lgarise" (in the wholly laudable sense of
that too often degraded word) the body of English literature. Only, such
a book would not have been what I was thinking of. To bring out the full
contrast-complement of these two strangely coincident masterpieces, both
must be read in the originals. Paradoxically, one might even say that a
French translation of Johnson, with the original of Voltaire, would show
it better than the converse presentment. _Candide_ is so intensely
French--it is even to such an extent an embodiment of one side of
Frenchness--that you cannot receive its virtues except through the
original tongue. I am personally fond of translating; I have had some
practice in it; and some good wits have not disapproved some of my
efforts. But, unless I knew that in case of refusal I should be ranked
as a Conscientious Objector, I would not attempt _Candide_. The French
would ring in my ears too reproachfully.

P. 396, last line.--Shift comma from after to before "even."

P. 399, l. 10.--_For_ "Rousseau" _read_ "his author."

P. 424, _note_, first line.--Delete quotes before "The."

P. 453, l. 15.--_For_ "Courray" _read_ "Cou_v_ray."

P. 468, l. 17.--_For_ "France has" _read_ "France had."

P. 477.--In the original preface I apologised--not in the idle hope of
conciliating one kind of critic, but out of respect for a very different
class--for slips due to the loss of my own library, and to the
difficulty (a difficulty which has now increased owing to circumstances
of no public interest, in respect of the present volume) of consulting
others in regard to small matters of fact. I have very gratefully to
acknowledge that I found the latter class very much larger than the
former. Such a note as that at Vol. I. p. xiii, will show that I have
not spared trouble to ensure accuracy. The charge of _in_accuracy can
always be made by anybody who cares to take "the other authority." This
has been done in reference to the dates of Prevost's books. But I may
perhaps say, without _outrecuidance_, that there is an _Art de negliger
les dates_ as well as one _de les verifier_. For the purposes of such a
history as this it is very rarely of the slightest importance, whether a
book was published in the year one or the year three: though the
importance of course increases when units pass into decades, and becomes
grave where decades pass into half-centuries. Unless you can collate
actual first editions in every case (and sometimes ev