Estimations of population density, which consider regional variability, are an important key variable in archaeology as they have consequences not only for the environmental but also for the economical and social domains. In this paper, a ten-step procedure of a consistent group of methods is described which deals with the data required for estimations of population density at different scale levels (from excavation to large-scale distribution maps). For distribution maps, a method is presented by which densities of sites are displayed using optimal isolines. These demarcate so called ‘settlement areas’ at scales of between 1:25,000 and 1:2.5 million. Our knowledge of the density of households from key areas with the most complete archaeological records is upscaled for the regions within these isolines. The results of this procedure are estimations of population density for the early Neolithic (Bandkeramik, 51st century BC) and the Roman period (2nd century AD) for regions with some 10,000 km2.
A simple statistical/graphical method is developed to analyse the relationship between settlement areas, soils, and precipitation. Taking into account the aspects of preservation of sites and the intensity of archaeological observations, an analysis of patterns of land use shows that in prehistory not all areas suitable for use were in fact incorporated into settlement areas. For prehistory, the idea of a most optimised use of land up to its carrying capacity (as it has been proposed for at least 50 years) can be falsified for specific areas. A large number of empty regions with good ecological conditions but lacking in settlement activity can be discussed as resulting from culture historical processes. As an example, the separation of areas inhabited by groups of different identities is discussed. The amount of used space (in terms of ‘settlement area’) however, increases from the early Neolithic to the 4th century BC from 5% to more than 40%. The increase between the Neolithic and the Iron Age is understood in terms of technological developments in farming systems. The percentage of areas with suitable conditions actually utilised between the Bandkeramik and Iron Age increases from 31.1% to 67.5% in the area covered by the Geschichtlicher Atlas der Rheinlande, and is much higher still in the Roman period (84.3%). State societies seem to use the land more efficiently compared to non-state systems. This is becoming even clearer on consideration of the intensity of human impact.
Large-scale distribution maps dividing the Neolithic in five periods were analysed. In each of the periods large settlement areas seem to be characterised either by the development of specific cultural innovations or by exchange of a specific raw material. In the course of time, the size of settlement areas in a specific region fluctuates markedly. It is most plausible to assume that this is due to a remarkable mobility of seemingly sedentary populations. Individual families recombine to new socio-cultural units every few hundred years.
The relationship between size of settlement areas and the number of households can be used to develop ideas relating to the flow of exchange goods. An example for the Bandkeramik considering the Rijckholt-Flint is presented. The combination of the number of households and the percentage of this raw material in the specific settlement areas visualises the amount needed and the amount transferred to other settlement areas in the neighbourhood. A future economical archaeology could use this information to develop ideas relating to the importance of the economic sector, ie, ‘procurement of flint’ in relation to the ‘production of foodstuffs’ according to the time required for each group of activities.
In the last section, the relationship between settlement areas and human impact is discussed. For the periods of subsistence economy, it is argued that the size of the population and its farming system are the two most important factors. For example, in Bandkeramik settlement areas, approximately 2% of the forest covering the landscape was cut down; in Roman times, and depending on the intensity of farming, this reaches magnitudes of between 20% and 50%. Although some of the methods and arguments used in this paper may be exchanged for better ones in the future, it is already apparent that a consistent system of methods is essential to transfer results of analyses on a lower scale level as input on a higher level and vice versa.