Jökull - 01.12.1981, Blaðsíða 7

released at a crack in the direction perpendi-
cular to it. The triple point intersections at
120° have originated there by the process of
continuous fracture growth, out of an initial
single fracture. Experiments in glass have led
to interpretation of this kind (.Ernsberger 1960).
It seems likely therefore that these discon-
tinuities have formed originally as fractures.
Hovvever the lack of strongly developed sets of
fractures might suggest a weak or complicated
stress field at the time of formation. This field
may have existed at a fairly advanced stage of
cooling when the temperature distribution
'vould hav’e been less even, and also when im-
perfections already developed on the major
joint surfaces could have acted as stress con-
centrators in the generally lower stress field.
Ihere is evidence on the cut surfaces that
fractures grow inwards from the prism edges
in this way. Very similar fracture patterns
have been observed in cooling lava blocks in
aa lava from the 1980 eruption of Hekla.
Origin of the bonding rnaterial
In the Icelandic rocks the primary ntinerals
are generally fresh but the interstitial glassy
patches often show alteration. The apparent
„weathering rind“ referred to above proves in
thin section to be mainly due to alteration of
the glassy patches to ferric iron rich products.
VVhere the discontinuities cut through
glassy patches they have the much rnore irre-
gular form and the result is to produce a more
sinuous discontinuity with „necking" and fluffy
edges (Fig. 3 (b)). These observations indicate
that the original fractures have not been ex-
posed to subaerial weathering for very long
and that the interstitial glass is largely the
source material of the bonding secondary
minerals. It seems very probable that the
alteration of the glass occurred during a low
temperature hydrothermal stage producing a
variety of amorphous material and smectite
and vermiculite. In the Icelandic examples
hvaloclastic deposits are associated with the
lavas (Friðleifsson 1973) solution of which
could also account for part of the bonding
material.
It is envisaged therefore that the rocks are
occupied by low temperature geothermal
fluids while still retaining small residual
stresses. Such conditions present a chemical
environment in which further fracture deve-
lopment may well be enhanced.
The process of stress corrosion has been well
studied in ceramics (Wiederhorn 1972) and more
recently in rocks (Anderson and Grew 1977). This
envisages that fracture extension can occur at
stresses well below the theoretical threshold if
a chemical reaction at the fracture tip reduces
bonding. Experiments suggest that only very
low residual stresses would be necessary
(,Speidel 1971). There is evidence of incipient
fracturing in about 20% of the rocks dealt with
in this paper. This tends to support the idea of
a low stress, partly chemical fracturing process
in the past.
Relevance to present day wealhering
One of the most obvious changes in the
bonded material is the oxidation of the iron.
This is restricted to the areas near the prism
edges or in sonle cases to a thin zone at the
junction between the bonding material and
the rock. Access of water and dissolved oxygen
is probably restricted to these sites.
New fractures are fairly widespread in all
the rocks within the bonding material. These
normally begin near the prism edges and ex-
tend inwards. There is sometimes a red coat-
ing along the new fractures which suggests
that fracturing precedes oxidation of the iron.
The new fractures could be interpreted as
tension fractures, perhaps related to fluctuat-
ing hydration states of the clay content. The
transverse lines encountered in some cases
(Fig. 3 (d)) may also be some kind of drying
crack, subsequently oxidised. Examination of
the scree fragments downslope shows that they
have developed to a very considerable extent
by failure along the discontinuities.
The development of fractures just behind
the „weathering rind“ on some prism faces in
the Icelandic rocks may have been caused by
small scale stress changes connected with the
reduction of density in the weathered zone but
JÖKULL 31.ÁR 5