Previous work on the Mesolithic of southern Germany has revealed a sharp decline in the number of sites known from the Early to the Late Mesolithic. The apparent decrease in site numbers may indicate a population decline, or alternatively, an adaptive change in settlement patterns that is less visible archaeologically. Current understandings of the Mesolithic in southern Germany rely heavily upon information provided by a few excavated sites, mostly in caves or on lake shores. This reliance upon only excavated sites has created a large bias in the way archaeologists view the Mesolithic. The lack of regional survey for much of southwest Germany has contributed greatly to this problem.

Surface sites can provide a much broader view of settlement patterns as well as give a more accurate view of the number of sites present for specific time periods. However, currently there are few available methods for dating surface assemblages beyond listing artifacts diagnostic for a particular time period. In many cases diagnostics are not available or do not accurately represent the time periods present at a site. These problems with the available data and current methods for dating surface sites make it impossible at present to determine if a population decline did occur.

An analysis of lithic debitage from well excavated assemblages within the project study area is used to evaluate differences between the Late Paleolithic, Early Mesolithic, and Late Mesolithic of southwest Germany. This data is then used to construct a classification model to date lithic assemblages. Seven test sites resulting from the Southwest German Regional Archaeological Survey Project are presented and classified to time period using the mode, which in turn reveals that analysis of diagnostic artifacts alone does not provide an accurate dating of surface assemblages. Only a thorough review of both diagnostic tools and debitage can accurately assign a potentially mixed surface assemblage to the correct time period or periods. With the inclusion of surface sites in our data base for the Mesolithic a more complete settlement pattern analysis will be possible that will answer the puzzle of site number decline in the Late Mesolithic.