DEPARTMENT OF MARINE, EARTH, AND ATMOSPHERIC SCIENCES, NORTH CAROLINA STATE UNIVERSITY, RALEIGH, NC 27695, USA

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 PAGES 581 624 1997 A View into the Subsurface of Mauna Kea Volcano, Hawaii: Crystallization Processes Interpreted through the Petrology and Petrography of Gabbroic and Ultramafic Xenoliths R. V. FODOR AND P. GALAR DEPARTMENT OF MARINE, EARTH, AND ATMOSPHERIC SCIENCES, NORTH CAROLINA STATE UNIVERSITY, RALEIGH, NC 27695, USA RECEIVED AUGUST 7, 1995 ACCEPTED NOVEMBER 26, 1996 Xenoliths from the southern flank of Mauna Kea volcano form two intrusions; e.g. in situ and gravity-settled crystallization, extensive broad categories. (1) Ultramafic: porphyroclastic dunite, wehrlite, differentiation, varieties of layering, mobilizations of late-stage, and olivine clinopyroxenite (Fo 89 4 83 6, clinopyroxene mg-number evolved liquids, compaction and convective disturbances in reservoirs. 90 3 86 3, spinel mg-number 57 42, spinel cr-number 70 52, no plagioclase); and granular wehrlite and olivine clinopyroxenite (Fo 83 76 ) with plagioclase (An 84 69 ) ±orthopyroxene, and Crmagnetite. (2) Gabbroic: granular gabbro, gabbronorite, and troctolite composed of olivine+clinopyroxene frameworks (Fo 82 74, mgpetrology; KEY WORDS: gabbroic xenoliths; ultramafic xenoliths; cumulate; Hawaii number 85 79) enclosing plagioclase (~An 79 69 ) ±orthopyroxene, in situ crystallization; gravity settling; Mauna Kea volcano and Fe Ti oxides; and plagioclase (<An 77 ) forming frameworks for, and fine-grained mosaics with, evolved olivine (Fo 75 61 ), clinopyroxene±orthopyroxene, and Fe Ti oxides. Most xenoliths are petrographically uniform, but some manifest modal, phase, INTRODUCTION cryptic, or grain-size layering, and some are composites of two rock Ultramafic and gabbroic xenoliths in cinder-cone ejecta types. Whole-xenolith incompatible elements are depleted, and there and lavas on Mauna Kea volcano, Hawaii, offer opare positive Eu anomalies; 87 Sr/ 86 Sr is 0 70360, and mineral portunities to study crystallization environments and proδ 18 O is 4 05 5 62. Porphyroclastic ultramafic xenoliths are grav- cesses for magmas that constructed the Hawaiian Islands. ity-settled and in situ cumulates from reservoir bottoms. Plagioclase- Jackson et al. (1982) first provided a reconnaissance study bearing xenoliths represent modal, phase, and cryptic layering (e.g. of the xenoliths, and Fodor & Vandermeyden (1988) wehrlite to gabbronorite) in reservoir-margin solidification zones later presented an account of mineral and whole-rock superimposed with small-scale (centimeter) modal, cryptic, phase, compositions of xenoliths at the summit cone. This exand grain-size layering. Mineral compositions point to tholeiitic pansion of Mauna Kea studies evaluates ultramafic and parentage for most xenoliths, but alkalic for some (e.g. clinopyroxene gabbroic xenoliths at three neighboring cinder cones on Al 2 O 3 >4 wt %). These Mauna Kea xenoliths are plutonic Mauna Kea s southern flank (Fig. 1). It reveals modal complements to postshield lavas (Hamakua Volcanics), and they assemblages, textures, and mineral compositions more identify that stage of volcano development with 15 5 wt % diverse than previously recognized, and phase, modal, MgO magmas that underwent processes intrinsic to mafic-layered rhythmic, and grain-size layerings all characteristics of Corresponding author. Oxford University Press 1997

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Fig. 1. Map showing the five shield volcanoes that make up the island of Hawaii, and the xenolith location on the southern flank of Mauna Kea volcano. The enlargement shows the topographic relief of the three cones, A, B, and C, sampled for xenoliths and their specific locations on the volcano. grand-scale layered intrusive bodies. Our study is an southern flank of the volcano at ~2900 m (Fig. 1). One effort to elucidate the internal magmatic workings of cone is called Puu Kalepeamoa and another cone ~1 Hawaiian volcanoes, as well as to understand how plu- km north is called Kilohana; we refer to them as cone tonic assemblages relate to the stages of Hawaiian magmatism. A and cone C, respectively (Fig. 1). Each is ~500 m The results augment our understanding of across at its base and ~125 m high. Nestled between magma-reservoir crystallization down to a centimetersize them is the rim and shallow crater (~25 m relief ) of a scale and complement current knowledge about third (unnamed) cone that we call B (Fig. 1). Hawaiian magma systems available from seismic and lava-composition studies. XENOLITH ROCK TYPES STUDY AREA Sixty-eight studied xenoliths range from ~4 to 30 cm. They are subrounded to angular, and they occur on the Mauna Kea is one of five volcanoes on the island of flanks and in the craters of the cones as individual blocks Hawaii and reaches an altitude of 4230 m (Fig. 1). Its or with thin lava rinds; some form cores of volcanic lavas are categorized as shield-building and postshield bombs. The rock types are dunite, wehrlite, olivine (Hamakua and Laupahoehoe volcanics) (Frey et al., 1990, clinopyroxenite, gabbro, gabbronorite, and troctolite 1991; Wolfe et al., 1997). The youngest postshield substage (Table 1a and b). We broadly categorize them as ultra- (hawaiitic pyroclastics and lavas of the Laupahoehoe mafic and gabbroic, where ultramafic has <10 vol. % Volcanics, ~66 4 ka) is pertinent here because it erupted plagioclase. We distinguish gabbronorite from gabbro by xenoliths at dozens of sites located between ~2800 m its interstitial and granular orthopyroxene. All rock types elevation and the volcano summit (e.g. Jackson et al., occur at cone C, and the proportions approximate those 1982). The sites we sampled are cinder cones on the noted by Jackson et al. (1982) for this location. Cone A 582

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Table 1a: Ultramafic xenoliths from the southern flank of Mauna Kea volcano, Hawaii: petrography, mineral endmembers, and layering Rock type Texture Fo cpx-mg-no. An opx-mg-no. Mode (vol. %) Layering ol cpx Cr-oxide pl opx Cone A A19 dun porphyroclastic 87 8 89 6 94 9 1 6 3 5 A11 dun porphyroclastic 87 6 89 4 92 6 3 1 4 3 A16 dun porphyroclastic 86 5 89 4 97 8 0 2 2 A6 dun porphyroclastic 86 5 88 7 97 8 0 4 1 8 A27d dun porphyroclastic 85 2 89 92 8 1 3 5 9 modal A27w wehr porphyroclastic 84 9 88 1 53 6 45 1 1 3 modal A45 wehr anhedral gran/poik 84 7 88 74 2 24 4 1 4 A17 wehr porphyroclastic 84 7 86 6 65 5 34 0 5 A18 dun porphyroclastic 76 8 80 9 92 5 0 2 7 3 A28 wehr anhedral gran/mosaic 76 81 9 70 4 43 4 51 8 1 3 3 2 Cone B B22 dun porphyroclastic 89 4 90 3 98 5 0 3 1 2 B12 dun porphyroclastic 84 4 86 9 97 3 0 7 2 B1 wehr anhedral gran 81 7 84 7 72 3 83 8 14 6 1 6 tr Cone C C27 dun porphyroclastic 87 8 89 9 97 6 0 4 2 C61 dun porphyroclastic 87 8 89 8 97 8 0 5 1 7 C70 dun porphyroclastic 87 8 89 8 92 7 2 9 4 4 C47 dun porphyroclastic 87 6 89 7 97 6 0 4 2 C46 wehr porphyroclastic 86 4 88 1 64 6 33 9 1 5 C28 dun porphyroclastic 86 1 89 8 98 0 2 1 8 C29 wehr porphyroclastic 84 6 87 6 85 5 12 5 2 C25d dun porphyroclastic 84 3 86 8 96 5 0 1 3 4 modal C25c ol cpxn t porphyroclastic 83 6 86 3 15 4 84 6 modal C99 wehr porphyroclastic 83 7 85 2 73 8 22 6 3 6 C78 ol cpxn t anhedral gran 82 7 86 68 5 13 9 86 0 2 C30 wehr poikilitic 82 85 8 70 8 67 5 29 3 1 7 1 5 C9 ol cpxn t anhedral gran/mosaic 81 4 85 78 5 82 2 34 7 61 4 0 6 2 9 0 4 C12 ol cpxn t anhedral gran/mosaic 81 3 85 2 83 7 82 9 33 8 65 4 0 1 0 6 0 1 C75 ol cpxn t anhedral gran/mosaic 81 84 29 5 70 4 0 1 phase gab nor anhedral gran 80 4 82 7 65 7 82 2 1 5 37 7 0 1 48 1 12 7 phase C20 wehr anhedral gran/mosaic 80 9 85 3 84 6 82 9 58 4 39 3 0 2 1 8 0 2 C52 wehr poikilitic 80 1 84 3 83 1 43 3 56 2 0 1 0 4 C21 ol cpxn t anhedral gran/mosaic 77 3 82 1 64 4 79 4 18 2 81 4 0 1 0 2 0 1 C77 wehr anhedral gran/mosaic 77 2 81 6 67 6 79 3 55 4 40 6 0 4 3 5 0 1 Modes are based on 1500 2000 counts per thin section. dun, dunite; wehr, wehrlite; ol cpxn t, olivine clinopyroxenite; gab nor, gabbronorite (see footnote ). C75 is phase layered, where one layer is gabbronorite and therefore not ultramafic, but included in this table none the less because the xenolith is a composite. Anhedral granular: mosaic refers to an overprint of polygonal grains in triple-point junctures. Lower-case d, w, and c in sample numbers respectively refer to dunite, wehrlite, and olivine clinopyroxenite portions of composite xenoliths. Modal and phase refer to the type of layering in composite xenoliths. 583

