Following our study of unpublished manuscripts of Hunain's translation of Galen's De Sectis ad eos qui introducuntur, a report of which appeared in JHS xcviii (1978) 167–9, we turned our attention to Hunain's translation of Galen's Ars Parva. Once again we have been impressed by the generally excellent quality of the Arabic version and we record a number of passages relevant to the problem of establishing the original Greek text. Perhaps our most important result is to confirm the existence, in Hunain's version of the Greek text, of an extensive preface including explanations of the terms ‘analysis’ and ‘synthesis’ which are missing from the printed text of Kϋhn.