Phase compositions were obtained with a Camebax (SX100) microprobe at Clermont-Ferrand University, using silicate, vanadate, and oxide standards (accelerating voltage = 15 kV; beam current = 20 nA at 1 mm diameter; counting time = 10 s). In the following section we will only consider Sites 1109, 1117, and 1118 because the rocks from Site 1114 are intensively altered and did not allow good mineral analyses.
At Site 1109 the rocks showing basaltic texture (Sample 180-1109D-41R-3, 72-74 cm) are composed of augitic clinopyroxene (Wo40En46-50Fs10-12) displaying low Ti content, ranging between 0.5 and 0.9 wt%. The Al content varies between 2.7 and 3.8 wt%. Iron, magnesium, and calcium content are, respectively, ~8, ~16, and 19.5 wt% (analyses 1 and 2 in Table T1). The clinopyroxene is associated with microlitic labradorite (An62) (analyses 1-3 in Table T2) retrogressed into albite (analysis 4 in Table T2).
In the coarse-grained diabase with well-preserved doleritic texture, one type of pyroxene is present in the samples located nearer the top of the hole. It displays a zonation from the core to the rim of the porphyroclast (Sample 180-1109D-48R-3, 63-69 cm [analyses 3-9 in Table T1]). According to Morimoto (1988), the pyroxene composition evolved from a magnesio- (Wo39-41En49-50Fs10) to a ferro- (Wo36En45Fs19) augite composition. The Al concentration decreases from 3.3 to 1.8 wt%; the iron content increases from the core to the rim. The titanium content is very low but increases from 0.5 to 0.9 wt%. In the coarse-grained diabase, the clinopyroxenes are associated with labradorite (An70) (analyses 5 and 6 in Table T2) depleted in calcium toward the rim (An55) (analyses 7 and 8 in Table T2). At the bottom of the hole in the coarse-grained diabase (Sample 180-1109D-51R-1, 58-67 cm), two types of clinopyroxenes are present, corresponding to subhedral porphyroclasts of augite associated with fibrous ferro-augite. The two clinopyroxenes are associated with unzoned phenocrysts of bytownite (An85) up to 1 cm in size (analyses 9 and 10 in Table T2) and labradorite laths.
In the fine-grained diabase (grain size < 2 mm), the clinopyroxene (Table T1) (Sample 180-1109D-51R-4, 6-9 cm) is a ferro-augite (Wo36En42-46Fs18-21). The Ti content is ~0.98 ± 0.2 wt%. The Al content varies between 2.03 and 3.62 wt% and does not change significantly with depth and texture (coarse- to fine-grained diabase). The Fe, Mg, and Ca contents are constant, at ~12, 15, and 19 wt% in clinopyroxenes (analyses 10-12 in Table T1). In the fine-grained diabase, the clinopyroxenes are associated with labradorite, with composition ranging between An75 and An55 (analyses 11-14 in Table T2).
The clinopyroxenes from the Site 1118 diabase have an augite composition (Wo40En46Fs14) (analyses 19-21 in Table T1). The Ti content is low (~0.5 wt%), and the Al content varies between 1.9 and 2.3 wt%. The Fe (7.5 wt%), Mg (16.5 wt%), and Ca (20 wt%) contents are constant. The associated plagioclases are either andesine (An45) or labradorite (An60-65).
In the coarse-grained zones, two distinct clinopyroxenes, either augites (Wo37En45Fs16) (analyses 22 and 27 in Table T1) or pigeonites (Wo9-13En46Fs40-44) (analyses 23-26 in Table T1), coexist, and the associated plagioclase is a labradorite (An55-65) (analyses 21-23 in Table T2) very often retrogressed into albite (An5) (analyses 19, 20, and 25 in Table T2).
At Site 1117, the clinopyroxene has a ferro-augite (Wo37-50En40-30Fs20-23) composition (analyses 13-18 in Table T1). Their compositions are characterized by very low aluminum (0.15 wt%) and titanium contents (<1 wt%). A few grains display higher Al contents, up to 2.3 wt%. The Ca content is constant at ~19 wt%, whereas Fe and Mg contents range, respectively, between 13 and 23 wt% and 6 and 15 wt%. The clinopyroxene was associated with a magmatic plagioclase; however, primary magmatic plagioclases were not analyzed because of the high degree of weathering of the rocks. Indeed, the plagioclase is totally transformed into albite (analyses 15-18 in Table T2).
