds, and this association is facilely explained by the distribution of the causative organisms, boring green algae. The remaining discoloring processes were associated much more with carbonate sites than terrigenous sites, and often, though not exclusively, with sites in deeper water. The association of gray-to-black discoloration with carbonate sites, particularly deepwater carbonate sites in the Bahamas, is particularly unexpected, as the oft remarked explanation for this discoloration requires exposure to sulfide and the presence of reduced iron (e.g., Pilkey et al. 1969a, Pilkey et al. 1969b). The visual appearance of these shells (e.g., Plate IA, 1C) suggests a chemical process, however, as compared with those with brown and orange discoloring suggestive of microbial production (Plate 1F, 2C).

The sources of the brown, brown-to-red, and orange discoloration are unknown, but the former two were geographically and bathymetrically widespread. Orange discoloration is a rarely reported color change (e.g., Lazo 2004), but occurred in a variety of EOPs in this study. Microbial biofilms are proving to be complex and locally variable (Aloisi et al. 2002, Nocker et al. 2004, Heijs et al. 2006). Further investigation of the surface films on skeletal remains and how these mediate the development of a taphonomic signature and the degradation of shell integrity is clearly needed. Advance in this area however is necessarily coupled with the recognition that coarsely defined descriptive terms for shell surface characteristics, such as discoloration and dissolution, by mixing disparate processes, do not provide the information necessary to decipher the details of the taphonomic milieu.

In SSETI deployments, deterioration of the inner shell surface did not proceed in step with deterioration of the outer shell surface. The distinctiveness of the inner and outer shell surface as mediators of taphonomic attack in eight-yr SSETI deployments is in keeping with other studies that have emphasized this dichotomy (e.g., Lescinsky 1993, Cutler 1995, Kidwell et al. 2001). The highest dissolution index was normally observed on the outer shell surface. This surface was exposed also throughout life and, as a consequence, the time course of deterioration was much longer (e.g., Kidwell et al. 2001), but the inner shell surface tended to be significantly different from controls more frequently, albeit not significantly so. For some deployed species such as Arctica islandica, Mercenaria mercenaria, and Codakia orbicularis, the time period of exposure of the outer shell surface is similar to or even many times greater than the eight-year deployment period, based on the estimated life spans of these species (e.g., Peterson 1983, Berg & Alatalo 1984, Powell & Mann 2005). The simple expectation that the shell area most degraded during life is also likely to degrade the fastest after death would produce the observed tendency for the outer shell surface to exhibit the greatest degree of degradation. The inner shell surface degraded in a different fashion also; chalkiness was most associated with the inner shell surface. Likely, this stems from three abetting factors. First, the inner shell surface was least degraded during life and so the least severe level of dissolution, a chalky surface, was most likely to show up on this shell surface. Second, many bivalves develop a slightly chalky inner shell surface during life (Crenshaw & Neff 1969, Wilkes & Crenshaw 1979) and almost certainly the process of tissue decomposition that likely occurred at least to some extent after death (even if the shell collectors removed the meats soon after capture) (Smith 1953, Berner 1969, Simon et al. 1994) would have imparted a light chalkiness on the inner surface. Finally, the presence of periostracum on the outer surface of some species tends to minimize minor dissolution because this type of alteration can only occur where the periostracum has been eroded and periostracal erosion likely rapidly leads to a more significant surface degradation than chalkiness.

The Multifariousness of the Taphonomic Process

Parsons-Hubbard (2005) recognized the inadequacy of any individual taphonomic characteristic in explaining the preservational process. Most taphonomic processes take place in most EOPs and the differential in rate among EOPs is not sufficient for any one taphonomic process to uniquely mark any EOP. Parsons-Hubbard (2005) stressed that uniqueness issued from the comingling of taphonomic processes in multifarious ways, so that EOPs are characterized by unique mixtures of taphonomic traits operating at different rates. It is this complexity that underlies the concept of taphofacies (Meldahl & Flessa 1990, Speyer & Brett 1991) and taphonomic grade (Brandt 1989, Martin et al. 1996). SSETI recovered experimental arrays at 41 locations and consolidated these into 24 EOPs based upon visual and a posteriori statistical evidence. SSETI measured a suite of taphonomic processes and focused on nine attributes, eight of which were resolved by PCA, plus the residual of wet weight. The majority of the EOPs bore unique taphonomic signatures when evaluated at this level of detail, supporting the inference of Parsons-Hubbard (2005) that the unique imprint of the taphonomic condition in an EOP is established by the comingling at various intensities of a suite of taphonomic processes. What is also clear is that these processes are inherently independent to some extent. Few correlations existed between them, the one between breakage and edge rounding and the one between orange discoloration and pitting being the two substantive exceptions. Different processes operating at different rates for different reasons in EOPs seems the most likely explanation for unique taphonomic signatures and, consequently, the search for the underlying cause will necessarily involve a multidimensional investigation into the biotic, chemical, and physical milieu of the EOP.

How Good Are Short-term Actualistic Experiments

Behrensmeyer et al. (2000) reviewed field experiments designed to measure the taphonomic process through the use of deployed skeletal remains. Such experiments, though often termed "long-term", have usually been conducted for one to two years, occasionally longer (e.g., Callender et al. 1994). The SSETI deployments considered here are, therefore, unusual in having been on the seafloor for a substantially longer period of time. Moreover, SSETI also recovered experiments after two years (Staff et al. 2002, Powell et al. 2002), a more typical time frame for field experiments; thus affording a comparison between the 'normal' experimental time frame and one of longer term.

The tendency for experiments to be conducted for one to two years is, of course, a function of financial support, the potential for experiments to remain undisturbed in many EOPs for a period of time, and the ability to find and recover experiments after years on the seafloor. Given such constraints, and the reasonable belief that the preservational process takes considerable time, the question arises: Can one learn anything substantive about the preservational process from experiments run for only a few years? Or, do the results from such experiments lead one astray by focusing one's attention on short-term processes that ultimately, over the longer term, may be nugatory or diversionary in nature?

The deployments recovered by SSETI after eight years on the seafloor are themselves not particularly long. A few sites have been buried below what might likely be the TAZ. But the vast majority of arrays are still exposed, partially buried, or not buried so deeply as to have escaped taphonomy's wrath. Nevertheless, in the context of the question, a comparison between the two-year and eight-year recoveries does provide some insight into the permanence of inferences made after one to two years. In doing so, we hope to avoid the quotidian, and perhaps egotistical, assertion that "longer is always better".

In 2002, Powell et al. (2002) summarized the early results of SSETI. "Overall (after two years), most shells deployed at most sites had relatively minor changes in shell condition. Most EODs generated relatively similar taphonomic signatures. A few sites did produce taphonomic signatures clearly distinguishable from the central group and these sites were characterized by one or more of the following: high rates of oxidation of reduced compounds, presence in the photic zone, and significant burial and exhumation events. Most sites, however, showed similar taphonomic signatures, despite the variety of EOD characteristics present. Discoloration and dissolution were by far the dominant processes. Periostracum breakdown, loss of shell weight, and chipping and breakage were less noticeable. EOD-specific edaphic factors often overrode the influence of geographic-scale environmental gradients. Taphonomic alteration was greater on hardgrounds and in brine-exposed sites than on terrigenous muds. Dissolution was less effective at sites where burial was greatest. Discoloration occurred most rapidly at shallower sites and on hardgrounds. Water depth was less influential in determining taphonomic signature than burial state or sediment type."

After eight years, the rank order of importance of taphonomic processes has not changed. Discoloration dominates, then dissolution, with other processes such as edge rounding and breakage playing a lesser role. The inference that loss of carbonate mass was inconsequential remained true after eight years. The inference that EOP-specific factors were more important than geographic-scale factors such as depth and to an important extent, light, though the latter clearly outweighed the former, was reinforced by the analysis of eight-year recovered shells. Thus, at a coarse level of resolution, the taphonomic processes active at SSETI sites were accurately assessed after two years of deployment time.

