Widespread Assimilation of a Seawater-Derived Component at Loihi Seamount, Hawaii

Home , Seawater

Geochimica et Cosmochimica Acta, Vol. 63, No. 18, pp. 2749–2761, 1999 Copyright © 1999 Elsevier Science Ltd Pergamon Printed in the USA. All rights reserved 0016-7037/99 $20.00 ϩ .00 PII S0016-7037(99)00215-X Widespread assimilation of a seawater-derived component at Loihi Seamount, Hawaii

1,2, 3 4 1 2 5 ADAM J. R. KENT, *DAVID A. CLAGUE, MASAHIKO HONDA, EDWARD M. STOLPER, IAN D. HUTCHEON, and MARC D. NORMAN 1Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA 2Lawrence Livermore National Laboratory, Livermore, CA 94551, USA 3Monterey Bay Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, CA 95039-0628, USA 4Research School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australia 5Centre for Ore Deposit Research, School of Earth Sciences, University of Tasmania, Hobart, TAS 7001, Australia

(Received February 4, 1999; accepted in revised form June 8, 1999)

Abstract—Many tholeiitic and transitional pillow-rim and fragmental glasses from Loihi seamount, Hawaii,

have high Cl contents and Cl/K2O ratios (and ratios of Cl to other incompatible components, such as P2O5, H2O, etc.) relative to other Hawaiian subaerial volcanoes (e.g., Mauna Loa, Mauna Kea, and Kilauea). We suggest that this results from widespread contamination of Loihi magmas by a Cl-rich, seawater-derived component. Assimilation of high-Cl phases such as saline brine or Cl-rich minerals (halite or iron–hydroxychlorides)

with high Cl/H2O ratios can explain the range and magnitude of Cl contents in Loihi glasses, as well as the variations in the ratios of Cl to other incompatible elements. Brines and Cl-rich minerals are thought to form from seawater within the hydrothermal systems associated with submarine volcanoes, and Loihi magmas could plausibly have assimilated such materials from the hydrothermal envelope adjacent to the magma chamber. Our model can also explain semiquantitatively the observed contamination of Loihi glasses with atmospheric-derived noble gases, provided the assimilant has concentrations of Ne and Ar comparable to or slightly less than seawater. This is more likely for brines than for Cl-rich minerals, leading us to favor brines as the major assimilant. Cl/Br ratios for a limited number of Loihi samples are also seawater-like, and show no indication of the higher values expected to be associated with the assimilation of Cl-rich hydrothermal minerals. Although Cl enrichment is a common feature of lavas from Loihi, submarine glasses from other Hawaiian volcanoes show little (Kilauea) or no (Mauna Loa, Mauna Kea) evidence of this process, suggesting that assimilation of seawater-derived components is more likely to occur in the early stages of growth of oceanic volcanoes. Summit collapse events such as the one observed at Loihi in October 1996 provide a ready mechanism for depositing brine-bearing rocks from the volcanic ediﬁce into the top of a submarine summit magma chamber. Copyright © 1999 Elsevier Science Ltd

1. INTRODUCTION to have experienced contamination by seawater-derived components (Michael and Schilling, 1989; Jambon et al., 1995; An important issue in the study of ocean island basalts (OIBs) Michael and Cornell, 1998). There are also several lines of is the degree to which chemical heterogeneity within and evidence from OIB, particularly from Hawaii, that suggest that between volcanoes is inﬂuenced by high-level (i.e., lithospheric such components can also be incorporated into magmas in and/or crustal) contamination processes, rather than represen- these environments. Noble gases with atmospheric or near- tative of geochemical variations in mantle source regions (Da- atmospheric isotopic compositions indicate the addition of at- vidson and Bohrson, 1998). Interactions between ascending mospheric components to magmas from Loihi seamount and to magma and the lithospheric mantle, the oceanic crust, and the submarine Kilauea magmas prior to eruption (Patterson et al., volcanic ediﬁce itself can all leave imprints on a magma’s 1990; Honda et al., 1991, 1993). In glasses from a suite of chemical and isotopic composition (e.g., Patterson et al., 1990; submarine Kilauea lavas, Kyser and O’Neil (1984) observed a Martin et al., 1994; Eiler et al., 1996; Garcia et al., 1998a; ␦ correlation between D and H2O, and Chaussidon and Jambon Reiners, 1998). Since the primary goal of many geochemical (1994) observed a correlation between ␦D and ␦11B. Both sets studies of OIB is to investigate the chemical and isotopic of authors suggested that the correlations resulted from assim- compositions of their mantle sources (e.g., Staudacher et al., ilation of seawater or seawater-altered basalt during magma 1986; Zindler and Hart, 1986; Honda et al., 1991; Hoffman, propagation along the East Rift Zone. Clague et al. (1995) 1997), it is important to be able to recognize and, if possible, to noted that Cl concentrations in some submarine glasses from correct for magmatic compositional variations that are due to Kilauea are higher than expected from crystal fractionation and assimilation and contamination. suggested that this could be due to assimilation of seawater- One potential source of contamination for oceanic magmas is altered basalt near the top of the summit magma chamber. sea water, and many mid-ocean ridge basalts (MORB) appear Finally, Kent et al. (1999) suggested that two unusually H2O-, Cl-, and B-rich samples from Loihi represent magmas that assimilated a seawater-derived component; olivine-hosted glass *Address reprint requests to Adam J. R. Kent, L-202, Lawrence Liv- ermore National Laboratory, Livermore, CA 94551 USA (kent5@ inclusions from these samples also show evidence of this llnl.gov). process, and variations in their H2O, Cl, and B contents suggest 2749 2750 A. J. R. Kent et al. variability in the composition of the assimilant. Kent et al. Norman and Garcia (1999). The noble gases in these samples (1999) also speculated that the high Cl contents reported in are of particular interest because their matrix glasses and many other pillow-rim glass samples from Loihi could be another glass inclusions in olivine phenocrysts have high H2O, Cl, and manifestation of contamination by Cl-rich, seawater-derived B concentrations (Kent et al., 1999). As expected (since a components and that assimilation of seawater-derived compo- seawater-derived assimilant would contain atmospheric isoto- nents could be a widespread process at this and other oceanic pic ratios of the noble gases), the 20Ne/22Ne and 40Ar/36Ar volcanoes. In this study, we examine this hypothesis using new ratios of these glasses are close to atmospheric values (Table 1). chemical and isotopic analyses of Loihi glasses and data com- In contrast, the 3He/4He ratios of these samples (as with many piled from the literature. other Loihi samples; e.g., Kurz et al., 1983; Honda et al., 1991, ϳ 1993) are unambiguously of mantle origin ( 27–32 RA for

