Jökull - 01.12.1984, Side 74

"PROFILES”
1981 1980
BOND NUMBERxlOOO
Fig. 5. Example of spurious but convincing
correlation between two random sequences of
“polarity zones“, generated from lottery
numbers.
5. mynd. Tenging milli tveggja “hraunstafla“ sem
búnir eru til upp úr vinningsnúmerum í Happ-
drœttisláni Ríkissjóðs.
ble random events such as the lengths of intervals
between the numbers of successive winning tick-
ets in a lottery draw.
I have accordingly constructed the left-hand
„profile" or polarity time scale fragment, from
the announcement of winning tickets for the
Class D Icelandic Government lottery draw of
July 1981. The polarity changes at each winning
ticket. I have used the first 17 tickets out of a
total of 55 winning 1000-kr prizes; altogether
125,000 tickets were issued. On the right-hand
side, there is a profile constructed similarly from
the equivalent 1980 announcement.
If each column is assumed to represent a lava
pile of 11/2—2 km thickness, one sees that there is
a good case for a „stratigraphic“ correlation to
exist between these; certainly it is no worse than
some correlations between polarity zones or
linear anomalies published in the literature,
where one has to take variable emplacement
rates into account.
A conclusion to be drawn from these results is
that one should primarily base one’s stratigraphic
ties on geological criteria, using paleomagnetism
only in conjunction with these rather than on its
own.
SHORT EVENTS
A special word of caution about events
apparently consisting only of one or two flows:
These should never be used for stratigraphic
correlation. Not only do many such events, when
real, have a very limited areal distribution in the
lava pile, but they are often equally likely to be
spurious. Besides V.R.M. and instabilities, it is
quite likely that short events are caused by faul-
ting or by hidden or eroded intrusions close to a
sampling site.
Examples already exist where events reported
in the literature (particularly, in the Appendix of
Watkins and Walker (1977)) have been subse-
quently found to be of this type, and the nature of
some others remains to be checked (e.g. some
events in profiles SK, JF, and BV in NW-Ice-
land).
Directional excursions of the field, even those
confirmed by laboratory measurements, are also
very rarely of any value in stratigraphic correla-
tion in Iceland.
DIFFERENCES BETWEEN FLUXGATE
AND LABORATORY MAGNETIC
RESULTS
It may be useful to give an overview of various
error sources inherent in the field fluxgate
method, and how these errors are eliminated by
magnetic measurements in the laboratory.
Error source in the field Laboratory method
Irregular shape of Regular cylindrical
sample samples
72 JÖKULL 34. ÁR