METAMORPHIC GEOLOGY AND SAMPLING STRATEGY

A thick carbonate sequence was sampled from the immediate contact zone around Cima Uzza to Monte Corona, over 1·5 km to the southeast (Figs 2 and 3). The sequence consists almost entirely of silica-poor dolomitic marble. Brack, (1984) divided the dolomitic marble into three stratigraphic units (Fig. 2), based on typical unit thicknesses mapped outside the contact aureole. The Calcare di Esino makes up the lower 300 m, and the Dolomia Principale the upper 200 m of the section. Between these units, Brack mapped 100 m of undifferentiated marble inferred to belong to the Breno, Gorno, or San Giovanni Bianco Formations. However, because of the homogeneity of the dolomitic marble, these stratigraphic divisions are typically equivocal in the field and neither Callegari, (1962) nor Bucher-Nurminen, (1982) applied stratigraphic distinctions to the dolomitic marble sequence. The following petrologic description focuses on the main carbonate lithologies from the dolomitic sequence sampled in this study: anchimetamorphic dolostone, silica-poor dolomitic marble, brucite marble, siliceous marble, and calcite veins and calc-silicate nodules and layers found in dolomitic marble host rock.

Figure 3. Geologic map showing sample locations and oxygen isotope ratios for carbonate minerals at Cima Uzza, Adamello contact aureole, northern Italy. Sample numbers are listed in parentheses after the [delta]¹⁸O values (per mil, SMOW). Intrusive samples analyzed in this study are also located. The forsterite isograd (Fo to the NW) is adapted from Bucher-Nurminen, (1982). Carbonate xenoliths and small mafic stocks and dikes are not mapped. Dolostone samples collected from Monte Corona (Fig. 2) are also not shown.

Outside the contact aureole on Monte Corona, the lowest ~100 m of the Calcare di Esino are preserved (Fig. 2). There, anchimetamorphic dolostone typically contains open pores up to several centimeters in diameter and <1% silicate minerals (this study; Callegari, 1962; Bucher-Nurminen, 1982). Calcite is typically present in trace quantities, and siliceous nodules or layers have not been observed (also Bucher-Nurminen, 1982).

The effects of metamorphism first appear north of Monte Corona, in the area around Passo del Frate (Figs 2 and 3). There, low-grade dolomitic marble contains rare grains of quartz, chlorite, talc, and tremolite, and is uniformly recrystallized. Porosity is visibly reduced compared with correlative rocks on Monte Corona. A distinctive but volumetrically minor feature of low-grade dolomitic marble is the presence of several dozen highly porous (10-30%), meter-scale enclaves that contain subequal amounts of calcite and dolomite. Bucher-Nurminen, (1982) may have interpreted these features to reflect a preservation of sedimentary porosity in the contact aureole (e.g. his fig. 1a). However, the abundant calcite in these porous enclaves occurs in a dense network of anastomosing veins that bisect coarse-grained dolomite crystals. Open pores are texturally associated with the veins. Because calcite was evidently introduced after metamorphic recrystallization, porosity seen at these low-grade localities seems unlikely to be a preserved sedimentary feature. Outside these larger porous enclaves in the outer aureole, millimeter-wide calcite veins cut host silica-poor dolomitic marble rarely.

At higher grades of metamorphism, silica-poor dolomitic marble remains the dominant lithology and occurs unchanged up to contacts with tonalite. This lithology contains <2% of non-carbonate minerals. Common accessory phases are humite group minerals (typically clinohumite), forsterite, spinel, phlogopite, calcite, chlorite, and serpentine (after forsterite). Chlorite occurs both as a fine-grained replacement of forsterite and spinel and as coarser-grained laths, possibly of prograde origin. Graphite is present only in forsterite-grade dolomitic marble in a limited area near the tonalite contact west of Cima Uzza, where it defines millimeter- to centimeter-scale layering. Opaque minerals include pyrite, pyrrhotite, geikielite, sphalerite, galena, hematite, ilmenite, and magnetite (this study; Callegari, 1962; Bucher-Nurminen, 1982).

Brucite marble is the most common lithology in carbonate xenoliths within gabbroic intrusions on the Uzza summit. Siliceous marble and much rarer silica-poor dolomitic marble also can occur within the same carbonate blocks. The distribution of siliceous marble and brucite marble within a xenolith does not appear to be spatially systematic. Brucite marble also commonly occurs near mafic dikes throughout the dolomitic contact aureole. In xenolith and aureole occurrences, brucite aggregates form up to 30 modal % of the rock. These aggregates are interpreted to be pseudomorphs after periclase, which is no longer present. A few brucite crystals display a tabular habit, suggesting that some brucite may also have formed as a rare prograde phase. Dolomite ( <= 30 modal %) is less abundant than calcite ( <= 60%). Most brucite marble samples also contain minor (<5%) forsterite, spinel, clinohumite, chlorite, and iron oxide and sulfide minerals (this study; Callegari, 1962).

Siliceous marble in xenoliths consists of an equigranular calcite matrix and up to 45 modal % of silicate and oxide minerals, mainly forsterite, spinel, and humite group minerals (typically clinohumite). Minor chlorite, serpentine, phlogopite, dolomite, pyrite, magnetite, galena, pyrrhotite, and pentlandite are also present (this study; Callegari, 1962; Bucher-Nurminen, 1982). Siliceous marble in xenoliths often is composed internally of millimeter- to centimeter-scale vein-like features defined by variable abundances of silicate minerals and spinel (Fig. 4a). In some samples, several coexisting spinel populations can be distinguished by size and grain shape (Fig. 4b).

Figure 4. (a) Plane-polarized photomicrograph of siliceous marble (sample 44a) from Cima Uzza showing twinned calcite matrix with alizarin stain, cut by numerous vein-like zones of forsterite and pleochroic green spinel. Scale bar represents 300 µm. (b) Plane-polarized photomicrograph of siliceous marble (sample 39b) from Cima Uzza showing twinned calcite matrix with alizarin stain, colorless round forsterite grains, and two coexisting pleochroic green spinel populations: large anhedral grains and small euhedral-subhedral grains. Scale bar represents 80 µm.

