Before the discussion on the origin of the high-temperature microscopic veins, we need to pay attention to the relation between the veins and the interfingering texture of clinopyroxene. As already mentioned, these two features contain the intergrowth of brown amphibole + orthopyroxene in clinopyroxene grains. These two kinds of the intergrowths are essentially similar to each other in texture (Figs. F6, F7) in chemical compositions of the involved minerals (Figs. F8, F9, F10, F11) and in the calculated temperatures (Fig. F12). These observations suggest that both intergrowth textures, and therefore both the high-temperature microscopic vein and the interfingering clinopyroxene, were formed under nearly the same conditions at the same time. We propose here that the interfingering texture of clinopyroxene and the high-temperature microscopic veins were formed through a common process. This is supported by an intimate association of these two features observed in several thin sections (Figs. F6A, F13A, F13C).
There are two candidates for the origin of the high-temperature microscopic vein, namely silicate melt origin or fluid origin. Although we have no direct evidence such as geochemical data from the veins, we can discuss it based on the occurrences of the veins themselves. The discontinuous and irregular occurrence (Fig. F5) and the along-vein variation in mineralogy suggest that these veins were formed by reaction of magmatic minerals with some kind of fluids under the conditions of low fluid/rock ratios. This is supported by the composition of the minerals in the veins. Al- and Ti-poor features of the secondary clinopyroxene (Fig. F9) compared to primary clinopyroxene are also reported from the Skaergaard intrusion as product of exchange between primary minerals and fluid (Manning and MacLeod, 1996). Orthopyroxenes and brown amphiboles from the veins within olivine grains are characterized by lower Ti, Ca, and Al contents and Ti content, respectively, compared to these minerals within clinopyroxene grains (Figs. F10, F11). Thus, the vein-constituting, secondary minerals inherited the chemical signatures from the primary, precursory phases, which suggests a reaction between the primary phases and migrating fluids.
Based on petrographical observations, the reactions between magmatic minerals and migrating fluids can be written as follows:
We cannot quantitatively determine these reaction equations because of the heterogeneity in mineral composition and the lack of accurate data on volumes of the involved phases and the actual volume changes in the examined samples. In any case, addition and/or subtraction of components by fluid migrations are required for all the above equations, although the compositions of the fluids are not elucidated in this study.
There are also, in this case, two candidates for the origin of the fluid phase, that is, magmatic fluid and hydrothermal fluid originated from seawater. Although it is very difficult to get unequivocal answer at present, we would like to try to discuss the origin of the fluid phase based on available information.
In general, it is accepted that the penetration of seawater does not predate ductile deformation due to lithospheric stretching within Layer 3 gabbros of slow-spreading ridges (Mével and Cannat, 1991). Therefore, the origin of the fluid phase cannot be attributed to the penetration of seawater, if there is evidence suggesting that the timing of the veins predated ductile deformation. As shown in the previous section (Figs. F3, F5), olivine, plagioclase and clinopyroxene were dissolved by the penetration of these veins and there is no mechanical offset along the high-temperature microscopic veins. This indicates that the penetration of the veins did not accompany the deformation and also did not take place through brittle fracture networks. Figure F13B shows the interfingering texture in deformed olivine gabbro (Sample 176-735B-148R-1 [Piece 2, 104-112 cm]). In this sample, the interfingering texture is clearly involved in the ductile deformation. This indicates that the formation of the interfingering texture predated ductile deformation. Furthermore there is evidence suggesting that the formation of the veins preceded the complete solidification of the gabbroic crystal mush. Most of the high-temperature microscopic veins clearly penetrate not only cumulus phases but subhedral to anhedral intercumulus phases as well (Figs. F3, F5, F6). However, some high-temperature microscopic veins occur along grain boundaries of magmatic minerals (Fig. F13C) (Sample 176-735B-101R-1 [Piece 7, 106-108 cm]). In this case, the high-temperature microscopic vein is running in the outer edge of a plagioclase grain, and a very fine-grained intergrowth of orthopyroxene + brown amphibole was formed in the edge of a clinopyroxene in contact with a high-temperature microscopic vein. These textures suggest that the timing of the penetration of the high-temperature microscopic veins might be before complete solidification of the gabbroic crystal mush. The boundary of the cumulus phases appears to have been still wet with trapped evolved melt, so that it was easier to move along the boundaries of magmatic phases. These lines of petrographic evidence suggest that the penetration of the high-temperature microscopic veins did not take place in completely solidified gabbros. Therefore, it is concluded that the penetration of the high-temperature microscopic veins predated ductile deformation. This conclusion indicates that the high-temperature microscopic veins cannot be related to the seawater migration. So by elimination, our preliminary conclusion is that the fluid phase responsible for the high-temperature microscopic veins and interfingering clinopyroxenes is of magmatic origin, although >90% crystallization of mid-ocean-ridge basalt (MORB) magma is necessary for the saturation and exsolution of fluid phase because of the nearly anhydrous nature of MORB (e.g., Nehlig, 1993; Dixon and Stolper, 1995; Dixon et al., 1995). This preliminary explanation is dependent on the hypothesis that the penetration of seawater does not predate ductile deformation within Layer 3 gabbros of slow-spreading ridges; therefore, we need additional data directly implying magmatic origin, such as isotopic and trace element data.
Kelley and Delaney (1986) and Kelley et al. (1993) mentioned, based on fluid inclusion data, that the exsolution of fluid from the latest-stage magma took place at a temperature >700°C in the MARK area. However, no obvious mineralogical expression for these magmatic fluids has been found until now. A plausible candidate for the mineralogical expression of the magmatic fluid is brown amphibole, which is observed within a corona of olivine (Kelley et al., 1993, for the MARK gabbros). Tribuzio et al. (2000) discussed the origin of interstitial brown amphibole from gabbros in the North Apennine ophiolites, which are considered to be formed at a slow-spreading ridge. They proposed that the brown amphibole might be a product of an interaction between infiltrating exsolved magmatic fluid and gabbroic crystal mush. We now propose an alternative explanation that the high-temperature microscopic veins we found are a mineralogical manifestation of the migration of fluids derived from late-stage magma beneath spreading ridges.
Because there is no axial fault and high-temperature ductile shear zone in the deep crust at fast-spreading ridges, it is difficult to assume seawater penetration at higher-temperature conditions. So it is proposed that the onset of metamorphism and hydrothermal penetration occurs at ~500°C and is triggered by downward propagation of a cracking front (Mével and Cannat, 1991). However, hydration at temperatures of ~800°-600°C is reported from the Hess Deep gabbros (e.g., Gillis. 1995; Manning and MacLeod, 1996; Manning et al., 1996). To the proposal by Manning and MacLeod (1996) that localized brittle deformation near the magma body under higher-temperature conditions promoted seawater-rock interaction, we can add another explanation that high-temperature hydration occurs by migration of fluid phases from evolved magmas at fast-spreading ridges without any incorporation of seawater. The exsolution of fluids from evolved magmas should not be restricted to slow-spreading environments, and therefore the veins described here should also be found at fast-spreading ridges.