After two years, abrasion indices tended to be inversely correlated with dissolution indices and discoloration was observed to be distinctive (orthogonal in PCA) to dissolution. After eight years, the second of these inferences remains true, with the curious exception of pitting and orange discoloration. The primary dissolution and discoloration factors, Factors 1-3 and 7-8, are defined solely by a single trait. On the other hand, abrasion is no longer inversely correlated with dissolution. Presumably, the overprinting process that dominated shortly after shell deployment has to a large extent been completed and the present distribution of abrasion is more in keeping with the importance of the process in the field.

Discoloration was revealed after eight years to be a complex mixture of a number of unique processes, including green discoloration from boring algae, brown discoloration, gray-to-black discoloration, and orange discoloration. Clues to this process were present after two years of deployment time, but insufficient discoloration had amassed to confidently separate any but the rapid development of a green color in shells exposed in the photic zone from other discoloring mechanisms. Hindsight reveals the accuracy of the tentative two-year inference.

The distribution of discoloration is markedly altered however. After two years, discoloration was most pronounced at shallow sites and on hardgrounds. Much of this was caused by the rapid influx of boring green algae, in keeping with expectation (Wilkinson & Burrows 1972, Kiene et al. 1995, Mao Che et al. 1996, Parsons-Hubbard 2005), and the general fading of the original shell color. After eight years, the processes resulting in brown and gray-to-black discoloration have increased in importance and these are neither concentrated on hardgrounds nor shallow sites (Fig. 24).

Dissolution was less effective where burial was greatest after eight years, as had been foreshadowed after two years; however, the significance level of the main effect was not strong and interaction terms were significant (Table 12). Thus, the importance of burial in modulating shell surface degradation by processes normally referred to as dissolution was not as large as one might surmise from literature speculation; of course, it is likely that few if any of the arrays have passed through the TAZ in these eight years.

After two years, only minor changes in shell condition had occurred at most sites and most EOPs were similar in their taphonomic signature. Increased shell degradation would be expected over the intervening five years and this was found to be the case. A greater spread in taphonomic signature was the concomitant, and anticipated, result (Figs. 23-25). After two years, a few sites had obtained unique taphonomic signatures. Brine-exposed sites and petroleum seeps were singled out as locations of greater that average dissolution. Petroleum seep GC234 and the East Flower Garden brine seep canyon continued to be sites of unusually active surface degradation by dissolution after eight years (Fig. 23). Discoloration was greatest after two years at two Bahamian sites, the platform top and the wall, and at two EFG sites, the brine canyon and the talus slope below it. Unlike the consistency in the suite of sites with highest dissolution scores, the degree of discoloration was dramatically altered from this simple picture after eight years (Fig. 24). Indeed, the two-year data would not have predicted in any way the eight-year observations. Most buried sites continued to be less discolored than many others. Most sites exposed in the photic zone continued to be more discolored than many others. However, discoloration was pervasive at a number of deeper-water carbonate sites in the Bahamas and the Gulf of Mexico after eight years (Figs. 24, 25). In most cases, these sites were distinguished by orange, brown, or gray-to-black discoloration that increased substantially over the intervening five years.

At the other end of the taphonomic spectrum, at the two sites with nearly pristine shells after two years, the hardgrounds of Parker Bank, and the upper talus slope in the Bahamas, shells became discolored over the intervening five years, although dissolution remained low. Shells at these sites no longer fell clearly below shells at other sites in the degree of taphonomic degradation. The site seemingly least influenced by taphonomic attach after eight years was the southern Texas shelf site, but this site was not sampled after two years. All remaining sites had some form of taphonomic imprint that fell near or above the mean of all EOPs for that taphonomic trait.

In summary, significant changes have occurred in the intervening six years since the two-year recoveries. Some trends identified after two years continued, but some did not. The most general conclusions were reinforced, on the whole, but the details behind them were sometimes greatly altered. This comparison suggests that some important revelations can be gleaned from experimental deployment of shells for one-to-two years, yet the same comparison discloses the spuriousness of other inferences formulated after two years. Longer deployments are better. The taphonomic process is, in general, slow, and the rates of processes are not a linear function of deployment time. Some processes begin early and become less important over time. Others increase in importance over time. Nonlinearity constrains the range of inferences that can be deduced accurately from short deployment times. Of course, we do not know how time-stable these conclusions are. Eight years is still a relatively short period of time. SSETI continues and longer-term deployments may yet be available to evaluate the verity of ratiocinations reached herein.

Although the number of SSETI sites is relatively large and the types of EODs diverse, the full ambit of preservation is unlikely thereby encompassed. It seems clear that an important actualistic focus should be placed on the initiation of long-term experiments in EOPs not encompassed by the SSETI program. Only time, and the deployment of shells in a wider range of EODs, can evaluate the seriousness of taphonomic bias in the full range of shelf and slope environments.

ACKNOWLEDGMENTS

The submersible work required for the deployment and recovery of experiments was made possible through a series of grants from NOAA's National Undersea Research Programs at the University of North Carolina at Wilmington and the Caribbean Marine Research Center, and the National Science Foundation-Geology and Paleontology Program. We would like to thank the NSF and these two NURP programs for the consistent funding of the nine major field efforts that permitted deployment and recovery over such a large regional area. The authors would like to thank the support crews of the Clelia, Johnson-Sea-Link, and Nekton Gamma submersibles and support vessels. The authors appreciate the efforts of the CMRC staff on Lee Stocking Island, the NURP personnel from UNCW and CMRC, and students and staff from Texas A&M University, Oberlin College, Rutgers University, University of Georgia, and University of Cincinnati that took part in these field programs.

LITERATURE CITED

Al-Agha, M. R., S. D. Burley, C. D. Curtis & J. Esson. 1995. Complex cementation textures and authigenic mineral assemblages in Recent concretions from the Lincolnshire Wash (east coast, UK) driven by Fe(0) to Fe(II) oxidation. J. Geol. Soc. Lond. 152:157-171.

Alexandersson, E. T. 1979. Marine maceration of skeletal carbonates in the Skagerrak, North Sea. Sedimentology 26:845-852.

Alexandersson, T. 1972. Micritization of carbonate particles: Processes of precipitation and dissolution in modern shallow-marine sediments. Bull. Geol. Inst. Univ. Upsala n.s. 3 7:201-236.

Allison, P. A. 1990. Variation in rates of decay and disarticulation of Echinodermata: Implications for the application of actualistic data. Palaios 5:432-440.

Aloisi, G., I. Bouloubassi, S. K. Heijs, R. D. Pancost, C. Pierre, J. S. Sinninghe Damste, J. C. Gottschal, L. J. Forney & J.-M. Rouchy. 2002. CH4-consuming microorganisms and the formation of carbonate crusts at cold seeps. Earth Planet. Sci. Lett. 203:195-203.

Balsam, W. L. & J. P. Beeson. 2003. Sea-floor sediment distribution in the Gulf of Mexico. Deep-Sea Res. I Oceanogr. Res. Pap. 50:1421-1444.

Barbieri, R. 2001. Taphonomic implications of foraminiferal composition and abundance in intertidal mud flats, Colorado River delta (Mexico). Micropaleontology 47:73-86.

Behrens, E. W. 1988. Geology of a continental slope oil seep, northern Gulf of Mexico. Am. Assoc. Petrol. Geol. Bull. 72:105-114.

Behrensmeyer, A. K., S. M. Kidwell & R. A. Gastaldo. 2000. Taphonomy and paleobiology. Paleobiology 26(suppl.): 103-147.

Berg, C. J, Jr. & P. Alatalo. 1984. Potential of chemosynthesis in molluscan mariculture. Aquaculture 39:165-179.

Bergquist, D. C., T. Ward, E. E. Cordes, T. McNelis, S. Howlett, R. Kosoff, S. Hourdez. R. Carney & C. R. Fisher. 2003. Community structure of vestimentiferan-generated habitat islands from Gulf of Mexico cold seeps. J. Exp. Mar. Biol. Ecol. 289:197-222.

Berner, R. A. 1969. Chemical changes affecting dissolved calcium during the bacterial decomposition of fish and clams in sea water. Mar. Geol. 7:253-274.

Bouma, A. H., M. H. Feeley, J. L. Kindinger, C. E. Stelting & T. W. C. Hilde. 1981. Seismic stratigraphic characteristics of upper Louisiana continental slope: an area east of Green Canyon. Offshore Technol. Conf. Pap. OTC 4098:283-291.

