Metamorphism and hydrothermal processes in axial magma chambers have been discussed with special reference to different spreading rates (e.g., Ito and Anderson, 1983; Mével and Cannat, 1991). According to the proposal of Mével and Cannat (1991), axial faults caused by lithospheric stretching extend into the lower crustal magma chambers in slow-spreading ridges. Through the high-temperature ductile shear zones formed by the axial faults, seawater penetration occurs under higher temperature (granulite and/or high amphibolite facies) conditions of >800°C. With cooling and an increase in penetrating seawater, metamorphic hydration proceeds into lower-temperature (greenschist facies and amphibolite/greenschist transitional) conditions. These synkinematic processes are followed by metamorphism under static conditions in which seawater penetrates through brittle failure at ~500°C. In contrast, there is no evidence for axial faults and high-temperature ductile shearing in deep crust at fast-spreading ridges (e.g., Hess Deep on the East Pacific Rise) (Gillis, 1995; Manning and MacLeod, 1996; MacLeod et al., 1996). Consequently, hydrothermal circulation and resultant metamorphism start with downward propagation of a cracking front caused by thermal contraction (Lister, 1974; Mével and Cannat, 1991). Thus, it has been thought that compared with slow-spreading ridges, the onset of metamorphism and hydrothermal penetration at fast-spreading ridges is later and under lower-temperature conditions of ~500°C (Mével and Cannat, 1991). However, Manning and MacLeod (1996) showed that the highest temperature recorded in the earliest amphibole veins is ~600°-750°C in the Hess Deep crust. So it is not clear whether the temperature of initial penetration of seawater is controlled by the difference in the spreading rate or whether these temperatures are higher in slow-spreading ridges than in fast-spreading ridges. Furthermore, fluid inclusion studies have revealed the exsolution of fluids from late-stage magmas at higher temperatures (>700°C) in oceanic gabbros of the slow-spreading Mid-Atlantic Ridge (Kelley and Delaney, 1987; Kelley et al., 1993; Gillis et al., 1993; Kelley, 1996, 1997). So it is necessary to distinguish the fluid phases of seawater origin and those of magmatic origin. On the other hand, based on thermodynamic calculations, McCollom and Shock (1998) mentioned a possibility that seawater circulation commenced at 750°-900°C with no mineralogical expression left in the oceanic gabbros, that is "cryptic alteration." Hart et al. (1999) pointed out that although the isotopic evidence for seawater circulation in Hole 735B gabbros is clear, magmatic effects dominate the trace element signatures.
In this context, the problem that should be addressed is what kind of mineral(s) are the expression of the earliest and highest-temperature metamorphism, and also, what kind of mineral(s) represents the earliest penetration of seawater and/or exsolution of magmatic fluids. Although it is not the purpose of this paper, the origin of amphibole, especially Ti- and Al-rich brown pargasitic amphibole, is one of the most critical problems for such a discussion of the magmatic/hydrothermal transition (e.g., Gillis, 1996; Tribuzio et al., 2000). Chemical composition of amphibole is sensitive to pressure, temperature, and other physical conditions and also to the chemistry of magmas, fluids, or protoliths from which the amphibole was formed. Therefore, in general, it is expected that the composition of amphibole will provide clues about its origin (e.g., Robinson et al., 1982). It is widely accepted that the origin of amphiboles contained in oceanic gabbros is very diverse, and amphiboles of magmatic, subsolidus metamorphic, and hydrothermal origin are included even in a single thin section. Actually, it is not easy to determine the origin of individual amphibole grains based on shape, texture, color, or chemical composition (e.g., Stakes et al., 1991; Vanko and Stakes, 1991; Gillis, 1996; Tribuzio et al., 2000). This is partly because amphibole formed at higher temperatures is easily overprinted or replaced by later-stage amphibole formed at lower temperatures. Thus, the overlapping by multiple generations of amphibole makes it difficult to discuss the origin of amphibole. As a result, it may not be meaningful to try to specify the origin of the individual amphibole grains.
Overall, although the metamorphic and hydrothermal processes that have occurred under relatively low temperature conditions have been described in detail for oceanic gabbros, those of the incipient stage under rather high temperature conditions have not been fully described, partly because of the difficulties due to overprint by later, lower-temperature products, as mentioned above. So we tried to focus on the transition from the latest stage of magmatism to the incipient stage of metamorphism recorded in the gabbroic rocks of Hole 735B. As a result, we found a new type of veins, the "high-temperature microscopic veins," from olivine gabbros. Information from the veins and related textures can reduce the difficulty in "amphibole-alone mineralogy" mentioned above and yields some constraints on fluid migration under high-temperature conditions in lower crustal olivine gabbros beneath slow-spreading ridges.