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 1b: Gabbroic xenoliths from the southern flank of Mauna Kea volcano, Hawaii: petrography, mineral endmembers, and layering Rock type General texture Fo cpx mg- An opx mg- pl occurrence Modes (vol. %) Layering no. no. ol cpx pl opx opaque amph Cone A A2 gab nor anhedral gran 80 7 84 9 78 3 82 7 interstit 33 6 45 4 20 7 0 2 0 2 A14 gab nor anhedral gran/mosaic 79 7 83 9 76 2 82 1 interstit 43 3 44 6 10 4 0 3 1 4 A21 gab nor anhedral gran 78 82 6 72 2 78 1 interstit 14 1 62 9 22 7 0 1 0 3 A22 gab (ol) anhedral gran 77 81 6 76 9 interstit 28 36 2 35 7 0 1 modal grain size cryptic gab (pl) fine-grained mosaic 75 2 80 2 75 1 gran/mosaic 12 6 21 4 65 9 0 1 A12 gab nor anhedral gran/mosaic 76 9 82 6 70 9 79 5 interstit 15 7 53 5 17 6 12 8 0 3 0 1 A7 gab nor anhedral gran 76 9 81 5 75 3 78 2 gran/mosaic 10 3 55 6 33 2 0 2 0 6 0 1 A13 gab nor anhedral gran 76 8 81 1 72 3 78 8 interstit 15 6 50 3 31 7 1 7 0 3 0 4 phase gab nor anhedral gran interstit 57 5 26 9 15 2 0 4 A5 gab nor anhedral gran 76 8 81 2 59 8 78 2 gran/mosaic 2 35 2 32 3 24 8 5 7 A3 gab nor anhedral gran 73 6 79 6 65 1 77 5 gran/mosaic 11 58 24 4 0 5 6 1 A9 gab (ol) anhedral gran/mosaic 73 4 79 3 66 5 interstit 52 5 32 5 11 9 3 modal grain size cryptic gab (pl) fine-grained mosaic 71 8 78 5 69 1 gran/mosaic 10 4 35 52 1 2 3 A10 gab nor anhedral gran 72 4 80 3 66 6 77 3 gran 27 9 23 3 30 2 15 8 1 8 1 A15 gab anhedral gran/mosaic 72 77 8 70 6 gran/mosaic 12 8 25 6 56 8 4 8 A2-2 gab nor fine-grained mosaic 69 4 77 8 60 1 73 8 gran/mosaic 4 5 38 1 47 9 2 6 6 8 Cone B B8 gab anhedral gran altered 83 5 80 1 interstit 62 8 14 9 21 7 0 5 B5 gab nor anhedral gran 79 1 62 8 74 4 gran/mosaic 44 3 46 4 0 1 9 1 B4 gab nor anhedral gran 68 3 78 63 1 73 7 gran/mosaic 3 30 65 7 0 2 1 1 B3 gab nor fine-grained mosaic 76 5 61 4 71 8 gran/mosaic 36 9 43 9 4 3 14 9 B2 gab nor anhedral gran 77 63 73 3 gran/mosaic 38 7 47 5 13 4 0 4 Cone C C7 gab anhedral gran/mosaic 81 8 84 5 79 1 interstit 67 3 19 5 11 5 1 5 C17 gab nor anhedral gran/mosaic 80 9 83 4 68 5 82 7 interstit 70 2 10 4 14 9 4 3 0 2 C8 gab anhedral gran/mosaic 77 4 80 8 69 1 interstit 63 4 19 8 14 1 2 7 C22 gab anhedral gran/mosaic 75 9 79 75 3 interst/gran 11 1 42 7 32 5 12 9 0 8 variable cpx ol opx props 584

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES C15 gab anhedral gran/mosaic 75 8 83 4 69 1 interstit 67 3 13 17 5 1 8 0 4 C50 gab nor anhedral gran/mosaic 75 6 81 6 71 4 78 4 interst/gran 8 7 28 7 48 4 8 1 6 1 modal C18 gab anhedral gran/mosaic 74 8 78 9 66 7 gran/mosaic 1 1 35 9 60 4 2 0 6 modal C3 gab nor anhedral gran/mosaic 74 7 79 3 77 3 73 8 gran/mosaic 12 6 21 6 65 1 0 5 0 1 C59 gab anhedral gran 74 4 81 2 73 5 gran 21 50 4 24 3 1 3 2 9 C16 gab anhedral gran/mosaic 73 9 80 62 gran/mosaic 10 2 33 8 54 4 1 1 C49 troc anhedral gran 73 8 81 2 64 1 gran 18 4 0 2 59 1 22 2 C10c gab anhedral gran 73 4 79 69 8 gran/mosaic 14 7 15 8 69 4 0 1 grain size C10f gab fine-grained mosaic 73 6 79 1 71 4 gran/mosaic 16 4 39 3 42 9 0 9 0 5 grain size C76 gab nor fine-grained mosaic 72 3 81 3 66 3 74 gran/mosaic 14 7 60 7 21 4 0 5 2 7 C6 troc anhedral gran/mosaic 72 3 79 6 67 1 gran/mosaic 27 3 2 7 68 2 1 8 C14 troc anhedral gran/mosaic 72 3 80 9 62 1 gran/mosaic 30 3 1 1 67 6 0 9 C2 gab nor fine-grained mosaic 71 9 78 7 63 4 75 2 gran/mosaic 5 8 35 1 50 3 0 1 8 7 C5 gab nor fine-grained mosaic 71 7 77 3 64 72 3 gran/mosaic 8 7 36 5 48 4 0 2 6 1 0 1 C4 gab nor fine-grained mosaic 70 6 78 7 63 2 71 7 gran/mosaic 0 4 30 2 55 2 0 2 14 C1c gab anhedral gran/mosaic idd 79 1 65 8 gran/mosaic 11 1 31 1 55 6 2 2 grain size C1f gab fine-grained mosaic idd 78 1 71 9 gran/mosaic 13 1 35 5 45 1 6 3 grain size C19c gab nor anhedral gran/mosaic 62 66 78 4 62 9 69 8 gran/mosaic 3 38 45 9 0 1 13 grain size C19f gab nor fine-grained mosaic 61 77 9 64 7 68 7 gran/mosaic 0 1 48 45 8 0 4 5 4 0 3 grain size Modes are based on 1500 2000 counts per thin section. gab, gabrro; gab nor, gabbronorite; troc, troctolite; (ol) and (pl) refer to olivine-rich and plagioclase-rich portions of modally layered gabbroic xenoliths. Anhedral granular/mosaic refers to an overprint of polygonal grains in triple-point junctures. idd refers to olivine altered to iddingsite. pl occurrence refers to the dominant way in which plagioclase occurs: interstitial to ol + cpx frameworks, granularly, and in mosaics of polygonal grains in triple-point relationships. Layering refers to types of layering observed in xenolith; this column includes reference to some variable clinopyroxene olivine plagioclase proportions, or varying proportions of these phases across a cut surface of the sample. Lower-case c, and f in sample numbers refer to coarse and fine portions of grain-size layered gabbroic xenoliths. 585

