Mesh-size effects on the ecological fidelity of death assemblages: a meta-analysis of molluscan live–dead studies

Actualistic comparison of death assemblages with local living communities is a standard approach to estimating the quality of paleoecological data, but wide variation in methods of data collection and analysis undermines attempts to draw general conclusions. Here, I apply both standard and meta-analytic statistics to a stringently constructed database of 19 molluscan live–dead studies in order to isolate methodological artifacts from true taphonomic bias, focusing on three paleoecologically relevant aspects of live–dead agreement: (1) the preservation potential of shelled mollusks in their life habitat (percentage of live species that are represented among dead shell material), (2) the habitat-level spatial fidelity of skeletal remains (percentage of dead individuals that are from species also collected alive in the same habitat), and (3) the reliability of species dominance (similarity in the rank-order abundance of live and dead species). Overall, agreement for all three metrics is high, but results are sensitive to dataset size, sediment grain size, and, most importantly, sieve mesh size. Coarse-mesh (sieves ≥ 1.5 mm) datasets are composed of late juvenile and adult specimens and have significantly higher live–dead agreement than fine-mesh datasets (sieves ≤ 1 mm), which include, and are probably dominated numerically by, larvae and newly settled juveniles. Dataset-size effects are present but not significant, and sediment grain-size effects are significant only if datasets are first partitioned by mesh size (live–dead agreement is higher in muds than in sands/gravels). This previously unrecognized mesh-size effect makes sense both ecologically and taphonomically, and identifies a simple protocol for isolating the most reliable information in molluscan death assemblages (i.e., focus on specimens ≥ 1.5 mm). The pervasiveness and magnitude of the effect indicates that mesh size needs greater consideration in future taphonomic studies and in the collection and interpretation of (paleo)ecological data. A post-juvenile focus may also be key to isolating high-fidelity data among other metazoan groups. © 2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.