Bouma, A. H., R. G. Martin & W. R. Bryant. 1980. Shallow structure of upper continental slope, central Gulf of Mexico. Offshore Technol. Conf. Pap. OTC 3913:583-592.

Brandt, D. S. 1989. Taphonomic grades as a classification for fossiliferous assemblages and implications for paleoecology. Palaios 4:303-309.

Brett, C. E. 1990. Obrution deposits. In: D. E. G. Briggs & P.R. Crowther, editors. Palaeobiology: a synthesis. Oxford: Blackwell Scientific Publishers. pp. 239-243.

Brett, C. E. & G. C. Baird. 1986. Comparative taphonomy: a key to paleoenvironmental interpretation based on fossil preservation. Palaios 1:207-227.

Bright, T. J. & E. N. Powell. 1983. The East Flower Garden brine seep, a unique ecosystem. In: Reefs and banks of the northwestern Gulf of Mexico: Their geological, biological, and physical dynamics. Northern Gulf of Mexico Topographic Features Synthesis, Final Report. Contract No. AA851-CT1-55 U.S. Dept. Interior, Minerals Management Service, Outer Continental Shelf Office. Pp. 277 310.

Brooks, J. M., T. J. Bright, B. B. Bernard & C. R. Schwab. 1979. Chemical aspects of a brine pool at the East Flower Garden Bank, northwestern Gulf of Mexico. Limnol. Oceanogr. 24:735-745.

Brooks, J. M., H. B. Cox, W. R. Bryant, M. C. Kennicutt, II, R. G. Mann & T. J. McDonald. 1986. Association of gas hydrates and oil seepage in the Gulf of Mexico. Adv. Org. Geochem. 10:221-234.

Cadee, G. C. 1999. Shell damage and shell repair in the Antarctic limpet Nucella concinna from King George Island. J. Sea Res, 41:149-161.

Cai, W.-J., F. Chen, E. N. Powell, S. E. Walker, K. M. Parsons-Hubbard, G. M. Staff, Y. Wang, K. A. Ashton-Alcox, W. R. Callender & C. E. Brett. 2006. Preferential dissolution of carbonate shells driven by petroleum seep activity in the Gulf of Mexico. Earth Planet. Sci. Lett. 248:227-243.

Callender. W. R. & E. N. Powell. 1992. Taphonomic signature of petroleum seep assemblages on the Louisiana upper continental slope: Recognition of autochthonous shell beds in the fossil record. Palaios 7:388-408.

Callender, W. R. & E. N. Powell. 1997. Autochthonous death assemblages from chemoautotrophic communities at petroleum seeps: Paleoproduction, energy flow, and implications for the fossil record. Hist. Biol. 12:165-198.

Callender, W. R. & E. N. Powell. 1999. Why did ancient chemosynthetic seep and vent assemblages occur in shallower water than they do today? Int. J. Earth Sci. 88:377-391.

Callender, W. R. & E. N. Powell. 2000. Long-term history of chemoautotrophic clam-dominated faunas of petroleum seeps in the northwestern Gulf of Mexico. Facies 43:177-204.

Callender, W. R., E. N. Powell & G. M. Staff. 1994. Taphonomic rates of molluscan shells placed in autochthonous assemblages on the Louisiana continental slope. Palaios 9:60-73.

Callender, W. R., E. N. Powell, G. M. Staff & D. J. Davies. 1992. Distinguishing autochthony, parautochthony and allochthony using taphofacies analysis: Can cold seep assemblages be discriminated from assemblages of the nearshore and continental shelf? Palaios 7:409-421.

Callender, W. R., G. M. Staff, K. M. Parsons-Hubbard, E. N. Powell, G. T. Rowe, S. E. Walker, C. E. Brett, A. Raymond, D. D. Carlson, S. White & E. A. Heise. 2002. Taphonomic trends along a forereef slope: Lee Stocking Island, Bahamas. I. Location and water depth. Palaios 17:50-65.

Canfield. D. E. & R. Raiswell. 1991. Carbonate precipitation and dissolution: its relevance to fossil preservation. In: P. A. Allison & D. E. G. Briggs, editors. Taphonomy: releasing the data locked in the fossil record. New York: Plenum Press. pp. 412-453.

Carney, R. S. 1994. Consideration of the oasis analogy for chemosynthetic communities at Gulf of Mexico hydrocarbon vents. Geo-Marine Lett. 14:149-159.

Christmas, J. F.. M. R. McGinty, D. A. Randle, G. F. Smith & S. J. Jordan. 1997. Oyster shell disarticulation in three Chesapeake Bay tributaries. J. Shellfish Res. 16:115-123.

Clark. R. R., 11. & R. A. Lutz. 1980. Pyritization in the shells of living bivalves. Geology (Boulder) 8:268-271.

Crenshaw, M. A. & J. M. Neff. 1969. Decalcification at the mantle-shell interface in mollusks. Am. Zool. 9:881-885.

Curray, J. R. 1960. Sediments and history of Holocene transgression, continental shelf, northwest Gulf of Mexico. In: F. P. Shepard, F. B. Phleger & T. H. van Andel, editors. Recent sediments, northwest Gulf of Mexico. Tulsa, Oklahoma: American Association of Petroleum Geologists. pp. 221-266.

Cochrane, J. D. & F. J. Kelly. 1986. Low-frequency circulation on the Texas-Louisiana continental shelf. J. Geophys. Res. C Oceans Atmosph. 91:10645-10659.

Cummins, H., E. N. Powell, R. J. Stanton, Jr. & G. Staff. 1986. The size-frequency distribution in palaeoecology: The effects of taphonomic processes during formation of death assemblages in Texas bays. Palaeont. (Lond.) 29:495-518.

Cutler, A. H. 1995. Taphonomic implications of shell surface textures in Bahia la Choya, northern Gulf of California. Palaeogeogr. Palaeoclimatol. Palaeoecol. 114:219-240.

Davies, D. J., E. N. Powell & R. J. Stanton, Jr. 1989. Relative rates of shell dissolution and net sediment accumulation--a commentary: Can shell beds form by the gradual accumulation of biogenic debris on the sea floor? Lethaia 22:207-212.

Davies, D. J., G. M. Staff, W. R. Callender & E. N. Powell. 1990. Description of a quantitative approach to taphonomy and taphofacies analysis: All dead things are not created equal. In: W. Miller III, editor, Paleocommunity temporal dynamics: the long-term development of multispecies assemblages. Spec. Publ. (Paleontol. Soc.) 5:328-350.

El-Sayed, S. Z. 1972. Primary productivity and standing crop of production. In: V. C. Bushnell, editor. Chemistry. primary productivity, and benthic algae of the Gulf of Mexico. New York: American Geographic Society. pp. 8-13.

Farrow, G. E. & J. A. Fyfe. 1988. Bioerosion and carbonate mud production on high-latitude shelves. Sediment. Geol. 6:281-297.

Fitzgerald. M. G., C. M. Parmenter & J. D. Milliman. 1979. Particulate calcium carbonate in New England shelf waters: Result of shell degradation and resuspension. Sedimentology 26:853-857.

Flessa, K. W., A. H. Cutler & K. H. Meldahl. 1993. Time and taphonomy: Quantitative estimates of time-averaging and stratigraphic disorder in a shallow marine habitat. Paleobiology 19:266-286.

Frerotte, B., A. Raguideau & J.-P. Cuif. 1983. Degradation in vitro d'un test carbonate d'invertebre, Crassostrea gigas. (Thunberg), par action de cultures bacteriennes. Interet pour l'analyse ultrastructurale. C. R. Acad. Sci. Ser, H Mech.-Phys.-Chim. Sci. 297:383-388.

Gallaway, B. J. 1988. Northern Gulf of Mexico continental slope study, final report, year 4, Vol. 1 Executive summary: Final Report, Minerals Management Service, New Orleans, Contract #14-12-0001-31212, Outer Continental Shelf Study, MMS-88-0052.69 pp.

Gardner, J. V., L. A. Mayer, J. E. H. Clarke & A. Kleiner. 1998. High-resolution multibeam bathymetry of East and West Flower Gardens and Stetson Banks, Gulf of Mexico. Gulf Mex. Sci. 16:131-143.

