Editiones Arnamagnæanæ. Series A - 01.06.1997, Síða 40
XXXVIII
more complete than it is now, though the end of the saga was already missing
from the manuscript. The two copies end in mid-sentence at the same point
(A 2494), and have space left for the ending.
In the first part of the saga, 179 is a closer copy of S6 than 18lg is, and for
that reason it has been chosen to supply the text for as much of the continua-
tion of the saga as it contains.
There are many points of difference between 179 and 181 g in the wording
of the second part of the saga. Some of them are bound to be due to innova-
tion in 179, but fewer presumably than are due to innovation in 181g. In some
instances it may be possible to see that an error has occurred in one or other
manuscript. For example, there is an evident haplography in 179 at A 1250.
But in rnost instances, individually considered, it would not be possible, on
the basis of the readings of 179 and 18lg alone, to tell which manuscript (if
either) has preserved the reading of S6.
The readings in other texts, i.e. B-F, might in theory sometimes indicate
which of 179 and 181 g is to be preferred, but generally these texts do not help
with such detail. Help is provided, however, by a group of manuscripts which
contain an A-type text, which indeed seem to be derived from S6 and derived
fronr it independently of 179 and 181 g. For this view of these manuscripts,
see the discussion below, pp.xxui-xux. Here it is sufficient to say that if it is
correct that the text they contain is a third witness to the text of the missing
part of S6, their agreement with either 179 or 181 g will confirm the correct-
ness of the one with which they agree.
It can be seen in the first part of the saga that these manuscripts are not tex-
tually close to S6 in the way that 179 and 18lg are, and the same must have
been true of the second part. This considerably reduces their usefulness for
the purpose of deciding between 179 and 181g, as they frequently agree with
neither, i.e. they have a variant of their own.
In this edition the second part of the A-text, ll97-2494, is presented as fol-
lows. The text is printed frorn 179 (= A2). This is to be assumed to represent
the text of S6, unless the select variants given below the text suggest other-
wise. Variants are given primarily from 181 g (= A3). The other manuscripts
(A4 and A5, representatives of the group just mentioned) are quoted when they
support A3, i.e. when the text printed is not or may well not be correct. They
are also quoted sometimes when they are neutral between A2 and A3, i.e. when
the choice in theory might fall on any one of the three readings. (But they are
not quoted if they are substantially different from both, for then there is little
likelihood that they represent S6.)
That is to say, the variants have been selected with the main purpose of
showing where the text (from 179) needs or may need correcting as the repre-
sentative of the missing S6.