Fallow Deer in Prehistoric Greece, and the Analogy between Faunal Spectra and Pollen Analyses

Article excerpt

Manipulating pollen and animal-bone records

Pollen analyses may have contributions from the pollen of trees, herbs, aquatic plants, fern and moss spores, and other plant groups and sub-groups. While the overall composition is of interest, the contributions of the different components often interfere with each other in percentage calculations. Individual pollen types may be excluded if their behaviour is grossly aberrant - whether for natural or unnatural reasons. For instance, bracken (Pteridium) and ivy (Hedera) are over-represented in some archaeological pollen analyses, and cause a dent in the other pollen curves which can be removed by excluding them from the calculations. Raw pollen counts are sub-divided and turned into percentages in ways that seek to minimize perturbing effects, and thereby to show underlying processes more clearly. To the unititiated, this may seem analogous to that famous oxymoron 'creative accountancy', but in fact it is respectable and legitimate.

The desire to define 'Minimum Numbers of Individuals' and to establish age-distributions encourages archaeozoologists to examine their raw evidence critically. However, they have neglected the possibilities of faunal diagrams (usually used in the context of pre-domestic faunas) as a source of ecological evidence analogous to pollen diagrams.

Analyses of pollen and of animal bones have some things in common. Faunal analyses may contain representatives from a number of autonomous or semi-autonomous ecological groups: bird and fish bones may be present, as well as carnivore and small mammal remains (rodents and insectivores), in addition to the bones of large herbivores. Just as it makes no sense to a pollen analyst to describe shrubs as percentages of the fern spore sum, or trees in terms of the aquatics, so the composition of each zoological sub-group needs to be treated (as far as possible) in isolation. At the other extreme, sub-divisions can be too fine. Palaeontological studies carried out on a family-by-family basis can make an ecological context almost impossible to recover.

In principle, the overall composition of a faunal spectrum is as important and interesting as the palynological equivalent, which allows the changing contributions of the different sub-groups to be assessed at a glance. In practice, the differences between pollen grains and animal bones constrain the analogy.

As differential excavation recovery causes notorious biases (Payne 1972; 1975a), comparison of faunal spectra is restricted to stratified sequences of three or more layers, where similar excavation biases may affect each unit. Inter-site comparisons, likewise, are restricted to comparing stratigraphic series of faunal spectra, each subject to its own perturbing processes.

A major complication which the pollen analyst is spared concerns the separation of wild and domestic specimens of the same species.

Since any trait that occurs in a wild population can also appear in domesticates, it makes little genetic sense to identify a given bone as 'wild' or 'domesticated'. The problem can be attacked from an ecological perspective, however. What is necessary is, firstly, a reasonably objective basis for categorizing bones as 'wild' or 'domestic'; and secondly, at least one animal that can safely be assumed to be wild, with whose behaviour the putatively wild animals can be compared. Principal components analysis of the faunal spectra will distinguish between 'wild' and 'domestic' behaviour - but a little thought will do the job adequately, if less glamorously. In prehistoric non-urban sites at least, wild animals will reflect the surrounding environment and therefore show a consistent pattern of relations with each other, while domestic animals will tend to show biologically irrational patterns reflecting their owners' preferences.

It is commonly observed that domestication leads to smaller animals. This seems to be a genetic consequence of the tendency of smaller animals to need less food. …