2. ANALYTICAL TECHNIQUES total fusion and up to 40 RA for He released during individual heating steps; Table 1). Major elements and Cl were analysed by a JEOL 733 electron probe at Lawrence Livermore National Laboratory and at the California Table 2 lists major element, H2O, and Cl analyses of pillow- Institute of Technology usinga5to10nAelectron beam, a 10 ␮m rim glasses from one alkalic and ﬁve tholeiitic and transitional beam size, and an accelerating voltage of 15 kV. Data were reduced basalts collected by submersible from 3 to 5 km depth on with the CITZAF correction algorithms using natural and synthetic stan- Loihi’s south rift. Measured Cl and H O contents for these dards (Armstrong, 1995). Analytical precision for electron probe anal- 2 yses, based on counting statistics for the compositions measured, were glasses range between 0.049 and 0.103 and 0.39 and 0.68 wt.%. Ͻ Ͻ 2% for SiO2,Al2O3, CaO, MgO, FeO; 10% for Na2O, K2O, TiO2; Sample locations and noble gas data for pillow-rim glass and Ͻ Ͻ 20% for Cl and P2O5 and 50% for MnO and Cr2O3. Data represent olivine separates from these samples are given in Valbracht et the average of 5 to 10 individual analyses taken from 3 to 5 separate al. (1997). glass chips from each sample. Glass chips free from alteration products were selected using the procedure outlined in Kent et al. (1999), and Table 3 reports on H2O, Cl, B, and Be contents of 15 secondary standards were monitored to ensure the comparability of pillow-rim glasses (from six alkalic lavas and nine tholeiitic data from the two different electron probes used. Cl analyses were and transitional lavas) dredged from ϳ1 to 3 km depth from made with reference to Cl-rich scapolite and sodalite with count times Loihi’s summit and north and south rift zones. The measured Cl of 90 to 120 s and background counts determined during each analysis. The detection limit for Cl was 70 to 80 ppm. All uncertainties stated in and H2O contents of these glasses range between 0.030 and this paper, unless otherwise noted, are Ϯ2␴. 0.122 and 0.38 and 0.97 wt.% and the Be and B contents range

H2O, Be, and B were analysed using a modiﬁed Cameca 3f ion between 0.7 and 2.0 and 1.2 and 4.9 ppm. Dredge locations for microprobe at Lawrence Livermore National Laboratory following the these samples are given in Moore et al. (1982) and Frey and procedures outlined in Kent et al. (1999). Uncertainties for H2O, Be, and B are based on the reproducibility of 3 to 6 individual analyses on Clague (1983). Noble gas and major element data for these the same glass chips analysed for major elements and are typically glasses were reported by Honda et al. (1993), and Frey and Ϯ Ϯ 15% for H2O and 20% for Be and B. Clague (1983) presented whole-rock major and trace element Analytical techniques for noble gas analyses are described in Honda data for many of the same samples. et al. (1993). Reproducibility of noble gas data was evaluated by analyzing two aliquots for each sample: one heated in several interme- We have also compiled unpublished data and data from the diate temperature steps (600, 1500, and 1600°C for LO-02-02; 700 and literature. Data for pillow-rim glasses from Loihi were com- 1600°C for LO-02-04), and one where all gas was extracted at 1500°C. piled from Moore et al. (1982), Byers et al. (1985), Staudacher Although variations exist between the amount of noble gases extracted et al. (1986), Sarda et al. (1988), Garcia et al. (1989, 1993, from different aliquots, there is generally good agreement among the 1995, 1998b), and Honda et al. (1993). Including the samples He, Ne, and Ar isotopic compositions from each aliquot. Uncertainties for the noble gas analyses include uncertainties in the correction factors studied here, we have major element electron microprobe anal- for mass discrimination and sensitivity determined by repeated analysis yses of pillow-rim glass from 110 Loihi samples (41 alkalic, 69 of the standard gases, as well as uncertainties in the blank correction tholeiitic and transitional). H2O and Cl analyses are available and, in the case of Ne, interference corrections. for 44 of these (12 alkalic, 32 tholeiitic and transitional); and

H2O, Cl, and noble gas data are available for 26 glasses (eight 3. SAMPLES AND RESULTS alkalic, 18 tholeiitic and transitional). These include two sam- Considerable geochemical data on Loihi lavas is available in ples analysed for Ne concentration and isotopic composition by the literature, but few samples have been analyzed for noble gas Staudacher et al. (1986) and one sample analysed for Ar concentration and isotopic composition by Sarda et al. (1988); the isotope ratios and Cl and H2O concentrations, the two classes of data on which the suggestion of a role for contamination by Cl contents of these were reported by Byers et al. (1985). seawater-derived components is usually based (e.g., Patterson Unpublished data are for fragmental glasses collected by piston et al. 1990; Honda et al., 1991; 1993; Michael and Schilling, core and submersible from glass-rich sand layers from the 1989; Michael and Cornell, 1998). We report here new analy- summit and ﬂanks of Loihi (Clague et al., 1997; Clague et al., ses for three sets of samples from Loihi, emphasizing obtaining in press). These glasses, which sample a wide range of magma both classes of data on the same samples. types, represent quench granulation and bubble wall fragments Table 1 reports noble gas concentrations and isotopic ratios produced during submarine volcanic eruptions. We have also for pillow-rim glass for two tholeiitic basalts dredged from ϳ4 compiled major element and Cl analyses of submarine glasses

km depth from Loihi’s south rift zone. Major element, H2O, Cl, from Kilauea (Garcia et al., 1989; Dixon et al., 1990; Clague et Be, and B contents and information on sample location are al., 1991, 1995; Honda et al., 1993); the North Arch volcanic given in Kent et al. (1999). Whole-rock Pb, Sr, and Os isotope ﬁeld (Dixon et al., 1997); and Mauna Kea and Mauna Loa ratios for these samples were reported by Bennett et al. (1996), (Garcia et al., 1989). Degassing of Cl is thought to be insig- and whole-rock major and trace element contents are given in niﬁcant at water depths of 700 m and greater (e.g., Michael and Table 1. Noble gas isotopic compositions of glass separates from two tholeiitic basalts from the south rift of Loihi from the study of Kent et al. (1999) and Norman and Garcia (1999). Following convention, stated errors are Ϯ1␴.

Sample 22 component seawater-derived a of Assimilation Temperature 4He 3He/4He Ne 36 Ar 84 Kr 130 Xe ϫ Ϫ9 ϫ Ϫ6 a ϫ Ϫ12 20 22 21 22 ϫ Ϫ12 38 36 40 36 ϫ Ϫ12 ϫ Ϫ15 (°C) [ 10 ] [ 10 ] R/RA [ 10 ] Ne/ Ne Ne/ Ne [ 10 ] Ar/ Ar Ar/ Ar [ 10 ] [ 10 ]