Lenses <20 cm in maximum dimension and discontinuous layers <10 cm thick of calc-silicate rock occur sporadically within silica-poor dolomitic marble in the contact aureole, primarily within the forsterite isograd (Fig. 5). Many of these calc-silicate features are similar to the siliceous marbles in xenoliths both in terms of mineralogy and modal abundances. Although they are volumetrically minor and altogether absent from many outcrops, the origin of the calc-silicate rocks within the aureole is of interest because siliceous layers are absent from Calcare di Esino dolostones on Monte Corona. Bucher-Nurminen, (1982) interpreted these layers as mineralized remnants of fluid channels that remained open during prograde metamorphism. However, clear veining is rarely seen in these occurrences and many calc-silicate nodules appear to be boudined marl layers (Fig. 5).

Figure 5. Calc-silicate nodules in silica-poor dolomitic marble host rock within the forsterite isograd zone from Cima Uzza, Adamello contact aureole. Individual calc-silicate segments (white arrows) are ~10 cm long, and could have formed from boudining of a marl layer.

Contact metamorphism of dolostone at Cima Uzza lacks a sequence of mappable isograds. Indeed, the forsterite-in isograd (Fig. 3; Bucher-Nurminen, 1982) denotes the metamorphic grade of most of the area. Furthermore, silicate minerals are too rare in low-grade dolomitic marble to map tremolite or talc isograds in such units. Additional complications are that retrograde tremolite (after forsterite) occurs within the regionally extensive forsterite zone, and that talc in the outer aureole may also be of retrograde origin (Bucher-Nurminen, 1982). Also, the presence of periclase (brucite) cannot be mapped as an isograd because it is present both in xenoliths and in centimeter- to meter-scale reaction rinds adjacent to the numerous mafic dikes throughout the contact aureole.

STABLE ISOTOPE DATA

Measurements of oxygen and carbon stable isotope ratios were made on minerals from representative carbonate rock lithologies from the contact aureole at Cima Uzza. Powder for bulk carbonate analyses was collected with a dental burr (1·0 mm diameter) or by powdering bulk samples (~1 cm³). Powdered samples were reacted with concentrated phosphoric acid (D = 1·92) at 50°C overnight to extract CO₂ gas from the carbonate powders (McCrea, 1950; Sharma & Clayton, 1965). Oxygen stable isotope ratios were also measured from silicate and oxide mineral separates from several samples of tonalite, gabbro, and siliceous marble. Such mineral separates were hand-picked to >95% purity after dissolving the carbonate fraction in HCl. Silicate analyses were made using the CO₂ laser probe extraction system at the University of Wisconsin (Valley et al., 1995). Isotopic ratios were measured with a Finnigan MAT 251 gas-ratio mass spectrometer. Oxygen isotopic ratios are given in standard per mil notation relative to SMOW, and carbon values are given relative to PDB. Eight analyses of NBS-19 yielded [delta]¹⁸O = 28·52 ± 0·04%° (1[sgr]) and [delta]¹³C = 1·88 ± 0·02%° (1[sgr]). Accepted values for NBS-19 are [delta]¹⁸O = 28·60%° and [delta]¹³C = 1·92%°. Raw silicate and oxide values were increased by +0·09%°, determined by adjusting repeat analyses of the UWG-2 standard (5·71 ± 0·09%°, n = 5) to its accepted value of 5·80%° (Valley et al., 1995). Stable isotope data are listed in Table 1 by rock type. Sample locations, rock type, and oxygen isotope data are shown in Fig. 3.

Table 1. Stable isotope ratios from Cima Uzza area, Adamello contact aureole

Anchimetamorphic dolostone

Eleven samples of anchimetamorphic silica-poor dolostone of the Calcare di Esino were collected for stable isotope analysis along the crest of Monte Corona, over 1·5 km southeast of Cima Uzza (Fig. 6). Four samples of anchimetamorphic dolostone from the Dolomia Principale were collected at Monte Columbine, 15 km southwest of Cima Uzza, the closest occurrence of Dolomia Principale outside the contact aureole. Analyses of dolomite from Calcare di Esino samples yield mean values of [delta]¹⁸O = 28·8 ± 1·1%° (1[sgr]) and [delta]¹³C = 2·5 ± 0·2%° (1[sgr]). Dolomia Principale dolomite analyses yield mean values of [delta]¹⁸O = 26·6 ± 1·2%° and [delta]¹³C = 2·4 ± 0·8%°.

Figure 6. Values of [delta]¹⁸O vs [delta]¹³C for dolomite from anchimetamorphic dolostone samples collected from the Calcare di Esino on Monte Corona, >1·5 km southeast of Cima Uzza, and from the Dolomia Principale on Monte Columbine, 15 km southwest of Cima Uzza.È

Dolomitic marble

Stable isotope analyses were made on 38 samples of silica-poor dolomitic marble at Cima Uzza (Fig. 7). Texturally homogeneous dolomitic marble is the most common lithology in the aureole. Dolomite analyses from samples of graphite-free dolomitic marble have a mean [delta]¹⁸O value of 29·7 ± 1·3%° and a mean [delta]¹³C value of 2·2 ± 1·3%° (n = 25). The single outlier in this group (66) occurs near a mafic dike. Dolomite analyses from samples of graphite-bearing dolomitic marble, found mainly on the western slope of Cima Uzza, have a mean [delta]¹⁸O value of 30·0 ± 0·9%° and a mean [delta]¹³C value of -0·9 ± 0·9%° (n = 13). Values of [delta]¹⁸O for dolomite from the two dolomitic marble groups are statistically indistinguishable, but [delta]¹³C values for graphite-bearing samples are ~3%° lower than those of the graphite-free group.

Figure 7. Values of [delta]¹⁸O vs [delta]¹³C for dolomite from silica-poor dolomitic marble samples from Cima Uzza. Dolomitic marble samples are subdivided into graphite-bearing and graphite-free populations. For comparison, the field for Calcare di Esino dolostone (Fig. 6) is also shown.

