The purpose of this article is to analyse the different categories of the lexical and syntactic variation which is common to two unpublished manuscripts of King Alfred's translation of Gregory's Cura Pastoralis, Cambridge, Trinity College R. 5. 22 (717), fols. 72–158, and Cambridge, University Library, Ii. 2. 4, referred to henceforth as T and U respectively. These manuscripts are an obvious choice for the following reasons: the relationship of the tenth- (or early-eleventh-)century T to the eleventh-century U (almost certainly written at Exeter) can be at least tentatively established as far as the history of the transmission of the test is concerned; the work has the status of royal origin and one can therefore expect it to have been treated with due respect by later copyists; and the work is of considerable length. My study is based on a complete collation of both manuscripts, since, in any analysis of this kind, numerical weight of evidence lends conviction to the conclusions, but the material, because of its quantity, cannot be presented here in its entirety. What I have attempted to do is to demonstrate, with selected examples as evidence, the clear pattern of variation that emerges. It is hoped that as a result some light is thrown upon linguistic developments within West Saxon and upon the scribal practices involved in the copying and recopying of earlier texts in late Old English times.