This information has been published in the Journal of Field Archaeology (JFA) in an article that can be read online at this link.
At KAP, we use a tripartite, nested system of ID numbers to identify the spaces (excavation areas and contexts) and objects (samples) recovered during excavation. The IDs for contexts and samples can be thought of as separate but related numbering systems.
Contexts (whether spatial, approximate, or group) are directly or indirectly associated with spatial data. The most common form of context is the spatial context, which corresponds to a spatially discrete, three-dimensional volume that contains a deposit or a feature. Group contexts refer either to features that are expected to remain in situ over the long term or to interpretive groupings of spatial contexts; they derive their spatial data from the spatial contexts they contain. Approximate contexts correspond to an object of interest that was recovered near its place of origin, but may no longer be in situ. Approximate spatial contexts are therefore given spatial data in the form of point coordinates that approximate the original find spot of the object. Contexts can be grouped, but cannot be subdivided.
Samples are not associated with spatial data, but can be thought of as “things that receive actions” that are located with volumetric spaces defined by contexts. They may be artifacts, ecofacts, bones, sediment samples, etc. Samples can be grouped or subdivided and may also contain multiple discrete objects (e.g., a bulk sample of ceramics, or a reconstructed vessel).
Excavation Area ID #
Excavation areas (EAs) are named in a way that is spatially significant, based on values derived from the truncated UTM coordinates (eastings and northings) of their centroids (formerly their southwest corners). Therefore, an excavation area with a centroid at truncated coordinates E24 and N534 would be given the identification of EA 24.534.
Context Numbers and the Context Key #
Contexts are numbered in a way that is not spatially significant; contexts within an excavation area are named sequentially in the order they are opened (and preferably, in the order they are excavated). This numbering continues from year to year within each excavation area, rather than restarting each year at 1. The database should prevent any instances in which a context number is used twice or skipped, but you should also remain vigilant that this does not happen.
All types of contexts are named in the same sequence. Therefore, you might encounter a situation in which context 98 is a spatial context, context 99 is an approximate context, and context 100 is a group context. Similarly, if a spatial context is subsequently included in a group context, no new number is given to either of the contexts. Rather, the record of the association between the contexts is maintained in the database through descriptions and through the creation of a “parent-child” relationship between the group context (the “parent”) and the spatial context (the “child”) or contexts (the “children”) they contain.
The context key refers to the combination of an excavation area ID and a context number that creates a unique identifier for each context. So, context number 286 in excavation area 99.526 is referred to by context key 99.526.286.
Samples, like contexts, are named sequentially within the unit that they belong to, that is, the context they are part of. Every sample should be associated with one context (either spatial or approximate); group contexts do not contain samples (but only “child” contexts). Samples can be subdivided after their original collection, in which case the original sample is redefined to describe its current contents, and a new sequential sample number is opened for the other part of the original sample. As is the case for contexts, the database should ensure that sample numbers are never duplicated or skipped. Nonetheless, be vigilant to prevent this from happening.
Each sample derives its unique identifier composed of a combination of its excavation area ID, context number, and sample number. So, the first sample from context 286 in area 99.526 is referred to as 99.526.286.1.
It is also possible to group samples for interpretive purposes, both across contexts and across excavation areas. For example, if sherds belonging to a single ceramic vessel were found in a context surrounding a pit (here, context 1), and then more sherds from the same vessel were found in the pit fill (here, context 2), each set of sherds could be made into a discrete sample in their own context, and then grouped into a new group sample number in only one of the contexts from which the sherds were recovered. The decision about which context should contain the group sample might be based on which context is “dominant,” which contained the most sherds, or which context the sherds were recovered from first, etc. Again, you should not associate the group sample with both contexts, neither should you create group samples containing the sherds in both contexts.
Samples should be grouped or subdivided only when there is interpretive or analytical value in doing so.