Rit (Vísindafélag Íslendinga) - 01.06.1984, Blaðsíða 281

PETROGENETIC REI.ATIONSHIPS
277
tion leaves a liquid of 80 to 63 wt. per cent of the starting material, generally
matching the compositions of the basalts of relatively high K content. Here
the misfit is, though, rather serious as regards P and Mn. In four out of six
calculations the “daughter liquid” is too high in Mn. Such a systematic
deviation is, in the case of a minor element, a serious matter, as it does not
count much in the total misfit figure. The high P could, in this case, be
“corrected” by a minor apatite subtraction, without seriously aífecting the
Ca content. The Mn values are more difficult to “correct”, as previously
mentioned. K in the calculated “daughter” is too low in three cases ofthe six
and Na too high in two cases. Ti turns out to be too high in two cases and
too low in one. Other elements do not show any noteworthy deviation.
Instead of carrying the calculations on from the high-Mg basalts all the
way down into the low-Mg ones, as in Table 45, a closer approximation to
the equilibrium relations can be obtained by taking the less magnesian
basalts as a separate group and try to fractionate within it, starting with a
basalt of a moderate Mg content as a “mother liquid”. This is shown in
Table 46, wherejan 14 has been chosen as a “mother”. In Table 44B Jan 14
has been derived from a high-Mg basalt through fractionation of a titan-sal-
ite and olivine but with a poor total fit (caused by deviation of Fe, Ca and
Ti). It is thus probably not the best chosen sample in this case, if one wants
to trace the path of fractionation of the whole rock suite from the most
primitive to the most evolved rocks. It is, nevertheless, as good a representa-
tive of the moderately evolved basalts as the others and can thus equally
well represent the “mother liquid” it is here supposed to do. A relatively
good total fit is obtained in this case by subtracting four minerals. A titan-
salite and plagioclase make up the bulk of the fractionate but smaller
amounts of both olivine and titanomagnetite are necessary accessories.
Clinopyroxene amounts to 43 to 50 per cent and plagioclase to 35 to 43 per
cent. The two other minerals only account for 13 to 15 per cent of the
fractionate and of this the titanomagnetite highly dominates the olivine,
which only accounts for 1 to 5 per cent of the whole fractionate. As the
compositional diíference between the “mother liquid and the derived
liquids” is increased the total misfit is kept low by allowing both titan-salite
and plagioclase to evolve. The approximation to equilibrium conditions is in
this case, therefore, better than in most of the previous cases. The misfit here
is primarily caused by P, K and Ti. In three out of six “daughter liquids” P
is too high and in two cases much too low. This discrepancy is caused by the
uneven distribution of apatite in the rocks. Apatite is found in greatest
amount in the low-magnesian basalts and the low-silica intermediate rocks,
but the amount diífers from one sample to another, indicating that the
apatite crystallization was somewhat discontinuous. The “daughter liquid”
is far too high in K in four cases of the six but shows a perfect fit in two cases.
In four cases of the six the Ti content of the “daughter” is similarly too high.
In one case the “daughter” is much too high in Na and in two cases Mg is