ARISTOTLE 



for P.A.'^ Some years ago, when working on P. A. for 

 the Loeb edition, my examination of the ms. Z at 

 several places led me to state {P. A. Introd. p. 46) 

 that a more reliable collation of the chief mss. than 

 Bekker's apparahis criticus afforded was clearl)' 

 needed. This view is amply confirmed by K. E. 

 Bitterauf, who has in fact undertaken such a collation 

 for G.A. (see below), and he shows that there are 

 several errors and misleading reports on every page 

 in Bekker's apparatus, 

 (h) the A comparison of the text of P. A. exhibited by our 

 "'3^i^[*P^ Greek mss. with the translation of Michael Scot 

 showed me that the former had all suffered identical 

 corruptions or losses (or both) in certain passages 

 {e.g., P. A. 684 b 22 ff.), by which the Greek ms. from 

 which Scot's Arabic original was translated had not 

 been affected ; and I found exactly the same when I 

 came to work oh G.A. (see, e.g., 722 a 20, 766 b 35). 

 My conclusion about the common origin of our Greek 

 MSS. is also supported by Bitterauf, who comes in- 

 dependently to the conclusion, based exclusively 

 upon a study of the Greek mss., that they are all 

 derived from a single archetype, Avhich, in his opinion, 

 contained a number of variant readings. 

 Modorn This brings us to a consideration of the ms. tradition 

 X-^l)*Z of G.A. After the pubhcation of Bekker's Berlin 

 edition in 1831, very little work was done on the mss. 

 of G.A. for about eighty years. Bussemaker, who 

 edited G.A. in the Didot edition (Paris, 1854), cites 

 many readings from the two Paris mss. E and m, and 

 several times quotes the authority of William of 

 Moerbeke, less frequently that of Michael Scot, and 



" Of the other two, U does not contain G.A., and in E 

 G.A. is written in a later hand. 