Loihi LO-02-02 a (glass, 2.0297 g) 600 31.0 Ϯ 0.30 45.6 Ϯ 1.1 32.6 Ϯ 0.8 10.70 Ϯ 0.25 9.99 Ϯ 0.07 0.0294 Ϯ 0.0004 125 Ϯ 4 0.1870 Ϯ 0.0009 306 Ϯ 1 2.74 Ϯ 0.13 18.40 Ϯ 0.77 1500 229.70 Ϯ 2.30 43.9 Ϯ 0.9 31.4 Ϯ 0.6 23.30 Ϯ 0.55 9.99 Ϯ 0.07 0.0309 Ϯ 0.0004 10919 Ϯ 319 0.1876 Ϯ 0.0003 302 Ϯ 1 12.54 Ϯ 0.59 137.39 Ϯ 4.70 1600 2.66 Ϯ 0.03 56.6 Ϯ 2.8 40.4 Ϯ 2.0 N.D.b N.D.b N.D.b 28 Ϯ 1 0.1878 Ϯ 0.0015 559 Ϯ 5 0.04 Ϯ 0.01 0.67 Ϯ 0.17 Total 263.00 Ϯ 2.00 44.3 Ϯ 0.8 31.6 Ϯ 0.6 34.00 Ϯ 0.58 9.99 Ϯ 0.05 0.0304 Ϯ 0.0003 11073 Ϯ 319 0.1875 Ϯ 0.0003 303 Ϯ 1 15.32 Ϯ 0.60 156.46 Ϯ 4.76 Loihi LO-02-02 b (glass, 1.2118 g) 1500 246 Ϯ 14 41.5 Ϯ 2.8 29.7 Ϯ 2.0 23.3 Ϯ 1.1 10.27 Ϯ 0.11 0.0308 Ϯ 0.0004 2728 Ϯ 181 0.1886 Ϯ 0.0005 338 Ϯ 1 7.64 Ϯ 0.22 89.5 Ϯ 6.9 Loihi LO-02-04 a (glass, 2.030 g) 700 23.3 Ϯ 0.24 43.2 Ϯ 1.3 30.8 Ϯ 0.9 25.77 Ϯ 0.63 9.94 Ϯ 0.05 0.0294 Ϯ 0.0004 241 Ϯ 7 0.1872 Ϯ 0.0007 302 Ϯ 1 5.25 Ϯ 0.25 74.38 Ϯ 2.75 1600 181.10 Ϯ 1.85 41.9 Ϯ 0.8 29.9 Ϯ 0.6 24.46 Ϯ 0.55 9.93 Ϯ 0.08 0.0292 Ϯ 0.0004 20410 Ϯ 596 0.1868 Ϯ 0.0003 298 Ϯ 1 9.39 Ϯ 0.45 157.30 Ϯ 5.98 Total 204.40 Ϯ 1.85 42.0 Ϯ 0.7 30.0 Ϯ 0.5 34.00 Ϯ 0.58 9.99 Ϯ 0.05 0.0293 Ϯ 0.0003 20651 Ϯ 596 0.1868 Ϯ 0.0003 298 Ϯ 1 14.64 Ϯ 0.51 231.68 Ϯ 6.58 Loihi LO-02-04 b (glass, 0.9842 g) 1500 164 Ϯ 10 38.5 Ϯ 2.6 27.5 Ϯ 1.9 18.8 Ϯ 0.9 10.00 Ϯ 0.14 0.0303 Ϯ 0.0002 10143 Ϯ 674 0.1884 Ϯ 0.0005 304 Ϯ 1 9.04 Ϯ 0.27 103.76 Ϯ 8.3 Air 1.4 1.0 9.8 0.029 0.1880 295.5

a 3 4 ϫ Ϫ6 RA , the He/ He ratio of the atmosphere, is 1.4 10 . b N.D. ϭ below blank. * Noble gas amounts are in cm3 STP/g. 2751 2752 A. J. R. Kent et al.

Table 2. Major element, Cl, and H2O contents of pillow-rim glasses from samples from the study of Valbracht et al. (1997) from the south rift zone of Loihi.

a Sample SiO2 TiO2 Al2O3 Cr2O3 FeOT MnO MgO CaO Na2O K2O P2O5 Cl H2O Total

2341-2 ac 47.28 3.14 14.54 0.01 11.85 0.18 5.19 10.86 3.46 1.07 0.45 0.098 0.61 98.74 2341-1 trd 48.66 3.18 13.81 0.01 12.00 0.17 5.41 10.74 3.01 0.72 0.38 0.103 0.60 98.80 2335-6 (i) trd 48.39 2.96 13.96 0.03 10.67 0.16 6.27 11.49 2.92 0.74 0.37 0.097 0.59 98.65 2335-10 trd 48.46 3.07 14.53 ndb 10.53 ndb 6.72 12.28 2.78 0.98 0.39 0.075 0.68 100.68 2338-2 te 48.79 2.79 13.12 0.04 11.29 0.17 6.97 11.74 2.54 0.55 0.32 0.078 0.39 98.79 2337-3b te 49.59 2.76 13.52 0.02 11.45 0.17 6.17 11.08 2.56 0.50 0.31 0.049 0.47 98.65

* All analyses reported in weight percent and are from glass chips provided by P. Valbracht and A. Malahoff. a Fe expressed as total FeO. b nd ϭ not determined. c a—alkali basalt. d tr—transitional basalt. e t—tholeiitic basalt.

Schilling, 1989), so we have focused our compilation on sam- rim and fragmental glasses from Loihi also have high Cl ples from depths greater than this. contents (up to 0.13 wt.%), however the overall range of Cl Note that throughout we refer to samples as either “alkalic” concentrations in Loihi alkalic glasses is similar to that of (including alkali basalts, basanites, and hawaiites) or “tho- submarine alkalic glasses from the North Arch ﬁeld (Fig. 1; leiitic and transitional” (all basalts), following the classiﬁ- Dixon et al., 1997). cations of Macdonald and Katsura (1964) based on total The large range of Cl contents evident in Loihi tholeiitic and alkalis vs. silica. transitional glasses, including the anomalously high values evident in many samples, are not matched by concomitant 4. DISCUSSION variations in the concentrations of other incompatible elements 4.1. Cl in Loihi Glasses (e.g., K, P, H, Be, and B), suggesting that Cl has been de- coupled from these elements in tholeiitic and transitional mag- A distinctive feature of many tholeiitic and transitional mas. This is evident both in large variations in the ratios of Cl glasses from Loihi is that they have higher Cl concentrations to other incompatible components (as shown for Cl/K O in Fig. than submarine glasses from other Hawaiian volcanoes. For 2 2) compared to ratios that do not involve Cl (e.g., P2O5/K2O example, numerous tholeiitic and transitional pillow-rim and ratios for all Loihi glasses vary by a factor of ϳ2.5, whereas fragmental glasses from Loihi have Cl concentrations in excess Cl/K O and Cl/P O ratios vary by factors of ϳ6–7) and in of 0.10 wt.% (up to a maximum of 0.18 wt.%), compared to a 2 2 5 poor correlations between Cl and other incompatible compo- range of Ͻ0.01 to 0.06 wt.% Cl in submarine glasses from nents for tholeiitic and transitional samples (compared to gen- Kilauea, Mauna Loa, and Mauna Kea (Fig. 1). Alkalic pillow- erally good correlations between incompatible components other than Cl; Fig. 3). Decoupling between Cl and other incompatible components Table 3. Cl, H2O, Be, B contents of Loihi pillow-rim glasses for which major element and noble gas compositions were analyzed by is less evident in alkalic glasses from Loihi; these have Cl/K2O Honda et al. (1993). ratios that are less elevated and variable than tholeiitic and transitional glasses (Fig. 2), and Cl is reasonably well corre- Cl H2O Be B Sample (wt.%) (wt.%) (ppm) (ppm) lated with other incompatible components in alkalic samples (Fig. 3). However, in comparison to alkalic glasses from the KK 15-4 ba 0.122 0.97 2.0 4.9 North Arch, alkalic glasses from Loihi have higher maximum KK 17-11 ab 0.070 0.74 1.2 2.6 b Cl/K2O ratios (0.15 compared to 0.12) and a larger range in KK 21-3 a 0.043 0.63 0.8 1.5 ϳ ϳ KK 29-13 ab 0.063 0.47 0.8 1.7 Cl/K2O (a factor of 4 compared to a factor of 2). KK 30-6 ab 0.066 0.60 1.1 2.2 The enrichment in Cl that we describe in tholeiitic and KK 31-9 ab 0.039 0.58 0.8 1.6 transitional lavas appears to be a widespread feature of Loihi c KK 20-13 tr 0.036 0.57 0.9 1.8 magmas. High Cl and Cl/K O ratios are evident in samples HW 2-16 td 0.029 0.41 0.7 1.3 2 KK 20-2 td 0.062 0.44 0.8 1.3 taken from throughout the summit region, as well as from the KK 22-2 td 0.042 0.57 0.8 1.6 southern and northern rift zones of the volcano. In addition, KK 22-20 td 0.030 0.47 0.8 1.5 high Cl and Cl/K O are common in fragmental glasses taken d 2 KK 22-4 t 0.041 0.56 0.9 1.7 from sand layers (Figs. 1 and 2), which, because of their large KK 23-9 td 0.040 0.42 0.8 1.5 KK 26-4 td 0.093 0.49 0.9 1.7 variations in major element composition, are thought to sample KK 29-4 td 0.039 0.38 0.7 1.2 many eruptions (Clague et al., in press). This suggests that high Cl and high Cl/K O (and ratios of Cl to other incompatible a 2 b—basanite. elements) in Loihi lavas, and whatever process or processes b a—alkali basalt. c tr—transitional basalt. produce them, are widespread features of magmatism at Loihi d t—tholeiitic basalt. seamount. Assimilation of a seawater-derived component 2753

