Mass analysis is a form of flake debris analysis that focuses on size, shape, and cortex characteristics of batches of flake debris as a means for measuring and quantifying variation in flakes debris aggregates (Ahler 1989a, 1989b; Ahler and Christensen 1983). In the mass analysis approach, all flake debris is size-graded by passing flakes through a series of nested screens of varying mesh sizes (1/4 inch, 1/2 inch, 3/4 inch, and 1 inch screens are used by Cultural Resource Analysts). The flakes in each size grade are then counted, weighed, and the number of cortical flakes recorded for raw material and provenience (i.e., unit and level). Cortical flakes consist of flakes with cortex, in any amount, on the platform, dorsal face, or both. This raw data is then used to determine a series of new variables (percentage of flakes in each size computed over all sizes, percentage of weight in each size computed over all sizes, average weight of flakes in each size grade, and the percentage of cortical flakes in each size) to be used in the analysis. Ahler (1989b:102) provides a list of attributes determined from mass analysis data that can be potentially useful for distinguishing various forms of reduction.
Ahler (1989b:89) notes two general theoretical observations regarding flintknapping that are relevant to mass analysis: 1) flintknapping is a reductive technology; therefore, there are predictable and repetitive size constraints on the byproducts, and 2) variation in load application (e.g., percussor used, placement of load) in the flintknapping procedure produces corresponding variations in flake size and shape. These observations build on principles of fracture mechanics. Cotterell and Kamminga (1987) define three major fracture initiations (flake types)
Conchoidal fractures (termed cone fractures by Tsirk 1979) can only be formed by a comparatively hard indenter such as a hammerstone (Cotterell and Kamminga 1987:686). Bending fractures (Cotterell and Kamminga 1979, 1987; Tsirk 1979) occur most often with soft hammer percussors such as antler and bone or in pressure flaking. Bipolar flaking often produces compression flakes. While these flakes are most commonly produced by the percussor noted, other factors also come into play. For example, hard hammers can produce bending flakes on edges with low edge angles (Cotterell and Kamminga 1987:689).
The main differences between the above fracture initiations, as viewed by the archaeologist, are the size and shape of the resulting flakes. In addition, as the parent piece (core or tool) is further reduced, there are certain restrictions on the size of flakes that can be removed. Differences in the size and shape of hard hammer percussion and pressure flaking, for example, are obvious. Subtle differences can also be examined for reduction modes that are not so extreme (e.g., hard hammer biface edging versus soft hammer biface thinning).
Application of mass analysis to flake debris from controlled experimental reductions often provides the baseline for the interpretation of archaeological samples. In the experiments, a number of replications (e.g., core, biface, etc.) are conducted. The resulting debris is then mass analyzed. Following this, a multivariate statistical method known as discriminant function analysis is conducted to determine the best way to separate the different experimental reductions based on the mass analysis variables. This information can then be applied to archaeological samples to provide an estimate of the technological/manufacturing stages represented by the archaeological samples. Even without the use of an experimental assemblage, mass analysis can provide general indications of the technologies present. Several general trends have been noted in mass analysis experiments (e.g., Ahler 1989a, 1989b; Ahler and Christensen 1983; Bradbury and Franklin 2000) that are applicable to any assemblage classified using this method. These general trends are
- As reduction continues, the average weight of flakes decreases
- As reduction continues, the percentage of flakes retained in size grade 2 (1/4 inch) increases
- As reduction continues, there is a decrease in the percentage of cortical flakes
In addition, discriminant functions derived from an experimental data set can be applied to an archaeological assemblage to provide additional information concerning the flake debris. These can also be checked using simple plots of the data along with the experimental data set.