Appropriate samples for rigorous, quantitative analyses of both morphological and paleoecological change were collected by one of the authors (CT) during an expedition to the Rio Gurabo and Rio Cana of the Dominican Republic in the summer of 2000. Bulk fossil sampling was employed where matrix and fossil materials were collected together without discrimination along restricted vertical and lateral intervals. Five sub-samples were taken at each locality; collectively these are referred to as one bulk fossil sample. These sampling localities are closely aligned to sampling localities of the Naturhistorisches Museum of Basel collection (Saunders et al. 1986).

By restricting the lateral extent from which the samples were collected, we minimize as much as possible the effect biological and taphonomic patchiness may have on assemblage composition. The average lateral extent of the samples was 1.02 m (median = 0.87 m); 26% of the samples extended less than 0.50 m horizontally. Similarly, by restricting vertical thicknesses, the time averaging represented within samples was minimized as much as possible. For the 31 bulk fossil samples collected from the Dominican Republic in 2000 (each with five sub-samples), the average thickness of samples was 0.156 m with the maximum thickness at 0.23 m.

The taphonomic condition of the fossil material is another proxy often used to assess time averaging (Best and Kidwell 2000). In all of the fossil samples we collected, there is little evidence of post-mortem boring or encrusting, thus suggesting that the fossils did not spend much time on the seafloor surface before being buried; virtually all borings in the samples are predatory borings, (Tang 2002). Not uncommonly, articulated corbulid bivalves are found and in several cases, very delicate spines (for example, on spondylid bivalves) and thin tubes of scaphapods and vermetid gastropods are present, again suggesting limited amounts of post-mortem exposure on the seafloor as well as limited transport. Preserved color patterns have been observed in associated Dominican Republic fossil samples (Costa et al. 2001). Thus, we interpret these fossil assemblages to be para-authochthonous, or representative of disturbed neighborhood assemblages (in this paper we will refer to these as "paleocommunities").

The two samples analyzed for this paper—060700-2A and 052500-1C—were collected from the Rio Gurabo section. Our fieldwork indicates that these samples are roughly equivalent to samples 18189 and 15907 of Saunders et al. (1986) respectively, in both stratigraphic and geographic location. Based on close correlation with the stratigraphic section prepared by Saunders et al. (1986), the samples are estimated to be separated by approximately 9 m vertically, with 060700-2A located at approximately 136 m in the section and 052500-1C at 127 m. Both are from within the NN 11 zone of the Late Miocene Cercado Formation. All specimens and unwashed sub-samples are deposited at the California Academy of Sciences Department of Invertebrate Zoology and Geology collections (San Francisco, California, USA).

The two samples analyzed are composed of skeletal grains in a matrix of poorly cemented siliciclastic grains. This type of muddy tropical depositional environment is considered to have minimal time averaging due to a higher rate of sedimentation and minimal physical reworking (Best and Kidwell 2000). Sample 060700-2A comes from blue-grey muddy sandstone horizons that contained thin shell beds, which were laterally continuous and traceable for some distance. The foraminifera Amphistigena was common in this sample as observed in the field, and it was also noted to contain medium to large bivalves, scaphapods, and gastropods. Sample 052500-1C was collected from a horizon containing an amalgamated, densely packed shell bed. In the field, it was noted that the siliciclastic and fossil components did not appear to be well size-sorted and that delicate molluscs were well preserved. Both samples appear to be storm-deposited shell beds and have similar lithological and taphonomic properties.

One complete sample bag from each sample analyzed was washed over a 2 mm sieve; most of the materials are fairly unconsolidated although occasional soaking and rewashing was necessary to disaggregate some portions of some samples. To clean individual fossils for accurate identification and morphometric analyses, specimens were soaked overnight in water and a sharp thin pick used to gently remove matrix particles.