Fig. 1. Variation of Cl with MgO for Loihi pillow-rim and fragmental glasses. Symbols are explained in the legend. Ϯ ␴ H2O-rich samples studied by Kent et al. (1999) are marked by small horizontal arrows. Representative 2 error bars are given at top right. Dashed lines show model Cl concentrations of a representative parental tholeiite (from Garcia et al., 1995 and Kent et al., 1999 and shown as a filled star) during crystal fractionation. The model is based on fractionation of olivine (by removal of increments of olivine in Fe-Mg equilibrium) between 16 and 8 wt.% and 75% clinopyroxene (using the composition “CPX4” from by Garcia et al., 1995), 20% anorthite (An70) and 5% olivine (Fo86) between 8% and 4% MgO. (A) Pillow rim glasses. Data for Loihi samples are from Tables 2 and 3 and Byers et al. (1985), Garcia et al. (1989, 1998b), and Kent et al. (1999). Fields for Mauna Loa and Mauna Kea pillow-rim glasses (shown as a single field), for glasses from the North Arch volcanic field, and for Kilauea pillow-rim glasses are from Byers et al. (1985), Garcia et al. (1989, 1998b), Honda et al. (1993), Clague et al. (1995), and Dixon et al. (1997). (B) Comparison between pillow rim glasses from (A) and fields for fragmental glasses from Loihi and Kilauea. Data for Loihi fragmental glasses from D. A. Clague and A. Davis (unpublished data); data for Kilauea fragmental glasses from Clague et al. (1995). 2754 A. J. R. Kent et al.

Fig. 2. Cl/K2O (wt.%) vs. MgO for Loihi pillow-rim and fragmental glasses. Dashed line shows Cl/K2O (which remains constant) during crystal fractionation of a representative parental Loihi tholeiite (Garcia et al., 1995; Kent et al., 1999). Symbols and data sources as for Fig. 1. Representative Ϯ2␴ error bars are given at top right. (A) Pillow-rim glasses; (B) Comparison between pillow-rim glasses from (A) and ﬁelds for fragmental glasses. Assimilation of a seawater-derived component 2755

Fig. 3. Variations of concentrations of selected incompatible elements and oxides in Loihi pillow-rim glasses. Symbols Ϯ ␴ and data sources as for Fig. 1, with additional data for K2O and P2O5 from Garcia et al. (1993, 1995). Representative 2 error bars are given at bottom right in each panel. (A) K2O vs. P2O5; (B) Cl vs. P2O5; (C) K2O vs. H2O; (D) Cl vs. H2O; (E) Be vs. B; (F) Cl vs. B. 2756 A. J. R. Kent et al.

4.1.1. Origin of Cl enrichment in Loihi glasses

High concentrations of Cl in basaltic lavas can be the result of magmatic processes such as low-degree partial melting and/or crystal fractionation. However, these processes also enrich the concentrations of other incompatible elements, and thus cannot explain the observed decoupling of Cl from other incompatible components at Loihi. As shown in Fig. 1, crystal fractionation of olivine, clinopyroxene, and plagioclase from a representative tholeiitic parental melt is incapable of producing the highest observed Cl contents of tholeiitic and transitional lavas (Fig. 1). Byers et al. (1985) and Garcia et al. (1989) suggested that Cl enrichment is a primary feature of the mantle sources of Loihi magmas. However, this would require an unusual kind of enrichment process affecting only Cl in order to explain the lack of correlation between Cl and other incompatible elements that typically accompany it in “enriched” sources (e.g., Schilling et al., 1980). Lavas from Loihi do not appear to have degassed significant water, S, or halogens (Dixon and Clague, 1997), and Dixon et al. (1990) suggested that Loihi lavas are not enriched in Cl, but rather that subaerial volcanoes, such as Kilauea and Mauna Loa, have been depleted in Cl by degassing of Cl from their summit magma reservoirs. However, if this were the case, then Fig. 4. Cl/K2O vs. H2O/K2O for pillow-rim glasses from Loihi. Data Cl would behave in Loihi glasses like an involatile incompat- sources and symbols as for Fig. 1. Fields show data for pillow-rim glasses from Mauna Loa and Mauna Kea (as a single field), Kilauea, ible element, and the observed lack of correlation between Cl and the North Arch volcanic field. Lines show mixing trends between and other incompatible components such as K2O and P2O5 a representative parental Loihi tholeiite (from Garcia et al., 1995 and (Figs. 2 and 3) could not be readily explained. Kent et al., 1999) calculated to MgO ϭ 8 wt.% (shown as a hollow star Submarine alteration and weathering is also known to in- and containing 0.44 wt.% H2O, 0.47 wt.% K2O, 0.013 wt.% Cl) and either altered basalt (containing 5 wt.% H2O, 1 wt.% K2O, 0.1 wt.% crease the concentrations of components such as K2O, H2O, Cl, Cl), seawater, 15 wt.% and 50 wt.% NaCl brine, or halite. Composi- and B in altered basaltic glasses (e.g., Staudigel and Hart, 1984; tions of the assimilant endmembers are detailed in Kent et al. (1999). Smith et al., 1995; Staudigel et al., 1996), and if alteration Numbers on mixing lines show percentages of Cl-rich components preferentially enriched Cl relative to other incompatible com- added to basalt in wt.%. Representative Ϯ2␴ error bars are given at top right. ponents (including K2O, P2O5, and H2O), this could explain the observed Cl enrichments. However, given that all glasses in this study were selected to be free of visible alteration phases, seawater (e.g., Bischoff and Rosenbauer, 1987), altered basalt, that the Cl contents of each glass were homogeneous to within and Cl-rich minerals such as halite and iron–hydroxychlorides

analytical uncertainty, and that there is an absence of H2O (e.g., Seyfreid et al., 1986; Berndt et al., 1988; Berndt and enrichment in nearly all samples, it is unlikely that the high Cl Seyfried, 1997) that may form within submarine hydrothermal concentrations in Loihi glasses directly reflect submarine systems. Highly saline brines (e.g., up to ϳ50 wt.% NaCl) and weathering. Moreover, high Cl contents and Cl/K2O ratios have iron–hydroxychloride and halite are thought to exist in the also been observed for melt inclusions within fresh olivine submarine hydrothermal systems associated with MORB (Sey- phenocrysts in some Loihi samples (Kent et al., 1999). freid et al., 1986; Kelley and Delaney, 1987; Bischoff and Our preferred explanation for the enrichment of Cl in Loihi Rosenbauer, 1987; Berndt et al., 1988; Berndt and Seyfreid, glasses is that Loihi magmas frequently assimilate seawater- 1990; 1997; Butterfield et al., 1997) and have also been pro- derived, Cl-rich components. Assimilation of Cl-rich seawater- posed to occur in hydrothermal systems associated with oce- derived components can explain the elevated Cl contents and anic volcanoes such as Loihi (e.g., Fournier, 1987). Loihi’s the large range in Cl concentrations and Cl/K2O ratios charac- summit is almost 1000 m below sea level (Moore et al., 1982), teristic of the Loihi data set, as well as the lack of correlation and hydrothermal systems in the volcanic edifice are dominated between Cl and elements such as K and P not enriched in by circulation of seawater (Fournier, 1987; Karl et al., 1988; seawater. This idea is not new: Clague et al. (1995) explained Davis and Clague, 1998). Interaction between magmas and elevated Cl contents in submarine Kilauea lavas in the same either seawater-derived fluids and/or altered rocks from the way. Michael and Schilling (1989), Jambon et al. (1995), and hydrothermal system could introduce a Cl-rich component into Michael and Cornell (1998) proposed that the same process the magma. occurs in mid-ocean ridge environments. For Loihi, Kent et al. The potential assimilants can be distinguished by their Cl/