Siliceous marble

Eight stable isotope analyses were made on calcite from six samples of siliceous marble in carbonate xenoliths from Cima Uzza (Fig. 8). Values of [delta]¹⁸O and [delta]¹³C for calcite range from 13·2 to 18·5%° and -2·6 to -6·7%°, respectively. These values are both significantly lower than those for silica-poor dolomitic marble. The degree of centimeter-scale isotopic heterogeneity was evaluated in two samples displaying large petrologic differences on this scale (Fig. 4a). Calcite analyses from siliceous (calcite 50-60%) and non-siliceous zones (calcite >90%) <2 cm apart (samples 39 and 44) are different by <1%° for [delta]¹⁸O (Fig. 8). In samples 39b and 44a, fractionations between coexisting calcite, forsterite, and spinel are different from each other and diverge from predicted equilibrium fractionations at all temperatures (Fig. 9). A 1·6%° difference occurs between coexisting populations of large ( >= 0·2 mm diameter) and small spinel grains (<0·1 mm diameter) from sample 39b (Fig. 4b).

Figure 8. Values of [delta]¹⁸O vs [delta]¹³C for carbonate minerals from samples of siliceous calcite marble and brucite marble from Cima Uzza. Siliceous marble samples were collected from xenoliths, whereas brucite marble samples are both from xenoliths and from within the contact aureole near mafic dikes. The fields for dolostone from the Calcare di Esino (Fig. 6) and graphite-bearing and graphite-free dolomitic marble from Cima Uzza (Fig. 7) are shown.

Figure 9. Values of [delta]¹⁸O for coexisting calcite, forsterite, and spinel in two siliceous calcite marble xenoliths (samples 39b and 44a) from Cima Uzza. Measured fractionations in both samples are smaller than equilibrium fractionations predicted over the range of possible formation temperatures [800°C equilibrium fractionations shown, based on Chiba et al., (1989) and Zheng, (1991)]. Analyses from sample 39b indicate a 1·6%° difference between large ( >= 0·2 mm diameter, 12·3%°) and small spinels (<0·1 mm diameter, 13·9%°) (Fig. 4b).

Brucite marble

Bulk carbonate powders (dolomite/calcite = 0·9-0·2) from nine samples of brucite marble were analyzed (Fig. 8). Four samples are of xenoliths from Cima Uzza, and five were collected near representative mafic dikes within the contact aureole. The range in [delta]¹⁸O and [delta]¹³C values for these samples is 28·2 to 15·6%° and 0·1 to -6·7%°, respectively.

Mineralogy and field relationships indicate a dolomitic marble protolith for both aureole and xenolith occurrences of brucite marble. Gray graphite-bearingdolomitic marble is preserved in some xenoliths and shows a sharp transition to white brucite marble. In comparison with dolomitic marble, however, samples of brucite marble display highly variable isotopic compositions. Some bulk [delta]¹⁸O carbonate values are up to 15%° lower than dolomitic marble samples, but other samples have [delta]¹⁸O carbonate values that are similar to those of the dolomitic marble. Samples of brucite marble generally show higher [delta]¹⁸O and [delta]¹³C carbonate values than the siliceous marble with which they are spatially associated in xenoliths.

Calc-silicate zones in the aureole

Carbonate minerals from calc-silicate layers and nodules within the contact aureole (Fig. 5) were analyzed for stable isotopes along with the marble hosting them (Fig. 10a). Two samples are from calc-silicate nodules containing open porosity in a host rock of silica-poor dolomitic marble. Sample 141 displays an ~2 cm diameter nodule with antigorite replacing forsterite as well as interstitial calcite (20-40%), and dolomite (~10%).Dolomite in the host rock has a [delta]¹⁸O value of 27·6%° and a [delta]¹³C value of 1·6%°. Calcite in the nodule, sampled 2·5 cm away from the analyzed host rock, yields slightly lower values of [delta]¹⁸O = 26·6%° and [delta]¹³C = 1·1%°. Sample 116 contains a ~8 cm diameter calc-silicate nodule that consists of calcite (~50%), dolomite, forsterite, retrograde tremolite, and retrograde antigorite. Calcite in the vug has a lower [delta]¹⁸O value of 26·2%° and a lower [delta]¹³C value of 0·7%°, whereas dolomite in the host rock has a [delta]¹⁸O value of 27·6%° and a [delta]¹³C value of 1·9%°.

Figure 10. (a) Values of [delta]¹⁸O vs [delta]¹³C for carbonate in small calc-silicate nodules and layers and adjacent host rock from the Cima Uzza area. Sample numbers are indicated next to analysis or series of analyses (see Table 1). Fields for Esino dolostone from Monte Corona (Fig. 6), graphite-bearing and graphite-free dolomitic marble from Cima Uzza (Fig. 7), and siliceous calcite xenoliths (Fig. 8) are shown. (b) Values of [delta]¹⁸O vs [delta]¹³C for calcite veins in dolomitic marble matrix from Cima Uzza area. Sample numbers are indicated next to each series of analyses (Table 1).

Five other calc-silicate layers and nodules do not display open porosity. Sample 940 is a silica-poor dolomitic marble with several sub-centimeter-sized calc-silicate nodules (calcite, forsterite, spinel, chondrodite, chlorite, antigorite). Calcite analyses from two nodules <3 cm apart show a 2·5%° difference in oxygen and a 3·5%° difference in carbon. Dolomite in the host rock has values intermediate between the two nodules of [delta]¹⁸O = 29·2%° and [delta]¹³C = 2·8%°. Sample 107 is a graphite-bearing dolomitic marble xenolith in tonalite that itself hosts a >8 cm diameter calc-silicate nodule (calcite, phlogopite, forsterite, spinel, antigorite, geikielite). Dolomite in the host rock is isotopically identical to other samples of graphite-bearing dolomitic marble from the aureole ([delta]¹⁸O = 29·7%°, [delta]¹³C = -0·4%°). Nodule calcite is 3·7%° lower in [delta]¹⁸O and 1·2%° lower in [delta]¹³C than the dolomitic host rock. Sample 49, collected near gabbro, consists of siliceous marble (calcite, dolomite, forsterite, chondrodite, spinel) hosting a 3 cm thick, dark green, folded siliceous layer (calcite, dolomite, spinel, forsterite, and chondrodite). The folded green layer is mineralogically distinct from the white marble host rock only in the higher abundance and darker color of its spinel. The marble and siliceous layer have nearly identical isotopic compositions: [delta]¹⁸O = 26·0%° and 25·8%° and [delta]¹³C = 1·2%° and 1·2%°, respectively. Sample 103 is a calc-silicate nodule in calcite marble obtained <1 m from a gabbroic stock. The nodule (>10 cm * >8 cm * 4 cm, diopside, phlogopite, chlorite, forsterite, calcite) is rimmed by a 1 cm thick, calcite-rich rim (calcite, dolomite, diopside, phlogopite) in a graphite-rich calcite marble (calcite, dolomite, graphite, phlogopite). Values of [delta]¹⁸O for calcite from the nodule (103c), calcite rim (103b), and graphitic calcite marble host rock (103a) increase systematically from 23%° to 24·5%°. Values of [delta]¹³C also increase systematically from -3·2%° to -2·2%°. Dolomite from a typical graphite-bearing dolomitic marble 1 m away (sample 104) yields isotopic values of [delta]¹⁸O = 30·6%° and [delta]¹³C = 1·3%°. Sample 108, which was collected ~10 m from a tonalite contact, consists of graphitic marble (calcite, dolomite, graphite, forsterite) hosting a clinopyroxene-tremolite nodule which contains no carbonate minerals. Calcite in the host marble ~1·5 cm away from the nodule yields [delta]¹⁸O and [delta]¹³C values of 15·8%° and -4·9%°, respectively.

