Jökull - 01.12.1989, Qupperneq 63

Table I. List of samples.
Tafla I. Listi yjir sýni.
Sample Eruption Environment Remarks
^ 769 Heimaey 1973
2 940 Katla 1625
3 941 Hekla 1693
4 942 Öræfajökull 1362
5 943 Surtsey 1964
6 985 Hekla5 (light colour)
7 986 Hekla5 (dark colour)
8 987 Karl
9 988 Hekla4 (11)
10 989 Hekla4 (8)
11 990 Hekla4 (6)
12 991 Hekla4 (4)
13 998 Unknown
14 999 Unknown
15 1000 Hekla 1947
16 1001 Hi
17 1002 H,
18 1003 H,
19 1004 Askja 1875
20 5933 Snæfellsjökull
21 5838/1 Snæfellsjökull
22 5838/2 Snæfellsjökull
Subaerial
Subglacial
Subaerial
Subaerial/subglacial
Submarine
Subaerial
Subaerial
Submarine
Subaerial
Subaerial
Subaerial
Subaerial
?
?
Subaerial
Subaerial
Subaerial
Subaerial
Subaerial
Subaerial/subglacial
Subaerial/subglacial
Subaerial/subglacial Duplicate analysis
Rhyolitic
Andesitic
Last phase
Late phase
Early phase
Initial phase
Skógar
Kúðafljót
parameters employed are elongation (I/L), thickness
(S/L), sphericity, and OP-index. Folk (1974) defined
sphericity (VP) by comparing the grain in question
w'th a reference sphere or equant with the same
volume as the grain:
'F =3 Vs2/LI
A perfectly spherical grain has a sphericity of 1.0
and so has a perfect cube. Dobkins and Folk (1968)
proposed the OP or oblate-prolate index:
OP=(10-((L—I)/(L—S) - 0.50))/(SA)
The OP index becomes negative for oblate (disc-
like) and positive for prolate (rod-shaped) grains.
STATISTICAL METHODS
Separate form diagrams were plotted for each size
class of every sample so that the form parameters of
every single grain have been inspected visually.
However, statistical methods were used to compare
the form parameters of many samples. The SYSTAT
statistical package was used for this purpose (Wil-
kinson, 1987; Wilkinson, 1988), and the selection of
procedures adopted was mainly based on the recom-
mendations of Till (1974) and Folk (1974). The first
step in the statistical analysis was a visual com-
parison and evaluation of the form diagrams. The
second step was to test the normality of the form
parameters by plotting histograms for each parame-
ter, and also by plotting the values of each parame-
ter against corresponding percentage points of a
JÖKULL, No. 39, 1989 61