(1999) proposed the same hypothesis based on a detailed study H2O ratios and absolute Cl concentrations. Figure 4 compares of two samples (including glass inclusions in olivine pheno- Cl/K2O and H2O/K2O ratios of Loihi pillow-rim glasses with crysts). fields for submarine glasses from Kilauea, Mauna Loa, Mauna Potential assimilants include not only seawater, but also the Kea, and the North Arch. Also shown are mixing lines between ϭ saline brines and complementary H2O-rich fluids derived from a representative Loihi tholeiite with MgO 8 wt.% and five Assimilation of a seawater-derived component 2757 potential assimilants: sea water, 15 and 50 wt.% NaCl brines, (Clague et al., 1995). However, this is not a universal feature of altered basalt, and halite. With the exception of the two unusu- Kilauea samples (for example numerous pillow-rim glasses and ally water-rich samples described by Kent et al. (1999), pillow- high-MgO fragmental glasses from Kilauea show no evidence rim glasses from Loihi form a near vertical array, with little or of Cl enrichment; Dixon et al., 1990; Clague et al., 1991, 1995) no variation in H2O/K2O with increasing Cl/K2O; i.e., the Loihi and it is less pronounced than in Loihi samples when it does data are not consistent with the addition of assimilants with low occur (Fig. 2).

Cl/H2O ratios such as seawater, 15% NaCl brine, a comple- The small amounts of assimilation of brines or Cl-rich min- Ͻ mentary H2O-rich ﬂuid formed by phase separation (not shown erals ( 1%) required to produce the Cl contents and Cl/K2O in Fig. 4), or altered basalt. Also, even if the absence of H2O ratios in tholeiitic, transitional, and alkaline glasses from Loihi enrichment was disregarded, bulk assimilation of altered basalt (Fig. 4), suggests that this process would be unlikely to affect is unlikely to explain the high Cl contents of Loihi glasses the concentrations and isotopic compositions of most petrolog- because of the large quantities of this material that would be ically important lithophile elements (e.g., rare-earth element, required to elevate Cl contents and the Cl/K2O ratios (Fig. 4). Sr, Pb). Tests of the assimilation hypothesis must instead focus Based on Fig. 4, both saline brines and Cl-rich minerals such on elements and/or isotopes that are signiﬁcantly enriched in

as halite have high enough Cl contents and Cl/H2O and Cl/K2O seawater-derived materials, such as B, Cl, and H (although, as ratios to qualify for the required assimilant. The total range of emphasized above, relatively few Loihi samples are enriched in

Cl/K2O ratios of tholeiitic and transitional basalts can be ex- H2O and B). Another possibility is the isotopic ratios of the plained by the addition of small amounts of these materials heavy noble gases (Ne, Ar, Kr, Xe), which are enriched in (e.g., Յ ϳ0.4 wt.% of 50 wt.% NaCl brine or halite, or Ͻ1 concentration in seawater by several orders of magnitude rel- wt.% of Fe2[OH]3Cl; Rucklidge and Patterson, 1977). Assim- ative to Loihi glasses (Patterson et al., 1990; Valbracht et al., ilation could result from direct incorporation of these phases 1997). In Sec. 4.2. we explore the consistency of noble gas into the magma or from assimilation of small amounts of rock isotopic data from Loihi samples with our hypothesis that there containing these phases. Note that except for the two samples has been widespread contamination of Loihi magmas by sea-

studied by Kent et al. (1999), H2O and B contents beyond those water-derived components. expected from crystal fractionation of parental melts have not been observed in other Loihi glasses; the assimilants required 4.2. Contamination of Noble Gases to explain these two samples would have to be anomalously

H2O- and B-rich, although still with Cl/H2O greater than that of Glasses and phenocrysts from Loihi have been contaminated seawater, suggesting a variability in the nature of the Cl en- by atmosphere-derived noble gases: i.e., the Ne and Ar isotopic richment process operating at Loihi. ratios of such samples range between atmospheric and mantle- Assimilation of a Cl-rich component can also explain why derived values (Patterson et al., 1990; Honda et al., 1991; 1993;

Cl/K2O ratios (and the ratios of Cl to other incompatible Valbracht et al., 1997). Although much of the study of noble elements) range to higher values and are more variable in gases from Loihi has been aimed at deﬁning the nonatmo- tholeiitic and transitional glasses than in alkalic glasses from spheric mantle-derived component (e.g., Staudacher et al., Loihi (Figs. 2 and 3). Alkalic magmas have higher primary 1986; Sarda et al., 1988; Honda et al., 1991; 1993; Valbracht et concentrations of Cl (and all other incompatible elements) than al., 1997), our interest is in whether the atmospheric noble gas tholeiitic and transitional magmas, and thus their Cl/K2O ratios signature in samples from Loihi is correlated with Cl enrich- will be less affected by the addition of Cl. For example, the ment, as would be expected if both were introduced during addition of 0.4 wt.% of a 50 wt.% NaCl brine (the estimated assimilation of seawater-derived components.

amount of assimilant required to produce the highest Cl/K2O In Fig. 5 we compare the Ne and Ar isotopic compositions ratios observed in tholeiitic and transitional magmas) to an and Cl/K2O ratios of the 26 pillow-rim glasses from Loihi for alkalic magma with preassimilation K2O and Cl concentrations which both classes of data are available. We have also plotted of 1.0 and 0.03 wt.% would produce a Cl/K2O ratio of 0.16, (shown as a field for clarity) data for samples from Kilauea compared to 0.30 in a tholeiitic magma with preassimilation from Honda et al. (1993). Although many Loihi glasses have 20 22 K2O and Cl contents of 0.50 and 0.015 wt.% (Kent et al., relatively low Cl/K2O and low, roughly atmospheric Ne/ Ne 40 36 1999). The maximum Cl/K2O ratios observed in alkalic glasses and Ar/ Ar on Fig. 5, the data extend both vertically (to high ϳ are 0.15, and are thus consistent with the addition of similar Cl/K2O at low noble gas isotopic ratios) and horizontally (to amounts of a 50 wt.% NaCl brine as required for tholeiitic and high noble gas isotopic ratios at low Cl/K2O). transitional glasses. Assimilation of such components by alka- Figure 5 shows the results of three simple mixing calcula- lic magmas is also consistent with indications that alkaline tions. All three calculations take a representative tholeiitic lavas contain atmosphere-derived noble gases (see below). Loihi basalt with MgO ϭ 8 wt.% (calculated from a represen- If our hypothesis is correct, assimilation of seawater-derived tative parental tholeiite composition taken from Garcia et al., Cl-rich components must be a widespread process at Loihi. In 1995; Kent et al., 1999) with mantle-like Ne and Ar isotopic 20 22 ϭ 40 36 ϭ contrast, the low Cl/K2O ratios of submarine glasses from compositions (solar Ne [ Ne/ Ne 13.8] and Ar/ Ar Mauna Loa and Mauna Kea (Fig. 2) suggest that lavas currently 7000) and Ne and Ar concentrations estimated for Loihi tho- exposed on the flanks of these volcanoes have not been signif- leiitic basalt as one end member. Details of the end members icantly affected by this process. Kilauea may be an intermediate used are given in the caption for Fig. 5. Three different Cl- and case, as there is some variation of the Cl/K2O ratio in pillow- rare-gas-rich end members were considered: (i) sea water rim glasses from the submarine portion of the East Rift Zone (model 1); (ii) a 50 wt.% NaCl fluid with seawater Ne and Ar (Fig. 2), consistent with assimilation of seawater-altered basalt concentrations (model 2); and (iii) a 50 wt.% NaCl fluid with an 2758 A. J. R. Kent et al.

