COMPOSITIONAL VARIATIONS IN METAMORPHOSED SEDIMENTS OF THE LITTLETON FORMATION, NEW-HAMPSHIRE, AND THE CARRABASSETT FORMATION, MAINE, AT SUB-HAND SPECIMEN, OUTCROP, AND REGIONAL SCALES

Rocks from several outcrops of low-grade Littleton Formation and later ally equivalent Carrabassett Formation were analyzed for major element s and Zr. Compositional variations well outside analytical uncertainti es were observed, in order of increasing magnitude, for the following: 10-g subsamples from a 2-kg hand sample of visually homogeneous slate ; 100-g samples taken 10 m apart in the same outcrop; 10-g subsamples from a visually homogeneous quartzite; samples of relict turbidite bed s in the same outcrop; and single 100-g samples from seven outcrops. W e regard these variations as inherited from the sedimentary protolith with little if any alteration over distance scales of centimeters or g reater. All could be modeled successfully as mixtures mainly of clay, quartz, and chlorite-mica, although the compositions of the clay and t he chlorite-mica components had to be adjusted for different sets of s amples, within reasonable bounds for sedimentary-diagenetic minerals. Compositional variations and trends in 2-kg samples of higher-grade Li ttleton rocks are essentially the same as for the low-grade samples. F or example, SiO2 and Zr correlate positively, as expected for sediment ary sorting of quartz and zircon from clay, but not for selective meta morphic mobilization of silica. Also found was centimeter-scale separa tion of K2O from Al2O3 and strong covariation of Zn with Al2O3 in simi lar to 10-g subsamples of a staurolite schist, compositional evidence of element transport over distances of a few centimeters. Because the protolith includes at least two major, Al2O3-bearing components in var iable proportions, element ratios to Al2O3 can vary such that variatio n in metamorphic rocks need not be a sensitive indicator of element mo bility relative to Al2O3 even if Al is immobile. The population of low -grade rocks encompasses a broad compositional range. Thus, it is not feasible to obtain a sample average with a small standard deviation. I t would also be difficult to demonstrate conclusively whether a set of samples was representative of its metamorphic grade, or whether rocks of all metamorphic grades developed from the same combination of prot olith compositions and physical characteristics.