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 yielded all but troctolite, and contains a preponderance of Wehrlite and olivine clinopyroxenite gabbronorite. Dunite, wehrlite, gabbro, and gabbronorite Wehrlites and olivine clinopyroxenites without plaare at cone B. Table 1(a and b) lists the rock types. gioclase are almost all porphyroclastic with olivine, commonly strained, and clinopyroxene, 3 5 mm, within mosaics of polygonal olivine and clinopyroxene (Fig. 3b); one plagioclase-free wehrlite is granular. Cr-spinel occurs within and between the silicate grains. PETROGRAPHY These rock types with interstitial plagioclase (~0 2 3 5 Uniform and nonuniform vol. %) are mainly anhedral granular, but some wehrlites Most xenoliths have uniform modes, textures, and grain are poikilitic. The granular samples contain areas where sizes, but some vary in these features. Plagioclase-bearing olivine and clinopyroxene are in mosaics of polygonal ultramafic xenoliths, for instance, have modal olivine/ grains, and some have interstitial orthopyroxene. Strained clinopyroxene ratios that change across a few centimeters olivine grains are present but not common. Cr-magnetite distance, and plagioclase occurs concentrated in areas, is the opaque phase. or pools, that may be a few millimeters thick (Fig. 2a). Depending on the cut portion of a xenolith examined, then, an overall olivine clinopyroxenitic or wehrlitic Gabbro, gabbronorite, and troctolite xenolith may locally (1 2 cm 2 ) be dunitic or gabbroic. Coarse anhedral granular. Most gabbroic xenoliths have Similarly, distributions of mafic minerals in plagioclaseanhedral granular textures (Fig. 3c f ) with grains ~0 5 4 rich xenoliths may be nonuniform across a surface of, mm, but some ~5 12 mm. The granularity of most of say, 10 cm 2 (Fig. 2b). Accordingly, single thin sections these xenoliths, which is usually manifested as interlocking for some xenoliths provide only good approximations of anhedral (occasionally subhedral) olivine, clinopyroxene, modal percentages. Nonuniformity also occurs as layers, ~0 3 3 cm, of and plagioclase, is overprinted to some extent by mosaics concentrated plagioclase, or olivine, or pyroxene to form of equigranular ~0 1 3 mm grains that share polygonal local leuco-gabbro, anorthosite, troctolite, wehrlite, or grain boundaries and triple-point junctures (Fig. 3c f ). pyroxenite (Fig. 2c). Modal layering occurs where dunites Strained olivine grains are rare. Orthopyroxene is usually are in composite xenoliths with wehrlite or olivine cliand enclosing other phases. Generally, Cr Fe Ti oxides intergranular, <0 5 mm, but sometimes large, 2 4 mm, nopyroxenite (Fig. 2d), and there is phase layering in a composite of gabbronorite and plagioclase-free olivine are <5 vol. % (one troctolite has ~22 vol. % ilmenite) clinopyroxenite (Fig. 2b), and in a gabbronorite with an and occur interstitially and as inclusions in pyroxene and olivine-free zone (Fig. 2f ). Some xenoliths have tapering, interstitial plagioclase. Some gabbroic xenoliths have contorted layers that describe small-amplitude folds, and amphibole intimately associated with clinopyroxene, and some xenoliths display small-scale faults (Fig. 2e). it is probably an alteration product thereof. Grain-size layering creates composite gabbroic xenof plagioclase abundances (~10 68 vol. %; Table 1b) We further evaluate coarse, gabbroic xenoliths in terms oliths that are uniform with respect to modes but have fine-grained (~0 5 1 mm) portions separated from coarsevariations. Specifically, one variety has up to ~30 35 and the petrographic aspects that attend those modal grained (~1 3 mm) portions by interfaces (Fig. 2g and h). Sometimes modal (and cryptic) layering attends grainolivine+clinopyroxene frameworks; this intra-framework vol. % plagioclase that is intergranular within size layering (olivine-rich portions, 1 5 mm; plagioclaseplagioclase often consists of mosaics of polygonal grains rich portions, 0 5 1 mm; Fig. 2f ). Table 1(a and b) identifies the layered xenoliths. (Fig. 3c and d). Another variety occurs when plagioclase exceeds ~35 vol. % and its coexisting olivine and clinopyroxene are too dispersed to constitute frameworks to contain the plagioclase (Fig. 3e). Ultimately, where plagioclase is >50 vol. %, as in troctolites, it forms Rock types and textures frameworks of polygonal grains enclosing olivine, clino- Dunite pyroxene, and Fe Ti oxide (Fig. 3f ). Dunites are porphyroclastic (Fig. 3a) and have large (3 5 mm) strained, or kink-banded, olivines within smaller, Fine-grained mosaic. Gabbro and gabbronorite xenoliths <2 mm, equigranular, polygonal olivine. Cr-spinel occurs also occur as fine-grained equigranular mosaics of anwithin and among olivine grains, and clinopyroxene is hedral and polygonal 0 2 1 0-mm plagioclase, clino- interstitial; both phases are generally <5 vol. %. No pyroxene, olivine, and Fe Ti oxides (Fig. 3g). They plagioclase was observed in dunite, but one sample, A18, generally have >40 vol. % plagioclase. Textures suggest contains a titaniferous magnetite vein with apatite. undercooling and co-crystallization of silicate phases; 586

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Fig. 2. (a) Wehrlite (C20) and olivine clinopyroxenite (C12) contain plagioclase interstitial to olivine+clinopyroxene frameworks, but plagioclase is sometimes concentrated (arrows), particularly in xenolith C20 as a layer several millimeters thick. Throughout, the olivine and clinopyroxene are in varying amounts and modal olivine/clinopyroxene ratios accordingly create local dunitic areas. (b) Plagioclase-free olivine clinopyroxenite C75 is capped by gabbronorite to create phase layering; plagioclase-rich gabbro C3 and troctolite C14 have uneven distributions of mafic minerals (arrows show depletions), creating areas unrepresentative of overall modes. (c) A pyroxenitic layer in gabbronorite C50 and anorthositic and clinopyroxenitic layers in gabbro C18. (d) Modal layering creating dunite wehrlite and dunite clinopyroxenite composite xenoliths. (e) Small-amplitude folds, tapered layers, and faulted layers (arrow) are evident in some gabbroic xenoliths of cone C (these two not examined quantitatively). (f ) Gabbros A22 and A9 are modally layered with olivine-rich and plagioclase-rich portions; there are also grain-size distinctions and mineral composition distinctions (cryptic layering) between the two types of portions in each gabbro. Gabbronorite A13 has phase layering, where olivine is absent from the lower zone as shown. (g) Fine- and coarse-grained composite gabbros (grain-size layering); the fine-grained portion of C10 is foliated. (h) A photomicrograph (planepolarized light) showing the interface between the fine and coarse portions of composite C1 in (g). 587

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Fig. 3. (a) Large kink-banded olivine adjacent to a mosaic of smaller olivine grains in porphyroclastic dunite C70; crossed-nicols. (b) Porphyroclastic wehrlite A17 with large olivine and clinopyroxene in a mosaic of smaller olivine; crossed nicols. (c) Anhedral granular texture of wehrlite C77 with interstitial plagioclase (one area labeled); plane-polarized light. (d) Gabbro C8 is a framework of olivine+clinopyroxene with ~14 vol. % interstitial plagioclase; plane-polarized light. (e) Gabbronorite A10 is composed of about one-third plagioclase, too large a volume to any longer be interstitial or intergranular within an olivine+clinopyroxene framework; plane-polarized light. (f ) Gabbro C16 is over one-half plagioclase, which creates a framework for olivine and clinopyroxene; plane-polarized light. (g) Fine-grained mosaic-textured gabbronorite C5; crossed-nicols. (h) Chadacrysts of olivine in an oikocryst of clinopyroxene in wehrlite C30. 588