22 20 40 36 Fig. 5. Cl/K2O (wt. %) vs. Ne/ Ne and Ar/ Ar for pillow-rim glasses from Loihi and Kilauea. Symbols for Loihi glasses as for Fig. 1. Open symbols show data points where noble gas and Cl/K2O ratios are taken from published data. Shaded ﬁeld shows data for samples from the East Rift Zone of Kilauea (from Honda et al., 1993 and Dixon et al., 1990). 22 20 40 36 Ϯ ␴ For clarity, individual error bars for Cl/K2O have been omitted and errors shown for Ne/ Ne and Ar/ Ar are 2 . Data sources as for Fig. 1 and Staudacher et al. (1986), Sarda et al. (1988), and Valbracht et al. (1997). Mixing curves are based on mixing a representative parental tholeiite (marked by hollow star; Garcia et al., 1995; Kent et al., 1999) calculated to MgO ϭ 8 wt.% with 22Ne and 36Ar concentrations of 5 ϫ 10Ϫ12 cc/g and 5 ϫ 10Ϫ10 cc/g (Valbracht et al., 1997), 22Ne/20Ne ϭ 13.8 and 40Ar/36Ar ϭ 7000 (Honda et al., 1993), and the following assimilants: Model 1, sea water (1.94 wt.% ϫ Ϫ7 20 Cl, 0.04 wt.% “K2O”; for sake of comparison, K in sea water and brines is calculated here as K2O), 1.74 10 cc/g Ne, 1.27 ϫ 10Ϫ6 cc/g 36Ar (Holland, 1984; Staudacher et al. 1986); Model 2, 50 wt.% NaCl brine (30.3 wt.% Cl, 0.78 wt.%

“K2O”; based on estimate given in Kent et al., 1999) with the same seawater Ne and Ar concentrations as model 1; Model 3, 50 wt.% NaCl brine with an order of magnitude less 20Ne and 36Ar than sea water (1.74 ϫ 10Ϫ6 cc/g 20Ne, 1.27 ϫ 10Ϫ5 cc/g 36Ar). Atmospheric 20Ne/22Ne (9.8) and 40Ar/36Ar (295.5) ratios were used for all the assimilant endmembers. Composition of assimilant end members are summarized in the associated table. Numbers on mixing trends show proportion of assimilant added to parental tholeiite in wt.%. Where several noble gas fractions have been extracted from a single aliquot, only the maximum Ne and Ar isotopic ratios are shown, in order to reduce the effects of adsorbed atmospheric gas; however, the conclusions of this study would be unchanged if we were to use the 20Ne/22Ne and 40Ar/36Ar of the total released gas. Where noble gases have been analysed in duplicate aliquots from the same sample, the noble gas composition

of each aliquot is shown as a separate point using the same measured Cl/K2O ratio. Note that the Cl/K2O ratios of tholeiitic–transitional magmas are more sensitive to the addition of a Cl-rich component than those of alkalic magmas, and although this does not affect the concave up nature of the mixing trajectories, in detail the curvatures and the mixing

proportions shown are only correct for tholeiitic (and transitional) samples with the assumed Cl and K2O contents. (A) 20 22 40 36 Cl/K2O vs. Ne/ Ne; (B) Cl/K2O vs. Ar/ Ar. order of magnitude less 20Ne and 36Ar than sea water (model Ozima and Podosek, 1983; Kennedy, 1988), although remixing 3). The assimilants in all three calculations have atmospheric between brine and seawater could result in fluids with high Cl Ne and Ar isotopic ratios. Both increased salinity and decreased and only moderately depleted noble gas concentrations. The noble gas concentrations result in more gently curved mixing noble gas concentrations of Cl-rich minerals from submarine trends in Fig. 5. The Cl contents of the Cl-rich assimilant in hydrothermal systems have not been measured, but are likely to models 2 and 3 are based on a 50 wt.% NaCl brine but could be lower than seawater (although we note that Cl-rich minerals also represent assimilation of highly Cl-enriched rocks and could conceivably contain noble gases adsorbed onto mineral minerals. The lower noble gas abundance of model 3 is an surfaces and/or trapped within fluid inclusions). attempt to illustrate the effects of the expected decrease in The Loihi data define overall concave upward fields in Fig. noble gas solubility in a high-temperature brine (noble gas 5. This requires that the Ne and Ar concentrations in the solubility in aqueous fluids decreases with increasing temper- high-Cl endmember are several orders of magnitude greater ature and salinity, and formation of a brine by phase separation than in the mantle endmember. This argues against a rare-gas- is expected to partition noble gases out of the brine phase; poor phase (as we expect Cl-rich minerals to be) as the source Assimilation of a seawater-derived component 2759 of high Cl and atmospheric noble gases in Loihi magmas. The are dominated by seawater. Planned deep drilling into Mauna Loihi data are bounded in Fig. 5 by models 1 and 3, suggesting Kea (Stolper et al., 1996) may allow for more detailed inves- that assimilation of Cl-enriched brines can account for atmo- tigation of the role of assimilation in the early stages of volca- spheric contamination of rare-gas isotopic ratios provided the nic development. brines have variable rare-gas concentrations that are similar to or somewhat lower than those of seawater. Some Kilauea 5. CONCLUSIONS submarine glasses also display Cl enrichment, although to a smaller extent than Loihi samples, and the largely horizontal Many pillow-rim and fragmental glasses from Loihi sea- trends of the Kilauea tholeiites in Fig. 5 are also consistent with mount have high Cl and high ratios of Cl to K (and other assimilation of a noble gas-bearing and seawater-derived com- incompatible elements) relative to submarine glasses from ponent, but to a consistently smaller degree than most Loihi other Hawaiian volcanoes (Mauna Loa, Mauna Kea, and samples, as inferred above based on their Cl/K2O ratios. Kilauea). We propose that this reflects widespread assimilation Our interpretation that seawater-derived brines are the most of a seawater-derived, Cl-rich component at this volcano. likely assimilant is also supported by Cl/Br ratios measured in Assimilation of seawater-derived Cl-rich phases such as sa- a limited number of pillow-rim glasses from Loihi (Jambon, line brine or Cl-rich minerals (halite or iron–hydroxychlorides)

