FIDELITY

Paleozoologists and taphonomists have long recognized various properties of quantification and sampling with respect to collections they study. Those same properties attend samples of modern owl pellets. The particular skeletal elements identified and the way in which prey remains are grouped for tallying both influence measures of relative prey abundance in a collection of 56 barn-owl (Tyto alba) pellets from southeastern Washington. As the number of prey and the number of pellets in a collection increases across 107 published collections of North American barn-owl pellets, the richness of mammalian genera per collection increases. As the size of the most abundant prey taxon in a pellet collection decreases, the average number of individual prey per pellet increases. Pellets with more identifiable mammalian remains contain more individual prey. Larger pellets contain more individual prey than smaller pellets. These observations indicate that the properties of quantification and sampling so well known to paleozoologists can be created during the biostratinomic phase of a taphonomic history. Modern owl pellets are an excellent educational resource for teaching principles of taphonomy and zooarchaeology.

The transport and processing of large mammal carcasses by humans seems to provide a perfect data-set for the application of foraging theory. However, such applications in archaeology have generally been unsuccessful in that the results diverge widely from the predictions of foraging theory, and ethnographic applications have been rare and the results mixed. These applications require good estimates of skeletal element return rates, but to date we have insufficient net return rate data. Using some basic parameters we can rank skeletal elements by gross return rate, and classify them into high cost and low cost elements. We examine three of the best data-sets on hunter-gatherer skeletal element transport (Hadza, Nunamiut, and Kua), and find that the Nunamiut and Kua data diverge significantly from the Hadza data. We argue that this difference is not due to differences in skeletal element transport, but rather that the Hadza data-set represents observed instances of transport while the Nunamiut and Kua data-sets represent discarded bone assemblages that were scavenged by carnivores. Thus the Nunamiut and Kua sets represent a first stage in bone destruction after discard by people, and this would be followed by further destruction as such assemblages are transformed into archaeological samples. This result, when joined to taphonomic data on skeletal element survival, leads to a general model of bone survival that separates skeletal elements into two groups: 1) a low-survival set defined by a lack of non-cancellous thick cortical portions, and 2) a high-survival set defined by the presence of thick cortical bone portions lacking cancellous bone. The archaeological representation of the low survival set is primarily the product of post-discard destructive processes, and most low survival elements also belong to the high cost set. The relative abundance of the high survival elements in archaeological contexts is primarily the product of what was discarded after processing, and most of these belong to the low cost set. Foraging theory needs to be linked to the realities of skeletal element survival and destruction as understood in taphonomy, connecting the general and middle range theory, respectively. We need a synthetic taphonomic-foraging theory model, and we provide some foundations for that model here.

Whereas a majority of previous fidelity studies have dealt exclusively with mollusks, this study evaluates the compositional fidelity of mixed brachiopod-mollusk benthic assemblages sampled from the San Juan Islands area (Washington State, USA). A total of ca. 2500 live specimens and over 7500 shells and shell fragments were recovered from nine samples dredged along a subtidal transect. The shell material was dominated by fragments; less than 500 dead specimens were represented by complete valves or shells. The compositional fidelity was high: over 60% of live species and over 70% of live genera were also found in the death assemblage and over 60% of dead species and genera were represented in the life assemblage. These high numbers were consistent for all analyzed size fractions (2.3, 4, and 12mm). The life and death assemblages displayed a significant Spearman rank correlation (r = 0.41, p = 0.0001) suggesting that, despite the biasing action of taphonomic processes and time-averaging, the relative abundance of species in the original communities is at least partly preserved in the resulting death assemblages. The results also indicate that a restrictive analytical approach, with fragments excluded from the datasets, appears to provide more credible estimates of diversity and fidelity than an exhaustive approach, which included all fragments. Differences between the two analytical strategies most likely reflect the presence of several genera (e.g., Chlamys), which were readily identifiable from fragments (the five most abundant species in the exhaustive death assemblage were all identifiable from even small and heavily altered fragments). The “Chlamys effect” illustrates a general principle, because species often vary in their morphological distinctness, the inclusion of fragments is likely to notably distort the taxonomic composition of the studied death (or fossil) assemblages and may depress estimates of diversity and evenness. This study suggests that mixed brachiopod-mollusk associations are reasonably well preserved in the death assemblage in terms of taxonomic composition and rank abundance of dominant taxa. Moreover, despite considerable microstructural and compositional differences between brachiopod and mollusk shells, the class-level fidelity is excellent when fragments are excluded from the analysis. The results are highly congruent with patterns observed previously in fidelity studies focused exclusively on mollusks.