"GL3" - читать интересную книгу автора (vol04)

which go back to the earliest version, these features argue only
a recollection of the Lost Tales (to be assumed in any case! -
and doubtless a very clear recollection), not a close derivation
from the actual text. If that had been the case, one might ex-
pect to find the re-emergence of actual phrasing here and there;
but that seems to be markedly lacking.

The history of the typescript becomes rather complex
towards the end (from $15), where my father expanded and re-
typed portions of the text (though the discarded pages were
not destroyed). But I see no reason to think that much time
elapsed between the two versions; for near the very end ($19)
the original typescript gives out, and only the second version
continues to the conclusion of the Quenta, which strongly sug-
gests that the revisions belong to the same time as the original
text.
Subsequently the whole text was revised throughout, the
corrections being made carefully in ink; these changes though
frequent are mostly small, and very often no more than slight
alterations of expression. This 'layer' of emendation was
clearly the first;* afterwards further changes were made at dif-
ferent times, often very hastily and not always legibly in pen-
cil. To present the text as first typed with annotation of every
small stylistic improvement is obviously quite unnecessary, and

(*The occurrence of Beleriand in the original typescript, first in $13, note
10, not as previously by emendation in ink from typescript Broseliand,
shows that some of this 'layer' was carried out while the typescript was still
in process of composition.)

would in any case require the introduction into the text of a
forest of reference numbers to the notes. The text given here
includes, therefore, without annotation, all minor changes that
in no way affect the course of the narrative or alter its impli-
cations. Those emendations that are not taken up into the text
but recorded in the notes are marked as 'late changes' if they
are clearly distinguishable, as is not always the case, from the
first 'layer' described above.
I have divided the text into the same 19 divisions made in
S (see p. 11); but since the opening of Q has nothing corre-
sponding in S this section is not given a number.

*

THE QUENTA

herein is

QENTA NOLDORINWA