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Fe Ti oxides, however (5 10 vol. %), are largely inter- State University Radiation Center. Before those analyses, stitial and probably late-forming. Some fine-grained xenoliths we cleaned the minerals in warm 10% HCl overnight. are foliated owing to <1 mm layers of concentrated Sr and Pb isotope compositions for xenoliths and clino- clinopyroxene, olivine, or plagioclase (Fig. 2g), and some pyroxene grains are from analyses we did at the University are portions of composite fine- and coarse-grained gab- of North Carolina isotope facility, using a VG Sector 54 broic xenoliths (grain-size layering; Fig. 2g and h). mass spectrometer. The 87 Sr/ 86 Sr ratio for SRM987 is 0 710254±15 (2σ), and Pb is corrected for mass Recrystallization and unmixing features discrimination by a factor of 0 15%/a.m.u. Krueger Enterprises provided the oxygen isotope analyses for Polygonal olivine, clinopyroxene, and/or plagioclase in mineral separates. nearly all xenoliths (Fig. 3c f ) suggest annealing or recrystallization histories. These grain boundaries can also be described as adcumulus and forming when interstitial liquids contributed to crystal growth and mutual interference MINERAL COMPOSITIONS boundaries (e.g. Hunter, 1987). Some gabbroic Olivine xenoliths have undergone subsolidus unmixing to create Ultramafic xenoliths without plagioclase have comexsolution lamellae of orthopyroxene or Fe Ti oxide. positionally primitive olivine, ~Fo 89 4 83 6, except dunite A18 with Fo 76 8 (Table 2; Fig. 4). Plagioclase-bearing wehrlite and olivine clinopyroxenite have olivine ( Table Weathering or alteration 3) that is slightly more evolved, Fo 82 7 76, and gabbroic There are three types of secondary modifications. (1) xenoliths have even more evolved olivine, Fo 81 61. The Olivine is oxidized to iddingsite in some cone-a and fine-grained variety invariably has the most evolved olcone-c xenoliths, and in most cone-b xenoliths. (2) ivine, <Fo 73 (Table 1b). Compositional zoning in any of Extreme alteration, perhaps deuteric, in cone-b xenoliths the xenoliths is generally <3 mol %. The Fo range we blackened some olivine; in these samples, olivine grains observe in the xenoliths collectively has also been obcan be entirely opaque or contain seemingly fresh areas served for olivine phenocrysts in Mauna Kea lavas (Frey mixed in with opaque areas. (3) Amphibole is an alteration et al., 1990, 1991; Yang et al., 1994; Baker et al., 1996). or reaction phase within some clinopyroxene grains in Modal- and phase-layered xenoliths (Fig. 2d and f; cone-a and cone-c gabbroic xenoliths. Table 1) generally have olivine ~0 3 1 8 mol % Fo higher in the olivine-dominant layers. As examples, modally ANALYTICAL TECHNIQUES layered gabbros A9 and A22 have the pairs Fo 73 4 and Fo 71 8, and Fo 77 and Fo 75 2, respectively, and olivine We determined mineral compositions with ARL-EMX clinopyroxenite gabbronorite composite C75 has Fo 81 and ARL-SEMQ electron microprobes at North Carolina and Fo 80 4. In gabbroic grain-size composites, however, State University using olivine, clinopyroxene, ortho- olivine compositions are essentially the same in each pyroxene, plagioclase, microcline, spinel, and ilmenite layer (Fig. 2g; Table 1b). provided by the Smithsonian Institution and Ni-doped NiO abundances in olivine correlate with Fo (Fig. 4). diopside as reference minerals. We applied the matrix Ultramafic xenoliths have the highest NiO, 0 42 0 25 corrections of Bence & Albee (1968) for EMX analyses wt %, and gabbroic samples have lower amounts, ~0 35 and phi rho Z for SEMQ analyses. Our data tables 0 05 wt % NiO. Similarly, the highest CaO contents are present average values for 8 15 analytical points on single in ultramafic xenoliths (0 06 0 20 wt %) and the lowest grains that we assessed as representative of a sample after are in gabbroic xenoliths (0 02 0 10 wt %), where they analyzing several of its grains. form a flat trend when plotted against Fo (Fig. 4). Whole-rock samples of selected xenoliths were analyzed The NiO values are characteristic of NiO in olivines of in duplicate for major- and trace-element abundances Hawaiian lavas and gabbros, but CaO abundances are (except REE) on a Phillips 1410 X-ray fluorescence low compared with olivine in Hawaiian lavas (Fig. 4). spectrometer at NCSU; procedures have been described by Fodor et al. (1992). Major elements, except K, Na, and P, were analyzed using glass disks; trace elements and K, Na, and P were determined on pressed powder. Clinopyroxene Reference samples for calibration curves include standxenoliths All samples have clinopyroxene, but in some gabbroic ards of the US, Canadian, South African, and Japanese it includes exsolved orthopyroxene and/or Fe- geological surveys. Neutron activation analyses for rare- oxide that complicate acquisition of compositional data. earth element abundances on whole-rocks, clinopyroxene, For grains with lamellae, we analyzed portions that are and plagioclase were performed at the Oregon optically clear, and our analyses, therefore, presumably 589

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 2: Olivine, pyroxene, plagioclase, and oxide compositions (in wt %) for representative ultramafic xenoliths of Mauna Kea volcano Dunite B22 C27 A19 A11 A6 A16 B12 A18 ol cpx sp ol cpx sp ol cpx sp sp ol cpx sp ol cpx sp ol cpx sp ol cpx sp ol cpx Cr-mt Cr-mt Ti-mt SiO 2 39 9 51 9 40 53 39 6 51 6 39 5 51 6 39 8 51 9 39 6 52 8 39 52 6 38 52 2 TiO 2 0 76 2 4 0 55 2 0 7 2 2 1 9 0 88 1 8 0 75 2 5 0 47 1 9 0 6 1 8 0 36 3 4 2 2 14 7 Al 2O3 3 15 2 2 5 13 2 2 8 17 4 15 5 4 2 16 4 3 6 14 2 2 2 12 4 2 9 13 7 2 7 9 7 10 5 8 Cr 2O3 1 2 39 3 0 93 40 9 0 97 39 9 42 0 71 41 5 0 91 37 5 1 41 4 0 87 32 5 0 8 21 2 24 5 3 3 FeO 10 3 3 2 29 1 12 3 4 29 5 11 8 3 6 25 4 26 1 11 9 3 5 26 7 13 3 7 32 9 13 3 5 32 15 4 5 40 9 21 5 6 7 55 2 51 67 1 MnO 0 14 0 06 0 31 0 09 0 04 0 26 0 16 0 08 0 22 0 23 0 14 0 1 0 2 0 14 0 1 0 28 0 15 0 1 0 24 0 22 0 09 0 29 0 41 0 33 0 46 0 42 0 51 MgO 48 9 16 7 11 5 48 2 17 11 7 47 5 17 4 12 6 12 1 47 3 16 6 12 6 46 7 16 3 10 9 46 6 16 6 10 1 45 4 16 6 9 39 9 15 9 6 4 6 3 6 1 CaO 0 09 22 6 0 14 22 8 0 1 21 8 0 13 21 8 0 17 22 6 0 17 22 9 0 14 22 8 0 14 20 8 Na 2O 0 54 0 44 0 62 0 5 0 42 0 46 0 42 0 62 K 2O NiO 0 42 0 35 0 38 0 39 0 38 0 35 0 29 0 25 Total 99 75 99 96 97 81 100 78 100 66 97 56 99 54 99 57 97 72 97 83 99 36 99 89 99 2 100 19 100 28 98 28 99 87 100 03 98 04 100 05 101 38 98 19 100 2 100 41 96 36 94 42 Recalc. FeO 18 2 17 2 16 7 17 16 7 19 3 19 7 21 6 25 8 24 2 36 2 Recalc. Fe 2O3 12 1 13 6 9 6 10 2 11 1 15 1 13 7 21 5 32 7 29 7 34 4 cr-no. 63 4 67 5 60 6 64 5 62 9 63 9 69 1 61 4 59 4 62 2 27 6 mg-no. (Fo) 89 4 90 3 52 9 87 8 89 9 54 8 87 8 89 6 57 3 56 87 6 89 4 57 86 5 88 7 50 2 86 5 89 4 47 8 84 4 86 9 42 6 76 8 80 9 30 7 31 6 23 1 Fs 5 2 5 4 5 6 5 7 6 5 6 7 1 10 7 Wo 46 8 46 5 44 7 45 8 47 47 45 9 42 9 An Or Equilib. T ( C) 590