DISCUSSION

Fluid infiltration through silica-poor dolomitic marble

Bucher-Nurminen, (1982) interpreted the scarcity of silicate and oxide minerals in dolomitic marble from Cima Uzza as an indication that magmatic fluids did not infiltrate pervasively during metamorphism. Instead, he proposed that magmatic fluid infiltration in the aureole was highly channelized through pre-metamorphic porosity structures. Such channelization would explain the origin of small calc-silicate nodules and layers in the dolomitic marble country rock at Cima Uzza and link them genetically with the larger-scale occurrences of siliceous marble found in carbonate xenoliths.

However, the scarcity of Si- and Al-bearing phases in Cima Uzza dolomitic marbles, excluding the small calc-silicate zones, does not rule out the possibility of more pervasive fluid infiltration through the sequence. The spatial invariance of [delta]¹⁸O values in samples of dolomitic marble argues more convincingly against any significant amount of fluid infiltration through the bulk of carbonate country rock (Fig. 11). Pervasive infiltration of externally derived fluids should have altered the high initial [delta]¹⁸O values of these rocks systematically along their flow path. The magnitude of such an isotopic shift depends on factors such as the temperature and mechanisms of isotopic exchange, fluid source and volume, and flow path. In the case that magmatic fluids infiltrated in a dominantly down-temperature direction, pervasive infiltration would generate regional oxygen alteration trends that varied systematically from protolith [delta]¹⁸O values `downstream' of the fluid source to `upstream' [delta]¹⁸O values equilibrated with the input composition of ~10%° from magmatic fluid (e.g. Baumgartner & Rumble, 1988; Bickle & Baker, 1990; Bowman et al., 1994; Gerdes et al., 1995). Such a trend is not seen in dolomitic marble samples.

Figure 11. Values of [delta]¹⁸O and [delta]¹³C for dolomite vs distance to nearest intrusion contact from graphite-free dolomitic marble samples (black circles) and for graphite-bearing samples (gray circles). Calcare di Esino dolostone samples from Monte Corona (Fig. 6) are plotted at a distance of 1500 m. Other sample distances have been measured from the sample locality to the closest occurrence of either tonalite or gabbro on Cima Uzza. Some samples are much closer to small dikes than indicated.

An alternative infiltration model of up-temperature flow has been proposed to explain mineral distributions at several carbonate contact aureoles (e.g. Ferry, 1994). Up-temperature infiltration of fluids that are initially in isotopic equilibrium with outer aureole carbonates should cause a progressive increase in rock [delta]¹⁸O values towards the intrusion contact as fluid-rock fractionations decrease and as fluid [delta]¹⁸O values increase because of fluid-rock interaction (e.g. Dipple & Ferry, 1992; Bowman et al., 1994). This increase could be on the order of 5%° for aureole-scale flow, depending on reaction rates and transport properties. Values of [delta]¹⁸O in graphite-free dolomitic marble at Cima Uzza show a slight increase of ~1%° towards the contact at Cima Uzza. It is possible that this shift results from a small volume of up-temperature fluid flow. However, it more probably reflects small [delta]¹⁸O variations in dolostone protolith across the aureole. Protolith variations are plausible because the silica-poor dolomitic marble in the outer aureole (Calcare di Esino) is interpreted to be stratigraphically below the silica-poor dolomitic marble exposed near intrusion contacts (Dolomia Principale) (Fig. 2).

Values of [delta]¹³C for dolomite in graphite-bearingdolomitic marble are consistently lower than dolomite [delta]¹³C values in graphite-free dolomitic marble. One possible explanation for both the lower [delta]¹³C values and the presence of graphite is that the graphite-bearing marble formed when carbon-bearing magmatic fluids infiltrated graphite-free marble. Magmatic fluid infiltration is suggested because graphite-bearing marble occurs almost exclusively near tonalite (Fig. 3). However, this possibility is inconsistent with the observation that graphite-bearing marble samples show no shift in dolomite [delta]¹⁸O values towards magmatic values. This indicates exchange only with a carbon reservoir and not also with an oxygen reservoir. The possibility of magmatic fluid infiltration is therefore ruled out because H₂O is the dominant volatile species released from typical felsic magmas and carbon-bearing species are minor constituents (Symonds et al., 1994). A more plausible explanation is that the difference in dolomite [delta]¹³C values between graphite-bearing and graphite-free dolomitic marble results from protolith differences in organic carbon content. Mass balance calculations show that the average 3%° difference in dolomite [delta]¹³C values in graphite-bearing and graphite-free samples can be attributed to closed-system isotopic exchange of dolomite with 2 modal % of low (-30%°) [delta]¹³C organic carbon (now graphite) during metamorphism. This large amount of graphite is found in some graphite-bearing marble samples at Cima Uzza. Similar correlations between [delta]¹³C values and graphite content have been observed for other cases of carbonate metamorphism (Nabelek et al., 1992; Kitchen & Valley, 1995; Bergfeld et al., 1996).