Gittings, S. R., T. J. Bright & E. N. Powell. 1984. Hard-bottom macrofauna of the East Flower Garden brine seep: Impact of a long-term, sulfurous brine discharge. Contrib. Mar. Sci. 27:105-125.

Glover, C. P. & S. M. Kidwell. 1993. Influence of organic matrix on the post-mortem destruction of molluscan shells. J. Geol. 101:729-747.

Halper, F. B., D. W. McGrail & W. J. Merrill. 1988. Seasonal variability in the currents on the outer Texas-Louisiana shelf. Estuar. Coast. Mar. Sci. 26:33-50.

Heijs, S. K., G. Aloisi, L. Bouloubassi, R. D. Pancost, C. Pierre, J. S. Sinninghe Damste, J. C. Gottschal, J. D. van Elsas & L. J. Forney. 2006. Microbial community structure in three deep-sea carbonate crusts. Microb. Ecol. 52:451-462.

Herrera-Duvault, Y. & M. Roux. 1986. Modalites et vitesse de dissolution de la coquille des modioles du site hydrothermal de 13[degrees]N sur la dorsale du Pacifique oriental. C. R. Acad. Sci. Ser. III Sci. Vie 302:251-256.

Hill, G. W., K. A. Roberts, J. L. Kindinger & G. D. Wiley. 1982. Geobiologic study of the south Texas outer continental shelf. US Geol. Surv. Prof. Pap. 1238:1-36.

Hook, J. E., S. Golubic & J. D. Milliman. 1984. Micritic cement in microborings is not necessarily a shallow-water indicator. J. Sediment. Petrol. 54:425-431.

Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54:187-211.

Isaji, S. 1995. Defensive strategies against shell dissolution in bivalves inhabiting acidic environments: The case of Geloina (Corbiculidae) in mangrove swamps. Veliger 38:235-246.

Kaehler, S. 1999. Incidence and distribution of phototrophic shell-degrading endoliths of the brown mussel Perna perna. Mar. Biol. (Bed.) 135:505-514.

Kennicutt, M. C., II, J. M. Brooks, R. R. Bidigare & G. J. Denoux. 1988a. Gulf of Mexico hydrocarbon seep communities--I. Regional distribution of hydrocarbon seepage and associated fauna. Deep-Sea Res. I Oceanogr. Res. Pap. 35:1639-1651.

Kennicutt, M. C., II, J. M. Brooks & G. J. Denoux. 1988b. Leakage of deep, reservoired petroleum to the nearsurface on the Gulf of Mexico continental slope. Mar. Chem. 24:39-59.

Kidwell, S. M. 1986a. Models for fossil concentrations: Paleobiologic implications. Paleobiology 12:6-24.

Kidwell, S. M. 1986b. Taphonomic feedback in Miocene assemblages: Testing the role of dead hardparts in benthic communities. Palaios 1:239 255.

Kidwell, S. M. 2001. Preservation of species abundance in marine death assemblages. Science (Wash. D.C.) 294:1091-1094.

Kidwell, S. M. & K. W. Flessa. 1995. The quality of the fossil record: Populations, species and communities. Ann. Rev. Ecol. Syst. 26:269-299.

Kidwell, S. M., F. T. Fursich & T. Aigner. 1986. Conceptual framework for the analysis and classification of fossil concentrations. Palaios 1:228-238.

Kidwell, S. M., T. A. Rothfus & M. M. R. Best. 2001. Sensitivity of taphonomic signatures to sample size, sieve size, damage scoring system, and target taxa. Palaios 16:26-52.

Kiene, W., G. Radtke, M. Gertidis, S. Golubic & K. Vogel. 1995. Factors controlling the distribution of microborers in Bahamian reef environments. Facies 32:176-188.

Killingley, J. S., W. H. Berger, K. C. MacDonald & W. A. Newman. 1980. [sup.18]O/[sup.16]O variations in deep-sea carbonate shells from the Rise hydrothermal field. Nature 287:218-221.

Kleeman, K. 1996. Biocorrosion by bivalves. P.S.Z.N.I: Mar. Ecol. 17:145-158.

Kowalewski, M. & W. Misniakiewicz. 1993. Reliability of quantitative data on fossil assemblages: A model, a simulation, and an example. N. Jahr. Geol. Palaontol. Abh. 187:243-260.

LaBarbera, M. 1981. The ecology of Mesozoic Gryphaea, Exogyra, and Ilymatogyra (Bivalvia: Mollusca) in a modern ocean. Paleobiology 7:510-526.

Lazo, D. G. 2004. Bivalve taphonomy: testing the effect of life habits on the shell condition of the littleneck clam Protothaca (Protothaca) staminea (Mollusca: Bivalvia). Palaios 19:451-459.

Lescinsky, H. L. 1993. Taphonomy and paleoecology of epibionts on the scallops Chlamys hastata (Sowerby 1843) and Chlamys rubida (Hinds 1845). Palaios 8:267-277.

Lin, S. & J. W. Morse. 1991. Sulfate reduction and iron sulfide mineral formation in Gulf of Mexico anoxic sediments. Am. J. Sci. 291:55-81. Lugo-Fernandez, A. 1998. Ecological implications of hydrography and circulation to the Flower Garden Banks, northwest Gulf of Mexico. Gulf Mex. Sci. 16:144-160.

Lutz, R. A., L. W. Fritz & R. N. Cerrato. 1988. A comparison of bivalve (Calyptogena magnifica) growth at two deep-sea hydrothermal vents in the eastern Pacific. Deep-Sea Res. 35:1793-1810.

Lutz, R. A., L. W. Fritz & D. C. R. Rhoads. 1985. Molluscan growth at deep-sea hydrothermal vents. Bull. Biol. Soc. Wash. 6:199-210.

Lutz, R. A., M. J. Kennish, A. S. Pooley & L. W. Fritz. 1994. Calcium carbonate dissolution rates in hydrothermal vent fields of the Guaymas Basin. J. Mar. Res. 52:969-982.

MacAvoy, S. E., R. S. Carney, C. R. Fisher & S. A. Macko. 2002. Use of chemosynthetic biomass by large, mobile, benthic predators in the Gulf of Mexico. Mar. Ecol. Prog. Ser. 225:65-78.

MacDonald, I. R., D. B. Buthman, W. W. Sager, M. B. Peccini & N. L. Guinasso, Jr. 2000. Pulsed oil discharge from a mud volcano. Geology (Boulder) 28:907-910.

Mao Che, L., T. le Campion-Alsumard, N. Boury-Esnault, C. Payri, S. Golubic & C. Bezac. 1996. Biodegradation of shells of the black pearl oyster, Pinctada margaritifera var. cumingii, by microborers and sponges of French Polynesia. Mar. Biol. (Berl.) 126:509-519.

Martin, R. E., J. F. Wehmiller, M. S. Harris & W. D. Liddell. 1996. Comparative taphonomy of bivalves and foraminifera from Holocene tidal flat sediments, Bahia la Choya, Sonora, Mexico (Northern Gulf of California): Taphonomic grades and temporal resolution. Paleobiology 22:80-90.

McNichol, A. P., C. Lee & E. R. M. Druffel. 1988. Carbon cycling in coastal sediments: 1. A quantitative estimate of the remineralization of organic carbon in the sediments of Buzzards Bay, MA. Geochim. Cosmochim. Acta 52:1531-1543.

Meldahl, K. H. & K. W. Flessa. 1990. Taphonomic pathways and comparative biofacies and taphofacies in a Recent intertidal shallow shelf environment. Lethaia 23:43-60.

Moir, B. G. 1990. Comparative studies of "fresh" and "aged" Tridacna gigas shell: Preliminary investigations of a reported technique for pretreatment of tool material. J. Archaeol. Sci. 17:329-345.

Morse, J. W., J. C. Cornwell, T. Arakaki, S. Lin & M. Huerta-Diaz. 1992. Iron sulfide and carbonate mineral diagenesis in Baffin Bay, Texas. J. Sediment. Petrol. 62:671-680.

