A statistical analysis of successional patterns in carrion-arthropod assemblages: implications for forensic...

A set of statistical protocols is proposed for analyzing carrion-arthropod succession in forensic entomology investigations. A total of 23 carrion-arthropod data sets from temperate, tropical, desert, and coastal environments were assembled in a standard format and analyzed using randomization tests and methods derived from quantitative community ecology. The data were analyzed in three ways. First, patterns of arthropod visitation on nonhuman carcasses were analyzed in each of the 23 cases. Analysis revealed two groups of taxa: those that persist on the carcass over a single time interval (= nonreoccurring taxa, 80% of the taxa) and those that appear, leave, and reappear over time on the carcass (= reoccurring taxa, 20%). Reoccurring taxa, which can confound estimation of the postmortem interval (PMI), were found in 6 classes, 12 orders, and 29 taxonomic families of arthropods, including some forensically important taxa (for example, calliphorids, sarcophagids, histerids). The recognition that reoccurring taxa exist and that they are found in forensically important groups is an early step toward factoring in their potential importance in future entomological investigations. Second, temporal changes in the taxonomic composition of the carrion-arthropod community were studied by quantifying the degree of taxonomic similarity between pairwise combinations of time-specific samples of the succession. In 13 of the 18 illustrated cases, the midsuccessional samples, owing to their higher species richness, were taxonomically more similar to all other pairwise samples, on the average, than early and late successional samples which were poorer in species. Variability in taxonomic composition is the norm for most periods of the succession; however, in 17 of the 23 cases, some successional periods (particularly endpoint samples) revealed no changes in arthropod species composition (= matching sample-pairs). When applied to medicolegal cases, it is suggested that data sets with large fractions of matching sample-pairs should produce wider-ranging PMI estimates, on the average, than data sets with smaller fractions of matching sample-pairs. Third, Monte Carlo simulation was applied to each of the 23 assemblages to test specific hypotheses about community-wide patterns of arthropod visitation times on nonhuman carcasses. Simulation results revealed that arthropod residence times in the majority of cases (13 or 56%) followed a "clumped" succession model, whereas, the remaining 10 cases (44%) showed a more "uniform" spacing pattern of residence times on carcasses. Comparison of species accumulation curves for observed and simulated data further revealed that among the 13 "clumped" cases, most (9 or 69%) followed a "clumped, early" model (rather than "clumped, midterm" or "clumped, late" models).(ABSTRACT TRUNCATED AT 400 WORDS)