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Wehrlites C46 A45 A17 C29 C99 C30 B1 ol cpx sp ol cpx sp ol cpx sp ol cpx sp ol cpx sp ol cpx cpx Cr-mt Cr-mt pl ol cpx Cr-mt pl SiO 2 39 6 51 8 39 4 51 9 39 1 52 39 6 52 6 39 3 51 2 39 52 2 51 1 49 7 38 7 51 7 50 1 TiO 2 0 77 2 3 0 73 3 0 62 2 5 0 54 2 3 0 98 3 2 0 67 1 6 5 8 4 0 86 7 4 Al 2O3 3 6 19 7 3 1 18 7 3 7 19 8 2 9 17 4 4 17 5 2 9 4 11 3 9 9 31 6 3 1 10 4 32 Cr 2O3 1 37 1 1 32 7 0 91 34 9 0 93 38 4 0 77 32 1 1 0 89 27 5 25 9 0 76 24 7 FeO 13 2 4 1 28 9 14 8 4 2 33 9 14 8 4 7 28 5 14 6 4 5 28 5 15 6 5 1 36 7 17 1 5 5 2 44 1 45 7 0 34 17 5 5 5 47 3 0 17 MnO 0 18 0 09 0 24 0 21 0 09 0 29 0 12 0 11 0 24 0 11 0 06 0 27 0 21 0 12 0 31 0 26 0 12 0 13 0 34 0 33 0 22 0 12 0 29 MgO 47 17 12 2 46 1 17 2 10 6 45 8 17 12 4 46 1 17 8 11 4 44 9 16 5 10 2 43 6 16 9 16 7 5 6 6 43 9 17 1 7 6 CaO 0 11 21 0 09 21 8 0 07 21 0 09 20 6 0 14 21 4 0 06 21 2 21 4 14 3 0 07 20 8 14 4 Na 2O 0 37 0 37 0 46 0 33 0 34 0 41 0 48 3 2 0 33 3 K 2O 0 11 0 08 NiO 0 34 0 33 0 32 0 29 0 3 0 3 0 34 Total 100 43 99 73 100 44 100 93 100 39 99 19 100 21 100 5 98 34 100 79 100 26 97 87 100 45 100 81 100 01 100 32 100 4 100 2 97 24 96 83 99 25 100 73 100 27 97 69 99 75 Recalc. FeO 18 6 21 2 17 9 18 6 22 27 3 30 1 28 1 Recalc. Fe 2O3 11 3 14 1 11 8 11 16 2 18 6 17 4 21 4 cr-no. 55 7 54 54 2 60 2 55 2 62 63 7 61 4 mg-no. (Fo) 86 4 88 1 53 9 84 7 88 47 1 84 7 86 6 55 2 84 6 87 6 52 2 83 7 85 2 45 2 82 85 8 84 6 32 9 28 1 81 7 84 7 32 5 Fs 6 7 6 7 6 1 7 2 8 2 8 8 5 8 8 Wo 43 9 44 5 44 2 42 2 44 3 43 6 44 9 42 6 An 70 8 72 3 Or 0 65 0 48 Equilib. T ( C) 591

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 2: continued C20 C52 C77 A28 ol cpx opx Cr-mt pl ol cpx Cr-mt pl ol cpx opx Cr-mt pl ol cpx Ti-mt pl SiO 2 38 3 52 3 55 46 7 38 5 51 1 47 5 38 1 51 6 54 2 51 37 4 50 6 49 8 TiO 2 0 7 0 34 5 4 0 86 8 9 0 65 0 27 17 5 1 2 15 Al 2O3 3 1 5 13 4 33 9 4 12 6 33 9 2 4 1 4 8 1 31 1 4 4 5 1 31 5 Cr 2O3 0 58 0 19 25 9 0 84 21 0 5 0 15 20 4 0 66 6 6 FeO 18 1 5 1 11 3 43 2 0 44 18 6 5 5 46 3 0 37 21 4 6 9 13 7 42 1 0 47 22 3 6 3 63 1 0 41 MnO 0 2 0 11 0 24 0 3 0 24 0 1 0 3 0 27 0 19 0 3 0 3 0 32 0 14 0 33 MgO 42 9 16 6 30 7 8 2 42 16 5 7 1 40 6 17 2 29 5 7 1 0 07 16 6 2 CaO 0 01 21 5 0 9 17 6 0 1 21 1 16 7 0 07 19 2 0 84 13 6 39 6 21 1 14 3 Na 2O 0 33 0 01 1 7 0 38 1 8 0 34 0 01 3 5 0 33 3 2 K 2O 0 08 0 12 0 17 0 2 NiO 0 25 0 25 0 26 0 22 Total 99 76 100 22 100 18 96 4 100 42 99 69 100 38 96 2 100 39 100 7 98 98 100 37 99 84 99 91 100 73 96 33 99 41 Recalc. FeO 25 3 30 36 6 35 9 Recalc. Fe 2O3 20 3 18 1 6 1 30 2 cr-no. 56 5 52 8 62 8 46 5 mg-no. (Fo) 80 9 85 3 82 9 36 5 80 1 84 3 29 7 77 2 81 6 79 3 25 7 76 81 9 23 5 Fs 8 2 16 8 8 9 11 1 20 3 7 10 2 Wo 44 3 1 7 43 7 39 6 1 6 43 7 An 84 8 83 1 67 6 70 4 Or 0 46 0 71 1 1 2 Equilib. T ( C) 1000 1087 592

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Olivine clinopyroxenites C78 C12 C21 ol cpx pl ol cpx opx Cr-mt pl ol cpx opx ilm pl SiO 2 39 3 52 3 51 38 9 52 53 2 46 5 38 2 52 55 52 TiO 2 0 6 0 61 0 32 4 8 0 91 0 35 51 Al 2O3 2 4 31 1 3 5 14 7 34 1 2 8 1 4 0 59 30 5 Cr 2O3 0 6 0 75 0 28 27 4 0 49 0 14 1 5 FeO 16 5 5 1 0 37 17 6 5 1 11 5 42 7 0 37 21 2 6 3 13 5 36 7 0 33 MnO 0 22 0 12 0 15 0 16 0 22 0 28 0 25 0 25 0 26 0 24 MgO 44 1 17 5 43 16 4 31 3 8 1 40 5 16 2 29 2 8 1 CaO 0 07 21 2 14 1 0 04 21 8 0 62 17 3 0 04 20 9 1 13 2 Na 2O 0 24 3 4 0 27 0 04 1 8 0 39 0 05 3 9 K 2O 0 3 0 1 0 21 NiO 0 27 0 28 0 24 Total 100 46 100 06 100 27 99 97 100 09 98 98 97 98 100 17 100 43 100 24 100 9 98 13 100 14 Recalc. FeO 25 7 31 2 Recalc. Fe 2O3 18 8 6 2 cr-no. 55 6 mg-no. (Fo) 82 7 86 81 3 85 2 82 9 36 77 3 82 1 79 4 Fs 8 8 2 16 9 10 2 20 1 Wo 42 8 44 9 1 2 43 3 1 9 An 68 5 83 7 64 4 Or 1 7 0 58 1 2 Equilib. T ( C) 969 1002 593

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 2: continued Composites A27 A27 C25 C25 C75 C75 (dunite) (wehrlite) (dunite) (ol (ol (ol cpx nt) cpx nt) gabn rt) ol cpx sp ol cpx sp ol cpx sp ol cpx ol cpx Cr-mt ol cpx opx ilm pl SiO 2 38 8 53 2 38 7 52 6 39 3 52 39 5 52 1 38 7 52 38 3 51 2 55 1 50 9 TiO 2 0 42 2 4 0 54 3 3 0 53 2 6 0 67 0 64 4 9 0 66 0 41 51 4 Al 2O3 2 2 13 7 2 3 14 1 3 3 17 9 3 6 3 1 16 4 3 1 2 0 2 30 9 Cr 2O3 0 57 34 6 0 78 38 2 1 32 2 0 66 0 77 24 7 0 71 0 08 0 39 FeO 14 2 3 8 34 2 14 3 4 2 31 7 15 1 4 6 33 9 15 8 5 3 17 9 5 9 44 18 3 6 4 11 6 34 2 0 3 MnO 0 2 0 12 0 28 0 2 0 12 0 28 0 12 0 09 0 25 0 14 0 16 0 24 0 13 0 19 0 23 0 17 0 24 0 42 MgO 45 8 17 2 10 3 45 2 17 5 10 45 5 17 11 45 17 1 42 8 17 4 8 3 42 2 17 2 30 10 1 CaO 0 1 22 3 0 09 21 1 0 06 20 7 0 07 20 6 0 04 20 6 0 04 19 8 1 2 13 3 Na 2O 0 3 0 33 0 38 0 36 0 32 0 36 0 03 3 7 K 2O 0 21 NiO 0 32 0 31 0 31 0 26 0 24 0 25 Total 99 42 100 11 95 48 98 8 99 47 97 58 100 39 99 6 97 85 100 77 100 55 99 92 100 86 98 49 99 32 99 5 99 86 96 71 99 31 Recalc. FeO 19 2 21 1 19 8 23 7 27 8 Recalc. Fe 2O3 16 7 11 8 15 7 17 7 7 1 cr-no. 62 9 64 5 54 8 48 7 mg-no. (Fo) 85 2 89 48 9 84 9 88 1 45 8 84 3 86 8 49 8 83 6 85 2 81 84 41 2 80 4 82 7 82 2 Fs 6 6 7 7 5 8 5 9 3 10 2 17 4 Wo 43 4 43 3 43 2 42 5 41 7 40 7 2 3 An 65 7 Or 1 2 Equilib. T ( C) 1071 Prefixes A, B, and C in sample numbers refer to locations of sample sites cones shown in Fig. 1. ol, olivine; cpx, clinopyroxene; opx, orthopyroxene; pl, plagioclase; sp, Cr-spinel; Cr-mt, Cr-rich magnetite; Ti-mt, Ti-rich magnetite; ilm, ilmenite. Recalc. refers to recalculated FeO and Fe 2 O 3 based on stoichiometry of oxide phases. Composites refers to two rock types forming a single xenolith. Equilibration T refers to clinopyroxene orthopyroxene geothermometry (Wells, 1997). 594