Nocker, A., J. E. Lepo & R. A. Snyder. 2004. Influence of an oyster reef on development of the microbial heterotrophic community of an estuarine biofilm. Appl. Environ. Microbiol. 7:6834-6845.

Olszewski, T. D. 2004. Modeling the influence of taphonomic destruction, reworking, and burial on time-averaging in fossil assemblages. Palaios 19:39-50.

Parker, R. H. 1960 Ecology and distribution patterns of marine macroinvertebrates, northern Gulf of Mexico. In: F. P. Shepard, F. B. Phleger & T. H. van Andel, editors. Recent sediments, northwest Gulf of Mexico. Tulsa, Oklahoma: American Association of Petroleum Geologists. pp. 302-337.

Parsons, K. M. & C. E. Brett. 1991. Taphonomic process and biases in modern marine environments: An actualistic perspective on fossil assemblage preservation. In: S. K. Donovan, editor. The processes of fossilization. London: Belhaven Press. pp. 22-65.

Parsons, K. M., E. N. Powell, C. E. Brett, S. E. Walker & W. R. Callender. 1997. Shelf and Slope Experimental Taphonomy Initiative (SSETI): Bahamas and Gulf of Mexico. Proc. 8th Int. Coral Reef Syrup. 2:1807-1812.

Parsons-Hubbard, K. 2005. Molluscan taphofacies in recent carbonate reef/lagoon systems and their application to sub-fossil samples from reef cores. Palaios 20:175-191.

Parsons-Hubbard, K. M., W. R. Callender, E. N. Powell, C. E. Brett, S. E. Walker, A. L. Raymond & G. M. Staff. 1999. Rates of burial and disturbance of experimentally-deployed molluscs: Implications for preservation potential. Palaios 14:337-351.

Parsons-Hubbard, K. M., E. N. Powell, G. M. Staff, W. R. Callender, C. E. Brett & S. E. Walker. 2001. The effect of burial on shell preservation and epibiont cover in Gulf of Mexico and Bahamas shelf and slope environments after two years: An experimental approach. In: J. Y. Aller, S. A. Woodin & R. C. Aller, editors, Organism-sediment interactions. Belle W. Baruch Library in Marine Science #21, Columbia, South Carolina: University of South Carolina Press. pp. 297-314.

Peterson, C. H. 1976. Relative abundances of living and dead molluscs in two California lagoons. Lethaia 9:137-148.

Peterson, C. H. 1983. A concept of quantitative reproductive senility: Application to the hard clam, Mercenaria mercenaria (L.)? Oecologia 58:164-168.

Pilkey, O. H., B. W. Blackwelder, L. J. Doyle & E. L. Estes. 1969a. Environmental significance of the physical attributes of calcareous sedimentary particles. Trans. Gulf Coast Assoc. Geol. Soc. 19:113-114. Pilkey, O. H., B. W. Blackwelder, L. J. Doyle, E. Estes & P. W. Terlecky. 1969b. Aspects of carbonate sedimentation on the Atlantic continental shelf off the southern United States. J. Sediment. Petrol. 39:744-768.

Plotnick, R. E. 1986. Taphonomy of a modern shrimp: Implications for the arthropod fossil record. Palaios 1:286-293.

Poulicek, M., M. F. Jaspar-Versali & G. Goffinet. 1981. Etude experimentale de la degradation des coquilles de mollusques au niveau des sediments marins. Bull Soc. R. Sci. Liege 50:513-518.

Poulicek, M. & C. L. Jeuniaux. 1982. Biomass and biodegradation of mollusk shell chitin in some marine sediments. In: S. Hirano & S. Tokura, editors, Chitin and chitosan. Proceedings of the 2nd International Conference on Chitin and Chitosan, Sapporo, Japan, pp. 196-199.

Powell, E. N. 1992. A model for death assemblage formation. Can sediment shelliness be explained? J. Mar. Res. 50:229-265.

Powell, E. N., T. J. Bright & J. M. Brooks. 1986. The effect of sulfide and an increased food supply on the meiofauna and macrofauna at the East Flower Garden brine seep. Helgol. Meeresunters. 40:57-82.

Powell, E. N., T. J. Bright, A. Woods & S. Gittings. 1983. Meiofauna and the thiobios in the East Flower Garden brine seep. Mar. Biol. (Berl.) 73:269-283.

Powell, E. N., H. Cummins, R. J. Stanton, Jr. & G. Staff. 1984. Estimation of the size of molluscan larval settlement using the death assemblage. Estuar. Coast. Shelf Sci. 18:367-384.

Powell, E. N. & J. M. Klinck. 2007. Is oyster shell a sustainable estuarine resource? J. Shellfish Res. 26:181-194.

Powell, E. N., J. N. Kraeuter & K. A. Ashton-Alcox. 2006. How long does oyster shell last on an oyster reef?. Estuar. Coast. Shelf Sci. 69:531-542.

Powell, E. N. & R. Mann. 2005. Evidence of recent recruitment in the ocean quahog Arctica islandica in the Mid-Atlantic Bight. J. Shellfish Res. 24:517-530.

Powell, E. N., K. M. Parsons-Hubbard, W. R. Callender, G. M. Staff, G. T. Rowe, C. E. Brett, S. E. Walker, A. Raymond, D. D. Carlson, S. White & E. A. Heise. 2002. Taphonomy on the continental shelf and slope: Two-year trends--Gulf of Mexico and Bahamas. Palaeogeogr. Palaeoclimatol. Palaeoecol. 184:1-35.

Powell, E. N., G. M. Staff, D. J. Davies & W. R. Callender. 1989. Macrobenthic death assemblages in modern marine environments: Formation, interpretation and application. Crit. Rev. Aquat. Sci. 1:555-589.

Rezak, R. & T. J. Bright. 1981. Seafloor instability at East Flower Garden Bank, northwest Gulf of Mexico. Geo-Marine Lett. 1:97-103.

Rezak, R., S. R. Gittings & T. J. Bright. 1990. Biotic assemblages and ecological controls on reefs and banks of the northwest Gulf of Mexico. Am. Zool. 30:23-35.

Roberts, H. H. & R. S. Carney. 1997. Evidence of episodic fluid, gas, and sediment venting on the northern Gulf of Mexico continental slope. Econ. Geol. 92:863-879.

Roberts, H. H., R. Sasson & P. Aharon. 1987. Carbonates of the Louisiana continental slope. Offshore Technol. Conf. Pap. OTC 5463:373-382.

Rothfus, T. A. 2004. How many taphonomists spoil the data? Multiple operators in taphofacies studies. Palaios 19:514 519.

Rowe, G. T. 1983. Biomass and production of the deep-sea macrobenthos. In: G. T. Rowe, editor, Deep Sea Biology. The sea, vol. 8. New York: Wiley-Interscience. pp. 97-121.

Rowe, G. T., M. Sibuet, J. Deming, J. Tietjen & A. Khripounoff. 1990. Organic carbon turnover time in deepsea benthos. Prog. Oceanogr. 24:141-160.

Sahl, L. E., D. A. Wiesenburg & W. J. Merrell. 1997. Interactions of mesoscale features with Texas shelf and slope waters. Cont. Shelf Res. 17:117-136.

Sartoretto, S. 1998. Bioerosion des concretions coralligenes de Mediterranee par les organismes perforants: Essai de quantification des processus. C. R. Acad. Sci. Ser. H Sci. Terre Planet. 327:839-844.

Scott, K. M. & C. R. Fisher. 1995. Physiological ecology of sulfide metabolism in hydrothermal vent and cold seep vesicomyid clams and vestimentiferan tube worms. Am. Zool. 35:102-111.

Shideler, G. L. 1981. Development of the benthic nepheloid layer on the south Texas continental shelf, western Gulf of Mexico. Mar. Geol. 41:37-61.

Simon, A., M. Poulicek, B. Velimirov & F. T. MacKenzie. 1994. Comparison of anaerobic and aerobic biodegradation of mineralized skeletal structures in marine and estuarine conditions. Biogeochemistry 5:167-195.

Siringan, F. P. & J. B. Anderson. 1994. Modern shoreface and inner-shelf storm deposits off the east Texas coast, Gulf of Mexico. J. Sediment. Res. B Stratigr. Global Stud. 64:99-110.

