Bibliotheca Arnamagnæana - 01.06.1965, Side 290

goda as well as notes by Ari concerning Porsteinn rauSr were used as
sources, the most reasonable explanation of this being that both notes
were found in the lost dttartala. The mention of Ketilbjprn Ketilsson’s
landnam known from the Haukdæla pdttr should no doubt be referred
to this, too, and probably still more members of the families with which
Ari was acquainted had such æv is assigned to them.
In these ævis, besides genealogical particulars, information must have
been given about a number of important events, which might well be
characterized by the term stortidendi used by Snorri Sturluson. On the
other hånd, it is not probable that Ari, as supposed by Bardi GuSmunds-
son should have composed actual annals. The references to Ari in later
literature are without difficulty explained as applying to the conunga ævi,
dttartala, and the material known from the extant tslendingabok, and
the faet that practically all the authors who referred to Ari used more
than one of these sections as source, suggests that they were all collected
in one work, which must be identified with the bok mentioned by Snorri
in the Prologue to the Heimskringla.
A feature characteristic of Ari’s authorship is the care with which he
accounts for his authorities. They are not only mentioned regularly by
name, but in several cases Ari refers to the special qualifications which
the person in question had for giving reliable information about a cer-
tain event. Ari mentions one written source, only, viz. a saga about the
English King Saint Edmund, which in the tslendingabok is adduced as
source of the dating of the death of King Edmund at 870. This saga
should probably be identified with the De miraculis Sancti Eadmundi,
which was written immediately before 1100 by the Archdeacon Herman-
nus. It is hardly reasonable to regard the Passio Sancti Eadmundi by
Abbo of Fleury as Ari’s source, as was previously generally the case. No
more does it seem called for to assume a lost Icelandic Jatmundar saga
hins helga as Ari’s source, as suggested by Hermann Pålsson. In previous
research Bede’s and Adam of Bremen’s writings have furthermore been
pointed out as possible sources to Ari, and investigations of Ari’s chrono-
logical information corroborate this supposition.
Like Sæmundr’s chronological Communications those of Ari’s consist
of the lengths of periods and of relative datings within such periods, but
Ari furthermore often gives information about a person’s age at a certain
time. These three categories of information everywhere appear in close
mutual connexion and obviously are the basis of Ari’s chronological