Ilmenites with variable composition were analyzed in the diabases and the gabbro. The low sum of oxide components is indicative of alteration (Table T3). Oxide grains have very variable compositions from one sample to another wherever they are present (diabase, gabbro, or basaltic rocks). They display Mn contents that range between 5.4 and 0.4 wt%, Ti contents that range between 18 and 32 wt%, and Fe contents that range between 61 and 71 wt%. At Site 1117, the titanomagnetite is rimmed by sphene.
The diabases from Sites 1109 and 1118 and the gabbro from Site 1117 contain various amounts of secondary phases, such as amphiboles and sphene (only observed in the Site 1117 gabbro), chlorite, epidote, zoisite, titanite, and zeolites, observed in all studied mafic rocks. These minerals are representative of low-grade metamorphism. Only the composition of some amphiboles, chlorites, and epidotes will be discussed in the following sections.
According to the classification of Leake (1978), the amphiboles developed after clinopyroxene at Site 1117 display a magnesiohornblende to actinolitic hornblende composition (Sample 180-1117A-12R-1, 29-34 cm) with high Cl and F contents, indicating the presence of water-rich fluid during their crystallization (Table T4). Amphibole porphyroclasts are zoned, the Ti content decreases from 1.054 to 0.67 wt%, and the Al and Na contents decrease from 7.25 and 2.5 wt% to 4.6 and 1.4 wt%, respectively. The rim composition of the hornblende porphyroclast is similar to the composition of the fibrous green actinolitic hornblende.
Chlorite (Table T5) is present in three distinctive structural sites. At Site 1117, chlorite develops in a symplectite texture at the contact between the clinopyroxene and the plagioclase; it consists of a very fine association of albite and brunsvigite. A ripidolitic chlorite present as unoriented patches replacing clinopyroxenes is largely developed in Site 1117 gabbro and in the diabase at Sites 1119 and 1118. Finally, a green oolitic chamoisite associated with glauconite developed in veins from all sites.
Epidote and clinozoisite (Table T4) are present as very small grains in the brecciated rocks from all sites. They also appear filling veins as euhedral grains associated with quartz.
The AlIV/AlVI diagram (Fig. F3) used to discriminate between pyroxenes from high- and low-pressure origins (Aoki and Shiba, 1973) clearly indicates a shallow level of intrusion for both the diabases and the gabbro. Indeed, the magmatic clinopyroxenes from all the studied samples plot in the medium-pressure granulite field and low-pressure field of igneous rocks. Clinopyroxenes from the Southwest Indian Ridge oceanic gabbro (Hébert et al., 1991) and from Leg 149, Site 900 gabbro from the Iberian Abyssal Plain (Cornen et al., 1996) are also reported on the diagram and display similar compositions.
At Site 1117, the petrological observations indicate that the retrograde evolution started with the development of brown to green hornblende amphibole. Concentrically zoned amphibole grains indicate disequilibrium but may provide some record of changing conditions during the growth of the amphiboles. The pale greenish overgrowths on the extreme rims of the brown porphyroclast grains are thought to have formed during the cooling of the rock body. Indeed, experimental data has shown that the Ti content is temperature dependent (Liou et al., 1974; Helz, 1973) and decreases with decreasing temperature. The amphibole compositions presented in Table T4 indicate a decrease in Ti content from the core to the rim of the porphyroclast.
The Al and Na content of amphiboles in metamorphic and magmatic rocks is pressure dependent, decreasing with a fall in pressure (Liou et al., 1974; Moody et al., 1983). Moreover, the buffering reaction that relates the crossite to the Ca amphibole-bearing greenschist facies transition has been determined to be (Brown, 1974)
The Na(M4) content in brown porphyroclasts of amphibole suggest low pressure of crystallization, probably <3 kbar. The very low Na(M4) content of the amphibole rim and of fibrous green amphibole (~1.4 wt% of Na) suggest low-pressure conditions, probably <2 kbar (Brown, 1977). These pressures are in agreement with the pressure obtained using the Altot barometer of Plyusnina (1982), indicating 4 to 3 kbar for the amphibole core composition and 3 to 2 kbar for the amphibole rim composition. The Otten (1984) thermometer applied on the core and rim amphibole composition gives, respectively, 610° and 550°C. The (Holland and Blundy, 1994) hornblende-plagioclase thermometer applied with the amphibole rim composition and the albitic plagioclase indicates a temperature between 560° and 590°C. Thus, the decrease in Ti, Na, and Altot from the core to the rim of the porphyroclast (Table T4) is coherent with a decrease in pressure and temperature from 3 to <2 kbar and 610° to <550°C. These estimates are coherent with the plagioclase-hornblende assemblage domain of stability that ranges between 750° and 530°C.