1994; A. Jambon, personal communication, 1999). These with high Cl/H2O ratios can explain the observed Cl contents glasses (with Cl contents of 0.03–0.06 wt.%; i.e., at the low end and large variations in the ratios of Cl to other incompatible of the range of samples considered here) have uniform Cl/Br elements in Loihi lavas. Brines and Cl-rich minerals are ratios (ϳ360) close to that of seawater (290) (and also to thought to form from seawater within the hydrothermal systems average MORB, 400 Ϯ 100; Jambon, 1994). Assimilation of associated with submarine volcanoes, so such phases are plau- halite or other Cl-rich mineral, even in small amounts, would sibly available for assimilation in the magmatic systems of be expected to substantially elevate Cl/Br (Br does not partition submarine volcanoes such as Loihi. into the halite structure; Berndt and Seyfried, 1997, and the Assimilation of seawater-derived, Cl-rich components can same is also probably true for other Cl-rich minerals such as also explain the contamination of Loihi lavas with atmospheric- iron–hydroxychlorides; Berndt and Seyfried, 1997). In con- derived noble gases if the assimilant had concentrations of Ne trast, formation of brines from sea water does not substantially and Ar comparable to, or slightly less than, seawater. This is fractionate the Cl/Br ratio (Berndt and Seyfried, 1997). more likely for brines than for Cl-rich minerals, leading us to favor brines as the major assimilant. In addition, seawater-like 4.3. Assimilation During Summit Collapse Br/Cl ratios in several Loihi pillow-rim glasses also favor seawater-derived brines as the primary assimilant as halite and In October 1996, a ϳ600 m diameter section of the summit other Cl-rich minerals strongly fractionate this ratio. of Loihi collapsed to form a crater about 300-m deep (Duen- Summit collapse events such as the one observed at Loihi in nebier, 1996; Davis and Clague, 1998; Garcia et al., 1998b). October 1996, provide a mechanism for depositing brine-bear- Summit collapse events such as this are thought to occur after ing rocks and/or Cl-rich hydrothermal minerals from the vol- removal of magma by eruption leaves the roof of the summit canic edifice into the top of a summit magma chamber. magma chamber unsupported (e.g., Decker, 1987). At Loihi, the collapsed region included several hydrothermally active Acknowledgments—The authors thank P. Valbracht and A. Malahoff areas (including the area formerly known as Pele’s vents), and for samples and P. Carpenter and D. Phinney for help with analyses. the collapse event changed both the geometry of fluid circula- Comments by A. Jambon and M. Palin and reviews by M. Garcia and tion in the existing summit hydrothermal system and the chem- P. Michael improved the quality of this contribution. We also thank istry of fluids released at summit hydrothermal vents (Duenne- Professor A. Jambon for sharing unpublished data. Supported by NSF grant EAR95-28594 and DOE grant DE-FG03-85ER13445. Performed bier et al., 1996; Davis and Clague, 1998). Collapse events like under the auspices of the DOE by LLNL under contract W-7405-Eng- this could be linked to assimilation of Cl-rich components as 48. Division of Geological and Planetary Science contribution 5708. they provide a means for brines and rocks (the latter potentially brine-bearing) from deep within the summit hydrothermal sys- REFERENCES tems to collapse into the top of the magma chamber. It is less likely that fluids from the hydrothermal system could enter the Armstrong J. T. (1995) CITZAF: a package of correction programs for magma chamber more directly, as semiductile flow in heated the quantitative electron microbeam analysis of thick polished materials, thin films and particles. Microbeam Anal. 4, 177–200. wallrocks around the magma chamber would probably “armor” Bennett V. C., Esat T. M., and Norman M. N. (1996) Two mantle- the magma chamber against the direct entry of fluids (Fournier, plume components in Hawaiian picrites inferred from correlated 1987). Os–Pb isotopes. Nature 381, 221–224. If summit collapse events are associated with assimilation, Berndt M. E. and Seyfreid W. E. J. (1990) Boron, bromine, and other then similar processes may operate in other Hawaiian volca- trace elements as clues to the fate of chlorine in mid-ocean ridges. Geochim. Cosmochim. Acta 54, 2235–2245. noes, where collapse craters are also common (Decker, 1987). Berndt M. E. and Seyfried W. E. (1997) Calibration of Br/Cl fraction- However, our observation that high Cl contents and Cl/K2O ation during subcritical phase-separation of seawater: possible halite ratios are largely limited to glasses from Loihi may indicate at 9 to 10°N East Pacific Rise. Geochim. Cosmochim. Acta 61, that assimilation of Cl-rich, seawater-derived components is an 2849–2854. Berndt M. E., Seyfreid W. E., and Janecky D. R. (1988) Trace element important process only in the early stages of the development evidence for formation of a Cl-rich mineral in mid-ocean ridge of oceanic volcanoes, when the summit and/or summit magma hydrothermal systems. Eos 69, 464. chamber is still submerged and summit hydrothermal systems Bischoff J. L. and Rosenbauer R. J. (1987) Phase separation in seafloor 2760 A. J. R. Kent et al.

