Venting fluids and sea floor mineralization represent the end result of a complex series of physical and chemical reactions that occur between crustal rocks and seawater in the subsurface portion of active hydrothermal systems. The fluids are generally considered to achieve their characteristics at the base of the hydrothermal cell in a reaction zone where temperatures are 375 -425 C and water/rock ratios are low [e.g. Seewald and Seyfried, 1990; Seyfried et al., 1991]. However, the link between the subsurface alteration assemblages and the fluid chemistry of active hydrothermal vents is not yet well constrained [ Gillis and Thompson, 1993]. As the oceanic crust is transported away from the ridge, the water-rock reactions evolve as the characteristics of the hydrothermal circulation respond to cooling, changes in the physical properties of the crust (in particular porosity and permeability), and reorganization of the pathways of fluid flow. Consequently, the mineralogy and chemistry of hydrothermally altered rocks recovered from the sea floor and from ophiolites reflect the integrated effects of all axial and ridge flank hydrothermal processes and provide a record of the extent of chemical exchange between the crust and seawater, the nature of the reactions that have taken place, and the variation in the composition of the circulating fluids.
Over the last fifteen years, studies of altered rocks from the sea floor and from ophiolites have been combined with experimental and theoretical studies to investigate the variations in water-rock interactions in different portions of submarine hydrothermal systems. These studies have recently been the subject of two detailed reviews: one that describes the alteration reactions within the recharge, reaction and discharge zones [ Alt, 1994], and one that addresses the physical and chemical characteristics of hydrothermal reaction zones based on experimental and theoretical considerations [ Saccocia et al., 1994]. Perhaps one of the most significant areas of progress in the last four years has been the investigation of alteration processes within lower crustal rocks. These studies have been enhanced by drilling into oceanic crust, which has recovered stratigraphic sections from the lower sheeted dikes and upper gabbros. In combination with studies of ophiolite sequences and experimental work, these investigations have resulted in the recognition that the reaction zone (where hydrothermal fluids are thought to acquire their final composition) may well lie within the lower part of the sheeted dikes and in the upper gabbros.
Based on geochemical and isotopic data from altered oceanic and
ophiolitic rocks, as well as modeling studies, hydrothermal convection
in the oceanic crust is most likely layered, with large volumes of
seawater circulating through the upper volcanics and reacting
at temperatures of <150 C, and only a small percentage penetrating
into the sheeted dikes and upper gabbros [e.g. Alt et al., 1986;
Gillis and Robinson, 1990; Hart et al., 1994; Rosenberg
et al., 1993]. This is supported by the observation that the
transition from volcanics to sheeted dikes at Deep Sea
Drilling Project(DSDP)/Ocean Drilling Program (ODP) Hole 504B, located
on the flank of the Costa Rica Rift, and in many ophiolites coincides with
a change from low temperature alteration to greenschist facies
metamorphism [ Alt et al., 1986; Gillis and Robinson, 1990].
In 1991, Hole 504B was extended to a depth of 2000 meters below
seafloor, and seismic and petrographic evidence indicates that the
bottom lies near the base of the sheeted dike complex [ Dick et
al., 1992]. The mineralogy and chemistry of the lower 500-600 m of
the sheeted dikes include secondary Ca-rich plagioclase (which is
locally replaced by anhydrite) and hornblende, increasing Al and Ti
contents of amphibole, and lower 
O values, suggestive of
high temperatures of alteration (400-500 C) consistent with
those generally attributed to the reaction zone [ Alt et al.,
1994]. However, a different mineral assemblage that includes
Na-rich plagioclase, amphibole (actinolite-magnesio-hornblende),
chlorite, and sphene, has been described in metabasalts recovered from
the Mid-Atlantic Ridge south of the Kane Fracture Zone (MARK) and is
also interpreted to have formed within the reaction zone [ Gillis
and Thompson, 1993]. The most problematic difference between these
two assemblages is the composition of the plagioclase. Experimental
studies suggest that the Ca/Na
concentration ratio of vent fluids
are consistent with plagioclase-quartz-fluid equilibrium
under pressure-temperature conditions resembling those in the reaction
zone for a plagioclase composition rich in Ca [ Berndt and
Seyfried, 1993; Saccocia et al., 1994]. Consequently, the
existence of Na-rich plagioclase in the MARK metabasalts is
inconsistent with this equilibrium model. This indicates that either (i)
a non-equilibrium process must be invoked to explain these variations,
(ii) that magmatically-derived fluids (which can undergo phase
separation) play an important, but as yet unquantified, role in
reactions within the reaction zone, or (iii) the MARK metagabbros are
not representative of the mineral assemblages in the reaction zone.