Stable isotopic evidence against pervasive fluid infiltration at Cima Uzza is supported by petrologic modeling of closed-system metamorphism, which represents the H₂O-CO₂ composition of pore fluid as it evolves as a result of buffering by mixed-volatile mineral reactions. Following an equilibrium thermodynamic method (Greenwood, 1975; Flowers & Helgeson, 1983; Ferry, 1994), the development of silica-bearing dolomitic marble assemblages during closed-system prograde metamorphism can be calculated for representative bulk compositions and porosities at Cima Uzza. Dolomitic marble at Cima Uzza is characterized by the absence of diopside and by the widespread distribution of dolomite + calcite + forsterite assemblages. For a range of porosity and initial quartz contents, dolomitic marble is predicted to develop a significant diopside zone during closed-system metamorphism (Bowman & Essene, 1982; Ferry, 1994; Dipple & Ferry, 1996). However, the low quartz content (~1%) of the dolomitic marble at Cima Uzza severely limits the progress of siliceous dolomite reactions. Assuming equilibrium between a binary H₂O-CO₂ fluid and mineral phases, the predicted closed-system fluid evolution path during prograde metamorphism is shown in Fig. 12 for a pressure of 1 kbar. This isobaric T-X_CO2 diagram represents selected mass transfer reactions in the system CaO-MgO-SiO₂-H₂O-CO₂. The T-X_CO2 fluid evolution path represents a typical Cima Uzza dolomitic marble that initially contained 1% quartz and 94% dolomite and maintained a 5% porosity. This porosity is representative of typical Calcare di Esino dolostone on Monte Corona, and an initial X_CO2 of 0·05 is taken as a likely composition of connate water. During prograde metamorphism, the protolith sequentially intersects univariant reaction curves 1, 2 and 3 to form talc, tremolite, and forsterite (reactions given in Fig. 12). Although internally buffered X_CO2 values increase as reactions progress, X_CO2 remains below that required to stabilize diopside. The petrologic model is consistent with the index mineral sequence in silica-poor dolomitic marble at Cima Uzza. The result supports the stable isotopic data in indicating that chemically reactive fluids during prograde metamorphism did not pervasively infiltrate dolomitic marble at Cima Uzza.

Figure 12. Isobaric T-X_CO2 phase diagram for siliceous dolomite reactions in the system CaO-MgO-SiO₂-H₂O-CO₂ calculated using the Holland & Powell, (1990) thermodynamic database. Reactions that are not plausible for a silica-poor bulk composition are omitted for clarity. The thick continuous line tracks the evolution of fluid composition during prograde metamorphism in a closed-system assuming a rock with 5% porosity and a starting composition of 94% dolomite and 1% quartz at X_CO2 = 0·05. Numbered reactions are referred to in the text.

Localized infiltration of dolomitic marble

Although fluid infiltration was not pervasive, some hydrothermal alteration of dolomitic country rock occurred on a centimeter to meter scale. This alteration almost always occurs in close spatial association with mafic stocks in the contact zone at the Uzza summit and around dikes within the contact aureole. Brucite marble in the contact aureole appears to be only associated with widely distributed small dikes and does not otherwise occur in a systematic regional distribution. Likewise, abundant carbonate xenoliths and inliers in gabbro at the Uzza summit show evidence for localized hydrothermal alteration of a dolomitic marble protolith.

Xenoliths are interpreted to have a protolith of locally derived silica-poor dolomitic marble for several reasons. Graphite-bearing dolomitic marble, which is indistinguishable from dolomitic marble in the aureole, is preserved as remnant blocks in several xenoliths. These xenoliths otherwise contain spatially complex distributions of siliceous marble and brucite marble. The stratigraphic correlation between carbonate xenoliths and nearby graphite-bearing dolomitic marble in the aureole is also indicated by several partially reacted carbonate inliers in gabbro that can be traced out into the aureole. Although it is possible that some xenoliths were transported from lower in the stratigraphic section, this seems physically unlikely because the gabbroic stocks are small relative to the xenoliths they contain. It also is unlikely that overlying strata were incorporated as xenoliths because the present level of exposure at Cima Uzza is believed to be the roof zone of the batholith (Bucher-Nurminen, 1982).

Siliceous marble

Siliceous marble in xenoliths consists primarily of calcite, forsterite, chondrodite, and spinel. Modal abundances vary on a scale of centimeters, but are commonly as high as 45 modal % for forsterite and humite group minerals and 7% for spinel. In contrast, Si and Al are essentially absent from dolomitic marble in the contact aureole (<2% non-carbonate minerals). Siliceous marble found in xenoliths also has no siliceous equivalent of an appropriate size (meter scale) anywhere else in the contact aureole. Siliceous marble therefore most probably formed by metasomatic addition of Si and Al to a silica-poor dolomitic protolith. Local vein-like concentrations of silicates and spinel (Fig. 4a) provide textural support for a hydrothermal origin.

Despite a likely hydrothermal origin, neither the spatial distribution of siliceous marble and brucite marble zones nor available mass balance constraints on Si, Al, and O-isotopes provides a consistent picture of metasomatic transport processes and fluid fluxes in the xenoliths. The spatial distribution of siliceous marble and brucite marble zones is characterized by irregular meter-scale domains including some veining. Models for reactive mass transport based on a pure infiltration mechanism predict the development of sharp and spatially distinct reaction fronts that develop perpendicular to the direction of fluid flow [a silica metasomatism front, an oxygen isotope front, etc. (Gerdes & Valley, 1994; Ferry & Gerdes, 1998)]. In contrast, reactive mass transport that is driven only by diffusive exchange would result in symmetrical mineralogical and chemical zonation on all sides of the xenolith. Not surprisingly, the spatial distribution of silicates, spinel, and periclase-brucite in xenoliths corresponds to neither end-member transport mechanism. The scale of Si and Al metasomatism inferred for some xenoliths (up to 100 m diameter) requires fluid infiltration, possibly along fractures or high-permeability bands, because the scale is too large to be accounted for by fluid-hosted diffusive transport only. However, deformation and diffusional exchange at xenolith margins and between high- and low-permeability bands within the xenolith probably took place as well.