At lower temperature, the plagioclase-hornblende association transforms into a plagioclase-actinolite-chlorite assemblage whose upper boundary was determined to be 430°C for ~1 kbar (Liou et al., 1987). The general presence of this assemblage indicates a common retrograde evolution below 430°C for all the sites. The chlorite, well developed in all sites, will help to precisely identify this evolution. Indeed, the chlorite composition can be described by a selection of major component and substitution vectors. The range of variation of composition is largely temperature dependent. Most representative relationships between temperature and site occupancy concern AlIV and octahedral vacancy. The Cathelineau and Nieva (1985) thermometer, based on the AlIV occupancy, applied on chlorite associated with albite in the symplectite gives temperatures between 250° and 220°C (Table T5). This thermometer applied on chlorites in replacing amphibole and clinopyroxene in the matrix gives temperatures of 204°-190°C for Sites 1117 and 1118 and 180°-233°C for Site 1109. The chlorite present in veins gave temperatures between 114° and 155°C for Sites 1117, 1118, and 1109, corresponding to subgreenschist facies conditions. To summarize, three generations of chlorite accompanied the low-temperature exhumation of the Moresby Seamount and growth during its deformation related to extension: a first one is present in the simplectite at 250°-220°C (these temperatures were obtained from Site 1117 chlorites only), the second generation of chlorite is pervasively present in the cataclastic rocks at 206°-190°C, and, finally, the third generation of chlorites is present in veins at 155°-114°C (these last two generations of chlorites are present at all sites).
As summarized in the pressure/temperature diagram (Fig. F4), the magmatic mineral association of diabase and gabbro corresponds to conditions of high temperature and fairly low pressure, conditions of igneous intrusion. From this point, retrograde metamorphism displays a drop in temperature from >610° to <300°C and the pressure decreases from 3 to 2 kbar to subsurface conditions. Following this path, the metamorphic associations are, successively, those of lower amphibolite and greenschist. According to the U/Pb and Ar/Ar dating, this evolution occurred between 66.4 ± 1.5 and 31.0 ± 0.9 Ma (Monteleone et al. and Brooks and Tegner, both this volume). Only the very end of the metamorphic evolution, occurring under subgreenschist to subsurface conditions, may be related to the rifting stage. In detail, during the rifting, the temperature evolution displays a discrete difference between the footwall (Site 1117) and the hanging wall (Sites 1118 and 1109) margins for the subgreenschist stage, showing higher temperature conditions (~30°-50°C) at Site 1117. These temperatures calculated on the metamorphic rocks give thermal paleogradients that are in good agreement with the present-day thermal gradients obtained at the sites during Leg 180 (see Fig. F2), indicating 95°-100°C/km above the active normal fault itself (Sites 1117 and 1108) and 31°-60°C/km on the hanging wall (Sites 1118 and 1109). This suggests that the different thermal regime between the footwall and the hanging wall, aided by the circulation of hydrothermal fluids, was efficient since the beginning of the rifting.
The texture developed by the metamorphic minerals in the metabasites does not vary a lot in the studied samples. The areal extent of rocks that contain aligned metamorphic minerals is much less than those in which minerals are randomly oriented or radiating. Aligned quartz or chlorite is evident only at Site 1117, which is located on the active normal fault. At this site, the brittle behavior of plagioclase and amphibole and the ductile behavior of quartz grains (Kirby, 1985) recrystallized in the foliation plane suggest that ductile deformation at Site 1117 occurred under greenschist to subgreenschist conditions. The presence of similar mineralogical assemblages in veins developed in all the sites suggests that the ductile deformation at Site 1117 occurred synchronously with the brittle deformation that is very well represented in the Moresby Seamount. It can be concluded that the last stages of the retrograde metamorphism must be attributed to the activity of the active normal fault that emerges at Site 1117.
Moreover, the geochronological data have shown that the Moresby Seamount basement was not thermally reset by subsequent rifting (Monteleone et al. this volume). That the crust remained cool during thinning is confirmed by the petrological data obtained on the diabase and gabbro from Sites 1117, 1118, and 1109, and this is true not only for the last million years but since the initiation of the rifting. This "cold" thermal regime in the upper crust is difficult to explain, considering the extensional setting. A solution is to consider that a major component of crustal thinning results from flow of the lower crust (Taylor and Huchon, this volume) and that the thermal effects of the thinning have yet to be transmitted to the upper crust. This is in agreement with the seismic tomography study, which indicates that the crust thins to an average thickness of 15-20 km flanking Moresby rift. This may help explain the geometry of numerous present-day passive margins where there is a lack of evidence for lower crust at the ocean-continent transition zone (Gardien et al., 2000).