Studia Islandica - 01.07.1963, Qupperneq 103

101
standard. Thus as Eyrbyggja, for example, comprises 38000 words,
its sum of pair words (24) is related to the size of Laxdœla according
to the following rule-of-three formula:
24 x
------- = --------; x = 37 (rounded off to whole number).
38000 58000
Correspondingly, the sum for the large Njála (97000 words) is re-
duced, and so on. In the column to the right are given in italics the
adjusted sums.
As is seen, no definite answer can be obtained from the lists as to
the specific affinity between Knýtlinga and Laxdœla. To be sure,
Laxdœla is far in advance of Eyrbyggja, Njála and Grettla. On an
average it has twice as many pair words — a proportion, by the
way, which recalls the position of Egla in its pair word series with
Heimskringla. But on the other hand, Egla's total in the present list
is not far behind that of Laxdœla. Thus a clear distinction between
these two sagas cannot be deduced.
In this case, however, the advanced position of Egla surely depends
on special conditions. As indicated above (p. 94) Knýtlinga has been
strongly influenced by Heimskringla. But if Snorri had composed not
only Heimskringla but Egla too — as can be safely assumed — it is
obvious that Egla would be highly “over-represented” in the pair word
series with Knýtlinga. Now it is possible to eliminate that source of
deviation with the aid of the distinction that was made above between
the pair words which are also to be found in Heimskringla and those
which are not. All the family sagas engaged in the comparison —
not only Egla — certainly depend more or less on Snorri’s literary
work. If one eliminates from the pair word series with Knýtlinga,
the numerous “Snorri words”, the remaining more qualified pair words
would be more likely to reveal any specific affinity between Knýt-
linga, on the one hand, and each of the five family sagas, on the other.
(This device, elimination of the “Snorri words”, was applied in the
Snorri-Egla paper too, in order to hring the mutual relations between
the family sagas into sharper relief.)
The Knýtlinga-table with “Snorri words” removed is shown on
p. 55. It changes the picture in a very suggestive way. First, how-
ever, it has to be admitted that such an elimination would obviously
“disfavour” Egla — if that saga was written by Snorri, nota bene.
In that case, Egla would contain especially many of the “Snorri words”
which have now been discarded. This assumption is rather drastically
verified by the new table. Egla's share of the pair words {10) with Knýt-
linga has now been reduced to about a seventh part of the former sum