Sampling and its effect on the composition of archaeobotanical assemblages from the Roman Period

The subject of the presented study is taken from a dissertation project by one of the authors who focused on the processing of archaeobotanical assemblages from the Roman Period. The main aim of the research was the reconstruction of selected aspects of the subsistence strategy of the population in the given period based on the evaluation of archaeobotanical data from various chronological and cultural contexts in a designated region, available to author. The analysed sets were obtained during field excavations primarily conducted in the last third of the 20th century and the beginning of the 21st century. Uniform methods of archaeobotanical sampling were not applied in the acquisition of these assemblages. Source information on the origin of the samples was considerably heterogeneous and, in many cases, distinctly fragmentary. This was the impulse behind the investigation into the question as to whether, and to what extent, the method of sampling affects the interpretive value of the investigated dataset and what are the limitations of the analysis of such a dataset. The principal aim of this study is not the archaeobotanical evaluation of samples, but rather to investigate a possible effect of their formal properties on the composition of archaeobotanical finds. The formal properties studied include the volume and the number of collected samples, and the spatial stratification of samples (context/feature). Intuitively, it would appear that the heterogeneous quality of this information may have a certain impact on the interpretive value of an archaeobotanical assemblage. We discuss the effect of the chosen method of sampling on the composition of macro-remains in archaeobotanical samples and assemblages with the use of statistical models.