Die babylonisch-assyrische Medizin in Texten und Untersuchungen. Band V: Keilschrifttexte aus Ninive 1

the Istamatis, but wrong in suggesting that the borrowing took place at a "pre-Arabic" stage (p. 69). The first of the parallels she points out (ibid.) is so close in the actual wording of the Arabic texts as to make it far more probable that it was the Arabic version of the Istamaits which was used in compiling the Arabic version of the Sirr. A similar query, but one of greater consequence, arises in connexion with (b). Weisser observes that there are parallels to Nemesius even in the shorter version A (p. 65). The most important of these parallels-let me call it NS-occurs at pp. 399.4-400.10 of her edition of version A. It is here that version B offers its longer extract from the De nat. hom., which I shall call NL. It is not entirely clear from her edition, which relegates NL to an appendix, whether B has NL in addition to NS or in its stead. The question is crucial, for NS is included in NL almost verbatim (a fact not sufficiently appreciated by Weisser). Hence if NS were duplicated in B, the secondariness of the longer version would, of course, be beyond doubt. But if, as one would expect, A simply has NS where B has NL, there is no earthly reason why A should not be an abbreviation of B. Indeed, this would then be the only natural assumption to make: for in no way can NS be regarded as anything but a minimally adapted extract from the beginning of NL (contra p. 65). In either case, the shorter version of the Sirr no less than the longer postdates the Arabic version of Nemesius on which they both draw. Prima facie, therefore, the most reasonable assumption would appear to be that the Sirr was first composed in Arabic, from sources including the Istamatits and a translation of the De nat. hom., by the translator of the latter. For the Sirr as a whole shares with NL some striking pecularities of diction (for examples see Ullmann, loc. cit., pp. 172f.). Minor differences (unduly emphasized by Weisser, p. 65) will be explained if we assume that much of the Sirr is a product of free composition. This would also go a long way to explain the deviations of NL from the text of Nemesius, which are clearly tendentious and cannot (pace W., p. 67) be passed off as …