Bibliotheca Arnamagnæana - 01.06.1970, Blaðsíða 346

320
Since the surviving substance in AM 564a suggests a total length
of 207 lines, the error involved in supposing one leaf to be lost here
is negligible.
From the end of f.6 to the beginning of f.7 of AM 564a, which
contains the end of the saga, the substance of 703 lines of Turville-
Petre’s text is missing (p. 29, 1. 19 to p. 50, 1. 31)30. This is roughly
the equivalent of five leaves (theoretically 725 lines), and the error
of 4,4 lines per leaf is smaller than that allowed by Jon Helgason
when he calculates that the 1207 lines of the saga after the end
of f.l of AM 445c,I would fill eight leaves31 (145 x 8 comes to 1160,
so that his error would be 47 lines, or nearly 6 per leaf). This,
the largest deviation from the theoretical figures in these calcula-
tions, is still less than 3 per cent.
5.4. One might object to this reconstruction that f.7 of AM
564a also contains the beginning of Hardar saga, in another hånd
and apparently added after the rest of AM 564a was complete32,
and the scribe of AM 445c,I would not be very likely to add
Gisla saga (ff.2-5) after the lapse of time needed for the work of
Hånds 2 and 3 of AM 564a, and for the pause which seems to have
come between them. But it is orthographically likely that the
Gisla saga fragment in AM 445c,I was written before the fragment
of Viga-Glums saga (see 2.7 and 2.8 above).
6.1. On p. ix of his introduction to Bårdar saga33, (1860), Gu5-
brandur Vigfusson says that eight leaves which then composed
30 The substance of f.6 of AM 564a is parallelled by only 110 lines in Turville-
Petre’s text (p. 26, 1. 14-p. 29, 1. 19), but this is because the Modruvallabok version
of the saga cuts out a passage in which Skuta’s instructions to the outlaw are
rather monotonously acted out.
31 J6n Helgason, op. cit., p. 12.
32 The end of Viga-Qlums saga, which filis the first twenty lines of f.7, was
apparently written in great haste—the degree of abbreviation is considerably
greater than in ff.5,6, and the lines are crooked and do not correspond to those
neatly marked for the remainder of f.7r. This suggests that the lines were not marked
out until the writing of Hardar saga was begun, and this would hardly have been
omitted if the intention had been to begin Hardar saga immediately. But Hånd 3
of AM 564a is in many respects more conservative than Hånd 2, and this suggests
that the interval which elapsed before Hånd 3 was begun was probably short.
33 Bdrdarsaga Snæfellsdss, VIglundarsaga, Por dar saga t Draumavitranir, Volsa-
pdttr (Nordiske Oldskrifter udgivne af det nordiske Literatur-Samfund, XXVII),
Copenhagen, 1860.