Bibliotheca Arnamagnæana - 01.06.2005, Síða 40
30
Jon Helgason
have begun later than at some point on EgFJ, p. 300 = EgEA
I, p. 157.13
(2) Whereas the nickname of Qnundr sjoni is spelled seoni in W,
it appears in 458 as sko ni (EgEA I, ch. 28,28; o = 6) or seoni
(ch. 67,10) because the scribe has read c instead of e. In
458(b) we find first skone at ch. 70,60 but later sioni at 80,2
(and 81,95 = parallel text from 458, line 34). The first form
surely comes from *W, the second from Y.
(3) W is the only [primary] manuscript that calls the earl in Verma-
land Armon (EgEA I, ch. 70,8), and this form is of course re-
produced at the corresponding place in 458. Elsewhere the
earl’s name is Arnfidr or Arnvidr (so also 463 and 560 Arn-
vidur). The name Armon recurs in 458(b) at ch. 74,10 (but
gen. sg. ArnuidaR at 74,6). Wherever else the earl is mentioned
by name in other manuscripts (ch. 70,17, 76,24, 76,31), the
name is left out in 458. The form Armon offers virtually defin-
itive proof that *W was in use at the beginning of 458(b).
Though it is accordingly very likely that 458(b) first followed *W and
later switched to Y, it is a more difficult matter to determine where the
one source ended and the other began. This is due to the faet that both
texts were of the B-class. Comparison with fragment i is of no help, for
wherever we have that fragment the textual status of 458 is the same: if
i agrees with M against 463/560, 458 always shares the reading of
463/560 irrespective of whether it would seem to be following *W or Y.
Agreement of 458 with M against the others is an absolute exception.
The instance most difficult to explain is EgEA I, ch. 77,18-19 gripi og
voru 458 (= M), where fragment i has gripi er i var and 463/560 gripe
er far voru\ two other examples, neither of mueh significance, are
EgEA I, 75,18 verid (= M), farit 1463/560, and ch. 77,24 sumer2 (= M),
svmir stvckv i 463/560.
One characteristic feature can, however, be attributed with some
probability to Y. 458 has small additions here and there that do not recur
in any other manuscript:
13 [It might have been explicitly stated at this juneture that Jon Helgason calculates with
an average of ca. 165 printed lines in EgFJ per lost leaf of W if there were seven such
leaves, but ca. 190 printed lines if there were only six (see below).]