Rumble et al. [1991] found that 
O
O
values for marbles from southern Maine are consistent with either
extensive fluid-rock interaction during metamorphism or premetamorphic
diagenesis. Until recently, the ambiguity suffered by Rumble et al. was
indicative of a nagging foundational uncertainty in the study of oxygen
isotopes in metamorphic rocks; indisputable evidence that observed
heterogeneities in mineral 
O
O indicative of fluid flow
occurred during metamorphism was lacking. Measurements of
small-scale intracrystalline 
O
O variations in metamorphic
minerals during this past quadrennium have provided us with this previously
lacking evidence.
Chamberlain and Conrad [1991] measured 
O
O in four
metamorphic garnets from three localities using the CO
laser-heating
fluorination method described by Sharp [1990]. One of
the analyzed
garnets, from the Gassetts schist of southeastern Vermont, exhibits a
decrease in 
of 2
from core to rim that can only be
explained by exchange with a mobile fluid phase. By documenting the
existence of this pattern of oxygen isotope zoning in a patently
metamorphic mineral, it was shown unequivocally that
externally derived fluids were the progenitors of significant shifts
in 
O
O during metamorphism. Chamberlain and Conrad suggested that
the zoning profile of this garnet resulted from exchange with a
low-
fluid phase that infiltrated the rock during
late stages of garnet growth.
Young and Rumble [1993], studying a sample of Gassetts
schist similar
to that examined by Chamberlain and Conrad [1991],
illustrated the
usefulness of millimeter-scale measurements of 
O
O in several
minerals for chronicling both fluid-rock reaction histories and
durations of fluid flow during metamorphism. Young and Rumble's
analysis suggests that oxygen carried by advecting fluid can find its
way into minerals even after cessation of flow by action of
net-transfer reactions. Thus the presence of fluid-derived oxygen in
a metamorphic mineral, as ascertained from measured 
, does not a priori constitute evidence for growth
during fluid infiltration. In the case of the Gassetts schist
sample, it was shown that the fluid-borne low-
oxygen evidenced by garnet zoning may have been sequestered by the
rock initially in the mineral chlorite during hydration reactions,
and that garnet acquired this low 
oxygen from
chlorite during subsequent post-fluid mineral-mineral reaction.
The small-scale isotope measurements suggest that fluid
infiltration during metamorphism of the sample studied must have been
short-lived with the total duration of flow being 
10
years.
Preservation of nonequilibrium 
values between two different
mica minerals with similar oxygen diffusivities was a crucial
observation in this work.
The importance of these types of measurements should not be
overlooked. Previously published geochronology data suggest garnet
growth in pelitic rocks may take approximately 
years. If so,
the study by Chamberlain and Conrad [1991] implies
protracted
continuous flow of water spanning millions of years. Alternatively,
the analysis of Young and Rumble [1993] indicates fluid
that was the
source of the low 
O
O oxygen flowed for 100,000 years or less.
Chamberlain and Conrad [1993] attributed meter-scale
spatial variations in the amplitude of metamorphic garnet 
O
O zoning
to growth of garnet during passage of an oxygen isotopic front
propagated by advecting fluid over a period of
10 Ma. Such a long period of flow of exogenous fluid has testable
consequences
for the 
O
O composition of other minerals in the rock. For
example, mica minerals, which have comparatively large oxygen
diffusivities, should show 
values reflective of
equilibrium with the fluid upstream of the front (for
micas, see below) regardless of their role in the garnet-producing
reaction.
The study by Chamberlain and Conrad [1993] raises the
issue of how to define 
O
O fronts in polymineralic rocks imparted by
flow of nonequilibrium fluid. Recalling that the displacement of a 
O
O
front is a function of the amount of exchangeable oxygen in the solid, and
considering that the amount of exchangeable oxygen is a function of oxygen
diffusivity, a parameter that varies over fifteen orders of magnitude from
mineral to mineral in some rocks, it is clear that there will be different
O
O front positions for different minerals. Moreover, mineral-specific
fronts may only be evident in portions of individual grains as a result of
diffusion-limited zoning.
The high spatial resolution that is unique to the ion microprobe also
contributed to our understanding of isotope exchange between fluid
and rock during metamorphism through the advent of new methods and
instrumentation. Valley and Graham [1991] used this
technology to
measure magnetite 
O
O on an unprecedented small scale.
Although the analytical
precision of the analyses is an order of magnitude larger than
laser-heating and conventional bomb analyses, the data reveal a
dramatic 9
decrease in 
at the outer
10 
m edge of the magnetite grains that could not have been
identified by other means. Valley and Graham concluded that the
narrow zones of 
depletion are the result of diffusive
exchange of oxygen with a hydrous fluid during the waning stages of
granulite facies metamorphism. The width of the depleted edges and
known oxygen diffusivities in magnetite were used to infer that
retrograde fluid flow apparently lasted between 
and
years.
Sample charging has hampered similar efforts in insulating materials.
New methods of analysis are nevertheless aiding in obtaining 
O
O
ion microprobe analyses of useful precision in silicates.
Jamtveit
and Hervig [1994] used ``extreme energy filtering'' to obtain
analyses of hydrothermal garnet with analytical
precision
on the order of 
. The data are sufficient to
delineate an abrupt decrease in 
of
5
from
core to rim with a spatial resolution of 50 
m. Jamtveit and
Hervig attributed the shift to a transition from magmatic fluid to
meteoric fluid compositions as the system evolved from a ductile to a
brittle strain regime.
During this last quadrennium, several laboratories [e.g., Weichert and Hoefs, 1993; Rumble et al., 1994] began investigations of the viability of substituting ultraviolet (UV) lasers for infrared lasers in order to obtain high-spatial resolution analyses of oxygen isotope ratios in minerals. Ultraviolet laser light is absorbed intrinsically by electronic transitions in minerals. Oxygen can therefore be liberated by UV irradiation without the localized heating that limits the spatial resolution of infrared laser methods. When combined with new He-entrainment mass spectrometry techniques [e.g., Hayes et al., 1990], the UV laser may substantially improve our ability to probe mass transfer between fluids and metamorphic rocks on the micrometer scale.
Studies of small-scale variations in 
O
O over the past several
years showed that exchange of oxygen isotopes between metamorphic
fluids and minerals can no longer be regarded as separate and independent
from advance of net-transfer reactions. Young [1993]
presented a methodology based on differential thermodynamics that
constitutes a necessary framework for evaluating small-scale
variations in 
O
O in the context of metamorphic parageneses formed
during fluid flow. Calculations of this type showed that net-transfer
reaction paths can strongly influence the ability of an external
fluid to effect changes in mineral 
values.
Oxygen and carbon isotopes are well-suited as tracers of mass transfer and transport in metamorphic rocks, but the need to identify the sources of metamorphic fluids with improved certainty has prompted development of other isotope tracers over the last several years. Notable among these are the isotopes of the noble gases He, Ne, Ar, Kr, and Xe. Irwin [1993] reported preliminary results of laser microprobe noble gas mass spectrometry that illustrate the potential of the isotopes of heavy noble gas elements (Ar, Kr) for constraining the origin of fluid trapped in individual metamorphic mineral grains.