he touch of restfulness to literary work. No original, it is said,
has yet been found for Book vii., and it is possible that none will ever
be forthcoming for chap. 20 of Book xviii., which describes the arrival
of the body of the Fair Maiden of Astolat at Arthur's court, or vii for
chap. 25 of the same book, with its discourse on true love; but the
great bulk of the work has been traced chapter by chapter to the
"Merlin" of Robert de Borron and his successors (Bks. i.-iv.), the
English metrical romance La Morte Arthur of the Thornton manuscript (Bk.
v.), the French romances of Tristan (Bks. viii.-x.) and of Launcelot
(Bks. vi., xi.-xix.), and lastly to the English prose Morte Arthur of
Harley MS. 2252 (Bks. xviii., xx., xxi.). As to Malory's choice of his
authorities critics have not failed to point out that now and again he
gives a worse version where a better has come down to us, and if he
had been able to order a complete set of Arthurian manuscripts from his
bookseller, no doubt he would have done even better than he did! But of
the skill, approaching to original genius, with which he used the books
from which he worked there is little dispute.

Malory died leaving his work obviously unrevised, and in this condition
it was brought to Caxton, who prepared it for the press with his usual
enthusiasm in the cause of good literature, and also, it must be added,
with his usual carelessness. New chapters are sometimes made to begin in
the middle of a sentence, and in addition to simple misprints there are
numerous passages in which it is impossible to believe that we have the
text as Malory intended it to stand. After Caxton's edition Malory's
manuscript must have disappeared, and subsequent editions are
differentiated only by the degree of closeness with which they follow
the first. Editions appeared printed by Wynkyn de Worde in 1498 and
1529, by William Copland in 1559, by Thomas East about 1585, and by
Thomas Stansby in 1634, each printer apparently taking the text of his
immediate predecessor and reproducing it with modifications. Stansby's
edition served for reprints in 1816 and 1856 (the latter edited by
Thomas Wright); but in 1817 an edition supervised by Robert Southey went
back to Caxton's text, though to a copy (only two are extant, and
only one perfect!) in which eleven leaves were supplied from Wynkyn
de Worde's reprint. In 1868 Sir Edward Strachey produced for the
present publishers a reprint of Southey's tex