geothermal systems: An experimental study of the effects on metal D. A. (1991) Possible solar noble gas component in Hawaiian transport. Am. J. Sci. 287, 953–978. basalts. Nature 349, 149–151. Butterfield D. A., Jonasson I. R., Massoth G. J., Feely R. A., Roe K. K., Honda M., MacDougall I., Patterson D. B., Doulgeris A., and Clague Embley R. E., Holden J. F., McDuff R. E., Lilley M. D., and Delaney D. A. (1993) Noble gases in submarine pillow basalt glasses from J. R. (1997) Sea-floor eruptions and the evolution of hydrothermal Loihi and Kilauea, Hawaii: A solar component in the Earth. fluid chemistry. Phil. Trans. Royal Soc. London A 355, 369–386. Geochim. Cosmochim. Acta 57, 859–874. Byers C. D., Garcia M. O., and Muenow D. W. (1985) Volatiles in Jambon A. (1994) Earth degassing and large-scale geochemical cycling pillow rim glasses from Loihi and Kilauea volcanoes, Hawaii. of volatile elements. In Volatiles in Magmas (eds. M. R. Carroll and Geochim. Cosmochim. Acta 49, 1887–1896. J. R. Holloway); Min. Soc. Am.—Rev. Mineral. 30, 479–517. Chaussidon M. and Jambon A. (1994) Boron content and isotopic Jambon A., De´ruelle B., Dreibus G., and Pineau F. (1995) Chlorine and composition of oceanic basalts: Geochemical and cosmochemical bromine abundance in MORB: The contrasting behaviour of the Mid implications. Earth Planet. Sci. Lett. 121, 277–291. Atlantic Ridge and the East Pacific Rise and implications for the Clague D. A., Davis A., and Dixon J. E. (1997) Formation of bubble- chlorine geodynamic cycle. Chem. Geol. 126, 101–117. wall fragments during submarine eruptions on Loihi seamount and Karl D. M., McMurtry G. M., Malahoff A., and Garcia M. O. (1988) Kilauea volcano. Eos 78, 792. Loihi seamount, Hawaii: A mid-plate volcano with a distinctive Clague D. A., Weber W. S., and Dixon J. E. (1991) Picritic glasses hydrothermal system. Nature 335, 532–535. from Hawaii. Nature 353, 553–556. Kelley D. S. and Delaney J. R. (1987) 2-Phase separation and fractur- Clague D. A., Moore J. G., Dixon J. E., and Friesen W. B. (1995) ing in Midocean Ridge gabbros at temperatures greater than 700°C. Petrology of submarine lavas from Kilauea’s Puna Ridge, Hawaii. J. Earth Planet. Sci. Lett. 83, 53–66. Petrol. 36, 299–349. Kennedy B. M. (1988) Noble gases in vent water from the Juan de Fuca Clague D. A., Bischoff J., Davis A., Dixon J. E., and Guyer R. (1999) Ridge. Geochim. Cosmochim. Acta 52, 1929–1935. Bubble-wall fragments formed by submarine steam explosions on Kent A. J. R., Norman M. D., Hutcheon I. D., and Stolper E. M. (1999) Loihi Seamount and Kilauea Volcano: Implications for phreatomag- Sea water assimilation in an oceanic volcano: Evidence from matrix matic eruptions. Bull. Volcan. (in press). glasses and glass inclusions from Loihi seamount, Hawaii. Chem. Davidson J. P. and Bohrson W. A. (1998) Shallow-level processes in Geol. 156, 299–319. ocean-island magmatism. J. Petrol. 39, 799–801. Kurz M. D., Jenkins W. J., Hart S. R., and Clague D. (1983) Helium Davis A. S. and Clague D. A. (1998) Changes in the hydrothermal isotopic variations in the volcanic rocks from Loihi seamount and the system at Loihi seamount after the formation of Pele’s pit in 1996. island of Hawaii. Earth Planet. Sci. Lett. 66, 388–406. Geology 26, 399–402. Kyser K. T. and O’Neil J. R. (1984) Hydrogen isotope systematics of Decker R. W. (1987) Dynamics of Hawaiian volcanoes: An overview. submarine basalts. Geochim. Cosmochim. Acta 48, 2123–2133. USGS Prof. Pap. 1350, 997–1018. Martin C. E., Carlson R. W., Shirey S. B., Frey F. A., and Chen C. Y. Dixon J. E. and Clague D. A. (1997) Evolving volcanoes and degassing (1994) Os-isotopic variation in basalts from Haleakala volcano, styles in Hawaii. GSA Ann. Meet. A-191 (abstr.). Maui: A record of magmatic processes in oceanic mantle and crust. Dixon J. E., Clague D. A., and Stolper E. M. (1990) Degassing history Earth Planet. Sci. Lett. 128, 287–301. of water, sulfur, and carbon dioxide in submarine lavas from Kilauea Macdonald G. A. and Katsura T. (1964) Chemical composition of volcano, Hawaii. J. Geol. 99, 371–394. Hawaiian lavas. J. Petrol. 5, 83–133. Dixon J. E., Clague D. A., Poreda R., and Wallace P. (1997) Volatiles Michael P. J., and Cornell W. C. (1998) Influence of spreading rate in alkalic basalts from the North Arch volcanic field, Hawaii: Ex- and magma supply on crystallization and assimilation beneath tensive degassing of deep submarine-erupted alkalic series lavas. J. mid-ocean ridges: Evidence from chlorine and major element Petrol. 38, 911–939. chemistry of mid ocean ridge basalts. J. Geophys. Res. 103, Duennebier F. K., et al. (1996) Researchers rapidly respond to subma- 18325–18356. rine activity at Loihi volcano, Hawaii. Eos 78, 229–232. Michael P. J. and Schilling J.-G. (1989) Chlorine in mid-ocean ridge Eiler J. M., Farley K. F., Valley J. W., Hofmann A. W., and Stolper magmas: Evidence for assimilation of seawater-influenced compo- E. M. (1996) Oxygen isotope constraints on the sources of Hawaiian nents. Geochim. Cosmochim. Acta 53, 3131–3143. volcanism. Earth Planet. Sci. Lett. 144, 453–468. Moore J. G., Clague D. A., and Normark W. R. (1982) Diverse basalt Fournier R. O. (1987) Conceptual models of brine evolution in mag- types from Loihi seamount, Hawaii. Geology 10, 88–92. matic-hydrothermal systems. USGS Prof. Pap. 1350, 1487–1506. Norman M. D. and Garcia M. O. (1999) Primitive magma compositions Frey F. A. and Clague D. A. (1983) Petrology of diverse basalt types and source characteristics of the Hawaiian plume: Constraints from from Loihi Seamount, Hawaii: Petrogenetic implications. Earth picritic lavas. Earth Planet. Sci. Lett. 168, 27–44. Planet. Sci. Lett. 66, 337–355. Ozima M. and Podosek F. A. (1983) Noble Gas Geochemistry. Cam- Garcia M. O., Jorgensen B. A., and Mahoney J. J. (1993) An evaluation bridge University Press. of the temporal and geochemical evolution of Loihi seamount lavas: Patterson D. B., Honda M., and McDougall I. (1990) Atmospheric Results from Alvin submersible dives. J. Geophys. Res. 98, 537–550. contamination: A possible source for heavy noble gases in basalts Garcia M. O., Muenow D. W., and Aggrey K. E. (1989) Major element, from Loihi seamount in Hawaii. J. Geophys. Res. 17, 705–708. volatile and stable isotope chemistry of Hawaiian submarine tholei- Reiners P. W. (1998) Reactive melt transport in the mantle and geo- itic glasses. J. Geoph. Res. 94, 10525–10538. chemical signatures of mantle-derived magmas. J. Petrol. 39, 1039– Garcia M. O., Foss D. J. P., West H. B., and Mahoney J. J. (1995) 1061. Geochemical and isotopic evolution of Loihi volcano, Hawaii. J. Rucklidge J. C. and Patterson G. C. (1977) The role of chlorine in Petrol. 38, 1647–1674. serpentinization. Contr. Mineral. Petrol. 65, 39–44. Garcia M. O., Ito E., Eiler J. M., and Pietruszka A. J. (1998a) Crustal Sarda P., Staudacher T., and Allegre C. J. (1988) Neon isotopes in contamination of Kilauea volcano magmas revealed by oxygen iso- submarine basalts. Earth Planet. Sci. Lett. 91, 73–88. tope analyses of glass and olivine from Puu Oo eruption lavas. J. Schilling J-G., Bergeron M. B., and Evans R. (1980) Halogens in the Petrol. 39, 803–817. mantle beneath the North Atlantic. Phil. Trans. Royal Soc. London A Garcia M. O., Rubin, K. H., Norman M. D., Rhodes J. M., Graham 297, 147–178. D. W., Muenow D. W., and Spencer K. (1998b) Petrology and Seyfreid W. E., Berndt, M. E. and Janecky, D. R. (1986) Chloride geochronology of basalt breccia from the 1996 earthquake swarm of depletions and enrichments in seafloor hydrothermal fluids: Con- Loihi seamount, Hawaii: Magmatic history of the 1996 eruption. straints from experimental basalt alteration studies. Geochim. Cos- Bull. Volcan. 59, 577–592. mochim. Acta 50, 469–475. Hoffman A. W. (1997) Mantle geochemistry: The message from oce- Smith H. J., Spivack A. J., Staudigel H., and Hart S. R. (1995) The anic volcanism. Nature 385, 219–229. boron isotopic composition of altered oceanic crust. Chem. Geol. Holland H. D. (1984) The Chemistry of the Atmosphere and the 126, 119–135. Oceans. J Wiley. Staudacher T., Kurz M. D., and Allegre C. J. (1986) New noble gas Honda M., MacDougall I., Patterson D. B., Doulgeris A., and Clague data on glass samples from Loihi seamount and Hualalai and on Assimilation of a seawater-derived component 2761

dunite samples from Loihi and Reúnion island. Chem. Geol. 56, Stolper E. M., DePaolo D. J., and Thomas D. M. (1996) Introduction to 193–205. special section: Hawaii Scientific Drilling Project. J. Geophys. Res. Staudigel H. and Hart S. R. (1984) Alteration of basaltic glass: Mech- 101, 11593–11598. anisms and significance for the oceanic crust-seawater budget. Valbracht P. J., Staudacher T., Malahoff A., and Allegre C. J. (1997) Geochim. Cosmochim. Acta 47, 337–350. Noble gas systematics of deep rift zone glasses from Loihi seamount, Staudigel H., Plank T., White W., and Schmincke H.-U. (1996) Geo- Hawaii. Earth Planet. Sci. Lett. 150, 399–411. chemical fluxes during seafloor alteration of the basaltic upper oce- Zindler A. and Hart S. (1986) Chemical geodynamics. Ann. Rev. Earth anic crust: DSDP Sites 417 and 418. AGU Monograph 96, 19–38. Planet. Sci. 14, 493–771.