A MULTIVARIATE METHOD FOR DETERMINING THE PROVENANCE AND PROTOLITH OFMETASEDIMENTARY ROCKS - AN EXAMPLE FROM THE FORK MOUNTAIN FORMATION, SOUTHWESTERN VIRGINIA PIEDMONT, USA
Pc. Ragland et al., A MULTIVARIATE METHOD FOR DETERMINING THE PROVENANCE AND PROTOLITH OFMETASEDIMENTARY ROCKS - AN EXAMPLE FROM THE FORK MOUNTAIN FORMATION, SOUTHWESTERN VIRGINIA PIEDMONT, USA, Geochemical Journal, 31(5), 1997, pp. 275-288
Metasedimentary and metavolcanic rocks within the Smith River allochth
on, southwestern Virginia Piedmont, are present in two units, the Bass
ett and the Fork. Mountain :Formations. The younger Fork Mountain Form
ation contains some metabasalt layers but primarily consists of a biot
ite gneiss and an overlying high-alumina mica schist; the biotite gnei
ss is wedge shaped and thins to the northwest. Fourteen major-oxide an
alyses were performed on samples of metasedimentary rocks from the For
k Mountain Formation. Compositions of both gneisses and schists fall o
n single linear trends on conventional Harker-type scattergrams, which
can be explained by sedimentary mixing lines between pure quartz and
a pelitic sedimentary assemblage. In addition, the negative correlatio
n between SiO2 and K2O is significant because in igneous rocks these t
wo oxides are normally positively correlated. Thus relatively coarse,
quartz-rich sediments apparently became the psammitic rocks that forme
d the paragneisses, and the finer, clay-rich sediments became che shal
es that are protoliths to the schists. A simple explanation of these r
elationships is a relatively fine-grained, deep-water sedimentary faci
es present in the northwest that transgressed to the southeast through
time; the source for these sediments would have been to the southeast
. Principal components analysis (PCA), ratio-ratio scattergrams, and t
-tests for differences of means, however, indicate that this two-compo
nent mixing is an over-simplification. Some subtle differences in chem
istry between the gneisses and schists exist that cannot be explained
by linear mixing alone and may imply more than one source region for t
he sediments. In addition, PCA and simultaneous solution of mass balan
ce equations estimate the following major rock-forming mineralogy for
the pelitic sediments, exclusive of quartz: illite -48%, montmorilloni
te -20%? chlorite -9%, K-feldspar - 20%. This mineral composition is c
onsistent with known patterns of mineral alteration during burial diag
enesis. Discriminant function analysis (DFA) was also performed; it su
ggests that an apparently large compositional gap is present between t
he gneisses and schists. Conventional bivariate scattergrams and PCA,
however, indicate that a compositional continuum exists for most oxide
s.