136 F. B. Ltiquiens, Chanson de Roland. 



to be a metrical experiment on the part of some copyist — technical 



considerations remove all doubt. ^ In short, to Miiller's formula should 



be added : Exclude from the Oxford manuscript whatever may be 



proved due to copyists. 



« * * 



If the foregoing arguments are cogent, the correct and complete 

 formula for the reconstruction of the original Chanson de Roland is : 

 (1) Never alter the Oxford manuscript to accord with the other 

 redactions except for an imperative reason ; (2) exclude from the 

 Oxford manuscript whatever may be proved due to copyists.^ 



Yale University, Frederick Bliss Luquiexs. 



1 Cf. supra^ p. 123. 



2 It is the author's intention to publish a series of articles setting forth 

 in detail (1) the instances in which there is an imperative reason for 

 altering O to accord with the other redactions ; (2) the instances in which 

 O may be proved to be due to copyists. 