Given the large uncertainties in the geometry and mechanisms of fluid flow, mass balance models that quantify Si and Al metasomatism, oxygen isotope exchange, and periclase (brucite) formation (below) are necessarily limited. Simple mass balance calculations, ignoring the effects of diffusion and dispersion, are given here to illustrate approximate time-integrated fluid fluxes required for a metasomatic formation of siliceous marble based on Si, Al, and O transport (e.g. Gerdes & Valley, 1994). Constraints on silica metasomatism can be made by assuming that quartz-saturated H₂O infiltrates unreacted dolomite along a constant flow path. Quartz-saturated H₂O contains between 4·9 * 10^-2 and 5·8 * 10^-2 molal silica at T = 600-750°C and at P = 1 kbar (Johnson et al., 1992). This temperature interval is constrained by periclase-brucite phase equilibria (below) and is also consistent with contact temperatures estimated elsewhere in the Adamello aureole (Matile & Widmer, 1993; Abart, 1995). For such input silica concentrations, time-integrated fluxes of (2·6-3·1) * 10³ mol H₂O/cm² rock provide enough silica to form 40 modal % forsterite along a 3 m column (a characteristic silicate abundance and transport length in Uzza xenoliths). This is a minimum flux estimate because it makes the conservative assumptions that infiltrating fluids were quartz saturated and that all aqueous silica was extracted from the fluid to form forsterite. It is also a minimum estimate because silica solubility decreases with increasing CO₂ content (Walther & Orville, 1983). Petrologic estimates of X_CO2 = 0·1-0·4 for siliceous marble assemblages (Bucher-Nurminen, 1982) imply that the mixed H₂O-CO₂ fluid in xenoliths was able to transport less silica than was assumed in making this flux calculation.

Aluminum is considerably less mobile than silica. At likely conditions of siliceous marble formation (T = 600-750°C, P = 1 kbar), corundum-saturated H₂O contains only 1-1·5 ppm Al in the form Al(OH)o3 (Ragnarsdóttir & Walther, 1985). Given a starting concentration of 1·5 ppm Al in an infiltrating fluid, a minimum flux of ~3·3 * 10⁵ mol H₂O/cm² rock is required to produce 5 modal % spinel along a 3 m flow path. This minimum estimate is problematic because it indicates that over two orders of magnitude more fluid is required to form the observed spinel abundances than is required to produce observed silica concentrations. However, this discrepancy may be resolved because aluminum has been shown to form complexes with Na, K, and Cl, which increases its solubility by an order of magnitude or more (Anderson et al., 1987; Woodland & Walther, 1987; Baumgartner & Eugster, 1988). Uncertainties in fluid composition and pH currently limit the quantitative constraints that can be made on fluid fluxes required for aluminum metasomatism in siliceous marbles at Cima Uzza. However, aluminum metasomatism has been observed in carbonate rocks elsewhere in the Adamello contact aureole near intrusive contacts (Abart, 1995), and it is possible that aluminum mobility was enhanced at Cima Uzza by complexation with other dissolved components and by rapid pH changes induced by reaction with carbonate rock.

Stable isotopic data from samples of siliceous marble also show evidence for fluid-rock interaction. Values of [delta]¹⁸O for calcite from siliceous marble are variable but are as much as 17%° lower than a typical dolomitic marble [delta]¹⁸O value of 30%° (Fig. 8, [delta]¹⁸O = 13·2-18·5%°). Analyses of coexisting forsterite and spinel from two siliceous xenolith samples show correspondingly low [delta]¹⁸O values. Although all phases in siliceous marble have been similarly affected by low [delta]¹⁸O hydrothermal fluids, fractionations between coexisting phases are smaller than predicted equilibrium fractionations at any likely equilibration temperature (<800°C, Fig. 9). Further evidence for local isotopic disequilibrium is a 1·6%° difference between coexisting large ( >= 0·2 mm diameter) anhedral and small ( <= 0·1 mm) euhedral-subhedral spinel populations (Figs 4b and Fig. 9). Diffusional exchange between coexisting phases during cooling has been shown to account for differences in [delta]¹⁸O between different grain size fractions of diopside in Adirondack marble (Edwards & Valley, 1998). This [delta]¹⁸O difference develops because a larger volume percentage of a small grain is able to maintain isotopic equilibrium with surrounding grains through diffusion than is possible for a larger grain size (Eiler et al., 1992). However, diffusional re-equilibration during cooling cannot explain the 1·6%° difference in spinel populations in siliceous marble because the apparent calcite-spinel fractionation is smaller for the small grain size population than it is for the larger spinel grain size. These apparent fractionations are opposite that predicted for diffusional exchange during cooling. More likely, the isotopic disequilibrium observed between coexisting phases in siliceous marble reflects variations in fluid composition during mineral growth or kinetic limits to isotope exchange during rapid grain growth (e.g. Cole, 1994).

Although [delta]¹⁸O values in siliceous marble are low relative to dolomitic marble, values are not sufficiently low to indicate equilibration with nearby gabbros. Given limited solubilities of silica and aluminum, gabbroic stocks that entrain the carbonate xenoliths at Cima Uzza are the most likely sources for the hydrothermal fluids becausethey are the closest reservoirs of these two elements. Abundant hydrous phases, such as hornblende and biotite, and local pegmatitic textures indicate that the mafic magma contained water and might have become fluid saturated during cooling. Plagioclase, hornblende, and biotite separates from a gabbro stock collected near the Uzza summit (sample 41, Table 1) yield [delta]¹⁸O values of 11·0%°, 9·2%°, and 9·2%°, respectively. The whole-rock [delta]¹⁸O composition of ~11%° implied by these analyses is in good agreement with a granodiorite whole-rock analysis of 11·5%° from the southwestern margin of the batholith (Abart, 1995) but it is high relative to a whole-rock [delta]¹⁸O of ~9%° from nearby tonalite (Table 1). Estimates of time-integrated fluid flux, analogous to those made above based on Si and Al solubility, can also be made for oxygen isotope exchange. Oxygen isotope mass balance calculations indicate a minimum fluid flux of only 45 mol H₂O/cm² rock to completely equilibrate the oxygen isotopic composition of 3 m of marble to that of the fluid source. This flux estimate does not require assumptions about initial fluid and rock [delta]¹⁸O values and is a minimum value because it assumes equilibrium exchange between fluid and rock (Baumgartner & Rumble, 1988). The flux is orders of magnitude smaller than that required to produce the necessary Si and Al metasomatism. It is therefore problematic that [delta]¹⁸O values for siliceous marble are not lower (~12%°) and more uniformly equilibrated with gabbro. Diffusional exchange with unreacted high [delta]¹⁸O marble or kinetic limits to oxygen isotope exchange are necessary to account for the high and scattered isotope data for siliceous marble, assuming a metasomatic formation. The isotope data are less problematic if an alternative scenario is adopted that the protolith itself had sufficient Si and Al concentrations, because the high fluxes needed for Si and Al metasomatism are not required. The absence of an appropriate protolith elsewhere in the contact aureole, however, remains a significant shortcoming to this scenario.

