Bibliotheca Arnamagnæana - 01.06.1970, Side 349

323
The ranges of measurements for AM 445c,I are smaller than those
for AM 564a, probably because the holes in the former are less badly
damaged. It seems likely from the figures within each fragment that
measurements between undamaged holes should be allowed to show
an error of about 3 mm. each way, and between damaged ones corres-
pondingly more as the extent of the damage increases. There is, how-
ever, a general eorrespondence here, even when none of the consider-
able actual damage is allowed for, for an extension of 3 mm. each
way to each range of measurements produces an almost complete
overlap of figures in every case. This amounts to agreement be-
tween the binding-holes of the two fragments.
6.5. How great is the theoretical chance of two fragments of
separate origin appearing to correspond with this sort of binding ?
I have examined 594 Icelandic membrane Mss. and fragments
in the Arnamagnæan Collection to find out how common it is.
Of these, 80 gave no result, either because they were fragments
too small or too badly damaged for the pattern of their binding-
holes to be discernible, or because I was unable to see the binding-
strings without damaging them. 2 appeared never to have been
bound at all. Of the remaining 512, 54 had “Double Binding”
(that is to say, all the binding strings were in pairs, whose total
number was three or more), and 55 showed an apparently related
type37, in which the binding consists simply of one pair of strings
at each end of the spine—I have called this the “small 4to” type.
37 These figures cannot be regarded as final or exact, because:
a) About 30 Icelandic membrane Mss. and fragments were for various reasons
not available while this work was being done.
b) Some Mss. have clearly been bound more than once, and it is possible that this
has given the appearance of “Double Binding” to some Mss. which have never
actually had it. 5 Mss. which now seem to have had eccentric binding-schemes
may also have been bound more than once.
c) Some Mss. have been so well restored that they may originally have had
“Double Binding” which does not now appear.
Some subjective judgments are therefore necessarily involved in these figures,
but the number of mistakes caused by these would have to be very large to in-
validate the general demonstration that the chance of a coincidence is extremely
remote, and factors b) and c) would in any case tend to cancel each other out.
In cases of serious doubt, where “Double Binding” seemed likely but was not
certain, I have assumed its presence, to avoid over-stating the coincidence involved
in the agreement of the fragments with which I am primarily concerned.