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Table 3: Olivine, clinopyroxene, orthopyroxene, and plagioclase compositions (wt %) in representative gabbroic xenoliths of Mauna Kea volcano C7 C17 A2 A14 A21 C8 A22 (ol) ol cpx pl ol cpx opx pl ol cpx opx pl ol cpx opx pl ol cpx opx pl ol cpx pl ol cpx pl SiO 2 39 3 51 3 47 8 38 7 51 8 54 8 51 2 38 9 52 2 55 4 47 5 38 6 51 6 54 7 48 5 38 1 51 6 53 6 49 9 38 2 51 4 49 9 37 9 51 1 48 5 TiO 2 0 81 0 96 0 4 0 61 0 78 0 89 0 41 0 86 0 38 1 0 98 Al 2O3 3 5 33 4 3 1 6 31 3 2 8 1 5 32 9 3 1 5 32 7 2 9 1 4 31 5 3 9 31 3 3 7 33 4 Cr 2O3 0 92 0 57 0 31 0 84 0 36 0 64 0 22 0 47 0 16 0 69 0 68 FeO 17 3 5 6 0 29 18 5 8 11 3 0 32 18 1 5 5 11 5 0 29 19 3 5 9 11 9 0 38 20 4 6 3 14 2 0 23 21 2 6 8 0 31 21 4 6 5 0 28 MnO 0 25 0 12 0 26 0 12 0 18 0 16 0 17 0 23 0 18 0 13 0 22 0 22 0 19 0 3 0 29 0 15 0 25 0 15 MgO 43 6 17 1 42 8 16 4 30 3 42 5 17 3 30 8 42 4 17 2 30 6 40 5 16 8 28 4 40 8 16 40 2 16 2 CaO 0 05 19 8 16 0 06 21 1 2 13 8 0 02 20 5 0 73 15 9 0 03 20 4 0 95 15 5 0 05 21 2 0 99 0 05 20 8 14 0 05 20 7 15 9 Na 2O 0 33 2 3 0 4 0 05 3 4 0 36 0 03 2 4 0 34 0 01 2 6 0 36 0 04 14 5 0 42 3 4 0 37 2 6 K 2O 0 06 0 18 0 06 0 13 3 0 1 0 07 NiO 0 25 0 35 0 28 0 27 0 24 0 14 0 24 0 25 Total 100 75 99 48 99 85 100 17 100 05 100 14 100 2 99 96 100 28 101 33 99 05 100 78 100 1 100 51 99 81 99 51 100 68 99 47 99 27 100 78 101 16 99 01 100 05 100 38 100 75 mg-no. 81 8 84 5 80 9 83 4 82 7 80 7 84 9 82 7 79 7 83 9 82 1 78 82 6 78 1 77 4 80 8 77 81 6 Fs 9 1 9 4 16 9 8 8 17 1 9 4 17 6 9 9 21 5 11 10 5 Wo 41 3 43 5 2 3 42 1 4 41 7 1 8 42 9 1 9 43 42 9 An 79 1 68 5 78 3 76 2 72 2 69 1 76 9 Or 0 35 1 06 0 35 0 76 0 83 0 59 0 4 Equilib. 1011 1055 1051 999 T ( C) 595

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 3: continued A22 (pl) A7 A13 A5 C22 C15 ol cpx pl ol cpx opx pl ol cpx opx pl ol cpx opx pl pl ol cpx pl pl ol cpx pl SiO 2 37 8 51 4 49 2 37 9 51 9 54 8 49 1 37 6 51 9 54 6 49 8 37 9 52 54 5 49 3 52 7 38 1 48 9 47 1 49 38 51 7 51 TiO 2 1 1 1 0 46 0 85 0 45 0 65 0 35 2 1 1 1 Al 2O3 3 3 32 7 3 1 5 32 4 2 6 1 4 31 8 2 6 1 3 31 3 28 5 5 2 34 1 32 2 3 3 31 5 Cr 2O3 0 49 0 36 0 15 0 29 0 15 0 28 0 11 0 02 0 32 FeO 22 8 7 1 0 31 21 5 6 7 14 2 0 41 21 8 6 8 13 5 0 31 21 8 6 3 14 2 0 33 0 35 22 6 7 0 34 0 29 22 4 5 6 0 33 MnO 0 26 0 17 0 2 0 15 0 18 0 26 0 11 0 18 0 26 0 13 0 23 0 42 0 12 0 28 0 13 MgO 38 8 16 1 40 1 16 6 28 6 40 4 16 4 28 2 40 4 15 9 28 5 39 4 14 1 39 4 15 8 CaO 0 06 21 2 15 6 0 02 20 4 0 74 15 3 0 06 20 9 1 7 14 7 0 06 21 1 1 3 14 7 11 9 0 09 21 6 17 1 15 3 0 04 22 3 14 Na 2O 0 41 2 8 0 35 0 01 2 7 0 4 0 05 3 0 4 0 03 2 9 4 3 0 46 1 7 2 7 0 39 3 4 K 2O 0 1 0 13 0 17 0 12 0 2 0 1 0 12 0 11 NiO 0 24 0 27 0 28 0 28 0 09 0 25 Total 99 96 101 27 100 71 99 99 100 46 100 64 100 04 100 4 100 25 100 23 99 78 100 7 99 36 100 52 98 65 97 95 100 1 99 2 100 44 99 61 100 37 100 64 100 34 mg-no. 75 2 80 2 76 9 81 5 78 2 76 8 81 1 78 8 76 8 81 2 78 2 75 9 79 75 8 83 4 Fs 11 3 10 7 21 5 10 8 20 5 10 2 21 3 11 2 9 Wo 43 2 41 9 1 4 42 7 3 3 43 9 2 5 46 5 45 9 An 75 1 75 3 72 3 73 6 59 8 84 3 75 3 69 1 Or 0 57 0 76 0 99 0 71 1 2 0 59 0 7 0 65 Equilib. 1040 1013 985 T ( C) 596

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES C50 C18 C3 C59 C16 C49 A3 ol cpx opx pl ol cpx pl ol cpx opx pl ol cpx cpx pl ol cpx pl ol cpx pl ol cpx opx pl SiO 2 38 1 51 5 54 3 50 7 37 5 49 8 51 1 37 5 50 9 53 2 48 6 37 6 50 8 50 6 49 5 37 5 51 5 51 3 37 8 53 3 52 6 37 2 51 7 54 8 52 2 TiO 2 0 92 0 48 1 3 1 0 33 1 1 1 1 1 1 0 43 0 73 0 28 Al 2O3 2 8 1 5 32 1 4 1 31 8 3 5 1 3 32 8 4 1 3 6 32 3 1 30 7 1 8 30 3 2 6 1 1 30 5 Cr 2O3 0 32 0 12 0 34 0 27 0 04 0 53 0 25 0 11 0 03 0 14 0 04 FeO 22 6 6 4 13 9 0 35 23 6 7 4 0 42 23 2 7 4 16 7 0 42 23 6 6 3 7 0 33 24 7 1 0 39 23 8 6 6 0 33 24 6 7 3 14 9 0 32 MnO 0 25 0 16 0 24 0 33 0 14 0 21 0 14 0 31 0 35 0 14 0 17 0 32 0 18 0 33 0 13 0 11 0 18 0 22 MgO 39 3 15 9 28 3 39 2 15 4 38 4 15 9 26 4 38 5 15 9 15 5 38 2 15 9 37 5 16 38 5 16 28 7 CaO 0 1 21 2 1 5 14 4 0 09 20 2 13 6 0 03 20 8 0 93 15 9 0 08 20 6 20 8 14 8 0 05 21 4 12 9 0 07 22 3 13 0 02 20 6 0 86 13 2 Na 2O 0 33 0 02 3 1 0 39 3 6 0 36 0 01 2 5 0 4 0 42 2 9 0 45 4 3 0 32 3 9 0 37 0 04 3 8 K 2O 0 14 0 23 0 14 0 09 0 12 0 2 0 18 NiO 0 3 0 18 0 24 0 17 0 22 0 12 0 29 Total 100 65 99 53 100 36 100 79 100 9 99 07 100 75 99 58 100 27 99 22 100 36 100 3 99 87 99 44 99 62 100 29 100 84 99 71 99 62 100 91 100 33 100 72 99 62 100 94 100 2 mg-no. 75 6 81 6 78 4 74 8 78 9 74 7 79 3 73 8 74 4 81 2 79 8 73 9 80 73 8 81 2 73 6 79 6 77 5 Fs 10 3 21 12 1 11 9 25 7 10 3 11 4 11 3 10 4 11 7 22 2 Wo 43 9 2 9 42 7 42 7 1 8 43 3 43 5 43 6 44 9 42 5 1 6 An 71 4 66 7 77 3 73 5 62 64 1 65 1 Or 0 83 1 3 0 81 0 53 0 69 1 2 1 06 Equilib. T 987 997 1000 ( C) 597