Another important observation is the scarcity of epidote in
hydrothermally altered oceanic rocks when compared with ophiolites
and experimental predictions [e.g. Seyfried et al., 1991;
Bettison-Varga et al., 1992; Nehlig et al., 1994]. However,
it must be emphasized that these comparisons are based on a very
limited sampling of rocks from the lower sheeted dikes of the oceanic
crust, and that wider variations in their mineralogy and chemistry than
are presently recognized should be anticipated as additional material
is recovered.
Hydrothermally altered plutonic rocks from the sea floor collected
by dredging, drilling and from submersibles indicate that the reaction
zone extends into the upper gabbros. The recovery of a 500 m-long section
of gabbro from ODP Hole 735B in the Southwest Indian Ocean has
enabled reconstruction of its alteration history and has provided
an excellent comparison with hydrothermally altered gabbros
sampled elsewhere on the sea floor [e.g. Vanko et al., 1992;
Gillis et al., 1993; Hekinian et al., 1993]. The
distribution and degree of alteration observed in both suites of rocks
are clearly related to permeability and deformation on a local scale,
and the drill core shows no depth-related metamorphism that reflects
an increase of temperature along a fixed gradient [ Stakes et
al., 1991]. Metamorphism in the upper portion of the core is
spatially related to deformation. Ductile shear zones exhibit
extensive dynamic recrystallization that began at temperatures of
granulite grade (900-700 C) and continued to amphibolite
grade (450 C). Penetration of seawater into the lower crust
occurred along zones of brittle-ductile deformation and through networks
of cracks adjacent to them, and is evidenced by depletions in
O and increasing abundance of both amphibole of
variable composition and metamorphic plagioclase of
intermediate composition. Downcore correlations of contemporaneous
mineral assemblages, oxygen isotopic compositions, and vein
abundance indicate that seawater was introduced by way of small cracks
and veins produced at the end of the phase of ductile deformation
and resulted in a hydrous ductile-deformation assemblage in the upper
parts of the section. In the lower parts of the section, circulation
of fluids is controlled by the distribution of highly permeable
magmatic hydrofracture horizons. The enhanced permeability of these
zones produced lower temperature greenschist and zeolite facies
assemblages as larger volumes of water penetrated the crust [ Stakes
et al., 1991; Vanko and Stakes, 1991] .
This general sequence of events---plastic deformation and metamorphism at high temperatures, followed by brittle deformation and the circulation of fluids, and then cooling with further alteration at lower temperatures---is observed in other sea floor metagabbros [e.g. Mével and Cannat, 1991; Vanko et al., 1992; Alt, 1994], although the details vary in response to the magmatic and tectonic interactions within each area. For example, the MARK metagabbros show a similar relation of hydrothermal alteration to deformation, although deformation was initiated at lower temperatures (700-550 C) of the amphibolite facies. Brittle fracturing on a variety of scales then provided pathways for more pervasive penetration of seawater and alteration to greenschist and lower amphibolite facies (up to 550 C). Alteration ceased at temperatures of between 80-300 C, suggesting that the lower crust became impermeable as it was transported off-axis [ Gillis et al., 1993; Kelley et al., 1993].
Calculations of pressures based on a combination of fluid inclusion
and isotopic analyses suggest that the gabbro section sampled by Hole
735B may have originated at a depth of 
2 km below the sea floor near
the top of the plutonic section [ Vanko and Stakes, 1991]. This
is consistent with depths estimated using quartz geobarometry on
axial hydrothermal fluids (summarized in Von Damm [1990]), and
with exposures in ophiolites, where only the top few hundred meters of
the gabbro section are altered [e.g. Nehlig et al., 1994].
However, the timing of fluid penetration, the pathways of fluid flow,
and the depth of penetration of hydrothermal fluids into the gabbros are
not yet well constrained and are questions that are just beginning to
be addressed by deep crustal drilling.