The only significant occasion that Cremna emerges from the shadows of historical obscurity is during its siege by Roman forces in the third century AD. Our sole surviving source for this incident is the account related by Zosimus which, despite covering two generous paragraphs of text, remains a collage of anecdotal observations of limited application to the student of siege operations. Notwithstanding the weaknesses of the literary evidence, an archaeological survey conducted with the assistance of the British Institute of Archaeology at Ankara in the mid 1980s, revealed that substantial elements of the Roman siege system still survived in situ. The recently published account of this survey (Mitchell et al 1995) presents a number of interesting arguments concerning the tactical role of these siege works and their relationship to the local topography. In the course of more general research into Roman siege works, the author undertook a field inspection of the site in September 1997 and would suggest that certain aspects of the published analysis require further refinement. This paper is intended to address these outstanding issues. The helpful comments made by Professors S Mitchell and J J Wilkes and Dr J J Coulton in respect of an earlier draft, are gratefully acknowledged.