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Table 3: continued C10c C10f A9 (ol) A9 (pl) A10 C6 C14 A15 ol cpx pl ol cpx pl ol cpx pl ol cpx pl ol cpx opx pl ol cpx pl ol cpx pl ol cpx pl SiO 2 37 1 51 6 50 8 37 6 51 1 50 3 38 2 51 51 4 37 4 50 6 51 37 2 51 6 54 51 7 37 2 52 1 51 3 37 4 52 1 52 5 37 2 49 6 50 7 TiO 2 1 1 1 1 0 96 1 1 0 82 0 48 0 7 0 63 1 3 Al 2O3 3 4 31 4 3 3 31 8 3 2 30 8 3 6 31 2 2 6 1 4 30 4 2 6 30 3 2 2 30 4 31 8 Cr 2O3 0 22 0 16 0 15 0 15 0 37 0 13 0 05 0 05 25 7 0 09 FeO 24 2 7 6 0 38 23 9 7 7 0 37 24 1 7 2 0 43 25 4 7 5 0 36 25 4 7 1 14 6 0 42 25 6 7 2 0 47 25 5 6 8 0 48 0 3 7 7 0 39 MnO 0 29 0 15 0 29 0 17 0 32 0 19 0 35 0 16 0 39 0 2 0 22 0 26 0 16 0 32 0 18 37 0 14 MgO 37 5 15 9 37 5 16 2 37 3 15 9 36 2 15 4 37 3 16 2 27 9 37 4 15 8 37 2 16 1 0 09 15 1 CaO 0 05 20 9 14 1 0 06 20 5 14 6 0 04 20 8 13 3 0 04 20 6 13 7 0 03 20 7 1 6 13 5 0 05 21 8 13 7 0 05 21 4 12 5 21 2 14 4 Na 2O 0 39 3 3 0 35 3 2 0 38 3 6 0 4 3 3 0 4 0 01 3 6 0 33 3 6 0 29 4 1 0 42 3 2 K 2O 0 13 0 06 0 16 0 14 0 22 0 19 0 2 0 19 NiO 0 17 0 23 0 19 0 19 0 15 0 12 0 13 0 13 Total 99 31 101 26 100 11 99 58 100 58 100 33 100 15 99 78 99 69 99 58 99 51 99 7 100 47 99 99 100 34 99 84 100 63 100 74 99 56 100 6 99 75 99 78 100 42 99 55 100 68 mg-no. 73 4 79 73 6 79 1 73 4 79 3 71 8 78 5 72 4 80 3 77 3 72 3 79 6 72 3 80 9 72 77 8 Fs 12 2 12 2 11 9 12 2 11 4 22 11 4 10 8 12 5 Wo 41 8 42 6 42 7 43 1 42 5 3 1 44 2 43 6 44 An 69 8 71 4 66 5 69 1 66 6 67 1 62 1 70 6 Or 0 76 0 35 0 95 0 84 1 3 1 1 1 2 1 1 Equilib. 1005 T ( C) 598

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES C4 A2-2 B4 C1c C1f C19c C19f ol cpx opx pl ol cpx opx pl ol cpx opx pl cpx pl cpx pl ol cpx opx pl ol cpx opx pl SiO 2 37 51 6 52 2 52 1 36 2 51 5 53 4 52 8 37 51 5 54 1 52 3 50 8 51 5 51 1 49 8 36 2 51 8 52 3 52 2 35 8 51 7 53 5 51 4 TiO 2 0 76 0 35 0 74 0 39 0 87 0 29 0 97 0 94 0 72 0 24 0 64 0 32 Al 2O3 0 27 1 2 30 6 2 4 1 4 30 3 2 8 1 3 30 8 3 2 30 9 3 1 32 2 7 1 1 30 3 2 3 1 3 30 6 Cr 2O3 0 03 0 03 0 06 0 03 0 08 0 01 0 01 0 01 0 01 0 01 FeO 26 6 7 5 18 2 0 38 28 2 7 9 16 8 0 34 28 8 7 8 16 8 0 31 7 5 0 4 7 6 0 45 33 7 6 20 0 28 32 8 7 9 19 1 0 25 MnO 0 34 0 18 0 36 0 46 0 23 0 42 0 41 0 14 0 26 0 21 0 21 0 45 0 37 0 43 0 49 0 35 0 39 MgO 35 9 15 5 25 8 35 8 15 5 26 5 34 8 15 5 26 5 15 9 15 9 29 9 15 5 24 6 29 2 15 6 24 7 CaO 0 03 21 3 0 99 12 7 0 04 19 9 1 2 12 0 11 21 1 1 2 12 8 20 7 13 2 20 9 14 6 0 06 21 1 0 86 12 6 0 11 21 1 1 12 9 Na 2O 0 38 0 05 4 0 35 0 03 4 3 0 36 0 01 4 0 38 3 7 0 37 3 1 0 34 0 01 4 0 35 0 02 3 8 K 2O 0 14 0 17 0 22 0 17 0 1 0 17 0 16 NiO 0 11 0 13 0 08 0 08 0 09 Total 99 98 99 95 99 18 99 92 100 83 98 58 100 17 99 91 101 2 100 15 100 47 100 43 99 66 99 87 100 12 100 05 99 69 100 14 99 55 99 55 98 49 99 85 100 44 99 11 mg-no. 70 6 78 7 71 7 69 4 77 8 73 8 68 3 78 73 7 79 1 78 8 61 8 78 4 68 7 61 4 77 9 69 6 Fs 12 27 8 12 9 25 6 12 5 25 6 12 12 1 12 2 30 8 12 6 29 6 Wo 43 8 1 9 41 8 2 4 43 3 2 4 42 6 42 7 43 5 1 7 43 2 2 An 63 2 60 1 63 1 65 7 71 9 62 9 64 7 Or 0 83 1 01 1 3 1 01 0 59 1 01 0 95 Equilib. 953 1024 974 966 972 T ( C) Prefixes A, B, and C in sample numbers refer to locations of sample sites cones shown in Fig. 1. ol, olivine; cpx, clinopyroxene; opx, orthopyroxene; pl, plagioclase. For modally layered and grain-size layered xenoliths, phases in each layer are presented with mineral reference to the dominant phase (ol or pl) or to grain-size (f, fine; c, coarse). Equilibrium T refers to clinopyroxene orthopyroxene geothermometry (Wells, 1997). 599

