Editiones Arnamagnæanæ. Series A - 01.06.1997, Qupperneq 50

XLVIII
Where the text is difficult to read on the present last page of S6, there are
five places where A4-A5 render S6 more accurately than A2 and A3 do (see the
notes to lines 54-5, 55-6, 63, 72 and 75), and two others where they agree
with A3 in rendering it more accurately than A2 does (lines 57 and 86). This
may suggest that A4-A7 descend from a copy of S6 made when it was less da-
maged than it was when A3 and A2 were copied from it. (There is no corre-
sponding evidence from the first page of S6 because the text of A4-A7 has
been shortened there.)
A2 and A3 end abruptly at A 2494, no doubt because when they were written
the lacuna in S6 began at this point. But A4-A7 continue without any sign of
difficulty, and have an ending for the saga. This ending is different from all
other extant endings, and has not been obtained from any of them. It may
have been obtained from some version of which there is no other trace. But it
is possible that there was no lacuna in S6 when the copy was made from
which A4-A7 are derived. Accordingly, the text of A4, with variants from A5-
A7, is presented in this edition as the possible ending of the A-text.
A4-A7 are derived throughout from a lost common original with marked
textual characteristics (above), and each has variants of its own and cannot be
the source of any of the others. Within the group there are a few large variants
common to A4, A6 and A7 where A5 has retained, essentially, the reading of A'
or A2-A3; these may show that these three texts are derived from a common
original later than that for the whole group. They are:
7.40-2 Vrn-sidann A1; og sídan gengo þeir fyrir Herm(ann) jarl og dróttn(ing) og
sagdi 0rendi sín. Herm(ann) jarl þagdi langa stund, og m(ælti) sídan af
megnum módi A5] Sídann sagdi hann af miklum módi A4, A6 (slightly
defective, but reading not in doubt), A7
17.46-50 . Nv-leikinn A2-A3; Nú m(æla) sumir í móti. Just(inus) sv(arar) þí skal ei
lofa jarli í leikin med þí han hefr latid marga riddara og latid hann hefna
manna sinna. Sva lykr at Placido jarli var lofat í leikin A5] -f A4, A6, A7.
(Haplography leikinn-leikinn)
18.5 þu-lagt A2-A3; A5] f and space left equivalent to one line A4, -f (also pre-
ceding ad -f) A6, A7
20.63-5 Vilhialmur-sfnumm A2-A3; A5] -f A4, A6, A7. (Haplography Vil-
hialmur-Wilhialmur).
It may seem an unlikely coincidence that these omissions should have oc-
curred independently in three out of four closely-related texts. If the three had
a common original, some twenty small shared variants could be attributed to
it, though most of them could hardly be said to support the supposition of its
existence, e.g.