Brucite marble

Brucite-bearing marble occurs both as the dominant lithology in xenoliths on Cima Uzza and as centimeter- to meter-wide reaction zones bordering mafic dikes or stocks in the aureole. Round aggregates of brucite are pseudomorphs after periclase, indicating at least two sequential reactions [reactions (4) and (5), Fig. 12]. At 1 kbar, periclase forms only at temperatures above 600°C; brucite is stable only below this temperature. At typical contact metamorphic temperatures (600-650°C), periclase formation also requires X_CO2 <= 0·1. Because the periclase-forming reaction [reaction (5), Fig. 12] produces pure CO₂, periclase growth will stop quickly at thesetemperatures unless internally generated CO₂ is diluted by the addition of externally sourced H₂O (e.g. Ferry, 1991). At T > 780°C and P = 1 kbar, dolomite breaks down to periclase + calcite regardless of the composition of the coexisting fluid. Although contact metamorphic temperatures as high as 900°C have been reported around gabbro bodies elsewhere in the Adamello aureole (Ulmer, 1982), the presence of unreacted dolomite within xenoliths requires T <= 780°C at Cima Uzza. Assuming no temperature gradient across the xenoliths, addition of at least some externally derived H₂O-rich fluids is therefore necessary to drive the formation of periclase in Uzza xenoliths. Silicate and oxide mineral abundances are low in brucite marble (forsterite/clinohumite <7%; spinel <2%), so that addition of Si and Al to the dolomitic marble protolith is not required. A second, possibly related, infiltration of H₂O-rich fluid is necessary to regress periclase to brucite during cooling.

Oxygen and carbon stable isotope values for brucite marble in xenoliths and near mafic dikes are more highly variable than siliceous marble found in xenoliths (Fig. 8). Values of [delta]¹⁸O in brucite marble range from 28%°, typical of unaltered dolomitic marble, to 16%°, typical of siliceous marble. Values of [delta]¹³C also range from a typical protolith value of 0·05%° to as low as -7%°. Some isotopic depletion can be attributed to Rayleigh distillation effects (e.g. Shieh & Taylor, 1969; Valley, 1986) but small fractionations between dolomite and CO₂ at high temperatures reduce this effect to <2%° for both carbon and oxygen if equilibrium is maintained, even for complete reaction of dolomite to periclase + calcite. Therefore, the isotopic data indicate exchange between brucite marble and a low [delta]¹⁸O and [delta]¹³C fluid. As with siliceous marble, a consistent spatial association suggests that the fluid source was dominantly mafic stocks and dikes.

The wide range of stable isotopic data for brucite marble samples can be understood by analogy with simple infiltration models for oxygen isotope exchange and for progress of the periclase-forming reaction (Baumgartner & Rumble, 1988; Bickle & Baker, 1990; Ferry & Rumble, 1997; Ferry & Gerdes, 1998). The calculations invoke the infiltration of isotopically distinct H₂O along a fixed flow path into an isothermal dolomite protolith and assume that isotopic and mineralogic reactions maintain local equilibrium. Under these circumstances, a fluid flux of 1000 mol H₂O/cm² rock will generate ~110 m of oxygen isotopic alteration over the range 600-750°C. The extent of oxygen isotope alteration along the flow path is insensitive to temperature over this range and scales proportionally with the cumulative fluid flux. As with the siliceous marble mass balance calculations, it is also completely independent of the initial [delta]¹⁸O values of fluid and rock, because the model assumes equilibrium exchange and therefore depends only on the relativeabundance of oxygen in the fluid and rock (Baumgartner & Rumble, 1988). In contrast, the extent of periclase formation depends strongly on temperature because of the influence of X_CO2 upon the reaction equilibria (Fig. 12, Ferry & Rumble, 1997; Ferry & Gerdes, 1998). Given the same input fluid flux of 1000 mol H₂O/cm² rock, periclase will form only in the first 13 m from the fluid inlet at 600°C (equilibrium X_CO2 = 0·02). In this case, all periclase-bearing rocks will be isotopically altered, as will an additional ~100 m of unreacted dolomitic marble along the flow path. In contrast, almost 1000 m of periclase will form for the same input flux along a constant 750°C flow path (equilibrium X_CO2 = 0·6). In this scenario, all oxygen isotopic alteration caused by fluid-rock reaction occurs in rocks that have reacted to form periclase, and additional periclase marble will form in advance of the oxygen isotope alteration front. These idealized calculations show that the amount of isotopic alteration within periclase-bearing rocks is strongly dependent on the temperature at which infiltration and reaction occurs. For a pure dolomite protolith at 1 kbar, the displacement distance of oxygen isotope alteration and periclase reaction fronts away from the fluid inlet coincide at ~670°C (equilibrium X_CO2 ~0·14). Isotope data for brucite marble samples both from xenoliths and near dikes include some isotopically altered and some unaltered (protolith) values. This indicates that oxygen isotope alteration was less widespread than periclase formation both within xenoliths and near dikes. The isotopic data for xenoliths and reaction rinds of brucite marble are therefore consistent with fluid infiltration at T > 670°C. At 700°C, a time-integrated fluid flux of 75 mol H₂O/cm² rock will completely react dolomite to periclase + calcite in a 15 m diameter xenolith by infiltration of pure H₂O, whereas as little as 0·5 mol H₂O/cm² rock will form a 10 cm thick rind of periclase next to a dike.