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 Fig. 4. Forsterite (Fo, mol %) variation with NiO and CaO contents (wt %) in olivine of the xenoliths from Mauna Kea volcano. Each data point represents an average value for an olivine grain acquired by 10 15 spot analyses per grain; multiple grains are plotted for some samples. For comparison, we show compositional fields for olivine in lavas and gabbros from Kilauea (Fodor & Moore, 1994), Maui (Fodor et al., 1977), Mauna Loa (Garcia et al., 1995), Mauna Kea (Baker et al., 1996), and Kahoolawe (Rudek et al., 1992; Fodor et al., 1993) volcanoes, Hawaii. represent post-exsolution compositions ( Tables 2 and 3; and b). For example, dunite olivine clinopyroxenite C25 Fig. 5). has mg-numbers 86 6 and 86 3, olivine clinopyroxenite Average mg-numbers for both interstitial and granular gabbronite C75 has mg-numbers 84 and 82 7, and olivinerich plagioclase-rich clinopyroxene in plagioclase-free (largely porphyroclastic) portions in gabbro A22 have clinoclinopyroxene dunite, wehrlite, and olivine clinopyroxenite xenoliths pyroxene mg-numbers 76 9 and 75 9. are in the range 90 3 83 6, with dunite occupying the Clinopyroxene mg-numbers plotted against coexisting high end of that range, except for evolved dunite A18 olivine Fo values form a continuum between ultramafic which has clinopyroxene beyond the range at mg-number and gabbroic xenoliths (Fig. 5a). However, clinopyroxene 80 9. Plagioclase-bearing wehrlite and olivine clino- mg-numbers are higher than coexisting Fo values, and pyroxenite have a lower range of clinopyroxene mgnumbers, differences between the coexisting values increase from 86 81 6, and gabbroic xenoliths have mg-num- ~2 to 8 mg-number units from ultramafic to gabbroic bers 84 9 77 3, where the fine-grained variety provides xenoliths. the majority of mg-numbers <80. Intra- and inter-grain Average Al 2 O 3 contents in clinopyroxene can be viewed variations are generally ~1 2 mg-number units but can as two subgroups (Fig. 5a). (1) One has Al 2 O 3 less than be up to ~7 mg-number units (Fig. 5b). ~3 wt %, and its clinopyroxene often coexists with Among modal- and phase-layered xenoliths, average orthopyroxene; compositions plot from mg-number ~90 mg-numbers in juxtaposed layers can be essentially ident- into and across the reference field for clinopyroxene in ical or differ by up to ~1 mg-number unit, where olivinedominant Kilauea tholeiitic lava. Troctolite, which does not have layers have the higher mg-numbers (Table 1a orthopyroxene, belongs to this subgroup by its low 600

FODOR AND GALAR MAUNA KEA CRYSTALLIZATION PROCESSES Fig. 5. Clinopyroxene mg-number [100 atomic Mg/(Mg+Fe)] plotted against various clinopyroxene components (wt %) and coexisting olivine forsterite (Fo) values for xenoliths from cones on the southern flank of Mauna Kea volcano (Fig. 1). (a) Each data point represents an average value for a clinopyroxene grain acquired by 10 15 spot analyses per grain; multiple grains are plotted for most samples. For comparison, we include compositional fields for clinopyroxene from Hawaiian tholeiitic magmas represented by Kilauea volcano lavas and gabbros (data from Moore et al., 1980; Ho & Garcia, 1988; Nichols & Stout, 1988; Garcia et al., 1989, 1992; Helz & Wright, 1992; Fodor & Moore, 1994), and for clinopyroxene in Hawaiian alkalic basalts (Fodor et al., 1975; Beeson, 1976; Chen et al., 1990; Frey et al., 1990). Data points for selected xenoliths are encircled. (b) Intra-sample compositional zoning for clinopyroxene in selected xenoliths to illustrate the ways by which Al 2 O 3 varies with respect to mg-number; C1 is a grain-size composite; small circles represent the fine-grained portion. 601

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 5 MAY 1997 clinopyroxene Al 2 O 3, ~2 wt %, which is among the clinopyroxene mg-numbers are ~85 and olivine ~Fo 81. lowest known for Hawaiian clinopyroxene. (2) The second Throughout its crystallization, orthopyroxene mg-num- subgroup has clinopyroxene Al 2 O 3 >3 wt %, no associated bers remain lower by ~3 mg-number units than the mgnumbers orthopyroxene, and plots above the tholeiitic reference of coexisting clinopyroxenes. field toward or in the alkalic basalt reference field. The ~6 wt % Al 2 O 3 of gabbro C22 is an extreme example of this alkalic subgroup. Only some clinopyroxenes show significant intra- and inter-grain Al 2 O 3 variations. These are usually increases Plagioclase in Al 2 O 3 with decreasing mg-number, but there are also Average plagioclase compositions ( Tables 2 and 3) are examples of varying Al 2 O 3 and nearly constant mg-num- An 87 Or 0 4 to An 60 Or 1 5 ; zoning extends the range to ber (e.g. C25c), and the reverse relationship (e.g. C18) ~An 50 Or 1 5 (Fig. 8). Wehrlite and olivine clinopyroxenite (Fig. 5b). contain the most calcic plagioclase, >An 81, and gabbroic Clinopyroxenes of all xenolith types collectively form xenoliths have comparatively evolved plagioclase, an essentially flat field for TiO 2 with decreasing mg- ~An 80 60, with the fine-grained variety <An 70 (Fig. 9a). number, where gabbronorites and troctolites overall have Compositional zoning within grains and variations within lower TiO 2 than gabbros. Compositions compare with samples can be anywhere from ~5 to 20 mol % An. those for Kilauea (tholeiitic) clinopyroxene, but >1 8 wt Figure 8b shows representative zoning patterns. % TiO 2 in clinopyroxene of C22 (Fig. 5a) is alkalic. Spot analyses expressed collectively show that An Or Clinopyroxene Cr 2 O 3 abundances among all xenoliths relationships in cone-a xenoliths and in most cone-c correlate positively with clinopyroxene mg-numbers, xenoliths conform to the compositional field in Kilauea where Cr detection limits are reached at mg-number ~77 tholeiitic lavas and gabbros (Fig. 8a). Some compositional (Fig. 5a). points for cone-c xenoliths fall outside the Kilauea field, Orthopyroxene-bearing xenoliths have clinopyroxene largely owed to plagioclase in fine-grained gabbroic xen- Na 2 O ~0 3 0 4 wt %, and most gabbro xenoliths have oliths. Plagioclase there began crystallization with Or Na 2 O >0 4 wt %; the highest is in C22, 0 45 0 60 wt values below the Kilauea trend and resembling the low %. Ultramafic xenoliths without orthopyroxene have a Or of mid-ocean ridge basalt ( MORB) plagioclase; proclinopyroxene Na 2 O range equal to that of all gabbroic gressive crystallization extended the compositions toward xenoliths. Collectively, Na 2 O abundances are higher than or into the Kilauea field (e.g. C4 in Fig. 8b). There are the ~0 20 0 35 wt % typically found in clinopyroxene also cone-c xenoliths with Or values greater than Kilauea of Hawaiian tholeiitic lavas and instead resemble values Or; olivine clinopyroxenite C78 has extreme Or, ~1 7 characteristic of clinopyroxene in alkalic basalt (Fig. 5a). mol%atan 65 (Fig. 8b). Figure 6A shows the average wollastonite (Wo) and Comparisons of plagioclase in fine- and coarse-grained ferrosilite ( Fs) components in clinopyroxenes. Ultramafic composites show that smaller grains, by and large, have and gabbroic xenoliths form a field of Wo decreasing lower Or and higher An than coexisting larger grains with increasing Fs. Only gabbro C22 is distinct by its (Fig. 8c). Plagioclase compositions in the olivine-rich and relatively high Wo, ~46 mol %. plagioclase-rich portions of modal- and phase-layered Rare-earth element (REE) abundances for clino- gabbroic xenoliths overlap (e.g. A9 and A22, Fig. 2f ). pyroxenes ( Table 4) separated from wehrlite, olivine REE compositions for plagioclase grains separated clinopyroxenite, and gabbroic xenoliths are plotted in from gabbro, gabbronorite, and troctolite are shown in Fig. 7a according to the clinopyroxene Al 2 O 3 subgroups Table 4 and Fig. 7a. The grains are respectively An 79 68, (e.g. Al 2 O 3 <3 wt % associates with orthopyroxene; Fig. An 77 55, and An 68 62. They are enriched in light REE 5a). The negative Ce anomaly for C21 clinopyroxene (LREE) when compared with the REE abundances of may reflect leaching owing to weathering (e.g. Fodor et plagioclase (An 68 62 ) from a Mauna Loa tholeiitic lava al., 1992, 1994). All patterns have shapes characteristic (Fig. 7a). of clinopyroxene (e.g. Irving & Frey, 1984), and, in Average An values in uniform (non-layered) general, clinopyroxenes in the low-al 2 O 3 subgroup have olivine+clinopyroxene framework and plagioclase-rich lower REE abundances. xenoliths have a rough inverse correlation with modal percentages of plagioclase (Fig. 9a). Plagioclase that deviates from the apparent trend is interstitial; namely, Orthopyroxene occurrences of <5 vol. % in cone-c xenoliths are either high An (85 83 mol %) or low An (72 65 mol %) (Fig. Orthopyroxene mg-numbers range from ~83 to 69 ( Table 9a). The An Fo plot (Fig. 9b) shows that cone-a xenoliths 3) and correlate positively with clinopyroxene mg-numbers conform well to the An Fo trend for Kilauea tholeiitic (Fig. 6b). Orthopyroxene begins forming when gabbro xenoliths, but cone-c xenoliths scatter from this 602