The direct exemplars used by early printers show us much about the printers' working practices--whether and how they "cast off" copy; how they handled space and layout--and about their responses to Middle English literature in the years around 1500--how far it seemed to need updating; how closely it deserved to be transmitted. So, happily for the history of printing, literature, and language, there are several exemplars recognized from the presses of the second generation of English printers, Wynkyn de Worde and Richard Pynson. (1) But from the first generation, such as William Caxton or the printers in Oxford and St. Albans, fewer exemplars are known. From Caxton's press, we know only the exemplar for Lorenzo Traversagni's Nova rhetorica (1478), and from the Oxford press of "the printer of Rufinus," who came there from Cologne, only that for Rufinus's Symbola apostolorum (1478). (2)
Besides these Latin texts, only one English manuscript has been firmly recognized as even spending time in Caxton's workshop--the "Winchester" manuscript of Malory's Arthurian writings--and this was seemingly not the exemplar for Caxton's 1485 edition. (3) It has also been suggested but ultimately doubted that Oxford, Magdalen College, MS 213 was a collateral source for Caxton's 1483 edition of Gower's Confessio Amantis, for checking against rather than printing from; (4) and it has been hypothesized but admitted unprovable that London, British Library, MS Harley 1900, was copied into a separate manuscript, now lost, in which the text was updated and rewritten to serve as Caxton's exemplar for Trevisa's English Polychronicon. (5) As yet no direct exemplar is known for one of Caxton's editions in English for which he is most famous.
It might, though, be possible to identify the exemplar for Caxton's first edition of the Middle English prose Brut, which he calls, more descriptively, The Chronicles of England (STC 9991). (6) This edition was printed in 1480 and was innovative in several respects: it was the first chronicle printed in English; it involved technical innovations in the use of justification and signatures; and it was the first dated use of Caxton's Type 4. It includes a blank leaf, a prologue, and a table of contents by Caxton in a first unnumbered quire of eight leaves (most likely added later). The rest of the book is in quires of eight leaves, a to y, apart from the final quire y, which has only six. Quire a begins with another blank leaf, and then the prose Brut, extending to the year 1419, runs through signatures a2r to u3v. This ends at the foot of signature u3v with the words "in rewle and in gouernaunce," but Caxton's edition then adds seamlessly to it from the top of signature u4r to signature y6v a continuation of the chronicle to 1461 found only in manuscripts descending from Caxton's edition (as is noted below). (7) Lister M. Matheson argues convincingly that Caxton composed the continuation himself. (8)
But where did Caxton get the prose Brut from? In a brilliant tracing of the textual tradition, Matheson deduces that Caxton got the prose Brut from a manuscript of the textual tradition that Matheson calls "CV-1419(r&g): B, subgroup (c)," that is, the Common Version, Group B, subgroup (c), extending to the year 1419, ending with the words "in rewle and in gouernaunce" and en route including or excluding various elements. (9) Matheson's exhaustive search allows him to note that this textual tradition is "represented" by only "a single manuscript," namely manuscript HM 136 in the Henry E. Huntington Library, San Marino, California. This manuscript is "important" for studying the textual tradition because "a text of this type formed the basis for William Caxton's." (10)
One might wonder whether manuscript HM 136 was in fact copied from Caxton's edition of The Chronicles of England. After all, there are six other manuscripts of which that is true, (11) and there are extant far more manuscripts copied from printed books than there are manuscript exemplars for such books. …