Smith, O. R. 1953. Observations on the rate of decay of soft-shell clams (Mya arenaria). Ecology 34:640-641.

Smyth, M. J. 1989. Bioerosion of gastropods shells: with emphasis on effects of coralline algal cover and shell microstructure. Coral Reefs 8:119-125.

Speyer, S. E. & C. E. Brett. 1991. Taphofacies controls background and episodic processes in fossil assemblage preservation. In: P. A. Allison & D. E. G. Briggs, editors. Taphonomy: releasing the data locked in the fossil record. Topics Geobiol. 9:501-545.

Staff, G. M., W. R. Callender, E. N. Powell, K. M. Parsons-Hubbard, C. E. Brett, S. E. Walker, D. D. Carlson, S. White, A. Raymond & E.A. Heise. 2002. Taphonomic trends along a forereef slope: Lee Stocking Island, Bahamas. II. Time. Palaios 17:66-83.

Staff; G. M. & E. N. Powell, 1990. Taphonomic signature and the imprint of taphonomic history: Discriminating between taphofacies of the inner continental shelf and a microtidal inlet. In: W. Miller III, editor, Paleocommunity temporal dynamics: the long-term development of multispecies assemblages. Spec. Publ. (Paleontol. Soc.) 5:370-390.

Stentoft, N. 1994. Early submarine cementation in fore-reef carbonate sediments, Barbados, West Indies. Sedimentology 41:585-604.

Tribble, G. W. 1993. Organic matter oxidation and aragonite diagenesis in a coral reef. J. Sediment. Petrol. 63:523-527.

Walker, S. E. 1988. Taphonomic significance of hermit crabs (Anomura; Paguridea): Epifaunal hermit crab-infaunal gastropod example. Palaeogeogr. Palaeoclimatol. Palaeoecol. 63:45-71.

Walker, S. E. 2001. Below the sediment-water interface: a new frontier in taphonomic research. Palaios 16: l 13-114.

Walker, S. E. & J. T. Carlton. 1995. Taphonomic losses become taphonomic gains: An experimental approach using the rocky shore gastropod, Tegula funebralis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 114:197-217.

Walker, S. E. & S. T. Goldstein. 1999. Taphonomic tiering: Experimental field taphonomy of molluscs and foraminifera above and below the sediment-water interface. Palaeogeogr. Palaeoclimatol. Palaeoecol. 149:227-244.

Walker, S. E., K. Parsons-Hubbard, E. N. Powell & C. E. Brett. 1998. Bioerosion or bioaccumulation? Shelf-slope trends for epi- and endobionts on experimentally deployed gastropod shells. Hist. Biol. 13:61-72.

Walker, S. E., K. Parsons-Hubbard, E. Powell & C. E. Brett. 2002. Predation on experimentally deployed molluscan shells from shelf to slope depths in a tropical carbonate environment. Palaios 17:147-170.

Warburton, F. E. 1958. Control of the boring sponge on oyster beds. Fish. Res. Board Can. Prog. Reps. Atl. Coast Stat. 69:7-11.

White, S. 2002. Encrusting foraminifera from Lee Stocking Island, Bahamas: taphonomy, shelf-to-slope distribution, and behavior. M.S. Thesis, University of Georgia, Athens, Georgia, 150 pp.

Wilkes, D. A. & M. A. Crenshaw. 1979. Formation of a dissolution layer in molluscan shells. Scan. Electron Mierosc. 2:469-474.

Wilkinson, M. & E. M. Burrows. 1972. The distribution of marine shell-boring green algae. J. Mar. Biol. Assoc. U.K. 52:59-65.

Zuschin, M. & P. Pervesler. 1996. Secondary hardground-communities in the northern Gulf of Trieste, Adriatic Sea. Senck. Marit. 28:53-63.

Zuschin, M., M. Stachowitsch, P. Pervesler & H. Kollman. 1999. Structural features and taphonomic pathways of a high-biomass epifauna in the northern Gulf of Trieste, Adriatic Sea. Lethaia 32:299-317.

Zuschin, M., M. Stachowitsch & R. J. Stanton, Jr. 2003. Patterns and processes of shell fragmentation in modern and ancient marine environments. Earth-Sci. Rev. 63:33-82.

ERIC N. POWELL, (1) * W. RUSSELL CALLENDER, (2) GEORGE M. STAFF, (3) KARLA M. PARSONS-HUBBARD, (4) CARLTON E. BRETT, (5) SALLY E. WALKER, (6) ANNE RAYMOND (7) AND KATHRYN A. ASHTON-ALCOX (1)

(1) Haskin Shellfish Research Laboratory, Rutgers University, 6959 Miller Ave., Port Norris, New Jersey 08349; (2) National Oceanic and Atmospheric Administration, Center for Coastal Monitoring and Assessment (N/SCI1), 1305 East-West Highway, Silver Spring, Maryland 20910; (3) Austin Community College, NRG Campus, Geology Department, 11928 Stonehollow Drive, Austin, Texas 78758 (4) Department of Geology, Oberlin College, Oberlin, Ohio 44074; (5) Department of Geology, University of Cincinnati, Cincinnati, Ohio 45221; (6) Department of Geology, University of Georgia, Athens, Georgia 30602; (7) Department of Geology and Geophysics, Texas A&M University, College Station, Texas 77843

* Corresponding author. E-mail: eric@hsrl.rutgers.edu.
TABLE 1.
Degrees of exposure for each SSETI site recovered in 1999-2001. Site
designations are those of Parsons-Hubbard et al. (1999) (their Tables
1 and 2), with the exception of East Flower Garden sites that have
been reordered. Parentheticals are EFG site names of Parsons-Hubbard
et al. (1999). E, exposed; D, dusted; P, partially buried; B, buried.

              Caribbean

Site Name   Site Type       Exposure

BA860TR     dune trough        D
BA850TN     dune trough        D
BA860DP     dune crest         P
BA830CU     dune crest         B
BA705BO     talus slope        D
BA740BO     talus slope        D
BA640CL     talus slope        E
BA230WA     wall               E
BA230WN     wall               E
BA100NO     sand channel       B
BA100SO     sand channel       B
BA50SND     sand channel       B
NORMANP     saline lagoon      B
AA50RID     patch reef         D
AA50SND     sand channel       B
AA100S1     sand channel       B
AA100S2     sand channel       B
AA240LE     wall               E
AA290WA     wall               E
AA7000R     talus slope        E
AA875TR     dune trough        E
AA865CR     dune crest         E

                    Gulf of Mexico

Site Name   Site Type                         Exposure

PARKR31     carbonate hardground                 P
PARKR41     carbonate hardground                 E
GB425S1     petroleum seep clam bed              P
GB425S2     petroleum seep clam bed              P
GB425S3     petroleum seep clam bed              P
GB425S4     petroleum seep clam bed              B
EFLGDN3     brine lake                           B
(EFLGDN2)
EFLGDN1     coralline algal deep reef            E
EFLGDN2     coralline algal deep reef            E
EFLGDN4     brine lake                           B
EFLGDN5     brine stream                         E
EFLGDN6     brine canyon mouth                   E
EFLGDN7     canyon talus fan                     B
EFLGDN8     talus slope                          B
STXSHS1     continental shelf                    D
STXSHS2     continental shelf                    B
GC234S2     petroleum seep tubeworm bed          B
GC234S3     petroleum seep carbonate ridge       P
GC234S4     petroleum seep carbonate ridge       E

TABLE 2.
The number of individuals used in statistical analyses for each
species, ocean basin, sediment type, depth range, and burial state.

