Colin J. Bray

Fluid chemistry of Archean seafloor hydrothermal vents: Implications for the composition of circa 3.2 Ga seawater

Seafloor hydrothermal vents of mid-Archean age (ca. 3230 Ma) have been identified and mapped in the Barberton greenstone belt, South Africa and are known as the Ironstone Pods. Fluid inclusion homogenization temperature data, when combined with gas chromatographic data, provide a minimum calculated water depth for the pods of 982 m. Ironstone Pod hydrothermal fluid endmember concentrations (Mg = 0) of various dissolved components derived from bulk fluid inclusion crush-leach experiments, include: Cl(730 mmol/L), Br (2.59), I (0.058), Na (822), NH4 (11.4), K (21.5), Ca (42.6), and Sr (0.15). This hydrothermal fluid also contains up to 1.07 mol% CO2, 0.03 mol% N2, 0.02 mol% CH4, 262 ppm COS, and minor amounts of C2–C4 hydrocarbons. Hydrothermal endmember Ca, Sr, and NH4, in particular, and to a lesser degree K, I, and CO2, commonly plot on, or very close to, modern vent fluid trends. By contrast, endmember Na and Br concentrations are distinct (higher) from modern vent fluids. High I and NH4 concentrations are consistent with contributions from sediments and/or organic matter. Calculated δ18OH2O values for the pod hydrothermal endmember fluid define a narrow range from 0.9 to 1.6‰ very similar to that of modern vent fluid values (0.4–2.1‰). A best estimate for the Ironstone Pod seawater endmember composition is Cl (920 mmol/L), Br (2.25), SO4 (2.3), I (0.037), Na (789), NH4 (5.1), K (18.9), Mg (50.9), Ca (232), and Sr (4.52). Barberton seawater components are commonly within an order of magnitude of modern seawater values, with the exception of significantly higher 1, NH4, Ca, and Sr in the inclusions. Sulfate concentrations are minimum estimates for Barberton seawater. Fluid inclusion samples containing the greatest amount of seawater component have higher N2 (up to 0.1 mol%) and low CO2, when compared to samples dominated by the hydrothermal endmember fluid. Barberton ambient seawater is considered to have been an evaporative brine of NaClCaCl2 composition during the time of pod deposition. Ironstone Pod fluid inclusion seawater endmember Br/Cl and I/Cl values of 2.45 × 10−3 and 40.2 × 10−6, respectively, are within error of bulk Earth (2.38 × 10−3 and 190 × 10−6) and are consistent with the chemistry of 3.23 Ga Barberton seawater being buffered by the mantle.

Fluid inclusion volatile analysis by heated crushing, on-line gas chromatography; applications to Archean fluids

Eighteen fluid inclusion volatile peaks have been detected and identified from 1–2-g samples (quartz) by gas chromatography using heated (∼ 105°C) on-line crushing, helium carrier gas, a single porous polymer column (HayeSep R; 10′ × 1/8″; 100/120#; Ni alloy tubing), two temperature programme conditions for separate sample aliquots, micro-thermal conductivity (TCD) and photoionization detectors (PID; 11.7 eV lamp) and off-line digital peak processing. In order of retention time these volatile peaks are: N2, Ar, CO, CH4, CO2, C2H4, C2H6, C2H2, COS, C3H6, C3H6, C3H4 (propyne), H2O (22.7 mins at 80°C), SO2, ±iso-C4H10±C4H8 (1-butene) ± CH3SH, C4H8 (iso-butylene), (?) C4H6 (1,3 butadiene), and ±n-C4H10 ±C4H8 (trans-2-butene) (80°C and −70°C temperature programme conditions combined). H2O is analysed directly. O2 can be analysed cryogenically between N2 and Ar, but has not been detected in natural samples to date. H2S, SO2, NH3, HCl, HCN and H2 cannot be analysed at present. Blanks determined by crushing heat-treated Brazilian quartz (800–900°C/4hrs) are zero for 80°C temperature programme conditions, except for a large, unidentified peak at ∼ 64 mins, but contain H2O, CO2 and some low molecular weight hydrocarbons at −70°C temperature conditions due to cryogenic accumulation from the carrier gas and subsequent elution. TCD detection limits are ∼30 ppm molar in inclusions; PID detection limits are ∼ 1 ppm molar in inclusions and lower for unsaturated hydrocarbons (e.g. ∼ 0.2 ppm for C2H2; ∼ 0.3 ppb for C3H6). Precisions are ∼±1–2%, except for H2O (∼±13%). Major fluid inclusion volatile species have been successfully analysed on a ∼50 mg fluid inclusion section chip (∼ 7 m × ∼ 10 m × ∼ 100 μm). Two distinct end-member Archean fluids, one internal and one external, have been found related to the Tanco zoned, granitic pegmatite, SE Manitoba. The former is an H2O (∼96%)-CO2 (∼40%)-CH4-N2 fluid (S species not included) with a moderate salinity of 6.6 ± 1.3 eq. wt.% NaCl which is interpreted to be magmatic in origin, whereas the latter is an H2O (∼97%)-CH4 (∼2%)-CO2 (∼0.4%)-C2H6-N2 fluid with a distinctly higher salinity of ∼10–20 eq. wt% NaCl which is interpreted to be of metamorphic/lower crustal (∼2,900 bar/∼10km) origin. The volatile compositions of H2O (∼87–94%)-CO2 (∼6–13%)-CH4-N2 fluids with ∼5–6 eq. wt.% NaCl from primary inclusions from three structurally controlled, mafic-ultramafic rock hosted Archean Au-quartz vein deposits in the Barberton greenstone belt, southern Africa (n=9) are distinctly different from the composition of the Tanco external fluid, but similar to the composition of primary fluids of interpreted

