Rit (Vísindafélag Íslendinga) - 01.07.1962, Page 189

187
formed, especially in a relatively viscous melt (Einarsson
1949).
Further insight was gained by studying the bread-crust
bombs formed in this eruption. Their crust is formed of light-
coloured translucent glass while inside this the glass grows
progressively darker towards the centre. In the same sense
the oxydation of the iron content increases. As the circum-
stances indicated that this oxydation had not heen caused by
atmospheric oxygen it was assumed that relief of pressure
caused a shift in the equilibrium hetween FeO and Fe203
towards the latter. Under pressure at depth the main form
of iron is FeO, according to this theory, and when such a
melt is chilled sideromelane is formed. Also if the melt is
relatively viscous (low temperature) extrusion and slow cool-
ing may lead to sideromelane. But if a more fluid magma
rises unexplosively to the surface there will be time for shift
of the equilibrium towards Fe203, and chilling then produces
only tachylite, the opacity being caused by the magnetite
molecule.
This explains the complete lack of sideromelane at ordi-
nary post-glacial volcanic craters in Iceland. If the difference
of sideromelane and tachylite was only a question of rapidity
of cooling this fact would be strange, for no doubt some of
the finer glassy material in the crater walls has been cooled
very rapidly.
Any theory of the Fragmental Rocks must account for the
predominance of sideromelane. And beside other ways of
formation of glassy and fragmental rocks the formation from
a viscous melt without chilling should be taken into account.
The main possibilities of explaining the various bodies of
which the Fragmental Rocks are composed are then the fol-
lowing:
1) Chilling by water. On p. 92 I have discussed how this
works, especially by preventing coalescence of lava that quite
normally is thrown out of a crater in the disperse state.
2) Explosive eruptions. I am convinced that very consider-