Plagioclase compositions and non-anatectic origin of migmatitic gneisses in Northern Cascade mountains of Washington State

Abstract

Regionally developed migmatitic gneisses make up most of the Skagit Gneiss, though there are some orthogneisses also, derived from pre- to late-metamorphic intrusives. The migmatites contain countless, though quantitatively subordinate, remnants of biotite schists, less abundant amphibolitic rocks, and minor varieties of metasediments. Biotite schists (fine-grained paragneisses) predominantly are plagioclase-rich, having the compositions of highly immature graywackes; quartz-rich varieties are minor, and alumina-excess rocks very rare. There are para- and ortho-amphibolites, the latter of basaltic-gabbroic parentage. Leucocratic hornblende schists are largely meta-sedimentary. Apart from subordinate varieties in which K-feldspar is a major constituent, the leucocratic migmatitic gneisses have (leuco-) trondhjemitic and, to a lesser extent, quartz-dioritic compositions.

The Skagit Gneiss comprises an epidote-bearing subfacies in which sodic andesine is the most calcic plagioclase, and a predominant epidote-free subfacies where plagioclase ranges from oligoclase to bytownite-anorthite. All of the former and at least most of the latter subfacies belongs in the staurolite-kyanite zone. Various mineralogical features suggest a low-P variant within the field of Barrovian-type metamorphism. Even the epidote-free subfacies did not reach the highest grades of regional metamorphism; temperatures are estimated to have been on the order of 600° C there. For the epidote-bearing subfacies, attainment of anatectic temperatures is still more improbable.

A systematic study of composition and zoning of plagioclase reveals, statistically as well as for lit-par-lit samples, close relationships between the plagioclase of leucocratic migmatitic gneisses and that of associated schists, amphibolites, etc., indicating that the former was directly derived from the latter by metamorphic recrystallization, with a small shift towards more sodic plagioclase compositions in most, but not all, of the gneisses. Plagioclase relationships and other data seem to rule out igneous injection as the chief agent of migmatization. An anatectic origin of the leucocratic migmatitic gneisses is unlikely for the following reasons. (1) The plagioclase of the gneisses ranges widely in composition (up to, and locally even beyond, calcic andesine), depending on the parent-rock plagioclase. (2) Major K-feldspar is lacking in most of the gneisses which, actually, are impoverished in K₂O compared to parent schists. (3) Most of the gneisses have bulk compositions far from those of lowest melting in the granite system. (4) There is no wholesale basification of schist and amphibolite remnants. (5) Ratios of leucocratic gneiss to remnant material commonly are far too high for all of the former to have been split off the latter. (6) The sequence of migmatization, such as amphibolite→(quartz)-dioritic gneiss→ trondhjemitic gneiss→leuco-trondhjemitic pegmatitic gneiss, is opposite to that to be expected in progressive anatexis. (7) A similar solid-state history of deformation and recrystallization is recorded in gneisses and remnants. (8) The metamorphic grade does not support anatexis. Points (4) + (5) also argue against metamorphic differentiation having been the sole agent of migmatization.

Small-scale metamorphic-differentiation features occur, but gradations from schist, etc., into leucocratic gneiss are widespread, not only across but also along the strike. Replacement commonly is associated with prominent development of porphyroblastic plagioclase. On a much larger scale, bulk-compositional changes are indicated where thick schist- or amphibolitederived leucocratic gneiss sequences contain only subordinate remnants of their parent rocks. Such changes imply introduction of material — if not by injection, then by metasomatism. An alternative model, namely, simple isoohemical recrystallization mimetic after patterns of primary sedimentary differentiation would require that (leuco)-trondhjemitic “arkoses” were intimately associated with and graded into highly immature graywackes and even basalts and gabbros. Nor can this model be reconciled with plagioclase relationships and other petrographic data, or with certain aspects of the actual geometry of the migmatites. Both metamorphic differentiation and metasomatism are believed to have contributed to migmatization. Plagioclase data point both ways. Overall compositional changes in large rock volumes suggest that metasomatism was a first-order agent of regional migmatization while metamorphic differentiation mostly played a more subsidiary role.

In the sequences from predominant types of schist and amphibolite to leucocratic gneisses, alumina contents are nearly constant. Assuming that alumina approaches something like an “internal standard”, the chemical balance of migmatization can be estimated. Small but systematic increase in Na₂O (by 1 to over 2 wt. %) is linked with marked increase in SiO₂ (on the order of 5 to 10% for schist-, and more for amphibolite-derived gneisses) and with removal of iron and magnesia. CaO does not change systematically, except for decline during the first stages of migmatization of amphibolite. Removal of iron and magnesia is coupled with loss of about 0.5 to about 1.5% K₂O in biotite-schist-derived gneisses (except in subordinate varieties with major K-feldspar). In the series amphibolite-gneiss, gain in K₂O due to progressive biotitization of hornblende is followed by loss of K₂O in leucocratic members. Some of the potash set free from migmatized schists is accounted for by late-metamorphic K activity (biotitization of hornblende in amphibolites and their derivatives, also biotitization of garnet, late growth of minor K-feldspar in many but not all gneisses).

Small-scale but widespread late-stage features, such as local intrusive motion of crystalline migmatite, crosscutting pegmatites, etc., are not discussed, as the topic of the paper is the main-stage regional migmatization.

Subordinate late-metamorphically intruded orthogneisses, most with major K-feldspar, postdate the autochthonous migmatites and are genetically unrelated to them. They rose from depths where T was high enough for melting to occur. They invaded a non-migmatized belt west of the Skagit Gneiss also.