Bibliotheca Arnamagnæana - 01.06.1970, Side 357
331
added to AM 564a as an afterthought32, and this would hardly
have happened if the large sagas represented in AM 445b had not
yet been written in the codex. It seems probable, then, that AM
445b was written before AM 445c,I, and its Hånd 3 is therefore
unlikely to be the same as the hånd in AM 445c,I.
Accepting that the two hånds are distinet, it remains possible
to place AM 445b either at the beginning of what survives of the
codex, or between f.4 of AM 564a and the Gisla saga fragment in
AM 445c,I. But the latter seems unlikely—it would seem more
natural to begin a large codex with the major sagas to be contained
in it, and with Land-ndmabok. The same principle can be seen
among the editors of the large editions of modem times43.
9.15. There seems to be no orthographic reason to think that
AM 445b is not of the same period as the other fragments. It
also seems likely from the developments within Hånd 1 of AM
445b (see 9.1, 9.7, 9.11, 9.13 above) that the leaves there are
in the correct order, although it is impossible to tell whether
there was any other material between the items represented there,
of which no trace now remains. It seems probable that AM 445b
came at the beginning of the codex, in so far as we are able to
reconstruct it.
10.1. If the foregoing argument is accepted, the surviving parts
of Psevdo-Vatnshyrna may be fitted together as follows—
1. The Melabok fragments (AM 445b, ff. 1,2). Since these leaves
are joined there was probably an even number of leaves
between them—from the material lost, two seems most
likely.
2. (after a number of lost leaves which cannot now be estimated)
The Vatnsdæla saga fragment (AM 445b f.3).
43 e.g. Islendinga sogur (Det Kongelige Nordiske Oldskrift-Selskab), Copen-
hagen, 1843—1875; Islendinga sogur, ed. Valdimar Åsmundarson, Reykjavik,
1891—1901; Origines Islandicæ, ed. GuQbrandur Vigfiisson and F. York Powell,
Oxford, 1905; fslenzk for nrit, Reykjavik, 1933- ; Islendinga sogur, ed. GuSni
Jonsson, Reykjavik, 1946; all these begin either with Landndmabok or with fs-
lendingabåk followed by Landndmabok. One late codex in Resen’s library seems
to have been arranged in a similar way, see Arne Magnussons Håndskriftfortegnelser,
ed. Kr. Kålund, Copenhagen, 1909, p. 113, last paragraph, and pp. 274-5 above.