At 10:34 AM 09/11/2001 +0100, you wrote:
> At 18:16 06/11/01 -0500, Dugald wrote:
> "At 10:54 AM 07/11/2001 +0100, Christian wrote:
> You seem to exclude the formation of granulites without fusion??
It is theoretically possible, but I have not seen it in the field, nor have
I been persuaded by reading the literature. Bob Newton's case for
infiltration of CO2 as a means of suppressing a(H2O) was hotly disputed by
John Valley and others.
> There are huge proterozoic terranes with basic magmatic complex in
>granulitic conditions; In SW of Madagascar, the parageneses are 1 or 2prxs,
>Pl, with or without Q, Brown Hb, Grt in a space of a few hundred square
>meters. All the parageneses of the transition zones of the 3facies:
>amphibolite, HP and IP granulite facies in the domain of the multivariant
>(because of the variations of MG/Fe , Na/Ca ratios, asiO2, aH2O and/or
>CO2?) subsolidus reactions Cpx+Grt+Q+V = Hb +Pl, Cpx+Grt+Q = Opx +Pl,
>Cpx+Ox+Pl+V = Hb, The magmatic textures are sometimes preserved and there
>no melting of these rocks.
?? How can preservation of magmatic textures be reconciled with inferred
absence of magma?
> Moreover, at first time, I have speak about the transition in the
>direction HP granulite to amphibolite facies.
Yes this echoes my own experience. For 30 years I have been wanting to
study a prograde transition from amphibolite to granulite facies, but none
is to be found! In every case the granulite-facies domain appears to have
been formerly more extensive and the amphibolite facies has encroached upon
it during cooling. The retrograde character of this transition is easily
explained if a small amount of unquenchable silicate liquid was the last
phase to finish crystallizing during cooling, because the reaction by which
it crystallized would consume anhydrous minerals and produce hydrous
minerals. Corresponding to each of the V-consuming retrograde reactions you
listed is a vapor-absent L-consuming reaction (e.g. Grt + Cpx + liquid = Hb
+ Pl + Qtz). It is known from experimental studies that the V-consuming
reactions are metastable in respect to the L-consuming reactions unless
a(H2O) has been suppressed to less than ~0.5 by infiltration of gaseous
species other than H2O from a hypothetical external source.
By contrast, in the H2O-vapor-present domain on the low-grade side of the
migmagrad it is common to find prograde textures across a reaction isograd,
suggesting that H2O vapour can readily escape from medium-grade rocks as
fast as it is produced.