IN TRYING TO piece together a coherent account of English secular and ecclesiastical history in the century after Bede, William was forced back upon the Anglo-Saxon Chronicle, supplemented by such miscellaneous sources as he could find either in his local library or in the course of his extensive bibliographical travels.Among them, charters and letters occupy an important place, and for a good many of these William is a valuable textual witness. This is true for letters of Aldhelm, Boniface and, to a greater extent than has been realized, of Alcuin.

In the first book of the Gesta Regum Anglorum William uses twelve of Alcuin's letters. Five of these reappear in the Gesta Pontificum, plus another four. I propose to study this material with two goals in view: firstly, to identify and describe the manuscript or manuscripts which William used and to assess his contribution to the reconstruction of Alcuin's text; secondly, to analyse and discuss the editorial treatment to which William subjected these letters in the process of incorporating them into his two major historical works.

Long ago Bishop Stubbs stated that William must have used two extant manuscripts of Alcuin's correspondence, BL Cott. Vesp. A. xiv (Dümmler's A2) and Cott. Tib. A. xv (A1). A2 was made for Archbishop Wulfstan (1002–23), perhaps at York; the provenance of A1 is uncertain. Stubbs argued solely on the basis of the letters contained in the Gesta Regum, some of which are found in no other manuscripts but A1 and A2. It is odd that he did not take into account the letters in the Gesta Pontificum or collate William's extracts with the most recent edition available to him, that of Jaffé and Wattenbach. But even on the basis of which letters are quoted in the Gesta Regum there was no reason to suppose that William knew A2, and the suspicion that he did not is confirmed by a scrutiny of the letters in the Gesta Pontificum and a collation of all of them with the Monumenta edition. In that edition Ernst Dümmler made somewhat desultory use of William's extracts without himself adding anything to Stubbs's argument.