AGRICOLTURA E SCAMBI NELLʼITALIA TARDO-REPUBBLICANA - © Edipuglia s.r.l. - www.edipuglia.it HELEN GOODCHILD & ROBERT WITCHER MODELLING THE AGRICULTURAL LANDSCAPES OF REPUBLICAN ITALY* Introduction In recent years, some scholars of Roman antiquity have turned to comparative data and modelling techniques in order to ‘fill in’ some of the gaps in the historical and archaeological evidence. For example, de Ligt has used cross-cultural comparisons to illuminate the character of Roman markets 1 , and Roman demographers have made extensive use of life tables based on early modern populations to gain a better understanding of the structure of Roman populations 2 . Such techniques have been particularly prominent in the study of ancient agriculture 3 ; a number of recent works have used comparative data to model farming practice, with broad implications for conventional interpretations of the social and economic organization of Roman Italy. In this paper, we review two recent models of Roman agriculture presented by Nathan Rosenstein and Paul Erdkamp. Firstly, we identify the similarities and differences between these models in terms of starting assumptions and variables, and evaluate the impact of these differences on their interpretation and significance. We then take these models and locate them within a ‘real world’ environment. This involves confronting theoretical ideals with the complexities of a specific landscape and, in particular, with the evidence of archaeological field survey. We take a small * We would like to thank the conference organizers for their invitation to speak. We are also grateful to all the conference participants for their stimulating comments and ideas. Thanks to Helga Di Giuseppe and Helen Patterson for ongoing discussion of the archaeology of the Tiber valley. Thanks also go to Alice Hiley, Niall McKeown and Tony Wilkinson who read drafts and provided valuable comments, as well as Vince Gaffney for discussions of the methodology and GIS applications. 1 De Ligt 1993. 2 E.g. Frier 2001; Parkin 1992. 3 E.g. Gallant 1991.