                       Arctica    Mytilus     Codakia     Strombus
                      islandica   edulis    orhicularis   luhuanus

Ocean basin
  Bahamas                115        218         198          225
  Gulf of Mexico         105        184         166          180
Sediment type
  Brine                   10         20          20           20
  Carbonate sand         100        167         178          180
  Hardground              25         40          40           40
  Terrigenous mud         55         65          67           70
  Carbonate mud            0         10           0            5
  Veneered
    hardground            30         80          59           80
  Terrigenous sand         0         20           0           10
Burial
  Exposed                 80        140         139          140
  Dusted                  20         62          40           60
  Partially buried        40         55          58           60
  Buried                  80        145         127          145
Depth range
  <50 m                   65         78          78           85
  50-<100 m               70        139         139          140
  100-<200 m               0         30          10           20
  200-<300 m              30         90          70           90
  [greater than or
    equal to] 300 m       55         65          67           70

                      Telescopium   Glycymeris   Argopecten
                      telescopium     undata     irradians

Ocean basin
  Bahamas                 65            82           15
  Gulf of Mexico           0            20           65
Sediment type
  Brine                    0             0           10
  Carbonate sand          45            45           20
  Hardground               5             5           15
  Terrigenous mud          0             0           15
  Carbonate mud            0             0           10
  Veneered
    hardground            15            32           10
  Terrigenous sand         0            20            0
Burial
  Exposed                 30            30           30
  Dusted                  10            37            0
  Partially buried         0             0           10
  Buried                  25            35           40
Depth range
  <50 m                   35            30           15
  50-<100 m               10            20           50
  100-<200 m               0            20            0
  200-<300 m              20            32            0
  [greater than or
    equal to] 300 m        0             0           15

                      Turritella   Mercenaria   Rhinoclavus
                       terebra     mercenaria    vertagus

Ocean basin
  Bahamas                  5           103          91
  Gulf of Mexico         184            20           0
Sediment type
  Brine                   20             0           0
  Carbonate sand          35            59          53
  Hardground              30             0           0
  Terrigenous mud         64             0           0
  Carbonate mud            0            10           9
  Veneered
    hardground            20            34          29
  Terrigenous sand        20            20           0
Burial
  Exposed                 59            24          18
  Dusted                  10            50          39
  Partially buried        45            10          10
  Buried                  75            39          24
Depth range
  <50 m                    5            20          19
  50-<100 m              100            14           9
  100-<200 m              20            30           9
  200-<300 m               0            59          54
  [greater than or
    equal to] 300 m       64             0           0

TABLE 3.
The simplified semiquantitative scales used in statistical analysis
derived from Davies et al. (1990). Data were originally recorded
using their more complex scales and then simplified in final
analysis. Note that periostracum loss, being a loss rather than a
gain, has the highest value assigned to the pristine attribute.

Taphonomic                                                Numerical
attribute                    Taphonomic state               value

Wholeness           Shell whole                               1
                    Shell broken                              2
Abrasion            Unabraded                                 0
                    Small nicks; frosted; sculpture           1
                      eroded, no holes
                    Highly polished; deeply eroded            2
                      with holes
Edge rounding       Edges natural                             0
                    Edges chipped                             I
                    Fragmented; edges sharp                   2
                    Fragmented; edges slightly worn           3
                    Fragmented; edges smooth                  4
Periostracum Loss   Missing                                   1
                    Partially missing                         2
                    Present                                   3
Discoloration       Original                                  1
                    Original, but faded                       2
                    Original color faded to white             3
                      or discolored (e.g., gray)
Dissolution         Undissolved                               0
                    Surface chalky; minor pitting             1
                    Surface pitting moderate to heavy;        2
                      surface soft; sculpture enhanced
                    Surface sculpture gone; deeply            3
                      dissolved

TABLE 4.
Taphonomic attributes treated as present (1) or absent (0)
in statistical analysis.

Taphonomic
attribute           Taphonomic state

Dissolution style   Chalkiness
                    Pitting
                    Soft surface
                    Deeply dissolved, surface gone
                    Deeply dissolved, surface gone,
                      shell perforated
Discoloration       Original color
                    Original color faded or faded to white
                    Gray, light to dark; Black or black mottled
                    Brown to red, brown mottled
                    Orange, orange mottled, orange mottled
                      with black or brown or white
                    Green, green mottled, green mottled with
                      black, brown or white

TABLE 5.
Example ANOVA results with EODs defined based on visual
details of habitat, biota, and sediment type. NS, not significant
at [alpha] = 0.05. Factor 1: maximum degree of dissolution, incidence
of deep dissolution, average degree of dissolution (outer
shell or spire). Factor 2: incidence of brown discoloration,
maximum degree of discoloration, incidence of original coloration
(negative). Factor 3: incidence of soft shell surface, average
degree of dissolution (inner shell surface or body whorl), incidence
of deep dissolution with holes, incidence of chalkiness.
Factor 4: incidence of faded coloration, incidence of brown
discoloration (negative).

                    Factor 1   Factor 2   Factor 3   Factor 4

EOD                  0.0001     0.0087     0.0001     0.0001
Species                --       0.0042       NS       0.0093
EOD*Species          0.0020     0.0051       NS       0.0004
Depth                  --         NS         NS         NS
Depth*EOD            0.0001     0.0081     0.0040     0.0001
Depth*Species          --         NS         NS       0.0022
Depth*EOD*Species    0.0006     0.0002       NS       0.0035

TABLE 6.
ANOVA results using environments of preservation for Factor 1:
maximum degree of dissolution, incidence of deep dissolution,
average degree of dissolution (outer shell or spire). NS, not
significant at a = 0.05. For Turkey Studentized Range tests,
different letters indicate significant differences at [alpha] = 0.05.
Sites are listed according to their rank order from highest
taphonomic index to lowest.

  ANOVA Results                     Tukey Studentized Range Test

EOP             NS   A                                       EFGCS
Species         NS   A   B                                   BACREST
EOP*Species     NS   A   B   C                               BA30M
Depth           NS   A   B   C   D                           BAWALL
Depth*EOP       NS   A   B   C   D                           BA183M
Depth*Species   NS   A   B   C   D                           BATALUS
Depth*EOP*
  Species       NS   A   B   C   D                           BATROFF
                     A   B   C   D                           GC234
                     A   B   C   D   E                       GB425
                         B   C   D   E   F                   PARKR
                         B   C   D   E   F                   AA15M
                             C   D   E   F   G               EFGCF
                             C   D   E   F   G               EFGDS
                                 D   E   F   G               STXSH
                                 D   E   F   G               EFGBR
                                     E   F   G               EFGHD
                                     E   F   G               EFGCM
                                         F   G   H           NORMP
                                             G   H   I       BA15M
                                                 H   I   J   AATROFF
                                                     I   J   AATALUS
                                                     I   J   AACREST
                                                         J   AAWALL
                                                         J   AA30M


TABLE 7.
ANOVA results using environments of preservation for Factor 2:
incidence of brown discoloration, maximum degree of
discoloration, and incidence of original coloration (negative).
NS, not significant at [alpha] = 0.05. For Tukey Studentized Range
tests, different letters indicate significant differences at
[alpha] = 0.05. Sites are listed according to their rank order
from highest taphonomic index to lowest.

ANOVA Results                    Tukey Studentized Range Test

EOP             0.0001   A                                   AACREST
Species           NS     A                                   BA15M
EOP*Species       NS     A   B                               EFGCM
Depth             NS     A   B                               EFGCF
Depth*EOP         NS     A   B                               AATROFF
Depth*Species     NS     A   B   C                           BA183M
Depth*EOP*
  Species         NS     A   B   C   D                       GC234
                         A   B   C   D                       NORMP
                         A   B   C   D   E                   PARKR
                         A   B   C   D   E                   AATALUS
                         A   B   C   D   E                   EFGDS
                         A   B   C   D   E                   GB425
                         A   B   C   D   E                   BATALUS
                             B   C   D   E   F               BAWALL
                             B   C   D   E   F               AA15M
                             B   C   D   E   F               EFGHD
                                 C   D   E   F   G           BACREST
                                 C   D   E   F   G           BATROFF
                                     D   E   F   G           BA30M
                                         E   F   G           STXSH
                                             F   G           AAWALL
                                                 G   H       EFGCS
                                                     H       EFGBR
                                                         I   AA30M

TABLE 8.
ANOVA results using environments of preservation for Factor 3:
incidence of soft shell surface, average degree of dissolution (inner
shell surface or body whorl), incidence of deep dissolution with
holes, incidence of chalkiness. NS, not significant at [alpha] = 0.05.
For Tukey Studentized Range tests, different letters indicate
significant differences at [alpha] = 0.05. Sites are listed according
to their rank order from highest taphonomic index to lowest.

