sumption of Surrey's indebtedness to the Hippolito version is hy-
pothetical.

However unsatisfactorily vague may seem the discussion of the
problem of the Second Book, the outlines are clearly drawn in
comparison with the questions involved in the problem of the
Fourth. For the Second Book we have a text admittedly inferior,
because there is very little doubt that it has been edited; for the
Fourth we have three texts, which do not agree among themselves
and the value of anyone of which depends upon its similarity
to an unknown original. These are (a) the Tottel edition of 1557,
reprinted by the Roxburghe Club in 1814, and the Fourth Book
alone by Fest; (b) the Hargrave MS. version, never printed at all,
but very carefully collated by Imelmann; and (c) the printed
edition of John Day, which exists today in an unique copy at Brit-
well Court, and which has never been either reprinted or collated. ¹
In the two accessible texts, the Tottel and the Hargrave MS., there
are certain curious differences. Imelmann has shown that the
second is more like the Douglas translation, and also the Italian
version of Picholomini, than is the Tottel. ² Granted that this
be true, it is not clear what it signifies. Either Tottel changed
the text, or the copyist of the Hargrave MS. changed the text, or
(what is more probable) both edited the manuscripts that they
received. The case is still more complicated by the fact that, as
Imelmann shows, there are apparent reminiscences of the Hargrave
MS. version in Phaer's translation, finished in April 1556. But
there is no proof that the copyist of the Hargrave MS. did not im-
prove his author in reference to Phaer! And these verbal similar-
ities are not more numerous than would happen by the doctrine

-539-