A discussion of the Jahns–Burnham proposal for the formation of zoned granitic pegmatites using solid-liquid-vapour inclusions from the Tanco Pegmatite, S.E. Manitoba, Canada

Jahns and Burnham (1969) proposed that the internal evolution of zoned granitic pegmatites could be explained by crystallisation from water-saturated melts which evolved to produce systems with a melt plus a separate aqueous fluid. Examination of microthermometric properties, chemical compositions and gas contents of solid-liquid-vapour inclusions from a number of the zones of the Tanco rare element granitic pegmatite places constraints on fluid evolution within the framework of the crystallisation history of the pegmatite, and contributes to an examination of the Jahns–Burnham proposal. Initial crystallisation at Tanco was from the wall rock inwards, producing the relatively unfractionated wall zone (potassium feldspar–quartz-albite-muscovite). Textural evidence, and an upward increase in the level of geochemical fractionation, indicate that much, but not all , of the subsequent crystallisation of the pegmatite was from the base upwards. Inclusions trapped by wall zone and metasomatic wall rock tourmaline indicate that the pegmatite was intruded as a 2 phase alumino-silicate melt/fluid mixture at ∼720°C, with an initial fluid composition of ∼98mol.% H 2 O (containing 2 equiv. mo1% NaCl) and 2 (containing 4 ). These observations indicate that both melt and fluid were present from the start of crystallisation (Jahns & Burnham 1969), but show that CO 2 and dissolved salts were important additional components of the fluid phase. The bulk of the pegmatite then crystallised in the range 600-470°C from melts and fluids with continued low levels of CO 2 (2-3mol.%) and approximately constant salinity (∼7 equiv. wt.% NaCl dissolved in the aqueous phase). Crystal-rich inclusions, which may represent trapped alumino-silicate melts, are present throughout pegmatite crystallisation down to temperatures as low as ∼262°C. The final stages of crystallisation resulted in the formation of the beryl fringe at 291 ± 33°C and the lower part of the quartz zone at 262 ± 29°C. By the later stages the fluid had cooled through an H 2 O-CO 2 – dissolved salt solvus resulting in H 2 O-CO 2 phase separation. Gas chromatographic analysis of the fluid components in the vug quartz, beryl fringe and lower part of the quartz zone shows that the inclusions contain H 2 O, CO 2 , CH 4 , N 2 , CO, Ar, and trace C 2 H 6 in the beryl fringe. Measured CH 4 :CO 2 ratios of 0·0060 (±0·0015) for the beryl fringe (twenty crushes on five samples) and 0·0042 (±0.0021) for the quartz zone (thirty crushes on six samples) yield f O 2 estimates of 1×10 −36 and 2 × 10 −38 , respectively, which are just above QFM at these temperatures.

Integrated cation–anion/volatile fluid inclusion analysis by gas and ion chromatography; methodology and examples

Abstract Combined gas and ion chromatographic analysis of well characterized, small (∼1 g) fluid inclusion-bearing samples is a powerful, but simple, means for obtaining integrated fluid concentrations of major and trace, volatile and ionic fluid constituents without using microthermometrically determined salinity for normalization. The methodology, which is described and assessed in detail, involves crushing a carefully cleaned sample at ∼105°C in a stainless steel crusher on-line to a gas chromatograph. After volatile analysis, the crushed sample is removed and leached with deionized water to produce a leachate solution, which is filtered and analyzed by ion chromatography, and other methods. For example, detailed procedures are given for I− analysis with a low detection limit of 0.5 ppb using a trace anion concentrator column and pulsed amperometric, rather than electrochemical, detection. The data are combined (calculation procedure given) to give whole fluid analyses in concentrations of mol% or mmol/l; this procedure removes a large systematic error of approximately a factor of 2, if separate volatile and hand crush cation/anion analyses are linked through sample weights. Results indicate a mean crushing efficiency of 82±6% for a mean sample mass of 0.97 g, based on 97 determinations. Tests on the efficiency of the leaching process showed that >90% of univalent ions are removed. Based on 145 analyses, charge balances are found to be close to 1.0 when all major positive and negative species are included, both analyzed and calculated (e.g., carbonate species). Analytical errors (coefficients of variation; percentage) for species in moles vary from 1.1 (Br−) to 17.9 (I−), from 0.02 (H2O) to 0.21 (I−) for species in mol%, and from 13.1 (F−, Br−) to 22.1 (I−) for species in mmol/l. Application of the combined GC/IC technique to well characterized sample sets from diverse settings, including the Tanco Li–Cs–Ta pegmatite, Archean Au–quartz vein systems, the Polaris MVT Pb–Zn deposit, and the ∼3.2 Ga Barberton sea-floor Fe-oxide deposits (ironstone pods) has demonstrated its utility for constraining fluid sources, identifying fluid types, and determining the effects of processes such as H2O–CO2 phase separation and fluid–wall rock interaction.

H2OCH4NaClCO2 inclusions from the footwall contact of the Tanco granitic pegmatite: Estimates of internal pressure and composition from microthermometry, laser Raman spectroscopy, and gas chromatography

Abstract Fluid inclusions in tourmaline and quartz from the footwall contact of the Tanco granitic pegmatite, S.E. Manitoba were studied using microthermometry (MT), laser Raman spectroscopy (LRS) and gas chromatography (GC). CH 4 -bearing, aqueous inclusions occur in metasomatic tourmaline of the footwall amphibolite contact. The internal pressures estimated from MT are lower than those obtained from LRS (mean difference = 54 ± 19 bars). The difference is probably due to errors in the measurement of Th CH 4 (V) and to the presence of clathrate at Th CH 4 (V) into which CO 2 had been preferentially partitioned. LRS estimates of pressure (125–184 bars) are believed to be more accurate. Aqueous phase salinities based on LRS estimates of pressure are higher than those derived using the data from MT: 10–20 eq. wt% NaCl. The composition of the inclusions determined by GC bulk analysis is 97.3 mol% H 2 O, 2.2 mol% CH 4 , 0.4 mol% CO 2 , 250 ppm C 2 H 6 , 130 ppm N 2 , 33 ppm C 3 H 8 , 11 ppm C 2 H 4 , and 3 ppm C 3 H 6 , plus trace amounts of C 4 hydrocarbons. The composition is broadly similar to that calculated from MT (92% H 2 O and 8% CH 4 , with 7 eq. wt% NaCl dissolved in the aqueous phase and 2 mol% CO 2 dissolved in the CH 4 phase), as expected due to the dominance of a single generation of inclusions in the tourmaline. However, two important differences in composition are: 1. (i) the CH 4 to CO 2 ratio of this fluid determined by GC is 5.33, which is significantly lower than that indicated by MT (49.0); and 2. (ii) the H 2 O content estimated from MT is 92 mol% compared to 98 mol% from GC. GC analyses may have been contaminated by the presence of secondary inclusions in the tourmaline. However, the rarity of the latter suggests that they cannot be completely responsible for the discrepancy. The differences may be accounted for by the presence of clathrate during measurement of Th CH 4 (critical), which would reduce CO 2 relative to CH 4 in the residual fluid, and by errors in visually estimating vol% H 2 O. The compositions of the primary inclusions in tourmaline are unlike any of those found within the pegmatite and indicate that the fluid was externally derived, probably of metamorphic origin. Inclusions in quartz of the border unit of the pegmatite are secondary and are either aqueous (18 to 30 eq. wt% CaCl 2 ; Th total = 184 ± 14° C ) or carbonic. Tm CO 2 for the carbonic inclusions ranges from −57.5 to −65.4°C and is positively correlated with Th CO 2 . Analyses of X CH 4 based on LRS agree within 5 mol% of those derived from MT and together indicate a range of compositions from 5 to 50 mol% CH 4 in the CO 2 phase. Bulk analysis by GC gives 99.0 mol% H 2 O, 0.6 mol% CO 2 , 0.4 mol% CH 4 , 160 ppm N 2 , 7 ppm C 2 H 6 , 4 ppm C 3 H 8 , and 2 ppm C 2 H 4 , with trace amounts of COS (carbonyl sulphide) and C 3 H 6 . The level of H 2 O in the analysis is consistent with the dominance of the aqueous inclusions in these samples, and the CH 4 : CO 2 ratios are consistent with estimates from MT and LRS. The preservation of variable ratios of CH 4 :CO 2 in inclusions 2 diffusion out of the inclusions nor reduction of fluids leaving the pegmatite were responsible for the more oxidized chemistries of the border unit inclusions relative to those in the tourmaline of the metasomatised amphibolite. The compositions of the inclusions in the quartz lie between those of the fluid trapped by the tourmaline (externally derived) and the measured composition of a CO 2 -bearing pegmatitic fluid, which indicates that the secondary fluids trapped in the border unit quartz were produced by late mixing.

Fluid inclusion volatile analysis by gas chromatography with photoionization micro-thermal conductivity detectors: Applications to magmatic MoS2 and other H2O-CO2 and H2O-CH4 fluids

Abstract Eighteen fluid inclusion volatile peaks have been detected and identified from 1–2 g samples (quartz) by gas chromatography using heated (~105°C) on-line crushing, helium carrier gas, a single porous polymer column (HayeSep R; 10′ × 1 8 ″: 100 120 #; Ni alloy tubing), two temperature programme conditions for separate sample aliquots, micro-thermal conductivity (TCD) and photoionization detectors (PID; 11.7 eV lamp), and off-line digital peak processing. In order of retention time these volatile peaks are: N 2 , Ar, CO, CH 4 , CO 2 , C 2 H 4 , C 2 H 6 , C 2 H 2 , COS, C 3 H 6 , C 3 H 8 , C 3 H 4 (propyne), H 2 O (22.7 min at 80°C), SO 2 , ± iso- C 4 H 10 ± C 4 H 8 (1-butene) ± CH 3 SH , C 4 H 8 (iso-butylene), (?) C 4 H 6 (1,3 butadiene) and ± n- C 4 H 10 ± C 4 H 8 (trans-2-butene) (80 and −70°C temperature programme conditions combined). H 2 O is analysed directly. O 2 can be analysed cryogenically between N 2 and Ar, but has not been detected in natural samples to date in this study. H 2 S, SO 2 , NH 3 , HCl, HCN, and H 2 ca nnot be analysed at present. Blanks determined by crushing heat-treated Brazilian quartz (800–900°C/4 h) are zero for 80°C temperature programme conditions, except for a large, unidentified peak at ~64 min, but contain H 2 O, CO 2 , and some low molecular weight hydrocarbons at −70°C temperature conditions due to cryogenic accumulation from the carrier gas and subsequent elution. TCD detection limits are ~30 ppm molar in inclusions; PID detection limits are ~ 1 ppm molar in inclusions and lower for unsaturated hydrocarbons (e.g., ~0.2 ppm for C 2 H 4 ; ~ 1 ppb for C 2 H 2 ; ~0.3 ppb for C 3 H 6 ). Precisions (1σ) are ~ ±1–2% and ~ ± 13% for H 2 O in terms of total moles detected; the latter value is equivalent to ±0.6 mol% at the 95 mol% H 2 O level. Major fluid inclusion volatile species have been successfully analysed on a ~50 mg fluid inclusion section chip (~7 mm × ~10 mm × ~100 μm ). Initial inclusion volatile analyses of fluids of interpreted magmatic origin from the Cretaceous Boss Mtn. monzogranite stock-related MoS 2 deposit, central British Columbia of ~97 mol% H 2 O, ~3% CO 2 , ~ 140–150 ppm N 2 , and ~16–39 ppm CH 4 (~300–350°C) are reasonable in comparison with high temperature (~400–900°C) volcanic gas analyses from four, active calc-alkaline volcanoes; e.g., the H 2 O contents of volcanic gases from the White Island (New Zealand), Mount St. Helens (Washington, USA), Merapi (Bali, Indonesia), and Momotombo (Nicaragua) volcanoes are 88–95%, >90% (often >95%), 88–95% and ~93%, respectively; CO 2 contents are ~3–10%, 1–10%, 3–8%, and ~3.5%. CO 2 /N 2 ratios for the Boss Mtn. MoS 2 fluids of ~ 190–220 are in the range for known volcanic gas ratios (e.g., ~ 150– 240; White Island). The ∑S content of the Boss Mtn. MoS 2 fluid prior to S loss by sulphide precipitation may have been ~2 mol% since CO 2 /∑S molar ratios of analysed high-temperature volcanic gases are ~ 1.5. This estimate is supported by ∑S contents for White Island, Merapi and Momotombo volcanic gases of ~2%, ~0.5–2.5%, and ~2%. COS has been determined in H 2 O-CO 2 fluid inclusions of interpreted magmatic origin from the Boss Mtn. MoS 2 deposit and the Tanco zoned granitic pegmatite, S.E. Manitoba at ~50–100 ppm molar levels, which are consistent with levels in volcanic gases. It appears that low, but significant, concentrations of C 2 -C 4 alkanes (~ 1–20 ppm), C 2 -C 4 alkenes (~ 1–480 ppb) and alkynes (e.g., C 3 H 4 ) have been detected in magmatically derived fluids (Boss Mtn. MoS 2 deposit; Tanco granitic pegmatite). Significantly higher, low molecular weight hydrocarbon concentrations have been determined in a CH 4 -rich (~ 2%), externally derived fluid of possible metamorphic or deep crustal origin trapped as inclusions in metasomatic wall-rock tourmaline adjacent to the Tanco pegmatite (e.g., 300/470 ppm C 2 H 6 ; 50 90 ppm C 3 H 8 ; 3–60 ppm C 2 H 4 C 3 H 6 n-C 4 H 10 ).

Fluid chemistry of veining associated with an ancient microearthquake swarm, Benmore Dam, New Zealand

Steeply dipping strata in the vicinity of Benmore Dam, Otago, New Zealand, are complexly deformed metasedimentary rocks of the Torlesse Supergroup (Middle Triassic age). Over an exposed area ∼100 m wide × 25 m high, these strata are disrupted by a fault-fracture mesh comprising conjugate Coulomb shears interlinked by extensional and extensional-shear fractures, all formed in a common stress field and hosting quartz + prehnite ± epidote ± calcite veining. The combined effect of these structures is shortening perpendicular to beddings and subvertical extension so that in their present attitude, they correspond to a set of conjugate thrust faults with associated extension fractures. On the evidence of incremental vein textures, the development of this distributed fault-fracture mesh is interpreted as resulting from a fluid-driven microearthquake swarm, which postdated regional low-grade metamorphism. Mechanical considerations suggest that the migrating hydrothermal fluids were significantly overpressured, possibly to approximately lithostatic values, if the mesh structure developed in its present attitude. Fluid-inclusion microthermometric studies show that Benmore vein quartz contains two-phase aqueous inclusions with salinities between 1.4 and 2.9 wt% NaCl equivalent and homogenization temperatures ( T h ) between 189 and 217 °C. The assemblage quartz + prehnite + epidote suggests trapping temperatures ( T t ) of ∼280 °C, requiring the addition of an ∼70 °C correction to T h values. Late calcite contains inclusions with noticeably lower salinity (0.0–0.9 wt% NaCl) and T h values (129–175 °C). Studies on quartz + pumpellyite ± calcite veins from nearby Lake Aviemore show similar fluid-inclusion salinity and T h values. Fluid-inclusion gas analyses show all the vein samples to be dominated by H 2 O (99.3–99.9 mol%) with few other gases apparent, including CH 4 (≤0.5%), N 2 (≤0.1%), CO 2 (≤0.1%), and C 2 -C 4 hydrocarbons. Cation and anion analyses, when combined with the gas data, show that NaCl dominates the fluid-inclusion salinities. Oxygen isotope results, when combined with calculated T t values, indicate that the water responsible for the deposition of Benmore and Aviemore quartz had δ 18 O compositions of 9.4‰ and 4.8‰, respectively. Calcite δ 13 C values between−25.3‰ and−38.0‰ are indicative of oxidation of CH 4 to CO 2 as a result of hydrothermal fluids interacting with organic- rich sediments. Fluid-inclusion \({\delta}D_{H_{2}O}\) values for Benmore range between −73‰ and −89‰ compared to−109‰ for the one Aviemore sample. This research has demonstrated that (1) water of meteoric origin, probably from subantarctic latitudes, penetrated to ≥6 km depth and underwent an oxygen isotope shift before depositing the Benmore-Aviemore veins; (2) the migrating hydrothermal fluids were likely overpressured well above hydrostatic to near lithostatic values if the mesh structure was active in its present orientation; and, (3) fluid migration was coupled to distributed brittle failure in the prevailing stress field, “self-generating” a permeable fault-fracture mesh.