   ANOVA Results                  Tukey Studentized  Range Test

EOP           0.0038   A                                       EFGCS
Species         NS         B                                   GC234
EOP*Species     NS         B   C                               AATROFF
Depth           NS             C                               AACREST
Depth*EOP       NS             C   D                           AA15M
Depth*
  Species       NS             C   D                           GB425
Depth*EOP
  Species       NS             C   D   E                       AATALUS
                               C   D   E   F                   EFGCF
                               C   D   E   F   G               EFGCM
                               C   D   E   F   G               EFGBR
                                   D   E   F   G   H           EFGDS
                                   D   E   F   G   H   I       AAWALL
                                   D   E   F   G   H   I       STXSH
                                   D   E   F   G   H   I       AA30M
                                       E   F   G   H   I   J   BATROFF
                                           F   G   H   I   J   EFGHD
                                               G   H   I   J   NORMP
                                               G   H   I   J   BA183M
                                               G   H   I   J   BATALUS
                                                   H   I   J   BACREST
                                                   H   I   J   BAWALL
                                                       I   J   BA15M
                                                       I   J   BA30M
                                                           J   PARKR

TABLE 9.
ANOVA results using environments of preservation for Factors
4-6. NS, not significant at [alpha] = 0.05. Factor 4: incidence
of faded coloration, incidence of brown discoloration (negative),
Factor 5: edge rounding, breakage. Factor 6: incidence of pitting,
incidence of orange discoloration.

                     Factor 4   Factor 5   Factor 6

EOP                     NS         NS         NS
Species                 NS         NS         NS
EOP*Species             NS         NS         NS
Depth                   NS         NS         NS
Depth*EOP               NS         NS         NS
Depth*Species           NS         NS         NS
Depth* EOP*Species      NS         NS         NS

TABLE 10.
ANOVA results using environments of preservation for Factor 7:
incidence of green discoloration. NS, not significant at
[alpha] = 0.05. For Tukey Studentized Range tests, different
letters indicate significant differences at [alpha] = 0.05.
Sites are listed according to their rank order from highest
taphonomic index to lowest.

  ANOVA Results                   Tukey Studentized Range Test

EOP             0.036   A                                    AA15M
Species          NS     A   B                                AAWALL
EOP*Species      NS         B   C                            BA15M
Depth            NS         B   C    D                       PARKR
Depth*EOP        NS             C    D   E                   EFGCM
Depth*Species    NS                  D   E   F               BAWALL
Depth*EOP*
  Species        NS                  D   E   F   G           EFGHD
                                         E   F   G   H       EFGDS
                                         E   F   G   H       BA30M
                                         E   F   G   H       AA30M
                                         E   F   G   H   I   BATROFF
                                         E   F   G   H   I   BACREST
                                             F   G   H   I   NORMP
                                             F   G   H   I   STXSH
                                             F   G   H   I   EFGBR
                                             F   G   H   I   BATALUS
                                             F   G   H   I   EFGCF
                                             F   G   H   I   EFGCS
                                             F   G   H   I   GC234
                                                 G   H   I   AATALUS
                                                 G   H   I   BA183M
                                                 G   H   I   GB425
                                                     H   I   AATROFF
                                                         I   AACREST

TABLE 11.
ANOVA results using environments of preservation for Factor 8:
incidence of gray/black discoloration. NS, not significant at [alpha] =
0.05. For Tukey Studentized Range tests, different letters indicate
significant differences at [alpha] = 0.05. Sites are listed according
to their rank order from highest taphonomic index to lowest.

  ANOVA Results                       Tukey Studentized Range Test

EOP           0.032    A                                       AACREST
Species       0.0034   A                                       BA183M
EOP*Species     NS     A                                       BA15M
Depth           NS     A   B                                   BACREST
Depth*EOP       NS     A   B   C                               AATROFF
Depth*
  Species       NS     A   B   C                               NORMP
Depth*EOP*
  Species       NS         B   C   D                           BA30M
                           B   C   D                           BAWALL
                           B   C   D   E                       AATALUS
                           B   C   D   E   F                   BATALUS
                           B   C   D   E   F   G               BATROFF
                               C   D   E   F   G   H           EFGCS
                               C   D   E   F   G   H   I       AA15M
                                   D   E   F   G   H   I   J   PARKR
                                   D   E   F   G   H   I   J   AA30M
                                   D   E   F   G   H   I   J   STXSH
                                   D   E   F   G   H   I   J   GB425
                                       E   F   G   H   I   J   EFGBR
                                           F   G   H   I   J   AAWALL
                                               G   H   I   J   EFGHD
                                                   H   I   J   EFGCF
                                                       I   J   GC234
                                                           J   EFGDS
                                                           J   EFGCM

TABLE 12.
ANOVA results using environments of preservation for the residual of
wet weight. NS, not significant at [alpha] = 0.05. For Tukey
Studentized Range tests, different letters indicate significant
differences at [alpha] = 0.05. Sites are listed according to their
rank order from highest taphonomic index to lowest.

ANOVA Results     Tukey Studentized   Range Test

EOP              NS     A                   AATROFF
Species          NS     A   B               BATROFF
EOP*Species      NS     A   B   C           AACREST
Depth            NS     A   B   C   D       BA183M
Depth*EOP       0.027   A   B   C   D       EFGCS
Depth*Species    NS     A   B   C   D   E   AATALUS
Depth*EOP*
  Species       0.20    A   B   C   D   E   BA30M
                        A   B   C   D   E   EFGHD
                        A   B   C   D   E   EFGCM
                        A   B   C   D   E   BACREST
                        A   B   C   D   E   AAWALL
                        A   B   C   D   E   EFGCF
                        A   B   C   D   E   BATALUS
                        A   B   C   D   E   STXSH
                        A   B   C   D   E   BAWALL
                        A   B   C   D   E   AA15M
                        A   B   C   D   E   GB425
                            B   C   D   E   AA30M
                            B   C   D   E   PARKR
                                C   D   E   NORMP
                                C   D   E   EFGDS
                                    D   E   GC234
                                    D   E   BA15M
                                        E   EFGBR

TABLE 13.
ANOVA results using sediment type and degree of exposure
as main effects. NS, not significant at [alpha] = 0.05. * indicates
not significant when the analysis was restricted to Mytilus edulis
and Strombus luhuanus only.

                          Factor 1   Factor 2   Factor 3   Factor 4

Sediment Type             0.006 *    0.0001     0.0001     0.0001
Species                   0.0001     0.0001     0.0001     0.0001
Sediment Type*Species     0.0001 *   0.0001     0.0001 *   0.0001
Degree of Exposure        0.04       0.0001     0.0014     0.0001
Exposure* Sediment Type   0.0001        NS      0.0001     0.0001
Exposure* Species         0.0001 *   0.0001     0.0001 *   0.0001 *
Exposure *Sediment
  Type*Species            0.0001 *      NS      0.0001 *   0.017 *

                          Factor 5   Factor 6   Factor 7   Factor 8

Sediment Type             0.0064     0.0001     0.0001     0.0001
Species                   0.0001     0.0001 *   0.0001     0.0001
Sediment Type*Species     0.0001     0.0001        NS      0.0001
Degree of Exposure           NS      0.0001     0.0003*    0.0001
Exposure* Sediment Type   0.0001     0.0002     0.0001     0.0027
Exposure *Species         0.0006 *   0.0008 *   0.0001 *   0.0001 *
Exposure* Sediment
  Type*Species            0.0005     0.011      0.014      0.0001 *
COPYRIGHT 2008 National Shellfisheries Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Powell, Eric N.; Callender, W. Russell; Staff, George M.; Parsons-Hubbard, Karla M.; Brett, Carlton
Publication:Journal of Shellfish Research
Article Type:Report
Geographic Code:1MEX
Date:Mar 1, 2008
Words:21594
Previous Article:Interrelationships between vent fluid chemistry, temperature, seismic activity, and biological community structure at a mussel-dominated, deep-sea...
Next Article:The taphonomic signature of a brine seep and the potential for burgess shale style preservation.
Topics:


Related Articles
Better ceramics through biology.
Shell shuffle.
The taphonomic signature of a brine seep and the potential for burgess shale style preservation.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters