Much scholarship devoted to the study of the text of Chaucer's Canterbury Tales has focused on the Hengwrt and Ellesmere manuscripts, attempting to reconcile their many differences in the content and presentation of the poem. In concentrating on these two manuscripts, and a small group of other witnesses dated to the first quarter of the fifteenth century, scholars have largely ignored over forty complete manuscripts copied throughout the remainder of the century. Study of the manuscripts has relied on features external to the text of the poem itself in order to chart the development of the tradition, such as the order of tales, while details of text, language and metre have remained relatively unconsidered. The subject of this study is a manuscript that has been neglected by scholars due to its date of copying, c. 1430-50, and certain idiosyncracies in the tale-order. Despite these factors this manuscript contains a text closely related to that of Hg, the earliest extant copy of the poem. In addition to preserving an accurate copy of an early exemplar, Ad3 also shows close links with El, particularly in its ordering of the tales and the inclusion of marginalia. This is therefore an important copy of the poem, highlighting the restrictions and limitations of current attitudes to the textual tradition, and with much to offer as an independent witness to an early exemplar, with unique access to materials used in the production of both Hg and El. This study draws on recent technological developments, such as the availability of electronic versions of Middle English texts and collation software, in order to provide a detailed and comprehensive analysis of Ad3. In addition to this an electronic version of the text is included to enable further research of this kind.