THE HYDROTHERMAL SYSTEM OF THE CIMA UZZA CONTACT AUREOLE

Hydrologic models of cooling plutons in the shallow crust predict that kilometer-scale hydrothermal systems should develop for a wide range of crustal permeability distributions, pluton sizes, and pluton and country rock temperatures (Ribando et al., 1976; Cathles, 1977; Norton & Knight, 1977; Norton & Taylor, 1979; Hanson, 1992, , 1995; Gerdes et al., 1998). Despite numerous simplifications, these hydrologic models appear to be broadly representative of many natural systems, because the time-integrated fluid fluxes predicted by them are similar to flux estimates based on petrologic and isotopic data from many contact aureoles (e.g. Baumgartner & Ferry, 1991; Ferry & Dipple, 1992; Manning et al., 1993; Bowman et al., 1994; Gerdes et al., 1998). Nevertheless, large-scale hydrothermal systems do not always develop as a result of pluton emplacement and cooling. Development of a hydrothermal system requires both permeable country rocks and a supply of fluid. One or both of these requirements appear to have been missing at Cima Uzza, because there is no evidence for large-scale, pervasive fluid flow in the thick dolomitic marble sequence of country rocks. Silica-poor dolomitic marble, which makes up the vast majority of outcrop at Cima Uzza, shows no statistically significant oxygen isotope shift relative to counterparts outside the contact aureole. Nearly all variations in carbon isotope ratios are correlated with the presence of graphite, reflecting sedimentary protolith rather than hydrothermal effects. In addition, tonalite contacts typically display a <20 cm thick skarn rim separating tonalite from a silica-poor dolomitic marble, which is isotopically unaltered and lacks brucite-periclase. Assuming that most calc-silicate rocks within the dolomitic marble sequence originated as marly layers, infiltration of tonalite-derived magmatic fluids appears to have been negligible even on a meter scale.

Local mineralogical and isotopic alteration is instead associated with the smaller gabbroic intrusions on Cima Uzza and with small mafic dikes that occur throughout the aureole. Alteration of carbonate xenoliths in gabbro is extensive. Brucite marble also develops on the centimeter to meter scale adjacent to mafic dikes and small gabbroic intrusions. Mafic intrusions therefore appear to have caused the limited amount of fluid-rock interaction observed in the contact aureole. Although gabbroic magmas are generally assumed to be low in volatiles, lenticular pods containing pegmatitic textures and the abundance of hydrous minerals in gabbros at Cima Uzza suggest that the magma was relatively rich in volatiles and could have exsolved a free aqueous phase during crystallization. In contrast, pegmatitic textures have not been observed in the tonalite. Gabbroic pegmatites have commonly been interpreted to form during the final stages of crystallization in the presence of high water activity, probably an exsolved aqueous fluid (e.g. Irvine, 1974; McBirney & Noyes, 1979; Viljoen & Scoon, 1985; Beard & Day, 1986; Larsen & Brooks, 1994). Gabbroic pegmatites have also been spatially linked to country rock mineralization in Greenland (Arnason et al., 1997). The spatial association of gabbro magmas and country rock alteration at Cima Uzza suggests that small gabbroic intrusions exsolved a reactive aqueous fluid during crystallization. The relatively small volume of these mafic bodies restricted country rock alteration to small areas near their contacts.

In many contact aureoles, high-temperature infiltration of meteoric fluids has been documented by stable isotope alteration patterns (e.g. Forester & Taylor, 1977; Norton & Taylor, 1979; Criss & Taylor, 1986). Hanson's, (1995) numerical models of fluid flow during contact metamorphism predict that meteoric fluid infiltrates an aureole in cases where magmatic fluid production is small and where country rocks are sufficiently permeable. At the Cima Uzza aureole, magmatic fluid production was limited, but stable isotopic evidence for meteoric fluid infiltration is also lacking. This implies that permeability in dolomitic marble country rock was low enough to prevent the development of a regional convection-type hydrothermal system that could introduce meteoric fluids (Norton & Knight, 1977; Hanson, 1995). Dolostone on Monte Corona contains large open pores and might have been rather permeable during early stages of metamorphism. However, gabbroic stocks have been interpreted as being slightly older than tonalite (Ulmer, 1982; Brack, 1984) and therefore may have been emplaced into permeable dolostone. This early stage of contact metamorphism associated with the gabbro could have driven recrystallization of dolomitic country rock, reducing permeability for the subsequent metamorphism associated with the tonalite. The small amount of initial quartz in dolomitic marble limits the extent of prograde reaction, so that permeability would not have increased during metamorphism in response to internal devolatilization reactions (e.g. Fyfe et al., 1978). In other contact aureoles, carbonate rocks with low silica contents have consistently been shown to sustain low permeabilities during metamorphism (Hover-Granath et al., 1983; Nabelek et al., 1984; Gerdes & Valley, 1994). Thus it seems likely that dolomitic marble permeability was low during the main pulse of thermal metamorphism associated with tonalite emplacement, such that a large-scale hydrothermal system could not develop despite long-lived thermal gradients. The occurrence of such a thick sequence of silica-poor dolomitic marble is relatively uncommon and may partially explain the limited hydrothermal alteration at Cima Uzza in contrast to other contact aureoles.

ACKNOWLEDGEMENTS

We thank G. Dipple, J. Fournelle, B. Grobéty, B. Hess, B. Klieforth, M. Kohn, N. Marchildon, and M. Spicuzza for assistance with many aspects of this project. Thoughtful reviews by M. Brandriss, J. Eiler, J. Farquhar and P. Nabelek improved this paper. This research was supported by NSF Grant EAR-93-16504 and by a National Young Investigator award (EAR-92-57160) to L.P.B., by NSF Grant EAR-93-04372 and DOE Grant 93ER14389 to J.W.V., and by Sigma Xi, Geological Society of America, and the Weeks Fund (Department of Geology and Geophysics, University of Wisconsin) fieldwork grants to M.L.G.

REFERENCES