Analysis of fluid inclusions in seafloor hydrothermal precipitates: testing and application of an integrated GC/IC technique

The principal research objective was to test an integrated gas (GC) and ion chromatographic (IC) technique for analysis of trapped fluids in seafloor hydrothermal precipitates and to compare the results with independently analyzed vent fluids. Twenty-three samples of chalcopyrite, pyrite, sphalerite, barite and anhydrite from hydrothermal chimneys from four seafloor sites, Axial Seamount (Juan de Fuca Ridge), the Vai Lili vent field (Lau Basin), the TAG hydrothermal field (Mid-Atlantic Ridge), and the 21°N vent field (East Pacific Rise), as well as of pyrite/marcasite and quartz from the TAG mound, were analyzed. A new type of blank for sulphides and sulphates was also developed. Water contents obtained by GC analysis are in agreement with known compositions of seafloor hydrothermal solutions. Also, the volatiles occur in the same order of abundance as in vent fluids CO2>N2(±CO±Ar±O2)≥CH4>COS>C2–C3 hydrocarbons. For all analyzed hydrothermal fluids, the mean Cl− concentrations are similar to those for the respective measured vent fluid compositions. Furthermore, the GC/IC results are directly comparable to salinity data determined by microthermometric methods. NH4+ can be enriched and is most probably formed by decomposition of organic matter. Volatile concentrations in fluid inclusions and data from vent fluids differ significantly for some of the investigated sites. Furthermore, the GC data indicate systematic variations regarding the volatile contents of trapped fluids from different active seafloor hydrothermal systems. Fluid inclusion volatile data of samples from the ASHES vent field at Axial Seamount define an inclined array consistent with phase separation.

Formation of tonalite in island arcs by seawater-induced anatexis of mafic rocks; evidence from the Lyngen Magmatic Complex, North Norwegian Caledonides

Abstract The Lyngen Magmatic Complex (LMC) of North Norway, consists of a western suite of layered gabbros of normal-mid oceanic ridge basalt (N-MORB) affinity and an eastern suite of layered gabbronorites, quartz-bearing gabbros and diorites/quartz-diorites of IAT (island-arc tholeiitte) to boninitic affinity. The boundary between the suites is defined by a large-scale ductile shear zone, the Rypdalen shear zone (RSZ). In this shear zone anatectic tonalites were generated by partial melting of the gabbro in the presence of an H 2 O bearing fluid phase. Quartz from the tonalites contains early secondary and secondary liquid-dominated inclusions (88–99 wt.% H 2 O), with an average salinity of 18 wt.% (calculated as NaCl eq ). Combined gas and ion chromatography shows that the major ions in the fluid are Cl − , Ca 2+ , Na + with smaller amounts of K + , Mg 2+ , Sr 2+ , Br − and NO 3 − . The dominant non-H 2 O volatile species is N 2 (0.5–10%), and small amounts of CO 2 , CH 4 and other hydrocarbons are also present. The cation concentrations in the fluid are variable, due to element exchange during interaction of the fluids with the tonalites, amphibolites and metagabbros of the RSZ. The fluid contributed Na + and K + to the melt and gained Ca 2+ in exchange, explaining the variable Na + /Ca 2+ ratio of the fluid. The Br − and Cl − contents of the fluid inclusions plot on the same line as evaporating sea water, which strongly suggests a seawater origin for the fluid phase, and a seawater source fits well with other geochemical signatures and the tectonic setting of the LMC. It is suggested that seawater escaped from a subducting slab and was channelled along the Rypdalen shear zone. This caused anatexis of the gabbro, generating tonalitic melts at 0.5–0.9 GPa and 680–800°C.